International Journal of Biological Macromolecules 26 (1999) 225 231 www.elsevier.

com/locate/ijbiomac

Role of galactomannan composition on the binary gel formation with xanthan

T.M.B. Bresolin a,b, M. Milas a, M. Rinaudo a, F. Reicher b, J.L.M.S. Ganter b,*
a

Centre de Recherches sur les Macromolecules Vegetales-CNRS, afliated with Uni6ersite Joseph Fourier, B.P. 53, 38041, Grenoble Cedex 9, France b Departamento de Bioqumica, Uni6ersidade Federal do Parana, C.P. 19046, 81531 -990 Curitiba (Pr), Brazil Received 14 August 1998; accepted 14 January 1999

Abstract The inuence of the galactomannan characteristic ratios (M/G) on the temperature of gelation (Tg) and the gel strength of mixtures of galactomannan with xanthan is reported. Two galactomannans were investigated: one highly substituted from the seeds of Mimosa scabrella (M/G=11), and the other, less substituted, from the endosperm of Schizolobium parahybae, with (M/G = 30) [Ganter JLMS, Zawadzki-Baggio SF, Leitner SC, Sierakowski MR, Reicher F. J Carbohydr Chem 1993;12:753]. The xanthan:galactomannan systems (4:2 g l 1, in 5 mM NaCl) showed a temperature of gel formation (Tg) of 24C for that of S. parahybae [Bresolin TMB, Milas M, Rinaudo M and Ganter JLMS. Int J Biol Macromol 1998;23:263] and 20C for the galactomannan of M. scabrella, determined by viscoelastic measurements and microcalorimetry. A Tg of 40 50C was found by Shatwell et al. [Shatwell KP, Sutherland IW, Ross-Murphy SB, Dea ICM. Carbohydr Polym 1991;14:29] for locust bean gum-LBG (M/G =43). Lundin and Hermansson [Lundin L, Hermansson AM. Carbohydr Polym 1995;26:129] reported a difference of 13C for Tg of two LBG samples with M/G=3 (40C) and 5 (53C), in mixtures with xanthan. It appears that the more substituted galactomannans have lower temperatures of gelation in the presence of xanthan. The mechanism of gelation depends also on the M/G ratio. For the lower values it involves only disordered xanthan chains in contrast to M/G ratios higher than 3. In addition, the presence of the galactomannan from M. scabrella increased slightly the temperature of the conformational change (Tm) of xanthan probably due to the ionic strength contribution of proteins (3.9%) present in the galactomannan. On the other hand, the galactomannans from S. parahybae, with 1.5% of proteins and M. scabrella, with 2.4% of protein, did not show this effect, the Tm of xanthan alone or in a mixture being practically unchanged. 1999 Elsevier Science B.V. All rights reserved.
Keywords: Xanthan:galactomannan interaction; Protein; Binary gel

1. Introduction Xanthan consists of a b-(1 4)-D-glucopyranosyl backbone substituted at O-3 on every second unit with a charged trisaccharide side-chain composed of a residue of glucuronic acid between two mannosyl units. The terminal mannosyl unit may be substituted at O-4 and O-6 by a pyruvate acetal. An O-acetyl group is frequently present at O-6 of the inner mannosyl [1]. This polymer shows a temperature characteristic of a conformational transition (Tm), with a disordered conformation above Tm and a helical conformation below
* Corresponding author. Tel.: +55-41-366-3144; fax: + 55-41-2662042. E-mail address: ganter@bio.ufpr.br (J.L.M.S. Ganter)

Tm, which depends on the ionic concentration of the solutions. Galactomannans are the energy-reserve polysaccharides in seeds of endospermic leguminous plants. These polymers generally consist of a main chain of (14) linked b-D-mannopyranosyl residues, most of which substituted at O-6 with single unit a-D-galactopyranosyl side-chain residues. The ratio of mannose to galactose (M/G) depends of the plant source and the method of extraction and can range from 1.0 to 5.0 [2,3]. The industrial uses of galactomannans and their presence in a wide variety of Brazilian trees directed attention to new sources, such as Senna multijuga [4], Stryphnodendron barbatimam [5], Schizolobium amazonicum [6], Schizolobium parahybae and Mimosa scabrella [7], amongst others. By the usual method

0141-8130/99/$ - see front matter 1999 Elsevier Science B.V. All rights reserved. PII: S 0 1 4 1 - 8 1 3 0 ( 9 9 ) 0 0 0 8 7 - 2

226

T.M.B. Bresolin et al. / International Journal of Biological Macromolecules 26 (1999) 225231

employed to obtain these polysaccharides, the whole seeds were milled and submitted to aqueous extraction at 4 or 25C, followed by centrifugation and precipitation of the supernatant with 75% (v/v) ethanol. As the products contained from 3 to 15% of protein, they were treated with organic solvent [8] in order to remove the free proteins (non-covalently linked). This treatment resulted in a decrease of approximately 40% of the initial protein in the galactomannan samples studied. The same effect was obtained when 50% ethanol (v/v) was used for precipitation of the polysaccharide. Free proteins are probably a reserve material of the embryo, coprecipitated during isolation of the galactomannan. In order to eliminate the residual protein, other methods were applied, such as enzymatic and alkaline hydrolysis, copper complexation and variation of ionic strength. The results showed that the galactomannan samples always retained small amounts of protein (around 12%), which may be covalently linked and/or adsorbed to the polysaccharide [9]. Galactomannans extracted from isolated endosperms also showed the presence of small amounts of proteins (1.5%) [10]. Mazzini and Cerezo [11] found proteins in galactomannans of Gleditsia triacanthos and other legume seeds submitted to different procedures of purication. This resistance to extensive efforts to separate the carbohydrate and protein indicated a covalent linkage. These authors suggested the existence of a glycopeptide having a 2-acetamido-2-deoxy-D-glucose-asparagine linkage [11,12]. Lopes et al. [13] found a large decrease in nitrogen and aggregate contents for welan gum when the welan broth was treated with a mild alkaline process before precipitation. Goycoolea et al. [14] observed a decrease in intrinsic viscosity of locust bean gum in alkali (1 M NaOH). These authors discussed the possibility that the reduction in [p] cannot be due solely to depolymerisation, and attributed these results to the dissociation of intermolecular interactions, in solution, but they did not discuss the effect of the presence of protein content in the samples. Commercial application of galactomannan is enhanced by its capacity to associate with xanthan [15]. The xanthan:galactomannan system is interesting because each polysaccharide alone does not form a gel but mixing results in a synergic gelation. For industrial exploitation it is important to know the conditions under which the interaction causes the gelation as well as the stability of the gel formed as characterized by Tg. The role of the galactose content of galactomannans on the interaction with xanthan has been previously recognised and is that the galactomannans which are less substituted show a greater synergistic effect [2,16]. However, even a highly substituted galactomannan from the seed of M. scabrella (M/G =1.1) interacted

with xanthan, as measured by viscometric analysis [10], in contradiction with some of the proposed molecular models [1619]. The inuence of the degree of galactose substitution of the galactomannan on Tg of mixtures and also the role of protein contents of the galactomannans samples on the Tm of xanthan is now described, as studied using viscoelastic measurements and calorimetry.

2. Materials and methods A commercial xanthan sample was used (Keltrol T, Kelco Division Merck) after purication by dissolution in NaCl 20 g l 1, centrifugation, ltration, precipitation, washing and drying as previously described [20]. The galactomannans were extracted from the isolated endosperm of S. parahybae (Vell.) Blake and from the whole seeds of M. scabrella Bentham [6]. The extractions were carried out in water at 25C for 4 h, which was then centrifuged at 2000 g for 20 min. The supernatants were precipitated, in the absence of NaCl, washed and dried as described above for the xanthan sample. The protein contents were determined by a modied Bradford method [21]. To test the role of proteins on the interaction, the galactomannan from M. scabrella was repuried as follows: the galactomannan was dissolved in 1 M NaCl at room temperature for 12 h. The solution was ltered successively through 8, 3, 1.2, 0.8 and 0.45 mm nitrate cellulose membranes. The ltrates were precipitated, washed and dried as for crude galactomannans. The xanthan (4 g l 1) and xanthan:galactomannan (X:G from 4:2 to 4:8 g l 1) solutions used in the physicochemical experiments were prepared at 5 or 10 mM NaCl and stirred for 14 h at room temperature. Under these conditions the Tm for xanthan is 43 and 51C, respectively [22]. Oscillatory rheological measurements were performed using cone and plate geometry (6 cm of diameter and 2 cone angle) on a Controlled Stress Rheometer Haake, RS 75. The temperature was controlled by a circulating water bath. Differential scanning calorimetry (DSC) measurements were performed using a Setaram Microcalorimeter Micro DSC III equipped with 1 cm3 batch vessels. The heating and cooling curves were obtained at a scan rate of 0.2C min 1. The experiments were carried out in triplicate.

3. Results and discussion It has previously been shown by calorimetric analysis that the presence of galactomannan of S. parahybae in mixtures with xanthan did not increase the temperature

T.M.B. Bresolin et al. / International Journal of Biological Macromolecules 26 (1999) 225231 Table 1 Analysis data of the galactomannansa Galactomannan source M. scabrella S. parahybae
a

227

Yieldb % (w/w) 17 25

M/G ratio 1.1 3.0

Protein % (w/w) 3.9; 2.4c 1.5

[h]D 20 +77 +32

[p] (ml g1) 740 1280

Bresolin et al. [10]. From whole seeds. c After purication.

b

Table 2 Calorimetric determination of xanthan Tm (C) in mixtures with different concentrations of galactomannansa Xanthan:galactomannan (g l1) S. parahybae 1.5% e 51b 52b 53b n.d. 53b 43.5c 44.3c n.d. n.d. n.d. M. scabrella 3.9% e 41d 43d 52d 54d n.d. 43.5c 47c 53c 55c n.d. M. scabrella 2.4%e 43.5c 44c n.d. n.d. n.d.

4:0 4:2 4:4 4:6 4:8

a

Heating or cooling rate: 0.2C min1; n.d. not determined. On heating in 10 mM NaCl. c On heating in 5 mM NaCl. d On cooling in 5 mM NaCl. e Percentage of protein.
b

of the conformational transition of xanthan [22]. This polysaccharide and another galactomannan obtained from the seeds of M. scabrella were now examined. Table 1 shows the characteristics of the galactomannan samples.

strength of the solutions and consequently the Tm of xanthan. The same effect with galactomannan has already been observed, but the authors [23,24] suggested that the modication in the Tm of xanthan was directly related to galactomannan and glucomannan concentrations.

3.1. Inuence of galactomannan and/or protein contents on the xanthan conformational transition
In order to observe the effect of galactomannan and/or protein concentrations on the xanthan conformational transition temperature (Tm) and its enthalpic transition value (DH), experiments with xanthan:galactomannan mixtures at different ratios (from 4:0 to 4:8 g l 1), and different degrees of purication were carried out with galactomannans from M. scabrella and S. parahybae (Table 2). It was observed that for galactomannans samples from S. parahybae and M. scabrella with the lowest protein content, the Tm of xanthan remains the same at the limit of the measurement accuracy. In contrast, the presence of M. scabrella galactomannan with 3.9% of proteins, increased signicantly the xanthan Tm. For example, at the highest concentrations of this sample in 5 mM NaCl, Tm values were reached similar to those obtained with S. parahybae galactomannan in 10 mM NaCl, showing a contribution of the proteins to the total ionic strength of the solution. Thus, the higher the content of proteins, which depends on the galactomannan purity and concentrations, the higher the ionic

3.2. Relation between xanthan conformation and the xanthan:galactomannan interaction

In Fig. 1 the DSC curves obtained on cooling and heating for xanthan:galactomannan mixtures are shown (M. scabrella, 3.9% protein content) at respective concentrations of 4 and 2 g l 1 in 5 mM NaCl. One can

Fig. 1. Differential scanning calorimetry (DSC) on cooling and heating of a xanthan:galactomannan from Mimosa scabrella (3.9% protein) mixture at 4:2 g l 1, in 5 mM NaCl.

228

T.M.B. Bresolin et al. / International Journal of Biological Macromolecules 26 (1999) 225231

Fig. 2. Differential scanning calorimetry (DSC) on cooling of a xanthan:galactomannan from Mimosa scabrella (3.9% protein) mixture, in 5 mM NaCl: (a) 4:2 g l 1; (b) 4:4 g l 1; (c) 4:6 g l 1.

observe a hysteresis in the conformational transition of xanthan as usually found at low ionic strength and xanthan concentration [25]. On heating, peak A which corresponds to the xanthan transition, is characterized by Tm = 47C and DH =4.9 J g 1 xanthan, in contrast with cooling when respective values of 43C and 1.2 J g 1 xanthan, were found. In the absence of galactomannan these were respectively 43.5C, 5 J g 1 and 41C, 1 J g 1. Thus, in the presence or in the absence of galactomannan, only approximately 20% of the ordered xanthan conformation is restored on cooling, in the Tm domain due to the kinetic process. The 80% of disordered xanthan is then available to interact with galactomannan. From a previous investigation [22], only the disordered chains of xanthan can interact with galactomannans with high galactose contents, at a temperature Tg, near to 20C, to form a complex in which the residual disordered xanthan chains become ordered. On heating

above the Tg, the complex is destroyed but as the temperature is below Tm, the entire xanthan chains are stabilised in their ordered conformation: at Tm, a DH value was again determined approximately the same as that of the enthalpic transition value of xanthan alone. In Fig. 2, the DSC curves obtained on cooling, are shown for different concentrations of galactomannan (M. scabrella, 3.9% protein content). One can observe a direct relation between the galactomannan concentration, not only on the enthalpic value of the B peaks, which refer to the galactomannan xanthan interactions, but also to the Tm of xanthan conformation and to the enthalpic value of the xanthan conformational transition (A peaks; Table 3). An increase of galactomannan concentration increases the protein content, and therefore the total ionic strength of the solution. Consequently, it stabilises the helical conformation of xanthan and increases the percentage of ordered xanthan chains. It is clear that the lower the absolute enthalpic values of the disorderorder xanthan transition, the lower is the percentage of ordered xanthan chains, and larger the absolute enthalpic values of the xanthangalactomannan interactions. Nevertheless, there is no quantitative relation between these values since is disordered xanthan chains can continue being ordered before reaching the Tg value. At the highest galactomannan concentrations of xanthan:galactomannan mixtures, the absolute value of the xanthan enthalpic peak (A) on cooling is not greatly different from the absolute value in the absence of galactomannan on heating (5 J g 1 xanthan) and consequently the peak of interaction (B) is very low. In order to distinguish the inuence of the galactomannan from that of the protein impurities, the same experiments were performed with M. scabrella which has a lower protein content. It is found (Table 3), for xanthan conformational transition on cooling, very low DH values, similar to those obtained in the absence of galactomannan, it follows that the enthalpic value of the interaction peak is larger compared with the mix-

Table 3 Characteristics of the xanthan (X), conformational transition (peak A*), and xanthan:galactomannans (X:G) interactions (peak B*) as a function of galactomannan concentration from Mimosa scabrella (3.9% protein) Galactomannan concentration (g l1) Tm (C) (peak A) 0 2 4 6
a

DH (J g1 xanthan) Xa (peak A) 19 0.5 1.29 0.3 190.5b 1.790.3 3.490.3

Tg (C) (peak B) 20 20b 21 21

DH (J g1 xanthan) Xa/G (peak B) 6.090.4 7.690.4b 290.3 0.79 0.3

41 43 41b 52 54

Xanthan concentration of 4 g l1 in 5 mM NaCl. Values obtained from M. scabrella galactomannan with 2.4% protein. * See Figs. 1 and 2. Results obtained from differential scanning calorimetry (DSC) measurements on cooling at 0.2C min1.

T.M.B. Bresolin et al. / International Journal of Biological Macromolecules 26 (1999) 225231 Table 4 Characteristics of gelation obtained from xanthan:galactomannan mixtures 4:2 (g l1)a Galactomannan M. scabrella (3.9% proteins) M. scabrella (2.4% proteins) S. parahybae (1.5% proteins) M/G 1.1 1.1 3 Tg (C) 20 20 24 DH (J g1 xanthan) 6.0 9 0.5 7.6 90.5 8.4 90.5 G% (Pa) at 10C 10 11 18 G%% (Pa) at 10C 2 2.2 3

229

a Tg is the gelling temperature, DH the enthalpic value corresponding to gelation; G% and G%%, respectively, the elastic and the loss modulus at 10C, after heating at 80C and cooling (1 Hz, 7% of strain).

ture containing galactomannan with 3.9% protein, which indicates a greater interaction. In Table 4 the values of G% and G%% are given for mixtures of xanthan and galactomannans from M. scabrella with different protein contents, and for galactomannan from S. parahybae. It can be seen that the decrease in the enthalpic value of the interaction peak of the sample with the higher protein content also corresponds to a proportional decrease in G%. This conrms the inuence of ionic impurities on xanthan conformation and especially, on the residual percentage of disordered chains left when the temperature, by cooling, reaches Tg, taking into account that for an M/G ratio lower then about 3 galactomannans interact preferentially with the disordered xanthan chains. Experiments with xanthan:galactomannan mixtures in 10 mM NaCl, at different ratios (from 4:0 to 4:8 g l 1) were carried out with the galactomannan from S. parahybae (1.5% protein). The temperature of Tm (Table 2) and the enthalpic value of the peak A (xanthan conformational transition) obtained on cooling is about the same, independent of the galactomannan concentration, 391 J g 1 xanthan, values which also correspond to that of xanthan alone in 10 mM NaCl. The corresponding enthalpic values of the synergistic interaction peaks (B) also remain constant (1 9 1 J g 1 xanthan). The higher salt concentration used for this sample and its lower protein content can explain these results, which are different from those obtained with M. scabrella galactomannan in 5 mM NaCl (Table 3). Nevertheless as the salt concentration is different it is not possible to attribute this variation only to the difference in protein content. One can observe that for xanthan galactomannan from M. scabrella (3.9% protein, ratio of 4:2 g l 1 in 5 mM NaCl), the maximum enthalpic value of the synergistic interaction peak is not appreciably different from that of the xanthan conformational transition obtained on heating, 8 and 5 J g 1 xanthan, respectively, and the gelling temperature Tg is signicantly lower than Tm (20C and 47C, respectively). Thus one cannot expect a modication of the conformation of the ordered xanthan chains to be adapted to that required in the heterotypic junction zones and to obtain, in this case,

synergistic interaction with low M/G galactomannan samples from ordered xanthan chains.

3.3. Inuence of M/G ratio on the xanthan:galactomannan interaction

As described in the literature [14,15], when M/G ratios are higher than :3, xanthan:galactomannan mixtures gel at : 60C. These interactions appear to occur either with the ordered or the disordered xanthan chains. The authors would like to emphasise, in this case, the inuence of the M/G ratio on the strength and stability of the gel obtained at temperatures lower than Tg. The results, from galactomannans M. scabrella (M/ G=1.1) and S. parahybae (M/G= 3), with a protein contents of 2.4 and 1.5%, respectively, are shown in Table 4 and Fig. 3. From these results, it appears that the M/G ratio has an inuence on the stability and the strength of the interaction between xanthan and galactomannan. The difference in the gelling temperature, Tg and in the G% modulus (Fig. 3) cannot be attributed to a difference in protein content (see also Table 3). The highest moduli obtained with the S. parahybae sample can be related to its high M/G ratio. In contrast to the M. scabrella sample an interaction with xanthan can be observed also to : 50C (Fig. 3). These interactions probably involve the xanthan chains according to the mechanism described in the literature for high M/G ratio. The increase in G% due to the second interaction at Tg = 24C (Fig. 3(c)) is equal to about 13 Pa being of the same order of magnitude as the increase obtained with the lower M/G ratio, from the M. scabrella sample (Fig. 3(b)), and in agreement with the enthalpic values observed for the two samples (Table 4). Thus the strength of the interaction at Tg seems to depend only on the fraction of disordered xanthan chains which can be assumed to be the same as that occurring with galactomannan samples with lower protein contents, and not to the M/G ratios of galactomannans (for M/G 53). For higher M/G ratio (: 5), under the same conditions, this interaction must becomes negligible, when compared with that observed at higher temperatures.

230

T.M.B. Bresolin et al. / International Journal of Biological Macromolecules 26 (1999) 225231

Fig. 3. G% variation on cooling at 1C min 1 at 1 Hz and 7% strain for xanthan:galactomannan mixtures of 4:2 g l 1, in 5 mM NaCl: (a) Mimosa scabrella, 3.9% protein; (b) Mimosa scabrella, 2.4% protein; (c) Schizolobium parahybae, 1.5% protein.

4. Conclusions The M/G ratio in galactomannans controls the mechanism and the temperature at which gelation occurs of xanthan galactomannans mixtures. For low galactose contents (M/G values higher than :3) interactions have been described at temperatures usually higher than 45C. The existence of these interactions does not depend on the ionic concentration of the solutions and seems to involve xanthan chains in their ordered as well as in their disordered conformations. For lower M/G values, interactions appear at lower temperatures but, in this case, only disordered xanthan chains can interact with galactomannan. As the percentage of disordered xanthan chains depends on the total ionic concentration, these interactions depend strongly on the external salt concentration and also on ionic impurities present in the solution or in galactomannan samples, such as like proteins. However, it is now shown that the strength of the xanthan:galactomannan interactions obtained in this case (M/G values lower than 3) does not depend greatly on the M/G ratio. In a previous investigation [26], the two mechanisms at high and low M/G values seem to be able to co-exist for a galactomannan from locust bean gum with M/G: 3.5, as discussed in Ref. [22]. A similar result has now been obtained for a galactomannan sample with M/G: 3. Recently Schorsch et al. [27] determined the viscoelastic properties of xanthan/guar gum mixtures with M/G ratios of 2 and 1.6. At 25C they observed an interaction which increased in the presence of salt in contrast to the

interaction of xanthan with the locust bean gum. This seems to indicate two different mechanisms of interaction. In the light of this observation and a previous one on xanthan/guar gum interactions [28] it cannot be excluded that an increase of salt content modies the mechanism of interaction from one involving disordered xanthan chains at low ionic strength to the more classic one, as usually found with higher M/G ratios. A determination of the thermal stability of these interactions, not performed in the work of Schorsch et al. [27] could resolve this point. It would also be interesting to analyse the inuence of the total ionic strength, xanthan:galactomannan and M/G ratios in order possibly to determine the preference of one or the other mechanism.

Acknowledgements This work was supported by CAPES, CNPq-UFPR, PRONEX-CARBOIDRATOS and CNRS (France).

References
[1] Stankowski JD, Mueller BE, Zeller SG. Carbohydr Res 1993;241:321. [2] Dea ICM, Morrison A. Adv Carbohydr Chem Biochem 1975;31:2412. [3] Dey PM. Adv Carbohydr Chem Biochem 1978;35:341. [4] Rechia-Vargas CG, Sierakowski MR, Ganter JLMS, Reicher F. Int J Biol Macromol 1995;17(6):409. [5] Reicher F, Leitner SCS, Sierakowski MR, Fontana JD, Correa JBC. Appl Biochem Biotechnol 1991;28/29:353. [6] Ganter JLMS, Heyraud A, Petkowicz CLO, Rinaudo M, Reicher F. Int J Biol Macromol 1995;17(1):13.

T.M.B. Bresolin et al. / International Journal of Biological Macromolecules 26 (1999) 225231 [7] Ganter JLMS, Zawadzki-Baggio SF, Leitner SC, Sierakowski MR, Reicher F. J Carbohydr Chem 1993;12:753. [8] Staub AM. Methods Carbohydr Chem 1965;5:5. [9] Bresolin TMB, Beltramini LM, Sander PC, Reicher F, Ganter JLMS. Latin Am Appl Res 1996;26:5. [10] Bresolin TMB, Sander PC, Reicher F, Sierakowski MR, Rinaudo M, Ganter JLMS. Carbohydr Polym 1997;33:131. [11] Mazzini MN, Cerezo AS. Anales Asoc Quim Argentina 1982;70:289. [12] Manzi AE, Mazzini MN, Cerezo AS. Carbohydr Res 1984;125:127. [13] Lopes L, Milas M, Rinaudo M. Int J Biol Macromol 1994;16:253. [14] Goycoolea FM, Morris ER, Gidley MJ. Carbohydr Polym 1995;27:69. [15] Williams PA, Phillips GO. In: Stephen AM, editor. Food Polysaccharides and their Application, vol. 14. New York: Marcel Dekker, 1995:463. [16] Morris ER, Rees DA, Young G, Walkinshaw MD, Darke A. J Mol Biol 1977;110:1. [17] Dea ICM, Morris ER, Rees DA, Welsh EJ, Barnes HA, Price J. Carbohydr Res 1977;57:249.

231

[18] Cairns P, Miles MJ, Morris VJ. Nature 1986;322:89. [19] Cheetham NWH, McCleary BV, Teng G, Lien F, Maryanto. Carbohydr Polym 1986;6:257. [20] Milas M, Rinaudo M. Carbohydr Res 1986;158:191. [21] Kresze GB. In: Bergmeyer J, GraL M, editors. Methods of Enzymatic Analysis, vol. III 92, 3rd edition. Weinheim: Verlag Chemie, 1983. [22] Bresolin TMB, Milas M, Rinaudo M, Ganter JLMS. Int J Biol Macromol 1998;23:263. [23] Williams PA, Clegg SM, Day DH, Phillips GO. In: Dickinson E, editor. Food Polym., Gels and Colloids, vol. 82. UK: RSC Special Publication, 1991:339. [24] Annable P, Williams PA, Nishinari K. Macromolecules 1994;27:4204. [25] Shatwell KP, Sutherland IW, Dea ICM, Ross-Murphy SB. Carbohydr Res 1990;206:87. [26] Goycoolea FM, Richardson RK, Morris ER, Gidley MJ. Macromolecules 1995;28:8308. [27] Schorsch C, Garnier C, Doublier JL. Carbohydr Polym 1997;34:165. [28] Lopes L, Andrade CT, Milas M, Rinaudo M. Carbohydr Polym 1992;17:121.