Nat. Cell Biol. 10, 902–911 (2008); published online 6th July 2008

The version of this Article initially published lacked sufficient detail on how samples for western blotting were processed. Many experiments involved analysis of multiple proteins of similar molecular weight, making stripping and re-probing of blots often problematic. For this reason, our preferred approach was to analyse the various proteins by parallel gels/blots, with immunoblotting of γ-tubulin or actin being performed to ensure good quality of samples and absence of protein degradation before or after cell extract preparations. For all immunoblot analysis performed in this paper, equal loading and transfer of proteins were verified by poinceau S staining of membranes. Specifically, the γ-tubulin blots shown in Figs 1c,d, 2a,e and 3b,c were from parallel gels, with comparable results being obtained by re-probing the blots. In Figs 1g, 2d and 6d, tubulin expression was assessed by stripping and re-probing blots. In each of these cases, results were confirmed by an independent experiment. In Fig. 5b, EGFR, c-Jun and actin blots were performed by sequential blotting of the same membrane, whereas p53 expression was assessed by a parallel gel/blot.