The Bacterium That Doesn't Know How To Tie Its Own Shoelaces

by Elio

A hackberry petiole gall psyllid just after
emerging from the gall on a hackberry. The
leaf is yellow because it is autumn. The
insect is only 3 to 4 mm long. Credit: Nancy
Moran, The University of Arizona. Source.

A greater microbial wonderland there isn't than the one inside cells of insects. Many, perhaps most, insects harbor bacterial endosymbionts that are full of physiological, genomic, and structural surprises. To confirm our predilection for the subject, go to Search on the right of this page and enter “endosymbionts.” Or read recent reviews (click here or here).

In many cases, bacterial endosymbionts are known to provide their host with needed nutrients, in others their role remains unknown. The classic example is Buchnera, the endosymbiont of aphids. Aphids feed on plant sap, which is rich in sugars but poor in needed amino acids. The bacterium provides them, and thus is essential to its host. Matters can get complicated, as when two endosymbionts partner to synthesize tryptophan, one making an early precursor in the biosynthetic pathway, the other taking it the rest of the way. Partnerships of this sort (actually, ménages á trois) seem not to be unusual. This too we discussed earlier. Other bacterial endosymbionts, notably Wolbachia, manipulate the sex life of their insect hosts for their own advantage. Thus, the world of bacterial endosymbionts of insects has turned out to be ever more varied and strange.

Enter Carsonella ruddii, or, if you’re a purist, Candidatus C. ruddii (because it hasn’t been cultivated outside its host—nor is it ever likely to be).

It has the puniest of genomes among entities that could be called bacteria. With a measly 160 kb or so of DNA, encoding some 182 proteins, it is the smallest “bacterial” genome published to date (but we’ve heard that even smaller ones are soon to be reported). The host of Carsonella is a psyllid, in particular the hackberry petiole gall psyllid. Psyllids are also known as the “jumping plant lice.”

A psyllid bacteriocyte with tubular
Carsonella surrounding the nucleus.
DNA is stained with DAPI. Source.

Is such a tiny genomic endowment sufficient to sustain life? Among the missing genes are some mightily essential-sounding ones, e.g., some needed for making ribosomes, for DNA replication, and for membrane functions. Can such a creature really be called a bacterium? Is Carsonella on the way to becoming an organelle? If so, what good is it to the host? Genes required for making several of the amino acids needed by the host (e.g., histidine, phenylalanine, and tryptophan) are absent or nonfunctional. On the other hand, if it doesn't pay for its upkeep in some manner, how come host cells are loaded with so many carsonellas?

It’s true that the Carsonella genome is extraordinarily streamlined, ergo efficient. For instance, 90% of its ORFs overlap in at least one direction. Ninety-two percent of them are tandem out-of-frame overlaps on the same strand. Carsonella ORFs are close to 20% shorter than those of other intracellular insect endosymbionts. But, efficient or not, the Carsonella genome is smaller than that of some mitochondria and chloroplasts, both of which require the products of many genes that they previously transferred to the host’s nucleus. Is this the case here? We don't know yet because no hosts of Carsonella have been completely sequenced. So, how does Carsonella get by with what is left in its genome? It hardly seems to be large enough to sustain life as we know it.

What are the other possibilities? Rather than getting required gene products from Carsonella genes in the nucleus, could some instead be provided (by a kind of prokaryotic empathy) by “nuclear” mitochondrial genes? Such proteins would have to be targeted to both the mitochondria and the symbionts. Another possibility is that there could be another bacterial endosymbiont lurking somewhere, but as yet none have been found. Or, as seems even more unlikely, could our catalog of required genes be in part erroneous, with the genes present in Carsonella actually being sufficient for life?

We have some information about gene transfer from endosymbionts to Buchnera-carrying aphids. However, the meaning of this is murky because these genes are derived from α-proteobacteria, whereas the Buchnera are γ-proteobacteria.

Transmission electron micrograph of a psyllid bacteriocyte.
A, bacteriocyte; B, endosymbiont; C, unidentified electron-
dense aggregate. Bar, 2 μm. Source.

Carsonella presents further enigmas on the cellular level. Its cells are extremely long tubules, many tens of μm in length and around 5 μm in width. They don't look like regular prokaryotes in EM sections because, among other things, there is no sign of a nucleoid. (Is their DNA content so small as to escape visualization?) They look like nearly uniform bags of ribosomes, except for some mysterious electron-dense inclusions. In fact, the creatures don’t look like anything, bacterial or otherwise, that we can recall. Let’s expect that further attention will be paid to the structure and morphology of these unusual creatures.

If these questions were not enough, let us add one more: Are such endosymbionts destined for further reduction, possibly with catastrophic consequences to themselves and, who knows, maybe even their hosts?

Here's a question for the astute among you (all our readers are astute by definition): What would it take to answer all those earlier questions? Which –omics would you go for first? More genomics of hosts, proteomics of the symbiont, transcriptomics of both? If we were grant-granting agencies, we’d be hard put to know which to fund first.

Comments

[Nathan Myers]

How does C. benefit its host? You might ask, what value does a blackmailer provide his or her victim? "That's a nice ribosome you've got there; it'd be a shame if anything should happen to it."

I haven't seen cases of intracellular blackmail identified. This might just be because I'm pig-ignorant. Or, is it a case of the dog not barking? A blackmailer never actually does anything, normally. Maybe there's a gene in C. that's never seen expressed.

[Robert G. E. Murray]

When is a bacterium not a bacterium? A first order response might be when it can no longer be recognised by structure and chemistry that it is bacterial in nature. Most (?all) cases involve existence in unrelated living cells and you rightly ask the question of exclusion: When does this pared down bacterial cell become an organelle? Presumably this state arises when it can go and grow nowhere else but an unrelated cytosol and has lost the essentials for being a fully functional bacterium. This one seems more than half-way to being an organelle!

[Mark O. Martin]

Great post, Elio! One of my favorite topics....

What kind of "omics" would the best approach? I suspect transcripto-somics, for convenience---the ability to easily remove/exclude eukaryotic mRNA from the equation is attractive. But "all of the above" may well be the approach used. Oh, and genomics of the symbiont, provided it can be isolated exclusively---this example you cite has such unusual morphology that I would worry a bit.

The whole issue of establishment of symbioses over time is fascinating. We generally study symbiotic associations with minimal integration (say, the Vibrio fischeri - Euprymna scolopes symbiosis) or extremely extensive integration (your own example above, well on the way to become faux-mitochondria).

KW Jeon of the University of Tennessee has long studied a phenomenon that happened a few decades ago in his lab: while investigating amoebae, Dr. Jeon had a "bacterial infection" of his stock cultures. The survivors seemed to integrate Legionella like bacteria into their cytoplasm, and have now become obligate. Here is a recent review:

Jeon KW. (2004). "Genetic and physiological interactions in the amoeba-bacteria symbiosis." J Eukaryot Microbiol. 51:502-508.

Abstract: Amoebae of the xD strain of Amoeba proteus that arose from the D strain by spontaneous infection of Legionella-like X-bacteria are now dependent on their symbionts for survival. Each xD amoeba contains about 42,000 symbionts within symbiosomes, and established xD amoebae die if their symbionts are removed. Thus, harmful infective bacteria changed into necessary cell components. As a result of harboring X-bacteria. xD amoebae exhibit various physiological and genetic characteristics that are different from those of symbiont-free D amoebae. One of the recent findings is that bacterial symbionts control the expression of a host's house-keeping gene. Thus, the expression of the normal amoeba sams gene (sams1) encoding one form of S-adenosylmethionine synthetase is switched to that of sams2 by endosymbiotic X-bacteria. Possible mechanisms for the switching of sams genes brought about by endosymbionts and its significance are discussed.

Elio, I wonder if Dr. Jeon would be interested in an essay for STC?

I can't help but wonder if this sort of thing could be "set up" and allowed to happen again in the lab (unless this case was purely luck!). In that case, data for the "pre-association" and "post-association" forms could be studied over time, as Rich Lenski has done with bacterial evolution in the laboratory, freezing "snapshots" of cultures at various time points in the process.

This could be relevant to far more than the study of symbioses, as many pathogens need to make similar genetic and biochemical adaptations. But what granting agency would support that kind of "blue sky" project?

Again, a great topic.