To the editor

The epidermal growth factor receptor (EGFR) is cotranslationally inserted into the endoplasmic reticulum (ER) membrane as a type I transmembrane protein. This is a highly regulated process involving inherent topogenic and anchoring sequences. Furthermore, the lumenal/extracellular part of the receptor is subsequently glycosylated and folded, and the extracellular domain eventually possesses several stabilising disulphide bonds. The transmembrane part of the receptor is very hydrophobic. In mutants of the insulin receptor, where one or more hydrophobic residues of the transmembrane segment were instead polar, the entire protein was transported into the ER lumen and eventually secreted. An event whereby the full-length EGFR after synthesis and membrane insertion could be translocated out of the membrane, into the cytoplasm and subsequently into the nucleus by way of a nuclear localization sequence, as proposed by Lin et al.1, would involve yet unresolved and exciting mechanisms. But no such mechanism is substantiated, nor suggested, in the paper published in the recent issue of Nature Cell Biology1.

We strongly doubt the interpretation of the authors' results. It is important to realise that the ER membrane is contiguous with the nuclear membrane. Given this fact, proteins localizing to the ER could easily be misinterpreted as nuclear. This could explain the observation of the EGFR localizing to nuclei. The immunofluorescence data provided do not convincingly demonstrate nuclear EGFR localization. It is difficult to distinguish between the nuclei and the cytosol, and the optical section could be too thick to give the required resolution. Furthermore, the purity of nuclear fractions is highly critical with respect to interpretations regarding transport of radiolabelled EGFR to the nucleus.

It is interesting that a new track from the plasma membrane to the ER was recently described2. Potentially, similar mechanisms for EGFR intracellular routing could be involved.