TTC26/DYF13 is an intraflagellar transport protein required for transport of motility-related proteins into flagella

1) The proteomic analysis. The quality of the analysis is essential for the conclusions that you draw from the paper. However as far as we can see, you have only done the analysis once. We feel that such a proteomic experiment should be repeated twice, and preferably three times to average the data. We would like to see the analysis repeated, or provided if you have already performed it.

This is a very interesting suggestion as we know any proteomic analysis will always have a certain rate of errors, both false positives and false negatives, particularly for proteomic analysis of complex cellular mixtures or organelle preparations. Because of this potential variability, the degree to which one should believe any particular protein identified in the proteomic analysis depends on the reproducibility of the experiment. To address this point, we repeated the entire experiment again, including growing new cultures, isolating flagella, and analyzing them by DIGE. The new results are highly similar to the previous results. In the first analysis, we recognized 2,324 spots in the 2D gel and 64 spots of them were significantly reduced in dyf13 mutant flagella. The second analysis recognized 3,358 spots and 106 spots of them were reduced. Of the 64 spots reduced in the first analysis, 53 were also reduced in the second analysis, and the ones that were not were all ones that were very close to the cutoff for being classified as having had their abundance changed. More importantly, out of the spots analyzed by mass spectrometry to determine protein identity, all but one (the metabolic enzyme METM) also showed consistent abundance changes in the second analysis, and so our conclusions about the set of proteins depleted in the dyf13 mutation are supported by this second analysis. We have added all of this information to the Results section, and we provide the ratio data for both experiments in separate columns in Supplementary file 1B. We have noted in the Results section that the abundance changes for these 16 spots are correlated between the two experiments with a correlation coefficient of 0.68, which is statistically significant (p=0.0013). We conclude that while there are differences in the exact number of spots classified as changing in abundance in the two experiments, these differences do not affect the identity of the spots that we actually analyzed. We have also provided more information about the number of experiments done for the dynein arm analysis, which were two experiments for outer dynein arms and five times for inner dynein arms.

2) We were surprised that the frequency of IFT trains was not reported since a possible reduction in the number of trains could explain the shorter length of the flagellum, at least in Chlamydomonas. This is also a feasible experiment given the quality of the kymographs shown at Figure 5A. Your group has measured IFT frequency using the same reporter in the Engel JCB2009 paper.

This is also an extremely good suggestion. Using our previously acquired kymograph data, we measured the frequency of IFT trains in control and dyf13 mutant cells. The frequency of IFT train did not significantly change in dyf13 mutant cells. This new data is now presented in Figures 6C and 6F. We thank the reviewers and editors for insisting on this point, because it supports our conclusion that DYF13 is not necessary for IFT behavior.

The reviewers were of the opinion that TEM could help interpretation of your data and that if you already have them; this would strengthen the paper. However, we decided that this was not a condition of publication, and leave this up to you.

We agree that this could be an interesting avenue for future exploration but feel that it is beyond the scope of the present paper.