Food Chemistry 156 (2014) 319–325

Contents lists available at ScienceDirect

Food Chemistry

journal homepage: www.elsevier .com/locate / foodchem

Changes in resistant starch from two banana cultivars duringpostharvest storage

http://dx.doi.org/10.1016/j.foodchem.2014.02.0120308-8146/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author. Tel.: +86 2087112851.E-mail address: fehhuang@scut.edu.cn (H.H. Huang).

Juan Wang, Xue Juan Tang, Ping Sheng Chen, Hui Hua Huang ⇑College of Light Industry and Food Science, South China University of Technology, Wushan Road No. 381, Guangzhou 510641, China

a r t i c l e i n f o

Article history:Received 5 September 2013Received in revised form 15 January 2014Accepted 3 February 2014Available online 12 February 2014

Keywords:Banana resistant starchPhysicochemical propertiesRipening stageStarch structure

a b s t r a c t

Banana resistant starch samples were extracted and isolated from two banana cultivars (Musa AAA group,Cavendish subgroup and Musa ABB group, Pisang Awak subgroup) at seven ripening stages during post-harvest storage. The structures of the resistant starch samples were analysed by light microscopy, polar-ising microscopy, scanning electron microscopy, X-ray diffraction, and infrared spectroscopy.Physicochemical properties (e.g., water-holding capacity, solubility, swelling power, transparency,starch–iodine absorption spectrum, and Brabender microviscoamylograph profile) were determined.The results revealed significant differences in microstructure and physicochemical characteristics amongthe banana resistant starch samples during different ripening stages. The results of this study providevaluable information for the potential applications of banana resistant starches.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Resistant starch (RS) is defined by EURESTA (European FLAIRConcerted Action on Resistant Starch) as ‘‘the sum of starch andproducts of starch degradation not absorbed in the small intestineof healthy individuals’’ (Asp, 1992). There are four types of RS: (1)physically entrapped, inaccessible starch within whole or partiallymilled seeds (RS1); (2) native granular starch, consisting of nongel-atinised granules (RS2); (3) retrograded starch produced by foodprocessing applications (RS3); and (4) chemically modified starch(RS4; Englyst, Kingman, & Cummings, 1992). RS has physiologicaleffects similar to those of prebiotics and dietary fibre: RS stimu-lates the growth of beneficial bacteria in the gut (e.g., bifidobacte-ria) and increases the production of short-chain fatty acidsassociated with gut immune function and microbiota modulation(Fuentes-Zaragoza et al., 2011; Johnson & Gee, 1996). Additionally,RS protects against several diseases, including type II diabetes,colorectal cancer, and other diet-related chronic diseases (Niba,2002; Topping & Clifton, 2001).

Several studies have focused on the functions of RS. Mutungi,Rost, Onyango, Jaros, and Rohm (2009) studied the crystallinityand the thermal and morphological characteristics of RS3 fromdebranched cassava starch. Garcia-Rosas et al. (2009) assessedthe changes in maize tortilla RS content and structure duringstorage. Aparicio-Saguilan et al. (2007) successfully prepared

slow-digestible cookies from RS-rich lintnerised banana starch. Inaddition, Aparicio-Saguilan, Gutierrez-Meraz, Garcia-Suarez, Tovar,and Bello-Perez (2008) investigated the physicochemical and func-tional properties of cross-linked banana RS. Ble-Castillo et al.(2008) reported that banana RS flour supplementation reducesbody weight and insulin resistance in obese individuals with typeII diabetes.

Unripe bananas are rich in RS2 (Johnson & Gee, 1996; Niba,2002). As tropical and subtropical fruits, bananas are mainlyplanted in tropical and subtropical zones. RS isolated from differ-ent banana cultivars may have different properties. Moreover, withpost-harvest storage, numerous enzymes transform the starches inthese fruits into different sugars. Consequently, the RS content ofbananas at different ripening stages may differ. However, few stud-ies have focused on the changes of banana RS throughout storage.This study assessed the changes in the content, physicochemicaland structural properties of RS from two banana cultivars (MusaAAA group, Cavendish subgroup and Musa ABB group, Pisang Awaksubgroup) at different stages of maturity.

2. Materials and methods

2.1. Materials

Two banana cultivars were used in this study: Musa AAA group,Cavendish subgroup and Musa ABB group, Pisang Awak subgroup.These cultivars were planted and sold in Guangdong province,China. Using the criteria reported by SH Pratt Co. (Luton, United

320 J. Wang et al. / Food Chemistry 156 (2014) 319–325

Kingdom; Soltani et al., 2011), the fruits were divided into sevenripening states: 1-entirely green; 2-green with a trace of yellow;3-more green than yellow; 4-more yellow than green; 5-yellowwith a trace of green; 6-entirely yellow; 7-entirely yellow withbrown speckles. Bananas at ripening stage 1 were selected andpurchased from a local market. Subsets of these bananas werestored at room temperature for different periods of time to attainripening stages 2–7. However, only bananas at stages 1–5 wereused in this study because RS samples from bananas at stages 6and 7 were colloid-like substances that were difficult to grind.All chemicals used in this study were of analytical grade.

2.2. Isolation and determination of banana resistant starch (BRS)content

Using the method reported by Cheng Yanfeng et al. (2008), ba-nanas were peeled, pulped, and digested with pectinase and amy-lase to remove pectin, cellulose, protein, and digestible starch. Thedigested banana pulp was centrifuged at 3 000 rpm for 15 min; theresulting precipitate was dehydrated at 50 �C, ground, and storedat 5 �C. RS content was determined by the method reported byGoni, Garcia-Diz, Manas, & Saura-Calixto (1996). Briefly, the meth-od consisted of the removal of protein and digestible starch, thesolubilisation and enzymatic hydrolysis of RS, and the quantifica-tion of RS. Human gastric and intestinal conditions (pH and transittime) were simulated.

2.3. Structural observations of BRS

2.3.1. Light microscopy and polarising microscopyBRS samples were dissolved in glycerol (50% concentration) and

observed under a microscope (Vanox BHS-2, Olympus Corporation,Japan) using both natural and polarised light.

2.3.2. Scanning electron microscopy (SEM)Particles of BRS powder were scanned, using a S3700N scanning

electron microscope (Hitachi, Japan). Samples were fixed on anobjective table coated with platinum (10–20 nm thickness).

2.3.3. X-ray diffraction (XRD)Cu Ka radiation was used to scan BRS samples over the 2h = 4–

60� range, with a step interval of 0.04�, a scanning rate of 17.7 s perstep, a voltage of 40 kV, and a current of 40 mA. The D8 ADVANCEX-ray diffractometer from Bruker Corporation (Germany) was usedfor the XRD analyses.

2.3.4. Infrared spectroscopyBRS samples were pressed in KBr. An infrared spectrometer

(VECTOR33, Bruker Corporation, Germany) was used to scan thesamples from 4000 cm�1 to 400 cm�1 of the infrared region.

2.4. Physicochemical properties of BRS

2.4.1. Water-holding capacity (WHC)WHC was measured by the method reported by Toyokawa,

Rubenthaler, Powers, and Schanus (1989). Briefly, 20 ml of starchsuspension (5 g/100 ml) were transferred to centrifuge tubes andheated in a water bath for 15 min at 50 �C, 70 �C, or 90 �C. Thetubes were centrifuged at 3,000 rpm for 15 min. The supernatantwas discarded; tubes containing sediment were placed at a 45� an-gle for 10 min to allow water drainage and weighed. WHC was cal-culated by Eq. (1).

WHC ð%Þ ¼ m2 �m1 �m0

m0� 100% ð1Þ

where m0 is the weight of the starch sample, m1 is the weight of thecentrifuge tube, and m2 is the weight of the starch sample and cen-trifuge tube following water drainage.

2.4.2. Solubility and swelling powerSolubility (S) and swelling power (SP) were determined, using

the method reported by Aparicio-Saguilán et al. (2005). In thisexperiment, 20 ml of starch suspension (5 g/100 ml) were trans-ferred to centrifuge tubes and heated in a water bath for 30 minat 50 �C, 70 �C, or 90 �C. After the tubes had cooled to room temper-ature, they were centrifuged at 3000 rpm for 15 min. The sedimentand supernatant were separated; the sediment was dried andweighed. S and SP were calculated using Eqs. (2) and (3),respectively.

S ð%Þ ¼ AW� 100% ð2Þ

SP ð%Þ ¼ DWð1� SÞ � 100% ð3Þ

where A is the weight of dry dissolved solids in the supernatant, Wis the weight of the sample, and D is the weight of the sediment.

2.4.3. TransparencyAn aqueous starch solution was preparing by mixing 1.0 g of

starch with 99.0 g of water. This solution was heated in a boilingwater bath for 15 min under continuous stirring and subsequentlycooled to room temperature. The transparency of the resultingstarch paste was detected at 620 nm (UV-1800 spectrophotometer,Shimadzu Co., Japan). Distilled water was used as a blank control,which was considered to have a transparency of 100%.

2.4.4. Starch–iodine absorption spectraSpectra of iodine-bound starch samples were determined, using

the method reported by Klucinec and Thompson (1998). BRS(50 mg) was dispersed into 10.0 ml of DMSO containing 10% of6.0 M urea. Subsequently, 2.0 ml of the dispersed solution, 25 mlof distilled water, and 1.0 ml of I2-KI (2.0 mg I2/ml and 20.0 mgKI/ml) were pipetted into a 50 ml volumetric flask and mixed.The mixed solution was brought to a volume of 50 ml with distilledwater. Control solutions were prepared without BRS. A UV–visiblespectrophotometer (UV-1800, Shimadzu Co., Japan) was used toscan each sample from 500 to 800 nm; kmax for each sample wasdefined as the wavelength that resulted in the highest absorbancevalue.

2.4.5. Pasting propertiesA microviscoamylograph (Visgraph-E, Brabender Instruments,

Inc., Germany) was used to determine the viscosity profiles (inBrabender units, BU) of the starch samples. Dispersions of BRS(6%, dry basis) were transferred to the microviscoamylographand subjected to thorough agitation. The dispersion was broughtto an initial temperature of 30 �C and subsequently to 95 �C at arate of 1.5 �C/min. The temperature of the dispersion was main-tained at 95 �C for 30 min; subsequently, the dispersion was cooledto 50 �C at a rate of 1.5 �C/min and maintained at 50 �C for 30 min(Aparicio-Saguilán et al., 2005).

2.5. Statistical analyses

Data were analysed by the SPSS statistical software package,v19.0 (IBM company). Data were expressed as means ± standarddeviation. One-way analysis of variance (ANOVA) was used tocompare the different BRS samples, Levene’s test was used to as-sess homogeneity of variances, and the Bonferroni test was usedfor multiple comparisons. Statistical significance was set at

J. Wang et al. / Food Chemistry 156 (2014) 319–325 321

P < 0.05. Values followed by the same letter in the same row or col-umn are not significantly different (P < 0.05) in the tables.

3. Results and discussion

3.1. Changes in BRS content during ripening

BRS content gradually decreased during storage. Cavendish BRScontent decreased rapidly during the first four ripening stages butdecreased slowly during the final three ripening stages. In contrast,Pisang Awak BRS content decreased slowly during the initial threeripening stages but decreased rapidly during the final four ripeningstages. At the same ripening stage, Pisang Awak bananas consis-tently had higher BRS content than had Cavendish bananas. Therewere significant differences in BRS content between the two bana-na cultivars, a result that may be attributed to differences in enzy-matic reactions that convert starches into sugars. The reaction rateof Cavendish bananas was possibly faster, thereby contributing to arapid reduction in BRS content.

3.2. Structural changes in BRS during ripening

3.2.1. Light microscopy and polarising microscopyMost starch particles in the Cavendish cultivar were oval in

shape, whereas others were spindly. In contrast, starch granulesin the Pisang Awak cultivar were generally round in shape, whileothers were triangular (Figs. 1a and b). In both cultivars, the edgesof the starch particles were completely intact and sharply definedduring the first ripening stage. A subset of starch particles de-graded at subsequent ripening stages; this phenomenon might beattributed to the enzymatic hydrolysis of banana starches.

Starch particles were observed by polarising microscopy(Figs. 1c and d). Maltese crosses were evident in these particles un-der polarised light, with certain particles exhibiting cross patternsand other particles displaying X-shaped patterns. The points ofthese crosses were located at the top and end of the starch parti-cles. In the Cavendish cultivar, the Maltese crosses became weakat ripening stages 4 and 5; in the Pisang Awak cultivar, the crossesbecame weak at ripening stage 5. This result revealed that therewere differences in the crystalline structures of RS between thetwo cultivars.

3.2.2. SEMFig. 1e and f shows the microscopic appearance of BRS from

Cavendish and Pisang Awak. Most starch particles had smooth sur-faces during the initial ripening stages, whereas enzymatic effectscaused starch particle surfaces to become rough and wrinkled dur-ing post-harvest ripening. At ripening stage 5, there were more de-graded and broken starch granules in the Cavendish cultivar thanin the Pisang Awak cultivar, suggesting that maturation processesoccurred more quickly in Cavendish bananas than in Pisang Awakbananas. This result revealed that enzymatic hydrolysis in Caven-dish bananas was faster than that in Pisang Awak bananas.

3.2.3. XRDXRD analyses revealed three major diffraction peaks for Caven-

dish BRS and four major diffraction peaks for Pisang Awak BRS. Areduction in the diffraction peaks during ripening indicates a lossin the crystallinity of starch structures. The crystallinities are sum-marised in Table 1. In the Cavendish cultivar, the relative degree ofcrystallinity of RS decreased very slowly from ripening stages 1–2and diminished rapidly during ripening stages 3 and 4; the differ-ences were significant. In the Pisang Awak cultivar, the relative de-gree of crystallinity of RS decreased from ripening stages 1–4; thedifferences were significant. In both cultivars, the crystallinity of

RS decreased slowly from ripening stages 4–5; however, the differ-ences were significant. Crystallinity, which represents the degreeof structural order, has significant effects on hardness, density,transparency, and diffusion (Oxford dictionary of science., 1999).In the two banana cultivars, RS crystallinity was not reduced atthe same rate, suggesting that the crystallization behaviour instarches is possibly different among banana varieties.

3.2.4. Infrared spectroscopySimilar IR spectra of RS were obtained at all five ripening stages

for Cavendish and Pisang Awak, indicating that the characteristicfunctional groups of RS were not affected by the maturationprocess (Fig. 2). The broad band at 3400 cm�1 is caused by OHgroups, whereas the band at 2800 cm�1 is generated by CH2 groups(Garcia-Rosas et al., 2009). The peak at 1600 cm�1 results fromcarboxylate ion (COO�) stretching vibrations in carboxylategroups. The IR bands at 1300 cm�1 and 1022 cm�1 are producedby C–O–H bending and C–O–H bending vibrations, respectively.The skeletal modes of the pyranose ring generate the peak at620 cm�1.

3.3. Physicochemical properties of BRS

3.3.1. WHCThe WHC of BRS at 50 �C, 70 �C, and 90 �C is shown in Table 2. In

general, higher WHC was obtained at higher temperatures. In theCavendish cultivar, higher WHC were observed during ripeningstage 4. In the Pisang Awak cultivar, BRS had the highest WHC dur-ing ripening stage 5 at 50 �C and 70 �C and during ripening stage 3at 90 �C. Therefore, different maturity levels of Cavendish and Pi-sang Awak bananas should be chosen for BRS extraction and appli-cations in accordance with the different temperature-dependentcharacteristics of banana fruits and relevant WHC requirements.

3.3.2. Solubility and swelling powerIn Cavendish, RS solubility increased with ripening (Table 2).

Therefore, at 50 �C, 70 �C, and 90 �C, higher BRS solubility was ob-served at ripening stage 5 than at the first four ripening stages. InPisang Awak, BRS solubility at 50 �C was higher at ripening stages 4and 5 than at ripening stages 1–3. However, complete BRS solubil-ity was observed at ripening stages 5 and 3 at 70 �C and 90 �C,respectively.

In Cavendish, there were no significant differences in BRS swell-ing power among the five ripening stages. In Pisang Awak, BRS hadthe highest swelling power during the first ripening stage.

3.3.3. TransparencyIn both cultivars, BRS transparency gradually decreased during

storage (Table 2). During ripening stage 5, transparency declined.At ripening stages 1 and 2, higher BRS transparency was observedfor the Cavendish cultivar than for the Pisang Awak cultivar; how-ever, the opposite was observed at ripening stages 3–5.

3.3.4. Starch–iodine absorption spectraIt has been reported that the maximum absorption wavelengths

are lower for amylopectin-iodine solutions than for amylose-io-dine solutions (Baldwin, Bear, & Rundle, 1944). Klucinec andThompson (1998), who focused on different fractions of high-amy-lose maize starches, concluded that the maximum absorptionwavelengths of amylose ranges from 643 to 655 nm but that themaximum absorption wavelengths of amylopectin ranges from559 to 583 nm. The starch–iodine absorption spectra of BRS areshown in Fig. 3. The maximum absorption wavelengths (kmax) ofCavendish BRS ranged from 560 to 580 nm, whereas the maximumabsorption wavelengths of Pisang Awak BRS ranged from 570 to610 nm, indicating that Cavendish BRS contains more amylopectin

1 2 3 4 5

a) Optical microscopy images (×200) of Musa AAA Cavendish BRS

1 2 3 4 5

b) Optical microscopy images (×200) of Musa ABB Pisang Awak BRS

1 2 3 4 5

c) Polarised light microscopy images (×200) of

) Polarised light microscopy images (×200) of

Musa AAA Cavendish BRS

1 2 3 4 5

Musa ABB Pisang Awak BRS

1 2 3 4 5

e) Scanning electron microscopy images (×500) of Musa AAA Cavendish BRS

1 2 3 4 5

f) Scanning electron microscopy images (×500) of Musa ABB Pisang Awak BRS

(

(

(

(

(

(

d

Fig. 1. Optical microscopy, polarised light microscopy and scanning electron microscopy images of resistant starch samples isolated from bananas at different ripeningstages. (The numbers 1–5 indicate the ripening stage of each banana sample.)

322 J. Wang et al. / Food Chemistry 156 (2014) 319–325

and less amylose than does Pisang Awak BRS. The maximumabsorption wavelengths of Cavendish BRS at ripening stages 1–5were similar at approximately 570 nm. However, the maximumabsorption wavelengths of Pisang Awak BRS shifted as bananasmatured through ripening stages 1–5. In particular, for ripening

stages 1, 2, 3, 4 and 5, the kmax of Pisang Awak BRS samples oc-curred at 585, 600, 605, 573, and 595 nm, respectively, suggestingthat the amylopectin content of Pisang Awak BRS decreasedquickly prior to ripening stage 3, causing kmax to shift to wave-lengths near the maximum absorption wavelength of amylose. It

Table 1The relative degrees of crystallinity (%) of banana resistant starch samples.*

Banana variety Ripe stage

1 2 3 4 5

Musa AAA Cavendish 33.8 ± 0.4a 33.6 ± 0.3a 33.1 ± 0.2b 24.4 ± 0.6c 23.5 ± 0.4d

Musa ABB Pisang Awak 43.1 ± 0.6a 40.6 ± 0.4b 36.9 ± 0.3c 33.5 ± 0.3d 32.9 ± 0.2e

Values followed by the same letter in the same row are not significantly different (P < 0.05).* Mean of three replicates ± standard error.

Fig. 2. Infrared spectra of resistant starch samples isolated from bananas at different ripening stages. (The numbers 1–5 indicate the ripening stage of each banana sample.)

Table 2The water-holding capacity, solubility, swelling power, and transparency of resistant starch samples isolated from bananas at different ripening stages.*

Banana variety Temperature (�C) Ripe stage

1 2 3 4 5

Water-holding capacity (%)Musa AAA Cavendish 50 1.131 ± 0.016d 1.197 ± 0.013c 1.395 ± 0.015b 1.598 ± 0.012a 1.569 ± 0.016a

70 1.093 ± 0.014e 1.201 ± 0.010d 1.459 ± 0.016c 1.816 ± 0.013a 1.700 ± 0.011b

90 2.513 ± 0.012d 2.606 ± 0.012c 3.006 ± 0.010b 3.716 ± 0.013a 3.034 ± 0.014b

Musa ABB Pisang Awak 50 1.219 ± 0.013b 1.130 ± 0.016c 1.130 ± 0.015c 1.159 ± 0.021c 1.432 ± 0.020a

70 1.274 ± 0.011b 1.273 ± 0.019b 1.296 ± 0.017b 1.306 ± 0.015ab 1.345 ± 0.016a

90 3.213 ± 0.014c 3.161 ± 0.020c 3.675 ± 0.019a 3.543 ± 0.023b 2.190 ± 0.013d

Solubility (%)Musa AAA Cavendish 50 1.50 ± 0.04e 2.19 ± 0.06 d 2.54 ± 0.08c 3.34 ± 0.08b 6.22 ± 0.06a

70 1.69 ± 0.05e 2.79 ± 0.07 d 3.77 ± 0.05c 4.82 ± 0.05b 6.93 ± 0.07a

90 5.31 ± 0.05e 5.84 ± 0.08 d 6.86 ± 0.05c 7.14 ± 0.07b 9.24 ± 0.05a

Musa ABB Pisang Awak 50 1.30 ± 0.04d 1.90 ± 0.03 c 2.57 ± 0.05b 3.30 ± 0.08a 3.19 ± 0.08a

70 1.50 ± 0.07d 1.90 ± 0.04 c 3.56 ± 0.06b 3.40 ± 0.05b 4.49 ± 0.10a

90 6.99 ± 0.06c 6.99 ± 0.06 c 9.70 ± 0.07a 8.40 ± 0.06b 6.98 ± 0.04c

Swelling power (%)Musa AAA Cavendish 50 88.32 ± 3.47 86.34 ± 3.01 84.86 ± 3.88 83.54 ± 3.61 84.17 ± 3.32

70 87.76 ± 3.27 89.00 ± 3.19 86.58 ± 4.35 84.36 ± 3.42 86.06 ± 3.5090 85.13 ± 4.18 88.36 ± 2.96 86.77 ± 3.13 84.54 ± 3.82 85.16 ± 3.78

Musa ABB Pisang Awak 50 96.98 ± 2.13a 91.98 ± 1.72ab 88.87 ± 1.81b 91.25 ± 2.51ab 89.46 ± 2.07b

70 100.86 ± 2.67a 91.91 ± 2.19b 88.86 ± 1.92b 91.80 ± 2.22b 89.80 ± 2.64b

90 98.37 ± 2.16a 93.19 ± 1.70ab 86.95 ± 1.89b 92.74 ± 2.14ab 89.34 ± 2.49b

Transparency (%)Musa AAA Cavendish 25 2.893 ± 0.017a 2.701 ± 0.014b 2.17 ± 0.036c 1.096 ± 0.039d 0.832 ± 0.013e

Musa ABB Pisang Awak 25 2.727 ± 0.021a 2.559 ± 0.023b 2.377 ± 0.026c 2.271 ± 0.031d 1.506 ± 0.017e

Values followed by the same letter in the same row are not significantly different (P < 0.05).* Mean of three replicates ± standard error.

J. Wang et al. / Food Chemistry 156 (2014) 319–325 323

is likely that amylose was rapidly degraded between ripeningstages 3 and 4, causing kmax to shift to 573 nm. At ripening stage5, both amylopectin and amylose contents continued to decreasein the Cavendish and Pisang Awak cultivars, as indicated by areduction in the starch–iodine absorption peaks with increasingbanana fruit maturity.

3.3.5. Pasting propertiesThe pasting properties of BRS were measured by a Brabender

microviscoamylograph; the results are shown in Table 3. The initialgelatinisation temperatures of Cavendish BRS were similar duringripening stages 1–3, whereas the gelatinisation temperatures of Pi-sang Awak BRS increased as banana fruits matured. In Cavendish

Fig. 3. Starch–iodine absorption spectra of resistant starch samples isolated from bananas at different ripening stages. (The numbers 1–5 indicate the ripening stage of eachbanana sample.)

Table 3Banana resistant starch viscosity parameters, as measured by the Brabender microviscoamylograph .*

Banana variety Ripe stage A(�C) B(BU) C(BU) D(BU) E(BU) F(BU) B-D(BU) E-D(BU) E-B(BU)

Musa AAA Cavendish 1 78.8 ± 0.5a 498 ± 35a 490 ± 36a 377 ± 40a 588 ± 33a 519 ± 26a 121 211 902 78.4 ± 0.7a 469 ± 31a 467 ± 32a 311 ± 30b 501 ± 35b 445 ± 34b 158 190 323 79.1 ± 0.7a 466 ± 23a 462 ± 28a 328 ± 39b 496 ± 31b 442 ± 33b 138 168 30

Musa ABB Pisang Awak 1 65.5 ± 1.0b 302 ± 26b 300 ± 26b 240 ± 35c 329 ± 22c 293 ± 20c 62 89 272 79.9 ± 0.8c 270 ± 31b 263 ± 37b 209 ± 24c 292 ± 30c 272 ± 32c 61 83 223 80.0 ± 0.7c 266 ± 37b 262 ± 44b 230 ± 40c 345 ± 37c 314 ± 31c 36 115 79

BU = Brabender unit.Values followed by the same letter in the same column are not significantly different (P < 0.05).* A-initial gelatinization temperature; B-peak viscosity (BU); C-viscosity at 95 �C (BU); D-viscosity after 30 min at 95 �C (BU); E-viscosity at 50 �C (BU); F-viscosity after30 min at 50 �C (BU); B-D: Breakdown; E-D: Consistency. E-B: Setback. Mean of three replicates ± standard error.

324 J. Wang et al. / Food Chemistry 156 (2014) 319–325

BRS, the properties of peak viscosity after gelatinisation (point B),viscosity at 95 �C (point C), viscosity at 50 �C (point E), and viscos-ity after 30 min at 50 �C (point F) decreased during ripening stages1–3, whereas the viscosity after 30 min at 95 �C (point D) was notaffected. In contrast, in Pisang Awak BRS, only the peak viscosity atpoint B and the viscosity at 95 �C (point C) decreased during ripen-ing stages 1–3. The viscosity of Cavendish BRS was consistentlyhigher than the viscosity of Pisang Awak BRS at points A–F. Caven-dish BRS had the highest peak viscosity (point B) during the firstripening stage.

The heat stability of BRS paste determines the viscosity after30 min of incubation at 95 �C (B-D; breakdown), whereas the coldstability of BRS determines the viscosity after 30 min of incubationat 50 �C (E-D; consistency). The setback refers to the difference be-tween the peak viscosity after gelatinisation and the viscosity at50 �C (Qian & Kuhn, 1999). In this study, Cavendish BRS had thehighest breakdown (B-D) and consistency (E-D) values at ripeningstage 2 and the maximum setback (E-B) value at ripening stage 1.Pisang Awak BRS had the highest breakdown value during the firstripening stage and the highest consistency and setback values dur-ing ripening stage 3. Moreover, higher breakdown (B-D), consis-tency (E-D), and setback (E-B) values were obtained fromCavendish BRS than from Pisang Awak BRS, with the exception ofthe setback (E-B) value at the third ripening stage.

4. Conclusions

The results of this study revealed that RS content decreasedduring postharvest storage. At the same ripening stage, BRS con-tent was consistently higher for the Pisang Awak cultivar thanfor the Cavendish cultivar.

Light microscopy and SEM observations revealed that BRS par-ticles in the two banana cultivars exhibited different microscopiccharacteristics. Starch particles from Cavendish BRS were oval,whereas starch granules from Pisang Awak BRS were round. TheCavendish banana fruits matured more quickly than did the Pisang

Awak banana fruits. The Maltese crosses of BRS were clear underpolarised light at the beginning of storage but became weak duringthe ripening process. BRS crystallinity gradually decreased duringthe ripening process. The infrared spectra of Cavendish and PisangAwak BRS were similar and remained relatively unaffectedthroughout the maturation of banana fruits, suggesting that thetypical functional groups of RS were well maintained at all fiveexamined ripening stages.

Differences in WHC, solubility, swelling power, and transpar-ency were observed in BRS from the two cultivars and at differentripening stages. Starch–iodine absorption spectra indicated thatCavendish BRS contained more amylopectin and less amylose thandid Pisang Awak BRS. The degradation processes of amylopectinand amylose in the two cultivars were asynchronous. Pasting prop-erties, determined by Brabender microviscoamylograph analyses,revealed differences in the heat stability (breakdown value), coldstability, and setback characteristics of BRS from the two cultivarsat ripening stages. In general, viscosity values were higher forCavendish BRS than for Pisang Awak BRS. These results reveal thatappropriate cultivars and ripening stages should be chosen for thepreparation of RS products.

Acknowledgments

This research was supported by The National Natural ScienceFoundation of China (Grant No. 31301530), the Fundamental Re-search Funds for the Central Universities, SCUT(Grant No.2013ZM0063), the Foundation for Distinguished Young Talents inHigher Education of Guangdong, China (Grant No. LYM10016)and the Bureau of Science and Information Technology of Guangz-hou, China (Grant No. 2013J4100056).

References

Aparicio-Saguilán, A., Flores-Huicochea, E., Tovar, J., García-Suárez, F., Gutiérrez-Meraz, F., & Bello-Pérez, L. A. (2005). Resistant starch-rich powders prepared by

J. Wang et al. / Food Chemistry 156 (2014) 319–325 325

autoclaving of native and lintnerized banana starch: Partial characterization.Starch-Starke, 57(9), 405–412.

Aparicio-Saguilan, A., Gutierrez-Meraz, F., Garcia-Suarez, F. J., Tovar, J., & Bello-Perez, L. A. (2008). Physicochemical and functional properties of cross-linkedbanana resistant starch. Effect of pressure cooking. Starch-Starke, 60(6),286–291.

Aparicio-Saguilan, A., Sayago-Ayerdi, S. G., Vargas-Torres, A., Tovar, J., Ascencio-Otero, T. E., & Bello-Perez, L. A. (2007). Slowly digestible cookies prepared fromresistant starch-rich lintnerized banana starch. Journal of Food Composition andAnalysis, 20(3–4), 175–181.

Asp, N. G. (Ed.). (1992). Resistant starch. In: Proceedings from the second plenarymeeting of EURESTA: European FLAIR Concerted Action No. 11 on physiologicalimplications of the consumption of resistant starch in man (p. 46, Suppl. 2, S1).European Journal of Clinical Nutrition.

Baldwin, R. R., Bear, R. S., & Rundle, R. E. (1944). Relation of starch iodine absorptionspectra to the structure of starch and starch components. Journal of the AmericanChemical Society, 66(1), 111–115.

Ble-Castillo, J. L., Aparicio-Trapala, M. A., Francisco-Luria, M. U., Cordova-Uscanga,R., Rodriguez-Hernandez, A., Mendez, J. D., & Diaz-Zagoya, J. C. (2008). Bananaresistant starch flour supplementation reduces body weight and insulinresistance in obese type 2 diabetics. Obesity, 16 (Supply), pp. 152–153.

Cheng Yanfeng, Y. G., Wang, J., Du, B., Bao, J., Wen, S., & Ding, W. (2008). Effects ofspray drying technologies on the retention rate of banana resistant starch.Transactions of the Chinese Society of Agricultural Engineering, 24(6), 282–286.

Englyst, H. N., Kingman, S. M., & Cummings, J. H. (1992). Classification andmeasurement of nutritionally important starch fractions. European Journal ofClinical Nutrition, 46(Suppl. 2), S33–S50.

Fuentes-Zaragoza, E., Sanchez-Zapata, E., Sendra, E., Sayas, E., Navarro, C.,Fernandez-Lopez, J., et al. (2011). Resistant starch as prebiotic: A review.Starch-Starke, 63(7), 406–415.

Garcia-Rosas, M., Bello-Perez, A., Yee-Madeira, H., Ramos, G., Flores-Morales, A., &Mora-Escobedo, R. (2009). Resistant starch content and structural changes inmaize (Zea mays) tortillas during storage. Starch-Starke, 61(7), 414–421.

Goni, I., Garcia-Diz, L., Manas, E., & Saura-Calixto, F. (1996). Analysis of resistantstarch: A method for food sand food products. Food Chemistry, 56(4), 445–449.

Johnson, I. T., & Gee, J. M. (1996). Resistant starch. Nutrition and Food Science, 96(1),20–23.

Klucinec, J. D., & Thompson, D. B. (1998). Fractionation of high-amylose maizestarches by differential alcohol precipitation and chromatography of thefractions. Cereal Chemistry, 75(6), 887–896.

Mutungi, C., Rost, F., Onyango, C., Jaros, D., & Rohm, H. (2009). Crystallinity, thermaland morphological characteristics of resistant starch type III produced byhydrothermal treatment of debranched cassava starch. Starch-Starke, 61(11),634–645.

Niba, L. L. (2002). Resistant starch: A potential functional food ingredient. Nutritionand Food Science, 32(2/3), 62–67.

Oxford dictionary of science. (1999). <http://en.wikipedia.org/wiki/Crystallinity>Accessed 13.10.14.

Qian, J. Y., & Kuhn, M. (1999). Evaluation on gelatinization of buckwheat starch: acomparative study of Brabender viscoamylography, rapid visco-analysis, anddifferential scanning calorimetry. European Food Research and Technology, 209,277–280.

Soltani, M., Alimardani, R., & Omid, M. (2011). Evaluating banana ripening statusfrom measuring dielectric properties. Journal of Food Engineering, 105, 625–631.

Topping, D. L., & Clifton, P. M. (2001). Short-chain fatty acids and human colonicfunction: Roles of resistant starch and nonstarch polysaccharides. PhysiologicalReviews, 81(3), 1031–1064.

Toyokawa, H., Rubenthaler, G. L., Powers, J. R., & Schanus, E. G. (1989). Japanesenoodle qualities. II. Starch components. Cereal Chemistry, 66(5), 387–391.