Electrochemical radiofluorination. 3. Direct labeling of phenylalanine derivatives with [18F]fluoride after anodic oxidation

Article abstract of DOI:10.1002/jlcr.918

As an example of applying the electrochemical method in radiofluorination a new procedure was developed for preparing [¹⁸F]fluorophenylalanine by direct use of [¹⁸F]fluoride. Several protecting groups for amino and carboxylic functio

Full text of DOI:10.1002/jlcr.918

JOURNAL OF LABELLED COMPOUNDS AND RADIOPHARMACEUTICALS
J Label Compd Radiopharm 2005; 48: 259–273.
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jlcr.918
Research Article
Electrochemical radiofluorination. 3. Direct labeling of
phenylalanine derivatives with [18F]fluoride after anodic
oxidation
.
Gabriele J. Kienzle, Gerald Reischl* and H.-Jurgen Machulla
.
.
PET-Zentrum, Radiopharmazie, Universitatsklinikum Tubingen, Germany
Summary
As an example of applying the electrochemical method in radiofluorination a new
procedure was developed for preparing [18F]fluorophenylalanine by direct use of
[18F]fluoride. Several protecting groups for amino and carboxylic function of
phenylalanine were investigated. Best results were obtained for N-trifluoroacetylphe-
nylalanine methyl ester (1) (10.5 ꢀ 2.5%) with a relative isomeric distribution of
50.5 ꢀ 1.5% ortho isomer, 11.5 ꢀ 1.5% meta isomer and 38.8 ꢀ 2.1% para isomer.
Optimized reaction conditions were with Et3N ꢁ 3HF as supporting electrolyte and
temperatures between 0 and –108C. By variation of working potential (1.5–2.0 V) and
chloride concentration (added as supporting electrolyte, 0.0–0.36 M) it was demon-
strated that an indirect electrolysis as reaction mechanism or a halogen exchange
(halex) reaction with primary formation of a chlorinated product followed by halogen
exchange are highly improbable. Copyright # 2005 John Wiley & Sons, Ltd.
Key Words: fluorophenylalanine; anodic fluorination; [18F]fluoride; nucleophilic
substitution; positron emission tomography
Introduction
Fluorine-18 labeled aromatic amino acids are considered to be of great value
for PET applications. This is especially true for 2- and 4-[18F]fluoropheny-
lalanine and 2-[18F]fluorotyrosine, which have been applied in tumor
detection1–4 and 6-[18F]fluorodopa and 6-[18F]fluoro-m-tyrosine which are
used measuring presynaptic dopaminergic function.5,6
The radiochemical syntheses of those radiopharmaceuticals are performed
mainly by electrophilic reactions using [18F]F2 as reagent. However, anionic
.
.
*Correspondence to: G. Reischl, PET-Center, Radiopharmacy, University Hospital Tubingen, Rontgenweg
.
15, D-72076 Tubingen, Germany. E-mail: gerald.reischl@uni-tuebingen.de
Contract/grant sponsor: Deutsche Forschungsgemeinschaft; contract/grant number: MA 1096/3-1
Copyright # 2005 John Wiley & Sons, Ltd.
Received 16 March 2004
Revised 17 June 2004
Accepted 20 November 2004

260
G. J. KIENZLE ET AL.
fluorine-18 can be produced more efficiently via the 18O(p,n)18F nuclear
reaction. In addition, that is the most common method for preparing no-
carrier-added radiotracers. Only a few nucleophilic labeling procedures exist
for aromatic radiopharmaceuticals. After the labeling step further reactions
and chiral purifications often are necessary. In general, these procedures rarely
are used, as they are time-consuming because of multi step syntheses and suffer
from low yields.7,8 Most recently, a catalytic enantioselective pathway was
described, demonstrating significant improvements.9
In aromatic compounds nucleophilic substitution generally is hampered by
the high electron density of the benzene ring activated by +M effect as in the
case of DOPA or tyrosine. In such cases, nucleophilic attack of fluoride is
difficult or even impossible. Substituents with electron withdrawing effects
enable the nucleophilic introduction of fluoride into the aromatic moiety.
Unfortunately, this application is limited.
By the methods of electrochemistry, an electron can be removed from the
phenylic ring, fluoride can easily react with the cation formed, and electrode
potentials can be selected for the precise oxidation of a wide range of aromatic
hydrocarbons. Thus, reactions are more selective when compared to reactions
with chemical oxidants.10,11
The first electrochemical fluorination of benzene and various substituted
benzenes with [18F]fluoride were recently reported by our group.12,13 Aim of
the present study was to develop a new method for preparing aromatic PET
radiopharmaceuticals by electrochemical 18F-fluorination using [18F]fluoride
and phenylalanine as a radiopharmaceutical example.
Results and discussion
This is the first literature report on the electrochemical synthesis of
[18F]fluorophenylalanine using [18F]fluoride (Figure 1). Several protecting
groups were applied for the amino group (acetyl, trifluoroacetyl, butoxycar-
bonyl and 2,4-dinitrophenyl) and the carboxylic group (methyl and ethyl). For
each of the derivatives 1–5 oxidation potentials were determined through
cyclic voltammetry. Interestingly, oxidation potentials depended on the kind
of protecting group (Table 1).
In the electrochemical 18F-labeling reactions of the derivatives 1–5 the
labeling yields also depended on the protecting groups (Table 1). Remarkably,
the phenylalanine derivative with the highest oxidation potential gave the best
radiochemical yield (1, 5.8 ꢀ 2.8%, total of o-, m- and p-isomers). Because the
aromatic ring and the protecting groups are far apart from each other,
inductive or mesomeric effects at the phenyl ring can be excluded to influence
yields of 18F-labeling. Most likely, the results are due to effects of
intermolecular forces and interactions on the surface of the anode. For 1
the isomeric distribution was determined with GC/MS (50.5 ꢀ 1.5% for ortho
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 259–273

ELECTROCHEMICAL RADIOFLUORINATION. 3
261
COOY
NHX
COOY
-2e , -H (-tBu
+ [18F]F
)
R
CH2 CH
R
CH2 CH
NHX
18F
1 - 6f
R
H
X
Y
1
2
TFA
TFA
Boc
Ac
Me
Et
H
3
H
Me
Me
Me
Me
4
H
5
H
DNP
TFA
6f
tBu
Figure 1. Generalized reaction scheme for the anodic oxidation of phenylalanine
derivatives and subsequent nucleophilic substitution with [18F]fluoride (TFA =
trifluoroacetyl, Boc = tert-butoxycarbonyl, DNP = 2,4-dinitrophenyl, Ac =
acetyl, Me = methyl, Et = ethyl)
Table 1. Oxidation potentials from cyclovoltammetry and radiochemical 18F-fluorina-
tion yields from phenylalanine derivatives 1–6f at 50 C in electrolyte A at 08C
Phenylalanine derivative
Oxidation potential (V)
Radiochemical yield (%)
n
1
2
3
4
5
6f
2.15 ꢀ 0.01
2.13 ꢀ 0.01
2.04 ꢀ 0.02
1.90 ꢀ 0.01
1.65 ꢀ 0.01
1.89 ꢀ 0.01
5.8 ꢀ 2.8
2.6 ꢀ 1.1
0.6 ꢀ 0.3
1.6 ꢀ 0.9
0.1 ꢀ 0.05
1.6 ꢀ 1.0
10
3
4
4
3
5
isomer, 11.5 ꢀ 1.5% for meta isomer and 38.8 ꢀ 2.1% for para isomer). The
observed selectivity indicated a reaction mechanism involving a phenylic
carbenium ion.
In our previous experiments with various substituted benzenes, radio-
chemical yields depended on the oxidation potential.13 Yields increased with
decreasing potential. As tert-butyl benzene gave the highest 18F-fluorination
yields on exchange of tert-butyl vs fluorine, we also investigated a tert-butyl
substituted phenylalanine derivative 6f. The molecule had to be built up in a
longer reaction sequence (Figure 2). Indeed, that substrate had a lower
oxidation potential due to its higher electron density in the aromatic ring as
consequence of the +I effect (Table 1). However, 6f turned out to be unstable
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 259–273

262
G. J. KIENZLE ET AL.
COOEt
1. NaOH
2. H2SO4
KI
K2CO3
H
C
NH(CHO)
+
COOEt
CH2
Cl
CH2
CH2
C(COOEt)2
C(COOH)2
NH(CHO)
NH(CHO)
6a
6b
aniline
-CO2
MeOH
SOCl2
CH2
CH2
CH2
H
COOMe
H
COOH
C
H
COOH
C
C
NH3Cl
NH2
NH(CHO)
6e
6d
6c
(F3CCO)2O
CH2
C
H
H
COOMe
COCF3
N
6f
Figure 2. Reaction scheme for the synthesis of 6f
under the electrochemical conditions and radiochemical yields were low
(1.6 ꢀ 1.0%). The relative amount of para isomer was somewhat, but not
significantly, higher (44 ꢀ 8%), while the ortho and meta isomers together
amounted to 56 ꢀ 8%.
When performing electrolyses in organic solvents, supporting electrolytes
are necessary for the transport of charge (Table 2). In order to optimize the
reaction the effect of different anions in the electrolyte was studied (electrolyte
B). For the different electrolytes the reaction times varied for certain amounts
of charge due to the conductivity of the solution and partly decomposition of
the supporting electrolyte at the high potential applied. Experiments with
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 259–273

ELECTROCHEMICAL RADIOFLUORINATION. 3
263
Table 2. Composition of electrolytes A–C, concentrations in mol/l in acetonitrile. In
addition to the salts listed, all solutions contained 0.033 mol/l of Et3N ꢁ 3 HF
Electrolyte A
Electrolyte B
Et3N ꢁ HCl
0.067
}
}
Et3N ꢁ 3 HF
0.022
(= 0.067 mol/l fluoride)
Et4NX (X = Cl, Br, I)
0.067
Bu4NPF6
0.067
}
}
Electrolyte C
Et4NCl
0.000
0.018
0.061
0.180
0.360
Bu4NPF6
0.400
0.382
0.339
0.220
0.040
Table 3. Radiochemical yields of fluorination of 1 with electrolytes A or B (0.067 M
with 0.033 M Et3N ꢁ 3 HF) as a function of charge [C]
Supporting electrolyte
Radiochemical yield (%)
25 C
n
50 C
Et3N ꢁ HCl
Et3N ꢁ 3HF
Et4NCl
Et4NBr
Et4NI
Bu4NPF6
3.4 ꢀ 1.6
10.5 ꢀ 2.5
2.9 ꢀ 0.7
1.5 ꢀ 0.6
0.01 ꢀ 0.01
2.9 ꢀ 1.4
6.4 ꢀ 3.7
}
4.2 ꢀ 1.6
3.0 ꢀ 1.2
0.3 ꢀ 0.2
}
7
5
5
5
3
3
Bu4NPF6 and Et3N ꢁ 3 HF were stopped after a charge of 25 Coulombs [C], all
other reactions after 50 C (reaction times were 60 to 90 min).
Variation of halogen showed that from chloride to iodide the supporting
electrolyte is oxidized more easily and in the same order radiochemical yields
were observed to decrease (Table 3). That may be due to an increasing
competition between the halogen ion and the aromatic substrate respecting
oxidation and adsorption on the electrode. Bu4NPF6 has an oxidation
potential of 3.02 V vs Ag/Ag+.14 Thus, this salt cannot be oxidized at the
working potential used, and a competition can be excluded between
supporting electrolyte and aromatic substrate. Indeed, fewer side reactions
were observed. Electrolyses with Bu4NPF6 as supporting electrolyte took
about 100 min for a charge of 25 C, in contrast to those reactions with Et4NCl
or Et3N ꢁ HCl, which required only 45 min for the same charge. Thus, in these
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 259–273

264
G. J. KIENZLE ET AL.
Table 4. Dependence of radiochemical yields on reaction temperature at 50 C for the
fluorination of 1
Temperature (8C)
Radiochemical yield (%)
n
25
0
ꢂ10
ꢂ20
3.2 ꢀ 2.1
5.8 ꢀ 2.8
5.5 ꢀ 1.5
4.2 ꢀ 0.1
6
10
4
3
cases it was possible to extend electrolyses to 50 C to obtain higher
radiochemical yields (e.g. Et3N ꢁ HCl at 50 C: 6.4 ꢀ 3.7%). Applying
Et3N ꢁ HCl the radiochemical yields were higher than with Et4NCl, demon-
strating that in the supporting electrolyte the protons did not reduce
nucleophilicity of [18F]fluoride. Best results were obtained with Et3N ꢁ 3HF
(10.5 ꢀ 2.5%, sum of isomers). This was not unexpected because higher
substrate concentrations often result in higher yields.
As shown in Table 4 the temperature effect was not significant. The optimal
temperatures were between ꢂ108C (5.5 ꢀ 1.5%, radiochemical yield) and 08C
(5.8 ꢀ 2.8%).
Reaction mechanism
For electrochemical fluorination of aromatic compounds the postulated
mechanism is an ECNECB process,15 with the following reaction steps. After
diffusion to the anode an aromatic substrate is oxidized to a radical cation
(electrochemical step E). In the following chemical step (CN) fluoride reacts
with the radical cation. After a second oxidation (electrochemical step E) and
loss of a proton (chemical step CB), the fluorinated product is formed.
If the electrolyte contains chloride or bromide ions, an indirect electrolysis is
conceivable as an alternative reaction mechanism. Thus, chloride for example,
is oxidized at the working potential and the chlorine then formed oxidizes the
aromatic substrate whereby chlorine itself is reduced to chloride again. In that
case, the working potential only has to be in the range of the oxidation
potential of chloride. That reaction path will result in fewer side reactions.
Another alternative mechanism could be a halex reaction, where, after primary
formation of a chlorinated product halogen exchange follows.
In our experiments it was possible to detect chlorinated products with GC/
MS. But with halex reactions the aromatic ring normally has to be activated
through s- or p-electron withdrawing groups.16 For that reason, this
mechanism seemed to be unlikely. Variation of working potentials and
chloride concentrations (added as supporting electrolyte) were used to
investigate which of the two mechanisms occurred.
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 259–273

ELECTROCHEMICAL RADIOFLUORINATION. 3
265
With cyclic voltammetry the oxidation potential of triethylamine hydro-
chloride was determined to be 0.65 V in electrolyte A. Working potential in
labeling reactions was reduced step by step from 2.0 to 1.5 V. In this range,
potentials are high enough to oxidize chloride ions of the supporting
electrolyte. At 1.7 and 1.5 V only a small amount of amino acid should be
oxidized at the electrode. In case the dominating mechanism is an indirect
electrolysis, radiochemical yields at 2.0 V should not decrease remarkably with
decreasing working potential. However, experiments showed labeling yields to
decrease significantly with decreasing potential (2.0 V:5.8 ꢀ 2.8%,
1.5 V:1.8 ꢀ 1.2%). Those results, at least, do not support the hypothesis of
an indirect mechanism as the dominating reaction pathway.
GC/MS investigations of product solutions showed that chlorinated
phenylalanine derivatives were formed to some extent. In the case of a halex
mechanism with a F- for Cl- exchange at primarily formed N-trifluoroacetyl-2-,
3- or 4-chloro-phenylalanine methyl esters, decreasing chloride concentrations
would be expected to decrease labeling yields.
Chloride concentrations in the electrolyte (electrolyte C) were reduced from
an excess over the aromatic substrate to a catalytic amount. In addition,
experiments without any chloride were performed. The reactions lasted for
15–110 min (high to low chloride concentration) until a charge of 25 C passed.
Decreasing chloride concentration was compensated by adding Bu4NPF6, so
that the overall concentration of electrolyte was 0.4M, constantly. All experi-
ments demonstrated that with decreasing chloride concentration radiochemical
yields increased clearly from 0.1 ꢀ 0.01% (0.36M of chloride) to 9.1 ꢀ 0.8%
(no chloride) (Table 5), indicating a halex reaction to be highly unlikely.
Experimental
General
Chemicals. Acetonitrile (Merck, Germany) was purified according to Sal-
beck17 by drying over CaCl2 for at least one week, distillation over P2O5, NaH
Table 5. Variation of chloride concentrations in electrosynthesis of N-trifluoroacetyl-2-,
3- or 4-[18F]fluoro-phenylalanine methyl esters and influence on radiochemical yields at
25 C (n=4)
Concentration
Ratio
Radiochemical yield (%)
Et4NCl (mol/l) Bu4NPF6 (mol/l) Aromatic substrate Chloride
0.000
0.018
0.061
0.180
0.360
0.400
0.382
0.339
0.220
0.040
1
10
3
1
1
0
1
1
1
2
9.1 ꢀ 0.8
5.8 ꢀ 0.7
1.8 ꢀ 0.4
0.1 ꢀ 0.04
0.1 ꢀ 0.01
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 259–273

266
G. J. KIENZLE ET AL.
and P2O5 and final chromatography with neutral Al2O3 which was dried in
vacuum at 1408C for about 2 days. Et3N ꢁ HCl (Merck-Schuchard, Germany)
was twice recrystallized from methanol and dried in vacuum at 408C for 2–3
days. Et4NCl, Et4NBr and Et4NI (all Fluka, Germany) were twice
recrystallized from acetonitrile under argon atmosphere and dried in vacuum
at 1008C for 2–3 days. Bu4NPF6 (Aldrich, Germany) was twice recrystallized
from a mixture of ethanol and water 3:1, washed with methanol and dried in
vacuum at 1008C for 2–3 days. Bu4NClO4 (Fluka, electrochemical grade) was
dried in vacuum at 1208C for 2–3 days. Et3N ꢁ 3 HF (Fluka) was distilled twice
under vacuum (808C, 5 Torr). AgNO3 (Roth, Germany) and all other
chemicals and solvents (Merck or Fluka) were of the highest purity available
and used as received.
Analytical procedures. NMR spectroscopy: Bruker AC 250 (1H: 250.1 MHz,
13C: 62.9 MHz) with tetramethylsilane as an internal standard. Mass spectro-
metry: Varian MAT 711 (70 eV), IR spectroscopy: Bruker IFS 48, melting
points: Gallenkamp MPG 350. Microanalyses: Chemisches Zentralinstitut,
.
University of Tubingen.
Phenylalanine derivatives
N-trifluoroacetyl-l-phenylalanine methyl ester (1) was either purchased
(Sigma, Germany) or synthesized from the unprotected amino acid,18
N-trifluoroacetyl-l-phenylalanine ethyl ester (2) was prepared analogously.
N-butoxycarbonyl-l-phenylalanine methyl ester (3) was from Aldrich and
N-acetyl-l-phenylalanine methyl ester (4) from Novachem (Germany). N-2,4-
dinitrophenyl-l-phenylalanine methyl ester (5) was synthesized according to
Fletcher.19 6a–6e were prepared analogously to tyrosine derivatives described
in literature.20–22 All phenylalanine derivatives 1–6f were dried in vacuum at
room temperature for 2–3 days before their use in electrochemistry. 2-, 3- and
4-fluoro-dl-phenylalanine were purchased (Fluka), protected as described
below and used as references.
Since the NMR data of many of the compounds prepared in this study have
not been reported at all, or were essentially incomplete, those data are
presented below.
1: The obtained yellow brown oil was distilled twice under reduced pressure,
resulting in colorless crystals. Yield: 11% (2.81 g), m.p.: 47–528C, b.p.: 1128C
1
at 3–4 Torr. MS (EI, m/z): 275.1 (M+), 162.0, 91.1, 131.1. H-NMR (CDCl3,
d[ppm]):3.15(1H, dd, J ¼ 5:5 Hz, 14.0 Hz), 3.24 (1H, dd, J ¼ 5:8 Hz, 14.0 Hz),
3.77 (3H, s), 4.88 (1H, ddd, J ¼ 5:5 Hz, 5.8 Hz, 7.6 Hz), 6.85 (1H, br d, NH),
7.07 (2H, dd, J ¼ 1:5 Hz, 7.6 Hz), 7.25–7.35 (3H, m). 13C-NMR (CDCl3, d
[ppm]): 37.3, 52.7, 53.5, 127.6, 128.8, 129.1, 134.8, 170.4, 115.5 (q,
J ¼ 288:0 Hz), 156.5 (q, J ¼ 37:2 Hz). IR (KBr [cmꢂ1]): 3292, 1751, 1703,
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 259–273

ELECTROCHEMICAL RADIOFLUORINATION. 3
267
1555, 1441, 1331, 1277. Calculated for C12H12F3NO3: C 52.37%, H 4.39%, N
5.09%, F 20.71%. Found: C 52.67%, H 4.31%, N 5.19%, F 20.36%.
2: The obtained orange oil could not be distilled. Upon addition of CH2Cl2
(5%) in n-hexane the product precipitated and was recrystallized. Yield: 15%
(4.8 g), m.p.: 70–748C. MS (EI, m/z): 289.1 (M+), 175.9, 91.1, 131.1. 1H-NMR
(CDCl3, d[ppm]): 1.23 (3H, t, J ¼ 7:0 Hz), 3.09 (1H, dd, J ¼ 5:8 Hz, 14.0 Hz),
3.19 (1H, dd, J ¼ 5:5 Hz, 14.0 Hz), 4.18 (2H, q, J ¼ 7:0 Hz), 4.81 (1H, ddd,
J ¼ 5:5 Hz, 5.8 Hz, 7.3 Hz), 6.85 (1H, br d, NH), 7.05 (2H, dd, J ¼ 1:5 Hz,
6.7 Hz), 7.21–7.29 (3H, m). 13C-NMR (CDCl3, d[ppm]): 14.0, 37.3, 53.5, 62.2,
127.5, 128.7, 129.2, 134.6, 170.0, 115.6 (q, J ¼ 288:0 Hz), 156.5 (q,
J ¼ 37:2 Hz). IR (KBr [cmꢂ1]): 3310, 1747, 1707, 1562, 1373, 1281, 1240,
1205, 1175. Calculated for C13H14F3NO3: C 53.98%, H 4.88%, N 4.84%, F
19.70%. Found: C 53.56%, H 4.90%, N 4.84%, F 20.03%.
5: The residue was recrystallized twice from ethanol. Yellow crystals were
obtained. Yield: 38% (1.2 g), m.p.: 113–1188C. MS (EI, m/z): 345.1 (M+),
1
286.1, 254.0, 193.8, 165.8, 91.1. H-NMR (CDCl3, d[ppm]): 3.22 (1H, dd,
J ¼ 7:3 Hz, 14.0 Hz), 3.36 (1H, dd, J ¼ 5:5 Hz, 14.0 Hz), 3.79 (3H, s), 4.58
(1H, ddd, J ¼ 5:5 Hz, 7.3 Hz, 7.3 Hz), 6.67 (1H, d, J ¼ 9:5 Hz), 7.19 (2H, dd,
J ¼ 1:8 Hz, 7.6 Hz), 7.25–7.36 (3H, m), 8.17 (1H, dd, J ¼ 2:4 Hz, 9.5 Hz), 8.87
(1H, d, J ¼ 7:3 Hz, NH), 9.08 (1H, d, J ¼ 2:4 Hz). 13C-NMR (CDCl3,
d[ppm]):38.4, 52.8, 57.4, 113.8, 124.0, 127.7, 128.9, 129.0, 130.1, 130.9, 134.5,
136.6, 147.0, 170.5. IR (KBr [cmꢂ1]): 3337, 1730, 1614, 1587, 1522, 1339, 1150,
1107. Calculated for C16H15N3O6: C 55.65%, H 4.38%, N 12.17%. Found: C
55.35%, H 4.27%, N 12.09%.
(4-tert-Butyl-benzyl)-formamidomalonic acid diethyl ester (6a). 4-tert-butyl-
benzylchloride (24 ml; 0.12 mol) and 25.6 g (0.12 mol) of formamidomalonic
acid diethyl ester were dissolved in 660 ml of acetone. Dry K2CO3 (220 g)
and half a tea-spoon of KI were added and the solution refluxed for 21 h.
After filtration of K2CO3 and evaporation of acetone, the residue was
dissolved in water and extracted with CH2Cl2. The combined organic extracts
were washed with water and dried over Na2SO4. After evaporation a yellow
brown oil was obtained. n-Hexane (160 ml) were added and the warmed
solution was shaken strongly. The yellow needles formed were filtered and
dried in vacuum at 408C for a few hours. Yield: 92% (38.8 g), m.p.: 99–1018C.
1
MS (EI, m/z): 349.0 (M+), 304.0, 289.0, 147.0. H-NMR (CDCl3, d[ppm]):
1.28 (9H, s), 1.29 (6H, t, J ¼ 7:1 Hz), 3.63 (2H, s), 4.27 (4H, q, J ¼ 7:1 Hz),
6.77 (1H, br s, NH), 6.96 (2H, d, J ¼ 8:4 Hz), 7.26 (2H, d, J ¼ 8:4 Hz), 8.15
(1H, s). 13C-NMR (CDCl3, d[ppm]): 14.0, 31.3, 34.4, 37.4, 62.8, 66.8, 125.3,
129.6, 131.7, 150.2, 159.9, 167.2. IR (KBr [cmꢂ1]): 3229, 2959, 2905, 2868,
1744, 1655, 1518, 1377, 1366, 1278, 1250, 1188, 1055. Calculated for
C19H27NO5: C 65.31%, H 7.79%, N 4.01%. Found: C 65.31%, H 7.79%,
N 4.02%.
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 259–273

268
G. J. KIENZLE ET AL.
(4-tert-Butyl-benzyl)-formamidomalonic acid (6b). 6a (38.3 g; 0.11 mol) was
dissolved in 600 ml of ethanol. NaOH (34.8 g) in 240 ml of water was added
and the solution refluxed for 2 h under inert gas atmosphere. Acetic acid was
added until pH 6 and H2SO4 (13%) until pH 2. (About 50 ml of acetic acid and
270 ml of H2SO4 were necessary.) After evaporation of about 600 ml of solvent
and addition of 200 ml of water, the solution had to be cooled for a few hours.
The white crystals formed were filtered, washed with water and dried in
vacuum at 708C. Yield: 90% (29.3 g), m.p.: 187–1908C. MS (EI, m/z): 249.0,
204.0, 189.0, 147.0, MS (FAB, NBA, m/z): 293.9 (M++H). 1H-NMR
(DMSO-d6, d[ppm]):1.23 (9H, s), 3.36 (2H, s), 6.95 (2H, d, J ¼ 8:2 Hz), 7.26
(2H, d, J ¼ 8:2 Hz), 7.97 (1H, s), 8.21 (1H, br s, NH), 13.40 (2H, br s). 13C-
NMR (DMSO-d6, d[ppm]): 31.2, 34.1, 36.9, 66.1, 124.8, 129.5, 132.8, 148.9,
160.7, 168.9. IR (KBr [cmꢂ1]): 3368, 2963, 2905, 2868, 1730, 1628, 1497, 1373.
Calculated for C15H19NO5: C 61.42%, H 6.53%, N 4.78%. Found: C 60.00%,
H 6.10%, N 4.64%.
4-tert-butyl-N-formyl-dl-phenylalanine (6c). 6b (29.3 g; 0.1 mol) was sus-
pended in 800 ml of petroleum ether 110/140 and refluxed for 3 h. The
product was filtered, washed with petroleum ether, recrystallized from
methanol and dried in vacuum at room temperature for some hours. Yield:
1
81% (17.5 g), m.p.: 186–1928C. MS (FAB, NBA, m/z): 250.1 (M++H). H-
NMR (CD3OD, d[ppm]):1.15 (9H, s), 2.48 (1H, dd, J ¼ 7:9 Hz, 14.0 Hz), 3.04
(1H, dd, J ¼ 5:2 Hz, 14.0 Hz), 4.60 (1H, dd, J ¼ 5:2 Hz, 7.9 Hz), 7.01 (2H, d,
J ¼ 8:2 Hz), 7.18 (2H, d, J ¼ 8:2 Hz), 7.88 (1H, s). 13C-NMR (CD3OD,
d[ppm]): 31.8, 35.2, 37.9, 53.6, 126.3, 130.3, 134.8, 150.8, 163.4, 174.2. IR (KBr
[cmꢂ1]): 3348, 2963, 2905, 2868, 1720, 1614, 1522, 1366, 1360. Calculated for
C14H19NO3: C 67.45%, H 7.68%, N 5.62%. Found: C 67.12%, H 7.86%, N
5.66%.
4-tert-Butyl-dl-phenylalanine (6d). 6c (9 g; 0.036 mol) was heated with 60 ml of
freshly distilled aniline at 90–958C for 1 h. After cooling, 80 ml of ether were
added. The crystals formed were filtered and washed with ether and ethanol.
Yield: 39% (3.3 g), m.p.: 255–2578C. MS (EI, m/z): 221.2 (M+), 204.2, 189.1,
147.0, MS (FAB, NBA, m/z):222.1 (M++H). 1H-NMR (CD3COOD,
d[ppm]): 1.25 (9H, s), 3.09 (1H, dd, J ¼ 8:2 Hz, 14.7 Hz), 3.28 (1H, dd,
J ¼ 4:9 Hz, 14.7 Hz), 4.28 (1H, dd, J ¼ 4:9 Hz, 8.2 Hz), 7.19 (2H, d,
J ¼ 7:9 Hz), 7.30 (2H, d, J ¼ 7:9 Hz). 13C-NMR (CD3COOD, d[ppm]): 31.6,
35.0, 36.7, 57.0, 126.6, 130.4, 132.7, 151.3, 174.8. IR (KBr [cmꢂ1]): 3450, 2963,
2905, 2868, 1624, 1587, 1512, 1360. Calculated for C13H19NO2: C 70.56%, H
8.65%, N 6.33%. Found: C 70.14%, H 8.45%, N 6.26%.
4-tert-Butyl-dl-phenylalanine methyl ester hydrochloride (6e). 6d (3 g;
0.014 mol) was dissolved in 35 ml of absolute methanol and cooled to
ꢂ788C. Freshly distilled thionyl chloride (2.5 ml; 0.034 mol) was added.
Thereby, temperature should not exceed 08C. The solution was stirred at room
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 259–273

ELECTROCHEMICAL RADIOFLUORINATION. 3
269
temperature for 48 h and then, evaporated to dryness. The residue was
dissolved in a minimal amount of methanol and insoluble particles were
filtered. After addition of dry ether white crystals were obtained. They were
filtered and dried in vacuum at room temperature. Yield: 62% (2.4 g), m.p.:
1
210–2148C. MS (EI, m/z): 236.2, 176.2, 147.2, 133.2, 88.1. H-NMR (DMSO-
d6, d[ppm]): 1.26 (9H, s), 3.06 (1H, dd, J ¼ 7:3 Hz, 14.4 Hz), 3.15 (1H, dd,
J ¼ 6:1 Hz, 14.4 Hz), 3.67 (3H, s), 4.21 (1H, dd, J ¼ 6:1 Hz, 7.3 Hz), 7.15 (2H,
d, J ¼ 8:2 Hz), 7.33 (2H, d, J ¼ 8:2 Hz), 8.71 (3H, br s, NH+3 ). 13C-NMR
(DMSO-d6, d[ppm]): 31.1, 34.2, 35.3, 52.5, 53.2, 125.3, 129.1, 131.6, 149.6,
169.4. IR (KBr [cmꢂ1]): 3450, 2959, 2905, 2868, 1747, 1495, 1364, 1236.
Calculated for C14H21NO2 ꢁ HCl: C 61.87%, H 8.16%, N 5.15%, Cl 13.04%.
Found: C 62.09%, H 8.74%, N 5.28%, Cl 12.76%.
N-Trifluoroacetyl-4-tert-butyl-dl-phenylalanine methyl ester (6f). The yellow
brown oil was purified by chromatography with silica gel (n-hexane/ethyl
acetate). Yield: 62% (1.7 g). MS (EI, m/z): 331.2 (M+), 218.1, 203.1, 147.1. 1H-
NMR (CDCl3, d[ppm]): 1.30 (9H, s), 3.14 (1H, dd, J ¼ 5:5 Hz, 14.0 Hz), 3.21
(1H, dd, J ¼ 5:5 Hz, 14.0 Hz), 3.79 (3H, s), 4.87 (1H, dt, J ¼ 5:5 Hz, 7.6 Hz),
6.81 (1H, br d, NH), 6.99 (2H, dd, J ¼ 1:8 Hz, 6.4 Hz), 7.32 (2H, dd,
J ¼ 1:8 Hz, 6.4 Hz). 13C-NMR (CDCl3, d[ppm]): 31.2, 34.5, 36.7, 52.7, 53.5,
125.7, 128.8, 131.4, 150.5, 170.4, 115.6 (q, J ¼ 288:0 Hz), 156.5 (q,
J ¼ 37:2 Hz). IR (film [cmꢂ1]): 3327, 2963, 2903, 2870, 1749, 1722, 1553,
1394, 1366, 1215, 1175. Calculated for C16H20F3NO3: C 58.00%, H 6.08%,
N 4.23%. Found: C 58.02%, H 5.87%, N 4.43%.
Syntheses of reference substances
2-, 3- and 4-fluoro-dl-phenylalanine were first derivatized to the methyl ester
hydrochloride as described for 6e and then to the N-trifluoroacetyl-2-, 3- and
4-fluoro-dl-phenylalanine methyl esters 7a–c analogously to the procedure
for 1.
N-Trifluoroacetyl-2-fluoro-dl-phenylalanine methyl ester (7a). Yield: 13%
(90 mg), m.p.: 49–538C (from n-hexane). MS (EI, m/z): 293.1 (M+), 234.0,
1
179.9, 109.0, 149.0. H-NMR (CDCl3, d[ppm]):3.12 (1H, dd, J ¼ 5:8 Hz,
14.0 Hz), 3.26 (1H, dd, J ¼ 5:5 Hz, 14.0 Hz), 3.72 (3H, s), 4.80 (1H, ddd,
J ¼ 5:5 Hz, 5.8 Hz, 7.6 Hz), 6.87 (1H, br d, NH), 6.93–7.05 (3H, m.), 7.15–7.21
(1H, m.). 13C-NMR (CDCl3, d[ppm]): 31.0, 52.8, 52.9, 170.3, 115.5 (q,
J ¼ 288:0 Hz), 115.5 (d, J ¼ 21:9 Hz), 121.7 (d, J ¼ 15:3 Hz), 124.4 (d,
J ¼ 2:9 Hz), 129.6 (d, J ¼ 7:6 Hz), 131.5 (d, J ¼ 4:8 Hz), 156.6 (q,
J ¼ 38:2 Hz), 161.3 (d, J ¼ 245:1 Hz). IR (KBr [cmꢂ1]): 3333, 3098, 2959,
1760, 1722, 1553, 1495, 1225, 1178, 760.
N-Trifluoroacetyl-3-fluoro-dl-phenylalanine methyl ester (7b). Yield: 53%
(367 mg), m.p.: 75–798C (from n-hexane). MS (EI, m/z):293.0 (M+), 233.9,
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 259–273

270
G. J. KIENZLE ET AL.
1
179.9, 109.1, 149.0. H-NMR (CDCl3, d[ppm]):3.15 (1H, dd, J ¼ 5:8 Hz,
14.0 Hz), 3.26 (1H, dd, J ¼ 5:5 Hz, 14.0 Hz), 3.81 (3H, s), 4.88 (1H, ddd,
J ¼ 5:5 Hz, 5.8 Hz, 7.3 Hz), 6.76–7.03 (4H, m, 3Harom.+NH), 7.24–7.33 (1H,
m.). 13C-NMR (CDCl3, d[ppm]): 37.0, 52.9, 53.4, 170.2, 114.6 (d, J ¼ 21:0 Hz),
115.5 (q, J ¼ 288:0 Hz), 116.1 (d, J ¼ 21:9 Hz), 124.9 (d, J ¼ 2:9 Hz), 130.3 (d,
J ¼ 8:6 Hz), 137.1 (d, J ¼ 6:7 Hz), 156.6 (q, J ¼ 38:2 Hz), 166.9 (d,
J ¼ 247:0 Hz). IR (KBr [cmꢂ1]): 3296, 3099, 3071, 2960, 1753, 1726, 1560,
1450, 1439, 1225, 1178.
N-Trifluoroacetyl-4-fluoro-dl-phenylalanine methyl ester (7c). Yield: 36%
(250 mg), m.p.: 73–768C (from n-hexane). MS (EI, m/z): 293.1 (M+), 234.0,
1
179.9, 109.1, 149.0. H-NMR (CDCl3, d[ppm]):3.14 (1H, dd, J ¼ 5:5 Hz,
14.0 Hz), 3.24 (1H, dd, J ¼ 5:8 Hz, 14.0 Hz), 3.79 (3H, s), 4.86 (1H, ddd,
J ¼ 5:5 Hz, 5.8 Hz, 7.3 Hz), 6.87 (1H, br d, NH), 6.98–7.06 (4H, m). 13C-NMR
(CDCl3, d[ppm]): 36.5, 52.9, 53.6, 170.3, 115.5 (q, J ¼ 288:0 Hz), 115.8 (d,
J ¼ 21:0 Hz), 130.3 (d, J ¼ 0:9 Hz), 130.7 (d, J ¼ 8:6 Hz), 156.5 (q,
J ¼ 38:2 Hz), 162.3 (d, J ¼ 247:0 Hz). IR (KBr [cmꢂ1]): 3310, 3090, 2960,
1751, 1726, 1568, 1510, 1441, 1205, 1165.
Cyclic voltammetry
Cyclic voltammetry was carried out in a 20 ml undivided glass cell with
platinum working and counter electrodes and a platinum and Ag/Ag+ double
reference electrode at room temperature. The solution for the reference
electrode contained AgClO4 (0.01 M) and Bu4NPF6 (0.1 M) in acetonitrile. As
electrolyte and salt bridge a solution of Bu4NPF6 (0.1 M) in acetonitrile was
used. The electrolyte was degassed with argon for 30 min. The measurements
were performed at room temperature and between the measurements the
solution was stirred magnetically. The potentiostat was a BAS 100 B/W (BAS,
Antwerp, Belgium). The BAS 100 W software was used. All cyclovoltammo-
grams of substrates were baseline corrected.
[18F]Fluoride production
[18F]Fluoride was produced through the 18O(p,n)18F nuclear reaction at a
PETtrace cyclotron (Ep=16.5 MeV, General Electric Medical Systems,
Sweden) in a high pressure silver target. [18O]H2O (1.5 ml; Rotem, Germany,
94–96% enrichment) was irradiated. The [18F]fluoride was trapped
on a pre-conditioned strong ion exchange cartridge (Waters Accel QMA
Plus, Waters, Germany), rinsed with 4 ml of acetonitrile for removing
water and eluted with 2 ml of electrolyte. Because the anion exchange
was performed by means of chloride, for the experiments with variable
chloride concentrations the cartridge was rinsed with 10 ml of KF
solution (0.4 M in water) for removing the chloride before trapping the
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 259–273

ELECTROCHEMICAL RADIOFLUORINATION. 3
271
[18F]fluoride. After rinsing with 3.5 ml of KF solution, chloride detection was
negative.
Electrochemical oxidation
Electrochemical oxidations were carried out in an undivided three-electrode
cell with platinum wire working (anode) and counter (cathode) electrodes
and an Ag/Ag+ reference electrode (Ag/0.01 M AgNO3 in acetonitrile). The
cell was already described elsewhere.12 Electrolyses were carried out
potential controlled with a Wenking STP 96 potentiostat together with a
Wenking EVI 95 current integrator (both Bank Elektronik GmbH, Germany).
A working potential of 2.0 V was used, in case of experiments with variable
working potentials the voltage was between 1.5 and 2.0 V. Periodical pulsing
between 2.0 and 0 V was necessary to minimize formation of a layer covering
the anode.
Bu4NClO4 was added to the solution of the reference electrode as
supporting electrolyte (either 0.1 or 0.4 M). A solution of Bu4NClO4 in
acetonitrile was applied as salt bridge (either 0.1 or 0.4 M). Solutions with the
higher concentrations were used in experiments with variable chloride
concentrations, as the electrolyte had the same concentration then. In all
other cases the lower concentrated electrolytes were applied.
As electrolyte, three different solutions were applied (Table 2), electrolyte A
for experiments with the different phenylalanine derivatives or varying
temperature or working potentials, electrolyte B for experiments with different
supporting electrolytes and electrolyte C for investigations with variation of
chloride concentration.
Syntheses were carried out under argon atmosphere. After the electrolyte
(2 ml) was placed in the cell, the phenylalanine derivative (0.4 mmol) was
added, the solution stirred magnetically and cooled down to 08C with a
cryostate (Lauda ecoline RE 104, Lauda GmbH & Co. KG, Germany). In
experiments with varying temperatures these ranged from ꢂ20 to +258C.
Except for investigations with various phenylalanine derivatives, 1 was used.
After certain amount of charge had passed, samples were taken and analyzed.
Electrolyses except for those with variation of supporting electrolyte and
chloride concentration were stopped after 50 C of current passed, correspond-
ing to a reaction time of 60–120 min, depending on the conductivity of the
solution.
Product analyses
Qualitative detection was carried out with HPLC (Gynkotek, Germany), a
UV-(Gynkotek, UVD 170S) and a NaI(Tl) scintillation detector (Scionix,
Germany, 25 B 25/1.5 EP). A Partisil 10 ODS 3 column (250 ꢃ 8 mm) with a
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 259–273

272
G. J. KIENZLE ET AL.
pre-column (Partisil 10 ODS 3, 40 ꢃ 8 mm, both CS Chromatographie
Service GmbH, Germany) were employed. Eluents contained water and
acetonitrile with 0.1% KF added, to avoid adsorption of fluoride on the
column. TLC and electronic autoradiography were applied for quantitative
assay of all fluorinated phenylalanine derivatives. For TLC silica gel
(Machery-Nagel, Germany) and ethyl acetate/n-hexane mixtures were used.
The radioactive products were detected with autoradiography (InstantImager,
Canberra-Packard, Germany). Determination of relative isomeric distribution
of 7a–c was carried out using GC and MS. For GC a FS-Lipodex D
column (25 m ꢃ 0.25 mm, Machery-Nagel) with a pre-column (5 m ꢃ 0.25 mm,
Hewlett-Packard, Germany) was applied. Helium (0.7 ml/min), split
injection and a constant temperature of 1708C were used. The system
contained an automatic injecting system, an oven (HP 5890 II), a GC/MS
interface and the MS (HP 5989 A). For MS the electron ionisation method was
used.
Conclusions
As an example of an aromatic amino acid, protected phenylalanine was 18F-
fluorinated successfully. For the first time [18F]fluorophenylalanine has been
prepared by electrochemical 18F-fluorination through nucleophilic substitu-
tion using [18F]fluoride. The reaction showed a strong dependence on the
nature of the protective groups at the amino and carboxylic group. Best
labeling results (i.e. 10.5 ꢀ 2.5%) were obtained with the methylester and the
TFA group for protection of the amino function, at temperatures between
ꢂ108C and 08C using Et3N ꢁ 3HF as supporting electrolyte at a potential of
2.0 V. These conditions resulted in a carrier-added product. Based on the
addition of 0.33 mmol fluoride for an activity of 400 MBq, the specific activity
is calculated to be 1.2 GBq/mmol.
Further improvements are in progress for increasing regioselectivity by use
of different activating groups in the aromatic ring, for minimizing the addition
of carrier and optimizing radiochemical yields by variation of supporting
electrolyte and solvent. Nevertheless, the presented results demonstrate the
value of the electrochemical 18F-fluorination for labeling of aromatic amino
acids to be used as metabolic tracers in PET.
Acknowledgements
This project was supported by Deutsche Forschungsgemeinschaft (MA 1096/
3-1). The platinum electrodes were a gift of Degussa AG, Germany. We are
.
grateful to Prof. Dr. Speiser, University of Tubingen, Institute of Organic
Chemistry for the use of his cyclovoltammetry equipment, his assistance in
performing the experiments and stimulating discussions.
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 259–273

ELECTROCHEMICAL RADIOFLUORINATION. 3
273
References
.
1. Bodsch W, Coenen HH, Stocklin G, Takahashi K, Hossmann KA. J Neurochem
1988; 50: 979–983.
.
2. Coenen HH, Kling P, Stocklin G. J Nucl Med 1989; 30: 1367–1372.
3. Vaalburg W, Coenen HH, Crouzel C, Elsinga PH, Langstrom B, Lemaire C,
Meyer GJ. Int J Rad Appl Instrum B 1992; 19: 227–237.
4. Ogawa T, Miura S, Murakami M, Iida H, Hatazawa J, Inugami A, Kanno I,
Yasui N, Sasajima T, Uemura K. Eur J Nucl Med 1996; 23: 889–895.
5. DeJesus OT, Endres CJ, Shelton SE, Nickles RJ, Holden JE. J Nucl Med 1997; 38:
630–636.
6. Firnau G, Garnett ES, Chirakal R, Sood S, Nahmias C, Schrobilgen G. Appl
Radiat Isot 1986; 37: 669–675.
7. Lemaire C, Damhaut P, Plenevaux A, Comar D. J Nucl Med 1994; 35: 1996–2002.
8. Kaneko S, Ishiwata K, Hatano K, Omura H, Ito K, Senda M. Appl Rad Isot 1999;
50: 1025–1032.
9. Krasikova RN, Zaitsev VV, Ametamey SM, Kuznetsova OF, Fedorova OS,
Mosevich IK, Belokon YN, Vyskocil S, Shatik SV, Nader M, Schubiger PA. Nucl
Med Biol 2004; 31: 597–603.
10. Shono T. Tetrahedron 1984; 40: 811–850.
11. Lund H, Baizer MM. Organic Electrochemistry (3rd edn). Marcel Dekker: New
York, 1991.
12. Reischl G, Kienzle GJ, Machulla HJ. J Radioanal Nucl Chem 2002; 254: 409–411.
13. Reischl G, Kienzle GJ, Machulla HJ. Appl Radiat Isot 2003; 58: 679–683.
14. Fleischmann M, Pletcher D. Tetrahedron Lett 1968; 60: 6255–6258.
15. Rozhkov IN. Russ Chem Rev (Engl Transl) 1976; 45: 615–629.
16. Adams DJ, Clark JH. Chem Soc Rev 1999; 28: 225–231.
17. Salbeck J. Dissertation, 1988, University of Regensburg.
18. Ukaji Y, Miyamoto M, Mikuni M, Tekeuchi S, Inomata K. Bull Chem Soc Jpn
1996; 69: 735–742.
19. Fletcher CM, Lowther AG, Reith WS. Biochem J 1954; 56: 106–111.
20. Gammon KL, Smallcombe SH, Richards JH. J Am Chem Soc 1972; 94:
4573–4580.
21. Teuber HJ, Krause H, Berarin V. Liebigs Ann Chem 1978; 757–770.
.
22. Hutinec A. Dissertation, 1996, University of Tubingen.
Copyright # 2005 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2005; 48: 259–273