Nature 439, 290–294 (2006)

Portions of the work repeated with respect to abscisic acid (ABA) binding have revealed errors in the calculations associated with Fig. 1, with the result that the molar ratio of ABA bound to FCA is substantially lower than claimed. There are also difficulties with the data in Fig. 2a, b that arose from the preparation of FY. We conclude that there is no effect of ABA on the FCA–FY interaction, and therefore requested to retract this paper on 14 July 2008. See the Brief Communication Arising in this issue1.