Sowing dates and sterile lines assessment
Results from pollen staining with 1% I-KI solution are illustrated in Figure 2. During two growing seasons, most pollens of all sterile lines appeared malformed. All observed pollens could not be stained by I-KI solution, indicating they were completely abortive. Additionally, all bagged spikes from four sterile lines yielded no seed. Consequently, four sterile lines performed 100% sterility, which means all seeds obtained in these sterile lines were from cross-pollination. Meanwhile, pollens of the restorer Yunmai112 were well-developed and could be darkly stained, thereby were fertile.
As a description of floral architecture attributes, the K41S and K64S showed exposed stigmas. Conversely, K68S exposed weak stigma exsertion, and K66S exhibited without exiting stigmas from the lemmas (Figure 3).
The mean performance of total florets per ten spikes on the male sterile line ranged from 87.00 to 98.20 florets. Therefore, regarding florets with exserted stigma number and SER floral traits of K41S and K64S recorded the same trend of about 21.87% and 22.81%, respectively, and revealed a significant difference from K66S and K68S recorded low florets with exserted stigma number and SER, with 0% and 0.82%, respectively (Table 1).
Floral architecture, especially stigma structure
The typical characteristics of stigmas in four sterile lines were illustrated in Figure 4. Each sterile line displayed a different stigma architecture during the flowering stage. The longest and shortest stigma separately came from K64S (4.25 mm, SER=22.81%) and K66S (3.14 mm, SER=0) (Figure 5A), and the correlation analysis results are clear - there is a significant and positive correlation between stigma length and SER (r=0.46, P<0.05, Table 2). However, we must consider that stigma length alone may not be the sole factor influencing SER. For instance, take K68S, which has only 1% SER - its stigma length is almost the same (4.19 mm) as that of K64S (4.25 mm). Therefore, it's crucial to examine other factors that could contribute to SER and not just rely on stigma length alone. In stigma branch angle, K68S was 137.33°, significantly bigger than other lines whilst K66S had the smallest value at 60.09° (Figure 5B), indicating the SER is independent of the stigma branch angle (Table 2). Regarding the stigma stretch width, K64S (5.99 mm) and K68S (5.89 mm) were significantly bigger than K66S (3.71 mm). Correlation analysis showed the stigma stretch width had a positive but not significant correlation with SER (Figure 5C, Table 2). In addition, three sterile lines K41S, K64S and K68S with stigma exsertion had significantly longer hairbrush lengths of stigma than K66S, but the correlation between hairbrush length and SER did not reach a significant level (Figure 5D, Table 2).
Generally, SER had a positive correlation with all observed stigma traits in this study, but only stigma length was significantly and positively correlated with SER.
Relationship of SER with glume characteristics of the sterile line
The differences in glume width among four sterile lines were displayed in Figures 6 and 7A. K41S, K64S and K66S had similar glume widths at 3.47-3.62 mm, significantly thinner than K68S (3.98 mm). Correlation analysis showed the glume width has a significantly negative correlation with SER (r=-0.46, p<0.05, Table 3), suggesting narrower glume is helpful for higher SER. However, glume width is not the only factor affecting SER because K66S did not have stigma exsertion.
In glume length, the minimum and maximum values were separately coming from K66S and K68S that had no or few exserted stigmas, suggesting glume length has no relation with SER in this study, which was also confirmed by the results from correlation analysis (Figure 7B, Table 3).
Furthermore, the length-width ratio of glume in K64S was significantly higher than that in K66S (Figure 7C), with values of 3.33 and 2.96, respectively. Meanwhile, it significantly and positively correlated with SER (r=0.67, p<0.01, Table 3). These results imply that wheat sterile lines with high SER generally would have a slender glume. Therefore, the length-width ratio of glume may be used as a helpful index for germplasm and offspring selection in breeding wheat sterile lines with a high stigma exsertion rate.
Stigma branch angle was a primary factor causing stigma exsertion in high SER line K41S and K64S
The SER of K41S and K64S were only 22-23% (Table 1), indicating most florets with non-exserted stigma though they have the same genotype as the stigma-exserted florets. Comparative observation from K41S and K64S revealed that there were significant differences between florets with exserted and non-exserted stigma in stigma branch angle and stigma stretch width (Table 4). Other floret traits such as stigma length, glume width and length-width ratio of glume that were significantly correlated with SER above, showed no significant differences between stigma exserted and non-exserted florets here. Considering the significant correlation between stigma branch angle and stigma stretch width (Table 2), these results suggested that the narrowing of the stigma branch angle from 123.26° to 96.34° in K64S and from 107.26° to 81.24° in K41S, reduced the stigma stretch width from 6.56 mm to 5.52 mm in K64S and from 5.34 mm to 4.13 mm in K41S, consequently caused stigmas unable to strew out the glume.
Exserted stigma impact to out-crossing seed setting rate
The highest total out-crossing grains (TOG) per ten spikes were recorded by K41S and K64S with 38.10 and 40.60 grains, respectively. On the contrary, values of 27.30 and 32.90 grains were recorded by K66S and K68S, respectively (Table 5).
The out-crossing seed setting rate of the florets with stigma exsertion was significantly higher than without. Hybrid seed proportions on female parents, e.g., K41S and K64S, recorded 41.37% and 42.11% of total tested spikes, respectively. However, K66S and K68S had values of 31.43% and 33.50%, respectively (Table 5).
Furthermore, the out-crossing ability (out-crossing seed setting rate with exserted stigma/out-crossing seed setting rate with non-exserted stigma) of the florets with exserted stigma was about 3.17 times that of the floret with non-exserted stigma in K41S and K64S(Table 5). Moreover, there is a significant correlation between out-crossing seed setting rate (OSSR) and the number of florets with exserted stigma, SER, grains of floret with exserted stigma and grains of floret with exserted stigma accounting for total hybrid seeds, with the r were 0.94, 0.93, 0.94 and 0.93, respectively (Figure 8). These results revealed that SER is an essential factor affecting the out-crossing seed setting rate in the wheat TPSGMS line.