Movies of primary cilia: GFP-Rab8A

I made some time-lapse movies this week of cilia in pig kidney epithelial cells (LLC-PK1) and saw some interesting things that I can’t explain and/or don’t really understand. I thought this would give a good chance to explore this as a route to getting some feedback. I would be very interested in any comments on these videos, especially those labelled 1 and 2. The major caveat is that this is from transient transfection so is at a relatively high level of expression. I absolutely do not rule out that we are inducing artefacts by doing this but it is just possible that we are highlighting structures better. Lots more we could do of course but in particular I am keen to know whether anyone knows of a similar tubular network to that in Movie 1 or to the odd “Trafficking” event seen in Movie 2.

Experimental detail:

Cells plated 24h before transfection and imaged 18h after that (so a total of 42h from plating to imaging).

Transfection was done using 1 microgram of DNA and Lipofectamine 2000 (standard protocol, premix LF2000 with Optimem for 5 mins then mix with DNA and 15 mins later drop carefully on to cells). GFP-Rab8a cDNA was a kind gift from Johan Peränen (Helsinki, Finland; construct described in Hattula, K., Furuhjelm, J., Tikkanen, J., Tanhuanpaa, K., Laakkonen, P., and Peranen, J., 2006. Characterization of the Rab8-specific membrane traffic route linked to protrusion formation. J. Cell Sci. 119, 4866-4877).

Cells were grown in complete medium throughout and imaged in DMEM supplemented with HEPES and sodium bicarbonate but in the absence of serum or phenol red. Over these short times, serum withdrawal does not seem to change ciliation of these cells – they grow cilia in complete medium on reaching confluence. Imaging was done on a wide-field microscope and hopefully these movies include scale bars and time stamps.

The videos are hosted by YouTube and are available through the the links below.


1. Tubular network visible below the primary cilium.

Is this a subset of endosomes? Part of the ER? Part of the cilium itself e.g. an expanded pre-ciliary vesicle? I wonder if this could be an expanded Rab11/Rab8 compartment as implicated by many papers including those of Chris Westlake and colleagues (. ).

2. Unusual looping out from base of cilia.

This looks odd, 3D reconstructions suggests that this “object” is indeed emerging from the base of the cilium and then seemingly heading back in again. Could this be something related to intraflagellar transport (IFT)? I need to read more on the way that Rab8 is trafficked to, and within, the cilium.

The next three are more just observations of things that one might expect to see.

3. Long wavy cilia.

This is a maximum intensity projection of a z-stack where the cilium is showing some substantial movements – not entirely sure how to describe it – we aren’t doing anything to these cells so this is really just diffusive/random motion. These are of course non-motile cilia so it isn’t beating. Just looks good! A key caveat is of course the speed of imaging versus the speed of movement.

4. 3-part cilia?

This video shows some retraction of a ling cilium, presumably induced by us imaging it (i.,e. photodamage). The interesting parts are the apparent 3-part labelling of the structure at the start of the movie suggesting a difference in thickness or accumulation of GFP-Rab8a along the length. The final structure is something we often see around the dish of cells – presumably “damaged” cilia. They are long and fragile in these in vitro cultures.

5. Cilia wrapping around one another

Here we see several cilia on adjacent cells interacting with one another. This seems similar to what the Lippincott-Schwartz lab has recently described in the journal Cilia:

Primary cilia utilize glycoprotein-dependent adhesion mechanisms to stabilize long-lasting cilia-cilia contacts Carolyn Ott, Natalie Elia, Suh Jeong, Christine Insinna, Prabuddha Sengupta, Jennifer Lippincott-Schwartz Cilia 2012, 1:3

It is hard from our video to define individual cilia or to determine the extent or effect of any interaction. As always it is hard to capture this to video – looking down the microscope, these structures clearly look like cilia and are certainly on top of the cells so do not represent the tubular network seen in video 1.

So, thoughts please. Comments here preferred but I’ll check YouTube as well.

PS Any alternative to this method of posting or problems people encounter would be welcome. Would this be better on Figshare? I haven’t explored that yet and these aren’t really figures, just observations on which I would welcome feedback.

If you got this far, thanks for checking my first contribution to the brave new world of Open Science also known as “crowd sourcing better understanding”!



7 thoughts on “Movies of primary cilia: GFP-Rab8A

  1. I am far from an expert in cilliogenesis but here are my thoughts. If Rab8 subfamily GTPases both cause the formation of cilia, and also positively mark them, it might be easier if a secondary marker was present, to see whether the primary effect are on ciliogenesis or on Rab8 localization.

    1) What is the releative distribution of these? For example I can really only see the tubular network in the one image.
    2) Are these roughly equivalent expression levels, or for example do you get the multiple cilia when you have really high Rab8A expression?
    3) Do you have any sense whether the transfected Rab8A is mostly GTP or GDP in these cells?

    • Great points Dave.
      1. Tubular network rare, clearly expression level dependent phenotypes here but all likely revealing I’d say.
      2. No apparent correlation between phenotypes and expression level.
      3. Its wild-type Rab8a so I would expect most to be empty (?).
      Always unclear in these transient expression experiments as to how much the probe reveals of a “naturally occurring process” and how much the probe itself induces. More work needed, just hoping someone might have seen the same….
      Thanks again for comments.

      • in our experience, the gtp loading state seems to vary… some gtpases are mostly active (GTP) and others are mostly inactive (GDP). I think because of that usually the CA/DN mutants can end up being more informative functionally. I wouuld guess that the wt would roughly phenocopy the DN if the gtpase is mostly GDP (i doubt much is empty).

  2. Hi David, I’d say the tubular network definitely does not look like ER! The major difference is in the angles of the junctions between tubules.

    • Thanks Viki,
      I agree – they just look wrong and anyway, why wouldn’t it spread more widely. My bet is on a sub-set of endosomes.
      Thanks for taking the time to take a look!

Comments are closed.