Up- and downregulation of BcMF27 resulted in pollen abnormality
To determine the function in pollen development, amiRNA and overexpression vectors of BcMF27 were constructed (Fig. 1a, b) and induced in B. campestris ssp. chinensis var. parachinensis. Positive transgenic lines were confirmed by PCR analysis and named BcMF27-amiR and BcMF27OE (Fig. 1c). The BcMF27 transcripts from the inflorescence of BcMF27-amiR and BcMF27OE plants were detected by real-time RT-PCR (Fig. 1d). The results showed that the expression of BcMF27 decreased in the inflorescence of BcMF27-amiR1-3 plants and increased in the inflorescence of BcMF27OE-1-3 plants although variation of the BcMF27 transcript was not remarkable.
No difference was observed in vegetative growth and flower organs development among BcMF27-amiR and BcMF27OE and control transgenic plants (Suppl. Fig. S1). However, SEM indicated that the pollen of BcMF27-amiR and BcMF27OE plants was defective (Fig. 2). Of pollen from BcMF27-amiR plants, 93.56% possessed collapsed walls (Fig. 2b, e, f, h). The surface of 29% of abnormal pollen walls locally accumulated unknown material (Fig. 2e). Of BcMF27OE pollen, 66.67% had collapsed walls (Fig. 2c, g, i). Pollen from control plants had an ellipsoid shape and normal reticular structure, and the deformity rate was only 13.75% (Fig. 2a, d, h, i).
To clarify anther and pollen development in BcMF27-amiR and BcMF27OE transgenic plants, semi-thin sections of BcMF27-amiR and BcMF27OE pollen were collected (Fig. 3). Tapetum development in BcMF27-amiR (Fig. 3f–j) and BcMF27OE (Fig. 3k–o) plants was identical to that of control plants (Fig. 3a–e). BcMF27-amiR (Fig. 3f–i) and BcMF27OE (Fig. 3k–n) pollen developed normally from the pollen mother cell stage to the binucleate stage. However, BcMF27-amiR pollen (Fig. 3j) exhibited two kinds of malformed shapes and BcMF27OE pollen (Fig. 3o) presented an irregular shape compared with control pollen at the trinucleate stage (Fig. 3e).
To further detail BcMF27-amiR and BcMF27OE pollen development, TEM was performed (Fig. 4). BcMF27-amiR pollen development normally proceeded to the uninucleate stage (Fig. 4f–h, r) in accordance with the control (Fig. 4a–c, q). At the binuclear stage, the intine thickened normally at germinal furrow regions in control pollen (Fig. 4d, t), but 95.45% of BcMF27-amiR pollen intines formed abnormally outside germinal furrows (Fig. 4i, u). Furthermore, 36.36% of BcMF27-amiR pollen with an additional germinal aperture displayed partial cytoplasm degradation and unknown material on the surface of pollen (Fig. 4j, x) and the remaining defective pollen still included four germinal furrows (Fig. 4p, z), while the control pollen contained a normal organized wall with three germinal furrows and a dense cytoplasm (Fig. 4e, w) at the trinucleate stage. BcMF27OE pollen showed normal development from the pollen mother cell stage to the binuclear pollen stage (Fig. 4k–n, s, v), whereas 52% of pollen exhibited aberrant intine deposition outside the germinal furrow regions at the mature pollen stage (Fig. 4o, y).
Functional disruption of BcMF27 caused morphologically unstable pollen tubes
Alexander staining was performed for pollen viability analyses (Suppl. Fig. S2). It was observed that, similar to control pollen (Suppl. Fig. S2a), BcMF27-amiR (Suppl. Fig. S2b) and BcMF27OE (Suppl. Fig. S2c) pollen grains were purplish red, which indicated normal BcMF27-amiR and BcMF27OE pollen viability.
In BcMF27-amiR and BcMF27OE transgenic plants, pollen germination in vitro was checked to determine the effect of BcMF27 expression imbalance on pollen germination and pollen tube growth (Fig. 5). The average germination rates of BcMF27-amiR and BcMF27OE pollen were 74.95% and 71.86%, respectively. The pollen germination of BcMF27-amiR and BcMF27OE was hardly affected compared with control pollen (80.68%; Fig. 5d). However, BcMF27-amiR (Fig. 5b) and BcMF27OE (Fig. 5c) had twisted tubes compared with the positive control (Fig. 5a). The percentage of twisted pollen tubes in BcMF27-amiR1, BcMF27-amiR2, and BcMF27-amiR3 (91.67%, 85.22%, and 94.58%, respectively) were significantly increased (Fig. 5d). Similarly, the percentage of twisted pollen tubes was significantly increased in BcMF27OE-1, BcMF27OE-2, and BcMF27OE-3 (50%, 78.33%, and 41.11%, respectively; Fig. 5d). However, pollen from BcMF27-amiR (Fig. 6b) and BcMF27OE (Fig. 6c) on control pistils germinated normally in vivo compared to that from control (Fig. 6a) at 4 h after pollination. The average seed number per silique from BcMF27-amiR, BcMF27OE, and control plants was not significantly different (data not shown).