Simultaneous detection of different glycosidase activities by 19F NMR spectroscopy

Article abstract of DOI:10.1016/S0008-6215(00)00060-4

A fast method for the simultaneous detection of different glycosidolytic activities in commercially available enzyme preparations and crude culture filtrates was found in using, as substrate, a mixture of different glycosyl fluorides and ¹⁹F NMR spectroscopy as a screening technique. Accompanying studies regarding the hydrolytic stability of these fluorides in various buffer systems, as well as conditions of their long-term storage, were carried out. A simple procedure for the preparation of β-D-mannopyranosyl fluoride in gram quantities is given. Copyright (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0008-6215(00)00060-4

www.elsevier.nl/locate/carres
Carbohydrate Research 326 (2000) 395–400
Simultaneous detection of different glycosidase activities
1
9
ꢀ
by F NMR spectroscopy
a
a
a
b
Martin Albert , Werner Repetschnigg , J o¨ rg Ortner , Joseph Gomes ,
Bernhard J. Paul , Carina Illaszewicz , Hansj o¨ rg Weber , Walter Steiner ,
a
a
a
b
a,
Karl Dax *
a
Institute of Organic Chemistry, Technical Uni6ersity Graz, Stremayrgasse 16, A-8010 Graz, Austria
b
Institute of Biotechnology, Technical Uni6ersity Graz, Petersgasse 12, A-8010 Graz, Austria
Received 12 October 1999; received in revised form 7 February 2000; accepted 21 February 2000
Abstract
A fast method for the simultaneous detection of different glycosidolytic activities in commercially available enzyme
preparations and crude culture filtrates was found in using, as substrate, a mixture of different glycosyl fluorides and
1
9
F NMR spectroscopy as a screening technique. Accompanying studies regarding the hydrolytic stability of these
fluorides in various buffer systems, as well as conditions of their long-term storage, were carried out. A simple
procedure for the preparation of b-D-mannopyranosyl fluoride in gram quantities is given. © 2000 Elsevier Science
Ltd. All rights reserved.
19
Keywords: Glycosyl fluorides; Glycosidases; F NMR spectroscopy; b-D-Mannopyranosyl fluoride
1
. Introduction
di- and oligosaccharides, as an alternative to
the chemical glycosylation, the enzyme-
catalysed glycoside synthesis constitutes a
promising option. Besides the powerful, but
expensive and often hardly accessible glycosyl
transferases [EC 2.4], the cheaper glycosidases
Increasing evidence of the essential role that
carbohydrates play in biological systems [2,3]
stimulates the search towards efficient meth-
ods for the synthesis of these compounds. For
[EC 3.2] have also been successfully used in
many instances [4–7]. The search for such
enzymes by screening crude enzyme prepara-
tions is time consuming since most methods
for determining glycosidase activities are
based on photometric measurements which
allow the testing of just one activity at a time
[
8]. In the course of an ongoing study towards
Scheme 1.
the enzymatic synthesis of di- and oligosac-
charides employing glycosidases from new
sources, a method for the simultaneous detec-
tion of individual glycosidolytic activities in
crude culture filtrates was sought.
ꢀ
See Ref. [1].
*
Corresponding author. Tel.: +43-316-8738244; fax: +
4
3-316-8738740.
E-mail address: dax@orgc.tu-graz.ac.at (K. Dax).
0008-6215/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0008-6215(00)00060-4

396
M. Albert et al. / Carbohydrate Research 327 (2000) 395–400
Table 1
19
a
F chemical shift values (l) of glycosyl fluorides used and of some fluorine containing salts in aq NaOAc–HOAc buffer (50
mmol, pH 5)
Glycosyl fluoride
l (ppm)
Glycosyl fluoride
l (ppm)
Salts
l (ppm)
a-
a-
a-
a-
D
D
D
-Man-F
-Glc-F
-GlcNAc-F
−138.6
−150.1
−146.0
−151.8
b-
b-Cell-F
b- -Gal-F
b-Lac-F
D
-Man-F
−146.6
−143.8
−143.1
−143.5
CF COONa
NaHF2
NaF
−75.4
−122
−133
3
D
L
a
-Fuc-F
Dependence (90.5 ppm) on the pH value was observed.
2
. Results and discussion
crystallisation; the same product had been ob-
tained previously in 7% yield by a similar
strategy [15]. Subsequent deacetylation with
Glycosyl fluorides are known to be accepted
as substrates by glycosidases [9] and methods
for the determination of a distinct glycosidase
activity, by monitoring the release of fluoride
with an ion-sensitive electrode, are elaborated
NH –MeOH furnished the desired b-
D-
3
mannopyranosyl fluoride (Scheme 2).
(
Scheme 1(b)) [10,11]. As an extension of this
principle, an approach for the simultaneous
detection of different glycosidase activities by
means of F NMR spectroscopy was investi-
gated. The resonance signals of the individual
1
9
19
glycosyl fluorides differ sufficiently in their F
chemical shift (see Table 1), consequently al-
lowing the detection of any glycosidolytic ac-
tivity by comparing the integral values over
the corresponding signals from a ‘blank’ and
an incubated sample (Scheme 1(a)).
Scheme 2.
Considering the lability of b-glycosyl
fluorides towards hydrolysis [16], the influence
of the buffer system (nature, concentration,
pH value) in this respect was examined. For
this purpose, a mixture of different glycosyl
The individual glycosyl fluorides were pre-
pared according to known methods [12], i.e.,
either by treatment of the corresponding per-
acetates with pyridine–HF followed by
deacetylation with a 0.01 M NaOMe–MeOH
solution to give the more thermodynamically
stable a-fluorides or by the reaction of per-O-
fluorides (plus CF COONa as an internal
3
standard) was dissolved in the respective
buffer system (containing 10% D O for NMR-
2
locking purposes) and the decrease in concen-
tration of the individual fluorides was
determined in appropriate time intervals by
acetylated glycosyl bromides with KHF in
2
CH CN under reflux followed by treatment
3
with a saturated solution of NH in MeOH to
19
3
F NMR spectroscopy. Half-life periods cal-
yield the corresponding b anomers. Nonethe-
culated therefrom are summarised in Table 2.
Based on these hydrolysis rates as well as
on general enzyme stability, we chose the
NaOAc–HOAc buffer (System 1 in Table 2)
for subsequent experiments. Similar studies
were carried out with regard to the long-term
storage of these fluorides. From the various
conditions tested for this purpose, solutions in
absolute methanol, stored at −80 °C, gave
less, b- -mannopyranosyl fluoride was not
D
accessible by this method. A published
procedure [13] for the synthesis of 2,3,4,6-
tetra-O-acetyl-b- -mannopyranosyl fluoride
D
using collidine-HF as fluorinating agent could
not be reproduced. However, fluorination
of 2,3,4,6-tetra-O-acetyl-b-
D
-mannopyranose
[
14] with diethylaminosulfur trifluoride
(
DAST) led to an anomeric mixture (a:b ratio
the best results, e.g., b- -galactopyranosyl
D
approximately 3:1), from which the pure b
anomer was isolated in 21% yield by simple
fluoride (the most labile) had a half-life period
of approximately 2 years.

M. Albert et al. / Carbohydrate Research 327 (2000) 395–400
397
Table 2
responding internal standard (IST (blank) and
IST (sample)) the proportion of unhydrolysed
glycosyl fluorides was obtained. From this, the
percentage of the enzyme catalysed hydrolysis
(EH) was calculated.
Half-life periods (t1/2 in h) of some glycosyl fluorides in
aqueous buffer systems
a
b
c
d
Glycosyl
fluoride
System 1 System 2 System 3 System 4
IGF (blank)
IGF (sample)
IST (sample)
a-
a-
b-
D
D
D
-Man-F 580
-Glc-F 290
190
230
7
29
13
13
39
290
290
6
39
14
17
29
290
290
8
39
11
11
29
V (blank)=
;
V (sample)=
;
IST (blank)
-Gal-F
7
V (blank)−V (sample)
V (blank)
b-Cell-F
b-
39
EH=
×100
D
-Man-F 14
b-Lac-F
14
29
As a general drawback, the accuracy in
quantitative determination of individual gly-
cosidolytic activities is limited by the fact that
certain glycosidases, in particular ‘inverting’
ones, are able to hydrolyse not only the stereo-
chemically corresponding glycosyl fluoride but
also its anomer; however, the hydrolysis rates
for these two reactions were found to differ by
two orders of magnitude [17,18]. Furthermore,
due to the high number of substrates present
in the sample, matrix and inhibition effects
may also affect the results.
An initial series of experiments using com-
mercially available glycosidases was carried
out. As an example, the results obtained from
a mixture of four glycosyl fluorides after incu-
bation (4 h) with a b-galactosidase from As-
pergillus oryzae are summarised in Table 3.
b-D
-Glc-F
a
b
c
NaOAc–HOAc, 90 mM, pH 5.
Tris–HCl, 90 mM, pH 7.5.
Phosphate–citrate, 90 mM, pH 7.0.
Phosphate–citrate, 180 mM, pH 7.0.
d
For screening experiments, defined solu-
tions containing comparable amounts of dif-
ferent glycosyl fluorides (and sodium
trifluoroacetate as internal standard) in dry
methanol were prepared. Therefrom, two
identical samples, containing approximately 3
mg of each fluoride, were taken and evapo-
rated to dryness. The samples were then redis-
solved in aqueous NaOAc–HOAc buffer
(
containing 10% D O for NMR referencing).
2
As expected, the corresponding b-
syl fluoride was completely hydrolysed. The
hydrolysis of a- -fucosyl fluoride and a- -glu-
D
-galacto-
One sample was either incubated directly with
an aqueous solution of a commercially avail-
able glycosidase or a defined volume of a
crude culture filtrate and the other — to take
into account the non enzymatic rate of
hydrolysis — was diluted with the same vol-
ume of neat buffer solution. The decrease of
the concentration of the individual glycosyl
fluorides (GF) was determined by the mea-
L
D
cosyl fluoride was attributed to side activities
in the enzyme preparation. The considerable
b-glucosidase activity, responsible for the par-
tial disappearance of the b-lactosyl fluoride,
could be confirmed by the conventional p-nitro-
phenyl method (using p-nitrophenyl b- -glu-
copyranoside as a substrate).
D
19
surement of the relative intensities of their F
NMR signals (IGF (sample) and IGF (blank))
after an appropriate interval of time (4–7 h).
By forming fractions (V (blank) and V (sam-
ple)) with the respective intensities of the cor-
The practical usability of this NMR method
was studied with crude culture filtrates of nine
different microorganisms. In each case, a solu-
tion of selected fluorides was incubated and
1
9
then analysed by means of F NMR spec-
Table 3
a
Enzyme catalysed hydrolysis (%) of the individual fluorides by an enzyme preparation from Aspergillus oryzae, with a
b-galactosidase as the main activity, after 4 h
a-
D
-Glc-F
a-L
-Fuc-F
b-Lac-F
b-
D
-Gal-F
Enzyme catalysed hydrolysis (%)
17
8
66
100
a
10 u b-galactosidase; NaOAc–HOAc buffer, 50 mM, pH 5.

398
M. Albert et al. / Carbohydrate Research 327 (2000) 395–400
Fig. 1. Blank sample. This spectrum was obtained 7 h after dissolving the set of glycosyl fluorides in acetate buffer (50 mmol, pH
). Reference signal (sodium trifluoroacetate) at −75.4 ppm is not shown.
5
Fig. 2. Study of a culture filtrate from Thermoascus aurantiacus. The spectrum was recorded 7 h after incubation with the crude
enzyme preparation. Reference signal (sodium trifluoroacetate) at −75.4 ppm is not shown.
Table 4
a
Individual rate of hydrolysis (%) of the glycosyl fluorides in the presence of a crude culture filtrate from Thermoascus aurantiacus
after 4 h
a-D
-Man-F b-
D
-Gal-F
b-Cell-F
a-
D
-GlcNAc-F b-
D
-Man-F a-
D
-Glc-F
a-L-Fuc-F
Enzyme catalysed
hydrolysis (%)
36
25
100
7
26
43
16
a
Addition of 5 vol% culture filtrate; NaOAc–HOAc buffer, 50 mM, pH 5.
troscopy. The results obtained with the ther-
mophilic fungus Thermoascus aurantiacus, af-
ter 7 h of incubation of a mixture of seven
glycosyl fluorides, are shown in Figs. 1 and 2
In most cases, the results obtained from
these experiments matched those obtained
from the photometric method. Due to mecha-
nistic differences in the hydrolysis of glycosi-
dases over glycosyl fluorides and p-nitro-
phenyl glycosides, a comparison can be under-
taken only qualitatively (low, medium or
(
Fig. 1 corresponds to the blank sample, Fig.
to the incubated one). The interpretation is
given in Table 4.
2

M. Albert et al. / Carbohydrate Research 327 (2000) 395–400
399
high activity). In some rare instances, photo-
metrically detected glycosidase activities could
agar) piece of an actively growing 4-day-old
PDA culture of T. aurantiacus and cultivated
at 47 °C on an orbital shaker (150 rpm) for 10
days. The crude broth was centrifuged and the
supernatant was used for further experiments.
19
not be verified with the F NMR method
and vice versa), which might be explained by
(
substrate specificities of the corresponding gly-
cosidases or attributed to matrix effects. Devi-
ations were also observed when coloured
culture filtrates were used.
2,3,4,6-Tetra-O-acetyl-i-
fluoride.—To a solution of 2,3,4,6-tetra-O-
acetyl- -mannopyranose (20.0 g, 57.4 mmol)
in abs CH Cl (300 mL) were added 10.5 mL
D-mannopyranosyl
D
Studies on the transglycosylation potential
of the crude culture filtrate of T. aurantiacus
are presented in a following paper [19].
2
2
DAST (80 mmol) at rt within 10 min. After 30
min the excess of DAST was quenched with
MeOH (20 mL) at 0 °C. The organic phase
was washed with water and satd aq NaHCO3
soln, then dried over Na SO . Concentration
2
4
3
. Experimental
under reduced pressure led to a brown oil
which was filtered through silica gel (100 g,
General procedures.—Melting points were
determined with a Tottoli apparatus (B u¨ chi
1
:1 EtOAc–cyclohexane eluent). The slightly
brown residue was redissolved in boiling
EtOH. 2,3,4,6-Tetra-O-acetyl-b- -manno-
3
00) and are uncorrected. Optical rotations
D
were measured with a Jasco DIP-360 digital
polarimeter at 589 nm at ambient tempera-
ture. NMR spectra were recorded at 300.13 or
pyranosyl fluoride precipitated as colourless
rhombic crystals. These were separated by
filtration (4.5 g, 21%) as soon as colourless
needles (the corresponding a anomer) started
to deposit; mp 106–106.5 °C, lit. 105.5–
1
13
1
99.98 MHz ( H), 75.47 or 50.29 MHz ( C)
19
and 282.4 MHz ( F), using a Bruker MSL
00 and a Varian Gemini 200 apparatus, re-
3
2
0
1
107 °C [15]; [h] +5.1° (c 1.1, CHCl ), lit.
D
3
spectively. As reference standards Me Si ( H
4
1
1
3
−1.0° (c 1.5, CHCl ) [15]; H NMR (199.98
3
and C NMR) and trichlorofluoromethane
1
9
MHz, CDCl ): l 5.53 (dd, 1 H, J 50.9 Hz,
3
1,F
(
F NMR) were used. TLC was performed on
Silica Gel 60 F 254 precoated aluminum plates
E. Merck 5554) with detection by charring
after spraying with vanillin–H SO (1%). For
J1,2 1.9 Hz, H-1), 5.45 (ddd, 1 H, J2,F 9.5 Hz,
J2,3 2.9 Hz, H-2), 5.1–5.3 (m, 2 H, H-3, H-4),
(
3
.90 (m, 1 H, J5,6 5.1 Hz, H-5), 4.2–4.4 (m, 2
2
4
13
H, H-6a, H-6b); C NMR (50.29 MHz,
CDCl ): l 104.0 (d, J 223 Hz, C-1), 66.6 (d,
column chromatography, Silica Gel 60, 230–
4
00 mesh (E. Merck 9385) was used.
3
1,F
J2,F 20 Hz, C-2), 68.3 (d, J3,F 6 Hz, C-3), 66.1
C4), 72.5 (d, J5,F 4 Hz, C-5), 62.5 (C-6); F
The crude culture filtrate used in this study
1
9
(
was that of the strain Thermoascus aurantiacus
Miehe, isolated from decomposed jute in
Bangladesh. It has been identified by Cen-
traalbureau voor Schimmelcultures (CBS),
Baarn, The Netherlands, and deposited in the
collection of the Institute of Biotechnology,
Technical University Graz, Austria, under the
stock number BT 2079. The basic mineral
NMR (282.4 MHz, CDCl ): l −142.6 (J
0.1 Hz, JF,H2 9.5 Hz).
3
F,H1
5
i-
tra-O-acetyl-b-
1.0 g, 2.85 mmol) was dissolved in a satd soln
of NH in abs MeOH (50 mL). After 15 h at
D
-Mannopyranosyl fluoride.—2,3,4,6-Te-
D-mannopyranosyl fluoride
(
3
4
°C, the solvent was removed under reduced
pressure at rt and b- -mannopyranosyl
fluoride was obtained as a colourless oil in
−
1
D
medium containing (g L
KH PO ; 0.5, MgSO ·7H O;
CaCl ·2H O; 40, Solka floc 40; 5, NH NO
tap water) 5,
0.05,
2
4
4
2
2
0
quantitative yield (0.52 g); [h]
−4.2° (c 2.15,
2
2
4
3
D
−
1
1
and 1 mL trace element solution (g L
tap
MeOH); H NMR (199.98 MHz, methanol-
water: 1.6, MnSO ·H O; 1.4, ZnSO ·7H O; 2,
d ): l 5.38 (dd, 1 H, J1,F 50.1 Hz, J1,2 1.0 Hz,
4
4
2
4
2
CoSO ·7H O) was used for growth. The initial
H-1), 3.76 (bd, 1 H, J2,F 11.8 Hz, H-2), 3.92
4
2
pH of the medium was adjusted to 5 prior to
sterilisation. The sterilised culture medium
(bs, 1 H, H-3), 3.5–4.5 (m, 3 H, H-4, H-6a,
1
3
H-6b), 3.33 (m, 1 H, H-5); C NMR (50.29
(
100 mL) in 300 mL Erlem neyer flasks was
MHz, methanol-d ): l 108.4 (d, J 210 Hz,
4
1,F
2
inoculated with a 1 cm PDA (potato dextrose
C-1), 71.2 (d, J2,F 18 Hz, C-2), 74.1 (d, J3,F 9

400
M. Albert et al. / Carbohydrate Research 327 (2000) 395–400
Hz, C-3), 68.1 (C4), 78.4 (d, J 4 Hz, C-5),
Wissenschaftlichen Forschung, Vienna, pro-
ject 11021-OECH. Dr Anna St u¨ tz is thanked
for proof-reading of the manuscript.
5
,F
1
9
6
2.7 (C-6);
NaOAc–HOAc buffer, 50 mM, pH 5): l
146.6 (d, J 48.3 Hz).
F NMR (282.4 MHz, aq
−
F,H1
1
9
F NMR enzyme assay.—A total of 0.33
mmol per glycosyl fluoride and 20 mg of the
internal standard sodium trifluoroacetate were
dissolved in abs MeOH (20 mL) for a series of
References
[
1] Parts of this work have been presented: M. Albert, W.
Repetschnigg, J. Ortner, B.J. Paul, H. Weber, W.
Steiner, K. Dax, Abstract of Papers, 19th International
Carbohydrate Symposium, San Diego, USA, August
2
0 experiments. For one sample, 1 mL of this
solution was taken out and the solvent was
removed under reduced pressure at 20 °C. The
samples were redissolved in the corresponding
1
998, AP090.
[
2] R.A. Dwek, Chem. Re6., 96 (1996) 683–720.
buffer (0.5 mL, 80 mM, 10 vol% D O) at the
[3] H. Lis, N. Sharon, Eur. J. Biochem., 218 (1993) 1–27.
2
[
4] D.H.G. Crout, G. Vic, Curr. Opin. Chem. Biol., 2 (1998)
8–111.
same time. A solution of the enzyme prepara-
tion (200 mL) or the crude culture filtrate (200
mL) was then added. The blank samples were
diluted with buffer (200 mL). In order to min-
imise errors, two identical samples were pre-
9
[
5] E.J. Toone, E.S. Simon, M.D. Bednarski, G.M. With-
esides, Tetrahedron, 45 (1989) 5365–5422.
[6] Y. Ichikawa, G.C. Look, C.-H. Wong, Anal. Biochem.,
02 (1992) 215–238.
2
[
7] K.G.I. Nilsson, ACS Symp. Ser., 466 (1991) 51–63.
8] H.U. Bergmeyer, Methods of Enzymatic Analysis, Vol.
IV, third ed., Verlag Chemie, Weinheim, 1984, pp. 145–
19
pared in each case. The F NMR spectra
were recorded 4–7 h later. The evaluation was
made after averaging the data of the corre-
sponding samples.
[
2
69.
[
9] J.E.G. Barnett, W.T.S. Jarvis, K.A. Munday, Biochem.
J., 105 (1967) 669–672.
Measurement of half li6es for spontaneous
hydrolysis of glycosyl fluorides.—The samples
were prepared as described above. They were
dissolved in the appropriate buffer system (4
[10] R. Belcher, M.A. Leonard, T.S. West, J. Chem. Soc.,
1959) 3577–3579.
11] G. Okada, D.S. Genghof, E.J. Hehre, Carbohydr. Res.,
1 (1979) 287–298.
12] T. Tsuchiya, Ad6. Carbohydr. Chem. Biochem., 48 (1990)
91–277 and Refs. cited therein.
13] Y.V. Voznij, I.S. Kalicheva, A.A. Galoyan, Bioorg.
Khim., 7 (1981) 406–409.
14] P. Kov a´ c, Carbohydr. Res., 153 (1986) 168–170.
(
[
[
[
[
7
19
mL, 10 vo1% D O) and the F NMR spectra
2
were recorded after 5, 40, and 80 min and 2.5,
4
, 18, and 40 h. The obtained hydrolysis
curves were in accordance to a first-order
expression over this period of time.
[15] K. Bock, C. Pedersen, Acta Chem. Scand. B, 29 (1975)
82–686.
6
[16] For a detailed study see: J.E.G. Barnett, Carbohydr.
Res., 9 (1969) 21–31.
[
17] T. Kasumi, Y. Tsumuraya, C.F. Brewer, H. Kersters-
Hilderson, M. Claeyssens, E.J. Hehre, Biochemistry, 26
(1987) 3010–3016.
Acknowledgements
[18] E.J. Hehre, H. Matsui, C.F. Brewer, Carbohydr. Res.,
1
98 (1990) 123–132.
We appreciate the financial support given
by the Austrian Fonds zur F o¨ rderung der
[
19] J. Ortner, M. Albert, K. Terler, W. Steiner, K. Dax,
Carbohydr. Res., in press.
.