*2.3. Phenotypic Analysis of Rice FWL Gene Mutants*

Phenotypes of allelic mutant lines that were homozygous, transgene-free, and with no detected off-target effects (see Section 2.4) of the *OsFWL1* and *OsFWL4* genes were analyzed. For *OsFWL1*, T2 plants of lines Osfwl1a#4 and Osfwl1b#11 (Table S6) were selected for Osfwl1a and Osfwl1b targets, respectively. For *OsFWL4*, T3 plants of lines Osfwl4a#7 and Osfwl4b#6 (Table S6) were analyzed for Osfwl4a and Osfwl4b targets, respectively.

The number of tillers per *osfwl4a* and *osfwl4b* mutant plants was 45.9% and 41.1% greater, respectively, than that of the wild type (WT; Figure 2A,B). The number of grains per panicle of mutants was not significantly changed (Figure 2C). Although 1000-grain weight of mutants was slightly reduced (Figure 2D), the grain yield per plant was increased by 25.6–35.8% (Figure 2E).

Additionally, the flag leaf width of *osfwl4a* and *osfwl4b* mutants was 7.7% and 6.3% greater, respectively, than that of the WT (Figure 3A,C). However, there was no marked difference in flag leaf length (Figure 3B). Analysis of leaf epidermal cell size revealed no significant difference in cell length and width between WT and mutants (Figure S4). This suggests that the increase in leaf width of mutants was caused by an increase in cell number but not in cell size. In addition, the plant height of mutants was slightly reduced compared with that of the WT (Figure 2A and Figure S5).

The expression profile of the *OsFWL4* gene during the life cycle of rice was examined using qRT-PCR. *OsFWL4* was mainly expressed in the developing endosperm and the stem at the heading stage (Figure 4); it was also expressed in the leaf, root, and panicle. To gain insights into the molecular function of *OsFWL4*, a gene co-expression analysis was performed using the Genevestigator program [33] with the mRNA-Seq datasets. Many positively correlated genes of *OsFWL4* were involved in cell signal transduction, disease resistance, and heavy metal resistance (Table S7). Interestingly, some negatively correlated genes of *OsFWL4* were found to encode the F-box domain- and BTB (bric-a-brac, tramtrack and broad complex) domain-containing proteins, which may play a role in protein ubiquitination [34] (Table S8).

Grain length of *osfwl1a* and *osfwl1b* mutants was 4.2% and 5.5% greater, respectively, than that of the WT (Figure 5A,B). However, there was no difference in grain width between the mutants and WT plants (Figure 5C). Grain thickness of mutants was slightly lower than that of the WT (Figure 5D). Finally, there was no change in the 1000-grain weight of mutants (Figure 5E). Additionally, plant height, leaf size, and grain yield per plant of mutants were not considerably different from those of the WT (Figure S6).

**Figure 2.** Analysis of yield traits of wild type (WT) and *OsFWL4* gene mutants. (**A**) WT and mutant plants, bar = 1 m. (**B**) Number of tillers per plant of the WT and mutants, *n* = 20. (**C**) Number of grains per main panicle of the WT and mutants, *n* = 10. (**D**) 1000-grain weight of the WT and mutants, *n* = 10. (**E**) Grain yield per plant of the WT and mutants, *n* = 12–15. Error bars are standard deviations. \*\*\**p* < 0.001.

**Figure 3.** Phenotypes of flag leaves of WT and *OsFWL4* gene mutants. (**A**) Flag leaves of the WT and mutants, bar = 2 cm. (**B**) Flag leaf length of the WT and mutants. (**C**) Flag leaf width of the WT and mutants. The values in (**B**) and (**C**) are means of 20 plants. Error bars are standard deviations. \*\*\**p* < 0.001.

**Figure 4.** qRT-PCR results of the *OsFWL4* gene in 14 tissue samples of *japonica* rice Zhonghua 11. The rice *Actin1* gene was used as the internal control. Legend: R1–R3, roots in the seedling, tillering, and heading stages, respectively; St1 and St2, stems in the jointing and heading stages, respectively; L1–L3, leaves in the seedling, tillering, and heading stages, respectively; P1–P3, panicles 5, 15, and 20 cm in length, respectively; En1–En3, endosperms 5, 14, and 21 days after pollination. Error bars are standard deviations of three technical repeats.

**Figure 5.** Analysis of grain shape of WT and *OsFWL1* gene mutants. (**A**) Grains of the WT and mutants, bar = 5 mm. (**B**) Grain length of the WT and mutants, *n* = 50. (**C**) Grain width of the WT and mutants, *n* = 50. (**D**) Grain thickness of the WT and mutants, *n* = 50. (**E**) 1000-grain weight of the WT and mutants, *n* = 10. Error bars are standard deviations. \*\**p* < 0.01; \*\*\**p* < 0.001.
