Accepted Manuscript
Effects of variety and growth location on the chain-length distribution of rice starches
Hongyan Li, Yingli Liu
PII: S0733-5210(18)30680-5 DOI: https://doi.org/10.1016/j.jcs.2018.11.009 Reference: YJCRS 2668
To appear in: Journal of Cereal Science
Received Date: 7 September 2018 Revised Date: 21 November 2018 Accepted Date: 21 November 2018
Please cite this article as: Li, H., Liu, Y., Effects of variety and growth location on the chain- length distribution of rice starches, Journal of Cereal Science (2018), doi: https://doi.org/10.1016/ j.jcs.2018.11.009.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. ACCEPTED MANUSCRIPT
1 Effects of variety and growth location on the chain-length distribution of rice
2 starches
3 Hongyan Li a,b *, Yingli Liu a, *
4 aBeijing Advanced Innovation Center for Food Nutrition and Human Health,
5 China-Canada Joint Lab of Food Nutrition and Health (Beijing), Beijing Engineering
6 and Technology Research Center of Food Additives, Beijing Technology and
7 Business University (BTBU), 11 Fucheng Road, Beijing 100048, China
8 bThe University of Queensland, Centre for Nutrition and Food Sciences, Queensland
9 Alliance for Agriculture and Food Innovation, Brisbane 4072, QLD, Australia.
10 11 *Corresponding author: MANUSCRIPT
12 Hongyan Li, E-mail: [email protected];
13 Yingli Liu, E-mail: [email protected];
ACCEPTED
1
ACCEPTED MANUSCRIPT
14 Abstract:
15 Five rice varieties with a wide range of amylose content harvested from three
16 different agro-climatic zones (Yanco, Mackay, and Darwin) of Australia are used to
17 explore effects of rice varieties and growth location on the fine structure of rice
18 starches. Number chain-length distributions (CLDs) of amylopectin branches are
19 characterized by fluorophore-assisted carbohydrate electrophoresis (FACE) and
20 parameterized by both empirical subdivision method and biosynthesis-based model.
21 This shows that amylopectin branches with degree of polymerization (DP) ~6-32
22 are not affected by both rice variety and growth location, but rice varieties from
23 Yanco tend to have smaller proportions of intermediate (DP~33-62) and long
24 (DP~63-100) amylopectin branches than those from Darwin and Mackay. The fitting MANUSCRIPT 25 results of starch biosynthesis model keep consistent with the above observations.
26 Weight CLDs of amylose are parameterized by size-exclusion chromatography (SEC),
27 showing that Yanco rices have significantly higher amylose content and higher
28 proportion of amylose branches with DP~500-5000. The significant interaction
29 between rice variety and growth location indicates that effects of grow location on
30 these fine structures are rice variety-specific. This study could provide implications of 31 environmentalACCEPTED effects on the cooking and eating quality of cooked rice for rice 32 breeders and industry.
33 Keywords: growth location, amylopectin, amylose, chain-length distribution
34
2
ACCEPTED MANUSCRIPT
35 1. Introduction
36 Rice ( Oryza sativa L.) is the second most widely produced cereal crop in the
37 world, which leads all cereal in supplying calories energy intake (Patindol,
38 Siebenmorgen and Wang, 2014). Rice is also adaptable and versatile, which is grown
39 in all continents (except Antarctica) and in more than 100 countries, between 40°S
40 and 53°N latitudes, from sea level to 3000 m in altitude, from dry land to under 1-to
41 2-m-deep water (De Datta, 1981).
42 Consumers consider the cooking and eating quality to be the most important
43 attribute. It is affected by a wide range of factors, such as amylose content,
44 postharvest processing, milling ratio, and cooking methods. Among these, starch 45 structure, especially that the CLDs have a crucialMANUSCRIPT role on rice texture (Li, Fitzgerald, 46 et al., 2017; Li et al., 2016). Starch, comprising ~90% of the dry weight of rice
47 grains, has two types of molecules: amylopectin (AP) and amylose (AM). AP
48 molecules are highly branched glucose polymers with a vast number of short branches
49 and large molecular weights ~10 7-8, whereas AM has a smaller molecular weight
50 (~10 5-6) with few long branches. There are several techniques for starch fine structural
51 analysis: FACE, high-performance anionic-exchange chromatography (HPAEC), and 52 SEC. FACEACCEPTED is the optimal method for determining amylopectin CLDs. It separates 53 molecules based on mass-to-charge ratio and provides baseline resolution between the
54 chains of different DPs, so it directly gives the number distribution of amylopectin. In
55 contrast to SEC, FACE does not suffer from problems of band-broadening, calibration,
3
ACCEPTED MANUSCRIPT
56 and inaccuracies in the Mark-Houwink relation, so FACE data are more accurate.
57 However, because of the inability to quantitatively detect chains above DP ~100,
58 FACE and HPAEC can only give information on amylopectin chains. SEC does not
59 have the same restriction and can therefore be used for the measurement of amylose
60 fine structure (Li and Gilbert, 2018).
61 Starch structure is varied between different rice varieties. Rice varieties are
62 generally classified by amylose content. Starch structure of different rices in terms of
63 amylose content has been extensively investigated (Syahariza et al., 2013). On the
64 other hand, environmental effects on rice structure are also significant. Aboubacar et
65 al. (2006) determined the amylopectin fine structure of long grain rice cultivars
66 planted in different locations in United States, and found that higher growing MANUSCRIPT 67 temperature results in more amylopectin branches with DP >10 to form consistent
68 crystallites, contributing to higher gelatinization temperatures and enthalpies. Two
69 medium grain rice cultivars from Arkansas and California were determined by
70 Cameron et al. (2007), rices from California had significantly higher amylose content
71 and smaller proportion of chains with DP 13-24. Sar et al. (2014) investigated three
72 rice cultivars with different amylose content from three different agro-climatic zones 73 of Cambodia,ACCEPTED found that starch fine structures are significantly different between 74 cultivars, but not significantly differ between different locations.
75 Rice is Australia’s third largest cereal grain export, and the ninth largest
76 agricultural export. Rice industry generates around $AUD 800 million per year, and
4
ACCEPTED MANUSCRIPT
77 Australia exports rice to 60 major international destinations including the Middle East,
78 the Pacific, North America, and Asia (Fransisca et al., 2015). In Australia,
79 commercial rice production primarily grow in southern New South Wales in areas
80 adjacent to the Murrumbidgee and Murray Rivers, with a small number of farms in
81 Northern Victoria (Sivapalan et al., 2007). In this study, five rice varieties with a wide
82 range of amylose content are harvested from three different agro-climates of Australia
83 (Darwin of Northern Territory: Tropical zone; Mackay of Queensland: Sub-tropical
84 zone; Yanco of New South Wales: Temperate zone); the fine structure of amylose and
85 amylopectin are measured by SEC and FACE, respectively; a biosynthesis model is
86 used to fit number CLDs of amylopectin to supply insights from the view of starch
87 biosynthesis; and effects of growth location on the starch fine structure are also 88 discussed. MANUSCRIPT
89 2. Materials and methods
90 2.1 Materials
91 Five rice varieties were selected for this study: Hom Mali Niaw (HMN), Kyeema
92 (KM), Doongara (DG), Pandan Wangi (PW), IR 64. All 5 rice varieties were planted
93 in 3 different locations of Australia in different seasons: the winter season of Darwin 94 (12.4°S) in NorthernACCEPTED Territory, the summer season of Mackay (21.1°S) in Queensland, 95 and the summer season of Yanco (34.6°S) in New South Wales. Table S1
96 summarizes the major weather conditions during rice growth in these three locations,
97 that is, the mean maximum temperature during the day, the mean minimum
5
ACCEPTED MANUSCRIPT
98 temperature during the night, mean daily sunshine, and mean rainfall. After harvesting,
99 all samples were air-dried and stored at room temperature for 1 month to maximize
100 the milling quality. The samples were milled to white rice with a lab-scale mill
101 (Satake Corp, Japan) and then stored in the fridge at 4 °C.
102 Protease from Streptomyces griseus (type XIV), and LiBr (ReagentPlus) were
103 purchased from Sigma-Aldrich Pty. Ltd. (Shanghai, China). Isoamylase (from
104 Pseudomonas sp. ) and D-glucose (glucose oxidase/peroxidase; GODOP) assay kit
105 were purchased from Megazyme International, Ltd. (Wicklow, Ireland). A series of
106 pullulan standards with peak molecular weights ranging from 342 to 2.35 × 10 6 were
107 purchased from Polymer Standards Service (PSS, Mainz, Germany).
108 8-Aminopyrene-1,3,6,-trisulfonate (APTS) was purchased from Beckman Coulter MANUSCRIPT 109 (Brea, USA). Dimethyl sulfoxide (DMSO, GR grade for analysis) was from Merck
110 Co. Inc. (Kilsyth, Australia). All other chemicals were reagent-grade and used as
111 received.
112 2.2 Extraction, Dissolution, and Debranching of Starch Molecules for
113 Structural Analysis
114 Starch isolation is conducted following the method described by Syahariza et al.
115 (2010). RiceACCEPTED grains were ground into flour with a cryogenic mill (Freezer/Mill 6850;
116 SPEX, Metuchen, NJ) in a liquid nitrogen bath as the cryogenic medium. All samples
117 were extracted and dissolved in DMSO solution with 0.5% (w/w) LiBr (DMSO/LiBr).
118 The milled flour was treated with protease and sodium bisulfite solution, followed by 6
ACCEPTED MANUSCRIPT
119 centrifugation to remove protein. The treated flour was dissolved in DMSO/LiBr
120 solution and precipitated using ethanol (each treatment followed by centrifugation) to
121 remove insoluble and soluble non-starch components, respectively. The precipitate
122 (extracted starch) was debranched using isoamylase for 3 h, heated in a water bath at
123 80 °C for 2 h, freeze-dried, and then dissolved in DMSO/LiBr solution for subsequent
124 analysis by SEC and FACE (Li, Wen, et al., 2017).
125 2.3 SEC
126 The SEC weight distribution of debranched starch was analyzed in duplicate
127 using an Agilent 1100 series SEC system (Agilent Technologies, Waldbronn,
128 Germany) equipped with a refractive index detector (RID, Shimadzu RID-10A, 129 Shimadzu Corp., Kyoto, Japan), as describedMANUSCRIPT elsewhere (Syahariza et al., 2013). A 130 series of SEC columns (GRAM precolumn and GRAM 100 and GRAM 1000
131 columns, PSS) placed in an oven at 80 °C were used to separate the debranched starch
132 molecules. DMSO/LiBr solution was used as the mobile phase with a flow rate of 0.6
133 mL/min. Pullulan standards were used for calibration to convert from the SEC elution
134 volume to the hydrodynamic volume Vh or the corresponding radius Rh using the
135 Mark-Houwink equation. The Mark-Houwink parameters K and α of pullulan in 136 DMSO/LiBrACCEPTED solution at 80 °C are 2.424 × 10 -4 dL g -1 and 0.68, respectively . The 137 SEC weight distribution, w(log X), obtained from DRI signal was plotted against X
138 (DP), related to M for starch by M=162.2( X-1)+18.0 (where 162.2 is the molecular
139 weight of the anhydroglucose monomeric unit and 18.0 that of the additional water in
7
ACCEPTED MANUSCRIPT
140 the end groups). For linear polymer (such as debranched starch), the number
141 distribution, N(X), is related to the corresponding weight distribution by: (log ) =
142 ( ). The K and α for linear starch chains in the eluent of DMSO/LiBr at 80 °C
143 are 1.5 × 10 -4 dL g -1 and 0.743, respectively (Wang et al., 2014).
144 2.4 FACE
145 The freeze-dried linear branches (0.2 mg) were labelled using APTS by the
146 method described by Wu et al. (2014). The size distribution of the labelled linear
147 glucans was analyzed with FACE to give the chain-length distribution (CLD),
148 denoted Nde (X) (the subscript “de” is used to indicate that the linear glucans were
149 obtained by debranching starch; X = DP). Separation of the labelled linear glucans 150 was performed on a PA-800 Plus SystemMANUSCRIPT and monitored with a solid-state 151 laser-induced fluorescence (LIF) detector with an argon-ion laser as the excitation
152 source (Beckman-Coulter, Brea, CA, USA). The capillary used was a 50-µm diameter
153 N-CHO coated capillary (included in the Carbohydrate Labelling and Analysis Kit).
154 Carbohydrate separation buffer, also included in the kit, was used as the separating
155 medium. The effective separation length of the capillary was 40 cm. The sample was
156 introduced into the capillary by pressure injection for 3 s at 0.5 psi (3.4 kPa above 157 atmospheric).ACCEPTED Separation of the labelled linear glucans was achieved using an applied 158 voltage of 30 kV (current ∼14 mA) at 25°C. 90 min of total separation time was used
159 to separate the first ∼160 peaks (Wu et al., 2014). The areas of the peaks give the
160 relative amount of glucans with different mass directly (the DPs of glucans in
8
ACCEPTED MANUSCRIPT
161 adjacent peaks differ by 1). Sample storage temperature in the PA-800 Plus System
162 was at 18°C. Diluted labelled glucans may require further dilution if the raw
163 electropherogram appears to be cut off as the detector response reaches a maximum,
164 which can lead to an underestimate of the overloaded signals.
165 2.5 Fitting Amylopectin Number CLDs with a Biosynthesis Model
166 The number CLDs of amylopectin which is characterized using FACE were
167 fitted with an amylopectin biosynthesis based model to obtain information on the
168 starch biosynthetic enzymes. The underlying theory is described elsewhere (Wu et al.,
169 2014). Fitting the amylopectin number CLDs with this model gives a series of
170 parameters that describe the relative activities of enzymes involved in amylopectin 171 biosynthesis. As explained by Wu et al. (2014),MANUSCRIPT the number CLD of amylopectin 172 branches with X≤100 is classified into 3 groups, which are 6< X≤32 (the single-lamella
173 chains), 32< X≤62 (trans-lamella chains), and 62< X≤100 (long trans-lamella chains).
174 The substrate-competing model is used to fit each region, and it is assumed that the
175 number CLDs of each region is governed by two enzymes sets, which denoted set i,
176 ii , iii , iv , v, and vi ; the most informative parameter is β, denoted β(i) , β(ii) , β(iii) , β(iv) , β(v) ,
177 and β(vi) , which describes the activity ratio of starch branching enzymes (SBE) to 178 starch synthaseACCEPTED (SS) in each region; the relative contributions of each region is also of
179 interest, which is termed h(i) , h(ii) , and h(iii) .
180 2.6 Statistical Analysis
9
ACCEPTED MANUSCRIPT
181 For each structural measurement, duplicated analysis was performed for each
182 sample. All of the data analyses were carried out using SPSS V. 16.0 software (SPSS
183 Inc., Chicago, IL). Analysis of variance (ANOVA) was performed to determine rice
184 variety and growth location effects on starch fine structural parameters by using the
185 general linear model procedure. Mean squares were used to calculate F statistics for
186 tests of significance. The total variation was the sums of all mean squares of the main
187 and interaction effects.
188 3. Results and Discussion
189 While FACE gives highly accurate CLDs, the technique cannot be used for large
190 chains (the present data go up to X ~100), and thus, amylopectin chains are analyzed 191 using FACE and amylose chains are analyzedMANUSCRIPT by SEC. As described elsewhere 192 (Vilaplana et al., 2012), the components with DP ≤100 are amylopectin chains, while
193 that with DP>100 are amylose chains. The nomination of amylose region could not
194 provide a definite separation between amylose and amylopectin, since there is no
195 rigorous definition of just how a molecule can be classified unambiguously as
196 amylose or amylopectin. Even though, this definition of amylose region is still
197 effective, which correlate well with iodine colorimetry method (Vilaplana et al., 198 2012). ACCEPTED
199 3.1 Effects of rice variety and growth location on the fine structure of
200 amylopectin
10
ACCEPTED MANUSCRIPT
201 Figure 1 shows the typical number CLDs of debranched amylopectin from
202 FACE. The number CLDs are plotted logarithmically, for reasons explained
203 elsewhere (Wu and Gilbert, 2010). All CLD number distributions are normalized to
204 share the same global maximum to compare between different rice varieties and
205 eliminate the effects of different sample concentrations. All starch samples display
206 typical number CLDs of amylopectin, with four peaks and/or shoulders observed (Wu
207 and Gilbert, 2010). The first peak is the global maximum at DP ~12, followed by a
208 small bump at approximately DP ~21. These two features, covering DP ~6-32, are
209 short amylopectin chains confined to one crystalline lamella (single-lamella). A local
210 minimum is observed at DP ~33, separating single- and trans-lamella branches, the
211 latter spanning two or more crystalline lamellae. The population with maximum at DP 212 ~44 are trans-lamellar branches that span throughMANUSCRIPT one crystalline lamella and the 213 adjacent amorphous lamella, while the other bump with a local maximum at DP ~75
214 are long amylopectin branches spanning at least two adjacent crystalline lamellae and
215 the amorphous lamellae in between (Wu et al., 2013). To quantify the differences
216 between samples, the maximum heights of peaks and/or shoulders at X=12, 21, 44,
217 and 75 in the number CLDs are denoted as hAP1 , hAP2 , hAP3 , and hAP4 (Table 1), the
218 number CLDs of amylopectin branches are further subdivided into three categories:
219 X=6-32, 33-62,ACCEPTED 63-100, which represents the region of short, intermediate, and long
220 amylopectin branches, respectively. The proportion (relative to the whole amylopectin
221 region ~DP 6-100) of each AP category for all rice samples is calculated and
222 summarized in Table 1. 11
ACCEPTED MANUSCRIPT
223 As shown in Figure 1 and Table 1, 6< X≤32, hAP1 , and hAP2 are parameters
224 representing the content of short AP branches. As presented in Table 1 and S2 ,
225 variations of parameters, 6< X≤32, hAP1 , and hAP2 , for rice variety are not significant
226 (p>0.05). It suggested that, even though all rice samples have different amylose content,
227 the similar proportion of short AP branches which is mainly located in the
228 single-lamella region is observed. On the other hand, parameters of 6< X≤32, hAP1 , and
229 hAP2 for locations are also not significant ( p>0.05). This indicates that, growth location
230 does not affect the proportion of short AP branches as well. Furthermore, no significant
231 variation for the interaction between rice variety and location is found ( Table S2 ). The
232 negative result that no statistically significant effect of both rice varieties and growth
233 locations on the content of short amylopectin branches is observed does not mean that 234 these short amylopectin branches are not geneticall MANUSCRIPTy determined. A previous report 235 showing no significant difference for the proportion of short amylopectin chains
236 (X≤32) between 12 different rice varieties supports the above result (Li et al., 2016).
237 Similarly, both 32 238 intermediate AP branches (Figure 1 and Table 1). 32 239 significantly different ( p<0.05) while hAP3 does not significantly differ ( p>0.05). This 240 indicates thatACCEPTED the content of intermediate AP branches is affected by rice variety, but 241 all rice samples have the similar content at DP~44. As shown in Table S2 , both 242 32 243 p<0.01 for hAP3 ). Besides, both 32 12 ACCEPTED MANUSCRIPT 244 similar trend that the corresponding values of Yanco rice are significantly smaller 245 than those of Darwin and Mackay rices, indicating Yanco rices have a smaller 246 proportion of intermediate AP branches. The interactions between rice variety and 247 location for both parameters are also significantly different ( p<0.01; Table S2 ), 248 revealing that the responses of intermediate AP branches to growth locations are rice 249 variety-specific. 250 62 251 branches (Table 1). 62 252 but hAP4 significantly differs between samples ( p<0.05) ( Table 1, S2 ), reflecting that 253 the total content of long amylopectin branches is not affected by the variation of variety, 254 but all rice varieties have different contents at DP~75. However, variations of growth MANUSCRIPT 255 location for both 62 256 S2 ), suggesting that Yanco rices have less long AP branches than Darwin and Mackay 257 rices. Furthermore, significant variations for the interaction between rice variety and 258 growth locations ( Table S2 ) reveal that the effects of growth location on the proportion 259 of long AP branches are rice variety-specific ( Figure 1 ). 260 As shown in Table S1, among these environmental conditions, the mean 261 maximum temperaturesACCEPTED in three locations are comparable, Australian irrigation system 262 can offset the water shortage due to the different rainfall among different climate 263 zones (Humphreys et al., 2006), but the minimum temperature in Yanco is 264 significantly lower than that in Darwin and Mackay. Thus, the lower night-time air 13 ACCEPTED MANUSCRIPT 265 temperature might be the main environmental factor contributing to the significant 266 structural features of Yanco rice, i.e. a smaller proportion of intermediate and long AP 267 branches. This is supported by other findings elsewhere, Suzuki et al. (2003) reported 268 that the relative amounts of amylopectin long chains (DP>25) increase in all starches 269 when maturation occurred at 28 °C instead of 21 °C. Counce et al. (2005) found that 270 amylopectin chain lengths with DP 13-24 are increased by the high temperature 271 treatment. Wei et al. (2012) observed that elevated temperature (32 versus 22 °C) 272 results in a decrease of DP~5-9 and DP~15-22 amylopectin chains and an increase in 273 DP~10-13 and DP>42 chains. Patindol, Siebenmorgen, Wang, et al. (2014) presented 274 that the elevated night-time air temperature causes an increase in the percentage of long 275 amylopectin branches (DP>19) by an average of 1.3%. MANUSCRIPT 276 3.2 Understanding effects of growth location on amylopectin biosynthesis by 277 fitting a biosynthesis model 278 AP number CLDs of all rice samples are well fitted with the biosynthesis model, 279 with all the features reproduced in the fitted number CLDs ( Figure 2 ). The whole 280 region of AP branches is classified into 3 groups, and each region is governed by two 281 enzyme sets, which means that enzyme sets i and ii contribute to the region of 6< X≤32, 282 enzyme setsACCEPTED iii and iv are responsible for the region of 32< X≤62, and enzyme sets v 283 and vi correspond to the region of 62< X≤100. In the substrate-competing model, 284 enzyme sets in each region are independent, and thus the parameters of h(i) , h(ii) , and 14 ACCEPTED MANUSCRIPT 285 h(iii) are defined to specify the relative abundance of each CLDs region generated from 286 all kinetics (Wu et al., 2013). 287 Table 2 and S3 show that, except h(i) and h(ii) , all other parameters for rice variety 288 are significantly different. Parameters of h(i) , β(i) , and β(ii) for different locations are not 289 significantly different, which is consistent with the results for parameters of 6< X≤32, 290 hAP1 , and hAP2 in Table 1. Correspondingly, h(ii) , β(iii) , and β(iv) for different locations are 291 consistent with parameters of 6< X≤32 and hAP3 in Table 1, showing that Yanco rices 292 have a smaller value of h(ii) and higher value of β(iii) . Similarly, h(iii) , β(v) , and β(vi) of 293 Yanco rice are also significantly different from those of Darwin and Mackay rices. 294 Parameters obtained by fitting amylopectin number CLDs with the biosynthesis 295 model provide insights into the starch synthesisMANUSCRIPT mechanism (Wu and Gilbert, 2010). 296 As reviewed by Jeon et al. (2010), several enzyme isoforms are involved in 297 amylopectin branching, elongation, or both in cereal endosperms, including SS, SBE, 298 and DBE. SS isoforms act in the elongation of linear glucan chain, SBE isoforms 299 introduce new branches on starch molecules, while DBE isoforms trim excess 300 branches during amylopectin biosynthesis. Inside the single-lamella region (with DP 301 6< X≤32), the comparison of β values shows that there are no significant differences in 302 the activity ratioACCEPTED of SBE to SS between different rice planting locations, reflecting that 303 growth locations do not affect the activities of enzyme sets i and ii , leading to the 304 similar CLDs in the single-lamella region of amylopectin. In the trans-lamella region 305 (32< X≤62), the significantly higher value of β(iii) for Yanco rice suggests a higher 15 ACCEPTED MANUSCRIPT 306 relative activity of SBE and/or a lower relative activity of SS, resulting in a smaller 307 proportion of intermediate AP chains. Similarly, in the long trans-lamella region 308 (62< X≤100), the high values of β(v) and β(vi) of Yanco rice also indicate higher relative 309 activity of SBE and/or lower relative activity of SS, which are in agreement with the 310 small proportion of 62< X≤100 and hAP4 for Yanco rices in Table 1. 311 3.3 Effects of rice variety and growth location on the fine structure of 312 amylose 313 Figure 3 shows SEC weight CLDs while Figure 4 presents SEC number CLDs. 314 Both SEC weight and number CLDs are normalized to yield the same height of the 315 highest amylopectin branch peak and eliminate the effects of different sample 316 concentrations (Syahariza et al., 2013). In FigureMANUSCRIPT 3 , the weight CLDs from all rice 317 samples show two large peaks of amylopectin branches and one smaller peak of 318 amylose branches. As mentioned above, DP~100 is used to divide the weight CLDs 319 into two parts: amylopectin branch region (DP ≤100) and amylose branch region 320 (DP>100). It is noteworthy here, there are always two peaks visible in amylose region, 321 which is reported elsewhere (Ward et al., 2006). However, in this study ( Figure 3) , 322 there are no clearly visible two peaks in amylose region. This may be attributed to the 323 band broadeningACCEPTED effect by SEC itself which makes two peaks overlapped. Thus, in this 324 study the height of these two peaks are not used as the quantitative parameters. In 325 Figure 3 , we can obviously see that amylose branches between DP 500 and 5000 of 326 Yanco rice are higher than others, especially for KM, PW, and IR64. Thus, the amylose 16 ACCEPTED MANUSCRIPT 327 region is further subdivided into 3 categories: X=100-500, 500-5000, 5000-20000, 328 which represent the region of short, intermediate, and long AM chains, respectively. 329 Amylose content is calculated from SEC weight CLDs as the ratio of the area under 330 the curve (AUC) of amylose branches to the AUC of overall amylopectin and 331 amylose branches (Syahariza et al., 2013). Similarly, the proportion (relative to the 332 whole starch region ~DP 6-20000) of each AM category of all samples is also 333 calculated and summarized in Table 1. 334 As presented in Table 1 and S2 , all parameters of AM content, 100< X≤500, 335 500< X≤5000, and 5000< X≤20000 for both rice variety and locations are significantly 336 different. Rice varieties selected for this study are obviously different in terms of 337 amylose content: one waxy rice (HMN, 3.87%), two low-amylose rices (KM, 19.59%; MANUSCRIPT 338 PW 20.31%), one intermediate-amylose rice (IR64, 25.29%), and one high-amylose 339 rice (DG 27.17%) ( Table S1), so it is not surprising that amylose content is 340 significantly affected by genotype variations. 341 As shown in Table 1, amylose content and the proportion of AM chains with DP 342 ~500-5000 of Yanco rice are higher than Darwin and Mackay rices ( p<0.05). Whereas, 343 Darwin rices have significantly higher proportion of short branches (DP ~100-500) 344 and smaller ACCEPTEDpercentage of long branches (DP ~5000-20000) than those of Mackay and 345 Yanco rices. Furthermore, parameters of AM content, 100< X≤500, and 500< X≤5000 346 also show significant interactions between rice variety and planting locations, even 17 ACCEPTED MANUSCRIPT 347 though rice variety × location variances only explain small fraction of total variation 348 for these parameters. 349 It has been reported that amylose content is affected by air temperature (day or 350 night). In greenhouse experiments, increased amylose to amylopectin ratio is observed 351 in kernels of rice plants with a cooler temperature during growing (Hirano and Sano, 352 1998). In field experiments, Aboubacar et al. (2006) reported that higher amylose 353 content are found in Missouri due to its lower air temperature. Patindol, 354 Siebenmorgen, Wang, et al. (2014) suggested the strong correlation between paste 355 viscosity and amylose content indicates that elevated night-time air temperature may 356 cause the lower amylose content. Since amylose is synthesized by GBSS. Of the two 357 currently identified GBSS isoforms, GBSSII functions in non-storage plant tissues, MANUSCRIPT 358 whereas the role of GBSSI is mostly confined to storage tissues such as the seed 359 endosperm. In rice endosperms, GBSSI is encoded by the Waxy (wx ) locus. Ahmed et 360 al. (2008) reported that the GBSS activity in endosperms developing at 12 °C is 361 significantly higher than the corresponding activity in endosperms developing at 362 22 °C. Similar result is also reported by Jiang et al. (2003). Therefore, based on the 363 above literature review and the local weather reports in 3 locations (as shown in Table 364 S1), the lowerACCEPTED night-time temperature during the growth of Yanco rices might be an 365 important factor contributing to the significantly higher amylose content and 366 proportion of amylose branches with DP~500-5000. 367 4. Conclusions 18 ACCEPTED MANUSCRIPT 368 This study reveals the effects of growth location on the first level structure, 369 CLDs of amylopectin and amylose. Growing location significantly affects 370 intermediate and long amylopectin branches, amylose content, and intermediate 371 amylose branches. The interactions between rice variety and location on these 372 structural features are also significant, suggesting the environmental effects on these 373 fine structures are rice variety-specific, especially for rices growing in Darwin and 374 Mackay. Fitting amylopectin number CLDs to a biosynthesis model gives new insight 375 into the enzymatic processes involved in amylopectin biosynthesis. The fitted 376 parameters support the structural parameters affected by rice variety and growth 377 location. Hence, the baseline information derived from this study will be useful in 378 explaining variability trends in starch properties, e.g. the eating quality of cooked rice 379 from different planting areas. MANUSCRIPT 380 Acknowledgement 381 This work was supported by National Key R&D Program of China 382 (2018YFD0400600, 2018YFD0400103), Beijing Excellent Talents Funding for Youth 383 Scientist Innovation Team (2016000026833TD01), High-level Teachers in Beijing 384 Municipal Universities (IDHT20180506), and Foundation for Young Scientist of 385 Beijing TechnologyACCEPTED & Business University (Grant PXM2018_014213_000033). 386 Rice samples are obtained during Hongyan Li’s Doctoral research, he also gratefully 387 acknowledges all sorts of helps from his PhD supervisors Robert Gilbert, Melissa 388 Fitzgerald, Sangeeta Prakash, and Timothy Nicholson. 19 ACCEPTED MANUSCRIPT 389 References: 390 Aboubacar, A., Moldenhauer, K.A., McClung, A.M., Beighley, D.H., Hamaker, B.R., 2006. 391 Effect of growth location in the United States on amylose content, amylopectin fine structure, 392 and thermal properties of starches of long grain rice cultivars. Cereal Chemistry 83, 93-98. 393 Ahmed, N., Maekawa, M., Tetlow, I.J., 2008. Effects of low temperature on grain filling, 394 amylose content, and activity of starch biosynthesis enzymes in endosperm of basmati rice. 395 Australian Journal of Agricultural Research 59, 599-604. 396 Cameron, D.K., Wang, Y.J., Moldenhauer, K.A., 2007. Comparison of starch 397 physicochemical properties from medium-grain rice cultivars grown in California and 398 Arkansas. Starch-Stärke 59, 600-608. 399 Counce, P., Bryant, R., Bergman, C., Bautista, R., Wang, Y.-J., Siebenmorgen, T., 400 Moldenhauer, K., Meullenet, J.-F., 2005. Rice milling quality, grain dimensions, and starch 401 branching as affected by high night temperatures. Cereal Chemistry 82, 645-648. 402 De Datta, S.K., 1981. Principles and practices of rice production. John Wiley and Sons, New 403 York. 404 Fransisca, Y., Small, D.M., Morrison, P.D., Spencer, M.J., Ball, A.S., Jones, O.A., 2015. 405 Assessment of arsenic in Australian grown and imported rice varieties on sale in Australia 406 and potential links with irrigation practises and soil geochemistry. Chemosphere 138, 407 1008-1013. 408 Hirano, H.-Y., Sano, Y., 1998. Enhancement of Wx gene expression and the accumulation of 409 amylose in response to cool temperatures during seed development in rice. Plant and Cell 410 Physiology 39, 807-812. MANUSCRIPT 411 Humphreys, E., Lewin, L., Khan, S., Beecher, H., Lacy, J., Thompson, J., Batten, G., Brown, 412 A., Russell, C., Christen, E., 2006. Integration of approaches to increasing water use 413 efficiency in rice-based systems in southeast Australia. Field Crops Research 97, 19-33. 414 Jeon, J.-S., Ryoo, N., Hahn, T.-R., Walia, H., Nakamura, Y., 2010. Starch biosynthesis in 415 cereal endosperm. Plant Physiology and Biochemistry 48, 383-392. 416 Jiang, H., Dian, W., Wu, P., 2003. Effect of high temperature on fine structure of amylopectin 417 in rice endosperm by reducing the activity of the starch branching enzyme. Phytochemistry 63, 418 53-59. 419 Li, H., Fitzgerald, M.A., Prakash, S., Nicholson, T.M., Gilbert, R.G., 2017. The molecular 420 structural features controlling stickiness in cooked rice, a major palatability determinant. 421 Scientific Reports 7, 43713. 422 Li, H., Gilbert, R.G., 2018. Starch molecular structure: The basis for an improved 423 understanding of cooked rice texture. Carbohydrate Polymers 195, 9-17. 424 Li, H., Prakash,ACCEPTED S., Nicholson, T.M., Fitzgerald, M.A., Gilbert, R.G., 2016. The importance of 425 amylose and amylopectin fine structure for textural properties of cooked rice grains. Food 426 Chemistry 196, 702-711. 427 Li, H., Wen, Y., Wang, J., Sun, B., 2017. The molecular structures of leached starch during 428 rice cooking are controlled by thermodynamic effects, rather than kinetic effects. Food 429 Hydrocolloids 73, 295-299. 20 ACCEPTED MANUSCRIPT 430 Patindol, J.A., Siebenmorgen, T.J., Wang, Y.-J., Lanning, S.B., Counce, P.A., 2014. Impact 431 of elevated nighttime air temperatures during kernel development on starch properties of 432 field-grown rice. Cereal Chemistry 91, 350-357. 433 Patindol, J.A., Siebenmorgen, T.J., Wang, Y.-J., 2014. Impact of environmental factors on 434 rice starch structure: A review. Starch-Stärke 66, 1-13. 435 Sar, S., Tizzotti, M.J., Hasjim, J., Gilbert, R.G., 2014. Effects of rice variety and growth 436 location in Cambodia on grain composition and starch structure. Rice Science 21, 47-58. 437 Sivapalan, S., Batten, G., Goonetilleke, A., Kokot, S., 2007. Yield performance and 438 adaptation of some Australian-grown rice varieties through multivariate analysis. Australian 439 Journal of Agricultural Research 58, 874-883. 440 Suzuki, Y., Sano, Y., Ishikawa, T., Sasaki, T., Matsukura, U., Hirano, H.-Y., 2003. Starch 441 characteristics of the rice mutant du2-2 Taichung 65 highly affected by environmental 442 temperatures during seed development. Cereal Chemistry 80, 184-187. 443 Syahariza, Z.A., Li, E., Hasjim, J., 2010. Extraction and dissolution of starch from rice and 444 sorghum grains for accurate structural analysis. Carbohydrate Polymers 82, 14-20. 445 Syahariza, Z.A., Sar, S., Hasjim, J., Tizzotti, M.J., Gilbert, R.G., 2013. The importance of 446 amylose and amylopectin fine structures for starch digestibility in cooked rice grains. Food 447 Chemistry 136, 742-749. 448 Vilaplana, F., Hasjim, J., Gilbert, R.G., 2012. Amylose content in starches: Toward optimal 449 definition and validating experimental methods. Carbohydrate Polymers 88, 103-111. 450 Wang, K., Hasjim, J., Wu, A.C., Henry, R.J., Gilbert, R.G., 2014. Variation in amylose fine 451 structure of starches from different botanical sources. Journal of Agricultural and Food 452 Chemistry 62, 4443-4453. 453 Ward, R.M., Gao, Q., de Bruyn, H., Gilbert, R.G., MANUSCRIPTFitzgerald, M.A., 2006. Improved methods 454 for the structural analysis of the amylose-rich fraction from rice flour. Biomacromolecules 7, 455 866-876. 456 Wei, K., Yang, W., Jilani, G., Zhou, W., Liu, G., Chaudhry, A., Cao, Z., Cheng, F., 2012. 457 Eeffect of high temperature on the enzymatic activities and transcriptional expression of 458 starch debranching enzyme (DBE) mutiple isoforms in developing rice endosperms. Journal 459 of Animal and Plant Sciences 22, 97-107. 460 Wu, A.C., Gilbert, R.G., 2010. Molecular weight distributions of starch branches reveal 461 genetic constraints on biosynthesis. Biomacromolecules 11, 3539-3547. 462 Wu, A.C., Li, E., Gilbert, R.G., 2014. Exploring extraction/dissolution procedures for analysis 463 of starch chain-length distributions. Carbohydrate Polymers 114, 36-42. 464 Wu, A.C., Morell, M.K., Gilbert, R.G., 2013. A parameterized model of amylopectin 465 synthesis provides key insights into the synthesis of granular starch. PloS One 8, e65768. ACCEPTED 466 467 21 ACCEPTED MANUSCRIPT Table 1. The structural parameters for amylopectin from number CLDs of FACE and amylose from weight CLDs of SEC. Rice Amylopectin variety 6 MANUSCRIPT Amylose Rice variety Am content 100 Table 2. Parameters obtained by fitting amylopectin number CLD with a biosynthesis model for different rice varieties at different locations. Rice Amylopectin variety h(i) h(ii) h(iii) β(i) β(ii) β(iii) β(iv) β(v) β(vi) HMN 0.99 ± 0.02 a 0.10 ± 0.01 a 0.006 ± 0.001 a 0.090 ± 0.004 a 0.035 ± 0.007 c 0.057 ± 0.003 a 0.029 ± 0.002 b,c 0.071 ± 0.015 c 0.023 ± 0.005 a,b KM 0.97 ± 0.02 a 0.10 ± 0.01 a 0.008 ± 0.001 b 0.123 ± 0.013 c 0.018 ± 0.010 a,b 0.058 ± 0.004 a,b 0.031 ± 0.002 c 0.057 ± 0.007 b 0.013 ± 0.005 a PW 0.98 ± 0.01 a 0.10 ± 0.01 a 0.007 ± 0.001 b 0.099 ± 0.009 a 0.026 ± 0.008 b 0.056 ± 0.006 a 0.029 ± 0.003 b,c 0.054 ± 0.008 b 0.035 ± 0.046 b IR64 0.99 ± 0.02 a 0.10 ± 0.02 a 0.007 ± 0.001 a,b 0.114 ± 0.011 b 0.015 ± 0.010 a 0.059 ± 0.004 a,b 0.016 ± 0.005 a 0.039 ± 0.004 a 0.019 ± 0.007 a,b DG 0.97 ± 0.02 a 0.10 ± 0.01 a 0.009 ± 0.001 c 0.118 ± 0.007 b,c 0.020 ± 0.005 a,b 0.061 ± 0.004 b 0.027 ± 0.002 b 0.053 ± 0.013 b 0.017 ± 0.02 a,b Location Darwin 0.98 ± 0.03 a 0.10 ± 0.01 b 0.008 ± 0.002 b 0.101 ± 0.011 a 0.024 ± 0.007 a 0.056 ± 0.004 a 0.027 ± 0.007 a 0.049 ± 0.014 a 0.017 ± 0.008 a Mackay 0.98 ± 0.01 a 0.10 ± 0.01 b 0.008 ± 0.001 b 0.108 ± 0.015 a 0.028 ± 0.009 a 0.057 ± 0.003 a 0.027 ± 0.008 a 0.052 ± 0.010 a 0.013 ± 0.009 a Yanco 0.99 ± 0.01 a 0.09 ± 0.01 a 0.007 ± 0.001 a 0.107 ± 0.016 a 0.026 ± 0.012MANUSCRIPT a 0.062 ± 0.003 b 0.026 ± 0.006 a 0.063 ± 0.015 b 0.035 ± 0.035 b ACCEPTED ACCEPTED MANUSCRIPT MANUSCRIPT ACCEPTED Figure 1. Number CLDs (arbitrary units) of all rice samples in the amylopectin range (X ≤100) from FACE analysis. ACCEPTED MANUSCRIPT Figure 2. Fitting a representative number CLD ofMANUSCRIPT amylopectin branches. Fitting plot ①: the short branches synthesized by enzyme sets (i) and (ii) covering the single-lamella region (6 MANUSCRIPT Figure 3 . SEC weight CLDs of debranched starch for HMN, KM, PW, IR64, and DG planted in Darwin, Mackay, and Yanco, respectively. ACCEPTED ACCEPTED MANUSCRIPT MANUSCRIPT Figure 4. SEC number CLDs of debranched starch for HMN, KM, PW, IR64, and DG planted in Darwin, Mackay, and Yanco, respectively. ACCEPTED ACCEPTED MANUSCRIPT ‹ Effects of variety and growth location on starch molecular structure are analyzed. ‹ The content of short amylopectin chains are not affected by variety and locations. ‹ Yanco rice tends to have more long-AP chains but less intermediate-AP chains. ‹ Yanco rice has higher amylose content, especially more branches with DP 500-5000. ‹ Environmental effects on starch molecular structure are rice variety-specific. MANUSCRIPT ACCEPTED