Fluorination Patterning: A Study of Structural Motifs That Impact Physicochemical Properties of Relevance to Drug Discovery

Article abstract of DOI:10.1021/acs.jmedchem.5b01455

The synthesis of a collection of 3-substituted indole derivatives incorporating partially fluorinated n-propyl and n-butyl groups is described along with an in-depth study of the effects of various fluorination patterns on their properties, such as lipophilicity, aqueous solubility, and metabolic stability. The experimental observations confirm predictions of a marked lipophilicity decrease imparted by a vic-difluoro unit when compared to the gem-difluoro counterparts. The data involving the comparison of the two substitution patterns is expected to benefit molecular design in medicinal chemistry and, more broadly, in life as well as materials sciences.

Full text of DOI:10.1021/acs.jmedchem.5b01455

Subscriber access provided by UNIV LAVAL
Article
Fluorination Patterning: A Study of Structural Motifs that Impact
Physicochemical Properties of Relevance to Drug Discovery
Quentin A. Huchet, Bernd Kuhn, Björn Wagner, Nicole A. Kratochwil, Holger
Fischer, Manfred Kansy, Daniel Zimmerli, Erick M Carreira, and Klaus Müller
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.5b01455 • Publication Date (Web): 02 Nov 2015
Downloaded from http://pubs.acs.org on November 3, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Fluorination Patterning: A Study of Structural
Motifs that Impact Physicochemical Properties of
Relevance to Drug Discovery
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
b
b
b
b
Quentin A. Huchet , Bernd Kuhn , Björn Wagner , Nicole A. Kratochwil , Holger Fischer ,
b
b
a*
b*
Manfred Kansy , Daniel Zimmerli , Erick M. Carreira , Klaus Müller
a
ETH Zürich, Laboratorium für Organische Chemie, Vladimir Prelog Weg 3 HCI H335, CH-
093 Zürich, Switzerland
8
b
Roche Pharma Research and Early Development, Roche Innovation Center Basel, F.
Hoffmann-La Roche AG, Grenzacherstrasse 124, CH-4070 Basel, Switzerland
Abstract: The synthesis of a collection of 3-substituted indole derivatives incorporating
partially fluorinated n-propyl and n-butyl groups is described along with an in-depth study of the
effects of various fluorination patterns on properties, such as lipophilicity, aqueous solubility and
metabolic stability. The experimental observations confirm predictions of a marked lipophilicity
decrease imparted by a vic-difluoro unit when compared to the gem-difluoro counterparts. The
data involving the comparison of the two substitution patterns is expected to benefit molecular
design in medicinal chemistry and more broadly in life as well as material sciences.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Introduction
In modern drug discovery, the process of improving biological activity runs in parallel to
optimization of physicochemical and pharmacokinetic properties. Integration of property
optimization at an early stage of the discovery process has resulted in overall reduction of early-
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
phase attrition due to inadequate pharmacokinetic (PK) profiles. Consequently, property-based
design is a top priority in medicinal chemistry. The availablility of automated, sophisticated
medium- to high-throughput assays enables rapid access to the full ADME (adsorption,
2
,3
distribution, metabolism, and excretion) profile of drug candidates. Accordingly and in line
with the active hits generated from high-throughput screening approaches and combinatorial
4
,5
chemistry, which may not lead to candidates with desirable drug-like profiles, the medicinal
chemist requires tools and approaches that enable selective fine-tuning of specific properties
6
7
such as basicity (pK ), lipophilicity (logP), solubility (Sol), and metabolic clearance rates
a
8
(
Cl ).
int
The unique characteristics of fluorine substituents and their effects on molecular properties have
6
,9
led to countless applications in lead optimization, ranging from pK modulation of amines,
a
1
0
11
alteration of lipophilicity, to selective blocking of oxidative metabolism. However, although
7
the modulation of lipophilicity plays a decisive role in the pharmacokinetic behavior of drugs,
1
1,12,13
the data documenting the impact of fluorine on lipophilicity is still incomplete.
In an effort
to address this gap and further investigate the potential of fluoroalkyl groups as modulators of
lipophilicity, we recently published a study on the polarity of partially fluorinated methyl
1
4,15
groups.
The data showed that the lipophilicity trend followed by saturated alkyl-퐶ꢀ
퐹
3−푛 푛
(ꢁ = 0 ꢂ 3) derivatives is 푅퐶ꢀ > 푅퐶퐹 ≫ 푅퐶ꢀ퐹 ~ 푅퐶ꢀ 퐹. Importantly, this study revealed
3
3
2
2
that the decrease in lipophilicity upon introduction of fluorine atoms at the terminal position of
ACS Paragon Plus Environment

Page 3 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
saturated propyl chains can be understood by the interplay of two parameters: polarity and
1
6
volume, with fluorine displaying opposing influences. In this regard, the gem-difluoropropyl
case was of particular interest in comparison to its monofluorinated counterpart. It suggested that
the polarity increase generated by the introduction of the second fluorine substituent was
compensated by the concomitant volume increase, overall resulting in essentially unchanged or
slightly increased lipophilicity. Thus, the lipophilicity decrease from a nonfluorinated or
trifluorinated methyl parent to the difluoromethyl derivative is similar to that observed for the
monofluorinated analog. In combination with DFT quantumchemical calculations at a standard
B3LYP level, which nicely reproduce experimental molecular dipole moments in the gas
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
4
phase, we found that a reliable qualitative estimate of the polarity increase for a gem-difluoro
1
5
unit can be understood on the basis a simple C-F bond polarity vector analysis. Accordingly,
the overall effect of fluorine substitution on the polarity of difluoroalkyl fragments was
correlated to the geometrical relationship of the local C–F bond dipole moments.
The well known proclivity of vicinal difluorides to adopt a gauche arrangement leads to a F-C-
17
C-F dihedral angle of approximately 70°, which is significantly smaller than the corresponding
1
8,19
angle (104°) subtented by two C-F bonds in the geminal arrangement.
Indeed, summation of
the local C-F bond polarity vectors leads to the prediction that the dipole moment resulting from
the vicinal arrangement of two C-F bonds is larger than for a geminal configuration.
Consequently, in the course of property optimization for a lead structure with lower lipohilicity,
exchange of a gem-difluoroalkyl group with its vicinal analog is expected to impart a substantial
polarity increase without volume alteration. To the best of our knowledge, there has been no
systematic investigation that experimentally substantiates the extent of increase in polarity in
going from geminal to vicinal difluoro analogs. Herein, we provide the results of our study of the
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
effects of such structural modifications on polarity as well as other pharmacokinetic parameters
such as lipophilicity, solubility, and metabolic stability.
Results and Discussion
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Examination of Polarity of Vicinal Fluoroalkyl Groups: Molecular Dipole Moment. The
1
5
simplified C-F bond dipole vector analysis previously reported for the estimate of polarity
effects in partially fluorinated alkyl units rests on a number of simplifying assumptions: (i) the
overall fragment dipole moment can be estimated by considering the individual bond-dipole
moments (휇 ) of the highly polarized C–F bonds, (ii) C–F dipole moments are essentially
ꢃꢄ
invariant with respect to their structural context, and (iii) the geometry of tetravalent carbon
centers are ideally tetrahedral. Although small deviations from exact tetrahedral geometries are
2
0
well documented in the literature for partially fluorinated alkane units in both the gas phase and
21
the solid state and slight dependence of polarity on structural contexts can be deduced from ab
14
initio calculations, the simplifying assumptions have nevertheless proven appropriate for
1
4,15
qualitative predictions.
As shown in Figure 1, comparison of vic-difluoroethane (1,2-
difluoroethane) with gem-difluoroethane (1,1-difluoroethane) is quite revealing. For gem-
substituted ethane the two bond-dipole moments (휇 ) sum to afford a net dipole moment for the
ꢃꢄ
fragment of approximately 1.16 휇ꢃꢄ (Figure 1, top). For the gauche conformation of the vic-
substituted ethane, the corresponding pair of bond-dipole moments combine to result in a net
dipole moment of ca. 1.63 휇ꢃꢄ for the fragment (Figure 1, bottom). Interestingly, for perfectly
staggered tetrahedral geometries the resultant polarity, µ , correspond to the small and large
res
diagonal of one and the same rhombus spanned by the two C-F bond vectors µ in, respectively,
CF
the geminal and the vicinal arrangements. Simple geometrical considerations then predict that
ACS Paragon Plus Environment

Page 5 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
the resultant polarity in the vicinal case is larger than that in the geminal arrangement by a factor
of √2 representing a substantial increase by approximately 40%. Slight deviations in tetrahedral
geometries or modulation of C-F bond polarities in different structural contexts will essentially
not affect the general conclusion that a vicinal difluoro alkane unit exhibits greater polarity than
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
5
its geminal difluoro counterpart. As we have illustrated previously, a very simple two-
parameter model can be used to relate differences in gas-phase molecular dipole moments to
magnitudes of lipophilicity changes upon conformational, configurational, topological, or
constitutional changes of partially fluorinated alkyl units. According to this model an increase in
∆
µ in units of µ , ∆µ = n∙µ , lowers logP by a polarity contribution of δlogP = -n, while
res
CF
res
CF
p
any difference in the number of F/H exchanges, m , results in a compensating volume
F/H
contribution δlogP = +0.3∙m . For pairs of fluorinated analogs with the same number of
v
F/H
fluorine ligands, mF/H = 0; hence ∆logP may be expected just to reflect the difference in polarity
µres. A vicinal difluoroalkyl unit may thus be expected to exhibit a significantly reduced
∆
lipophilicity by ~0.4 logP units compared to the corresponding alkyl unit with a geminal
difluorination pattern. This ∆logP value may be considered to be an estimated maximum
difference given all the simplifying assumptions
Fluorinated molecules that are more elaborate than difluoroethane are not expected to strictly
1
7,18,19
adopt the geometry of simple hydrocarbons.
Thus, the polarity increase determined by the
convenient algorithm described above would be expected to deviate from estimates based on
quantumchemical calculations on molecules of relevance to the drug discovery process.
Although it would be informative to examine and compare dipole moments for molecules
measured in the gas phase, unfortunately, no experimental data are available for the fluorinated
alkyl groups of interest in this study. This led us to consider dipole moments computed for the
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
14
gas phase by standard density functional theory (DFT) as previously described and shown in
Figure 2. The calculated results reproduce the known tendency for the gauche conformation of
vic-difluoroethane to prevail in the gas phase over the trans conformation by 1.3 kcal/mol
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
Figure 2, top). Significantly, the analysis of the computations lead to the conclusion that the
gauche conformation displays a marked polarity increase (ca. 20%) compared to gem-
difluoroethane. The in silico results thus qualitatively confirm the estimates based on the simple
C-F bond dipole vector analysis (Figure 1).
Figure 1. Vector analysis comparing the resulting dipole moments (휇푟ꢅꢆ) of gem- and vic-
difluoroethane.
We next examined the homologous gem- and vic-difluorinated n-propane isomers (Figure 2,
bottom). For each of these more than one conformation is possible. For 1,1-difluoropropane
these are designated as gem-gauche-gauche (both C-F bonds gauche to the carbon backbone)
and gem-gauche-trans (C-F bonds gauche and antiperiplanar to the carbon backbone). For 1,2-
difluoropropane the two conformations are shown as vic-gauche-gauche (the two C-F bonds
ACS Paragon Plus Environment

Page 7 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
gauche to each other and the terminal C-F bond gauche to the carbon backbone) and vic-gauche-
trans (the two C-F bonds gauche to each other and the terminal C-F bond antiperiplanar to the
carbon backbone). The calculations summarized in Figure 2 reveal that each displays
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
significantly different dipole moments. In all cases the energy difference between gauche-gauche
(gg) and gauche-trans (gt) conformations in the gas phase is small (∆E 0-0.3 kcal/mol), which
rel
would suggest that both conformations are significantly populated at equilibrium. The polarity
increase from gem- to vic-difluoropropane can thus be estimated to be ∼0.53 D, corresponding to
∼
0.3 µ , which is close to the qualitative estimate based on the simple CF bond vector analysis.
CF
Figure 2. Computed gas phase dipole moments (휇푐ꢇꢈ푐) for gem- and vic-difluoroalkyl groups
(ethane or propane). Relative energies calculated in the gas phase (퐸 ) are also provided for
푟ꢅꢈ
various conformations of each compound (for computational details see Experimental Section).
Translating these results to lipophilicity considerations in the series of partially fluorinated n-
propylbenzene derivatives 1–3 (Figure 3), we previously reported that the lipophilicity of gem-
1
4
difluoro derivative 4 is predicted to be similar to, if slightly higher than that of monofluoride 2.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Considering the two gauche conformations of vic-difluoropropane as isoenergetic, log 푃 for
compound 5 can be estimated to vary between 2.7-2.9, in support of the notion that a vic-
difluoropropyl moiety imparts a substantial polarity increase onto alkyl fragments, which would
translate further into a marked lipophilicity decrease.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. Predicted Clog 푃 values for vic-difluoropropylbenzene 5 based on gas-phase
calculations.
Solvation Energy. While the use of gas-phase molecular dipole moments is quite convenient for
back-of-an-envelope lipophilicity estimates, consideration of solvation energies provide a more
direct link between the two quantities. In a previous study of a series of 3-propylindole
derivatives it was shown that calculated solvation energies can serve as good predictor of
14
polarity trends. Accordingly, it was of interest to examine further the solvation energies
calculated for prototypical geminal and vicinal difluoroalkanes units of interest in this study
(Figures 4 and 5). The comparison of values of solvation energies computed for the trans and
gauche conformations of vic-difluoroethane (Figure 4, top) is consistent with the conclusions
ACS Paragon Plus Environment

Page 9 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
drawn from calculated dipole moments in the gas phase. Note, however, that the solvation
energy for the non-polar trans-conformation (in which bond dipole moments compensate each
other by symmetry) does not vanish since the solvation energy reflects the interaction of the
polar medium with partial charges on the molecular surface, but is significantly smaller in
absolute value than that for the polar gauche-conformation. As a result, the energy difference for
the two conformations is much more pronounced in water solution than in the gas phase. While
this result is not surprising since more polar conformations are expected to be more stabilized in
a polar medium, the marked solvation energy difference between trans and gauche
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
conformations of vic-difluoroethane ∆E (gauche-trans) = -1.4 kcal/mol suggests that such
solv
fluorinated scaffolds could exhibit substantial changes in polarity effects as a consequence of
conformational flexibility. Comparison of the calculated solvation energies of gem- and vic-
difluoropropane reveals a substantial solvation energy difference of ∆E (vic-gem) = -2.2
solv
kcal/mol between these two doubly fluorinated alkyl groups, taken in their respective favored
conformations (gem-gauche-trans and vic-gauche-trans, Figure 4, bottom). This difference is
particularly striking when considering the minor structural variation between the two
constitutional isomers.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. Solvation energies, E , and relative energies, E , for difluoroethyl and
solv
rel
difluoropropyl analogs in different conformations, estimated by the Poisson-Boltzmann
approximation for B3LYP/cc-pVDZ++ geometry-optimized structures in water solution (ε = 80;
14
for computational details see ).
Very similar results were observed for the corresponding 1,1- and 1,2- difluorobutane
homologues. The calculated difference in solvation energies ∆E (vic-gem) = -2.0 kcal/mol is
solv
slightly smaller than that of the difluoropropane analogs (Figure 5). Comparison of the results for
corresponding propane and butane derivatives suggests that the effects of partial fluorination on
solvation energy are reduced in absolute magnitude with increased size of the alkyl group. We
may thus expect that a vic-difluoropropyl moiety should have a stronger impact on polarity than
the corresponding vic-difluorobutyl unit.
ACS Paragon Plus Environment

Page 11 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. Solvation energies (퐸ꢆ푠ꢈꢉ) and relative energies (E ) for different conformers of 1,1-
rel
and 1,2-difluorobutane, estimated by the Poisson-Boltzmann approximation for B3LYP/cc-
pVDZ++ geometry-optimized structures in water solution (ε = 80; for computational details see
1
4
). .
We subsequently examined the effect of gem- versus vic-fluorine substitution in a series of
partially fluorinated 3-propylindole derivatives (Figure 6). There are two key features that form
the basis of this exercise. Firstly, the contribution of the indole nucleus to the overall polarity is
presumed to be constant regardless of the alkyl substituent at C3. Secondly, the energy
differences from Figure 4 provide a basis for a prediction for the solvation energy difference
between gem-difluoro derivative 8 and its vic-difluoro analog 10 of ∆E (vic-gem) = -2.2
solv
kcal/mol. Gratifyingly, a systematic change of ∆Esolv ∼ -2.2 kcal/mol is computed for
corresponding conformations of geminal and vicinal difluorinated 3-propylindole derivatives.
Hence, vic-difluoropropyl indole derivative 10 is anticipated to be stabilized by solvation in the
range of -12.3 to -12.6 kcal/mol (Figure 6). The polarity trend is thus in qualitative agreement
with the one shown in Figure 3.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 6. Polarity pattern based on solvation energies (퐸ꢆ푠ꢈꢉ) computed for 3-substituted indole
derivatives 6–10. Upper and lower limits of computed solvation energy ranges refer to
conformations where the terminal fluorine in 7 and 10 is, respectively, gauche or trans to the
propyl backbone, or both fluorine atoms in 8 are, respectively, both gauche or gauche, trans to
the propyl backbone.
Synthetic Chemistry. A set of 3-substituted indoles were synthesized as model compounds
bearing a n-propyl or n-butyl chain incorporating the fluorination patterns of interest (Figure 7).
1
4
Our previous results involving 3-propylindole derivatives 6–9 led us to prepare and study
compounds 10–14 in order to assess physicochemical properties. The n-butyl group also allows a
shift of the vicinal difluoro-substitution pattern from the terminal to an inner position (15 and 16)
as well as the incorporation of a third fluorine atom to generate a trifluoro-substitution pattern
with a bis-vicinal arrangement of the three fluorine ligands in either a threo- or an erythro-
stereochemistry (compounds 17 and 18, respectively).
ACS Paragon Plus Environment

Page 13 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 7. Set of model compounds 10–18 discussed in this study.
The synthesis of terminal vic-difluoro derivatives 10 and 14 necessitated access to olefins 19 and
2
0, respectively (Scheme 1). Olefin 21 was isolated following Pd-catalyzed allylation of indole
2
2
(
22) with allyl alcohol (23) promoted by Et B. The corresponding butenyl substituted
3
homologue 24 was obtained by indium-mediated allylation reaction of indole-3-carbaldehyde
23
(
25) with allyl bromide (26) in presence of Hantzsch’s ester (23). Each olefin, 21 and 24, was
then subjected to protection as the corresponding N-tosylamide under phase-transfer catalysis to
2
4
afford 19 and 20 in yields of 90% and 97%, respectively.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 1. Synthesis of terminal olefin derivatives 19 and 20.
The terminal olefins 19 and 20 were converted to their respective epoxides 28 and 29 in high
2
5
yield over two steps (Scheme 2), by first dihydroxylation to the vicinal diols 30 and 31,
19,26
followed by epoxide formation
by treatment with a combination of nonaflyl fluoride and
27
28
triethylamine in acetonitrile. In line with a procedure by Hamatani et al., epoxides 28 and 29
were heated in neat triethylamine trihydrofluoride at 150 °C to afford fluorohydrins 32 and 33 as
mixtures of regioisomers. These were then advanced to vic-difluoro derivatives 34 and 35 using
29
deoxofluorination conditions described by Yin et al. Removal of the N-tosyl groups finally
3
0
afforded model compounds 10 and 14.
ACS Paragon Plus Environment

Page 15 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 2. Synthetic access to vic-difluoro derivatives 10 and 14.
The synthesis of gem-difluoro derivatives 11–13 started with Weinreb amides 36 and 37, which
2
4,31
were obtained from the corresponding carboxylic acids 38 and 39 (Scheme 3).
N-tosyl-
protected amides 36 and 37 were then converted to ketones 40-42, respectively, upon treatement
31
with MeMgBr or EtMgBr. Conversion of the ketones to the corresponding gem-difluorinated
32
indole derivatives 43–45 was effected by their treatment with DAST. Finally, the targeted
model compounds 11–13 were isolated after deprotection upon exposure to potassium hydroxide
3
0
in boiling ethanol.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 3. Synthesis of gem-difluoro-substituted derivatives 11–13.
The threo and erythro vicinal difluoro- and bis-vicinal trifluorobutyl derivatives (15–18) were
derived from a single common precursor, alkyne 49 (Scheme 4). An optimized multigram-scale
33
three-step synthesis of 49 started from 3-methylindole 46, N-tosylation to 47, and subsequent
3
4
Wohl-Ziegler bromination to the intermediate 48, which was subjected to Cu(I)-promoted
3
5
coupling with methyl propiolate to afford alkyne 49 in 47% yield over three steps after
crystallization.
ACS Paragon Plus Environment

Page 17 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 4: Three-step preparation of the key alkyne intermediate 49 and its conversion to trans-
and cis-allylic alcohol derivatives 51 and 52, respectively.
Alkyne 49 was then be selectively converted to the trans-allylic alcohol 51 by reduction with
3
6
lithium aluminum hydride. For the cis-allylic alcohol 52, the alkyne ester 49 was first
selectively reduced with diisobutylaluminium hydride to the propargylic alcohol 50 and then
converted to the cis-allylic alcohol 52 by hydrogenation using Lindlar’s catalyst in excellent
3
7
yield.
For the transformation of both allylic alcohols 51 and 52 into the corresponding fluoroepoxides
7 and 58 (Scheme 5), an two-step procedure was established that relies on the selective
epoxidation of aliphatic 1,2,3-triols. Thus, the threo- and erythro-triol derivatives 53 and 54,
5
2
5
obtained from 51 and 52, respectively, by standard olefin dihydroxylation, were subjected to
the epoxidation conditions described earlier for the synthesis of epoxides 28 and 29 (Scheme
1
9,26,27
2).
This provided the syn- and anti-epoxyalcohols 55 and 56, in good isolated yields over
two steps. The subsequent deoxofluorination to the fluoroepoxides 57 and 58, respectively, were
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
29
achieved under the mild conditions reported by Yin et al. in acceptable yields, whereas
treatment with DAST resulted in complete decomposition of the starting materials within
minutes at -78 °C.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 5: Synthesis of fluoroepoxide derivatives 57 and 58 and their conversion to vicinal
difluorobutyl derivatives 15 and 16 and bis-vicinal trifluorobutyl derivatives 17 and 18.
The anti- and syn-fluoroepoxides 57 and 58 were selectively opened either with a hydride or a
38
fluoride source. Hydride attack at the terminal position afforded the fluorohydrins 59 and 60,
which were further converted into threo and erythro vic-difluoro derivatives 61 and 62,
respectively, in good yields over two steps. Terminal opening of epoxides 57 and 58 with
triethylamine trihydrofluoride at elevated temperature resulted in the fluorohydrins 63 and 64,
2
8
respectively, which were then subjected to mild deoxofluorination conditions, leading to the
corresponding threo and erythro bis-vic-trifluoro derivatives 65 and 66 in excellent yields over
two steps. Finally, removal of the tosyl protection group was achieved with magnesium in
ACS Paragon Plus Environment

Page 19 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
39
methanol to produce the desired vicinal difluoro-derivatives 15 and 16 and the bis-vicinal
trifluoro-derivatives 17 and 18, respectively, in clean reactions and good yields.
Lipophilicity. Comparison of logP for 1,1-difluoropropyl derivative 8 with the 2,2-
difluoropropyl isomer 11 reveals that the position of the gem-difluoromethyl unit within the
propyl side chain has little influence on lipophilicity (logP=2.9 versus logP=2.8 for 8 and 11,
respectively). However, the fact that the 2,2-difluoropropyl derivative 11 is slightly less
lipophilic than the terminal isomer 8 is in line with the computed dipole moments for 2,2-
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
4
difluoropropane (2.40 D, exp. 2.39 D) and 1,1-difluoropropane (2.34 D and 2.20 D, Figure 2).
By contrast, the 1,2-vic difluorinated isomer 10 exhibits a substantial lipophilicity decrease
(
∆logP ~ -0.4 to -0.3) compared to gem-difluoropropyl analogs 8 and 11, respectively, in good
agreement with our theoretical predictions. Comparing the vic-difluoro propyl derivative to the
non-fluorinated parent compound 6, a polarity increase close to 1 log unit is achieved by
introduction of the vic-difluorination pattern.
Figure 8. Experimentally measured logP values for model compounds 6–18.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Similar but slightly diminished ∆logP effects are observed for the corresponding butyl series 12–
14 corroborating the notion that larger alkyl groups result in a diminuation of the expected
fluorine substitution effects. Again, it is the vic-difluorobutyl derivative which exhibits the most
pronounced polarity in this series.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The n-butyl group allows the vicinal difluoro substitution pattern to be moved from the terminal
to an inner position without involving the position directly adjacent to the indole unit which may
not be available for a stable monofluoro-substituted derivative. It is gratifying to see that the
terminal as well as the internal vicinal difluorobutyl derivatives 14-16 exhibit identical
lipophilicities, confirming the lipophilicity-lowering effect of the vicinal difluoro-substitution
pattern independently of its position within a linear alkyl group. The case of 16 is of particular
interest, as it represents the erythro-diastereomer. Unlike the threo-diastereomer 15, for which,
4
0
based on theoretical calculations for the 2,3-difluorobutane model system, the all-trans carbon
backbone with the two fluorine ligands gauche to each other represents the lowest-energy
conformation both in the gas-phase and in polar water medium, the erythro-diastereomer 16 has
to adopt either of two (degenerate) gauche carbon backbone conformations in order to bring the
two fluorine ligands into a gauche arrangement. These two conformations are predicted to be
only slightly disfavored in the gas-phase, but clearly favored in polar (water) medium, over the
all-trans butyl backbone with the two C-F bonds in a less favored antiplanar arrangement.
Interestingly, these conformational aspects do not translate into a measurable lipophilicity
difference, although we found that the erythro-diastereomer 16 exhibits a marginally higher
lipophilicity (logP = 2.82) compared to its threo counterpart 15 (logP = 2.77); however, the
small deviation may not be significant, given the experimental uncertainties in logP
determinations.
ACS Paragon Plus Environment

Page 21 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Comparing n-butylindol derivatives with a vicinal difluoro-substitution pattern (e.g., 14) to the n-
propylindol derivatives carrying a geminal difluoro group (e.g., 11) reveals another interesting
aspect: The lipophilicity increase upon alkyl chain homologation is just about compensated by
replacing a gem-difluoro by a vic-difluoro-substitution pattern. In this way, a ‘liponeutral’ bulk
increase can be achieved if desired in the exploration of target binding pockets.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 9: 1,2,3-Trifluorobutane with bis-vicinal arrangement of the three fluorine ligands, as
exemplified for the threo-diastereomer; C-F bond vector analysis; erythro-diastereomer with
gauche butane backbone.
The n-butyl group also allows a third fluorine to be incorporated in such a way that the three
fluorine substituents would form a bis-vicinal arrangement with both adjacent pairs of fluorines
in gauche conformations. This is particularly true for the threo-diastereomer (17) for which a
gauche-gauche arrangement of all three fluorine ligands is achieved for the low-energy all-trans
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
butyl backbone conformation (Figure 9), which would be expected also to be the preferred
conformation in polar water medium according to theoretical predictions for threo-2,3-
4
0
15
difluorobutane. Based on simple C-F bond vector considerations, such an arrangement should
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
result in a significant increase in the total dipole moment, going up from 1.63 µ for a single
CF
vicinal gauche difluoro to 1.91 µ for the bis-vicinal gauche trifluoro substitution pattern.
CF
However, this polarity gain would be expected to be largely mitigated by the concomitant
1
5
volume increase by the additional fluorine substituent. It is therefore interesting to note that the
experimental lipophilicity of threo-2,3,4-trifluorobutyl indole derivative 17 is quite comparable
to the vicinal 3,4-difluorobutyl analogs 15 and 16 (∆logP ~ -0.1). Furthermore, both the threo-
and erythro-diastereomers 17 and 18 exhibit identical lipophilicity within experimental error
limits. As pointed out above for compound 16, the n-butyl chain of the erythro-derivative 17
may adopt a gauche conformation in order to accommodate the neighboring fluorine ligands in
the preferred gauche arrangements, which would then lead to the prediction of essentially
identical polarity and hence lipophilicity of the two diastereoisomers (Figure 9).
Solubility. The influence of the vic- and gem-difluoro units on the aqueous solubility of
substituted indoles was assessed by measuring thermodynamic solubility (Sol) of model
compounds 6, 9-14 in an aqueous buffer at pH 6.5 (Figure 10, for experimental details see
Supplementary Material). From these data it is evident that the difluoropropyl indole derivatives
are more soluble than either the non-fluorinated parent 6 or the trifluoro derivative 9. This
correlates qualitatively with the observation that the difluoropropyl derivatives are more polar
than either the parent or the terminally trifluorinated propyl derivatives (Figure 8). The vic-
difluoropropyl derivative 10 outperforms its gem-difluoro counterpart 11 and is substantially
more soluble than the non-fluorinated parent 6 by approximately an order of magnitude. A
ACS Paragon Plus Environment

Page 23 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
similar, albeit much less pronounced result is also observed for the vic-difluorobutyl derivative
1
4 in comparison to its gem-difluoro-substituted isomers 12 and 13, which illustrates the
mitigating effect of an increasing size of the alkyl group.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 10. Thermodynamic solubilities for model compounds 6, 9–14. Only a lower limit of the
solubility could be determined for compound 10 due to lack of sufficient material.
Some of the solubility data cannot be compared directly because compounds 6, 9, 10 are oils
whereas the derivatives 11-14 are solids with relatively low melting points between 38 °C and 50
°
C. Thus, only the high solubility of 10 compared to its non-fluorinated and trifluorinated
congeners 6 and 9, respectively, and the better solubility of the vic-difluorobutylindole derivative
4 compared to its two gem-difluoro counterparts 12 and 13 can be taken at face value. Given
1
that for compound 10 only a lower limit of solubility could be established due to lack of
sufficient material, it may well be that even with consideration of crystal packing effects for the
solid 11, the outperformance by the vic-difluoropropyl derivative 10 would remain. In this
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
respect it is gratifying to note that the speculative solubility increase by a factor of ∼2 for the vic-
difluoro over the gem-difluoro substitution pattern in the n-propylindole series is qualitatively
reproduced in the n-butylindole series with melting points in the narrow range of 38-42 °C.
Taken together, these observations provide first indications of beneficial solubility modulation
by the vic-difluoro substitution motif. The solubility gain by the introduction of a vic-difluoro-
substitution pattern into an alkyl group may be modest relative to what can typically be achieved
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
1
by introducing polar hydrogen bond donor or even ionic functional groups. However, such
structural modifications may be counter-indicated for reasons of potential incompatibility to the
42
target binding pocket, violations of the Lipinski Rules or due to an inverse correlation of
4
3
aqueous solubilty with membrane permeability. On the other hand, small alkyl groups are
commonly used in medicinal chemistry to explore hydrophobic pockets of target binding sites
with concomitant increase in lipophilicity and reduction of aqueous solubility. Therefore,
partially fluorinated alkyl groups may serve to back-compensate lipophilicity increase and
reduction in solubility. In this regard, the vic-difluoro-substitution pattern is of particular interest.
Remarkably, in terms of volume increase, a difluoro-substituted alkyl group is between that of
1
5
the parent alkyl unit and its next higher homologue. Thus, the geminal or vicinal difluoropropyl
units occupy van der Waals volumes between those of a propyl and an n-butyl or isobutyl unit,
with a volume almost identical to that of a cyclopropylmethyl group. However, unlike the latter
the difluoro-substituted alkyl unit, and in particular the vic-difluoro derivative will exhibit a
pronounced increase in polarity.
Metabolic Stability. Metabolic stability of some of the novel partially fluorinated n-propyl and
n-butyl-substituted indoles were measured in human and rat liver microsomal degradation
assays. The intrinsic pseudo-first-order decay rate constants (Cl ) are given in Figure 11. While
int
ACS Paragon Plus Environment

Page 25 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
most Cl data come with reasonably narrow error limits, some exhibit rather large variations due
int
to difficulties in experimental compound detection and thus have to be considered with due
caution. Nevertheless, the overall patterns based on the mean Cl values remain essentially
int
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
unaffected.
All indole derivatives of this study are quite rapidly degraded by both rat and human liver
microsomes. The mode(s) of oxidative degradation have not been determined, and more than one
mechanism may be operating. This makes detailed comparisons and conclusions difficult.
Attempts to reduce oxidative attack at the indole unit by blocking the 5-position with either
fluorine or alkoxy groups failed to reduce metabolic degradation significantly, and further N-
tosylation produced largely insoluble compounds (results not shown). Nevertheless, and in spite
of quite some scatter in the Cl data, it is evident that the vic-difluoroalkyl compounds are
int
among the least affected derivatives within their respective series, the relative stabilization effect
being more pronounced in the n-propyl than in the n-butyl series. It is gratifying to note that in
our test series the introduction of the vic-difluoro-substitution pattern does not result in a
detectable increase in metabolic liability. This can be taken as an encouragement to consider this
fluorination pattern as a promising concept to modulate lipophilicity and solubility of alkyl
substituents in other structural contexts.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
-
1
Figure 11. Clearance rates (퐶ꢊ , in [min /(mg/µl) ]) measured in human (filled bars) and rat
푖푛ꢋ
prot
(
open bars) liver microsomal assays for model compounds 6, 9–14. Upper and lower bounds of
Cl values, indicated by vertical full and broken bars, respectively, represent 95%-confidence
int
levels from linear regressions of experimental compound concentrations at up to 8 successive
time points (see experimental part).
The investigation of fluorine-18 labeled radiopharmaceuticals as in vivo diagnostic imaging
agents has led French et al. to observations concerning the stability of the β-fluoroethyl group
4
4,45
compared to that of the fluoroethoxy unit.
In the study, ethers bearing a fluorine at the β-
position are less prone to metabolic defluorination than the corresponding hydrocarbon. It was
suggested that the presence of a β-heteroatom reduces the metabolic liability of the fluoroalkyl
unit presumably due to stabilization by an inductive effect. Considering the strong
electronegativity of fluorine, our observations of increased stability of vic-difluorides may be a
manifestation of a similar effect. Further extrapolation of these results may lead to considering a
vic-difluoropropoxy group as a particularly robust scaffold with significantly reduced
ACS Paragon Plus Environment

Page 27 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
lipophilicity. This will be discussed in the context of a more detailed investigation into alkoxy
groups with partial fluorine-substitution patterns.
Conclusion
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A set of 3-substituted indole derivatives incorporating a collection of partially fluorinated n-
propyl and n-butyl groups have been synthesized in order to study the effects of various
fluorination patterns on lipophilicity as well as solubility and metabolic stability. Collectively,
the experimental data confirm qualitative and more quantitative in silico predictions of a marked
lipophilicity decrease imparted by a vic-difluoro unit by contrast to its gem-difluoro counterparts.
Additional data obtained on aqueous solubility and metabolic stability underscore the importance
of both fluorination pattern and alkyl chain length making the vic-difluoropropyl moiety
substantively effective in comparison to other partially fluorinated alkyl groups. Lessons learned
regarding the polarity of vic-difluoro versus gem-difluoro-substituted alkyl groups may benefit
molecular design not only in medicinal chemistry but also in the life and material sciences.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Experimental Section
Materials and Methods
Chemistry. All non-aqueous reactions were carried out using oven-dried or flame-dried
glassware under inert atmosphere (dry nitrogen) unless otherwise stated. Stainless steel syringes
were used to transfer air- and moisture-sensitive liquids. Reactions were magnetically stirred if
not indicated otherwise and monitored by thin-layer chromatography using Merck Silica Gel 60
F254 TLC aluminium plates and visualized by fluorescence quenching under UV light. In
addition, TLC plates were stained with ceric ammonium molybdate, potassium permanganate, or
p-anisaldehyde stain. Column chromatographic purification was performed as flash
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
chromatography with 0.3–0.5 bar of air pressure on silica gel (60 Å, 230–400 mesh, Fluka
Analytical). Technical grade solvents were employed, which were distilled prior to use.
Concentration under reduced pressure was performed by rotary evaporation at 40 °C at the
appropriate pressure. Purified compounds were further dried under high vacuum. Yields refer to
chromatographically purified and spectroscopically pure compounds, unless otherwise stated.
Tetrahydrofuran, acetonitrile, toluene, diethylether and dichloromethane were dried by passage
over two 4 x 36 inch columns of anhydrous neutral A-2 alumina (8 x 14 mesh; Macherey und
Nagel; activated under a flow of nitrogen at 300 °C overnight; solvent drying system) under an
argon atmosphere (H O content < 30 ppm as determined by Karl-Fischer titration). Analytical
2
grade methanol, ethanol, and N,N-dimethylformamide from Aldrich or Acros were used without
further purification unless otherwise stated. Triethylamine and pyridine were distilled over
potassium hydroxide under dry nitrogen atmosphere. N,N-Diisopropylethylamine was distilled
over sodium hydride under dry nitrogen atmosphere. All other chemicals were purchased from
ACS Paragon Plus Environment

Page 29 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Acros, Aldrich, Fluka, Merck, Lancaster, ABCR, Fluorochem, Combi-Blocks, SynQuest Labs or
TCI and used without further purification unless otherwise stated.
Analyses. Melting points were measured using a Büchi B-540 melting-point apparatus in open
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
capillaries and are uncorrected. Proton nuclear magnetic resonance ( H NMR) spectra, carbon
1
3
19
nuclear magnetic resonance ( C NMR) spectra, and fluorine nuclear magnetic resonance ( F
NMR) spectra were recorded on a Bruker AV400 spectrometer operating at 400 MHz. All
measurements were carried out at room temperature (ca. 22 °C) unless otherwise specified.
Chemical shifts (δ) are reported in parts per million (ppm), and are referenced to the residual
1
solvent signal as internal standard (deuterated chloroform at 7.26 ppm for H NMR, and 77.16
1
3
19
ppm for C NMR). F NMR spectra are referenced relative to CFCl (δ = 0.00 ppm) in
3
deuterated chloroform. Coupling constants (J) are reported in Hz. Multiplicities are abbreviated
as follows: s (singlet), bs (broad singlet), d (doublet), t (triplet), q (quartet), p (quintet), h (sextet),
13
19
and m (multiplet). All C NMR spectra were recorded with proton decoupling. All F NMR
spectra were recorded either with or without proton decoupling. Infrared spectra were recorded
on a Perkin Elmer Spectrum RX-I FT-IR spectrometer as thin film. Absorptions are given in
–
1
wavenumbers (cm ). Mass spectrometric analyses were recorded by the mass spectrometry
service at ETH Zurich under the direction of Dr. Xiangyang Zhang. EI-MS (m/z): Waters
Micromass AutoSpec Ultima spectrometer. ESI-MS (m/z): Bruker Daltonics maXis spectrometer.
Chemical names and structures of the reported molecules were generated with ChemBioDraw
Ultra 12.0 (Cambridgesoft) and modified where appropriate.
Computational Details. Initial geometries of all model compounds were built with the modeling
4
6
software MOLOC and their structures subsequently optimized with the B3LYP hybrid density
functional and Dunning’s correlation-consistent polarized split-valence double-zeta basis set
ACS Paragon Plus Environment