Addressing the Structural Complexity of Fluorinated Glucose Analogues: Insight into Lipophilicities and Solvation Effects

Article abstract of DOI:10.1002/chem.202002825

In this work, we synthesized all mono-, di-, and trifluorinated glucopyranose analogues at positions C-2, C-3, C-4, and C-6. This systematic investigation allowed us to perform direct comparison of ¹⁹F resonances of fluorinated glucose analogues and also to determine their lipophilicities. Compounds with a fluorine atom at C-6 are usually the most hydrophilic, whereas those with vicinal polyfluorinated motifs are the most lipophilic. Finally, the solvation energies of fluorinated glucose analogues were assessed for the first time by using density functional theory. This method allowed the log P prediction of fluoroglucose analogues, which was comparable to the C log P values obtained from various web-based programs.

Full text of DOI:10.1002/chem.202002825

Accepted Article
Accepted Article
Title: Addressing the structural complexity of fluorinated glucose
analogues: Insight into lipophilicities and solvation effects
Authors: Jacob St-Gelais, Émilie Côté, Danny Lainé, Paul A. Johnson,
and Denis Giguére
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.202002825
Link to VoR: https://doi.org/10.1002/chem.202002825
01/2020

10.1002/chem.202002825
Chemistry - A European Journal
Addressing the structural complexity of fluorinated glucose analogues: Insight into
lipophilicities and solvation effects
Jacob St-Gelais, Émilie Côté, Danny Lainé, Paul A. Johnson, Denis Giguère*
Département de Chimie, 1045 av. De la Médecine, Université Laval, Québec City, Qc, Canada G1V 0A6,
E-mail: denis.giguere@chm.ulaval.ca
Graphical abstract
Abstract
In this work, we synthesized all mono-, di-, and trifluorinated glucopyranose analogues at positions C-2, C-3, C-4,
and C-6. This systematic investigation allowed us to perform direct comparison of 19F resonances of fluorinated
glucose analogues and also to determine their lipophilicities. Compounds with a fluorine atom at C-6 are usually
the most hydrophilic whereas those with vicinal polyfluorinated motifs are the most lipophilic. Finally, the solvation
energies of fluorinated glucose analogues were assessed for the first time using the density functional theory. This
method allowed the log P prediction of fluoroglucose analogues, which was comparable with the Clog P values
obtained from various web-based programs.
This article is protected by copyright. All rights reserved.

10.1002/chem.202002825
Chemistry - A European Journal
Introduction
Carbohydrates are a structurally-diverse group of polyhydroxylated aldehydes or ketones natural products.
Their biological significance includes immune regulation, infection, and cancer metastasis. In addition, sugars are
important components of genetic materials and glycoproteins. Unfortunately, carbohydrates suffer from inherently
reduced binding affinities and poor pharmacokinetic properties. This can be explained by the nature of hydroxyl
groups forming weak hydrogen bonds with proteins and the energetically costly desolvation effects (i.e. the breaking
of watercarbohydrate interactions and desolvation of the binding site).1
In the unbound state, carbohydrates are part of a strong hydrogen-bonding network with solvent molecules.
Removal of water upon desolvation is energetically costly, leading to a positive associated free-energy difference.
It is estimated that the transfer of a hydroxyl group from an aqueous solution to gas phase has an energetic penalty
of G = 25 kJ mol-1.2 This outcome can be explained by an enthalpic penalty of H = 35 kJ mol-1 and a beneficial
entropic term of S = 10 kJ mol-1, due to the release of solvating water into bulk.3 Moreover, carbohydrates exhibit
a number of adjacent hydroxyl groups and a free-energy penalty of G = 34 kJ mol-1 is attributed to the desolvation
of vicinal diols (G = 17 kJ mol-1 per hydroxyl groups). This nonlinear desolvation effect can be explained by
shared hydration of neighboring functional groups. Taking all this into consideration, the hydrogen-bonding
interaction energy is about G = 1821 kJ mol-1 (maximal energy gain) and cannot compensate for the cost of its
desolvation.4 Hence, the reduction of desolvation penalties can improve binding affinity, exemplified by
carbohydrate-binding proteins E-selectin5 and galectin-3.6
Clearly, the desolvation penalties attributable to hydroxyl groups present an obstacle that has proven
challenging to overcome. Approaches to improve binding affinities and pharmacokinetic properties of
carbohydrates have been proposed via the use of bioisosteres of carbohydrates functional groups7 for the design of
glycomimetic drug candidates.8 As such, deoxygenation has been shown to reduce carbohydrates’ polar surface
area, allowing an increase in binding affinities by formation of new hydrophobic contacts with proteins, along with
reducing the enthalpic cost of desolvation.2b Similarly, deoxyfluorination of carbohydrates is a useful strategy for
improving biological molecular recognition through hydrophobic desolvation and electrostatic interactions through
polarized CF bonds.9 Important properties of the CF bond are depicted in Figure 1a. The carbon-fluorine bond
is a strong bond in organic chemistry, but is also highly polarised allowing the participation in multipolar
interactions (such as CFCO, CFNR4, and CFHC interactions).10 Moreover, fluorine can act as a
hydrogen-bond acceptor11 and its desolvation energies are smaller than for hydroxyl groups (G = 5 kJ mol-1 and
H = 2 kJ mol-1).3 Because of these unique properties, many pharmaceuticals containing fluorine atoms are
This article is protected by copyright. All rights reserved.

10.1002/chem.202002825
Chemistry - A European Journal
developed each year.12 Incorporation of fluorine can prevent oxidative metabolism and increase metabolic stability.
Nevertheless, there is a limited amount of research discussing the desolvation effects of organofluorines and to the
best of our knowledge there is no study reporting the desolvation energies of fluorinated carbohydrates. Moreover,
in relation with carbohydrates, bioisosteric replacement of hydroxyl groups with fluorine atoms afforded new
chemical entities with enhanced or specific affinities with immunoglobulin,13 glycogen phosphorylase,14 and UDP-
Gal mutase.15 Finally, fluorosugars have been used as a standard method for the stabilization of glycosidic bonds16
as well as probes for sugar epitope mapping.17
Figure 1. a) Properties of the CF bond; b) Medically relevant fluorinated glucose analogues 1 and 2; and c)
Fluorinated glucose analogues 316 prepared in this study.
Over the years, there has been much interest in the preparation of fluorinated glucose analogues. This class
of carbohydrates are now considered as valuable biological tools for biochemical investigations. For example, 2-
deoxy-2-[18F]fluoroglucose (18F-FDG) 1 is a well-known radiopharmaceutical used in medical imaging18 and
This article is protected by copyright. All rights reserved.

10.1002/chem.202002825
Chemistry - A European Journal
SAR7226 2 is a 4-deoxy-4-fluoroglucose sodium-glucose transporter inhibitor (Figure 1b).19 A vast amount of
research has also been implemented for the biological investigations and physical properties of polyfluorinated
glucose analogues.9 20 As part of our program related to the synthesis of fluorinated carbohydrates,21 we developed
a flexible strategy that allowed us the synthesis of a range of fluorinated carbohydrate analogues,22 enabling the
determination of their lipophilicity.21a, 23 In this work, we wish to report the first systematic investigation involving
the synthetic pathways to generate all mono-, di-, and trifluorinated glucopyranose analogues at position C-2, C-3,
C-4, and C-6 (Figure 1c), along with the determination of their corresponding lipophilicities (in the context of this
study, the term fluoroglucose refers to deoxyfluoroglucose). Moreover, systematic studies of the contribution of
fluorine atoms to desolvation energies are relatively infrequent. Therefore, we wish to report solvation energies of
fluorinated glucose analogues in order to study the impact of the position and the number of integrated fluorine
atoms on a pyran core.
Results and discussions
Our synthetic endeavors started with the preparation of all the monofluoroglucose analogues at position C-
2, C-3, C-4, and C-6 (Scheme 1). Thus, acetylated glucose 17 was transformed in two steps into known 3,4,6-tri-
O-acetyl-D-glucal 18 in 88% yield.24 Treatment of compound 18 with Selectfluor® led to a near 1:1 mixture of 2-
fluoroglucose and 2-fluoromannose analogues in 70% yield.25 Deprotection under basic conditions furnished the
desired 2-deoxy-2-fluoro-D-glucose 3 in 91% yield (/ = 1.1:1 in acetone-d6). Our next challenge involved the
preparation of 3-fluoroglucose. Previous synthesis of this compound used allofuranose derivatives as starting
material and the overall yields were quite low.26 Therefore, we used commercially available 1,6-anhydro--D-
glucopyranose (levoglucosan) 19 as a starting point to prepare compound 4. Benzylation of compound 19 at O-2
and O-4 proceeded in 80% yield and was followed by a deoxyfluorination step using Et3N3HF as practical fluorine
source.27 Thus, preparation of the trifluoromethansulfonate intermediate preceded nucleophilic fluorination with
neat Et3N3HF allowing formation of compound 20 with complete retention of configuration, presumably via
neighboring group participation of the benzyloxy at C-4.28 The configuration of the fluorine atom was ascertained
2
3
3
based on 19F NMR spectroscopy (19F NMR (470 MHz): JF-3,H3 = 45.9 Hz, JF-3,H-2 = JF-3,H-4 = 17.5 Hz). Direct
cleavage of the 1,6-anhydro bridge with BCl3 furnished the desired 3-fluoroglucose 4 in 82% yield (/ = 1:1 in
acetone-d6). We next turned our attention to the synthesis of 4-fluoroglucose 5. Mono-O-p-toluenesulfonylation of
compound 19 furnished known intermediate 21.29 The latter underwent epoxide formation under basic conditions
and benzylation, allowing formation of the known Cerny epoxide in 89% over 2 steps. Treatment with a mixture of
KHF2 and TBAF3H2O allowed formation of compound 22 as the sole isomer in 52% yield (81% based on recovered
This article is protected by copyright. All rights reserved.

10.1002/chem.202002825
Chemistry - A European Journal
starting material).20b Concomitant benzyl ether deprotection and cleavage of the anhydro bridge with BCl3 gave 4-
fluoroglucose 5 in 85% yield (/ = 2.3:1 in acetone-d6). Finally, the 6-fluoroglucose analogue 6 was prepared
according to known procedure.30 Thereby, DAST-mediated deoxyfluorination of methyl -D-glucopyranoside 23
yielded the corresponding 6-fluorinated glucoside 24 in 81% yield and treatment with HCl generated compound 6
in 93% yield (/ = 1:1.5 in acetone-d6).
c) Selectfluor
a) HBr/AcOH
then Ac2O
33%
OH
O
OAc
O
OAc
O
b) Zn, NH4Cl
HO
HO
AcO
AcO
AcO
AcO
d) NaOMe
91%
88%
F
OH
OH
AcO OAc
(2 steps)
18
17
3
F
: 2 -Glu
= 1.1:1
O
O
O
e) BaO, BnBr
80%
OH
O
OH
O
F
h) BCl3
HO
F
f) Tf2O, Pyr.
g) Et3N.3HF
88% (2 steps)
82%
= 1:1
HO
HO
OH
BnO
OBn
20
19
4
F
: 3 -Glu
i) TsCl (1 equiv)
40%
j) NaOMe
k) NaH, BnBr
89% (2 steps)
O
O
OH
O
OH
O
OH
O
h) BCl3
F
HO
85%
= 2.3:1
l) TBAF, KHF2
HO
OH
F
OBn
TsO
OH
52% (81% brsm)
22
F
21
5
F
: 4 -Glu
OH
F
O
n) HCl
O
m) DAST
HO
HO
O
HO
HO
HO
HO
81%
93%
= 1.5:1
HO
HO
HO OH
OMe
OMe
24
23
6
F
: 6 -Glu
Scheme 1. Synthesis of monofluoroglucose analogues 36. Reagents and conditions: a) HBr/AcOH (15 equiv),
CH2Cl2, 0 °C, 1 h; b) Zinc dust (7.5 equiv), NH4Cl (7.5 equiv), CH3CN, 60 °C, 1 h, 88% over 2 steps; c) Selectfluor®
(1.5 equiv), CH3NO2/H2O (5:1), 0 °C to rt, 16 h, then Ac2O (5 equiv), pyridine (10 equiv), CH2Cl2, rt, 2 h, 33% ; d)
1 M NaOMe, CH2Cl2/MeOH, rt, 1 h, 91% (/ = 1.1:1 in acetone-d6); e) BaO (4 equiv), BnBr (5 equiv), DMF, 60
°C, 4 h, 80%; f) Tf2O (2 equiv), CH2Cl2/pyridine (4:1) 0 °C to rt, 1 h; g) Et3N3HF (15 equiv), microwave heating,
100 °C, 2 h, 88% over 2 steps; h) BCl3 (6 equiv), CH2Cl2, 0 °C to rt, 2.5 h, 82% (/ = 1:1 in acetone-d6) for 4,
85% (/ = 2.3:1 in acetone-d6) for 5; i) TsCl (1 equiv), pyridine, toluene, 0 °C to rt, 18 h, 40%; j) 1 M NaOMe,
CH2Cl2/MeOH, rt, 2 h; k) BnBr (1.5 equiv), NaH (1.5 equiv), TBAI (0.3 equiv), DMF, 0 °C to rt, 2 h, 89% over 2
steps; l) TBAF3H2O (3 equiv), KHF2 (1.5 equiv), 180 °C, 3 h, 52% (81% brsm); m) DAST (6 equiv), CH2Cl2, 40
°C to rt, 2 h, 81% (/ = 1.5:1 in acetone-d6); n) HCl, water, 55 °C, 18 h, 93% (/ = 1:1.5 in acetone-d6).
This article is protected by copyright. All rights reserved.

10.1002/chem.202002825
Chemistry - A European Journal
Levoglucosan is a useful synthetic starting material for the synthesis of polyfluorinated carbohydrates. We
thus tackled the preparation of all difluoroglucose analogues at position C-2, C-3, C-4, and C-6 from levoglucosan
(Scheme 2). The synthesis of compound 11 started with triethylsilyl triflate-catalyzed acetolysis of intermediate 20
in high yield. Then, in order to successfully fluorinate at C-6, we protected the anomeric position. Glycosylation
with allyloxytrimethylsilane using our previously described protocol afforded intermediate 26.23 Deoxyfluorination
at C-6 was accomplished with acetyl deprotection under basic conditions, followed by treatment with DAST. Final
deprotection with BCl3 afforded the desired 3,6-difluoroglucose analogue 11 in 75% (/ =1.6:1 in acetone-d6). We
next turned our attention to the synthesis of 4,6-difluoroglucose 12 and 3,4-difluoroglucose 10. Compound 22 was
benzylated at O-3 in 97% yield and intermediate 28 was treated with a mixture of TESOTf and acetic anhydride.
Compound 29 was prepared in 94% yield and subjected to glycosylation with TMSOTf and allylOTMS affording
the corresponding allyl 4-fluoroglucoside 30 in 66% yield. Deprotection of the acetyl group was followed by DAST-
mediated fluorination in 84% yield over two steps. Final deprotection with BCl3 afforded 4,6-difluoroglucose 12 in
95% yield (/ =1.8/1 in acetone-d6). Additionally, fluorination at C-3 of intermediate 22 with retention of
configuration was possible via a triflate intermediate and subsequent treatment with Et3N3HF. Compound 32 was
isolated in 80% yield over two steps and deprotected using BCl3 yielding 3,4-difluoroglucose 10 in 89% yield (/
= 1.7/1 in acetone-d6). To complete our set of difluorinated glucose analogues, we next focus on the synthesis of
2,3-difluoroglucose 7, 2,4-difluoroglucose 8, and 2,6-difluoroglucose 9 from ditosylate 33.31 A known 4-step
sequence allowed the generation of intermediate 34 on gram scale.20a, 21a, 32 Benzylation of the latter intermediate in
high yield was followed by Lewis acid-catalysed acetolysis (36: 94% yield) and glycosylation (37: 66% yield).
Fluorination at C-6 followed a similar strategy than above: acetyl deprotection under basic conditions and
deoxyfluorination with DAST afforded compound 38 in 69% yield over two steps. Ether deprotections leaded to
2,6-difluoroglucose 9 in 70% yield (/ = 1.3/1 in acetone-d6). Furthermore, compound 7 was readily accessible
from intermediate 34 following a fluorination with retention of configuration. Generation of a triflate at C-3
followed treatment with Et3N3HF allowing the preparation of 39 in high yield. Benzyl ether deprotection and
cleavage of the 1,6-anhydro bridge was achieved with BCl3 yielding 2,3-difluoroglucose 7 in 97% yield (/ =
1.2/1 in acetone-d6). Lastly, 2,4-difluoroglucose 8 was accessed following our one-pot 4-step process involving a
series of epoxide formation-opening sequence in 60% yield on gram scale21a and treatment with BCl3 in 97% yield
(/ = 1.2/1 in acetone-d6).
This article is protected by copyright. All rights reserved.

10.1002/chem.202002825
Chemistry - A European Journal
Scheme 2. Synthesis of difluoroglucose analogues 712. Reagents and conditions: a) TESOTf (0.1 equiv), Ac2O
(excess), 30 °C to rt, 0.5 h, 98% (/ = 5.5:1) for 25, 94% (/ = 5.3:1) for 29, 98% (/ = 5.3:1) for 36; b)
AllylOTMS (10 equiv), TMSOTf (1 equiv), CH3CN, microwave heating, 90 °C, 0.75 h, 64% (/ = 1:1.1) for 26,
66% (/ = 1.5:1) for 30, 57% (/ = 1:1.6) for 37; c) 1M NaOMe, CH2Cl2/MeOH, rt, 1 h; d) DAST (3 equiv),
2,4,6-collidine (6 equiv), CH2Cl2, microwave heating, 100 °C 1 h, 78% over 2 steps (/ = 1:1) for 27, 84% over 2
steps (/ = 1.4:1) for 31, 69% over 2 steps (/ = 1:1.6) for 38; e) BCl3 (6 equiv), CH2Cl2, 0 °C to rt, 2.5 h, 75%
(/ = 1.6:1 in acetone-d6) for 11, 95% (/ = 1.8:1 in acetone-d6) for 12, 89% (/ = 1.7:1 in acetone-d6) for 10,
70% (/ = 1.3:1 in acetone-d6) for 9, 97% (/ = 1.2:1 in acetone-d6) for 8, 97% (/ = 1.2:1 in acetone-d6) for 7;
f) BnBr (1.5 equiv), NaH (1.5 equiv), TBAI (0.3 equiv), DMF, 0 °C to rt, 2 h, 97% for 28, 98% for 35; g) Tf2O (2
equiv), pyridine (10 equiv) 0 °C to rt, 1 h; h) Et3N3HF (15 equiv), 80 °C, 24 h, 80% over 2 steps for 32, 95% over
2 steps for 39; i) KHF2 (4 equiv), TBAF3H2O (8 equiv), 180 °C, 24 h, 60%.
This article is protected by copyright. All rights reserved.

10.1002/chem.202002825
Chemistry - A European Journal
The synthesis of all the trifluorinated glucose analogues at C-2, C-3, C-4, and C-6 have been previously
described by us and is summarized in Scheme 3.22, 23 Briefly, trifluoroglucose 1416 were accessed from compound
39, 43, and 32 respectively following a 4-step sequence: step 1 – TESOTf-catalyzed acetolysis; step 2 –
glycosylation with allyloxytrimethylsilane; step 3 – C-6 fluorination, and step 4 – global deprotection under acidic
conditions. As for the 2,3,4-trifluoroglucose 13, we first opted for an epimerization at C-3 using a Dess-Martin
oxidation followed by a reduction with NaBH4 starting from compound 50. Then, nucleophilic fluorination at C-3
proceeded with TBAF via a triflate intermediate, leading to volatile intermediate 52. Finally, acetolysis under acidic
conditions and acetyl deprotection with HCl afforded the desired compound.
Scheme 3. Synthesis of trifluoroglucose analogues 1316. Reagents and conditions: a) TESOTf (0.1 equiv), Ac2O
(excess), 0 °C to rt, 1 h, 92% (/ = 17:1) for 40, 94% (/ = 6.5:1) for 44, 63% (/ = 22:3) for 47; b) AllylOTMS
(10 equiv), TMSOTf (1 equiv), CH3CN, microwave heating, 90 °C, 0.75 h, 44% (/ = 1:1.7) for 41, 61% (/ =
1:1) for 45, 48% (/ = 1:1) for 48; c) 1M NaOMe, CH2Cl2/MeOH, rt, 1 h; d) DAST (3 equiv), 2,4,6-collidine (6
equiv), CH2Cl2, microwave heating, 100 °C 1 h, 85% over 2 steps (/ = 1:1.7) for 42, 95% over 2 steps (/ =
1:1) for 46, 85% over 2 steps (/ = 1:1) for 49; e) HCl (37% in water), H2O/acetone, 70 °C, 16 h, 66% (/ = 3:2
in acetone-d6) for 14, 85% (/ = 7:3 in acetone-d6) for 15, 66% (/ = 3:1 in acetone-d6) for 16, quant. (/ = 5:1
in acetone-d6) for 13; f) DMP (1.5 equiv), CH2Cl2, 0 °C, 1 h; g) NaBH4 (8 equiv), MeOH, 20 °C, 6 h, 61% over 2
steps; h) Tf2O (1.5 equiv), pyridine (3 equiv), CH2Cl2, rt, 0.5h; i) TBAF3H2O (3 equiv), 50 °C, 18 h; j) Ac2O (30
equiv), H2SO4 (10 equiv), 0 °C to rt, 16 h, then NaOAc (20 equiv), rt, 0.3 h, 14% over 3 steps (/ = 5.2:1).
This article is protected by copyright. All rights reserved.

10.1002/chem.202002825
Chemistry - A European Journal
With this unique set of fluoroglucose analogues in hand, we used 19F NMR spectroscopy to probe the
4
conformation of fluoro sugars in deuterated acetone. Firstly, all glucose analogues adopt standard C1-like
conformations (Figure 2). Comparison of the vicinal and geminal coupling constants for each anomeric pair
suggests that there is little change in the conformation of the molecules. 33 Also, the  anomer is favored in solution
for all analogues. This is in comparison to the fact that the  anomer is preferred for D-glucopyranose and
monofluoroglucopyranoses in deuterated water.34 Moreover, analogues 6, 916 with a fluorine atom at C-6 adopt a
GG conformation for both anomers.35 This information could be extracted from the 19F NMR spectrum (470 MHz,
acetone-d6), coupling constants between F-6 and H-5 of about J = 24  28 Hz. Hyperconjugation effects can explain
the GG conformation, the CH bond at C-5 can donate electron density into the *CF orbital.10, 36 Furthermore, 19F
resonance for the -anomer occurs at lower field than that of the  anomer for all fluorine atoms except for F-4.
Also, F-3 is the most deshielded signal and there are large chemical shift differences between the - and  anomers.
This indicates that there are considerable interactions between 1,3-diequatorial groups with electronegative
substituents (i.e. F-3 and hydroxyl group at C-1 on  anomers). This observation is in agreement with a shielding
of F-3 and a spin-spin coupling between H-1 and F-3 (4JF3-H1 = 3.3 Hz) for the  anomers, as exemplified for the
propose conformation of 16 (Figure 2). Other long-range coupling can be observed, for example F-4 and H-6
4
holds a coupling constant of JF4-H6 = 3.4 Hz (W arrangement). Moreover, there is a decrease in chemical shift as
more fluorine atoms are incorporated on the pyran core, except for F-2 which is fairly constant. This is exemplified
with a downfield shift of F-6 with a chemical shift of 234.6 ppm for 6 to 236.1 ppm for 16. Finally, in acetone-
d6, our set of fluoroglucoses showed no intramolecular OHFC hydrogen bonding as demonstrated for other
polyfluorinated carbohydrates.37
This article is protected by copyright. All rights reserved.

10.1002/chem.202002825
Chemistry - A European Journal
Figure 2. Direct comparison of 19F resonances of fluorinated glucose analogues 316 (19F NMR (470 MHz, acetone-
d6)) and proposed conformation of 3,4,6-trifluoroglucose 16 showing the GG conformation of F-6 and long-range
(4J) couplings in bold.
We then proceeded to perform a systematic investigation of lipophilicity in order to quantify the impact of
deoxyfluorination on a carbohydrate scaffold. Thus, we investigated the influence of the number and the position
of integrated fluorine atom over the log P. In this regard, we used a log P determination method based on 19F NMR
spectroscopy.38 Briefly, this method measures the distribution of organofluorine between n-octanol and water
relative to an internal reference. Then, an aliquot of each phase is analyzed by 19F NMR and the integration ratio to
reference (known log P) in each phase is cross-correlated. This study complements previous work by the group of
Linclau38 and our group21a, 23 on the lipophilicity of fluorinated carbohydrates. As such, the group of Linclau reported
the log P of monofluoroglucose analogues [2-fluoroglucose (log P 2.21) and 6-fluoroglucose (log P 2.36)] and
polyfluoroglucose analogues [2,3-difluoroglucose (log P 1.11) and 2,3,4-trifluoroglucose (log P 0.17)].38 For our
part, we already reported the lipophilicity of all trifluorinated glucose analogues and showed that the inclusion of a
fluorine atom at C-6 can substantially decrease the lipophilicity.23 Figure 3a shows the lipophilicity of all
fluorinated glucose analogues at position C-2, C-3, C-4, and C-6. Among all difluoroglucose analogues, compounds
with a fluorine atom at C-6 are the most hydrophilic except for 2,4-difluoroglucose 8 (log P 1.57). This holds true
This article is protected by copyright. All rights reserved.

10.1002/chem.202002825
Chemistry - A European Journal
also for monofluoroglucose analogues, with 6-fluoroglucose 6 (log P 2.36) being the most hydrophilic compound
of this set. Moreover, 3,6-difluoroglucose 11 (log P 1.60) is more lipophilic than 2,4-difluoroglucose 8 and 4,6-
difluoroglucose 12, which have comparable log P values (1.57 and 1.56 respectively). Also, vicinal
polyfluorinated analogues are the most lipophilic of their respective group. Indeed, 2,3-difluoroglucose 7 (log P
1.11) and 3,4-difluoroglucose 10 (log P 1.29) are the most lipophilic of all difluoroglucose analogues. Similarly,
multivicinal 2,3,4-trifluoroglucose 13 (log P 0.18) is more lipophilic than vicinal 3,4,6-trifluoroglucose 16 (log P
0.42) and 2,3,6-trifluoroglucose 14 (log P 0.47), which in turned are more lipophilic than 2,4,6-trifluoroglucose
15. Vicinal polyfluorinated analogues may have a larger hydrophobic surface area enabling an increase in log P
values. Finally, as for monofluoroglucose analogues 36, the position of the fluorine atom has a strong impact over
the lipophilicity with 4-fluoroglucose 5 being the most lipophilic (log P 2.14).
Recently, the group of Linclau compared experimental log P of aliphatic organofluorines with quantum
chemical log P calculations based on solvent-dependent 3D conformational analysis, and with Clog P based on 2D
structural motif.39 Similarly, we evaluated the correlation between the Clog P values using web-based calculators
with all of the experimental log P of fluorinated glucose analogues (Figure 3b). The ALOGPS40 program gives the
best correlation, but Molinspiration41 and Biobyte42 suffered to predict adequately the log P of fluoroglucose
analogues. Although the root mean square deviation for the ALOGPS program is 0.97, some points are distributed
vertically. This means that the program is giving the same Clog P value to various fluoroglucose analogues and
cannot discriminate on the position of fluorine atoms (i.e. Clog P of 0.91 for compound 7, 8, and 10, which is far
from experimental values, log P of 1.11, 1.57, and 1.29 respectively). This trend was also noticeable for
Molinspiration, calculating the same log P values to all the monofluoroglucoses (log P 0.29) and almost the same
log P values to all difluoroglucoses (log P 1.22).
Solvation free energies can be associated to a broad number of physical properties, such as solubilities or
distribution of chemical species between different phases. They are the free energy change accompanying the
transfer of a molecule between gas and solvent. Similarly, the hydration free energies have an impact on the design
of more drug-like glycomimetics. Carbohydrate hydration has been the subject of much research over the years.43
The aqueous solvation behavior of glucose is well studied44 with the impact of aqueous solvation over
conformational energy mapping45 or over the anomeric equilibrium.46 Glucose solvation stabilizes the β anomer in
water and has little impact upon conformation of the pyran ring. Interestingly, the hydroxymethyl group adopts
either a TG or GT conformation in solution with the absence of the GG conformer. This can be explained by the
gauche effect (a solvent-dependent phenomenon) and stabilizing hydrogen bonding networks.47 In the case of all 6-
fluoroglucose analogues, we observed GG conformations for F-6 and this might impact the solvation behavior of
This article is protected by copyright. All rights reserved.

10.1002/chem.202002825
Chemistry - A European Journal
fluoroglucose analogues. To the best of our knowledge, there is currently no data related to the solvation energies
of fluorinated carbohydrates. Therefore, we first computed the dehydration energies (i.e. from water to gas phase)
of our set of fluoroglucose analogues in order to gain more insight of the impact of the position and the number of
integrated fluorine atoms. For this purpose, we employed the range separated hybrid density functional CAM-
B3LYP-D348 with the double-zeta basis 6-31+G(d,p) and the SMD solvation model on the dominant solution-state
 anomer.49 Calculations were performed with Gaussian 16 revision B01.50 Results are presented in Figure 3a in
brackets for each fluoroglucose analogue. Values were all negative and increased with the number of integrated
fluorine atoms: Ghyd 1F  Ghyd 2F  Ghyd 3F. Moreover, all 6-fluoroglucose analogues gave the highest values
for each of their respective sets of integrated fluorines. This underlines the importance of the conformation about
the C5C6 bond over the desolvation energies. We then proceeded to compute the solvation free energies with n-
octanol as solvent. This allowed us to calculate the free energies of our fluoroglucose analogues to transfer between
water to octanol (see supporting information). It was then possible to perform a correlation of these data with the
experimental log P (Figure 3c). There is good correlation with a root mean square deviation of 0.93. This indicates
that by simply calculating solvation energies with density functional theory corrected by a solvation model, it is
possible to predict the log P of fluoroglucose analogues and this could probably be applied to other fluorinated
carbohydrates.
This article is protected by copyright. All rights reserved.

10.1002/chem.202002825
Chemistry - A European Journal
( Ghyd, kcal mol-1)
Compound
a.
trifluoroglucose
difluoroglucose
monofluoroglucose
logP
-0.9
logP
logP
-1.7
0.1
OH
O
HO
F
-1.0
-1.1
-1.2
-1.3
-1.4
-1.5
-1.6
-1.7
-1.8
-1.8
-1.9
-2.0
-2.1
-2.2
-2.3
-2.4
-2.5
-2.6
0.0
F
OH
OH
O
7
(-20.6)
F
F
-0.1
F
-1.11
F
OH
OH
O
OH
O
13
(-17.8)
F
-0.18
F
F
-0.2
HO
HO
(-20.8)
OH
HO
OH
(-25.9)
10
5
F
O
OH
O
-1.29
-1.42
-0.3
F
O
HO
F
-2.14
F
HO
HO OH
(-17.4)
HO
HO
OH
16
14
F
4
3
(-23.6)
OH
O
OH
-2.19
-2.21
-2.36
-0.4
9
-0.42
(-19.1)
F
O
HO
HO
-0.47
F
F
O
F
-0.5
OH
HO
(-23.1)
HO OH
(-19.6)
OH
HO
F
-1.56
-1.57
-1.60
12
F
F
OH
O
HO
HO
-0.6
-0.7
-0.8
(-16.7)
O
F
HO
HO
OH
(-21.7)
-0.64
F
OH
6
8
(-21.1)
F
O
F
O
F
HO
HO
F
F
OH
15 (-17.4)
HO OH
(-19.6)
logP glucose -3.24
11
b.
c.
Figure 3. a) Lipophilicities of fluorinated glucose analogues 316; b) Correlation between Clog P values obtained
from web-based programs and the experimental log P of fluorinated glucose analogues; c) Correlation between
solvation energies (Gwater-octanol) obtained from CAM-B3LYP-D3/6-31+G(d,p) with SMD solvation and
experimental log P of fluorinated glucose analogues.
This article is protected by copyright. All rights reserved.

10.1002/chem.202002825
Chemistry - A European Journal
Conclusions
In this study, the synthesis of all mono-, di-, and trifluorinated glucopyranose analogues at position C-2, C-
3, C-4, and C-6 has been realised. The lipophilicities of fluoroglucoses were measured, and it was determined that
analogues with a fluorine atom at C-6 are usually the most hydrophilic and compounds with vicinal polyfluorinated
motifs are the most lipophilic. Solvation energies were computed with CAM-B3LYP-D3 corrected by the SMD
solvation model. With these results, we deepen the knowledge of individual fluorinated glucose analogues. By
selecting the appropriate fluorinated carbohydrates, we hope to develop novel personalized fluoroglycoconjugates
with optimal pharmacokinetics properties.
Acknowledgements
This work was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC)
and Université Laval. J. St-G. thanks the Fonds de Recherche du Québec-Nature et Technologies for a postgraduate
fellowship. E. C. and D. L. thank NSERC for an Undergraduate student research award and a Postgraduate
Scholarships respectively. Finally, this research was enabled in part by Calcul Québec and Compute Canada.
Conflict of interest
The authors declare no conflict of interest.
References
1 J. Cramer, C. P. Sager, B. Ernst, J. Med. Chem. 2019, 62, 89158930.
2 a) R. A. Bryce, I. H. Hillier, J. H. Naismith, Biophys. J. 2001, 81, 13731388; b) C. P. Sager, D. Eris, M. Smiesko,
R. Hevey, B. Ernst, Beilstein J. Org. Chem. 2017, 13, 25842595.
3 S. Cabani, P. Gianni, V. Mollicia, L. Lepori, J. Solution Chem. 1981, 10, 563595.
4 a) T. Steiner, Angew. Chem. Int. Ed. 2002, 41, 4876; b) A. Vedani D. W. Huhta, J. Am. Chem. Soc. 1990, 112,
47594767; c) A. Vedani, J. D. Dunitz, J. Am. Chem. Soc. 1985, 107, 76537658.
5 F. P. C. Binder, K. Lemme, R. C. Preston, B. Ernst, Angew. Chem. Int. Ed. 2012, 51, 73277331.
6
R. Kumar, M. M. Ignjatovic, K. Peterson, M. Olsson, H. Leffler, U. Ryde, U. J. Nilsson, D. T. Logan,
ChemMedChem 2019, 14, 15281536.
7 R. Hevey, Biomimetics 2019, 4, 53.
8 R. Hevey, Pharmaceuticals 2019, 12, 55.
This article is protected by copyright. All rights reserved.

10.1002/chem.202002825
Chemistry - A European Journal
9 a) B. Linclau, A. Arda, N.-C. Reichardt, M. Sollogoub, L. Unione, S. P. Vincent, J. Jiménez-Barbero, Chem. Soc.
Rev. 2020, doi:10.1039/c9cs00099b; b) J. C. Biffinger, H. W. Kim, S. G. DiMagno, ChemBioChem 2004, 5,
622627.
10 D. O’Hagan, Chem. Soc. Rev. 2008, 37, 308319.
11 C. Dalvit, C. Invernizzi, A. Vulpetti, Chem. Eur. J. 2014, 20, 1105811068.
12 a) E. P. Gillis, K. J. Eastman, M. D. Hill, D. J. Donnelly, N. A. Meanwell, J. Med. Chem. 2015, 58, 83158359;
b) B. M. Johnson, Y.-Z. Shu, X. Zhuo, N. A. Meanwell, J. Med. Chem. 2020, in press.
13
A. Hoffmann-Roder, M. Johannes, Chem. Commun. 2011, 47, 99039905; b) T. Oberbillig, C. Mersch, S.
Wagner, A. Hoffmann-Roder, Chem. Commun. 2012, 48, 14871489.
14 I. P. Street, C. R. Armstrong, S. Withers, Biochem. 1986, 25, 60216027.
15 a) I. N’Go, S. Golten, A. Arda, J. Canada, J. Jiménez-Barbero, B. Linclau, S. P. Vincent, Chem. Eur. J. 2014, 20,
106112; b) K. E. van Straaten, J. R. A. Kuttiyatveetil, C. M. Sevrain, S. A. Villaume, J. Jiménez-Barbero, B.
Linclau, S. P. Vincent, D. A. R. Sanders, J. Am. Chem. Soc. 2015, 137, 12301244.
16 a) S. G. Withers, I. P. Street, M. D. Percivalin, Fluorinated Carbohydrates: Chemical and Biochemical Aspects,
In ACS Symposium Series, Ed.: N. F. Taylor, No. 374, American ChemicalSociety , Washington DC, 1988, pp.
5977.; b) S. S. Lee, I. R. Greig, D. J. Vocadlo, J. D. McCarter, B. O. Patrick, S. G. Withers, J. Am. Chem. Soc.
2011, 133, 15826–15829; c) M. N. Namchuk, J. D. McCarter, A. Becalski, T. Andrews, S. G. Withers, J. Am. Chem.
Soc. 2000, 122, 1270–1277.
17 a) C. P. J. Glaudemans, Chem. Rev. 1991, 91, 25–33; b) R. U. Lemieux, Chem. Soc. Rev. 1989, 18, 347–374.
18 S. M. Ametamey, M. Honer, P. A. Schubiger, Chem. Rev. 2008, 108, 1501–1516.
19
F. J. Weiberth, H. S. Gill, Y. Jiang, G. E. Lee, P. Lienard, C. Pemberton, M. R. Powers, W. Subotkowski, W.
Tomasik, B. J. Vanasse, Y. Yu, Org. Process Res. Dev. 2010, 14, 623631.
20
a) L. Mtashobya, L. Quiquempoix, B. Linclau, J. Fluorine Chem. 2015, 171, 9296; b) C. Q. Fontenelle, D.
Shishmarev, P. W. Kuchel, B. Linclau, Trends Carbohydr. Res. 2017, 9, 2934; c) R. S. Timofte, B. Linclau, Org.
Lett. 2008, 10, 36733676; d) J. Pacak, J. Podesva, Z. Tocik, M. Cerny, Chem. Commun. 1972, 37, 25892599; e)
B. P. Rempel, S. G. Withers, Aust. J. Chem. 2009, 62, 590599; f) S. Bresciani, T. Lebl, A. M. Z. Slawin, D.
O’Hagan, Chem. Commun. 2010, 46, 54345436; g) M. J. Corr, D. O’Hagan, J. Fluorine Chem. 2013, 155, 7277;
h) L. Quiquempoix, Z. Wang, J. Graton, P. G. Latchem, M. Light, J.-Y. Le Questel, B. Linclau, J. Org. Chem. 2019,
84, 58995906; i) S. G. Withers, M. D. Percival, I. P. Street, Carbohydr. Res. 1989, 187, 4366.
21 a) V. Denavit, D. Lainé, J. St-Gelais, P. A. Johnson, D. Giguère, Nat. Commun. 2018, 9, 4721; b) V. Denavit, D.
Lainé, C. Bouzriba, E. Shanina, É. Gillon, S. Fortin, C. Rademacher, A. Imberty, D. Giguère, Chem. Eur. J. 2019,
25, 44784490; c) D. Lainé, V. Denavit, D. Giguère, J. Org. Chem. 2017, 82, 49864992; d) V. Denavit, D. Lainé,
G. Le Heiget, D. Giguère, Fluorine-containing carbohydrates: Synthesis of 6-deoxy-6-fluoro-1,2:3,4-di-O-
isopropylidene--D-galactopyranose. In, Carbohydrate Chemistry: Proven Synthetic Methods, Eds.: Vogel, C.;
Murphy, P. 2018, Vol. 4, pp 247253.
22 V. Denavit, J. St-Gelais, T. Tremblay, D. Giguère, Chem. Eur. J. 2019, 25, 92729279.
23 J. St-Gelais, M. Bouchard, V. Denavit, D. Giguère, J. Org. Chem. 2019, 84, 85098522.
24
H. Chen, T. Xian, W. Zhang, W. Si, X. Luo, B. Zhang, M. Zhang, Z. Wang, J. Zhang, Carbohydr. Res. 2016,
431, 4246.
25 I. Fokt, S. Szymanski, S. Skora, M. Cybulski, T. Madden, W. Priebe, Carbohydr. Res. 2009, 344, 14641473.
26
a) I. Johansson, B. Lindberg, Carbohydr. Res. 1966, 1, 467473; b) A. B. Foster, R. Hems, J. M. Webber,
Carbohydr. Res., 1967, 5, 292301; c) T. J. Tewson, M. J. Welch, J. Org. Chem. 1978, 43, 10901092.
This article is protected by copyright. All rights reserved.

10.1002/chem.202002825
Chemistry - A European Journal
27 K. Kobayashi, T. Kondo, Macromolecules, 1997, 30, 65316535.
28
S. Hornik, L. C. Stastna, P. Curinova, J. Sykora, K. Kanova, R. Hrstka, I. Cisarova, M. Dracinsky, J. Karban,
Beilstein J. Org. Chem. 2016, 12, 750759.
29 T. B. Grindley, R. Thangarasa, Carbohydr. Res. 1988, 172, 311318.
30 P. J. Card, J. Org. Chem. 1983, 48, 393395.
31 T. Trnka, M. Cerny, Collect. Czech. Chem. Commun. 1971, 36, 22162225.
32 B. T. Grindley, G. J. Reimer, J. Kralovec, Can. J. Chem. 1987, 65, 10651071.
33 See Supporting Information for more details related to assignation of NMR signals.
34
a) L. Phillips, V. Wray, J. Chem. Soc. (B) 1971, 16181624; b) V. Wray, J. Chem. Soc. Perkin II, 1976,
15981605.
35 K. Bock, J. O. Duus, J. Carbohydr. Chem. 1994, 13, 513543.
36 L. Hunter, Beilstein J. Org. Chem. 2010, 6, No. 38.
37 a) G. T. Giuffredi, V. Gouverneur, B. Bernet, Angew. Chem. Int. Ed. 2013, 52, 1052410528; b) G. T. Giuffredi,
B. Bernet, V. Gouverneur, Eur. J. Org. Chem. 2011, 38253836; c) G. T. Giuffredi, L. E. Jennings, B. Bernet, V.
Gouverneur, J. Fluorine Chem. 2011, 132, 772778.
38
B. Linclau, Z. Wang, G. Compain, V. Paumelle, C. Q. Fontenelle, N. Wells, A. Weymouth-Wilson, Angew.
Chem., Int. Ed. 2016, 55, 674678.
39 B. Jeffries, Z. Wang, H. R. Felstead, J.-Y. Le Questel, J. S. Scott, E. Chiarparin, J. Graton, B. Linclau, J. Med.
Chem. 2020, 63, 10021031.
40 ALOGPS 2.1 program : http://www.vcclab.org/web/alogps/
41 Molinspiration: https://www.molinspiration.com/cgi-bin/properties
42
Daylight/Biobyte ClogP Pomona College and BioByte, Inc., Claremont, CA (bio-loom version 1.6, program
version 5): http://www.biobyte.com/
43 a) S. A. Galema, E. H, J. B. F. N. Engberts, J. R. Grigera, Carbohydr. Res. 1994, 265, 215225; b) S. A. Galema,
J. B. F. N. Engerts, H. Hoeiland, G. M. Foerland, J. Phys. Chem. 1993, 97, 68856889; c) S. A. Galema, H.
Hoeiland, J. Phys. Chem. 1991, 95, 53215326; d) S. A. Galema, M. J. Blandamer, J. B. F. N. Engerts, J. Am.
Chem. Soc. 1990, 112, 96659666; e) S. A. Galema, M. J. Blandamer, J. B. F. N. Engberts, J. Org. Chem. 1992,
57, 19952001.
44 J. W. Brady, J. Am. Chem. Soc. 1989, 111, 51555165.
45 G. Lelong, M.-L. Saboungi, J. W. Brady, Mol. Simulat. 2012, 38, 11861197.
46 S. Ha, J. Gao, B. Tidor, J. W. Brady, M. Karplus, J. Am. Chem. Soc. 1991, 113, 15531557.
47 J. L. Alonso, M. A. Lozoya, I. Pena, J. C. Lopez, C. Cabezas, S. Mata, S. Blanco, Chem. Sci. 2014, 5, 515522.
48
a) T. Yanai, D. Tew, N. Handy, Chem. Phys. Lett. 2004, 393, 5157; b) D. Becke, J. Chem. Phys. 1993, 98,
56485652; c) C. Lee, W. Yang, R. G. Parr, Phys. Rev. B, 1988, 37, 78589; d) S. Grimme, J. Antony, S. Ehrlich,
H. Krieg, J. Chem. Phys., 2010, 132, 154104.
49 A. V. Marenich, C. J. Cramer, D. G. Truhlar, J. Phys. Chem. B, 2009, 113, 637896.
50 Frisch, M. J., et al. Gaussian, Inc., Wallingford CT (2016).
This article is protected by copyright. All rights reserved.