Local translation of ribosomal proteins in axons

A recent bioRxiv pre-print publication from Christine Holt’s lab suggests that ribosomes may be remodeled in axons by locally translated ribosomal proteins. This is surprising because we know that ribosomes are assembled in the nucleolus. Well, I have some concerns about a few of the experiments depicted there.

Briefly, Shigeoka et al first show that axons (derived from xenopus embryonic retina) contain ribosomal proteins-coding mRNAs, that these are enriched at a specific branching step of axon growth and that their translation is induced by the growth factor Netrin-1. These mRNAs contain a short cis-element in their 5’UTR that enhances translation upon Netrin-1 treatment. The authors suggest that local translation of these ribosomal proteins (RPs) allows them to locally assemble with ribosomes or ribosome-containing granules and that this is important for local axonal translation and for the branching of the axons.

This is a very important finding, thought provoking and could be relevant to numerous biological systems. But I have a few problems with a number of their experiments.


  1. A simple control is missing:

In figure 2D-F they show that the translation of a reporter mRNA (Venus) with the full 5’UTR of Rps4x (the most abundant ribosomal protein mRNA) is induced upon netrin treatment (by FRAP).  The translation of a construct that is missing the cis-element (or missing the 5’UTR) is not induced.

However, they do not show that the mRNA expression level and localization of all these reporters is similar. If somehow the  deletion reporters are mislocalized or expressed 2-fold lower – that could explain their results as well.


2. Maturation time of Venus is not accounted for

In figures 2E-F and 3C, they photobleach Venus or Venus-Rps4x in the axon, then show the re-appearance of Venus fluorescence withing 2-10 minutes.

However, the maturation time of Venus (at least in vitro – I didn’t find data in vivo) is 40 min. So it seems unlikely that detectable levels of fluorescence appear within the short time of the experiment. To me it seems more likely that it diffused from un-bleached regions of the axon. Adding cycloheximide should resolve this issue.

3. Morpholino experiments – questions & controls

In Fig 5 they show that morpholino (MO) against Rps4x leads to inhibition of total translation in the axon cone (by puromycin labeling and anti-puro antibody).

Its difficult to say from the plot but they show a reduction of ~20% in Rps4x protein after 24hr MO, and close to 40% reduction in nascent translation. How is this accounted for?

Actually, by looking closer at the data of 5C, the major difference between control MO and Rps4x MO is ~10 /62 axons (~16%) with very high Rps4x levels. If you take out this sub-group, then the control and Rps4x MO look very similar.

So, what’s going on here?

Missing controls:

cycloheximide or other translation inhibitor as control for total translation inhibition.

MO against Rps4x in the soma – does it have the same effect on total translation or is this unique to the axon?

MO against another RP which is not enriched in axons.

why is Rpl17 level not reduced, if total translation goes down?

Why do they show data in IF intensity for the axons, but data in Western blot for the soma? is it just a matter of getting enough material for the WB? too much fluorescent signal in the soma? Can WB detect the ~20% reduction seen in the axons?

4. Show me the mRNA

The mRNAs are detected only by RNA-seq. Show me by FISH or live imaging (e.g. MS2) the localization of the mRNAs, that they are in granules that open upon Netrin-1, that there is sufficient number or mRNAs to account for the amount of proteins that they detect e.g. in the photobleaching experiments [which are with CMV promoter], that MO actually affect the mRNA levels.


5. Summary

And after all that, I still think that they are on the right path. It seems likely that at least Rps4x plays a role in axonal branching and it is regulated by the 5’UTR cis element. The SILAC experiments and the IF in Fig 1 both support axonal translation of the RPs upon Netrin-1 treatment and incorporation of them into the ribosomes (although – in Fig 4D they show data for RPL19, but not RPS4X, why only that and not [also] the other?)

This is a pre-print so I do hope that in the next version of the paper some of my questions will be answered and my concerned addressed.


Toshiaki Shigeoka, Max Koppers, Hovy Ho-Wai Wong, Julie Qiaojin Lin, Asha Dwivedy, Janaina de Freitas Nascimento, Roberta Cagnetta, Francesca van Tartwijk, Florian Strohl, Jean-Michel Cioni, Mark Carrington, Clemens F. Kaminski, William A. Harris, Hosung Jung, Christine E. Holt (2018) On-site ribosome remodeling by locally synthesized ribosomal proteins in axons bioRxiv 500033


2 responses to “Local translation of ribosomal proteins in axons

  1. So, the maturation time for venus is well below 40 minutes.


    The estimate for t 1/2 maturation is on the order of minutes, which is why they used venus for these experiments.


    • Hi PDA,
      This 2002 paper indeed shows a maturation rate of t1/2 <5 min.
      However, in the link i provided there's a ref to a 2011 paper which claims its 40min. it also referencing another paper showing t1/2 of 100 min(!). which is correct? I honestly don't know. maybe the 2002 paper and you are right and the interpretation of the experiment is correct. but what if its not?


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s