Identi cation Of New Sources Of Rice Genotypes (Oryza Sativa L.) Tolerant To Cold Stress


 Rice crop is adversely affected by cold stress which is a common problem in the highland areas of central East Africa (Uganda, Rwanda, north east Tanzania and Kenya) and Madagascar. Cold stress is known to cause 26% to 80% yield loss. Currently, rice production in Uganda is mostly growing to the low altitude areas as there is lack of cold-tolerant varieties that can grow in the higher altitudes of the country. None of the released varieties in Uganda are tolerant cold. The few cold-tolerant lines that have been identified are short-grain types, though Ugandan consumers generally prefer long-grain size. This research aimed at identifying new sources of cold tolerant rice genotypes with preferred traits. A total of 50 lines arranged in a 5x10 alpha lattice design with two replications were used for both stressed and non-stressed experiments. Twenty one (21) days old seedlings were stressed at 10oC in cold air for 10 days and the other remained under normal ambient temperature. IRRI Standard Evaluation System (SES, 2013) was used to score the effect of cold stress on leaf morphology (leaf wilting score) and on leaf color (leaf yellowing score) after 7 day recovery. This study revealed eight highly cold tolerant (SCRID091-20-2-2-4, GIZA 177, NERICA 1, MET P27, MET P23, MET P60, MET P20 and MET P5) in which only GIZA 177 was short grain and fifteen tolerant varieties (MET P32, MET P11, MET P2, MET P17, MET P3, MET P24, MET P16, MET P37, MET P18, MET P9, MET P40, MET P31, MET P39, MET P22, MET P36) were identified.


Introduction
Rice is a sensitive crop that frequently exposed to variable biotic and abiotic stress which adversely affect growth and production (GRiSP, 2013). Biotic constraints include; weeds, blast, rice yellow mottle virus (RYMV) and African rice gall midge (AfRGM). Major abiotic constraints include; extreme temperatures (heat and cold) (Dramé et al., 2013;Zhang et al., 2014), flood and drought (Odogola, 2006 ;Namazzi et al., 2010), and variable rainfall (Akongo et al., 2016) as well as soil problems (salinity, nutrient deficiencies and toxicities) (NEWEST, 2012;Dramé et al., 2013). However, cold stress is the most devastating abiotic production constraint which affects the rice plant at all growth stages from germination to grain filling, leading to high economic yield losses (Gothandam, 2012, Shinada et al., 2013. It is observed in the high altitude areas around the globe. Cold stress that occur at critical reproductive stages can adversely affect grain quality and cause yield reductions in high-altitude regions of the world (Singh et al., 2017). The loss in quality and quantity is as a result of severe reduction in seedling establishment, growth retardation, decrease tiller number, delay and incomplete grain maturation (Satake and Hayase, 1974).
In Uganda, identification of rice genotypes tolerant to cold stress is still in its exploratory stage and all the varieties under production are sensitive to the stress. Some cold tolerant rice genotypes were identified by Nyiramugisha et al. (2016) but they were short grained, sticky and non-aromatic; characteristics which is not preferred by farmers. In this study, the susceptibility and tolerance of 50 rice genotypes to cold stress was investigated. The aim was to identify rice genotypes with new sources of tolerant to cold stress and with acceptable grain size for the improvement of rice in current breeding programme.

Genetic materials and description of study area
Fifty (50) genotypes, including a tolerant and a susceptible check, were obtained from NaCRRI, in which their sources were from: Africa Rice (Ibadan, Nigeria), African rice (Benin), Egypt, Tanzania and Madagascar were used for the study (

Experimental design and data collection
A total of fifty genotypes were tested for their reaction to cold stress following the protocol described by (Lee, 2001). Seedlings of 3-4 leaf stage (21 days after planting) were exposed to 12 o C/10 o C (day/night) cold air temperature for 10 days in a cold growth chamber following a 12/12 hour day/night artificial lighting cycle. Ten days after stress treatment, the seedlings were taken back to normal growth conditions to allow the genotypes to recover, for seven days. The experiments were laid out in a 5 by 10 alpha lattice design, replicated twice. Two sets of experiments were established with one group exposed to cold stress while the other group was made to grow at normal temperature conditions (stress free). The seeds were planted in bin plastic pots of 24 cm x 36 cm width filled with forest soil. Each genotype was planted in a pot containing twelve hills.. Di-ammonium phosphate (DAP) fertilizers (18-46-0) was applied at the rate of 0.4 g/pot during planting while urea (46-0-0) was top dressed at the rate of 0.4 g/pot fifty days after germination. All agronomic management practices were done regularly as needed.
Data based on visual rating (leaf growth/leaf wilting score and seedling color/leaf yellowing score) using Standard Evaluation System (SES) (IRRI, 2013), survival rate, seedling height and tiller number were collected 7 days after recovery for the stressed plants. Similar data were also collected from the non-stressed plants. However, for the non-stressed experiment, genotypes were not different for visual rating data and survival rate. As a result these parameters were not considered for further analyses. Instead, seedling height, tiller number and grain yield per pot were used for comparison. Survival rate was calculated using the following formula from (Nohrman, 1953), i.e.
Survival rate(%) = Surviving plants Total number of treated plants * 100 Seedling color and leaf growth were taken based on scale of 1-9 developed by IRRI (2013). The scales for these two traits were basically given by the extent of cold damage on leaf coloration and leaf morphology respectively. For seedling color, a score of 1 represents dark green seedling while a score of 9 represents dead seedling. For leaf growth, a score of 1 represents no damage on leaf and normal leaf color and hence highly tolerant while 9 represents most seedlings are dead or dying and hence highly sensitive. Therefore, based on leaf growth damage, seedling with scores 1-3 are considered as cold tolerant, 4-6 are intermediate and 7-9 are susceptible. The detailed description for the scales is presented in Table 2.

Data analysis
Data were analyzed using GenStat software, 18 th edition (Payne et al., 2015) Restricted Maximum Likelihood (ReML) approach was used to generate analysis of variance (ANOVA) and to examine differences in the performance of different rice genotypes for their tolerance to cold stress. Fisher's LSD test was used to separate the means. Pearson correlation was used to examine the association among the different tolerance parameters. The model used was as follows; = Y ̿ … + + + + Where, = The observed value of trait from the i th genotype from the k th block nested in the j th replicate, Y ̿ … = Grand mean, Gi = the effect of i th genotype, =Effect of j th replication, ( ) is the effect of the k th block nested within the j th replicate , = random residual or error term Base index was estimated by the linear combination of the mean phenotypic values of the traits weighted directly by their respective economic weights (Williams 1962). Base index value was calculated as: Where, y=Mean values, a= is economic weight of the economic characters under study Tolerance indices and reduction percentage were calculated for yield per pot for the data collected from stress and non-stress experiment. Cold tolerance indices were calculated as follow; = (1 − ( ⁄ ))/(1 − ( ̅̅̅ ̅̅̅̅ ⁄ ) …. (Fischer and Maurer, 1978) = − ……… (Rosielle and Hamblin, 1981) = ⁄ ………… (Bouslama and Schapaugh, 1984) (%) = ( − ) * 100 …. (Choukan et al., 2006) Where Ys= yield under stress, Yp= Yield under non-stress, ̅ s=Yield grand mean under stress, ̅ p=Yield grand mean under non-stress, SSI= stress susceptibility index, TOI=Tolerance index

Response of rice genotypes to cold stress
The analysis of variance revealed a highly significant difference (p<0.001) among genotypes for traits collected under stress (leaf growth, seedling color, survival rate, seedling height and tiller number) (Table 3), indicating genetic variability among tested genotypes in their reaction to cold stress. However, the genotypes did not show significant difference under non-stress conditions to all the traits except seedling height significant at p<0.01 (Table 4). 9

Performance of rice genotypes under stress and non-stress conditions
The tested genotypes were classified into five different cold tolerance categories based on the (IRRI) (SES) scale with a little modification using leaf growth/leaf wilting score. Eight genotypes recorded score ranging from 1.0 to 1.3 were classified as highly tolerant, fifteen tolerant were classified as tolerant (score 1.7-3.1), nine were classified as moderately tolerant (score 3.2-4.8), thirteen were classified as susceptible (scores 5.5-6.6) and five classified as highly susceptible (7.6-9.0) ( Table 5, Figure 1 and Appendix 1).

Figure 1: Reaction categories of 50 rice genotypes to cold stress at seedling stage in terms leaf wilting score after recovery
Seedling color score varied significantly (p<0.001) among the tested genotypes after stress treatment recovery (Table 3). The seedling of the tested genotypes fell into five color categories.
Five genotypes exhibited a dark green color with the mean score ranging from 1.0 to 1.39, eighteen were light green (mean score ranging 1.6-2.5), ten yellowish (mean score ranging 3.6-5.4), thirteen brownish (mean score ranging 5.6-6.9) while four seedlings dried and died (mean score ranging 8.3-9.1) ( Figure 2 and Appendix 1). Genotypes ARC39-145-P-2 (5) and (ARS126-3-B-1-2 (11) which were the most susceptible were completely dead after recovery. Survival rate results showed a significant variation among the tested genotypes after stress recovery. It varied from 0% (completely dead) to 100% (complete seedlings survival). Highly cold tolerant genotypes exhibited 100% survival rate whereas the highly susceptible genotypes revealed low survival rate (0-33%) (Appendix 1). The cold tolerant check (GIZA177) exhibited low score (little damage on leaf and dark green color) and high survival rate, whereas susceptible check TOX 3058-28-1-1-1/WITA 9 showed higher Score (most seedling dried and reddish brown in color) and low survival rate ( Figure 3) The evaluated genotypes varied significantly in their seedling heights both under stress and nonstress conditions (Table 3 and Table 4). Seedling height under stress ranged from nothing to 37.2 cm while that under non-stress conditions ranged from 24.9 cm to 45.1 cm (Appendix 1). The tallest genotype under stress was SCRID091-20-2-2-4-4 (37.2 cm) followed by MET P9 (34.1 cm) while the shortest genotype was ARC36-2-1-2 (8.5 cm). Likewise, genotypes MET P9 and SCRID091-20-2-2-4-4 were the tallest under non-stress (45 cm each) (Appendix 1). In general, lower seedling heights were recorded from the stressed plants compared to the non-stressed plants (Appendix 1). Genotypes varied significantly in their reduction in seedling heights.
The number of tillers per hill varied significantly for genotypes evaluated under cold stress (Table 3). However, there was no significant difference among genotypes for tiller number under non-stress conditions. The number of tiller per hill ranged from 1 to 3 for stressed genotypes and 2 to 5 for non-stressed genotypes. The highest numbers of tillers were recorded for MET P60, followed by MET P27 GIZA 177, NERICA 1 and SCRID091-20-2-2-4-4, while lowest tiller number was recorded for ARC36-2-1-2 in the stressed seedlings. Reduction in tiller number for non-stress and stress condition varied significantly for the evaluated genotypes (Table 3).

Comparison of rice genotypes based on tolerance indices
Performance differences between stressed and non-stressed genotypes were clearly seen after recovery and in later growth stages (Figure 3). and yield stability index (YSI), are presented in Table 8. The reduction rate ranged from 7.33 to 100% for grain yield. The lowest reduction in grain yield was recorded for genotype SCRID091-20-2-2-4-4 (7.33%) followed by MET P20 (7.97%), indicating tolerance to cold stress. Highest yield reduction percentage was recorded for ARC39-145-P-2 (5) and ARS126-3-B-1-2 (11) at 100%, suggesting high susceptibility to cold stress. The SSI value ranged from 0.24 to 3.29. The lower the SSI value the more tolerant the genotypes is. The lowest SSI value, was calculated for C D

Discussion
The study demonstrates the existence of new sources of tolerance to cold stress in the new rice genotypes which could be used to introgress cold tolerance into farmers' preferred but cold susceptible rice cultivars. Substantial variations were observed among the tested rice genotypes for cold stress parameters such as leaf growth/leaf wilting and seedling color/leaf yellowing, survival rate, tiller number and seedling height (Table 5).
Leaf growth/leaf wilting score was able to distinguish the tolerant and susceptible genotypes.
Similarly, Andaya and Mackill (2003), were able to identify highly cold tolerant rice genotypes at seedling stage using leaf grow/leaf wilting score and suggested that leaf growth could be the most suitable parameter in evaluating cold tolerance at seedling stage. The evaluated genotypes were grouped into five categories based on their reaction to cold treatments. The categories include; highly tolerant (8 genotypes (2014) and Wainaina et al. (2015). In addition, GIZA 177 is a Japonica subspecies. Japonica subspecies are known to be cold tolerant as confirmed by Cruz and Milach (2004) and Park et al. (2013). SCRID091-20-2-2-4-4 has background of origin of Madagascar, which is one of the places were rice is grown mostly in the highland parts of the country. The newly identified highly cold tolerant genotypes were from African rice Ibadan Nigeria and they were from the O.barthi interspecific line. Even though, GIZA 177 is a highly cold tolerant genotype, it is of a short grain type which is not preferred by Ugandan farmer. However, it can be a good source of cold tolerance for breeding programme improving farmer preferred varieties. SCRID091-20-2-2-4-4 is a medium grain size variety. The other newly identified genotypes (i.e. MET P27, MET P23, MET P60, MET P20 and MET P5 have long grain sizes preferred by Ugandan farmers. Leaf growth/leaf wilting scores were significantly (p<0.001) and negatively correlated survival rate, tiller number, seedling height and grain yield (g) per pot (Table 9), suggesting the lower the score the better performance of genotypes. There was strong positive correlation between leaf growth and seeding color scores (r =0.93 *** ) suggesting that selection can be done using any of the traits.
The loss of chlorophyll content observed in rice seedlings was evaluated using seedling color/leaf yellowing scores. The tested genotypes were grouped into five different leaf color categories based on seeding color observed after recovery. The categories included dark green (with 5 genotypes), light green (with 18 genotypes), yellowish (with 10 genotypes), brownish (with 13 genotypes) and mostly dried and dead seedlings (with 4 genotypes). Cold stress heavily affects chlorophyll content, mainly exhibited though leaf discoloration which finally weakens the photosynthetic ability of the seedling (Lou et al., 2007;Kim et al., 2012;. Studies conducted by Kim et al. (2012) and Donoso et al. (2015) confirmed that seedling color scores was related to chlorophyll content. In this study, seeding color was significantly and negatively correlated with survival rate, seedling height, tiller number and finally grain yield.
This suggested that seedlings with green color (i.e lower score) had good survival, height, tiller and grain yield. In other words, genotypes with lower seedling color scores had better photosynthetic ability than seedlings with higher scores. Similarly, Ranawake and Nakamura (2011) reported that genotypes with green color exhibited high height in their evaluation of inbred lines for cold tolerance at seedling stage.
Assessment of survival rate was effective in identifying degree of recovery in genotypes. Cold tolerant genotypes recovered up to 100%, whereas susceptible genotypes were unable to recover and/or had low recovery from the applied cold stress. These results were similar to Kim et al. (2012) in their study conducted to evaluate rice seedlings tolerance to constant and intermittent low temperature stress observed good recovery. Survival rate was significantly and positively correlated with seedling height, tiller number and grain yield per pot. The higher survival rate the more tolerant the genotypes and the better the performance in later growth.
Seedlings were stunted under applied cold treatment. This was clearly observed by comparing seedlings under stress and non-stressed conditions. Seedling height under stress ranged from 8.51 to 37.2 cm and 24.9 to 45.2 cm for those in non-stress conditions. Stunting at seedling stage is of the major indicators of cold stress, as reported by Yoshida (1981);  and Zhang et al. (2014). In this study, seedling height was positively correlated with survival rate, tiller number and grain yield per pot.
Examination of tillering is one useful way to examine the growth status of rice plants (Yoshida, 1981). In this study, tiller number per hill was ranged from 1 to 3 and 2 to 5 for the stressed and non-stressed experiment respectively. Lower tiller number was exhibited under stressed conditions suggesting that cold stress affected tillering ability. Similar result were obtained Shimono et al. (2007) and Ndour et al. (2016) who reported that low temperature at seedling stage reduced tillering ability.
Twenty three best cold tolerant rice genotypes were selected by base index selection using five traits, for consistent performance under highland parts of Uganda. Base index is a method of selection where indices are established by the linear combination of the mean phenotypic values of the characters weighted directly by their respective economic weights (Williams, 1962).
Economic weight was given to the traits based on their importance to identify cold tolerance. As Biosci et al. (2013) reported economic weight is given to certain traits based on breeding objectives and their importance or contribution in selecting genotypes of interest. Traits with high genetic variance had the power to distinguish tolerance and susceptibility therefore received high economic weights (i.e LG, SC, SR and TN). Traits with low genetic variance had less power in distinguishing the cold tolerance and susceptible genotypes and received low economic weight (i.e SH). Based on the index (the sum value of the five traits), SCRID091-20-2-2-4-4 was ranked first followed by MET P27. The lowest value was obtained for ARS126-3-B-1-2 (11).
The highest index value for SCRID091-20-2-2-4-4 showed that this genotype performed relatively better for the five collected traits and revealed its tolerance to cold stress. The lower index value of ARS126-3-B-1-2 (11) genotype pointed to the susceptibility of the genotype to cold stress. Base index selection was able to identify and group the cold tolerant and susceptible rice genotypes in this screening condition which fitted well with visual selection and tolerance indices selection.
Tolerance indices tell us the amount of yield reduction due to cold stress as compared to nonstress conditions. The cold stress treatment used in this study caused up to 100% yield reduction for highly susceptible genotypes. Low Stress Susceptibility Index (SSI) value implied tolerance of the genotypes under applied stress. SCRID091-20-2-2-4-4 followed by MET P20 had low SSI value. Numerous studies confirmed that the lower the SSI value, the more tolerant the genotype is and it was reported as a good index to identify stress tolerance and susceptible genotypes (Guttieri et al., 2001;Talebi et al (2009) ;Shiranirad and Abbasian, 2011;Zdravković et al., 2013). High Tolerance index (TOL) value indicated high loss due to the applied stress and susceptibility of the genotypes. The lower the TOL value, the more desirable the genotypes (Rosielle and Hamblin, 1981). Based on this approach the desired genotypes in this study was SCRID091-20-2-2-4-4 which had low TOL value, followed by MET P20 and the undesirable genotype was ARC39-145-P-2 (5). The higher yield under stress to non-stress ration (YSI) suggested tolerance of genotypes (Bouslama and Schapaugh, 1984). In all values of tolerance indices, SCRID091-20-2-2-4-4 ranked first followed by MET P20 while the worst genotype was ARS126-3-B-1-2 (11) which ranked last for tolerance indices. According to this study, yield reduction rate, TOL, SSI and YSI were the best fitted tolerance indices to identify cold tolerant genotypes. A similar study reported that these indices were the more powerful stress discriminators in severe cold stress condition (Talebi et al., 2009). In general, tolerance indices were able to distinguish genotypes which tolerate cold stress and showed small performance difference as compared to non-stressed treatment. This revealed that genotypes with low TOL, SSI and YSI can give better yield in presence of cold stress. In other words, farmers may face lower yield penalty if they grew these varieties with low TOL, SSI, and YSI.

Abbreviations
Df. Degree of freedom

Consent for publication
No applicable

Availability of supporting data
All data generated and analyzed in this study included in the figures and tables presented in the manuscript

Competing interest:
No competing interest

Funding
This work was funded by scholarship from Alliance for a Green Revolution in Africa (AGRA)