Slow development restores the fertility of photoperiod-sensitive male-sterile plant lines

of fertility of photoperiod the plant growth under short-day photoperiod. fertility Photoperiod- and thermo-sensitive genic male sterility (P/TGMS) lines are widely used in crop breeding. The fertility conversion of Arabidopsis thaliana TGMS lines including cals5-2 , which is defective in callose wall formation, relies on slow development under low temperatures. In this study, we discovered that cals5-2 also exhibits PGMS. Fertility of cals5-2 was restored when pollen development was slowed under short-day photoperiods or low light intensity, suggesting that slow development restores the fertility of cals5- 2 under these conditions. We found that several other TGMS lines with defects in pollen wall formation also exhibited PGMS characteristics. This similarity indicates that slow development is a general 36 mechanism of PGMS fertility restoration. Notably, slow development also 37 underlies the fertility recovery of TGMS lines. Further analysis revealed the pollen 38 wall features during the formation of functional pollens of these P/TGMS lines 39 under permissive conditions. We conclude that slow development is a general 40 mechanism for fertility restoration of P/TGMS lines and allows these plants to take 41 different strategies to overcome pollen formation defects.


44
Hybrid breeding, using three-line and two-line systems, is applied in increasing 45 crop yields to meet the rising needs for food worldwide (Li et al., 2007). The three-

56
Much progress about TGMS mechanism has been made in recent years. A number 57 of TGMS loci have been identified in rice. In AnnongS-1, a pre-mature stop codon 58 is introduced in the RNase Z S1 gene (Zhou et al., 2014). The genome of 59 HengnongS-1 contains a mutation in the homolog of Arabidopsis MALE 60 STERILITY1 (MS1) (Qi et al., 2014). An artificial TGMS rice line was created by 61 silencing a UGPase (Chen et al., 2007). In addition, TGMS in indica rice Peiai64S 62 is determined by the mutation of the genetic locus for a long noncoding RNA 63 (PMS1T and PMS3) (Zhou et al., 2012). Similar to rice, several TGMS loci have 64 been identified in Arabidopsis, including the PLANT U-BOX 4 (PUB4), MYB33, lipase that hydrolyzes triglycerides into glycerol and hexadecanoic acid which are 68 the components of plasma membrane (Zhu et al., 2020). The male sterile mutants 69 acyl-coa synthetase5-2 (acos5-2), cyp703a2, callose synthase 5-2 (cals5-2), and 70 ruptured pollen grain (rpg1) have defective pollen walls (Dong et al., 2005;Morant 71 et al., 2007;Guan et al., 2008;de Azevedo Souza et al., 2009). These lines also 72 show the TGMS phenotype (Zhu et al., 2020). Analysis of the underlying 73 mechanism revealed that slow development induced by low temperature is involved 74 in the fertility conversion in these TGMS lines (Zhu et al., 2020). 75 The first PGMS line was discovered in 1973 in the japonica rice variety NongKen 76 58S (Shi, 1985). Then, a number of PGMS lines derived from NongKen 58S with 77 significant hybrid vigor, were created for application in agriculture (Fan et al., 78 2016). Photoperiod was considered to be the only environmental regulator of 79 fertility conversion of the sterile lines until the discovery of TGMS in the late 1980s 80 (Sun et al., 1989;Chen 2001). In the summer of 1989 in China, low temperatures 81 unusually led to unexpected fertility restoration in several rice sterile lines 82 including W6541S, AnnongS-1, and HengnongS-1 (Chen, 2001). In 1991, the 83 unusual high autumn temperatures caused unintentional sterility in the 7001S rice 84 line (Chen, 2001). Later on, it was found that both photoperiod and temperature 85 may contribute to the sterility or fertility recovery (He et al., 1987). In addition, 86 researchers found that in O. sativa, japonica varieties are mainly affected by 87 photoperiod, while indica varieties are mainly affected by temperature (Sun et al., 88 1991). Scientists also discovered that mutation of a noncoding RNA shows 1C), similar to that observed under low temperature ( Figure 1B). In addition, results 115 from Alexander staining showed the formation of mature pollen grains (purple) in 116 cals5-2 under SD condition ( Figure 1C). Thus, SD photoperiod restores pollen 117 formation and fertility of cals5-2. 118 The pollen development process, microgametogenesis, depends on timely 119 coordination of meiosis, mitosis, cell growth and expansion (Sanders et al., 1999).

120
Our previous work showed that the pollen development is slowed under low 121 temperature (Zhu et al., 2020). To analyze whether the development is also slowed 122 under SD conditions, we compared the pollen growth under LD and SD conditions 123 using two different approaches. First, we analyzed the growth rate of cals5-2 from 124 tetrad to mature pollen grain under LD and SD conditions. One tetrad encloses four 125 microspores in callose wall, which can be easily observed using light microscopy.

126
In Arabidopsis, mature pollen grains contain three nuclei which are easy to identify.

127
Under the LD condition, a few fertile pollen grains can be occasionally observed in 128 cals5-2 anthers ( Figure 1A). It took about three days for tetrads to develop into 129 mature pollen grains under LD photoperiod (n=10; Figure 1D), compared to about 130 four days under the SD photoperiod (n=12; Figure 1D). The second approach was 6 and their sizes were 17.0 μm to 19.3 μm after two and four days' growth, 137 respectively ( Figure 1E). These results demonstrated that SD photoperiod 138 significantly slowed microspore development. A similar delay in microspore 139 development was observed under the low temperature conditions which can restore 140 the fertility of cals5-2 and other TGMS lines (Zhu et al., 2020). Therefore, both low 141 temperatures and SD photoperiod restore the fertility of cals5-2 plants through 142 slowing microspore development.

144
Low light intensity restores the fertility of cals5-2. 145 In addition to photoperiod, light intensity is another important factor that affects 146 light signaling and general energy production. Thus, we further investigated 147 whether the fertility of cals5-2 can be restored by low light intensity. We placed 148 cals5-2 in the growth chamber with low light intensity (45 μmol m -2 s -1 ) for ten light intensity took about three and four days, respectively ( Figure 2C). By 155 measuring microspore diameters at two and four days after release from tetrad, we 156 found the microspores were significantly smaller at both time points under the low 7 microspore development. Taken together, low light intensity restores cals5-2 159 fertility through slowing its microspore growth, which is the same as the low 160 temperature and SD photoperiod.   Figure 3N), but some electron dense (pe-like) materials were 193 visible between the plasma membrane and the thin wall (TW; Figure 3N and O).  The expression level of CalS5 in cals5-5 was found to be about 54.9% of that of 215 wild-type ( Figure S2B). cals5-5 was fully fertile under normal conditions ( Figure   216 4A). In cals5-5, pollen production and callose wall formation were similar to those 217 of wild type ( Figure 4B; Figure S2D). To find out whether high-temperature could 218 cause male sterility in the mutant, cals5-5 was placed under high temperature 219 (28 o C). The wild type control was fully fertile at both 23 o C and 28 o C ( Figure 4A).

221
The average seed set of the cals5-5 was significantly decreased ( Figure 4C 13 remained rough and irregular, similar to what was observed for cals5-5 ( Figure 5A   272 and B). Thus, despite the pollen wall defects in these lines under low temperature, 273 SD photoperiod, and low light intensity, the pollen fertility can be restored through 274 slow development (Figure 1 and 2).    cycles, 72°C for 5 min. The qPCR was performed as described (Peng et al., 2019).

390
All reactions were repeated three times with β-TUBULIN as the normalizing gene. grains were taken using an Olympus optical microscope with an Olympus digital 396 camera. Aniline blue staining of callose was performed as described in a previous 397 report ( Zhang et al., 2007). The anthers were fixed in Carnoy's fixative for 2h and 398 then the tetrads were separated to a glass slide. The tetrads were stained with 0.1% 399 (m/v) aniline blue. The pictures of callose staining were taken with an Olympus 400 BX-51 microscope (Olympus). T&D staining assay was performed as described in 401 a previous report (Lou et al., 2014). The inflorescence of WT and mutant were 402 embedded into spur resin, and the sections of pollen were put on the surface of the 403 50°C dryer. Then the desiccation of sections was performed, and the sections were 404 stained with toluidine blue for 5 min (10 mg/ml), Tinopal for 15 min (10 mg/ml) 405 (Sigma, USA) and DiOC 2 for 5 min (5 mg/ml) (Sigma). measured using an Olympus optical microscope. To determine the microspore or 427 pollen diameters, images were captured using a BX51 Olympus microscope and 428 measured using ImageJ software (NIH). The F-test was performed to compare the 429 diameters between normal condition and low light intensity conditions.      conditions, but its fertility was largely restored under SD photoperiod or low light 525 intensity. (C) npu-2 showed male sterility under normal conditions, but its fertility 526 was significantly restored under SD photoperiod or low light intensity. SEM 527 observations revealed that the pollen surface was restored to a normal reticulated