New Hypothesis on the Role of Alternating Sequences in Calcium

Biomacromolecules 2005, 6, 1031-1040

1031

New Hypothesis on the Role of Alternating Sequences in Calcium-Alginate Gels Ivan Donati,*,† Synnøve Holtan,† Yrr A. Mørch,† Massimiliano Borgogna,‡ Mariella Dentini,§ and Gudmund Skjåk-Bræk† Institute of Biotechnology, Norwegian University of Science and Technology, Sem Sælands V. 6/8, 7491 Trondheim, Norway, Department of Biochemistry, Biophysics and Macromolecular Chemistry, University of Trieste, Via Licio Giorgieri 1, 34127 Trieste, Italy, Department of Chemistry, University of Rome “La Sapienza”, Rome, Italy Received November 2, 2004; Revised Manuscript Received December 3, 2004

The availability of mannuronan and mannuronan C-5 epimerases allows the production of a strictly alternating mannuronate-guluronate (MG) polymer and the MG-enrichment of natural alginates, providing a powerful tool for the analysis of the role of such sequences in the calcium-alginate gel network. In view of the calcium binding properties of long alternating sequences revealed by circular dichroism studies which leads eventually to the formation of stable hydrogels, their direct involvement in the gel network is here suggested. In particular, 1H NMR results obtained from a mixed alginate sample containing three polymeric species, G blocks, M blocks, and MG blocks, without chemical linkages between the block structures, indicate for the first time the formation of mixed junctions between G and MG blocks. This is supported by the analysis of the Young’s modulus of hydrogels from natural and epimerized samples obtained at low calcium concentrations. Furthermore, the “zipping” of long alternating sequences in secondary MG/MG junctions is suggested to account for the shrinking (syneresis) of alginate gels in view of its dependence on the length of the MG blocks. As a consequence, a partial network collapse, macroscopically revealed by a decrease in the Young’s modulus, occurred as the calcium concentration in the gel was increased. The effect of such “secondary” junctions on the viscoelastic properties of alginate gels was evaluated measuring their creep compliance under uniaxial compression. The experimental curves, fitted by a model composed of a Maxwell and a Voigt element in series, revealed an increase in the frictional forces between network chains with increasing length of the alternating sequences. This suggests the presence of an ion mediated mechanism preventing the shear of the gel. Introduction Alginate is a collective term for a family of polysaccharides produced by brown algae1 and bacteria.2,3 Chemically they are linear copolymers of 1f4 linked β-D-mannuronic acid (M) and R-L-guluronic acid (G) arranged in a blockwise pattern along the chain with homopolymeric regions of M (M blocks) and G (G blocks) residues interspersed with regions of alternating structure (MG blocks). In its biosynthetic pathway, alginate is produced via a postpolymerization reaction involving a C-5 inversion on the M residues of mannuronan. This reaction is catalyzed by the mannuronan C-5 epimerases. Recently, it has been found that the genome of the alginate-producing bacterium Azotobacter Vinelandii encodes at least seven different mannuronan C-5 epimerases genes. These genes have been sequenced and cloned and expressed in Escherichia coli; the enzymes thus produced have been designed AlgE1-AlgE7.4,5 Since all natural alginates are produced from homopolymeric mannuronan by the same basic C-5 inversion from M to G, the remarkable * Corresponding Author. Tel: 0039 040 558 3692. Fax: 0039 040 558 3691. E-mail: [email protected]. † Norwegian University of Science and Technology. ‡ University of Trieste. § University of Rome “La Sapienza”.

variability in composition and sequence found in the polysaccharide is solely due to the different catalytic properties of the various epimerases. For example, the AlgE4 epimerase forms alginates with long strictly alternating sequences.6,7 The availability of these alginate-modifying enzymes and their use makes it possible to produce alginates with tailored structure and physical properties.8 The most relevant application of alginate in the biotechnology field is connected with its ability to form stable gels when in contact with solutions of divalent cations such as calcium. The ionotropic gelation of alginate has been described, based on X-ray fiber diffraction data from dehydrated specimens,9,10 using the so-called “egg-box” model.11,12 According to this model, chain-chain associations are induced by the presence of calcium and junction zones are formed between 2/1 helical chains of G sequences that present cavities suitable to accommodate divalent cations in a chelate type of binding. Circular dichroism and electron microscopy evidences describe junctions between G sequences as microcrystalline dimers12 or composed of very few laterally associated chains.13 Although the model originally proposed by Rees has been widely accepted and is regarded as basically correct, experimental evidence has prompted a new inspection of the

10.1021/bm049306e CCC: $30.25 © 2005 American Chemical Society Published on Web 01/25/2005

1032

Biomacromolecules, Vol. 6, No. 2, 2005

core structure of calcium alginate gels. Peculiar features triggered by the presence of an excess of the cross-linking cation have been explained by invoking lateral association of preformed junctions composed of long G sequences. In particular, a large lateral association between G junctions was reported to account for the syneresis of calcium-alginate gels14 (nevertheless, using computational methods Perez and co-workers15 recently pointed out the lower specificity of dimer-dimer associations compared to the junction forming chain-chain associations). In addition, it should be stressed that the “egg-box” model dramatically fails to account for the (albeit limited) gel forming properties of bacterial alginates lacking of long G blocks, such as those isolated from P. aeruginosa. In considering (the nature of) alginate gels, the alternating sequences and their effect within the network have been almost completely disregarded. Although the calcium binding ability of alternating sequences has first been suggested16 and then proved,12,17 the undoubtedly lower affinity toward the cation has confined MG blocks to a mere part of the elastically active chain in the gel network, without any direct involvement in the formation of the junction zones. In this respect the reduction of the stiffness of the polysaccharide chain, due to the intrinsically higher flexibility compared to GG and MM diads,18 represents the only role allocated to alternating sequences. Very recently, the production of new tailor-made alginates has been prompted by the availability of C-5 epimerases, which allow an extremely efficient tuning of both composition and physicochemical properties of the polysaccharide. In particular, the epimerase AlgE4 which enables the conversion of M blocks into alternating sequences in a processive mode of action,19 has provided new alginates with interesting properties. In this respect, it is to be mentioned that, besides the remarkable increase in syneresis displayed by the AlgE4treated samples, a much higher stability of the gel is directly correlated with the presence of long alternating sequences.7,20 In the present paper, a new hypothesis on the sequences involved in junction formation in alginate gels is proposed. Based on the evidence of gel formation by a polymer composed of strictly alternating sequences, a new insight into the role of MG blocks in the alginate gel network has been attempted. Experimental data from alginates enzymatically enriched in alternating sequences allowed the proposal of the direct involvement of MG blocks in the formation of both mixed GG/MG and MG/MG junctions. The effect of the presence of secondary MG/MG junctions on the viscoelastic properties of the gels was investigated by measuring their creep compliance, which suggested the presence of an ion mediated mechanism preventing slipping within the network of the gel. Therefore, the analysis of the effect of the epimerase AlgE4 on different alginate samples has introduced the idea that G blocks are not the only sequences involved in junction formation. Experimental Section Materials. Sodium alginate samples isolated from Laminaria hyperborea stipe (L. hyp.), Ascophyllum nodosum (A. nod.), and Macrocystis pyrifera (M. pyr.) were provided by

Donati et al. Table 1. Chemical Composition and Intrinsic Viscosity of PolyME6, Polyalternating, and Mannuronan sample PolyME6a polyalternatinga mannuronan

FG

FM

FGG

FGM,MG

FMM

[η] (dL/g)b

0.84 0.45 0

0.16 0.55 1

0.77 0 0

0.07 0.45 0

0.09 0.1 1

9.86 ( 0.03 6.95 ( 0.02 9.25 ( 0.03

a PolyME6 and polyalternating have been prepared from mannuronan using epimerases AlgE6 and AlgE4, respectively, following the conditions reported elsewhere.33 b Solvent: NaCl 0.1 M, 20 °C. Detail on the definition of FG, FM, FGG, FGM,MG, and FMM are reported in caption to Table 2.

FMC Biopolymer (Norway). High molecular weight mannuronan (fraction of guluronic residues, FG, a r Figure 8. 400 MHz - 1H NMR (anomeric region) of M. pyr. and A. nod.

concentration) of hydrogels obtained under these conditions can be related to the presence of long alternating blocks rather than to their overall (diad) content. To prove the correlation between syneresis and length of the MG blocks, alginate samples from L. hyp., M. pyr., and A. nod. were treated with the epimerase AlgE4 yielding the samples LhypE4, MpyrE4, and AnodE4 (Table 2). Due to its processive mode of action,19 this enzyme leads to an elongation of alternating sequences as demonstrated by the increase in the ratio FMGM/FGGM (see Table 2) for the treated samples LhypE4, MpyrE4, and AnodE4 with respect to the native samples. These three epimerized alginates have then been used to prepare homogeneous calcium gels cylinders and the syneresis was measured in the same calcium concentration range as the native alginate samples, i.e. L. hyp., M. pyr., and A. nod. It is important to underline that, in all of the cases analyzed, the syneresis-calcium concentration gradient was increased by the epimerization (Figure 7a-c), displaying values of 1.8, 3.4, and 2.8 for LhypE4, MpyrE4, and AnodE4, respectively. The latter result supports the conclusion that long alternating sequences are responsible for the deswelling of the gel via a mechanism that is likely to involve their calcium induced “zipping” to form secondary junctions (MG/MG junctions) similar to those obtained in the case of pure polyalternating hydrogels (see Figure 3b). An additional proof of the direct involvement of alternating sequences in the syneresis of calcium alginate gels is provided by the analysis of the behavior of a galactosesubstituted alginate from L. hyp. stipe.31,32 The introduction of the bulky side chain moieties on the polysaccharide chain, largely affecting the G residues in alternating sequences, hinders their association preventing the formation of secondary MG/MG junctions. As a consequence, no syneresis is displayed by gel cylinders obtained from such galactose-Gsubstituted alginate in the same calcium concentration range.33 The involvement of long alternating sequences in the formation of secondary MG/MG junctions leads to an additional consideration on the effect on the mechanical properties of

(1)

with β)

( ) 6 lkBT

1/2

where l is the step length of a Gaussian chain connected to a source from which chain may be drawn at a cost of  per unit length. From a qualitative point of view, the free energy (G1) of the network configuration characterized by a grater span of the chains compared to the junctions (r > a) can be written as (eq 2) G1 ) (2Φ - 1)G(r)

(2)

where Φ represents the functionality of the junction and G(r) is the free energy of each Gaussian chain. The total free energy of the network can be lowered, under certain conditions, allowing the collapse of some chains connecting the junctions, obtaining the free energy G2 (eq 3) 2Φ - 2

G2 ) G(a) +

G(r′) ∑ i)1

(3)

where r′ is the length of the ith chain stretched after the collapse. As depicted in Figure 6 of ref 34, the feasibility of the collapse depends on the ratio r/a. In particular, the decrease in the ratio r/a could bring about a situation in which G2 < G1 causing a collapsed state with a reduction in the number of cross-links. We focused on this latter aspect of the reel-chain-rod model to study the dependence of the Young’s modulus on the calcium concentration for gel cylinders obtained from M. pyr. and A. nod. and from their AlgE4 epimerised products (Figure 9a,b). In the case of M. pyr. (Figure 9a), a clearly detectable decrease in the Young’s modulus (∆E ∼ 0.9) is experienced upon increasing the amount of calcium carbonate from 20 to 25 mM. The same behavior was shown by sample MpyrE4 (which is characterized by an AlgE4catalyzed elongation of the alternating sequences) that showed a drop in the Young’s modulus (∆E ∼ 0.9) between 17.5 and 20 mM of calcium carbonate. On the other hand, due to the limited length of the existing alternating sequences in the native sample, the formation of long secondary

1038

Biomacromolecules, Vol. 6, No. 2, 2005

Donati et al.

Figure 10. Schematic representation of a possible partial collapse of alginate gel network upon calcium induced “zipping” of long alternating blocks.

Figure 11. Example of a creep compliance curve for an alginate gel.

Figure 9. Young’s modulus-calcium concentration dependence for (a) M. pyr. (squares) and MpyrE4 (circles), and (b) A. nod. (squares) and AnodE4 (circles). Values are reported as mean ( s.d. (n ) 8).

MG/MG junctions and the collapse of the gel network are prevented in the case of A. nod. (Figure 9b). However, the enzymatic treatment of the native alginate, i.e., in sample AnodE4, introduces sufficiently long alternating sequences to enhance the binding of calcium in MG/MG junctions and therefore to induce the energetically driven collapse of the gel network, as demonstrated by the decrease in the Young’s modulus (∆E ∼ 0.9) between 17.5 and 20 mM of calcium carbonate (Figure 9b). A schematic (and oversimplified) representation of this process, re-drawn from ref 34 and adapted to the case of alginate network, is shown in Figure 10. L. hyp. and LhypE4 samples (not reported) can be considered as a negative control where the limited length of the MG blocks prevents the collapse of the network. The effect of the formation of secondary MG/MG junctions on the viscoelastic properties of the calcium-alginate gels was studied with time-domain rheological experiments. In particular, creep compliance for gel cylinders obtained from samples listed in Table 2 was evaluated by applying an “instantaneous” stress and measuring the time dependent increase of the strain. The applied stress was selected in order to induce an initial 8% strain under uniaxial compression. Such a strain was found to be within the linear viscoelastic

region for all of the samples analyzed. The creep compliance J ()/σ) was recorded for approximately 30 min before removing the load. During the experiment, no loss of water from the specimen was detected. Figure 11 shows a typical creep curve obtained from alginate gels, showing the viscoelastic behavior of this material. The creep compliance was analyzed by means of a model composed of a Maxwell element in series with one Voigt element (eq 4), which accurately fitted all experimental curves J(t) ) J0 + J1(1 - e-t/τ) +

t ηN

(4)

where J(t) is the measured compliance, J0 and J1 are the compliances of the Maxwell and Voigt springs respectively, ηN is the so-called Newtonian viscosity of the Maxwell dashpot, and τ is the retardation time associated with the Voigt element. The results obtained for the different samples are reported in Table 4, and although deserving more extended investigations, some conclusions can be drawn. It is noteworthy that the trend exhibited by the instantaneous spring component E0 ()1/J0) parallels that obtained for the Young’s modulus for all of the samples analyzed (Table 4), supporting the use of this calculation. Moreover, from the present data, it can be concluded that high M samples lead to the formation of highly elastic gels and that their AlgE4-treatment increases the rigidity.35

Biomacromolecules, Vol. 6, No. 2, 2005 1039

Alternating Sequences in Calcium-Alginate Gels Table 4. Results from the Analysis of the Creep Compliance Curves by Means of Equation 4a sample

E (kPa)b

J0(Pa-1) × 105

J1(Pa-1) × 105

τ (s)

ηN (×107 Pa s)c

L. hyp. LhypE4 M. pyr. MpyrE4 A. nod. AnodE4

8.8 9.1 6.0 8.6 2.9 8.1

8.5 8.2 13.5 9.7 24.9 10.4

1.9 1.6 1.5 0.6 2.6 0.7

192 188 148 138 141 120

4.7 5.4 4.7 5.9 3.1 6.1

a Gel formation: polymer solution: 1%, CaCO : 17.5 mM; GDL: 35 3 mM. b Young’s modulus measured for the different samples. c Error associated to the value obtained from the fitting (0.1 (n ) 3).

A further comment can be added from the analysis of the Newtonian viscosities ηN reported in Table 4. Bearing in mind that the value of ηN is influenced by both the concentration of the polymer in the gel and the cross-link density, the frictional forces between the chains connecting the junctions in the gel can account for the resistance of the network toward slipping.36 It follows that the higher the interchain interactions, the higher the Newtonian viscosity of the hydrogel. In this respect, it should be noted that the AlgE4 epimerization of natural samples leads to hydrogels characterized by a higher value of ηN in all of the cases analyzed. Although the applied stress has been corrected taking into account the increase in the polymer concentration in the gel due to syneresis, no similar correction can be performed to rule out the dependence of ηN on the cross-link density. Therefore, even if the higher variations of ηN are detected for MpyrE4 and AnodE4, with respect to the gels obtained from the native alginates, the contribution of the AlgE4 treatment cannot be clearly discerned from the present data due to the concomitant increase of the cross-link density. However, in the case of samples L. hyp. and LhypE4, the high similarity of the J0 value allows a direct comparison of ηN (Table 4). By showing a higher value of the Newtonian viscosity, the gel obtained from LhypE4 can be described as more compliant than the one from the natural sample (Figure 12a). Considering that the treatment of the L. hyp. sample with AlgE4 brings about only an elongation of MG blocks, as revealed by the increase in the ratio FMGM/FGGM, one could safely conclude that the higher Newtonian viscosity stems from the formation of extended secondary MG/MG junctions that, by bridging alternating sequences in the gel network via an ion mediated mechanism, enhance the frictional forces between polymer chains reducing their slipping. To support the latter conclusion, it is important to underline that no difference in the value of ηN between L. hyp. and LhypE4 can be detected when such “zipping” process is avoided using a lower calcium concentration (13.3 mM; Figure 12b). The effect of the formation of extended secondary MG/MG junctions on the Newtonian viscosity is also revealed from the comparison between the gels obtained from L. hyp., A. nod., and M. pyr. In particular, the latter one, characterized by the longest alternating sequences among the nonepimerized samples, displays a remarkably high value of ηN. Considering the lower cross-link density of the gel obtained from M. pyr. compared to that from L.

Figure 12. Creep compliance curves for hydrogels from L. hyp. and LhypE4 obtained by using (a) 17.5 mM and (b) 13.3 mM of CaCO3.

hyp., it is reasonable to assume a notable contribution to ηN stemming from the ion mediated “zipping” process on long MG blocks. Conclusions The availability of the AlgE4 epimerase, enabling the production of new alginates enriched in alternating sequences, represents a powerful tool to identify the role of such sequences in calcium-alginate gel networks. The calcium-binding ability, resulting eventually in gel formation of a polymer composed of strictly alternating sequences has prompted a new insight on the involvement of long MG blocks in the gel network. In particular, the formation of mixed GG/MG junctions, with the macroscopic outcome of a remarkable increase in the Young’s modulus for epimerised samples, has been proposed for epimerized alginate samples characterized by the presence of long alternating sequences. By parallelism with the formation of junctions in polyalternating sample, the formation of pure MG/MG junctions via an ion mediated “zipping” mechanism has been suggested that accounts for the dependence of syneresis on the length of alternating sequences and for its impairment due to the introduction of bulky groups. The formation of extended secondary MG/MG junctions, inducing a partial collapse of the gel network, brings about only a notable decrease in the storage modulus, as expected for a biopolymer gel network according to the theory developed by Higgs and Ball.34 Finally, time-domain rheological experiments were used to study the relationship between the formation of secondary MG/MG junctions and the increase in frictional forces between chains connecting the junctions in the gel network.

1040

Biomacromolecules, Vol. 6, No. 2, 2005

In conclusion, the results reported in the present paper assign a new role to the MGM sequences introducing the possibility of their direct involvement in alginate gel network. Acknowledgment. This work has been financed by the Norwegian Research Council. The authors thank Prof. Olav Smidsrød and Prof. Sergio Paoletti for helpful discussions. References and Notes (1) Painter, J. In The polysaccharides; Aspinall, G. O., Ed.; Academic Press: New York, 1983; p 195. (2) Gorin, P. A. J.; Spencer, J. F. T. Can. J. Chem. 1966, 44, 993. (3) Govan, J. R. W.; Fyfe, J. A. M.; Jarman, T. R. J. Gen. Microbiol. 1981, 125, 217. (4) Ertesvåg, H.; Doset, B.; Larsen, B.; Skjåk-Bræk, G.; Valla, S. J. Bacteriol. 1994, 176, 2846. (5) Svanem, B. I. G.; Skjåk-Bræk, G.; Ertesvåg, H.; Valla, S. J. Bacteriol. 1999, 181, 68. (6) Høidal, H. K.; Ertesvåg, H.; Skjåk-Bræk, G.; Stokke, B. T.; Valla, S. J. Biol. Chem. 1999, 274, 12316. (7) Strand, B. L.; Mørch, Y. A.; Syvertsen, K. R.; Espevik, T.; SkjåkBræk, G. J. Biomed. Mater. Res. 2003, 64A, 540. (8) Draget, K. I.; Strand, B. L.; Hartmann, M.; Valla, S.; Smidsrød, O.; Skjåk-Bræk, G. Int. J. Biol. Macromol. 2000, 27, 117. (9) Atkins, E. D. T.; Nieduszynski, I. A.; Mackie, W.; Parker, K. D.; Smolko, E. E. Biopolymers 1973, 12, 1865. (10) Atkins, E. D. T.; Nieduszynski, I. A.; Mackie, W.; Parker, K. D.; Smolko, E. E. Biopolymers 1973, 12, 1879. (11) Grant, G. T.; Morris, E. R.; Rees, D. A.; Smith, P. J. C.; Thom, D. FEBS Lett. 1973, 32, 195. (12) Morris, E. R.; Rees, D. A.; Thom, D.; Boyd, J. Carbohydr. Res. 1978, 66, 145. (13) Smidsrød, O. J. Chem. Soc., Faraday Trans. 1 1974, 57, 263. (14) Stokke, B. T.; Draget, K. I.; Smidsrød, O.; Yuguchi, Y.; Urakawa, H.; Kajiwara, K. Macromolecules 2000, 33, 1853. (15) Braccini, I.; Perez, S. Biomacromolecules 2001, 2, 1089. (16) Smidsrød, O.; Haug, A. Acta Chem. Scand. 1972, 26, 79.

Donati et al. (17) Wang, Z.-Y.; Zhang, Q.-Z.; Konno, M.; Saito, S. Biopolymers 1993, 33, 703. (18) Smidsrød, O.; Glover, R. M.; Whittington, S. G. Carbohydr. Res. 1973, 27, 107. (19) Campa, C.; Holtan, S.; Nilsen, N.; Bjerkan, T. M.; Stokke, B. T.; Skjåk-Bræk, G. Biochem. J. 2004, 381, 155. (20) Mørch, Y. A.; Skjåk-Bræk, G. Personal communication. (21) Gimmestad, M.; Sletta, H.; Ertesvåg, H.; Bakkevig, K.; Jain, S.; Suh, S.; Skjåk-Bræk, G.; Ellingsen, T. E.; Ohman, D. E.; Valla, S. J. Bacteriol. 2003, 185, 3515. (22) Grasdalen, H.; Larsen, B.; Smidsrød, O. Carbohydr. Res. 1979, 68, 23. (23) Smidsrød, O.; Haug, A. Acta Chem. Scand. 1972, 26, 79. (24) Martinsen, A.; Skjåk-Bræk, G.; Smidsrød, O. Biotech. Bioeng. 1989, 33, 79-89. (25) Lattner, D.; Flemming, H.-C.; Mayer, C. Int. J. Biol. Macromol. 2003, 33, 81. (26) Emmerichs, N.; Wingender, J.; Flemming, H.-C.; Mayer, C. Int. J. Biol. Macromol. 2004, 34, 73. (27) Flory, P. J. Proc. R. Soc. London A 1976, A 351, 351. (28) Draget, K. I.; Gåserød, O.; Aune, I.; Andersen, P. O.; Storbakken, B.; Stokke, B. T.; Smidsrød, O. Food Hydrocolloid 2001, 15, 485. (29) Grasdalen, H. Carbohydr. Res. 1983, 118, 255. (30) Grasdalen, H.; Larsen, B.; Smidsrød, O. Carbohydr. Res. 1981, 89, 179. (31) Donati, I.; Vetere, A.; Gamini, A.; Skjåk-Bræk, G.; Coslovi, A.; Campa, C.; Paoletti, S. Biomacromolecules 2003, 4, 624. (32) Donati, I.; Coslovi, A.; Gamini, A.; Skjåk-Bræk, G.; Vetere, A.; Campa, C.; Paoletti, S. Biomacromolecules 2004, 5, 186. (33) Donati, I.; Draget, K. I.; Borgogna, M.; Paoletti, S.; Skjåk-Bræk, G. Biomacromolecules 2005, 6, 88. (34) Higgs, P. G.; Ball, R. C. Macromolecules 1989, 22, 2432. (35) Ainsworth, P. A.; Blanshard, J. M. V. J. Texture Stud. 1980, 11, 149. (36) Mitchell, J. R.; Blanshard, J. M. V. J. Texture Stud. 1976, 7, 219. (37) FG + FM ) 1. FGG + FMM + FGM + FMG ) 1, with FGM ) FMG. FG ) FGGG + FMGM + FGGM + FMGG, with FGGM ) FMGG.

BM049306E