Conformational control enabled by the fluorine gauche effect in a model of the β2-AR agonist salbutamol (Ventolin)

Article abstract of DOI:10.1016/j.jfluchem.2018.02.007

The bronchodilator salbutamol adopts a characteristic gauche conformation about the ?_O-C-C-N torsion angle. This topology is predicated on stabilizing stereoelectronic interactions of the type σ → σ_C-X*. X-ray crystallographic analysis of salbutamol also indicates that an intramolecular hydrogen bond reinforces this intuitive conformation (?_O-C-C-N ≈ 60°). In this study, we demonstrate that single site OH → F substitution in model salbutamol systems preserves the gauche conformational preference by virtue of reinforcing hyperconjugative interactions of the type σ_C-H → σ_C-F* and σ_C-C → σ_C-N*. Since the amine remains fully protected throughout this conformational analysis, intramolecular hydrogen bonding can be discounted. Conformational mimesis is confirmed by NMR spectroscopy in solution, and also in the solid state.

Full text of DOI:10.1016/j.jfluchem.2018.02.007

Accepted Manuscript
Title: Conformational Control Enabled by the Fluorine
Gauche Effect in a Model of the ␤2-AR Agonist Salbutamol
(VentolinTM
)
Authors: Charlotte S. Teschers, Constantin G. Daniliuc,
Gerald Kehr, Ryan Gilmour
PII:
S0022-1139(18)30039-3
DOI:
Reference:
https://doi.org/10.1016/j.jfluchem.2018.02.007
FLUOR 9121
To appear in:
FLUOR
Received date:
Revised date:
Accepted date:
24-1-2018
8-2-2018
8-2-2018
Please cite this article as: Teschers CS, Daniliuc CG, Kehr G, Gilmour R,
Conformational Control Enabled by the Fluorine Gauche Effect in a Model of the
␤2-AR Agonist Salbutamol (VentolinTM), Journal of Fluorine Chemistry (2010),
https://doi.org/10.1016/j.jfluchem.2018.02.007
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.

Conformational Control Enabled by the Fluorine Gauche Effect in a Model of the 2-AR Agonist
Salbutamol (VentolinTM)
Charlotte S. Teschers,a Constantin G. Daniliuc,†a Gerald Kehr,a Ryan Gilmour*a,b
a Organisch-Chemisches Institut, Westfälische Wilhelms-Universität Münster, Corrensstrasse 40, 48149
Münster, Germany.
b Cluster of Excellence ‘Cells in Motion’, Westfälische Wilhelms-Universität Münster, Münster, Germany.
E-mail: ryan.gilmour@uni-muenster.de
Phone: +49 251 83 33 271
† X-ray crystallographer
Graphical abstract
Highlights


Synthesis of two fluorinated derivatives of salbutamol
Both derivatives are effective conformational mimics of salbutamol in solution as confirmed by
NMR conformer population analysis

Single crystal X-ray structure analysis of one fluorinated derivative allows the torsion angle to be
determined and clearly demonstrates the gauche effect

Abstract
The bronchodilator salbutamol adopts a characteristic gauche conformation about the ϕO-C-C-N torsion
angle. This topology is predicated on stabilizing stereoelectronic interactions of the type σ→σC-X*. X-ray
crystallographic analysis of salbutamol also indicates that an intramolecular hydrogen bond reinforces
this intuitive conformation (ϕO-C-C-N ≈ 60°). In this study, we demonstrate that single site OH → F
substitution in model salbutamol systems preserves the gauche conformational preference by virtue of
reinforcing hyperconjugative interactions of the type σC-H→σC-F* and σC-C→σC-N*. Since the amine remains
fully protected throughout this conformational analysis, intramolecular hydrogen bonding can be
discounted. Conformational mimesis is confirmed by NMR spectroscopy in solution, and also in the solid
state.
Keywords Conformation; Fluorine gauche effect; Molecular editing; NMR spectroscopy; X-ray
crystallography
1. Introduction
β2-Adrenoreceptor (AR) agonists constitute a venerable class of bronchodilators for the clinical
management of pulmonary diseases such as asthma and chronic obstructive pulmonary disease.[1, 2]
Common to several β2-AR agonists is the phenolethanolamine structural motif as indicated in Figure 1.
Whilst this privileged scaffold containing a characteristic β-amino alcohol is essential for biological
efficancy, the physicochemical profile of these compounds can be tailored by subtle structural
permutations, thus rendering this core highly modular.[3, 4] A pertinent example of this development
process is evident from a structural comparison of the natural hormone epinephrine (adrenalin) with
salbutamol. Marketed as VentolinTM, sales for this WHO listed essential medicine topped $1.5 billion in
2011.[2, 5]
A structural study of salbutamol reveals a dihedral angle around the C(sp3)-C(sp3) OCCN ≈ 60° with the
aromatic core positioned anti to the amine.[8] Crystallographic analyses also reveal an intramolecular
hydrogen bond between the amine and hydroxyl groups. This gauche arrangement of the β-amino
alcohol is a manifestation of geometrically predisposed, stabilising hyperconjugative interactions
involving electron rich donor orbitals with low-lying acceptors: σC-H→σC-O* and σC-C→σC-N*. These
reinforcing interactions are the basis of the gauche effect that is synonymous with 1,2-difluoroethane.[9-
14]
Given the clinical prominence of β2-adrenoreceptor (AR) agonists containing a β-amino alcohol motif,
coupled with our interest in β-fluoroamines,[15-20] we sought to emulate the conformational behaviour
inherent to salbutamol with a model fluorinated system. Molecular editing with fluorine continues to
inform the drug development process,[21-23] but in this instance, engaging the low-lying anti-bonding
orbital of the C-F bond (σC-F*) also offers the possibility to mimic the 3D topology of the original scaffold

whilst retaining the electronic character. Speficially, it was envisaged that a single site OH
F
substitution would preserve the gauche conformation due to σC-H→σC-F* and σC-C→σC-N* interactions
(Figure 2). Whilst this would provide a strategy to control torsional rotation about the C(sp3)-C(sp3) bond,
rotation around the aryl C(sp2)-C(sp3) bond required clarification due to the possibility for π→σC-F*.[24]
Of course, it is important to recognize that stereoelectronic effects only manifest themselves in the
absence of overriding steric constraints such as those imposed by an enzyme.[25-28] In this case, the
active binding conformation of slabutamol is gauche [29, 30] such that this effect likely plays a
reinforcing role.[26-28] For the purposes of this conformational study, compound 1 was initially selected.
The highly functionalised core was deemed prudent to mitigate the risk of fluoride expulsion from a free
p-OH, and the benzylated tertiary amine would eliminate the possibility of an intramolecular hydrogen
bond,[31, 32] thereby simplifying the interpretation.
In the initial conformer population analysis, three species partitioned by 120° (Figure 3) were
considered.[33, 34] Whilst gauche-1 and gauche-2 satisfy the stereoelectronic requirements of the
gauche effect,[13, 35, 36] it is likely that gauche-1 will be less populated on account of de-stabilising
non-bonding interactions between the tertiary amine and Ar group. Moreover, the anti conformer is
sterically and stereoelectronically disfavoured such that conformational refinement is expected. Finally,
since the C(sp3)-F bond is benzylic it is conceivable that the π-system will align with the σC-F* orbital
providing an additional degree of structural pre-organisation. However, given the relatively small
energetic advantage, this effect will only manifest itself in the absence of overriding steric interactions.
Herein we report the synthesis and conformational analysis of a salbutamol analogue. Single site OH → F
substitution conserves the conformational preference inherent to the parent scaffold and demonstrates
that β-fluoroamines are excellent structural mimics of β-amino alcohols.
2. Results and Discussion
Synthesis: The synthesis of fluorinated salbutamol derivative 1 was accomplished in six steps, starting
from commercially available methyl 5-acetylsalicylate 2 (Scheme 1). Derivative 3 was prepared according
to literature procedures [37-39] in 58 % yield (over four steps) in preparation for fluorine introduction.
Deoxyfluorination was achieved by exposing substrate 3 to a mixture of diethylaminosulfur trifluoride
(DAST) (10 eq.) and triethylamine tris-hydrofluoride (1 eq.). Unfortunately, attempts to isolate the
compound via chromatography on SiO2 were unsuccessful and significant degradation was observed.
Consequently, following work-up the product was re-dissolved in EtOAc and triturated with pentane to
yield the expected product in 67% yield. Analysis of compound 4 revealed a fluorine resonance at
F
= -177.8 as a doublet of doublet of doublets with the respective coupling constants (2JFH = 47.6 Hz, 3JFH
=
29.2, 15.9 Hz). In the 1H spectrum, the characteristic ABMX spin system of the CFH-CH2 motif was
observed at 4.94 ppm with the typical 2JFH (47.6 Hz) and both 3JHH (8.1 and 3.9 Hz) coupling constants to
the diastereotopic methylene hydrogens [ H = 3.15 (3JFH = 15.9 Hz, 2JHH = 14.9 Hz, 3JHH = 8.1 Hz) and 2.75
(3JFH = 29.2 Hz, 2JHH = 14.9 Hz, 3JHH = 3.9 Hz)]. The respective NMR resonances for the 13C doublets were
recorded at 93.7 (1JFC = 173.0 Hz; CFH) and 57.8 (2JFC = 25.9 Hz; CH2) ppm (for details see the Supporting
Information).

Finally, compound 4 was converted to target 1 in 83% yield via DIBAL-H reduction of the ester.
Unfortunately, attempts to purify this oil by chromatography led to significant decomposition.
Fortunately, the reduction proceeded smoothly and cleanily to allow the conformational analysis to be
perfomed directly. The NMR data obtained are comparable to those of compound 4, but with the
substitution of the methyl ester group ( H = 3.91, C = 52.2, 166.6) by the diastereotopic hydrogens of
the hydroxymethyl group [ H = 4.70/4.68 (2JHH = 13.0 Hz), C = 62.1].
Solution phase conformer population analyses: NMR conformational analyses of the fluorinated
salbutamol derivatives 1 and 4, were conducted in CDCl3 together with an analysis of the hydroxyl
precursor 3 for comparison. For the three staggered conformers (gauche-1, gauche-2 and anti), 3J(HH)
and 3J(FH) coupling constants were calculated according to Forrest [40] and Hamman et al.[41] based on
model compounds that closely resemble 1, 3 and 4 (Figure 5). Model-1, model-3 and model-4, differ
from the experimental scaffolds only in the N-substitutent (Et versus Bn) and phenolic position (Me
versus Bn) (for further details see the Supporting Information).
This analysis required calculated 3J values which ultimately depend on the relative electronegativity E of
the substituents [40, 42]. E values were obtained directly from literature sources ( E(H) = 0.00 [43],
E(F) = 1.70 [43], E(NR2) = 1.12 [44], E(OH) = 1.30 [43] or 1.56 [43]) or calculated from published NMR
data [45] ( E(R) = 0.49 [46] for 4 and 3, E(R) = 0.52 [47] for 1, E(NR2) = 0.56 [48]). To account for a
slight discrepancy in literature data for the hydroxyl and the amine substituents, two different values for
each of these substituents were employed in the calculations to assess the effect of the parameter
choice on the conformational analysis. The conformer population distribution was determined according
to Hamman et al.[41] using the average ø 3J(HHn)g coupling constants. A summary of the coupling
constants used for the evaluation of conformer populations is provided in Table 1, together with the
experimental values.
Immediately obvious from this comparative analysis is that the calculated proton-proton gauche
coupling constants are smaller than the experimental values (see Table 1). This is expected and in line
with the gauche conformation having the smallest coupling constant of the three staggered
conformations according to the Karplus equation.[49] Consequently, experimental coupling constants
can only approach the calculated ∅ 3J if the molecule adopts the gauche conformation. Any deviation
from a 60° dihedral angle, or the population of other conformers, leads to higher experimental coupling
constants. Thus, calculations yielding coupling constants that are greater than the experimental values
would be unsuitable for estimating conformation population distributions.
For 1 and 4, conformer populations P were determined from both 3J(HH) and 3J(FH) coupling constants,
respectively [42]. The calculated population distributions for compounds 1, 4 and 3 are summarised in
Table 2.
Qualitatively, the calculated population distributions are in good agreement with each other. It is striking
that one set of parameters ( E(NR2) = 1.12 with E(OH) = 1.56 or with 3J(FH) used for the calculations)
affords higher deviations, even resulting in a negative value for the population of the anti conformer of
compound 1 (entries 4, 8 and 12). This suggests that either the literature value for E(NR2) and/or the

calculation via 3J(FH) is not approrpiate for determining the conformation population distribution for the
molecules under investigation. Consequently, these conformer population distributions have been
omitted for the remainder of the discussion.
The calculated conformer population distributions confirm the designated gauche-2 preference of
compounds 1, 4 and 3. The overall conformer population distribution of the fluorinated salbutamol
derivatives 1 and 4 follows the trend P(gauche-2) >> P(gauche-1) > P(anti). This is fully consistent with
the fluorine-nitrogen gauche effect hypothesis delineated in the introduction.
Clearly, parameter choice influences the calculation of the conformer populations less than the choice of
coupling constants [3J(HH) versus 3J(FH)] used for the calculations (Table 2, entries 1 to 3 and entries 5 to
7). It is tempting to speculate that the data obtained for calculations using 3J(HH) are more accurate, as
the influence of the trans substituents of the coupled nuclei is taken into account in the calculation of
3J(HH)g, but not for the fluorine-proton coupling constants.
In compound 1, the anti conformer is less populated (ø P(anti) = 12 %) than in the case of the
corresponding ester 4 (ø P(anti) = 18 %). Conformer gauche-1 is equally populated by both molecules
(ø P(gauche-1) = 28 % and 27 % for compounds 1 and 4, respectively). Pleasingly, the target compound
of this study, salbutamol derivative 1, has an indisputable preference for the gauche-2 conformation in
solution (ø P(gauche-2) = 60 %).
For the population distribution of the hydroxyl derivative 3, similar results were obtained with three
different data sets (entries 9 to 11) indicating a high robustness of the calculation method towards
parameter changes. Compared to the fluorinated derivatives 1 and 4, hydroxyl derivative 3 exhibits a
more pronounced preference to adopt conformation gauche-2 [ø P(gauche-2) = 70 % (3) versus
ø P(gauche-2) = 60 % (1) and ø P(gauche-2) = 55 % (4)]. Surprisingly, the gauche-1 and anti conformers
are equally populated for the hydroxyl derivative. None the less, the population distribution of all three
molecules are sufficiently similar to consider compounds 1, 4 and 3 as effective conformational mimics.
The solution phase conformation analysis of hydroxyl derivative 3 is complemented by the reported
crystal structure of salbutamol .[8] In the solid state, salbutamol adopts the gauche-2 conformation,
which is also the preferred solution conformation of salbutamol derivatives 1, 4 and 3.
The results of the solution phase conformational analysis of the fluorine derivatives 1 and 4 are
complemented by single crystal X-ray diffraction analysis of compound 4 [Figure 6 (a)]. As expected, a
gauche conformation of F1 and N1 can be observed, with a dihedral angle F1C7C8N1 of 62.6(2)°. This
allows for a stabilising hyperconjugative interaction (σC-H→σC-F*). The dihedral angle C5C7C8N1 (-176.0(2)°)
reveals an anti arrangement of the aromatic core and the amine head group, thereby allowing for the
postulated σC-C→σC-N* interaction. Thus, the conformation about the C(sp3)-C(sp3) bond is, as expected,
the gauche-2 conformation. The X-ray analysis also reveals that the C7-F1 bond is not perpendicular to
the aryl plane. For derivative 4, angles of 34.4(2)° ( C6C5C7F1) and -151.1(2)° ( C4C5C7F1) were found. Hence,
the expected π  σC-F* interaction is not observed in the solid state.
Finally, overlaying the X-ray crystal structres of salbutamol and the fluorinated derivative 4 confirms that
β-fluoroamines are excellent conformational mimics of the β-amino alcohols that are intrinsic to a large
family of β2-AR agonists [Figure 6 (b)].

3. Conclusions
Single site OH → F substitution has enabled the design of a conformational mimic of salbutamol
(VentolinTM). Solution phase conformation analysis of the target structure 1 and its precursor 4 revealed
a clear preference to adopt the gauche-2 conformation. The dominant solution phase conformer closely
resembles the conformation reported for salbutamol in the solid state. Finally, crystal structure overlay
analysis of synthesis intermediate 4 and the parent drug further underscores the success of the approach
and the value of the C(sp3)-F bond in structural chemistry.
4. Experimental
Full experimental details are provided in the Supporting Information.
4.1. methyl 5-(2-(benzyl(tert-butyl)amino)-1-fluoroethyl)-2-(benzyloxy)benzoate (4)
Optimized procedure: DAST (1.30 mL, 9.84 mmol, 9.7 equiv.) and NEt3•3HF
(0.17 mL, 1.0 mmol, 1.0 equiv.) were added to a solution of compound 3
(450 mg, 1.01 mmol, 1.0 equiv.) in dry CH2Cl2 (5 mL). After the mixture was
stirred at ambient temperature overnight, it was carefully poured into a
saturated aq NaHCO3, extracted with CH2Cl2 (3 × 20 mL), washed with brine, and dried over MgSO4. The
solvents were removed in vacuo. The obtained residue was dissolved in EtOAc and triturated with
pentane to give compound 4 as an off-white solid (304 mg, 0.676 mmol, 67 %);
m.p. 89 °C; IR (ATR): ̃ = 3031(w), 2957(w), 2852(w), 1724(s), 1612(w), 1585(w), 1499(m), 1452(m),
1438(m), 1379(w), 1357(w), 1272(m), 1226(s), 1194(s), 1163(m), 1125(m), 1080(s), 1052(m), 1039(m),
1028(s), 994(m), 950(w), 929(m), 912(m), 844(m), 810(m), 784(m), 760(m), 728(s), 691(s), 672(m) cm-1;
1H NMR (600 MHz, CDCl3) δ 7.56 (d, 4JHH = 2.3 Hz, 1H, H-C6), 7.47 (m, 2H, o-PhO), 7.40 (m, 2H, o-PhN), 7.37
(m, 2H, m-PhO), 7.30 (m, 3H, p-PhO/m-PhN), 7.23 (m, 1H, p-PhN), 7.14 (dd, 3JHH = 8.7 Hz, 4JHH = 2.3 Hz, 1H,
H-C4), 6.92 (d, 3JHH = 8.7 Hz, 1H, H-C3), 5.17 (s, 2H, OCH2Ph), 4.94 (ddd, 2JFH = 47.6 Hz, 3JHH = 8.1, 3.9 Hz,
1H, H-C7), 3.96/3.69 (each d, 2JHH = 14.8 Hz, each 1H, NCH2Ph), 3.91 (s, 3H, OCH3), 3.15 (ddd, 3JFH =
15.9 Hz, 2JHH = 14.9 Hz, 3JHH = 8.1 Hz)/2.75 (ddd, 3JFH = 29.2 Hz, 2JHH = 14.9 Hz, 3JHH = 3.9 Hz)(each 1H,
H-C8), 1.13 (s, 9H, CH3tBu);
13C{1H} NMR (151 MHz, CDCl3) δ 166.6 (C9), 158.0 (d, 5JFC = 1.7 Hz, C2), 142.5 (br, i-PhN), 136.8 (i-PhO),
131.8 (d, 2JFC = 21.0 Hz, C5), 130.9 (d, 3JFC = 6.4 Hz, C4), 129.5 (d, 3JFC = 6.6 Hz, C6), 128.7 (m-PhO), 128.5
(br m, o-PhN), 128.3 (br, m-PhN), 127.9 (p-PhO), 126.9 (o-PhO), 126.7 (br, p-PhN), 120.5 (C1), 113.8 (C3),
93.7 (d, 1JFC = 173.0 Hz, C7), 70.8 (OCH2Ph), 57.8 (br d, 2JFC = 25.9 Hz, C8), 55.7 (NCH2Ph), 55.4 (br, CtBu),
52.2 (OCH3), 27.6 (CH3tBu);
19F NMR (564 MHz, CDCl3) δ -177.8 (ddd, 2JFH = 47.6 Hz, 3JFH = 29.2, 15.9 Hz);
+
HRMS (ESI), m/z calcd. for C28H33FNO3 450.2439 [M+H]+; found: 405.2453.
4.2. (5-(2-(benzyl(tert-butyl)amino)-1-fluoroethyl)-2-(benzyloxy)phenyl)methanol (1)

Optimized procedure: DIBAL-H (1.0 M in toluene, 4.35 mL, 4.35 mmol,
2.1 equiv.) was added dropwise to a solution of compound 4 (932 mg,
2.07 mmol, 1.0 equiv.) in dry THF (20 mL) at -78 °C. After the mixture was stirred
at -78 °C for 3 h, water was added at ambient temperature followed by
treatment with saturated aq Rochelle salt, extraction with Et2O (3 x 15 mL), washing with brine, and
drying over MgSO4. The solvents were removed in vacuo. Compound 1 was obtained as a brown-orange
oil (879 mg, 70% pure from 1H NMR, calculated yield: 725 mg, 1.72 mmol, 83 % [from 19F NMR using
ethyl fluoroacetate as internal standard]);
IR (ATR): ̃ 3402(br), 3030(w), 2966(w), 2925(w), 2869(w), 1729(w), 1611(w), 1496(m), 1454(m),
1379(w), 1362(w), 1247(m), 1299(m), 1160(w), 1124(m), 1080(w), 1041(m), 1016(m), 950(w), 899(w),
844(w), 810 (m), 731 (s), 695(s) cm-1;
1H NMR (599 MHz, CDCl3) δ 7.42 (m, 2H, o-PhN), 7.41–7.39 (m, 4H, o,m-PhO), n.o. (p-PhO), 7.31 (m, 2H,
m-PhN), 7.23 (m, 1H, p-PhN), 7.04 (d, 4JHH = 2.3 Hz, 1H, H-C6), 6.99 (dd, 3JHH = 8.4 Hz, 4JHH = 2.3 Hz, 1H,
H-C4), 6.86 (d, 3JHH = 8.4 Hz, 1H, H-C3), 5.10 (s, 2H, OCH2Ph), 4.98 (ddd, 2JFH = 47.8 Hz, 3JHH = 8.5, 3.3 Hz,
1H, H-C7), 4.70/4.68 (each d, 2JHH = 13.0 Hz, each 1H, H-C9), 4.00/3.67 (each d, 2JHH = 15.0 Hz, each 1H,
NCH2Ph), 3.18 (ddd, 3JFH = 16.5 Hz, 2JHH = 15.0, 3JHH = 8.5 Hz)/2.72 (ddd, 3JFH = 31.4 Hz, 2JHH = 15.0 Hz, 3JHH
=
3.3 Hz)(each 1H, H-C8), 1.15 (s, 9H, CH3tBu);
13C{1H} NMR (151 MHz, CDCl3) δ 156.5 (d, 5JFC = 1.9 Hz, C2), 142.8 (i-PhN), 136.7 (i-PhO), 132.1 (d, 2JFC =
20.4 Hz, C5), 129.5 (C1), 128.8 (m-PhO), 128.4 (o-PhN), n.o. (p-PhO), 128.3 (m-PhN), 127.4 (o-PhO), 126.6
(p-PhN), 126.5 (d, 3JFC = 6.7 Hz, C4), 126.4 (d, 3JFC = 6.2 Hz, C6), 111.4 (C3), 94.3 (d, 1JFC = 172.7 Hz, C7),
70.3 (OCH2Ph), 62.1 (C9), 58.0 (d, 2JFC = 25.6 Hz, C8), 55.6 (NCH2Ph), 55.3 (CtBu), 27.6 (CH3tBu);
19F NMR (564 MHz, CDCl3) δ -176.8 (ddd, 2JFH = 47.8 Hz, 3JFH = 31.4, 16.5 Hz);
+
HRMS (ESI), m/z calcd. for C27H33FNO2 422.2490 [M+H]+; found: 422.2487.
Notes
The authors declare no competing financial interest.
Acknowledgements
We acknowledge generous financial support from the WWU Münster, the Deutsche
Forschungsgemeinschaft (SFB 858, and Excellence Cluster EXC 1003 “Cells in Motion—Cluster of
Excellence), and the Studienstiftung des deutschen Volkes (scholarship to Charlotte S. Teschers). We are
grateful to Dr. Klaus Bergander (WWU Münster) for assistance with NMR spectroscopy.

References
[1] M. Cazzola, C.P. Page, L. Calzetta, M.G. Matera, Pharmacol. Rev. 64 (2012) 450-504.
[2] The selection and use of essential medicines: report of the WHO Expert Committee, 2007:(including
the 15th model list of essential medicines), World Health Organization, 2007.
[3] L.L.H. Charles, T. Paul, Phenylaminoethanol Derivatives, 1971, US Patent App. 3705233
[4] R.H. Uloth, J.R. Kirk, W.A. Gould, A. Larsen, J. Med. Chem. 9 (1966) 88-97.
[5] E. Palmer, Top 20 generic molecules worldwide. http://www.fiercepharma.com/special-report/top-
20-generic-molecules-worldwide. (Accessed 18/09/2017).
[6] J. Beier, H. Pujol, B. Seoane, E. Jimenez, C. Astbury, E. Massana, S. Ruiz, G. de Miquel, BMC Pulm.
Med. 16 (2016) 102.
[7] F.M. Rackemann, J. Am. Med. Assoc. 114 (1940) 1998-2002.
[8] J.P. Beale, C.T. Grainger, Cryst. Struct. Commun. 1 (1972) 71.
[9] S. Wolfe, Acc. Chem. Res. 5 (1972) 102-111.
[10] D. O'Hagan, Chem. Soc. Rev. 37 (2008) 308-319.
[11] L. Hunter, Beilstein J. Org. Chem. 6 (2010).
[12] L.E. Zimmer, C. Sparr, R. Gilmour, Angew. Chem. Int. Ed. 50 (2011) 11860-11871.
[13] C. Thiehoff, Y.P. Rey, R. Gilmour, Isr. J. Chem. 57 (2017) 92-100.
[14] F. Scheidt, P. Selter, N. Santschi, M.C. Holland, D.V. Dudenko, C. Daniliuc, C. Mück‐Lichtenfeld, M.R.
Hansen, R. Gilmour, Chem. Eur. J. 23 (2017) 6142-6149.
[15] C. Sparr, W.B. Schweizer, H.M. Senn, R. Gilmour, Angew. Chem. Int. Ed. 48 (2009) 3065-3068.
[16] C. Sparr, R. Gilmour, Angew. Chem. 122 (2010) 6670-6673.
[17] E.-M. Tanzer, L.E. Zimmer, W.B. Schweizer, R. Gilmour, Chem. Eur. J. 18 (2012) 11334-11342.
[18] I.G. Molnár, E.M. Tanzer, C. Daniliuc, R. Gilmour, Chem. Eur. J. 20 (2014) 794-800.
[19] I. Sagamanova, C. Rodríguez-Escrich, I.G. Molnár, S. Sayalero, R. Gilmour, M.A. Pericàs, ACS Catal. 5
(2015) 6241-6248.
[20] I.G. Molnar, M.C. Holland, C. Daniliuc, K. Houk, R. Gilmour, Synlett 27 (2016) 1051-1055.
[21] S. Purser, P.R. Moore, S. Swallow, V. Gouverneur, Chem. Soc. Rev. 37 (2008) 320-330.
[22] I. Ojima, J. Org. Chem. 78 (2013) 6358-6383.
[23] E.P. Gillis, K.J. Eastman, M.D. Hill, D.J. Donnelly, N.A. Meanwell, J. Med. Chem. 58 (2015) 8315-8359.
[24] T. Schaefer, R.W. Schurko, R. Sebastian, F.E. Hruska, Can. J. Chem. 73 (1995) 816-825.

[25] J.M. Silla, D.R. Silva, M.P. Freitas, Mol. Inf. 36 (2017) 1700084
[26] G. Deniau, A.M. Slawin, T. Lebl, F. Chorki, J.P. Issberner, T. van Mourik, D. O’Hagan, ChemBioChem 8
(2007) 2265-2274
[27] M.D. Clift, G.P. Deniau, D. O’Hagan, R.B. Silverman, Biochemistry 46 (2007) 13819-13828
[28] F.M. de Rezende, L.A. Andrade, M.P. Freitas, J. Mol. Struct. 1093 (2015) 8-12
[29] V. Katritch, K.A. Reynolds, V. Cherezov, M.A. Hanson, C.B. Roth, M. Yeager, R. Abagyan, J. Mol.
Recognit. 22 (2009) 307-318
[30] P.L. Freddolino, M.Y.S. Kalani, N. Vaidehi, W.B. Floriano, S.E. Hall, R.J. Trabanino, V.W.T. Kam, W.A.
Goddard, Proc. Natl. Acad. Sci. U S A 101 (2004), 2736-2741
[31] J.A. Howard, V.J. Hoy, D. O'Hagan, G.T. Smith, Tetrahedron 52 (1996) 12613-12622.
[32] J.D. Dunitz, R. Taylor, Chem. Eur. J. 3 (1997) 89-98.
[33] S. Paul, W.B. Schweizer, G. Rugg, H.M. Senn, R. Gilmour, Tetrahedron 69 (2013) 5647-5659.
[34] C. Thiehoff, L. Schifferer, C.G. Daniliuc, N. Santschi, R. Gilmour, J. Fluor. Chem. 182 (2016) 121-126.
[35] C. Thiehoff, M. Holland, C. Daniliuc, K. Houk, R. Gilmour, Chem. Sci. 6 (2015) 3565-3571.
[36] D.Y. Buissonneaud, T. van Mourik, D. O'Hagan, Tetrahedron 66 (2010) 2196-2202.
[37] D.T. Collin, D. Hartley, D. Jack, L.H. Lunts, J. Press, A.C. Ritchie, P. Toon, J. Med. Chem. 13 (1970) 674-
680.
[38] T. Jen, J.S. Frazee, C. Kaiser, D.F. Colella, J.R. Wardell Jr, J. Med. Chem. 20 (1977) 1029-1035.
[39] M. Mammen, S. Dunham, Compounds having beta2 adrenergic receptor agonist and muscarinic
receptor antagonist activity, 2005, US20050113417 A1
[40] T. Forrest, Org. Magn. Reson. 6 (1974) 355-357.
[41] S. Hamman, C. Béguin, C. Charlon, C. Luu‐Duc, Magn. Reson. Chem. 21 (1983) 361-366.
[42] A. Dahbi, S. Hamman, C.G. Béguin, Magn. Reson. Chem. 24 (1986) 337-342.
[43] L.A. Donders, F.A. De Leeuw, C. Altona, Magn. Reson. Chem. 27 (1989) 556-563.
[44] C. Altona, J.H. Ippel, A.J. Hoekzema, C. Erkelens, M. Groesbeek, L.A. Donders, Magn. Reson. Chem.
27 (1989) 564-576.
[45] J.R. Cavanaugh, B. Dailey, J. Chem. Phys. 34 (1961) 1099-1107.
[46] L.A. Sharp, S.Z. Zard, Org. Lett. 8 (2006) 831-834.
[47] B. Wang, H.-X. Sun, Z.-H. Sun, G.-Q. Lin, Adv. Synth. Catal. 351 (2009) 415-422.
[48] H.C. Brown, J.V. Kanth, P.V. Dalvi, M. Zaidlewicz, J. Org. Chem. 64 (1999) 6263-6274.
[49] M. Karplus, J. Chem. Phys. 30 (1959) 11-15.

Figure 1: Selected 2-AR agonists. The conserved phenolethanolamine core is highlighted in blue.[1, 5-7]
Figure 2: Molecular editing approach to obtain a conformational mimic of the top-selling drug
salbutamol.[5]
Figure 3: The fluorine-nitrogen gauche effect in salbutamol derivative 1 restricts the rotational freedom
around the C(sp3)-C(sp3) bond in favour of the gauche-2 conformer. The -system might align in plane
with the antibonding C-F* orbital.

Figure 5: The model compounds used to calculate data for the solution phase conformer population
analysis.
Figure 6: Single X-ray diffraction analysis of compound 4 (CCDC 1581462). a) Thermal ellipsoids are
shown with 30 % probability. b) The overlay of the crystal structures of derivative 4 (grey) and
salbutamol (black) [8] demonstrates a closely similar conformation of the phenolethanolamine core in
both molecules. Atoms chosen for the overlay are C7 and C8.

Scheme 1: Synthesis of fluorinated salbutamol derivative 1. Conditions: a) BnBr, K2CO3, acetonitrile,
75 °C, 16 h, quantitative; b) Br2, CHCl3, 0 °C, 5 h, 97%; c) N-benzyl-tert-butylamine, acetonitrile, 85 °C, 6 h,
60%; d) sodium borohydride, MeOH, rt, 16 h, quantitative; e) (diethylamino)sulfur trifluoride,
triethylamine trihydro-fluoride, CH2Cl2, rt, 16 h, 67%; f) diisobutylaluminium hydride, THF, -78 °C to rt,
16 h, 83%.

Table 1: Experimental 3J coupling constants of compounds 1, 4 and 3 and calculated 3J coupling constants
of their respective model compounds model-1, model-4 and model-3.
compound
entry
experimental 3J coupling constants
3J(HH1)
8.5 Hz
8.1 Hz
9.9 Hz
3J(HH2)
3.3 Hz
3.9 Hz
4.2 Hz
3J(FH1)
16.5 Hz a
15.9 Hz
--
3J(FH2)
31.4 Hz a
29.2 Hz
--
1
4
3
1
2
3
calculated 3J coupling constants
ø 3J(HH1)g
2.69 Hz
2.68 Hz
2.73 Hz
2.73 Hz
3.00 Hz
2.80 Hz
2.71 Hz
2.79 Hz
ø 3J(HH2)g
2.27 Hz
1.81 Hz
2.30 Hz
1.83 Hz
2.72 Hz
2.46 Hz
2.30 Hz
2.00 Hz
3J(HH)t
12.9 Hz
12.4 Hz
12.9 Hz
12.5 Hz
13.2 Hz
13.0 Hz
12.8 Hz
12.6 Hz
3J(FH)g
7.33 Hz
6.31 Hz
7.41 Hz
6.38 Hz
--
3J(FH)t
44.8 Hz
40.7 Hz
45.2Hz
40.9 Hz
--
4[a,c]
model-1
model-4
5[b,c]
6[a,d]
7[b,d]
8[a,d,e]
9[a,d,f]
10[b,d,e]
11[b,d,f]
--
--
model-3
--
--
[c]
t = trans, g = gauche, [a] E(NR2) = 0.56, [b] E(NR2) = 1.12,
E(R) = 0.52 , [d] E(R) = 0.49 , [e] E(OH) =
1.30, [f] E(OH) = 1.56

Table 2: Calculated solution phase conformer population distributions of compounds 1, 4 and 3 at
ambient temperature.
compound
entry
1[b,d]
3J(XH) [a]
3J(HH)
3J(FH)
3J(HH)
3J(FH)
3J(HH)
3J(FH)
3J(HH)
3J(FH)
3J(HH)
3J(HH)
3J(HH)
3J(HH)
P(gauche-1)
33 %
24 %
26 %
30 %
32 %
23 %
25 %
28 %
18 %
14 %
10 %
6 %
P(gauche-2)
57 %
P(anti)
10 %
2[b,d]
64 %
11 % [h]
14 %
1
3[c,d]
60 %
4[c,d]
73 %
-3 %[h]
15 %
5[b,e]
53 %
6[b,e]
58 %
20 % [h]
20 %
4
3
7[c,e]
55 %
8[c,e]
66 %
6 %
9[b,d,f]
10[b,e,g]
11[c,d,f]
12 [c,e,g]
68 %
14 %
70 %
17 % [h]
18 % [h]
21 %
71 %
73 %
[a] coupling constants [3J(HH) or 3J(FH)] used for the calculation of population distributions, [b] E(NR2) =
[d]
0.56, [c] E(NR2) = 1.12,
E(R) = 0.52 , [e] E(R) = 0.49 , [f] E(OH) = 1.30, [g] E(OH) = 1.56, [h] negative
∑
values or P ≠ 100 % can result from indirect determination of the population density.