Metabolism study and biological evaluation of bosentan derivatives

Article abstract of DOI:10.1016/j.ejmech.2016.06.006

Bosentan, the first-in-class drug used in treatment of pulmonary arterial hypertension, is principally metabolized by the cytochromes P450, and it is responsible for cytochromes induction and drug-drug interaction events with moderate to severe consequences. A strategy to reduce drug-drug interactions consists of increasing the metabolic stability of the perpetrator, and fluorinated analogues are often designed to block the major sites of metabolism. In this paper bosentan analogues were synthesized, and their metabolism and biological activity were evaluated. All synthesized compounds showed an improved metabolic stability towards CYP2C9, with one maintaining a moderate antagonist effect towards the ET_A receptor.

Full text of DOI:10.1016/j.ejmech.2016.06.006

Accepted Manuscript
Metabolism study and biological evaluation of bosentan derivatives
Susan Lepri, Laura Goracci, Aurora Valeri, Gabriele Cruciani
PII:
S0223-5234(16)30483-4
10.1016/j.ejmech.2016.06.006
EJMECH 8667
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 2 February 2016
Revised Date: 31 May 2016
Accepted Date: 4 June 2016
Please cite this article as: S. Lepri, L. Goracci, A. Valeri, G. Cruciani, Metabolism study and biological
evaluation of bosentan derivatives, European Journal of Medicinal Chemistry (2016), doi: 10.1016/
j.ejmech.2016.06.006.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT
Metabolism Study and Biological Evaluation of Bosentan Derivatives
Susan Lepri, ‡ Laura Goracci, ‡ Aurora Valeri and Gabriele Cruciani.*
Department of Chemistry, Biology and Biotechnology, University of Perugia, Via Elce di Sotto 8,
06123 Perugia, Italy.
*Corresponding author. Gabriele Cruciani: phone, +39 075 585 5629; Fax, +39 075 45646; E-mail,
gabriele.cruciani@unipg.it.
‡ These authors contributed equally to this work.
Abstract
Bosentan, the first-in-class drug used in treatment of pulmonary arterial hypertension, is principally
metabolized by the cytochromes P450, and it is responsible for cytochromes induction and drug-
drug interaction events with moderate to severe consequences. A strategy to reduce drug-drug
interactions consists of increasing the metabolic stability of the perpetrator, and fluorinated
analogues are often designed to block the major sites of metabolism. In this paper bosentan
analogues were synthesized, and their metabolism and biological activity were evaluated. All
synthesized compounds showed an improved metabolic stability towards CYP2C9, with one
maintaining a moderate antagonist effect towards the ETA receptor.
Keywords: bosentan, perfluorinated analogues, metabolic specificity, phase I metabolism, HLM.
1

ACCEPTED MANUSCRIPT
List of abbreviations
Cl Clearance, CYP Cytochrome P450, DDI drug-drug interactions, ERA Endothelin Receptor
Antagonist, ET-1 Endothelin 1, ETA Endothelin receptor A, ETB Endothelin receptor B, PAH
pulmonary arterial hypertension.
1. Introduction
The molecular mechanism of pulmonary arterial hypertension (PAH) is still uncertain, but it is
likely that the excessive vasoconstriction of the pulmonary arterioles related to endothelial
dysfunctions results in an imbalance of various endothelial vasoactive mediators (such as
endothelin-1, or ET-1), affecting the growth of the smooth muscle cells, and may facilitate the
structural remodeling characteristic of PAH [1]. Since this condition arises from a complex
interplay of molecular and genetic abnormalities, in clinical practice PAH can be controlled by co-
administration of multiple drugs [2]. Combination of several treatments on the one hand might
increase their clinical effectiveness but, on the other hand, it is also likely to increase the risk of
adverse drug-drug interactions (DDIs) with a reduced treatment efficacy or increased side effects
[3]. As mentioned above, with high levels of ET-1 detectable in lung tissues of PAH patients, the
administration of endothelin receptor antagonists (ERAs) helps to reduce symptoms of disease [4].
The first-in-class ERA used in PAH treatment is bosentan (1), a sulfonamide-based drug, antagonist
for both endothelin receptor A (ETA) and B (ETB) [5, 6]. Bosentan is a well-known strong
perpetrator in pharmacokinetic drug-drug interaction, as it is an inducer of drug transporters and
metabolizing enzymes, and an inhibitor of P-glycoprotein and organic anionic transporting
polypeptides [7-9]. Moreover, a possible perpetrator effect of bosentan metabolites has been
recently studied [7].
2

ACCEPTED MANUSCRIPT
Thus, to decrease the risk of adverse events associated with DDI during bosentan administration, a
thorough knowledge of the pharmacokinetic profile of this drug is mandatory. Unfortunately,
despite its large and valuable therapeutic use, the bosentan metabolism profile is not optimal.
Bosentan is commonly reported to be metabolized in the liver by the cytochrome P450 enzymes
CYP2C9 and CYP3A4 producing three metabolites (Figure 1): the aliphatic hydroxylation at the
tert-butyl group produces Ro 48-5033 (2), a pharmacologically active metabolite that may
contribute to 10 - 20 % of the total activity of the parent compound; the oxidative demethylation of
guaiacol moiety by CYP3A4 forms phenol derivative Ro 47-8634 (3), whereas metabolite Ro 64-
1056 (4) is a minor product formed from both primary metabolites [10]. Recently, a new metabolite
[11] was identified in human liver microsomes (HLM) as shown in Figure 1. In bosentan
metabolism the involvement of CYP2C9 may represent an issue when the metabolic pathway of a
drug is investigated. Indeed, CYP2C9 is a polymorphically expressed enzyme showing high human
variability in the frequency of variant CYP2C9 allele (about 60 allelic variants have been reported)
[12]. Recently, several studies have been conducted on allelic variants of CYP2C9 [13-16],
demonstrating a great interethnic and intra-ethnic variability; most recently, due to this CYP2C9
polymorphism, Chen et al. [17] have highlighted the importance of directing more attention when
administering bosentan to a subject carrying some infrequent CYP2C9 alleles. The genetic
polymorphism of CYP2C9 has been also associated to bosentan-induced liver injury [18] although
this point is still under debate [11].
In addition, clinical trials have also demonstrated that bosentan is an inducer of isoenzymes
CYP2C9 and CYP3A4, making DDI studies even more complex when bosentan is co-administrated
with drugs that are metabolized by these CYP isoforms (e.g. warfarin [19], cyclosporine [20], oral
estrogens [21], simvastatin [22], and sildenafil [23]). Although the mechanism of isoenzyme
induction by bosentan is still unclear, the optimization of the metabolic stability of bosentan
towards CYP3A4 and mainly towards CYP2C9 could represent a valid strategy to reduce DDI. In
3

ACCEPTED MANUSCRIPT
this study four bosentan derivatives were designed and synthesized to evaluate the impact of slight
changes on the bosentan scaffold in phase I metabolism in human liver microsomes (HLM), and in
CYP3A4 and CYP2C9 isoforms. Metabolism by CYP2D6 was also investigated to explore a
possible role of this enzyme in the metabolism of the synthesized compounds. For the most
interesting compounds the possible agonist and antagonist activities against endothelin receptors
was studied.
4

ACCEPTED MANUSCRIPT
Figure 1. Proposed metabolic pathway of bosentan in human. The site of metabolism is highlighted
by blue arrows [10]. The major metabolic pathway, indicated by a thicker arrow, leads to the
5

ACCEPTED MANUSCRIPT
formation of metabolite 2. The formation of a recently identified metabolite is indicated by a dashed
arrow [11].
2. Results and Discussion
2.1 Design
An enhancement of the metabolic stability can be achieved by modifying the chemical structure of
the parent compound. For example, a common strategy requires blocking the site of detrimental
metabolism with the introduction of fluorinated groups. The benefits of fluorine addition are
numerous: the size of fluorine is similar to that of hydrogen (fluorine radius is just 20% larger and
H3C
not distinguish between a natural substrate and its fluorinated analogue [25]. Moreover, since H3C
F bond is stronger than that of H3C H by 5.0 Kcal/mol [24], the metabolic oxidation by CYP450 is
̶ F bond is 27% longer than H3C ̶ H bond [24]) and indeed it has been shown enzymes often do
̶
̶
reduced. This blocking effect has been widely used to prevent deactivation of active molecules due
to their metabolic instability. For instance, fluorinated analogues of buspirone [26] and vitamin D3
[27] have shown improved metabolism respect to their hydrogenated compound; moreover,
incorporation of fluorine is reported to enhance the drug potency (e.g. the ezetimibe, fluorinated
antitumor benzothiazoles 5F 203 and GW 610, fluroquinolones, etc. [28]). In this study, since the
tert-butyl group of bosentan is the main site of metabolism, this group was replaced with fluorinated
substituents. Despite their different volumes (see Table S1), a large number of fluorinated isosteres
of the tert-butyl group have been used in drug discovery, ranging from trifluoromethyl (CF3) [29],
pentafluorosulfanyl (SF5) [30, 31], bicyclo[1.1.1]pentanyl (BCP) [32, 33], and cyclopropyl-
trifluoromethyl (cyclopropyl-CF3) [34]. Thus, rigid rules for the replacement of the tert-butyl group
in metabolic stability studies cannot be applied. In the present study, we investigated the effect of
hexafluoro-2-hydroxyprop-2-yl, hexafluoro-2-methoxyprop-2-yl, and heptafluoro isopropyl, to give
6

ACCEPTED MANUSCRIPT
compounds 5a-c, respectively (Figure 2, block I). In particular, in compound 5a a further acidic
function adjacent to the two trifluoromethyl functionality, with a pKa value of 8.60, was introduced.
Block I
F3C
5a
HO
CF3
F3C
5b
H3CO
CF3
O2
S
CYP 3A4
CYP 2C9
F3C
F
NH
O
5c
Block I
CYP 3A4
O
O
CF3
N
N
OH
Block II
N
Block II
N
H
5d
O
1
Figure 2. Proposed modification of bosentan moieties to evaluate the effect on metabolism and the
biological response.
The fourth analogue (5d, Figure 2 block II) was designed to test if the metabolism of bosentan by
CYP2C9 could be reduced by a slight change in scaffold decoration, without modifying the
structure at the site of metabolism (i.e. the tert-butyl group, as shown in Figure 1).
The MetaSiteTM software [35, 36] was used to obtain information about which interactions play a
role in the exposition of bosentan towards the heme in the CYP2C9 cavity. Indeed, MetaSiteTM
allows selection of the desired site of metabolism (SoM) of bosentan and to orient the molecule in
the CYP2C9 cavity, computing the bosentan-CYP hydrophobic complementarity and the
complementarity of charges and H-bonds between the ligand and the enzyme. Such
complementarities are then used to assign a contribution score (for the selected SoM) to the
different atoms in the substrate. This innovative strategy to design derivatives with tailored
7

ACCEPTED MANUSCRIPT
metabolism towards a specific enzyme recently proved to be effective to modulate the metabolism
of bepridil by FMO3 [37].
In the case of bosentan, the pose of this compound within the CYP2C9 cavity was modeled by
fixing the tert-butyl group orientation towards the heme. The contribution scores reported in Figure
3 highlight that the molecular groups influencing the bosentan aliphatic hydroxylation at the tert-
butyl group are the nitrogen of the sulfonamide function (interacting with Arg 108) and the
pyrimidinoxyethanol group, which interacts with the carbonyl group of Val 292. Based on these
considerations, the 2-hydroxyethoxy moiety at the 6 position of the substituted pyrimidine ring was
replaced by an ethoxy group, attempting to disfavor the exposition of the tert-butyl group towards
the heme, and to consequently increase metabolic stability.
Figure 3. Contribution scores (circles) generated by MetaSiteTM when the tert-butyl group of
bosentan is exposed to the heme in CYP2C9 cavity. The two CYP2C9 amino acids Val 292 and Arg
108 are predicted to be responsible for the favourable exposition of the tert-butyl group towards the
heme, interacting with the 2-hydroxyethoxy moiety and the sulfonamide moiety, respectively.
8

ACCEPTED MANUSCRIPT
2.2 Synthesis
Schemes 1−3 illustrate the synthesis of compounds described herein. To obtain the common
intermediate dichloropyrimidine 9 we followed the known literature protocols [38, 39]. Briefly,
condensation of aryloxymalonate 6 with the pyrimidylamidine 7 afforded pyrimidinepyrimidone 8,
which was converted into the chlorinated intermediate 9 upon treatment with PCl5 in refluxing
acetonitrile. In particular, the aryloxymalonate 6 and pyrimidylamidine 7 were prepared by O-
alkylation of guaiacol and amination of 2-cyanopyrimidine, respectively (Scheme 1).
Scheme 1a
a Reagents and conditions: (a) Dimethyl 2-chloromalonate, MeONa, MeOH, 45 °C, 1 h [38]; (b) i.
MeONa, MeOH, rt, 24 h; ii. AcOH, NH4Cl, rt, 24 h [39]; (c) MeONa, MeOH, rt, 10 h [38]; (d) PCl5,
N,N-diisopropylethylamine, CH3CN, reflux, 20 h [38].
As shown in Scheme 2, the nucleophilic attack of N-lithiated sulfonamides 10a-b to
dichloropyrimidine 9 afforded the monochlorosulfonamides 11a-d (respectively), afterwards
converted into the target bipyrimidines 5a-d by nucleophilic substitution of the second chlorine
atom by sodium glycolate (for 5a-c) or sodium ethoxide (for 5d).
9

ACCEPTED MANUSCRIPT
Scheme 2a
aReagents and conditions: (a) i. n-BuLi, THF,
̶ 78 °C, 15 min ; ii. DMF, 50 °C, 16 h; (b) Na,
ethylene glycol 110 °C (for 5a-c) or EtOH 75 °C (for 5d), 5 h [38].
The synthesis of sulfonamides 10a-c is worth describing briefly (Scheme 3), whether sulfonamide
10d is commercial available.
The diazotization of commercial 2-(4-aminophenyl)-1,1,1,3,3,3-hexafluoropropan-2-ol with sodium
nitrite in a hydrogen chloride-acetic acid mixture, followed by reaction with CuCl-SO2 complex in
acetic acid solution, gave the arylsulfonyl chloride intermediate 12a which was afterwards
transformed in the desired sulfonamide 10a by simple addition of aqueous ammonium hydroxide to
the reaction mixture (Scheme 3). The sulfonamide 10c was prepared analogously from 4-
(perfluoroisopropyl)aniline 15 which, in turn, was obtained by alkylation of aniline with 2-
iodoperfluoropropane [40]. On the contrary, a different strategy was adopted to obtain the
methoxylated sulfonamide 10b. Butyllithium-promoted bromine/lithium exchange followed by the
addition of gaseous hexafluoroacetone, transformed 1,4-dibromobenzene into fluorinated
bromoarene 13. The latter was methylated by methyl iodide upon treatment with sodium hydride in
DMF to give the methoxylated bromoarene 14. Bromoarene 14 was smoothly transformed into the
desired sulfonamide 10b in 4 steps in 20–25% overall yield. In particular, lithiation of bromo
compound 14 by butyllithium, followed by addition of gaseous sulfur at –60 °C, gave the lithium
10

ACCEPTED MANUSCRIPT
sulfinate intermediate [41]; the latter was converted into the desired sulfonamide 10b upon
treatment with N-chlorosuccinimide and, eventually, with aqueous ammonia in acetone at 50 °C
[41].
Scheme 3a
OH
CF3
OH
CF3
OH
F3C CF3
F3C
F3C
a
b
NH2
SO2Cl
12a
SO2NH2
10a
OH
F3C CF3
O
O
Br
F3C
CF3
F3C
CF3
c
d
e
Br
Br
SO2NH2
Br
13
14
10b
F
F
F
F3C
CF3
F3C
CF3
F3C
CF3
f
a
b
NH2
NH2
SO2Cl
12c
SO2NH2
15
10c
a Reagents and conditions: (a) i. NaNO2, HCl, AcOH,
̶ 10 °C, 45 min; ii. SO2, CuCl, 0 °C, 10 min
[42]; (b) NH4OH aq., THF, 0 °C, 15 min; (c) i. n-BuLi, THF, -95 °C, 15 min; ii. Hexafluoroacetone
sesquidrate H2O, -95 °C, 10 min; (d) i. NaH, DMF, 0 °C, 40 min; ii. CH3I, 40 min, rt; (e) i. n-BuLi,
THF, -78 °C, 15 min; ii. SO2, -60 °C, 5 min; iii. NCS, CH2Cl2, 0 °C for 15 min, then 25 min rt; iv.
NH4OH aq, acetone, 40 °C, 12 h [41]; f) 2-iodoperfluoropropane, Na2(S2O4), NaHCO3,
N(Bu)4HSO4, MTBE, water, rt, 8 h [40].
2.3 Metabolism profile
Metabolism of bosentan (1) and its analogs 5a-d was evaluated upon incubation with HLM and
recombinant isoforms (3A4, 2D6, 2C9) at three incubation times (0, 15, 30 minutes; see the
11

ACCEPTED MANUSCRIPT
Experimental Section). Although bosentan is a not substrate for CYP2D6, this isoform was also
tested to evaluate if this enzyme could be involved in the metabolism of the synthesized derivatives.
Samples were analyzed by LC-MS/MS and the raw data were processed with the Mass-MetaSiteTM
software (version 3.0.2, Molecular Discovery Ltd., UK [43-45]) to assist in the metabolism
elucidation process. Table 1 shows the percentage of unchanged substrate in all matrices and the
nature of metabolites found, while a more detailed characterization of the detected metabolites
based on the LC-MS/MS parameters is provided in Table 2. In addition, the main MS/MS fragment
assignments used for structure elucidation are reported in Supplementary Data (Table S1). MS/MS
spectra of 1 and its metabolites were congruent with literature data [46]. Bosentan (1) and
compounds 5a and 5c were minimally metabolized by HLM. For compound 5a, the protection of
the tert-butyl group induces the formation of the O-demethylated metabolite only, although this
does not result in an improved metabolic stability. Interestingly, for compound 5c, the most stable
compound, the formation of the corresponding 5c-OH (aromatic)-O-demethyl(guaiacol) metabolite
(Table 1) was observed, with an accurate mass of 666.0888 (M+H+). The formation of this
metabolite was confirmed by the presence of MS/MS fragments 622.0637, 313.0806, and 189.0651
(Tables 2 and S1).
On the contrary, compounds 5b and 5d were less stable than bosentan in HLM (with 73.9 and
68.5% of parent compound remaining after 30 minutes of incubation, respectively). It is noteworthy
that for compound 5b, which differs from the analogue 5a by the additional methylation of
hexafluoro-2-methoxyprop-2-yl moiety, a second O-demethylation process was also observed to
give 5b-O-demethyl(arylsulfonamide). In particular, 5b-O-demethyl metabolites exhibiting a
retention time (RT) of 4.72 min and 5.14 min have been attributed to hexafluoro-2-methoxyprop-2-
yl arylsulfomamide and guaiacol O-demethylation, respectively (see Table S1 for main fragment
assignment). Compound 5d, it resulted the least stable in HLM, possessing a similar metabolism of
12

ACCEPTED MANUSCRIPT
bosentan in this matrix, although the corresponding 5d-OH(t-Bu)-O-demethyl(guaiacol) was not
observed (Table 1).
Metabolism studies in specific CYP isoforms revealed that none of the investigated compounds
were substrates of CYP2D6, with the level of unchanged substrates being greater than 99%.
Concerning bosentan (1), the obtained results for CYP3A4 and CYP2C9 are congruent with
literature data [10]. In particular, three metabolites have been detected with CYP3A4: 1-OH(t-Bu)
(Table 1), referring to the aliphatic oxidation on tert-butyl moiety as for incubation in CYP2C9; 1-
O-demethyl(guaiacol), generated by the O-demethylation reaction, which matches the literature
metabolite 3; 1-OH(t-Bu)-O-demethyl(guaiacol), which can be associated to 4, being generated by
the combination of O-demethylation and aliphatic oxidation. Only one metabolite, 1-OH (t-Bu),
could be detected with CYP2C9, with a m/z of 568.1861 (M+H+), attributable to an aliphatic
oxidation on tert-butyl moiety, according to the literature metabolite 2. As previously mentioned
bosentan (1) was also confirmed as a non-substrate for CYP2D6.
Concerning the fluorinated analogues 5a-c, all of them were found to be substrates for CYP3A4,
with the nature of the formed metabolites being identical to that in HLM. Indeed, for all the
fluorinated analogues the O-dealkylation reaction occurred at the methoxy guaiacol group to give
5a-c-O-demethyl(guaiacol). In addition, for compound 5b, which differs from the analogue 5a by
the additional methylation of the hexafluoro-2-hydroxyprop-2-yl moiety, a second O-demethylation
process was also observed to give 5b-O-demethyl(arylsulfonamide), as for the HLM data. This
suggests that, despite the introduction of fluorinated moieties to replace the tert-butyl group of
bosentan, that region of the molecule is still favorable to be exposed to the heme in CYP3A4, and
thus our modifications to give 5a-c only affect the reactivity factor. Also in CYP3A4, only
compound 5c underwent the aromatic oxidation in combination with O-demethylation to produce
the corresponding 5c-OH(aromatic)-O-demethyl (guaiacol) metabolite.
13

ACCEPTED MANUSCRIPT
In our study, the stabilization of the synthesized compounds towards CYP2C9 metabolism was the
primarily goal. A significant impact of the fluorination strategy was observed analyzing the
metabolism by CYP2C9. Indeed, since bosentan is metabolized by this isoform only at the tert-
butyl group through an aliphatic hydroxylation, this reaction was impeded for compounds 5a-c and
these compounds were totally stable after 30 min incubation in the presence of the CYP2C9
isoform. Differently from what was observed in CYP3A4, the O-demethylation of the hexafluoro-2-
methoxypropan-2-yl moiety was not observed in CYP2C9, suggesting that for this enzyme this
reaction is less favorable.
Concerning compound 5d, which was designed to reduce metabolism by CYP2C9 without altering
the tert-butyl moiety, it resulted to be practically stable after incubation with CYP2C9 (with 99% of
substrate remaining and only traces of the hydroxylated metabolite formed), proving that the
replacement of the 2-hydroxyethoxyl moiety with an ethoxy group disfavors the exposition of the
tert-butyl group towards the CYP2C9 heme, as suggested by MetaSiteTM (Figure 3).
For completeness, the binding modes for the synthesized analogues 5a-d in CYP3A4 and CYP2C9
are available as Supplementary Material (see Figure S1). In particular, the in silico predictions in
CYP3A4 are in agreement with the experimental results showing that the guaiacol group is the most
favourable SoM. The similar analysis for CYP2C9 cannot be applied because MetaSiteTM assumes
that the docked compounds are substrates, while compounds 5a-c are metabolically stable.
Table 1. Bosentan and its analogue metabolites analysis in HLM, CYP3A4, and CYP2C9.
Comp.
Structure
HLM
CYP3A4
CYP2C9
% sub.
after
30’
% sub.
after
30’
% sub.
after
30’
Metabolites
analysis
Metabolites analysis
Metabolites analysis
1-OH (t-Bu)
1-O-demethyl
(guaiacol)
1-OH (t-Bu)
1-O-demethyl
(guaiacol)
1-OH (t-
Bu)
1
91.6
20.9
75.8
1-OH (t-Bu)-O-
demethyl(guaiacol)
1-OH (t-Bu)-O-
demethyl(guaiacol)
14

ACCEPTED MANUSCRIPT
5a-O-demethyl
(guaiacol)
5a-O-demethyl
(guaiacol)
5a
5b
90.1
73.9
72.8
19.2
100
100
ND
ND
5b-O-demethyl
(arylsulfonamide)
5b-O-demethyl
(guaiacol)
5b-O-demethyl
(arylsulfonamide)
5b-O-demethyl
(guaiacol)
5c-O-demethyl
(guaiacol)
5c-OH (aromatic)-
O-demethyl
5c-O-demethyl
(guaiacol)
5c-OH (aromatic)- O-
demethyl (guaiacol)
5c
99.0
68.5
16.3
15.0
100
ND
(guaiacol)
5d-OH (t-Bu)
5d-O-demethyl
(guaiacol)
5d-OH (t-Bu)
5d-O-demethyl
(guaiacol)
5d-OH (t-
5d
99.0
Bu)
Table 2. Compound 1, and 5a-d and their main metabolites identified by LC-MS/MS and their
revealed characteristic fragments.
Exact
Elemental
composition
Accurate
Characteristic
fragments
Comp.
Structure assigned
∆ppm
RT (min)
[M+H]+
[M+H]+
1
508.1649, 311.1013
280.0829, 202.0723
_
552.1911
568.1860
C27H29N5O6S
C27H29N5O7S
552.1912
568.1861
0.06
5.10
(bosentan)
1-OH (t-
Bu)
524.1591, 311.1022,
280.0818, 202.0722
-0.03
0.37
3.79
4.73
1-O-
demethyl
(guaiacol)
538.1755
C26H27N5O6S
538.1753
494.1493, 189.0643
510.1442, 189.0643
1-OH (t-
Bu)-O-
554.1704
662.1139
648.0983
676.1295
662.1139
C26H27N5O7S
C26H21F6N5O7S
C25H19F6N5O7S
C27H23F6N5O7S
C26H21F6N5O7S
554.1703
662.1140
648.0971
676.1296
662.1141
0.12
0.16
0.77
-0.19
-0.31
4.16
4.72
4.39
5.48
4.72
demethyl
618.0875, 311.1015,
280.0832, 202.0727
5a
_
5a-O-
demethyl
(guaiacol)
604.0728, 297.0847,
280.0588, 189.0645
632.1033, 311.1013,
280.0831, 202.0727
5b
_
5b-O-
demethyl
(sulfonamid
e)
618.0810, 311.1009,
280.0826, 202.0727
5b-O-
demethyl
(guaiacol)
618.0876, 297.0860,
280.0640, 189. 0643
662.1139
C26H21F6N5O7S
662.1136
0.35
5.14
15

ACCEPTED MANUSCRIPT
620.0833, 311.1008,
280.0822, 202.0725,
5c
_
664.1095
650.0939
C26H20F7N5O6S
C25H18F7N5O6S
664.1096
650.0939
-0.13
0.01
5.60
5.27
5c-O-
demethyl
(guaiacol)
606.0671, 297.0858,
280.0617, 189.0642,
HO
O2
S
5c-OH
NH
OH
(aromatic)-
O-demethyl
(guaiacol)
622.0637, 313.0806,
189.0651
666.0888
C25H19F7N5O7S
666.0888
-0.02
4.74
F3
C
O
O
N
F
CF3
N
OH
N
N
508.1645, 308.1142,
280.0836, 202.0725
5d
_
536.1962
552.1911
522.1803
C27H29N5O5S
C27H29N5O6S
C26H27N5O5S
536.1963
552.1912
522.1803
-0.07
-0.05
-0.05
5.96
4.99
5.52
5d-OH (t-
524.1598, 308.1142,
280.0831, 202.0727
Bu)
O2
S
NH
OH
5d-O-
demethyl
(guaiacol)
297.0858, 280.0617,
189.0642,
O
O
HO
N
N
N
N
2.4 Biological assays
Compounds metabolized by CYP2C9 and CYP3A4 (5a-c) were further investigated for their
cytotoxicity and for the agonist and antagonist response towards ETA and ETB. Compound 5d,
although very interesting for demonstrating that the tailored metabolic stabilization towards a
specific isoform is possible, was not further investigated because it exhibited the worst overall
metabolic profile in HLM (Table 1).
2.4.1 Cell viability
Cells viability of human recombinant ovarian (CHO-k1) cells incubated with compounds 1, and
5a-c was assessed by MTT assay following 24 h exposure to the compounds. The dose–
response curves of each compound in CHO-k1 cells are shown in Figure 4. The IC50 values,
graphically determined from each concentration–response curves, were higher than 150 µM.
Thus, the tested compounds start to be slightly cytotoxic at micromolar concentration, which
exceeds the nanomolar activity range on ET receptors.
16

ACCEPTED MANUSCRIPT
1
5a
5b
5c
Comp.
195.9 ± 1.14
> 200
152.6 ± 1.05
159.4 ± 1.08
IC5
0
Figure 4. Cytotoxicity of compounds 1, 5a-c on CHO-k1 cells and IC50 determination by MTT test.
Results are expressed as percentages of OD versus vehicle control cells. IC50 values were calculated
on the best fitting of dose–response inhibition type curves using GraphPad Prism 5 software.
2.4.2 Biological activity
Since one of the most important abnormalities of PAH is an overexpression of ET-1 that acts upon
the receptors ETA and ETB, which mediate its physiological and pathological effects, the possible
agonist and antagonist activities of bosentan analogues towards endothelin receptors were
determined by Cerep, France [47]. The in vitro cellular function assays were performed for
compounds 5a-c by fluorimetric method using two cells lines: human neuroblastoma cell-line
(SKNMC) for ETA receptor and CHO-k1 cells for ETB receptor. Agonist and antagonist activities
were evaluated by comparing the response of different test compound solutions with a reference
compounds. In particular, endothelin-1 was the reference compound used for agonist assays in both
receptors (IC50 = 0.67 nM and 0.1 nM in ETA and ETB, respectively). The reference compound used
17

ACCEPTED MANUSCRIPT
for antagonist effect in ETA was BQ-123 (IC50 = 0.62 nM, KB = 0.082 nM) [48], while BQ-788 (IC50
= 0.41 nM, KB = 0.50 nM) was used in ETB [49]. In the agonist assay the ability of test compound
in stimulating the receptor was evaluated in term of % of control agonist response; in turn,
antagonist effect was evaluated in terms of % inhibition of control agonist response and IC50 for
active compounds was calculated.
Figure 4. Evaluation of % inhibition of ETA and ETB for compounds 5a-c.
None of the bosentan analogues resulted in agonist activities towards endothelin receptors
(supplementary data, Table S2). As shown in Figure 4 compound 5a was completely inactive
towards ETA, and really low for ETB (% of inhibition of control agonist response ~ 13) suggesting
that the acidity of hexafluoro-2-hydroxy-2-propyl group heavily influences compound binding
affinity towards ETA. Compound 5b, which differs from 5a for the replacement of acidic hydroxyl
group with a methoxy moiety, has a low activity, slightly better for ETA (IC50 = 693 ꢀM) than ETB.
Compound 5c, where the 4-position of benzensulfonamide is occupied by a heptafluoroisopropyl
group, was found to be the most active in this series with an interesting moderate effect selective for
ETA with an IC50 of 75.4 nM. Interestingly, although compound 5c has a lower potency with respect
to bosentan, it was found to be more selective towards the ETA receptor (for comparison, bosentan
binds to human ETA and ETB receptor with IC50 values of 4.7 and 95 nM, respectively [50]).
18

ACCEPTED MANUSCRIPT
3. Conclusions
At present the use of bosentan as part of a combination therapy in PAH treatment has resulted in
improvements of life quality and survival. However, due to bosentan ability to induce CYP3A4 and
CYP2C9, belonging to CYP450 family, the risk associated for DDI is elevated, and hepatic events
are known to occur and patients require regular monitoring with liver function tests [51]. In
addition, the interaction of bosentan with the poly-morphically expressed enzyme CYP2C9 is cause
of concern, due to the possible inter-ethnic and intra-ethnic variability of the drug response. To
investigate possible strategies to enhance the metabolic stability of bosentan, especially towards
CYP2C9, four synthetic analogues of bosentan have been designed and synthesized. These
compounds were tested for their metabolism, agonistic and antagonistic activities.
The replacement of the tert-butyl group with perfluorinated bioisosteres (compounds 5a-c)
impeded, as expected, the tert-butyl aliphatic hydroxylation reaction, which is one of the major
metabolism pathways of bosentan by CYP3A4 and CYP2C9. Interestingly, in HLM or in CYP3A4
this did not lead to a systematic increase of the metabolic stability, and new metabolic reactions
occurred for 5b and 5c. In particular, 5b was metabolized to a greater extent than bosentan in HLM,
and a novel O-demethylation reaction at the sulfonamide moieties was observed. The occurrence of
the O-demethylation reaction at the sulfonamide group of 5b was also observed after incubation
with CYP3A4, indicating that the methylated hexafluoro-2-hydroxyprop-2-yl can still be oriented
towards the heme of CYP3A4. Despite the variable behaviour observed for fluorinated analogues in
HLM and 3A4, these compounds resulted to be totally stable after 30 minutes incubation not only in
CYP2D6 (like bosentan) but also in CYP2C9. Although the metabolic stability towards CYP2C9
was our primarily goal, the antagonist activity of compounds 5a-c on ETA and ETB was also tested,
with compound 5c exhibiting a moderate antagonist activity on ETA. To date there is no evidence
that the selective inhibition of ETA has a clinical advantage over the nonselective one but both
19

ACCEPTED MANUSCRIPT
sitaxsentan and ambrisentan, two selective ETA antagonists already marketed, have shown lower
development of abnormal liver function [52].
Finally, an innovative in silico driven strategy to improve metabolic stability using MetaSiteTM was
applied to design and synthesize an additional analogue of bosentan aiming at specifically reducing
metabolism by CYP2C9 without modifying the SoM. The resulting compound 5d, differing from
bosentan by a change in the lateral chain only, proved to be much more stable than bosentan when
incubated in the presence of CYP2C9 for 30 minutes, with 99% of unchanged substrate remaining.
Thus, although compound 5d was the least stable compound in HLM, this study has demonstrated
that the MetaSiteTM predictions can be used as an efficient approach for tailoring the phase I
metabolic pathway of a drug.
4. Experimental section
4.1 Chemistry
4.1.1 Generalities
13
19
1H, C, and F NMR spectra were recorded at 400, 100.6 and 376.3 MHz, respectively. Spectra
were recorded at 298 K if not specified otherwise. Chemical shifts (δ) are given in ppm relative to
the internal standards tetramethylsilane and trichlorofluoromethane. HRMS spectra were registered
on Agilent Technologies 6540 UHD Accurate Mass Q-TOF LC/MS, HPLC 1290 Infinity. Melting
points were corrected using a calibration curve established with authentic standards. Purities of the
final compounds were ≥98% pure and were determined by UHPLC: column, Phenomenex AERIS
Peptide 1.2 mm × 1000 mm (1.7 ꢀm); flow rate, 0.8 mL/min; acquisition time, 20 min; DAD 190-
650 nm; oven temperature, 45 °C; gradient of acetonitrile in water containing 0.1% of formic acid
(0-100% in 20 minutes).
Tetrahydrofuran was stored over potassium hydroxide pellets in the presence of cuprous chloride,
from which it was distilled, before being redistilled from sodium wire after the characteristic blue
20

ACCEPTED MANUSCRIPT
color of in situ generated sodium biphenyl ketyl (benzophenone-sodium "radical anion") had been
found to persist. Air and moisture sensitive compounds were stored in Schlenk tubes or Schlenk
burettes. They were protected by and handled under an atmosphere of 99.995 % pure nitrogen,
using appropriate glassware. All the commercial available products were bought and used without
any further purification.
4.1.2 Products
Dimethyl 2-(2-methoxyphenoxy)malonate (6) [38]. A sodium methoxide solution in methanol
(2.3 g of Na, 0.10 mol, 1.0 eq in 75 ml of MeOH) was slowly added to a stirred mixture of
guaiacole (12.5 g, 0.100 mol, 10 eq) and dimethyl 2-chloromalonate (19.0 g, 0.114 mol, 1.14 eq).
After having reacted for 1 h at 45 °C, the volatile material were evaporated and the residue was
taken up in toluene, washed with water and NaHCO3 (1% aqueous solution). The concentrated
organic phase afforded 20.0 g (yield 78.7 %) of a viscous oil that was characterized as follows: bp
1
131 °C (lit. [38] bp 128 °C); H NMR (CDCl3) δ 7.12 – 7.00 (m, 2H), 6.98 – 6.90 (m, 1H), 6.90 –
6.82 (m, 1H), 5.27 (s, 1H), 3.85 (s, 3H), 3.84 (s, 6H); 13C NMR (CDCl3) δ 166.2 (2C), 150.4, 146.0,
124.4, 120.8, 119.0, 112.5, 78.4, 55.8, 53.0 (2C); HRMS calcd for C12H14O6 255.0869 (M+H+),
found 255.0870 (M+H+).
2-Amidinopyrimidine hydrochloride (7) [38]. A MeONa solution in methanol (0.35 g of Na, 15
mmol, 0.10 eq in 112 ml of MeOH) was slowly added to 2-cyanopyrimidine (13.8 g, 0.131 mol, 1.0
eq) and the solution was stirred at rt for 24 h. Acetic acid (0.75 ml, 11.9 mmol, 0.09 eq) was added,
followed by NH4Cl (7.80 g, 0.146 mol, 1.1 eq) and the suspension was stirred for further 24 h. The
solid was filtered, the solution concentrated and the residue was triturated with diethyl ether. The
solid was solved in methanol (25 ml) and filtered. Ethyl acetate (500 ml) was added and the
resulting crystalline precipitate filtered off and dried at 80 °C to afford 7.8 g (yield 37.9%) of a
white solid. Mp 126-127 °C; 1H NMR (D2O) δ 8.95 (s, 1H), 7.72 (s, 1H), 4.79 (s, HDO); 13C NMR
21

ACCEPTED MANUSCRIPT
(D2O) δ 159.8, 158.3, 152.3, 124.8 (2C); HRMS calcd for C5H6N4 123.0671 (M+H+), found
123.0666 (M+H+).
4,6-Dihydroxy-5-(o-methoxyphenoxy)-2,2'-bipyrimidine (8) [38]. A methanol solution (25 ml) of
dimethyl 2-(2-methoxyphenoxy) malonate 6 (11.0 g, 430 mmol, 1.0 eq) was added to a stirred
MeONa solution in methanol (4.85 g of Na, 210 mmol, 5 eq in 50 ml of MeOH). After 30 minutes,
2-amidinopyrimidine hydrochloride 7 (7.50 g, 47.0 mmol, 1.1 eq) was added and the reaction
stirred for further 20 hours. The solvent was evaporated, the residue was taken up in NaOH 1 N,
acidified with HCl 1 N and the precipitate was filtrated off, washed with water and dried under high
vacuum at 80 °C to afford 13.4 g (yield 55.1%) of a yellow solid that was characterized as follows.
1H NMR (DMSO-d6) δ 12.34 (s, 2H), 9.02 (d, J = 4.9 Hz, 2H), 7.71 (t, J = 4.7 Hz, 1H), 7.10 – 6.97
(m, 1H), 6.92 (t, J = 7.6 Hz, 1H), 6.77 (t, J = 7.6 Hz, 1H), 6.65 (d, J = 7.6 Hz, 1H), 3.81 (s, 3H); 13C
NMR (DMSO-d6) δ: 158.4 (2C), 157.1 (2C), 149.6, 148.7 (2C), 146.8, 123.2, 122.2, 121.5, 120.8,
113.5, 113.2, 56.0; HRMS calcd for C15H12N2O4 313.0937 (M+H+), found 313.0939 (M+H+).
4,6-Dichloro-5-(o-methoxyphenoxy)-2,2'-bipyrimidine (9) [38]. A mixture of 4,6-dihydroxy-5-
(o-methoxyphenoxy)-2,2'-bipyrimidine 8 (2.08 g, 6.40 mmol, 1.0 eq), N,N-diisopropylethylamine
(2.20 ml, 22.9 mmol, 3.5 eq) and PCl5 (2.75 g, 13.2 mmol, 2.0 eq) was refluxed in acetonitrile (20
ml) for 20 h. The cooled mixture was dropped in iced water and extracted with ethyl acetate. The
organic phase was evaporated and the residue was chromatographed on Silica gel with diethyl ether,
followed by ethyl acetate, to afford 1.68 g (yield 74.9 %) of a yellow solid that was characterized as
follows: mp 80-81 °C [lit. [38] mp 79-80 °C]; 1H NMR (CDCl3) δ 9.05 (d, J = 4.8 Hz, 2H), 7.46 (t,
J = 4.9 Hz, 1H), 7.21 – 7.05 (m, 1H), 7.00 (dd, J = 8.1, 1.5 Hz, 1H), 6.95 – 6.84 (m, 1H), 6.80 (dd, J
13
= 8.0, 1.6 Hz, 1H), 3.85 (s, 3H); C NMR (CDCl3) δ 156.8, 154.8 (2C), 153.1, 152.0 (2C), 145.8,
141.6, 141.2, 121.5, 118.4, 117.5, 113.4, 109.5, 52.8. HMRS: calcd for C15H10Cl2N4O2 349.0259
(M+H+), found 349.0261 (M+H+).
22

ACCEPTED MANUSCRIPT
N-(6-Chloro-5-(2-methoxyphenoxy)-2,2'-bipyrimidin-4-yl)-4-(1,1,1,3,3,3-hexafluoro-2-
hydroxypropan-2-yl)benzenesulfonamide (11a). Sulfonamide 10a (2.0 g, 6.2 mmol, 1.1 eq) was
solved in dry THF (8 ml) and cooled to -78 °C under a nitrogen atmosphere. n-BuLi (1.62 N
solution in hexane, 7.70 ml, 2.2 eq) was added drop-wise. After the addition was completed, the
solution was stirred for 15 minutes, then the temperature was allowed to raise at -60 °C and the
THF was evaporated under very high vacuum. DMF was added (10 ml), followed by
dichloropyrimidine intermediate 9 (2.0 g, 5.7 mmol, 1.0 eq) and the mixture was stirred overnight at
50 °C. The reaction mixture was poured onto a 1:1 mixture of water-ice and the aqueous phase was
washed with DEE, then acidified with HCl and the resulting precipitate was filtrated off, washed
with water and dried at 110 °C to afford a bright yellow solid (2.22 g, yield 61.2%). Mp 226-229
°C; 1H NMR (DMSO-d6, T = 320 K) δ 11.82 (bs, 1H), 9.00 (d, J = 4.9 Hz, 2H), 8.74 (s, 1H), 8.15
(d, J = 7.5 Hz, 2H), 7.68 (d, J = 8.8 Hz, 2H), 7.63 (t, J = 4.9 Hz, 1H), 7.08 (d, J = 8.2 Hz, 1H), 7.05
– 6.96 (m, 1H), 6.86 – 6.77 (m, 1H), 6.56 (d, J = 8.0 Hz, 1H), 3.79 (s, 3H); 13C NMR (DMSO-d6) δ
161.9, 158.1 (2C), 156.4, 150.0, 149.1(2C), 146.2, 146.0, 134.4, 133.6, 128.5 (2C), 126.7 (2C),
123.2 (q, J = 288.7 Hz, 2C), 123.2, 122.0, 121.1, 115.0, 114.0, 77.4 (hept, J = 28.7, 27.4 Hz), 56.4;
19F NMR (DMSO-d6) δ -74.22 (s, 6F). HMRS: calcd for C24H16ClF6N5O5S 636.0543 (M+H+),
found 636.0548 (M+H+).
N-(6-Chloro-5-(2-methoxyphenoxy)-2,2'-bipyrimidin-4-yl)-4-(1,1,1,3,3,3-hexafluoro-2-
methoxypropan-2-yl)benzenesulfonamide (11b). Following the same procedure used for 11a, but
starting from sulfonamide 10b (150 mg, 0.45 mmol), the titled compound was obtained as a white
solid (100 mg, 35.1 % yield. Mp > 290 °C; 1H NMR (DMSO-d6) δ 8.93 (d, J = 4.8 Hz, 2H), 8.08 –
7.97 (m, 2H), 7.58 (t, J = 4.9 Hz, 1H), 7.47 (d, J = 8.2 Hz, 2H), 7.06 (dd, J = 8.1, 1.4 Hz, 1H), 6.96
(td, J = 7.7, 1.5 Hz, 1H), 6.78 (td, J = 7.7, 1.5 Hz, 1H), 6.42 (dd, J = 8.1, 1.4 Hz, 1H), 3.81 (s, 3H),
13
3.36 (s, 3H); C NMR (DMSO-d6) δ 162.6, 160.0, 157.9 (2C), 157.1, 149.5, 148.8, 147.8, 146.2,
133.8, 128.7 (2C), 128.1, 127.4 (2C), 122.5, 122.5 (q, J = 288.8 Hz, 2C), 121.7, 120.9, 113.8,
23

ACCEPTED MANUSCRIPT
113.5, 82.7 (hept, J = 28.7 Hz) 56.1, 54.7; 19F NMR (DMSO-d6) δ -70.71 (s, 6 F); HRMS: calcd. for
C25H18ClF6N5O5S 650.0700 (M+H+), found 650.0700 (M+H+).
N-(6-Chloro-5-(2-methoxyphenoxy)-2,2'-bipyrimidin-4-yl)-4-(11,1,1,2,3,3,3-
heptafluoropropan-2-yl)benzenesulfonamide (11c). The titled compound was obtained following
the same procedure used for 11a, but starting from sulfonamide 10c. Mp 287-289 °C; Yield 35% 1H
NMR (acetone + 5% MeOD) δ 9.00 (d, J = 4.9 Hz, 2H), 7.78 (d, J = 8.3 Hz, 2H), 7.66 (t, J = 4.9
Hz, 1H), 7.61 (d, J = 8.3 Hz, 2H), 7.00 (dd, J = 8.1, 1.6 Hz, 1H), 6.94 (td, J = 8.1, 7.7, 1.5 Hz, 1H),
6.66 (td, J = 7.7, 7.2, 1.7 Hz, 1H), 6.55 (dd, J = 8.0, 1.5 Hz, 1H), 3.79 (s, 3H); 13C NMR (CDCl3) δ
160.9, 159.4, 158.0 (2C), 154.8, 150.3, 149.4, 148.5, 146.5, 136.6, 127.6 (d, J = 20.6 Hz), 126.5
(2C), 125.5 (d, J = 10.9 Hz, 2C), 122.6, 121.9, 121.7 (qd, J = 283.7 and 30.7 Hz, 2C), 120.4, 114.9,
113.5, 79.2 – 77.6 (m), 55.6; 19F NMR (acetone-d6 + 5% MeOD) δ -76.83 (m, 6F), -183.30 (m, 1F);
HRMS: calcd. for C24H15ClF7N5O4S 638.0500 (M+H+), found 638.0509 (M+H+).
4-(tert-Butyl)-N-(6-chloro-5-(2-methoxyphenoxy)-[2,2'-bipyrimidin]-4-yl)benzenesulfonamide
(11d). Following the same procedure described for 11a, but starting from commercial 4-t-
butylbenzenesulfonamide and dichloropyrimidine 9, a pale yellow solid was obtained in 75.1% of
yield. Mp 150-151 °C [lit. [38] mp 152 °C]; 1H NMR (DMSO-d6) δ 11.81 (bs, 1H), 8.96 (d, J = 4.9
Hz, 2H), 7.71 (d, J = 8.4 Hz, 2H), 7.62 (t, J = 4.9 Hz, 1H), 7.33 – 7.25 (m, 2H), 7.10 – 7.04 (m,
1H), 7.00 – 6.93 (m, 1H), 6.76 (td, J = 7.7, 1.5 Hz, 1H), 6.41 (d, J = 8.0 Hz, 1H), 3.80 (s, 3H), 1.22
(s, 9H); 13C NMR (DMSO-d6) δ 162.3, 160.0, 158.1, 156.8, 153.2, 149.4, 148.8, 146.2, 142.6,
134.2, 127.2, 124.6, 122.6, 121.8, 120.8, 113.9, 113.6, 56.2, 34.8, 31.3; HRMS: calcd. for
C25H24ClN5O4S 526.1316 (M+H+), found 526.1313 (M+H+).
4-(1,1,1,3,3,3-Hexafluoro-2-hydroxypropan-2-yl)-N-(6-(2-hydroxyethoxy)-5-(2-
methoxyphenoxy)-2,2'-bipyrimidin-4-yl)benzenesulfonamide (5a). Chloropyrimidine 11a (0.317
g, 0.50 mmol, 1.0 eq) was added to a stirred solution of sodium glycolate (from 0.11 g of Na in 3 ml
of ethylene glycol) and the resulting mixture was stirred at 110 °C for 5 h. The cooled mixture was
24

ACCEPTED MANUSCRIPT
dropped in ice, the aqueous phase washed with diethyl ether and then acidified. The resulting
precipitate was filtrated off, washed with water and chromatographed (MeOH/EA 3:7, SiO2) to
afford 302.0 mg as a yellow solid (90.7% yield). Mp 99-100 °C; 1H NMR (DMSO-d6, T = 325 K) δ
9.00 (d, J = 4.4 Hz, 2H), 8.72 (bs, 1H, OH), 8.14 (bs, 2H), 7.73 – 7.65 (m, 3H, CH2 arom + NH),
7.65 – 7.59 (m, 1H), 7.08 – 7.01 (m, 1H), 6.95 (bt, J = 7.9 Hz, 1H), 6.76 (bt, J = 7.6 Hz, 1H), 6.61 –
6.46 (m, 1H), 4.72 (bs, 1H, OH), 4.35 (bs, 2H), 3.82 (s, 3H), 3.56 (bs, 2H); 13C NMR (DMSO-d6, T
= 325 K) δ 162.7, 161.4, 158.1 (2C), 157.5, 155.5, 149.2, 147.8, 147.3, 133.3, 128.9, 128.4 (2C),
126.7 (2C), 123.10 (q, J = 291.5, 289.0 Hz, 2C), 122.5, 121.7, 121.0, 115.0, 114.0, 77.4 (h, J = 29.4
Hz), 68.6, 60.1, 56.5; 19F NMR (DMSO-d6) δ -74.17 (s, 6F). HRMS: calcd. for C26H21F6N5O7S
662.1144, found 662.114.
4-(1,1,1,3,3,3-Hexafluoro-2-methoxypropan-2-yl)-N-(6-(2-hydroxyethoxy)-5-(2-
methoxyphenoxy)-2,2'-bipyrimidin-4-yl)benzenesulfonamide (5b).
Following the same
procedure used for 5a, but starting from 11b (150.0 mg, 0.45 mmol), The titled compound was
1
obtained as a white solid (22 mg, 22.9 % yield). Mp 163-168 °C H NMR (DMSO-d6, 325 K) δ
11.51 (bs, 1H, NH), 9.18 – 8.92 (m, 2H), 8.64 – 8.37 (m, 2H), 7.92 – 7.44 (m, 3H), 7.16 – 6.91 (m,
2H), 6.93 – 6.58 (m, 2H), 4.38 (bs, 2H), 3.80 (s, 3H), 3.52 (bs, 2H), 3.44 (s, 3H), 1.54 (bs, 1H, OH);
13C NMR (DMSO-d6) δ 161.8, 158.5 (2C), 158.2, 151.2, 149.4, 149.1, 146.1, 143.6, 131.1, 128.6
(2C), 124.2, 123.2, 122.3 (q, J = 290.5 Hz), 122.3, 122.1, 116.8, 114.0 (2C), 113.8, 82.8 (hept, J =
27.9 Hz), 69.2, 59.6, 56.5, 55.1; 19F NMR (DMSO-d6) δ -70.64 (s, 6F); HRMS: calcd. for
C27H23F6N5O7Si 676.1301 (M+H+), found 676.1303 (M+H+).
4-(1,1,1,2,3,3,3-Heptafluoropropan-2-yl)-N-(6-(2-hydroxyethoxy)-5-(2-methoxyphenoxy)-2,2'-
bipyrimidin-4-yl)benzenesulfonamide (5c). Following the same procedure described for 5a, but
starting from chloro intermediate 11c, the titled compound was obtained as a white solid (105 mg,
yield 65.3 %). Mp 243-245 °C 1H NMR (DMSO-d6, 325 K) δ 11.60 (s, 1H, NH), 9.07 (d, J = 5.0
Hz, 2H), 8.73 – 8.28 (m, 2H), 7.90 – 7.81 (m, 2H), 7.77 – 7.52 (m, 1H), 7.17 – 6.94 (m, 2H), 6.90 –
25

ACCEPTED MANUSCRIPT
6.63 (m, 2H), 4.52 (bs, 1H, OH), 4.49 – 4.23 (m, 2H), 3.79 (s, 3H), 3.54 (s, 2H); 13C NMR
(acetone-d6) δ 161.6 (d, J = 20.6 Hz), 158.4, 157.8 (2C), 155.5, 151.0 (d, J = 137 Hz), 149.8, 146.0 ,
143.5, 130.7 (2C), 130.1 (d, J = 20.6 Hz), 126.6, 126.0, 125.9, 124.6, 120.9, 120.3 (qd, J = 314.0,
27.8 Hz, 2C), 118.5 , 112.8 (2C), 91.4 (dhept, J = 201.9, 32.5 Hz), 70.1, 60.6, 55.6. 19F NMR
(DMSO-d6) δ -74.82 (d, J = 8.6 Hz, 6 F), -181.01 (m, 1F); HRMS: calcd. for C26H20F7N5O6S
664.1101 (M+H+), found 664.1104 (M+H+).
4-(tert-Butyl)-N-(6-ethoxy-5-(2-methoxyphenoxy)-[2,2'-bipyrimidin]-4-yl)benzenesulfonamide
(5d). Following the same procedure used for 5a, but starting from chloro intermediate 11d and
ethanol, the titled compound was obtained as a white solid (yield 69.3 %) that was characterized as
follows: mp 140-142 °C. 1H NMR (CDCl3) δ 8.69 (d, J = 4.9 Hz, 2H), 7.84 (s, 1H), 7.52 (d, J = 8.0
Hz, 2H), 7.06 – 6.98 (m, 3H), 6.92 – 6.72 (m, 2H), 6.60 (t, J = 7.7 Hz, 1H), 6.37 (d, J = 8.1 Hz,
1H), 4.38 (q, J = 7.3 Hz, 2H), 3.77 (s, 3H), 1.15 (t, J = 7.0 Hz, 3H), 1.13 (s, 9H); 13C NMR (CDCl3)
δ 163.5, 162.1, 161.8, 157.7 (2C), 155.3, 153.6, 148.7, 146.8, 140.9, 126.3 (2C), 125.1, 124.9, 124.7
(2C), 122.1, 120.7, 120.6, 112.5, 62.4, 56.2, 34.6, 31.0 (3C), 14.4; HRMS: calcd. for C27H29N5O5S
536.1968 (M+H+), found 536.1973 (M+H+).
4-(1,1,1,3,3,3-Hexafluoro-2-hydroxypropan-2-yl)benzene-1-sulfonyl chloride (12a) [42]. To a
mixture of 2- (4-aminophenyl)-1,1,1,3,3,3-hexafluoropropan-2-ol (10.0 g, 38.6 mmol, 1.0 eq), HCl
(37 % in water, 20 ml) and acetic acid (6 ml), NaNO2 (2.93 g, 42.5 mmol, 1.1 eq) in water (3.3 ml)
was added drop-wise at -15 °C. The temperature of the reaction was kept below -5 °C for 45
minutes. Meanwhile, a saturated solution of sulfur dioxide in acetic acid was prepared, by bubbling
into acetic acid SO2 for 15 minutes. Freshly prepared CuCl (0.920 g, 9.2 mmol, 0.25 eq) was added
to the solution at rt and while stirring, SO2 introduction was continued for 20 minutes. At 0 °C, the
diazotization reaction mixture was added in small portion to the SO2-CuCl complex solution. After
addition was completed, stirring was continued for 10 minutes while temperature was maintained
under 10 °C. The reaction mixture was poured onto a 1:1 mixture of water-ice (330 ml) and stirred
26

ACCEPTED MANUSCRIPT
until ice was melted. The mixture was then extracted with DEE (3x70 ml) and the combined
organic extracts were washed with water (2x70 ml), saturated NaHCO3 (caution, vigorous gas
evolution) and brine. The organics were then dried over Na2SO4 and concentrated under reduced
pressure. Flash chromatography of residue (100% DCM, SiO2) gave 4-(1,1,1,3,3,3-hexafluoro-2-
1
hydroxypropan-2-yl)benzene-1-sulfonyl chloride (7.10 g, 55.5 % yield). H NMR (CDCl3) δ 8.13
(d, J = 8.1 Hz, 1H), 8.01 (d, J = 9.0 Hz, 2H), 3.97 (s, 1H); 13C NMR (CDCl3) δ 142.4, 133.1, 125.0
(2C), 123.7 (2C), 118.73 (q, J = 287.1, 2C), 73.85 (hept, J = 28.6 Hz); 19F NMR (CDCl3) δ
̶ 75.67
(s, 6 F).
4-(1,1,1,3,3,3-Hexafluoro-2-hydroxypropan-2-yl)benzenesulfonamide (10a). A solution of
sulfonyl chloride 12a (5.00 g, 14.4 mmol, 1.0 eq) in THF (25 ml) was cooled with an ice bath.
Aqueous ammonium hydroxide was added drop-wise (25% water solution, 4.5 mL). After the
addition was completed, the solution was stirred at rt for 15 min. The solvent was removed in
vacuum and the residue was dissolved in NaOH 2.0 M solution, washed with DEE. The aqueous
phase was acidified with H2SO4 and then extracted with EA, washed with water and the organic
phase was dried over Na2SO4. The solvent was evaporated to afford pure 4-(1,1,1,3,3,3-hexafluoro-
1
2-hydroxypropan-2-yl)benzenesulfonamide (3.50 g, yield 74.2 %) as a white solid. H NMR
13
(DMSO-d6) δ 9.00 (s, 1H), 7.95 (d, J = 8.8 Hz, 2H), 7.87 (d, J = 8.4 Hz, 2H), 7.47 (s, 2H); C
NMR (DMSO-d6) δ 142.7, 131.2, 124.7 (2C), 123.0 (2C), 119.71 (q, J = 288.4 Hz, 2C), 73.80
(hept, J = 28.3 Hz); 19F NMR (DMSO-d6) δ
̶ 74.25 (s, 6F); HMRS: calcd for C9H7F6NO3S
324.0129 (M+H+), found 324.013 (M+H+).
2-(4-Bromophenyl)-1,1,1,3,3,3-hexafluoropropan-2-ol (13). Gaseous hexafluoroacetone was
generated by adding cautiously the sesquihydrate (6.56 g, 34.0 mmol, 2.0 eq) to concentrated (98%
aq.) sulfuric acid at 50 °C and it was condensed, through Teflon tubing, into a cold trap cooled to -
75 °C. In a second Schlenk tube, n-butyllithium (1.62 N hexane solution, 10.5 ml, 17.0 mmol) was
added drop-wise, in 15 min, to a solution of 1,4-dibromobenzene (4.00 g, 17.0 mmol, 1.0 eq) in
27

ACCEPTED MANUSCRIPT
tetrahydrofuran (40 ml) at -95 °C. Warming the cold trap slowly to 25 °C made the
hexafluoroacetone stream through a Teflon canola, ending at 5 mm above the liquid surface, into
the Schlenk tube containing the reaction intermediate. When the addition was complete, the mixture
was kept for further 10 min at -95 °C. Upon chromatography on silica gel, using a 1:20 (v/v)
mixture of diethyl ether and hexanes as the eluent, a viscous colorless oil was collected; 3.68 g
13
(67.1% yield). 1H NMR (CDCl3) δ 7.60 (s, 4H), 3.57 (s, 1H); C NMR (acetone-d6) δ 131.9 (2C),
128.3 (2C), 126.7, 125.0, 122.4 (q, J = 288.4 Hz, 2C), 74.4 (hept, J = 30.3 Hz); 19F NMR (CDCl3) δ
-76.09 (s, 6F).
1-Bromo-4-(1,1,1,3,3,3-hexafluoro-2-methoxypropan-2-yl)benzene (14). To a stirred suspension
of hexane washed NaH (60% in mineral oil, 238 mg, 6.20 mmol, 1.15 eq) in dry DMF (5 ml)
alcohol 13 (1.72 g, 5.4 mmol, 1.0 eq) solution in dry DMF (5 ml) was added under nitrogen
atmosphere at 0 °C. After 40 minutes, methyl iodide (1.2 ml, 16.2 mmol, 3.0 eq) was added and the
mixture was further reacted at rt for 40 minutes. Then, the mixture was added to ice and extracted
with DEE (3x30 ml). The organic phases were reunited, the solvent removed under vacuum and the
residue distilled at 19 mmHg (96 °C) to afford a colorless oil (1.630 g, 89.6% yield) of 1-bromo-4-
1
(1,1,1,3,3,3-hexafluoro-2-methoxypropan-2-yl)benzene. H NMR (CDCl3) δ 7.70 – 7.52 (m, 2H),
7.57 – 7.35 (m, 2H), 3.48 (hept, J = 1.0 Hz, 3H); 13C NMR (Acetone-d6) δ 132.2 (2C), 130.1 (2C),
19
126.7, 124.7, 122.3 (q, J = 289.6 Hz, 2C), 74.5 (hept, J = 30.3 Hz), 54.0; F NMR (CDCl3) δ -
71.41 (s, 6 F).
4-(1,1,1,3,3,3-Hexafluoro-2-methoxypropan-2-yl)benzenesulfonamide
(10b)
[41].
n-
Butyllithium (1.62 N hexane solution, 2.75 ml, 4.45 mmol) was added drop-wise, in 15 min, to a
solution of bromo derivate 14 (1.50 g, 4.45 mmol, 1.0 eq) in tetrahydrofuran (15 ml) at -78 °C and
the mixture wasreacted, while stirring, for 10 min before the bath temperature was allowed to rise to
–60 °C. Sulfur dioxide was bubbled into the mixture until its pH value reached 6-7. The cold bath
was removed and the temperature was allowed to rise to 25 °C. Hexane (20 mL) was added and the
28