- Email: [email protected]
The supramolecular structure of LPS–chitosan complexes of varied composition in relation to their biological activity
Carbohydrate Polymers 123 (2015) 115–121
Contents lists available at ScienceDirect
Carbohydrate Polymers journal homepage: www.elsevier.com/locate/carbpol
The supramolecular structure of LPS–chitosan complexes of varied composition in relation to their biological activity V.N. Davydova a,∗ , A.V. Volod’ko a , E.V. Sokolova a , E.A. Chusovitin b , S.A. Balagan b , V.I. Gorbach a , N.G. Galkin b,c , I.M. Yermak a , T.F. Solov’eva a a G.B. Elyakov Pacific Institute of Bioorganic Chemistry, Far Eastern Branch of Russian Academy of Sciences, Prospect 100 let Vladivostoku 159, Vladivostok 690022, Russia b Institute of Automation and Control Processes, Far Eastern Branch of Russian Academy of Sciences, Radio Str. 5, Vladivostok 690041, Russia c Far Eastern Federal University, Sukhanova Str. 8, Vladivostok 690091, Russia
a r t i c l e
i n f o
Article history: Received 24 September 2014 Received in revised form 13 January 2015 Accepted 15 January 2015 Available online 21 January 2015 Keywords: Lipopolysaccharide (LPS) Chitosan LPS–chitosan complexes -Potential Supramolecular structure
a b s t r a c t The complexes of chitosan (Ch) with lipopolysaccharides (LPSs) from Escherichia coli O55:B5 (E-LPS) and Yersinia pseudotuberculosis 1B 598 (Y-LPS) of various weight compositions were investigated using quasielastic light scattering, -potential distribution assay and atomic force microscopy. The alteration of potential of E-LPS–Ch complexes from negative to positive values depending on Ch content was detected. The Y-LPS–Ch complexes had similar positive -potentials regardless of Ch content. The transformation of the supramolecular structure of E-LPS after binding with to Ch was revealed. Screening of E-LPS and Y-LPS particles by Ch in the complexes with high polycation was detected. The ability of LPS–Ch complex to induce biosynthesis of TNF-␣ and reactive oxygen species in stimulated human mononuclear cells was studied. A significant decrease in activity complexes compared to that of the initial LPS was observed only for E-LPS–Ch complexes. © 2015 Elsevier Ltd. All rights reserved.
Chemical compounds studied in this article: Lipopolysaccharide (PubChem CID: 11970143) Chitosan (PubChem CID: 71853)
1. Introduction Lipopolysaccharide (LPS or endotoxin) is one of the strongest elicitors of the immune system of macroorganisms. It is responsible for induction of many cytokines and chemokines but also the other immune mediators in human immune cells, such as monocytes, macrophages, dendritic cells, and others (Ulevitch, 2000). Activation of these cells may be beneficial at low amounts of LPS but at higher LPS concentrations lead to pathophysiological reactions such as sepsis (Bone, 1996). LPS consists of the hydrophobic part, known as lipid A, the repeating O-antigen polysaccharide, and the core oligosaccharide (Raetz, 1990). Many of the immune activation abilities of LPS can be attributed to the lipid A unit, which binds to the tolllike receptor 4 (TLR-4) and activates the host defense effector
∗ Corresponding author. Tel.: +7 423 2310719; fax: +7 423 2314050. E-mail address: [email protected] (V.N. Davydova). http://dx.doi.org/10.1016/j.carbpol.2015.01.028 0144-8617/© 2015 Elsevier Ltd. All rights reserved.
system by rapidly triggering proinflammatory processes (Peri, Piazza, Calabrese, Damore, & Cighetti, 2010). Therapies for Gram-negative sepsis remain unsatisfactory despite the concerted effort to develop new treatments for this life-threatening syndrome. Currently, no drugs are specific for endotoxin-induced clinical syndromes. Hence, compounds with therapeutic potentials as anti-endotoxin agents may include those that either bind LPS at high affinity and neutralize its toxicity or those that competitively interact with LPS receptors on host cells without activating biological responses (receptor antagonists). Both groups of agents, acting through different mechanisms, block endotoxin binding to specific receptors on target cells, thus preventing the synthesis of proinflammatory cytokines (Van Amersfoort, Van Berkel, & Kuiper, 2003). One of the promising ways to inhibit the harmful inflammatory/septic effects of endotoxin is to bind LPS to certain polycations, which can interact with lipid A bearing the negatively charged phosphate groups and, as a result, block this toxic centre of the endotoxin.
116
V.N. Davydova et al. / Carbohydrate Polymers 123 (2015) 115–121
Chitosan, the cationic (1-4)-2-amino-2-deoxy--d-glucan, with degree of acetylation typically close to 0.20, is prevalently produced from marine chitin. Chitosan and its partially depolymerized derivatives and oligomers have a wide spectrum of biological activities and are particularly useful in the fields of wound healing, oral delivery, and food industry (Aider, 2010; Muzzarelli, 2009, 2010). The systematic studies of chitosan binding to LPS and the biological properties of the chitosan–LPS complexes showed that it interacts specifically with LPS to form stable water-soluble complexes of various stoichiometric compositions and that the formation of LPS–chitosan complexes is accompanied by significant modification of immunological properties of LPS (Solov’eva, Davydova, Krasikova, & Yermak, 2013). Due to the amphiphilic nature of their molecules and the very low critical micellar concentration (CMC), LPS exist as aggregates in aqueous solutions and their aggregate state has been found to be an important factor of interaction between LPS and Ch (Davydova, Naberezhnykh, Yermak, Gorbach, & Solov’eva, 2006). As shown previously, the LPS–chitosan complexes are 10–20 times less toxic than LPS alone (Yermak et al., 2006). This effect depends on LPS structure, the ratio between the components in the complex and chitosan molecular weight. A substantial reduction of LPS toxicity in LPS–chitosan complexes may be explained by the blocking of the toxophoric centre of endotoxin or the alterations in the molecular charge and/or the structure of LPS aggregates by chitosan (Solov’eva et al., 2013). The development of modern polymers analysis techniques offers great opportunities for studying the supramolecular organization of macromolecules. The atomic force microscopy can obtain a high resolution image of the biomolecules under nearphysiological conditions. It allows us to study complexes of different stoichiometry between endotoxins of different structure and chitosan and to try to find a correlation between their structure and biological activity. For this purpose two different structural types of LPS (commercial Escherichia coli LPS O55 (E-LPS) and LPS isolated from bacteria Yersinia pseudotuberculosis 1B598 (Y-LPS)) and Ch with molecular weight of 110 kDa and N-acetylation degree of 1% have been chosen to prepare their complexes with different stoichiometry.
2. Materials and methods
calculated according to IR-spectroscopy data (Domszy & Roberts, 1985). The molecular weight of the chitosan was determined by viscosimetry in 0.1 M AcOH/0.2 M NaCl according to (Harding, 1997). 2.2. Preparation of LPS–Ch complexes LPS (1 mg) and Ch (7–1.4 mg) (according to its contents in the complexes) were dissolved separately in 5 ml of pyrogen-free deionized water. The solutions were stored for 48 h at 37 ◦ C; equal aliquots of LPS and Ch were then mixed and incubated for 18 h at 37 ◦ C. The LPS and chitosan solutions were decontaminated of bacteria by filtration using a “Millex GS” filter unit (Millipore, Ireland). 2.3. Dynamic light scattering (DLS) and electrophoretic properties of the LPS–chitosan complexes The sizes and -potentials of the LPS aggregates and LPS–Ch complexes in solution were determined using a ZetaSizer Nano ZS system (Malvern, UK) operating at 633 nm. Prior to measurements, the samples were left for 1 h to allow the large aggregates to settle, as they can interfere with the measurements even if their content does not exceed a few percent. The measurements were performed at 25 ◦ C for E-LPS and at 37 ◦ C for Y-LPS. The hydrodynamic diameters of the particles were automatically calculated with the instrument’s software based on analysis of the autocorrelation function. The -potentials were calculated from the experimentally determined electrophoretic mobility using the Henry equation (Henry, 1931). 2.4. Atomic force microscopy (AFM) LPSs were dissolved in distilled de-ionized water at concentration of 0.1 mg/ml; Ch samples—at concentrations of 0.01, 0.1 or 0.5 mg/ml. The LPS–Ch complexes were prepared at the same concentrations. Aliquots (12 l) of the aqueous solutions of each complex and their initial component were deposited onto freshly cleaved mica and dried at 37 ◦ C for 24 h or at 70 ◦ C for 30 s (for LPS). The morphology of Ch, LPS and their complexes was studied in air by AFM (Solver P47) in the tapping contact mode using a tip with the radius of 10 nm. The topographic parameters of the macromolecular structure have been automatically calculated using the self-developed Balagan’s Grain Analysis programme.
2.1. Isolation of lipopolysaccharides and chitosan 2.5. Ethical approval Cells of Y. pseudotuberculosis (serovar 1B, strain 598) isolated from a patient suffering from the far-eastern scarlatina-like fever (Institute of Epidemiology and Microbiology, Vladivostok, Russia) were grown at 4 ◦ C in a previously described nutrient medium (Ovodov et al., 1971). LPS from bacterial cells was isolated using the phenol-chloroform-petroleum ether procedure (Galanos, Luderitz, & Westphal, 1969). Nucleic acids were removed by ultracentrifugation at 105,000 × g. The purified LPS (yield: 1.2%) contained less than 1% of protein. Protein content was determined by Lowry’s method (Lowry, Rosebrough, Farr, & Randall, 1951), nucleic acid contents were estimated in accordance with Spirin et al. (1958). Structural characterization of Yersisnia pseudotuberculosis LPS elucidated previously (Krasikova, Gorbach, Solov’eva, & Ovodov, 1978; Tomshich, Gorshkova, Elkin, & Ovodov Yu, 1976; Tomshich, Gorshkova, & Ovodov, 1985). A preparation of LPS from E. coli 055:B5 was purchased from Sigma (Sigma Chemicals, St. Louis, MO, USA). A chitosan (Ch) sample with molecular weight of 110 kDa and 1% degree of N-acetylation was obtained by alkaline treatment of crab chitin according to the published protocol (Wolfrom & Shen Han, 1959). The degree of N-acetylation of the chitosan sample was
The study protocol was approved by the medical ethics committee of the local hospital (Vladivostok, Russian Federation). Informed consent was obtained from all subjects who participated in the study. All donors were free of medicines administration for 14 days prior to blood sampling. Blood was drawn from the antecubital vein of normal healthy human volunteers and anticoagulated in plastic tubes (Greiner Bio-One International AG, Kremsmuenster, Austria) with 30 IU lithium heparinate used as an anticoagulant. 2.6. TNF-inducing activity Human peripheral blood was collected by venipuncture into sterile siliconized tubes and diluted at a ratio of 1:5 with sterile Medium 199 (Sigma, USA) containing 300 mg/l of glutamine (Gibco, Life Technology, Germany) and 50 g/ml of gentamicin. The diluted blood (0.1 ml) was transferred into sterile polypropylene plates and then incubated with the corresponding LPS, Ch, or LPS–Ch complex (37 ◦ C, 5% CO2 ). Control incubation with 1 g/ml of LPS from E. coli (strain 055:B5) was performed for each experiment. After 24 h the supernatants were collected and frozen, followed
V.N. Davydova et al. / Carbohydrate Polymers 123 (2015) 115–121
by cytokine determination using specific ELISA (DuoSet developing system, Genzyme, Boston, USA). 2.7. Leukocytes We rapidly separated leukocytes from venous blood by lyzing erythrocytes in a solution containing 0.15 M NH4 Cl, 10 mM NaHCO3 and 0.1 mM EDTA for 10 min (Lehmann, Sørnes, & Halstensen, 2000). The leucocyte fraction was washed twice, resuspended in PBS, and adjusted to a concentration of ≈107 non-lymphocytes (polymorphonuclear leukocytes and monocytes) per millilitre. 2.8. Detection of production of reactive oxygen species Production of reactive oxygen species was assayed by flow cytometry using APF as described (Setsukinai, Urano, Kakinuma, Majima, & Nagano, 2003). Cells were plated at 100,000 cells/well in a 96-well tissue culture plate (Costar, Cambridge, MA) and stimulated for 1 h with samples (50 and 100 g/ml, final value); APF (10 M) was added to the media for the final 30 min. The cells were lifted, kept on ice (10 min), and analyzed immediately by flow cytometry. Flow cytometric measurements were performed using a BD FACSCalibur four-colour analyzer (Becton Dickinson). Forward and side scatter light was used to identify cell populations and to measure cell size and granularity. 2.9. Statistical analysis All data were represented as mean ± standard deviation. Statistical analysis was done by one-way ANOVA. Differences were considered to be statistically significant if p < 0.05. 3. Results and discussion 3.1. Size and -potential of the LPS–chitosan complexes Several LPS–Ch complexes of varying compositions obtained at E. coli O55:B5 LPS (E-LPS) and Y. pseudotuberculosis 598 LPS (YLPS) concentrations of 0.1 mg/ml were prepared. The weight ratios between the components in the complexes were chosen according to the previously obtained results (Davydova, Yermak, Gorbach, Krasikova, & Solov’eva, 2000; Yermak et al., 2006). The study of the complexes by DLS and electrokinetic methods showed that their sizes and surface potentials depended on the endotoxin structure and composition of the complexes (Table 1). It is known that LPS as amphiphilic molecules tend to form supramolecular aggregates in aqueous solutions at concentrations above the critical micellar concentration. The length of the hydrophilic carbohydrate moiety of a LPS molecule affects size and structure of its aggregates (Jucker, Harms, & Zehnder, 1998). ELPS possessing the long O-specific chains and hexa-acylated lipid A formed negatively charged aggregates homogeneous in size. According to the results E-LPS formed the negatively charged homogeneous in size aggregates in the aqueous solution. Rather homogeneous complexes with varied surface potential changing from negative to positive with increase in content of Ch in complex were also registered after binding LPS with Ch. The sizes of the complexes with weight ratios of E-LPS:Ch 7:1, 1:1, 1:7 were close to that of LPS aggregates and increased slightly with raising Ch content in complexes. In the case of E-LPS–Ch 4:1 w/w complex we detected the formation of two types of particles with much larger sizes (3110 and 658 nm) compared to those of other complexes and initial E-LPS aggregates and the low negative value of the -potential. These large particles were probably formed by aggregation of smaller ones due to a dramatic decrease in an electrostatic
117
repulsion between them because the negative LPS charge was neutralized by the positive charge of Ch. Y-LPS having the short O-specific chains and penta-acylated lipid A and containing some amount of molecules with the hexaacylated lipid A moiety showed multimodal size distribution forming three groups of particles differing in size with the same negative charge. In contrast to E-LPS the endotoxin from Y. pseudotuberculosis formed only the positively charged complexes with Ch, being heterogeneous in size regardless of the Ch content in them. These findings let us assume a different interaction mechanism of E-LPS and Y-LPS with Ch. In the case of the complex with high Y-LPS content (Y-LPS–Ch 5:1 w/w) there was no apparently any transformation of the aggregate structure of the initial Y. pseudotuberculosis endotoxin: the sizes of the complex were similar to those of YLPS (Table 1). A significant change in sizes was observed for the 1:1 Y-LPS–Ch complex. The particles of this complex were significantly smaller than those of the initial Y-LPS and had practically an unimodal size distribution. The formation of large particles with diameter 779 nm having high positive -potential (21.4 mV), as well as small particles with low positive charge (9.6 mV) for the YLPS–Ch 1:5 w/w complex, was observed. These findings may result from intercalation of Ch into the large aggregates (1038 nm) of the initial Y-LPS and their disintegration into smaller units (240.5 nm for Y-LPS–Ch 1:1 or 779 nm for Y-LPS–Ch 1:7). Another interaction mechanism can also be proposed. According to size distribution of the initial Y-LPS (Table 1), a dynamic equilibrium can be expected to exist between large (1038 nm) and small (108.5 nm) aggregates in the Y-LPS solution. This process strongly depends on concentration: particles of smaller sizes become dominant in dilute solutions (Santos, Silva, Castanho, Martins-Silva, & Saldanha, 2003; Sasaki & White, 2008). Ch binding Y-LPS decreases its concentration in the solution and shifts the equilibrium from large aggregates to smaller ones. A comparative analysis of the parameters of Y-LPS and their complexes with Ch presented in Table 1 confirmed our previous assumption (Davydova et al., 2008) that chitosan binds to the surface of particles during complex formation. Some differences in particle sizes and the -potential values for LPS and their complexes with Ch presented here and published earlier for these LPS and their complexes with Ch (Davydova et al., 2008) can be explained by different concentrations of the initial components and conditions of samples preparation. 3.2. AFM of LPS and their complexes A more detailed study of the structures of the LPS aggregates and the LPS–Ch complexes were performed by atomic force microcopy. 3.2.1. Initial components The AFM image of E-LPS show micelle-like formations typical for amphiphilic polymers. A smooth sublayer composed of very small particles (<10 nm) with its structure being indiscernible by AFM is also detected (Fig. 1a). The E-LPS sample deposited on mica formed several layers of different height (Fig. 1a): a 5 nm high sublayer and the upper layer structures up to 30 nm high. The AFM images of Y-LPS differed significantly from those of E-LPS. Y-LPS formed micelle-like structures of various size: separate particles uniformly distributed on mica and extended irregularly shaped aggregates. The aggregates were composed of particles 95 ± 18 nm in size, which could also be identified among the arranged particles (Fig. 1b). There were a ∼30 nm high layer composed of densely-packed sphere-like particles and individual particles with height up to 7 nm on the mica surface (Fig. 1b). This fact confirms our earlier assumption that large extended LPS formations resulted from aggregation of small particles. The
118
V.N. Davydova et al. / Carbohydrate Polymers 123 (2015) 115–121
Table 1 -Potential and particle size distribution for LPS and LPS–Ch complexes. Sample
-Potential measurements
Size measurements
-Potential (mV)
Contents (%)
d (nm)
Area intensity (%)
−33.5 ± 6.4
100
132.5 ± 0.3
100
Ch
+36.1 ± 1.6
100
E-LPS–Ch 7:1 w/w E-LPS–Ch 4:1 w/w
−29.0 ± 0.2 −6.9 ± 0.6
100 100
153.8 ± 2.3 3110 ± 207 658.9 ± 2.3
100 62.3 37.7
E-LPS
–
–
E-LPS–Ch 1:1 w/w
+23.1 ± 1.1
100
202.7 ± 2.6
100
E-LPS–Ch 1:7 w/w
+39.0 ± 1.2
100
280.8 ± 1.5 1038 ± 7.8
100 54.5
Y-LPS
−27.7 ± 1.4
100
108.5 ± 6.1 34.5 ± 3.0
41.1 4.4
Y-LPS–Ch 5:1 w/w
+20.3 ± 2.1 +6.2 ± 1.6
1203 ± 225.1 179.2 ± 7.1 56.0 ± 17.9
41.6 53.6 4.8
Y-LPS–Ch 1:1 w/w
+23.7 ± 0.9
240.5 ± 19.6 55.0 ± 7.6
81.5 18.5
Y-LPS–Ch 1:5 w/w
+21.4 ± 1.6 +9.6 ± 0.8
779.9 ± 24.3 170.8 ± 24.6
47.3 52.8
45.9 54.1 100 47.9 53.1
heterogeneity of Y-LPS previously detected by DSL (Table 1), electron microscopy and centrifugation (Davydova et al., 2000, 2008) was confirmed in this study by AFM. Chitosan was studied at concentrations of 0.01, 0.1 and 0.5 mg/ml. According to the AFM images shown in Fig. 2, the supramolecular structure of Ch was strongly depended on concentration: 0.01 mg/ml Ch (Fig. 2a) produced a homogeneous and very smooth (root-mean-square roughness (RMS)—0.13 nm) surface that is consistent with the previously obtained results (Volod’ko et al., 2014). The 0.1 mg/ml polysaccharide (Fig. 2b) had a structure of densely packed round-shaped particles. We also observed closely packed round-shaped particles associated into lineaments for 0.5 mg/ml Ch (Fig. 2c). All the experiments illustrated that Ch was uniformly distributed on the surface of mica. These finding are in agreement with the earlier published results demonstrating that Ch can associate in solutions and form extended assemblies (Philippova et al., 2001). The microfibrillar structure of N-deacetylated Ch preparations from crab shells has also been
detected earlier by scanning electron microscopy (Yen, Yang, & Mau, 2009). 3.2.2. LPS–chitosan complexes The supramolecular structure of the E-LPS–Ch complexes was completely different from that of the initial LPS (Fig. 3). At the ratios of 4:1 w/w (Fig. 3a) the E-LPS–Ch complex was represented by spherical particles with a halo around each particle, which were uniformly distributed on over the mica surface. The internal particle diameter was 150 ± 30 nm whereas the particles with external halo were 360 ± 60 nm in size. At the 1:1 w/w ratio between the components (Fig. 3b) the image of the complex showed long cross-linked worm-like structures that were going in the random coil. The AFM image of the E-LPS–Ch 1:7 w/w complex (Fig. 3c) showed the structures uniformly distributed on the mica surface, but greater aggregates with the average size of 648 ± 118 nm were formed at some points. In this case the sample covered the whole surface of mica, in the same
Fig. 1. AFM topography images of E-LPS (a) and Y-LPS (b), CLPS = 0.1 mg/ml; the insert shows the portion of the image with the contrast adjusted to show the fine structure of the layer.
V.N. Davydova et al. / Carbohydrate Polymers 123 (2015) 115–121
119
Fig. 2. AFM topography images: (a) Ch, 0.01 mg/ml; (b) Ch, 0.1 mg/ml; (c) Ch, 0.5 mg/ml.
manner as it was for the unbound Ch. The increase in Ch content probably promotes screening of endotoxin particles thus forming a positively charged hydrophilic corona on their surface according to the theory of formation of polyelectrolyte complexes proposed by Schatz, Domard, Viton, Pichot, and Delair (2004). This assumption agrees well with the DLS data: the -potential of this complex was similar to that of unbound Ch (Table 1). Some discrepancies in sizes observed of for the E-LPS–Ch complexes by AFM and DLS can be attributed to artefacts arising from sample drying on mica or can be explained by different conditions of sample preparation and samples recording. Spherical-shape particles were detected for the Y-LPS–Ch complexes. Individual point units 10 nm in height and 49 nm in diameter and two layers with different heights (10 nm and ∼20 nm) were observed for the Y-LPS−Ch 5:1 w/w complexes (Fig. 3d). The macromolecular structure of the Y-LPS–Ch 1:1 w/w complex was represented by several types of structures. There were
individual spherical structures 53 ± 12 nm high and 287 ± 61 nm in diameter that have accumulated in some places to form denser and larger formations, such as branching clusters and a sublayer (Fig. 3e). The similar particles were also detected in the AFM images of initial LPS but they were smaller and tighter packed in the extended structures. These findings support the previously stated assumption that the small size of the complex results from dissociation of large particles of the initial LPS due to a shift in the equilibrium between large and small LPS aggregates towards smaller units. It should be noted that the sizes of the Y-LPS–Ch 1:1 w/w complex determined by AFM and DLS coincide. Beam- or star-shaped branching structures with different size were registered for the Y-LPS–Ch 1:5 w/w complex (Fig. 3f). These particles have a constriction in the middle that let us suppose that association starts on the LPS surface. These branching structures probably consist of circular segments that are similar to those observed for the initial Ch but are arranged in a different way.
Fig. 3. AFM topography images of the complexes: (a) E E-LPS–Ch 1:1 w/w; (c) E-LPS–Ch 1:7; (d) Y-LPS–Ch 5:1 w/w; (e) Y-LPS–Ch 1:1 w/w; the insert shows the portion of the image with the contrast adjusted to show the fine structure of the layer; (f) Y-LPS–Ch 1:5 w/w. CLPS = 0.1 mg/ml.
120
V.N. Davydova et al. / Carbohydrate Polymers 123 (2015) 115–121
Fig. 4. TNF-␣ level stimulated by E-LPS (a) and Y-LPS (b) and their complexes. Concentration of the free LPS and LPS in a complex—10 ng/ml (black column), 100 ng/ml (light grey column); concentration of alone Ch—10 ng/ml (black column), 100 ng/ml (light grey column). Mean (± SD) contents of cytokine in serum are presented. Whole blood was obtained from 5 healthy subjects and incubated with the samples at different concentrations. The serum level of cytokines in healthy donors was considered to be the negative control and was used to calculate statistical significance. * Differences between the samples and the control were significant, p < 0.05.
Thus, the DLS, electrokinetic and AFM studies allow us to conclude that the mechanisms of complex formation differ for E-LPS and Y-LPS complexes. The binding of Ch to Y-LPS is not accompanied by any significant transformation in the supramolecular structure of LPS. The changes in size and the positive charge of the Y-LPS–Ch complexes are most likely to be due to Ch binding to the surface of LPS particles. 3.3. Biological activity Biological activities of various LPS differ greatly depending on their structures. Furthermore, the physical state of LPS aggregates in aqueous solutions significantly affects its biological properties (Caroff, Karibian, Cavaillon, & Haeffner-Cavaillon, 2002). The ability of LPS and its complexes with Ch to induce cytokine synthesis in blood cells and to generate reactive oxygen species (ROS) in neutrophils were studied. 3.3.1. Cytokine-inducing activity The influence of composition of the LPS–Ch complexes on their cytokine-inducing activity was studied by a test on induction of the key proinflammatory cytokine, tumour necrosis factor-alpha (TNF-␣). Fig. 4 shows that E-LPS possesses a high ability to stimulate synthesis of TNF-␣. Its cytokine-inducing activity in the complexes decreased significantly (at least threefold) and depended on content of Ch in the complex. The low activity was shown for the E-LPS–Ch complexes with component weight ratios of 7:1; 4:1, 1:1 at low LPS concentration (10 ng/ml) and 4:1 (LPS concentration of 100 ng/ml). The complex with the highest Ch content (E-LPS–Ch 1:7 w/w) possessed the highest cytokine-inducing activity; however its activity was lower
than that for the initial LPS. A significant decrease in activity of the LPS bound with chitosan could be result from transformation of the supramolecular structure of LPS detected by AFM and DLS. In contrast to E-LPS, the endotoxin from Y. pseudotuberculosis was less active in this assay due to the penta-acyl structure of its lipid A moiety. After the complex with Ch had been formed, the cytokine-inducing activity of this LPS was changed insignificantly. A statistically significant decrease in the TNF-␣ level was observed only for the LPS–Ch (1:5 w/w) complex with the predominance of Ch (1:5 w/w) at LPS concentration of 100 ng/ml. This fact may be connected with the low activity of the initial LPS. At the same time sphere-like particles were detected in AFM images of the initial Y-LPS as well as its complexes with Ch, which may argue in favour of the fact that binding of Ch does not significantly affect the supramolecular structure of Y-LPS, although dissociation of large aggregates of the initial LPS occurs. 3.3.2. ROS-inducing activity The ability of LPS and their complexes to induce synthesis of reactive oxygen species was estimated by flow cytometry for a model based on white blood cells derived from healthy donors. According to the data of Fig. 5, the ability of LPS to induce reactive oxygen species changed after binding to Ch. This effect was more pronounced for E-LPS. The E-LPS–Ch 4:1 w/w complex exhibited the lowest activity compared to other complexes of this LPS type. The ability of Y-LPS to induce ROS after binding to Ch was changed insignificantly. A slight rise in the ROS level was observed after the cells had been stimulated with complexes with high Ch content (Y-LPS–Ch 1:5 w/w and E-LPS–Ch 1:7 w/w). These complexes were characterized by high positive -potentials and surface localization of Ch according to the AFM data. Hence, we suggested that this effect
Fig. 5. The effect of E-LPS (a) and Y-LPS (b) and their complexes on ROS formation in human neutrophils. Concentrations of the initial LPS and LPS in complexes are 25 g/ml (black column), 50 g/ml (light grey column); the concentrations of Ch alone are 25 g/ml (black column), 50 g/ml (light grey column) and 350 g/ml (grey column). The mean (±SD) of ROS levels (%) are presented. White blood cells were incubated with the samples at different concentrations or with saline solution (the control column). * Differences between the samples and the control were significant, p < 0.05.
V.N. Davydova et al. / Carbohydrate Polymers 123 (2015) 115–121
could be due to the action of the Ch component, which is known to possess high activity in this test on its own. Our assumption regarding the transformation of the supramolecular structure of E-LPS after binding to Ch agrees well with the data obtained by studying the biological activity of the resulting complexes. The mechanism of biological response of the cells to LPS is complicated and can involve both soluble and cell surface receptors (Tobias, Soldau, Gegner, Mintz, & Ulevitch, 1995). Mueller, Lindner, Dedrick, Schromm, and Seydel (2005) proved that stimulation of human mononuclear cells with an LPS aggregate suspension was significantly higher than that using a LPS monomer solution of the identical concentration. Our findings demonstrate that dissociation of LPS aggregates took place during the formation of the Y-LPS–Ch complexes; however there were no considerable changes in their biological activity. An increase in size of E. coli LPS–Ch complexes was conversely accompanied by a decrease in their activity. These facts allow supposing that detoxication of LPS bound to Ch may result from inhibition of the interaction of LPS with serum endotoxin-binding proteins and/or cell receptor proteins; however further research is needed. 4. Conclusions The data presented here show that sizes, surface potentials and the supramolecular structure of the LPS–Ch complexes depend both on the endotoxin structure and on the ratios between LPS and Ch. A detailed study of the supramolecular structures of the LPS–Ch complexes using atomic force microscopy clearly indicated that chitosan affected the structure of aggregates of LPS isolated from various bacteria and having different molecular structures in different ways. The transformations in the supramolecular structure of LPS after binding to Ch showed a good correlation with the modification of biological activity of endotoxins in the complexes under study. These findings create the preconditions for further detailed study of the mechanisms of chitosan-induced reduction of LPS toxicity in view of the supramolecular structure of endotoxin. Acknowledgement This work was financially supported by the Russian Foundation for Basic Research, project no. 13-04-00786. References Aider, M. (2010). Chitosan application for active bio-based films production and potential in the food industry: Review. LWT—Food Science and Technology, 43, 837–842. Bone, R. C. (1996). The sepsis syndrome: Definition and general approach to management. Clinics in Chest Medicine, 17, 175–181. Caroff, M., Karibian, D., Cavaillon, J. M., & Haeffner-Cavaillon, N. (2002). Structural and functional analyses of bacterial lipopolysaccharides. Microbes and Infection, 4, 915–926. Davydova, V. N., Bratskaya, S., Yu Gorbach, V. I., Solov’eva, T. F., Kaca, W., & Yermak, I. M. (2008). Comparative study of electrokinetic potentials and binding affinity of lipopolysaccharides–chitosan complexes. Biophysical Chemistry, 136, 1–6. Davydova, V. N., Naberezhnykh, G. A., Yermak, I. M., Gorbach, V. I., & Solov’eva, T. F. (2006). Determination of binding constants of lipopolysaccharides of different structure with chitosan. Biochemistry (Moscow), 71, 332–339. Davydova, V. N., Yermak, I. M., Gorbach, V. I., Krasikova, I. N., & Solov’eva, T. F. (2000). Interaction of bacterial endotoxins with chitosan. Effect of endotoxin structure, chitosan molecular mass, and ionic strength of the solution on the formation of the complex. Biochemistry (Moscow), 65, 1082–1090. Domszy, J. G., & Roberts, G. A. F. (1985). Evaluation of infrared spectroscopic techniques for analysing chitosan. Macromolecular Chemistry, 186, 1671–1677.
121
Galanos, C., Luderitz, O., & Westphal, O. (1969). A new method for the extraction of R lipopolysaccharides. European Journal of Biochemistry, 9, 245–249. Harding, S. E. (1997). The intrinsic viscosity of biological macromolecules. Progress in measurement, interpretation and application to structure in dilute solution. Progress in Biophysics and Molecular Biology, 68, 207–262. Henry, D. C. (1931). The cataphoresis of suspended particles. Part I. The equation of cataphoresis. Proceedings of the Royal Society of London, 133, 106–129. Jucker, B. A., Harms, H., & Zehnder, A. J. (1998). Polymer interactions between five Gram-negative bacteria and glass investigated using LPS micelles and vesicles as model systems. Colloids and Surfaces B: Biointerfaces, 11, 33–45. Krasikova, I. N., Gorbach, V. I., Solov’eva, T. F., & Ovodov, Yu. S. (1978). Studies on lipid A from Yersinia pseudotuberculosis lipopolysaccharide. Isolation and general characterization. European Journal of Biochemistry, 89, 287–289. Lehmann, A. K., Sørnes, S., & Halstensen, H. (2000). Phagocytosis: Measurement by flow cytometry. Journal Immunological Methods, 243, 229–242. Lowry, O. H., Rosebrough, N. J., Farr, A. L., & Randall, R. J. (1951). Protein measurement with the Folin phenol reagent. Journal of Biological Chemistry, 193, 265–275. Mueller, M., Lindner, B., Dedrick, R., Schromm, A. B., & Seydel, U. (2005). Endotoxin: Physical requirements for cell activation. Journal of Endotoxin Research, 11, 299–303. Muzzarelli, R. A. A. (2010). Chitins and chitosans as immunoadjuvants and nonallergenic drug carriers. Marine Drugs, 8, 292–312. Muzzarelli, R. A. A. (2009). Chitins and chitosans for the repair of wounded skin, nerve, cartilage and bone. Carbohydrate Polymers, 76, 167–182. Ovodov Yu, S., Gorshkova, R. P., & Tomshich, S. V. (1971). Chemical and immunochemical studies on Pasteurella pseudotuberculosis lipopolysaccharides—I. and general characterization. Immunochemistry, 8, Isolation 1071–1078. Peri, F., Piazza, M., Calabrese, V., Damore, G., & Cighetti, R. (2010). Exploring the LPS/TLR4 signal pathway with small molecules. Biochemical Society Transactions, 38, 1390–1395. Philippova, O. E., Volkov, E. V., Sitnikova, N. L., Khokhlov, A. R., Desbrieres, J., & Rinaudo, M. (2001). Two types of hydrophobic aggregates in aqueous solutions of chitosan and its hydrophobic derivative. Biomacromolecules, 2, 483–490. Raetz, C. R. (1990). Biochemistry of endotoxins. Annual Review of Biochemistry, 59, 129–170. Santos, N. C., Silva, A. C., Castanho, M. A. R. B., Martins-Silva, J., & Saldanha, C. (2003). Evaluation of lipopolysaccharide aggregation by light scattering spectroscopy. ChemBioChem, 4, 96–100. Sasaki, H., & White, H. S. (2008). Aggregation behavior of an ultra-pure lipopolysaccharide that stimulates TLR-4 receptors. Biophysical Journal, 95, 986–993. Setsukinai, K. I., Urano, Y., Kakinuma, K., Majima, H. J., & Nagano, N. (2003). Development of novel fluorescence probes that can reliably detect reactive oxygen species and distinguish specific species. The Journal of Biological Chemistry, 278, 3170–3175. Schatz, C., Domard, A., Viton, C., Pichot, C., & Delair, T. (2004). Versatile and efficient formation of colloids of biopolymer-based polyelectrolyte complexes. Biomacromolecules, 5, 1882–1892. Spirin, A. S., Belozerskii, A. N., Kudlai, D. G., Skavronskaya, A. G., & Mitereva, V. G. (1958). Changes in the composition of nucleic acids of experimentally evolved saccharolytic intestinal bacteria. Biokhimiya (Moscow), 23, 154–163. Solov’eva, T. F., Davydova, V. N., Krasikova, I. N., & Yermak, I. M. (2013). Marine compounds with therapeutic potential in gram-negative sepsis. Marine Drugs, 11, 2216–2229. Tobias, P. S., Soldau, K., Gegner, J. A., Mintz, D., & Ulevitch, R. J. (1995). Lipopolysaccharide binding protein-mediated complexation of lipopolysaccharide with soluble CD14. Journal of Biological Chemistry, 270, 10482–10488. Tomshich, S. V., Gorshkova, R. P., & Ovodov, Y. (1985). Structural studies on the core of the lipopolysaccharides from Yersinia pseudotuberculosis. Khimiya Prirodnykh Soedinenij, 6, 751–754. Tomshich, S. V., Gorshkova, R. P., Elkin, Y. N., & Ovodov Yu, S. (1976). Lipopolysaccharide from Yersinia pseudotuberculosis type 1B. A structural study of O-specific chains. European Journal of Biochemistry, 65, 193–199. Ulevitch, R. J. (2000). Molecular mechanisms of innate immunity. Immunological Research, 21, 49–54. Van Amersfoort, E. S., Van Berkel, T. J., & Kuiper, J. (2003). Receptors, mediators, and mechanisms involved in bacterial sepsis and septic shock. Clinical Microbiology Reviews, 16, 379–414. Volod’ko, A. V., Davydova, V. N., Chusovitin, E. V., Sorokina, I. V., Dolgikh, M. P., Tolstikova, T. G., et al. (2014). Soluble chitosan–carrageenan polyelectrolyte complexes and their gastroprotective activity. Carbohydrate Polymers, 101, 1087–1093. Wolfrom, M. L., & Shen Han, T. M. (1959). The sulfonation of chitosan. Journal of the American Chemical Society, 81, 1764–1766. Yen, M. T., Yang, J. H., & Mau, J. L. (2009). Physicochemical characterization of chitin and chitosan from crab shells. Carbohydrate Polymers, 75, 15–21. Yermak, I. M., Davidova, V. N., Gorbach, V. I., Luk’yanov, P. A., Solov’eva, T. F., Ulmer, A. J., et al. (2006). Forming and immunological properties of some lipopolysaccharide-chitosan complexes. Biochimie, 88, 23–30.
Contents lists available at ScienceDirect
Carbohydrate Polymers journal homepage: www.elsevier.com/locate/carbpol
The supramolecular structure of LPS–chitosan complexes of varied composition in relation to their biological activity V.N. Davydova a,∗ , A.V. Volod’ko a , E.V. Sokolova a , E.A. Chusovitin b , S.A. Balagan b , V.I. Gorbach a , N.G. Galkin b,c , I.M. Yermak a , T.F. Solov’eva a a G.B. Elyakov Pacific Institute of Bioorganic Chemistry, Far Eastern Branch of Russian Academy of Sciences, Prospect 100 let Vladivostoku 159, Vladivostok 690022, Russia b Institute of Automation and Control Processes, Far Eastern Branch of Russian Academy of Sciences, Radio Str. 5, Vladivostok 690041, Russia c Far Eastern Federal University, Sukhanova Str. 8, Vladivostok 690091, Russia
a r t i c l e
i n f o
Article history: Received 24 September 2014 Received in revised form 13 January 2015 Accepted 15 January 2015 Available online 21 January 2015 Keywords: Lipopolysaccharide (LPS) Chitosan LPS–chitosan complexes -Potential Supramolecular structure
a b s t r a c t The complexes of chitosan (Ch) with lipopolysaccharides (LPSs) from Escherichia coli O55:B5 (E-LPS) and Yersinia pseudotuberculosis 1B 598 (Y-LPS) of various weight compositions were investigated using quasielastic light scattering, -potential distribution assay and atomic force microscopy. The alteration of potential of E-LPS–Ch complexes from negative to positive values depending on Ch content was detected. The Y-LPS–Ch complexes had similar positive -potentials regardless of Ch content. The transformation of the supramolecular structure of E-LPS after binding with to Ch was revealed. Screening of E-LPS and Y-LPS particles by Ch in the complexes with high polycation was detected. The ability of LPS–Ch complex to induce biosynthesis of TNF-␣ and reactive oxygen species in stimulated human mononuclear cells was studied. A significant decrease in activity complexes compared to that of the initial LPS was observed only for E-LPS–Ch complexes. © 2015 Elsevier Ltd. All rights reserved.
Chemical compounds studied in this article: Lipopolysaccharide (PubChem CID: 11970143) Chitosan (PubChem CID: 71853)
1. Introduction Lipopolysaccharide (LPS or endotoxin) is one of the strongest elicitors of the immune system of macroorganisms. It is responsible for induction of many cytokines and chemokines but also the other immune mediators in human immune cells, such as monocytes, macrophages, dendritic cells, and others (Ulevitch, 2000). Activation of these cells may be beneficial at low amounts of LPS but at higher LPS concentrations lead to pathophysiological reactions such as sepsis (Bone, 1996). LPS consists of the hydrophobic part, known as lipid A, the repeating O-antigen polysaccharide, and the core oligosaccharide (Raetz, 1990). Many of the immune activation abilities of LPS can be attributed to the lipid A unit, which binds to the tolllike receptor 4 (TLR-4) and activates the host defense effector
∗ Corresponding author. Tel.: +7 423 2310719; fax: +7 423 2314050. E-mail address: [email protected] (V.N. Davydova). http://dx.doi.org/10.1016/j.carbpol.2015.01.028 0144-8617/© 2015 Elsevier Ltd. All rights reserved.
system by rapidly triggering proinflammatory processes (Peri, Piazza, Calabrese, Damore, & Cighetti, 2010). Therapies for Gram-negative sepsis remain unsatisfactory despite the concerted effort to develop new treatments for this life-threatening syndrome. Currently, no drugs are specific for endotoxin-induced clinical syndromes. Hence, compounds with therapeutic potentials as anti-endotoxin agents may include those that either bind LPS at high affinity and neutralize its toxicity or those that competitively interact with LPS receptors on host cells without activating biological responses (receptor antagonists). Both groups of agents, acting through different mechanisms, block endotoxin binding to specific receptors on target cells, thus preventing the synthesis of proinflammatory cytokines (Van Amersfoort, Van Berkel, & Kuiper, 2003). One of the promising ways to inhibit the harmful inflammatory/septic effects of endotoxin is to bind LPS to certain polycations, which can interact with lipid A bearing the negatively charged phosphate groups and, as a result, block this toxic centre of the endotoxin.
116
V.N. Davydova et al. / Carbohydrate Polymers 123 (2015) 115–121
Chitosan, the cationic (1-4)-2-amino-2-deoxy--d-glucan, with degree of acetylation typically close to 0.20, is prevalently produced from marine chitin. Chitosan and its partially depolymerized derivatives and oligomers have a wide spectrum of biological activities and are particularly useful in the fields of wound healing, oral delivery, and food industry (Aider, 2010; Muzzarelli, 2009, 2010). The systematic studies of chitosan binding to LPS and the biological properties of the chitosan–LPS complexes showed that it interacts specifically with LPS to form stable water-soluble complexes of various stoichiometric compositions and that the formation of LPS–chitosan complexes is accompanied by significant modification of immunological properties of LPS (Solov’eva, Davydova, Krasikova, & Yermak, 2013). Due to the amphiphilic nature of their molecules and the very low critical micellar concentration (CMC), LPS exist as aggregates in aqueous solutions and their aggregate state has been found to be an important factor of interaction between LPS and Ch (Davydova, Naberezhnykh, Yermak, Gorbach, & Solov’eva, 2006). As shown previously, the LPS–chitosan complexes are 10–20 times less toxic than LPS alone (Yermak et al., 2006). This effect depends on LPS structure, the ratio between the components in the complex and chitosan molecular weight. A substantial reduction of LPS toxicity in LPS–chitosan complexes may be explained by the blocking of the toxophoric centre of endotoxin or the alterations in the molecular charge and/or the structure of LPS aggregates by chitosan (Solov’eva et al., 2013). The development of modern polymers analysis techniques offers great opportunities for studying the supramolecular organization of macromolecules. The atomic force microscopy can obtain a high resolution image of the biomolecules under nearphysiological conditions. It allows us to study complexes of different stoichiometry between endotoxins of different structure and chitosan and to try to find a correlation between their structure and biological activity. For this purpose two different structural types of LPS (commercial Escherichia coli LPS O55 (E-LPS) and LPS isolated from bacteria Yersinia pseudotuberculosis 1B598 (Y-LPS)) and Ch with molecular weight of 110 kDa and N-acetylation degree of 1% have been chosen to prepare their complexes with different stoichiometry.
2. Materials and methods
calculated according to IR-spectroscopy data (Domszy & Roberts, 1985). The molecular weight of the chitosan was determined by viscosimetry in 0.1 M AcOH/0.2 M NaCl according to (Harding, 1997). 2.2. Preparation of LPS–Ch complexes LPS (1 mg) and Ch (7–1.4 mg) (according to its contents in the complexes) were dissolved separately in 5 ml of pyrogen-free deionized water. The solutions were stored for 48 h at 37 ◦ C; equal aliquots of LPS and Ch were then mixed and incubated for 18 h at 37 ◦ C. The LPS and chitosan solutions were decontaminated of bacteria by filtration using a “Millex GS” filter unit (Millipore, Ireland). 2.3. Dynamic light scattering (DLS) and electrophoretic properties of the LPS–chitosan complexes The sizes and -potentials of the LPS aggregates and LPS–Ch complexes in solution were determined using a ZetaSizer Nano ZS system (Malvern, UK) operating at 633 nm. Prior to measurements, the samples were left for 1 h to allow the large aggregates to settle, as they can interfere with the measurements even if their content does not exceed a few percent. The measurements were performed at 25 ◦ C for E-LPS and at 37 ◦ C for Y-LPS. The hydrodynamic diameters of the particles were automatically calculated with the instrument’s software based on analysis of the autocorrelation function. The -potentials were calculated from the experimentally determined electrophoretic mobility using the Henry equation (Henry, 1931). 2.4. Atomic force microscopy (AFM) LPSs were dissolved in distilled de-ionized water at concentration of 0.1 mg/ml; Ch samples—at concentrations of 0.01, 0.1 or 0.5 mg/ml. The LPS–Ch complexes were prepared at the same concentrations. Aliquots (12 l) of the aqueous solutions of each complex and their initial component were deposited onto freshly cleaved mica and dried at 37 ◦ C for 24 h or at 70 ◦ C for 30 s (for LPS). The morphology of Ch, LPS and their complexes was studied in air by AFM (Solver P47) in the tapping contact mode using a tip with the radius of 10 nm. The topographic parameters of the macromolecular structure have been automatically calculated using the self-developed Balagan’s Grain Analysis programme.
2.1. Isolation of lipopolysaccharides and chitosan 2.5. Ethical approval Cells of Y. pseudotuberculosis (serovar 1B, strain 598) isolated from a patient suffering from the far-eastern scarlatina-like fever (Institute of Epidemiology and Microbiology, Vladivostok, Russia) were grown at 4 ◦ C in a previously described nutrient medium (Ovodov et al., 1971). LPS from bacterial cells was isolated using the phenol-chloroform-petroleum ether procedure (Galanos, Luderitz, & Westphal, 1969). Nucleic acids were removed by ultracentrifugation at 105,000 × g. The purified LPS (yield: 1.2%) contained less than 1% of protein. Protein content was determined by Lowry’s method (Lowry, Rosebrough, Farr, & Randall, 1951), nucleic acid contents were estimated in accordance with Spirin et al. (1958). Structural characterization of Yersisnia pseudotuberculosis LPS elucidated previously (Krasikova, Gorbach, Solov’eva, & Ovodov, 1978; Tomshich, Gorshkova, Elkin, & Ovodov Yu, 1976; Tomshich, Gorshkova, & Ovodov, 1985). A preparation of LPS from E. coli 055:B5 was purchased from Sigma (Sigma Chemicals, St. Louis, MO, USA). A chitosan (Ch) sample with molecular weight of 110 kDa and 1% degree of N-acetylation was obtained by alkaline treatment of crab chitin according to the published protocol (Wolfrom & Shen Han, 1959). The degree of N-acetylation of the chitosan sample was
The study protocol was approved by the medical ethics committee of the local hospital (Vladivostok, Russian Federation). Informed consent was obtained from all subjects who participated in the study. All donors were free of medicines administration for 14 days prior to blood sampling. Blood was drawn from the antecubital vein of normal healthy human volunteers and anticoagulated in plastic tubes (Greiner Bio-One International AG, Kremsmuenster, Austria) with 30 IU lithium heparinate used as an anticoagulant. 2.6. TNF-inducing activity Human peripheral blood was collected by venipuncture into sterile siliconized tubes and diluted at a ratio of 1:5 with sterile Medium 199 (Sigma, USA) containing 300 mg/l of glutamine (Gibco, Life Technology, Germany) and 50 g/ml of gentamicin. The diluted blood (0.1 ml) was transferred into sterile polypropylene plates and then incubated with the corresponding LPS, Ch, or LPS–Ch complex (37 ◦ C, 5% CO2 ). Control incubation with 1 g/ml of LPS from E. coli (strain 055:B5) was performed for each experiment. After 24 h the supernatants were collected and frozen, followed
V.N. Davydova et al. / Carbohydrate Polymers 123 (2015) 115–121
by cytokine determination using specific ELISA (DuoSet developing system, Genzyme, Boston, USA). 2.7. Leukocytes We rapidly separated leukocytes from venous blood by lyzing erythrocytes in a solution containing 0.15 M NH4 Cl, 10 mM NaHCO3 and 0.1 mM EDTA for 10 min (Lehmann, Sørnes, & Halstensen, 2000). The leucocyte fraction was washed twice, resuspended in PBS, and adjusted to a concentration of ≈107 non-lymphocytes (polymorphonuclear leukocytes and monocytes) per millilitre. 2.8. Detection of production of reactive oxygen species Production of reactive oxygen species was assayed by flow cytometry using APF as described (Setsukinai, Urano, Kakinuma, Majima, & Nagano, 2003). Cells were plated at 100,000 cells/well in a 96-well tissue culture plate (Costar, Cambridge, MA) and stimulated for 1 h with samples (50 and 100 g/ml, final value); APF (10 M) was added to the media for the final 30 min. The cells were lifted, kept on ice (10 min), and analyzed immediately by flow cytometry. Flow cytometric measurements were performed using a BD FACSCalibur four-colour analyzer (Becton Dickinson). Forward and side scatter light was used to identify cell populations and to measure cell size and granularity. 2.9. Statistical analysis All data were represented as mean ± standard deviation. Statistical analysis was done by one-way ANOVA. Differences were considered to be statistically significant if p < 0.05. 3. Results and discussion 3.1. Size and -potential of the LPS–chitosan complexes Several LPS–Ch complexes of varying compositions obtained at E. coli O55:B5 LPS (E-LPS) and Y. pseudotuberculosis 598 LPS (YLPS) concentrations of 0.1 mg/ml were prepared. The weight ratios between the components in the complexes were chosen according to the previously obtained results (Davydova, Yermak, Gorbach, Krasikova, & Solov’eva, 2000; Yermak et al., 2006). The study of the complexes by DLS and electrokinetic methods showed that their sizes and surface potentials depended on the endotoxin structure and composition of the complexes (Table 1). It is known that LPS as amphiphilic molecules tend to form supramolecular aggregates in aqueous solutions at concentrations above the critical micellar concentration. The length of the hydrophilic carbohydrate moiety of a LPS molecule affects size and structure of its aggregates (Jucker, Harms, & Zehnder, 1998). ELPS possessing the long O-specific chains and hexa-acylated lipid A formed negatively charged aggregates homogeneous in size. According to the results E-LPS formed the negatively charged homogeneous in size aggregates in the aqueous solution. Rather homogeneous complexes with varied surface potential changing from negative to positive with increase in content of Ch in complex were also registered after binding LPS with Ch. The sizes of the complexes with weight ratios of E-LPS:Ch 7:1, 1:1, 1:7 were close to that of LPS aggregates and increased slightly with raising Ch content in complexes. In the case of E-LPS–Ch 4:1 w/w complex we detected the formation of two types of particles with much larger sizes (3110 and 658 nm) compared to those of other complexes and initial E-LPS aggregates and the low negative value of the -potential. These large particles were probably formed by aggregation of smaller ones due to a dramatic decrease in an electrostatic
117
repulsion between them because the negative LPS charge was neutralized by the positive charge of Ch. Y-LPS having the short O-specific chains and penta-acylated lipid A and containing some amount of molecules with the hexaacylated lipid A moiety showed multimodal size distribution forming three groups of particles differing in size with the same negative charge. In contrast to E-LPS the endotoxin from Y. pseudotuberculosis formed only the positively charged complexes with Ch, being heterogeneous in size regardless of the Ch content in them. These findings let us assume a different interaction mechanism of E-LPS and Y-LPS with Ch. In the case of the complex with high Y-LPS content (Y-LPS–Ch 5:1 w/w) there was no apparently any transformation of the aggregate structure of the initial Y. pseudotuberculosis endotoxin: the sizes of the complex were similar to those of YLPS (Table 1). A significant change in sizes was observed for the 1:1 Y-LPS–Ch complex. The particles of this complex were significantly smaller than those of the initial Y-LPS and had practically an unimodal size distribution. The formation of large particles with diameter 779 nm having high positive -potential (21.4 mV), as well as small particles with low positive charge (9.6 mV) for the YLPS–Ch 1:5 w/w complex, was observed. These findings may result from intercalation of Ch into the large aggregates (1038 nm) of the initial Y-LPS and their disintegration into smaller units (240.5 nm for Y-LPS–Ch 1:1 or 779 nm for Y-LPS–Ch 1:7). Another interaction mechanism can also be proposed. According to size distribution of the initial Y-LPS (Table 1), a dynamic equilibrium can be expected to exist between large (1038 nm) and small (108.5 nm) aggregates in the Y-LPS solution. This process strongly depends on concentration: particles of smaller sizes become dominant in dilute solutions (Santos, Silva, Castanho, Martins-Silva, & Saldanha, 2003; Sasaki & White, 2008). Ch binding Y-LPS decreases its concentration in the solution and shifts the equilibrium from large aggregates to smaller ones. A comparative analysis of the parameters of Y-LPS and their complexes with Ch presented in Table 1 confirmed our previous assumption (Davydova et al., 2008) that chitosan binds to the surface of particles during complex formation. Some differences in particle sizes and the -potential values for LPS and their complexes with Ch presented here and published earlier for these LPS and their complexes with Ch (Davydova et al., 2008) can be explained by different concentrations of the initial components and conditions of samples preparation. 3.2. AFM of LPS and their complexes A more detailed study of the structures of the LPS aggregates and the LPS–Ch complexes were performed by atomic force microcopy. 3.2.1. Initial components The AFM image of E-LPS show micelle-like formations typical for amphiphilic polymers. A smooth sublayer composed of very small particles (<10 nm) with its structure being indiscernible by AFM is also detected (Fig. 1a). The E-LPS sample deposited on mica formed several layers of different height (Fig. 1a): a 5 nm high sublayer and the upper layer structures up to 30 nm high. The AFM images of Y-LPS differed significantly from those of E-LPS. Y-LPS formed micelle-like structures of various size: separate particles uniformly distributed on mica and extended irregularly shaped aggregates. The aggregates were composed of particles 95 ± 18 nm in size, which could also be identified among the arranged particles (Fig. 1b). There were a ∼30 nm high layer composed of densely-packed sphere-like particles and individual particles with height up to 7 nm on the mica surface (Fig. 1b). This fact confirms our earlier assumption that large extended LPS formations resulted from aggregation of small particles. The
118
V.N. Davydova et al. / Carbohydrate Polymers 123 (2015) 115–121
Table 1 -Potential and particle size distribution for LPS and LPS–Ch complexes. Sample
-Potential measurements
Size measurements
-Potential (mV)
Contents (%)
d (nm)
Area intensity (%)
−33.5 ± 6.4
100
132.5 ± 0.3
100
Ch
+36.1 ± 1.6
100
E-LPS–Ch 7:1 w/w E-LPS–Ch 4:1 w/w
−29.0 ± 0.2 −6.9 ± 0.6
100 100
153.8 ± 2.3 3110 ± 207 658.9 ± 2.3
100 62.3 37.7
E-LPS
–
–
E-LPS–Ch 1:1 w/w
+23.1 ± 1.1
100
202.7 ± 2.6
100
E-LPS–Ch 1:7 w/w
+39.0 ± 1.2
100
280.8 ± 1.5 1038 ± 7.8
100 54.5
Y-LPS
−27.7 ± 1.4
100
108.5 ± 6.1 34.5 ± 3.0
41.1 4.4
Y-LPS–Ch 5:1 w/w
+20.3 ± 2.1 +6.2 ± 1.6
1203 ± 225.1 179.2 ± 7.1 56.0 ± 17.9
41.6 53.6 4.8
Y-LPS–Ch 1:1 w/w
+23.7 ± 0.9
240.5 ± 19.6 55.0 ± 7.6
81.5 18.5
Y-LPS–Ch 1:5 w/w
+21.4 ± 1.6 +9.6 ± 0.8
779.9 ± 24.3 170.8 ± 24.6
47.3 52.8
45.9 54.1 100 47.9 53.1
heterogeneity of Y-LPS previously detected by DSL (Table 1), electron microscopy and centrifugation (Davydova et al., 2000, 2008) was confirmed in this study by AFM. Chitosan was studied at concentrations of 0.01, 0.1 and 0.5 mg/ml. According to the AFM images shown in Fig. 2, the supramolecular structure of Ch was strongly depended on concentration: 0.01 mg/ml Ch (Fig. 2a) produced a homogeneous and very smooth (root-mean-square roughness (RMS)—0.13 nm) surface that is consistent with the previously obtained results (Volod’ko et al., 2014). The 0.1 mg/ml polysaccharide (Fig. 2b) had a structure of densely packed round-shaped particles. We also observed closely packed round-shaped particles associated into lineaments for 0.5 mg/ml Ch (Fig. 2c). All the experiments illustrated that Ch was uniformly distributed on the surface of mica. These finding are in agreement with the earlier published results demonstrating that Ch can associate in solutions and form extended assemblies (Philippova et al., 2001). The microfibrillar structure of N-deacetylated Ch preparations from crab shells has also been
detected earlier by scanning electron microscopy (Yen, Yang, & Mau, 2009). 3.2.2. LPS–chitosan complexes The supramolecular structure of the E-LPS–Ch complexes was completely different from that of the initial LPS (Fig. 3). At the ratios of 4:1 w/w (Fig. 3a) the E-LPS–Ch complex was represented by spherical particles with a halo around each particle, which were uniformly distributed on over the mica surface. The internal particle diameter was 150 ± 30 nm whereas the particles with external halo were 360 ± 60 nm in size. At the 1:1 w/w ratio between the components (Fig. 3b) the image of the complex showed long cross-linked worm-like structures that were going in the random coil. The AFM image of the E-LPS–Ch 1:7 w/w complex (Fig. 3c) showed the structures uniformly distributed on the mica surface, but greater aggregates with the average size of 648 ± 118 nm were formed at some points. In this case the sample covered the whole surface of mica, in the same
Fig. 1. AFM topography images of E-LPS (a) and Y-LPS (b), CLPS = 0.1 mg/ml; the insert shows the portion of the image with the contrast adjusted to show the fine structure of the layer.
V.N. Davydova et al. / Carbohydrate Polymers 123 (2015) 115–121
119
Fig. 2. AFM topography images: (a) Ch, 0.01 mg/ml; (b) Ch, 0.1 mg/ml; (c) Ch, 0.5 mg/ml.
manner as it was for the unbound Ch. The increase in Ch content probably promotes screening of endotoxin particles thus forming a positively charged hydrophilic corona on their surface according to the theory of formation of polyelectrolyte complexes proposed by Schatz, Domard, Viton, Pichot, and Delair (2004). This assumption agrees well with the DLS data: the -potential of this complex was similar to that of unbound Ch (Table 1). Some discrepancies in sizes observed of for the E-LPS–Ch complexes by AFM and DLS can be attributed to artefacts arising from sample drying on mica or can be explained by different conditions of sample preparation and samples recording. Spherical-shape particles were detected for the Y-LPS–Ch complexes. Individual point units 10 nm in height and 49 nm in diameter and two layers with different heights (10 nm and ∼20 nm) were observed for the Y-LPS−Ch 5:1 w/w complexes (Fig. 3d). The macromolecular structure of the Y-LPS–Ch 1:1 w/w complex was represented by several types of structures. There were
individual spherical structures 53 ± 12 nm high and 287 ± 61 nm in diameter that have accumulated in some places to form denser and larger formations, such as branching clusters and a sublayer (Fig. 3e). The similar particles were also detected in the AFM images of initial LPS but they were smaller and tighter packed in the extended structures. These findings support the previously stated assumption that the small size of the complex results from dissociation of large particles of the initial LPS due to a shift in the equilibrium between large and small LPS aggregates towards smaller units. It should be noted that the sizes of the Y-LPS–Ch 1:1 w/w complex determined by AFM and DLS coincide. Beam- or star-shaped branching structures with different size were registered for the Y-LPS–Ch 1:5 w/w complex (Fig. 3f). These particles have a constriction in the middle that let us suppose that association starts on the LPS surface. These branching structures probably consist of circular segments that are similar to those observed for the initial Ch but are arranged in a different way.
Fig. 3. AFM topography images of the complexes: (a) E E-LPS–Ch 1:1 w/w; (c) E-LPS–Ch 1:7; (d) Y-LPS–Ch 5:1 w/w; (e) Y-LPS–Ch 1:1 w/w; the insert shows the portion of the image with the contrast adjusted to show the fine structure of the layer; (f) Y-LPS–Ch 1:5 w/w. CLPS = 0.1 mg/ml.
120
V.N. Davydova et al. / Carbohydrate Polymers 123 (2015) 115–121
Fig. 4. TNF-␣ level stimulated by E-LPS (a) and Y-LPS (b) and their complexes. Concentration of the free LPS and LPS in a complex—10 ng/ml (black column), 100 ng/ml (light grey column); concentration of alone Ch—10 ng/ml (black column), 100 ng/ml (light grey column). Mean (± SD) contents of cytokine in serum are presented. Whole blood was obtained from 5 healthy subjects and incubated with the samples at different concentrations. The serum level of cytokines in healthy donors was considered to be the negative control and was used to calculate statistical significance. * Differences between the samples and the control were significant, p < 0.05.
Thus, the DLS, electrokinetic and AFM studies allow us to conclude that the mechanisms of complex formation differ for E-LPS and Y-LPS complexes. The binding of Ch to Y-LPS is not accompanied by any significant transformation in the supramolecular structure of LPS. The changes in size and the positive charge of the Y-LPS–Ch complexes are most likely to be due to Ch binding to the surface of LPS particles. 3.3. Biological activity Biological activities of various LPS differ greatly depending on their structures. Furthermore, the physical state of LPS aggregates in aqueous solutions significantly affects its biological properties (Caroff, Karibian, Cavaillon, & Haeffner-Cavaillon, 2002). The ability of LPS and its complexes with Ch to induce cytokine synthesis in blood cells and to generate reactive oxygen species (ROS) in neutrophils were studied. 3.3.1. Cytokine-inducing activity The influence of composition of the LPS–Ch complexes on their cytokine-inducing activity was studied by a test on induction of the key proinflammatory cytokine, tumour necrosis factor-alpha (TNF-␣). Fig. 4 shows that E-LPS possesses a high ability to stimulate synthesis of TNF-␣. Its cytokine-inducing activity in the complexes decreased significantly (at least threefold) and depended on content of Ch in the complex. The low activity was shown for the E-LPS–Ch complexes with component weight ratios of 7:1; 4:1, 1:1 at low LPS concentration (10 ng/ml) and 4:1 (LPS concentration of 100 ng/ml). The complex with the highest Ch content (E-LPS–Ch 1:7 w/w) possessed the highest cytokine-inducing activity; however its activity was lower
than that for the initial LPS. A significant decrease in activity of the LPS bound with chitosan could be result from transformation of the supramolecular structure of LPS detected by AFM and DLS. In contrast to E-LPS, the endotoxin from Y. pseudotuberculosis was less active in this assay due to the penta-acyl structure of its lipid A moiety. After the complex with Ch had been formed, the cytokine-inducing activity of this LPS was changed insignificantly. A statistically significant decrease in the TNF-␣ level was observed only for the LPS–Ch (1:5 w/w) complex with the predominance of Ch (1:5 w/w) at LPS concentration of 100 ng/ml. This fact may be connected with the low activity of the initial LPS. At the same time sphere-like particles were detected in AFM images of the initial Y-LPS as well as its complexes with Ch, which may argue in favour of the fact that binding of Ch does not significantly affect the supramolecular structure of Y-LPS, although dissociation of large aggregates of the initial LPS occurs. 3.3.2. ROS-inducing activity The ability of LPS and their complexes to induce synthesis of reactive oxygen species was estimated by flow cytometry for a model based on white blood cells derived from healthy donors. According to the data of Fig. 5, the ability of LPS to induce reactive oxygen species changed after binding to Ch. This effect was more pronounced for E-LPS. The E-LPS–Ch 4:1 w/w complex exhibited the lowest activity compared to other complexes of this LPS type. The ability of Y-LPS to induce ROS after binding to Ch was changed insignificantly. A slight rise in the ROS level was observed after the cells had been stimulated with complexes with high Ch content (Y-LPS–Ch 1:5 w/w and E-LPS–Ch 1:7 w/w). These complexes were characterized by high positive -potentials and surface localization of Ch according to the AFM data. Hence, we suggested that this effect
Fig. 5. The effect of E-LPS (a) and Y-LPS (b) and their complexes on ROS formation in human neutrophils. Concentrations of the initial LPS and LPS in complexes are 25 g/ml (black column), 50 g/ml (light grey column); the concentrations of Ch alone are 25 g/ml (black column), 50 g/ml (light grey column) and 350 g/ml (grey column). The mean (±SD) of ROS levels (%) are presented. White blood cells were incubated with the samples at different concentrations or with saline solution (the control column). * Differences between the samples and the control were significant, p < 0.05.
V.N. Davydova et al. / Carbohydrate Polymers 123 (2015) 115–121
could be due to the action of the Ch component, which is known to possess high activity in this test on its own. Our assumption regarding the transformation of the supramolecular structure of E-LPS after binding to Ch agrees well with the data obtained by studying the biological activity of the resulting complexes. The mechanism of biological response of the cells to LPS is complicated and can involve both soluble and cell surface receptors (Tobias, Soldau, Gegner, Mintz, & Ulevitch, 1995). Mueller, Lindner, Dedrick, Schromm, and Seydel (2005) proved that stimulation of human mononuclear cells with an LPS aggregate suspension was significantly higher than that using a LPS monomer solution of the identical concentration. Our findings demonstrate that dissociation of LPS aggregates took place during the formation of the Y-LPS–Ch complexes; however there were no considerable changes in their biological activity. An increase in size of E. coli LPS–Ch complexes was conversely accompanied by a decrease in their activity. These facts allow supposing that detoxication of LPS bound to Ch may result from inhibition of the interaction of LPS with serum endotoxin-binding proteins and/or cell receptor proteins; however further research is needed. 4. Conclusions The data presented here show that sizes, surface potentials and the supramolecular structure of the LPS–Ch complexes depend both on the endotoxin structure and on the ratios between LPS and Ch. A detailed study of the supramolecular structures of the LPS–Ch complexes using atomic force microscopy clearly indicated that chitosan affected the structure of aggregates of LPS isolated from various bacteria and having different molecular structures in different ways. The transformations in the supramolecular structure of LPS after binding to Ch showed a good correlation with the modification of biological activity of endotoxins in the complexes under study. These findings create the preconditions for further detailed study of the mechanisms of chitosan-induced reduction of LPS toxicity in view of the supramolecular structure of endotoxin. Acknowledgement This work was financially supported by the Russian Foundation for Basic Research, project no. 13-04-00786. References Aider, M. (2010). Chitosan application for active bio-based films production and potential in the food industry: Review. LWT—Food Science and Technology, 43, 837–842. Bone, R. C. (1996). The sepsis syndrome: Definition and general approach to management. Clinics in Chest Medicine, 17, 175–181. Caroff, M., Karibian, D., Cavaillon, J. M., & Haeffner-Cavaillon, N. (2002). Structural and functional analyses of bacterial lipopolysaccharides. Microbes and Infection, 4, 915–926. Davydova, V. N., Bratskaya, S., Yu Gorbach, V. I., Solov’eva, T. F., Kaca, W., & Yermak, I. M. (2008). Comparative study of electrokinetic potentials and binding affinity of lipopolysaccharides–chitosan complexes. Biophysical Chemistry, 136, 1–6. Davydova, V. N., Naberezhnykh, G. A., Yermak, I. M., Gorbach, V. I., & Solov’eva, T. F. (2006). Determination of binding constants of lipopolysaccharides of different structure with chitosan. Biochemistry (Moscow), 71, 332–339. Davydova, V. N., Yermak, I. M., Gorbach, V. I., Krasikova, I. N., & Solov’eva, T. F. (2000). Interaction of bacterial endotoxins with chitosan. Effect of endotoxin structure, chitosan molecular mass, and ionic strength of the solution on the formation of the complex. Biochemistry (Moscow), 65, 1082–1090. Domszy, J. G., & Roberts, G. A. F. (1985). Evaluation of infrared spectroscopic techniques for analysing chitosan. Macromolecular Chemistry, 186, 1671–1677.
121
Galanos, C., Luderitz, O., & Westphal, O. (1969). A new method for the extraction of R lipopolysaccharides. European Journal of Biochemistry, 9, 245–249. Harding, S. E. (1997). The intrinsic viscosity of biological macromolecules. Progress in measurement, interpretation and application to structure in dilute solution. Progress in Biophysics and Molecular Biology, 68, 207–262. Henry, D. C. (1931). The cataphoresis of suspended particles. Part I. The equation of cataphoresis. Proceedings of the Royal Society of London, 133, 106–129. Jucker, B. A., Harms, H., & Zehnder, A. J. (1998). Polymer interactions between five Gram-negative bacteria and glass investigated using LPS micelles and vesicles as model systems. Colloids and Surfaces B: Biointerfaces, 11, 33–45. Krasikova, I. N., Gorbach, V. I., Solov’eva, T. F., & Ovodov, Yu. S. (1978). Studies on lipid A from Yersinia pseudotuberculosis lipopolysaccharide. Isolation and general characterization. European Journal of Biochemistry, 89, 287–289. Lehmann, A. K., Sørnes, S., & Halstensen, H. (2000). Phagocytosis: Measurement by flow cytometry. Journal Immunological Methods, 243, 229–242. Lowry, O. H., Rosebrough, N. J., Farr, A. L., & Randall, R. J. (1951). Protein measurement with the Folin phenol reagent. Journal of Biological Chemistry, 193, 265–275. Mueller, M., Lindner, B., Dedrick, R., Schromm, A. B., & Seydel, U. (2005). Endotoxin: Physical requirements for cell activation. Journal of Endotoxin Research, 11, 299–303. Muzzarelli, R. A. A. (2010). Chitins and chitosans as immunoadjuvants and nonallergenic drug carriers. Marine Drugs, 8, 292–312. Muzzarelli, R. A. A. (2009). Chitins and chitosans for the repair of wounded skin, nerve, cartilage and bone. Carbohydrate Polymers, 76, 167–182. Ovodov Yu, S., Gorshkova, R. P., & Tomshich, S. V. (1971). Chemical and immunochemical studies on Pasteurella pseudotuberculosis lipopolysaccharides—I. and general characterization. Immunochemistry, 8, Isolation 1071–1078. Peri, F., Piazza, M., Calabrese, V., Damore, G., & Cighetti, R. (2010). Exploring the LPS/TLR4 signal pathway with small molecules. Biochemical Society Transactions, 38, 1390–1395. Philippova, O. E., Volkov, E. V., Sitnikova, N. L., Khokhlov, A. R., Desbrieres, J., & Rinaudo, M. (2001). Two types of hydrophobic aggregates in aqueous solutions of chitosan and its hydrophobic derivative. Biomacromolecules, 2, 483–490. Raetz, C. R. (1990). Biochemistry of endotoxins. Annual Review of Biochemistry, 59, 129–170. Santos, N. C., Silva, A. C., Castanho, M. A. R. B., Martins-Silva, J., & Saldanha, C. (2003). Evaluation of lipopolysaccharide aggregation by light scattering spectroscopy. ChemBioChem, 4, 96–100. Sasaki, H., & White, H. S. (2008). Aggregation behavior of an ultra-pure lipopolysaccharide that stimulates TLR-4 receptors. Biophysical Journal, 95, 986–993. Setsukinai, K. I., Urano, Y., Kakinuma, K., Majima, H. J., & Nagano, N. (2003). Development of novel fluorescence probes that can reliably detect reactive oxygen species and distinguish specific species. The Journal of Biological Chemistry, 278, 3170–3175. Schatz, C., Domard, A., Viton, C., Pichot, C., & Delair, T. (2004). Versatile and efficient formation of colloids of biopolymer-based polyelectrolyte complexes. Biomacromolecules, 5, 1882–1892. Spirin, A. S., Belozerskii, A. N., Kudlai, D. G., Skavronskaya, A. G., & Mitereva, V. G. (1958). Changes in the composition of nucleic acids of experimentally evolved saccharolytic intestinal bacteria. Biokhimiya (Moscow), 23, 154–163. Solov’eva, T. F., Davydova, V. N., Krasikova, I. N., & Yermak, I. M. (2013). Marine compounds with therapeutic potential in gram-negative sepsis. Marine Drugs, 11, 2216–2229. Tobias, P. S., Soldau, K., Gegner, J. A., Mintz, D., & Ulevitch, R. J. (1995). Lipopolysaccharide binding protein-mediated complexation of lipopolysaccharide with soluble CD14. Journal of Biological Chemistry, 270, 10482–10488. Tomshich, S. V., Gorshkova, R. P., & Ovodov, Y. (1985). Structural studies on the core of the lipopolysaccharides from Yersinia pseudotuberculosis. Khimiya Prirodnykh Soedinenij, 6, 751–754. Tomshich, S. V., Gorshkova, R. P., Elkin, Y. N., & Ovodov Yu, S. (1976). Lipopolysaccharide from Yersinia pseudotuberculosis type 1B. A structural study of O-specific chains. European Journal of Biochemistry, 65, 193–199. Ulevitch, R. J. (2000). Molecular mechanisms of innate immunity. Immunological Research, 21, 49–54. Van Amersfoort, E. S., Van Berkel, T. J., & Kuiper, J. (2003). Receptors, mediators, and mechanisms involved in bacterial sepsis and septic shock. Clinical Microbiology Reviews, 16, 379–414. Volod’ko, A. V., Davydova, V. N., Chusovitin, E. V., Sorokina, I. V., Dolgikh, M. P., Tolstikova, T. G., et al. (2014). Soluble chitosan–carrageenan polyelectrolyte complexes and their gastroprotective activity. Carbohydrate Polymers, 101, 1087–1093. Wolfrom, M. L., & Shen Han, T. M. (1959). The sulfonation of chitosan. Journal of the American Chemical Society, 81, 1764–1766. Yen, M. T., Yang, J. H., & Mau, J. L. (2009). Physicochemical characterization of chitin and chitosan from crab shells. Carbohydrate Polymers, 75, 15–21. Yermak, I. M., Davidova, V. N., Gorbach, V. I., Luk’yanov, P. A., Solov’eva, T. F., Ulmer, A. J., et al. (2006). Forming and immunological properties of some lipopolysaccharide-chitosan complexes. Biochimie, 88, 23–30.