Elemental Content in Brown Rice by Inductively Coupled Plasma Atomic Emission Spectroscopy Reveals the Evolution of Asian Cultivated Rice

Mean and parameters of macroelement contents in brown rice

In general, mean contents (mg/kg) of 863 accessions in Yunnan brown rice has rather high concentrations of nutritional elements such as P (3 797.72 ± 490.94), K (2 631.70 ± 403.27), Ca (155.90 ± 58.04), Mg (1 512.86 ± 211.10), Fe (32.56 ± 25.37), Zn (32.49 ± 13.56) and Cu (13.94 ± 11.39) (Tables 1, 2).

Mean and parameters of macroelement contents in brown rice from various classification traits in Yunnan rice are given in Table 1. Overall, the most significant differences occurred (P < 0.01), using one-way anova via SPSS software, for four macroelements (P, K, Ca, and Mg) in brown rice between rice landrace and improved cultivar, as well as improved cultivar of indica vs. japonica. There was no significant difference between indica and japonica among core collections of rice landraces. There were significant differences (P < 0.05) between different types among the landraces for some elements: lowland vs. upland (P and Mg), glutinous vs. non-glutinous (K), early-mid vs. late (Ca), white vs. red (K), white vs. purple (K), non-nuda vs. nuda (P and Mg). The largest difference (mg/kg) of macroelements between landrace and improved cultivar were K (746.07) > P (432.75) > Mg (304.09) > Ca (26.02) and P (300.01) > K (273.06) > Mg (196.86) > Ca (22.48); For P in brown rice of landraces: nuda vs. non-nuda (126.67) > lowland vs. upland (98.33) > red vs. purple (60.05) > white vs. purple (50.67) > glutinous vs. non-glutinous (39.83) > early-mid vs. late (31.40) > indica vs. japonica (24.41) > white vs. red (9.39); K: red vs. purple (257.01) > white vs. purple (160.30) > white vs. red (96.70) > glutinous vs. non-glutinous (93.34) > early-mid vs. late (40.19) > nuda vs. non-nuda (36.04) > lowland vs. upland (14.23) > indica vs. japonica (9.24); Ca: white vs. purple (21.13) > early-mid vs. late (12.82) > red vs. purple (10.74) > white vs. red (10.19) > lowland vs. upland (3.97) > nuda vs. non-nuda (3.65) > indica vs. japonica (2.19) > glutinous vs. non-glutinous (1.46); Mg: nuda vs. non-nuda (76.23) > lowland vs. upland (43.21) > glutinous vs. non-glutinous (27.55) > indica vs. japonica (21.04) > white vs. purple (20.69) > early-mid vs. late (16.76) > white vs. red (15.36) > red vs. purple (5.32).

Mean and parameters of microelement contents in brown rice

Mean and parameters of microelement contents in brown rice from various classification traits in Yunnan rice are given in Table 2. Overall, the most significant differences (P<0.01) occurred, using one-way anova via SPSS software, for three microelements (Zn, Cu and Mn) in brown rice between rice landrace and improved cultivars, but only for Fe between indica vs. japonica of improved cultivars. There was no significant difference between indica and japonica among core collections of rice landraces. Significant differences (P < 0.05) between different types among rice landraces for some elements were: glutinous vs. non-glutinous (Zn), early-mid vs. late (Fe and Cu), non-nuda vs. nuda (Zn). The largest difference (mg/kg) of Fe: improved cultivar of indica vs. japonica (20.76) > landrace and improved cultivar (5.57) > white vs. purple (3.68) > white vs. red (2.05) > early-mid vs. late (4.92) > lowland vs. upland (2.38) > glutinous vs. non-glutinous (2.01) > red vs. purple rice (1.63) > nuda vs. non-nuda (1.87) > indica vs. japonica (0.14); Zn: landrace and improved cultivar (9.95) > improved cultivar of indica vs. japonica (3.68) > red vs. purple (3.45) > nuda vs. non-nuda (3.14) > glutinous vs. non-glutinous (2.21) > white vs. purple (1.83) > white vs. red (1.61) > lowland vs. upland (1.55) > early-mid vs. late (1.27) > indica vs. japonica (0.74); Cu: red vs. purple (3.73) > landrace and improved cultivar (3.18) > white vs. purple rice (2.20) > early-mid vs. late (1.98) > improved cultivar of indica vs. japonica (1.64) > white vs. red (1.54) > lowland vs. upland (0.92) > glutinous vs. non-glutinous (0.44) > indica vs. japonica (0.30) > nuda vs. non-nuda (0.21); Mn: landrace and improved cultivar (2.55) > red vs. purple (1.65) > nuda vs. non-nuda (0.97) > white vs. red (0.96) > white vs. purple (0.69) > improved cultivar of indica vs. japonica (0.48) > early-mid vs. late (0.25) > indica vs. japonica (0.19) > lowland vs. upland (0.18) > glutinous vs. non-glutinous (0.07). From the results it can be concluded that the largest differences among classification traits in brown rice of core collections from Yunnan rice are Fe (improve indica and japonica), Cu (white and purple rice), Zn and Mn (rice landrace and improved cultivar), but Fe and Zn (indica and japonica), Cu (red vs. purple rice), Mn (glutinous and non-glutinous) have much less significant differences.

Phylogenetic relationship of elemental content in brown rice

According to a cluster analysis of the mean contents (mg/kg) and anova of eight mineral elements in brown rice for 863 accessions using SPSS 10.5 software, the most significant differences are between rice landrace and improved rice. Among rice landraces, we analyzed the genetic relationships of 14 rice types according to mean contents (mg/kg) of eight mineral elements for 789 accessions in Yunnan brown rice using SPSS 10.5 (see Figure 1). The rice landraces formed five distinct major subgroups (A: purple rice; B: nuda and upland; C: non-glutinus; D: red rice; E: nine other types), which seemed to be correlated with the provenance of the samples. Most of rice landraces and purple rice were clearly grouped (on the two sides of the dendrogram). Interestingly, purple rice does not cluster with either of the groups. Among group E (nine types), the dendrogram demonstrates that subgroups I (japonica, non-nuda, early-mid and glutinous) and II (rice landraces, indica, late and white rice) have close associations with red rice (group D), but they are more distantly associated with group C (non-glutinous), group B (upland and nuda) based on the mean of eight mineral elements in brown rice (see Figure 1).

Mean elemental content in Yunnan brown rice

Mean contents (mg/kg) of 863 accessions in Yunnan brown rice have rather high concentrations of nutritional elements such as P (3 797.72), K (2 631.70), Ca (155.90), Mg (1 512.86), Fe (32.56), Zn (32.49) and Cu (13.94). According to other studies carried out, the mineral content in brown rice (P, K, Ca, Mg, Fe, Zn and Cu) from Vietnamese and Australian rice is relatively lower than that of Yunnan rice (Phuong et al. 1999; Yasui and Shindoh 2000). Further research of the mineral element differences in brown rice between Yunnan rice from China and Vietnamese or Australian rice is indispensable. The mineral element data of the core collections (863) in Yunnan rice obtained from this study can be regarded as a reasonable representation of the mineral levels of the rice germplasm from a large part of the Province; the Vietnamese rice was selected from the Red River region, which is one of the most important rice production areas of Vietnam. The geographic origin of rice can be determined based on the trace-element composition in brown rice (Yasui and Shindoh 2000).

Mean elemental contents in brown rice between subspecies and rice types

Some results indicate that elemental profiling is a useful functional genomics tool (Lahner et al. 2003). The most significant difference (P < 0.01) for seven mineral elements (except Fe) in brown rice is found between rice landrace and improved cultivar, and it is relatively greater than that of subspecies or classification traits. K, Ca, Mg, Fe and Mn in brown rice of improved cultivars is higher than that of landraces, whereas P and Zn is lower. The most significant difference (P < 0.01) for five mineral elementals (P, K, Ca, Mg and Fe) is found between indica vs. japonica of improved cultivars; however, there are no significant differences between indica and japonica among the core collections of rice landraces. Rice landraces contain greater haplotype diversity than that of elite varieties, suggesting that landraces represent an intermediate stage in domestication between wild and elite cultivars (Londo et al. 2006). P, K, Mg, Fe, Zn, Cu and Mn in brown rice of improved indica is higher than that of improved japonica, whereas Ca is lower (see Tables 1, 2); P, Ca, Mg, Fe and Cu in brown rice of landrace japonica is higher than that of landrace indica, whereas K, Zn and Mn is lower (Tables 1, 2). There are significant differences (P < 0.05) of mean content (mg/kg) for eight mineral elements in brown rice of landrace classification traits, such as lowland vs. upland (P and Mg), glutinous vs. non-glutinous (K and Zn), early-mid vs. late rice (Ca, Fe and Cu), white vs. purple (K), nuda vs non-nuda (P, Mg, Zn and Mn). The antiquity of the divide between indica and japonica subspecies has been estimated at more than 100 000 years (Sweeney and McCouch 2007). Rice (O. sativa) was cultivated by Asian Neolithic farmers >11 000 years ago, although the date of rice domestication is a matter of continuing debate. In general, Ca content in brown rice of japonica is higher than that of indica; cold tolerance at booting stage of japonica rice is stronger than that of indica (Zeng et al. 2005b). Ca content was found to be significantly correlated with panicle length, 1–2 node length, unfilled grain number per panicle, total grain number per panicle, seed setting rate, and grain density, implying that Ca content of brown rice was closely related with panicle traits, especially with seed setting rate (Zeng et al. 2005a). The domesticated forms have accumulated genes for adaptation to cultivated fields, which might be associated with the loss of adaptive genes to survive in natural habitats (Onishi et al. 2007). These results demonstrate that indica-japonica differentiation of five mineral elements (P, K, Ca, Mg and Fe) in brown rice enhance the eurytopicity from landrace to improve cultivar.

The evolution of elemental contents in brown rice for Asian cultivated rice

For the first time, a core collection of Yunnan rice was investigated and the phylogenetic relationship was based on its mineral element content in brown rice, which reveals the evolution of Asian cultivated rice. Elemental dating of the divergence time between indica and japonica rice has yielded intriguing results. Hierarchical cluster analysis of dendrogram using average linkage from SPSS software showed that Yunnan rice could be grouped into rice landrace or improved cultivar, with the rice landraces being further clustered into five subgroups. As stated previously, purple rice does not cluster with either of the groups. Our present data confirms that indica is the closest relative of late and white rice, and they constitute rice landraces together, whereas japonica is the closest relative of non-nuda, early-mid and glutinous rice. According to our research, japonica might be more primitive than indica, non-nuda might be more primitive than nuda, early-mid rice might be more primitive than late rice, glutinous might be more primitive than non-glutinous, and white or red rice might be more primitive than purple rice. Waxy DNA sequence analyses indicate that the splice donor mutation is prevalent in temperate japonica rice varieties, but is rare or absent in tropical japonica, indica, aus, and aromatic varieties (Olsen et al. 2006). Sequencing of the alleles from both mapping parents as well as from two independent genetic stocks of Rc revealed that the dominant red allele differed from the recessive white allele by a 14-bp deletion within exon 6 that knocked out the basic-helix-loop-helix (bHLH) domain of the protein (Sweeney et al. 2006). Nuclear and chloroplast data support a closer evolutionary relationship between indica and aus and among the tropical japonica, temperate japonica, and aromatic groups (Garris et al. 2005). Seven model-based populations were identified on the basis of the genetic structure of 692 rice landraces in Yunnan revealed by 36 simple sequence repeats (SSR) markers, and are partly consistent with Ding's five-level taxonomic system (Zhang et al. 2006). The five macroelements (P, K, Ca, Mg and S) of brown rice from O. rufipogon and O. officinalis and O. granulata were higher than that of the six cultivars (Cheng et al. 2005). The origin of cultivated rice has puzzled plant biologists for decades (Sang and Ge 2007). The genomic variation in rice has revealing implications for studying the genetic basis of indica-japonica differentiation under rice domestication and subsequent improvement (Tang and Shi 2007). Therefore the phylogenetic relationship of elemental content in brown rice gives insight into the domestication process and the relationship between the subspecies or ecotypes of classification traits, and reveals the evolution of Asian cultivated rice. These results support the notion that Yunnan is the center of genetic differentiation of indica and japonica subspecies of Asian cultivated rice (Zeng et al. 2007).

Plant materials

The present study was conducted at Bamen village (altitude 500 m) of Yaojie town in Xingping county (24°08′ N, 101°99′E) of Yunnan Province in China, and belongs to the double-cropping rice zone and enjoys a very warm climate. Because of a strong reaction of de-silicate and Al accumulation, the red mudstone paddy soil there is acid (pH 5.53) with high levels of Fe and Al. Soil analysis showed organic matter 29.6 g/kg, available P 10.08 mg/kg (total soil P 530 mg/kg), available N 108.47 mg/kg (total soil N 1 760 mg/kg), available K 61.42 mg/kg (total soil K 19 900 mg/kg), available Ca 1 505 mg/kg, and available Mg 310 mg/kg. Average available micronutrient levels in the soil were 3.80 mg/kg of Cu, 1.45 mg/kg of Zn, 297 mg/kg of Fe, and 22.75 mg/kg of Mn. The previous crop of the experimental plots was cucumber. In order to reduce field error and workload, we mixed the brown rice with the same quantities for two replications per accession, for fertility and management uniformity in the examination field. A total of 863 accessions of rice core collections from Yunnan were grown together at the same site and time. High temperature and rich rainfall were characteristic of the weather from the tillering to ripening stages, and the highest water temperature in the fields was determined to be 41°C during the early tillering stage. Double irrigation during the tillering period, together with rich rainfall, provided sufficient water throughout the growing periods. Classification methods were used as follows: the four taxonomic characters (indica vs. japonica; early-mid or late rice; lowland vs. upland; and non-glutinous vs. glutinous) were divided by Ding's taxonomic system (Ding 1959), and determined some traits (nuda vs. non-nuda, white vs. red or purple rice) according to the published reports (Zeng et al. 2003).

Assay methods

Apparatus, analysis condition of spectrum, reagents, standards and sample preparation of determination of ICP-AES were used according to Zeng et al. (2008). P, K, Ca, Mg, Fe, Zn, Cu and Mn contents of rice seeds in 863 accessions harvested from the experimental plots (50 plants/plot) were tested with two replications in the Supervision and Testing Center for Farm Products Quality, Ministry of Agriculture, China. These samples were then de-husked using the laboratory de-hulling machine. Half a gram of each brown rice sample was precisely weighed and put into a beaker. Nitric acid (5 mL) and 1 mL of perchloric acid was added, the beaker was heated with electricity to nitrify and decompose rice until the solution became clear. The clear solution was continuously heated and evaporated to dry. The residue was dissolved with 5 mL of 1:1 hydrochloric acid and the solution was transferred into a graduated bottle of 50 mL. The sample was reduced to ashes with high temperature or decomposed by warm acid decomposition. The residue was dissolved in acid solution to transform the elements to test into inorganic ion. This solution was then properly diluted and analyzed with an ICP-AES spectrometer. Mean, standard deviation (SD), t-test and mean cluster were used in the analysis using SPSS software (Release 10.5, SPSS Inc. Chicago, USA). Significance analysis between different types for each element was carried out using one-way anova with SPSS software at the 5% or 1% probability level.

(Handling editor: Song Ge)