Synthesis, bioconjugation and stability studies of [18F]ethenesulfonyl fluoride

Article abstract of DOI:10.1002/jlcr.3667

Fluorine-18 labelled prosthetic groups (PGs) are often necessary for radiolabelling sensitive biological molecules such as peptides and proteins. Several shortcomings, however, often diminish the final yield of radiotracer. In an attempt to provide higher yielding and operationally efficient tools for radiolabelling biological molecules, we describe herein the first radiochemical synthesis of [¹⁸F]ethenesulfonyl fluoride ([¹⁸F]ESF) and its Michael conjugation with amino acids and proteins. The synthesis of [¹⁸F]ESF was optimised using a microfluidic reactor under both carrier-added (c.a.) and no-carrier-added (n.c.a.) conditions, affording, in a straightforward procedure, 30-50% radiochemical yield (RCY) for c.a. [¹⁸F]ESF and 60-70% RCY for n.c.a. [¹⁸F]ESF. The conjugation reactions were performed at room temperature using 10?mg/mL precursor in aqueous/organic solvent mixtures for 15?min. The radiochemical stability of the final conjugates was evaluated in injectable formulation and rat serum, and resulted strongly substrate dependent and generally poor in rat serum. Therefore, in this work we have optimised a straightforward synthesis of [¹⁸F]ESF and its Michael conjugation with model compounds, without requiring chromatographic purification. However, given the general low stability of the final products, further studies will be required for improving conjugate stability, before assessing the use of this PG for PET imaging.

Full text of DOI:10.1002/jlcr.3667

Synthesis, Bioconjugation and Stability
Studies of [18F]Ethenesulfonyl Fluoride
Bo Zhang,[a],[b] Giancarlo Pascali,*[a],[c] Naomi Wyatt,[a] Lidia Matesic,[a] Mitchell A.
Klenner,[a],[d] Tiffany R. Sia,[a],[c] Adam J. Guastella,[c] Massimiliano Massi,[d] Andrea J.
Robinson,[b] and Benjamin H. Fraser*[a]
[a] Australian Nuclear Science and Technology Organisation (ANSTO), New Illawarra Rd,
Lucas Heights, NSW, Australia
[b] Monash University, Wellington Rd, Clayton, Victoria, Australia
[c] Brain and Mind Centre – The University of Sydney, Mallett St, Camperdown, NSW,
Australia
[d] Curtin University, Kent St, Bentley, Western Australia, Australia
*E-mail: gianp@ansto.gov.au; bfr@ansto.gov.au;
Phone: +61 2 9717 7611
Short title
[18F]ethenesulfonyl fluoride: synthesis, conjugation and products stability
Keywords
fluorine-18, sulfonyl fluoride, Michael acceptor, peptide labelling, prosthetic group.
This article has been accepted for publication and undergone full peer review but has not
been through the copyediting, typesetting, pagination and proofreading process which may
lead to differences between this version and the Version of Record. Please cite this article as
doi: 10.1002/jlcr.3667
This article is protected by copyright. All rights reserved.

Abstract
Fluorine-18 labelled prosthetic groups (PGs) are often necessary for radiolabelling sensitive
biological molecules such as peptides and proteins. Several shortcomings, however, often
diminish the final yield of radiotracer. In an attempt to provide higher yielding and
operationally efficient tools for radiolabelling biological molecules, we describe herein the
first radiochemical synthesis of [18F]ethenesulfonylfluoride ([18F]ESF) and its Michael
conjugation with amino acids and proteins. The synthesis of [18F]ESF was optimised using a
microfluidic reactor under both carrier-added (c.a.) and no-carrier-added (n.c.a.) conditions,
affording, in a straightforward procedure, 30-50% radiochemical yield (RCY) for c.a. [18F]ESF
and 60-70% RCY for n.c.a. [18F]ESF. The conjugation reactions were performed at room
temperature using 10 mg/mL precursor in aqueous/organic solvent mixtures for 15 min. The
radiochemical stability of the final conjugates was evaluated in injectable formulation and
rat serum, and resulted strongly substrate dependent and generally poor in rat serum.
Therefore, in this work we have optimised a straightforward synthesis of [18F]ESF and its
Michael conjugation with model compounds, without requiring chromatographic
purification. However, given the general low stability of the final products, further studies
will be required for improving conjugate stability, before assessing the use of this PG for PET
imaging.
This article is protected by copyright. All rights reserved.

Introduction
Positron emission tomography (PET) is a widely used molecular imaging tool that monitors
the uptake and distribution of radiolabelled molecules in vivo.1 It is used both in medical
diagnostic procedures and as a powerful tool for investigating the fundamental mechanisms
of disease pathology.2 Fluorine-18 is one of the most commonly used PET radionuclides due
to its clean decay profile, low positron energy, ideal half-life (109 minutes)3,4 and propensity
to be incorporated into target molecules either directly or indirectly. The more direct and
efficient method usually involves [18F]fluoride displacement of a suitable leaving group from
a precursor molecule, requiring the use of harsh reaction conditions (high temperature and
strong base) which are not suitable for biological molecules (proteins, peptides,
antibodies).5 The indirect method is typically a two-step process which involves the
preliminary synthesis of a radiofluorinated prosthetic group (PG), followed by incorporation
of the PG into the target molecule. The second step is run under mild conditions to avoid
degradation of the biological vector.
Indirect fluorine-18 radiolabelling can be achieved by three methodologies. In the first
approach, the radiolabelled PG is conjugated under mild conditions with an unmodified
macromolecule. In this case, the radiolabelling precursor would be a bifunctional molecule,
which undergoes selective radiolabelling on one site and specific bioconjugation on another
site. Typical examples of these PGs are 4-nitrophenyl-2-[18F]fluoropropionate and N-
succinimidyl-4-[18F]fluorobenzoate, which react with free amine functionalities naturally
available within the target.6,7 The second approach requires prior modification of the target
macromolecule with a selected PG that can be radiolabelled under mild conditions using
“raw” forms of fluorine-18. Typical examples are PGs that consist of B, Si or NOTA (for
This article is protected by copyright. All rights reserved.

[18F]AlF2+).8–11 The third approach involves the use of click-chemistry reactions, and
therefore requires modification of the macromolecule with an unnatural moiety (e.g. alkyne
or azide) and the synthesis of a radiolabelled PG bearing the “partner” click functionality.12
Even if the radiochemistry of the second and third approaches is relatively simple, their
application might be hindered by the need to modify the target biomolecule. In general, the
first approach would be preferred as it does not require modification of the target
biomolecule and thus widens the applications. However, several shortcomings limit the use
of this approach, such as an often complicated synthesis of the PG, the sensitivity of the
biomolecule to optimised conjugation conditions, the need for HPLC pre-purification of the
PG, post-purification of the conjugate, and the potential pharmacokinetic impact of adding
the PG to target molecules. Given this, our goal was to develop a PG which can overcome
most of these shortcomings, can be shipped to an imaging centre, and can be used in a
simple “shake and bake” procedure to synthesize a wide range of radiotracers for PET
imaging.
PGs that consist of radioactive sulfonyl fluorides were first reported in 2012. They have been
prepared in high radiochemical yields (RCYs) even at room temperature and possess high
stability in aqueous environments, although the decomposition of some of the studied
analogues have been observed in mouse serum.13,14 Among the molecules with sulfonyl
fluoride moieties, ethenesulfonyl fluoride (ESF) is one of the strongest Michael acceptors
reported in the literature.15 It reacts with various soft nucleophiles, including amines, thiols,
enamines, dienes, phosphines and activated methylene groups, under mild conditions with
typically high yields, but is unreactive towards water and other hard nucleophiles.16
Consequently, [18F]ESF has significant potential to be used as a PG to radiolabel targets that
This article is protected by copyright. All rights reserved.

contain lysine or cysteine residues or other ‘soft’ nucleophilic groups naturally present in
biomolecules. The small size and hydrophilic property of [18F]ESF makes it ideal for
radiolabelling oligopeptides since the resulting conjugation adducts may be less likely to
adversely affect target binding and radiotracer pharmacokinetics, especially when compared
to some larger and more hydrophobic PGs currently known. In this work, we aimed to
systematically explore and optimise a simple and high yielding radiochemical synthesis of
[18F]ESF and investigate its utility for radiolabelling model structures and biological vectors.
Experimental
Materials & Instruments
All the reagents used were purchased from Sigma-Aldrich (analytical grade or higher) and
used without further purification, unless stated otherwise. All solvents used for
chromatography were purchased from Merck and were of HPLC grade. MP1 cartridges were
purchased from either ORTG or Medchem Imaging. Solid Phase Extraction (SPE) cartridges
were purchased from Waters.
HPLC for quality control of non-radioactive products consisted of a Waters 626 pump,
Waters in-line degasser AF, Waters 717 plus autosampler, Waters 2996 photodiode array
detector, Waters 600S controller and Alltima C18 column (3 μm pore size, 53 mm length and
7 mm internal diameter). The Radio-HPLC used for quality control of radioactive products
was carried out using a Shimadzu system comprising of a CBM-20 controller, LC-20AD pump,
SIL-20AHT auto injector SPD-M20A PDA and a Lablogic Posi-RAM gamma detector. The
eluent and stationary phases of HPLC runs are reported in SI (Conditions A, B, C).
This article is protected by copyright. All rights reserved.

1
A Bruker Advance DPX-400 NMR spectrometer was run at 400.13 MHz for H NMR, 100.61
MHz for 13C NMR and 376.50 MHz for 19F NMR. A Waters Micromass ZQ quadrupole mass
spectrometer was used to record mass data, and a Bruker Alpha Fourier Transform
spectrometer was used to record IR data. A Buchi Reveleris Flash Chromatography System,
with UV detection at both 254 nm and 280 nm was used to purify the synthesised
compounds using a Silica column.
Aqueous [18F]fluoride of 10-20 GBq was produced from an IBA Cyclone 18 Twin cyclotron by
the 18O(p,n)18F nuclear reaction and delivered to a NanoTek microfluidic system from
Advion.17 Precursor solution and [18F]fluoride complexes were pre-loaded onto two separate
storage loops, and were delivered (10-20 µL for optimisation, 100-200 µL for production) at
a set rate into a microreactor (2 m × 100 µm, 15.6 µL, fused silica) heated at predetermined
temperatures. Reaction conditions which were varied during optimisation runs were:
precursor to [18F]fluoride ratio, [18F]fluoride volume, flow rate and microreactor
temperature. After the reaction mixture passed though the microreactor, the total volume
of the microfluidic system tubing (400 µL) was swept with inert solvent (typically H2O for
carrier-added (c.a.) and DMSO for no-carrier-added (n.c.a.)), and the final product mixture
was delivered to a collection vial. The outlet could be connected for HPLC analysis or
directed towards purification to either the c.a. or the n.c.a. route in as shown Figure 1.
Synthesis of non-radioactive precursors and adducts
Ethenesulfonyl chloride (ESCl)
The synthesis process was an adaptation of a previously reported procedure.18 2-
Chloroethane-1-sulfonyl chloride (5.00 g, 30.7 mmol) was dissolved in diethyl ether (20 mL)
in a round bottom flask and cooled at -50 °C. A solution of 2,6-lutidine (3.94 g, 36.8 mmol) in
This article is protected by copyright. All rights reserved.

diethyl ether (7 mL) was slowly added dropwise, maintaining the bath temperature at -50 °C.
Once the addition was complete, the mixture was stirred for 45 min and warmed to room
temperature. The reaction was quenched by cooling at 0 °C and the additional of ice-cold 1%
H2SO4 (15 mL). The organic layer was separated, washed once more with ice-cold 1% H2SO4
(5 mL), followed by water (5 mL) and 20% NaCl solution. This worked-up organic phase was
then fractionally distilled under vacuum at 30 °C to get 1.73 g clear liquid (yield 45%).
2,4,6-Trichlorophenyl ethenesulfonate (TCPE)
The synthesis process was an adaptation of a previously reported procedure.19 2-
Chloroethane-1-sulfonyl chloride (1.81 g, 11.1 mmol) was dissolved in dichloromethane
(12.5 mL) and cooled to -50 °C. 2,4,6-trichlorophenol (2.00 g, 10.1 mmol) and triethylamine
(2.35 g, 23.3 mmol) were pre-mixed in dichloromethane (2.5 mL) and added dropwise to the
cooled sulfonyl chloride solution. During addition, the temperature was kept below 0 °C.
Following complete addition, the reaction was allowed to warm to room temperature and
stirred for additional 1 h. The reaction mixture was then diluted with diethyl ether (0.5 mL)
and was washed with HCl (5 mL, 2 M) and saturated NaHCO3 (5 mL) solution. The organic
layer was dried over MgSO4 and the solvent removed under reduced pressure. The crude
product was purified by flash chromatography (Silica, gradient of ether/ hexane) to give the
title compound as a white solid (2.00 g, yield 69%).
Non-radioactive adducts
The synthesis procedure was adapted from the literature.16 Ethenesulfonylfluoride (ESF)
(200 mg, 1.81 mmol) was added to amino acid analogues (1 eq) in DMF (1 mL). The mixture
was stirred for 1 hour at room temperature, before being quenched with water (5 mL) and
This article is protected by copyright. All rights reserved.

extracted with diethyl ether (5 mL). The organic layer was rotary evaporated to give the
crude product. Flash chromatography (Silica, gradient of hexane/ethyl acetate) was used to
purify the compounds.
Synthesis of [18F]ESF
The synthesis of both c.a. and n.c.a. [18F]ESF was performed in a microfluidic reactor system
as described in the previous section. The purification of c.a. [18F]ESF was achieved by
connecting the outlet of the microreactor to a neutral alumina Light (nAl) SPE. The
purification of n.c.a. [18F]ESF was performed using the distillation setup in Figure 1
(distillation time 20-30 min). Distillation progress was monitored using a radiodetector near
SPE D and typically complete after 20-30 min. Radiochemical purity (RP) was measured by
Radio-HPLC using Conditions B.
Conjugation of [18F]ESF
Small scale reactions
On each experimental day, all conjugation reactions were performed in parallel with all the
tested substrate solutions, directly in the inserts of HPLC autosampler vials, by adding 20 µL
(around 20 MBq) of [18F]ESF to 80 µL of amino acid (AA) analogues solution (1 mg/mL in
EtOH, DMF, DMSO, saline, water and PBS pH 7.4, with or without adding triethylamine
(TEA)). After completion of [18F]ESF addition, the test vials were transferred in the HPLC
system and the whole series have been analysed sequentially up to 3 times (i.e. over-night
analysis). Radio-HPLC: Conditions B for 10 mg/mL reaction, Conditions C for 1 mg/mL
reactions.
This article is protected by copyright. All rights reserved.

Large scale reactions
Large scale reactions were performed by adding 100 µL of [18F]ESF (around 100 MBq) to 400
µL of amino acid analogue solutions in their best solvent conditions (determined from
previous small scale reactions) in a 4 mL vial. After 15 min reaction time (45 min for Ani, 30
min for Insulin), the reaction mixture was purified by passing through a nAl SPE. The activity
of the purified eluents and the nAl SPE was then measured, which enabled the calculation of
the isolated RCY. The RP was measured by Radio-HPLC using Conditions B.
Stability of [18F]ESF conjugates
Stability in injectable formulation
The purified adducts from large scale reactions were diluted in saline to make injectable
formulations (saline containing 10% v/v of EtOH or DMSO), and then analysed by Radio-
HPLC at 0, 15, 30, 45, 60 and 120 min using Conditions B.
Stability in rat serum
10 µL injectable formation of each adduct (c.a. and n.c.a. Ani, c.a. Cys, n.c.a. Insulin, n.c.a.
Bovine serum albumin (BSA)) was added into 100 µL of rat serum in a shaker at 37 °C.
Samples were taken at 15, 30, 45 and 60 min and then analysed by Radio-HPLC using
Conditions B.
Results and discussion
Non-radioactive conjugation reactions
Conjugation reactions of non-radioactive ESF with selected amino acid analogues and aniline
(Figure 2) were performed to evaluate the conjugation ability of ESF and to synthesize the
This article is protected by copyright. All rights reserved.

non-radioactive adduct standards for characterisation. This set of compounds was chosen
because it consisted of a range of functionalities (i.e. aliphatic amine, aromatic amine and
thiol). Among the AA analogues, side chains with different electron withdrawing abilities
were selected to evaluate their effects on the amine group reactivity. All analogues reacted
with ESF at room temperature with satisfying yields, except for Lys which could not be
synthesized as the reaction always provided the double conjugated Lys-bis(ESF).
Synthesis of [18F]ESF
The synthesis of [18F]ESF was first attempted based on the reported synthesis of non-
radioactive ESF using 2-chloroethanesulfonyl chloride (ClSCl, Scheme 1A) or ethenesulfonyl
chloride (ESCl, Scheme 1B).14,16 Unfortunately both routes were not successful, most likely
due to the instability of ESCl and the degradation of intermediates under basic radiolabelling
conditions. We then changed our strategy and tested a simple 18F/19F isotopic exchange on
non-radioactive ESF to obtain carrier added (c.a.) [18F]ESF (Scheme 1C).
The exchange reaction was performed in a microfluidic system, which enabled fast
optimisation of reaction parameters (i.e. temperature and reaction time) and minimisation
of operational variation. We first tested the effect of different forms of 18F fluorination
complex, and we obtained a maximum RCY of 63% at 160 °C using the traditional
azeotropically dried K2.2.2/K2CO3 [18F]fluoride complex. Since we suspected that the basicity
of the fluorinating complex was leading to the degradation of [18F]ESF, we repeated the
reaction with non-azeotropically dried (wet) tetraethylammonium bicarbonate (TEAB)
[18F]fluoride complex, and we were able to increase the maximum RCY to 83% at 130 °C.
Following this lead, we directly tested untreated irradiated target water as a source of
[18F]fluoride, and a maximum RCY of 32% was obtained at 100 °C. This source, obtained
This article is protected by copyright. All rights reserved.

directly from cyclotron target production, although operationally simple, could be
problematic due to cross-site differences in targetry and the presence of longer-lived
radioactive cationic species. Encouraged by this result, we reverted back to trapping the
[18F]fluoride on an anion exchange resin (MP1 or QMA) followed by elution with isotonic
saline, which resulted in a maximum RCY of 90% at 100 °C.
Upon testing the effect of ESF concentration, we observed an increased RCY at higher ESF
concentration, with a maximum RCY of 90% at 20 mg/mL. However, we chose 5 mg/mL as
the optimised concentration because of little RCY difference between 20 mg/mL and 5
mg/mL. We also noticed the optimised temperature was 100 °C; further increasing or
decreasing the temperature resulted in less RCY, probably due to decomposition of the
product and incomplete reaction respectively.
Lastly, we tested the effect of residence times (equivalent to reaction time in flow chemistry,
controlled by flow rate), and we obtained the best RCY of 81% with a residence time of 47
seconds. Longer and shorter residence times did not improve the RCY, most likely due to the
same reason as the increase or decrease of temperature (decomposition of the product and
incomplete reaction respectively).
In order to achieve the best flexibility in the subsequent conjugation step, we reduced the
ionic strength in the reaction mixture by employing hypotonic saline (0.45% NaCl) as
[18F]fluoride eluent and 5 mg/mL ESF in water as precursor. This gave us a satisfying RCY of
64% at 130 °C with a residence time of 47 seconds.
A vessel reaction was also performed under similar conditions (5 mg/mL ESF in water,
[18F]fluoride in isotonic saline, 130 °C for 15 min, 1 mL total volume), providing a lower RCY
This article is protected by copyright. All rights reserved.

of 57%. Since the only radioactive impurity in this reaction was unreacted [18F]fluoride,
purification was straightforwardly performed by connecting the microreactor outlet to a nAl
SPE. An isolated RCY of 30-50% was obtained by measuring the activity of the [18F]ESF
product and the nAl SPE. This gave a final product with >95% RP assessed by Radio-HPLC
(Radio-TLC could not be used as [18F]ESF adsorbs strongly to Silica), and was stable for
transportation over 4 h in the reaction media.
Since ESF has a very low molar extinction coefficient and is not detectable by the UV
component of HPLC, GC-MS was instead used to identify the presence of ESF. We injected
the mixture of radioactive [18F]ESF spiked with an excess amount of non-radioactive ESF into
the HPLC, and collected the radioactive peak suspected to be [18F]ESF. After decay, we then
analysed the collected peak by GC-MS and the results confirmed the expected identity (SI
3.3).
Having successfully produced c.a. [18F]ESF and verified its potential as a PG, we investigated
the development of a method to produce no-carrier-added (n.c.a.) [18F]ESF. Examination of
the literature showed that radiofluorination on the sulfur atom of tosylate esters can be
achieved.20,21 However, this reaction has traditionally been regarded as an undesired side
process, as the tosylate group is typically used on alkyl chains as a leaving moiety to be
displaced by the [18F]fluoride. We therefore envisaged exploiting this possibility by building
a precursor in which the only site of likely fluoride attack would be the sulfur atom. Our
n.c.a. radiolabelling precursor is a stable white solid 2,4,6-trichlorophenylethenesulfonate
(TCPE) which is synthesized by reaction of ClSCl with the anion generated by base treatment
of 2,4,6-trichlorophenol. The TCPE precursor was then subjected to nucleophilic
radiofluorinations to produce n.c.a. [18F]ESF (Scheme 1D).
This article is protected by copyright. All rights reserved.

The synthesis of n.c.a. [18F]ESF was performed on the microfluidic system using the same
strategy for optimisation, and we found that use of wet TEAB/[18F]fluoride complex in
CH3CN and TCPE precusor concentration of 20 mg/mL gave a maximum RCY of 86% at
110 °C, irrespective to residence time and solvent. We picked 24 seconds as our optimised
residence time as it was operationally straightforward and time saving.
A vessel reaction was also performed under similar conditions (0.5 mL of both 20 mg/mL
TCPE and TEAB/[18F]fluoride in DMSO heated at 80 °C for 30 min), however this resulted in a
reduced RCY of 35%. Exploiting the fact that ESF is volatile while TCPE is non-volatile, we
employed a distillation approach to purify the n.c.a. [18F]ESF from a high-boiling solvent
(DMSO). This was achieved by connecting the microreactor to a closed vial which was
heated at 80 °C and pressurized with N2 gas. The gaseous effluents from this vial were then
directed through a Silica SPE, which trapped the [18F]ESF gas. At this point, [18F]ESF could be
eluted from this Silica SPE by using either aqueous (water, saline, PBS pH7.4) or organic
solvents (EtOH, DMSO, CH3CN). However, we noticed that when using higher activities (>500
MBq), only EtOH and CH3CN could be used for elution, as [18F]ESF was not stable in other
solvents, most likely due to radiolysis. A RCY of 60-70% was achievable by this distillation
method and the final product was >95% RP with no UV impurities detectable. We also
verified the RP of [18F]ESF in the cartridge was still >95% after 4 h. This means we can
transport the cartridge to other sites where [18F]ESF can be eluted with desired solvent and
conjugated with molecules of interest.
Conjugation of [18F]ESF
Having optimised a simple way of producing c.a. [18F]ESF, we studied its utilization in
conjugation reactions with selected AA analogues and aniline (Figure 2) in various solvents.
This article is protected by copyright. All rights reserved.

Radio-TLC and Radio-HPLC were both used to measure conjugation RCY and the results were
comparable;22 however, we chose Radio-HPLC to assess conjugation efficiency as it allows
the measurement of both unreacted [18F]ESF and [18F]fluoride. In a first set of reactions
utilizing AA at 1 mg/mL concentration, all the HPLC vials were queued for overnight analysis
to get quick proof of principle results; in this case, the reaction time was not controlled and
could be up to 8 h (complete data reported in SI). In this initial screening, all analogues
tested in organic solvents (i.e. EtOH, DMSO, DMF) reacted with [18F]ESF and gave RCYs over
35%, while the use of aqueous solvents did not provide any conjugate, except for Pro and
Ani. This is most likely due to the use of the hydrochloride salts of the AA substrates, which
may not have sufficient availability of the free amine form for conjugation in aqueous
environment. Pro and Ani were exceptions, which may be due to their different basic
characters (secondary and aromatic amines respectively).
Due to the short half-life of fluorine-18, we modified our procedure to obtain satisfactory
yields in shorter reaction times. We discovered that increasing the concentration of
substrates to 10 mg/mL gave improved conjugation yields for a 15 min reaction time. Table
1 reports the example of Trp which gave a 64% RCY under these new conditions (entry 2),
compared to 41% in 6 h for the 1 mg/mL case (entry 1). However, in most reactions, we
noticed the presence of a small amount of unreacted [18F]ESF (Figure 3A). In order to drive
the reaction to completion, we tested the addition of triethylamine (TEA) to transiently
convert the amine salt to its free base form. We first started adding 1 equivalent (eq) of TEA
with respect to AA, and obtained significantly higher RCYs with no unreacted [18F]ESF left
except for Ani. As for Ani, increasing the reaction time to 45 min allowed us to increase RCY
to 85% with no unreacted [18F]ESF. As for the other AA, we gradually decreased the amount
This article is protected by copyright. All rights reserved.

of TEA and we found that 0.25 eq of TEA was sufficient to consume all the unreacted
[18F]ESF and gave Trp conjugate a higher RCY of 91% (Table 1, entry 3; Figure 3B). We also
verified that this amount of TEA did not affect the stability of the conjugate by re-analysing
the conjugate over time.
Larger scale conjugation reactions for each substrate under optimised conditions were also
performed. Since all [18F]ESF was consumed in these conditions, final passage of the product
mixture through a nAl SPE removed the only radioactive impurity, [18F]fluoride, providing
conjugation products with >95% RP (Table 1, entry 4). Since we wanted to assure the
highest concentration of the final product, we did not rinse the cartridge with additional
solvent; as consequence of this, the isolated RCYs (Table 2) were slightly lower than the
small scale reactions. However, we noticed that it was possible to obtain similar RCYs by
rinsing the cartridge with further 1 mL of solvent. The identity of conjugate products was
confirmed through HPLC retention time (Rt) comparison with the non-radioactive standards
(see Figure 4 for example of [18F]Pro-ESF). The identification of [18F]Lys-ESF was tentative
since we did not have the non-radioactive standard and could only speculate from its Rt
which was between Lys and Lys-bis(ESF). These purified products were assessed for stability
in their production solvent mixture (i.e. 80% organic), and were >90% RP after 4 h, except
for [18F]Lys-ESF which resulted in complete defluorination.
The conjugation reactions of n.c.a. [18F]ESF were conducted using a similar strategy as for
the c.a. experiments, with the additional tests of BSA and Insulin. Also in this case, we first
tried the longer time reactions with AA concentrations of 1 mg/mL, and then shortened the
reaction time to 15 min while increasing the AA concentrations to 10 mg/mL (SI and Table 1,
entries 6-7). In general, most AA analogues afforded higher RCY with n.c.a. [18F]ESF than
This article is protected by copyright. All rights reserved.

with c.a. [18F]ESF, probably due to the increased stoichiometric ratio of precursor towards
the PG. Since the amount of unreacted [18F]ESF was less, we re-evaluated the amount of
TEA required and found that 0.1 eq of TEA relative to AA was enough to drive the
conjugation reaction to completion (Table 1, entries 7-8). Surprisingly, DMF gave a
dramatically lower RCY (Table 1, entry 5), which is most likely due to the presence of small
amounts of dimethylamine impurity in the solvent. We verified this hypothesis by testing
the stability of n.c.a. [18F]ESF in 0.04% aqueous dimethylamine and confirmed its
degradation to [18F]fluoride within 15 min. This different behaviour between n.c.a. and c.a.
[18F]ESF is a textbook example of the effect of carrier amount in radiochemistry.23 The two
additional protein candidates, BSA and Insulin, were tested in saline and H2O (at 10 mg/mL),
and showed satisfying RCYs of 34% and 70%, respectively, after 15 min.
Larger scale reactions of n.c.a. [18F]ESF on each substrate in their optimised conditions were
then performed. Products were obtained with >95% RP (Table 1, entry 9) and the purity
remained unchanged for 4 h in the conjugation media, except for [18F]Lys-ESF. The molar
activity (MA) of n.c.a. PG conjugates could only be estimated as a lower bound, as the mass
amount of adduct in the HPLC injection was below the detection limit; despite this, n.c.a. MA
were at least >1000 times higher than the c.a. ones (Table 2). Since the [18F]ESF conjugates
of BSA and Insulin could not be separated from the unreacted substrates, the effective MA
(Activity/Protein amount) was instead calculated.
Stability of [18F]ESF conjugates
Further stability tests were conducted on all of the purified conjugates (i.e. both c.a. and
n.c.a.) in injectable formulations (saline containing 10% v/v EtOH or DMSO) over a 2 h time
frame. Our aim was to understand the time needed to reduce the RP of the conjugates to a
This article is protected by copyright. All rights reserved.

threshold level of 10%. Figure 5 shows a selection of the results from n.c.a. conjugates. In
this case, we found that after 2 h [18F]Ani-ESF was the most stable candidate with 92% RP,
followed by [18F]Cys-ESF with 43%; all the other AA conjugates degraded (i.e. <10% RP)
within 1 h through defluorination. BSA and Insulin conjugates were both 54% pure after 2 h.
Of the c.a. adducts, [18F]Ani-ESF and [18F]Cys-ESF proved to be the better candidates, with
purities of 85% and 12% respectively after 2 h.
Having evaluated the stability of all conjugates in injectable formulations, we selected the
best candidates and tested their stabilities in rat serum. [18F]Ani-ESF approached the test
threshold after one hour in both c.a. and n.c.a. cases (16% and 10% RP, respectively),
whereas [18F]Cys-ESF completely degraded within the first 15 mins. For the two proteins,
the Insulin conjugate reached 13% RP after 15 min, while the BSA conjugate was completely
degraded within the same time.
Conclusions
In this work, we have developed the first radiosynthesis of [18F]ESF, under both c.a. and
n.c.a. conditions that do not require chromatographic purification and are operationally
straightforward. We have also successfully tested the conjugation of [18F]ESF with selected
analogues at room temperature, and developed a simple cartridge-based method for
purification. Importantly, the conjugation does not require any modification of the targets,
but employs naturally occurring “soft” nucleophilic functional groups (i.e. SH, NH2). Whilst
the conjugates are stable in their preparation mixture, their stabilities in injectable
formulation and rat serum are generally poor and strongly substrate dependent. Further
work on improving conjugate stability is needed to allow use of this PG in PET imaging. We
This article is protected by copyright. All rights reserved.

are currently investigating other strategies to decrease S-F bond reactivity by inhibiting
hydrogen bond activation and using different formulation strategies.14,24
Acknowledgements
We thank N. Paneras (ANSTO) for cyclotron production of 18F, and G. Perkins for performing
vessel batch reactions. Financial supports from Monash University (MGS and MIPRS), The
Australian Institute of Nuclear Science and Engineering (AINSE PGRA 12074) and the
Australian Research Council (ARC-Linkage LP1590101307) are gratefully acknowledged.
This article is protected by copyright. All rights reserved.

References
1.
2.
Schwaiger M, Kunze K, Rischpler C, Nekolla SG. PET/MR: Yet another Tesla? J Nucl
Cardiol. 2017;24(3):1019-1031. doi:10.1007/s12350-016-0665-2.
Tochon-Danguy HJ, Sachinidis JI, Egan GF, et al. Positron emission tomography:
radioisotope and radiopharmaceutical production. Australas Phys Eng Sci Med.
1999;22(4):136-144. http://www.ncbi.nlm.nih.gov/pubmed/10740886. Accessed
February 17, 2018.
3.
4.
5.
6.
Miller PW, Long NJ, Vilar R, Gee AD. Synthesis of11C, 18F, 15O, and 13N radiolabels
for positron emission tomography. Angew Chemie - Int Ed. 2008;47(47):8998-9033.
doi:10.1002/anie.200800222.
Jacobson O, Chen X. PET designated flouride-18 production and chemistry. Curr Top
Med Chem. 2010;10(11):1048-1059.
http://www.ncbi.nlm.nih.gov/pubmed/20388116. Accessed February 17, 2018.
Olberg DE, Hjelstuen OK. Labeling strategies of peptides with 18F for positron emission
tomography. Curr Top Med Chem. 2010;10(16):1669-1679.
http://www.ncbi.nlm.nih.gov/pubmed/20583991. Accessed February 17, 2018.
Haubner R, Wester HJ, Burkhart F, et al. Glycosylated RGD-containing peptides: tracer
for tumor targeting and angiogenesis imaging with improved biokinetics. J Nucl Med.
2001;42(2):326-336. http://www.ncbi.nlm.nih.gov/pubmed/11216533. Accessed
February 17, 2018.
7.
8.
9.
Vaidyanathan G, Zalutsky MR. Labeling proteins with fluorine-18 using N-succinimidyl
4-[18F]fluorobenzoate. Int J Radiat Appl Instrumentation Part B Nucl Med Biol.
1992;19(3):275-281. doi:10.1016/0883-2897(92)90111-B.
Schirrmacher R, Bradtmöller G, Schirrmacher E, et al. 18F-Labeling of Peptides by
means of an Organosilicon-Based Fluoride Acceptor. Angew Chemie Int Ed.
2006;45(36):6047-6050. doi:10.1002/anie.200600795.
Chansaenpak K, Vabre B, Gabbaï FP. [ 18 F]-Group 13 fluoride derivatives as
radiotracers for positron emission tomography. Chem Soc Rev. 2016;45(4):954-971.
doi:10.1039/C5CS00687B.
10. Da Pieve C, Allott L, Martins CD, et al. Efficient [ 18 F]AlF Radiolabeling of Z HER3:8698
Affibody Molecule for Imaging of HER3 Positive Tumors. Bioconjug Chem.
2016;27(8):1839-1849. doi:10.1021/acs.bioconjchem.6b00259.
11. Smith GE, Sladen HL, Biagini SCG, Blower PJ. Inorganic approaches for radiolabelling
biomolecules with fluorine-18 for imaging with Positron Emission Tomography. Dalt
Trans. 2011;40(23):6196. doi:10.1039/c0dt01594f.
12. Liu S, Shen B, T. Chin F, Cheng Z. Recent Progress in Radiofluorination of Peptides for
PET Molecular Imaging. Curr Org Synth. 2011;8(4):584-592.
doi:10.2174/157017911796117197.
This article is protected by copyright. All rights reserved.

13. Inkster JAH, Liu K, Ait-Mohand S, et al. Sulfonyl fluoride-based prosthetic compounds
as potential 18F labelling agents. Chem - A Eur J. 2012;18(35):11079-11087.
doi:10.1002/chem.201103450.
14. Matesic L, Wyatt N a., Fraser BH, Roberts MP, Pham TQ, Greguric I. Ascertaining the
suitability of aryl sulfonyl fluorides for [18F]radiochemistry applications: A systematic
investigation using microfluidics. J Org Chem. 2013;78(22):11262-11270.
doi:10.1021/jo401759z.
15. Zheng Q, Dong J, Sharpless KB. Ethenesulfonyl Fluoride (ESF): An On-Water Procedure
for the Kilogram-Scale Preparation. J Org Chem. 2016;81(22):11360-11362.
doi:10.1021/acs.joc.6b01423.
16. Krutak JJ, Burpitt RD, Moore WH, Hyatt JA. Chemistry of ethenesulfonyl fluoride.
Fluorosulfonylethylation of organic compounds. J Org Chem. 1979;44(22):3847-3858.
doi:10.1021/jo01336a022.
17. Pascali G, Matesic L, Collier TL, et al. Optimisation of nucleophilic 18F
radiofluorinations using a microfluidic reaction approach. Nat Protoc. 2014;9(9):2017-
2029. doi:10.1038/nprot.2014.137.
18. Rondestvedt CS. Unsaturated Sulfonic Acids. IV. Preparation and Properties of α-
Bromoalkenesulfonyl Chlorides. J Am Chem Soc. 1954;76(7):1926-1929.
doi:10.1021/ja01636a056.
19. Edetanlen-Elliot O. Vinyl sulfonates: a platform for novel substrates of biological
importance. 2011. http://discovery.ucl.ac.uk/1301991/.
20. Neal TR, Apana S, Berridge MS. Improved synthesis of [18F]fluoromethyl tosylate, a
convenient reagent for radiofluoromethylations. J Label Compd Radiopharm.
2005;48(8):557-568. doi:10.1002/jlcr.949.
21. Pascali G, Nannavecchia G, Pitzianti S, Salvadori PA. Dose-on-demand of diverse 18F-
fluorocholine derivatives through a two-step microfluidic approach. Nucl Med Biol.
2011;38(5):637-644.
22. Ory D, Van den Brande J, de Groot T, Serdons K, Bex M, Declercq L, Cleeren F, Ooms M, Van
Laere K, Verbruggen A, Bormans G, Retention of [(18)F]fluoride on reversed phase HPLC
columns. J Pharm Biomed Anal. 2015;111:209-14. doi: 10.1016/j.jpba.2015.04.009
23. Lapi SE, Welch MJ. A historical perspective on the specific activity of
radiopharmaceuticals: What have we learned in the 35 years of the ISRC? Nucl Med
Biol. 2012;39(5):601-608. doi:10.1016/j.nucmedbio.2011.11.005.
24. Dong J, Krasnova L, Finn MG, Barry Sharpless K. Sulfur(VI) fluoride exchange (SuFEx):
Another good reaction for click chemistry. Angew Chemie - Int Ed. 2014;53(36):9430-
9448. doi:10.1002/anie.201309399.
This article is protected by copyright. All rights reserved.

Table 1. Examples of [18F]ESF conjugation reactions
Time
AA[a]
Solvent[b]
[18F]ESF[c]
RCY (%)[d]
(min)
360
15
1
2
Trp (1)
Trp (10) DMSO (80)
DMSO (80)
20
20
41
64
20 + 0.25 eq
TEA
100 + 0.25
eq TEA
20
3
4
Trp (10) DMSO (80)
15
15
91
DMSO
Trp (10)
97[d]
(400)
5
6
7
Phe (1)
Phe (1)
DMF (80)
DMSO (80)
120
300
15
2
80
77
20
20
Phe (10) DMSO (80)
20 + 0.1 eq
TEA
20 + 0.1 eq
TEA
8
9
Phe (10) DMSO (80)
15
15
90
DMSO
Phe (10)
97[d]
(400)
[a] Concentrations in parentheses, expressed in mg/mL; [b]
Volumes in parentheses, expressed in µL; [c] Volumes expressed
in µL, c.a. for Trp, n.c.a. for Phe; [d] Assayed by RadioHPLC; [d]
RP reported.
This article is protected by copyright. All rights reserved.

Table 2. Summary of optimised large scale conjugation reactions.
Activity
(MBq)[d]
4.9
Substrates (solvent)
RCY (%)[c]
MA (MBq/nmol)[e]
c.a.
Ani (EtOH)
Cys (EtOH)
57
61
61
35
65
67
64
63
51
39
73
71
59
71
71
72
67
62
55
29
0.03
0.05
0.03
n/a
0.04
0.04
0.03
0.02
0.06
>11.23
>6.86
>1.59
n/a
>14.88
>32.04
>7.91
>15.73
>21.38
0.66[f]
0.03[f]
34
10
5.2
8.5
7
6.4
5.1
2.8
19.8
24.9
21.2
14.2
16
15.1
14.9
12.4
10.9
9.6
Gln (DMSO)[a]
Lys (DMSO)[a]
Pro (DMSO)
Phe (DMSO)[a]
Ser (DMSO)[a]
Trp (DMSO)[a]
Tyr (EtOH)
n.c.a.
Ani (EtOH)
Cys (DMSO)
Gln (DMSO)
Lys (DMSO)[b]
Pro (DMSO)
Phe (DMSO)[b]
Ser (DMSO)[b]
Trp (DMSO)[b]
Tyr (EtOH) [b]
BSA (saline)
Insulin (H2O)
4
[a] 0.25 eq of TEA added; [b] 0.1 eq of TEA added; [c] Isolated RCY calculated for single
conjugation reaction as: (Product activity)/(Starting activity), all RP were >95%; [d] Isolated
activity of final product; [e] Corrected to end of [18F]ESF synthesis; [f] effective MA calculated
as: (Activity)/(Protein amount).
This article is protected by copyright. All rights reserved.

Figure 1. Schematics of microfluidic system used for [18F]ESF production. c.a. [18F]ESF: The
outlet of the microfluidic system is connected to a neutral alumina SPE which removes the
unreacted [18F]fluoride. n.c.a. [18F]ESF: The outlet of the microfluidic system is connected to
a closed vial (A) heated at 80 °C and pressurized with N2 gas (0.1-0.2 L/min). The gaseous
effluents from this vial are then directed through silica light SPE (C) into a second vial (B).
The outlet of vial B is connected a Silica plus SPE (D) to trap the [18F]ESF.
This article is protected by copyright. All rights reserved.

Figure 2. Aniline and amino acid analogues radiolabelled in this work. The conjugation site
with [18F]ESF are highlighted in italic (red).
This article is protected by copyright. All rights reserved.

Figure 3. RadioHPLC profiles of c.a. [18F]ESF conjugation with Trp (A), and effect of TEA
addition (B)
This article is protected by copyright. All rights reserved.

Figure 4. Identification of [18F]Pro-ESF. The UV Rt of radioactive [18F]Pro-ESF (4.9 min, blue
solid line, inverted scale) is identical to the UV Rt of non-radioactive Pro-ESF standard
(purple dash line, inverted scale).
This article is protected by copyright. All rights reserved.

Figure 5. Stability results of selected n.c.a. [18F]ESF adducts in injectable formulations.
This article is protected by copyright. All rights reserved.

Scheme 1. Routes to the radiosynthesis of [18F]ESF. <18F−> represents a form of
radiofluorinating complex.
This article is protected by copyright. All rights reserved.

Graphical table of contents
We
[18F]ethenesulfonyl fluoride [18F](ESF), that proceeds
straightforwardly without the need of
report
herein the
radiosynthesis of
chromatographic purification. We then investigated
the use of [18F]ESF as a new prosthetic group on
amino acids, insulin and BSA. While the conjugation
methodology and the purification processes were
simple and efficient, the stabilities of the final products are strongly substrate-dependent, and
generally low in serum.
This article is protected by copyright. All rights reserved.