Biomacromolecules 2003, 4, 799-807 799

Pressure Cell Assisted Solubilization of Xyloglucans: Tamarind

Seed Polysaccharide and Detarium Gum
David R. Picout,*,† Simon B. Ross-Murphy,† Neil Errington,‡ and Stephen E. Harding‡
Biopolymers Group, Division of Life Sciences, King’s College London, Franklin-Wilkins Building,
150 Stamford Street, Waterloo, London SE1 9NN, U.K., and National Centre for Macromolecular
Hydrodynamics, School of Biosciences, University of Nottingham, Sutton Bonington,
Loughborough LE12 5RD, U.K.
Received December 17, 2002; Revised Manuscript Received March 14, 2003

To improve the solubilization of two water-soluble xyloglucans, tamarind seed polysaccharide and detarium
Downloaded via KUMHO PETROCHEMICAL LABORATORIES on November 24, 2023 at 06:44:49 (UTC).

gum, by reducing substantially molecular aggregation, a “pressure cell” heating method was used. Conditions
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

allowing solubilization and chain depolymerization were produced by varying appropriately the pressure,
time, and temperature applied. The various MW fractions of solubilized xyloglucans were characterized by
capillary viscometry and light scattering techniques in order to extract, with reliability, fundamental
macromolecular parameters. Mark-Houwink and Flory exponents were found to be 0.67 ( 0.04 and 0.51
( 0.06, respectively for both xyloglucan data combined, consistent with linear random coil behavior. A
detailed analysis of the data seems to suggest that tamarind gum solutions are slightly perturbed by the
effect of excluded volume, whereas detarium gum samples are close to the θ state. Chain flexibility parameters
such characteristic ratio, C∝, and persistence length, Lp, were calculated for tamarind and detarium using
the Burchard-Stockmayer-Fixman (BSF) geometric method. Lp values of 6-8 nm were estimated for
xyloglucans. The seemingly linear structure of tamarind and detarium, as suggested by the value of the
Mark-Houwink and Flory exponents obtained, follows from analysis of the data by the classical Zimm
method but not when employing the square root or Berry method which suggests a more branched chain
profile. This was the approach adopted in our previous work on the characterization of detarium samples.

Introduction Table 1. Results of Dionex HPAE Chromatography Showing

Quantitatively Oligosaccharide Fractions Converted from Detarium
and Tamarind Xyloglucan when Treated with
Xyloglucans are a major class of structural polysaccharides endo-1,4-β-Glucanase [From Wang et al. (1996)]5
found in the primary cell walls of higher plants. Cell growth Oligosaccharide ratioa Deduced monosaccharides
and enlargement are controlled by the looseness of a thin XXXG XLXG XXLG XLLG xylose galactose glucose
net of microfibrils made of cellulose. Xyloglucan cross-links tamarind 1.00 0.42 2.07 6.20 1.00 0.51 1.34
these cellulose microfibrils and provides the flexibility detarium 1.00 0.30 5.60 6.20 1.00 0.46 1.33
necessary for the microfibrils to slide.1,2 Xyloglucan polysac- a Standard nomenclature used.31-33
charide has the same skeleton, β-1,4 D-glucan, as cellulose
and therefore is not digested by human digestive enzymes in the paper industry as a replacement for starches and
(it acts as dietary fiber). The cellulose backbone of xylo- galactomannans.
glucan is partially substituted by R-(1 6)-linked xylose units. Detarium senegalense Gmelin, the seed flour of an African
Some of the xylose residues are β-D-galactosylated at O-2. leguminous plant traditionally used in Nigeria for its food
It is known that the distribution of side chain residues is thickening properties in soups and stews, was also found to
different in the xyloglucans from different species and it has contain a high proportion of xyloglucans. Its structural
even been shown to differ slightly, but significantly, between composition is very similar to tamarind xyloglucan,4 the main
two natural populations of the same species growing in difference in terms of simple composition being in the
different environments.3 proportion of galactose, relative to xylose and glucose. Table
1 from Wang and co-workers5 shows the structural differ-
Tamarind seed polysaccharide is the major polysaccharide
ences between these two xyloglucans.
constituent of seeds from the tree Tamarindus indica. The Like many other polysaccharides, tamarind xyloglucan and
seeds contain xyloglucans and are used extensively as food detarium xyloglucan are water-soluble, but their individual
thickeners, stabilizers and gelling agents in Japan. In the macromolecules tend not to fully hydrate and consequently
USA, its major industrial use has been as a wet end additive supramolecular aggregated species remain present even in
very dilute solutions. This is because the β(1 f 4), cellulose-
* To whom correspondence should be addressed. Phone: +44 (0) 20 like backbone promotes interchain interactions, and so the
7848 4246. Fax: +44(0) 20 7848 4082. E-mail: [email protected].
† King’s College London. polymers show a balance between hydrophobic and hydro-
‡ University of Nottingham. philic character. Light scattering techniques employed to
10.1021/bm0257659 CCC: $25.00 © 2003 American Chemical Society
Published on Web 04/11/2003
800 Biomacromolecules, Vol. 4, No. 3, 2003 Picout et al.

characterize such polymers in solution have not been very Table 2. Summary of Intrinsic Viscosity [η] and Static Light
successful in providing good reliable data because of the Scattering Results (Using the ALV System) for Tamarind Seed
Polysaccharide.
nonhomogeneity of these solutions at the molecular level.
[η] Mw Rg b A2
Certainly, poor reproducibility may be the reason few light -6
sample treatment (dL/g) (× 10 ) (nm) (× 104) c
a
scattering results have been published so far. The architecture
of tamarind seed xyloglucan, for example, has been inves- tamarind untreated 6.15 IR IR IR
100 °C, 30 min, 4 bar 6.1 0.83 136 21
tigated by light scattering, small-angle X-ray scattering
0.79 133 18
(SAXS) and synchrotron radiation.6 The data appeared to
show that tamarind in aqueous solution consisted of multi- 130 °C, 10 min, 4 bar 5.9 0.77 114 11
stranded aggregates, with a high degree of particle stiffness 0.75 115 12
but no reproducibility of the molar mass was achieved.
Various Mws values for tamarind, 115 0007 or 650 0008 130 °C, 10 min 6.05 0.73 103 6
(GPC) or 880 0009 (light scattering) or even 2 500 0006 were 0.71 99 7
reported in the literature. In this work as in our earlier
130 °C, 20 min 5.85 0.67 101 6
contributions, we define Mw as relative weight average
0.68 100 7
molecular mass, without units. 130 °C, 30 min 5.65 0.64 94 12
These variations are due to a strong tendency to self- 0.65 100 7
association of the polysaccharide, and it is possible that these 130 °C, 60 min, 4 bar 5.45 0.56 94 10
studies may have been affected by using non “molecular” 0.54 97 10
solutions in which the materials were not fully solubilized. 130 °C, 60 min 5.35 0.56 92 9
0.58 95 10
The problem of the characterization of the solution properties
of water soluble polymers is long-standing, and in order to 160 °C, 10 min, 4 bar 4.7 0.52 93 10
circumvent such problem, the pressure cell solubilization 0.56 98 12
method originally employed by Vorwerg and co-workers on 160 °C, 10 min 4.35 0.50 83 11
solutions of starch10,11 has been proven to be appropriate and 0.52 93 6
effective. This pressure/temperature solubilization approach 160 °C, 30 min 4.0 0.49 89 14
subsequently applied with success to nonstarch polysaccha- 0.45 89 11
rides such as detarium gum12 and a series of galactoman- a Mw corresponds to the zero concentration and zero angle extrapola-

nans13,14 is now employed in the present work to obtain tions of the Zimm plot. b Rg ) z average root-mean-square radius of
gyration. c A2 ) second virial coefficient in units of reciprocal concentration,
“molecular” solutions of tamarind seed polysaccharide. viz. mol mL/g2. IR ) irreproducible results.
Detarium xyloglucan will be recharacterized in this work
but using a more complete range of pressure cell treatments in the same way as for tamarind; the extraction procedure
and in order to be compared directly with tamarind using of the detarium flour from the seed samples purchased at a
the same experimental conditions and characterization tech- local market in Nsukka, Enugu State, Nigeria is described
niques. By achieving full solubilization and by reducing, in ref 5.
substantially, time-dependent aggregation phenomena in this Methods. Dilute solutions of tamarind and detarium gum
study, we are able to characterize the two polymers men- were prepared at 0.05 wt %, i.e., below C* defined as 1/[η],
tioned above. In this way, we hope to obtain reliable light by adding known weights of the freeze-dried samples to
scattering data, which tend to be scarce for such xyloglucans. deionized water at 50 °C, containing 0.02% sodium azide
Using appropriate models, we intend to calculate various as bactericide. The temperature was raised to 80 °C as dry
parameters such as the Mark-Houwink and Flory exponents, powder was added with stirring. The heating was stopped
the characteristic ratio, and the persistence length of both as soon as 80 °C was reached and the solutions were left
polysaccharides in order to obtain molecular information such covered overnight, with stirring, at room temperature to allow
as chain flexibility and shape. further hydration to occur.17 The solution at this stage was
used as a reference material (“untreated sample”). A total
Experimental Section of 30 mL of this solution was then added to the reaction
chamber of a pressure/heating cell (HEL Ltd, Barnet, Herts.
Materials. Purified samples from commercial grade U.K.). These solutions were then subjected, under stirring,
tamarind xyloglucan polysaccharide (cold water soluble) to a range of temperature and pressure conditions between
were supplied by Dainippon Pharmaceutical Co. Ltd., Tokyo, 70 and 160 °C and 0-4 bar added pressure. Added pressure
Japan. The samples were purified from the flour using an was applied using nitrogen gas while the reaction chamber
isolation procedure devised by Girhammar and Nair,15 and was at a temperature of 50 °C. These conditions were applied
modified by Rayment et al.16 to allow complete hydration for a range of times from 10 to 60 min. The general protocol
of the gum. The moisture content of the extracted polymer is described in more detail in ref 12.
was determined by incubation overnight in an oven at 103 The treated solutions were then analyzed using several
°C to a constant weight. Freeze-drying was carried out using techniques:
an Ehrist ALPHA I-5 Freeze-dryer (DAMON/IEC (U.K.) 1. Capillary Viscometry. The intrinsic viscosities [η] given
Ltd.), and samples were stored in a desiccator until used. in Tables 2 and 4 were determined using the Viscosity
The detarium xyloglucan used in this study has been purified Measuring Unit AVS 350 (Schott-Geräte, Hofheim, Ger-
Tamarind Seed Polysaccharide and Detarium Gum Biomacromolecules, Vol. 4, No. 3, 2003 801

Table 3. Summary of Intrinsic Viscosity [η] and Static Light ments of solution and solvent flow times. The viscometer
Scattering Results (SEC/MALLS) for Tamarind Seed was immersed in a precision water bath at (25.00 ( 0.05)°C
Polysaccharide.
(CT1450 water bath with DLK400 refrigeration unit, Schott-
[η] Mw Rg
Geräte, Germany) and 10 replicate measurements made on
sample treatment (dL/g) (× 10-6) (nm)
each solution. Results were analyzed as described above.
tamarind 70 °C, 10 min 6.7 0.80 115
2. Size Exclusion Chromatography Coupled to Multiangle
0.73 97
100 °C, 10 min 6.5 0.76 106
Laser Light Scattering (SEC/MALLS). The size exclusion
100 °C, 10 min, 4 bar 6.15 0.74 101 chromatography system used consisted of Jasco HPLC pump,
130 °C, 10 min 6.0 0.64 101 a guard column and TSK G5000 and G4000 columns. An
130 °C, 10 min, 4 bar 5.4 0.64 101 on-line degasser was used to remove gas from the eluent. A
0.68 104 flow rate of 0.8 mL/min for the mobile phase was used at
130 °C, 30 min, 4 bar 5.15 0.61 104 room temperature. A DAWN-DSP multiangle laser light
0.56 92 scattering detector and an Optilab 903 refractometer (Wyatt
Technologies, Santa Barbara, CA) were used for light
Table 4. Summary of Intrinsic Viscosity [η] and Static Light scattering intensity and concentration detection, respectively.
Scattering Results (Using the ALV System) for Detarium Gum.
The mobile phase was 0.02-weight % sodium azide in
[η] Mw Rgb A2 distilled de-ionized water. 100 µL samples of the tamarind
sample treatment (dL/g) (× 10-6)a (nm) (× 104)c solutions were injected into the size exclusion system after
detarium untreated 8.7 IR IR IR filtering through 0.45 µm filters (Whatman Ltd., Maidstone,
100 °C, 10 min, 4 bar 8.8 1.66 119 3 England). Repeat injections were made for each sample. Data
1.63 114 3
were captured and analyzed using the software package
130 °C, 10 min, 4 bar 8.3 1.25 120 3
ASTRA (v. 4.20). Data returned are the number, weight, and
1.20 110 5 z averages for molecular weight and root-mean-square radius
of gyration. These measurements were performed only for
130 °C, 10 min 8.0 1.25 103 -1 some of the tamarind samples (Table 3).
1.27 107 -1 3. Static and Dynamic Light Scattering. These measure-
ments were performed simultaneously, i.e., both in static and
130 °C, 20 min 7.8 1.15 109 -1
dynamic mode, on the same photons, at 20 °C with a fully
1.10 110 -2
130 °C, 30 min 7.25 1.08 130 -1
computerized ALV-5000 System comprising a compact
1.07 127 -1 goniometer system and a multi-τ real-time digital correlator
130 °C, 60 min, 4 bar 7.35 1.07 123 -1 (ALV-Laser Vertriebsgesellschaft m.b.H, Langen, Ger-
1.11 126 2 many). The angular range applied was from 30° to 150° in
160 °C, 10 min, 10 bar 6.5 0.8 102 -5 steps of 10°; the duration of single measurements was
0.82 99 -5 typically 10 s averaged over a minimum number of 3 runs
until a statistically significant result was obtained in static
160 °C, 10 min 5.5 0.75 84 -6
0.67 80 -8
mode (ALV/Static & Dynamic Fit and Plot Program used).
160 °C, 30 min 4.9 0.63 73 -5 Although dynamic data were calculated our approach was
0.64 75 -5 to concentrate on the quality of the static data. A He-Ne
a M corresponds to the zero concentration and zero angle extrapola-
laser (λ0 ) 632.8 nm) was the light source, and the scattering
w
tions of the Zimm plot. b Rg ) z average root-mean-square radius of of toluene was used as the primary standard. The refractive
gyration. c A2 ) second virial coefficient in units of reciprocal concentration, index increment, dn/dc, was chosen as 0.146 mL g-1. 18
viz. mol mL/g2. IR ) irreproducible results.
Solutions used for light scattering were solutions of 0.05%
many), connected to a ViscoDoser AVS 20 Piston Buret (for polymer (prepared as described previously and treated in the
automatic dilutions). This makes automated measurements pressure cell appropriately) and serial dilutions (0.04%,
of the flow-through times in a capillary viscometer (Ubbel- 0.03%, 0.02%, and 0.01%). These solutions were filtered 3
hode viscometer for dilution sequences). The viscometer was times directly into the cylindrical light scattering cuvettes
immersed in a precision water bath (transparent thermostat (Pyrex disposable culture tubes, Corning Incorporated, Corn-
CT 1650, Schott-Geräte, Hofheim, Germany) to maintain the ing, New York) (total volume ∼3 mL) using Acrodisc PF
temperature at 25 ( 0.05 °C. All polymer concentrations 0.8/0.2 µm syringe filters (Gelman Laboratory, Michigan).
ranged from 0.01 to 0.05% (w/v) so that the viscosity relative All solution preparation stages were carried out in a laminar
to that of the solvent (water) lay in the range 1.2 < ηr < airflow cabinet to minimize contamination with dust. Here
2.0. Results were analyzed using separate Huggins and both tamarind and detarium gum samples were used.
Kramer extrapolations (linear regression, 99% confidence
intervals) and the final result quoted in dL/g (1 dL/g ) 100 Results and Discussion
mL/g ) 0.1 m3/kg).
The intrinsic viscosities [η] in Table 3 were determined [η] Determination. The intrinsic viscosities [η] of the
using an alternative Viscosity Measuring Unit (AVS 400, tamarind samples after various pressure cell treatments (or
Schott-Geräte, Germany) and a capillary viscometer (Ubbe- none) are shown in Tables 2 and 3. Table 4 displays the
lohde type for dilution series) to make automatic measure- intrinsic viscosities obtained for detarium gum after being
802 Biomacromolecules, Vol. 4, No. 3, 2003 Picout et al.

subjected to similar treatments under combined temperature,

time, and pressure regimes. The data clearly show that the
intrinsic viscosity variation for both xyloglucans is minimal
with increasing temperature up to around 130 °C (with or
without excess pressure) unless the time of treatment is
increased considerably (>10 min). When temperature is
increased to 160 °C, [η] values decrease more sharply. This
phenomenon is not surprising and has been previously
reported for a series of galactomannans in our two previous
papers;13,14 the observed effect is consistent with thermal
degradation of the polymers chains. When excess pressure
is applied (4-10 bar), in most cases, [η] is higher than for
the samples treated at the same temperature and time
conditions but without the excess pressure. This is observed
especially at the highest temperatures used (160 °C), whereas
at lower temperatures, [η] does not vary much as mentioned
before. The addition of excess pressure in the treatment of
xyloglucans seems to “protect” the polysaccharide chains
from some of the degradation effect, as also observed for
galactomannans.13,14 Speculations have been made on the Figure 1. Mark-Houwink-Sakurada plot (log[η] vs log Mw) for
tamarind (O) and detarium (b) samples treated under various
effect of pressure on chain degradation,13 but the reasons temperature, time, and pressure conditions. The slope (the exponent)
still remain unclear. is 0.67 ( 0.04.
Mw and Rg Determination. Light scattering techniques obtained from the slope for the double logarithmic plot of
were used in order to collect Mw (weight-average molecular intrinsic viscosity against Mw are 0.76 ( 0.07 and 0.62 (
weight) and Rg (z average root-mean-square radius of 0.04 for tamarind and detarium, respectively. These values
gyration or more formally 〈S2〉Z1/2) data on all of the samples. are approximately within the bounds of the Mark-Houwink
Our facilities enabled us to use two types of light scattering equation for linear flexible macromolecules.
instruments. As mentioned above, the ALV-5000 system was To get a value of the MHS exponent for both xyloglucans
used to characterize the detarium samples and some of the combined, a statistical analysis of covariance was carried
tamarind samples. From the light scattering static mode Mw, out on the two sets of data using Minitab statistical software
Rg, and A2 (the second virial coefficient) were obtained from release 13.1 (Minitab Inc., Pennsylvania). It was found that
the appropriate Zimm plots by extrapolation of the experi- there were no significant differences in slopes between the
mental data to c ) 0 and q2 ) 0 using fitted polynomials two sets of data, but a significant difference in the two
(Table 2 and Table 4). SEC/MALLS was used only for the intercepts (the usual goodness of fit, P < 0.0005) was found.
Mw and Rg determination of a few treated tamarind solutions To obtain the best slope representative of all of the data, a
(Table 3). Light scattering measurements were also made set of two parallel lines was fitted to the data. The slope
on six untreated samples of tamarind and three untreated calculated was found to be 0.67 ( 0.04 for both xyloglucans
samples of detarium. Particle molar mass values obtained combined, giving an MHS exponent R well within the known
ranged from 0.48 × 106 to 1.14 × 106 for tamarind and from limits 0.5-0.8 found in the literature for a polymer chain in
1.5 × 106 to 2.4 × 106 for detarium gum, showing the the so-called flexible coil conformation. This linear confor-
difficulty in obtaining reproducible data and, therefore, their mation is further supported by Figure 2, which represents
reliability. Rg values obtained were also very different from the double log plot of Rg against Mw for the tamarind and
each other leading us to the observation that untreated detarium data. The Flory exponents obtained for tamarind
samples are practically impossible to characterize by light and detarium separately are 0.54 ( 0.07 and 0.49 ( 0.09,
scattering because of the strong presence of aggregates. respectively, and lie close to or within the expected range
Indeed, aggregates can cause distortions in the angular 0.5-0.6. The lines shown in Figures 1-2 are for guidance
dependence of scattered light and thus lead to errors in the only. They are representative of the best slope for all the
determination in the intercept on the scattering intensity (Kc/ data and do not represent the best fit as in fact not one line
Rθ) axis, leading to a false and nonreproducible estimation but a set of parallel lines was fitted into the data.
of Mw and Rg. As can be seen in the various tables (2-4), The statistical analysis of covariance of the two sets of
pressure cell treated samples, however, give good and data (no significant differences between the two slopes but
reproducible light scattering results. Solubilization has differences between the intercepts) enabled the calculation
improved with the pressure cell heating method, suggesting of the best-fitted slope for tamarind and detarium data
that the aggregates have been reduced or even been fully together. The Flory exponent was found to be 0.51 ( 0.06,
hydrated. being close to the Flory-theta chain, 0.5, value which seems
Mark-Houwink-Sakurada (MHS) and Flory Expo- to suggest that xyloglucans have a linear coil conformation
nents. Figure 1 shows the MHS plot (log [η] vs log Mw) but without excluded volume. As we discuss later, there is
constructed using all of the experimental data on the samples an inconsistency between R, and this exponent in terms of
of tamarind and detarium combined. The MHS exponents R the excluded or no excluded volume concept.
Tamarind Seed Polysaccharide and Detarium Gum Biomacromolecules, Vol. 4, No. 3, 2003 803

Figure 2. Values of radius of gyration, here Rg plotted as a function

of (weight average) Mw for the tamarind (O) and detarium (b) data.
The Flory exponent is 0.51 ( 0.06.

Structural Architecture of Xyloglucans. In the literature,

there is no strong evidence about the structural architecture
of xyloglucans, and it is not clear whether xyloglucan
macromolecules are linear or long chain branched. Detarium
xyloglucan solutions were studied by Wang and co-workers,5
and they found that semidilute solution characterization work
was very consistent with much of the published data for the
rheology of other polysaccharide solutions. The data sug-
gested that detarium gum was a well-behaved linear polymer
entanglement network system. However, light scattering
measurements carried out by the same group on dilute
solutions of detarium showed that the scattering profile was
Figure 3. (a) Kratky plot constructed using conventional Zimm plot
not consistent with that of a linear macromolecule, but instead data for tamarind samples treated under various conditions: (O) 130
strongly suggested a small degree of long chain branching. °C, 10 min, 4 bar; (0) 130 °C, 30 min; (4) 130 °C, 60 min; (f) 160
By contrast, the MHS and Flory exponents calculated in °C, 10 min. Theoretical curves calculated for models as in Burchard.20
The dashed line represents linear Gaussian chains and the dotted
our study tend to suggest that tamarind and detarium line is for homogeneous branching. (b) As Figure 3a for detarium
macromolecules are linear. Because the quality of the zero samples: (O) 130 °C, 10 min, 4 bar; (0) 130 °C, 30 min; (4) 130 °C,
concentration data obtained from the Zimm plots for both 60 min, 4 bar; (f) 160 °C, 10 min.
xyloglucans is highly acceptable, we adopt a well-known
procedure to examine the shape of the macromolecules using scattering factor, which reflects the angular dependence of
the so-called Kratky plot.19 It is important to note that this the scattered light, and q is the magnitude of the scattering
procedure is widely employed for both synthetic and vector ()4πλ/sin(θ/2)). The dimensionless parameter u,
biopolymers using small-angle X-ray light scattering (or, measures the intramolecular probe distance relative to the
even light scattering for high molecular weight polymers) incident light wavelength, and P(u) can be calculated for
but that it is not such an ideal approach to use for light different chain architectures. The two different curves
scattering of polysaccharides or other water soluble polymers. represented in Figure 3, parts a and b, reflect fits to different
This is because the overall data are usually not reliable for models for the chain behavior, the upper dashed line
the angular region above 60° or at most 90° because of a representing the theoretical profile for a flexible Gaussian
lack of scattering and because the Mws of the macromolecules chain and the lower dotted line illustrates the corresponding
are generally not high enough (W. Burchard, personal profile for a high degree of random homogeneous branch-
communication). However, this approach has been success- ing.19 Experimental data are shown with different symbols.
ful20 with very high Mw polysaccharides (>2 × 107) such as Rg values were obtained from light scattering measurements
branched amylopectin and glycogen or with very stiff using the Zimm plot. From the plots, it can be seen that the
macromolecules such as xanthan. experimental data, both for tamarind and detarium, follow
Parts a and b of Figure 3 show a Kratky plot for four the linear chain model rather well. However, if good Zimm
tamarind and four detarium samples, respectively, treated plot data are reanalyzed in terms of the Berry plot, and then
under various temperature, time, and pressure conditions. zero angle data replotted in terms of u2P(u) versus u, the
Here u ) qRg, P(u) t Rθ/Rθ)0 is the so-called particle same experimental data plotted in Figure 3, parts a and b,
804 Biomacromolecules, Vol. 4, No. 3, 2003 Picout et al.

Figure 5. BSF plot ([η]/Mw1/2 vs Mw1/2) for tamarind (O) and detarium
(b) data. Dotted lines indicate 99% confidence intervals.

from Figures 3 and 4. It seems therefore that depending upon

the method used for the determination of Mw and Rg (Zimm
or Berry plots) very different final conclusions can be drawn.
We wish to point out here that although the angular
distribution is linear in the Zimm plot, we also applied the
Berry plot method simply for comparison with the data
obtained in our initial paper on detarium.12 In this paper, we
deduced the presence of branching from the Berry plot.
However, comparison of the results from this paper with the
present set seems to suggest that the earlier samples had too
low an [η] value for the measured Mw. The branching may
therefore be actually due to insufficient deaggregation,
whereas the same is not seen here. To clarify whether
branching or insufficient deaggregation is achieved we would
probably need small-angle X-ray rather than light scattering.
Figure 4. (a) As Figure 3a, but with the Kratky plot constructed using
Berry plot analysis for tamarind samples: (O) 130 °C, 10 min, 4 bar; The reasons for these differences in the two methods are
(0) 130 °C, 30 min; (4) 130 °C, 60 min; (f) 160 °C, 10 min. Again unclear, but in any case, it can be said that if only the data
the dashed line and the dotted line represent linear Gaussian chains in the angular region below 60° were to be reliable, in the
and homogeneous branching models, respectively. (b) As Figure 4a Kratky approach, then all data would almost lie in the
for detarium samples: (O) 130 °C, 10 min, 4 bar; (0) 130 °C, 30
min; (4) 130 °C, 60 min, 4 bar; (f) 160 °C, 10 min. overlapping region of the two theoretical models. Therefore,
no strong conclusions can be drawn on the shape of these
now appear to follow the homogeneous branching model xyloglucan macromolecules from the Kratky approach. This
more closely, as shown in Figure 4, parts a and b. follows because the wavelength of the HeNe laser used in
Zimm and Berry Plot Analysis. The classical approach21 the present work is too high (λ0 ) 632.8 nm) for the size of
to processing light scattering data is to produce the so-called the macromolecules measured (∼80 nm for Rg) and hence
Zimm plot of Kc/Rθ versus a linear function of sin2(θ/2) there is almost no shape information.
(angular dependence) and c (concentration dependence). The Chain Flexibility of Xyloglucans. To determine the chain
linear function is chosen in such a way that the three- characteristic ratio, C∞, and the persistence length, Lp, of
dimensional surface of sin2(θ/2) and c is compressed into a flexible to semiflexible polymers, one common plot used is
plane. This approach is still widely employed, but for very that due to Burchard et al. (BSF plot)23 in which [η]/(Mw)1/2
high molecular weight, or less flexible polymers, the angular is plotted against (Mw)1/2 . The BSF method was applied to
dependent “upswing” means the usual simultaneous linear the tamarind and detarium data separately and to the
extrapolation to sin2(θ/2)f0 and c f 0 to give Kc)0/Rθ)0 ) combined data. Figure 5 shows the BSF plot for all tamarind
(1/Mw) is distorted. An alternative approach by Berry22 plots and detarium data combined. Statistical analysis was carried
(Kc/Rθ)1/2 vs sin2(θ/2) and c. This does linearize the upturn out on the two sets of xyloglucan data. Both sets showed
effect described above, but this is an explicit weighing of significant differences between the slopes and between the
the data. The ALV system can extrapolate data either in intercepts, but this was only due to a few tamarind data
Zimm or Berry coordinates, but results obtained, particularly (samples treated at 160 °C), which were in our opinion
for c f 0 are, of course, not the same. This actually distorts increasing the slope artificially. Because we have less
the apparent angular dependence of Kc)0/Rθ, in a way which confidence in these data sets, we consider that the best way
appears to alter the profile very significantly as we can see of analyzing the overall data is in fact to regard them as one
Tamarind Seed Polysaccharide and Detarium Gum Biomacromolecules, Vol. 4, No. 3, 2003 805

Table 5. Determination of MHS and Flory Exponents, Characteristic Ratios C∞ and Chain Persistence Lengths Lp for Detarium and
Tamarind Xyloglucans.
Burchard-Stockmayer-Fixman
(BSF) method
samples residue mra MHS exponent Flory exponent C∞ Lp (nm)
detarium 434 0.62 ( 0.04 (SE) 0.49 ( 0.09 (SE) 25 ( 4 (SE) 5-8
tamarind 445 0.76 ( 0.07 (SE) 0.54 ( 0.07 (SE) 19 ( 4 (SE) 4-6
both xyloglucans ∼440 0.67 ( 0.04 (SE) 0.51 ( 0.06 (SE) 26 ( 2 (SE) 6-8
a Calculated from the structure reported by Wang et al.5 SE ) standard error.

population. Using a simple regression analysis will obtain More seriously, perhaps, is the argument whether these
the best slope and intercept representing all of the data chains are indeed perturbed by excluded volume or whether
combined. all of the effects seen are due to intrinsic chain stiffness? In
From the BSF plots, the intercept Kθ (which corresponds practice, it is extremely difficult to separate these two effects
to the chain in the θ state, where there is no excluded for systems when the persistence length itself is compara-
volume) was obtained and from the MHS equation and the tively low (say <∼10 nm). By contrast for the ultra-stiff
Flory Fox equation, C∞ and then Lp were calculated for both polymers xanthan and schizophyllan polymers (Lp ∼ 120
polymers individually and also for all the data combined. and 180 nm, respectively26,28,29), it is reasonable to assume
Detailed methods are given, for example, in ref 13. there is almost no excluded volume. For these polymers, the
Table 5 summarizes the MHS and Flory exponents, chains adopt a helical structure, this in turn results in the
characteristic ratio, and chain persistence lengths calculated high Lp. Here the number of persistence lengths in the chain
for tamarind and detarium xyloglucans. The molar masses contour length L is typically small. We observe it is more
of the polymer residues used in the calculation of the usual to revert to Kuhn lengths, Lk, where Lk ) 2Lp, and
characteristic ratios, C∞, were calculated for tamarind and then define the number of Kuhn segments nk as L/Lk. For
detarium using the deduced monosaccharides ratio attributed xanthan and schizophyllan systems, nk is typically ∼2-6,
to these two polymers (see Table 1) and are shown in Table so by the usual criteria, i.e., nk g 6-10, these are not
5. For all xyloglucan data combined, the persistence length, Gaussian chains.
Lp, calculated was found to be 6-8 nm. This is reasonably If we consider an ideal polymer chain with n segments
close to, if slightly larger than, the Lp values found in the each of a length l, the contour length L is defined by L ) nl
literature for cellulose and derivatives,24 which is not too ) nkLk. The number of segments n represents the number
surprising because xyloglucans have a cellulosic backbone. of repeat units in the polymer chain and can be defined as n
Compared to the Lp values obtained for galactomannans,14,25 ) Mw/m, where Mw is the weight-average molecular weight
the xyloglucan Lp values calculated here are slightly higher of the polymer and m is the relative molar mass of a residue
suggesting that xyloglucan chains are stiffer (in relative (repeat unit). From all of the equations above, nk can be
terms) than galactomannan chains but may still be considered rewritten as nk ) Mwl/2mLp. When calculated for the
relatively flexible compared to very stiff macromolecules tamarind and detarium samples using23 l ) 0.54 nm (the
such as xanthan (Lp ∼ 120 nm26). O-O virtual bond length for a 1,4-diequatorially linked
Validity of Methods Employed. In this study as in our residue), we found 55 < nk < 100 for tamarind and 60 < nk
earlier papers, we have employed the Burchard-Stock- < 160 for detarium.
mayer-Fixman method to extrapolate to lower chain length, In the present case, for both tamarind and detarium, we
to estimate such parameters as the persistence length Lp and have a more than sufficient value of nk (nk far greater than
the chain characteristic ratio C∞. There is no doubt this 6-10) to consider the chains as coil polymers, so we have
approach could be criticized on a number of grounds. First, to use other criteria to examine the contribution or, otherwise,
the method itself is only one (albeit the simplest) of several of excluded volume effects. One of these is to consider the
methods and has, in the past, been censured on a number of values of either the Mark-Houwink exponent R, or the Flory
grounds. For example, it is well-known27 that, with decreas- exponent of Rg and Mw. For the first exponent, we have
ing chain length, chain hydrodynamic effects, the so-called values of ∼0.76 for tamarind and ∼0.62 for detarium. Both
draining term in the Flory-Fox equation, change, and for are significantly >0.5, the Flory or θ-state value, but less
rods and semiflexible chains, the effect can be marked. The than the excluded volume asymptotic limit of 0.8. For the
effect of increased draining causes a decrease in Flory’s corresponding Flory exponent, we have values of 0.54 and
viscosity constant Φ or draining parameter, which can, in 0.49, respectively (equivalent Flory limits 0.5 in the θ state
turn produce a distortion in the BFS plot. We also comment and 0.6 in the excluded volume limit). This would suggest
that direct measurements of Lp, for example from small-angle that tamarind gum solutions are indeed slightly perturbed
X-ray scattering measurements, are often somewhat larger by the effects of excluded volume, whereas detarium gum
than the BSF extrapolated values, but currently, such samples are close to the θ state. Qualitatively, at least, data
measurements have to be made at substantially greater for the second virial coefficient, A2, support this assertion,
concentrations. Nevertheless, the lower chain stiffness found corresponding values are typically ∼9 × 10-4 mol mL/g2
from BFS plots has been known historically. That said, the for tamarind, whereas the values for detarium are lower, and
difference is usually not so great, typically ∼30%, which is indeed some are slightly negative. However, all values are
about the absolute error of the present measurements. quite small, and the data are certainly not good enough, nor
806 Biomacromolecules, Vol. 4, No. 3, 2003 Picout et al.

is the range of Mw sufficiently wide, that we can investigate, because xylose has one less OH group than glucose and
for example, the Mw dependence of A2. The BSF plot for galactose, it can form fewer hydrogen bonds with water,
detarium, Figure 5, can be approximated to a straight line being therefore more hydrophobic (E. R. Morris, personal
parallel to the abscissa, even though that the data are communication). Besides, the main reason for a decrease in
scattered. For a clear θ system, with R and the Flory exponent the entropy of mixing and in solubility is that the hydro-
both equal to 0.5, the plot is, of course, constrained to give phobic interaction causes a stronger clustering of the water
a horizontal line. in the neighborhood of the hydrophobic groups.
For the tamarind data, there is clearly a finite slope,
although the range of Mws is even lower. This then poses a Conclusion
further question. Values of A2 are quite low, which suggests
there is little effect of excluded volume, but both R and Flory The method of solubilizing highly aggregated water-
exponents tend to support the opposite conclusion. Could soluble polysaccharides using the pressure cell approach has
this simply be the effect of intrinsic chain stiffness alone? been again applied successfully, this time with xyloglucan
As we hinted above, there can be no clear answer to this polysaccharides from tamarind and detarium legumes. The
without performing other experiments, perhaps using small- pressure-cell treated samples produced high quality repro-
angle X-ray scattering, SAXS, classical hydrodynamics or ducible light scattering data because of the absence of (or
from more detailed dynamic light scattering, giving access highly reduced) aggregates in solution. Fundamental macro-
to Burchard’s F parameter, the ratio of Rg to Rh, the Stokes molecular parameters relating to chain structure and chain
radius. For this sample, dynamic light scattering (DLS) flexibility were calculated. Mark-Houwink and Flory ex-
results were so scattered that we could come to no definitive ponents with values of 0.67 ( 0.04 and 0.51 ( 0.06,
conclusions on F. However, all of the Rh values (calculated respectively, were obtained for both xyloglucan data com-
from the Stokes-Einstein relationship) were significantly bined, and calculated Lps of 6-8 nm suggest that tamarind
lower than the Rg values, but not lower than ∼Rg/2. This, at and detarium behave as linear flexible (to semiflexible) coil
least, reduces the possibility of either rigid rod or homoge- polysaccharides, with the tamarind sample showing some
neous sphere architecture for xyloglucans. Future work ought perturbation from random coil behavior from excluded
to investigate the dynamic scattering behavior in more detail. volume effects.
However, such measurements can be extremely time- Interestingly, we failed to confirm the branched structure
consuming. reported for detarium gum in earlier work.12 Indeed evidence
This may, in fact, not be a real issue, as has been from this work suggests that such deductions are highly
demonstrated in recent measurements by Norisuye and co- model dependent, because analysis of the data by the classical
workers30 for a semiflexible microbial polysaccharide system. Zimm method suggested essentially linear chains, whereas
Here, analysis with and without excluded volume produced employing the square root or Berry method, also commonly
estimates of ∼9 ( 1 and ∼11 ( 1 nm, respectively, used in this area, suggests a more branched chain profile.
suggesting that, in this range of stiffness (and, presumably, We hope to have an opportunity to explore this apparent
for lower values such as we estimate here), the overall ambiguity in future. Despite this, the achievement of obtain-
difference is within experimental error. ing consistent and apparently reliable Mark-Houwink and
Flory parameters for a second class of “difficult” polysac-
Although some of the discussions here are, by their nature,
charides was highly gratifying.
speculative, it is interesting to consider what we can make
of the values obtained. For detarium on its own, we have Acknowledgment. We thank the Biotechnology and
Lps around 5-8 nm. This is slightly larger than for the Biological Sciences Research Council (BBSRC) for financing
galactomannans13,14 (∼4 nm). We also have some indication this project under Grant 29/D10446. We are also grateful to
from Figure 5 that the value for tamarind is slightly lower Dr. Peter Ellis, Division of Life Sciences, King’s College
than that for detarium. All this is, in fact, quite consistent London, for helpful discussions as well as Prof. E. R. Morris
with our earlier conclusions. For the relatively unsubstituted (Cork, Ireland), Dr. S. Radosta (Golm, Germany), Prof. T.
galactomannans, values reflect only the stiffness of the β-(1 Norisuye (Osaka, Japan), Prof. W. Burchard (Freiburg,
f 4) backbone and, consequently, lie quite close to those Germany), and Prof. K. Kajiwara (Tokyo, Japan) for valuable
for cellulose in nonaqueous solvents. We would then expect comments.
a slightly higher value for the more side-chain packed
xyloglucans and with a slightly greater value for detarium References and Notes
than for tamarind. This reflects the greater proportion of the (1) Hayashi, T. Ann. ReV. Plant Physiology Plant Molecular Biol. 1989,
XXLG oligosaccharide fraction in detarium compared to 40, 139.
(2) Whitney, S. E. C.; Brigham, J. E.; Darke, A. H.; Reid, J. S. G.; Gidley,
tamarind [G ) unsubstituted glucose residue; X ) xylose M. J. Plant J. 1995, 8, 491.
substituted glucose residue; L ) galactosylxylose substituted (3) Reid, J. S. G.; Edwards, M. E. Food polysaccharides and their
glucose residue; sequences always read toward the reducing applications; Marcel Dekker: New York, 1995; p 155.
(4) Reid, J. S. G. AdV. Botanical Res. Incorporating AdV. Plant Pathology
end of the molecule31-33].5 Because there is some evidence 1985, 11, 125.
that xylose units are more hydrophobic than galactose and (5) Wang, Q.; Ellis, P. R.; Ross-Murphy, S. B.; Reid, J. S. G. Carbohydr.
glucose units, our conclusions are consistent with expectation. Res. 1996, 284, 229.
(6) Lang, P.; Kajiwara, K. J. Biomater. Sci.-Polym. Ed. 1993, 4, 517.
Indeed, in xylose, the absence of the C(6) hydroxymethyl (7) Glicksman, M. Food hydrocolloids; CRC Press: Boca Raton, FL,
group causes a big reduction in conformational entropy; also 1986; p 191.
Tamarind Seed Polysaccharide and Detarium Gum Biomacromolecules, Vol. 4, No. 3, 2003 807

(8) Tamarind Seed Polysaccharide GLYCOID; Dainippon Pharmaceutical (22) Berry, G. C. J. Chem. Phys. 1966, 44, 4550.
Co., Ltd., 1989. (23) Morris, E. R.; Ross-Murphy, S. B. Tech. Carbohydr. Metab. 1981,
(9) Gidley, M. J.; Lillford, P. J.; Rowlands, D. W.; Lang, P.; Dentini, B310, 1-46.
M.; Crescenzi, V. Carbohydr. Res. 1991, 214, 299. (24) Ross-Murphy, S. B. In Cellulose Chemistry and its applications;
(10) Aberle, T.; Burchard, W.; Vorwerg, W.; Radosta, S. Starch-Starke Nevell, T. P., Zeronian, S. H., Eds.; Ellis Horwood Ltd.: Chichester,
1994, 46, 329. U.K., 1985; p 202.
(11) Vorwerg, W.; Radosta, S. Macromol. Symp. 1995, 99, 71. (25) Robinson, G.; Ross-Murphy, S. B.; Morris, E. R. Carbohydr. Res.
(12) Wang, Q.; Ellis, P. R.; Ross-Murphy, S. B.; Burchard, W. Carbohydr. 1982, 107, 17.
Polym. 1997, 33, 115. (26) Sato, T.; Norisuye, T.; Fujita, H. Polym. J. 1984, 16, 341.
(13) Picout, D. R.; Ross-Murphy, S. B.; Errington, N.; Harding, S. E. (27) Yamakawa, H. Modern Theory of Polymer Solutions; Harper and
Biomacromolecules 2001, 2, 1301. Row: New York, 1971.
(14) Picout, D. R.; Ross-Murphy, S. B.; Jumel, K.; Harding, S. E. (28) Norisuye, T.; Yanaki, T.; Fujita, H. J. Polym. Sci.: Part B: Polym.
Biomacromolecules 2002, 3, 761. Phys. 1980, 547.
(15) Girhammar, U.; Nair, B. M. Food Hydrocoll. 1992, 6, 285. (29) Yanaki, T.; Norisuye, T.; Fujita, H. Macromolecules 1980, 1462.
(16) Rayment, P.; Ross-Murphy, S. B.; Ellis, P. R. Carbohydr. Polym. (30) Norisuye, T.; Xu, X.; Zhang, L.; Nakamura, K. Dilute-Solution
1995, 28, 121. Behavior of Aeromonas Gum, a Heteropolysaccharide, Conference
(17) Ellis, P. R.; Morris, E. R. Diabetic Med. 1991, 8, 378. Proceeding, Euro-Japanese Workshop on Functional Polysaccharides,
(18) Theisen, A.; Johann, C.; Deacon, M. P.; Harding, S. E. RefractiVe Kyoto, 2002.
increment data-book for polymer and biomolecular scientists; Not- (31) Buckeridge, M. S.; Rocha, D. C.; Reid, J. S. G.; Dietrich, S. M. C.
tingham University Press: Nottingham, U.K., 2000. Physiol. Plant. 1992, 86, 145.
(19) Burchard, W. AdV. Polym. Sci. 1983, 48, 1. (32) Fanutti, C.; Gidley, M. J.; Reid, J. S. G. Planta 1991, 184, 137.
(20) Burchard, W. In Physical Techniques for the Study of Food (33) York, W. S.; Vanhalbeek, H.; Darvill, A. G.; Albersheim, P.
Biopolymers; Ross-Murphy, S. B., Ed.; Blackie Academic and Carbohydr. Res. 1990, 200, 9.
Professional: Glasgow, 1994; p 151.
(21) Zimm, B. H. J. Chem. Phys. 1948, 16, 1099. BM0257659