A convenient synthesis of [11C]paraquat and other [N-methyl-11C]bisquaternary ammonium compounds

Article abstract of DOI:10.1002/jlcr.564

[¹¹C]Paraquat was synthesized by the reaction of [¹¹C]methyl triflate with the mono-triflate salt of 1-methyl-[4,4′]bipyridinyl. The product was selectively separated from the precursor by a microcolumn of Chelex 100 ion exchange resin. The method was applied to the synthesis of a variety of [N-methyl-¹¹C]bisquaternary ammonium compounds. This is the first reported use of a chelating cation exchange resin for the selective purification of organic dications. Copyright

Full text of DOI:10.1002/jlcr.564

JOURNAL OF LABELLED COMPOUNDS AND RADIOPHARMACEUTICALS
J Label Compd Radiopharm 2002; 45: 281–289.
DOI: 10.1002/jlcr.564
Research Article
1
A convenient synthesis of [ C]paraquat and
1
11
other [N-methyl- C]bisquaternary
ammonium compounds
Douglas M. Jewett* and Michael R. Kilbourn
Division of Nuclear Medicine, Department of Radiology, University of Michigan
Medical Center, Ann Arbor, MI 48109-0552, USA
Summary
1
1
11
C]Paraquat was synthesized by the reaction of [ C]methyl triflate with the
[
0
mono-triflate salt of 1-methyl-[4,4 ]bipyridinyl. The product was selectively
separated from the precursor by a microcolumn of Chelex 100 ion exchange
resin. The method was applied to the synthesis of a variety of [N-
1
1
methyl- C]bisquaternary ammonium compounds. This is the first reported
use of a chelating cation exchange resin for the selective purification of organic
dications. Copyright # 2002 John Wiley & Sons, Ltd.
1
1
11
Key Words: [ C]paraquat; Chelex-100; [ C]hexamethonium; [ C]methyl
11
triflate
Introduction
Bisquaternary ammonium compounds may be selectively concentrated
1
2
in tissues by cation transport and by affinity for anionic sites. The
availability of carbon-11 labelled bisquaternary ammonium compounds
as PET tracers might help to explain the toxicity of such compounds as
wel1 as to suggest possible diagnostic uses based on dication transport
*Correspondence to: D. M. Jewett, Division of Nuclear Medicine, Department of Radiology,
University of Michigan Medical Center, Ann Arbor MI 48109-0552, USA.
E-mail: doogie@umich.edu
Contract/grant sponsor: Department of Energy; contract/grant number: DE-FG02-87ER60561
Conract/grant sponsor: NIH; contract/grant number: MH47611
Copyright # 2002 John Wiley & Sons, Ltd.
Received 9 October 2001
Revised 4 December 2001
Accepted 6 December 2001

2
82
D. M. JEWETT AND M. R. KILBOURN
systems or specific tissue affinities. Paraquat 1c is a mitochondrial toxin
3
that is avidly concentrated in the lungs, heart and other organs. It has
been implicated in some studies as a possible environmental cause of
4,5
neurodegenerative disorders, though the mechanisms by which it may
,7
6
cross the blood–brain barrier are not well explained.
Carbon-11
labelled paraquat may help to identify conditions resulting in enhanced
CNS uptake in intact animals.
Hexamethonium 3c is a ganglionic blocker, which also binds avidly to
2
collagen. Once used clinically, it has been implicated in a recent
8
fatality apparently as the result of unexpected lung damage from an
aerosol preparation. Carbon-11 labelled hexamethonium may be useful
as a diagnostic agent as well as to shed light on possible pulmonary
toxicity. A synthesis of carbon-11 labelled hexamethonium has been
9
reported, however, the procedure required several steps, starting with
0
0
N,N,N ,N -tetramethyl-1,6-hexanediamine, and necessarily resulted in a
product with low specific activity.
We observed the following considerations in developing a new,
general synthesis of carbon-11 labelled bisquaternary ammonium com-
pounds: (1) In order to obtain the highest possible specific activity the
bisquaternary compounds should be obtained in one step by carbon-11
methylation of the monocationic precursor. (2) As hydrophilic dications
such compounds should be capable of isolation by trapping on
microcolumns of chelating ion exchange resin. (3) The higher reactivity
1
1
11
of [ C]methyl triflate compared to [ C]methyl iodide should allow
instantaneous methylation at room temperature. (4) Use of a triflate salt
of a monocationic precursor should offer two advantages over halide
salts. Triflate salts tend to be highly soluble in polar solvents such as
1
1
DMSO, and the triflate anion cannot attack [ C]methyl triflate as a
competitive nucleophile as halide ions may. We describe here a simple
11
synthesis of [ C]paraquat which incorporates the above considerations.
Other carbon-11 labelled bisquaternary ammonium compounds (see
Figure 1) were prepared in the same way in high radiochemical purity
11
and satisfactory yields. [ C]Hexamethonium presented a special
problem for purification in that it was not as strongly retained on the
chelating resin relative to the monocationic precursor.
Results and discussion
11
11
The formation of [ C]paraquat from [ C]methyl triflate occurred
within a few seconds in DMSO or acetonitrile at room temperature. The
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 281–289

1
1
SYNTHESIS OF [ C]PARAQUAT
283
Figure 1. Synthesis of carbon-11 bisquaternary ammonium compounds and their
precursors
subsequent steps – dilution with aqueous ammonia, trapping on the
chelating cation exchange resin, selective elution of the precursor with
aqueous nPrOH/NaCl/NH OH and recovery of the product in normal
4
saline required about 4 min. In a representative experiment approxi-
11
mately 300 mCi [ C]paraquat with a specific activity of 2200 Ci/mmole
11
were obtained from 1400 mCi [ C]CO . The radiochemical purity was
2
>
99%. The precursor was present at the level of 0.13 mg/ml, compared
to 2.6 mg/ml for paraquat. Table 1 summarizes the results obtained for
1
1
[
C]paraquat and other carbon-11 labelled bisquaternary compounds
obtained by the same method.
Figure. 2 shows elution profiles of paraquat 1c and its precursor 1b
from microcolumns of two different forms of chelating ion exchange
+
resin, Chelex-100 (Na form, 200–400 mesh, Bio-Rad) and a chelating
cation exchange fiber (TIN-600, Toray), simulating conditions for the
radiochemical synthesis. Since it was necessary to use widely different
UV detector settings to accommodate the large excess of precursor over
product encountered in the radiochemical synthesis, the base line was
adjusted after the elution of the precursor. Of the compounds labelled,
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 281–289

2
84
D. M. JEWETT AND M. R. KILBOURN
Table 1. Summary of results of radiosynthesis of carbon-11 bisquaternary
ammonium compounds
Compound
n) experiments
Counter
anion
Reaction
solvent
Radiochem.
yield
Radiochem.
purity (%)
Resin
a
(
1
1
1
1
1
2
2
3
4
5
c (4)
c (2)
c (1)
c (1)
c (1)
c (1)
c (1)
c (1)
c (1)
c (1)
OTfꢀ
OTf
MeCN
MeCN
MeCN
Water
69
63
74
58
15
63
26
74
52
1.6
>99
>99
>99
97
>99
>99
>98
>99
>99
88
TIN-600
Chelex-100
TIN-600
TIN-600
TIN-600
Chelex-100
TIN-600
TIN-600
Chelex-100
TIN-600
ꢀ
ꢀ
I
OTf
ꢀ
ꢀ
OTf
OTf
DMF
ꢀ
MeCN
MeCN
MeCN
MeCN
DMSO
ꢀ
I
OTf
ꢀ
ꢀ
OTf
OTf
ꢀ
a
11
Approximate uncorrected radiochemical yield based on the amount of [ C]methyl triflate
distilled.
only paraquat had a sufficiently strong chromophore to permit
obtaining an elution profile based on ultraviolet absorbance. However,
compounds 2c and 4c behaved qualitatively the same as paraquat in
their affinity for the chelating cation exchange resin, as judged by the
well-defined trapping of radioactivity on the microcolumn and
1
1
subsequent displacement by normal saline. In contrast [ C]hexametho-
nium 3c was not as strongly retained on the resin. Apparently, steric
constraints interfered with specific binding to the dianionic site of the
1
1
resin. For [ C]hexamethonium, conditions have not been established
for the selective removal of the precursor from the resin. While the
11
results show that [ C]hexamethonium can be obtained in high
radiochemical purity, additional work must be done to find a weak
ionic eluant to selectively remove the large excess of precursor for
applications requiring a high specific activity.
The elution profiles indicate that both the fibrous and spherical forms
of chelating cation exchanger are satisfactory for the separation of
1
1
[
C]paraquat from its monocationic precursor. The peak for paraquat
is sharper for the Chelex-100 resin at the low flow rate used to obtain the
1
0
UV elution profiles. In practice, the fibrous resin was preferred
because of the ease of packing and compatibility with high flow rates.
This resin, however, is not commercially available at present.
Apparently, this is the first reported use of a chelating resin to isolate
organic dications such as paraquat.
The use of an ion-pairing HPLC method for monitoring the synthesis
11
of hydrophilic organic dications simplifies the interpretation of
experiments, in that the retention times are substantially determined
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 281–289

1
1
SYNTHESIS OF [ C]PARAQUAT
285
Figure 2. Elution profiles of paraquat (d) and its monocationic precursor (b) from
chelating ion exchange resins. Upper trace: Chelex-100 resin. Lower trace: TIN-
6
00 fibrous resin. (a) Inject simulated reaction mixture while eluting with a
solution of 15% n-propanol, 0.5% NaCl and 0.3% ammonium hydroxide in
water. (c) Begin elution with normal saline
by the charge alone. All of the carbon-11 labelled bisquaternary
ammonium compounds had essentially the same retention times. Thus,
we were able to obtain evidence for the stability of the previously
1
2
unreported hexamethyl hydrazinium dication 5c by reacting com-
1
pound 5b with [ C]methyl triflate.
3
11
With the general method described above, high specific activities are
possible provided that the monocationic precursors are not contami-
nated by small amounts of the unlabelled dication. In the case of
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 281–289

2
86
D. M. JEWETT AND M. R. KILBOURN
paraquat the presence of a strong UV chromophore allowed the
1
precursor to be highly purified by recrystallization. While H NMR
analysis of the other precursors indicated them to be substantially pure,
contamination by small but significant amounts of the dications could
not be ruled out. For some applications of these compounds, as tracers,
it would be necessary to develop HPLC methods able to detect them at
trace levels, so that the precursors could be rigorously purified by
recrystallization.
Experimental
General
Starting bisdiamines 1a, 2a, 3a and 4a were purchased from Aldrich
0
0
Chemical Co. N,N,N N -tetramethylhydrazine 5a was purchased from
Fluka Chemical Co. Proton NMR spectra were obtained on a Bruker
DRX500 instrument using a TMS internal standard. Melting points
were obtained with a Mel-Temp apparatus (Laboratory Devices,
Cambridge, MA). The HPLC system consisted of a Spectra Physics
SP8800 ternary pump, Hitachi L4000 detector and HP3395 integrator.
The radioactivity of the HPLC effluent was measured by a sodium
iodide crystal coupled to a photomultiplier module and an Ortec 449
ratemeter.
General procedure for monocationic triflate precursors
(Compounds 1b–4b). Methyl iodide was reacted at room temperature in
acetonitrile with a ten-fold molar excess of the corresponding
bisdiamine. In this way, the monocationic product was favored over
the dication, albeit at the expense of a low chemical yield. The resulting
monocationic iodide salt was washed with ether and recrystallized
repeatedly from isopropanol, ethyl acetate, diisopropyl ether or
mixtures thereof. This salt was then dissolved in water and reacted
with a slight molar excess of aqueous silver triflate. Silver iodide was
removed by filtration through diatomaceous earth. The aqueous filtrate
was concentrated to dryness and the resulting solid was washed with
ether. The precuresor for paraquat 1b was recrystallized as the triflate
salt from isopropanol. Due to the high solubility of the other triflate
salts 2b–5b in organic solvents, these were not recrystallized at the
triflate salt stage.
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 281–289

1
1
SYNTHESIS OF [ C]PARAQUAT
287
0
1
Compound 1b, triflate salt (from 4,4 -dipyridyl 1a). H-NMR
(500MHz, DMSO) d 9.13 (2H, d, J=6.75Hz), 8.88 (2H, d,
J=6.14Hz), 8.62 (2H, d, J=6.9 Hz), 8.04 (2H, d, J=6.2Hz), 4.39 (3H,
14
s, N-Me). HPLC UV 300 nm, >99%; m.p. 130–1328C (lit. 1358C).
Compound 2b, triflate salt (from diazabicyclooctane 2a). H-NMR
15
1
(
500 MHz, DMSO) d 3.25 (6H, t), 3.02 (6H, t), 2.96 (3H, s, N-Me).
Compound 3b, triflate salt (from N,N,N ,N -tetramethylhexanedia-
1
0
0
mine 3a). H-NMR (500 MHz, DMSO) d 3.24 (2H, m), 3.05 (9H, N-Me,
s), 2.24 (2H, m), 2.15 (6H, N-Me, s), 1.68 (2H, m), 1.43 (2H, m), 1.33
(
4H, CH , m).
2
0
0
Compound 4b, triflate salt (from N,N,N ,N -tetramethylethanedia-
1
mine 4a). H-NMR (500 MHz, DMSO) d 3.55 (2H, t), 3.31 (9H,
N-Me, s), 2.76 (2H, m), 2.29 (6H, N-Me, s).
Compound 5b. (from tetramethylhydrazine 5a). Iodide salt m.p.
3
1
1
2
238C, lit. 223–2248C. H-NMR (500 MHz, DMSO) d 3.21 (9H,
N-Me, s), 2.66 (6H, N-Me, s).
Elution profile for paraquat and monocationic precursor
A microcolumn containing 6 mg of Chelex-100 resin, or alternatively,
Toray TIN-600 fibrous resin (sodium form, see below) was connected to
an HPLC pump and injector via 30 m ꢁ 0.1 mm i.d. tubing. The latter
provided a pressure drop necessary for proper operation of the pump
and also provided a convenient delay between the time of injection and
the beginning of the elution of the monocation. The outlet of the
microcolumn was connected to a UV absorbance detector. Elution was
begun with solution of 15% n-propanol, 0.05% NaCl and 0.3%
ammonium hydroxide in water. A solution containing 1.0 mg precursor
1
b and 50 mg paraquat hydrochloride in 50 mL water was injected to
simulate the amounts of product and precursor in a carbon-11 labelling
reaction. Due to the large amount of monocation, the detector was
initially set at 326 nm on the shoulder of the UV absorbance. After
elution of the monocation was complete, the elution solvent was
changed to 0.9% NaCl and the detector was set at 310 nm to monitor
the elution of paraquat from the column.
Ion exchange column
The chelating ion-exchange column consisted of 6 mg iminodiacetate
functionalized fibrous resin (Toray TIN-600, sodium form) slurry
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 281–289

2
88
D. M. JEWETT AND M. R. KILBOURN
packed in a 3 mm i.d. polyethylene column (bed length 5 mm). The resin
was held in place by small plugs of polypropylene wool (Aldrich). The
end fittings were miniature barbed polypropylene male luer adapters
(Value Plastics, Fort Collins, CO 80525, USA).
Radiosynthesis
The precursor (triflate salt, 1 mg) was dissolved in 0.1 ml anhydrous
11
acetonitrile. [ C]Methyl triflate, prepared by the gas-phase reaction of
1
1
16
C]methyl iodide with silver triflate at 1808C, was distilled into this
[
solution in a stream of nitrogen at room temperature. When maximum
radioactivity had accumulated the reaction mixture was diluted with
4
ml water containing 0.3% NH OH (pH 11) and passed through the
4
ion exchange column. The column was washed with 10 ml of a solution
of 15% n-propanol, 0.05% NaCl and 0.3% ammonium hydroxide in
water, followed by 2 ml of water to remove the precursor. The
radioactive product was eluted from the column with 4–10 ml 0.9%
11
NaCl. In the case of [ C]hexamethonium the column was washed with
water but not with nPrOH/NH OH/NaCl, since this solution was
4
sufficiently strong to elute the product. The radioactivity of the product
and waste fractions were measured in a Capintec ionization counter.
Approximate uncorrected radiochemical yields were obtained by
dividing the radioactivity in the pure product by that of the total
11
amount of radioactivity trapped as [ C]methyl triflate.
Analysis by HPLC
The radioactive product was analyzed by reversed phase, ion-pair
HPLC on a Keystone BetaBasic 8 column (4.6 ꢁ 150 mm). The solvent
consisted of 7.5 mM sodium octanesulfonate and 20 mM NH H PO
4
2
4
in 20% aqueous acetonitrile, pH 4.0, 1.5 ml/min. Radioactivity was
1
1
detected by a sodium iodide crystal flow detector. For [ C]paraquat
specific activity was determined by UV absorbance at 254 nm. Retention
times were 6.1 min for paraquat and 2.7 min for the precursor.
Acknowledgements
This work was supported Department of Energy Grant no. DE-FG02-
8
7ER60561 and by NIH Grant no. MH47611.
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 281–289

1
1
SYNTHESIS OF [ C]PARAQUAT
289
References
1
. Rose MS, Smith LL, Wyatt I. Nature 1974; 252: 314–315.
2
. Korn N, Huang CC, Seevers RH, Rothwell C, Counsell RE. Int J Nucl
Med Biol 1979; 6: 153–161.
3
. Houze P, Baud FJ, Scherrman JM. Toxicokinetics of paraquat. In
Paraquat Poisoning, Bismuth C, Hall AH (eds). Marcel Decker, Inc: New
York, 1995; 161–193.
4
5
6
7
. Brooks AI, Chadwick CA, Gelbard HA, Cory-Slecchta DA, Federoff HJ.
Brain Res 1999; 823: 1–10.
. Thiruchelvam M, Richfield EK, Baggs RB, Tank AW, Cory-Schlechta
DA. J Neurosci 2000; 20: 9207–9214.
. Widdowson PS, Farnworth MJ, Simpson MG, Lock EA. Hum Exp
Toxicol 1996; 15: 231–236.
. Corasaniti MT, Defilippo R, Rodino P, Nappi G, Nistico G. Funct Neurol
1
991; 6: 385–391.
8
. Anon. Science 2001; 292: 2226–2227.
9
. Qureshi MA. Proc Int Symp Appl Ion Radiat 1982; 2: 1035–1041.
1
0. Yoshioka T, Shimamura M. Bull Chem Soc Jpn 1983; 56: 3726–3729.
1. Carneiro MC, Puignou L, Galceran MT. J Chromatogr A 1994; 669:
1
2
17–224.
12. Klages F, Wolf H. Chem Ber 1959; 92: 1842–1849.
13. Beltrami RT, Bissell ER. J Am Chem Soc 1956; 78: 2467–2468.
14. Banks RE, Besheesh MK, Mohialdin-Khaffaf SN, Sharif I. J Fluorine
Chem 1997; 81: 157–161.
1
1
5. Banks RE, Besheesh MK, Mohialdin-Khaffaf SN, Sharif I. J Fluorine
Chem 1996; 78: 43–50.
6. Jewett DM. Appl Radiat Isot 1992; 43: 1383–1385.
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2002; 45: 281–289