Partially fluorinated alkoxy groups ? Conformational adaptors to changing environments

Article abstract of DOI:10.1016/j.jfluchem.2017.02.003

Lipophilicities of partially fluorinated n-propyloxy indole derivatives and their rates of oxidative metabolic degradation are presented. Comparison of the lipophilicity data with those of compounds containing the same partially fluorinated propyl groups attached to carbon or nitrogen reveals remarkable similarities and some distinct differences. A further striking difference in lipophilicity pattern is noted between terminally fluorinated n-propyloxy and corresponding methoxy derivatives. The lipophilicity patterns are rationalized in a consistent way by application of a simple rule-of-thumb based on polar-bond vector superposition, taking into account conformational aspects deduced from X-ray crystal structures and quantum chemical calculations. Several of these groups can switch between polar and non-polar conformations of comparable energies and may thus be regarded as potentially effective conformational adaptors to changing chemical environments. All compounds exhibit relatively high rates of metabolic degradation with a moderate correlation between degradation rate and lipophilicity.

Full text of DOI:10.1016/j.jfluchem.2017.02.003

Accepted Manuscript
Title: Partially Fluorinated Alkoxy Groups − Conformational
Adaptors to Changing Environments
Authors: Quentin A. Huchet, Nils Trapp, Bernd Kuhn, Bjo¨rn
Wagner, Holger Fischer, Nicole A. Kratochwil, Erick M.
Carreira, Klaus Mu¨ller
PII:
S0022-1139(16)30508-5
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.jfluchem.2017.02.003
FLUOR 8944
To appear in:
FLUOR
Received date:
Accepted date:
17-12-2016
7-2-2017
Please cite this article as: Quentin A.Huchet, Nils Trapp, Bernd Kuhn, Bjo¨rn Wagner,
Holger Fischer, Nicole A.Kratochwil, Erick M.Carreira, Klaus Mu¨ller, Partially
Fluorinated Alkoxy Groups − Conformational Adaptors to Changing Environments,
Journal of Fluorine Chemistry http://dx.doi.org/10.1016/j.jfluchem.2017.02.003
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.

Partially Fluorinated Alkoxy Groups – Conformational Adaptors to Changing
Environments
Quentin A. Hucheta, Nils Trappa, Bernd Kuhnb, Björn Wagnerb, Holger Fischerb, Nicole A.
Kratochwilb, Erick M. Carreiraa, Klaus Müllera,*
a
ETH Zürich, Laboratorium für Organische Chemie, Vladimir-Prelog-Weg 3, 8093 Zürich,
Switzerland
- Quentin A. Huchet: quentin.alban.huchet@gmail.com
- Nils Trapp : nils.trapp@org.chem.ethz.ch
- Erick M. Carreira: erickm.carreira@org.chem.ethz.ch
- Klaus Müller : klaus.mueller@org.chem.ethz.ch
b
F. Hoffmann-La Roche AG, Pharmaceutical Research & Early Development, Roche
Innovation Center Basel, Grenzacherstrasse 124, 4070 Basel, Switzerland
- Dr. Bernd Kuhn: bernd.kuhn@roche.com
- Björn Wagner: bjoern.wagner@roche.com
- Dr. Holger Fischer: holger.fischer@roche.com
- Dr. Nicole A. Kratochwil: nicole_a.kratochwil@roche.com
1

* Corresponding Author:
Prof. emer. Klaus Müller
-
Address: Prof. emer. Klaus Müller, Laboratory for Organic Chemistry,
ETHZ, CH-8093 Zurich, Switzerland
- Phone: +41 44 632 2890
- Fax: +41 44 632 1328
- E-mail: klaus.mueller@org.chem.ethz.ch
2

Text to Graphical Abstract
Depending on the substitution pattern, partially fluorinated propyloxy groups can alter their
lipophilicity by low-barrier conformational changes and thus act as adaptors to chemical
environment. Such groups should be of particular interest in drug discovery.
(259 characters, incl spaces)
Short version:
Partially fluorinated propyloxy groups can alter their lipophilicity by conformational changes
and thus act as adaptors to chemical environment. They should be of particular interest in
drug discovery.
(201 characters, incl spaces)
3

Highlights
 New lipophilicity and metabolic stability (partially fluorinated propyloxy groups)
 Successful application of simple rule-of-thumb based on a polar bond vector
superposition method as a means for the qualitative assessment of local
molecular polarity and assessment relative lipophilicity change
 Transparent interpretation of the different lipophilicity profiles for fluorinated
methoxy (exponential) versus propyloxy (parabolic) groups
 Documentation/discussion of the conformation-dependent modulation of
lipophilicity
 Recognition of some of these fluoroalkoxy groups as potential ‘conformational
adaptors’ (‘lipophilicity chameleons’) providing interesting opportunities in Drug
Discovery
 Comprehensive Table of vector superposition patterns and derived qualitative
assessments of lipophilicity changes for partially fluorinated ether moieties as a
tool for innovative drug discovery.
Abstract
Lipophilicities of partially fluorinated n-propyloxy indole derivatives and their rates of
oxidative metabolic degradation are presented. Comparison of the lipophilicity data with
those of compounds containing the same partially fluorinated propyl groups attached to
carbon or nitrogen reveals remarkable similarities and some distinct differences. A further
striking difference in lipophilicity pattern is noted between terminally fluorinated n-
propyloxy and corresponding methoxy derivatives. The lipophilicity patterns are rationalized
in a consistent way by application of a simple rule-of-thumb based on polar-bond vector
superposition, taking into account conformational aspects deduced from X-ray crystal
structures and quantum chemical calculations. Several of these groups can switch between
polar and non-polar conformations of comparable energies and may thus be regarded as
potentially effective conformational adaptors to changing chemical environments. All
compounds exhibit relatively high rates of metabolic degradation with a moderate correlation
between degradation rate and lipophilicity.
4

Keywords
5-alkoxyindoles; fluorinated alkoxy groups; lipophilicity; metabolic stability; bond vector
model; X-ray structures; conformations
1. Introduction
Incorporation of fluorine into organic compounds to modulate their physicochemical and
pharmacological properties has become a widely adopted strategy in lead optimization
programs [1]. Fluoroaromatic and trifluoromethyl units are among the most frequently used
fluorine-containing moieties [1c][2][3]. Likewise, the trifluoromethoxy group has often been
used to block potential metabolic degradation of anisole-type methoxy groups [4]. However,
the replacement of aryl methoxy by trifluoromethoxy generally results in a substantial
increase in lipophilicity [4][5] often with undesired pharmacological consequences [4]. This
lipophilicity increase can be qualitatively rationalized on the basis of a simple polar bond
vector analysis [5] suggesting that the polarity of the ether unit partially compensates the
local polarity exerted by the trifluoromethyl unit so that the volume increase from methyl to
trifluoromethyl essentially dominates the change in lipophilicity (logP). Based on this simple
rule-of-thumb analysis (Figure 1) the increase in lipophilicity upon replacement of an anisole-
type methoxy by a trifluoromethoxy group can thus be estimated to be logPest ~ +0.8. A
search in the Roche compound property database for matched molecular pairs (MMPs) with
the only structural difference being a trifluoromethoxy group in place of a methoxy group
resulted in 36 neutral matched pairs for which the average experimental lipophilicity
difference was logP ~ +1.0 ± 0.3 [5]. Very similar results have been reported by Li Xing et
al. by an MMP analysis of corresponding data in the corporate database of Pfizer [4].
5

In recent years and fostered by significant developments in synthetic methodology [6], the
difluoromethyl group has come more into focus. Unlike the axially isotropic trifluoromethyl
group, the difluoromethyl counterpart is axially anisotropic and can adopt different
conformations with characteristically different polarity depending on its immediate
topological environment [2]. Difluoromethoxy is of particular interest as a synthetically
easily accessible [7], chemically and pharmacologically rather stable unit [8].
Difluoromethoxy groups attached to aryl units are known to adopt, in the absence of steric
encumbrance, either of two typical conformations [1d][5]. One is essentially orthogonal to
the aryl unit and has the two fluorine ligands in endo positions, i.e., gauche to the COC
backbone thus antiperiplanar to the two oxygen sp3-type lone pair orbitals. This
conformation resembles that of the orthogonal aryl trifluoromethoxy group. The other
conformation has one fluorine ligand in an endo, the other in an exo position, i.e.,
antiperiplanar to the Caryl-O bond (see Figure 2). Based on a survey of published X-ray
crystal structures available in the Cambridge Structure Database (CSD) [9], this exo-endo
arrangement for the difluoromethyl unit is encountered with in-plane, twisted, or even
orthogonal difluoromethoxy groups. Density functional theory (DFT) calculations (for details
see Experimental Part) predict that the two types of aryl-OCHF2 conformations
(conformations I and II, Figure 2) are of similar energy in the gas-phase. While the
orthogonal fluorine endo,endo-conformation I profits from two (n*CF) anomeric effects at
the cost of anisole-type -conjugation, the in-plane or twisted fluorine exo,endo-conformation
has only one anomeric interaction, but regains part of the anisole-type -conjugation. In the
absence of aryl conjugation, i.e. by replacing the phenyl unit by a methyl group, the doubly
anomeric conformation is predicted to be favored by 2 kcal/mol in the gas-phase, but only
0.7 kcal/mol in polar (water) medium. Based on a simplified vector analysis the endo-endo
6

conformation of the difluoromethoxy group is estimated to be very lipophilic, similar to the
trifluoromethoxy group. By contrast, the exo-endo conformation is predicted to be rather
polar and significantly more polar than the trifluoromethoxy group (Figure 2). These findings
led to the hypothesis that a difluoromethoxy group may be able to adapt to changing
environments simply by a low-energy barrier conformational switch [5]. An indirect support
for this idea came from Li Xing and her group at Pfizer carrying out MMP analyses for a
large set of molecular pairs differing exclusively in a difluoromethoxy in place of a
trifluoromethoxy group [5]: The difluoromethoxy compounds showed a clear tendency for
improved membrane permeation compared to the congeners with a trifluoromethoxy group.
The same MMP analysis also indicated a substantial down-shift of lipophilicity by replacing
an OCF3 by an OCHF2 group. This is in line with our own findings on 15 neutral matched
molecular pairs resulting in logP = -0.7 ± 0.1 for an exchange of the lipophilic OCF3 group
by the more polar OCHF2 unit suggesting a significant contribution of the polar exo-endo
conformation of the OCHF2 group in polar (water) medium.
Little is known regarding the monofluoromethoxy group. This may be the result of a general
notion in Medicinal Chemistry that an OCH2F group may not be sufficiently stable either
chemically or pharmacologically [4]. Our search in the Roche compound database revealed
only one OCF3/OCH2F neutral matched molecular pair with logP = -1.0. This would
indicate that an OCH2F group would be similar in lipophilicity as the parent aromatic
compound or its methoxy-substituted derivative, whereas computed logP (clogP) would
predict a slight increase in lipophilicity (clogP  0.0 ÷ 0.1 depending on the method [10])
for the transition from OCH3 to OCH2F. No X-ray structural data are currently available from
the literature, but DFT calculations on the model compound monofluoromethoxybenzene at
7

the B3LYP/cc-pVDZ++ level indicate that the OCH2F group may almost exclusively adopt
the anomerically stabilized endo (gauche) conformation (Figure 3) whereas the ‘non-
anomeric’ exo conformation would be destabilized by at least 4 kcal/mol and would thus
probably not contribute significantly to the overall lipophilicity. Interestingly, the simple
vector analysis for the OCH2F group indicates a very small lipophilicity change (logPest

+0.2) in going from OCH3 to OCH2F in the anomeric gauche conformation, while the
‘elusive’ exo conformation would be predicted to be significantly more polar (Figure 3).
Of the three fluorinated aryl methoxy groups, the difluoromethoxy unit stands out as the most
interesting group exhibiting a relatively balanced conformational equilibrium with a marked
conformation-dependent lipophilicity. Environmental changes may thus easily induce shifts
in equilibrium with concomitant modulation of lipophilicity. Neither the trifluoromethoxy nor
the monofluoromethoxy group can induce such effects, the former because it is axially
symmetric, the latter because it has essentially only one conformation at its disposition.
In our search for other potentially effective conformational adaptor units to environmental
changes we turned to partially fluorinated n-propyloxy groups. The combined polarity
contributions by the C-O and C-F bonds has been examined in a few cases by the simple
bond vector analysis [4], identifying interesting candidates. In the current study, we report
our results for a series of indole derivatives carrying partially fluorinated n-propyloxy groups
in 5-position (Figure 4). These studies complement our earlier work on 3-alkyl substituted
indole derivatives [2][11] and N-alkyl-substituted piperidine carboxamide derivatives [12]
with the partially fluorinated alkyl groups attached to carbon and nitrogen atoms, respectively.
8

2. Synthesis of partially fluorinated 5-propyloxy-indoles.
The partially fluorinated 5-propyloxy-indoles 2-4 and the reference compound 1 are readily
available through coupling of corresponding n-propyl nonaflates [12] to N-tosyl protected 5-
hydroxy-indoles under mild basic conditions and subsequent indole deprotection under
somewhat harsher basic conditions (Scheme 1). For the vic-difluoro derivative 5 somewhat
milder conditions for the final deprotection of the indole unit had to be used in order to avoid
undesired side reactions. The key substrate, N-tosyl-5-hydroxyindole 10, was readily
available in an optimized three-step procedure in 83% overall yield from commercially
available 5-hydroxyindole 7, based on the combined results of previous reports [13][14]. For
the 2,2-difluoropropyloxy derivative 6 a different route involving late-stage fluorination had
to be chosen as nucleophilic substitution of the corresponding 2,2-difluoropropyl nosylate
with N-tosyl-protected 5-hydroxyindole (10) could not be effected. Thus, the latter was first
coupled with chloroacetone 22 to yield the N-tosyl-protected 2-oxo-propyloxyindole 23,
which was fluorinated by treatment with DAST and then deprotected under mild conditions
to give the desired 6 in very good yields.
With the exception of 5, intermediate N-tosyl-protected 5-propyloxy-indole derivatives 1-4
and 6 were solids and could be crystallized from methylene chloride/n-hexane solution
mixtures by slow evaporation at room temperature, offering information on conformational
aspects of the propyloxy units by X-ray crystal structure analyses (the details of the X-ray
crystal structures are compiled in the Supplementary Information).
3. Results and Discussion
9

3.1 Lipophilicity
The experimental lipophilicities (logP) of the 5-propyloxy indole derivatives 1-6 are
displayed in Figure 5. The logP pattern for the parent n-propyloxy indole 1 and its terminally
fluorinated analogues 2-4 resembles that of 3-propyl indole derivatives 10-13 [2] with logP
values systematically reduced by 0.3 to 0.4 units. The logP reduction can be understood on
the basis that indole 30 and 5-methoxyindole 31 have identical lipophilicities (2.1) [2], and
the presence of an aliphatic C3 unit in 3-propylindole (24) gives a lipophilicity up-shift of
logP  3∙0.4, while the transition from methoxyindole 31 to propyloxyindole 1 involves
homologation by an aliphatic C2 unit increasing lipophilicity by logP  2∙0.4 (Figure 5).
Furthermore, the vicinal 2,3-difluoropropyl unit manifests itself as the most polar group in
both series, being substantially more polar than the respective terminally mono- or
difluorinated analogs. A difference in relative lipophilicity is seen for the internal geminal
difluorinated propyl group, which is more polar in the case of C-attachment (29), similar to
the terminal monofluorinated analog 25, but considerably less polar in the case of O-
attachment (6), coming close to the lipophilicity of the terminal trifluorinated analog 4.
Interestingly, a very similar discrepancy in lipophilicity has been noted for the 2,2-
difluoropropyl group attached to nitrogen [12].
However, the marked parabolic dependency of logP observed for compounds 1-4 on the
degree of terminal fluorination of the n-propyloxy group is in stark contrast to the exponential
increase of logP displayed in the series of fluorinated methoxy groups (Figure 6). Thus,
introduction of a single terminal fluorine atom into the n-propyloxy group results in a
significant lipophilicity lowering by almost half a logP unit. Addition of a second fluorine at
the terminal methyl group keeps the lipophilicity still low, while the third fluorine completing
10

the trifluoromethyl group brings the lipophilicity essentially back to that of the non-
fluorinated parent compound. These findings are quite remarkable and a priori unexpected in
view of the identical bond vector superposition patterns for corresponding fluorinated
methoxy and n-propyloxy derivatives (Figure 6, top). However, the results can be
rationalized on the basis of the decoupling of anomeric effects when moving from the
methoxy to the corresponding n-propyloxy cases and from additional conformational
flexibility in the n-propyl chain.
As discussed above the monofluoromethyl unit can adopt two conformers with the fluorine
atom either endo (gauche) or exo (antiperiplanar) to a given main chain (Figure 4), differing
significantly in local polarity. In the monofluoromethoxy case the non-polar endo conformer
is highly favored by an anomeric nO  *CF hyperconjugation, essentially to the exclusion of
the polar exo conformer (Erel  4 kcal/mol), whereas in the 3-fluoropropyloxy case the two
conformations (Figure 7) are closer in energy so that the more polar exo conformation may
significantly contribute to overall lipophilicity. Likewise, the propyloxy group may adopt
low-energy gauche conformations that may also contribute significantly; hence, by
comparison to their respective non-fluorinated parent compounds, the overall lipophilicity
logP, upon introduction of one fluorine atom, remains the same or is slightly up-shifted for
the methoxy group, but substantially down-shifted for the n-propyloxy unit (Figure 7).
Moving to the difluoromethyl case, DFT calculations in the gas phase suggest that both
difluoromethoxy and 3,3-difluoro-n-propyloxy groups exhibit more balanced equilibria
between a highly non-polar endo-endo and a moderately polar endo-exo conformation. In
both cases we may then expect the moderately polar endo,exo-conformation with an
11

estimated logPest  2.9 (logPest  0, Figure 8) to contribute significantly to overall
lipophilicity in polar medium. However, this would not account for the significant logP
down-shift observed for the 3,3-difluoro-n-propyloxy case (3). A possible rationalization may
come from the X-ray crystal structure of the N-tosyl derivative 18: This structure displays a
gauche conformation for the n-propyloxy backbone (Figure 8). With an unchanged polar
endo,exo-conformation of the terminal CHF2 group, its polar C-H bond points back to the
ether oxygen with a favorable intramolecular O∙∙∙H-CF2 contact. Thus, the terminal
difluoromethyl group can act as an intramolecular weak Lewis acid polarizer at the oxygen
with two consequences, (i) increasing the C-O bond polarities (strengthening the polar logP
component), and (ii) reducing the local van der Waals volume of the difluorinated n-
propyloxy group (reducing the compensatory volume contributions to logPest). Taking both
effects into account in our simplified bond vector analysis (Figure 8) a significant logP
downshift can be estimated. The relatively low logP measured for compound 3 may then be
taken as an indication for the co-existence of such a gauche conformation in solution.
Interestingly, B3LYP calculations using a Poisson-Boltzmann approximation for solvation
effects (for details see Supplementary Information) predict this gauche conformation to be
the global minimum in both the gas-phase and polar medium and further support the notion
of its increased polarity as considered by the simple bond vector analysis.
For the trifluoromethyl case, polar bond vector analyses predict significant compensation of
the C-F by the C-O bond moments (Figure 9) so that the relatively small polar contribution to
logPest is largely overcompensated by the volume contribution to logPest resulting in a
significant upshift of logPest  +0.8 for both the trifluoromethoxy and the 3,3,3-
trifluoropropyloxy cases. While true for the former, this finding contrasts the latter, as the
12

lipophilicity of trifluoropropyloxy derivative 4 is essentially the same or even slightly lower
than that of the non-fluorinated parent compound 1. A possible explanation of this
discrepancy may come again from an X-ray crystal structure of the corresponding N-tosyl-
protected indole derivative 19 (Figure 9): the n-propyloxy unit adopts a gauche OCCC
backbone conformation. This is quite unexpected particularly in view of the relatively large
size of a CF3 group and a concomitant approximately (1,3)-parallel arrangement of polar C-O
and C-F bonds. While this conformation may be favored by crystal packing effects, it may
nevertheless also be a conformation that could contribute to overall lipophilicity in solution.
In the crystal the three fluorine atoms engage only in three probably weak contacts to H-C
units with non-bonded distances between fluorine and carbon atoms, dF∙∙∙C, at or slightly
above van der Waals contact distance (3.32 Å): they involve a phenyl C-H unit (dF∙∙∙C = 3.33
Å) and a CH3 group of neighboring tosyl moieties (dF∙∙∙C = 3.34 Å), as well as a C(2)-H unit
of an neighboring indole ring (dF∙∙∙C = 3.44 Å); all other F∙∙∙HC contacts between the CF3
group and neighboring molecules are in the range of 3.5-4.1 Å (for details see Supplementary
Information). The gauche conformation is predicted to be quite polar (Figure 9) according to
the simple bond vector analysis and may thus contribute significantly to logP in particular
since, in water the (1,5)-dispositioned fluorine and oxygen atoms may be well solvated.
B3LYP calculations predict the gauche conformation to be very close in energy to the all-
trans conformation, and its high polarity is also theoretically confirmed (see Supplementary
Information). Furthermore, the gauche conformation is doubly degenerate and may thus
accrue significant population relative to the singular all-trans conformation. Thus, the
observed relatively low logP of the trifluoropropyloxy derivative 4 may indicate the presence
of significant amounts of symmetry equivalent gauche conformations in solution.
13

Interestingly, all three terminally fluorinated n-propyloxy groups in 2-4 exhibit relatively
well-balanced equilibria between conformations of significantly different polarity. Thus, all
these groups appear to be viable candidates as conformational adaptors to changing
environments, and their incorporation into lead compounds may positively influence relevant
pharmacological properties, such as, e.g., membrane permeation.
The application of simple polar bond vector superposition and consideration of
conformational flexibility is further validated by an examination of the remaining cases
represented by the most polar compound 5 in this series and the surprisingly lipophilic
compound 6. For the vic-2,3-difluoropropyloxy derivative 5, the simple bond vector analysis
predicts relatively polar conformations assuming an all-trans OCCC backbone with the
terminal fluorine in either an exo- or endo-position (Figure 10). The two conformations are
predicted by B3LYP calculations to be of almost equal energy in polar medium. Their
associated estimated logPest values nicely embrace the experimental logP value. As for the
previous difluoro- and trifluoro-propyloxy groups in 3 and 4, a low-energy gauche OCCC
backbone conformation is predicted by B3LYP calculations for both the gas-phase and polar
water medium, which could also be considered to contribute to the relatively low lipophilicity
of 5.
For the 2,2-difluoropropyloxy group in 6 a rather low lipophilicity (logPest  2.3; logPest  -
0.6) would be predicted in the all-trans OCCC backbone conformation (Figure 11). However,
this conformation exhibits an unfavorable clustering of negative partial charges of the
fluorine and oxygen atoms and is thus predicted to be energetically rather unfavorable in the
gas-phase relative to a gauche OCCC backbone conformation in which one fluorine atom
adopts an exo position (trans to the CCO backbone). This is supported by the X-ray crystal
14

structure of the N-tosyl-protected compound 6 (21) which shows the 2,2-difluoropropyloxy
group in a gauche conformation (Figure 11). It appears that the functional unit RCF2-CX (X
= F, O, N) intrinsically favors a conformation with the two fluorine atoms in an exo (trans)
and endo (gauche) rather than a double-endo orientation relative to an electronegative atom X
(for X=F, see [15]; for X=N, see [16]) in order to avoid clustering of negatively polarized
atoms. However, in water medium this highly polar conformation may be significantly
stabilized by solvation and, thus, is predicted to be even slightly more stable than the two
symmetrically related gauche conformations. In the latter there is partial compensation of
polar bond vectors resulting in a somewhat less polar conformation for which logPest  0 by
application of the simple rule-of-thumb.
3.2 Metabolic Stability
All compounds of the new series 1-6 are relatively rapidly metabolized. The pseudo-first-
order rate constants for intrinsic clearance, Clint, as measured in human liver microsomal
assays, are shown in Figure 12. While the specific degradation mechanisms have not been
determined, we note that three out of the five partially fluorinated n-propyloxy derivatives
exhibit somewhat reduced degradation rates although rate reductions remain within a factor
of 2. Thus, the two most polar derivatives of the series, 3-fluoro- and 2,3-difluoropropyloxy
derivatives 2 and 5 are clearly more stable than the non-fluorinated parent compound 1. On
the other hand, the more lipophilic derivatives 4 and 6 are more readily oxidized than 1.
Overall this short series shows a weak trend of increased oxidative metabolism with
enhanced lipophilicity in accord with common expectation [17]. Consistent with earlier
studies of partially fluorinated n-propyl groups attached to carbon [11] or nitrogen atoms [12],
15

the terminally mono-fluorinated or vicinal difluorinated propyl groups stand out as most
promising metabolism-modulating candidates.
4. Conclusion
Experimental lipophilicities for a series of partially fluorinated n-propyloxy indole
derivatives and their rates of oxidative metabolic degradation in human liver microsomes are
presented. The lipophilicity data for this series complement earlier data on series in which the
same partially fluorinated propyl groups are attached to aromatic carbon (3-substituted indole
derivatives) or aliphatic nitrogen (N-substituted piperidine derivatives). The lipophilicity
pattern of O-alkyl derivatives display remarkable similarities to those observed for the other
series, but also some distinct differences. However, the most notable difference in
lipophilicity pattern is seen between terminally fluorinated n-propyloxy-indole derivatives
and corresponding methoxy-aryl derivatives, the latter showing an exponential increase of
lipophilicity, the former the typical parabolic dependency of intrinsic lipophilicity with
increasing degree of terminal fluorination. This discrepancy can be rationalized in a
consistent way by examination of simple polar-bond vector superposition with concomitant
consideration of conformational properties of the flexible n-propyl unit. The conformational
aspects are backed by X-ray crystal structures and quantum chemical calculations including
estimates of solvation effects. The simple polar bond vector superposition method allows
identification of distinct conformational dependency of local polarity for all partially
fluorinated n-propyloxy groups, which are nicely confirmed by quantum chemical
computations. Those with relatively narrow energy gaps between conformations of different
lipophilicity can be regarded as ‘lipophilicity chameleons’ that can easily switch from polar
to non-polar conformations depending on changing chemical environment. While interesting
16

pharmacological properties have been reported for the difluoromethoxy group, the
corresponding use and beneficial pharmacological effects of partially fluorinated n-propyloxy
groups remain to be explored. The remarkable difference between the exponential and
parabolic logP profiles of, respectively, fluorinated methoxy and n-propyloxy groups can be
traced essentially to two factors: the decoupling of anomeric effects with concomitant shifts
of equilibria between non-polar and polar conformations and to additional conformational
backbone flexibility of the n-propyloxy unit resulting in further modulation of lipophilicity.
This then leads to a consistent rationalization of lipophilicity for all members of the current
series, including the very polar vicinal difluoro- and rather lipophilic endo-geminal
difluoropropyloxy cases, by application of the same polar bond vector superposition method,
which lends further support to the validity of this simple approach.
An interesting aspect of the polar bond vector superposition method is the fact that, for
tetrahedral, well-staggered structures, a given bond vector arrangement may result from
different fluoro-substitution patterns at various topological separations from the ether unit.
Thus, different fluoroalkoxy groups in specific conformations may share one and the same
vector superposition pattern for which then the same lipophilicity change relative to their
respective non-fluorinated reference structures is predicted. Likewise, a given bond vector
arrangement may translate into different topological and conformational solutions for diverse
partially fluorinated alkyl ethers. For such analyses a comprehensive look-up table of all
possible superposition patterns with calculated resultant dipole moments, res, and estimated
lipophilicity shifts, logPest, for fluoroalkyl ethers with one up to four fluorine atoms is very
convenient and can be a source of innovative structure design in Drug Discovery. Such a
Table is included in Chapter 2 of the Supplementary Information.
All compounds in this series exhibit relatively high rates of metabolic degradation. A
moderate correlation is found between increasing degradation rate and enhanced lipophilicity.
17

Thus, the more polar derivatives, particularly the terminally mono-fluorinated and vicinal
difluorinated n-propyloxy derivatives, are somewhat more stable than the non-fluorinated
parent, whereas the more lipophilic congeners are more rapidly degraded. These findings are
consistent with corresponding findings in the earlier series of compounds with partially
fluorinated C- and N-attached n-propyl derivatives.
5. Experimental data
5.1. General information
Proton nuclear magnetic resonance (1H NMR) spectra, carbon nuclear magnetic resonance
spectra (13C NMR), and fluorine nuclear magnetic resonance (19F NMR) spectra were
recorded on Bruker AV400 spectrometer. All measurements were carried out at room
temperature (ca. 22 °C). Chemical shifts () are reported in parts per million (ppm), and
referenced to the residual solvent signal as internal standard (deuterated chloroform at 7.26
13
ppm for 1H NMR, and 77.16 ppm for C NMR). 19F NMR spectra are referenced relative to
CFCl3 in CDCl3. Coupling constants (J) are reported in units of hertz (Hz). Multiplicities are
abbreviated as follows: s (singlet), bs (broad singlet), d (doublet), dd (doublet of doublet), t
(triplet), dt (doublet of triplet), td (triplet of doublet), tt (triplet of triplet), q (quartet), qd
(quartet of doublet), p (quintet), h (sextet), and m (multiplet). 13C NMR spectra were recorded
1
1
with H-decoupling, unlike 19F NMR spectra which are reported with and without H-
decoupling. IR Spectra were recorded on a Perkin Elmer Spectrum RX-I FT-IR spectrometer
as thin film; absorptions are given in wavenumbers (cm-1). Mass spectrometric analyses
were performed by the mass spectrometric service of the Laboratorium für Orgnaische
Chemie (LOC) at ETH Zürich by L. Bertschi and O. Greter under the direction of Dr. Zhang.
HRMS measurements using positive electrospray ionization (ESI+) on Bruker Daltonics
18

maXis and Varian IonSpec ESI FT-ICR at 4.7 T. Melting points were measured using a
Büchi B-540 melting-point apparatus in open capillaries and are uncorrected. Thin-layer
chromatographic (TLC) analyses were carried out on Merck Silica Gel 60 F254 alumina
plates stained with ceric ammonium molybdate (CAM) and p-anisaldehyde.
5.2. Characterization of indole derivatives 1-6 and 16-21
5-Propyloxy-1H-indole 1: TLC: Rf = 0.53 (2:1 hexanes/EtOAc; UV, CAM, p-anisaldehyde).
1H NMR (400 MHz, CDCl3): 8.03 (bs, 1H), 7.31-7.24 (m, 1H), 7.17 (t, J = 2.8 Hz, 1H),
7.13 (d, J = 2.4 Hz, 1H), 6.89 (dd, J = 8.8 Hz, J = 2.4 Hz, 1H), 6.51-6.46 (m, 1H), 3.99 (t, J =
13
6.6 Hz, 2H), 1.91-1.80 (m, 2H), 4.34-4.20 (m, 2H), 1.07 (t, J = 7.4 Hz, 3H) ppm. C NMR
(100 MHz, CDCl3): 153.8, 131.1, 128.4, 124.9, 113.1, 111.7, 103.7, 102.5, 70.6, 22.9, 10.8.
IR (thin film): ῦ 3428, 3107, 2971, 1530, 1348, 1183, 936, 855, 834, 753, 744, 735, 613 cm-1.
HRMS (EI+): exact mass calculated for C11H13NO ([M]+), 175.0992; found 175.0994.
5-(3-Fluoropropyloxy)-1H-indole 2: TLC: Rf = 0.15 (5:1 hexanes/EtOAc; UV, CAM, p-
anisaldehyde). 1H NMR (400 MHz, CDCl3): 8.04 (bs, 1H), 7.28 (d, J = 8.7 Hz, 1H), 7.18 (t,
J = 2.8 Hz, 1H), 7.14 (d, J = 2.4 Hz, 1H), 6.88 (dd, J = 8.8 Hz, J = 2.4 Hz, 1H), 6.49 (ddd, J
= 3.1 Hz, J = 2.0 Hz, J = 0.9 Hz, 1H), 4.70 (dt, J = 47.0 Hz, J = 5.9 Hz, 2H), 4.16 (t, J = 6.1
Hz, 2H), 2.21 (dp, J = 25.8 Hz, J = 6.0 Hz, 2H) ppm. 13C NMR (100 MHz, CDCl3): 153.4,
131.3, 128.4, 125.1, 112.9, 111.8, 103.8, 102.5, 81.2 (d, J = 164.1 Hz), 64.5 (d, J = 5.5 Hz),
19
19
30.8 (d, J = 19.9 Hz) ppm. F NMR (376 MHz, CDCl3, decoupled): -221.80 (s) ppm. F
NMR (376 MHz, CDCl3, not decoupled): -221.80 (tt, J = 47.0 Hz, J = 25.8 Hz) ppm. IR
(thin film): ῦ 3412, 2967, 1471, 1454, 1282, 1223, 1155, 1124, 1047, 986, 950, 895, 838, 801,
19

756, 723, 605 cm-1. HRMS (EI+): exact mass calculated for C11H13FNO ([M]+), 194.0976;
found 194.0972.
5-(3,3-Difluoropropyloxy)-1H-indole 3: TLC: Rf = 0.24 (10:3 hexanes/EtOAc; UV, CAM, p-
anisaldehyde). 1H NMR (400 MHz, CDCl3): 8.05 (bs, 1H), 7.29 (d, J = 8.8 Hz, 1H), 7.19 (t,
J = 2.8 Hz, 1H), 7.13 (d, J = 2.5 Hz, 1H), 6.87 (dd, J = 8.8 Hz, J = 2.5 Hz, 1H), 6.50 (ddd, J
= 3.1 Hz, J = 2.0 Hz, J = 0.9 Hz, 1H), 6.14 (tt, J = 56.8 Hz, J = 4.8 Hz, 1H), 4.18 (t, J = 6.0
13
Hz, 2H), 2.35 (ttd, J = 16.5 Hz, J = 6.0 Hz, J = 4.8 Hz, 2H) ppm. C NMR (100 MHz,
CDCl3): 153.0, 131.4, 128.4, 125.2, 115.9 (t, J = 238.4 Hz), 112.8, 111.9, 103.9, 102.6,
62.8 (t, J = 6.9 Hz), 34.7 (t, J = 27.1 Hz) ppm. 19F NMR (376 MHz, CDCl3, decoupled): -
117.83 (s) ppm. 19F NMR (376 MHz, CDCl3, not decoupled): -117.83 (dt, J = 57.1 Hz, J =
16.9 Hz) ppm. IR (thin film): ῦ 3412, 2940, 1473, 1455, 1223, 1183, 1157, 1119, 1081, 978,
802, 756, 724 cm-1. HRMS (ESI+): exact mass calculated for C11H11F2NNaO ([M+Na]+),
234.0701; found 234.0705.
5-(3,3,3-Trifluoropropyloxy)-1H-indole 4: TLC: Rf = 0.29 (5:1 hexanes/EtOAc; UV, CAM,
p-anisaldehyde). 1H NMR (400 MHz, CDCl3): 8.06 (s, 1H), 7.29 (d, J = 8.8 Hz, 1H), 7.20
(t, J = 2.8 Hz, 1H), 7.14 (d, J = 2.5 Hz, 1H), 6.88 (dd, J = 8.8 Hz, J = 2.4 Hz, 1H), 6.50 (ddd,
J = 3.1 Hz, J = 2.0 Hz, J = 0.9 Hz, 1H), 4.26 (t, J = 6.7 Hz, 2H), 2.65 (qt, J = 10.6 Hz, J =
13
6.7 Hz, 2H) ppm. C NMR (100 MHz, CDCl3): 152.8, 131.5, 128.4, 126.2 (q, J = 276.6
19
Hz), 125.2, 113.0, 112.0, 104.2, 102.6, 62.1 (q, J = 3.7 Hz), 34.3 (q, J = 28.5 Hz) ppm. F
19
NMR (376 MHz, CDCl3, decoupled): -64.66 (s) ppm. F NMR (376 MHz, CDCl3, not
decoupled): -64.66 (t, J = 10.8 Hz) ppm. IR (thin film): ῦ 3414, 2928, 1474, 1455, 1285,
1251, 1220, 1134, 1069, 1008, 956, 895, 875, 839, 803, 756, 724, 603 cm-1. HRMS (ESI+):
exact mass calculated for C11H11F3NO ([M+H]+), 230.0787; found 230.0793.
20

(±) 5-(2,3-Difluoropropyloxy)-1H-indole 5: TLC: Rf = 0.27 (5:2 hexanes/EtOAc; UV, CAM,
p-anisaldehyde). 1H NMR (400 MHz, CDCl3): 8.08 (bs, 1H), 7.32-7.28 (m, 1H), 7.20 (t, J
= 2.8 Hz, 1H), 7.14 (d, J = 2.5 Hz, 1H), 6.89 (dd, J = 8.8 Hz, J = 2.5 Hz, 1H), 6.49 (ddd, J =
3.0 Hz, J = 2.0 Hz, J = 0.9 Hz, 1H), 5.17-4.92 (m, 1H), 4.88-4.65 (m, 2H), 4.34-4.20 (m, 2H)
ppm. 13C NMR (100 MHz, CDCl3): 152.8, 131.6, 128.4, 125.3, 112.8, 112.0, 104.2, 102.6,
89.8 (dd, J = 176.2 Hz, J = 19.8 Hz), 82.2 (dd, J = 172.7 Hz, J = 23.1 Hz), 66.9 (dd, J = 25.8
Hz, J = 8.0 Hz) ppm. 19F NMR (376 MHz, CDCl3, decoupled): -196.75 (d, J = 13.0 Hz), -
19
233.91 (d, J = 13.0 Hz) ppm. F NMR (376 MHz, CDCl3, not decoupled): -196.51 -
196.99 (m), -233.91 (tdd, J = 47.8 Hz, J = 21.9 Hz, J = 13.2 Hz) ppm. IR (thin film): ῦ 3415,
2957, 1478, 1452, 1285, 1225, 1162, 1124, 1105, 1032, 860, 807, 758, 727 cm-1. HRMS
(EI+): exact mass calculated for C11H11F2NO ([M]+), 211.0804; found 211.0804.
5-(2,2-Difluoropropyloxy)-1H-indole 6: TLC: Rf = 0.45 (5:2 hexanes/EtOAc; UV, CAM, p-
anisaldehyde). 1H NMR (400 MHz, CDCl3): 8.08 (bs, 1H), 7.30 (d, J = 8.8 Hz, 1H), 7.21 (t,
J = 2.8 Hz, 1H), 7.13 (d, J = 2.5 Hz, 1H), 6.91 (dd, J = 8.8 Hz, J = 2.5 Hz, 1H), 6.50 (ddd, J
= 3.1 Hz, J = 2.1 Hz, J = 0.9 Hz, 1H), 4.16 (t, J = 11.5 Hz, 1H), 1.81 (t, J = 18.8 Hz, 3H)
13
ppm. C NMR (100 MHz, CDCl3): 152.9, 131.7, 128.4, 125.3, 122.0 (t, J = 239.1 Hz),
19
112.9, 111.9, 104.3, 102.7, 71.0 (t, J = 34.2 Hz), 21.1 (t, J = 26.0 Hz) ppm. F NMR (376
MHz, CDCl3, decoupled): -98.02 (s) ppm. 19F NMR (376 MHz, CDCl3, not decoupled): -
98.02 (qt, J = 19.3 Hz, J = 11.5 Hz) ppm. IR (thin film): ῦ 3417, 2938, 1478, 1454, 1391,
1220, 1183, 1158, 1147, 1125, 1058, 904, 757, 725 cm-1. HRMS (EI+): exact mass
calculated for C11H11F2NO ([M]+), 211.0804; found 211.0802.
N-Tosyl-protected 5-propyloxy-1H-indole 16: mp = 136 °C. TLC: Rf = 0.29 (10:1
1
hexanes/EtOAc; UV, CAM). H NMR (400 MHz, CDCl3):  7.87 (d, J = 8.9 Hz, 1H), 7.73
(d, J = 8.3 Hz, 2H), 7.50 (d, J = 3.7 Hz, 1H), 7.19 (d, J = 8.2 Hz, 2H), 6.96 (d, J = 2.5 Hz,
21

1H), 6.92 (dd, J = 8.9 Hz, J = 2.5 Hz, 1H), 6.56 (d, J = 3.6 Hz, 1H), 3.91 (t, J = 6.6 Hz, 2H),
13
2.32 (s, 3H), 1.80 (h, J = 7.2 Hz, 2H), 1.03 (t, J = 7.4 Hz, 3H) ppm. C NMR (100 MHz,
CDCl3):  156.1, 144.9, 135.4, 131.9, 129.9, 129.6, 127.2, 126.9, 114.5, 114.3, 109.3, 104.6,
70.1, 22.8, 21.7, 10.7 ppm. IR (thin film): ῦ 2965, 1456, 1369, 1224, 1188, 1174, 1156, 1143,
1091, 994, 813, 720, 675, 594, 572, 538 cm-1. HRMS (ESI+): exact mass calculated for
C18H20NO3S ([M+H]+), 330.1158; found 330.1163.
N-Tosyl-protected 5-(3-Fluoropropyloxy)-1H-indole 17: mp = 93 °C. TLC: Rf = 0.25 (2:1
hexanes/EtOAc; UV, CAM). 1H NMR (400 MHz, CDCl3):  7.88 (d, J = 9.0 Hz, 1H), 7.76-
7.67 (m, 2H), 7.51 (d, J = 3.6 Hz, 1H), 7.20 (d, J = 8.2 Hz, 2H), 6.97 (d, J = 2.5 Hz, 1H), 6.92
(dd, J = 8.9 Hz, J = 2.5 Hz, 1H), 6.57 (dd, J = 3.7 Hz, J = 0.7 Hz, 1H), 4.64 (dt, J = 47.1 Hz,
J = 5.8, 2H), 4.09 (t, J = 6.1 Hz, 2H), 2.33 (s, 3H), 2.16 (dp, J = 26.1 Hz, J = 5.9 Hz, 2H)
ppm. 13C NMR (100 MHz, CDCl3):  155.7, 145.0, 135.4, 131.9, 129.9, 129.8, 127.3, 126.9,
114.6, 114.2, 109.3, 104.7, 80.9 (d, J = 164.4 Hz), 64.1 (d, J = 5.2 Hz), 30.6 (d, J = 20.0 Hz),
21.7 ppm. 19F NMR (376 MHz, CDCl3, decoupled): -222.28 (s) ppm. 19F NMR (376 MHz,
CDCl3, not decoupled): -222.28 (tt, J = 47.0 Hz, J = 26.1 Hz) ppm. IR (thin film): ῦ 2965,
1457, 1368, 1224, 1174, 1154, 1091, 1050, 994, 812, 758, 720, 703, 675, 593, 538 cm-1.
HRMS (ESI+): exact mass calculated for C18H19FNO3S ([M+H]+), 348.1064; found
348.1065.
N-Tosyl-protected 5-(3,3-Difluoropropyloxy)-1H-indole 18: mp = 133 °C. TLC: Rf = 0.40
1
(2:1 hexanes/EtOAc; UV, CAM, p-anisaldehyde). H NMR (400 MHz, CDCl3):  7.88 (d, J
= 9.0 Hz, 1H), 7.78-7.68 (m, 2H), 7.52 (d, J = 3.6 Hz, 1H), 7.21 (d, J = 8.0 Hz, 2H), 6.96 (d,
J = 2.5 Hz, 1H), 6.91 (dd, J = 9.0 Hz, J = 2.5 Hz, 1H), 6.57 (d, J = 3.6 Hz, 1H), 6.08 (tt, J =
13
56.6 Hz, J = 4.7 Hz, 1H), 4.11 (t, J = 6.0 Hz, 2H), 2.39-2.24 (m, 5H) ppm. C NMR (100
MHz, CDCl3):  155.2, 145.0, 135.4, 131.9, 130.1-129.9 (m), 127.4, 126.9, 115.7 (t, J =
22

238.6 Hz), 114.7, 114.1, 109.2, 104.8, 62.4 (t, J = 7.0 Hz), 34.5 (t, J = 21.9 Hz), 21.7 ppm.
19
19F NMR (376 MHz, CDCl3, decoupled): -118.06 (s), ppm. F NMR (376 MHz, CDCl3,
not decoupled): -118.06 (dt, J = 57.1 Hz, J = 17.0 Hz) ppm. IR (thin film): ῦ 2931, 1457,
1367, 1224, 1174, 1155, 1142, 1121, 1087, 1048, 982, 813, 760, 674, 587, 563, 538 cm-1.
HRMS (ESI+): exact mass calculated for C18H18F2NO3S ([M+H]+), 366.0970; found
366.0971.
N-Tosyl-protected 5-(3,3,3-Trifluoropropyloxy)-1H-indole 19: mp = 107 °C. TLC: Rf = 0.25
(10:3 hexanes/EtOAc; UV, CAM, p-anisaldehyde). 1H NMR (400 MHz, CDCl3):  7.89 (d, J
= 9.0 Hz, 1H), 7.76-7.70 (m, 2H), 7.53 (d, J = 3.6 Hz, 1H), 7.21 (d, J = 8.3 Hz, 2H), 6.97 (d,
J = 2.5 Hz, 1H), 6.92 (dd, J = 8.9 Hz, J = 2.5 Hz, 1H), 6.58 (dd, J = 3.6 Hz, J = 0.7 Hz, 1H),
4.19 (t, J = 6.6 Hz, 2H), 2.62 (qt, J = 10.5 Hz, J = 6.6 Hz, 2H), 2.33 (s, 3H) ppm. 13C NMR
(100 MHz, CDCl3):  155.0, 145.0, 135.4, 131.9, 130.1, 130.0, 127.5, 126.9, 126.1 (q, J =
276.6 Hz), 114.7, 114.2, 109.2, 105.0, 61.6 (q, J = 3.7 Hz), 34.2 (q, J = 28.8 Hz), 21.7 ppm.
19F NMR (376 MHz, CDCl3, decoupled): -64.75 (s), ppm. 19F NMR (376 MHz, CDCl3, not
decoupled): -64.75 (t, J = 10.8 Hz) ppm. IR (thin film): ῦ 2927, 1458, 1368, 1253, 1221,
1173, 1134, 1091, 1011, 995, 811, 759, 720, 675, 590, 577, 547, 536 cm-1. HRMS (ESI+):
exact mass calculated for C18H17F3NO3S ([M+H]+), 384.0876; found 384.0877.
(±) N-Tosyl-protected 5-(2,3-Difluoropropyloxy)-1H-indole 20: TLC: Rf = 0.27 (5:2
hexanes/EtOAc; UV, CAM, p-anisaldehyde). 1H NMR (400 MHz, CDCl3):  7.89 (d, J = 8.9
Hz, 1H), 7.75-7.69 (m, 2H), 7.53 (d, J = 3.6 Hz, 1H), 7.21 (d, J = 8.2 Hz, 2H), 6.99 (d, J =
2.5 Hz, 1H) 6.94 (dd, J = 9.0 Hz, J = 2.5 Hz, 1H), 6.58 (dd, J = 3.5 Hz, J = 0.7 Hz, 1H), 5.12-
4.88 (m, 1H), 4.85-4.59 (m, 2H), 4.22 (ddd, J = 18.5 Hz, J = 5.1 Hz, J = 1.2 Hz, 2H), 2.34 (s,
13
3H) ppm. C NMR (100 MHz, CDCl3):  155.0, 145.1, 135.3, 131.9, 130.2, 130.0, 127.6,
126.9, 114.7, 114.2, 109.2, 105.0, 89.5 (dd, J = 176.6 Hz, J = 20.0 Hz), 81.9 (dd, J = 173.1
23

Hz, J = 23.4 Hz), 66.4 (dd, J = 26.0 Hz, J = 7.9 Hz), 21.7 ppm. 19F NMR (376 MHz, CDCl3,
19
decoupled): -196.94 (d, J = 13.3 Hz), -234.34 (d, J = 13.3 Hz) ppm. F NMR (376 MHz,
CDCl3, not decoupled): -196.73 -197.16 (m), -234.34 (tdd, J = 47.6 Hz, J = 21.7 Hz, J =
13.4 Hz) ppm. IR (thin film): ῦ 3144, 2958, 1447, 1367, 1224, 1188, 1174, 1157, 1142, 1091,
1038, 996, 812, 721, 703, 671, 591, 538 cm-1. HRMS (ESI+): exact mass calculated for
C18H18F2NO3S ([M+H]+), 366.0970; found 366.0971.
N-Tosyl-protected 5-(2,2-Difluoropropyloxy)-1H-indole 21: mp = 94 °C. TLC: Rf = 0.33 (5:2
hexanes/EtOAc; UV, CAM, p-anisaldehyde). 1H NMR (400 MHz, CDCl3):  7.90 (d, J = 8.9
Hz, 1H), 7.75-7.70 (m, 2H), 7.54 (d, J = 3.7 Hz, 1H), 7.21 (d, J = 8.4 Hz, 2H), 6.99 (d, J =
2.3 Hz, 1H), 6.96 (dd, J = 8.9 Hz, J = 2.6 Hz, 1H), 6.58 (dd, J = 3.6 Hz, J = 0.8 Hz, 1H), 4.10
13
(t, J = 11.4 Hz, 2H), 2.33 (s, 3H), 1.77 (t, J = 18.8 Hz, 3H) ppm. C NMR (100 MHz,
CDCl3):  154.9, 145.1, 135.3, 131.8, 130.3, 130.0, 127.6, 126.9, 121.7 (t, J = 239.3 Hz),
19
114.7, 114.2, 109.2, 105.2, 70.4 (t, J = 34.6 Hz), 21.7, 21.1 (t, J = 25.9 Hz) ppm. F NMR
(376 MHz, CDCl3, decoupled): -98.07 (s), ppm. 19F NMR (376 MHz, CDCl3, not
decoupled): -97.94 -98.19 (m) ppm. IR (thin film): ῦ 3144, 1449, 1369, 1221, 1188, 1174,
1156, 1145, 1125, 1092, 1059, 993, 908, 813, 676, 586, 541 cm-1. HRMS (ESI+): exact mass
calculated for C18H18F2NO3S ([M+H]+), 366.0970; found 366.0970.
5.3. Determination of lipophilicity (log DpH7.4
)
In the high-throughput assay method, which is derived from the conventional “shake flask”
method, the compound of interest is distributed between a 50 mM aqueous TAPSO buffer at
pH 7.4 and 1-octanol. The distribution coefficient is then calculated from the difference in
concentration in the aqueous phase before and after partitioning and the volume ratio of the
24

two phases. To measure logD values within the range of -1 to 3.5, it is necessary to carry out
the procedure at four different octanol/water ratios. The “one-phase analysis” experiment
starts with 2-9 mL of a pure DMSO solution of the compound, which is dispensed into,
respectively, 38 or 171 mL of the aqueous buffer solution, bringing the compound
concentration to approximately 0.5 mM. A small part of this solution is then analyzed by UV.
The observed optical density corresponds to the concentration of the substrate before
partitioning. To a measured aliquot of the aqueous solution, a matching aliquot of 1-octanol is
added, and the mixture is incubated by quiet shaking for 2 h at 23 ± 1 °C. The emulsion is
allowed to stand overnight at the same temperature to ensure that partition equilibrium is
reached. Then, thorough centrifugation at 3000 rpm for 10 min is applied to separate the
layers, and the concentration of the compound in the aqueous phase is determined again by
measuring the UV absorption under the same conditions as the reference. For the neutral
compounds 1-6, logD = logP. Each logP determination was carried out at least three times
with absolute standard deviations typically < 0.1 in all cases. Nevertheless, reported logP
values rounded to one digit after the decimal point.
5.4. Determination of metabolic stability in human liver microsomes (hClint)
Microsomal incubations were carried out in 96-deep-well plates with a final incubation
volume of 600 μL. Each incubation contained 2 μM of test compound, 0.5 mg/mL rat or
human liver microsomes, and an NADPH regenerating system, which contained potassium
phosphate buffer (50 mM, pH 7.4), MgCl2 (10 mM), EDTA (1 mM), NADP+ (2 mM),
glucose-6-phosphate·2H2O (20 mM), and glucose-6-phosphate dehydrogenase (4 units/mL).
Test compounds were incubated for up to 45 min at 37 °C under vortexing at 800 rpm.
Aliquots of 50 μL were removed after 1, 3, 6, 9, 15, 25, 35, and 45 min and quenched in 150
25

μL of acetonitrile containing internal standard. Samples were then cooled and centrifuged
before analysis by high-performance liquid chromatography (HPLC) coupled with tandem-
mass spectrometry (LC-MS/MS). Log peak area ratios for the test compound and internal
standard were plotted against incubation time, and a linear fit was made with an emphasis on
the initial data points. The slope of the fit is used to estimate a pseudo-first-order rate
constant of intrinsic clearance, Clint, in units of min−1/(mg/μL of protein concentration) with a
95% confidence interval from the measurements of up to eight successive data points.
5.5. Determination of X-ray crystal structures
Crystals for compounds 16-19 and 21 were grown by vapor diffusion using dichloromethane
as solvent and n-hexane as anti-solvent. Details of the X-ray crystal structures are given in the
Supplementary Information. CCDC 1520274 (16), CCDC 1520275 (17), CCDC 1520276
(18), CCDC 1520277 (19), and CCDC 1520278 (21) contain the supplementary
crystallographic data which can be obtained free of charge from The Cambridge
Crystallographic Data Centre.[18]
5.6. Computational details
Initial geometries of all model compounds were built de novo or derived from available X-ray
structures with the modeling software MOLOC[19] and their structures subsequently
optimized with the B3LYP hybrid density functional and Dunning’s correlation-consistent
polarized split-valence double-zeta basis set with diffuse functions (cc-pVDZ++), as
implemented in the software package Jaguar[20]. Default convergence criteria were applied.
Estimates of solvation energies in water were computed with the Poisson-Boltzmann solver[21]
26

in Jaguar. B3LYP-calculated molecular dipole moments were scaled by a factor of 0.946 to
fit experimental gas-phase dipole moments of smaller fluoroalkanes.[2] The estimated local
polarities, res, from the polar bond vector superposition method correlate reasonably well (R2
= 0.879) with the B3LYP-caluclated molecular dipole moments (see Figure S1 of the
Supplementary Information).
27

Acknowledgments
We are grateful to the ETH (grant: ETH-13-10-2) and F. Hoffmann-La Roche AG for support
of the research described in this manuscript.
28

References
[1]
a) S. Swallow, Prog. Med. Chem. 54 (2015) 65–133; b) E.P. Gillis, K.J. Eastman,
M.D. Hill, D.J. Donnelly, N.A. Meanwell, J. Med. Chem. 58 (2015) 8315–8359; c)
W.K. Hagmann, J. Med. Chem. 51 (2008) 4359-4369; d) K. Müller, C. Faeh, F.
Diederich, Science 317 (2007) 1881–1886.
[2]
[3]
[4]
Q.A. Huchet, B. Kuhn, B. Wagner, H. Fischer, M. Kansy, D. Zimmerli, E.M. Carreira,
K. Müller, J. Fluorine Chem. 152 (2013) 119-128.
Thomson Reuters Integrity Database; see http://thomsonreuters.com/products_-
services/science/science_products/a-z/integrity/.
L. Xing, D.C. Blakemore, A. Narayanan, R. Unwalla, F. Lovering, R.A. Denny, H.
Zhou, M.E. Bunnage, ChemMedChem 10 (2015) 715-726.
[5]
[6]
K. Müller, Chimia 68 (2014) 356-362.
a) J. Rong, L. Deng, P. Tan, C. Ni, Y. Gu, J. Hu, Angew. Chem., Int. Ed. 55 (2016)
2743-2747; b) C. Ni, J. Liu, L. Zhang, J. Hu, Angew. Chem., Int. Ed. 46 (2007) 786-
789; c) T. Fukuzumi, N. Shibata, M. Sugiura, H. Yasui, S. Nakamura, T. Toru,
Angew. Chem. Int. Ed. 45 (2006) 4973-4977.
[7]
a) K. Levchenko, O.P. Datsenko, O. Serhiichuk, A. Tolmachev, V.O. Iaroshenko, P.K.
Mykhailiuk, J. Org. Chem. 81 (2016) 5803-5813; b) C.-B. Yue, J.-H. Lin, J. Cai, C.-P.
Zhang, G. Zhao, J.-C. Xiao, H.F. Li, RCS Adv. 6 (2016) 35705-35708; c) W.R.
Dolbier, Jr., F. Wang, X. Tang, C.S. Thomoson, L. Wang, J. Fluorine Chem. 160
(2014) 72-76; d) C.S. Thomoson, W.R. Dolbier, Jr., J. Org. Chem. 78 (2013) 8904-
8908.
29