Biomacromolecules 2004, 5, 2384-2391
2384
Solution Properties of the Xyloglucan Polymer from Afzelia africana† Yilong Ren, David R. Picout, Peter R. Ellis, and Simon B. Ross-Murphy* Department of Life Sciences, King’s College London, Franklin-Wilkins Building, 150 Stamford Street, London SE1 9NN, United Kingdom Received June 3, 2004; Revised Manuscript Received July 29, 2004
In this paper we describe the solution properties of a new xyloglucan polysaccharide extracted from the African legume Afzelia africana Se. Pers. The polysaccharide is of high weight-average molecular weight (Mw), but application of the “pressure cell” method enabled a range of Mw fractions to be prepared. Results from the light scattering/intrinsic viscosity measurements on these fractions suggest that like other xyloglucans from tamarind and detarium it occurs in solution as a polymeric coil, with a small amount of excluded volume. Measurement of dilute and semidilute solution rheology suggests that, like these polymers, and the related galactomannan series, it forms viscous solutions at higher concentrations via entanglements. Introduction As described in more detail in a separate paper,1 which describes a full fine-structural and biochemical characterization of the nonstarch polysaccharide derived from it, Afzelia africana Se. Pers. is a legume belonging to the subdivision of Caesalpinioideae. This is one of the three classes of Leguminosae.2-4 The species is widespread in West African countries including Ghana, Gambia, the Ivory Coast, and Nigeria, but like many such leguminous crop seeds in this part of the world, it remains underexploited and underutilized. In Africa the polysaccharide-rich flour, a light brown powder extracted from the cotyledon of the seed, is traditionally used as a food condiment for thickening soups and stews.2 Although little is known about the botanical characteristics and chemical composition of the seed, an early study showed that it is rich in nonstarch polysaccharides.2 In the previously mentioned paper1 we were able to employ both biochemical and chemical techniques to establish the primary structure of the main nonstarch polysaccharide found in afzelia seed flour. This demonstrated that it was a xyloglucan, similar to, but differing in detailed substituent composition from, tamarind and detarium polysaccharides. These details are summarized in Table 1. Such polymers have a range of potential applications including in the food and pharmaceuticals sectors, especially as a nutritional supplement in the treatment of diabetes.5 Since they are plant products, they are also a valuable renewable resource. * To whom correspondence should be addressed. Phone: +44 (0) 20 7848 4081. Thermal fax: +44 (0) 20 7848 4082. Plain paper fax: +44 (0) 20 7848 4500. E-mail:
[email protected]. † Abbreviations: In this paper an unambiguous shorthand nomenclature for xyloglucan oligosaccharides is used. Each (1f4)-β-linked D-glucosyl residue in the backbone is given a one-letter code according to its substituents. Thus, G ) an unsubstituted glucose residue, X ) a xylosesubstituted glucose residue, and L ) a galactosylxylose-substituted glucose residue; sequences always read toward the reducing end of the molecule. Chart 1 illustrates these different substituent units in a representative tetrasaccharide unit.
Table 1. Oligosaccharide Fractions Produced from Tamarind, Detarium, and Afzelia Polysaccharides oligosaccharide ratios
deduced monosaccharides
polymer
XXXG
XLXG
XXLG
XLLG
Xyl
Gal
Glu
afzelia detarium6 tamarind8
1.00 1.00 1.00
0.21 0.30 0.37
2.80 5.60 2.06
1.68 6.20 3.03
1.00 1.00 1.00
0.45 0.46 0.51
1.33 1.33 1.34
A number of workers, including ourselves, have investigated detarium6-8 and tamarind8-12 nonstarch polysaccharides in some detail over the past few years, particularly from the viewpoint of rheological13,14 and macromolecular (i.e., light scattering/intrinsic viscosity) approaches. By contrast there appear to be only very limited studies on the afzelia polysaccharide (hereafter referred to simply as afzelia). The problem of the macromolecular characterization of water-soluble polymers, especially polysaccharides, is longstanding. To circumvent this problem, the so-called “pressure cell” method15 has proven to be both appropriate and effective.7,8,16-19 Indeed, in our recent papers we suggest that reliable macromolecular datasfor example, for chain persistence length and chain characteristic ratioscould be obtained by producing Mw “fractions” simply by “tuning” the excess pressure and the heating time and temperature. More details are given in these other papers. By using this approach and “classical” light scattering techniques, it was possible therefore to obtain the basic macromolecular parameters, that is to say, the Mark-Houwink and Flory exponents. These exponents provide a fuller understanding of the effect of primary structure on chain flexibility. Semidilute solution rheology is related to chain flexibility, because less flexible chain systems produce viscous solutions at low concentrations. There are well-documented approaches to solution rheology, including combining steady and dynamic oscillatory shear measurements, and correlating the
10.1021/bm049678n CCC: $27.50 © 2004 American Chemical Society Published on Web 09/11/2004
Solution Properties of a Xyloglucan Polymer
Biomacromolecules, Vol. 5, No. 6, 2004 2385
Chart 1. Structure of a Typical Xyloglucan Tetrasaccharide Repeat Unit, in This Case XLXG
zero-shear rate viscosity with the “volume-filling” term C[η].20,21 These approaches make up the second part of this paper. Materials and Methods Sample Preparation. The pure polysaccharide used in this study was extracted from afzelia seeds. The method is described in detail in another paper,1 but briefly involved soaking the seeds at room temperature for 42 h, removing the seed coat, cutting the cotyledon into small cubes, and drying the cubes at room temperature. This material is then ground into a fine light brown powder, further dried at room temperature for 24 h, and stored in a -20 °C freezer. The pure polysaccharide is then purified from the flour using a modified version of the isolation procedure of Girhammar and Nair.22 Polysaccharide solutions (0.2-6%, w/w) for viscoelastic measurements were prepared on the basis of dry weight. The sample was dispersed in deionized water for 30 min at 80 °C on a hot plate, and then mixed overnight using a magnetic stirrer at room temperature. The xyloglucan content of the purified material was determined by quantification of the monosaccharides using a gas-liquid chromatographic method with alditol acetate derivatized samples after the method of Englyst and co-workers,23 but modified to allow more complete hydration of the water-soluble xyloglucan.22 For intrinsic viscosity and light scattering measurements, 30 mL of a 0.1% concentration of this solution was added to the reaction chamber of a pressure/heating cell (HEL Ltd., Barnet, Herts, U.K.). This solution was then subjected, with stirring, to a range of temperature and pressure conditions between 100 and 160 °C, and with an overpressure of 3 bar of N2. This added pressure was applied while the reaction chamber was at a temperature of 50 °C. Such conditions were applied for a series of hold times at the highest temperature, varying from 10 to 120 min. (Although some of these conditions may appear arbitrary, they draw upon
our recent experience on the application of this technique to a range of both galactomannans18,19 and other xyloglucans.8) Intrinsic Viscosity [η]. The intrinsic viscosity had been determined previously using a manual method,1 but the measurement was repeated here and for all other pressure cell treated afzelia samples using an automated system consisting of a viscosity measuring unit, AVS 350 (SchottGera¨te, Hofheim, Germany), connected to a ViscoDoser AVS 20 piston buret (for automatic dilutions). This makes automated measurements of the flow times in an Ubbelohdetype capillary viscometer immersed in a precision water bath (CT 1650, Schott-Gera¨te) to maintain the temperature at 25 ( 0.05 °C. All polymer concentrations ranged from 0.01% to 0.1% (w/v) so that the viscosity relative to that of the solvent (water) lay in the range 1.2 < ηr < 2.0. Results were analyzed using separate Huggins and Kramer extrapolations (linear regression, 99% confidence intervals), and the final result is quoted in dL g-1 (1 dL g-1 ) 100 mL g-1 ) 0.1 m3 kg-1). Light Scattering Measurements. These measurements were performed at 20 °C with a fully computerized ALV5000 system (ALV-Laser Vertriebsgesellschaft mbH, Langen, Germany). The angular range applied was from 30° to 150° in steps of 10°; the duration of single measurements was typically 10 s averaged over a minimum number of three runs until a statistically significant result was obtained in static mode (the ALV/Static & Dynamic Fit and Plot Program was used). A He-Ne laser (λ0 ) 632.8 nm) was the light source, and the scattering of toluene was used as the primary standard. The refractive index increment, dn/dc, was chosen as 0.146 mL g-1. Solutions used for light scattering were solutions of 0.1% polymer (prepared as described previously and treated in the pressure cell appropriately) and serial dilutions (0.08%, 0.06%, 0.05%, 0.04%, 0.03%, 0.02%, and 0.01%). These solutions were filtered three times directly into the cylindrical light scattering cuvettes (Pyrex disposable culture tubes, Corning Inc., Corning, NY) (total volume ∼3 mL) using
2386
Biomacromolecules, Vol. 5, No. 6, 2004
Acrodisc PF 0.8/0.2 µm syringe filters (Gelman Laboratory, Michigan). All solution preparation stages were carried out in a laminar airflow cabinet to minimize contamination with dust. Steady Shear Flow Measurements. Steady shear experiments on solutions of different concentrations (0.2-6.0%, w/w) were performed on a Rheometrics fluids spectrometer, RFSII (TA Instruments, Crawley, U.K.), with a cone and plate configuration (diameter 50 mm, cone angle 0.02 rad). The transducer system has dual torque ranges of 0.002-10 and 0.02-100 g cm, which correspond to a stress range of 0.006-300 Pa with this geometry. Most of the steady shear measurements were carried out at shear rates of 0.05-1000 s-1 with a reduction in rate at the higher concentrations. All the measurements were conducted at 25 °C. Oscillatory Shear Measurements. Oscillatory experiments were conducted on the RFSII using the same geometry as that used in the steady shear measurements described above. A sinusoidal strain wave with frequency ω was applied to the lower plate, and the response of the sample exerted on the upper plate was detected by the transducer system. Strain sweep experiments were carried out to determine the linear viscoelastic range of the polysaccharide solutions studied. The complex shear modulus (G*) and dynamic viscosity (η*) were measured at a strain range of 0.1-100% in 5% increments at a frequency of 1 or 10 rad‚s-1. In frequency sweep measurements the strain was selected to be 40% for the 1.5%, 3%, and 6% w/w solutions, within the linear viscoelastic limits established as above, although some extra measurements were made both at higher strains (for lower concentration samples) and vice versa. For frequency sweep experiments, the frequency of the applied strain wave was varied typically from 0.1 to 150 rad‚s-1 with 3-5 points per decade. The viscoelastic parameters, namely, the dynamic viscosity (η*), storage modulus (G′), and loss modulus (G′′), were extracted. Results Intrinsic Viscosity [η]. The result for the untreated afzelia sample, i.e., without pressure cell treatment, was 5.35 ( 0.10 dL g-1 () 535 mL g-1) using the automated viscometer system, and agreed very well with the previous manual result, 5.3 ( 0.5 dL g-1. The intrinsic viscosities [η] of the samples after various pressure cell treatments (or none) are shown in Table 2. Few data were measured below 130 °C, since previous experience with galactomannans and the other xyloglucans had shown that the change in intrinsic viscosity variation is minimal up to around 130 °C (with or without excess pressure) unless the time of treatment is increased considerably (>10 min). As mentioned above, pressure cell measurements were made at 130 and 160 °C, with a fixed overpressure of 3 bar of N2, and the main variable was the time held at this maximum temperature. It should be noted however that, using our current apparatus, the minimum overall temperature cyclesheat to say 130 °C, hold, and coolsis typically around 90 min, and ca. 120 min for 160 °C experiments. The results in Table 2 follow the
Ren et al. Table 2. Summary of Intrinsic Viscosity [η] and Static Light Scattering Results for Afzelia Polysaccharide treatment
[η]/dL g-1
untreated 100 °C, 10 min, 3 bar
5.35 5.25
130 °C, 10 min, 3 bar
4.5
130 °C, 30 min, 3 bar
4.3
130 °C, 60 min, 3 bar
3.8
130 °C, 90 min, 3 bar
3.55
130 °C, 120 min, 3 bar
2.85
160 °C, 10 min, 3 bar 160 °C, 30 min, 3 bar
2.55 1.95
160 °C, 60 min, 3 bar
1.75
Mwa (×10-6) IRc 0.84 0.81 0.73 0.72 0.68 0.65 0.56 0.51 0.48 0.45 0.29 0.26 0.24 0.21 0.23 0.22
Rgb/nm IR 79 80 96 90 114 109 74 72 98 86 64 70 65 38 45 33
a M corresponds to the zero concentration and zero angle extrapolaw tions of the Zimm plot. b Rg ) z-average root-mean-square radius of gyration. c IR ) irreproducible results as a result of the presence of supramolecular aggregates.24
expected pattern, in that [η] values decrease with both the maximum temperature and the “hold time” at this temperature. For example, heating at 130 °C for 120 min produces a reduction in [η] slightly less than that of heating for 10 min at 160 °C. Light Scattering. Integrated light scattering was used to determine Mw (weight-average molecular weight) and Rg (z-average root-mean-square radius of gyration (more formally 〈S〉z1/2)) data on all the samples. Mw, Rg, and A2 (the second virial coefficient) were obtained from the appropriate Zimm plots by extrapolation of the experimental data to c ) 0 and q2 ) 0 using fitted polynomials. A2 values were small, suggesting that the system may not be far from Θ conditions, but the values are not tabulated here. Supramolecular aggregates cause distortions in the angular dependence of scattered light and lead to errors in the determination of the intercept on the scattering intensity (Kc/Rθ) axis.24 Such errors result in very scattered values of Mw and Rg for the untreated samples, and no reliable results could be obtained. This is indicated by the code “IR” in Table 2. On the other hand, and as indicated in Table 2, pressure cell treated samples give much more reproducible light scattering results for Mw. This is completely consistent with our previous experiences.8 The radii of gyration, Rg, show similar trends, although these data are naturally more scattered. For the most highly degraded samples, those heated for >∼30 min at 160 °C, the Rg values are significantly below the usual, albeit conservative, measurement cutoff limit of λ0/10, i.e., ∼25% greater Lp values; i.e., they are >∼25% more stiff than the galactomannans. In terms of the precision with which Lp can usually be estimated, here and elsewhere, this is a small but significant difference. Nevertheless, as we pointed out previously,8 there is still an inconsistency in that the second virial coefficient A2 is around zero, and yet the MHS exponent is significantly >0.5. This raises the question as to whether these chains are indeed perturbed by excluded volume or whether all the effects seen are due to some intrinsic chain stiffness. In this earlier paper we pointed out how difficult it is to separate these two effects for systems when the persistence length itself is comparatively low (say,
