18F-glycosylation using Koenigs-Knorr conditions: A comparative study

Article abstract of DOI:10.1002/jlcr.1054

We compared two ¹⁸F-glycosylation methods, the BF ₃-mediated ¹⁸F-glycosylation versus the newly developed AgOTf-activated ¹⁸F-glycosylation procedure. The AgOTf-activated ¹⁸F-glycosylation makes use of 3,4,6-tri-O-acetyl-2-deoxy-2-[ ¹⁸F]fluoro-glucopyranosyl bromide and revealed an improved radiochemical yield of 67 ± 6% in the case of a protected serinyl precursor as compared to 27 ± 4% obtained by the BF₃-method. This suggests the suitability of Koenigs-Knorr conditions for ¹⁸F-glycosylation of protected bioactive peptides. Copyright

Full text of DOI:10.1002/jlcr.1054

JOURNAL OF LABELLED COMPOUNDS AND RADIOPHARMACEUTICALS
J Label Compd Radiopharm 2006; 49: 101–108.
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jlcr.1054
Research Article
18F-glycosylation using Koenigs–Knorr conditions:
a comparative study
Simone Maschauer, Torsten Kuwert and Olaf Prante*
Laboratory of Molecular Imaging, Clinic of Nuclear Medicine, Friedrich-Alexander
University, Schwabachanlage 6, D-91054 Erlangen, Germany
Summary
We compared two 18F-glycosylation methods, the BF3-mediated 18F-glycosylation
versus the newly developed AgOTf-activated 18F-glycosylation procedure. The
AgOTf-activated 18F-glycosylation makes use of 3,4,6-tri-O-acetyl-2-deoxy-2-
[18F]fluoro-glucopyranosyl bromide and revealed an improved radiochemical yield
of 67 ꢀ 6% in the case of a protected serinyl precursor as compared to 27 ꢀ 4%
obtained by the BF3-method. This suggests the suitability of Koenigs–Knorr
conditions for 18F-glycosylation of protected bioactive peptides. Copyright # 2006
John Wiley & Sons, Ltd.
Key Words: 18F-glycosylation; F-18; Koenigs–Knorr; glycosyl donor; PET
Introduction
The use of 18F-labelled prosthetic groups for the radiofluorination of
biomolecules has gained considerable importance for the synthesis of
radiopharmaceuticals suitable for positron emission tomography (PET), since
the presence of sensitive functional groups excludes a direct nucleophilic
radiofluorination reaction under basic conditions. In particular, N-succinimi-
dyl-4-[18F]fluorobenzoate ([18F]SFB),1–3 4-nitrophenyl-2-[18F]fluoropropio-
nate,4 4-[18F]fluorobenzaldehyde5,6 or 18F-labelled maleimide derivatives6,7
have found widespread use for 18F-labelling of peptides and proteins. As an
extension of the available 18F-labelling agents, we recently examined the
applicability of tetra-O-acetylated 2-[18F]fluoro-2-deoxy-glucopyranose
([18F]1) as 18F-glycosylation agent.8 However, this approach implies problems
*Correspondence to: Olaf Prante, Laboratory of Molecular Imaging, Clinic of Nuclear Medicine, Friedrich-
Alexander University Erlangen-Nurnberg, Schwabachanlage 6, D-91054 Erlangen, Germany.
¨
E-mail: olaf.prante@nuklear.imed.uni-erlangen.de
Contract/grant sponsor: Friedrich-Alexander University, Erlangen
Copyright # 2006 John Wiley & Sons, Ltd.
Received 25 November 2005
Accepted 7 January 2006

102
S. MASCHAUER ET AL.
when applied to the labelling of peptides: the use of BF3 as a Lewis acid
promoter could cause side-reactions, such as cleavage of protecting groups,
and limits the choice of the reaction solvent. In order to circumvent these
drawbacks, we aimed at the development of an alternative 18F-glycosylation
procedure.
We herein report the radiosynthesis of 3,4,6-tri-O-acetyl-2-deoxy-2-
[18F]fluoroglucopyranosyl bromide ([18F]2) and its use as 18F-labelled glycosyl
donor under Koenigs–Knorr9 conditions in silver trifluoromethanesulphonate
(AgOTf) promoted 18F-glycosylation reactions on protected serinyl derivatives
as model compounds. In addition, we compared this approach to the BF3-
promoted reaction sequence with regard to radiochemical yield, reaction time,
mechanistic implications and dependence on the labelling precursor.
Results and discussion
Starting from the tetra-O-acetylated 18F-labelled glycosyl donor [18F]1, which
was synthesized by nucleophilic 18F-for-OTf substitution on the corresponding
tetra-O-acetylated mannopyranoside following the method of Hamacher and
others10 and isolated by HPLC in a radiochemical yield (RCY) of 68% (decay-
uncorrected) in 30 min as reported previously,8 we developed two different
routes for 18F-glycosylation of amino acids: the BF3-method and the AgOTf-
method (Scheme 1). First, we reinvestigated the BF3-method8 using 200 mM
BF3 ꢁ Et2O and 10 mM Fmoc-Ser-OH or Z-Ser-OBn as glycosyl acceptors in
acetonitrile as solvent. This procedure gave a decay-uncorrected radiochemical
yield (RCY) of 23% for [18F]3 after HPLC-separation in a total synthesis time
of about 60 min (Scheme 1). Using Z-Ser-OBn as a more sterical demanding
glycosyl acceptor, a decreased RCY of about 11% for [18F]4 was obtained.
Studying the reaction of [18F]1 with Fmoc-Ser-OH at early time points
(1–6 min) by analytical radio-HPLC, we observed the formation of glycoester
[18F]5, which was almost completely converted into the desired O-glycosylated
product [18F]3 during the progress of the reaction (Figure 1). Thus, the use of
BF3 in high concentration (200 mM) ensured completed formation of [18F]3,
nicely confirming the observation of Salvador et al. for the synthesis of
glycopeptide building blocks.11 Therefore, we conclude, that glycoester [18F]5
was cleaved in the presence of BF3 leading to an oxocarbenium intermediate as
depicted in Scheme 2, which could subsequently be attacked at the
electrophilic C-1 by the hydroxyl side chain of Fmoc-Ser-OH.
In order to provide an alternative 18F-glycosylation method, we developed
and optimized a procedure based on the classical Koenigs–Knorr conditions
using [18F]2 as glycosyl donor, AgOTf as activator and Fmoc-Ser-OH or
Z-Ser-OBn as glycosyl acceptor. Peracetylated 2-deoxy-2-[18F]fluoroglucopyr-
anose ([18F]1) was completely converted into its anomeric bromide [18F]2 with
HBr/CH3COOH in a reaction time of 5 min followed by solid phase extraction
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 101–108

KOENIGS–KNORR 18F-GLYCOSYLATION
103
Scheme 1. Time course of the BF3-mediated 18F-glycosylation method starting
from [18F]1 compared to the AgOTf-activated 18F-glycosylation (Koenigs–Knorr
method) using glycosyl donor [18F]2. RCY are expressed as decay-uncorrected
yields of the HPLC-isolated 18F-labelled products and referred to [18F]fluoride
on a RP-18 cartridge (Scheme 1). The AgOTf-method was optimized in regard
to solvent, temperature, reaction time, activator and the ratio of glycosyl
acceptor (Z-Ser-OBn) to activator (AgOTf) (Table 1). The most favourable
reaction parameters turned out to be the ratio of labelling precursor Z-
Ser-OBn to AgOTf of 0.5–6 and dichloromethane as solvent affording [18F]4 in
a RCY of 67% after a reaction time of 1–2 min at room temperature (Table 1).
Applying these Koenigs–Knorr reaction conditions to the 18F-glycosylation of
Fmoc-Ser-OH to give [18F]3, the RCY decreased to about 21% (Table 2), due
to the formation of glycoester [18F]5 as major by-product in equivalent RCY
compared to the desired product [18F]3. HPLC analysis showed the formation
of a radiolabeled hydrophilic compound, possibly indicating the presence of
an oxocarbenium trifluoromethanesulphonate intermediate, which equably
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 101–108

104
S. MASCHAUER ET AL.
Figure 1. Dependence of the radiochemical yield of the 18F-labelled products on
reaction time for the BF3-mediated 18F-glycosylation of Fmoc-Ser-OH using
[18F]1 as glycosyl donor (200 mM BF3 ꢁ Et2O in CH3CN, V=500 ml, 8 mM
Fmoc-Ser-OH, T=808C)
.
Scheme 2. (i) HBr/CH3COOH, CH2Cl2; (ii) BF3 Et2O, CH3CN, Fmoc-
Ser-OH; (iii) AgOTf, CH2Cl2, Fmoc-Ser-OH
decreased during the progress of the reaction in support of both products, the
18F-glycosylated product ([18F]3 or [18F]4) and the by-product [18F]5. In
contrast to the BF3-mediated glycosylation method, no consequential
conversion of [18F]5 into product [18F]3 occurred under Koenigs–Knorr
conditions (Scheme 2).
In summary, the two investigated 18F-glycosylation methods differed in
maximum RCY for the 18F-glycosylation key step depending on the labelling
precursor (Table 2). The Koenigs–Knorr 18F-glycosylation method revealed
an improved radiochemical yield of 67 ꢀ 6% in the case of the Na- and
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 101–108

KOENIGS–KNORR 18F-GLYCOSYLATION
105
Table 1. Optimization of the Koenigs–Knorr 18F-glycosylation of Z-Ser-OBn using
[18F]2
Activator Ratio/(Z-Ser-OBn Solvent
activator)a
Temperature Reaction time RCYb
(%)
(min)
Hg(CN)2
HgBr2
5
1
CH2Cl2
CH2Cl2
rt
rt
10
2
7
0
AgOTf
AgOTf
AgOTf
AgOTf
0.1–0.25
0.5–6
10
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
rt
rt
rt
rt
1–2
1–2
1–2
1–2
22 ꢀ 10c
67 ꢀ 6c
55 ꢀ 13c
0
24–120
AgOTf
AgOTf
AgOTf
AgOTf
AgOTf
3
3
3
1
1
Et2O
rt
2
2
2
5
5
30
38
7
7
0
dichloroethane 808C
CH3NO2
CH3CN
DMF
908C
rt
1308C
AgOTf
AgOTf
AgOTf
3
1
1
CH2Cl2
CH2Cl2
CH2Cl2
08C
ꢂ208C
ꢂ208C
5
1
10
46
22
53
a n (Z-Ser-OBn) = 1.5–50 mmol, V ¼ 500 ml.
b RCY are expressed as mean of two independent experiments ðn ¼ 2Þ determined by analytical radio-HPLC
from a sample withdrawn from the reaction mixture and referred to the 18F-labelled glycosyl donor [18F]2.
c Mean ꢀ standard error of the mean (SEM, n ¼ 4).
Table 2. BF3-method vs AgOTf-method: comparison of maximum radiochemical yields
(RCY) of the 18F-glycosylation key step
Precursor
Product
AgOTf-method using
BF3-method using
glycosyl donor [18F]2
glycosyl donor [18F]1
RCYa (%)
t (min)
RCYa (%)
t (min)
Fmoc-Ser-OH
Z-Ser-OBn
[18F]3
[18F]4
21 ꢀ 3
67 ꢀ 6
2
1–2
47 ꢀ 8
27 ꢀ 4
5
10
a RCY are expressed as mean ꢀ standard error (SEM, n ¼ 4) determined by analytical radio-HPLC from a
sample withdrawn from the reaction mixture and referred to the 18F-labelled glycosyl donor.
C-protected serinyl precursor (Z-Ser-OBn) as compared to 27 ꢀ 4% obtained
by the BF3-method (Table 2). In contrast to this result, a serinyl labelling
precursor containing a free carboxyl group (Fmoc-Ser) was advantageously
labelled by 18F-glycosylation applying the reaction conditions of the BF3-
method. Thus, the use of [18F]2 for 18F-glycosylation under Koenigs–Knorr
conditions should particularly be of interest for labelling precursors without
interfering free C-terminus, such as bioactive cyclic peptides.
As shown in Scheme 1, we also performed a comparison of both 18F-
glycosylation methods with regard to the reaction time and overall decay-
uncorrected radiochemical yields. Starting from [18F]fluoride, the BF3-method
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 101–108

106
S. MASCHAUER ET AL.
gave [18F]3 or [18F]4 in an overall decay-uncorrected RCY of 23 and 11%,
respectively, within a total reaction time of 60 min including 2 HPLC
purification steps (Scheme 1). The AgOTf-method permitted the use of
dichloromethane as solvent, lasted 15 min longer affording [18F]3 in an overall
RCY of 8%, but was more advantageous for the radiosynthesis of [18F]4
revealing an overall decay-uncorrected RCY of 27% (Scheme 1). Experiments
concerning the deacetylation of the 18F-glycosylated compounds ([18F]3,
[18F]4) were beyond the scope of the present study, since this issue was
successfully addressed elsewhere.8
Experimental
General
Na-(9-Fluorenylmethoxycarbonyl)-l-serine (Fmoc-Ser-OH) and Na-(phenyl-
methoxycarbonyl)-l-serine phenylmethyl ester (Z-Ser-OBn) were purchased
from Bachem (Germany). [18F]Fluoride was obtained from PET Net GmbH
(Erlangen, Germany). Radio-thin layer chromatography (radio-TLC) was
carried out on plastic sheets (Polygram1, Sil G/UV254, Macherey Nagel) using
ethyl actetate/n-hexane (1:1, v/v) as eluent. Analytical HPLC was performed
on the following system: HPLC Agilent 1100 with a quarternary pump and
variable wavelength detector and radio-HPLC-detector D505TR (Canberra
Packard). Computer analysis of the HPLC data was performed using FLO-
One software (Canberra Packard). Electronic autoradiography was used to
analyse radio-TLC data (Instant ImagerTM, Canberra Packard). The syntheses
of 1,3,4,6-tetra-O-acetyl-2-deoxy-2-fluoro-d-glucopyranose ([18F]1, [19F]1),
1,3,4,6-tetra-O-acetyl-2-O-trifluoromethanesulfonyl-b-d-mannopyranose and
all other reference compounds ([19F]2, [19F]3, [19F]4, [19F]5) were described
18
elsewhere.8 Each F-labelled compound was identified by k0 values on
the radio-HPLC system and co-injection of the corresponding reference
compound.
3,4,6-Tri-O-acetyl-2-deoxy-2-[18F] fluoroglucopyranosyl bromide ([18F]2)
Three hundred microlitre 33% HBr/CH3COOH were added to a reaction
vessel containing [18F]1 in 50 ml dichloromethane at room temperature. The
reaction was monitored by radio-TLC. After 5 min conversion was complete
and the mixture was diluted with 10 ml H2O and passed through a C18-
cartridge (Merck, 100 mg). The cartridge was washed with H2O, dried and
eluted with 1 ml CH3CN. Starting from 250 MBq [18F]1 this procedure yielded
205 MBq [18F]2 within 15 min. Radio-TLC: Rf ¼ 0:77. HPLC (Kromasil C8,
250 ꢃ 4.6, 1.5 ml/min, 40–100% CH3CN in water (0.1% TFA) in 50 min):
k0 ¼ 6:2.
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 101–108

KOENIGS–KNORR 18F-GLYCOSYLATION
107
Na-(9-Fluorenylmethoxycarbonyl)-O-(3,4,6-tri-O-acetyl-2-deoxy-2-
[18F]fluoro-d-glucopyranosyl)-l-serine ([18F]3) and Na-(phenylmethoxycar-
bonyl)-O-(3,4,6-tri-O-acetyl-2-deoxy-2-[18F]fluoro-d-glucopyranosyl)-l-serine
phenylmethyl ester ([18F]4)
18F-glycosylation using BF3. A solution of the labelling precursor (10 mM
Fmoc-Ser-OH or Z-Ser-OBn) in 200 ml anhydrous CH3CN and 5 ml BF3
etherate were added to a reaction vessel containing dry [18F]1 at 808C. The
radioactive products ([18F]3 or [18F]4) were obtained in a radiochemical yield
of 47% ([18F]3) or 27% ([18F]4) after
a
reaction time of
5–10 min. The reaction mixture was diluted with H2O (1:10) and passed
through a C18-cartridge (Merck, 100 mg). The cartridge was washed with
CH3CN/H2O (30:70) to remove hydrolytic by-products, dried and eluted with
1.0 ml CH3CN. The solvent was evaporated and the residue (diluted in 500 ml
acetonitrile/water (40:60)) was separated by gradient reversed-phase HPLC
(Kromasil C8, 125 ꢃ 8 mm, 4 ml/min, 40–100% CH3CN in H2O (0.1% TFA)
within 50 min). The product fraction was diluted with water (1:10), fixed on a
C18-cartridge (Merck, 100 mg) and eluted with 1 ml acetonitrile in a reaction
vessel. Starting from 250 MBq [18F]1 this procedure yielded 85 MBq of [18F]3
or 40 MBq of [18F]4 within 30 min.
18F-glycosylation using AgOTf. A solution of the labelling precursor (10 mM
Fmoc-Ser-OH or Z-Ser-OBn) and 10 mM AgOTf in 500 ml anhydrous CH3CN
were added to a reaction vessel containing dry [18F]2 at room temperature. The
radioactive products ([18F]3 or [18F]4) were obtained in a radiochemical yield
of 21% ([18F]3) or 67% ([18F]4) after a reaction time of 1–2 min. The reaction
mixture was passed through a Si-cartridge (Merck, 200 mg) and seperated by
gradient reversed-phase HPLC (Kromasil C8, 125 ꢃ 8 mm, 4 ml/min, 40–100%
CH3CN in H2O (0.1% TFA) within 50 min). The product fraction was diluted
with water (1:10), fixed on a C18-cartridge (Merck, 100 mg) and eluted with
1 ml acetonitrile in a reaction vessel. Starting from 205 MBq [18F]2 this
procedure yielded 30 MBq [18F]3 or 100 MBq [18F]4 within 30 min.
[18F]3 and [18F]4 were characterized by radio-HPLC and radio-TLC for
both methods ([18F]3: Radio-TLC: Rf ¼ 0:10. HPLC (Kromasil C8, 250 ꢃ 4.6,
1.5 ml/min, 40–100% CH3CN in water (0.1% TFA) in 50 min): k0 ¼ 10:3.
[18F]4: Radio-TLC: Rf ¼ 0:71. HPLC (Kromasil C8, 250 ꢃ 4.6, 1.5 ml/min,
40–100% CH3CN in water (0.1% TFA) in 50 min): k0 ¼ 13:8).
Optimization of the 18F-glycosylation procedure using the AgOTf-method
The 18F-glycosylation procedure was optimized by repeating the reaction with
varying parameters as indicated in Table 1.
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 101–108

108
S. MASCHAUER ET AL.
Conclusion
The 18F-labelled glycosyl bromide [18F]2 was easily available without the
necessity of an additional HPLC purification step as compared to the
radiosynthesis of the corresponding tetra-O-acetylated glycosyl donor [18F]1.
Applying Koenigs–Knorr reaction conditions, [18F]2 proved its suitability for
the AgOTf-activated 18F-glycosylation of Z-Ser-OBn with improved radio-
chemical yield. Therefore, the 18F-glycosylation agent [18F]2 should be
preferentially applied to O-glycosylation reactions on suitably protected
bioactive peptides in further studies, in order to develop 18F-glycopeptides as
potential PET radiopharmaceuticals.
Acknowledgements
The authors gratefully acknowledge Wilhelm Hamkens and the cyclotron staff
for producing and delivering [18F] fluoride (PET Net GmbH, Erlangen). This
work was supported by a grant of the Hochschul- und Wissenschaftspro-
gramm (HWP, Chancengleichheit fur Frauen in Forschung und Lehre) at the
¨
Friedrich-Alexander University.
References
1. Wester HJ, Hamacher K, Stocklin G. Nucl Med Biol 1996; 23: 365–372.
¨
2. Bergmann R, Scheunemann M, Heichert C, Mading P, Wittrisch H,
Kretzschmar M, Rodig H, Tourwe D, Iterbeke K, Chavatte K, Zips D,
´
Reubi JC, Johannsen B. Nucl Med Biol 2002; 29: 61–72.
3. Vaidyanathan G, Zalutsky MR. Bioconjug Chem 1994; 5: 352–356.
4. Haubner R, Kuhnast B, Mang C, Weber WA, Kessler H, Wester HJ,
Schwaiger M. Bioconjug Chem 2004; 15: 61–69.
5. Poethko T, Schottelius M, Thumshirn G, Hersel U, Herz M, Henriksen G,
Kessler H, Schwaiger M, Wester HJ. J Nucl Med 2004; 45: 892–902.
6. Toyokuni T, Walsh JC, Dominguez A, Phelps ME, Barrio JR, Gambhir SS,
Satyamurthy N. Bioconjug Chem 2003; 14: 1253–1259.
7. de Bruin B, Kuhnast B, Hinnen F, Yaouancq L, Amessou M, Johannes L,
Samson A, Boisgard R, Tavitian B, Dolle F. Bioconjug Chem 2005; 16: 406–420.
´
8. Maschauer S, Pischetsrieder M, Kuwert T, Prante O. J Label Compd Radiopharm
2005; 48: 701–719.
9. Koenigs W, Knorr E. Chem Ber 1901; 34: 957–981.
10. Hamacher K, Coenen HH, Stocklin G. J Nucl Med 1986; 27: 235–238.
¨
11. Salvador LA, Elofsson M, Kihlberg J. Tetrahedron 1995; 51: 5643–5656.
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 101–108