Extended Data Fig. 10: Additional considerations.

a–c, CaMV35S::PMEI5-transformed plants are severely compromised in female fertility. a, Constitutive expression of PMEI5 severely affected plant development²⁰. b, Flower organogenesis appeared normal, but pistil development was suppressed; at maturity, pistils were at most 1/3 of the normal size. Pollen production appeared normal. c, The stigmatic papillae of these plants were under-developed; pollinated stigma did not retain any pollen grains, reflecting their inability to support pollen germination to penetrate the stigmatic papillae. The aniline blue-stained ovules showed high levels of callose deposition, a symptom of stress. Even in normal pistils, ovules with ectopic callose accumulation do not attract pollen tubes (Extended Data Fig. 2j, k), precluding these PMEI5-overexpressing ovules from being used in ovule penetration studies. Very few seeds were produced from these PMEI5-overexpressing plants (about 10 μl of seeds from a full pot was a good yield under our growth condition). Similar observations were made in several plantings, as we obtained the seeds in 2014.d, e, aca9-1 pollinated pistils. d, Flowering aca9-1 plants produced many under-developed siliques (black); even elongated siliques (white arrows) had a substantially reduced number of fertilized ovules⁴⁰. e, aca9-1 pollination of wild-type pistils. Aniline blue staining showed few elongating pollen tubes and their arrival at ovules (arrowheads), even at 30 h after pollination. Exudates from pistils pollinated by aca9-1 pollen showed a basal level of de-esterified pectin (Extended Data Fig. 4c). These observations are linked to the experiment presented in Extended Data Fig. 4; together, four independent preparations showed comparable results. f–i, NO as a gaseous agent to block the entrance of supernumerary pollen tubes. Observations below are included here to relate how findings from the present study might be linked to several topics that are not yet fully understood. f, Confocal images of mature ovules from unpollinated pistils expressing a moderate level of LURE1–GFP (top; a large majority of ovules in the transformed pistils), showing the typical filiform-apparatus (arrow) localization. Occasionally, an ovule overexpressed LURE1–GFP (bottom), showing the presence of the protein in the inner integuments (i) that envelope the micropylar (m) chamber. The working distance of the LURE1 gradient diffused from the filiform apparatus is not known, although histoimmunodetection suggests that it reaches the micropylar region¹. NO, as a gaseous molecule, should diffuse readily and reach the micropylar region (and possibly beyond), although its working distance is unknown and difficult to determine. These observations are included for the consideration of plausible functional linkage between the two gradients that would be expected to exist in the system. g, LURE1–GFP localization at completion of pollination (20 h after pollination) in LURE1p::LURE1–GFP pistils by wild-type pollen. Typically fewer than 50% of the ovules showed notable GFP signal in the synergid cell (category 1) (Fig. 4a, b); others retained a weak synergid cell signal (category 2). A low percentage of ovules showed LURE1–GFP localization in both the filiform apparatus (arrow) and the synergid cell (category 3). When pollinated by pollen from hap2/+ mutant plant, the sperm from half of the applied pollen (hap2 pollen) was incompetent for fusion; therefore, half of the ovules were not fertilized¹². A higher percentage of ovules from hap2/+ pollinated pistils showed category-3 localization, with notable synergid cell as well as filiform apparatus LURE1–GFP signal. h, Confocal image of a category-3 ovule. A maximum projection from 4 images (1 μm total thickness) (left) and a single optical section (right), showing LURE1–GFP localization in a synergid cell and across the filiform apparatus. It could be envisioned that category 3 ovules in hap2/+ pollinated pistils could be candidates for fertilization recovery (that is, having the propensity to salvage fertilization). In the hours after failed fertilization, NO induced by the arrival of the sterile hap2 pollen tube could have dissipated (for example, see data plot in Fig. 2c), allowing some level of secretion of newly synthesized LURE1 and pollen tube attraction to be reactivated. The phenomenon in g was observed in every experiment reported in Fig. 4a (leftmost two panels). The data presented here were pooled from two independent experiments, designated to collect the numerical data presented here for discussion. i, Unpollinated but ageing ovules maintained LURE1–GFP at their filiform apparatus and thus should continue to be competent in pollen tube attraction. These data also demonstrate that NO accumulation at the filiform apparatus does not correlate with ovule age, providing further support for the notion that NO accumulates in response to pollen tube arrival. These observations were similar to those in Fig. 4a (left) and control experiments shown in Extended Data Fig. 6a–c. Scale bars, 50 μm. Arrows, filiform apparatus region.