The effect of deoxyfluorination and: O -acylation on the cytotoxicity of N -acetyl-d-gluco- And d-galactosamine hemiacetals

Article abstract of DOI:10.1039/d1ob00497b

Fully acetylated deoxyfluorinated hexosamine analogues and non-fluorinated 3,4,6-tri-O-acylated N-acetyl-hexosamine hemiacetals have previously been shown to display moderate anti-proliferative activity. We prepared a set of deoxyfluorinated GlcNAc and GalNAc hemiacetals that comprised both features: O-acylation at the non-anomeric positions with an acetyl, propionyl and butanoyl group, and deoxyfluorination at selected positions. Determination of the in vitro cytotoxicity towards the MDA-MB-231 breast cancer and HEK-293 cell lines showed that deoxyfluorination enhanced cytotoxicity in most analogues. Increasing the ester alkyl chain length had a variable effect on the cytotoxicity of fluoro analogues, which contrasted with non-fluorinated hemiacetals where butanoyl derivatives had always higher cytotoxicity than acetates. Reaction with 2-phenylethanethiol indicated that the recently described S-glyco-modification is an unlikely cause of cytotoxicity.

Full text of DOI:10.1039/d1ob00497b

Organic &
Biomolecular Chemistry
View Article Online
View Journal
PAPER
The effect of deoxyfluorination and O-acylation
on the cytotoxicity of N-acetyl-D-gluco- and
D-galactosamine hemiacetals†
Cite this: DOI: 10.1039/d1ob00497b
a,b
Vojtěch Hamala,a,b Lucie Červenková Šťastná, a Martin Kurfiřt,
Petra Cuřínová,a Martin Balouch, c Roman Hrstka,d Petr Voňkad and
a
Jindřich Karban
*
Fully acetylated deoxyfluorinated hexosamine analogues and non-fluorinated 3,4,6-tri-O-acylated
N-acetyl-hexosamine hemiacetals have previously been shown to display moderate anti-proliferative
activity. We prepared a set of deoxyfluorinated GlcNAc and GalNAc hemiacetals that comprised both fea-
tures: O-acylation at the non-anomeric positions with an acetyl, propionyl and butanoyl group, and deox-
yfluorination at selected positions. Determination of the in vitro cytotoxicity towards the MDA-MB-231
breast cancer and HEK-293 cell lines showed that deoxyfluorination enhanced cytotoxicity in most ana-
logues. Increasing the ester alkyl chain length had a variable effect on the cytotoxicity of fluoro analogues,
which contrasted with non-fluorinated hemiacetals where butanoyl derivatives had always higher cyto-
toxicity than acetates. Reaction with 2-phenylethanethiol indicated that the recently described S-glyco-
modification is an unlikely cause of cytotoxicity.
Received 15th March 2021,
Accepted 26th April 2021
DOI: 10.1039/d1ob00497b
rsc.li/obc
GlcNAc 1),9 4-deoxy-4-fluoro-N-acetyl-D-glucosamine (4F-Ac3-
GlcNAc 2),10 4-deoxy-4-fluoro-N-acetyl-D-galactosamine (4F-Ac3-
Introduction
Fluorinated carbohydrates have been long recognized as GalNAc 3),10 and 4,6-dideoxy-4,6-difluoro-N-acetyl-D-galactosa-
mechanistic probes and inhibitors for carbohydrate-processing mine (4,6-diF-Ac2-GalNAc 4)10 displayed moderate cytotoxicity
enzymes and carbohydrate-binding proteins.1 The similarity against L1210 murine leukaemia cells (IC50 24–35 μM, Fig. 1).
between fluorine and a hydroxyl group or hydrogen with Their non-acetylated counterparts were inactive due to poor
respect to size has also stimulated research into the potential cellular permeability. Fully acetylated 6,6-difluoro-L-fucose (6,6-
of fluorosugars to act as glycan chain modifiers by inhibiting diF-Ac4-Fuc 5) and 6,6,6-trifluoro-L-fucose (6,6,6-triF-Ac4-Fuc 6)
carbohydrate-processing enzymes and/or by affecting the inhibited proliferation of the human colon cancer cells
balance between carbohydrate metabolites.1,2 While this HCT116 (IC50 43 μM and 58 μM, respectively, Fig. 1).11
research identified several potent metabolic inhibitors,3–8
notably acetylated derivatives of (3R)-3-fluoro-N-acetylneurami- example, fully acetylated
The cytotoxicity depended on the acetylation pattern.8 For
showed very low cytotoxicity
2
nic acid3,4 and 2-deoxy-2-fluoro-L-fucose,3,5 it also revealed that towards the human prostate cancer cell line PC-3
some fluorosugars exhibited cytotoxicity, which complicated (IC50 277 μM). However, its hemiacetal (or lactol) derivative 7
their use for remodelling cell-surface glycans, but opened the with an unprotected anomeric hydroxyl (C1-OH) was cytotoxic
possibility of therapeutic use.2 For example, fully acetylated (IC50 61 μM) whereas 1,3-di-O-acetylated derivative 8 and fully
derivatives of 3-deoxy-3-fluoro-N-acetyl-D-glucosamine (3F-Ac3- de-O-acetylated derivative 9 were inactive (Fig. 1, the inset).
We clearly observed enhancement of cytotoxicity towards
PC-3 and ovarian A2780 cell lines when going from fully acetyl-
aInstitute of Chemical Process Fundamentals of the CAS, v. v. i., Rozvojová 135,
ated 3F-GalNAc analogue 10 to its hemiacetal 13 (Fig. 1). The
16502 Praha 6, Czech Republic. E-mail: karban@icpf.cas.cz
bUniversity of Chemistry and Technology Prague, Technická 5, 16628 Praha 6,
cytotoxicity enhancement for hemiacetals 11 and 12 relative to
their acetylated counterparts 1 and 3, respectively, occurred with
A2780 cells for 11 and PC-3 cells for 12.12 Lactol 7 was more cyto-
toxic to ovarian A2780 cells than fully acetylated 2 was.12
These results were consistent with the findings of the
Yarema group who showed that non-fluorinated 3,4,6-tri-O-
butanoyl derivatives of N-acetyl-D-glucosamine (GlcNAc) and
Czech Republic
cDepartment of Chemical Engineering, University of Chemistry and Technology,
Technická 3, 166 28 Prague 6, Prague, Czech Republic
dResearch Centre for Applied Molecular Oncology, Masaryk Memorial Cancer
Institute, Žlutý kopec 7, Brno, 65653, Czech Republic
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
d1ob00497b
This journal is © The Royal Society of Chemistry 2021
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
Fig. 1 The structures and cytotoxicity expressed as IC50 [µM] of cyto-
toxic acetylated fluorinated hexoses.8,10–12 Cytotoxicity against PC-3
and A2780 cells were determined using 24 h exposure if not otherwise
indicated.
Fig. 2 Structures of the target deoxyfluorinated analogues.
N-acetyl-D-mannosamine (ManNAc) hemiacetals were more
cytotoxic toward MDA-MB-231 breast cancer cells and Jurkat
cells (at the concentration range approximately 30–50 μM, IC50
values were not given) than their fully O-butanoylated counter-
parts.13 The corresponding 1,3,4-tri-O-butanoyl derivatives
having an unprotected primary (O6) hydroxyl group were non-
toxic. Lower cytotoxicity of the acetylated ManNAc hemiacetal
relative to its butanoylated analogue was attributed to dimin-
ished cell membrane permeability.13 Ensuing studies by the
Yarema group confirmed the importance of the 3,4,6-tri-O-acyl
pattern for the antiproliferative effect of non-fluorinated hexo-
samine hemiacetals.14–16
These results motivated us to investigate more systematically
the cytotoxicity of hexosamine analogues that would combine
structural features contributing to in vitro cytotoxicity: (1) deoxy-
fluorination, (2) an unprotected anomeric hydroxyl group, and
(3) protection of the non-anomeric OH groups with short C2–C4
fatty acid esters. We hypothesized that by combining these
structural features we might obtain fluoro analogues possessing
cytotoxic activity in low micromolar range with potential utility
in the development of cancer chemotherapy.
protected with acetyl, propionyl, or butanoyl (butyryl) esters
(Fig. 2). We determined the in vitro cytotoxicity of these deoxy-
fluorinated hexosamine hemiacetals using the cancerous
MDA-MB-231 cell line and non-cancerous HEK-293 cell line
and compared it with the cytotoxicity of the corresponding
acetylated and butanoylated non-fluorinated hemiacetals.
Deoxyfluorination enhanced cytotoxicity in the majority of
cases but it remained lower than that of the cisplatin bench-
mark. The influence of the ester alkyl chain length differed
from the trend reported for non-fluorinated derivatives. The
possibility that the cytotoxicity resulted from the recently
described reaction of GlcNAc hemiacetals with the thiol group
of cysteine residues17 was investigated by reaction with
2-phenylethanethiol.
Results
Apart from fluorine introduction, the main synthetic challenge
Therefore, we prepared a panel of mono-, di-, and trifluori- was orthogonal protection of the anomeric position. Building
nated GlcNAc and GalNAc analogues with an unprotected on our previous results,12,18,19 we designed a synthesis that
anomeric hydroxyl group and the remaining hydroxyl groups relies on the use of a thiophenyl moiety for the temporary pro-
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2021

View Article Online
Organic & Biomolecular Chemistry
Paper
tection of the anomeric position and an azide as a masked trifluoride (DAST) in the next step as reported by Lin et al.25
2-acetamido group. Synthesis of the majority of the target ana- The separation was possible for the galacto-configured pro-
logues follows the approach in Scheme 1. The C-2 azido group ducts 56–60 and 4-fluoro gluco-derivatives 53–55, giving frac-
and C-3 and C-4 fluorine substituents have been introduced tions enriched in the α-anomer (α/β ≈ 10 : 1) but failed for
using 1,6-anhydropyranose chemistry as we previously 3-fluoro gluco-derivatives 50–52.
described.12 Opening of the internal acetal by treatment with
The opening of the intramolecular acetal by reaction with
phenyl trimethylsilyl sulfide/ZnI2 introduced the thiophenyl PhSTMS was accompanied by the formation of low quantities
moiety at the anomeric carbon and liberated the C-6 hydroxyl (≤11% by 19F NMR analysis) of side-products detectable by
for acylation or deoxyfluorination.18,20 Standard hydrolysis of TLC very near the spot of the phenyl thioglycoside. Some of
the thioglycosidic bond and transformation of the azido group them (see ESI,† the synthesis of compounds 54–56, and 59)
into an acetamido functionality completed the synthesis. The resulted from the pyranose ring-contraction discussed in our
6-fluoro analogues were prepared by modification of reported previous report.26 Some of the side-products were only partially
methods.21,22 Selection of the deoxyfluorination and separable by chromatography, but were removed via thioglyco-
O-acylation patterns was partly the matter of synthetic conven- side hydrolysis in the subsequent steps.
ience. We prepared 3-, 4-, and 6-monodeoxyfluorinated, 3,6-
and 4,6-difluorinated, and 3,4,6-trifluorinated analogues 58–60 was deoxyfluorinated by treatment with DAST under
shown in Fig. 2.
microwave conditions.27 Treatment of the resulting difluoro
The primary C6 hydroxyl of hexopyranoses 50, 53–55, and
The synthesis is detailed in Scheme 2. The starting deoxy- derivatives with NBS in acetone/water effected thioglycoside
fluorinated 1,6-anhydro-2-azido-pyranoses 35–38 were pre- hydrolysis and yielded hemiacetals 61–67 (Scheme 2c). The
pared as previously described.12 The O-acylation of the 3- and product of DAST-mediated migration of the β-configured thio-
4-positions provided compounds 39–49 (Scheme 2a). Cleavage phenyl moiety was not formed from the β-anomer of 50 in
of the internal acetal and installation of the thiophenyl moiety accordance with our previous findings about the O4-benzy-
19,20,23
at C1 on reaction with PhSTMS/ZnI2
yielded intermedi- lated variant of 50 (R2 = OBn).26 As thioglycosides 53–55 and
ates 50–60 as a mixture of anomers (Scheme 2b). 3-O-Acyl 58–60 were subjected to deoxyfluorination as the α-anomers
gluco-derivatives 53–55 produced a relatively high proportion only, thiophenyl migration could not occur. Thioglycosides 50,
of the α-anomer (α/β = 5.3 : 1, 4.5 : 1 and 3.8 : 1 for acetyl, pro- 51, 54–57 and 59–60 were O6-acylated and then hydrolysed at
pionyl and butyryl, respectively). Comparable α-selectivity (α/β C1 to afford hemiacetals 68–75 (Scheme 2d). Hemiacetals
= 6 : 1) was also reported by Demchenko et al. for a non-fluori- 61–75 were subjected to azide/acetamide transformation by
nated counterpart of 53 (R2 = OBn) and might indicate a treatment with thioacetic acid in pyridine28 to produce the
remote participation of the O3-acyl in the acetal cleavage.24 We target 2-acetamido hemiacetals 11, 14, 16–17, 20–23, 25–28
attempted chromatographic separation of the thioglycoside and 31–33 (Scheme 3). The products of this reaction were
anomers resulting from 1,6-anhydro bridge cleavage because occasionally contaminated by traces of the anomeric
the β-anomer was expected to undergo unwanted thiophenyl O-acetates (see below formation of 87/88, Scheme 4b) that were
migration from C1 to C6 on reaction with diethylaminosulfur removed by chromatography and/or recrystallization.
Reversing the order of the last two steps and conducting
azide/acetamide transformation prior to thioglycoside hydro-
lysis (as with thioglycosides 51 and 52, Scheme 4a) was incon-
venient because hydrolysis of the thiophenyl moiety vicinal to
C2 acetamide in intermediates 76 and 77 was sluggish and
incomplete as judged by thin-layer chromatography and the
moderate yields (Scheme 4a).
The approach outlined in Scheme 1 was initially adopted
also for the preparation of 6-fluoro D-gluco analogues. Thus,
known
2-azido-4-O-benzyl-1,6-anhydropyranose
7829,30
(Scheme 4b) was O3-acylated, de-O-benzylated by treatment
with NaBrO3/Na2S2O4,31 and the liberated C4 hydroxyl was acy-
lated to give di-O-acyl derivatives 81 and 82. Compounds 81
and 82 were converted to hemiacetals 85 and 86 via thioglyco-
sides 83 and 84 by reaction with PhSTMS, followed by deoxy-
fluorination at C6,32 and hydrolysis at the anomeric position
(Scheme 4b). The final step, however, was problematic because
azide/acetamide transformation gave expected hemiacetals 18
and 19 in mixture with O1-acetylated products 87 and 88,
respectively, which were difficult to separate chromatographi-
cally. The structures of the inseparable C1-OAc by-products are
supported by NMR assignment (see ESI†) and also by the
Scheme 1 An overview of the synthesis of 3F, 4F, 3,6-diF, and 4,6-diF
analogues.
This journal is © The Royal Society of Chemistry 2021
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
Scheme 2 Synthesis of deoxyfluorinated 2-azido precursors.
finding that a sample of compound 18 was converted to com- corresponding esters. Hydrogenolytic debenzylation afforded
pound 87 by acetylation in pyridine. the target non-fluorinated acylated hemiacetals 99, 100, 102,
Although the anomeric acetates 87 and 88 could probably and 103. (Scheme 5C) Monofluorinated acetylated derivatives 7
be converted to desired hemiacetals by established methods,33 and 12 and trifluorinated analogues 24 and 34 were prepared
we investigated an alternative route, in which reaction with by published procedures.12,26
thioacetic acid was avoided, and the anomeric centre was pro-
The in vitro cytotoxicity of fluorinated analogues towards
tected as a benzyl glycoside (Scheme 5A). Readily available the MDA-MB-231 cell line derived from human triple negative
benzyl 2-acetamido-D-glucopyranoside 8934 was tert-butyldiphe- breast adenocarcinoma and the human embryonic kidney
nylsilylated at O6 and benzoylated at O3 and O4 in one step to HEK-293 cell line39,40 was determined using MTT cell viability
give fully protected glucosamine 90. Acid-catalysed desilyla- assay after 72 h treatment (Table 1) and expressed as IC50
tion35 provided alcohol 91. Compound 91 was first deoxyfluori- values. The MDA-MB-231 cell line was chosen because Yarema
nated under microwave conditions and the resulting 6-fluoro et al. reported its sensitivity to acylated hexosamine hemiace-
analogue 92 subjected to debenzoylation under Zemplén con- tals.13 The HEK-293 cell line was chosen to investigate the
ditions to yield diol 93. O-Acylation and hydrogenolytic deben- effects of fluoro analogues on non-cancerous cells. Fluorinated
zylation provided the desired target fluoro analogues 18 and 19 analogues showed cytotoxicity in the range 15–92 μM, indicat-
(Scheme 5A). Very recently, butane-2,3-diacetal protection was ing that all of our analogues were less cytotoxic than the cispla-
effectively applied to shorten a similar reaction sequence.36 tin benchmark (entry 29). Furthermore, there was no selectivity
The 6-fluoro D-galacto analogues 29 and 30 were obtained from for cancerous MDA-MB-231 cells over HEK-293 cells. In fact,
readily available benzyl 2-acetamido-galactoside 9437 by the syn- the HEK-293 cell line was often more sensitive to GlcNAc ana-
thetic sequence comprising deoxyfluorination with DAST to logues than the MDA-MB-231 line was (e.g., entries 13–15).
produce 95, subsequent hydrolysis of the isopropylidene ketal Increasing the length of the ester alkyl chain increased the
to diol 96, O-acylation, and hydrogenolytic debenzylation to cytotoxicity of non-fluorinated hexosamines in accordance
yield hemiacetals 29 and 30 (Scheme 5B). A similar approach with the literature data (entries 1 vs. 2 and 16 vs. 17). However,
was utilized to prepare non-fluorinated derivatives used as this trend did not translate to fluorinated analogues because
reference compounds for the cytotoxicity study. Hence, known we did not observe the expected positive correlation between
benzyl 2-acetamido-glucoside 8934 and 2-acetamido-galactoside the length of the ester alkyl chain and cytotoxicity. The effect
10138 were acylated at all available hydroxyls to produce the on cytotoxicity going from acetyl to propionyl esters depended
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2021

View Article Online
Organic & Biomolecular Chemistry
Paper
Scheme 3 N3 to NHAc transformation to obtain target fluoro analogues 11, 14, 16–17, 20–23, 25–28 and 31–33.
on the cell line (entries 12 vs. 13) and configuration (entries 6 and 32 on migration ability of MDA-MB-231 cells using a
and 7 vs. 20 and 21). Butanoylation, however, often decreased scratch assay.42 Compounds 20 and 32 were chosen because of
cytotoxicity, particularly in the MDA-MB-231 line (entries 7 vs. their good cytotoxicity whereas compounds 15 and 30 rep-
8, 13 vs. 14, 18 vs. 19, 23 vs. 24, and 26 vs. 27).
resented O-butanoyl analogues. None of the tested compounds
Deoxyfluorination usually enhanced cytotoxicity relative to exhibited anti-migratory properties in the scratch assay (data
non-fluorinated hemiacetals, especially for fluorinated GalNAc not shown).
analogues—all of these had lower IC50 values than their non-
We attempted to simulate membrane permeation of com-
fluorinated counterparts 102 and 103 (entries 16 and 17 vs. pounds 20–22 (4,6-diF-GlcNAc analogues), 31–33 (4,6-
18–28). The increase in cytotoxicity upon fluorination was pro- diF-GalNAc analogues), and 99 and 102 (acetylated non-fluori-
nounced for Ac3-GalNAc-1-OH that was almost non-toxic in the nated analogues) using the COSMOPerm43 approach.
MDA-MB-231 line (entry 16). The contribution of fluorination Dipalmitoylphosphatidylcholine (DPPC) and dioleoylphospha-
at the 4-position to cytotoxicity was usually superior to fluori- tidylcholine (DOPC) were used as common model membranes.
nation at the 6-position (entries 7 vs. 9, 21 vs. 23, and 22 vs. No clear correlation between the measured cytotoxicity and cal-
24) and quite frequently also to fluorination at the 3-position culated permeation coefficients log Perm was found. The mole-
(entries 12 vs. 11, 6 vs. 3, 22 vs. 19). Increasing the number of cules tend to permeate freely through the chosen model mem-
fluorine substituents improved cytotoxicity in about a half of branes considering the free permeation threshold approxi-
the cases, and trifluorinated (entries 15 and 28) and some 4,6- mately log Perm = −8 (see ESI† for details).
difluorinated (e.g. entries 12 and 26) analogues belonged to
Qin et al. have recently suggested that 3,4,6-tri-O-acetylated
the most cytotoxic compounds.
GlcNAc non-specifically and non-enzymatically reacts with the
Since per-O-butanoylated ManNAc reduced cell migration,41 thiol group of cysteine residues to produce proteins covalently
we determined the inhibitory effect of analogues 15, 20, 30 labelled with thiolated amino sugars.17 Only cysteine residues
This journal is © The Royal Society of Chemistry 2021
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
located in a basic microenvironment of proteins rich in lysine
residues are likely to react. This reaction, which was termed
S-glyco-modification, was modelled by incubation of hemiace-
tal 99 with 2-phenylethanethiol under alkaline conditions
(aqueous Na2CO3) for 1 h (Scheme 6A).17 Following acetylation,
a mixture of 3-deoxy-3-thio amino sugars was isolated in which
compounds 104–106 were identified by NMR.17 These products
resulted from a conjugate addition of 2-phenylethanethiol to
enals 108 and 109 followed by closure of the pyranose or fura-
nose ring. Enals 108 and 109 interconvert by acetyl migration
and were formed by E1cB elimination of acetic acid from the
open chain form 107 of lactol 99. In addition, base-catalysed
deacetylation of 99 proceeded in parallel with elimination.17
We hypothesized that S-glyco-modification of proteins may
contribute to the mechanisms by which our fluoro analogues
exerted their cytotoxicity. However, cytotoxic difluorinated ana-
logue 20 reacted with 2-phenylethanethiol much more slowly
than 99 under conditions described in ref. 17, and no products
were detectable by TLC after 1 h (Scheme 6B, conditions (a)).
After 24 h, the NMR analysis indicated deacetylated analogue
112 as the main component together with a low quantity of
the expected elimination-addition products 110 and 111 and
other unidentified products (Scheme 6B, conditions (b)).
Using sodium methanolate in methanol accelerated the for-
mation of the thiolated products 110 and 111 (Scheme 6B, con-
ditions (c)) so that they could be isolated in about 90% and
60% purity, respectively. Attempts to induce formation of 110
or 111 using excess Et3N in THF failed, most likely because the
Scheme 4 (a) Slow hydrolysis of 2-acetamido-thioglycosides 76 and
77; (b) attempted synthesis of 6-fluoro analogues 18 and 19.
Scheme 5 (A) Synthesis of 6F-GlcNAc analogues 18 and 19; (B) synthesis of 6F-GalNAc analogues 29 and 30; (C) synthesis of non-fluorinated
derivatives 99, 100, 102, and 103, and structures of the remaining fluorinated target analogues 7, 12, 24 and 34 prepared by known procedures.12,26
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2021

View Article Online
Organic & Biomolecular Chemistry
Paper
Table 1 Cytotoxicity towards MDA-MB-231 and HEK-293 cell lines
expressed as IC50 [μM], obtained after 72 h treatment using the MTT
assay
Entry Analogue Substitution pattern
MDA-MB-231 HEK-293
GlcNAc analogues
1
2
3
4
5
6
7
8
99
100
11
14
15
7
16
17
18
19
23
20
21
22
24
Ac3-GlcNAc-1-OH
96 10
61 10
Bu3-GlcNAc-1-OHa
3F-Ac2-GlcNAc-1-OH
3F-Pr2-GlcNAc-1-OHb
3F-Bu2-GlcNAc-1-OH
4F-Ac2-GlcNAc-1-OH
4F-Pr2-GlcNAc-1-OH
4F-Bu2-GlcNAc-1-OH
6F-Pr2-GlcNAc-1-OH
6F-Bu2-GlcNAc-1-OH
49
49
50
46
28
49
74
9
9
7
6
7
2
5
42
51
39
21
30
27
2
7
6
3
8
5
44 10
46
52 10
47
33 16
9
61 13
55 11
5
10
11
12
13
14
15
3,6-F2-Ac-GlcNAc-1-OH 48
4,6-F2-Ac-GlcNAc-1-OH 25 10
4,6-F2-Pr-GlcNAc-1-OH
4,6-F2-Bu-GlcNAc-1-OH 54
3,4,6-F3-GlcNAc-1-OH
9
7
41
7
3
7
17
19
15
5
3
8
29
GalNAc analogues
16
17
18
19
20
21
22
23
24
25
26
27
28
29
102
103
25
26
12
27
28
29
30
31
32
33
34
Ac3-GalNAc-1-OH
Bu3-GalNAc-1-OH
414 130
136 40
37 13
123 26
69 17
3F-Pr2-GalNAc-1-OH
3F-Bu2-GalNAc-1-OH
4F-Ac2-GalNAc-1-OH
4F-Pr2-GalNAc-1-OH
4F-Bu2-GalNAc-1-OH
6F-Pr2-GalNAc-1-OH
6F-Bu2-GalNAc-1-OH
4,6-F2-Ac-GalNAc-1-OH 61
4,6-F2-Pr-GalNAc-1-OH 20
4,6-F2-Bu-GalNAc-1-OH 54
35
34
4
6
61 13
38
27
38
60
4
6
3
9
50 11
40
42
38
4
9
5
92 12
43 11
59 11
5
4
8
3
30
37 12
26
3.8 0.5
6
3,4,6-F3-GalNAc-1-OH
28
2
Scheme 6 Reaction of 2-phenylethanethiol with (A) non-fluorinated
hemiacetal 99,17 (B) 4,6-difluorinated analogue 20, (C) 3-fluorinated
analogue 11.
cisPt
3.7 0.6
a Bu = butanoyl. b Pr = propionyl (in Table 1 only).
stereochemistry and protecting group pattern.14,46–51
Replacing an ester with a fluorine was shown to bring new
biological activity and ameliorate pharmacokinetic pro-
interconversion between the cyclic and open-chain forms was
very slow under aprotic conditions.
Acylated sugars are hydrolyzed by cytoplasmic esterases and
the S-glyco-modification of proteins can only proceed as long
as the ester at the 3-position remains intact because it func-
tions as the leaving group in the elimination step. Fluorine at
the 3-position can also act as a leaving group and, unlike an
ester, cannot be readily hydrolyzed. Accordingly, reaction of
3-fluoro analogue 11 with the MeONa/MeOH/2-phenyletha-
nethiol and subsequent acetylation mostly yielded the
expected thiolated product 104 together with a low amount of
manno-configured product 113, both arising from elimination
of HF and subsequent addition of 2-phenylethanethiol
(Scheme 6C). Formation of thiolated N-acetyl-mannosamine
derivatives 111 and 113 could also result from a base-catalysed
C2 epimerization of glucosamine products 110 and 104,
respectively.44,45 NMR analysis did not detect signals of pre-
viously identified compounds 105 and 106.17
perties of hexosamines.7,8,52–55
A subset of hexosamine
hemiacetals carrying short fatty acid esters (C2 to C4) at
the 3-, 4- and 6-positions and an amide at the 2-position
are cytotoxic; some of them have other antitumor pro-
perties including antimetastatic effects.41 The effect of
fluorination on cytotoxicity of these derivatives have not yet
been systematically studied except for 4-fluoro GlcNAc.8,53
This is the first comparative study of cytotoxicity using a
panel of deoxyfluorinated hexosamine lactols.56 The use of
1,6-anhydropyranose chemistry for regio- and stereoselective
installation of azide and fluorine at C2–C4 and a thiophe-
nyl group at C112,18,19,57 was adapted in this report for the
synthesis of selectively fluorinated, ester-protected hexo-
samine hemiacetals. This methodology is flexible and can
be adjusted for the regioselective synthesis of fluorinated
hexosamines carrying different acyl groups at preselected
positions (including N-acyl) because these groups were
installed sequentially at different stages of the synthesis. It
should be noted that intermediate fluorinated 2-azido
Discussion
Derivatives based on the hexosamine scaffold display an thioglycosides and hemiacetals obtained by this method-
array of useful bioactivities depending on the scaffold ology are also valuable glycosyl donors for the synthesis of
This journal is © The Royal Society of Chemistry 2021
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
glycoconjugates containing fluorinated glucosamine and For example, 4-fluoro GlcNAc was converted to nucleotide sugar
galactosamine.19,58 4-fluoro-UDP-GlcNAc inside cells,8 and Keenan et al. reported
Fluorination of 3,4,6-O-acylated GlcNAc/GalNAc moderately anomeric phosphorylation of some mono and multi-fluorinated
enhanced their cytotoxicity except for essentially non-cytotoxic hexopyranoses by the action of anomeric kinases.61
3,4,6-O-acetylated GalNAc where the increase in cytotoxicity
The lack of correlation between in vitro modelled S-glyco-
upon fluorination was pronounced. The most cytotoxic fluoro modification and growth inhibition suggests that other mecha-
analogues had IC50 values approximately five-fold higher than nisms may contribute to the cytotoxicity of fluoro analogues.
cisplatin did. Although cisplatin was more cytotoxic in the They can include the effect of fluorination on the membrane
tested cell lines, it also causes severe adverse effects in chemo- permeability or the rate at which the esters are hydrolyzed by
therapy and has to be administered in low doses while fluoro cytosolic esterases. Furthermore, fluorinated hexosamines or
sugars might be tolerated at higher concentrations. In vivo products of their enzymatic conversions can trigger mecha-
experiment will be necessary to investigate this possibility.
nisms leading to cell death by inhibiting key carbohydrate-pro-
We confirmed that for non-fluorinated analogues, increasing cessing enzymes or other critical targets.
the ester alkyl chain length from acetate to butanoate resulted
in a parallel increase in cytotoxicity in accordance with the lit-
erature.16 Improved efficiency of longer alkyl chains was attribu-
ted to improved diffusion through the cell membrane due to
higher lipophilicity.16 However, this trend did not operate for
our fluorinated analogues and was reversed in the case of
4-fluoro GlcNAc and 4,6-difluoro GlcNAc analogues using
MDA-MB-231 cell line (Table 1, entries 6–8 and 12–14). Recently
Stephenson et al. observed an even more pronounced reversal
of cytotoxicity of acetylated, propionylated, and butanoylated
4-fluoro GlcNAc hemiacetals 7, 16 and 17, respectively, against
human neurons at 100 μM.53 The authors suggested that exces-
sive lipophilicity added by longer alkyls might have slowed
down the diffusion from cell membrane to cytoplasm although
it remains unclear why this argument should not also apply to
non-fluorinated hemiacetals. Unlike neurons, astrocytes dis-
played good cell viability at a concentration of 100 μM of
4-fluoro GlcNAc analogue 7.53 This suggests that higher suscep-
tibility of the HEK-293 cell line relative to MDA-MB-231 cells to
the cytotoxic effect of our GlcNAc analogues could result from a
close similarity between HEK-293 cells and neuronal cells.39,59
The 3,4,6-O-acetylated N-acyl-hexosamines can react with
thiol groups of cysteine residues17,49,60 by an elimination-
addition mechanism (S-glyco-modification) as long as the
3-position remains O-acetylated.17 An increased cysteine
modification by 3,4,6-O-acetylated N-azidoacetyl-galactosamine
(Ac3-GalNAz) relative to the corresponding per-O-acetate (Ac4-
GalNAz) was reported to correlate with the increased cytotoxicity
Conclusions
A set of 3,4,6-O-acylated mono-, di- and trifluorinated GlcNAc/
GalNAc lactols has been prepared. The synthetic methodology
allows for the synthesis of regioselectively protected multifluori-
nated 2-azido-thioglycosides having potential in the synthesis of
fluorinated glycoconjugates. Cell-based assay showed that deox-
yfluorination of acylated GlcNAc/GalNAc hemiacetals generated
compounds cytotoxic against the MDA-MB-231 and HEK-293
cell lines, suggesting that they may also inhibit tumour growth
in vivo. A follow-up in vivo study can decide whether the lower
in vitro cytotoxicity compared to cisplatin is compensated by
better tolerability. Model reaction with 2-phenylethanethiol
indicated that cysteine modification is an unlikely mechanism
leading to cell death. Glycoproteomic analysis is required to
confirm this hypothesis. Future research can also focus on the
investigation of the influence of fluorination on the lipophilicity
and the possible correlation between the lipophilicity, mem-
brane permeability and cytotoxicity using a 19F NMR-based
method for lipophilicity measurement.62 Furthermore, 19F NMR
methods may be adapted to determine the rate of transmem-
brane transport of fluorinated analogues and give insight into
their transformations inside cells.63–67
Author contributions
of Ac3-GalNAz compared to Ac4-GalNAz suggesting that S-glyco- Conceptualization: J.K. and V.H.; organic synthesis: V.H., J.K.
modification was responsible for the cell death.49 However, for and M.K.; spectroscopic characterization: L.C. and P.C.; tests
the fluoro analogues tested in this work the extent of S-glyco- of cytotoxicity: P.V. and R.H.; computation M.B.; article
modification did not correlate with cytotoxicity provided that writing: J.K. with input from all authors.
the reaction with 2-phenylethanethiol correctly modelled this
process. The cytotoxic difluoro analogue 20 reacted with 2-phe-
nylethanethiol sluggishly whereas the less cytotoxic non-fluori-
nated counterpart 99 was significantly more reactive. In
addition, 3-fluoro analogue 11 was predominantly converted to
Conflicts of interest
There are no conflicts to declare.
3-thio-products and therefore should be more cytotoxic than
compound 20 that mostly gave a deacetylated product, whereas
the opposite was true. Alternatively, compound 11 might have
Acknowledgements
been quickly removed by cellular metabolism once the O4 and We thank the Czech Science Foundation (grant no. 17-18203S)
O6 esters were hydrolysed before it could react with cysteine and Ministry of Education, Youth and Sport (INTER COST, no.
thiol groups. Fluorinated sugars can indeed be metabolized. LTC20052) for supporting our research. This publication is
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2021

View Article Online
Organic & Biomolecular Chemistry
Paper
partially based on research done within the framework of 16 R. T. Almaraz, U. Aich, H. S. Khanna, E. Tan,
COST Action CA18103 (INNOGLY) supported by European
Cooperation in Science and Technology. RH and PV were sup-
R. Bhattacharya, S. Shah and K. J. Yarema, Biotechnol.
Bioeng., 2012, 109, 992–1006.
ported by MH CZ - DRO (MMCI, 00209805). MB acknowledges 17 K. Qin, H. Zhang, Z. Zhao and X. Chen, J. Am. Chem. Soc.,
support from Karel Berka (Palacky University Olomouc) for the 2020, 142, 9382–9388.
access to TURBOMOLE, MacroModel, LigPrep, COSMOThermX 18 M. Kurfiřt, L. Červenková Št’astná, M. Dračínský,
and COSMOperm calculations.
M. Müllerová, V. Hamala, P. Cuřínová and J. Karban, J. Org.
Chem., 2019, 84, 6405–6431.
19 V. Hamala, L. Červenková Šťastná, M. Kurfiřt, P. Cuřínová,
M. Dračínský and J. Karban, Org. Biomol. Chem., 2020, 18,
5427–5434.
Notes and references
1 B. Linclau, A. Arda, N. C. Reichardt, M. Sollogoub, 20 L.-X. Wang, N. Sakairi and H. Kuzuhara, J. Chem. Soc.,
L. Unione, S. P. Vincent and J. Jimenez-Barbero, Chem. Soc.
Rev., 2020, 49, 3863–3888.
Perkin Trans. 1, 1990, 1677–1682.
21 Z. A. Morrison and M. Nitz, Carbohydr. Res., 2020, 495, 108071.
2 S. Goon and C. R. Bertozzi, J. Carbohydr. Chem., 2002, 21, 22 M. Sharma, G. G. Potti, O. D. Simmons and W. Korytnyk,
943–977. Carbohydr. Res., 1987, 162, 41–51.
3 C. D. Rillahan, A. Antonopoulos, C. T. Lefort, R. Sonon, 23 M. Kurfiřt, L. Č. Šťastná, P. Cuřínová, V. Hamala and
P. Azadi, K. Ley, A. Dell, S. M. Haslam and J. C. Paulson,
Nat. Chem. Biol., 2012, 8, 661–668.
4 T. Heise, J. F. A. Pijnenborg, C. Büll, N. van Hilten,
J. Karban, J. Org. Chem., 2021, 86, 5073–5090.
24 A. V. Demchenko, M. A. Wolfert, B. Santhanam, J. N. Moore
and G.-J. Boons, J. Am. Chem. Soc., 2003, 125, 6103–6112.
E. D. Kers-Rebel, N. Balneger, H. Elferink, G. J. Adema and 25 P.-C. Lin, A. K. Adak, S.-H. Ueng, L.-D. Huang, K.-T. Huang,
T. J. Boltje, J. Med. Chem., 2019, 62, 1014–1021.
J.-a. A. Ho and C.-C. Lin, J. Org. Chem., 2009, 74, 4041–
5 N. M. Okeley, S. C. Alley, M. E. Anderson, T. E. Boursalian,
4048.
P. J. Burke, K. M. Emmerton, S. C. Jeffrey, K. Klussman, 26 V. Hamala, L. Červenková Št’astná, M. Kurfiřt, P. Cuřínová,
C.-L. Law, D. Sussman, B. E. Toki, L. Westendorf, W. Zeng,
X. Zhang, D. R. Benjamin and P. D. Senter, Proc. Natl. Acad.
Sci. U. S. A., 2013, 110, 5404–5409.
M. Dračínský and J. Karban, Beilstein J. Org. Chem., 2021,
accepted, Beilstein Arch. 2021, 202110. DOI: 10.3762/
bxiv.2021.10.v1.
6 J. G. Allen, M. Mujacic, M. J. Frohn, A. J. Pickrell, 27 C. Mersch, S. Wagner and A. Hoffmann-Röder, Synlett,
P. Kodama, D. Bagal, T. San Miguel, E. A. Sickmier, 2009, 2167–2171.
S. Osgood, A. Swietlow, V. Li, J. B. Jordan, K.-W. Kim, A.-M. 28 R. V. Kolakowski, N. Shangguan, R. R. Sauers and
C. Rousseau, Y.-J. Kim, S. Caille, M. Achmatowicz, O. Thiel, L. J. Williams, J. Am. Chem. Soc., 2006, 128, 5695–5702.
C. H. Fotsch, P. Reddy and J. D. McCarter, ACS Chem. Biol., 29 J. Karban, M. Budesinsky, M. Cerny and T. Trnka, Collect.
2016, 11, 2734–2743. Czech. Chem. Commun., 2001, 66, 799–819.
7 S. R. Barthel, A. Antonopoulos, F. Cedeno-Laurent, 30 D. Hesek, M. Lee, W. Zhang, B. C. Noll and S. Mobashery,
L. Schaffer, G. Hernandez, S. A. Patil, S. J. North, A. Dell, J. Am. Chem. Soc., 2009, 131, 5187–5193.
K. L. Matta, S. Neelamegham, S. M. Haslam and 31 M. Adinolfi, G. Barone, L. Guariniello and A. Iadonisi,
C. J. Dimitroff, J. Biol. Chem., 2011, 286, 21717–21731. Tetrahedron Lett., 1999, 40, 8439–8441.
8 S.-I. Nishimura, M. Hato, S. Hyugaji, F. Feng and 32 Thioaglycon migration did not occur in an appreciable
M. Amano, Angew. Chem., Int. Ed., 2012, 51, 3386–3390. extent in agreement with our findings reported in ref. 26.
9 M. Sharma, R. J. Bernacki, M. J. Hillman and W. Korytnyk, 33 S. M. Andersen, M. Heuckendorff and H. H. Jensen, Org.
Carbohydr. Res., 1993, 240, 85–93. Lett., 2015, 17, 944–947 and references therein.
10 M. Sharma, R. J. Bernacki, B. Paul and W. Korytnyk, 34 T. Yamaguchi, D. Hesek, M. Lee, A. G. Oliver and
Carbohydr. Res., 1990, 198, 205–221. S. Mobashery, J. Org. Chem., 2010, 75, 3515–3517.
11 Y. Dai, R. Hartke, C. Li, Q. Yang, J. O. Liu and L. X. Wang, 35 E. M. Nashed and C. P. J. Glaudemans, J. Org. Chem., 1987,
ACS Chem. Biol., 2020, 15, 2662–2672. 52, 5255–5260.
12 S. Horník, L. Červenková Šťastná, P. Cuřínová, J. Sýkora, 36 S. J. Richards, T. Keenan, J. B. Vendeville, D. E. Wheatley,
K. Káňová, R. Hrstka, I. Císařová, M. Dračínský and
J. Karban, Beilstein J. Org. Chem., 2016, 12, 750–759.
13 U. Aich, C. T. Campbell, N. Elmouelhi, C. A. Weier,
S. G. Sampathkumar, S. S. Choi and K. J. Yarema, ACS
Chem. Biol., 2008, 3, 230–240.
14 Z. Wang, J. Du, P.-L. Che, M. A. Meledeo and K. J. Yarema,
Curr. Opin. Chem. Biol., 2009, 13, 565–572.
15 C. T. Saeui, L. Liu, E. Urias, J. Morrissette-McAlmon,
H. Chidwick, D. Budhadev, C. E. Council, C. S. Webster,
H. Ledru, A. N. Baker, M. Walker, M. C. Galan, B. Linclau,
M. A. Fascione and M. I. Gibson, Chem. Sci., 2021, 12, 905–
910.
37 L. Rochepeau-Jobron and J.-C. Jacquinet, Carbohydr. Res.,
1998, 305, 181–191.
38 K. J. Gregg, M. D. L. Suits, L. Deng, D. J. Vocadlo and
A. B. Boraston, J. Biol. Chem., 2015, 290, 25657–25669.
R. Bhattacharya and K. J. Yarema, Mol. Pharm., 2018, 15, 39 G. Shaw, S. Morse, M. Ararat and F. L. Graham, FASEB J.,
705–720.
2002, 16, 869–871.
This journal is © The Royal Society of Chemistry 2021
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
40 K. C. Tjandra, C. R. Forest, C. K. Wong, S. Alcantara,
H. G. Kelly, Y. Ju, M. H. Stenzel, J. A. McCarroll,
M. Kavallaris, F. Caruso, S. J. Kent and P. Thordarson, Adv.
Healthcare Mater., 2020, 9, 2000261.
E. Hostetler, C. Sur, L. Zhang, J. B. Schachter, J. F. Hess,
H. G. Selnick, D. J. Vocadlo, E. J. McEachern, J. M. Uslaner,
J. L. Duffy and S. M. Smith, J. Pharmacol. Exp. Ther., 2020,
374, 252–263.
41 C. T. Campbell, U. Aich, C. A. Weier, J. J. Wang, S. S. Choi, 55 H. G. Selnick, J. F. Hess, C. Tang, K. Liu, J. B. Schachter,
M. M. Wen, K. Maisel, S.-G. Sampathkumar and
K. J. Yarema, J. Med. Chem., 2008, 51, 8135–8147.
42 J. E. N. Jonkman, J. A. Cathcart, F. Xu, M. E. Bartolini,
J. E. Amon, K. M. Stevens and P. Colarusso, Cell Adh. Migr.,
2014, 8, 440–451.
43 J. A. H. Schwöbel, A. Ebert, K. Bittermann, U. Huniar,
K.-U. Goss and A. Klamt, J. Phys. Chem. B, 2020, 124, 3343–
3354.
J. E. Ballard, J. Marcus, D. J. Klein, X. Wang, M. Pearson,
M. J. Savage, R. Kaul, T. S. Li, D. J. Vocadlo, Y. Zhou,
Y. Zhu, C. Mu, Y. Wang, Z. Wei, C. Bai, J. L. Duffy and
E. J. McEachern, J. Med. Chem., 2019, 62, 10062–10097.
56 The cytotoxic activities of compounds 7 and 11–13 were
previously determined by us using 24 h incubation with
PC-3 and A2780 cells (ref. 12). These results are not com-
parable with the results presented here due to short incu-
bation time of 24 h, which does not allow for cells to
divide.
44 M. Hanchak and W. Korytnyk, Carbohydr. Res., 1976, 52,
219–222.
45 W. L. Salo, M. Hamari and L. Hallcher, Carbohydr. Res., 57 J. Karban, J. Sykora, J. Kroutil, I. Cisarova, Z. Padelkova and
1976, 50, 287–291. M. Budesinsky, J. Org. Chem., 2010, 75, 3443–3446.
46 N. Elmouelhi, U. Aich, V. D. P. Paruchuri, M. A. Meledeo, 58 A. F. G. Bongat and A. V. Demchenko, Carbohydr. Res.,
C. T. Campbell, J. J. Wang, R. Srinivas, H. S. Khanna and
K. J. Yarema, J. Med. Chem., 2009, 52, 2515–2530.
2007, 342, 374–406.
59 B. He and D. M. Soderlund, Neurosci. Lett., 2010, 469, 268–
47 C. Kim, O. H. Jeon, D. H. Kim, J. J. Chae, L. Shores,
272.
N. Bernstein, R. Bhattacharya, J. M. Coburn, K. J. Yarema 60 W. Qin, K. Qin, Y. Zhang, W. Jia, Y. Chen, B. Cheng,
and J. H. Elisseeff, Biomaterials, 2016, 83, 93–101.
48 J. Ma, C. Wu and G. W. Hart, Chem. Rev., 2021, 121, 1513–
1581.
49 Y. Hao, X. Fan, Y. Shi, C. Zhang, D.-e. Sun, K. Qin, W. Qin,
W. Zhou and X. Chen, Nat. Commun., 2019, 10, 4065.
50 N. Darabedian, B. Yang, R. Ding, G. Cutolo, B. W. Zaro,
C. M. Woo and M. R. Pratt, Front. Chem., 2020, 8, 318.
L. Peng, N. Chen, Y. Liu, W. Zhou, Y.-L. Wang, X. Chen and
C. Wang, Nat. Chem. Biol., 2019, 15, 983–991.
61 T. Keenan, F. Parmeggiani, J. Malassis, C. Q. Fontenelle,
J.-B. Vendeville, W. Offen, P. Both, K. Huang, A. Marchesi,
A. Heyam, C. Young, S. J. Charnock, G. J. Davies,
B. Linclau, S. L. Flitsch and M. A. Fascione, Cell Chem.
Biol., 2020, 27, 1199–1206.e5.
51 S.-U. Lee, C. F. Li, C.-L. Mortales, J. Pawling, J. W. Dennis, 62 B. Linclau, Z. Wang, G. Compain, V. Paumelle,
A. Grigorian and M. Demetriou, PLoS One, 2019, 14,
C. Q. Fontenelle, N. Wells and A. Weymouth-Wilson,
e0214253.
Angew. Chem., Int. Ed., 2016, 55, 674–678.
52 X. M. v. Wijk, R. Lawrence, V. L. Thijssen, S. A. v. d. Broek, 63 J. R. Potts, A. M. Hounslow and P. W. Kuchel, Biochem. J.,
R. Troost, M. v. Scherpenzeel, N. Naidu, A. Oosterhof, 1990, 266, 925–928.
A. W. Griffioen, D. J. Lefeber, F. L. v. Delft and 64 S. Bresciani, T. Lebl, A. M. Z. Slawin and D. O’Hagan,
T. H. v. Kuppevelt, FASEB J., 2015, 29, 2993–3002. Chem. Commun., 2010, 46, 5434–5436.
53 E. L. Stephenson, P. Zhang, S. Ghorbani, A. Wang, J. Gu, 65 R. E. London and S. A. Gabel, Biophys. J., 1995, 69, 1814–
M. B. Keough, K. S. Rawji, C. Silva, V. W. Yong and
C.-C. Ling, ACS Cent. Sci., 2019, 5, 1223–1234.
54 X. Wang, W. Li, J. Marcus, M. Pearson, L. Song, K. Smith,
1818.
66 J.-X. Yu, R. R. Hallac, S. Chiguru and R. P. Mason, Prog.
Nucl. Magn. Reson. Spectrosc., 2013, 70, 25–49.
G. Terracina, J. Lee, K.-L. K. Hong, S. X. Lu, L. Hyde, 67 D. Shishmarev, C. Q. Fontenelle, I. Kuprov, B. Linclau and
S.-C. Chen, D. Kinsley, J. P. Melchor, D. J. Rubins, X. Meng,
P. W. Kuchel, Biophys. J., 2018, 115, 1906–1919.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2021