Distribution of methyl substituents over branched and linear regions in methylated starches

Article abstract of DOI:10.1016/S0008-6215(98)00238-9

Granular potato starch and amylopectin potato starch were methylated in aqueous suspension to molar substitutions (MS) up to 0.29. A method was developed to determine the MS of both branched and linear regions. After exhaustive digestion of the methylated starches with α-amylase, the highly branched fraction with a degree of polymerisation (DP) > 8 was separated from the linear oligomers by selective precipitation of the former in methanol. The substitution levels of branched and linear regions were determined. It was found that methylation takes place preferably at the branched regions of amylopectin and that amylose is higher substituted than linear regions of amylopectin. The distribution of methyl substituents in trimers and tetramers was determined by FABMS and compared to the outcome of a statistically random distribution. The results provided evidence for heterogeneous substituent distributions. Quantification of the degree of heterogeneity of the branched and linear regions showed a much larger deviation from random distribution in the linear regions. Copyright (C) 1998 Elsevier Science Ltd.

Full text of DOI:10.1016/S0008-6215(98)00238-9

Carbohydrate Research 312 (1998) 201±208
Distribution of methyl substituents over
branched and linear regions in methylated
starches
Yuri E.M. van der Burgta, Jack Bergsmab, Ido P. Bleekerb,
Paul J.H.C. Mijlandb, Anca van der Kerk-van Hoof c,
Johannis P. Kamerlinga,*, Johannes F.G. Vliegentharta
a Bijvoet Center, Department of Bio-Organic Chemistry, Utrecht University, PO Box 80.075, NL-3508
TB Utrecht, The Netherlands
b AVEBE Research & Development, AVEBE-weg 1, NL-9607 PT Foxhol, The Netherlands
c Bijvoet Center, Department of Mass Spectrometry, Utrecht University, Sorbonnelaan 16, 3584 CA
Utrecht, The Netherlands
Received 5 June 1998; accepted 27 August 1998
Abstract
Granular potato starch and amylopectin potato starch were methylated in aqueous suspension
to molar substitutions (MS) up to 0.29. A method was developed to determine the MS of both
branched and linear regions. After exhaustive digestion of the methylated starches with ꢀ-amylase,
the highly branched fraction with a degree of polymerisation (DP) > 8 was separated from the
linear oligomers by selective precipitation of the former in methanol. The substitution levels of
branched and linear regions were determined. It was found that methylation takes place preferably
at the branched regions of amylopectin and that amylose is higher substituted than linear regions
of amylopectin. The distribution of methyl substituents in trimers and tetramers was determined
by FABMS and compared to the outcome of a statistically random distribution. The results pro-
vided evidence for heterogeneous substituent distributions. Quanti®cation of the degree of het-
erogeneity of the branched and linear regions showed a much larger deviation from random
distribution in the linear regions. # 1998 Elsevier Science Ltd. All rights reserved
Keywords: Methylated starches; Amylopectin potato starch
1. Introduction
textile, adhesives, plastic, pharmaceutical and pet-
rochemical industries. It can be isolated from sev-
eral plants and is composed of the ꢀ-d-
glucopyranans amylose and amylopectin. Amylose
is essentially linear with (1!4) linkages, whereas
amylopectin is branched due to additional (1!6)
linkages. For many applications, starch is modi®ed
Starch is widely used as a renewable, raw mate-
rial in products, that are applied in the food, paper,
* Corresponding author. Fax: +31-30-254-0980; e-mail:
kame@boc.chem.puu.nl
0008-6215/98/$Ðsee front matter # 1998 Elsevier Science Ltd. All rights reserved
PII S0008-6215(98)00238-9

202
Y.E.M. van der Burgt et al./Carbohydrate Research 312 (1998) 201±208
to improve its functionality. This can be done
enzymatically, physically or chemically. In the
chemical derivatisation of starch the hydroxyl
groups are partially substituted [1]. Although such
processes have been carried out on an industrial
scale for over 50 years, the information on topo-
chemical aspects is scarce. Detailed information on
the distribution of substituents can contribute to
the understanding of relationships between mole-
cular structure and functional properties, thus
opening ways to `tailored derivatives' by selecting
derivatisation procedures. To acquire structural
data about starch derivatives, di€erent levels of
organisation in terms of substituent distribution
have to be studied, namely, branched and linear
regions, crystalline and amorphous domains, amy-
lose and amylopectin, and distribution in granules.
The aim of this study is the determination of
substitution level and substituent distribution in
branched and linear regions of methylated star-
ches. This aspect is of interest since functional
properties of starch derivatives such as pasting
temperature, gelatinisation and cooking character-
istics, solids±viscosity relationships and retro-
gradation tendencies are determined by the type of
substituent, the level of substitution and topo-
chemical aspects. For example, retrogradation of
gelatinised starches is amongst other aspects deter-
mined by interactions between linear starch chains.
In particular, the long-chain amylose molecules
play a predominant role in this process. Chemical
derivatisation is essential for many applications to
avoid retrogradation. In this context the chemical
derivatisation of linear starch regions of both
amylopectin and amylose is of importance.
degrees of methylation are listed in Table 1. The
molar substitution (MS) values were determined by
the Morgan±Zeisel method [3] and via a mono-
saccharide analysis [4]. The Morgan±Zeisel assay is
based on ether-cleavage with hydrogen iodide
resulting in the formation of methyl iodide, which
is led by a CO2-¯ow into a trap of a Ag(I)±metha-
nol solution; afterwards the remaining Ag(I) is
determined. For the monosaccharide analysis, the
samples were subjected to methanolysis and tri-
methylsilylation yielding seven (partially methyl-
ated) methyl glucoside monomers (methyl 2,3,6-tri-
O-methylglucoside was only found in trace
amounts and therefore neglected). The monomers
were characterised by GLC-mass spectrometry and
quanti®ed by GLC (Table 2). The quantitative
data were used for the calculation of the MS values
of all methylated starches. As is evident from
Table 1 both methods provide similar results.
Using the monomer compositions of P4±P30
and A4±A30 (Table 2), statistical distributions of
the methyl substituents over the eight theoretically
possible monomers were calculated, assuming con-
stant and independent reactivities of HO-2, HO-3
and HO-6 [5]. For all methylated starches more
non- and disubstituted monomers and less mono-
substituted monomers were observed than expec-
ted on the basis of a statistically random
distribution (de®ned as homogeneous [5]). This can
Table 1
Methylated potato starch and methylated amylopectin potato
starch
Samples
Code
MSa
(Morgan±Zeisel) (GLC)
MSa
Here, we report a method for the separation of
mixtures of branched and linear regions, obtained
from methylated potato starch and methylated
amylopectin potato starch after ꢀ-amylolysis [2].
Furthermore, the isolated branched and linear
regions were studied with regard to substitution
level and distribution of substituents.
Methylated potato starch
Methylated potato starch
Methylated potato starch
Methylated potato starch
Methylated potato starch
Methylated potato starch
Methylated amylopectin
potato starch
Methylated amylopectin
potato starch
Methylated amylopectin
potato starch
Methylated amylopectin
potato starch
Methylated amylopectin
potato starch
Methylated amylopectin
potato starch
P4
P6
P8
P10
P20
P30
A4
0.042
0.057
0.080
0.102
0.202
0.270
0.038
0.044
0.063
0.076
0.103
0.211
0.296
0.048
A6
A8
0.066
0.078
0.097
0.191
0.290
0.067
0.089
0.097
0.191
0.293
2. Results and discussion
A10
A20
A30
Monomer compositions of methylated starches.Ð
Methylated potato starch and methylated amylo-
pectin potato starch samples were prepared by
methylation of starch granules in an aqueous sus-
pension using dimethyl sulfate. The various starch
derivatives P4±P30 and A4±A30 with varying
aThe molar substitution (MS) is de®ned as mol of substituents/
mol of Glc monomer. The MS varies from 0 (native starch) to
3 (permethylated linear starch).

Y.E.M. van der Burgt et al./Carbohydrate Research 312 (1998) 201±208
203
Table 2
Monomer compositions of methylated potato starch (P4±P30)
and methylated amylopectin potato starch (A4±A30)
obtained DB values are higher than the actual
degree of branching, because 2-O-methylated
(1!6)-linked Glc monomers can not be taken into
account due to overlap of signals with those of
nonsubstituted (1!4)-linked Glc monomers.
Samples
Glca Glc2 Glc3 Glc6 Glc2,3 Glc2,6 Glc3,6
Me Me Me
Me2
Me2
Me2
P4
P6
P8
P10
P20
P30
A4
A6
A8
95.7 2.5 0.9 0.8
93.9 3.8 1.2 0.9
92.6 4.7 1.4 1.1
90.1 6.3 1.9 1.3
80.7 12.2 3.3 2.0
73.7 16.4 4.2 2.4
95.4 2.7 0.9 0.8
93.5 4.1 1.3 0.9
91.5 5.3 1.7 1.1
90.7 5.9 1.8 1.2
82.5 11.0 3.1 1.8
74.0 16.2 4.2 2.3
0
0.1
0.1
0.1
0.2
0.7
1.3
0.1
0.1
0.2
0.2
0.6
1.2
0
0
0
0
0.2
0.3
0
0
0
Characterisation of products obtained by a-
amylolysis of methylated starches.ÐThe methyl-
ated gelatinised starches P4±P30 and A4±A30 were
extensively digested with ꢀ-amylase from Bacillus
subtilis [2], yielding mixtures of (1!4)-ꢀ-d-gluco-
pyranans of di€erent sizes with varying degrees of
branching. The dextrose equivalents (DE) of the ꢀ-
amylase digests were determined by the method of
Lu€±Schoorl [10,11]. The underlying principle is
based on oxidation of all reducing sugars with
Cu(II). For each sample the average degree of
polymerisation (DP) was calculated from the DE
value [12]. With increasing MS of the starting
materials the average DP values of the digests after
ꢀ-amylolysis increase (Fig. 1). Apparently, methy-
lation of hydroxyl groups decreases the ability for
binding of ꢀ-amylase at subsites, thus limiting the
number of (1!4) linkages that can be cleaved.
For the investigation of the possible occurrence
of partially methylated reducing Glc units in the
0.1
0.1
0.2
0.9
1.7
0.1
0.1
0.2
0.2
0.8
1.8
A10
A20
A30
0
0.2
0.3
aAll monomers are characterised as trimethylsilylated methyl
glucopyranosides (ꢀ+ꢁ) and are given in molar percentages.
be explained by assuming that some areas in a
starch granule are better accessible during deriva-
tisation than others, resulting in a heterogeneous
substituent distribution along the polymer. How-
ever, other parameters could be responsible for this
di€erence between observed and expected mono-
mer compositions, such as vicinal substituent
e€ects and variable reactivities of hydroxyl groups
during methylation. Therefore, heterogeneity in the
distribution of substituents should be determined
in oligomer fragments.
As is clear from Table 2, in all methylated star-
ches the HO-2 group had by far the highest posi-
tionial degree of substitution. This preference, with
respect to the position in glucose, is independent of
the MS. It means that topochemical e€ects could
not be demonstrated at the level of monomers, as
was already reported [6].
Methodology for the analysis of branched and
linear regions.ÐAll steps in the separation of
branched and linear regions of methylated starches
and the characterisations thereof are depicted in
Scheme 1.
To determine the average degree of polymerisa-
tion (DP) and degree of branching (DB) of mix-
1
tures of ꢀ-d-glucopyranans H NMR spectroscopy
was used. The DP is calculated by dividing the
integral of the anomeric proton signals by the sum
of the integrals of the H-1ꢀ and H-1ꢁ signals of the
glucose unit at the reducing end [7±9]. The DB is
estimated by dividing the integral of the anomeric
proton signals by the integral of the signal of
(1!6)-linked Glc monomers. Note that thus
Scheme 1. Methodology for the analysis of branched and lin-
ear regions in P4±P30 and A4-A30.

204
Y.E.M. van der Burgt et al./Carbohydrate Research 312 (1998) 201±208
nor is it complete. As a result, the methanol pre-
cipitate is `contaminated' with low amounts of
small oligomers, whereas the supernatant is `con-
taminated' with substantial amounts of large oli-
gomers. To analyse this cross-contamination the
methanol precipitates and the methanol super-
natants of P4±P30 and A4±A30, respectively, were
subjected to HPLC. As a typical example in Fig. 2
the fractionation of P10 is shown.
In pilot experiments it was established that after
11.5 min only (1!4)-linked linear oligomers elute
(fraction 2; DP G8±G1). Up to 11.5 min (fraction
1; DP>G8) large and highly branched oligomers
elute. After ꢀ-amylolysis of methylated amylo-
pectin potato starch no long linear oligomers are
expected, because of the absence of the highly
polymeric amylose. Therefore, fraction 1 isolated
from A4±A30 corresponds only with branched
regions. However, in fraction 1 isolated from
methylated potato starch (P4±P30) some linear
glucopyranans, originating from incomplete diges-
tion of amylose, might occur. Di€erences found
between substitution levels of fraction 1 and frac-
Fig. 1. Average DP values of digests of the starting methyl-
ated starches after ꢀ-amylolysis versus MS. The MS values
were measured by Morgan±Zeisel [3] (Table 1) and the DP
values via Lu€±Schoorl [10,11]. Native potato starch and
amylopectin potato starch (MS=0) were used as references.
ꢀ-amylase digests of the methylated starches with
the highest MS values (P20, P30, A20 and A30),
generated oligosaccharides were converted into
their corresponding oligosaccharide alditols [13].
Then, the mixture of reduced oligomers obtained
was hydrolysed, the liberated (partially methy-
lated) glucose residues were converted into aldo-
nonitriles [4,14,15], and the mixtures of glucitols
and glucononitriles were analysed by GLC-mass
spectrometry after acetylation. It was found that
methyl substituents were present mainly in the
glucononitriles; only a trace of 6-O-methylglucitol
was found. It is therefore concluded that ꢀ-amylo-
lysis hardly occurs at methylated glucose sites; only
6-O-methylation is tolerated.
Separation of branched and linear starch re-
gions.ÐIn order to analyse the substitution level
of branched and linear starch regions, the glucan
mixtures obtained after ꢀ-amylolysis of methylated
starches were treated with methanol, yielding in
each case a precipitate and a supernatant (Scheme
1). 1H NMR spectroscopy of the precipitate
showed the presence of mainly large highly bran-
ched d-Glc oligomers, whereas the supernatant
contained only smaller (1!4)-linked d-Glc oligo-
mers, hardly containing (1!6) linkages. The pre-
sence of highly branched oligomers in the methanol
precipitate was further supported by 1H NMR
studies after incubation with the debranching
enzyme pullulanase [2].
Fig. 2. HPLC elution pro®le of methanol supernatant (upper)
and methanol precipitate (lower) isolated from P10. See
Experimental for the analysis parameters.
It should be noted, however, that precipitation of
the larger glucans in methanol is not fully selective,

Y.E.M. van der Burgt et al./Carbohydrate Research 312 (1998) 201±208
205
tion 2 in potato starch and amylopectin potato
starch, respectively, may therefore be attributed
mainly to di€erent substitution levels of amylose
and amylopectin.
glucopyranans (MSlr) can be calculated. In this
way, the MS values of the branched and the linear
regions of 12 di€erent methylated starches were
acquired and the results are summarised in Table 3.
In methylated potato starch (P4±P30) the MS of
the branched regions is 60±140% higher than the
MS of the linear regions. This preference for the
branched regions is probably related to a better
accessibility or higher reactivity of these regions.
An even larger di€erence in levels of substitution
between branched and linear regions is observed in
amylopectin potato starch (A4±A30; 140±260%).
The latter ®nding may be attributed to a high
reactivity of the amylose fraction in potato starch,
which is predominantly present in the amorphous
starch domains [16]. It implies that in amylopectin
molecules of methylated starch the level of sub-
stitution in branched regions is higher than in linear
regions. Previously, it was proposed that 2-hydro-
xypropyl and phosphate groups are located at
branched regions of chemically modi®ed starch
[17]. However, no quanti®cation could be made,
whereas with the method described here (Scheme 1)
the MS of branched and linear regions can be
determined separately.
Substitution level of branched and linear regions
from methylated starches.ÐThe MS values of the
methanol precipitates and the methanol super-
natants of P4±P30 and A4±A30, respectively, were
measured by monosaccharide analysis [4] (Table 3).
The molar ratios of the partially methylated mono-
mers in the glucans of the methanol precipitates and
the methanol supernatants of P4±P30 and A4±A30,
respectively, did not di€er from the ones in the cor-
responding starting materials (Table 2).
The cross-contamination as mentioned above
(Fig. 2) of branched and linear regions in the
methanol precipitates and the methanol super-
natants has been quanti®ed by HPLC for each
sample (Table 3), making use of the following
equations:
MSmp ˆ MSbrÁfraction1 ꢀin mp†
ꢀ1†
‡ MSlrÁfraction2 ꢀin mp†
Determination of distribution of substituents in
methylated highly branched d-glucopyranans and
methylated linear glucopyranans with FABMS.Ð
The distribution of substituents in oligomer frag-
ments represents the situation in the corresponding
polymer backbone [5,18]. In order to prepare sui-
table derivatives for FABMS analysis, the oligo-
mers in the methanol supernatants derived from
P20, P30, A20 and A30 were permethylated with
CD3I [19], whereas those in the methanol pre-
cipitates derived from P20, P30, A20 and A30 were
permethylated with CD3I after partial methano-
lysis. Although oligomers of speci®c DP are che-
mically identical, mass spectrometry can
distinguish between CH3 groups present in the
starch derivative and newly introduced CD3
groups. It can be anticipated that all per(deuter-
io)methylated trimers show the same behaviour in
terms of ionisation. Therefore, quanti®cation of all
FAB intensities in the trimer pseudo-molecular ion
region ([M+Na+]) is valid. The same holds for
per(deuterio)methylated tetramers. It should be
noted that quanti®cation is more accurate in the
trimer pseudo-molecular ion region than in the
tetramer one, because of higher signal to noise
ratios. Since resolution decreases using larger oli-
gomers, FABMS analysis was restricted to trimers
MSms ˆ MSbrÁfraction1 ꢀin ms†
ꢀ2†
‡ MSlr Áfraction2 ꢀin ms†
where mp=methanol precipitate, ms=methanol
supernatant, br=branched regions, lr=linear
regions.
Combining eqs (1) and (2) the MS of the bran-
ched glucopyranans (MSbr) and the MS of the linear
Table 3
Molar substitutions of branched and linear regions from
methylated potato starch and amylopectin potato starch
Sample % fraction % fraction MS MS MS MS
1 (in mp)
1 (in ms) (mp) (ms) (br) (lr)
P4
P6
P8
P10
P20
P30
A4
A6
A8
85.7
84.3
86.0
94.3
96.8
98.1
91.8
87.5
89.0
89.1
98.4
99.2
15.4
17.8
20.0
22.1
45.0
51.6
20.9
26.3
27.6
23.6
40.7
44.2
0.059 0.043 0.062 0.039
0.087 0.059 0.094 0.052
0.125 0.073 0.136 0.057
0.139 0.100 0.142 0.088
0.282 0.200 0.287 0.129
0.377 0.292 0.380 0.198
0.077 0.044 0.081 0.034
0.113 0.063 0.123 0.042
0.133 0.075 0.143 0.049
0.162 0.099 0.172 0.076
0.284 0.171 0.287 0.091
0.419 0.251 0.421 0.116
A10
A20
A30
mp=methanol precipitate, ms=methanol supernatant.
br=branched regions, lr=linear regions.

206
Y.E.M. van der Burgt et al./Carbohydrate Research 312 (1998) 201±208
and tetramers. The data obtained from FABMS
analysis were compared to the expected distribu-
tions according to binomial statistics [5]. In the
calculation methyl substitutents are distributed
randomly over equal and independent glucose
monomers. A typical analysis of trimers derived
from P20 is shown Fig. 3. For the methanol
supernatant (mainly linear regions) the observed
amount of nonsubstituted oligomers is lower than
expected on the basis of a random distribution,
whereas the amount of monosubstituted oligomers
is higher. This forms a clear indication for the het-
erogeneous substitution during methylation of
granular starch. For the methanol precipitate
(mainly branched regions) the observed contribu-
tion of non- and highly substituted oligomers is
slightly lower than calculated, whereas the amount
of mono- and trisubstituted oligomers is slightly
higher. The trends in branched regions are less
pronounced than those in linear regions.
Previously, evidence was obtained for the block-
wise substitution in granular methylated starches
by GPC analysis after ꢀ-amylolysis followed by
glucoamylolysis, however, without distinguishing
branched from linear regions [16]. Our results from
FABMS studies render quanti®cation of the degree
of heterogeneity in the substitution pattern of both
branched and linear regions possible (Table 4). The
degree of heterogeneity is expressed in the para-
meter Hn, where n corresponds with the DP of the
per(deuterio)methylated oligomers [5]. If the
observed and calculated mole fractions are iden-
tical (Fig. 3), the substituent distribution is homo-
geneous (Hn ˆ 0). Increasing di€erences between
observed and calculated mole fractions give rise to
an increase of the value of Hn. All heterogeneity
values in Table 4 are normalised by dividing by
their corresponding MS values as calculated from
FABMS data. In this way heterogeneities of samples
Fig. 3. Calculated (striped bars) and observed (white bars)
substituent distribution in trimers from the methanol super-
natant (mainly linear regions) and the methanol precipitate
(mainly branched regions) of P20. Note that the branched
regions contain more substituents than the linear regions.
with di€erent degrees of substitution can be
de®ned. It is evident from Table 4 that methyl
substituents in branched regions are almost dis-
tributed randomly (i.e. according to binomial sta-
tistics), whereas in the linear regions a higher
degree of heterogeneity is observed.
3. Conclusion
Methylation of granular starch in an aqueous
suspension using dimethyl sulfate takes place pre-
ferably at the branched regions of amylopectin.
However, the linear regions of amylopectin are
substituted more heterogeneously (i.e. not ran-
domly) than the branched regions. Amylose is
higher substituted than linear regions of amylo-
pectin as can be concluded from di€erences in
substitution level between methylated potato
starch and amylopectin potato starch.
4. Experimental
Table 4
Heterogeneity of substituent distribution in methylated potato
and amylopectin potato starches
Digestion of methylated starches.ÐMethylated
starch (14 g; prepared from potatoes or amylo-
pectin potatoes by methylation of starch granules
in an aqueous suspension using dimethyl sulfate,
see Table 1) was suspended in Millipore water
(300 mL), and CaCl2 (30 mg) was added. The sus-
pension was gelatinised by cooking for 15 min at
Sample
P20
P30
A20
A30
H3/MS(trimers) ms
H3/MS(trimers) mp
H4/MS(tetramers) ms
H4/MS(tetramers) mp
0.847 1.029 0.950 1.016
0.063 0.141 0.069 0.183
0.926 0.956 0.749 0.941
0.258 0.467 0.310 0.433
ꢀ
120 C. After cooling down, the pH was set at 6.2.
mp=methanol precipitate, ms=methanol supernatant.
Calculation of H3 and H4 is described in ref [5]. Homogeneous
distribution results in Hn ˆ 0; with increasing Hn hetero-
geneity of substituent distribution increases.
Then, the gelatinised starch was digested for 21 h
with 0.1 mass% BAN 240L (Novo Nordisk A/S) in
a shaking bath at 68.5 ꢀC and constant pH (6.2). In

Y.E.M. van der Burgt et al./Carbohydrate Research 312 (1998) 201±208
207
the ®rst 4 h the pH was re-adjusted several times.
ꢀ-Amylase-digested methylated starch was de-
branched with 2 mass% Optimax L300 (Genencor)
at pH 4.5 and 57 ꢀC for 24 h.
derivatives were analysed as trimethylsilylated
methyl glucosides with the same system, using the
same temperature program. FAB mass spectro-
metry was performed on a JEOL JMS-SX/SX
102A four-sector mass spectrometer, operated at
10 kV accelerating voltage, equipped with a JEOL
MS-FAB 10 D FAB gun operated at a 10 mA
emission current, producing a beam of 6 keV xenon
atoms. The per(deuterio)methylated oligosaccharide
samples were measured over a mass range of m/z 10±
1200 in a matrix of m-nitrobenzyl alcohol saturated
with NaI.
Methanol precipitation of highly branched
in
d-glucopyranans
a-amylolysis
digests.Ð
MeOH(1.5 L) was slowly added to an aqueous
solution of ꢀ-amylase-digested methylated starch
(300 mL). The suspension was stirred at room
temperature for 3 h. Then, the supernatant was
separated from the sticky precipitate, and both
fractions were concentrated.
Analytical procedures.ÐMonosaccharide analy-
sis including GLC-mass spectrometry was carried
out as described [4]; for the quanti®cation by GLC
empirical molar response factors were applied
[20,21].
Prior to NMR analysis samples were exchanged
once in D2O (99.9 atom% D, Isotec), then lyophi-
lised and dissolved in 99.96 atom% D2O (Isotec). 1D
1H NMR spectra were recorded on a Bruker AC-
300 spectrometer (Department of Organic Chem-
istry, Utrecht University), equipped with a 5 mm
broad-band probe, at a probe temperature of 27 ꢀC.
Chemical shifts are expressed in ppm down®eld
from the signal for external Me4Si, but were
actually measured by reference to external acetone
(ꢂ 2.225).
For the detection of (partially methylated) glu-
cose units at the reducing end, ®rst partially
methylated oligoglucopyranan samples, obtained
after ꢀ-amylolysis of the di€erent methylated
starches, were reduced with NaBD4 in 0.5 M
NH4OH [13]. The reduced partially methylated
oligoglucopyranan samples were hydrolysed with
ꢀ
2 M CF3CO2H (2 h, 100 C), and the generated
saccharide mixtures were converted into glucono-
nitrile derivatives by treatment with NH2OH in N-
methylimidazole (30 min, 80 ꢀC) [14,15]. The
mixtures of (partially methylated) glucitols and
glucononitriles were acetylated with Ac2O (10 min,
Acknowledgements
This study was supported by the PBTS Research
Program with ®nancial aid from the Ministry of
Economic A€airs and the Integral Structure Plan
for the Northern Netherlands from the Dutch
Development Company.
ꢀ
20 C), and analysed by GLC-mass spectrometry
after working-up.
Partially methylated oligoglucopyranans from
methanol precipitates (after partial methanolysis;
ꢀ
1.5 h, 60 C, methanolic 0.5 M HCl) and methanol
References
supernatants were permethylated with CD3I as
described previously (Me2SO/NaOH) [19], and
analysed by FABMS.
[1] O.B. Wurzburg (Ed.), Modi®ed Starches: Proper-
ties and Uses, CRC Press, Boca Raton FL, 1986.
[2] J.F. Robyt, Starch, 2nd ed., Academic Press, New
York, 1984, Chapter 4.
GLC analyses were performed on a WCOT CP-
column
SIL
5CB
fused-silica
capillary
[3] R.J. Smith, Characterization and Analysis of Star-
ches, in H.J. Roberts, R.L. Whistler, and
E.F. Paschall (Eds.), Starch Chemistry and Tech-
nology, Academic Press, New York, 1965, pp 610±
635 and references cited therein.
[4] J.P. Kamerling and J.F.G. Vliegenthart, Mass
Spectrometry, in A.M. Lawson (Ed.), Clinical Bio-
chemistryÐPrinciples, Methods, Applications, Vol.
1, Walter de Gruyter, Berlin, 1989, pp 176±263.
[5] P.W. Arisz, H.J.J. Kauw, and J.J. Boon, Carbo-
hydr. Res., 271 (1995) 1±14.
(25 mÂ0.32 mm) using a temperature program of
ꢀ
ꢀ
110±230 C at 4 C/min. HPLC analyses were per-
formed on an Aminex HPX-42A column
(300Â7.8 mm) at 85 ꢀC, eluted with water at a ¯ow
rate of 0.4 mL/min and RI detection. GLC±mass
spectrometry of partially methylated glucitol acet-
ates and glucononitrile acetates was carried out on
an MD800/8060 system (Fisons instruments;
electron energy, 70 eV), equipped with a DB-1
fused-silica capillary column (30 mÂ0.32 mm,
J&W Scienti®c), using a temperature program of
110±230 ꢀC at 4 ꢀC/min. (O-Methylated) Glc
[6] P.A.M. Steeneken and E. Smith, Carbohydr. Res.,
209 (1991) 239±249.

208
Y.E.M. van der Burgt et al./Carbohydrate Research 312 (1998) 201±208
[7] M.J. Gidley, Carbohydr. Res., 139 (1985) 85±93.
[8] J.D. Blake, M.L. Clarke, and J. Littlemore, Car-
bohydr. Res., 138 (1985) 161±167.
[9] G.S. Nilsson, K. Bergquist, U. Nilsson, and
L. Gorton, Starch, 48 (1996) 352±357.
[10] N. Schoorl, Chemisch weekblad, 5 (1929) 130.
[11] L.M. Marchal, J. Jonkers, and J. Tramper, Starch,
48 (1996) 220±224.
[15] A.D. French, E.E. Smith, and W.J. Whelan, Car-
bohydr. Res., 22 (1972) 123±134.
[16] P.A.M. Steeneken and A.J.J. Woortman, Carbo-
hydr. Res., 258 (1994) 207±221.
[17] L.F. Hood and C. Mercier, Carbohydr. Res., 61
(1978) 53±66.
[18] P. Mischnick and G. Kuhn, Carbohydr. Res., 290
(1996) 199±207.
[12] J.J.M. Swinkels, Karakterisering van zetmeelhy-
drolysaten Glucose informatieband, 8 (1985) 71±93.
[13] J. Lonngren and S. Svensson, Adv. Carbohydr.
Chem. Biochem., 29 (1974) 41±106.
[19] I. Ciucanu and F. Kerek, Carbohydr. Res., 131
(1984) 209±217.
[20] J.T. Scanlon and D.E. Willis, J. Chromatogr. Sci.,
23 (1985) 333±339.
[14] F.R. Seymour, E.C.M. Chen, and S.H. Bishop,
Carbohydr. Res., 73 (1979) 19±45.
[21] A.D. Jorgensen, K.C. Picel, and V.C. Stamoudis,
Anal. Chem., 62 (1990) 683±689.