Chemical , morphological , rheological and thermal properties of Solanum lycocarpum phosphorylated starches

Em vista da necessidade de amidos com características específicas, é fundamental o estudo de amidos não convencionais e de suas modificações, de acordo com as exigências do mercado consumidor . O objetivo deste trabalho foi estudar as características físico-químicas de amidos nativo e fosfatado da S. lycocarpum. O amido foi fosfatado com tripolifosfato de sódio (de 5 a 11%), sob agitação. Determinaram-se a composição química, a morfologia, a densidade, a capacidade de ligação à água fria, o poder de inchamento e o índice de solubilidade, a turbidez e a sinerese, as propriedades reológicas e calorimétricas. Não se detectou fósforo na amostra nativa, porém, a fosfatação gerou amidos modificados com teores de fósforo de 0,015, 0,092 e 0,397%, que se caracterizaram por absorver maior quantidade de água, a frio e a quente. Os resultados reológicos mostraram a forte influência do teor de fósforo na viscosidade dos fosfatados, apresentando temperatura de empastamento menor e viscosidade de pico maior que as do amido nativo. A


INTRODUCTION
Starch is an important ingredient for the industrial sector and is a major component of the human diet.It has diverse applications and its marketing is on the rise in many fields of industry (Limberger et al., 2008).Even so, this natural polymer requires modifications.According to Bemiller (1997), modified starches have better functional properties than natural starches, as they tend to retrograde less, their pastes are more stable during cooling and thawing and more transparent, their gels have good adhesion, improved texture and tend to form films.
Research on new unconventional starch sources and modifications is needed to find starches with rheological properties providing better quality to the final product to meet the demands of the consumer market.
Fruits of the Amazonian regions and the Brazilian Cerrado that have potential as starch sources are still little explored.For example, despite intense pharmacological use of Solanum lycocarpum Saint Hillaire, the wolf's fruit, there is little information on its starch characteristics and chemical modification, therefore, studies are necessary on its technological usefulness.
The fruit of this plant species may contain more than 20 g of starch per 100 g pulp (Junior et al., 2004).These starch granules are rounded shaped and relatively small (16.6 µm), containing 29.3% amylose and their pastes have regular stability, either cold or hot (Mota et al., 2009).These characteristics are comparable with those of rice and cassava starches.
However, the major disadvantage of unconventional starches is their restricted use in industry, because of some undesirable properties.These starches may have organic (lipids, proteins, pigments, etc.) and minerals compounds derived from the extraction process, which when interacting with amylose and amylopectin, can influence their properties and performance (Chan et al., 2010).
Water absorption and solubility, for example, depend on the starch crystalline structure, i.e., inter or intramolecular interaction of hydrogen bonds, which may break in hot water, reducing micelle strength by breaking hydrogen bonds and greatly increasing water absorption, resulting in swelling and solubilization of the starch granule (Swinkels, 1996).
However, starch properties can be improved by chemical modifications.Phosphorylation with sodium tripolyphosphate acid (STPA) is one of the most commonly used chemical modification of starches, as it is a salt of a relatively low cost and a simple process (Batista et al., 2010).Phosphate groups introduced into the starch chains causes repulsion between phosphate groups on adjacent chains and increases hydration.Phosphate groups are covalently bonded to molecules of amylopectin (Noda et al., 2007).The phosphorus content is an important factor in the variation starch functional properties, including gelatinization and retrogradation (Karim et al., 2007).

MATERIALS AND METHODS
Native starch was extracted from unripe S. lycocarpum fruits.Fruits were washed, peeled, chopped and ground in a circular rotor blade mill (MA-580, Marconi, São Paulo, Brazil), with water added with 5 g L -1 sodium metabisulfite (Synth) to prevent browning.The resultant slurry was separated by sieving through 75-250 micron stainless steel screen (Bertel, São Paulo, Brazil), followed by decantation and washing with absolute ethanol (Synth) to remove fats and excess water, drying in oven with air renewal and circulation (Marconi MA-035, São Paulo, Brazil), at 45 °C for 6 h.
The phosphorus (P) content was determined according to Brazil (2005), by spectrophotometry using a UV/visible (SP-2000UV) spectrophotometer at 420 nm.The contents of ash (method 923.03), moisture (method 925.10), ether extract (method 920.39) and crude protein (method 960.52, 5.83 conversion factor) were analyzed according to the methods described by AOAC (2005).Crude fiber was determined according to Brazil (2005).Starch content was determined using the technique described by Cereda et al. (2004).Amylose content was determined by blue value iodometric analysis (McCready & Hassid, 1943).
The shape and size of the starch granules were analyzed by a LEICA EC3 optical microscope (Wetzlar, Germany) and the images analyzed by the software Leica Application Suite v. LAS EZ. 2.0.0 (Leica Microsystems, 2010), which enables clear visualization of shape and estimation of size of the starch granules.
Absolute density (ρ) was determined as described by Leach & Schoch (1964) with some modifications.A 10 ml pycnometer with known mass was used to measure the pycnometer mass with xylene (b), xylene density (d) and the pycnometer mass with xylene and starch (c), using 5 g starch (a) (dry basis).The absolute density was calculated by Equation 1.
The binding capacity in cold water (BCCW) was measured as described by Gilles & Medcalf (1965).A 2.5 g sample was weighed into a centrifuge tube with 40 ml of distilled water and stirred in a Dubnoff pendulum shaker (ET-053, Tecnal Piracicaba, Brazil) for 1 h.Samples were centrifuged at 2200 rpm for 10 min.The supernatant was removed and the tube containing the pellet was weighed.The water bound to the starch was determined by equation 2: Eq 2.
The swelling power (SP) and water solubility index (WSI) were determined according to Schoch & Leach (1964) at temperatures ranging from 60 to 90 °C.
Turbidity and syneresis analyses of starch gel were performed according to Oliveira & Cereda (2003).A suspension of 8% starch in water was heated to obtain a translucent gel, which was distributed into five containers of 100 ml (approximately 20 ml of gel per container), cooled and stored at 4 °C.Turbidity was determined by absorbance at 640 nm in a spectrophotometer.Syneresis was determined as the percentage of water released by the gel in relation to the total mass after centrifugation at 3000 rpm for 15 min, during 5 days.
Viscosity was determined using a Rapid Visco Analyzer 4 (RVA Newport Scientific PTY LTD, Sydney, Australia), according to the manufacturer's recommendations.Suspensions of 2.5 g starch in 25 ml of distilled water were adjusted to 14% moisture and analyzed according to the following time/temperature regime: 50 ºC min -1 , heating from 50 to 95 °C at a 6 °C-1 min rate, held at 95 ° C for 5 min and cooled from 95 to 50 °C at a 6 °C min -1 rate.Viscosity was expressed as centiPoise (cP).From the profiles generated by the RVA, we evaluated the following parameters: maximum peak viscosity, minimum viscosity after peak, breakdown viscosity (difference between the maximum viscosity and minimum viscosity of paste maintained at 95 °C for 5 min), final viscosity and retrogradation (difference between final viscosity and minimum viscosity at 95 °C for 5 min) and pasting temperature (°C).
The calorimetric analysis was performed in a differential scanning calorimeter DSC Q200 (TA Instruments, New Castle, USA), according to Fakirov et. al. (1997).The calorimeter was calibrated using indium standard.For the starch gelatinization temperature, approximately 5 mg of sample of known moisture were placed in a hermetically sealed aluminum crucible.The scanning profile with balance consisted of reading temperatures between 5 and 110 ºC, with heating rate of 10 °C min -1 and nitrogen flow of 50 ml m -1 .The gelatinization enthalpy was calculated using the Universal Analysis software version 4.3A.
The experiment was arranged in a completely randomized design (CRD), with only one factor.Treatments were the phosphorus levels determined, in addition to the native starch, in the different starches prepared with 5, 7 and 11% of STPA.
The statistical analysis of the experimental data was according to Gomes (2009).Measurements of chemical composition (means of five replicates) were expressed as mean ± standard deviation and correlation coefficient (r).The effect of different phosphate levels on the chemical composition was qualitatively determined using one-way analysis of variance (ANOVA) at 5% probability level and significant responses were compared using the Tukey's test at 5% probability level.
Experimental data from other physicochemical properties were designed based on the phosphorus content and, in some cases, adjusted to a non-linear regression exponential model according to equation 3: where y is generic response function, x is the real variable, a, b and n represent the coefficients estimated by the method of least squares, with significance assessed by the t test at 5% probability level.The fit of the nonlinear regression model was evaluated by comparing the standard error of the estimate (SEE) using the mathematical Quasi-Newton method.
A cluster analysis by type of phosphate starch was performed with the means of the variables involved in the physicochemical properties using the procedure Cluster Analysis: Joining (tree clustering) of the Statistica 8.0 software (Statsoft, 2007), which is based on the Euclidean distance.The complete linkage method was selected for grouping.Additionally, the cluster analysis was complemented by the Principal Component Analysis (PCA).
The relationship between ash, starch, amylose, BCCW, breakdown viscosity, retrogradation and variation of gelatinization enthalpy with phosphorus content was obtained with the Pearson correlation coefficient (r), at 5% probability level.The correlation coefficient was interpreted as follows (Callegari-Jacques 2003): -If 0.00 <r <0.30, there is a weak linear correlation; -If 0.30 <r <0.60, there is a moderate linear correlation; -If 0.60 <r <1.00, there is a strong linear correlation.
The statistical analyses and the graphs were performed using the Statistica 8.0 software (Statsoft, 2007).

RESULTS AND DISCUSSION
In this study, the S. lycocarpum native starch showed a high purity (99.31% starch) and low levels of other fractions in its chemical composition (Table 1).Phosphorus was not detected and the content of amylose and amylopectin were 28.79 and 70.52%, respectively, agreeing with the findings of Mota et al. (2009).
Table 1 shows that contents of ether extract, fiber and protein remained constant, indicating that phosphorylation was successful.However, the ANOVA detected differences in moisture content (10.97 to 8.26%), starch (from 99.31 to 95.96%) and amylose (28.79 to 14.05%), which decreased with phosphorus content.
There were significant differences for ash content among the modified starches, increasing approximately 10, 13 and 16 times, when the native starch was phosphorylated up to 0.015, 0.092 and 0.397% phosphorus, respectively (Table 1), showing a strong positive correlation (r = 0.921).These results confirm the observations made by Limberger et al. (2008) with rice starch, who argued that the relationship between ash and phosphorus levels is caused by the introduction of phosphate groups into the starch chains.
Proportionally, the ash content due to the phosphorus introduction was responsible for the decrease in starch and amylose, showing strong inverse correlation (-0.999 <r < -0.975) and, consequently, the increase the proportion of amylopectin (r = 0.969) .
Because the introduction of phosphorus into S. lycocarpum starch changed the chemical composition, it also accounts for the changes in the physico-chemical characteristics, forming products with different properties.
The native starch granules showed a smooth surface with different sizes and shapes (Figure 1).Starch granules were rounded, elliptical, truncate and irregular in shape.In addition to these shapes, swollen granules were observed in phosphorylated starch (Figures 1B,1C and 1D), demonstrating that the native starch undergoes apparent damage after phosphorylation, which can influence paste properties.
The average diameter and density of native starch was 27.3 µm and 1.87 g mL -1 (Figure 2A), respectively.However, after phosphorylation, the diameter increased and the density decreased (up to approximately 42 µm and 1.35 g mL -1 , respectively) because of the swelling of the granules.The binding capacity in cold water of phosphorylated starches was also greater than the native starch (Figure 2B), from an initial BCCW value of 125.3% to 138.7, 153.4 and 175.1% for starches containing 0.015, 0.092 and 0.397%, of phosphorus, respectively.The correlation between BCCW and P was positively strong (r = 0.928).
However, these starches gelatinize when heated in a large amount of water and increase in size and partially solubilize, which can be seen by the swelling power values (Figure 3A) and the water solubility index (Figure 3B) respectively, increasing with temperature.Hoover (2001) explained that this fact occurs because the starch structure breaks, leading to weakening of hydrogen bonds and interaction of water molecules with the hydroxyl groups Rev. Ceres, Viçosa, v. 61, n.4, p. 458-466, jul/ago, 2014

Component (%)
Figure 2. Effect of phosphorus content on diameter, density and binding capacity to cold water (BCCW) of starch extracted from unripe fruits of S. lycocarpum: A) smaller diameter ( ), larger diameter ( ) and density ( ).B) BCCW.The introduction of phosphorus caused the swelling power and solubility index of native starch to increase (Figure 3).Daniel et al. (2006) reported that this phenomenon is due to the ability of phosphate groups to absorb larger quantities of water, that is, they have negative charges that repel each other, thus facilitating penetration and absorption of water (Wang et al., 2003).
Phosphorylated starches showed lower turbidity and syneresis than the native starch because of their higher binding capacity in cold water (Figure 4).According to Limberger et al. (2008), these starches form pastes that can be clearer and prevent retrogradation and hence syneresis, because they restrain greater contact between the amylose molecules.These molecules are solubilized during heating and leave the granule, preventing the formation of micro crystals responsible for retrogradation.
The viscosity profiles (Figure 5) confirm the strong influence of phosphorus introduction on the rheological properties of phosphorylated starches.The pasting temperature of these starches was lower and the viscosity peak was higher (around 95 °C and between 6260 and 6440 cP respectively), while for the native starch, the corresponding values were 70.7 °C and 4407 cP.This viscosity profile was also reported for phosphorylated starches of wheat and corn (Batista et al., 2010) and starches of corn, potato and beans (Chan et al., 2010).Batista et al. (2010) discussed that the reduction in paste temperature was due to starch gelatinization, while the maximum viscosity indicates the presence of intermolecular forces which strengthen the amorphous region of starch granules (Karim et al., 2007).Although it has greater stability to heat and breakdown (breakdown viscosity of 1104 cP), the native starch showed higher retrogradation (535 cP), with final viscosity of 3845 cP, whereas the phosphorylated starches showed higher breakdown viscosity (1705-2556 cP) and lower retrogradation (211-302 cP) with final viscosity between 4170 and 4816 cP.
For Chan et al. (2010), the evident breakdown viscosity in modified starches compared with native starches is probably due to the weak structure of the granules during the chemical modification, which facilitates the breakdown of the granular structure, while the tendency to retrogradation is influenced by the amylose content (Singh et al., 2005), since the amylose inhibits the rearrangement of the granular structure during the cooling of the gelatinized starch paste (Singh et al., 2003).
This may have occurred in this study, however, the breakdown viscosity decreases with the phosphorus content and the relationship between the breakdown viscosity and phosphorus content was strongly negative, with r = -0.826.In the case of retrogradation, it showed a strong negative correlation (r = -0.935)and positive correlation (r = 0.999) with the amylose content.
Starch thermal properties are described in Table 2.In this study, the variation in enthalpy was negatively correlated with phosphorus (r = -0.933),requiring energy of 9.7, 8.5, 8.1 and 6.4 J g -1 for the transition of starch granules with contents of 0, 0.015, 0.092 and 0.397% from the crystalline to the amorphous state, respectively.Bonds with phosphorous may have caused increase in the interaction forces.According to Acquarone & Rao (2003), inter and intra molecular bonds, in random positions in the starch granule, stabilize and strengthen the granule, and this occurs because during phosphorylation a double modification takes place, i.e., a combination of replacement with cross-linking, which hinders retrogradation (Wurzburg, 1986).
The hierarchical clustering dendrogram shows the formation of groups of genotypes with some degree of similarity and the dissimilarity among groups (Figure 6A).For a distance equal to 3, we observe the formation of three groups: group I formed by the native starch with 0% phosphorus; Group II with starches containing 0.015 and 0.092% of phosphorus; and group III with the phosphorylated starch with 0.397% phosphorus.
In the principal components analysis all studied variables contributed to component 1, with correlation greater than 0.70.However, the variables that contributed most to component 2 were phosphorus content and pasting temperature.
The cluster analysis and principal component analysis were in agreement on determining similarity between starches, forming three distinct groups (Figure 6).

Figure 3 .
Figure 3. Variation in swelling power (A) and water solubility index (B) of S. lycocarpum starch as a function of phosphorus content and temperature.

Figure 4 .
Figure 4. Variation in turbidity and syneresis of starch extracted from unripe fruits of S. lycocarpum as a function of phosphorus content.

Table 2 .P
Variation in thermal properties of starch extracted from unripe fruits of S. lycocarpum as a function of phosphorus content initial temperature, T p : peak temperature, T f : final temperature, r r r r rH: variation in enthalpy and r r r r rt: variation in temperature.

Figure 6 .
Figure 6.Cluster analysis and principal component analysis of S. lycocarpum starches as a function of phosphorus content.A) Similarity dendrogram and B) score dispersion graph of principal components (PC).

Table 1 .
Variation in chemical composition (dry basis) of S. lycocarpum starch as a function of phosphorus content* * Means of 5 repetitions ± standard deviation, ** Difference between starch and amylose contents.