Biomacromolecules 2001, 2, 126-133
126
Screening of Chitosans and Conditions for Bacterial Flocculation Sabina P. Strand,*,† Marianne S. Vandvik,† Kjell M. Vårum,‡ and Kjetill Østgaard† Department of Biotechnology, The Norwegian University of Science and Technology (NTNU), N-7034 Trondheim, Norway, and Norwegian Biopolymer Laboratory (NOBIPOL), Department of Biotechnology, The Norwegian University of Science and Technology (NTNU), N-7034 Trondheim, Norway Received August 22, 2000
Chitosans with different chemical compositions and molecular weights have been evaluated as flocculants of Escherichia coli suspensions. The flocculation performance of chitosans at different conditions (pH, ionic strength) was followed by residual turbidity measurements. For precise comparison, the chitosan concentrations corresponding to 75% flocculated bacteria (x75) were calculated from a mathematical function fitted to the measured data. At all conditions, an increase in the fraction of acetylated units (FA) resulted in lower x75 and thereby better flocculation efficiency. Especially the most acetylated chitosans (FA 0.49 and FA 0.62) were excellent flocculants. An increase in FA from 0.002 to 0.6 caused a 10-fold reduction in necessary concentrations, at both pH 5 and 6.8. pH was a rather insignificant factor in the range 4-7.4, further pH increase led to either increase of necessary doses at low FA or sudden ceasing of flocculation at high FA. The chitosans flocculated in a broad range of molecular weights, although an increase in molecular weight was a favorable factor. Increase in ionic strength caused a severalfold reduction in x75 for all chitosans and considerable broadening of flocculation intervals. Introduction Biopolymers receive more and more attention as possible alternatives for synthetic polymers in many technological processes involving separation or stabilization of dispersed systems, such as in drinking water and wastewater treatment, sludge dewatering, or downstream processing. The prevailing use of synthetic polymers and/or aluminum as major water treatment chemicals has raised questions about their possible adverse effects on human health. One of the most promising natural polymers is chitosan, made commercially by heterogeneous alkaline de-N-acetylation of chitin abundant in the exoskeleton of crustaceans. The term chitosan refers to the whole family of acidic soluble linear heteropolysaccharides composed of varying amounts of (1f4)-linked 2-acetamido-2-deoxy-β-D-glucopyranose (GlcNAc) and 2-amino-2-deoxy-β-D-glucopyranose (GlcN). The chemical composition is described by the molar fraction of GlcNAc units (FA), which can be varied from 0 to 0.6. Chitosans are usually polydisperse with respect to molecular weight and the average molecular weights of commercial products typically range from 104 to 106 g/mol. The cationic nature of chitosans makes them especially interesting for flocculation of negatively charged particles. Adsorption and flocculation by synthetic cationic polymers have been intensively studied in many different systems.1-6 One of the conclusions is that highly charged polymers cause aggregation by way of charge neutralization, and their efficiency is independent of molecular weight,3,4 while † ‡
Department of Biotechnology. Norwegian Biopolymer Laboratory, Department of Biotechnology.
weakly charged polymers adsorbing in loops and tails conformation flocculate predominantly by bridging.2,4 The third suggested mechanism is patch flocculation due to the attraction between oppositely charged domains on particles partially covered with adsorbed polymer.5,7 The actual mechanism is highly dependent on the relative rates of polymer adsorption, polymer chain rearrangements, and particle collision, which are in turn affected by particle concentration and mixing conditions.6,8 At high particle concentrations, the polymer adsorption rate may be slow compared to particle collision rate, and fast nonequilibrium bridging9,10 by the polymer adsorbed in a transient conformation with loops and tails is often observed. Despite rapidly growing interest in chitosan, there have been only a few fundamental studies on flocculation by chitosans. A thorough investigation of flocculation of Escherichia coli disintegrates revealed that nonequilibrium bridging was the main mechanism of flocculation in this system and that chitosan exhibited a hydrogen bonding capacity toward cell debris.11 Further, the flocculation dosages were shown to increase with increasing FA. Molecular weight and ionic strength had only weak effects on flocculation in this study. In a recent report, latex particles were flocculated by a variety of chitosans by charge neutralization enhanced by the charged patch mechanism, and the efficiency of chitosan was greatly reduced upon increasing the pH from acidic to neutral.12 Chitosans were also tested as flocculants of kaolin suspensions, showing that flocculation efficiency depended on FA and molecular weight.13 Demarger and Domard have also studied the interactions between chitosan and undecylenic acid disper-
10.1021/bm005601x CCC: $20.00 © 2001 American Chemical Society Published on Web 11/29/2000
Biomacromolecules, Vol. 2, No. 1, 2001 127
Chitosans and Conditions for Bacterial Flocculation
sions, resulting in either flocculation or redispersion. These were shown to be mainly electrostatic, and the effects of both physicochemical parameters and structural parameters related to chitosan were examined in this system.14,15 Most other studies dealing with chitosan are based on the use of only one particular chitosan type, usually some commercial product with values of FA 0-0.2. Chitosan and its derivatives were shown to be good flocculants of E. coli suspensions, flocculating mainly by bridging.16 In another study, chitosan was reported to be as effective as, or even more effective than synthetic organic polymers and appeared to work over a wider range of turbid waters, especially highly turbid waters.17 Chitosan has been shown to be a potential coagulant of bentonite suspensions,18 and different pretreatment conditions to prepare optimal chitosan coagulants for bentonite suspensions were evaluated.19 Chitosan has also been used for stabilization and destabilization of oil in water emulsions.20,21 A standard jar apparatus is routinely applied to evaluate the performance of flocculants in batch systems.3,19,22 Their major disadvantages are large sample volumes and limited number of samples that can be run simultaneously. These can become important limitations when working with bacterial cultures or another living material or with particular flocculants that are not commercially available. For screening purposes as in this study, a down-sized assay in tubes was therefore used. It is often reported that high charge density is the crucial factor for effective flocculation by cationic polymers. However, preliminary experiments in our laboratory showed that especially low charge chitosans, that is, with high FA values, were excellent flocculants of E. coli suspensions. To our knowledge, a systematic wide-range screening of different chitosans at different conditions for flocculation of bacteria has not yet been reported in the literature. The objective of this work was to compare the performance of a wide range of different chitosans for flocculation of E. coli suspensions to investigate the importance of variables such as the chemical composition, degree of polymerization, pH, and ionic strength. This will form the basis for subsequent studies on adsorption behavior and electrokinetic properties in order to elucidate the actual flocculation mechanisms and interactions involved. Materials and Methods Bacteria and Culture Conditions. E. coli DSM 498 purchased from German Collection of Microorganisms (DSM) was chosen as a model organism. The cells were grown on Luria-Bertani medium (10 g of trypton, 5 g of yeast extract, 5 g of NaCl, 1000 mL of distilled water, pH 7.0) at 37 °C in 500 mL shake flasks overnight. Chitosans. Chitosan samples were either provided by Pronova Biopolymers (Drammen, Norway) or were prepared by homogeneous deacetylation of shrimp chitin. All of them were converted into water-soluble hydrochloride salts (chitosan-HCl)23 and the freeze-dried chitosans were stored at -20 °C. The fractions of acetylated units (FA) were determined by 1H NMR spectroscopy.24 To obtain a series
Table 1. Characteristic Properties of the Chitosan Samples Applieda fraction of acetylated units (FA)
intrinsic viscosity (η, mL/g)
number-average degree of polymerization (DPn)
0.002 0.009 0.01 0.01 0.13 0.37 0.49 0.62
670 390 610 800 790 620 1270 820
880 370 740 1110 1020 670 1330 820
a Fraction of acetylated units (F ) was determined by 1H NMR A spectroscopy.24 Intrinsic viscosities were measured according to Draget 26 et al.
of chitosans with the same FA and different molecular weight, depolymerization of chitosan with nitrous acid was performed as described by Allan and Peyron25 and the chitosans were reduced conventionally with NaBH4. The number-average degrees of polymerization (DPn) were calculated from molecular weights determined from Mark-HouwinkSakurada equations23 on the basis of measured intrinsic viscosities26 using a calibration against chitosan samples with known Mn (from osmometry) and intrinsic viscosity.23 The chemical compositions, intrinsic viscosities, and the numberaverage degree of polymerizations of chitosans used in this study are shown in Table 1. Solutions of chitosans with concentration 1-2 mg/mL were prepared by gentle shaking in Milli-Q grade water at 5 °C overnight and adjusted to appropriate ionic strength with NaCl (usually 0.1 M, except for the experiment where ionic strength was a variable). They were further diluted with 0.1 M NaCl to a desired concentration series. Only freshly prepared solutions (less than 24 h) stored at 5 °C were used in the experiments. Bacterial Suspensions. Living E. coli cells in the stationary phase were harvested by centrifugation (6000 rpm, 5 min), washed twice, and resuspended in an appropriate buffer, unless otherwise stated. Acetate buffers (10 mM) were used at pH 4 and pH 5, and 10 mM phosphate buffered saline solutions (PBS) were used at pH 6.2 and higher. All buffers were prepared in Milli-Q grade water and had a constant ionic strength of 0.1 M adjusted by NaCl. In the ionic strength experiment, cells were washed and resuspended in NaCl solutions with concentrations from 10-4 to 1.0 M. The pH of these unbuffered suspensions was within 6-6.5. The cells were always resuspended to give a optical density (OD) at 620 nm of 3.0-3.2 when recorded on a Medispec III UV/VIS spectrophotometer, zero-set against a buffer. The working concentration in the flocculation experiments was 1.0-1.3 g/L dry weight, which corresponded to 3-4 × 109 cells/mL as determined from both colony-forming units counts and direct microscopic counting. The dry weights were obtained after filtration of 5 mL of culture through a membrane filter (Millipore), washing with distilled water, and drying at 105 °C to constant weight. Flocculation Procedure. All experiments were performed at ambient temperature (20-23 °C). The flocculation assay was performed in 13 mL graduated polystyrene tubes with
128
Biomacromolecules, Vol. 2, No. 1, 2001
conical bottoms (Sarstedt). Eight milliliters of bacterial suspension was pipetted in the tubes, and 1 mL of chitosanHCl solution was added under stirring on a Vortex mixer (2000 rpm, 5 s) to ensure proper mixing. A corresponding blank was prepared with 1 mL of 0.1 M NaCl. The tubes were shaken for 10 min on a rotary shaker at 200 rpm (Heidolph) and allowed to stand for 120 min before the first sample of supernatant for OD measurement was taken from the top layer. The tubes were then left overnight to complete the sedimentation process, and the second supernatant sample was pipetted from the middle of the tube after 24 h. The optical densities of the supernatants were measured at 620 nm on a Medispec III UV/VIS spectrophotometer, zero-set against the actual buffer. The flocculation was expressed as the decrease of turbidity relative to blank (referred to as % bacteria flocculated), calculated as (1 - (OD sample/OD blank)) × 100. In the ionic strength experiment, flocculation was expressed as percent decrease of initial turbidity to include the possible aggregation of bacteria by salt at high ionic strengths. The nonzero values at zero chitosan concentration thus represent the fraction of bacteria flocculated by salt alone. All samples were run in duplicates. Determination of Critical Concentrations x75. The flocculation curves were obtained by plotting the relative fraction of bacteria flocculated as a function of chitosan concentration, expressed as milligrams of chitosan-HCl per gram of cell dry weight. Fitting a suitable mathematical function to the data set on a purely empirical basis generated smooth continuous graphs. When the number of parameters were restricted to a minimum, it was found that a sigmoidal Gompertz curve of the type y ) a exp(-exp[(x - x0)/b]) was superior compared to other three-parameter models. Using Sigma plot 5.0, Gompertz curves were fitted to the initial points up to the maximal value of flocculation, provided at least one point was known in the steepest part of the curve, i.e., between 20 and 80% bacteria flocculated. Generally, a nearly perfect fit with regression coefficient R2 > 0.99 was obtained. To simplify comparison of different data sets, the chitosan concentrations corresponding to 75% bacteria flocculated were then calculated from the plot equations and referred to as x75. Results Method of Evaluation. The flocculation assay was previously thoroughly tested and optimized (data not shown). The typical difference between samples run in duplicate was generally below 1%. Higher deviations, up to 10-15%, were occasionally recorded in the steepest parts of the flocculation curves, where flocculation was just initiated or cells became restabilized. Reproducibility of independent experiments can be assessed from x75 determinations presented in Figure 3. In all experiments reported, the maximal flocculation obtained reached 95-98% after 2 h and over 98% after 24 h of sedimentation unless otherwise stated. Effect of Bacterial Concentration. To test how the changes in bacterial concentration affected the flocculation, four suspensions of E. coli with dry weights corresponding to 0.19, 0.50, 1.02, and 2.06 g/L were flocculated by chitosan
Strand et al.
Figure 1. Flocculation of E. coli with chitosan FA 0.13 (DPn 1020) as a function of bacterial concentration after 2 h (A) and 24 h (B, C) of sedimentation. Bacteria were resuspended in acetate/NaCl buffer with pH 5.0 and ionic strength 0.1 M.
FA 0.13 (DPn 1020). The flocculation curves after 2 and 24 h of sedimentation are presented in parts A and B of Figure 1 on a linear abscissa scale, and the latter also in Figure 1C on a logarithmic scale. The three distinct regions of flocculation: initial increase, optimal flocculation, and restabilization, became more apparent in the logarithmic plots. The flocculation profile of the most diluted suspension (0.19 g/L) differed most clearly from the others. The increase in the bacterial concentration to 0.5 g/L resulted in better flocculation performance as shown by the reduction of the initial effective chitosan concentration and broadening of the curves (Figure 1). Further increase in the bacterial concentration affected only the chitosan concentration at the onset of restabilization. Also, clearer supernatants were obtained at the higher concentrations. The negative values of percent flocculated at high chitosan concentrations demonstrate the ability of chitosan to prevent aggregation and, consequently, to reduce the sedimentation velocity compared to blank.
Chitosans and Conditions for Bacterial Flocculation
Figure 2. Flocculation of E. coli with different chitosans varying in FA, after 2 h (A) and 24 h (B) of sedimentation. Bacteria (3-4 × 109 cells/mL) were resuspended in PBS with pH 6.8 and ionic strength 0.1 M. The symbol explanation is given in the lower graph.
Although only limited information about the kinetics could be obtained from this batch method, by comparing part A of Figure 1 and parts B and C of Figure 1, it is evident that the flocculation rate decreased with decreasing bacterial concentration. After 24 h, the most diluted suspension (0.19 g/L) approached the others, giving a broader flocculation range (Figure 1B,C). However, on comparison of estimated x75 values, an amount of chitosan that was twice as large was necessary to flocculate 75% of bacteria in suspension with 0.19 g/L dry weight as in suspensions with 0.5-2 g/L (calculations not included). Effect of Chemical Composition. Examples of flocculation of E. coli with chitosans with FA values from 0.002 to 0.62 and DPn from 670 to 1320 at pH 6.8 and ionic strength 0.1 M are shown in parts A (2 h) and B (24 h) of Figure 2. The chitosans with high FA were effective flocculants of E. coli at much lower concentrations compared to those with low FA. As FA decreased, both the onsets of flocculation and restabilization moved toward higher chitosan concentrations. Consequently, there were clear differences between the concentration ranges where the optimal flocculation occurred for different chitosans. All chitosans flocculated in a broad range of concentrations. The relative width of the optimal flocculation region increased with time and at 2 h also with increasing FA (Figure 2A). After 24 h, most chitosans reached the same width of optimal flocculation interval, except FA 0.002 showing a more narrow flocculation band after both 2 and 24 h of sedimentation (Figure 2B). To compare the chitosans and establish a more general relationship between FA and flocculation efficiency, a
Biomacromolecules, Vol. 2, No. 1, 2001 129
Figure 3. Chitosan-HCl concentration at 75% flocculation (x75) as a function of FA, after 24 h of sedimentation. Bacteria (3-4 × 109 cells/ mL) were resuspended in PBS pH 6.8 (A) and acetate/ NaCl pH 5.0 (B) with ionic strength 0.1 M.
variable denoted x75 was calculated from each flocculation curve as described in the methods section. Figure 3 A shows the x75 plotted against FA at pH 6.8. Each point represents the results of an independent experiment. An increase in FA resulted in a decrease of the chitosan concentration necessary to flocculate E. coli cells. The magnitude of differences among x75 for different chitosans is evident on the logarithmic scale, x75 of chitosan FA 0.62 (DPn 820) was about a 10fold lower than that of FA 0.01 (DPn 1110). It is less apparent from Figure 3A that the chitosan FA 0.002 (DPn 880) had about twice as large x75 as FA 0.01 (DPn 1110), since they are difficult to distinguish on the linear abscissa scale. The relationship between FA and x75 was also examined at pH 5, where all chitosans are fully soluble and their charge density is high. The results are given in Figure 3B. A trend very similar to that at pH 6.8 (Figure 3A) was found, with a dramatic increase in effective chitosan concentrations associated with the decrease in FA. Effect of Degree of Polymerization. It is important to note that the chitosans used to establish the relationships between FA and concentration in Figures 2 and 3 had DPn ranging from 670 to 1330. To study the effect of degree of polymerization on flocculation efficiency in more detail, two series of chitosans with FA 0.01 and FA 0.49 were partially degraded and used as flocculants of E. coli suspensions. The results are presented in Figure 4. The two chitosans exhibited reasonably similar response. In both cases, the x75 concentrations decreased with increasing DPn.
130
Biomacromolecules, Vol. 2, No. 1, 2001
Figure 4. Chitosan-HCl concentration at 75% flocculation (x75) as a function of DPn for chitosans FA 0.01 and FA 0.49, after 24 h of sedimentation. Bacteria (3-4 × 109 cells/mL) were resuspended in PBS with pH 6.8 and ionic strength 0.1 M.
Strand et al.
Figure 6. Flocculation of E. coli with chitosan FA 0.49 at different pH, after 24 h of sedimentation. Bacteria (3-4 × 109 cells/mL) were resuspended in acetate/NaCl buffers pH 4.0-5.0 or PBS buffers pH 6.2-7.6 with ionic strengths of 0.1 M.
Figure 5. Chitosan-HCl concentration at 75% flocculation (x75) as a function of pH for chitosans with varying FA, after 24 h of sedimentation. Bacteria (3-4 × 109 cells/mL) were resuspended in acetate/ NaCl buffers pH 4.0-5.0 or PBS buffers pH 6.2-7.8 with ionic strengths of 0.1 M.
This effect was more pronounced for samples with the lowest DPn and declined toward a more stable level as the DPn increased. DPn also affected the width of the optimal flocculation interval for both chitosan types. As DPn decreased, wider flocculation profiles were observed and the onset of restabilization moved to higher chitosan concentrations (data not shown). Effect of pH. To determine how pH may influence the flocculation of E. coli by chitosans, several experiments were carried out in the range of pH 4-7.8. The ionic strength of 0.1 M was kept constant by addition of NaCl to the buffer. The x75 concentrations as a function of pH are shown in Figure 5. The chitosan with FA 0.002 flocculated without any significant pH dependence in the pH region 4-7.4. At pH 7.8, much higher doses were abruptly necessary to initiate flocculation, and no restabilization occurred at this pH. A very similar trend was observed also for chitosans with FA 0.009 (data not shown) and FA 0.13 (Figure 5). The chitosan with FA 0.37 flocculated from pH 4 to pH 7.4, while no flocculation was observed as pH was raised to 7.8. It was found that it still flocculated at pH 7.6, however less efficiently, with less than 90% bacteria flocculated (data not shown).
Figure 7. Flocculation of E. coli with chitosans FA 0.002 (A) and FA 0.49 (B) at different ionic strengths, after 24 h of sedimentation. Bacteria (3-4 × 109 cells/mL) were resuspended in NaCl solutions (pH 6-6.5).
The original flocculation curves for the chitosan with FA 0.49 at four different pH values are shown in particular in Figure 6. This chitosan will remain soluble in this pH range.27 At pH 7.6, only a partial flocculation in a very narrow concentration range occurred, and at higher pH no flocculation was observed at the concentration scale examined. The x75 concentrations slightly decreased as the pH increased from 4 to 7 (Figure 5). Effect of Ionic Strength. The flocculation of E. coli suspensions by four different chitosans at NaCl concentrations ranging from 10-4 to 0.5 M was performed, with data for chitosan FA 0.002 and FA 0.49 shown in Figure 7. The ionic strength of the medium influenced strongly both bacteria and chitosans. The increase in the ionic strength to
Chitosans and Conditions for Bacterial Flocculation
Figure 8. Chitosan-HCl concentration at 75% flocculation (x75) as a function of ionic strength for chitosans with varying FA, after 24 h of sedimentation. Bacteria (3-4 × 109 cells/mL) were resuspended in NaCl solutions (pH 6-6.5).
0.5 M and higher (data not shown) resulted in nearly complete removal of E. coli cells from supernatant within 24 h, even without any chitosan present (Figure 7). Upon addition of chitosan, the flocculation by electrolyte was reversed and the bacteria were restabilized in suspension. This is apparent only for FA 0.49 (Figure 7B), but a similar trend was observed also for other chitosans at higher concentrations (data not shown). At 0.1 M, approximately 15% of the cells were flocculated after 24 h, while at ionic strengths lower than 0.05 M, no flocculation by salt was observed at all. Another obvious effect of salt is broadening of optimal flocculation range intervals. As expected, the electrolyte behaved as an aid flocculant by lowering the x75 concentrations; see Figure 8. Comparing the x75 of chitosans at 10-4 M with that at 10-1 M, a factor of 2 reduction for FA 0.002 and factor of 10 reduction for FA 0.49 were observed. It is also apparent that chitosan with FA 0.002 was considerably less efficient than the others, especially at low ionic strength. Discussion Method. On evaluation of both the standard deviations of the duplicates and the results of independent experiments (Figure 3), our flocculation assay gave satisfyingly reproducible results. Many samples could be tested simultaneously, allowing one batch of bacteria for the whole series examined, and with very low amounts of flocculants consumed. The determination of a critical flocculation concentration by extrapolation to 100%11 in order to compare different chitosans, was rather difficult and imprecise due to the sigmoidal shape of the flocculation curves. Similarly, the broad regions of optimal flocculation complicated explicit determination of an optimal flocculation concentration. Consequently, another approach was applied, based on the interpolation of smooth curves fitted to the measured data. The Gompertz curve was chosen solely because it gave the best fit to all data with a restricted number of parameters, and there is no particular theoretical basis for this choice. Bacterial Concentration. It is generally accepted that increase in the particle concentration in a given system leads to a higher frequency of collisions and thereby a faster
Biomacromolecules, Vol. 2, No. 1, 2001 131
flocculation rate. It has also been observed that different flocculation mechanisms may operate at different particle concentrations due to the kinetic aspects of the polymer adsorption.1,6 The main goal of the tests carried out in our work was to assess how critical it is to keep the bacterial concentration constant. It is evident that at least a 10-fold decrease in bacterial concentration was needed to get a significant deviation from the typical flocculation pattern, showing higher x75 concentration and lower optimal range (Figure 1). Since the cell dry weights during all our experiments were kept between 1 and 1.3 g/L, no significant effects on flocculation performance should be expected due to variations in bacterial concentrations at standard test conditions. Chemical Composition: FA. The chemical composition of the chitosan, i.e., FA, determines chitosan solubility, charge density, and conformation. The acetyl group exhibits also a certain degree of hydrophobicity and is rather bulky. As pointed out in the introduction, charge density has been found to be one of the most important factors for performance of chitosans, generally with a more effective flocculation at high charge density, that is at lower FA.11-13 Hence, it was rather unexpected that chitosans in our investigations showed the opposite pattern; the higher FA, the better flocculation performance (Figures 2-3). At pH 6.8, one could expect that the decrease in solubility associated with lower FA27 may account for higher doses, but the same relationship between FA and chitosan concentration was obtained also at pH 5.0 where all chitosans are fully soluble (Figure 3B). Regarding solubility, it is important to stress that chitosan is adsorbed to the bacterial cells and, consequently, the solubility is not a necessary condition for flocculation, as also suggested by Parazak et al.22 The limited solubility of chitosans may become important when the excess amounts of polymer appear in the solution and may result in sweep-out flocculation. The relationship between FA and charge density at a given pH is not so straightforward, for several reasons. First, the literature reporting pKa values for chitosans does not seem consistent. Anthonsen and Smidsrød28 found a pKa of 6.6, irrespective of FA. On the other hand, Sorlier et al.29 reported increase of pKa with FA in the range from 6.2 to 7.1. Another discrepancy is found concerning the changes in the apparent pKa with degree of ionization and the values of the intrinsic pKa for different chitosans.28,29 This could partly be due to differences in molecular weight of chitosans used. Further, at higher ionic strengths or when interacting with other charged species, the long-range effects on chitosan may be screened and, consequently, the apparent pKa value may increase. In any case, the presence of acetylated units in chitosan clearly seemed to be a highly favorable factor for flocculation of E. coli cells both at pH 5.0, where the chitosans are highly charged, and at pH 6.8, where a major part of the amino groups is in free amino form. The influence of factors other than purely electrostatic forces, such as contribution from hydrophobic interactions or possibly even hydrogen bonding, cannot be neglected. The latter was also emphasized by Agerkvist.11
132
Biomacromolecules, Vol. 2, No. 1, 2001
The effect of FA on the chitosan conformation in solution was studied by Anthonsen et al.23 The authors concluded that the presence of bulky acetylated units increased the stiffness and extension of the chain at all ionic strengths. The possible consequences of this are beyond the scope of this paper and will be discussed elsewhere. Degree of Polymerization. The effect of DPn of a polymer on flocculation is closely connected to the dispersions studied and mechanisms involved. It has been found that low molecular weight polymers flocculate predominantly by charge neutralization, while for high molecular weight polymers the bridging mechanism often dominates.1,6 In the case of charge neutralization, no effect of DPn on flocculation was observed. The relationship between the DPn and x75 concentration was strikingly similar for both chitosans examined (Figure 4), and accordingly, the effect of DPn appears unrelated to FA. The small increase in efficiency with increasing DPn cannot therefore explain the differences observed between chitosans discussed above. The finding that chitosans were more effective flocculants of E. coli at higher DPn may point at least to a contribution from the bridging mechanism, proposed also by other authors in similar systems.11,16 However, considering data for short polymers with DPn < 250 (Figure 4), another mechanism may seem more likely. pH. Chitosan is a weak polyelectrolyte, and its charge density is thus greatly influenced by pH. The complex relationship between the charge density, FA, and pH is discussed above, and consequently, one should be rather cautious estimating the true charge on the different chitosans under actual conditions. As mentioned earlier, the flocculation potential of chitosans is traditionally connected to their charge density. Thus, it is often found that the best flocculation results are obtained in the pH region where the biopolymer is highly charged, usually below pH 5.12,16,18 As shown in Figure 5, no such pH dependency was observed. The chitosans with low FA (FA 0.002, FA 0.009, FA 0.13) had almost identical x75 concentrations for the whole range of pH 4.0-7.5, where the charge density shows an extreme variation. One possible explanation, at least for low pH, may be offered by the counterion condensation theory developed by Manning.30,31 If the stoichiometric charge density of a polymer is higher than a certain critical value, determined by the minimal allowed distance between charges, it is reduced by counterion condensation so that the net effective charge density due to polymer and counterions is kept constant. In our case, the effective charge density of low acetylated chitosans at low pH may be much lower than the expected stoichiometric value. It may also be mentioned that the pH measured in bulk is usually very different from that close to the surface.32 Consequently, the adsorbed chitosan layer may face a microenvironment rather different from bulk conditions and the true charge density may deviate from simple theoretical predictions. The much higher concentrations of chitosans with low FA needed to initiate the flocculation at pH 7.8 and the absence of restabilization observed resemble the so-called sweep out flocculation with excess insoluble chitosan, where cells
Strand et al.
become physically entrapped in precipitated mass. The chitosan FA 0.49 is soluble at neutral pH values, and no such type of flocculation at pH 7.8 was observed. It is more complicated to interpret the results of highacetylated chitosans (Figure 5). The slight decrease in x75 concentrations associated with increase in pH up to 7.4, if significant, may be merely the result of higher adsorption and stronger contribution from hydrophobic interactions due to acetylated residues. Similar observations were made by Parazak et al.22 The lower restabilization concentrations at higher pH (Figure 6) may also be due to higher amounts of chitosan adsorbed resulting in charge reversal or sterical hindrance. The sudden ceasing of flocculation at pH > 7.4 remains unexplained, but it seems that a certain amount of charged residues is essential for flocculation by all chitosan types. Ionic Strength. According to classical theories of colloidal stability, bacteria as any other charged colloids are stabilized in suspension due to the electrostatic repulsion arising from the extended double layers. The extension of the double layer and, consequently, the range of repulsion forces is a function of the ionic strength in the surrounding medium. With increasing electrolyte concentration, the double layer becomes more compressed and the range of the repulsive forces is shortened. As bacteria can approach each other closer, the attractive van der Waals and other forces may overcome the electrostatic barrier and cause aggregation more easily. This is clearly demonstrated in our study by flocculation of E. coli at ionic strengths higher than 0.5 M without any chitosan present (Figure 7). The ionic strength greatly influences the chitosan conformation. The increasing counterion concentration will screen the charged groups, and reduced segment-segment repulsion will result in a more flexible chain conformation.23 The screening of charges also caused the pKa of chitosan to increase from 6.6 to 7 as the ionic strength increased from 0.04 to 0.24 M.28 It has been shown that flocculation by cationic polymers at low ionic strengths occurs over a very narrow range of polymer concentrations.5 Further, the optimal flocculation concentrations decreased with increasing ionic strength.1,4 Both findings are in agreement with our results shown in Figures 7 and 8. Due to the decreased electrostatic repulsion at higher ionic strengths, the cells may more easily approach each other and thereby all kinds of attractive interactions become more likely at low amounts of adsorbed polymer. The fact that FA 0.002 had much higher x75 than all the other chitosans at any ionic strength (Figure 8) illustrates the importance of GlcNAc units, regardless of salt concentration. For chitosans with high FA, the increase in ionic strength brought about a larger reduction of x75 (Figure 8). Possibly, the attractive interactions connected with high FA chitosans may be of relatively short-range nature and will be operative first after the compression of the double layer. Another interesting point is the ability of chitosan to reverse the flocculation by salt (Figure 7B). Since this ability strongly increased with the FA, it is probably connected to sterical hindrance due to the bulky N-acetyl groups.
Chitosans and Conditions for Bacterial Flocculation
Conclusions The presence of the acetylated residues in the chitosan chain was a favorable factor for flocculation of E. coli. The increase in FA from 0.002 to 0.6 caused approximately 10fold reduction in necessary concentrations, at both pH 5 and 6.8. The chitosans with higher molecular weights flocculated better, but no dramatic effect was observed in the range of DPn 700-1300. Flocculation was observed also for chitosans with DPn much lower than 200. pH was an insignificant factor in the interval from 4 to 7.4; further increase in pH resulted in either sudden ceasing of flocculation (high FA chitosans) or a large increase of doses (low FA). The increase in ionic strengths led to a decrease in effective chitosan concentration and broadening of the flocculation ranges. Finally, it is important to note that the results presented here cannot be directly generalized to other bacteria without considering bacterial variations in surface properties such as charge and hydrophobicity. However, a better understanding of the interactions between bacteria and chitosans is essential as a background for possible future applications. In future studies, the adsorption and electrokinetic properties in this system will be investigated in order to gain more insight into the flocculation mechanism and the nature of the interactions involved. Acknowledgment. This work was in part supported by Jotun a.s., Norway, and by the Norwegian Research Council. We also thank to Prof. O. Smidsrød for valuable discussions. References and Notes (1) Baran, A. A.; Gregory, D. Colloid J. 1996, 58 (1), 9-14. (2) Chaplain, V.; Janex, M. L.; Lafuma, F.; Graillat, C.; Audebert, R. Colloid Polym. Sci. 1995, 273, 984-993. (3) Durand-Piana, G.; Lafuma, F.; Audebert, R. J. Colloid Interface Sci. 1987, 119 (2), 474-480. (4) Eriksson, L.; Alm, B.; Stenius, P. Colloids Surf., A 1993, 70, 4760. (5) Gregory, J. J. Colloid Interface Sci. 1973, 42, 448-456. (6) Gregory, J.; Sheiham, I. Br. Polym. J. 1974, 6, 47-59.
Biomacromolecules, Vol. 2, No. 1, 2001 133 (7) Leong, Y. K. Colloid Polym. Sci. 1999, 277, 299-305. (8) Gregory, J. Colloids Surf. 1988, 31, 231-253. (9) Adachi, Y.; Cohen Stuart, M. A.; Fokkink, R. J. Colloid Interface Sci. 1994, 167, 346-351. (10) Pelssers, E. G. M.; Cohen Stuart; M. A.; Fleer, G. J. Colloids Surf. 1989, 38, 15-25. (11) Agerkvist, I. PhD Thesis, Royal Institute of Technology, Stockholm, Sweden, 1992. (12) Ashmore, M.; Hearn, J. Langmuir 2000, 16, 4906-4911. (13) Domard, A.; Rinaudo, M.; Terrassin, C. J. Appl. Polym. Sci. 1989, 38, 1799-1806. (14) Demarger-Andre, S.; Domard, A. Carbohydr. Polym. 1993, 22, 117126. (15) Demarger-Andre, S.; Domard, A. Carbohydr. Polym. 1994, 24, 177184. (16) Baran, A. A. Colloids Surf. 1987, 31, 259-264. (17) Kawamura, S. J. Am. Water Works Assoc. 1991, 83 (10), 88-91. (18) Huang, CH.; Chen, Y. J. Chem. Technol. Biotechnol. 1996, 66, 227232. (19) Huang, CH.; Chen, S.; Pan, J. R. Water Res. 1999, 34 (3), 105762. (20) Del Blanco, L. F.; Rodriguez, M. S.; Schulz, P. C.; Agullo, E. Colloid Polym. Sci. 1999, 277, 1087-1092. (21) Pinotti, A.; Bevilacqua, A.; Zaritzky, N. Scanning 1999, 21, 354358. (22) Parazak, D. P.; Burkhardt, CH. W.; McCarthy, K. J.; Stehlin, M. P. J. Colloid Interface Sci. 1987, 123 (1), 59-72. (23) Anthonsen, M. W.; Vårum, K. M.; Smidsrød, O. Carbohydr. Polym. 1993, 22, 193-201. (24) Vårum, K. M.; Anthonsen, M. W.; Grasdalen, H.; Smidsrød, O. Carbohydr. Res. 1991, 211, 17-23. (25) Allan, G. G.; Peyron, M. In Chitin and Chitosan; Skjåk-Braek, G., Anthonsen, T., Sandford, P. A., Eds.; Elsevier Applied Science: London, U.K., 1989; pp 443-466. (26) Draget, K. I.; Vårum, K. M.; Moen, E.; Gynnild, H.; Smidsrød, O. Biomaterials 1992, 13, 635-638. (27) Vårum, K. M.; Ottøy, M. H.; Smidsrød, O. Carbohydr. Polym. 1994, 25, 65-70. (28) Anthonsen, M. W.; Smidsrød, O. Carbohydr. Polym. 1995, 26, 303305. (29) Sorlier, P.; Denuziere, A.; Hartmann, D.; Domard, A. Poster presented at the 8th International Chitin & Chitosan Conference, Yamaguchi, Japan, Sept. 21-23, 2000. (30) Manning, G. S. Q. ReV. Biophys. 1978, 11, 179-246. (31) Manning, G. S. Ber. Bunsen-Ges. Phys. Chem. 1996, 100, 909922. (32) Claesson, P. M.; Ninham B. W. Langmuir 1992, 8, 1406-1412.
BM005601X