Chemical and Rheological Properties ofan Extracellular Polysaccharide Producedby the Cyanobacterium Anabaenasp. ATCC 33047

Jose Moreno,1 M. Angeles Vargas,1 Jose M. Madiedo,2 Jose Munoz,2

Joaquın Rivas,1 Miguel G. Guerrero1

1Instituto de Bioquımica Vegetal y Fotosıntesis, Consejo Superior deInvestigaciones Cientıficas-Universidad de Sevilla, Centro deInvestigaciones Cientıficas Isla de la Cartuja, Avenida Americo Vespucio,s/n, 41092 Sevilla, Spain; telephone: +34-954489513; fax: +34-954460065;e-mail: jmfernan@cica.es2Departamento de Ingenierıa Quımica, Facultad de Quımica, Universidadde Sevilla, Sevilla, Spain

Received 7 January 1999; accepted 25 July 1999

Abstract: The cyanobacterium (blue-green alga) Ana-baena sp. ATCC 33047 produces an exopolysaccharide(EPS) during the stationary growth phase in batch cul-ture. Chemical analysis of EPS revealed a heteropolysac-charidic nature, with xylose, glucose, galactose, andmannose the main neutral sugars found. The infrared(IR) spectrum of EPS showed absorption bands of car-boxylate groups. The average molecular mass of thepolymer was 1.35 MDa. Aqueous dispersions at EPS con-centrations ranging from 0.2% to 0.6% (w/w) showedmarked shear-thinning properties (power-law behavior).Linear dynamic viscoelastic properties showed that theelastic component was always higher than the viscouscomponent. Viscous and viscoelastic properties demon-strated the absence of conformational changes withinthe concentration range studied. Stress-growth experi-ments revealed that 0.4% and 0.6% (w/w) EPS disper-sions showed thixotropic properties. A detailed compari-son of the linear dynamic viscoelasticity, transient flow,and decreasing shear rate flow curve properties wasmade for 0.4% (w/w) dispersions of xanthan gum (XG),Alkemir 110 (AG), and EPS. Viscoelastic spectra demon-strated that the EPS dispersion turned out to be more“fluidlike” than the AG and XG dispersions. The flowindexes indicated that the EPS dispersion was less shear-sensitive than that of XG, showing essentially the sameviscosity, that is, >50 s−1. The fact that viscosities of EPSand AG dispersions were not substantially differentwithin the shear-rate range covered must be empha-sized, in relation to EPS potential applications. The rhe-ological behavior of EPS dispersions indicates the forma-tion of an intermediate structure between a random-coilpolysaccharide and a weak gel. © 2000 John Wiley & Sons,Inc. Biotechnol Bioeng 67: 283–290, 2000.

Keywords: blue-green algae; cyanobacteria; exopolysac-charide; chemical properties; rheology; viscoelasticity;viscosity

INTRODUCTION

Microbial exopolysaccharides have significant commercialvalue, particularly in regard to the production of gels andmodification of the rheological properties of aqueous sys-tems. In the future they could possibly replace the plant andmacroalgae exopolysaccharides traditionally used in thefood, pharmaceutical, textile, and oil industries (Linton,1990; Sutherland, 1990). Among microbial exopolysaccha-rides, one of the most important is that fromXanthomonascampestris, known as xanthan. The characteristic rheologi-cal properties induced by this polysaccharide have deter-mined its industrial success (Ash, 1985). The rheologicalbehavior of xanthan gum reflects the ordered conformationand consequent intermolecular interactions usually adoptedin aqueous solutions (Cuvelier and Launay, 1986; Richard-son and Ross-Murphy, 1987b; Rochefort and Middleman,1987; Ross-Murphy et al., 1983). Rheology provides sig-nificant information in regard to: (a) quality control of eitherraw materials or final products; (b) process engineering; (c)optimization of formulations on the basis of the relation-ships existing between microstructure and physical proper-ties; and (d) objective texture analysis of commercial prod-ucts.

The physical properties of exopolysaccharides from thecyanobacteriaPhormidium J1, Anabaenopsis circularis,Calothrix desertica, Synechocystissp. PCC 6803, andCya-nospira capsulatahave been investigated only recently(Bertocchi et al., 1990; Lapasin and Pricl, 1995; Navarini etal., 1992). Novel flocculating and gelling properties havebeen suggested that might be exploited in the food andpharmaceutical industries. Additional applications proposed

Correspondence to:J. MorenoContract grant sponsors: CICYT, Spain; Plan Andaluz de Investigacı´on;

Spain and Acciones Integradas Hispano Brita´nicasContract grant numbers: BIO940661; CVI 0131; 1996-0232

© 2000 John Wiley & Sons, Inc. CCC 0006-3592/00/030283-08

for cyanobacterial exopolysaccharides include, among oth-ers, the conditioning of soils for improvement of their wa-ter-holding capacity, specific binding of metal cations(which could prove useful in wastewater management anddetoxification of metal-contaminated media), and elimina-tion of solid matter in water reservoirs (Bender et al., 1994;Bertocchi et al., 1990; De Philippis and Vincenzini, 1998;Plude et al., 1991).

Anabaenasp. ATCC 33047 is a halotolerant, filamen-tous, nitrogen-fixing cyanobacterium that releases into theculture medium large amounts (up to 15 g L−1) of an exo-polysaccharide during the stationary growth phase. Exo-polysaccharide production by this cyanobacterium was in-duced in the absence of a combined nitrogen source, andwas markedly enhanced in response to an increase in eitherair flow rate, temperature, or irradiance, both in batch and incontinuous culture (Moreno et al., 1998). On the other hand,Anabaenasp. ATCC 33047 also produces other commer-cially valuable chemicals, such as phycobiliproteins(Moreno et al., 1995), which are accumulated intracellular-ly. Thus, cultures of this organism appear to be suitable forcarrying out an integrated production of some importantintracellular and extracellular compounds.

In this work, data concerning some chemical and rheo-logical properties of the purified exopolysaccharide releasedby Anabaenasp. ATCC 33047 in diazotrophic batch cul-tures are presented.

MATERIALS AND METHODS

Microorganism and Culture Conditions

Anabaenasp. ATCC 33047 was grown in batch culture at40°C in 5-cm-deep, 1-L capacity Roux flasks on the me-dium previously described (Moreno et al., 1995), containing85 mM NaCl, 50 mM NaHCO3, 8 mM KCl, 1 mM K2HPO4,0.5 mM MgSO4, and 0.35 mM CaCl2, as well as a supply ofessential micronutrients and Fe-EDTA (Arnon et al., 1974).Cultures were continuously bubbled with air (250 L L−1

h−1) supplemented with 6% (v/v) CO2, and laterally andcontinuously illuminated with mercury-halide lamps at asurface irradiance of 460mE m−2 s−1.

Isolation and Purification of EPS

Cell suspensions were harvested at the late stationary phaseof growth (6-day culture), heated at 50°C for 10 min in thepresence of 0.3M NaCl and 0.03M Na2-EDTA (pH 10),and centrifuged at 27,000g for 45 min at room temperature.The supernatants were dialyzed (Visking dialysis tubing)against 10 to 20 volumes of deionized water for 96 h at 4°C,with two daily changes. Three volumes of 2-propanol wereadded to the dialyzed supernatants. The precipitated poly-mer was redissolved in deionized water, and the precipita-tion procedure was repeated twice more. The EPS was then

suspended in deionized water, lyophilized to dryness, andthe material used for analytical work.

Analytical Procedures

The total carbohydrate content of the EPS was measured bythe phenol–sulfuric acid method (Dubois et al., 1956), usingglucose as a standard, and the total protein content wasestimated by the method previously described by Herbert etal. (1971), with bovine serum albumin as a standard. Eachvalue given is the mean of three independent determina-tions. Pellets for infrared (IR) analysis were obtained bycarefully grinding 2 mg of polysaccharide with 200 mg ofdry KBr and then pressing in a mold. The IR spectrum wasobtained using a Perkin-Elmer 398 IR spectrophotometer.Elemental analysis was carried out using an elemental ana-lyzer (Model 1106/R, Carlo Erba Strumentazione) con-nected to an integrator (Model 3390A, Hewlett-Packard).

For sugar determination, the EPS was hydrolyzed with4N trifluoroacetic acid (TFA) at 100°C in sealed tubes for 4h (6 mg EPS per milliliter TFA). The solution was evapo-rated to dryness with additions of methanol to remove TFA.The hydrolysate was dissolved in deionized water, neutral-ized with NaOH to pH 8.2, and passed through an AmberliteIR-120 column (Sigma Co.) (pretreated with 50 mM HCl),using water as eluant to isolate the neutral and acid sugars.The elute was again passed through an Amberlite IRA-400column (Sigma) (pretreated with 10% [v/v] acetic acid),using water as eluant for the neutral sugars and 10% (v/v)acetic acid to elute the uronic acids.

Uronic acids were analyzed by the carbazole method(Bitter and Muir, 1962), using galacturonic acid as a stan-dard. Neutral sugars were identified by descending paperchromatography on Whatman No. 3 paper, using acetone–butanol–water (7:2:1) as the solvent system. The sugarswere visualized by spraying the chromatograms withAgNO3 (1 mL of an AgNO3-saturated solution in 200 mL ofacetone) followed by alcoholic 0.5M NaOH solution. Theneutral sugar composition was confirmed and quantified byderivatization to alditol acetates as previously described(Blakeney et al., 1983). The analysis of alditol acetate wascarried out (using inositol as a internal standard) by gas–liquid chromatography using a Hewlett-Packard 5710Achromatograph, on a stainless-steel column (0.3 × 200 cm)filled with Gaschrom Q (100–200 mesh). Column, injector,and detector temperatures were 235°, 250°, and 250°C, re-spectively. The carrier gas was nitrogen at a flow rate of 30mL min−1.

Molecular Mass Determination

The average molecular mass of EPS was determined bygel-permeation chromatography as previously described(Petersen et al., 1989). Samples (1 mL) containing 5 mg ofEPS were layered on a glass Bio-Rad Econo-column (2 × 87cm) packed with Sigma Sephacryl S-400-HR resin, andeluted with 10 mM morpholine-propane/sulfonic acid

284 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 67, NO. 3, FEBRUARY 5, 2000

(MOPS) buffer containing 0.1M NaCl at a flow rate of 9mL h−1. The collected fractions were analyzed for totalcarbohydrate content. Sigma Dextran standards used wereof 71, 260, 580, and 2000 kDa. Data shown are the averagesof two independent determinations, with standard devia-tions being <10%.

Rheological Properties

Aqueous dispersions of EPS produced byAnabaenasp.ATCC 33047 were prepared at 0.2%, 0.4%, and 0.6% (w/w). Aqueous dispersions of either xanthan gum (XG;Sigma) and a commercial synergistic mixture of gums, Al-kemir 110 (AG; Vedeqsa), both at a concentration of 0.4%(w/w), were also characterized for comparison purposes.XG was of “practical grade” quality and AG contained xan-than, guar, and locust-bean gums in a mass ratio that wasnot provided by the manufacturer. Sodium azide was addedto the dispersions to prevent growth of microorganisms.Dispersions were stored in sealed flasks kept at 4°C untilrheological measurements were performed. Before startingany experiment the samples were allowed to remain for 2 hat room temperature, and then carefully placed into the sen-sor system, where they were left for 30 min for equilibra-tion. The samples in the sensor system were placed in awater atmosphere to prevent them from drying out duringthe experiments and were kept at 20°C while carrying outthe rheological measurements. Data shown represent meanvalues of three independent measurements, with standarddeviations being <5%.

The flow behavior was determined in a Haake RotoviscoRV-20/CV-20N rheometer using an ME-31 Mooney–Ewartsensor system, with an inner radius (Ri) of 14.465 mm andan outer radius (Re) of 15 mm (measurement gap 0.535mm). As the radius ratio was <1.1, it was not necessary toapply any shear-rate correction (Prentice, 1983). The shearprogram applied consisted of an increasing shear-rate rampfrom 0.1 to 300 s−1 for 15 min. Then, the maximum shearrate was maintained for 10 min to check for any shear-timeeffect. Finally, the flow curves (see Results) were deter-mined by applying a decreasing shear-rate ramp from 300s−1 to 0.1 s−1.

Shear-stress growth experiments were conducted at 0.6s−1 in a Haake Rotovisco RV-100/CV-100 apparatus, usingthe ME-31 sensor system. The shear-stress response wasmonitored for approximately 103 s.

Oscillatory shear experiments within the linear viscoelas-tic domain were carried out in a Haake Rotovisco RV-20/CV-20N with the ME-31 sensor system. A frequency rangefrom 0.2 to 60 rad s−1 was covered.

RESULTS AND DISCUSSION

Chemical Properties

The purified EPS produced byAnabaenasp. ATCC 33047consisted predominantly of phenol-reacting material, with a

total carbohydrate content of 72.3% of the EPS dry weight.The polysaccharide also contained a polypeptidelike mate-rial, which accounted for about 7% of the dry weight. Paperchromatography of the neutral sugars fraction isolated fromthe hydrolyzed EPS revealed a composition of xylose, ga-lactose, glucose, and mannose. Gas–liquid chromatographyof alditol-acetate derivatives of neutral sugars showed thatgalactose and glucose were present in virtually equimolecu-lar amounts, with xylose being the major component andmannose found in small amounts. The molar ratio foundwas 11:5:5:1 for xylose, galactose, glucose, and mannose,respectively.

The acid fraction of hydrolyzed EPS analyzed by thecarbazole assay showed a uronic acids content of 19.4% ofthe EPS dry weight. Such a composition is not a particularlynoteworthy feature of theAnabaenaEPS, because manydifferent monosaccharidic compositions have been reportedso far for polysaccharides released by cyanobacteria, withregard to both the number of constitutive sugars (rangingfrom two to ten) and their typology (different arrangementsof pentoses, hexoses, and uronic acids) (Vincenzini et al.,1990). Most cyanobacterial exopolysaccharides are com-posed, like that fromAnabaena, of at least one uronic acidand several neutral sugars, sometimes in combination withpolypeptidelike material. Glucose, galactose, xylose, andmannose, the neutral sugar components ofAnabaenaEPS,are also common in cyanobacterial polysaccharides, in sev-eral proportions. The uronic acid content of most cyanobac-terial exopolysaccharides is about 20% to 30%, similar tothat fromAnabaena, but an exopolysaccharide fromMicro-cystis flos-aquaecontains levels of uronic acid as high as83% of the dry weight (Bender et al., 1994; Bertocchi et al.,1990; Plude et al., 1991).

Carbon and oxygen were the most abundant elements inthe elemental analysis of the EPS produced byAnabaenasp.ATCC 33047, representing 36.6% and 43.1% of the dryweight, respectively. The presence of nitrogen (1.5%) indi-cated the participation of nitrogenous compounds in theEPS composition, whereas sulfur was below the detectionlimit.

The IR spectrum of the EPS produced byAnabaenasp.ATCC 33047 exhibited a broad O—H stretching absorptionband centered around 3470 cm−1, a minor C—H stretchingband at 2922 cm−1, and several C—O absorption bands inthe 1400 to 900 cm−1 region, including a set of strong C—Ostretching bands at 1042 and 1072 cm−1 (Silverstein et al.,1991). The spectrum also displayed absorption bands ofcarboxylate groups from 1607 to 1663 and near 1400 cm−1

(data not shown). However, the spectrum did not displayabsorption bands due to sulfate groups (i.e., bands at 1240and 820 cm−1, attributable to S4O and C−O−S stretchingvibrations), which is in agreement with the absence of sulfurin the elemental analysis. Sulfate residues, however, seem tobe present in most of the cyanobacterial polysaccharidesstudied (De Philippis and Vincenzini, 1998), as well as inexopolysaccharides from some eukaryotic algae (Arad etal., 1985; Gudin and Thepenier, 1986).

MORENO ET AL.: PROPERTIES OF ANABAENA EXOPOLYSACCHARIDE 285

When the polysaccharide was subjected to gel-permeation chromatography, it eluted as a single peak. Theaverage molecular mass ofAnabaenaEPS was reproduciblydetermined to be 1.35 MDa (data not shown). This value issimilar to what has been reported for most cyanobacterialpolysaccharides studied (De Philippis and Vincenzini,1998). In general, the size of the molecular mass of micro-bial exopolysaccharides is positively correlated with theirapparent viscosity. The relatively high viscosity and theviscoelastic properties exhibited by aqueous dispersions ofAnabaenaEPS (see later) may be related to its high mo-lecular mass.

Rheological Properties

Flow Properties

Flow curves of aqueous solutions of the EPS produced byAnabaenasp. ATCC 33047 are shown in Figure 1. Withinthe available shear-rate range, the aqueous dispersions ofEPS exhibited non-Newtonian, shear-thinning behavior.These pseudoplastic properties are important in helping toprovide good sensory qualities, such as mouth feeling, fla-vor releasing, and suspending properties of food products(Enriquez et al., 1989); however, a food grade quality re-mains to be established for theAnabaenaEPS.

The shear-rate dependence of viscosity (h) was shown tofit power-law equation [Eq. (1)] fairly well:

h = h1~g. )n-1 (1)

whereh1 is the viscosity at 1 s−1 derived from the fittingequation,g

.is the shear rate, andn is the so-called flow

index. Asn tends to 1, the shear-thinning properties are lessand less pronounced, so that Newtonian behavior isachieved whenn 4 1.

The values of the power-law parameters obtained by lin-ear regression are shown in Table I. It is clear that, whileh1

increased with EPS concentration as a consequence of theexpected development of a stronger structure, the flow in-dex did not follow any definite trend, demonstrating that noconformational changes occurred from 0.2% to 0.6% (w/w)EPS. In any case, all dispersions characterized exhibited amarked shear-thinning response (n < 0.4). Furthermore, itshould be pointed out that the different flow index valuesfound were responsible for the fact that the influence of EPSconcentration on viscosity was dependent on the shear rate,as can be seen in Figure 1. It is well-known that, in con-centrated polysaccharide dispersions, the increase in viscos-ity with concentration is less pronounced as shear rate in-creases (Lapasin and Pricl, 1995). This is necessary for theirpractical applications because marked differences in viscos-ity must be expected at the different operative shear ratesassociated with their performance as stabilizers, or withoperations such as pouring, mixing, and pumping.

Then value found for the 0.2%AnabaenaEPS dispersionwas comparable to that obtained for the correspondingCya-nospira capsulataEPS dispersion (Navarini et al., 1990).However, at higher EPS concentrations, theAnabaenadis-persions turned out to be more thixotropic than thoseformed byCyanospira capsulataEPS.

Figure 2 shows a comparison of the shear-rate depen-dence of viscosity for 0.4% (w/w) aqueous dispersions ofEPS, AG, and XG. EPS and AG dispersions exhibited quitesimilar viscosity and/or shear thinning properties, as dem-onstrated by their fitting parameters. This appears promis-ing in regard to EPS applications, because AG is widelyused in the food industry as a stabilizer. Furthermore, com-parison of the shear-rate flow curves demonstrated that theEPS dispersion turned out to be less shear-sensitive than theXG dispersion. Indeed, the highest viscosities and the mostmarked shear-thinning properties were shown by the XGdispersion, even though differences in viscosities decreasedsubstantially as shear rate increased to >50 s−1. It is ofinterest to note that the slope of the straight line plotted inFigure 2 is −0.84 for the XG dispersion, typical of weakgels, whereas the corresponding values for EPS and AG areeven lower than −0.7, a typical value for entanglement net-works (Clark and Ross-Murphy, 1987; Ross-Murphy,1995a,b).

Transient Flow Properties

Stress-growth experiments at the inception of shear werecarried out to assess time-dependent behavior. These testsconsist of fixing a certain shear rate and monitoring theresponse of shear stress with time. A thixotropic materialinitially shows an increase in shear stress until a peak value,the so-called stress overshoot, is reached. Thereafter, the

Figure 1. Effect of shear rate on viscosity of dispersions of the exopoly-saccharide fromAnabaenasp. ATCC 33047 at different concentrations.Symbols: (n) 0.2; (h) 0.4; (s) 0.6% (w/w), respectively.

286 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 67, NO. 3, FEBRUARY 5, 2000

shear stress decays with time until reaching the steady-statevalue. The shear time defining the stress overshoot (tp) aswell as, logically, the associated strain (gp) decreased asEPS concentration increased (Table I). That is, both param-eters showed lower values, as the polysaccharide dispersionseemed to become more structured, as confirmed by thelowest value obtained for XG dispersion. This appears to bea general behavior in polysaccharide dispersions (Grassi etal., 1996; Navarini et al., 1990).

Figure 3 shows shear-stress (t) decay at the shear time at0.6 s−1 for EPS, AG, and XG dispersions.The results cor-responding to 0.4% and 0.6% (w/w) EPS dispersions are

compared with those of 0.4% (w/w) XG and AG disper-sions. It is noteworthy that EPS and mainly XG dispersionsexhibited a small region of constant shear-stress values ataround 30 s, which may be attributed to the occurrence of aresidual structure that later collapses under greater strain.

The shear-stress decay with shear time was fitted to a sumof two first-order kinetic equations. Fort $ tp:

t − te = A exp@−k1~t − tp!# + B exp@−k2 ~t − tp!# (2)

wheret is shear stress, gkte is the steady-state shear stresspredicted by the equation,A andB are adjustable preexpo-

Table I. Comparison of rheological parameters of the exopolysaccharide produced byAnabaenasp. ATCC 33047 with xanthan (XG) and Alkemir 110 (AG) gums.

Dispersions% (w/w)

Flow properties: power-law parameters

h1

(mPa s) n − 1 R2

0.2 EPS 372 −0.63 0.99970.4 EPS 630 −0.61 0.99930.6 EPS 961 −0.64 0.99880.4 AG 517 −0.59 0.99880.4 XG 2143 −0.84 0.9990

Dispersions% (w/w)

Transient flow properties

tp

(Pa)tp(s)

gp tea

(Pa)k1

(s−1)k2

(s−1)A

(Pa)B

(Pa)tp/te S+ e

0.4 EPS 0.69 7.0 4.2 0.34 0.10 3.8z 10−3 0.22 0.14 2.10 1.10 6.1z 10−6

0.6 EPS 1.95 5.5 3.3 1.10 0.23 3.2z 10−3 0.35 0.51 1.77 0.77 1.6z 10−4

0.4 AG 1.70 7.9 4.7 0.62 0.09 2.1z 10−3 0.95 0.21 2.65 1.65 9.2z 10−4

0.4 XG 3.24 4.0 2.4 2.30 0.23 2.3z 10−1 1.03 1.43 1.41 0.41 1.9z 10−3

aEquilibrium value estimated by the fitting equation. Data are mean values of three independentmeasurements, with SD being <5%.

Figure 2. Effect of shear rate on viscosity of dispersions of the exopoly-saccharide fromAnabaenasp. ATCC 33047, xanthan gum (XG), andAlkemir 110 (AG) at a concentration of 0.4% (w/w). Symbols: (h) EPS;(n) AG; (s) XG.

Figure 3. Time-dependent shear stress decay of dispersions of the exo-polysaccharide fromAnabaenasp. ATCC 33047, xanthan gum (XG), andAlkemir 110 (AG). Symbols: (h) EPS; (L) AG; (s) XG at a concentra-tion of 0.4% (w/w), respectively, and (n) EPS at 0.6% (w/w). Lines plottedshow the trends predicted by the fitting model.

MORENO ET AL.: PROPERTIES OF ANABAENA EXOPOLYSACCHARIDE 287

nential factors,tp is the shear time at the stress overshoot,k1

andk2 are first-order kinetic coefficients, andt is shear time.Nonlinear regression analysis was used to fit the experi-

mental results to Eq. (2). Lines plotted in Figure 3 show thetrends predicted by the model, allowing a visual comparisonwith experimental results. The results obtained fit the ki-netic model fairly well, except for the XG dispersion due tothe occurrence of a small undershoot at about 50 s. Theparameter that provides information about the fitting qualityis the sum of the square residuals,e. The best fit is thatassociated with the lowest value ofe:

e = S ~texp − teq!2

wheretexp is the experimental shear stress andteq is theshear stress derived from the fitting equation.

k1 andk2 should explain the kinetic decay of shear stress.Nevertheless, ask2 values are less than those ofk1 (Table I),fortunately only one parameter,k1, will be associated withthe breakdown kinetics of macromolecular junction zonesas well as with the orientation velocity in the flow direction.This process was clearly faster for the more structured dis-persions (see also oscillatory shear results), which contained0.4% XG and 0.6% EPS. No significant differences werefound for eitherk1 or k2 between 0.4% EPS and 0.4% AGdispersions. Two parameters can be often found in the lit-erature to describe, directly from the stress growth curve,the structural breakdown induced by shear when at least anapparent steady state is available: (1) the overshoot ratio,tp/te (Navarini et al., 1992; Richardson and Ross-Murphy,1987a); and (2) the amount of overshoot,S+ 4 [tp − te]/te

(Franco et al., 1995; Ganani and Powell, 1985). Both pa-rameters showed that the highest structural breakdown wasachieved with 0.4% AG dispersion, whereas the lowestbreakdown was observed for the 0.4% XG dispersion, withthe EPS dispersions presenting intermediate results.

It has been claimed that the stress overshoot is indicativeof the structure level of the dispersion in the quiescent state(Rochefort and Middleman, 1987). However, it is advisableto rely on linear viscoelastic functions rather than on non-linear parameters when trying to establish relationships withthe undisturbed dispersion state at rest, hence the usefulnessof small-amplitude oscillatory shear experiments.

Linear Dynamic Viscoelastic Properties

Small-amplitude oscillatory shear experiments were con-ducted to gain a better understanding of the rheology ofthese dispersions in shear conditions close to the undis-turbed state. These experiments basically consist of apply-ing the strain as a sinusoidal time function. They makepossible the determination of viscoelastic functions, such asthe storage modulus (G8) and the loss modulus (G9), whichare functions of frequency. The former is related to theelastic response, and the latter to the viscous response (Ma-cosko, 1994). Oscillatory shear results may be considered as

a sort of fingerprint of the material structure as long as theyare obtained within the linear viscoelastic domain.

Figure 4 shows that the strain values chosen to performfrequency sweeps (7.5%) were within the linear viscoelasticdomain, as demonstrated by the fact that the complex vis-cosity remained essentially constant when conductingstrain-sweep experiments at 1 Hz. No critical strains for theonset of nonlinear viscoelastic response could be observedwithin the strain range studied.

Figure 5 shows the frequency dependence of the storageand loss moduli for theAnabaenaEPS dispersions. This isthe so-called viscoelastic or mechanical spectrum. The elas-tic component was always higher than the viscous onewithin the available frequency range, indicating thatAna-baenaEPS dispersions can find suitable applications as sta-bilizers and that a sort of faint weak gel structure wasformed at EPS concentrations as low as 0.2% (w/w). More-over, bothG8 andG9 assumed a power-law dependence withfrequency (G8 ~ va, G9 ~ vb). As far as the influence ofEPS concentration on the viscoelastic response is con-cerned, it is worth noting that: (a)G8 was consistentlyhigher thanG9; and (b) the frequency dependence of theviscoelastic functions was weaker as EPS concentration in-creased (a ranged from 0.55 to 0.36 andb from 0.56 to0.44). Both of these findings demonstrate that the EPS dis-persions increasingly more elastic as concentration rises.

Figure 6 compares the viscoelastic spectra of 0.4% (w/w)EPS, AG, and XG dispersions. AG and particularly XGdispersions turned out to be more structured than EPS dis-persions in the linear viscoelastic domain, as demonstratedby their lowera andb values.a was around 0.25 for bothgums, whereasb was 0.31 for AG and 0.18 for XG, indi-

Figure 4. Complex viscosity at a frequency of 6.28 rad s−1 as a functionof strain, within the linear viscoelastic domain, of dispersions of the exo-polysaccharide fromAnabaenasp. ATCC 33047, xanthan gum (XG), andAlkemir 110 (AG). Symbols: (n) EPS at 0.2% (w/w); (h) EPS; (m) AG;(d) XG at a concentration of 0.4% (w/w), respectively, and (s) EPS at0.6% (w/w).

288 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 67, NO. 3, FEBRUARY 5, 2000

cating that the latter could be characterized by a slowerrelaxation mechanism. Further evidence can be found uponcomparing their respectiveG9/G8 ratio values, which wereclearly higher for the EPS dispersions than for the AG andXG dispersions, showing that the EPS dispersions are more“fluidlike.”

The dynamic viscoelastic results of XG dispersions havebeen attributed to the occurrence of attractive interactionsbetween helices formed by xanthan-rigid macromolecules,which may give rise to a weak network by means of either

hydrogen bonding or possibly hydrophobic effects (Carnali,1992). The dynamic viscoelastic results of the AG disper-sion studied seem to be dominated by XG content, because,despite also containing glucomannans (such as locust beangum or guar gum) in a ratio not provided by the manufac-turer, its viscoelastic spectrum is closer to that of XG dis-persions rather than those typical of the aforementionedglucomannans (Ross-Murphy, 1995a,b).

The rheological properties ofAnabaenaEPS dispersionsreveal they have an intermediate behavior between that typi-cal of random coil polysaccharides and that of weak gels. Asimilar interpretation was found forCyanospira capsulataEPS dispersions (Navarini et al., 1992). However, accordingto the small-amplitude oscillatory shear results obtained, thestructure ofAnabaenaEPS dispersions must be closer tothat of weak gels, like xanthan gum, than that ofCyanospiracapsulataEPS dispersions. This is supported by the fact thatthe response ofAnabaenadispersions correspond to theplateau relaxation zone in the available frequency windowwith no evidence of the pseudoterminal zone appearance.Indeed, the lack of a “crossover” between the curves ofG8and G9 at low frequency is quite clear. Nevertheless, theflow curve (see flow index values) and transient flow (strainvalues associated with the stress overshoot) results supportthe interpretation that the behavior ofAnabaenaEPS dis-persions is not typical of weak gels, but it is closer to thatfound for random-coil polysaccharides (Ross-Murphy,1995b). Thus, the rheological results obtained are consistentwith the occurrence of a structure based on topologicalphysical interactions among macromolecules forming weakjunction zones, which may coexist with specific intermo-lecular interactions, cooperatively giving rise to a transientnetwork.

CONCLUSIONS

AnabaenaEPS is a heteropolysaccharide containing xylose,glucose, galactose, and mannose, as well as carboxylategroups. These carboxylate groups confer to the EPS a poly-electrolyte nature and could serve as binding sites for diva-lent metal ions. However, further work is needed to deter-mine whether the EPS has high and/or selective affinitiesfor metals ions, thus being useful as a metal-chelating agent.On the other hand, the carboxylate groups of the EPS ofAnabaenamight be also used for linking to starch or syn-thetic polymers to generate new polysaccharides with spe-cial properties. The high molecular weight found for EPS isconsistent with the viscoelastic, shear-thinning, and time-dependent flow properties of their aqueous dispersions. Thepseudoplasticity of this EPS is an advantageous flow prop-erty that may be useful, particularly when mixed with othermaterials in commercial applications. Flow curves revealedthat viscosity values ofAnabaenaEPS dispersions are simi-lar to those of a food-grade commercial mixture of poly-saccharides and not far from the values exhibited by xan-than gum. While at high shear rates, the low viscosities ofAnabaenaEPS dispersions enhances unit operations like

Figure 5. Viscoelastic spectra of dispersions of the exopolysaccharidefrom Anabaenasp. ATCC 33047 at different concentrations. Symbols:G8

(storage modulus) (n, 0.2;h, 0.4;s, 0.6% [w/w], respectively);G9 (lossmodulus) (m, 0.2; j, 0.4; d, 0.6% [w/w], respectively).

Figure 6. Viscoelastic spectra of dispersions of the exopolysaccharidefrom Anabaenasp. ATCC 33047, xanthan gum (XG), and Alkemir 110(AG) at a concentration of 0.4% (w/w). Symbols:G8 (storage modulus) (h,EPS;n, AG; s, XG); G9 (loss modulus) (j, EPS;m, AG; d, XG).

MORENO ET AL.: PROPERTIES OF ANABAENA EXOPOLYSACCHARIDE 289

pumping or mixing, and the large increase in viscosity asshear rate decreases favors their application as a dispersionstabilizer and thickening agent. These promising propertiesare confirmed by their viscoelastic spectra. A comparisonwith the viscoelastic spectra of Alkemir 110 and xanthangum dispersions shows thatAnabaenaEPS dispersions pre-sent a more “fluidlike” relaxation mechanism. The stressdecay with shear time can be modeled by the addition oftwo first-order kinetic equations. The rheology of the EPSdispersions seems to be consistent with the formation of anintermediate structure between that of a random-coil poly-saccharide and a weak gel.

On the other hand, the rheological stability of this exo-polymer to temperature, pH, and salt concentration are alsobeing studied, and preliminary results suggest thatAna-baenaEPS may offer certain advantages over other com-mercially available gums.

The authors thank Dr. A. Gil (Departamento Quı´mica Organica,Facultad de Quı´mica, Universidad de Sevilla) for providing fa-cilities for obtaining the infrared spectrum. The valuable techni-cal assistance of Miss M. Jose´ Figueroa Gonza´lez is also verymuch appreciated.

References

Arad SM, Adda M, Cohen E. 1985. The potential of production of sulfatedpolysaccharides fromPorphyridium. Plant Soil 89:117–127.

Arnon DI, McSwain BD, Tsujimoto HY, Wada K. 1974. Photochemicalactivity and components of membrane preparations from blue-greenalgae. I. Coexistence of two photosystems in relation to chlorophyllaand removal of phycocyanin. Biochim Biophys Acta 357:231–245.

Ash SG. 1985. Polymers from microbes: the polysaccharide gums. ShellPolymers 9:18–21.

Bender J, Rodrı´guez-Eaton S, Ekanemesang UM, Phillips P. 1994. Char-acterization of metal-binding bioflocculants produced by the cyano-bacterial component of mixed microbial mats. Appl Environ Microbiol60:2311–2315.

Bertocchi C, Navarini L, Cesa`ro A. 1990. Polysaccharides from cyanobac-teria. Carbohydr Polym 12:127–153.

Bitter T, Muir HM. 1962. A modified uronic acid carbazole reaction. AnalBiochem 4:330–334.

Blakeney AB, Harris PJ, Henry RJ, Stone BA. 1983. A simple and rapidpreparation of alditol acetates for monosaccharide analysis. CarbohydrRes 113:291–293.

Carnali JO. 1992. Gelatian in physically associating biopolymer systems.Rheol Acta 31:399–412.

Clark AH, Ross-Murphy SB. 1987. Structural and mechanical properties ofbiopolymer gels. Adv Polym Sci 83:57–192.

Cuvelier G, Launay B. 1986. Concentration regimes in xanthan gum so-lutions deduced from flow and viscoelastic properties. CarbohydrPolym 6:321–333.

De Philippis R, Vincenzini M. 1998. Exocellular polysaccharides fromcyanobacteria and their possible applications. FEMS Microbiol Rev22:151–175.

Dubois M, Gilles KA, Hamilton JK, Rebers PA, Smith F. 1956. Colori-metric method for determination of sugars and related substances.Anal Chem 28:350–356.

Enriquez LG, Hwang JW, Hong GP, Bati NA, Flick GJ. 1989. Food emul-sifiers: chemistry, technology, functional properties and applications.In: Charalambous G, Doxastakis G, editors. Amsterdam: Elsevier. p335–416.

Franco JM, Mun˜oz J, Gallegos C. 1995. Transient and steady flow of a

lamellar liquid crystalline surfactant/water system. Langmuir 11:669–673.

Ganani E, Powell RL. 1985. Transient flow viscometry. J Rheol 29:931–941.

Grassi M, Lapasin R, Pricl S. 1996. A study of the rheological behaviourof scleroglucan weak gel systems. Carbohydr Polym 29:169–181.

Gudin C, Thepenier C. 1986. Bioconversion of solar energy into organicchemicals by microalgae. Adv Biotechnol Proc 6:73–110.

Herbert D, Phipps PJ, Strange RE. 1971. Chemical analysis of microbialcells. In: Norris JR, Ribbons DW, editors. Methods in microbiology,vol 5B. London: Academic Press. p 209–344.

Lapasin R, Pricl S. 1995. Rheology of industrial polysaccharides: theoryand applications. London: Chapman & Hall. p 324–338.

Linton JD. 1990. The relationship between metabolite production and thegrowth efficiency of the producing organisms. FEMS Microbiol Rev75:1–18.

Macosko CW. 1994. Rheology. Principles, measurements and applications.New York: VCH. p 121–126.

Moreno J, Rodrı´guez H, Vargas MA, Rivas J, Guerrero MG. 1995. Nitro-gen-fixing cyanobacteria as source of phycobiliprotein pigments.Composition and growth performance of ten filamentous heterocyst-ous strains. J Appl Phycol 7:17–23.

Moreno J, Vargas MA, Olivares H, Rivas J, Guerrero MG. 1998. Exo-polysaccharide production by the cyanobacteriumAnabaenasp.ATCC 33047 in batch and continuous culture. J Biotechnol 60:175–182.

Navarini L, Bertocchi C, Cesa`ro A, Lapasin R, Crescenzi V. 1990. Rhe-ology of aqueous solutions of an extracellular polysaccharide fromCyanospira capsulata. Carbohydr Polym 12:169–187.

Navarini L, Cesa`ro A, Ross-Murphy SB. 1992. Viscoelastic properties ofaqueous solutions of an exocellular polysaccharide from cyanobacte-ria. Carbohydr Polym 18:265–272.

Petersen GR, Nelson GA, Cathey CA, Fuller GG. 1989. Rheologicallyinteresting polysaccharides from yeasts. Appl Biochem Biotechnol 20:845–867.

Plude JL, Parker DL, Schommer OJ, Timmerman RJ, Hagstrom SA, JoersJM, Hnasko R. 1991. Chemical characterization of polysaccharidefrom the slime layer of the cyanobacteriumMicrocystis flos-aquaeC3-40. Appl Environ Microbiol 57:1696–1700.

Prentice JH. 1983. Precision and accuracy on viscometric measuremenst onfoods. J Rheol 27:621–635.

Richardson RK, Ross-Murphy SB. 1987a. Non-linear viscoelasticity ofpolysaccharide solutions. 1: Guar galactomanannan solutions. Int JBiol Macromol 9:250–256.

Richardson RK, Ross-Murphy SB. 1987b. Non-linear viscoelasticity ofpolysaccharide solutions. 2: xanthan polysaccharide solutions. Int JBiol Macromol 9:257–263.

Rochefort WE, Middleman S. 1987. Rheology of xanthan gum: salt, tem-perature and strain effects in oscillatory and steady shear experiments.J Rheol 31:337–369.

Ross-Murphy SB. 1995a. Rheology of biopolymer solutions and gels. In:Dickinson E, editor. New physico-chemical techniques for the char-acterization of complex food systems. Glasgow, UK: Blackie Aca-demic. p 139–156.

Ross-Murphy SB. 1995b. Structure–property relationships in food biopol-ymers gels and solutions. J Rheol 39:1451–1463.

Ross-Murphy SB, Morris VJ, Morris ER. 1983. Molecular viscoelasticityof xanthan polysaccharide. Faraday Symp Chem Soc 18:115–129.

Silverstein RM, Bassler GC, Morril TC. 1991. Spectrometric identificationof organic compounds. New York: John Wiley & Sons.

Sutherland IW. 1990. Biotechnology of microbial exopolysaccharides.Cambridge studies in biotechnology, vol 9. Cambridge, UK: Cam-bridge University Press.

Vincenzini M, De Philippis R, Sili C, Materassi R. 1990. A novel exo-polysaccharide from a filamentous cyanobacterium: production,chemical characterization and rheological properties. In: Dawes EA,editor. Novel biodegradable microbial polymers. Dordrecht: Kluwer. p295–310.

290 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 67, NO. 3, FEBRUARY 5, 2000