Extended Data Figure 5: AtSWEET9 is necessary for nectar secretion.

a, Identification of three Atsweet9 T-DNA insertion lines. RT–PCR (40 cycles) performed on RNA isolated from leaves (L) and whole flowers (F) of wild-type, Atsweet9-1 and Atsweet9-2 plants. AtSWEET9 expression can only be detected in wild-type flowers. Actin was used as a constitutively expressed control. b, RT–PCR performed on RNA isolated from wild-type and Atsweet9-3 (SALK_202913C) plants. GAPDH (At3g04120) was used as a constitutively expressed control. c, Lack of nectar in nectaries of Atsweet9-2 mutants. d, e, Nectar secreted from nectaries of complemented Atsweet9 mutants under its native promoter: AtSWEET9 (d) or AtSWEET9–eGFP (e). f, Nectar production in wild-type and Atsweet9-3 flowers. Atsweet9-3 mutant lines do not secrete nectar, similar to Atsweet9-1 and Atsweet9-2. g, h, Nectar production in Atsweet9 mutants complemented with AtSWEET9, 11 (g) or 12 (h) expressed under the control of the AtSWEET9 promoter. Nectar production was restored by expression of AtSWEET11 and AtSWEET12 under the AtSWEET9 promoter in the Atsweet9 mutant plants. Arrows indicate nectar droplet on the peeled-down sepals. i, Ultrastructure of lateral nectaries of wild-type and Atsweet9-1 mutant flowers. The morphology of wild-type and Atsweet9-1 mutant flowers was observed by scanning electron microscopy. Sepals were removed before imaging. As judged by scanning electron microcopy, mutant nectaries appeared normal, indicating that loss of nectar secretion was not caused by physical defects of nectaries. c–h, Original magnification, ×10.