Correction to: Nature https://doi.org/10.1038/nature04681 Published online 4 May 2006

In the Letter “Downstream nuclear events in brassinosteroid signalling”, the following mistakes were made during figure mounting.

The published Fig. 3a depicts a western blot experiment monitoring the phosphorylation status and levels of BES1 at different time points (1, 2, 4 hours) after treatment with brassinolide (BL) or mock. Figure 3a contains undisclosed cropping to organize samples differently from the original experiment and to show a single untreated control, as no change in BES1 phosphorylation or levels was observed with time in mock-treated plants. In addition, an inversion of lanes between the 2 h and 4 h +BL signals occurred during figure preparation. We have retrieved the original blot (shown in the Supplementary Information) and have produced a corrected panel for Fig. 3a in which cropped regions are shown with dotted lines (Corrected Fig. 3a). The correction does not alter the original conclusions that BL treatment leads to dephosphorylation of the BES1 transcription factor, with little or no change in total BES1 levels in our conditions.

Time course analyses of BL treatment on BES1 protein. Plants were treated with mock (-BL) or BL (+BL) for the time indicated and subjected to western blot analysis using anti-BES1 antibody. The blot was cropped to show the -BL 1 h control only. The dotted lines indicate cropping. PBES1, phosphorylated BES1 ; *, non-specific band.

The published Fig. 3b shows a western blot experiment for BES1 in different genetic backgrounds following mock or BL treatments, as well as exposure of wild-type (WT) plants to the BL biosynthetic inhibitor brassinazole (BRZ) to deplete endogenous BRs. The original blot was modified to remove background shadows above P-BES1 signals without mention in the corresponding legend. We have recovered the original blot (shown in the Supplementary Information), and have produced a new panel supporting the prominent role of BL in controlling BES1 phosphorylation status, with no obvious change in the accumulation of BES1 protein in our conditions (Corrected Fig. 3b below). Additionally, the signal corresponding to WT plants grown on BRZ results from the duplication of the BL-deficient det2-1 background lane. The fact that BES1 accumulates under its P-BES1 form in absence of BL and that BL addition promotes the dephosphorylation of P-BES1 into BES1 had already been observed in several independent studies1,2,3, and by us in other unpublished experiments showing the effect of BL and BRZ now shown below (Unpublished Fig. 1). The western blot shown on the corrected Fig. 3b therefore lacks the WT+BRZ lane and, consistently, the phenotype of a BRZ-grown plant was removed. We apologize for such alteration of the blot. The conclusions of this experiment remain unchanged.

BES1 is constitutively nuclear and its levels are unaffected by BRs. BES1 phosphorylation status in response to BL treatment in wild-type (WT) plants or genotypes affected in BR biosynthesis (det2-1) or signalling (bri1-5, bin2-1 and a BRI1-OX plant which overexpresses BRI1(ref. 27 [ref. 6 below]). Plants were treated with mock or BL as described in the legend to Fig. 1d and subjected to western blot analysis using anti-BES1 antibody. P-BES1, phosphorylated BES1 ; *, non-specific band. The phenotype of the corresponding plants is shown. Only the wild-type Col is shown, as both Col and Ws ecotypes responded in the same way to BL treatment (data not shown).
BRZ promotes BES1 phosphorylation. BES1 phosphorylation status in response to BL or BRZ treatments in wild-type (WT). Plants were treated with mock (-BL) or BL (+BL) as described in the legend to Fig. 1d, or grown on 1 μM BRZ (Grown on +BRZ) or 1 μM BL (Grown on +BL), and subjected to western blot analysis using anti-BES1 antibody. PBES1, phosphorylated BES1; *, non-specific band.

There are several additional duplications:

  1. 1.

    Duplications of Ubq10 RT-PCR DNA loading control gels between Supplementary Fig. S1b and Fig. S5 (lanes 1, 2), and between Supplementary Fig. S3c and Fig. S5 (lanes 1, 2; lanes 4, 5, respectively).

  2. 2.

    Possible duplications of lanes 1, 2 between the two subpanels of Supplementary Fig. S3c.

  3. 3.

    Duplications in Supplementary Fig. S5: lanes 3 and 4, BIN2-GFP; and lanes 4 and 5, actin.

Loss-of-function of BIN2 displays constitutive BR responses. Detection of BIN2 mRNA. RT-PCR was performed on total mRNA extracted from whole seedlings using BIN2 specific primers. ACTIN2 (ACT2) amplification served as a control.
Expression analysis of transgenic lines expressing BIN2-GFP and bin2-GFP. Expression of BIN2-GFP and bin2-GFP transgenes. RT-PCRs using BIN2- and GFP-specific primers were performed on several transgenic lines carrying either BIN2-GFP or bin2-GFP constructs. Amplification of Ubq10 was used as loading control. BIN2-GFP and bin2-GFP lines were also subjected to western blot analysis with a GFP antibody. Transgenic lines harboring similar levels of RNA and protein (BIN2::BIN2-GFP #4 and the dwarf line BIN2::bin2-GFP #1) were selected for further analyses. The dotted line indicates cropping. *, non-specific band used for loading control.

Careful examination of other RT-PCR DNA gels highlighted some similarities for Ubq10 between the two panels of Supplementary Fig. S2b.

The seeds corresponding to most lines failed to germinate, except for bin2-3, for which we were able to repeat the experiment presented in the published Supplementary Fig. S1b (full DNA gels are found in the Supplementary Information of this correction file). We therefore built a new panel and confirmed our original conclusions that the bin2-3 allele is a knock-out mutant for the BIN2 gene, and that BIN2 acts as a negative regulator of brassinosteroid signalling (Corrected Supplementary Fig. S1b). In contrast to WT plants, no amplification of BIN2 was indeed observed using cDNAs from bin2-3 while the ACT2 loading control was amplified in both backgrounds. The ACT2 primers used are 5′-GCCCAGAAGTCTTGTTCCAG-3′ and 5′-TCATACTCGGCCTTGGAGAT-3′.

The published Supplementary Fig. S5 depicts the accumulation of BIN2-GFP or bin2-GFP (carrying the E263K gain-of-function mutation) proteins and corresponding mRNA, to confirm that the transgenic lines used expressed transgenes to comparable levels. The protein blot probed with GFP antibodies, aiming at detecting both BIN2-GFP and bin2-GFP, was cropped to remove a lane corresponding to a BIN2-GFP transgenic line that failed to accumulate the corresponding protein and that in addition had a severe phenotype unrelated to brassinosteroids. A similar cropping was performed on the corresponding BIN2-GFP RT-PCR DNA gel. We apologize for not having clearly described such croppings in the corresponding legend. We found the original BIN2-GFP protein blot and BIN2-GFP DNA gel (shown in the Supplementary Information) and have highlighted cropped regions with dotted lines. The published Ubq10 DNA corresponded to a different replicate experiment and has therefore been replaced by the appropriate one in the corrected Supplementary Fig. S5. Additionally, when building the published figure, lane 3 (BIN2::bin2-GFP#1) from the BIN2-GFP protein blots was accidentally used for lanes 4 and 5 (BIN2::BIN2-GFP#4 and #7) as well as lane 4 from the actin blot duplicated as lane 5. The actin loading control blot could not be recovered. We have therefore built a corrected figure panel below, using a non-specific band as loading control (Corrected Supplementary Fig. S5).

The authors apologize deeply for the inconvenience caused. The abovementioned corrections do not change the main conclusions of the published article. In addition, the findings reported in our original article regarding the role of BIN2 and related GSK3 kinases in brassinosteroid signalling, as well as the role of BIN2 phosphorylation in regulating DNA-binding of BES1-related transcription factors, have been fully validated by independent groups4,5.