A Positive Feedback Loop Mediated by CsERF31 Initiates Female Cucumber Flower Development: ETHYLENE RESPONSE FACTOR31 mediates a positive feedback loop that initiates female cucumber flower development

Abstract

Sex determination is a crucially important developmental event that is pervasive throughout nature and enhances the adaptation of species. Among plants, cucumber (Cucumis sativus L.) can generate both unisexual and bisexual flowers, and the sex type is mainly controlled by several 1-aminocyclopropane-1-carboxylic acid (ACC) synthases. However, the regulatory mechanism of these synthases remains elusive. Here, we used gene expression analysis, protein-DNA interaction assays and transgenic plants to study the function of a gynoecium-specific gene, ETHYLENE RESPONSE FACTOR31 (CsERF31), in female flower differentiation. We found that in a predetermined female flower, ethylene signalling activates CsERF31 by CsEIN3, and then CsERF31 stimulates CsACS2, which triggers a positive feedback loop to ensure female rather than bisexual flower development. A similar interplay is functionally conserved in melon (Cucumis melo L.). Knockdown of CsERF31 by RNAi causes defective bisexual flowers to replace female flowers. Ectopic expression of CsERF31 suppresses stamen development and promotes pistil development in male flowers, demonstrating that CsERF31 functions as a sex switch. Taken together, our data confirm that CsERF31 represents the molecular link between female-male determination and female-bisexual determination, and provide mechanistic insight into how ethylene promotes female flowers, rather than bisexual flowers, in cucumber sex determination.

Keywords: ERF; cucumber; ethylene signalling; positive-feedback; sex determination; unisexual flower.

Figure 1

The expression pattern of CsERF31. A, The relative expression level of CsERF31 in different organs and treatments in the main stem androecious (CsACS1/CsCS2 genotype), gynoecious (CsACS1G/CsACS2), and hermaphroditic (CsACS1G/csacs2) cucumber plants. Error bars indicate SD (n = 3). The letters in (A) were used to indicate the difference between each other P < 0.01 in (a1 and a2), (a3 and a4), (b1 and b3), (b2 and b4), (b1 and c1), (b2 and c2), (b3 and c3), (b4 and c4), (d1 and d2), (d3 and d4), and (d5 and d6) from the Student’s t test. B–D, In situ hybridization of CsERF31 mRNA during different stages of bud development; the red arrow represents the hybrid signal. Pe: petal; St: stamen; Pi: pistil. Scale bar = 200 μm (B–D).

Figure 2

Expression of CsERF31 is induced by ethylene signaling. A, GUS staining of proCsERF31:GUS transgenic Arabidopsis plants treated with ethephon and ddH₂O. B, Y1H assays showing the interaction of CsEIN3 and AtEIN3 with the EBS1 (EIN3 binding site1) box region in the CsERF31 promoter. C, EMSA showing the interaction of CsEIN3 and the EBS1 box region in the CsERF31 promoter. D, Schematic diagram of the constructs using dual-LUC analysis in cucumber cotyledon transient expression systems. TSS, Transcription start site. E, Effects of CsEIN3 on the activity of the CsERF31 promoter (pCsERF31). Asterisks indicate significant differences in different groups. Error bars indicate SD (n = 8). ^*P < 0.001, Student’s t test.

Figure 3

CsERF31/CmERF1 binds the CsACS2/CmACS7 promoter in vitro and activates CsACS2/CmACS7 transcription in vivo. A, Y1H assays showing the interaction of CsERF31/CmERF1 with the CsACS2/CmACS7 promoter and the D-ERE box. B, EMSA showing the interaction of CsERF31 and the D-ERE box region in the CsACS2 promoter. C, Schematic diagram of the constructs using dual-LUC analysis cucumber cotyledon transient expression systems. D, Effects of CsERF31/CmERF1 on the activity of the CsACS2/CmACS7 promoter. Asterisks indicate significant differences in different groups. Error bars indicate SD (n = 8). ^*P < 0.001, Student’s t test. D and E, Schematic diagram of the constructs using GUS reporter analysis in transgenic tobacco lines. F, GUS staining of the hybridization line of pCsACS2:GUS and oxERF31 transgenic tobacco and ethephon treatment in proCsACS2:GUS leaves.

Figure 4

Knockdown of CsERF31 leads to unarrested stamen development on female flowers. A, Representative images of CsERF31-RNAi transgenic (CsACS1G/CsACS2), vector (transgenic negative control, CsACS1G/CsACS2), and WI1983H (CsACS1G/csacs2) plants. B, The variety flowers on CsERF31-RNAi transgenic plants, vector, and WI1983H. C, Floral organ of the flowers in (B). The red arrow shows induction of unarrested stamen. D, The malformed pistils of the flowers in (B). E, Box-plot of the statistical data showing the ovary length of the bisexual flowers on the flowering day. The letters in (E) were used to indicate the difference between each other P < 0.01 in (a, b), (a, c), and (b, c) from the Student’s t test. Boxes show the first quartile, median, and third quartile; whiskers show minimum and maximum values; dots show outliers data. F, Expression analysis of CsERF31 and CsACS2 in the lateral buds of CsERF31-RNAi lines (1, 2, and 4), vector and WI1983GM (CsACS1/CsACS2). Error bars represent SD (n = 3). G, Ethylene production in CsERF31-RNAi transgenic cucumber, vector, WI1983H, and WI1983GM. Error bars represent SD (n = 4). The letters in (G) indicate significant differences at each other P < 0.01 in (a, b), (a, c), and (b, c) by the Student’s t test. Scale bar = 10 cm (A), 1 cm (B), 2 mm (C), and 1 mm (D).

Figure 5

Ectopic expression of CsERF31 suppresses the stamen and promotes pistil development in male flowers. A, Transgenic plant overexpressing CsERF31 (oxCsERF31, left); empty vector transgenic plants as the negative control (vector, right). Green fluorescence was used as a marker to confirm the expression of the target genes in the plants. Scale bar = 1 cm. B, The expression level of floral development-related genes in the stamen of male flowers in the negative control plants (vector stamen), the pistil of transgenic plants (oxCsERF31 pistil), and the pistil of female flowers in the negative control plants (vector pistil). Error bars represent SD (n = 3). The letters in (B) indicate significant differences at each other P < 0.01 in (a, b) and (b, c) by the Student’s t test.

Figure 6

A working model to explain how CsERF31 mediates female flower development in cucumber. In this model, the upstream ethylene (C₂H₄ in black ovals) represses CsWIP1 in an unknown way and activates CsEIN3 in the ethylene signaling pathway to promote CsERF31 expression. CsERF31 stimulates CsACS2 to produce more ethylene (C₂H₄ in red oval, left part of figure), forming the “ethylene-CsERF31-CsACS” positive feedback loop (ACC, the precursor of ethylene). High expression of CsERF31 arrests the stamen and promotes complete pistil initiation of female bud development. However, in csacs2 mutants, loss-of-function CsACS2 could not form a positive feedback loop. Low expression of CsERF31 (the light red text) can barely arrested the stamen and partially promotes the pistil, producing a bisexual bud only (middle part of figure). When lacking upstream ethylene (right part of the figure), the whole pathway is turned off, and the expression of CsWIP1 induces exclusive male bud development. The red text represents the gene or the process was activated, gray represents the inactivation.