Amino Acids Bearing Aromatic or Heteroaromatic Substituents as a New Class of Ligands for the Lysosomal Sialic Acid Transporter Sialin

Article abstract of DOI:10.1021/acs.jmedchem.9b02119

Sialin, encoded by the SLC17A5 gene, is a lysosomal sialic acid transporter defective in Salla disease, a rare inherited leukodystrophy. It also enables metabolic incorporation of exogenous sialic acids, leading to autoantibodies against N-glycolylneuraminic acid in humans. Here, we identified a novel class of human sialin ligands by virtual screening and structure-activity relationship studies. The ligand scaffold is characterized by an amino acid backbone with a free carboxylate, an N-linked aromatic or heteroaromatic substituent, and a hydrophobic side chain. The most potent compound, 45 (LSP12-3129), inhibited N-acetylneuraminic acid 1 (Neu5Ac) transport in a non-competitive manner with IC50 ≈ 2.5 μM, a value 400-fold lower than the KM for Neu5Ac. In vitro and molecular docking studies attributed the non-competitive character to selective inhibitor binding to the Neu5Ac site in a cytosol-facing conformation. Moreover, compound 45 rescued the trafficking defect of the pathogenic mutant (R39C) causing Salla disease. This new class of cell-permeant inhibitors provides tools to investigate the physiological roles of sialin and help develop pharmacological chaperones for Salla disease.

Full text of DOI:10.1021/acs.jmedchem.9b02119

Subscriber access provided by the University of Exeter
Article
Amino acids bearing aromatic or heteroaromatic substituents as a new
class of ligands for the lysosomal sialic acid transporter sialin (SLC17A5)
Lilian Dubois, Nicolas Pietrancosta, Alexandre Cabaye, Isabelle Fanget, Cécile Debacker, Pierre-André
Gilormini, Patrick M Dansette, Julien Dairou, Christophe Biot, Roseline Froissart, Anne Goupil-Lamy, Hugues-
Olivier Bertrand, Francine Clemy Acher, Isabelle McCort-Tranchepain, Bruno Gasnier, and Christine Anne
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.9b02119 • Publication Date (Web): 01 Jul 2020
Downloaded from pubs.acs.org on July 2, 2020
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Amino acids bearing aromatic or heteroaromatic
substituents as a new class of ligands for the
lysosomal sialic acid transporter sialin (SLC17A5)
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2,3
1,4
5
Lilian Dubois, Nicolas Pietrancosta, Alexandre Cabaye, Isabelle Fanget, Cécile
5
6‡
1
1
6
Debacker, Pierre-André Gilormini, Patrick Dansette, Julien Dairou, Christophe Biot,
7
4
4
1
Roseline Froissart, Anne Goupil-Lamy, Hugues-Olivier Bertrand, Francine Acher,
1*
5*
5*
Isabelle McCort-Tranchepain, Bruno Gasnier, and Christine Anne
1
Université de Paris, Laboratoire de Chimie et Biochimie Pharmacologiques et Toxicologiques,
CNRS, UMR 8601, F-75006 Paris, France
2
Sorbonne Université, École Normale Supérieure, PSL University, CNRS, Laboratoire des
Biomolécules, LBM, F-75005 Paris, France.
3
Sorbonne Université, INSERM, CNRS, Neurosciences Paris Seine - Institut de Biologie Paris
Seine (NPS - IBPS), F-75005, Paris, France.
4
BIOVIA, Dassault Systèmes, F-78140 Velizy-Villacoublay, France
5
Université de Paris, SPPIN - Saints-Pères Paris Institute for the Neurosciences, CNRS, F-
7
5006 Paris, France
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
6
Université de Lille, CNRS, UMR 8576, UGSF, Unité de Glycobiologie et Fonctionnelle, F-
5
9650 Lille, France
7 Service de Biochimie et Biologie Moléculaire Grand Est, Centre de Biologie et de Pathologie
Est, Hospices Civils de Lyon, F-69677 Bron, France
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ABSTRACT: Sialin is a lysosomal sialic acid transporter defective in Salla disease, a rare inherited
leukodystrophy. It also enables metabolic incorporation of exogenous sialic acids, leading to
autoantibodies against N-glycolylneuraminic acid in humans. Here we identified a novel class of
human sialin ligands by virtual screening and structure-activity relationship studies. The ligand
scaffold is characterized by an amino acid backbone with a free carboxylate, an N-linked aromatic
or heteroaromatic substituent and a hydrophobic side chain. The most potent compound, 45
(
LSP12-3129), inhibited N-acetylneuraminic acid 1 (Neu5Ac) transport in a non-competitive
manner with IC ≈ 2.5 µM, a value 400-fold lower than the K for Neu5Ac. In vitro and molecular
5
0
M
docking studies attributed the non-competitive character to selective inhibitor binding to the
Neu5Ac site in cytosol-facing conformation. Moreover, compound 45 rescued the trafficking
defect of the pathogenic mutant (R39C) causing Salla disease. This new class of cell-permeant
inhibitors provide tools to investigate the physiological roles of sialin and to help develop
pharmacological chaperones for Salla disease.
2
ACS Paragon Plus Environment

Page 3 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Introduction
SLC17 transporters, a subgroup from the major facilitator superfamily (MFS) of secondary
active transporters, are involved in the excretion of urate and other organic anions in the kidney,
the uptake of anionic neurotransmitters (glutamate, ATP) into synaptic vesicles, and the export of
sialic acids and acidic hexoses from lysosomes for reuse in metabolism.1–7 The latter function is
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ensured by sialin, a H /sialic acid symporter encoded by the SLC17A5 gene. Owing to its H+
coupling and the steep pH gradient (~ 2.5 units) across the lysosomal membrane, sialin actively
exports sialic acids from the lysosomal lumen, thus contributing with de novo biosynthesis to their
cytosolic availability. This in turn impacts the sialylation of cell surface glycoconjugates and their
role in cell-cell, cell-extracellular matrix and cell-pathogen interactions. Sialin has also been
+
5–7
8
implicated in the uptake of nitrate at the plasma membrane and the uptake of aspartate, glutamate
and N-acetyl-aspartyl-glutamate into synaptic vesicles.9,10
In addition to its role in recycling endogenous sialic acids, lysosomal export by sialin enables
the metabolic incorporation of exogenous sialic acids internalized by fluid phase endocytosis,11
thus circumventing the absence of sialic acid transporter at the plasma membrane. This endocytic
route impacts human health by incorporating dietary-derived N-glycolylneuraminic acid (Neu5Gc)
into glycoconjugates. Owing to a mutation in the cytidine monophosphate-Neu5Ac hydroxylase
(
CMAH) gene, humans do not synthesize this sialic acid which is abundant in other species.12
Neu5Gc is thus incorporated into human cell glycans, inducing the production of anti-Neu5GC
autoantibodies associated with chronic inflammation, cancer and atherosclerosis.13,14 This
endocytic route is also exploited in glycoengineering approaches to alter cell surface glycans with
synthetic sialic acid derivatives and thereby: modulate immune responses; inhibit pathogen
binding; target toxins to tumors; or track cells in vivo after transplantation.15,16
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The essential role of sialin in sialic acid homeostasis is also highlighted by its inactivation in
two autosomal recessive genetic diseases, Salla disease and Infantile free Sialic acid Storage
Disease (ISSD).5,17,18 Both diseases are associated with defective lysosomal export of sialic
acids,6,7,19 leading to their accumulation in diverse tissues and their excretion in urine. However,
they strongly differ in clinical course and severity. While ISSD affects multiple organs from birth,
causing death within a few months or years, Salla disease selectively affects the white matter of
the brain in a progressive manner.17,20,21 Salla patients show hypotonia, ataxia, nystagmus and
delayed motor development during the first year. Psychomotor milestones, including speech,
progressively worsens during infancy and childhood, leading to severe motor and cognitive deficits
in adulthood. Magnetic resonance imaging shows nonspecific brain hypomyelination with a
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
22
thinning of the corpus callosum. Current treatment is supportive.
The pathophysiology of Salla disease remains poorly understood. Sialin-defective mice studies
have shown that brain hypomyelination results from defective maturation of oligodendrocytes23,24
.
However, the link between this defect and defective lysosomal sialic acid export, or other transport
9
25
activities of sialin (but see ref. ), is unclear. Decreased downregulation of the polysialylated
Neural Cell Adhesion Molecule (PSA-NCAM), which could delay oligodendrocyte/axon contacts,
2
3
may be involved. However, this decrease is limited, indicating that other mechanisms, such as
2
6
impaired ganglioside metabolism, should contribute to the myelination defect.
On the other hand, genotype-phenotype relationship studies have shown a clear correlation
between the level of Neu5Ac transport and the clinical severity induced by SLC17A5 mutations.
Salla disease is almost exclusively associated with one missense mutation, R39C, in homozygous
or compound-heterozygous state which partially impairs lysosomal targeting of sialin6,27 and
which decreases its transport activity to ~20% of wild-type level6,7 whereas ISSD-associated
4
ACS Paragon Plus Environment

Page 5 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
mutations abolish either the expression or the transport activity of sialin.6,7,18 The association of
Salla disease with residual sialic acid transport has been confirmed in a rare case of patient with a
2
8
homozygous K136E mutation, which also partially preserves sialic acid transport and lysosomal
localization.6,7 The trafficking defect may result from destabilization of the R39C and K136E
mutants, and their retention by the protein quality control system. This feature along with the
residual transport activity and the overwhelming predominance of the R39C mutation suggest
pharmacological chaperone therapy as a promising option to treat Salla disease.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
In contrast with other SLC17 transporters such as VGLUT29–31 and VNUT proteins, we lack
32
pharmacological tools to study the cellular and physiological roles of sialin or to help rescue the
3
3
R39C mutant causing Salla disease. In a previous study, we characterized the interaction of
human sialin with synthetic sialic acids and identified one compound with IC50 ≈ 40 µM, a value
2
5-fold lower than the K for Neu5Ac. We also built a cytosol-facing homology model of human
M
sialin based on a distantly homologous MFS crystallographic structure, the glycerol-3-phosphate
34
transporter (GlpT) from E. coli. After validating the sialic acid-binding site of this model by site-
directed mutagenesis, virtual screening against the ZINC database and Neu5Ac transport studies
led us to identify the artificial tripeptide FR139317 as a new sialin ligand unrelated to sialic acids.33
In this study, we disclose other virtual hits from this screen and identify a novel scaffold of sialin
ligand well suited for chemical modifications. After optimization, the new ligands obtained inhibit
sialin with a µM affinity in a non-competitive manner and one of them could partially rescue the
trafficking defect of the pathogenic R39C mutant.
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Results
Screening and Scaffold Selection. Our previous virtual screening of a GlpT-based homology
3
5
model of human sialin against a subset of the ZINC database (~12,000 compounds) identified
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
33
the endothelin-A receptor antagonist FR139317 2 and eight other molecules (Figure 1A) as
candidate sialin ligands. Transport studies confirmed 2 as a competitive inhibitor of N-
acetylneuraminic acid 1 (Neu5Ac) transport by human sialin with a K of 20 µM, a value ~50-fold
i
3
3
lower than the K for Neu5Ac. We thus tested the other compounds using the same transport
M
6
assay, which is based on a sorting mutant (L22G/L23G) redirected to the plasma membrane. In
3
this approach, the poorly tractable lysosomal export is replaced by a whole-cell [ H]Neu5Ac
uptake in acidic extracellular buffer to mimic the lysosomal lumen and facilitate transport
measurements (Figure 1B). This technique provides higher signal-to-noise ratios than lysosomal
assays, with highly similar Neu5Ac transport properties.6
Among the novel virtual hits, four compounds inhibited Neu5Ac uptake and three of them,
Fmoc-Lys(Cbz)-OH 3, Fmoc-D-Phe(pCH NHBoc)-OH 4 and calcein 5 were more efficient
2
inhibitors than Neu5Ac (Figure 1C). Interestingly, 3 and 4 are both amino acid derivatives with
common features: the  amine is protected by the same hydrophobic carbamate,
(
Fluoren-9-yl)methoxycarbonyl (Fmoc); the carboxylic acid is free; and a 4 to 6-carbon side chain
bears a distal amine protected by a benzyloxycarbonyl (Cbz) or tert-butoxycarbonyl (Boc) group.
A
6
ACS Paragon Plus Environment

Page 7 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
OH
HO
N
COOH
OH
HO
AcHN
O
O
H
H
O
HO
N
N
CO H
2
O
O
H
N
1
N
CO2H
H
N
O
Neu5Ac
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
N
FR139317
O
HO C HO
O
OH CO H
2
2
HN
N
N
O
H
O
O
3
N
CO H
2
O
CO H
CO H
2
2
Fmoc-Lys(Cbz)-OH
O
5
Calcein
O
O
NH
O
O
O
O
CO2H
CO2H
N
H
O
CO H
2
HO C
O
2
N
4
N
HO C
H
2
Fmoc-D-Phe(pCH NHBoc)-OH
O
2
O
6
7
H N
N
N
NH2
2
N
NH2
O
N
H
N
CO H
2
HO C
N
H
CO H
2
2
HO C
N
O
2
N
S
S
O
H
N
O
CO2H
O
H
N
N
CO H
2
HO C
CO H
2
2
O
N
H
CO H
2
NH2
O
8
9
10
BAPTA
CO H
2
Methotrexate
GSSG
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figure 1. Identification of novel sialin inhibitors (A) Structures (2 to 10) of the best hits from a
3
3
previous virtual screening. (B) Whole-cell assay of human sialin. Mutation of its lysosomal
sorting motif redirects sialin to the plasma membrane to facilitate transport measurements. (C)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
Human sialin was assayed for [ H]Neu5Ac uptake at pH 5.0 in the absence (control) or presence
of these compounds (means ± SEM of 2 to 4 independent experiments).
These similarities suggested a novel scaffold recognized by sialin. To test this hypothesis, we
evaluated a series of 14 N-Fmoc amino acids 1124 in our transport assay (see structures in
Supplementary Information, pp. S3-S4). N-Fmoc amino acids 11, 12, 16, 22 and 24 were prepared
according to classical method (see Supporting information), whereas compounds 1315, 1721
and 23 were commercially available. Interestingly, 10 out of 14 compounds inhibited Neu5Ac
uptake with an efficiency greater than, or similar to, compounds 3 and 4 (Figure 2), thus confirming
the existence of a novel scaffold characterized by an N-Fmoc group, a free carboxylate, and a
hydrophobic side chain. The lower activity of compounds 22-24 may result from the presence of
a bulkier side chain (compare Fmoc-Lys(Fmoc)-OH 24 with 3) or from a constrained ring structure
of the side chain in Fmoc-Pro-OH 23. All tested amino acids, except Fmoc-DL-Leu-OH 12,
belonged to the L-series. Comparison of the racemic 12 and L- 13 forms of Fmoc-Leu-OH suggests
that the  carbon configuration is not critical to bind sialin (Figure 2).
To characterize further this sialin ligand scaffold, we examined the influence of the proximal
amine substitution. The 9-fluorenylmethoxycarbonyl group of compound 3 was thus replaced by
a coumarinyl; quinolinyl; xanthenyl; or anthracenyl group linked to the  amine by an amide or
carbamate bond. In another set of compounds, we examined the influence of substituents attached
to the side chain. For this purpose, we first introduced a 4-methylcarbonylcoumarin instead of Cbz
8
ACS Paragon Plus Environment

Page 9 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
in compound 3 and then selected compound 16 among those with a good activity to alter the side
chain because the sulfanyl group allows alkylation with substituted coumarins in a metabolically
stable manner (resistance to cytosolic and lysosomal hydrolases). The synthesis and biological
activity of the resulting compounds are described in the next sections.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 2. N-Fmoc amino acids (30 µM) were tested for inhibition of human sialin as in Figure 1.
The dotted lines correspond to the inhibition by compounds 3 and 4.
Chemistry. All 7-hydroxy 4-substituted coumarins were synthetized via Pechmann
condensation in good yields using either Brønsted acid with resorcinol or Lewis acid with an
aminophenol derivative (Scheme 1). Compounds 2528 were prepared according to conventional
methods with slight improvements36–38 (see Supporting information).
3
9
40
Using chlorotriisopropoxytitanium (IV) was more efficient than zinc (II) chloride for the
synthesis of 7-(dialkylamino)coumarins 29 and 30. The xanthenyl derivative 32 was obtained
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
41
according to a known procedure as the 3-(3,6-dihydroxy-9H-xanthen-9-ylidene)propionic acid
tautomer, unambiguously shown by NMR in aprotic solvent (DMSO).
Regioselective protection of the 7-hydroxy-4-hydroxymethyl-coumarin 27 with methoxymethyl
chloride before introducing the 9H-fluoren-9-yl methyl carbonate in the benzylic position and
deprotection of the 7-hydroxyl led to 35.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 1. Synthesis of functionalized coumarinsa
Cl
(ii)
MeO
HO
O
O
Cl
Cl
2
6
7
(i)
OH
HO
OH
OH
HO
O
O
O
O
2
5
8
(iii)
O
O
2
CO H
2
Cl
Cl
(iv)
(v)
R
N
R
HO
HO
OH
N
R
O
O
R
2
R= Me
R= Et
29: R= Me
30: R= Et
CO H
CO H
2
2
(v)
(vi)
N
OH
N
O
O
HO
OH
HO
HO
O
O
OH
O Fmoc
3
1
3
32
35
O Fmoc
(ix)
OH
(vii)
(viii)
27
MOMO
O
O
MOMO
O
O
O
3
34
a
Reagents and conditions: (i) Ethyl-4-chloroacetoacetate, 70% aq H SO , 0 °C, 1.5 h, 97%; (ii)
K CO , acetone, Me SO , RT, overnight, 94%; (iii) H O, reflux, 48 h, 69%; (iv) Ethyl-4-
chloroacetoacetate, CH SO H RT, 4 h, 97%; (v) (a) ClTi(OiPr) , dimethyl 3-oxoglutarate, toluene,
reflux, overnight; (b) NaOH, MeOH, RT, overnight, 3566% (two steps); (vi) Succinic anhydride,
2
4
2
3
2
4
2
3
3
3
1
0
ACS Paragon Plus Environment

Page 11 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
7
0% aq H SO , 140 °C, 4 h, 92%; (vii) MOM-Cl, DIEA, CH Cl , 0 °C, 45 min, 71%; (viii) Fmoc-
2
4
2
2
Cl, pyridine, RT, 2 h, 34%; (ix) TFA, CH Cl , 0 °C then RT, 1.5 h, 83%.
2
2
Amide bond formation between the amine of a lysine-OMe derivative and acids 29-32 as well
as commercially available 7-hydroxy-4-carboxymethyl-coumarin and 4-carboxyquinolin-2-one
were carried out42 (Scheme 2). Attempts to condense previously activated 7-hydroxy-4-
carboxymethyl coumarin as an acid chloride or HOBt ester in presence of EDC with Lys (Cbz)-
OMe failed. In contrast, amide bond formation occurred in presence of HBTU, HOBt via uronium
salt activation43,44 with this 4-carboxymethyl-coumarin or with xanthene 32, leading after
saponification of the methyl ester to 36 and 40 in poor yields (19% and 8%, respectively). The
condensation of 29, 30, 4-carboxyquinolin-2-one and 31 with Lys(Cbz)-OMe performed in the
presence of HBTU in CH Cl also led to the expected products, generally in better yields
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
2
4
5
(1250%) . In contrast, under these conditions, the coupling of 29 to the distal amine of lysine 43
was poorly efficient (5%, not optimized). The carbamate analog 44 of 36 was obtained from 27
with carbonyl diimidazole activation.
Scheme 2. Synthesis of N-acyl and N-carbamate diprotected lysinea
1
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
CO H
2
O
O
CO2H
(CH2)4
NHCbz
NHCbz
.
(i)
N
H
HCl Lys(Cbz)-OMe
+
36
HO
O
O
HO
O
O
O
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
CO H
2
N
H
(CH2)4
(ii)
R = Me2 37
R = Et2 38
29
3
.
HCl Lys(Cbz)-OMe
+
+
0
R
N
R
O
CO H
2
O
CO H
2
(ii)
NHCbz
.
HCl Lys(Cbz)-OMe
39
40
N
H
(CH2)4
N
H
O
HN
O
HO
O
CO H
2
(i)
.
NHCbz
NHCbz
32
HCl Lys(Cbz)-OMe
+
N
H
(CH )
2 4
O
OH
O
CO H
2
N
H
(CH2)4
(ii)
.
HCl Lys(Cbz)-OMe
+
31
41
N
O
O
O
CO H
2
NHCbz
O
N
H
(CH2)4
(iii)
.
HCl Lys(Cbz)-OMe
+
+
27
29
42
44
HO
O
O
CO H
2
H
O
N
Fmoc-Lys-OMe
FmocHN
(CH )
2 4
(ii)
43
Me N
O
O
2
aReagents and conditions: (i) (a) HBTU, HOBt, DMF, DIEA, RT, overnight; (b) aq LiOH, THF,
RT, 2 h, 19% for 36, 8% for 40 (two steps); (ii) (a) HBTU, CH Cl , DIEA, RT, 2 h; (b) aq LiOH,
THF, RT, 2 h, 50% for 37, 19% for 38, 43% for 39, 12% for 41, 5% for 44 (two steps); (iii) (a)
CDI, DMF, DIEA, RT then 80 °C, 1.5 h, 21%; (b) aq LiOH, THF, RT, 2 h 21% (two steps).
2
2
1
2
ACS Paragon Plus Environment

Page 13 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Alkylation of Fmoc-Cys-OH by coumarins 25, 26 and 28 proceeded at room temperature
overnight in presence of an excess of base to give compounds 4547 around 30% yields (Scheme
3).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 3. Synthesis of S-alkyl cysteinea
Cl
R1
1
2
25: R = R = H
Fmoc-Cys-OH
+
1
2
2
6: R = H, R = Me
16
2
1
2
R O
O
O
28: R = Cl, R = H
(i)
FmocHN
S
CO H
2
1
2
4
4
4
5: R = R = H
R1
1
2
6: R = H, R = Me
1
2
7: R = Cl, R = H
2
R O
O
O
aReagents and conditions: (i) from 25 and 28; DIEA, THF, RT, 16 h, 35% and 32% respectively;
from 26: Et N, DMF, RT, 16 h, 28%.
3
Structure-Activity Analysis of the Heterocycle-Substituted Amino Acids. Compounds
3542, and 44 (see structures in Supplementary Information, pp. S3-S4) were tested in our whole-
cell assay of sialin at 30 µM (Figure 3) as for the Fmoc-amino acids 1124. When 9-
fluorenylmethyl carbonate is directly linked to the coumarin 27, the resulting compound 35 does
not significantly inhibit sialin, showing that removal of the amino acid backbone and the free
carboxylate abolishes binding. Replacing the 9-fluorenylmethyl carbamate linked to the  amine
in 3 by diverse heterocycles in compounds 3641 reduced activity, ranging from a moderate loss
with the bulky coumarin derivative 41 to virtually no inhibition with the quinolinyl compound 39.
No significant difference was observed among coumarinyl derivatives 3638 bearing different
1
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
substituents at the 7-position (hydroxy, dimethyl or diethylamino). Of note, when the amide bond
of 36 was converted into a more flexible carbamate bond in 42, the affinity for sialin increased to
a level similar to that of compound 3.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The influence of substituting the side chain was tested in compounds 4447. Replacing the
benzyloxycarbonyl group present in 3 by a substituted coumarin in 44 did not significantly alter
activity. In contrast, the Fmoc-Cys(substituted coumarin)-OH 4547 had better activities than the
cysteine analog Fmoc-Cys(pMeOBzl)-OH 20 (Figure 3, red dotted line) or the lysine analog 3
(brown line), showing the length of the side-chain is not critical for recognition.
Figure 3. Screening of heterocycle-substituted amino acids. Compounds (30 µM) were tested for
inhibition of human sialin as in Figure 1. Each derivative is colored according to, and compared
with (dotted lines), its reference compound, 3 (brown) or 20 (red).
1
4
ACS Paragon Plus Environment

Page 15 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The most active compounds 11, 13, 4547 inhibited Neu5Ac uptake by sialin in a concentration-
dependent manner (Figure 4), in agreement with a specific interaction. Fmoc-Cys[Coum-7-
OH]-OH 45 (LSP12-3129) was the most potent inhibitor with an IC of 2.4 ± 0.7 µM (n = 3), a
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
0
3
3
value ~400-fold lower than the K for Neu5Ac. To assess its selectivity, compound 45 was tested
M
+
on two other H -driven lysosomal transporters, LYAAT1 and cystinosin, measured in whole-cell
assays similar to that of sialin.46,47 Interestingly, it had no effect on LYAAT1 and partially inhibited
cystinosin (43.7 ± 9.1% inhibition, n = 3) at a concentration (30 µM) that fully inhibits sialin
(Figure 4C, D).
1
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. Dose-response relationship and selectivity of selected compounds. (A) Inhibition of
human sialin. Means ± SEM of 2-3 independent experiments. (B) Representative sialin inhibition
curves for compounds 13 (blue) and 45 (green) (triplicate measurements). (C, D) Effect of 13 and
+
4
5 on two other H -driven transporters, LYAAT1 and cystinosin. (C) Representative experiments
(triplicate measurements). (D) Means ± SEM of 3 independent experiments where the 3
transporters were assayed in parallel.
Compounds 45 and 13 (Fmoc-Leu-OH) are non-competitive inhibitors. We next studied
how 45 and, for comparison, 13 (Fmoc-Leu-OH) interact with human sialin. Fmoc-Leu-OH was
1
6
ACS Paragon Plus Environment

Page 17 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
chosen because it efficiently inhibits Neu5Ac uptake (Figure 4B; IC = 23.3 ± 3.5 µM, n = 4) and
5
0
is commercially available as a radiolabeled compound. Of note, Fmoc-Leu-OH had limited effects
on LYAAT1 and cystinosin at a concentration (100 µM) that fully inhibit sialin (Figure 4C, D).
We first asked whether these compounds are substrates (translocated ligands) or blockers (not
translocated) of the lysosomal transporter. When sialin-expressing cells (EGFP-Sialin L22G/L23G
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
3
construct) and mock cells (EGFP) were incubated with [ H]Neu5Ac or [ H]Fmoc-Leu-OH at
similar saturation levels (IC /10, corresponding to 100 and 2 µM, respectively), no sialin-
5
0
3
dependent [ H]Fmoc-Leu-OH uptake could be detected in acidic medium, in contrast with the
3
strong [ H]Neu5Ac uptake signal (Figure S1). Fmoc-Leu-OH thus acts as a blocker of sialin rather
than a translocated substrate.
To extend these experiments to 45, we developed another whole-cell assay where sialin activity
+
2
is detected by the co-transport of H . We co-expressed the ratiometric fluorescent pH probe E GFP
4
8
with a mRFP-tagged human sialin L22G/L23G construct in HEK293 cells to detect the cytosolic
acidification associated with the uptake of sialin substrates. Neu5Ac (10 mM) applied at pH 5.5
2
induced a significant E GFP fluorescence response consistent with the expected acidification. In
contrast, a saturating concentration (30 µM) of 45 did not induce any detectable acidification
(Figure S2). Therefore, neither Fmoc-Leu-OH nor 45 are significantly translocated by human
sialin, implying they both act as blockers.
We next determined the inhibition mode of these compounds and studied the saturation kinetics
3
of [ H]Neu5Ac uptake at varying inhibitor concentrations. Unexpectedly, despite the ‘active site’
focus of our initial virtual screening, both Fmoc-Leu-OH and 45 acted as non-competitive rather
than competitive inhibitors since they decreased the Vmax of uptake while leaving the K for
M
Neu5Ac essentially unchanged (see representative experiments in Figure 5A and 5B). At 10 µM,
1
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Fmoc-Leu-OH decreased Vmax by 56 ± 5% (n = 5 independent experiments) with K values of 1.6
M
±
0.3 and 1.3 ± 0.2 mM in the absence and presence of inhibitor, respectively. Similarly, 45 (3
µM) decreased Vmax by 31 ± 8% (n = 3), with K values of 1.4 ± 0.5 and 1.7 ± 0.5 mM in the
M
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
absence and presence of inhibitor. Mean K values of 10.2 ± 3.0 µM (n = 3) and 7.8 ± 1.8 µM (n
i
=
4) were obtained for Fmoc-Leu-OH and 45, respectively.
Potential mechanism for the non-competitive inhibition. Conceivably, the non-competitive
+
inhibition by Fmoc-Leu-OH and 45 could reflect a decrease in the H electrochemical gradient
driving Nue5Ac uptake by sialin. However, the lack of effect of 45 on LYAAT1 and the moderate
inhibition of cystinosin by Fmoc-Leu-OH (Figure 4C, D) ruled out this possibility. As our virtual
33
screening was conducted on a cytosol-facing model of sialin, we reasoned that the lack of
competition between Neu5Ac, on one hand, and 45 or Fmoc-Leu-OH, on the other hand, might
result from a selective blockade of sialin in cytosol-facing state. Like any secondary active
transporter,49,50 sialin should operate through alternating-access structural transitions which expose
the substrate binding site to either the luminal (extracellular in our assay) or the cytosolic side of
the membrane.51 If an inhibitor binds to the Neu5Ac-binding site of the outward-facing
conformation (Figure 5C, inhibitor 1), the inhibitor and Neu5Ac should exclude each other and an
excess of Neu5Ac should displace the inhibitor binding equilibrium, and rescue inhibition. In
contrast, if the inhibitor permeates the cell membrane (for instance, by diffusing across the lipid
bilayer) and binds to the inward-facing state, the competition between Neu5Ac (on the
extracellular side) and the inhibitor (on the cytosolic side) should be biased toward inhibitor
binding (Figure 5C, inhibitor 2) because structural transitions are much slower than binding
equilibriums: when the cytosolic inhibitor dissociates, the protein in apo, inward-facing state will
1
8
ACS Paragon Plus Environment

Page 19 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
rebind another inhibitor molecule before it can switch to the apo, outward-facing state.
Extracellular Neu5Ac thus cannot displace the inhibitor binding equilibrium in this case, resulting
in non-competitive inhibition. The non-competitive inhibition of sialin by Fmoc-Leu-OH and 45
may thus result from their selective, or preferential, binding to the cytosol-facing state. It may be
noted that Fmoc-Leu-OH does enter cells in a sialin-independent manner (Figure S1), enabling its
action on the cytosolic side of the transporter.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 5. Inhibition mode of Fmoc-Leu-OH and 45. (A, B) Representative Eadie-Hofstee graphs
3
(
left) of [ H]Neu5Ac uptake saturation kinetics in presence or absence of Fmoc-Leu-OH 13 and
4
5 (means ± SEM of triplicate measurements). The reciprocal maximal velocity is plotted against
inhibitor concentration (right graphs) to determine inhibition constants. (C) Potential mechanism
accounting for the non-competitive inhibition by aromatic N-substituted amino acids (inhibitor 2).
2
0
ACS Paragon Plus Environment

Page 21 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Black arrows depict the Neu5Ac transport cycle in our assay. Red arrows depict two inhibition
mechanisms.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
We used two approaches to test this model. First, we examined how fast sialin recover its
transport activity upon washing of N-substituted amino acid inhibitors. If these inhibitors act on
the extracellular side, washing should quickly restore Neu5Ac transport. On the contrary, if they
predominantly act on the cytosolic side, recovery should be rate-limited by the cellular efflux of
cytosolic inhibitors. Cells expressing sialin at the plasma membrane were thus preincubated with
3
Neu5Ac or N-substituted amino acids at pH 5.0, washed at neutral pH and tested for [ H]Neu5Ac
uptake at pH 5.0 in the absence of inhibitor. For comparison, another set of cells was inhibited
3
during [ H]Neu5Ac uptake, as done previously. When this protocol was applied to Neu5Ac 1, the
3
neutral wash fully restored [ H]Neu5Ac transport to a level (130%) above that of the uninhibited
control, probably reflecting the trans-stimulation of sialin by the cytosolic pool of Neu5Ac
5
2
accumulated during the pre-incubation step (see ref. for a similar trans-stimulation in lysosomal
membrane vesicles). In contrast, when sialin was pre-inhibited by Fmoc-Leu-OH or 45, only
partial recovery was achieved after a 15-min wash (55% and 49% persisting inhibition,
respectively) (Figure 6), in agreement with our hypothesis of an action on the cytosolic, rather than
the extracellular, side of the transporter. Of note, when this experiment was repeated with
3
3
FR139317 2, which inhibits sialin in a competitive manner , the 15-min wash fully restored
[3H]Neu5Ac transport without trans-stimulation (Figure S3). The competitive vs. noncompetitive
nature of sialin inhibitors thus correlates with the rate of Neu5Ac transport recovery upon washing,
in agreement with an action on distinct sides of the membrane. The lower lipophilicity of
2
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
FR139317 (logP = 2.29 in protonated state) as compared to Fmoc-Leu-OH and 45 (logP = 4.1 and
3
.91, respectively) is also consistent with this model.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 6. Persistence of 13- and 45-induced inhibition upon washing. HEK293 cells expressing
sialin at the plasma membrane were pre-incubated or not with Neu5Ac 1 (5 mM), Fmoc-Leu-OH
13 (120 µM) or 45 (12 µM) for 15 min at pH 5. After 15- or 30-min wash at pH 7 at room
3
temperature, [ H]Neu5Ac uptake was measured and compared to the signals obtained in absence
or presence of competitors at the same concentrations (means ± SEM of 4 independent
experiments). t-test relative to control: **= p<0.01; *** = p<0.001; **** = p<0.0001; NS, ≥0.05).
2
2
ACS Paragon Plus Environment

Page 23 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Secondly, we examined the docking of 45 to a novel 3D homology model of human sialin based
on a closer template. Very recently, two crystal structures of E. coli H+/D-galactonate symporter
(DgoT) in outward-facing occluded and inward-facing open conformations have been
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
51
determined . DgoT is a closer homologue of human sialin (25.5% sequence identity and 45.0%
similarity over 420 amino acids; BlastP alignment) than E. coli GlpT (21.0% identity and 37.9
similarity over 309 amino acids). Moreover, the DgoT structures have a higher resolution and a
better global model quality estimation than the GlpT structure34,51. Accordingly, we built two new
sialin homology models based on the inward-facing open (6E9N) and outward-facing occluded
(6E9O) conformations of DgoT and performed docking experiments with flexible side-chains
5
3
located at the potential binding site, using GOLD software as implemented in Discovery Studio
019.
5 could not be docked to the outward-facing occluded sialin model because the binding site
2
4
was too small to accommodate the inhibitor. In contrast, 45 was well accommodated in the binding
site of the inward-facing model (Figure 7A), providing a putative structural explanation for the
non-competitive inhibition by this compound. The selected pose was further submitted to a 20-ns
molecular dynamics simulation in a membrane environment to optimize the 3D structure.
2
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 7. Docking of 45 to a DgoT-based homology model of human sialin. (A) Side view of the
inward-facing homology model with the docked inhibitor (in green). (B, C) Residues involved
(purple in B) and nature of their interaction with 45 shown in a 2D-depiction.
The interaction of 45 with the inward-facing sialin model is depicted in Figure 7B, C. The
carboxylic group of sialic acids and sialin inhibitors, a feature shared with all substrates of SLC17
transporters, is bound to a conserved arginine (R57 in sialin) and two tyrosine (Y119 and Y306).
Similar interactions of D-galactonate with this conserved arginine (R47) and two other tyrosine
Y44, Y79) are essential for substrate recognition and selectivity in DgoT.51
(
2
4
ACS Paragon Plus Environment

Page 25 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Two hydrophobic moieties of 45 are expected to block the alternating-access transition of the
transporter: a Fmoc group on the amino function of the cysteine core and a coumarinyl group on
the side chain of that amino acid. On one side, the 9H-fluorene part of the Fmoc group makes
hydrophobic interactions with F179, P180, L426 and several van der Waals contacts with I124,
H183, G427 and N430. The carbonyl of the Fmoc group establishes a hydrogen bond with Q123.
On the other side of the ligand, the coumarinyl group forms hydrogen bonds with S407, a 
stacking with Y203 and van der Waals contacts with L199 and Y301. In between, F179 contacts
the sulfur atom and the carboxylate (Figure 7B, C). In MFS transporters, helical discontinuities in
specific TMs act as molecular hinges for the rotation of the N and C domains during alternating-
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
49
access transitions . In DgoT, TM4 has a kink at P135 (P180 in sialin) and TM10 is continuous in
the outward-facing conformation whereas TM4 is continuous and TM10 has a kink in the inward-
5
1
facing conformation . The tight interactions of 45 with residues close to these kinks (Y179, P180
in TM4; S407 in TM10) in the sialin model should block alternating-access movements, providing
an explanation for its inhibition of Neu5Ac transport.
Compound 45 partially rescues the sialin R39C mutant. The R39C mutation causing Salla
disease induces a trafficking defect of sialin6,27 possibly due to its destabilization and its
recognition by chaperones of the protein quality control system. We reasoned that 45 binding to
the cytosolic side of sialin might stabilize nascent R39C polypeptides in the endoplasmic reticulum
and rescue their delivery to lysosomes.
4
5 (12 µM) efficiently inhibited the residual transport activity of sialin R39C in our whole-cell
assay (Figure S4), showing 45 binding is not impaired by the pathogenic mutation. We thus
transiently expressed wild-type (WT) or R39C human sialin fused to EGFP (with an intact sorting
2
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
motif) in HeLa cells in the absence or presence of 45 (70 µM) and examined their intracellular
distribution by fluorescence microscopy. As this distribution varies across cells in a given
condition (Figure 8A), sialin was compared to a lysosomal marker (LAMP1) and the dual staining
in individual cells was classified into three categories: lysosomal (≥70% overlap between sialin
and LAMP1 puncta); non-lysosomal (≤30% overlap); and mixed. In agreement with earlier
studies, the lysosomal category predominated in untreated cells expressing WT sialin (83%)
whereas this category dropped to 11% for the R39C mutant with a predominance of the mixed
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(55%) and non-lysosomal (34%) categories. However, when sialin R39C was expressed in the
presence of 45 (72 µM), the lysosomal category increased to 32% at the expense of the non-
lysosomal category (15%) while this treatment did not alter the distribution of WT sialin (Figure
8
B). Similar results were observed when sialin was expressed by lipofection instead of
electroporation (Figure S4). 45 thus partially rescues the trafficking defect of the R39C mutant.
To confirm this effect, we repeated these experiments and quantitated the colocalization between
EGFP-sialin and LAMP1 using scatter plots and Pearson’s correlation coefficient analysis of the
fluorescence intensities. This analysis showed a much lower level of sialin R39C/LAMP1
colocalization relative to the wild-type. When sialin R39C-expressing cells were treated with a
high dose (300 µM) of 45, colocalization was rescued to a level close to that of the wild-type
(Figure 8C).
Next, we tested whether this increased delivery of sialin R39C to lysosomes could decrease sialic
acid storage in patient cells. We used Salla fibroblasts from a compound-heterozygote patient
R39C and L336W+frameshift alleles) because these cells accumulate sialic acid to higher levels
(
5
4
than homozygous R39C cells. Compound-heterozygote Salla fibroblasts and control fibroblasts
were thus treated, or not, with 45 (30 to 170 µM) for 2 days and their level of free sialic acid was
2
6
ACS Paragon Plus Environment

Page 27 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
assayed by mass spectrometry. Salla fibroblasts cultured in standard medium accumulated free
sialic acid ~8-fold as compared to control fibroblasts. The 45 treatment did not reduce, but instead
slightly increased, this accumulation (Figure 8C), probably reflecting an inhibition of sialin R39C
by 45 at the lysosomal membrane. To circumvent this effect, we used a pulse-chase protocol in
which the drug was applied for 2 days to correct the trafficking defect followed by a 6-hr chase in
drug-free medium to remove lysosomal inhibition. However, this treatment did not reverse sialic
acid storage (data not shown).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 8. Effect of compound 45 on the R39C mutant. (A, B) Wild-type or R39C human sialin
tagged with EGFP was transiently expressed in HeLa cells by electroporation in the presence or
absence of 45 (70 µM). After two days, cells were fixed and analyzed under fluorescence
2
8
ACS Paragon Plus Environment

Page 29 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
microscopy using LAMP1 immunostaining (red) to detect late endosomes and lysosomes. Sialin
localization was categorized as illustrated in (A) in a blind manner. The graph (B) synthesizes the
2
outcome of 3 independent experiments (162 to 184 cells per condition;  test: ***, p<0.001; ****,
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
p<0.0001). (C) In independent experiments, colocalization was quantitated using scatter plots of
the sialin and LAMP1 pixel intensities. The graphs show the distribution of Pearson’s correlation
coefficient across 20 to 25 cells per condition in a representative example of three independent
experiments (Kruskal-Wallis one-way ANOVA with post hoc Dunn’s test: **, p<0.01;***,
p<0.001). (D) Skin fibroblasts from a healthy subject (WT) and a compound heterozygote
(p.Arg39Cys/Leu336Trpfs) Salla patient were cultured for 2 days with or without 45 (170 µM)
and assayed for free sialic acid level by micro-LC/ESI MRM-MS3 (means ± SEM of 4 independent
experiments; t-test: * = p<0.05; ** = p<0.01).
DISCUSSION
3
3
In this study, we exploited our previous virtual screening of human sialin and report chemical
substitutions of validated virtual hits. This led to the identification of a novel ligand scaffold
unrelated to sugar substrates, which is characterized by an amino acid backbone, a free
carboxylate, an N-linked aromatic or heteroaromatic substituent and a hydrophobic side chain.
Two ligands with this scaffold, Fmoc-Leu-OH 13 and 45, were characterized in more detail.
They inhibited Neu5Ac transport with IC of 23 and 2.4 µM, respectively, as compared to an IC
50
5
0
of ~1 mM for Neu5Ac. Neither compound was translocated by sialin (Figures S1 and S2) and,
somewhat surprisingly, both inhibited Neu5Ac transport in a non-competitive, rather than a
competitive, manner (Figure 5), notwithstanding the substrate-binding pocket focus of our virtual
2
9
ACS Paragon Plus Environment