2023 Volume 70 Issue 1 Pages 25-32
Starch is comprised of very large α-glucan molecules composed primarily of linear amylose and highly branched amylopectin. Most methods for analyses of starch structure use hydrolytic enzymes to cleave starch. When undegraded, whole starch structures can be analyzed by gel-permeation chromatography (GPC), but this typically yields a single peak each for amylopectin and amylose. The objective of this study was to stably separate amylopectins in whole starch based on their molecular weight using GPC, and to determine the structure of each peak. When alkali-gelatinized whole starch was applied to GPC columns (Toyopearl HW75S × 2, HW65S, and HW55S), it was separated into three peaks. Iodine staining and chain length distribution analyses of debranched samples showed that peaks were mainly composed of high-molecular weight (MW) amylopectin consisting of many clusters, low-MW amylopectin consisting of a small number of clusters, and amylose.
GPC, gel-permeation chromatography; MW, molecular weight; Mn, number-average MW; Mw, weight-average MW; DP, degree of polymerization.
Starch is a carbohydrate produced in plants by photosynthesis and mainly stored in storage organs (seeds, cotyledons, and tubers). Starch is a homopolymer of glucose units composed of amylose, a primarily linear molecule connected by α-1,4 glucosidic bonds, and amylopectin, a highly branched molecule linked via α-1,4 and α-1,6 glucosidic bonds. The higher-order structure of starch is still unknown, but there are two models of the molecular structure of amylopectin: the cluster model1)2) proposed by many researchers and the building block and the backbone model.3) Nakamura and Kainuma4) reviewed these models, explored inconsistencies in the building block and the backbone model, and showed that the cluster model is less inconsistent. According to the cluster model, unit chains of amylopectin are organized in a cluster structure, in which branching positions are concentrated in an amorphous region. Clusters consist of amorphous and crystalline regions, and the latter are filled with double helices of α-glucan chains and a few branching points.5) The nonrandom nature of branching positions has been recognized as the origin of an alternating lamellar structure of crystalline and amorphous regions, and results in multiple clusters connected in a tandem fashion.6) However, the detailed starch structure, such as arrangement of amylose and amylopectin, and size of amylopectin molecules, remains unknown.7)
An analytical method that uses undegraded whole starch can be used to assess the macrostructure and molecular weight (MW) of starch. However, in most starch structural analyses, starch branches are cleaved by various kinds of amylolytic enzymes with well-established specificities, and the resulting small molecules are analyzed.
In previous studies, gel-permeation chromatography (GPC) of whole starch using a Sephacryl S1000 column (Pharmacia Biotech, Uppsala, Sweden) yielded two peaks, one each for amylopectin and amylose in order of decreasing MW.8)9) GPC of whole starch using a Shodex OH pak KB-G guard column and KB-806 and KB-804 analytical columns (Showa Denko K.K., Tokyo, Japan) also yielded two peaks, one each for amylopectin and amylose.10)11) Yoo and Jane11) estimated that the weight-average MW (Mw) of rice amylopectin using high-performance size-exclusion chromatography with multiangle laser-light scattering (HPSEC-MALLS) was 27 × 108. GPC analyses of whole corn starch using Toyopearl HW75S × 2, HW65S, and HW55S columns revealed two peaks derived from amylopectin and amylose.12) Takeda et al.13) found that when undegraded amylopectin purified from different plant materials was separated by GPC (HW40S, HW50S, and HW75S packed in three layers), only one broad peak was obtained using a refractive index (RI) detector, while three peaks were obtained using a fluorescence detector. Thus, in most previous studies, GPC of whole starch yielded only two peaks, one each for amylopectin and amylose, indicating that size separation of amylopectin is insufficient to assess its MW distribution, an important factor for characterizing macromolecules with varying MW.
In the present study, we succeeded in stably separating whole starch into two peaks of amylopectins with different MW and a third peak composed of small amylopectin and amylose by GPC, using Toyopearl HW75S × 2, HW65S, and HW55S columns with an RI detector. The debranched starch structure, and chain-length distributions of amylopectins in each peak were further analyzed. Although the separation by these columns is not perfect, we tried to estimate the molecular weight at the peak-top position using pullulan standards with known MW and amylopectin-derived Cluster Dextrin with defined MW. Furthermore, we estimated average number of clusters of unit chains in an amylopectin molecule in relation to its MW.
Materials. The wild-type Japonica rice cultivar ‘Nipponbare’ harvested in Akita Prefecture in 2018 or 2019 as well as corn starch (J-oil mills Inc., Tokyo, Japan) was used in the present study.
A pullulan series comprising P-5 (Mw 0.58 × 104), P-10 (Mw 1.22 × 104), P-20 (Mw 2.37 × 104), P-50 (Mw 4.80 × 104), P-100 (Mw 10.0 × 104), P-200 (Mw 18.6 × 104), P-400 (Mw 38.0 × 104), P-1600 (Mw 166 × 104), Shodex STANDARD P-82 (Showa Denko, Tokyo, Japan) were used as MW markers. Cluster Dextrin (Glico Nutrition Co., Ltd., Osaka, Japan) and maltotetraose (Kikkoman Co., Chiba, Japan) were also used as a standard sample.
Purification of rice starch. Starch was extracted and purified from polished rice grains (5 g) using the cold-alkaline method as described previously.14)15)
Toyopearl HW75S × 2, HW65S, and HW55S GPC columns chromatography. Purified starch (30 mg for fractionation of rice starch and 20 mg for corn starch) was suspended in 1.6 mL of distilled water in a 15 mL centrifuge tube, and gelatinized with 400 µL of 5 M NaOH for 30 min at 37 °C. Distilled water (2 mL) and 2 mL of eluent (0.2 % NaCl / 0.05 M NaOH) were added, and samples were filtered through a 5 µm Durapore polyvinylidene fluoride (PVDF) membrane (Merck KGaA, Darmstadt, Germany). A 5 mL sample of filtrate was applied to a 4-column system comprising Toyopearl HW75S × 2, HW65S, and HW55S columns (2.2 cm diameter, 30 cm × 4; Tosoh Corp., Tokyo, Japan) pre-equilibrated with 0.2 % NaCl / 0.05 M NaOH, and carbohydrate was detected using a RI-8020 RI detector (Tosoh Corp.).16) Columns were incubated at 40 °C, and samples were eluted at a flow rate of 1 mL/min by using a DP-8020 pump (Tosoh Corp.).
Starch analysis of GPC fractions. Each 10 mL fraction was collected by a fraction collector, and 180 µL of each fraction was neutralized with 10 µL of 1 M HCl and stained with 10 µL of 1 % KI / 0.1 % I2. Starch in each fraction (#1-6 in Fig. 1) was precipitated with three volumes of ethanol at －30 °C overnight. Precipitates were washed with 70 % ethanol and dried. α-glucans of each fraction were debranched at 37 °C for 24 h by 354 U of isoamylase derived from Pseudomonas amyloderamosa (Hayashibara Co., Ltd., Okayama, Japan) according to the previous report.17) GPC analysis of debranched starch was performed using a Toyopearl HW55S column and three columns of HW50S connected in series (2.2 cm diameter × 30 cm each; Tosoh Corp.) following the method reported by Toyosawa et al.18) Chain length distribution analysis of each fraction was performed by fluorophore-assisted capillary electrophoresis (FACE) with a P/ACE MDQ Carbohydrate System (Sciex LLC, Framingham, MA, USA) according to the methods of the previous reports.17)19) The difference in chain length distribution was expressed as molar change (%), equivalent to the rate of molar change of each chain (%) in Fujita et al.,20) and was calculated for each DP as the differences in molar (%) at each DP value vs. that of Fraction #1. Differences are expressed by the percentage of the molar % value of a given DP for Fraction #1.
Typical patterns among at least two reproducible data were shown in the figure. Images show iodine staining of respective fraction. The # numbers indicate the fractions collected from each peak (#1, #3, and #5), valleys between Peak 1 and Peak 2 (#3) and Peak 2 and Peak 3 (#4) and tail of Peak 3 (#6).
Iodine staining of whole starch fractions separated by GPC.
GPC separation of whole rice starch using Toyopearl HW75S × 2, HW65S, and HW55S columns yielded three peaks with elution times of 135, 170, and 215 min (Fig. 1A). Peak 1 was the sharpest and largest, while Peaks 2 and 3 were broader in shape. Each fraction was collected over 5 min from 95 min to 275 min, and aliquots were stained with iodine because starch and iodine react to form a colored complex (Fig. 1A). Frequently branched chains of amylopectin bind to iodine, yielding a red color,21) while long linear chains of amylose bind to iodine to give a blue color. Fraction #1 (Peak 1) and Fraction #2 (between Peaks 1 and 2) of whole rice starch were red following iodine staining (Fig. 1A). Therefore, it could be inferred that these fractions contain a large amount of amylopectin. By contrast, Fraction #3 (close to Peak 2) was purple, indicating that both amylopectin and amylose were eluted together (Fig. 1A). Fraction #4 and subsequent fractions were blue, indicating that they contained a large amount of amylose (Fig. 1A).
GPC analysis of corn starch by the same column also showed three peaks, although the height of Peak 1 was comparable to that of Peak 2 (Fig. 1B). Previous GPC analysis of whole corn starch separated by the same column series (Toyopearl HW75S × 2, HW65S, and HW55S) revealed only two peaks derived from amylopectin and amylose.12) One of the explanations of the inconsistency between the previous12) and the current studies might be the different preparation of the sample, the amount of application and concentration of starch, conditions of elution, etc., although the detailed condition was unclear in the previous work. In this work, we chose relatively low concentration of starch (0.4 %) and high amount for applied volume (5 mL). This condition might have achieved higher resolution of amylopectin molecules through the long columns (total of 120 cm).
Structural analysis of debranched α-glucans fractions separated by GPC.
Next, starch in fractions of whole starch obtained by GPC (Fig. 1A) was precipitated with three volumes of ethanol, debranched with Pseudomonas isoamylase, and separated using Toyopearl HW55S and HW50S × 3 columns (Fig. 2). The GPC pattern of debranched purified whole starch from the ‘Nipponbare’ rice cultivar was in three peaks (Frac. I－III), of which Frac. I mainly contained amylose, Frac. II mainly contained long-chain of amylopectin, and Frac. III mainly contained short-chain of amylopectin (Fig. 2A). Based on a calibration curve using pullulan standards with known MW (P-100 to P-400), peak-top MW of Frac. I was 29.9 × 104. Based on a calibration curve using pullulan standards (P-5 to P-50) and maltotetraose (G4), peak-top MW of Frac. II and III was 1.0 × 104, and 0.3 × 104, respectively.
The percentages in graphs indicate the ratios of each peak to total carbohydrates (%). Amylose, long-chain of amylopectin, and short-chain of amylopectin were mainly eluted in Fractions I, II, and III, respectively. (A) Debranched whole starch in endosperm of Japonica rice cultivar ‘Nipponbare’. Filled circles indicate the MW of pullulan standards (P-400, 38.0 × 104; P-200, 18.6 × 104; P-100, 10.0 × 104; P-50, 4.80 × 104; P-20, 2.37 × 104, P-10, 1.22 × 104; P-5, 0.58 × 104) and maltotetraose (G4). (B) Fraction #1, (C) Fraction #2, (D) Fraction #3, (E) Fraction #4, and (F) Fraction #5 from GPC separation of whole starch (Fig. 1A). P-100 to P-400 were used for estimation of peak-top MW of Frac. I. P-5 to P-50 and G4 were used for estimation of peak-top MW of Frac. II and III.
The Fraction #1 of the GPC-separated whole starch (Fig. 1A) contained a small amount of amylose (Frac. I: 6.1 %, Fig. 2B) and a high levels of long amylopectin chains (Frac. II: 22.4 %) and short amylopectin chains (Frac. III: 71.5%), indicating that a large proportion of the Peak 1 eluted from whole starch separated by GPC (Fig. 1A) was amylopectin (Fig. 2B). A higher percentage of amylopectin (Frac. II: 21.3 %, Frac. III: 70.2 %; Fig. 2C) than amylose (Frac. I: 8.45 %; Fig. 2C) was also observed in Fraction #2 of GPC-separated whole starch (between Peaks 1 and Peak 2; Fig. 1A). In Fractions #1 and #2 of the GPC-separated whole starch, the small Frac. I peak was assigned as amylose, because the GPC columns could not completely separate amylopectin from amylose due to its insufficient resolution. However, amylopectin also contains unbranched long chains similar to amylose, known as extra-long chains (ELC).22) ELC of amylopectin may also be present in Peak 1 of GPC-separated whole starch, but the amount of ELC in endosperm starch in wild-type Japonica rice is known to be very low.23) Therefore, the percentage of ELC in Peak 1 is much lower than the true amylose content of starch, which could not be separated. A higher percentage of amylopectin (Frac. II: 23.9 % and Frac. III: 65.5 %) than amylose (Frac. I: 10.6 %) was also present in Fraction #3, which corresponds to Peak 2 of GPC-separated whole starch (Fig. 1A). The percentage of amylose in Fraction #3 was higher than that in Fractions #1 and #2, suggesting that these fractions contained more amylose (Fig. 2D). This is consistent with the fact that iodine staining of Fraction #3 yielded a purple color (Fig. 1A, Fraction #3).
A large amount of amylose (Frac. I: 56.1 %) was present in Fraction #4 (between Peaks 2 and Peak 3 of GPC-separated whole starch) (Fig. 2E), and even a larger amount of amylose (Frac. I: 70.l %) was present in Fraction #5 (Peak 3 of GPC-separated whole starch), indicating that these fractions contained abundant amylose (Fig. 2F). Fraction #5 also yielded Frac. II and Frac. III of GPC-separated debranched starch, indicating that it contained some amylopectin. This may be because amylopectin could not be clearly separated from whole starch by GPC due to its low resolution. Since amylose contains a small number of branches,24)25) it is possible that some amylose branches were present in Fracs. II and III of GPC-separated debranched starch. No peak was detected in Fraction #6 because the amount of starch that could be fractionated from whole starch by GPC was small and below the detection limit for debranched starch (data not shown).
The ratio of Frac. III to Frac. II (III/II) for non-fractionated ‘Nipponbare’ whole starch was 3.2 (Fig. 2A); the III/II ratio for Fractions #1 and #2 was 3.2 and 3.3, respectively, similar to non-fractionated ‘Nipponbare’ whole starch (Figs. 2A-C). However, these values were higher than those for Fraction #3 (2.7), Fraction #4 (2.6), and Fraction #5 (2.3), which contained a higher percentage of amylose (Figs. 2D-F). The reason for the decrease in the III/II ratio may be because some portion of amylose was also detected in Frac. II, although amylose is mainly detected in Frac. I.
Chain length distribution of whole starch fractions separated by GPC.
To investigate the differences in the fine structure of amylopectin in each whole starch fraction separated by GPC, we performed chain length distribution analysis of debranched starch by capillary electrophoresis with precolumn labeling of the reducing termini of unit chains with the charged fluorophore 8-aminopyrene-1,3,6-trisulfonic acid (Fig. 3A). The chain length distributions of Fractions #1‒6 (Fig. 1A) were almost identical, especially Fractions #1‒3. These results are well explained by the cluster model; each of these GPC size-separated fractions of whole starch consisted of amylopectins with different numbers of cluster units per molecule, but the structure of cluster units was essentially identical.
(A) Molar percentage for each liberated chain to the total chains after debranching starch. (B) Rate of molar change in chain length distribution for each fraction compared to Fraction #1.
The chain length distributions by capillary electrophoresis of Fractions #1‒3, which contained mainly amylopectin (Fig. 2) were almost identical. By contrast, Fractions #4 and #5 were considered to be mainly amylose with co-eluted small amylopectins, and Fraction #6 was composed primarily of amylose, based on the results of iodine staining (Fig. 1A) and the results shown in Fig. 2. In more detailed comparisons of molar change values (Fig. 3B), Fractions #5 and #6 differed significantly from the other fractions consisting of larger molecules. Compared to Fraction #1, the amounts (by mole) of unit chains with DP > approximately 20 were smaller in Fractions #5 and #6. The longer the unit-chain length, the lesser the amounts of unit chains in Fractions #5 and #6 in the DP range of approximately 20-30. The chains with DP 35-50 decreased by 10 % from those of Fraction #1. The presence of fewer chains with DP < approximately 20 needs careful consideration. A straightforward assumption is that these chains were released from branched amylose molecules since Fractions #5 and #6 contained mostly amylose. Amylose is a mixture of linear molecules and slightly branched molecules,24)25) and the side chain distribution of amylose can differ from that of amylopectin from the same starch, with slightly fewer long chains and more short chains,26)27) consistent with the results for Fractions #5 and #6. Alternatively, the side chains detected in Fractions #5 and #6 could be small amylopectins. Takeda et al.13) revealed that upon GPC separation of amylopectins isolated from starch, such small amylopectins gave rise to a tail on the single prominent amylopectin peak by RI detection, in which < 3 % of total carbohydrate was detected. However, small amylopectins accounted for 20‒40 % of all amylopectins by fluorescence detection. Thus, it is reasonable to conclude that Peak 3 in Figure 1 is comprised of both amylose and small amylopectins, and the side chains detected in Fractions #5 and #6 (Fig. 3) are the sum of those from both amylose and small amylopectins, although the contribution of short amylose chains is relatively small. The number of side chains in ‘Nipponbare’ rice amylose was previously estimated to be 1.4,27) while the small amylopectins of Japonica rice is composed of 35 units chains per molecule.13) The number of unit chains per molecule for the small amylopectins is 25 times higher than that of amylose.
Based on the general values for amylose content (18 %) and small amylopectin content (3 % of amylopectin) for Japonica rice in the literature,13) the weight ratio of amylose and small amylopectin was calculated to be 18:2.5. On the other hand, the amylose and small amylopectin were reported to have number-average DP values of 860 and 700, respectively.13)27) The number-average DP implies a molar ratio of 1.2:1.4, provided the amounts of amylose and amylopectin by weight are the same. Multiplying the respective values of the estimated weight and molar ratio and the number of side chains per molecule reported previously of amylose (1.4) and small amylopectin (35)13)27) yields a ratio for the origin of short unit chains of 30:123 (18 × 1.2 × 1.4 for amylose, 2.5 × 1.4 × 35 for small amylopectin). This value indicates that the contributions of amylose and small amylopectin to the short-chain profile are roughly 20 and 80 %, respectively. Thus, we concluded that the chain length distribution of Fractions #5 and #6, which contained fewer long chains and more short chains than Fractions #1‒#3, essentially reflects small amylopectins eluted in these fractions. Hizukuri proposed that unit chains of amylopectin can be classified into five groups, A, B1, B2, B3, and B4, according to their chain length.5) A and B chains differ in their branching state: A chain carries no side chain, and B chain carries other side chains. A and B chains also differ in their average chain-length. The subscripts of B-chain groups denote the number of cluster units in which a B chain is involved. For example, B2 chain connects two cluster units and has enough length to span the two clusters. The DP value (chain length) of the decreased amount (by mole) of chains in small amylopectin molecules in Fractions #5 and #6 was approximately 20-60 (Fig. 3B), and the chain lengths correspond to B2 or B3 chains of amylopectin. Because these B chains are recognized as cluster-connecting chains, it seems rational that the cluster-connecting chains are less abundant in small amylopectin molecules.
Estimation of MW of whole starch fractions separated by GPC.
Estimating molecular weights of solutes by GPC generally requires column calibration with suitable polymer standards, but no such standard for GPC separation of amylopectins is commercially available. Therefore, instead of column calibration, estimation of MW for these fractions was attempted as follows. Firstly, a calibration curve for the GPC column was constructed using pullulan standards with known MW. Next, for Peak 3, which comprised essentially linear amylose, the MW of amylose was estimated directly by applying the calibration curve (Fig. 4). Secondly, Cluster Dextrin, details of which are given below, was analyzed using the same columns. Finally, the defined MW of Cluster Dextrin and the calibration curve based on pullulan standards were compared to verify the differences in MW for highly branched molecules and linear molecules at the same elution volume. Pullulan, an extracellular α-glucan produced by Aureobasidium pullulans, has been commonly used as a MW marker in GPC of starch-related α-glucans because some commercial pullulan products with a defined average MW and narrow MW distribution are readily available. Pullulan consists of maltotriosyl residues connected via α-1,6 linkages and is thus regarded as a liner polysaccharide. Cluster Dextrin is produced by the unique intramolecular cyclization reaction of a bacterial branching enzyme on amylopectin.28) The enzymatic end-product has been reported to retain the structure of the cluster unit of the amylopectin substrate, from less than 1 unit to approximately 2 units.28) In an amylopectin molecule, unit chains of α-1,4 glucosyl chains are organized in a cluster fashion, and multiple clusters combine to build a molecule of amylopectin. Thus, Cluster Dextrin, which retains the original amylopectin cluster(s), is expected to behave in a similar manner to amylopectin during separation by GPC.
The black line shows the peak from starch in endosperm from Japonica rice cultivar ‘Nipponbare’ (Fig. 1) and the gray line shows the peak from Cluster Dextrin. Filled circles indicate the MW of pullulan standards (P-1600, 166 × 104; P-400, 38.0 × 104; P-200, 18.6 × 104; P-100, 10.0 × 104; P-50, 4.80 × 104; P-20, 2.37 × 104; P-10, 1.22 × 104; P-5, 0.58 × 104. The MW for each peak-top of Peaks 1-3 (open circles) was estimated from the extended (dashed line) calibration curve of pullulan standards.
A linear relationship between elution volume and MW was obtained by applying eight pullulan standards (P-5 to P-1600) onto the GPC column, with MW ranging from 5.8 × 103 to 166 × 104 (Fig. 4). The linear region covers Peak 3, which is essentially composed of amylose, and 100 % or 90 % (excluding the highest 10 %), respectively, of the reported values for Mw of HPSEC fractions of isolated amyloses from Japonica rice29) or various botanical sources;30) therefore, the established GPC conditions allow for assessing the MW distribution of amylose without amylose isolation from starch granules.
Cluster Dextrin was eluted from the column as a slightly broad single peak with a peak-top between those of pullulan P-50 (MW 4.80 × 104) and P-100 (10.0 × 104), and the MW at the peak-top was tentatively calculated to be 5.2 × 104 according to the calibration curve (Fig. 4). This value can be regarded as ‘pullulan equivalent,’ and the actual value should be higher than 5.2 × 104 due to structural differences, especially in branching, between linear molecules of pullulan and highly branched molecules of Cluster Dextrin. In GPC separation, solutes are separated according to their hydrodynamic volumes, a product of limiting viscosity number and MW, [η]MW.31) Since [η] is dependent on the branching of an analyte, α-1,4 glucans with different degrees of branching yield different calibration curves.32)Therefore, Cluster Dextrin, which has a highly branched structure, should exhibit higher MW than linear pullulan when compared at the same elution volume. The weight-average DP of Cluster Dextrin was determined to be 900 by the HPSEC-MALLS method.28) The reported DP value can be converted, by multiplying by 162, to a Mw of 14.6 × 104, which is 2.8-fold higher than the MW obtained from pullulan calibration. Takata et al.28) further revealed that Cluster Dextrin could be separated by HPSEC into two major components, each accounting for approximately 50 % by weight, and their Mw and peak-top MW were 19.6 × 104 and 12.5 × 104 for the larger component and 6.2 × 104 and 7.5 × 104 for the smaller component, respectively. Assuming that the peak-top of a single peak obtained in this study would be positioned between the two peaks reported by Takata et al.,28) the Mw at the peak-top in this study would be in the range of 7.5‒12.5 × 104, which is 1.4‒2.4-fold higher than the MW estimated from pullulan calibration.
From a factor of approximately 2 for the Mw of pullulan and Cluster Dextrin eluted in the same volume, the MW of Peaks 1 and 2 was estimated by extrapolating the calibration curve from pullulan standards (Fig. 4). Extrapolation should be valid because a high correlation between Mw by HPSEC using pullulan standards and that by the HPSEC-MALLS-RI method has been reported for corn amylopectins.33) For Peak 3, MW of small amylopectin eluted in the same region was similarly estimated since Peak 3 contained small amylopectin molecules as judged by chain length distribution analysis. The peak-top MW of Peaks 1, 2, and 3, in terms of ‘pullulan equivalent,’ was 3.0 × 108, 1.7 × 107, and 4.2 × 105, respectively; after multiplying by a factor of 2, they were 6.0 × 108, 3.4 × 107, and 8.4 × 105, respectively. These peak-top values can be regarded as Mw for peaks detected on a weight basis that are symmetrical in shape, and the results in this study (Fig. 1A) almost met these conditions. These values are much larger than the MW of the three amylopectin peaks obtained by Takeda et al.,13) and The MW of Peak 1, which represents the largest amylopectin, was smaller than that of the peak obtained by HPSEC-MALLS-RI (27 × 108).10) These inconsistencies are probably due to differences in separation procedure (e.g., different separation matrixes, different detection methods, weight or molar basis). The MW estimated for Peak 1 was in the order of 108, typical for MW of amylopectin, and MW for Peak 2 was roughly 1/18 that of Peak 1. Peak 2 accounted for a significant amount of amylopectin by weight (Fig. 1A). Although the exact reason why this peak was 1/18 the size of amylopectin is unknown, it might be the result of fragmentation of amylopectin molecules during preparation of starch from endosperm prior to the GPC experiment. Regardless, it is interesting how limited fragmentation can give fragments of a specific size. However, it should be noted that the species with MW in the order of 108 might be stable aggregates.33) The peak-top MW of Cluster Dextrin estimated by pullulan calibration in this study fell within the range defined by the peak-top weight-average DP of two peaks by HPSEC.28)
Based on the peak-top MW estimated above, the number of clusters was calculated as the peak-top MW/MW of a cluster unit. The DP of a cluster unit was estimated as 190 by Takeda et al.,13) and thus 190 multiplied by 162 gives the MW of a cluster unit. The calculated numbers of clusters for Peaks 1‒3 and Cluster Dextrin were 19,480, 1,120, 27, and 3.4, respectively. The value for Cluster Dextrin was in the same order as the estimation (< 1 to 2) by Takata et al.,28) indicating the validity of the estimate. In terms of size of α-glucan components, Peaks 1‒3 obtained in this study were considered to be different from those of L, M, and S fractions reported by Takeda et al.,13) which were detected only on a molar-basis (detection by fluorescence label introduced into reducing residues of α-glucans) but not on a weight-basis (detection by a refractive index).
In this study, whole starch was successfully separated by GPC with RI detection into three peaks containing amylopectin with different MW values. Iodine staining and chain length distribution analyses of each peak revealed that Peaks 1, 2, and 3 mainly contained high-MW amylopectin, low-MW amylopectin, and amylose with very-low-MW amylopectin and/or possible fragments of amylopectin, respectively. The chain length distribution pattern by capillary electrophoresis was almost identical for amylopectin-rich fractions, indicating that amylopectins separated by size differed largely in the number of cluster units per molecule, but possessed the same structure for one cluster unit. On the other hand, amylose-rich fractions were shown to have fewer cluster-connecting chains (B2－4 chains), attributed to the short-chain composition of branched amylose. Previous studies have shown that the rice flours obtained by different milling conditions have different physical properties as well as different patterns of gel filtration of whole starch, even though the debranched starch structure is not changed.16)34) This indicates that the differences in MW based on the number of clusters of amylopectin have a significant effect on the physical properties of starch, and that the GPC method for whole starch in this study provides important information indicating differences in physical properties of starch.
The MW of amylopectin in Peaks 1, 2, and 3 was estimated to be 6.0 × 108, 3.4 × 107, and 8.4 × 105, respectively. This is the first report on the three well-separated peaks of amylopectins with different size on a weight-based analysis by GPC. Information on the size distribution of amylopectin is expected to provide insight into other important aspects of starch, including biosynthetic mechanisms and physicochemical properties.
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
We thank Prof. Naoyoshi Inouchi, Fukuyarna University, for providing information on columns for separation of whole starch. We also thank Dr. Naoko Crofts, Dr. Satoko Miura, Ms. Yuko Hosaka, Ms. Naoko F. Oitome, Ms. Misato Abe, Ms. Yuko Nakaizumi, and Ms. Rika Takahashi (Akita Prefectural University) for technical support and growing rice plants. We thank Bioedit for English language editing. The Pseudomonas isoamylase used for debranching amylopectin was a kind gift from Hayashibara Co., Ltd.
