Heterocyclic derivatives of sugars: The formation of 1-glycosyl-3-methylpyrazol-5-ones from hydrazones

Article abstract of DOI:10.1016/S0008-6215(00)00165-8

Conditions to effect the conversion of monosaccharide and disaccharide hydrazones to 1-glycosyl-3-methylpyrazol-5-ones were examined. The sugar pyrazolone derivatives were sensitive to oxidation, but high yields were achieved with 2,2,2-trifluoroethyl acetoacetate in mildly acidic solution. Azo coupling of the pyrazolones produced highly coloured azopyrazolone derivatives that prevented further degradation, and these may prove useful labels for chromatographic analysis of carbohydrates. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0008-6215(00)00165-8

www.elsevier.nl/locate/carres
Carbohydrate Research 329 (2000) 169–177
Heterocyclic derivatives of sugars: the formation of
1-glycosyl-3-methylpyrazol-5-ones from hydrazones
Warren C. Kett *, Michael Batley, John W. Redmond
Department of Chemistry, Di6ision of En6ironmental and Life Sciences, Macquarie Uni6ersity, North Ryde,
NSW 2109, Australia
Received 27 September 1999; accepted 21 May 2000
Abstract
Conditions to effect the conversion of monosaccharide and disaccharide hydrazones to 1-glycosyl-3-methylpyrazol-
5-ones were examined. The sugar pyrazolone derivatives were sensitive to oxidation, but high yields were achieved
with 2,2,2-trifluoroethyl acetoacetate in mildly acidic solution. Azo coupling of the pyrazolones produced highly
coloured azopyrazolone derivatives that prevented further degradation, and these may prove useful labels for
chromatographic analysis of carbohydrates. © 2000 Elsevier Science Ltd. All rights reserved.
Keywords: Sugar hydrazone; Pyrazolone; Azopyrazolone; Carbohydrate labelling
ring may offer better detection characteristics
1. Introduction
and result in improved separations, and we
have now investigated the use of methyl pyra-
zolones. Furthermore, pyrazolones open up
possibilities for further derivatisation using
methods that have been developed during
their long history in the preparation of dyes
and colourants [5,6].
Previously, the feasibility of preparing pyra-
zole derivatives of sugars from saccharide hy-
drazones was demonstrated [1] as part of a
search for alternatives to common procedures
for the analysis of glycans [2,3]. The derivati-
sation exploited the tautomerism of acyclic
saccharide hydrazones to cyclic glycosylhy-
drazines [4], which enabled the use of familiar
hydrazine chemistry to generate the hetero-
cyclic ring. The pyrazoles were formed in a
simple two-step process from glycan hydra-
zones, but a major disadvantage was that
dimethylpyrazole absorbs in the near-UV
spectrum and has only a modest extinction
coefficient. Modification of the heterocyclic
The archetypal pyrazolone synthesis is the
condensation of a b-ketoester, such as ethy-
lacetoacetate, with a hydrazine [5,6]. Previous
attempts at the formation of glycosylpyra-
zolones in this manner yielded an intractable
mixture [7], but we have been able to produce
pyrazolone derivatives of selected mono- and
disaccharides in good yield by optimising the
reaction conditions and azo-coupling the
pyrazolones to improve their stability. The
highly coloured azo-pyrazolone derivatives
may prove useful for the labelling of glycans,
which appear to be especially attractive for the
analysis of the products of hydrazinolysis of a
glycoprotein, or b-elimination in the presence
of hydrazine [8–10].
* Corresponding author. Present address: Molecular Im-
munology Laboratory, School of Biomedical Sciences, Curtin
University of Technology, Level 5, Rear 50 Murray Street,
Perth, WA 6000, Australia. Tel.: +61-8-92240356; fax: +61-
8-92240360.
E-mail address: wkett@waimr.uwa.edu.au (W.C. Kett).
0008-6215/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0008-6215(00)00165-8

170
W.C. Kett et al. / Carbohydrate Research 329 (2000) 169–177
about 220 nm. The absorption spectra of the
2. Results
glycosylpyrazolones separated by HPLC indi-
cated that the principal species was the 5-pyra-
zolone isomer 2 as expected (Scheme 1), but
small amounts of the 3-pyrazolone isomer, 10,
were also present.
Reactions with ethyl acetoacetate.—The ad-
dition of ethyl acetoacetate to freshly prepared
aqueous solutions of lactose hydrazone pro-
duced deep red-coloured and complex mix-
tures. Analysis by HPLC using porous
graphitic carbon (PGC) or hydrophilic inter-
action chromatography with diode array de-
tection or ESIMS ([M+H]+=423 m/z;
[M−H]−=421 m/z) enabled the detection of
the 1-lactosylpyrazolone products. The reac-
tion was not instantaneous, and a reaction
intermediate was observed as described below.
Both isomers of the pyrazolone ring were
detected. Each isomer can exist in more than
one tautomeric form, with the proportions
depending on the solvent [6,11]. In aqueous
solution, the main form of 5-pyrazolones is
the NH tautomer, which has an absorption
maximum in the range of 250–260 nm, while
3-pyrazolones exist predominantly as the OH
tautomer with an absorption maximum at
The glycosylpyrazol-5-one derivatives de-
graded rapidly after chromatographic isola-
tion and had to be further derivatised using
azo coupling before they could be character-
ised. Crude preparations of lactosylpyra-
zolones and fucosylpyrazolones were reacted
at moderately alkaline pH with diazotised 4-
methyl aminobenzene and yellow 1-glycosyl-
4-(4-methylbenzeneazo)-3-methylpyrazol-5-
ones, and1-glycosyl-4-(4-methylbenzeneazo)-
5-methyl-pyrazol-3-ones (Scheme 2) formed
almost instantly. The products were purified
by reversed-phase HPLC and characterised by
visible and NMR spectroscopy and collision-
induced dissociation ESIMS/MS.
The mass spectra contained peaks corre-
sponding to the expected pseudo-molecular
Scheme 1. Reaction pathways for the formation of 1-lactosylpyrazol-5-one (2) and 1-lactosylpyrazol-3-one (10) from lactose
hydrazone. The azine intermediate 1 must adopt the Z conformation to enable heterocyclic ring closure [17].

W.C. Kett et al. / Carbohydrate Research 329 (2000) 169–177
171
to the glycan moiety and an abstracted hydro-
gen. The lactosyl derivatives gave additional
ions corresponding to loss of (i) the galactose
residue and (ii) loss of the azo groups. The
UV absorption maximum of the major isomer
was at 418 nm, about a 70 nm longer wave-
length than that for the other isomer, confirm-
ing that the predominant product was the
tolylazopyrazol-5-one isomer [11].
The predominant sugar ring form was b-
pyranosyl for both the glucosyl and fucosyl
derivatives, as shown by the chemical shift
and coupling constant of the anomeric proton
1
signal in the H NMR spectra of the com-
pounds. The spectrum of the lactosyl com-
pound, recorded in D2O, displayed broad
resonances until base was added. The broad-
ening was probably the result of interconver-
sion between four tautomers [11], which was
eliminated when the pyrazolone was converted
1
to the anion [12,13]. The H NMR spectrum
of the fucosyl azopyrazolone derivative was
acquired in chloroform solution and was not
broadened, presumably because a single tau-
tomer is favoured in this solvent.
The distribution of the other sugar ring
forms produced by selected mono- and disac-
charide hydrazones was determined after con-
version to azopyrazolone derivatives (Table
1). The reaction mixtures were separated by
reversed-phase HPLC with diode array detec-
tion, and the pyrazolone ring isomers were
identified by their characteristic UV–Vis ab-
sorption spectra. The number of isomers and
their proportions were determined, but no at-
tempt was made to establish the configuration
of each isomer. Previously, we showed that
the distribution of sugar-ring isomers in pyra-
zole derivatives [1] was determined by a com-
plex equilibrium involving intermediates
produced by the first carbonyl condensation
step. It was, therefore, expected that the distri-
bution of ring forms for the pyrazolone
derivatives would be similar, and the isomeric
proportions were broadly similar to those for
the pyrazoles, with a slightly greater prepon-
derance of the major isomer.
Scheme 2. Formation of azopyrazolones.
Table 1
Proportions of tolylazopyrazol-5-one isomers from six aldoses
and a ketose
Isomer proportions a
Glucose
Galactose
Mannose
Ribose
Fructose
Lactose
94:2:1:3
69:6:7:18
2:4:93:1
6:36:37:21
74 b:13:13
97:3
Melibiose
92:4:3:1
a The proportions were determined by reversed-phase
HPLC, and the isomers are listed in order of elution.
b The major peak appears to be two incompletely resolved
isomers.
Optimisation of pyrazolone formation.—The
effects of using other b-ketoesters that pro-
vided a systematic variation in the reactivities
of the ester moieties and of varying reaction
ions, and the principal collision-induced frag-
ments of these ions had lost mass equivalent

172
W.C. Kett et al. / Carbohydrate Research 329 (2000) 169–177
conditions were evaluated by derivatising lac-
tose hydrazone. The reaction products were
analysed by HPLC using two separation
modes, PGC [14,15] and hydrophilic interac-
tion media [16]. Peak areas were determined
with diode-array detection, while on-line ES-
IMS was used for peak identification.
The reaction of 2,2,2-trifluoroethyl acetoac-
etate with lactose hydrazone was much faster,
and the yield was sensitive to the buffer used
(Fig. 1). At longer reaction times, degradation
of the product became more noticeable, and
when the derivatisation was repeated with lac-
tose hydrazone at 100 times greater concentra-
tion in methanol–water solvent, 50% of the
lactosylpyrazolone had degraded within 4 h. It
is believed that oxidation of the pyrazolone
ring is responsible [18–21]. Under optimal
conditions (0.2 M CH3CO2Na buffer, pH 4.6),
the reaction was complete within 2 h (Fig. 1)
and degradation was negligible.
Precise quantitation of the amount of pyra-
zol-3-one was not possible, but assuming that
the extinction coefficient at 227 nm for the
pyrazol-3-one derivative is approximately half
that for the pyrazol-5-one derivative at 246
nm [11], the proportion of the pyrazol-3-one
isomer formed was less than 1% for all three
alkyl esters.
When N-hydroxysuccinimidyl acetoacetate
was reacted with lactose hydrazone, approxi-
mately equal amounts of the pyrazol-5-one
and pyrazol-3-one isomers were formed. It is
likely that the increased reactivity of the N-
hydroxysuccinimidyl moiety [22] resulted in
acylation of the saccharide hydrazone
(Scheme 1, paths D–F) rather than enamine
formation. Hence, 2,2,2-trifluoroethyl acetoac-
etate is a better reagent for our purposes as it
provides a compromise between the two ob-
jectives of increasing ester reactivity to effect
fast pyrazolone ring closure and the formation
of a single pyrazolone isomer.
The reaction of lactose hydrazone with ei-
ther methyl acetoacetate or ethyl acetoacetate
produced an intermediate that was detected by
hydrophilic interaction chromatography. The
amount of intermediate decreased with time in
a first-order manner, while the amount of
lactosylpyrazol-5-one product increased. The
intermediate formed from methyl acetoacetate
decayed with a half-life of approximately 5 h,
whilst the half-life for that formed from ethyl
acetoacetate was 7.5 h. The intermediates are
believed to be the cyclic azines, 1 (R=methyl
or ethyl), and not the open-chain diazines 3,
because ketone condensation with the outer
nitrogen of a saccharide hydrazone (Scheme 1,
path A) has been shown to lead to rapid sugar
ring closure [1]. Heterocyclic ring closure of
these azines is much slower than reported for
similar azines formed from alkyl and phenyl
hydrazines [17].
Deri6atisation efficiency.—An estimate of
the derivatisation efficiency was obtained for
the formation of the tolylazopyrazolone
derivative from lactose by using HPLC analy-
sis with detection at 418 nm. For calibration,
the N-unsubstituted analogue, 4-(4-methyl-
benzeneazo)-1-H-3-methylpyrazol-5-one (12c),
was prepared, purified and checked by NMR
and HPLC. In view of the strong qualitative
similarities between the visible absorption
spectra of the glycosylated and non-glycosy-
lated azopyrazolones, it was assumed that the
extinction coefficients at 418 nm for the glyco-
syl derivatives were the same as that for the
standard compound.
Fig. 1. Time course of the formation of 11a from lactose
hydrazone by reaction with 2,2,2-trifluoroethyl acetoacetate
at room temperature in (A) 30:70 MeOH–water; (B) 30:70
MeOH–0.2 M CH3CO2Na pH 4.6; (C) 30:70 MeOH–0.2 M
Na2HPO4 PH 6; (D) 30:70 MeOH–25 mM Na2HPO4 pH 6;
(E) 30:70 MeOH–25 mM Na2B4O7 pH 8. Yields are ex-
pressed relative to the maximum yield observed. The initial
concentration of lactose hydrazone was 1 mM. HPLC analy-
sis was performed on a PGC column.

W.C. Kett et al. / Carbohydrate Research 329 (2000) 169–177
173
The HPLC peak area was a linear function
of the amount of 4-(4-methylbenzeneazo)-1-
H-3-methylpyrazol-5-one injected, over the
range of 20–5000 pmol. The peak areas for
the tolylazo derivative 12a, obtained from
derivatising amounts of lactose in the range of
4–200 nmol were fitted by least-squares analy-
sis, and a straight line was obtained with a
slope that was 8097% of the slope obtained
for the calibrant solution [21].
high efficiency are the use of an active ester
like 2,2,2,-trifluoroethyl acetoacetate [24] and
the suppression of oxidation in mildly acidic
solution [21]. We have avoided more forcing
conditions that might accelerate heterocyclic
ring closure, for fear that this might compro-
mise acid-sensitive substituents like sialic
acids.
4. Experimental
General.—2,2,2-trifluoroethyl acetoacetate
[25] and 1,3-dimethylpyrazol-5-one [26] were
synthesised according to published proce-
dures. 4-(4-Methylbenzeneazo)-1,3-dimethyl-
pyrazol-5-one and 4-(4-methylbenzeneazo)-1-
H-3-methylpyrazol-5-one was prepared by
coupling diazotised 4-methyl aminobenzene in
the usual manner [27] followed by recrystalli-
sation of the products from EtOAc. No sig-
nificant impurities were detected in the
3. Discussion
Veibel and co-workers have shown that 5-
pyrazolones are susceptible to oxidation to
bispyrazolones and 4-hydroxypyrazolones
[18–21]. The reaction occurs via a free-radical
mechanism, and the kinetic order was found
to vary according to the substrate [18,23].
Oxidation was faster in alkali and was more
complex in aqueous solution, resulting in un-
characterised oils [19]. Pyrazol-4,5-diones and
hydroxy-bispyrazolones have also been iden-
tified as oxidation products of pyrazolones
[23]. Although the pyrazolone ring is normally
stable in either strong acid or base [5], pyra-
zol-4,5-diones have been shown to undergo
pyrazolone ring opening to generate oxobu-
tanoic acid hydrazones [23].
Glycosylpyrazolones appear to be especially
susceptible to oxidation. Indeed, in a previous
study, ethyl acetoacetate produced an in-
tractable mixture when added to galactose
hydrazone [7]. It seems likely that the condi-
tions used in that study (basic conditions with
extended time at reflux) caused extensive oxi-
dation of the galactosylpyrazolone initially
formed. In the present investigation we were
able to use high-resolution chromatography
with mass spectrometry to show that the pyra-
zolones were present and were able to convert
them to stable azopyrazolones.
1
products using either H NMR spectroscopy
or reversed-phase HPLC with multiwave-
length detection. Glycosylpyrazole derivatives
were prepared as described previously [1]. All
other materials were purchased from Sigma-
Aldrich (Sydney, Australia). All commercial
products were used without prior purification.
NMR spectra were acquired at 27 °C on a
Varian XL-400 spectrometer operating at 400
1
MHz for H and 100 MHz for 13C observa-
tion. Proton decoupling was used to assist in
spectral assignments. Chemical shifts are re-
ported downfield from Me4Si, which was used
as internal reference in deuterochloroform so-
lutions, or from sodium 3-(trimethylsil-
yl)propionic acid, which was used in D2O
solutions.
The HPLC system was comprised of two
LC-10 pumps, SGU-104 solvent degassing
unit, SPD-10MA photodiode array detector,
10Axl autosampler, manual Rheodyne injec-
tor (Rheodyne, Cotari, USA), and FRC-10
fraction collector. The pumps were configured
for either binary high-pressure mixing or qua-
ternary low-pressure mixing as required. Data
acquisition and instrument control was
achieved using Shimadzu CLASS 5000
software.
Hence, it has been shown that saccharide
hydrazones can be converted efficiently to 1-
glycosyl-3-methylpyrazol-5-one derivatives un-
der conditions suitable for use in routine
analysis. In order to maintain high yields,
however, the pyrazolones must be further
derivatised as quickly as possible, for example,
by azo coupling. The key requirements for

174
W.C. Kett et al. / Carbohydrate Research 329 (2000) 169–177
concd on a rotary evaporator, and lyophilised
Electrospray-ionisation mass spectrometry
(ESIMS) was performed with a Fisons VG
Quattro spectrometer (VG Analytical,
to give a yellow gum (70 mg; 0.13 mmol;
1
15%). H NMR: (ꢀ0.02 M ND3 in D2O): l
2.37 (s, 6 H, 2 methyl groups), 3.55 (dd, 1 H,
H-2%, J2%3% 9.7 Hz), 3.65–4.00 (m, 10 H, H-3–
H-6 and H-3%- H-6%), 4.10 (dd, 1 H, H-2, J23
9.0 Hz), 4.52 (d, 1 H, H-1%, J1%2% 7.9 Hz), 5.35
(d, 1 H, H-1, J12 9.2 Hz), 7.33 (d, 2 H, J 5.9
Hz), 7.55 (d, 2H, J 6.3 Hz). Positive-ion ES-
IMS/MS: [M+H]+ =m/z 541. Daughters of
m/z 541 (CID, 16 V): m/z 379 (relative inten-
sity (RI)=90%, loss of galactose with hydro-
gen abstraction), m/z 259 (RI=60%, loss of
azo moiety and two hydrogens from m/z 379),
m/z 217 (RI=100%, loss of lactose with hy-
drogen abstraction). Negative-ion ESIMS:
[M−H]−=m/z 539.
Preparation of 1-i-lactosyl-4-(4-methylben-
zeneazo) - 5 - methylpyrazol - 3 - one (14a).—
Freshly prepared lactose hydrazone (0.25 g,
0.75 mmol) was dissolved in 3:7 dioxane–wa-
ter (5 mL), and NHS acetoacetate (0.5 g, 2.5
mmol) was added. After 30 min the sample
was diluted with water (5 mL) and applied to
a 50×20 mm column packed with Hypercarb
(Alltech Associates, Sydney, Australia). The
column was washed with 9:1 MeOH–13 mM
TFA (300 mL at 10 mL/min) to elute most of
the 3-methylpyrazol-5-one formed from excess
hydrazine. The lactosylpyrazolones were
eluted with 5:2:13 MeCN–MeOH–13 mM
TFA (50 mL) and lyophilised. The residue
was dissolved in 0.5 M Na2CO3 (5 mL), diazo-
tised 4-methyl aminobenzene (2 mmol) was
added, and the solution was stirred for 30
min. Methyl acetoacetate (0.4 mL) was added,
and stirring was continued for a further 10
min. The pH was adjusted to 6, and the
mixture was applied to a preconditioned 1-g
C18 solid-phase extraction cartridge, which
was then washed with water (10 mL). The
lactosylazopyrazolones were eluted with 3:7
MeCN–water (6 mL) and lyophilised. The
product was analysed by HPLC on a 150×
4.6 mm Hypersil ODS column. The gradient
was 3:17 MeCN–0.026 M HCO2H for 1 min,
followed by a linear gradient to 7:13 MeCN–
0.026 M HCO2H at 21 min. The eluent was
monitored at 350 and 420 nm. The mixture
contained mostly 1-lactosyl-4-(4-methylben-
Manchester, UK). Data acquisition and in-
strument control was achieved using Mass
Lynx software (VG Analytical). The Shi-
madzu HPLC was coupled with the mass
spectrometer for LC/MS experiments. A 10:1
split ratio between the detector effluent and
the mass spectrometer inlet was achieved us-
ing a T-piece.
Preparation of 1-i- -lactosyl-4-(4-methyl-
D
benzeneazo)-3-methylpyrazol-5-one (12a).—
Lactose (1 g, 3 mmol) was converted to the
hydrazone as described previously [1]. The
residue was dissolved in 1:3 MeCN–water (20
mL) that had been purged with argon, and
2,2,2-trifluoroethyl acetoacetate (2 mL) was
added. Argon was bubbled through the mix-
ture for a further 3 h, and the sample was
lyophilised to yield 2.2 g of a yellow powder
containing mostly 3-methylpyrazol-5-one (11c)
and 1-lactosyl-3-methylpyrazol-5-one (11a). A
portion of the crude product (0.5 g, corre-
sponding to approximately 0.7 mmol of 11a)
was dissolved in 0.5 M NaHCO3 (50 mL).
Diazotised 4-methyl aminobenzene (5 mmol)
was added with stirring, the solution turned
yellow immediately, and a yellow precipitate
began to deposit. After 10 min, 2,4-pentane-
dione (1 mL) was added, and the solution was
stirred for a further 10 min before the pH was
adjusted to pH 6 by the dropwise addition of
glacial AcOH. The mixture was washed with
CHCl3 (3×30 mL), and the washings were
discarded. The product was purified from the
aq liquor by HPLC chromatography on a
100×10 mm polymeric Jordi column using
repetitive injection on a BIOCAD chromatog-
raphy workstation (Perseptive Biosystems,
MA, USA). The gradient was 3:17 MeCN–
0.026 M HCO2H for 5 min, followed by a
linear gradient to 35:65 MeCN–0.026 M
HCO2H at 25 min, then to 6:4 MeCN–0.026
M HCO2H at 27.5 min where the composition
was maintained for 3 min. The flow rate was
1.8 mL/min, and the eluted species were mon-
itored at 420 and 250 nm. Fractions corre-
sponding to 1-lactosyl-4-(4-methylbenzen-
eazo)-3-methylpyrazol-5-one were pooled,

W.C. Kett et al. / Carbohydrate Research 329 (2000) 169–177
175
gen at 2 L/min and was placed in-line after the
diode-array detector.
(i ) Preparation of lactose hydrazone stan-
dard.—Two stock solutions containing 7.36
mM lactose in 2:3 EtOH–water were pre-
pared, with one also containing approximately
zeneazo)-3-methylpyrazol-5-one, but 1-lactosyl-
4-(4-methylbenzeneazo)-5-methylpyrazol-3-
one was also present, eluting 0.8 min later
than the major isomer and having a visible
absorption maximum at 348 nm. This peak
was collected and analysed by ESIMS/MS.
[M+H]+=m/z 541. Daughters of m/z 541
(collision voltage 16 V): m/z 379 (RI=16%,
loss of galactose with hydrogen abstraction),
m/z 259 (RI=13%, loss of tolylazo moiety
and two hydrogens from m/z 379), m/z 217
(RI=100%, loss of lactose with hydrogen ab-
straction), m/z 163 (RI=18%, galactose
residue), m/z 145 (RI=22%, loss of water
from galactose residue). Negative-ion ESIMS:
[M−H]−=m/z 539.
1 mg/mL 1- -glucosyl-3,5-dimethylpyrazole as
D
internal standard, and were stored at −20 °C.
Aliquots (100 mL) of either solution were
placed into autosampler vials, hydrazine
monohydrate (20 mL) was added, and the
samples were incubated at rt overnight. After
drying in a centrifugal evaporator, toluene
(0.5 mL) was added, and the samples were
re-evaporated. Finally, water (100 mL) was
added, and the samples were evaporated again
before capping and storage at 4 °C for up to
10 days.
Time-course studies of the formation of
lactosylpyrazolones
Analysis by HPLC on PGC. The reaction
mixtures were separated on a Shandon Hyper-
sil 5 mm, 100×4.6 mm PGC column (Alltech
Associates) using a flow rate of 1 mL/min.
Quaternary or ternary gradients (formed by
configuring the Shimadzu LC-10 for low-pres-
sure mixing) were used depending on the
acidic component of the eluent. For mass
spectrometry the volatile eluents were: A., wa-
ter; B., MeCN; C., 0.26 M trifluoroacetic acid;
and D., MeOH. For quantitative analysis the
eluents were A., 8 mM H2SO4; B., MeCN; C.,
MeOH. Aliquots (20 mL) of the sample were
withdrawn at intervals by the autosampler.
The column was maintained at 35 °C, the flow
rate was 1 mL/min and the detection wave-
lengths were 230 and 250 nm.
Analysis by hydrophilic interaction HPLC.
Either a PolyLC 5 mm, 100×2.1 mm
polyethylaspartamide column (PolyLC Inc.,
Columbia, USA) or a Lichrosorb diol 5 mm,
200×2.1 mm column (Alltech Associates)
was used. Injections of 10 mL were made at
intervals using an autosampler. Gradients of
MeCN and water (formed by configuring the
Shimadzu LC-10 for binary high-pressure
mixing) were used at a flowrate of 0.22 mL/
min. The columns were maintained at 30 °C,
and the detection wavelengths were 210, 230
and 250 nm. For some experiments an evapo-
rative light scattering detector (Alltech Associ-
ates) was used to detect lactose. This detector
was operated at 110 °C, with a flow of nitro-
(ii ) Formation of lactose pyrazolones using
i-ketoesters.—The formation using 2,2,2-tri-
fluoroethyl acetoacetate was performed as fol-
lows. The procedure for other esters was simi-
lar. A sample of lactose hydrazone (0.75
mmol), prepared as in (i), was dissolved in
water (160 mL). Aliquots (50 mL) were placed
into two 800-mL crimp-top autosampler vials
and buffer or solvent (300 mL) was added. A
solution of 1:4 2,2,2-trifluoroethyl acetoac-
etate–2,2,2-trifluoroethanol (15 mL) was
added, the vial capped and placed into the
HPLC autosampler. HPLC analysis was per-
formed using either a PGC column or a
PolyLC column. The PGC column was eluted
with a ternary gradient at 1 mL/min. The
gradient was 90:3:7 8 mM H2SO4–MeCN–
MeOH for 1 min, then changed to 79:14:7 at
14 min, then to 43:50:7 at 18 min. The PolyLC
column was eluted at 0.22 mL/min with a
gradient comprised of 47:3 MeCN–water for
1 min, then changed linearly to 7:3 at 18 min,
and then to 3:7 at 21 min. The columns were
equilibrated between analyses for 6 min
(PGC) or 10 min (PolyLC).
(iii ) Monitoring the degradation of lacto-
sylpyrazolones in concentrated solutions.—
Lactose hydrazone (25 mg, 80 mmol) was
dissolved in 3:7 MeOH–water (1 mL), 2,2,2-
trifluoroethyl acetoacetate (20 mL) was added,
and the vials were capped and placed into the
autosampler. The autosampler was pro-

176
W.C. Kett et al. / Carbohydrate Research 329 (2000) 169–177
added and incubated at rt overnight. After
drying in a centrifugal evaporator, 0.5 mL of
toluene was added, and the samples were re-
evaporated. The residue was dissolved in wa-
ter (200 mL), and an aliquot (50 mL, 350 nmol)
was taken for conversion to the azopyrazolone
derivative and HPLC analysis in the manner
described for the estimation of yield (above).
grammed to dilute an aliquot of this sample at
intervals, immediately prior to analysis.
Aliquots (10 mL) were added to water (1.0
mL) and mixed, and 10-mL samples were in-
jected. Analysis was performed using PGC as
described above for lactosylpyrazolones. Peak
areas were measured relative to that of 3-
methyl-1H-pyrazolone.
Preparation of 1-i- -fucopyranosyl-4-(4-
L
Estimation of the yield of 1-lactosyl-4-tolyl-
azopyrazol-5-one.—An aliquot (2–100 mL) of
a stock solution of lactose containing between
1.6 and 400 nmol was placed into an autosam-
pler vial and evaporated to dryness. Hydra-
zine hydrate (25 mL) and water (25 mL) were
added, and after 4 h the sample was evapo-
rated to dryness. Toluene was added (50 mL),
and the samples were evaporated, dissolved
in water (25 mL), and re-evaporated before
being dissolved in 3:7 MeOH–0.2 M sodium
acetate pH 4.6 (100 mL). After the addition
of 1:4 2,2,2-trifluoroethyl acetoacetate–2,2,2
trifluoroethanol (20 mL), the vial was capped
and incubated at rt for 1 h. The sample was
evaporated to dryness (centrifugal evapora-
tor), redissolved in water (25 mL), and re-
evaporated before being dissolved in satd
NaHCO3 (0.5 mL), followed by the addition
of diazotised aminotoluene (2–20 mmol). Af-
ter 10 min, the reaction was quenched by
adding acetylacetone (20 mL), the sample was
left for 10 min, then it was quantitatively
transferred to appropriate volumetric flasks to
produce solutions containing between 0.3 and
20 nmol/mL of the lactose derivative. The
insoluble precipitates formed from reagent ex-
cess were observed to redissolve upon dilu-
tion. Replicate samples were prepared.
The solutions were analysed by HPLC us-
ing a Keystone C18 column, 5mm, 150×4.6
mm. The gradient was 16:85 MeCN–(1:19
MeCN–0.02 M ammonium acetate pH 5.9)
for 1 min, increasing linearly to 7:18 at 17
min, then increased 3:2 at 20 min and main-
tained for 3 min. Aliquots (25 mL) of each
sample were injected using an autosampler,
and the detection wavelength was 420 nm.
Estimation of the number of isomers formed
for selected monosaccharides and disaccha-
rides.—Approximately 1.4 mmol of sugar dis-
solved in water (100 mL) was placed into a
vial, 20 mL of hydrazine monohydrate was
methylbenzeneazo) - 3 - methylpyrazol - 5 - one
(12b).—Fucose hydrazone was prepared ac-
cording to the method of Williams [4]. Fucose
hydrazone (100 mg, 0.56 mmol) was dissolved
in 1:3 MeCN–water (5 mL) and 2,2,2-tri-
fluoroethyl acetoacetate (0.2 mL) was added.
After 3 h, the sample was lyophilised, the
residue was dissolved in 0.5 M NaHCO3 (10
mL), and diazotised 4-methyl aminobenzene
(1 mmol) was added in aliquots (4×0.5 mL)
within 2 min. After 5 min, 2,4-pentanedione
(0.2 mL) was added, and the mixture stirred
for a further 10 min. The mixture was aci-
dified with AcOH to pH 6 and washed with
CHCl3 (2×10 mL). It was purified by re-
versed-phase HPLC as described above for
1-lactosyl-4-(4-methylbenzeneazo)-3-methyl-
pyrazol-5-one, concd on a rotary evaporator,
and lyophilised to give a yellow gum (50 mg;
1
0.14 mmol; 25%). H NMR (CDCl3): 1.34 (d,
3 H, H-6), 2.26 (s, 1 H, methyl), 2.24 (s, 1 H,
methyl), 3.61 (dd, 1 H, H-3, J3,4 3.5 Hz), 3.67
(d, 1 H, H-4, J4,5 0 Hz), 3.82 (q, 1 H, H-5, J5,6
6.5 Hz), 4.17 (dd, 1 H, H-2, J2,3 9.2 Hz), 5.07
(d, 1 H, H-1, J1,2 9.2 Hz), 7.16 (d, 2 H, H-2%,
J2%3% 8.4 Hz), 7.26 (d, 2 H), 13.3 (br s, OꢀH).
13C NMR (CDCl3): 11.8, 16.5 and 21.0 (3
Me), 68.0, 72.0, 73.2 and 75.0 (Fuc C-2, C-3,
C-4 and C-5), 82.1 (Fuc C-1), 115.9 (Ph C-2/
6), 127.2 (Ph C-4), 130.2 (Ph C-3/5), 136.2 (Ph
C-1), 138.7 (Pyr C-4), 148.8 (Pyr C-3), 159.3
(Pyr C-5). Positive-ion ESIMS: [M+H]+=
m/z 363. Negative-ion ESIMS: [M+H]+=
m/z 363.
References
[1] W.C. Kett, M. Batley, J.W. Redmond, Carbohydr. Res.,
299 (1997) 129–141.
[2] S. Hase, Pre- and post-column detection oriented
derivatisation techniques in HPLC, in Z. El Rassi (Ed.),
Carbohydrate Analysis: High Performance Liquid Chro-

W.C. Kett et al. / Carbohydrate Research 329 (2000) 169–177
177
matography and Capillary Electrophoresis, Elsevier, Am-
sterdam, 1995, pp. 555–575.
tion chromatography of carbohydrates and glycoconju-
gates, in Z. El Rassi (Ed.), Carbohydrate Analysis: High
Performance Liquid Chromatography and Capillary Elec-
trophoresis, Elsevier, Amsterdam, 1995, pp. 41–101.
[15] K. Koizumi, J. Chromatogr. A, 720 (1996) 119–126.
[16] S.C. Churms, J. Chromatogr. A, 720 (1996) 75–92.
[17] A.R. Katritizky, P. Barcynski, D.L. Ostercamp, J.
Chem. Soc., Perkin Trans. 2, (1987) 969–975.
[18] S. Veibel, Acta Chem. Scand., 26 (1972) 3685–3690.
[19] S. Veibel, S.C. Linholt, Acta Chem. Scand., 9 (1955)
970–974.
[20] S. Veibel, S.C. Linholt, Acta Chem. Scand., 9 (1955)
963–969.
[21] S. Veibel, G. Westoo, Acta Chem. Scand., 7 (1953)
119–127.
[22] G.W. Cline, S.B. Hanna, J. Org. Chem., 53 (1988) (1988)
3583–3586.
[23] Y. Yamamoto, T. Kuwahara, K. Watanabe, K. Watan-
abe, Redox Report, 2 (1996) 333–338.
[3] S. Honda, J. Chromatogr. A, 720 (1996) 337–353.
[4] J.M. Williams, Carbohydr. Res., 117 (1983) 89–94.
[5] R.H. Wiley, P. Wiley, The Chemistry of Heterocyclic
Compounds, Vol. 20, Interscience, New York, 1964.
[6] J. Elguero, Pyrazoles and their benzo derviatives, in
A.R. Katritzky (Ed.), Comprehensi6e Heterocyclic
Chemistry, Vol. 5, Pergamon, Oxford, 1984, pp. 167–
303.
[7] D. Mackiewiczowa, J. Jasinska, J. Sokolowski, Pol. J.
Chem., 55 (1981) 2333–2339.
[8] T. Patel, J. Bruce, A. Merry, C. Bigge, M. Wormald, A.
Jaques, R. Parekh, Biochemistry, 32 (1993) 679–693.
[9] T. Patel, R.B. Parekh, Methods Enzymol., 230 (1993)
57–66.
[10] C. Cooper, N. Packer, J. Redmond, Glycoconjugate J.,
11 (1993) 163–167.
[11] J. Elguero, C. Marzin, A.R. Katritzky, P. Linda, The
Tautomerism of Heterocycles, Ad6. Het. Chem. Supple-
ment 1, Academic Press, New York, 1976.
[24] A.C. Satterthwait, W.P. Jencks, J. Am. Chem. Soc., 96
(1974) 7018–7031.
[12] E.P. Mazzola, S.A. Turujman, S.J. Bell, J.N. Barrows,
J.E. Bailey, E.G. Mouchahoir, S. Prescott, M.D. Archer,
M. Oliver, W.F. Reynolds, K.W. Nielsen, Tetrahedron,
52 (1996) 5691–5698.
[13] H. Zollinger, Color Chemistry: Synthesis, Properties and
Applications of Organic Dyes and Pigments, VCH, Wein-
heim, 1991.
[25] R.F. Pratt, T.C. Bruice, J. Am. Chem. Soc., 92 (1970)
5956–5964.
[26] A.R. Katritzky, F.W. Maine, Tetrahedron, 20 (1964)
299–314.
[27] B.S. Farniss, A.J. Hannaford, P.W.G. Smith, A.R.
Tatchell, Vogel’s Textbook of Practical Organic Chem-
istry, fifth ed., Longman Scientific and Technical, Har-
low, UK, 1989, pp. 946–949.
[14] Z. El Rassi, Reversed-phase and hydrophobic interac-
.