The endocyclic oxygen atom of D-mannopyranose is involved in its binding to pradimicins

Article abstract of DOI:10.1016/j.tetlet.2019.151530

Methyl 5-thio-α-D-mannopyranoside (1) and six inositols were evaluated for their ability to bind to pradimicins (PRMs) via molecular modeling and three binding assays. In all the experiments, the binding affinity of 1 was slightly lower than that of methy

Full text of DOI:10.1016/j.tetlet.2019.151530

Journal Pre-proofs
The endocyclic oxygen atom of D-mannopyranose is involved in its binding to
pradimicins
Yasunori Watanabe, Fumiya Yamaji, Makoto Ojika, Takahiro Sugawara, Dai
Akase, Misako Aida, Yasuhiro Igarashi, Yukishige Ito, Yu Nakagawa
PII:
S0040-4039(19)31329-2
DOI:
https://doi.org/10.1016/j.tetlet.2019.151530
TETL 151530
Reference:
Tetrahedron Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
1 November 2019
12 December 2019
13 December 2019
Please cite this article as: Watanabe, Y., Yamaji, F., Ojika, M., Sugawara, T., Akase, D., Aida, M., Igarashi, Y., Ito,
Y., Nakagawa, Y., The endocyclic oxygen atom of D-mannopyranose is involved in its binding to pradimicins,
Tetrahedron Letters (2019), doi: https://doi.org/10.1016/j.tetlet.2019.151530
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will
undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing
this version to give early visibility of the article. Please note that, during the production process, errors may be
discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
© 2019 Published by Elsevier Ltd.

Graphical Abstract
To create your abstract, type over the instructions in the template box below.
Fonts or abstract dimensions should not be changed or altered.
The endocyclic oxygen atom of D-manno-
pyranose is involved in its binding to pradimicins
Leave this area blank for abstract info.
Yasunori Watanabe, Fumiya Yamaji, Makoto Ojika, Takahiro Sugawara, Dai Akase, Misako Aida, Yasuhiro Igarashi,
Yukishige Ito, Yu Nakagawa
Man-OMe
H
H
H
MeO
H
MeO
OH
OH
H
O
O
O
S
O
S
O
O
O
Ca2+
Ca2+
substitution
OMe
OMe
H
OH
OH
OH
O
O
OH
O
O
Slight decrease
in binding affinity
O
O
O
H
H
H
HO
HO
H
N
H
N
O
O
O
O
HO
Disaccharide
HO
Disaccharide
H
Me
Me
O
O
O
O
Me
Me
Pradimicin

1
Tetrahedron Letters
journal homepage: www.elsevier.com
The endocyclic oxygen atom of D-mannopyranose is involved in its binding to
pradimicins
Yasunori Watanabea , Fumiya Yamajia, Makoto Ojikaa, Takahiro Sugawarab, Dai Akaseb, Misako Aidab,
Yasuhiro Igarashic, Yukishige Itod, Yu Nakagawaa,d,*
aDepartment of Applied Biosciences, Graduate School of Bioagricultural Sciences, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8601, Japan
bCenter for Quantum Life Sciences, and Department of Chemistry, Graduate School of Science, Hiroshima University, 1-3-1 Kagamiyama, Higashi-Hiroshima,
Hiroshima 739-8526, Japan
cBiotechnology Research Center, Toyama Prefectural University, 5180 Kurokawa, Imizu, Toyama 939-0398, Japan
dSynthetic Cellular Chemistry Laboratory, RIKEN, 2-1 Hirosawa, Wako, Saitama 351-0198, Japan
ARTICLE INFO
ABSTRACT
Article history:
Received
Received in revised form
Accepted
Methyl 5-thio--D-mannopyranoside (1) and six inositols were evaluated for their ability to bind
to pradimicins (PRMs) via molecular modeling and three binding assays. In all the experiments,
the binding affinity of 1 was slightly lower than that of methyl -D-mannopyranoside (Man-
OMe) and inositols hardly bound to PRMs. These results indicate that the endocyclic oxygen
atom of Man-OMe is involved in its binding to PRMs.
Available online
2009 Elsevier Ltd. All rights reserved.
Keywords:
Antibiotic
Carbohydrate recognition
Molecular interaction
Pradimicin
Introduction
anthraquinone and D-alanine moieties of PRM-A are in close
contact with the C2 and C3 positions of Man-OMe, respectively,
Pradimicins (PRMs, Fig. 1) [1–3] are a unique class of
antibiotics that show anti-fungal, antiviral, and anti-parasitic
activities by binding to the glycans of pathogenic species [4–8].
Their glycan binding ability originates from specific interactions
with D-mannose (Man) in the presence of Ca2+. Since such C-
type lectin-like behavior is never observed in any other small
molecule, considerable attention has been directed towards the
molecular basis of Man recognition by PRMs. Earlier studies
suggested that PRMs bind Man in Ca2+-containing neutral
solutions to form insoluble aggregates of the ternary
[PRM2/Ca2+/Man4] complex [9]. Subsequent spectroscopic
analysis showed that PRMs initially form the binary [PRM2/Ca2+]
complex, which subsequently binds four molecules of Man in
two independent steps (Scheme 1) [10–12]. However, during the
two decades since the discovery of PRMs, the aggregate-forming
propensity of the [PRM2/Ca2+/Man2] and [PRM2/Ca2+/Man4]
complexes has hampered their structural analysis, leaving the
question as to how PRMs bind Man unanswered.
in the complex [15]. Recently, we also determined the crystal
structure of the [PRM2/Ca2+] complex using a water-soluble PRM
derivative [16]. Based on these findings, we performed density
functional theory (DFT) calculations to estimate the structure of
the [PRM-A2/Ca2+/Man-OMe2] complex (Fig. 1). In our binding
model, the 2-, 3-, and 4-hydroxyl groups of Man-OMe are
involved in Ca2+ coordination and/or hydrogen bonding with
PRM-A, while the 1-methoxy and 6-hydroxyl groups do not
interact with either Ca2+ or PRM-A. The endocyclic oxygen atom
in the pyranose ring of Man-OMe acts as an acceptor for
hydrogen bonding with the Ca2+-coordinated water molecule.
The roles of the hydroxyl groups of Man-OMe in our binding
model are in good agreement with our previous finding that
PRM-A binds 6-deoxy-Man-OMe but not with 2-, 3-, and 4-
deoxy-Man-OMe [15]. Regarding its 1-methoxy group, Man and
 anomer of Man-OMe were found to co-precipitate with PRM-A
in a manner similar to Man-OMe [17], supporting the trivial role
of the methoxyl group and configuration at the C1 position of
Man-OMe. However, there is a lack of experimental evidence to
support the involvement of the endocyclic oxygen atom of Man-
OMe in binding to PRM-A. In this study, we evaluated binding
of PRMs with methyl 5-thio--D-mannopyranoside (1, Fig. 1)
[18] to examine the role of the endocyclic oxygen atom of Man-
OMe. We also revealed that PRMs are insensitive to six inositols,
which mimic the spatial configuration of the 2-, 3-, and 4-
Recognizing this issue, our group started, in 2009, to explore
an analytical strategy based on solid-state NMR technology [13–
15]. Making use of the inherent aggregate-forming property of
PRM-A (Fig. 1) [2], the first member of PRMs, we prepared
solid aggregates mainly composed of a 1:1 complex of PRM-A
and methyl -D-mannopyranoside (Man-OMe), i.e. the [PRM-
A2/Ca2+/Man-OMe2] complex. Structural analysis by dipolar-
assisted rotational resonance (DARR) revealed that the

2
Tetrahedron
hydroxyl groups of Man while lacking an endocyclic oxygen
Next, we calculated the interaction energy (E) between the
atom.
[PRM-A2/Ca2+] complex and Man-OMe or 1. For interaction
energies, BSSE (basis set superposition error) was taken into
consideration. The E value for 1 was –139.96 kcal/mol, while
that for Man-OMe was –144.30 kcal/mol. The absolute value of
E for 1 is slightly smaller than that for Man-OMe, implying a
lower binding affinity of 1. The main reason for this is the weak
interaction of the endocyclic sulfur atom of 1 with the Ca2+-
coordinated water molecule, compared with that of the
corresponding oxygen atom of Man-OMe. These results of our
preliminary calculations collectively suggest that 1 would bind to
the [PRM2/Ca2+] complex in a manner similar to PRM-A, but
with lower potency.
1
5
D-Ala
moiety
6
1
3
2
4
13
Figure 2. Energy-minimized structure of the [PRM-A2/Ca2+/12]
complex obtained by DFT calculations at the -B97X-D/6-31G(d)
level of theory. Carbon, oxygen, nitrogen, sulfur, hydrogen, and
calcium atoms are shown in gray, red, blue, yellow, white, and green,
respectively. Blue and green dotted lines represent possible hydrogen
bonding and Ca2+ coordination, respectively. Two water molecules
are coordinated to Ca2+ in the complex.
Figure 1. Structures of pradimicins, Man-OMe, and 1 (upper) and
energy-minimized structure of the [PRM-A2/Ca2+/Man-OMe2]
complex obtained by solid-state NMR analysis and DFT calculations
at the -B97X-D/6-31G(d) level of theory (lower). Carbon, oxygen,
nitrogen, hydrogen, and calcium atoms are shown in gray, red, blue,
white, and green, respectively. Blue and green dotted lines represent
possible hydrogen bonding and Ca2+ coordination, respectively. Two
water molecules are coordinated to Ca2+ in the complex.
As the first validation of our computational prediction, we
attempted to determine a binding constant for 1 to PRM-A by
isothermal titration calorimetry (ITC). However, unfortunately,
complex formation of PRM-A with 1 caused severe aggregation,
resulting in complicated ITC signals. We, therefore, performed
co-precipitation experiments [17] as an alternative evaluation.
Our established procedure consists of four steps: (1) aggregate
formation of PRM-A with CaCl2 and sugars in aqueous solutions
at pH 6.0, (2) extensive washing of the aggregate, (3) acid
treatment to dissociate the aggregate, and (4) quantification of the
sugar/PRM-A molar ratio by 1H-NMR on the basis of their
integration values. Since step (2) involves the complete release of
sugars from the secondary binding site of the [PRM-A2/Ca2+]
complex, this co-precipitation experiment can quantify the
primary binding of sugars (Scheme 1). Table 1 shows the
1/PRM-A molar ratio in the aggregate, along with those for Man-
OMe and methyl -D-glucopyranoside (Glc-OMe). While only a
negligible amount of Glc-OMe was detected in the aggregate, 1
showed significant binding to PRM-A. However, the detected
amount of 1 was slightly lower than that of Man-OMe. This
result reflects lower affinity of 1 than that of Man-OMe,
supporting our modeling study.
Scheme 1. Complex-forming process of PRM with Ca2+ ion and
Man.
Results and discussion
The binding evaluation of PRMs with 1 started with DFT
studies centered on estimating the effect of the sulfur atom on our
proposed binding model. An initial structure was constructed for
the [PRM-A2/Ca2+/12] complex by replacing Man-OMe with 1,
followed by optimization through DFT calculations at the
B97X-D/6-31G(d) level of theory. The obtained energy-
minimized structure (Fig. 2) shows a remarkable similarity to the
[PRM-A2/Ca2+/Man-OMe2] complex in terms of the hydrogen-
bond network and Ca2+ coordination. The 2-hydroxyl group of 1
is involved in both Ca2+ coordination and hydrogen bonding with
the 13-carbonyl group of PRM-A, and the 3- and 4-hydroxyl
groups interact with the D-alanine moiety of PRM-A. Moreover,
a hydrogen bond is also found between the Ca2+-coordinated
water molecule and the endocyclic sulfur atom of 1, suggesting
that 1 could behave similarly to Man-OMe in binding to PRM-A.
To obtain more experimental support, we evaluated the
antagonistic effects of 1 on the antifungal activity of PRM-A
against Candida rugosa (MIC of PRM-A = 4 g/mL). Earlier
studies revealed that the antifungal activity of PRM-A is closely
related to its binding to cell wall mannans and is suppressed by
the exogenous addition of oligomannoses [9]. Thus, we
determined the minimum antagonistic concentrations (MACs) of
Man-OMe, Glc-OMe, and 1 required for suppressing the PRM-A

3
(32 g/mL)-induced growth inhibition of C. rugosa cells. As
shown in Table 2, Man-OMe exerted a significant antagonistic
effect, while Glc-OMe did not suppress the antifungal activity of
PRM-A up to a 64 mM concentration. Mirroring the results of the
co-precipitation experiments, the MAC value of 1 was slightly
higher than that of Man-OMe, supporting weaker binding of 1 to
PRM-A compared with that of Man-OMe.
validate our binding model, in which the endocyclic oxygen
atom of Man is involved in its binding to PRMs.
These results further led us to predict that PRMs would be
insensitive to inositols, which mimic the spatial configuration of
2-, 3-, and 4-hydroxyl groups of Man while lacking the
endocyclic oxygen atom (Fig. 3). To confirm this assumption, six
inositols were evaluated for their binding to PRM-A and PRM-
FA1. Results are shown in Tables 1–3. As expected, these
inositols exhibited negligible binding in all three binding
experiments. These results indicate that PRMs can discriminate
Man not only from common monosaccharides but also from
inositols.
Table 1.
Molar ratio of Man-OMe, Glc-OMe, 1, and inositols relative to
PRM-A in aggregates.
Sugar
Man-OMe
Glc-OMe
1
Sugar/PRM-A ratio
1.02 (0.05)a
< 0.01
Table 3.
Aggregate formation of PRM-FA1 induced by Man-OMe, Glc-OMe,
1, and inositols.
0.85 (0.01)
< 0.01
myo-Inositol
allo-Inositol
epi-Inositol
1D-Inositol
1L-Inositol
muco-Inositol
Sugar
Man-OMe
Glc-OMe
1
AC50a (mM)
5.8
< 0.01
< 0.01
> 10
< 0.01
6.8
< 0.01
myo-Inositol
allo-Inositol
epi-Inositol
1D-Inositol
1L-Inositol
muco-Inositol
> 10
< 0.01
> 10
a Standard deviation of at least three separate experiments.
> 10
> 10
Table 2.
> 10
Antagonistic effect of Man-OMe, Glc-OMe, 1, and inositols on
antifungal activity of PRM-A against Candida rugosa.
> 10
a
Concentration that induces 50% aggregation of PRM-FA1. AC50
values were determined from corresponding dose-response curves
(see ESI).
Sugar
Man-OMe
Glc-OMe
1
MACa (mM)
16
> 64
32
myo-Inositol
allo-Inositol
epi-Inositol
1D-Inositol
1L-Inositol
muco-Inositol
> 64
> 64
> 64
> 64
> 64
> 64
a
Minimum antagonistic concentration required for suppressing
PRM-A (32 g/mL)-induced growth inhibition of C. rugosa
cells. The experiments were performed as duplicates.
Figure 3. Structures of inositols used in this study.
Further experimental support was obtained by aggregation
experiments using PRM-FA1 (Fig. 1), a D-serine analog of PRM-
A
[19]. Unlike the [PRM-A2/Ca2+] complex, the [PRM-
In summary, we evaluated the binding of methyl 5-thio--D-
mannopyranoside (1) to PRMs through molecular modeling and
three binding experiments. The results collectively indicate that 1
binds to PRMs with slightly lower affinity than Man-OMe,
supporting our binding model, in which the endocyclic oxygen
atom of Man-OMe participates in binding to PRMs. Additionally,
we demonstrated that PRMs hardly form a ternary complex with
inositols, although they have a spatial configuration mimicking
that of the 2-, 3-, and 4-hydroxyl groups of Man. This emerging
selectivity for Man over inositols further underscores the
potential of PRMs as superior lectin mimics. Investigation into
the medicinal and glycobiological applications of PRMs is
underway, and its results will be reported in due course.
FA12/Ca2+] complex hardly aggregates in neutral aqueous
solutions. However, when it binds Man, the resulting [PRM-
FA12/Ca2+/Man2] and [PRM-FA12/Ca2+/Man4] complexes easily
aggregate to produce precipitates. Assuming that the
aggregability of PRM-FA1 depends on its binding affinity for
sugars, we evaluated the binding affinities of Man-OMe, Glc-
OMe, and 1 by measuring their concentrations (AC50) that induce
50% aggregation of PRM-FA1 (Table 3). Similar to the results
obtained in the above two tests using PRM-A, 1 induced the
aggregation of PRM-FA1, but with less potency than that of
Man-OMe. Inertness of PRM-FA1 towards Glc-OMe suggests
that its aggregation truly originates from the specific binding of
Man-OMe or 1. Taken together, all the three abovementioned
experiments support our computational prediction, and thereby

4
Tetrahedron
Acknowledgments
[7] K.O. François, J. Balzarini, Med. Res. Rev. 32 (2012) 349–387.
[8] Y. Igarashi, T. Oki, Adv. Appl. Microbiol. 54 (2004) 147–166.
[9] T. Ueki, K. Numata, Y. Sawada, T. Nakajima, Y. Fukagawa, T. Oki, J.
Antibiot. 46 (1993) 149–161.
This work was partly supported by a Grant-in-Aid for JSPS
Fellow (No. 17J03048) and JSPS KAKENHI Grant (No.
JP15H04496). Candida rugosa AJ 14513 (NBRC 0750) was
provided from Ajinomoto Co., Inc. (Kanagawa, Japan). The
DFT calculations were partly carried out at the Research Center
for Computational Science, Okazaki, Japan. We thank Mr.
Kazushi Koga for NMR spectroscopic assistance.
[10]M. Hu, Y. Ishizuka, Y. Igarashi, T. Oki, H. Nakanishi, Spectrochim. Acta
A 55 (1999) 2547–2558.
[11]M. Hu, Y. Ishizuka, Y. Igarashi, T. Oki, H. Nakanishi, Spectrochim. Acta
A 56 (1999) 181–191.
[12]K. Fujikawa, Y. Tsukamoto, T. Oki, Y.C. Lee, Glycobiology 8 (1998)
407–414.
[13]Y. Nakagawa, Y. Masuda, K. Yamada, T. Doi, K. Takegoshi, Y.
Igarashi, Y. Ito, Angew. Chem. Int. Ed. 50 (2011) 6084–6088.
[14]Y. Nakagawa, T. Doi, Y. Masuda, K. Takegoshi, Y. Igarashi, Y. Ito, J.
Am. Chem. Soc. 133 (2011) 17485–17493.
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://xxxx.
[15]Y. Nakagawa, T. Doi, K. Takegoshi, Y. Igarashi, Y. Ito, Chem. Eur. J. 19
(2013) 10516–10525.
[16]Y. Nakagawa, T. Doi, K. Takegoshi, T. Sugahara, D. Akase, M. Aida, K.
Tsuzuki, Y. Watanabe, T. Tomura, M. Ojika, Y. Igarashi, D. Hashizume,
Y. Ito, Cell Chem. Biol. 26 (2019) 950–959.
References
[1] T. Takeuchi, H. Hara, H. Naganawa, M. Okada, M. Hamada, H.
Umezawa, S. Gomi, M. Sezaki, S. Kondo, J. Antibiot. 41 (1988) 807–
811.
[17]Y. Nakagawa, Y. Watanabe, Y. Igarashi, Y. Ito, M. Ojika, Bioorg. Med.
Chem. Lett. 25 (2015) 2963–2966.
[18]M. Izumi, Y. Suhara, Y. Ichikawa, J. Org. Chem. 63 (1998) 4811–4816.
[19]Y. Sawada, M. Hatori, H. Yamamoto, M. Nishio, T. Miyaki, T. Oki, J.
Antibiot. 43 (1990) 1223–1229.
[2] T. Oki, M. Konishi, K. Tomatsu, K. Tomita, K. Saitoh, M. Tsunakawa,
M. Nishio, T. Miyaki, H. Kawaguchi, J. Antibiot. 41 (1988) 1701–1704.
[3] Y. Fukagawa, T. Ueki, K. Numata, T. Oki, Actinomycetologia 7 (1993)
1–22.
[4] M.M.F. Alen, T.D. Burghgraeve, S.J.F. Kaptein, J. Balzarini, J. Neyts, D.
Schols, PLoS ONE 6 (2011) e21658.
[5] J. Balzarini, Nat. Rev. Microbiol. 5 (2007) 583–597.
[6] V.M. Castillo-Acosta, L.M. Ruiz-Pérez, J. Etxebarria, N.C. Reichardt, M.
Navarro, Y. Igarashi, S. Liekens, J. Balzarini, D. González-Pacanowska,
PLoS Pathog. 12 (2016) e1005851.
C
l
i
c
k
h
e
r
e
t
o
r
e
m
o
v
e
i
n
s
t
r
u
c
t
i
o
n
t
e
x
t
.
.
.

5
Highlights
• DFT calculations provide a binding model of
pradimicin (PRM) with 5-thio-mannose
• Binding affinity of 5-thio-mannose for PRM is
slightly lower than that of mannose
• Six inositols hardly bind to PRM