IC4R004-GWAS-2015-25371505

Project Title

Genetic dissection of ozone tolerance in rice (Oryza sativa L.) by a genome-wide association study

The Background of This Project

• Tropospheric ozone causes various negative effects on plants and affects the yield and quality of agricultural crops. Here, we report a genome-wide association study (GWAS) in rice (Oryza sativa L.) to determine candidate loci associated with ozone tolerance. A diversity panel consisting of 328 accessions representing all subgroups of O. sativa was exposed to ozone stress at 60 nl l –1 for 7 h every day throughout the growth season, or to control conditions. Averaged over all genotypes, ozone significantly affected biomass-related traits (plant height –1.0%, shoot dry weight –15.9%, tiller number –8.3%, grain weight –9.3%, total panicle weight –19.7%, single panicle weight –5.5%) and biochemical/physiological traits (symptom formation, SPAD value –4.4%, foliar lignin content +3.4%). A wide range of genotypic variance in response to ozone stress were observed in all phenotypes. Association mapping based on more than 30 000 single-nucleotide polymorphism(SNP) markers yielded 16 significant markers throughout the genome by applying a significance threshold of P<0.0001. Furthermore, by determining linkage disequilibrium blocks associated with significant SNPs, we gained a total of 195 candidate genes for these traits. The following sequence analysis revealed a number of novel polymorphisms in two candidate genes for the formation of visible leaf symptoms, a RING and an EREBP gene, both of which are involved in cell death and stress defence reactions. This study demonstrated substantial natural variation of responses to ozone in rice and the possibility of using GWAS in elucidating the genetic factors underlying ozone tolerance.

Plant Culture & Treatment

• The experiment was conducted in a greenhouse at the University of Bonn, Germany, from May to November 2013. A mapping population consisting of 328 rice cultivars was obtained from the International Rice Research Institute (The Philippines). The seeds were germinated in the dark for 3 d at 28 °C and then transferred to a greenhouse under natural light. Three-week-old seedlings were transplanted into 2 × 6 m ponds filled with soil (a local luvisol: 16% clay, 77% silt, 7% sand, 1.2% organic carbon, pH 6.3; Schneider, 2005) at 16.5 × 19 cm spacing. A constant water level of at least 3 cm was maintained from 10 d after transplanting throughout the growth season. Each of the six plots contained one replicate of all 328 cultivars in a completely randomized distribution. The plots were randomly assigned to ozone and control treatments, and open-top chambers (height 1.3 m) were built around all plots to ensure an identical microclimate. To avoid nutrient limitations, 107 g of K 2 SO 4 and 98 g of Ca(H 2 PO 4 ) 2 were applied to each plot as basal fertilizer at the beginning of the season, and 155 g of urea was applied in three splits during the season. Temperature, air humidity, and CO 2 concentration were constantly monitored at 12 min intervals using sensors installed at 2 m height in the greenhouse. The average temperature was 25/19 °C (day/night), the average humidity was 60/80% (day/night), and the average CO 2 concentration was 460/600 ppm (day/night). Supplementary lighting was installed above the plots to ensure a minimum light intensity of 12.5 klux even on cloudy days. Water was removed from the ponds in week 19, and the plants were harvested in week 21. Panicles were separated from the shoots and dried at 50 °C for at least 72 h to complete dryness. The shoot samples were dried for 10 weeks in the greenhouse until they reached a constant moisture content of around 11% and then weighed.

Figure 1 Box plots for relative phenotypic values. The median of each trait is shown as the horizontal bar in the box, and the upper and lower sides of a box represent the first and third quartile values of the distribution, respectively. Whiskers extended to 1.5 times the interquartile range (box size) or to the maximum/minimum values.

• Five weeks after transplanting, ozone fumigation was initiated to induce chronic stress at a target level of 60 ppb for 7 h every day. Comparable concentrations are already being observed in some areas and are expected to be reached in many countries in the future. Ozone was produced using custom-made ozone generators (UB 01; Gemke TechnikGmbH, Ennepetal, Germany) with dried air passing through silica gels as input. Ozone output was regulated by an ozone monitor (K100 W; Dr A. Kuntze GmbH, Meerbusch, Germany) with an ozonesensor (GE 760 O3; Dr A. Kuntze GmbH) placed inside the chambers. The generated ozone was blown into a central pipe running above the plant canopy, from which three parallel perforated side pipes for ozone distribution branched off at a distance of 40 cm from each other. The pipes were calibrated for even ozone distribution prior to transplanting of rice seedlings using a handheld ozone monitor (series 500; Aeroqual Ltd, Auckland, New Zealand). The fumigation lasted from 9:00 until 16:00 each day for 15 weeks until the end of the growth season. During the growth season, acute ozone stress was applied three times in weeks 8, 10, and 14 after transplanting. The average concentration of acute stress was 150 ppb and it lasted for 7 h (9:00 to 16:00). The ozone concentration was constantly monitored by the handheld ozone monitor placed within the canopy during the fumigation. The average ozone con-centration recorded was 63 ppb in the ozone treatment (excluding the episodes of acute stress), while in the control the concentration was 12 ppb on average.

Figure 2 Association mapping results for relative SPW. (A) Frequency distribution of observed relative SPW. (B) QQ plot of expected and observed Pvalues. (C) Manhattan plot from association mapping using the MLM. The top 50 SNPs are shown in blue and the SNPs exceeding the significance threshold of P<0.0001 are shown in red. (D) The peak region on chromosome 2. (E) The peak region on chromosome 10. In (D) and (E), pair-wise LD between SNP markers is indicated as D’ values: dark red indicates a value of 1 and white indicates 0. The dotted squares in (D) and (E) denote the linkage blocks that contain significant SNPs. (This figure is available in colour at JXB online.)

Research Findings

• We tested the effect of ozone on nine traits, including leaf cell death as represented by LBS; growth parameters such as plant height, shoot DW, and tiller number; grain yield component parameters such as TKW, TPW, and SPW; and biochemical parameters such as chlorophyll content (SPAD value) and foliar lignin content. We also analysed foliar lignin content as an agronomically important parameter, which may represent apoplastic stress, since the coupling of monolignol molecules requires the oxidation of hydroxyl group and therefore is highly dependent on the apoplastic redox status. ANOVA analysis (Table 1) demonstrated that all of these traits were significantly affected by the ozone concentration employed, i.e. 60 ppb for 7 h daily plus three additional episodes of 150 ppb for 1 d. In plant height, shoot DW, SPW, and SPAD value, we also observed significant interaction between genotype and treatment (G×T). On average, plant height decreased by 1.0%, DW decreased by 15.9%, tiller number decreased by 8.3%, TKW decreased by 9.3%, TPW decreased by 19.7%, SPW decreased by 5.5%, SPAD value decreased by 4.4%, and lignin content increased by 3.4%. Box plots for relative phenotypic values [e.g. relative plant height=(plant height ozone /plant height control )×100] indicated substantial genotypic variation in ozone responses, which was particularly pronounced in the case of relative DW, relative TPW, and relative SPW (Figure. 1). Growth parameters and grain yield components mostly correlated significantly with each other, and LBS showed a significant correlation with relative tiller number. We also compared the ozone response among the five subpopulations (aromatic, aus, indica, temperate japonica, and tropical japonica) plus the admixed group, which cannot be clearly assigned to any of these subpopulations， Subpopulations indica and temperate japonica showed a significantly lower LBS than the other subpopulations. Constitutive lignin content and relative lignin content also showed significant differences among the subpopulations.

Table 1 Effect of ozone stress on phenotypic values

• Relative SPW Relative SPW was distributed approximately normally(Figure. 2A). The –log 10 P values showed deviation from the expected values only in the most significantly associated markers (Figure. 2B), suggesting reliable performance of the MLM for this trait. By applying the threshold of –log 10 P>4.0, we identified three significant SNPs on chromosomes 2 and 10 (Figure. 2C). On both chromosomes, the LD blocks consisted of a relatively small number of SNPs (Figure. 4D, E), and a total of 38 genes were located in the LD blocks containing these three SNPs. The LD block on chromosome 10 contained 20 genes (excluding retrotransposons and a non-expressed gene), among which six genes had leucine-rich repeat regions including five with a GO annotation of ‘signal transduction’, which could be involved in ozone sensing and triggering downstream reactions.

Labs working on this Project

• Institute of Crop Science and Resource Conservation (INRES), University of Bonn, Karlrobert-Kreiten Strasse 13, 53115 Bonn, Germany
• Key Laboratory of Crop Genetics & Physiology of Jiangsu Province, Yangzhou University, Yangzhou 225009, PR China
• Campus Klein-Altendorf, University of Bonn, Klein-Altendorf 2, 53359 Rheinbach, Germany

Corresponding Author

• Michael Frei (mfrei@uni-bonn.de)