Correction to: Nature https://doi.org/10.1038/nature14067 Published online 17 December 2014

It was brought to our attention via a comment in PubPeer that images in Extended Data Fig. 1c, Extended Data Fig. 1i and Extended Data Fig. 7k in our paper contain duplicated sub-panels. After extensive internal investigation of the source data (raw, uncropped immunoblots and microscopy images), we can confirm that figure preparation errors (copy-pasting placeholder images without updating them with correct panels) were responsible for the mistakes. The errors are in presentation only, and do not affect our scientific conclusions. Specific explanations related to each duplication are described in the following sections.

1. Extended Data Fig. 1c

The immunoblots for the mAchR4 TIL R230E and the mAchR4 TIL P331E mutants were the same (see Supplementary Information: Correction Fig. 1).

The duplication was most likely introduced upon copy-pasting of a sub-panel to be used as a template/placeholder for a subsequent sub-panel. Unfortunately, the swap with the relevant image for one of the mutants was forgotten, leaving the duplication error. Inspection of the original, uncropped, immunoblot scans revealed that the one for the mAchR4 TIL R230E mutant was the duplicate, and that for the mAchR4 TIL P331E mutant was genuine (see Supplementary Information: Correction Fig. 1). The duplicate image has been replaced by a correct representative immunoblot for the mAchR4 TIL R230E mutant (Supplementary Information: Correction Figs. 1 and 2).

This correction does not affect our scientific conclusions as both mAchR4 TIL R230E and mAchR4 TIL P331E mutants were pulled down by the Endo A2 SH3 domain.

2. Extended Data Fig. 1i

The XY confocal microscopy images for control and Endo TKD were the same (Supplementary Information: Correction Fig. 3).

As above, the duplication was most likely introduced upon a similar copy-pasting of a sub-panel to be used as a template/placeholder for a subsequent one. Unfortunately, the swap with the relevant image for one of the conditions was forgotten, leaving the duplication error. Inspection of the original, uncropped, confocal microcopy images revealed that the image for the control was genuine and that of Endo TKD was the duplicate. The duplicate image has been replaced by a correct representative confocal microcopy image for Endo TKD (Supplementary Information: Correction Fig. 3).

This correction does not affect our scientific conclusions as AP2 was recruited to the plasma membrane of Endo TKD cells.

3. Extended Data Fig. 7k

The bright field microscopy images for Endo1+2+3 TKD, 2 days, –NGF and +NGF were the same, and those for Endo1+2+3 TKD, 4 days, –NGF and +NGF were also identical to each other (Supplementary Information: Correction Fig. 4).

As above, the duplication was most likely introduced upon a similar copy-pasting of a sub-panel to be used as a template/placeholder for subsequent ones. Unfortunately, the swap with the relevant images for some of the conditions (here, –NGF vs +NGF rows) was forgotten, leaving the duplication errors. Inspection of the original, uncropped, bright field micrographs revealed that the images for the Endo1+2+3 TKD +NGF (2 and 4 days) were genuine, as they showed some neurite extensions that were absent in the –NGF condition (Supplementary Information: Correction Figs. 5 and 6). Thus, the images for Endo1+2+3 TKD, –NGF after at 2 and 4 days were the duplicates. The duplicate images have been replaced by correct representative bright field microscopy images for Endo TKD, –NGF, 2 and 4 days (Supplementary Information: Correction Fig. 4).

This correction does not affect our scientific conclusions as, in absence of NGF (–NGF) after two or four days, Endo1+2+3 TKD did not affect neurite extension by PC12 cells (a process which requires +NGF).

4. Source data

The source data for all the figure panels involved contain the original, uncropped images and are available as Supplementary Information in the online version of this amendment.

5. Same image in Fig. 1e and Extended Data Fig. 2b—not a duplication error

Although not mentioned in the PubPeer comment, we also noted that the same microscopy image of a BSC1 cell from a resting (i.e., no denopamine added), control RNAi sample was present both in Fig. 1e and in Extended Data Fig. 2b (albeit at lower magnification in the latter to show the whole lamellipodia) without a specific mention in the figure legends. This was not a duplication error, but in the main figure we were showing a magnification (Fig. 1e), while in Extended Data Fig. 2b we provide a more complete image to show the context. Thus, we are using the same control BSC1 image in both places.

Supplementary information is available in the online version of this amendment.