NMR Tube Degradation Method for Sugar Analysis of Glycosides

Article abstract of DOI:10.1021/acs.jnatprod.6b00180

The sugar subunits of natural glycosides can be conveniently determined by acid hydrolysis and ¹H NMR spectroscopy without isolation or derivatization. The chemical shifts, coupling constants, and integral ratios of the anomeric signals allow e

Full text of DOI:10.1021/acs.jnatprod.6b00180

Note
pubs.acs.org/jnp
NMR Tube Degradation Method for Sugar Analysis of Glycosides
Jose-Luis Giner,* Ju Feng, and David J. Kiemle
́
Department of Chemistry, State University of New York-ESF, Syracuse, New York 13210, United States
S
* Supporting Information
ABSTRACT: The sugar subunits of natural glycosides can be
1
conveniently determined by acid hydrolysis and H NMR
spectroscopy without isolation or derivatization. The chemical
shifts, coupling constants, and integral ratios of the anomeric
signals allow each monosaccharide to be identified and its
molar ratio to other monosaccharides to be quantified. The
NMR data for the anomeric signals of 28 monosaccharides and
three disaccharides are reported. Application of the method is
demonstrated with the flavonoid glycoside naringin (1), the
aminoglycoside antibiotics kanamycin (2) and tobramycin (3),
and the saponin digitonin (4).
any natural products occur as glycosides. The combina-
torial diversity generated by glycosylation is impressive,
M
often overwhelming. Dozens of types of sugars can potentially
be incorporated, each of which typically has multiple sites for
the attachment of other sugars. NMR spectroscopy using 2D
methods can be used to determine the structures of natural
glycosides, but overlapping signals often make it challenging.1
A common strategy used to delimit the component sugars is to
hydrolyze a sample and to determine which sugars are present.
This is generally accomplished by GC after TMS derivatization
or LC, both of which require authentic standards.2 An NMR
method was developed that allows sugar determination with-
out isolation or derivatization. Furthermore, once the NMR
measurements have been made and the data recorded, authentic
standards are no longer needed. This method has been used
extensively in the analysis of polysaccharides from wood
products3,4 and has been applied to the analysis of glycolipids.5
The method is general and can be advantageously applied
to many glycosidic natural products. Herein we describe the
method and demonstrate its application to some representative
examples: the flavonoid glycoside naringin (1), the aminoglyco-
side antibiotics kanamycin (2) and tobramycin (3), and the
steroidal saponin digitonin (4) (Figure 1).
RESULTS AND DISCUSSION
■
1
Observation of the anomeric region by H NMR in D2O is
usually difficult due to the overlapping water signal. However,
this signal can be shifted by changing the pH. By measuring the
spectra in 2 M sulfuric acid, the water peak is shifted downfield
from ca. 4.7 ppm to ca. 6 ppm, where it does not interfere with the
observation and integration of the anomeric protons (Figure S6,
Supporting Information). The use of deuterated sulfuric acid
helps to minimize the water signal, and presaturation can be used
to greatly reduce its intensity, which improves the dynamic range
of the signals of interest. In practice, when presaturation is used, it
is often unnecessary to use deuterated solvents beyond what is
Figure 1. Natural product glycosides.
Received: March 1, 2016
© XXXX American Chemical Society and
American Society of Pharmacognosy
A
DOI: 10.1021/acs.jnatprod.6b00180
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Note
Table 1. 1H NMR Data for the Anomeric Signals of Reference Sugars
δ (J in Hz; % total anomeric signals)
Hexoses
glucose (5)
galactose (6)
mannose (7)
allose (8)
5.26, d (3.6; 38%)
5.30, d (3.8; 33%)
5.21, s (62%)
4.68, d (7.9; 62%)
4.63, d (7.9; 67%)
4.94, s (38%)
5.17, d (3.4; 18%)
4.90, d (8.3; 82%)
a
b
allose (8)
5.54, d (2.2; 14%)
5.17, d (3.4; 16%)
5.15, s (59%)
4.90, d (8.3; 70%)
5.02, d (8.3; 41%)
altrose (9)
a
b
altrose (9)
5.41, d (2.2; 69%)
5.15, s (18%)
5.02, d (8.3; 13%)
5.21, d (3.7; 14%)
gulose (10)
4.90, d (8.4; 87%)
a
b
gulose (10)
5.48, d (2.4; 62%)
5.21, d (3.7; 5%)
5.10, s (52%)
4.90, d (8.4; 33%)
5.02, d (5.8; 48%)
idose (11)
a
b
idose (11)
5.38, s (94%)
5.10, s (3%)
5.02, d (5.8; 3%)
4.85, s (45%)
talose (12)
5.30, s (55%)
Uronic Acids
glucuronic acid (13)
glucuronic acid (13)
5.32, d (3.7; 49%)
4.7, d (8.0; 51%)
a
c
c
5.59, v br s (6%)
5.51, s (27%)
5.32, d (3.7; 32%)
5.59, v br s (20%)
5.37, d (3.6; 44%)
4.75, d (8.0; 35%)
5.51, s (80%)
glucuronolactone (14)
galacturonic acid (15)
Pentoses
4.68, d (8.0; 56%)
arabinose (16)
ribose (17)
5.27, d (3.5; 34%)
4.96, d (6.2; 69%)
5.22, d (3.4; 35%)
5.03, d (4.6; 69%)
4.56, d (7.8; 66%)
4.88, s (31%)
xylose (18)
4.61, d (7.9; 61%)
4.90, d (1.3; 31%)
lyxose (19)
Deoxy Sugars
2-deoxyribose (20)
2-deoxyglucose (21)
fucose (22)
5.32, v br s (47%)
5.40, br s (47%)
5.23, d (3.7; 29%)
4.84, v br s (53%)
4.96, br d (9.3; 53%)
4.60, d (8.0; 71%)
d
d
1.26, d (6.5; 71%)
1.22, d (6.5; 29%)
rhamnose (23)
quinovose (24)
5.14, s (57%)
4.91, s (43%)
d
d
1.31, d (6.2; 43%)
1.29, d (6.3; 57%)
5.20, d (3.7; 31%)
4.68, d (8.0; 69%)
d
d
1.30, d (6.2; 69%)
1.27, d (6.2; 31%)
Amino Sugars
glucosamine (25)
N-acetylglucosamine (26)
galactosamine (27)
kanosamine (28)
6-glucosamine (29)
nebrosamine (30)
Disaccharides
5.49, d (3.4; 63%)
5.25, d (3.5; 55%)
5.51, d (3.6; 59%)
5.30, d (3.6; 50%)
5.30, d (3.7; 41%)
5.40, d (3.4; 61%)
4.99, d (8.5; 37%)
4.80, d (8.4; 45%)
4.93, d (8.5; 41%)
4.76, d (7.7; 50%)
4.72, d (8.0; 59%)
4.96, d (8.4; 39%)
cellobiose (31)
5.26, d (3.7; 20%)
5.40, d (3.8; 50%)
5.26, d (3.6; 20%)
4.70, d (7.9; 30%)
5.26, d (3.8; 20%)
4.71, d (8.0; 30%)
4.53, d (8.0; 50%)
4.69, d (8.0; 30%)
4.48, d (7.9; 50%)
maltose (32)
lactose (33)
Methylated Sugars
2,3,4,6-tetramethylglucose (34)
2,3,6-trimethylglucose (35)
5.40, d (3.5; 60%)
5.43, d (3.5; 60%)
4.64, d (8.0; 40%)
4.68, d (7.9; 40%)
a
b
c
Sample was not heated. (All other samples were heated at 100 °C for 1 h.) Signal from 1,6-anhydro sugar. Signal from glucuronolactone (14).
d
Methyl signal.
required to maintain the lock signal. Since hydrolysis of glyco-
sides to monosaccharides generally involves heating the sample
in the presence of acid, the hydrolysis and analysis can be
combined in the same NMR tube.
Every monosaccharide has a pair of characteristic chemical shifts
and coupling constants for the anomeric (H-1) signals (Table 1).
These signals are not concentration dependent (Figure S5,
Supporting Information). Furthermore, they are found in a
specific α/β ratio. Thus, glucose (5) can be identified by its sig-
nals at 5.26 ppm (J = 3.6 Hz, α-pyranoside, 38%) and 4.68 ppm
(J = 7.9 Hz, β-pyranoside, 62%) (Figure 2d). However, for the
purpose of identification, it is not necessary to know which
B
DOI: 10.1021/acs.jnatprod.6b00180
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Note
decomposed in the presence of 2 M D2SO4 at 37 °C for 4 days.
This would allow its analysis by NMR in mixtures generated by
the mild hydrolysis conditions typically used for 2-deoxysugars. It
should be noted that the 2-deoxysugars 20 and 21 showed broad
peaks, especially 2-deoxyribose (20), which may limit the utility
of this method. The 6-deoxysugars fucose (22), rhamnose (23),
and quinovose (24) did not pose any problems. Besides the
anomeric signals, their methyl signals are also well separated from
other signals and could also be diagnostic (Figure S2, Supporting
Information).
As expected, the acetyl group of N-acetylglucosamine (26)
was lost when heated for 1 h at 100 °C. However, at lower
temperatures, 26 was relatively stable, undergoing only 43%
deacetylation after 4.5 h at 65 °C.
As a demonstration of this method, a 250 μg sample of the
flavonoid glycoside naringin (1) was hydrolyzed in the NMR
tube. The spectrum (Figure 1a) was acquired over 30 min at
600 MHz using presaturation of the residual solvent peak. The
sum of the integrals of the anomeric signals of glucose (5) is
equal to the sum of the anomeric signals of rhamnose (23),
indicating a 1:1 ratio. It was noticed that at early times during the
hydrolysis of 1 the signals of the terminal α-rhamnose subunit
predominated (Figure 1b). This is thought to be due to the more
rapid hydrolysis of α-linkages and terminal glycosides.8 The
observation that one sugar is liberated before another might be of
use for structure determination, but the extent of its generality
remains to be demonstrated. Because of the insolubility of
the flavonoid group, neither naringin, its monodeglycosylated
product, nor its aglycone was observed in 2 M D2SO4. This
simplifies the analysis of the free sugars, but if required, we
found that DMSO-d6 can be added to solubilize the nonpolar
substances without interfering with either the hydrolysis or the
NMR analysis (Figure S9, Supporting Information). However,
the shifts and ratios of the anomeric signals do change in the
presence of DMSO, and care must be taken in their assignment
(Figure S10, Supporting Information).9
The data for three rare amino sugars were obtained through
hydrolysis of the aminoglycoside antibiotics kanamycin (2)
and tobramycin (3). Both of these antibiotics incorporate a
kanosamine unit (3-amino-3-deoxyglucose, 28) in their struc-
tures. The signals corresponding to 28 could therefore be
assigned through their occurrence in the hydrolysis products
of both 2 and 3 (Figure 2). The remaining signals from the
hydrolysis of 2 and 3 therefore arise from 6-glucosamine (29)
and nebrosamine (30), respectively. The hydrolyses of both ami-
noglycosides were slow, but that of the nebrosamine glycosidic
Figure 2. Anomeric protons of products from naringin (1) hydrolysis.
(a) Hydrolysis of 250 μg of 1, 1 h, 95 °C. (b) Hydrolysis of 12 mg of 1,
15 min, 95 °C. (c) Rhamnose (23). (d) Glucose (5).
signals belong to the α- and β-anomers, respectively, as long
as the pattern matches. Furthermore, knowing whether the sig-
nals originate from pyranose or furanose forms is unnecessary.
Although the equilibrium between the furanoside and pyranoside
forms could potentially double the number of anomeric signals,6
in general, monosaccharides show only two predominant forms.
It should be noted that sugars lacking anomeric hydrogens (e.g.,
ketoses) cannot be identified by this method.
After measuring the reference spectra, the sugars were heated
to 100 °C in 2 M D2SO4/D2O for 1 h to check their stability to
conditions that would be sufficient to hydrolyze most glycosides.
Perhaps somewhat surprisingly, the monosaccharides remain
unchanged after heating in acid even for prolonged periods in
almost all cases (Figure S4, Supporting Information). Changes
were seen, however, for some sugars. The rare or unnatural
hexoses allose (8), altrose (9), gulose (10), and idose (11) were
found to equilibrate with their 1,6-anhydro forms upon heating.7
These represented 14%, 69%, 62%, and 94% of the equilibrium
mixtures, respectively, and their signals would have to be taken
into account in a quantitative analysis. The more common
glucuronic acid (13) also displayed new signals upon acid treat-
ment. These signals were due to the formation of glucurono-
lactone (14), which was formed as 33% of the equilibrium
mixture under these conditions. Again, this needs to be taken into
account, but does not interfere with the analysis. Galacturonic
acid (15) does not react under these conditions.
Although most sugars tolerated heating with acid, some sugars
were unstable. The pentoses, arabinose (16), ribose (17), xylose
(18), and lyxose (19), all displayed slow degradation to furfural
(9.50, 7,94, 7.61, and 6.79 ppm). These were relatively minor
amounts, representing 2%, 9%, 5%, and 4% of the mixtures,
respectively, after heating for 1 h at 100 °C, and could be
minimized by milder hydrolysis conditions.
Complete decomposition was observed for 2-deoxysugars
(20 and 21). However, their glycosides are much more easily
hydrolyzed than those of other sugars and do not require such
harsh conditions. Stability tests showed that 21 was only 50%
Figure 3. Anomeric protons of the hydrolysis products of kanamycin
(2) and tobramycin (3). (a) Hydrolysis of 7.8 mg of 2, 17 h, 95 °C. (b)
Hydrolysis of 7.3 mg of 3, 113 h, 95 °C.
C
DOI: 10.1021/acs.jnatprod.6b00180
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Note
linkage in tobramycin (3) was extremely slow due to the 2-amino
group. After more than 4 days at 95 °C, only 10% of this linkage
had been hydrolyzed. This, however, was sufficient for the mea-
surement of the anomeric signals of 30.
The hydrolysis of the steroidal glycoside digitonin (4) gave
a 2:2:1 mixture of glucose (5), galactose (6), and xylose (18)
(Figure 4a). At an early point in the hydrolysis, it was noticed that
Figure 5. Anomeric protons of the hydrolysis products (1 h, 100 °C) of
(a) fully O-methylated glucose and (b) fully O-methylated maltose.
Bruker Avance III HD with an Ascend 800 MHz magnet and a 5 mm
TCI cryoprobe. All spectra were measured using 500 μL of 2 M D2SO4
as the solvent. Because of the high ionic strength of the solvent, it is not
always possible to tune and match these samples in all NMR probes. To
1
obtain quantitative data, the H NMR spectra were acquired with a
relaxation delay set to 10 s, 30° pulse angle, 2.7 s acquisition time, and
spectral widths of 16 ppm. Since the T1 of the anomeric signals of four
representative sugars [galactose (6), glucose (5), mannose (7), and
xylose (18)] were measured to be 1.3−2.0 s (Figure S8, Supporting
Information), integration of the signals is quantitative with a 30° pulse
according to the Ernst equation.10 For discussion of quantitative NMR
see Pauli et al.11 Chemical shift calibration was with 0.05% sodium
2,2-dimethyl-2-silapentane-5-sulfonate (DSS, (CH3)3Si = 0.015 ppm).
Spectra were acquired with 16K data points and processed using zero
filling, exponential multiplication, and line broadening of 0.3 Hz. To
prepare 2 M D2SO4/D2O, 20 g of sulfuric acid-d2 (99.5% D, Cambridge
Isotope Laboratories) was diluted to a final volume of 100 mL with
deuterium oxide (99.9% D, Cambridge Isotope Laboratories). Glyco-
sides and sugars were obtained commercially. Typically, ca. 10 mg
samples were analyzed. Heating of the reference sugars was carried out in
a boiling water bath; otherwise, heating blocks were used. The anomeric
ratios given in Table 1 were obtained by manual integration of the spectra
shown in the Supporting Information. Methylation of glucose and
maltose was carried out as described by Ciucanu and Costello.12
Figure 4. Anomeric protons of the hydrolysis products of digitonin (4).
(a) Hydrolysis of 0.8 mg of 4, 11 h, 100 °C. (b) Hydrolysis of 11.1 mg of
4, 12 min, 90 °C (800 MHz).
the terminal xylose (18) group was liberated faster than the other
sugars (Figure 4b). As was the case for naringin (1), neither
digitonin (4), its aglycone, nor its intermediate hydrolysis
products are soluble in water and are therefore not observed. In
this case, unidentified anomeric signals were also detected early
in the hydrolysis; these later vanished. The observed signals
suggest a disaccharide fragment, having signals consistent with
the α- and β-anomers of a terminal moiety, as well as a β-linkage,
although additional anomeric signals indicative of a trisaccharide
or higher oligomer may have been obscured. This observation
should contain information about the linkage pattern and
suggests that a library of disaccharide fragments might be useful
in determining the sequence of sugars. The signals of three
common disaccharides (31−33) are included in Table 1.
Methylation analysis has been a useful technique in classical
carbohydrate analysis for determining linkage patterns.2,10 Tests
were carried out to ascertain whether this technique could be
combined with the NMR method. Exhaustive methylation
of glucose (5) was followed by hydrolysis in the NMR tube with
2MD2SO4/D2Oat100°C for 1 h to give 2,3,4,6-tetramethylglucose
(34). Repeating the process with maltose (32) yielded a 1:1 mix-
ture of 34 and 2,3,6-trimethylglucose (35), the anomeric signals
of which are well separated (Figure 5). This indicates the
presence of a 1,4-linkage. Besides the anomeric signals, the
methoxy signals of the different methylated sugars were also
different, and, because they are strong signals found in a region
of the spectrum without much overlap, these may also be
diagnostically useful for sugar analysis by NMR.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the ACS
Publications website at DOI: 10.1021/acs.jnatprod.6b00180.
■
S
1H NMR spectra of the anomeric regions of sugars 5−35
and the methyl regions of sugars 22−34, a spectrum of
furfural, spectra of 5−7 heated for varying lengths of time
and different concentrations of 5−7, spectra of 5 at
different acid concentrations, T1 measurements of 5−7
and 18, hydrolysis of naringin (1) in the presence of
DMSO, and spectra of glucose (5) in the presence of
varying concentrations of DMSO (PDF)
AUTHOR INFORMATION
Corresponding Author
*Tel (J.-L. Giner): +1 (315) 470-6895. E-mail: jlginer@syr.edu.
■
Some drawbacks of the NMR tube hydrolysis method are that
it is destructive and that NMR is less sensitive than GCMS.
However, adequate amounts for analysis by this method are often
available. In these cases, the experimental simplicity makes this a
useful and convenient method.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We acknowledge the upgrade of the 600 MHz NMR spectro-
meter under NSF grant CHE-1048516. The acquisition of the
800 MHz NMR spectrometer was made possible by NIH grant
S10 OD012254.
EXPERIMENTAL SECTION
NMR spectra were obtained at 30 °C using a Bruker 600 Avance III
instrument using either a 5 mm BBFO SmartProbe or a 5 mm BBFO
Prodigy Cold probe, except for Figure 4b, which was obtained on a
■
D
DOI: 10.1021/acs.jnatprod.6b00180
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Note
REFERENCES
■
(1) Lowicki, D.; Czarny, A.; Mlynarski, J. Nucl. Magn. Reson. 2013, 42,
383−419.
(2) Chaplin, M. F.; Kennedy, J. F. Carbohydrate Analysis: A Practical
Approach; IRL Press: New York, 1994.
(3) Kiemle, D. J.; Stipanovic, A. J.; Mayo, K. E. In Hemicelluloses: Science
and Technology, ACS Symp. Series; Gatenholm, P., Tenkanen, M., Eds.;
ACS: Washington, D.C., 2004; Vol. 864, Chapter 9; pp 122−139.
(4) Bose, S. K.; Barber, V. A.; Alves, E. F.; Kiemle, D. J.; Stipanovic, A.
J.; Francis, R. C. Carbohydr. Polym. 2009, 78, 396−401.
(5) Sallans, L.; Giner, J.-L.; Kiemle, D. J.; Custer, J. E.; Kaneshiro, E. S.
Biochim. Biophys. Acta, Mol. Cell Biol. Lipids 2013, 1831, 1239−1249.
(6) Angyal, S. J. Angew. Chem., Int. Ed. Engl. 1969, 8, 157−166.
(7) Angyal, S. J.; Dawes, K. Aust. J. Chem. 1968, 21, 2747−2760.
(8) BeMiller, J. N. Adv. Carbohyd. Chem. 1967, 22, 25−108.
(9) Angyal, S. J. Carbohydr. Res. 1994, 263, 1−11.
(10) Ernst, R. R.; Anderson, W. A. Rev. Sci. Instrum. 1966, 37, 93−102.
(11) Pauli, G. F.; Godecke, T.; Jaki, B. U.; Lankin, D. C. J. Nat. Prod.
̈
2012, 75, 834−851.
(12) Ciucanu, I.; Costello, C. E. J. Am. Chem. Soc. 2003, 125, 16213−
16219.
E
DOI: 10.1021/acs.jnatprod.6b00180
J. Nat. Prod. XXXX, XXX, XXX−XXX