Direct, nucleophilic radiosynthesis of [18F]trifluoroalkyl tosylates: Improved labelling procedures

Article abstract of DOI:10.1039/c2ob25802a

A rapid and efficient protocol to afford the title compound 2-[ ¹⁸F]-fluoro-2,2-difluoroethyl tosylate ([¹⁸F]7b) is described. Starting from [¹⁸F]fluoride ion, labelling reagent 7b was obtained in good yields and a high specific radioactivity. Compound ([ ¹⁸F]7b) was then used to synthesise a prospective radiotracer for PET-imaging in dementia.

Full text of DOI:10.1039/c2ob25802a

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Organic &
Biomolecular
Chemistry
Cite this: Org. Biomol. Chem., 2012, 10, 6980
www.rsc.org/obc
PAPER
Direct, nucleophilic radiosynthesis of [18F]trifluoroalkyl tosylates: improved
labelling procedures
Patrick J. Riss,* Valentina Ferrari, Laurent Brichard, Paul Burke, Robert Smith and Franklin I. Aigbirhio
Received 26th April 2012, Accepted 27th June 2012
DOI: 10.1039/c2ob25802a
A rapid and efficient protocol to afford the title compound 2-[18F]-fluoro-2,2-difluoroethyl tosylate
([18F]7b) is described. Starting from [18F]fluoride ion, labelling reagent 7b was obtained in good yields
and a high specific radioactivity. Compound ([18F]7b) was then used to synthesise a prospective
radiotracer for PET-imaging in dementia.
batches of up to 370 GBq (103 human doses) as [18F]fluoride
Introduction
ion are realistically achievable on standard medical cyclotrons
from the 18O(p,n)18F nuclear reaction via irradiation of H218O
Quantitative imaging of radiotracer distribution with positron
emission tomography (PET) provides invaluable information
liquid targets. Using [18F]fluoride ion, high specific radioactivity,
about molecular transactions in living subjects. PET is based on
the detection of coincident 511 keV photons originating from the
positron decay process in proton rich radionuclides. This allows
for spatial localisation and quantification of the decaying radio-
nuclide in tissue.1
The true potential of PET, however, lies in the combination of
a positron-emitting radionuclide with a molecular entity that
participates in a biochemical process of interest. With the
increasing application of PET, spanning preclinical small animal
imaging, applications in drug development, biomedical research
and clinical diagnostic imaging, there is a strong demand for
novel radiotracers for a variety of biological targets. Conse-
quently, a wide portfolio of reliable chemical methodology is
required, allowing access to novel radiotracers suitable for
increasingly complex imaging studies.2
i.e. a high ratio of radioactive to non-radioactive molecules
(>150 GBq μmol−1) in the radiotracer formulation, is easily
feasible. This allows for PET-imaging of saturable biological
systems under genuine tracer conditions.1d Alternative pro-
duction routes for 18F as electrophilic fluorination agents exist
but their comprehensive adaptation into routine application has
been hampered by low specific radioactivity, inconvenient hand-
ling, and low production yields.1,4 As a result, nucleophilic fluor-
ination using [18F]fluoride ion is by far the most common and
important route for fluorine-18 labelling.1–4
The trifluoromethyl (CF3) group is a common motif in small
molecule drugs, cf. Fig. 1. Trifluoromethylation of drug scaffolds
is a common method of lead optimisation in drug development,
thus a constant feed of novel bioactive compounds containing
this function can be anticipated.5
18F is the most frequently employed PET nuclide3 providing a
unique combination of beneficial physical and chemical charac-
teristics; high positron yield from an almost exclusive decay via
the β+ decay branch (97%) is paired with a very low positron
energy (638 keV) limiting the linear range of the emitted posi-
tron to a few millimetres in aqueous tissue. This accounts for a
high image resolution compared to most other PET nuclides.
This profile is rounded by an expedient half-life (109.7 minutes)
rendering 18F suitable for multi-step reactions, transport of radio-
tracer formulations over moderate distances and convenient
handling of the tracer in imaging studies.1 Moreover, its ability
to readily form stable bonds with carbon atoms promotes the
straightforward minimally invasive introduction of F atoms into
most organic molecules. A key advantage is that production
The CF3 group is a bioisostere for iodo, bromo, carbonyl, car-
boxyl, sec-propyl and tert-butyl groups.5 With an electronegativ-
ity in the range between chlorine and fluorine, it is a strong
electron withdrawing group. Consequently, CF3 substitution is
commonly used to manipulate the acidity of neighbouring
groups. Notably, CF3-groups coordinate to two main group
metals of biological relevance.5b
However, despite the high abundance of CF3 groups in phar-
maceuticals and drug candidates and its high metabolic stability,
only a few approaches for 18F labelling of CF3 groups have been
reported. The radiosynthesis of [18F]CF3-groups has been further
complicated by the low reactivity of leaving groups in difluoro-
methylene precursors, elaborate multi-step synthesis, and low
specific activity in particular with carrier added [18F]CF3-
labelled radiotracers obtained via electrophilic fluorination.8 So
far, nucleophilic 18F-fluorination of CF3 groups has mainly been
achieved using harsh reaction conditions. This is a major limit-
ation in 18F radiochemistry, since the CF3 group is present in a
The Wolfson Brain Imaging Centre, Box 65 Addenbrooke’s Hospital,
CB2 0QQ Cambridge, United Kingdom. E-mail: pr340@wbic.cam.ac.uk;
Fax: +44 1223 331826; Tel: +44 1223 761394
6980 | Org. Biomol. Chem., 2012, 10, 6980–6986
This journal is © The Royal Society of Chemistry 2012

View Article Online
Results and discussion
To develop this method, substrate 7a was used for optimisation
of the labelling conditions, investigating the influence of the 18F-
fluoride source, reaction time, stoichiometry, temperature and
solvent on the radiochemical yields (RCY) as determined by
radio-HPLC and/or radio-TLC (Table 1). An alternative sub-
strate, 1,1-difluoroprop-1-en-2-yl tosylate (8a, Scheme 2), was
labelled with [K+ ⊂ K222][18F]F− under optimised reaction con-
ditions to obtain 1,1,1-[18F]trifluoroprop-2-yl tosylate as a label-
ling agent.
Synthesis of labelling precursors
Fig. 1 Examples of aliphatic trifluoromethyl groups in radiotracers,
current pharmaceuticals and prospective drugs in clinical trials. 1: [18F]-
TFMISO, a hypoxia PET radiotracer;6a 2: [11C]methyl lansoprazole;6b 3:
Flecainide, a cardiovascular sodium channel inhibitor for the treatment
of tachycardia;6c 4: LGD-2226, a prospective selective androgen recep-
tor mediator in clinical trials;6d 5: BAY 38-7271 a cannabinoid receptor
agonist in clinical trials;6e 6: Quazepam, an α1 selective benzodiazepine
derivative.6f
The desired difluorovinyl-functionalised labelling precursors are
readily accessible using established synthesis methods e.g. via
sulphur or phosphorus based ylide-reagents, respectively, or via
HF elimination reactions. In our case, the model compounds
were obtained in good yields using known transformations by
elimination and Horner–Wadsworth–Emmons olefination.7
Compounds 7a and 8a were obtained in good yields as
described previously.
large number of well-characterised drug candidates and pharma-
ceuticals (Fig. 1). Therefore, an efficient method to radiolabel
the CF3 group in high specific radioactivity would open up a
wide range of compounds for PET-imaging applications in drug
development and medical research. This will also avoid the issue
of having to modify leads that already contain ‘native’ C–F
bonds in CF3 groups by introducing further fluorine atoms in
radiotracer development.
Recently, we have reported a rapid, one-step method for
nucleophilic radiosynthesis of aliphatic [18F]CF3-groups in high
specific radioactivity under mild conditions.7 Incorporation of
the radiolabel is considered to be achieved via the equivalent
nucleophilic addition of H[18F]F to 2,2-difluorovinyl groups as
shown in Scheme 1.
Optimisation of the labelling reaction
We chose the direct reaction of 7a with potassium
4,7,13,16,21,24-hexaoxa-1,10-diazabicyclo[8.8.8]hexacosane
(K222) cryptate [18F]fluoride ion complex in DMSO at 130 °C
as a starting point for our investigations.
Our effort was then focused on the time course of the reaction,
and the time dependent outcome of the reaction from the [18F]
fluoride source was investigated (Table 1, entries 1 to 5).
Notably all of these can be readily obtained from the target water
Table 1 Optimisation of the reaction conditions for the formation of
[
18F]7b
In addition to direct labelling under mild conditions, second-
ary labelling agents such as [18F]7b provide a novel route for the
introduction of metabolically insensitive 18F-labels in heteroatom
bound prosthetic groups. In particular defluorination reactions
are known to occur to a lower extent than in aliphatic 18F
fluorides due to the extraordinarily high stability of C–F bonds
in trifluoromethyl groups. This is to the best advantage of brain
imaging studies, where uptake of [18F]fluoride ion into the skull
will seriously confound PET data. Apart from direct labelling
with [18F]F−, indirect labelling using prosthetic groups is often
the only way to label sensitive small molecules or peptides for
PET. The present study is concerned with synthesis and appli-
cation of novel [18F]trifluoroalkyl prosthetic groups for PET
chemistry.
Entry M[18F]F
Solvent T/°C Time/min RCYa/%
1
2
3
4
5
6
7
8
K[K222][18F]F
DMSO 130
DMSO 130
DMSO 130
DMSO 130
6
6
6
6
6
3
3
3
3
3
3
3
61.7 0.4
60.3 0.6
54.2 0.5
58.9 2.2
47.6 0.4
77.9 3.0
91.0 1.0
91.9 1.4
24.8 3.9
54.4 2.9
63.4 2.6
77.2 0.9
Cs[18F]F
TBA[18F]F
Ag[18F]F
K[K222][18F]F + CuOTf DMSO 130
K[K222][18F]F
K[K222][18F]F
K[K222][18F]F
K[K222][18F]F
K[K222][18F]F
K[K222][18F]F
K[K222][18F]F
DMSO 110
DMSO 90
DMSO 85
DMSO 70
9
10
11
12
DMF
MeCN
THF
90
82
66
a n = 3.
Scheme 1 Addition of H[18F]F.
Scheme 2 Synthesis of labelling precursors 7a–8a.
Org. Biomol. Chem., 2012, 10, 6980–6986 | 6981
This journal is © The Royal Society of Chemistry 2012

View Article Online
by solid phase extraction (SPE) in a straightforward sequence on
a QMA cartridge, elution in a small volume, concentration and
azeotropic removal of excess water. This is highly advantageous
for general application of 18F. Currently being the established
basis for 18F-chemistry throughout the world, this process has
been automated on a variety of commercial systems.
130 °C. Prolonged reaction times gave lower RCY of [18F]7b,
due to an increase in the formation of 18F-labelled precursor 7a.
Radiochemical yields of [18F]7b were found to improve with
decreasing reaction temperature, which was clearly attributable
to suppression of a side reaction leading to an unidentified, more
polar radioactive product. Highest yields were achieved at 85 °C
after a reaction time of 3–6 minutes. Further reduction of the
reaction temperature to 70 °C gave unsatisfyingly low yields.
Since rapid chromatographic separation of unlabelled precur-
sor is a key component of radiotracer production processes, and
low precursor concentrations generally result in less demanding
conditions for separation, we investigated the concentration
dependent RCY as a function of reaction time. Highest radio-
chemical yields were obtained using 5 0.5 mg of precursor 7a
per millilitre of DMSO, equivalent to a final concentration of
about 2 mM. Reduction of the concentration of precursor to
1 mM still resulted in a moderate to good labelling yield. Even
using a final concentration of 0.4 mM of 7a in DMSO, [18F]7b
was obtained in sufficient yield to facilitate production of radio-
tracers in clinically relevant doses, although at the expense of
high variances in the RCY (Fig. 4).
Moderate to good yields were obtained for all [18F]fluoride
sources with no clear preference. It was therefore decided to
optimise the conditions using potassium 4,7,13,16,21,24-
hexaoxa-1,10-diazabicyclo[8.8.8]hexacosane (crypt-222) cryp-
tate [18F]fluoride ion complex, the most commonly used method
for [18F]fluoride ion, which would then enable a more straight-
forward transfer of the methodology into general application.
In order to determine the best solvent for optimal performance
of the radiolabelling step, the dipolar aprotic solvents aceto-
nitrile, DMF, DMSO, and THF were investigated as reaction
media. However, substitution of DMSO for either one of the
alternative solvents did not have a beneficial effect (Fig. 2). In
the case of THF, degradation of the labelled product was acceler-
ated. For this reason, DMSO was the reaction medium of choice
in all further experiments.
In the next step the time dependent radiochemical yield was
investigated as a function of reaction time at different temp-
eratures in DMSO (Fig. 3).
As noted earlier, 18F-labelled “precursor” [18F] 7a was
observed as a by-product of the reaction in some cases (Table 1
entries 1–6, 10–12). The most likely mechanism for its for-
mation is by an addition–elimination mechanism (Scheme 3),
rather than nucleophilic substitution of 19F for 18F.
[18F]fluoride was rapidly incorporated into the labelling pre-
cursor and high yields were achieved within 3 minutes at
Since this mechanism could also result in an undesired
dilution of the final radioactive product with unlabelled 7b due
to the incorporation of 19F− into 7a, we determined the specific
radioactivity (AS) of [18F]7a obtained using the conditions
described in Table 1 entry 8. For this experiment, a batch of
approximately 5 GBq [18F]fluoride ion was used for the radiosynth-
esis. The AS of [18F]7a was determined to be 86 MBq nmol−1,
which is in good accordance with the typical values obtained in
Fig. 2 Radiochemical yield as a function of the solvent.
Fig. 4 Plot of radiochemical yield as a function of labelling precursor
concentration.
Fig. 3 Radiochemical yield as a function of temperature.
6982 | Org. Biomol. Chem., 2012, 10, 6980–6986
Scheme 3 Addition–elimination mechanism leading to [18F]7a.
This journal is © The Royal Society of Chemistry 2012

View Article Online
low activity 18F-production batches. It can therefore be con-
cluded that there is no significant deteriorating effect by equili-
bration of trifluoro-carbanion and difluorovinyl species. On the
other hand, addition of stoichiometric amounts of potassium
fluoride to the reaction mixture led to a significant reduction in
the reaction rate.
good yields when the intermediate carbanion is rapidly proto-
nated, i.e. by trace amounts of water in the reaction mixture.
Conversely, at very low water concentrations, an addition–
elimination mechanism as outlined in Scheme 3 will gradually
lead to increasing amounts of [18F]7a being formed.
Whereas too high concentration of water leads to the loss of
reactivity of the n.c.a. [18F]fluoride ion under our optimised reac-
tion conditions (Table 1), excellent radiochemical yields were
obtained for [18F]7b. However, fine-tuning of the water content
might prove difficult in practical application and impede wide-
spread adaptation of the protocol. Hence, we investigated
alternative additives that would suppress unwanted side-reactions
without affecting the reactivity of 18F−, avoiding the necessity of
careful adjustment of their concentration. We surmised that an
acidic compound with an appropriate pKa value would be appro-
priate, simply by protonating the carbanion intermediate without
affecting the reactivity of 18F−.
Potassium fluoride cryptate and crown ether complexes are
known to provide sufficient basicity to deprotonate nitroalkanes
and carbonyl compounds in dipolar aprotic solvents and facilitate
controlled reactions in condensation reactions. The pKa value of
water in DMSO (31.4) is in the same range as the pKa values of
low molecular weight alcohols (Table 2).11a Moreover, alcohols
and diols, in particular secondary and tertiary alkyl alcohols, are
known to promote reactivity when used in conjunction with
alkali metal fluorides of low solubility in aprotic solvents.2e,11b,c
Considering these facts, we surmised that alcohols would be
tolerated in much higher, hence easily controllable concen-
trations. In contrast, water interferes with the reaction at higher
concentration. We therefore focused our effort on low molecular
weight alcohols as additives for the reaction mixture, which have
been shown to allow for 18F-labelling. In order to investigate the
contribution of the pKa of our additives onto the outcome of the
reaction, compounds spanning the pKa range from 11 to 32 were
included at a concentration of 1 M. Based on our earlier
findings, DMSO was used as a solvent and [K+ ⊂ K222][18F]F−
was used as source of activated 18F−.
In order to elucidate potential 18F–19F isotopic exchange with
fluorine atoms in the product, a control experiment was con-
ducted. Non-radioactive 7b was reacted with [18F]fluoride ion
for 15 minutes and the formation of [18F]7b was monitored.
Only trace amounts of about 1 percent were formed after
5 minutes and even after 15 minutes at 110 °C only minor
amounts of [18F]7b were found (<3%). These findings indicate a
low susceptibility of aliphatic CF3 groups to isotopic exchange
under our reaction conditions.
Due to the nanomolar quantities of [18F]fluoride ion in the
reaction mixture, normally traces of water would strongly affect
the labelling efficiency due to solvation of the fluoride ion.
Nevertheless, in our assumption, a small amount of water would
benefit ‘trapping’ of the anionic intermediate in the course of the
reaction by protonation, thus avoiding excessive equilibration of
18F and 19F addition–elimination products. A high variance
between different batches of dried fluoride and between different
DMSO qualities gave rise to the assumption that the residual
water content might play a key role in the observed yields. For
this reason, we investigated the dependency of the radiochemical
yield of [18F]7a on the water content of the reaction medium
(Fig. 5).
When anhydrous DMSO8 was used, a significant amount of
labelled precursor 7a was formed at 130 °C. In contrast, the ratio
between [18F]7b and [18F]7a gradually improved in favour of the
desired product [18F]7b when the water content in DMSO was
increased.
Maximum yield was obtained with about 5 ppm of water
being added to the reaction mixture with only trace amounts of
[18F]7a being formed. As expected, increasing hydration then
impedes the formation of both products at higher water concen-
trations with the [18F]7b being the almost exclusive product at
10 ppm. Only traces of [18F]7a were formed at 100 ppm and no
labelled product was found at 1000 ppm.8 We reasoned that the
no-carrier-added (n.c.a.) reaction proceeds via an intermediate
equilibrium where the final labelling product is only obtained in
Using a 1 M concentration of malononitrile and nitromethane
did not furnish the desired product in considerable yields. Higher
RCY in the range of 20% was found when indole and benzothia-
zole were used as proton donors, however, the ratio between
Table 2 Influence of protic additives on labelling outcome
Time/
min
[
[
18F]7b : RCY
18F]7a
(product)/%
11
Entry Additive
pKa
1
2
3
4
5
6
7
8
Isoamyl alcohol
2-Propanol
2-Methyl-2-propanol 32.2
n.a.
30.3
5
5
5
5
5
5
5
5
6 : 1
10 : 1
4 : 1
5 : 1
2 : 1
3 : 2
1 : 1
Trace
7 : 1
10 : 1
5 : 1
6 : 1
10 : 1
n.a.
62.2 3.8
66.9 5.8
56.5 7.95
45.6 1.4
22.7 7.35
18.6 0.7
3.7 0.5
Ethanol
Benzothiazole
Indole
Nitromethane
Malononitrile
Isoamyl alcohol
2-Propanol
29.8
27.0
21.0
17.2
11.1
n.a. 10
30.3
Trace
9
62.3 4.2
72.5 8.3
55.75 2.0
50.35 3.9
75.0 11.0
Trace
10
11
12
13
14
15
10
10
10
5
5
5
2-Methyl-2-propanol 32.2
Ethanol
29.8
n.a.
30.3
31.4
DMSO (neat)
2-Propanol (neat)
2-Methylaniline
Fig. 5 Dependence of the radiochemical yield on the water content.
X-axis denotes amounts of water being added to the solvent.
3 : 1
48.5 2.2
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 6980–6986 | 6983

View Article Online
Table 3 Two step labelling yields
Compound Precursor Product
RCY/%
9a
4-CNC6H4OH 4-CNC6H4OCH2CF2[18F]F (9b) 93
2-ClC6H4CO2H 2-ClC6H4CO2CH2CF2[18F]F (10b) 82
HN(CH2C6H5) (C6H5CH2)2CH2CF2[18F]F (11b) 77
10a
11a
12a
13a
(C6H5)CSH
(C6H5)CSCF2[18F]F (12b)
86
91
[
18F]13b
Scheme 4 Radiotracer synthesis.
[18F]7b and [18F]7a was not satisfactory (Table 2). More striking
results were observed when alcohols were used as reaction
media (Table 2 entries 1–4). Their use afforded the desired 18F-
labelled product [18F]7b in much better yields of up to 67%
after 5 minutes, combined with a very good ratio between 18F-
labelled 7b and 7a. Prolonged reaction times did not have a sig-
nificant effect on the reaction outcome. (Table 2 entries 9–12).
Since highest yields were found using 1 M 2-propanol in DMSO
and 2.5 mg of precursor in 500 ml, we briefly investigated the
use of only 2-propanol. However, only traces of [18F]7b were
obtained in this case.
Inc., Washington DC, USA) dual BGO metabolite detector
system with Flow-Count B-FC-4000 analogue/digital interface
and a Bioscan 1′′ NaI(Tl) detector with Flow-Count B-FC-4000
analogue/digital interface were used for radioactivity detection.
Lablogic Laura 3 and Laura 4 software (Lablogic Systems Ltd,
Sheffield, UK) was used for data acquisition and evaluation. For
screening of reaction conditions, a Chromolith RP18e (5 μm)
0.4 mm × 100 mm column (Merck KGaA, Darmstadt, Germany)
at a flow rate of 2 ml min−1 (7 mM NH4OH–acetonitrile gradi-
ent), a Phenomenex Primesphere RP-18 (5 μm) 0.46 × 250 mm
column at a flow rate of 1 ml min−1 (70–80% MeCN in 0.05 M
ammonium formate pH 6.8), a Phenomenex Gemini RP-18
(5 μm) 0.46 × 250 mm column at a flow rate of 1 ml min−1
(70–80% MeCN in 0.05 M ammonium formate pH 6.8), and a
Chromolith RP18e (5 μm) 0.4 mm × 100 mm column (Merck
KGaA, Darmstadt, Germany) at a flow rate of 4 ml min−1
(water–acetonitrile, 67 : 33) were used as stationary phase. A GE
Healthcare BAS-IP MS storage phosphor screen 35 cm × 43 cm
was used for radioTLC (Fisher Scientific UK Ltd, Lough-
borough, UK). Detection and evaluation were performed using a
Duerr CR 35 NDT (raytest Isotopenmessgeraete GmbH, Strau-
benhardt, Germany) and raytest AIDA QWBA software. NMR
spectra were recorded on BrukerAvance III 400 QNP Ultrashield
Plus Cryo or a BrukerBrukerAvance 500 CryoUltrashield
(Bruker UK Ltd, Coventry, UK). Chemical shifts are reported
downfield from TMS, relative to the solvent residual signal.
Melting points were determined using a Kofler melting point
apparatus. Low resolution mass spectrometry was conducted
using a Bruker Esquire (Bruker UK Ltd, Coventry, UK) electron
spray ion source and detector. Accurate, high resolution mass
spectra were recorded on an Orbitrap spectrometer using electron
spray ionisation. Flash chromatography was conducted using a
Gilson PLC 2020 chromatography system and normal phase
silica gel cartridges.
Using the optimised conditions (Table 2 entry 2), [18F]7b was
obtained in an AS of 56 MBq nmol−1, 5 hours from end-of-bom-
bardment (EOB) for a production scale irradiation.12 This would
correspond to an AS of approximately 200 MBq nmol−1 for a
fully automated process of about 90 minutes duration, indicating
the AS has not been diluted under these conditions.
Compound [18F]7b as labelling reagent
As a prosthetic group, the 2-[18F]fluoro-2,2-difluoroethyl group
provides great potential as a metabolically insensitive, readily
available supplement to PET chemistry. With n.c.a. [18F]7b in
hand we briefly examined the reactivity of this alkylating agent
towards different nucleophiles. For this reason, [18F]7b was
reacted with N, O and S nucleophiles 9a–12a in DMF using
Cs2CO3 as base. The results are summarised in Table 3. Moder-
ate to good yields were obtained within 10 minutes for 9b–12b.
[18F]7b was also used to synthesise radiotracer candidate 13b as
a proof of concept for radiotracer synthesis.10 The results pres-
ented in Table 3 clearly show that [18F]7b is well suited as a lab-
elling agent.
Compound 13b is a potential imaging agent for neurofibrillar
tangles formed by hyperphosphorylated human τ-protein in
several forms of dementia, for example Alzheimer’s disease
(Scheme 4).
General procedure for radiolabelling
Experimental section
[18F]Fluoride ion was produced using the 18O(p,n)18F nuclear
reaction via proton bombardment (16 → 3 MeV) of an H218O
liquid target on a GE PETtrace cyclotron at a beam current of
30–40 μA for 5 to 10 minutes. The radionuclide was extracted
from the enriched target water via solid phase extraction on a
Waters Accell plus light QMA strong anion exchanger cartridge
(CO22−-form). Reactive [18F]F− was obtained by elution of the
trapped radioactivity using a mixture of appropriate bases
(20 μmol) in acetonitrile (300 μl) and water (300 μl). Six ali-
quots of the eluate (100 ml) were transferred to 5 ml conical
All solvents and reagents were obtained from Alfa Aesar (Alfa
Aesar UK Ltd, Heysham, Morecambe, UK), Sigma-Aldrich
(Sigma-Aldrich Co. Ltd, Poole, UK) and Fisher Scientific
(Fisher Scientific UK Ltd, Loughborough, UK). Solid phase
extraction cartridges were obtained from Waters (Waters Ltd,
Elstree, UK). Analytical HPLC was performed on an Agilent
1100 series HPLC system (Agilent Technologies UK Ltd,
Wokingham, UK), consisting of a G1312 A binary pump and a
G1314 variable wavelength UV-detector. A Bioscan (Bioscan
6984 | Org. Biomol. Chem., 2012, 10, 6980–6986
This journal is © The Royal Society of Chemistry 2012

View Article Online
bottom reaction tubes and the mixtures were concentrated in a
stream of nitrogen. Remaining free water was removed by
azeotropic co-evaporation with 3 portions of anhydrous aceto-
nitrile (3 × 1 ml). Labelling precursor dissolved in the appropri-
ate solvent was added to the residue and heated to the desired
temperature. Aliquots were withdrawn from the reaction mixture
(100 ml) at multiple time points and transferred into water
(0.5 ml). The resultant sample was directly injected into
radioHPLC or used for radioTLC (1 ml, CHCl3).
1H NMR (400 MHz, CDCl3) δ (ppm): 1.91 (t, J = 3 Hz, 3H,
CH3), 2.46 (s, 3H, ArCH3), 7.31 (d, J = 8.5 Hz, 2H, ArH), 7.81
(d, J = 9 Hz, 2H, ArH). 13C NMR (100 MHz, CDCl3) δ (ppm):
13.1, 21.7, 109.6 (dd, JCF = 15.5 Hz, JCF = 49.5 Hz), 128.3,
129.9, 132.5, 145.7. 19F NMR (376 MHz, CDCl3) δ (ppm):
−108.5 (d, J = 55.6 Hz), −94.99 (d, J = 55.6 Hz). MS (ESI) =
247.0, C10H10F2O3S requires: 247.0235. HRMS C10H9F2O3S
requires: 247.0235, found: 247.0241; C10H10F2O3S requires C,
48.38; H, 4.06; F, 15.31; O, 19.33; S, 12.92, found C, 48.41; H,
3.93%.
2,2,2-Trifluoroethyl 4-methylbenzenesulfonate (7b).1 2,2,2-
Trifluoroethanol (10g, 100 mmol) and triethylamine (14.3 g,
140 mmol) are dissolved in anhydrous diethyl ether (120 ml).
Toluenesulfonyl chloride (17.2 g, 90 mmol) was added in por-
tions at room temperature. The mixture was stirred for approxi-
mately 48 hours, until the toluenesulfonyl chloride had been
consumed. The solids were filtered off and the filter cake is
washed with diethyl ether (2 × 30 ml). The filtrate was concen-
trated in vacuo and the residue was purified by flash chromato-
graphy on silica gel (hexane–diethyl ether; 9 : 1). Product 6a was
Procedure for screening of the proton source
Stock solutions of additive (10 mmol) and 7a (50 mg,
0.2 mmol) in DMSO (10 ml) were prepared. Aliquots (500 ml)
of these solutions were added to [K+ ⊂ K222][18F]F− residues in
2 ml Wheaton reactivials at the desired reaction temperature.
Samples were withdrawn at the desired time point, diluted with
water and analysed by radio-HPLC.
1-(4-Fluorobenzyl)-N-(1-(4-(2,2,2-trifluoroethoxy)phenethyl)-
piperidin-4-yl)-1H-benzo[d]imidazol-2-amine (13b). 1H NMR
(400 MHz, CDCl3) δ (ppm): 1.38 (q, J = 10 Hz, 2H), 1.85 (t,
J = 11 Hz, 2H), 2.02 (d, J = 11 Hz, 2 H), 2.37 (t, J = 8 Hz, 1H),
2.39 (d, J = 6 Hz, 1H), 2.62 (d, J = 6 Hz, 1 H), 2.64 (t, J =
8 Hz, 1H), 2.79 (d, J = 11 Hz, 2 H), 3.79–3.95 (m, 2 H), 5.03 (s,
2H), 6.82 (d, J = 9 Hz, 2H), 6.94–7.14 (m, 8 H), 7.52 (d, J =
8 Hz, 2 H). 13C NMR (100 MHz, CDCl3) δ (ppm): 163.7,
161.3, 155.7, 153.1, 141.5, 134.2, 130.8, 130.7, 130.6, 129.6,
129.5, 128.3, 128.2, 121.9, 120.2, 116.4, 116.3, 116.1, 113.9,
107.3, 61.4, 52.2, 50.0, 45.1, 32.7, 32.5, 30.9. 19F NMR
(376 MHz, CDCl3) δ (ppm): −77.7, −113.6. MS (ESI) = 526.2,
C29H30F4N4O requires 526.2356, HRMS C29H30F4N4O requires
526.2356, found: 527.2360 [M + H], C29H30F4N4O requires C,
66.15; H, 5.74; F, 14.43; N, 10.64; found C, 66.09; H, 5.72, N,
10.86%.
1
obtained as colourless crystals in 89% (20.3 g) yield. MP = H
NMR (400 MHz, CDCl3) δ (ppm): 2.46 (s, 3H, ArCH3), 4.38
(q, J = 8 Hz, 2 H, CH2), 7.38 (d, J = 9 Hz, 2 H, ArH), 7.81 (d,
J = 9 Hz, 2 H, ArH). 13C NMR (100 MHz, CDCl3) δ (ppm):
21.7, 64.5 (q, JCF = 37.8 Hz, 120.5, 123.2, 128.1, 130.1, 131.8,
145.9. 19F NMR (376 MHz, CDCl3) δ (ppm): −74.06. MS (ESI)
= 254.0, C9H9F3O3S requires 254.0224.
1,1,1-Trifluoropropan-2-yl 4-methylbenzenesulfonate (8b).
Synthesised as described for 7b, from 5.7 g (50 mmol) 1,1,1-
trifluoropropan-2-ol. Product 8b was obtained as colourless oil
1
in 84% (10.1 g) yield. H NMR (400 MHz, CDCl3) δ (ppm):
1.45 (d, J = 7 Hz, 3 H, CH3), 2.44 (s, 3H, ArCH3), 4.82 (p, J =
6 Hz, 1 H, CH), 7.35 (d, J = 9 Hz, 2 H, ArH), 7.79 (d, J = 9 Hz,
2 H, ArH). 13C NMR (100 MHz, CDCl3) δ (ppm): 14.7, 21.7,
73.2 (q, JCF = 34.2 Hz, 118.7, 121.5, 124.3, 127.9, 132.9, 145.6.
19F NMR (376 MHz, CDCl3) δ (ppm): −78.66. MS (ESI) = 267.0,
C10H10F3O3S requires 267.0297, HRMS C10H10F3O3S requires
267.0297, found: 267.0303; C10H11F3O3S requires C, 44.77; H,
4.13; F, 21.25; O, 17.89; S, 11.95; found C, 45.07 H, 4.07%.
2,2-Difluorovinyl 4-methylbenzenesulfonate (7a).9 2,2,2-Tri-
fluoroethyl tosylate (2.57g, 10 mmol) was dissolved in anhydrous
THF (15 ml) and cooled to −78 °C. n-Butyl lithium (1.6 M in
hexanes, 12.5 ml, 20 mmol) was added dropwise and the resul-
tant mixture was stirred at −78 °C for 40 minutes. A mixture of
water (4.5 g, 25 mmol) and THF (10 ml) was added and the
mixture was allowed to warm to room temperature. The reaction
mixture was diluted with diethyl ether and the phases were separ-
ated. The aqueous phase was extracted with diethyl ether (20 ml)
and discarded. The organic phases were combined, dried over
MgSO4 and concentrated. The residue was purified by flash
chromatography on silica gel (pentane–diethyl ether, 19 : 1).
Compound 7a was obtained as a colourless oil in 79% (1.86 g)
yield. 1H NMR (400 MHz, CDCl3) δ (ppm): 2.46 (s, 3H,
ArCH3), 4.38 (q, J = 8 Hz, 1H, CH), 7.38 (d, J = 9 Hz, 2H,
ArH), 7.81 (d, J = 9 Hz, 2H, ArH). 13C NMR (100 MHz,
CDCl3) δ (ppm): 21.7, 64.5 (q, JCF = 37.8 Hz), 120.5, 123.2,
128.1, 130.9 (d, JCF = 171 Hz), 131.8, 145.9. 19F NMR
(376 MHz, CDCl3) δ (ppm): −74.06. MS (ESI) = 234.0,
C9H8F2O3S requires: 234.0162.
Conclusions
We have demonstrated a novel, versatile methodology for the
nucleophilic radiosynthesis of [18F]fluoro-1,1-difluoromethyl-
labelled compounds which have not been accessible so far.
Direct nucleophilic 18F-fluorination of CF3 groups in high
specific radioactivity opens up a wide range of potential candi-
dates for radiotracer studies. Furthermore, accessing such
“native” functionalities in biologically well-characterised mole-
cules will be highly favourable to the invasive introduction of
additional fluorine atoms or fluorinated prosthetic groups into
radiotracer candidates. Our novel approach is an advancement of
utmost utility for the field of radiochemistry and in particular for
PET-imaging. Limitations of our protocol involving H2O
addition to the labelling medium have now been overcome by
replacement of water by organic proton donors, which do not
impair reactivity of fluoride ion at a concentration of 1 M.
1,1-Difluoroprop-1-en-2-yl 4-methylbenzenesulfonate (8a).
Synthesised from 6b (2.7 g, 10 mmol) as described for 7a. Com-
pound 7b was obtained as a colourless oil in 83% (2.05 g) yield.
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 6980–6986 | 6985

View Article Online
(b) P. J. H. Scott, personal communication; (c) H. Watanabe, N. Chopra,
D. Laver, H. S. Hwang, S. S. Davies, D. E. Roach, H. J. Duff,
D. M. Roden, A. A. M. Wilde and B. C. Knollmann, Nat. Med., 2009,
15, 380; (d) A. van Oeveren, M. Motamedi, N. S. Mani, K. B. Marschke,
F. J. López, W. T. Schrader, A. Negro-Vilar and L. Zhi, J. Med. Chem.,
2006, 49, 6143; (e) F. Mauler, J. Mittendorf, E. Horváth and J. De Vry,
J. Pharmacol. Exp. Ther., 2002, 302, 359; (f) L. G. Miller,
W. R. Galpern, J. J. Byrnes and D. J. Greenblatt, Pharmacol., Biochem.
Behav., 1992, 43, 413.
Notes and references
1 (a) H. J. Wester, in Handbook of Nuclear Chemistry, ed. A. Vértes,
S. Nagy and Z. Klencsár, Kluwer Academic Publishers, Amsterdam,
2003, vol. 4, 167; (b) P. W. Miller, N. J. Long, R. Vilar and A. D. Gee,
Angew. Chem., Int. Ed., 2008, 47, 8998; (c) L. Cai, S. Lu and V. W. Pike,
Eur. J. Org. Chem., 2008, 2853; (d) S. P. Hume, R. N. Gunn and
T. Jones, Eur. J. Nucl. Med. Mol. Imaging, 1998, 25, 173; (e) S. P. Hume
and R. Myers, Curr. Pharm. Des., 2002, 8, 1497.
7 P. J. Riss and F. I. Aigbirhio, Chem. Commun., 2011, 47, 11873.
8 (a) M. R. Kilbourn, M. R. Pavia and V. E. Gregor, Int. J. Rad. Appl.
Instrum. A, 1990, 41, 823; (b) O. Josse, D. Labar, B. Georges,
V. Grégoire and J. Marchand-Brynaert, Bioorg. Med. Chem., 2001, 9,
665; (c) W. R. Dolbier Jr, A.-R. Li, C. J. Koch, C.-Y. Shiue and
A. V. Kachur, Appl. Radiat. Isot., 2001, 54, 73.
2 (a) P. J. H. Scott, Angew. Chem., Int. Ed., 2009, 48, 6001; (b) P. J. Riss,
S. Lu, S. Telu, F. I. Aigbirhio and V. W. Pike, Angew. Chem., Int. Ed.,
2012, 51, 2698–2702; (c) L. Li, M. N. Hopkinson, R. L. Yona, R. Bejot,
A. D. Gee and V. Gouverneur, Chem. Sci., 2011, 2, 123; (d) J. Castillo
Meleán, J. Ermert and H. H. Coenen, Org. Biomol. Chem., 2011, 9, 765;
(e) P. J. Riss and F. Roesch, Org. Biomol. Chem., 2008, 6, 4567.
9 T. M. Gøgsig, L. S. Søbjerg, A. T. Lindhardt (neé Hansen), K. L. Jensen
and T. Skrydstrup, J. Org. Chem., 2008, 73, 3404.
3 G. J. Meyer, S. L. Waters, H. H. Coenen, A. Luxen, B. Maziere and
B. Langström, Eur. J. Nucl. Med., 1995, 22, 1420.
10 Sigma-Aldrich 41647 DMSO puriss, absolute, over molecular sieve
(H2O ≤0.005%), ≥99.5% (GC) was used.
11 (a) W. M. Olmstead, Z. Margolin and F. G. Bordwell, J. Org.
Chem., 1980, 45, 3295; (b) D. W. Kim, H.-J. Jeong, S. T. Lim,
M.-H. Sohn, J. A. Katzenellenbogen and D. Y. Chi, J. Org.
Chem., 2008, 73, 957; (c) C. M. Sharts and W. A. Sheppard,
Org. React., 1974, 21, 125.
4 (a) H. Teare, E. G. Robins, E. Arstad, S. K. Luthra and V. Gouverneur,
Chem. Commun., 2007, 2330; (b) R. Chirakal, R. M. Adams, G. Firnau, G.
J. Schrobilgen, G. Coates and E. S. Garnett, Nucl. Med. Biol., 1995, 22, 111.
5 (a) H. L. Yale, J. Med. Pharm. Chem., 1959, 1, 121; (b) A. G. M. Barret,
M. R. Crimmin, M. S. Hill, P. B. Hitchcock and P. A. Procopiou, Angew.
Chem., Int. Ed., 2007, 46, 6339; (c) D. O’Hagan and C. Isanbor, J. Fluor-
ine Chem., 2006, 127, 303; (d) T. Furuya, A. S. Kamlet and T. Ritter,
Nature, 2011, 473, 470; (e) S. Purser, P. R. Moore, S. Swallow and
V. Gouverneur, Chem. Soc. Rev., 2008, 37, 320.
12 Fluoride was produced in the morning and used for multiple experiments
throughout the day; typical radiotracer productions in our laboratories
afford the formulated product in specific activities of 150–800 MBq nmol−1
,
6 (a) M. Suehiro, G. Yang, G. Torchon, E. Ackerstaff, J. Humm,
J. Koutcher and O. Ouerfelli, Bioorg. Med. Chem., 2011, 19, 2287;
1 hour from end of bombardment (EOB).
6986 | Org. Biomol. Chem., 2012, 10, 6980–6986
This journal is © The Royal Society of Chemistry 2012