ARTICLE pubs.acs.org/Biomac
SANS Measurements of Semiflexible Xyloglucan Polysaccharide Chains in Water Reveal Their Self-Avoiding Statistics Franc-ois Muller,† Sabine Manet,‡ Bruno Jean,‡ Gerard Chambat,‡ Franc-ois Boue,† Laurent Heux,‡ and Fabrice Cousin*,† † ‡
Laboratoire Leon Brillouin, CEA Saclay, 91191 Gif sur Yvette Cedex, France CERMAV, CNRS UPR 5301, BP 53 38041 Grenoble Cedex, France ABSTRACT: We explored the behavior and the characteristics of xyloglucan polysaccharide chains extracted from tamarind seeds in aqueous media. The initial solubilization is achieved by using a 0.01 M NaOH solution. The absence of compact aggregates in the solution and the average molecular mass of the individual chains were unambiguously demonstrated by size exclusion chromatography with multi-angle light scattering detection. The composition and the stability of the solution were quantitatively checked over weeks by using liquid state nuclear magnetic resonance with DMSO as internal standard. The conformational characteristics of the chains were measured using nondestructive small-angle neutron scattering (SANS). The unambiguous determination of the Flory exponent (ν = 0.588) by SANS enabled us to directly prove that xyloglucan chains in water behave like semiflexible worm-like chains with excluded volume statistics (good solvent), contrary to most of the neutral water-soluble polymer chains that rather exhibit Gaussian statistics (θ-solvent). In addition to the Flory exponent, the persistence length lp and the cross section of the chains were also determined by SANS with utmost precision, with values of 80 and of 7 Å, respectively, which provides a complete description of the conformational characteristics of XG chains at all relevant length scales.
’ INTRODUCTION Xyloglucans (XG) are polysaccharides that belong to the class of hemicelluloses and play a major structural role in the primary cell walls of dicotyledons1 and thick storage walls of some seeds.2 XG chains consist of a glucopyranose backbone ((1,4)-β-linked D-glucopyranosyl), similar to cellulose, with substitution groups making the dissolution of high molecular mass possible. For nonsolenaceous dicotyledons, the side chains are present statistically on three of every four glucopyranosyl groups and the sequence is repeated all along the chain, independent of the side chain number and nature. For XG from tamarind seeds, according to the nomenclature,3 most of the repeating units are XLLG (around of 50% molar), while 9, 28, and 13% molar are found for XLXG, XXLG, and XXXG, respectively,46 where G is (1,4)-β-D-Glc, X is R-(1,6)-D-Xyl, and L is β-DGal-(1,2)-R-D-Xyl. The high affinity for cellulose of this sophisticated structure has been shown to slightly depend on structural details;511 in particular, the interactions have been suggested to be mediated by the presence of the unsubstituted glucopyranose sequences in both substances through hydrogen bonding and van der Waals forces.12 Accordingly, this interaction has previously been exploited in many systems of practical interest involving bacterial cellulose or natural fibers (textiles, pulp, and paper)13,14 and on the nanoscale involving cellulose nanocrystals, where XG are used as dispersing agents15 or to build nanostructured surfaces.16,17 In addition, XG on their own have numerous industrial uses, for example, as a food thickener, because they are commercially available as a powder product in large quantities with a reasonable price. r 2011 American Chemical Society
Nevertheless, prior to the design of complex architectures involving XG chains based on the self-assembly concept and using XG chain’s properties in an optimal manner, the question of XG conformation and organization in aqueous solvent still needs to be addressed. Indeed, the difficulty and the limitations of using XG chains in aqueous media in a controlled way first rely on their solubilization. As XG chains are extracted from natural compounds, they are highly polydisperse and only slightly soluble in water since they can easily form aggregates via hydrogen bonding.18 Different solutions have been proposed to overcome this tendency to self-association, like XG carboxylation that obviously changes the molecular architecture19 or pressure cell-assisted solubilization,20 that decreases the molecular weight and improves XG solubility. The conformational behavior of xyloglucan chains has thus been extensively studied in various experimental situations. So far, the most complete data have been obtained from hydrodynamical behaviors studied by viscosimetry and/or rheological experiments, mostly combined with static and dynamic light scattering on series of xyloglucan of different origins and solubilized in various conditions.1921 Even if the overall conformation behavior is coherent with the picture of a relatively stiff polymer, the results yielded large approximation of local parameters, mainly because they were deduced from Received: June 27, 2011 Revised: July 29, 2011 Published: August 01, 2011 3330
dx.doi.org/10.1021/bm200881x | Biomacromolecules 2011, 12, 3330–3336
Biomacromolecules macroscopic data assuming structural models. In particular, the MarkHouwinkSakurada (MHS) exponents were found in a range of values where it is not possible to clearly discriminate between a conformation in good or θ solvent.22 In fact, to the best of our knowledge, only Picout et al.20 attempted to determine using light scattering, the most relevant parameter of a macromolecule: the Flory exponent, ν, which is representative of the interactions between the chemical groups of the polymer chain and the solvent. Unfortunately, the result was still ambiguous, particularly for XG from taramind as they found ν = 0.54 ( 0.07, a value that could correspond either to a Gaussian (θ-solvent statistic, ν = 0.50) or to a swollen coil (good solvent statistic, ν = 0.588). Besides, the very small experimental value obtained for the second virial coefficient A2 did not allow the authors to conclude on the chain statistics. To our opinion, small-angle scattering methods based on X-rays (SAXS) or neutrons (SANS) are more suitable tools than light scattering methods in order to unambiguously determine the XG conformation in water. Indeed, their high statistical averaging and their short wavelength make direct structural investigations possible on characteristic length scales of a polymer chain, typically from 1 to 100 nm. Here, we will use SANS, demonstrating that this technique allows us to obtain exhaustive and direct determinations of the chain conformation with an utmost precision and all the structural local parameters of the chains. The sample preparation is based on a simple solubilization process that enables us to obtain stable solutions of XG in water. Its relevance is demonstrated using size exclusion chromatography coupled to multiangle laser light scattering (SEC-MALLs) and liquid-state nuclear magnetic resonance (NMR) prior to the SANS measurements.
’ MATERIALS AND METHODS Materials and Solubilization Process. XG chains from tamarind seeds (XG) were obtained from Dainippon Pharmaceutical Co., Ltd. (Osaka, Japan), grade 3S. A mass of 3 g of the product was heated at 100 °C under reflux for 3 h in a 500 mL round-bottom flask with 300 mL of 0.01 M NaOH (with a few milligrams of NaN3 to prevent microbial spoilage) under constant agitation to achieve a good dispersion of the powder. The flask is then allowed to cool to room temperature and its content is centrifuged for 30 min at 20000 g (here, g = 9.8 m.s2). The supernatant is dialyzed against water until the conductivity reaches a value below 5 μS 3 cm1. Chloroform drops are added to the cell dialysis and kept during all the dialysis process to prevent microbial spoilage during the removing of all salts including NaN3. From our observations, microbial spoilage of XG in such solution is very quick; hence this point requires a careful attention. The dry mass of the final solution has been determined and found typically between 2 and 8 g/L, depending on the dialysis conditions. The solution is finally stored in a flask with again a few milligrams of sodium azide. For neutron experiments (see below), the aqueous dispersions were dialyzed in three successive baths of pure D2O, to enhance the contrast of the hydrogenated XG chains as compared with the aqueous solvent. Nuclear Magnetic Resonance (NMR). The experiments have been carried out with a Tecmag spectrometer operated at a 1 H frequency of 300 MHz. The 90° pulse length was 12.3 μs. DMSO has been used as an internal standard for quantitative measurements. The total repeating time was fixed at 34 s, in order to make possible the total relaxation of the DMSO signal, which T1(1H)
ARTICLE
was measured at 7 s. Temperature in the probe was kept constant at 300 K with a SHAKE temperature controller. The number of transients was fixed at 256 to get the best signal-to-noise ratio. The quantitative measurements have been done as follows. DMSO exhibits a very sharp signal in proton NMR spectra (at 2.7 ppm), which is located well outside the complex xyloglucan resonances. The procedure consists of dissolving known amounts of XG and anhydrous DMSO in D2O to reach a constant final weight concentration of 0.1 g/L of DMSO and 8 g/L of XG in order to record spectra with lines of comparable intensity. It has been assumed that the presence of DMSO at this very low concentration does not affect the overall polarity of the medium. Moreover, it has been checked out that DMSO does not interact with XG by comparing spectra with and without the internal standard, which turned out to be identical. The proportion of solubilized XG has been measured by integrating the signals arising from the anomeric protons between 4.5 and 5.2 ppm. Size Exclusion Chromatography Coupled to Multiangle Laser Light Scattering (SEC-MALLs). The solubilized XG chains have been characterized by size exclusion chromatography (SEC) with triple detection using a Waters Alliance GPCV2000 (U.S.A.) equipped with a differential refractometer, a differential viscosimeter, and a multiangle laser light scattering (MALLS) detector from Wyatt (U.S.A.). The concentration injected was 0.5 g/L, with an injection volume of 100 μL using two columns in series (Shodex OH-pack 803 and 805). All samples were filtered on a 0.2 μm pore membrane (Sartorius AG; cellulose acetate filter) before injection in order to retain large aggregates, if any. It was also checked that the overall elution profiles remained globally symmetric. The eluent used was 0.1 M NaNO3, at 30 °C elution temperature and using a flow rate of 0.5 mL/min. The value of dn/dc adopted is equal to 0.142, as used by Picout et al.20 The molecular weight distribution, weight average molecular weight (Mw), polydispersity index (Mw/Mn), and the intrinsic viscosity ([η], mL/g) have been obtained as characteristics of the biopolymer. Small-Angle Neutron Scattering (SANS). The experiments have been carried out at the Orphee reactor in Saclay on the PACE spectrometer. Three different configurations were used (neutron wavelength λ = 13 Å, sample to detector distance d = 4.7 m; λ = 5 Å, d = 4.7 m; λ = 5 Å, d = 1 m), giving access to exploitable data in the Q-range from Q = 3.18 103 Å1 to Q = 0.42 Å1. Samples were enclosed in quartz cells of 2 mm inner thickness. All the measurements were performed at atmospheric pressure and room temperature. The scattered intensities were corrected for the detector background by cadmium scattering, for the parasitic intensity scattered by quartz cell by subtraction and normalized to the water scattered intensity. Standard procedures for data reduction23 were done using the Pasinet software24 to obtain the scattered intensity, or more precisely the differential cross section per unit of volume, in absolute scale (cm1). The scattering intensity of the solute was then obtained by subtraction of the intensity of the solvent (measured independently) to the one of the solution. In the following, this final intensity will be noted I(Q). Owing to the very small scattered intensity expected for the sample, very long counting rates were used for the sample as well as for the standards and solvent, up to 8 h for the configuration corresponding to the lowest Q. The error bars, calculated from the Poisson counting statistics of the raw 2D data, strongly depend on the chosen configuration as the incoming neutron flux is highly dependent on the wavelength and sample-to-detector distance. Hence they appear to not propagate regularly on the whole measured Q-range (three configurations). 3331
dx.doi.org/10.1021/bm200881x |Biomacromolecules 2011, 12, 3330–3336
Biomacromolecules
ARTICLE
Table 1. Characteristics of the XG Chains from SEC-MALLSa measurement
Mn (106 g/mol)
Mw (106 g/mol)
I (Mw/Mn)
Rg (Å)
[η] (dL/g)
K (mL/g)
a
1
0.33
0.47
1.42
508
3.47
7.3 102
0.67
2
0.32
0.46
1.44
507
3.34
7.5 102
0.67
a
Number-average molecular mass Mn, weight-average molecular mass Mw, polydispersity index I, radius of gyration Rg, intrinsic viscosity [η], scattering constant K, and MarkHouvinkSakurada parameter a.
’ RESULTS AND DISCUSSION Testing the Consistency of the Solubilization Process. We have developed an alternative approach to XG carboxylation19 and to pressure cell-assisted solubilization.20 This potentially makes it possible to obtain stable solutions of free XG in water without aggregation or chain scission, hence, possible molecular solubilization of individual XG chains. This approach is based on the use of a solvent with a chaotropic character during the solubilization process to prevent intermolecular interactions between XG chains (a NaOH solution here). Therefore, prior to the SANS measurements, the relevance of this approach is shown. (a). Characteristics of the XG Chains by SEC-MALLs Experiments. The good solubilization of the XG chains was first checked by SEC-MALLS. The experiment has been repeated at different times after preparation on a sample with an XG concentration of 7.8 g/L. The results were found reproducible. The obtained chromatogram did not show any presence of aggregates and the recovery mass was greater than 95% for all measurements. The characteristic parameters of the chains have been derived from the viscosimetry and static light scattering data. The parameters obtained on two measurements are presented in Table 1. The rather low values of the weight-average molecular mass obtained tend to indicate that the pre-existing large and compact aggregates formed by strong intermolecular association of XG chains have been removed by the solubilization process, as already pointed out in previous studies.20 Although indicative on this sample, the MarkHouvinkSakurada parameter of a = 0.67 obtained from the viscosimetric data is within the range obtained for linear macromolecules (0.5 e a e 0.8). Furthermore, it is in perfect agreement with the ones proposed by Picout et al.20 This is a very good indication of the reliability of the dissolution process. The recovering of more than 95% of the initial materials at the end of the experiments definitely rules out the presence of aggregates formed by strong links. Therefore, these results are compatible with a good solubilization at the molecular level of the XG chains. Please note that SEC-MALLs does not allow to conclude on the in situ formation of fragile aggregates based on H-bonds because such aggregates can be easily broken by the high shear applied on the solution during its passage within the size exclusion column under flow. (b). Composition and Stability of the XG Chains by Quantitative NMR. The 1H NMR spectra also allowed us to determine the effective molar ratio of saccharides in the sample.4 In that specific case, the experiments were conducted at 353 K to avoid the superimposition of the anomeric and water proton signal. The ratios were found to be 1.00, 0.45, and 1.30 for xylose, galactose, and glucose, respectively. The slight difference with previously reported values for tamarind seeds (1.00, 0.51, and 1.34) obtained from chromatography experiments25 could arise from different response factors depending on the sugars and yet they stay within the experimental uncertainties.
Figure 1. Evolution in time of the ratio of intensity between the integrals of XG and DMSO at 3 g/L (0.3 wt %; squares) and 5 g/L (0.5 wt %; diamonds). The full lines are guides to the eyes, showing the invariance in time for both concentrations.
To check out the long-term stability in time of the obtained solutions, which is a prerequisite for further use and characterization, including scattering experiments (see below), we recorded quantitative nuclear magnetic resonance (NMR) spectra over a period of a few weeks. We first checked the quantitative character of the experiments by independently measuring the concentration within the NMR tube using the DMSO internal standard. Assuming the molar ratio measured previously and the numbers of protons for each species, the theoretical ratio between the proton integral of DMSO and XG should be 0.90, whereas an experimental value of 0.94 was measured within the experimental error. The stability of the solutions was then checked by recording a series of spectra over more than 30 days at regular intervals. The basis of this experiment relies on the molecular weight detection limit of liquid state NMR due to the long tumbling times of large aggregates or microgels that make their detection impossible with this technique. The overall shape of the spectra remained identical, showing that the XG structure does not evolve with time. Taking into account that large aggregates are not detected by liquid-state NMR owing to their very long tumbling times, this favorable comparison means that the entire polymer was dissolved at a molecular scale. Figure 1 shows the evolution of the intensity ratio of the integrals of XG and DMSO peaks, which was found constant within the experimental precision over more than one month. This behavior interestingly indicates the overall stability in time of the sample, within the precision and time scale of the experiments. Thus, once the initial solubilization is achieved, the structure formed by XG chains in water no longer evolves into large aggregates or microgels over at least a month. Therefore, the interpretation of SANS will not in principle be affected by kinetics or structural evolution. Structural Behavior of XG Chains at Different Scales: SANS Experiments. For the aforementioned reasons, SAXS and SANS techniques are perfectly suited tools for the determination of the structural local parameters of a polymer chain (Flory exponent, 3332
dx.doi.org/10.1021/bm200881x |Biomacromolecules 2011, 12, 3330–3336
Biomacromolecules
Figure 2. SANS data of a 5 g/L XG solution in D2O: (a) I(Q) vs Q representation. The full lines indicate the scaling law of the three different observed behaviors. Alternative representation of the SANS data: (b) Q1.7I(Q) vs Q representation, and (c) QI(Q) vs Q representation. In both panels, the dashed lines represent the Q-independent behavior for each representation. The onsets in Q of the different behaviors discussed in the text are displayed in each panel.
persistence length, and cross-section of the chains) that have largely been used for synthetic polymers. The only work where these techniques were used on XG chains was performed by Gidley et al.4 who demonstrated that XG chains are not branched chains but linear ones using SAXS. It was thus limited to a restricted high Q-range (very local length scales); hence, these experiments provided only the contour radius of the chains and not other local parameters. Our present aim is to use the SANS technique to fill the lack of structural information of XG chains from tamarind seeds dispersed in water on the relevant length scale of 580 nm. (a). Description of the Scattering Curve. The structural behavior of the chains at the local scale has been explored using SANS. The scattering curve of a XG solution at 5 g/L in D2O is shown in Figure 2. Please first note that the long counting rates used during the experiments result in very good statistics and very low error bars down to intensities of the order of 103 cm1, an intensity range that is rarely attained in classical SANS experiments for which the lower limit is usually of the order of 102 cm1. It is the quality of these results that allows us to propose a detailed description of the microscopic behavior of the XG chains, which has not yet been done to our knowledge. The scattering curve clearly displays three distinct standard behaviors corresponding to polymer chains in good solvent with a persistence length: at low Q, that is, at large scale in direct space, the scattering curve decays with a scaling power law of 1.7 within the experimental errors. There is a first break of the slope around Q0 = 0.042 Å1 and the scattering decays at intermediate Q with a new scaling power law of 1. There is a second break of slope around Q1 ∼ 0.2 Å1. It is difficult to unambiguously conclude on the decay behavior above such scattering wavevectors for two reasons. First, the probed Q-range is very narrow, making it difficult to determine the exponent of a potential power law. Second, the order of magnitude of the obtained values of I(Q) at such high Q-values is
ARTICLE
103 cm1 and are thus very sensitive to the data treatment. Indeed, any small changes in the measured intensity, both for the signal of the suspension and for the solvent, could induce a big change on the intensity variation. On the contrary, there is no ambiguity on the behavior at low and intermediate Q-values, as proven by Figure 2, which displays the scattering curves obtained in both Q1.7I(Q) and QI(Q) versus Q representations. (b). Determination of the Conformational Statistics and Flory Exponent. At low Q, the Q1.7 decay of the scattered intensity remarkably shows that the XG chains are well dispersed and that they behave as polymer chains in good solvent with excluded volume interactions. Indeed, at this characteristic length scale in direct space, one probes the Flory exponent ν describing their conformation26,27 that links the radius of gyration Rg and the number of repetitions units N via Rg ∼ Nν. Thus, in the reciprocal space, the chains scatter with a Q-dependent slope of Q1/ν. The Q1.7 decay enables the recovery of the exact value of the Flory exponent ν of 1/1.7 = 0.588 (sometimes noted as 3/5) of polymer chains in good solvent with excluded volume interactions. It is noteworthy that the representation of Figure 2b unambiguously proves that the experimental scaling law is not compatible with a Q2 decay, which should be obtained for polymer chains in θ-solvent with a Flory exponent of 0.5 (Gaussian statistic). (c). Structural Behavior at Large Scale. It should be noted that a down-turn from a Q1.7 scaling law could be observed for the smallest Q-values investigated (Figure 2b). As the Q-range for which it is observed is very narrow, it is difficult to assess at this stage if it comes from the reaching of the Guinier regime, that is, the Q-regime where one probes the gyration radius Rg of the whole chain, or if it comes from the existence of a specific structure of the system at large scale. However, the value of Rg ≈ 510 Å obtained by SEC-MALLs technique (Table 1), in fact, rules out the possibility to have reached the Guinier regime. Such regime, indeed, takes place for QRg < 1, thus, below Q ≈ 2 103 Å1, which is much lower than the observed down-turn onset (Figure 2b) and out of our Q-window measurement. Therefore, the down-turn comes from the presence of a specific structure in the system at large scale, which has a typical correlation distance ξ (with a value of ∼900 Å derived from ξ = 2π/Qξ, with Qξ ≈ 7 103 Å1) In addition, let us point out that this down-turn at low Q shows that there are no compact aggregates in the solution. Indeed, in case of compact aggregates, the slope of the scattering decay at low Q would have been higher than 1.7; hence, an intensity up-turn rather than a down-turn should be observed. This confirms the SEC-MALLs results. If the large scale structure is due to the formation of labile H-bonds, it does not affect the average conformation of the XG chains at length scales below ξ, that is, below 900 Å. Hence, an isolated XG chain would follow the similar selfavoiding behavior as its constitutive repeating units. (d). Structural Behavior at Intermediate and Local Scale. Above Q0 = 0.042 Å1, at intermediate scale, the local conformation of the chains, that is, their form factor, are probed. The Q1 scaling law (Figure 2c) proves that the XG chains have a local rodlike behavior because they exhibit a fractal dimension of 1 (here, I(Q) ∼ Qd, where d is the fractal dimension). This indicates that the XG chains are semiflexible chains. Namely, at this local scale, the chains are stiff to some extent. The persistence length lp characterizes this stiffness. The value of Q0 corresponds to the Q-limit of the regime for which the scattering exclusively arises from the stiffness. The onset of this latter regime depends strongly on the conformation of the chains. From the relation for worm-like chains in good solvent (ν = 0.588), Qonset ≈ 3.5/lp,28,29 we can deduce an 3333
dx.doi.org/10.1021/bm200881x |Biomacromolecules 2011, 12, 3330–3336
Biomacromolecules
ARTICLE
estimation of lp of around 83 Å. This result shows that the stiffness of XG chains in water corresponds to 1617 sugar units, which corresponds to four repeating blocks because the repeating unit for XG has an extended length of 20.6 Å. When going toward large Q, the Q1 regime is observed as long as the monomers can be considered as punctual. At higher Q, deviations from the stiff rod-like behavior are observed because the thickness of monomers (or cross-section) contributes to the scattering. This arises here at Q1. However, as pointed out before, it is difficult to assess a clear exponent to the scattering power law. We refrain, thus, ourselves to determine the thickness by this direct observation because the experimental errors do not allow us to obtain a reliable value, despite the apparent Q2 scaling law. (e). Quantitative Determination of lp and Chain CrossSection from the Modeling of the Scattering Curve. Beyond the previous description based on power laws and characteristic scattering wavevector crossovers, an independent quantitative analysis can be achieved by applying a unique adequate model to adjust the whole scattering curve. For semiflexible polymer chains, also called KratkyPorod (KP) wormlike chains, the key parameter is the determination of the persistence length parameter lp.30 The modeling of KP chains is a priori tricky because the chain cannot be analytically described at all length scales. In the case of infinite KP chains, des Cloizeaux has given an exact calculation of the form factor.31 In the case of finite KP chains, an analytical expression has been given by Sharp and Bloomfield32 but is only valid for Qlp e 2. Pedersen and Schurtenberger33 have given expressions of the form factor for the whole Q-range both for chains with Gaussian and with excluded volume statistics. Such expressions have been derived from Monte Carlo simulations where the results were parametrized numerically using the procedure of Yoshizaki and Yamakawa.34 The model consists of calculating the form factor for a flexible cylinder of length L (contour length) with a circular cross section of diameter 2R and a uniform scattering length density. The flexible cylinder is made of subcylinders of length lp. This model is, thus, a discrete representation of the worm-like chain model of Kratky and Porod applied in a pseudocontinuous limit. In the Pedersen Schurtenberger model the total form factor is expressed as Ptot ðQ Þ ¼ ½ð1 χðQ , L, bÞPchain ðQ , L, bÞ þ χðQ , L, bÞProd ðQ , LÞΓðQ , L, bÞ
ð1Þ
where Pchain(Q,L,b) is the form factor of a flexible chain without interactions between monomers or with excluded volume interactions (depending on the investigated case) and Prod(Q,L) is the form factor of an infinitely thin rod.35 At small Q-values, Pchain(Q,L,b) dominates and at large Q-values it is Prod(Q,L). The transition between the two regimes is described by the crossover function (Q,L,b), and Γ(Q,L,b) is a correction function for the crossover region. As we unambiguously showed that the chains have excluded volume interactions, we used this model to fit our data thanks to the SASfit software36 where it is included. The model has four different free parameters: (1) the intensity scattered when Q f 0, I0, linked to the mass of the chains, (2) the contour length, L, of the chains, (3) the persistence length, lp, and (4) the chain cross section radius, R. Because the average mass of the XG chains measured using SEC-MALLS is large, the accessible Q-range of our SANS data is too limited to permit a determination of L and I0. Therefore, these two parameters were first fixed at arbitrary large values to shift the Guinier regime at
Figure 3. I(Q) vs Q representation of the SANS data with the adjustment using the PedersenSchurtenberger model of a Kratky and Porod worm-like chain model with excluded volume interactions (full line). The parametersL and I0 due to the accessible Q-range were fixed at arbitrary large values.
very low Q. As we believe that there exists a structure at large scale in the system (see previous section), we have not considered in the fit the data points below 7 103 Å1, because the scattering curve is not equal to a single form factor in this Q-range. The best fit obtained is shown in Figure 3 and clearly demonstrates that the model used gives very good agreement with the experimental data. The structural parameters found for such a fit are lp = 76.2 Å and R = 3.16 Å. The value of the persistence length is in remarkable agreement with the one obtained previously (≈83 Å) by the simple estimation from the scattering curves. This confirms to some extent the validity of the model. The fit also gives us a chain’s cross section of 2R ≈ 6.32 Å with enough confidence to be relevant as compared to simple observation. It is in good agreement with the one obtained by Gidley et al.4 through a Guinier plot of high Q-values of SAXS experiments. We have unambiguously demonstrated that XG chains dispersed in water can be described as semiflexible worm-like chains with a self-avoiding statistic (good solvent). The direct measurement of the Flory exponent (ν = 0.588) is clearly an important and central feature in the field of the XG study. Clearly, XG chains from tamarind seeds dispersed in water do not exhibit Gaussian statistics (θ-solvent, ν = 0.50), but have a self-avoiding conformation. Furthermore, it has been also proved that XG chains admit a persistence length that is larger than their statistical basic units, whose value has been evaluated with an utmost precision thanks to the direct measurement of the crossover from chain conformation to rigid rod scaling behavior. Finally, this indicates that the value of ν = 0.54 ( 0.07 obtained by Picout et al20 should correspond to swollen chains (ν = 0.588). Therefore, this confirms the role of the hydrophobicity of the side residues of XG chains on their behavior in water, as a slight decrease of xylose content enables to pass from a θ-solvent behavior (as for detarium seeds) to a good solvent behavior (as for tamarind seeds).20 The local parameters determined by modeling are in very good agreement with the ones deduced by direct observation of the scattering curve. This allows us to be confident about the utmost precision of their values. Such precision is of main importance because, although difficult to measure or simply evaluate, the persistence length of a bare cellulosic backbone is in the 110130 Å range, whereas it has been shown to drop below 100 Å when substituents decorate the main chain.37 Therefore, the 80 Å persistence length derived from our experimental results 3334
dx.doi.org/10.1021/bm200881x |Biomacromolecules 2011, 12, 3330–3336
Biomacromolecules
ARTICLE
use of SANS has been shown to be of central interest for the unambiguous determination of the physical characteristics in the case of XG chains from tamarind seeds. This technique has been widely used for other types of macromolecules and appears as an ideal tool to investigate the local structure of such biological macromolecules, as made for instance for charged polysaccharides such as pectin44 or hyaluronan,4547 and could generally be applied, in our opinion, to different types of hemicelluloses.
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected]. Figure 4. Representation of an isolated XG chain in water: The chain is a self-avoiding chain (ν = 0.588) with a persistence length of 80 Å.
is reinforced with the idea that the decoration of XG chains is a trick used by Nature to make the dissolution of cellulose-like backbone chains in water possible. A schematic representation of the XG chains conformation is shown in Figure 4.
’ CONCLUSIONS In this contribution, it was shown that it was possible to disperse unfractionated XG chains extracted from natural seeds in aqueous medium by applying an adequate solubilization process that uses a chaotropic solution as solvent to minor H-bonding formation. The obtained solutions were stable over weeks. The physical parameters of the chains were determined by SANS, which was used for a first time on a broad Q-range in the case of XG chains in order to probe all of the relevant scales of this polymeric system. At the lowest Q-range probed during our SANS experiments, we observed a correlation that may come from a structure of XG chains at a large scale although such chains are perfectly swollen at the local scale. In a further work, we will investigate such larger scales by static and dynamic light scattering experiments in order to achieve the description of such a large-scale structure and the mechanisms of its formation. Remarkably, it was unambiguously demonstrated that the XG chains are semiflexible worm-like chains in good solvent. The chain’s stiffness was determined with a much higher precision than the values previously reported in the literature and corresponds to 1617 glucose backbone monomers. The chain’s cross section (∼6.3 Å) is statistically as thick as the average substituent length (Figure 4). Interestingly, the local stiffness corresponds to the minimum length needed for the interaction between XG and cellulose,7,8 whereas no preferential backbone orientation has been evidenced, indicating a rather nonspecific interaction between cellulose and the XG chains.10 All together, these results emphasize the possible role of the local stiffness on the interaction between chains in solution and cellulose surfaces. It could also explain why usual much more flexible hydrosoluble polymers (including for example dextran) do not interact with cellulose surfaces. Conversely, more rigid βf4 linked polysaccharides like (gluco)mannan38,39 or to a certain extent (arabino)xylan40 share common conformational features41 and do interact with cellulose surfaces,42,43 despite their rather different sugar composition. Additionally, a precise conformational description of the polysaccharides potentially interacting with cellulose will gain more and more interest for both the production of reliable models and the design of biomimetic materials. In that sense, the
’ ACKNOWLEDGMENT This work is funded by the French National Project ANR-08NANO-P235-36. Quentin Serra is acknowledged for performing the liquid-state NMR experiments. ’ REFERENCES (1) Carpita, N. C.; Gibeaut, D. M. Plant J. 1993, 3, 1–30. (2) Fry, S. C. The growing plant cell wall. Chemical and metabolic analysis; Longman Scientific & Technical Ed.: Essex, U.K., 1988. (3) Fry, S. C.; York, W. S.; Albersheim, P.; Darvill, A.; Hayashi, T.; Joseleau, J. P.; Kato, Y.; Lorences, E. P.; Maclachlan, G. A.; Mcneil, M.; Mort, A. J.; Reid, J. S. G.; Seitz, H. U.; Selvendran, R. R.; Voragen, A. G. J.; White, A. R. Physiol. Plant. 1993, 89, 1–3. (4) Gidley, M. J.; Lillford, P. J.; Rowlands, D. W.; Lang, P.; Dentini, M.; Crescenzi, V.; Edwards, M.; Fanutti, C.; Reid, J. S. G. Carbohydr. Res. 1991, 214, 299–314. (5) Vincken, J. P.; Dekeizer, A.; Beldman, G.; Voragen, A. G. J. Plant Physiol. 1995, 108, 1579–1585. (6) Faik, A.; Chileshe, C.; Sterling, J.; Maclachlan, G. Plant Physiol. 1997, 114, 245. (7) Hayashi, T; Ogawa, K.; Mitsuishi, Y. Plant Cell Physiol. 1994, 35, 1199–1205. (8) Lopez, M.; Bizot, H.; Chambat, G.; Marais, M.-F.; Zykwinska, A.; Ralet, M.-C.; Driguez, H.; Buleon, A. Biomacromolecules 2010, 11, 1417–1428. (9) Lima, D. U.; Bruckeridge, M. S. Carbohydr. Polym. 2001, 46, 157–163. (10) Lima, D. U.; Loh, W.; Bruckeridge, M. S. Plant Physiol. Biochem. 2004, 42, 389–394. (11) Chambat, G.; Karmous, M.; Costes, M.; Picard, M.; Joseleau, J. P. Cellulose 2005, 12, 117–125. (12) Hanus, J.; Mazeau, K. Biopolymers 2006, 82, 59–73. (13) Zhou, Q.; Rutland, M. W.; Teeri, T. T.; Brumer, H. Cellulose 2007, 14, 625–641. (14) Kumar, C. S.; Bhattacharya, S. Crit. Rev. Food Sci. Nutr. 2008, 48, 1–20. (15) Winter, H. T.; Cerclier, C.; Delorme, N.; Bizot, H.; Quemener, B.; Cathala, B. Biomacromolecules 2010, 11, 3144–3151. (16) Jean, B.; Heux, L.; Dubreuil, F.; Chambat, G.; Cousin, F. Langmuir 2009, 25, 3920–3923. (17) Cerclier, C.; Cousin, F.; Bizot, H.; Moreau, C.; Cathala, B. Langmuir 2010, 26, 17248–17255. (18) Lang, P.; Burchard, W. Makromol. Chem. 1993, 194, 3157–3166. (19) Lang, P.; Kajiwara, K.; Burchard, W. Macromolecules 1993, 26, 3992–3998. (20) Picout, D. R.; Ross-Murphy, S. B.; Errington, N.; Harding, S. E. Biomacromolecules 2003, 4, 799–807. (21) Ren, Y.; Picout, D. R.; Ellis, P. R.; Ross-Murphy, S. B. Biomacromolecules 2004, 5, 2384–2391. 3335
dx.doi.org/10.1021/bm200881x |Biomacromolecules 2011, 12, 3330–3336
Biomacromolecules
ARTICLE
(22) Patel, T. R.; Morris, G. A.; Ebringerova, A.; Vodenicarova, M.; Velebny, V.; Ortega, A.; Garcia de la Torre, J.; Harding, S. E. Carbohydr. Polym. 2008, 74, 845–851. (23) Br^ulet, A.; Lairez, D.; Lapp, A.; Cotton, J. P. J. Appl. Crystallogr. 2007, 40, 165–177. (24) Lairez, D. Pasinet; software available free of charge at http:// didier.lairez.fr/. (25) Wang, Q.; Ellis, P. R.; Ross-Murphy, S. B.; Reid, J. S. G. Carbohydr. Res. 1996, 284, 229–239. (26) Flory, P. J. Principles of polymer chemistry; Cornell University: New York, NY, 1953. (27) De Gennes, P.-G. Scaling concepts in polymer physics; Cornell University Press: Ithaca, NY, 1979. (28) Rawiso, M.; Duplessix, R.; Picot, C. Macromolecules 1987, 20, 630–648. (29) Rawiso, M.; Aime, P.; Fave, J. L.; Schott, M.; M€uller, M. A.; Schmidt, M.; Baumgartl, H.; Wegner, G. J. Phys. (Paris) 1988, 49, 861–880. (30) Kratky, O; Porod, G. Recl. Trav. Chim. Pays-Bas 1949, 68, 1106. (31) des Cloizeaux, J. Macromolecules 1973, 6, 403. (32) Sharp, P.; Bloomfield, V. A. Biopolymers 1968, 6, 1201. (33) Pedersen, J. S; Schurtenberger, P. Macromolecules 1996, 29, 7602–7612. (34) Yoshizaki, T.; Yamakawa, H. Macromolecules 1980, 13, 1518. (35) Pedersen, J. S. In Neutrons, X-rays and Light: Scattering Methods Applied to Soft Matter; Lindner, P., Zemb, T., Eds.; Elsevier: NorthHolland, The Netherlands, 2002; Chapter 16, p 401. (36) Kohlbrecher, J. SASfit; software available free of charge at http://kur.web.psi.ch/sans1/SANSSoft/sasfit.html. (37) Kroon-Batenburg, L. M. J.; Kruiskamp, P. H.; Vliegenthart, J. F. G.; Kroon, J. J. Phys. Chem. B 1997, 101, 8454–8459. (38) Picout, D. R.; Ross-Murphy, S. B.; Jumel, K.; Harding, S. E. Biomacromolecules 2002, 3, 761–767. (39) Morris, G. A.; Patel, T. R.; Picout, D. R.; Ross-Murphy, S. B.; Ortega, A.; Garcia de la Torre, J.; Harding, S. E. Carbohydr. Polym. 2008, 72, 356–360. (40) Dervilly-Pinel, G.; Thibault, J-F; Saulnier, L. Carbohydr. Res. 2001, 330, 365–372. (41) Picout, D. R.; Ross-Murphy, S. B. Carbohydr. Res. 2002, 337, 1781–1784. (42) Whitney, S. E. C.; Brigham, J. E.; Darke, A. H.; Reid, J. S. G.; Gidley, M. J. Carbohydr. Res. 1998, 307, 299–309. (43) Linder, A. P.; Bergman, R.; Bodin, A.; Gatenholm, P Langmuir 2003, 19, 5072–5077. (44) Cros, S.; Garnier, C.; Axelos, M. A. V.; Imberty, A.; Perez, S. Biopolymers 1996, 39, 339–352. (45) Buhler, E; Boue, F Eur. Phys. J. E 2003, 10, 89–92. (46) Buhler, E; Boue, F. Macromolecules 2004, 37, 1600–1610. (47) Bonnet, F.; Schweins, R.; Boue, F.; Buhler, E. Europhys. Lett. 2008, 83, 48002.
3336
dx.doi.org/10.1021/bm200881x |Biomacromolecules 2011, 12, 3330–3336
