**Table 1.** *Cont.*

### *3.2. Whole Genome Sequence Comparison between CSSL104 and KDML10' and Co-Expression Network Analysis Revealed that Major Hub Genes Have a Role in Photosynthesis*

We compared whole genome sequences of CSSL104 and its drought susceptible parental line KDML105 to define the genes responsible for drought tolerance in CSSL104. A total of 101,950 SNPs located on 3440 genes were detected. The regions with a high density of SNPs were on chromosomes 1, 8, 9, and 11 (Figure 2).

**Figure 2.** Genetic regions introgressed into the KDML105 genome. Single nucleotide polymorphisms (SNPs) between CSSL104 and KDML105 rice. The blue lines show 100 SNPs within 5000 background nucleotides. All loci containing SNPs were subjected to a co-expression network analysis. The results, shown in Figure 3A, revealed 18 major nodes with a high connection to other genes. The gene ontology of these 18 genes is listed in Table 2. The map position of these genes is shown in Figure 3B. Based on quantitative trait loci (QTL) data from the Qtaro database [23], six loci are located in QTL regions for drought-stress tolerance on chromosomes 1, 3, and 8 (Figure 3B). The high-density SNP on chromosome 1 was consistent with the location of the drought-tolerant (DT)-QTL, which is flanked by markers RZ14 and R117. In this QTL, co-expression network analysis identified two genes, *LOC\_Os01g72800* and *LOC\_Os01g72950*, as the major nodes. Chromosome 3 did not display high-density SNPs. On this chromosome, *LOC\_Os03g02590* and *LOC\_Os03g03910* were located in two drought-tolerance QTLs mapped between markers RM7332, RM545, and RG104, RZ329. Another node gene, *LOC\_Os03g52460,* is located between markers C136 and R1618, corresponding to another drought-tolerance QTL. Chromosome 8 displays several major nodes: *LOC\_Os08g16570* is located between markers RM72 and RM331, while *LOC\_Os08g41040* and *LOC\_Os08g41460* are located between RM5353 and RM 3480. This region on chromosome 8 also displayed high-density SNPs between CSSL104 and KDML105 (Figures 2 and 3B).

The eighteen rice genes reported here have homologs in *Arabidopsis* for which tagged mutants are available (Table 2). Nine of them (*CPFTSY*, *NDH-O*, *SOQ1*, *LHCB3*, *RRF*, *PGRL1B*, *HCF244*, *NAD(P)-linked oxidoreductase,* and *MRL1*) were annotated to be involved in the photosynthesis process [24–32]. Moreover, the homolog of *LOC\_Os11g43600* is *CPRF1,* an *Arabidopsis* gene required for chloroplast development [33]. These findings sugges<sup>t</sup> that these rice genes are involved in photosynthesis adaptation during drought stress.

Therefore, we obtained seven homozygous, T-DNA tagged *Arabidopsis* mutant lines corresponding to *ndhO* (*at1g74880*), *lhcb3* (*at5g54270*), *rrf* (*at3g63190*), *pgrl1b* (*at4g11960*), *pgrl1a* (*at4g22890*), *at2g27680*, and *mrl1* (*at4g34830*). These lines were drought stressed by growing them in MS medium supplemented with 0 mM, 75 mM, or 150 mM mannitol. Their growth response was assessed by measuring the green pixel area per plant and compared to the wild type (WT).

**Figure 3.** Candidate genes for drought tolerance. Gene co-expression network was analyzed by using the Rice Oligonucleotide Array Database [19], showing the major node genes with yellow dots (**A**), while DT-QTLs from the Q-TARO database [23] and Kanjoo et al. [11] are shown in red and yellow boxes on the chromosome, respectively. (**B**) Loci written in blue letters indicate the proposed drought-tolerance loci based on this study.


**Table 2.** Rice gene candidates for drought tolerance, their *Arabidopsis* homologs, and the inferred function of the genes in rice.


**Table 2.** *Cont.*

Under normal conditions, all mutant lines displayed a significantly lower number of green pixels than WT, suggesting lower growth than WT (Figure 4A). Both drought-stress treatments decreased growth in all lines, with the 150 mM causing the more severe reduction. At 75 mM mannitol, *pgrl1b* had a significantly lower number of green pixels than WT, while *pgrl1a* showed similar growth to WT. Other mutant lines showed better growth than WT (Figure 4B). Under the severe drought-stress conditions induced with 150 mM mannitol, the growth of *lhcb3*, *at2g27680*, *mrl1,* and WT were similar. Mutants *pgrl1b* and *pgrl1a* had a significantly lower growth than WT, while *rrf* had a lower growth at the beginning of the treatment but displayed better growth than WT after 5 days of the treatment. However, similar growth between WT and *rrf* was found after seven days of drought stress. Among the mutant lines, *ndhO* was the only mutant line that had significantly better growth than WT under severe drought stress (Figure 4C).

The stability indexes of *Arabidopsis* mutants and WT were calculated to compare drought tolerance after six days of drought stress. After the intermediate drought stress (75 mM mannitol), all mutant lines except *pgrl1a* displayed significantly higher stability than WT, suggesting the contribution of *NDH-0*, *LHCB3*, *RRF*, *PGRL1b, at2g27680,* and *MRL1* to drought-tolerance adaptation (Figure 5A). Under severe drought (150 mM mannitol), significantly higher stability than WTs was detected for the *ndh-o, rrf*, *pgrl1b, at2g27680,* and *mrl1* mutants (Figure 5B). The *rrf* mutant line displayed the highest stability under both intermediate and severe drought-stress conditions.

**Figure 4.** Growth response of seven mutant *Arabidopsis* lines to drought stress. Comparison of growth (green pixels per plant) of wild type (WT) and the T-DNA insertion mutant lines *ndhO* (*at1g74880*), *lhcb3* (*at5g54270*), *rrf* (*at3g63190*), *pgrl1b* (*at4g11960*), *pgrl1a* (*at4g22890*), *at2g27680*, and *mrl1*(*at4g34830*) grown in (**A**) normal Murashige and Skoog (MS) medium, (**B**) under intermediate drought stress (MS medium supplemented with 75 mM mannitol), and (**C**) under severe drought stress (MS medium supplemented with 150 mM mannitol). Statistical analysis was by *t*-test. \* and \*\* above the name of the mutant line represent significant difference (*p* < 0.05) and highly significant difference (*p* < 0.01) between WT and mutant, respectively.

**Figure 5.** Stability during drought stress. The stability index, which is the ratio between the values from stressed plants and normal growth plants, is shown in ( **A**) 75 mM mannitol and (**B**) 150 mM. This figure represents the mean ± SE of WT and mutant lines. The \* shows a significant di fference at *p* values ≤ 0.05 and \*\* indicates *p* value ≤ 0.01.
