New Evidence for the Endosymbiotic Origin of the Centrosome

A new paper provides evidence that certain RNAs associate with centrosomes and may represent a centrosomal RNA genome. Furthermore this potential genome includes an enzyme that could copy the centrosomal associated RNA. But first some background.

i-84b415d1811ad9919a82464ec98ebe4a-MTarray.jpg
For many years, there have been scientists interested in the mysterious entity called the centrosome. Often refered to as the black hole of the eukaryotic cell, these intracellular structures (see red dots) are often located next to the nucleus (blue), where the microtubule cytoskeleton (green) converges (called the microtubule organizing center).

Another fact for you: the primary cilium, a long skinny cellular extension found in most cells, grows off of one of the centrosomes. This primary cilium is probably involved in some sensing mechanism. In fact in some cells (such as your rods and cones) the primary cilium is where most of the cellular sensory apparatus is located. Mutations that result in short primary cilliums cause kidney problems and probably involve a miss-sensing of some property of fuid within the loops of Henle.

Centrosome are composed of two centrioles, barrel like structures made mostly of tubulin + other more unique proteins (see image below the fold). The number of centrosomes per cell is tightly regulated and the duplication of centrosomes is quite bizarre. You start off with a single centrosome. At some point (S phase?) the two centrioles come apart and new (or "daughter") centrosomes seem to grow off of the old (or "mother") centrosomes. Very reminiscent of semi-conservative replication of DNA. But it get stranger.

i-025f1b51ca6e3e0f507830e8518f2d18-centrosome.jpg

If you watch the centrosomes throughout cell division, right before cytokinesis (the physical cleavage of two daughter cells) occurs, one of the two centrosomes from one of the two cells, migrates to the cleavage furrow and (in the words of Michel Bornens) gives it a "kiss". This event is thought to stimulate the final braking off of the two daughter cells.

Then about 5-6 years ago Alexey Khodjakov performed a remarkable experiment. He labeled the centrosomes with a green fluorescent protein (I think it was GFP-gamma-tubulin). After centrosome duplication he blasted one of the two centrosomes with a laser and got cells with lopsided mitotic spindles. The microtubules one side of the spindle converged on a centrosome while the microtubules on the other side, having no centrosome, became unorganized. Incredibly if he blasted both centrosomes, the cells could form a spindle (mostly due to the fact that motor dynein can organise the microtubules into a spindle without centrosomes). The cells then progressed through mitosis and even divided (although sometimes they were trapped in cytokinesis - see Bornens' kiss). After division they got a huge surprize, the cells stoped dividing further. They discovered that centrosomes are needed to initiate a new round of DNA synthesis (or the G1-S transition). So just before DNA is copied, a cell asks itself "is my centrosome okay?" If the answer is yes => proceed with DNA copying. If not => stop cell division. You can advance through this "check point" if you get rid of p53, a key protein that regulates the cell cycle.

But why all the fuss? What is so special about the centrosome? Why are cells checking it? Why is it kissing stuff? Why?

Michel Bornens came up with this individuation theory. Centrosomes mark an individual cell. At certain points, cells check up on their centrosome to check on their individualistic identity. If something is wrong, such as improper cell division, cells can sense their non-individual state and activate programed death and/or other drastic measures. Screwed up cells (such as cancer cells) have all sorts of centrosome defects ... many have an overabundance of centrosomes and centrosome associated proteins suck as pericentrin. i-5384c7bf1a8a6324e01ecae71e38efc8-spirochaete.jpg

And there's more! Many years ago, Lynn Margulis (famous for her theory of the endosymbiotic origin of mitochondria) proposed that centrosomes originated from the fusion of a spirochaete-like-organism (right, image of a spirochaete) and the prototypical eukaryotic cell. The spirochaete may have helped the host cell to move around (I never found this explanation very satisfying - the spirochaete must have offered some other benefit). In support of this idea, various publications since the 50s claimed that the centrosome contained its own nucleic acid, just like mitochondria, but these experiments were never definitive.

The new paper.

Now out of Bob Palazzo's lab (no relation ... see extra comment at the end) Mark Alliegro's Lab (click here for an explanation) comes a truly remarkable finding. His group purified centrosomes from clam oocytes ... for some reason clams have huge centrosomes and thus makes this experiment a whole lot easier. Along with the purified centrosomes came ... certain specific RNAs. (But wait, people have claimed this in the past, and one of the enriched RNA was 18s ribosomal RNA, so perhaps this mRNA was contamination?) What makes this finding believable is that only certain RNAs (five in this paper) were enriched. Besides the 18s rRNA, the other four RNAs were not found in any genomic database. (For you RNA enthusiasts these RNAs have now been called cnRNAs - centrosomal RNAs.)

And one of these cnRNAs is QUITE intriguing. It encodes a protein that has similarity to 'RNA-dependent nucleotide polymerase'. Furthermore, this RNA probably gets translated into protein. (Warning: technical jargon coming up) The group made antibodies against a peptide derived from the hypothetical protein, and found an immunoreactive band in lysates from oocytes and adult clam (end jargon).

So let me rephrase this. Clam centrosomes contain RNAs not found in any genome (so far), and one of these RNAs encodes an enzyme that could potentially copy RNA? Holly replication, Batman!

So could it be? Centrosomes have their own genome (an RNA genome at that?) And this genome encodes an enzyme that could potentially replicate an RNA genome? Wow!

Ref:
Mark C. Alliegro, Mary Anne Alliegro, and Robert E. Palazzo
Centrosome-associated RNA in surf clam oocytes
PNAS (2006) 103:9034-9038

(Note: This whole Palazzo/centrosome thing is very quirky. He knows of me, I know of him, but we've never met. Bob Palazzo has worked in the centrosome field for quite a while, I've published papers on a related structure, the microtubule organizing center. At meetings, some people think I have his long publication records, others who know Bob Palazzo well, ask me whether I'm his son -- he in fact does have a son named Alex. I almost went into the centrosome field by going to Steve Doxsey's lab (great guy!) but just couldn't live in Worcester. Now I'm studying (in part) RNA ... and Bob Palazzo is into cnRNA ... I know ... very strange.)

More like this

The story about about centrosomes possibly having their own RNA-based genome is one that I am watching rather closely. There are numerous apparent relics of the RNA world, such as the reverse transcriptase at the core of the telomerase complex, type I and type II self-splicing introns (ribozymes all), but if Palazzo is right, the RNA-genome would seem to be a fairly important piece of evidence. Of course, nowadays the view that such an RNA world existed is more or less mainstream. However, this world appears to have been fairly short-lived: evidently the replication initiators in all three kingdoms have a common ancestor - which would have existed in the DNA world, meaning that LUCA would have come after the RNA world.

For the articles on the homologies between the replication initiators, please see:

Structural basis for ATP-dependent DnaA assembly and replication-origin remodeling
Jan P Erzberger, Melissa L Mott & James M Berger
Nature Structural & Molecular Biology - 13, 676 - 683 (2006)

Nucleotide-dependent conformational changes in the DnaA-like core of the origin recognition complex
Megan G Clarey, Jan P Erzberger, Patricia Grob, Andres E Leschziner, James M Berger, Eva Nogales, & Michael Botchan
Nature Structural & Molecular Biology - 13, 684 - 690 (2006)

PS I consider your blog a real resource. Didn't care much for that soccer-thingy a while back, but my wife got a kick out of it.

Thanks for those refs.

The RNAi machinery is also looking very ancient. I haven't read the papers yet, but it would seem that Argonauts (the main components of many RNAi based gene silencing complexes) are conserved in eukaryotes + archea. It would appear that the original role may have been to regulate transcription rather than translation. I'll post something on it in the near future.