Amphiumas: gigantism, extended parental care and freaky morphology in a group of eel-like salamanders

There's a biologist called D. B. Means who's published on amphiumas: I bet he or she is tired of being the butt of constant jokes.

Unless I'm much mistaken, that would be Dr D. Bruce Means.

And speaking of names...

Auffengerg (1959)

Auffenberg, surely?

Thanks - corrections now made.

There actually was a documentary on Animal Planet regarding the Amphiumas which actually wasn't half bad. It didn't touch upon the single or multiple evolutions of gigantism aspect, but it did provide some good information on their biology and whatnot. Trouble is, I can't remember the name.

Amphiumid vertebrae are extremely distinctive and possess large, parallel crests on their dorsal surfaces and a prominent ventral keel.

Photo or it never happened.

They have internal gills even as adults.

Apparently they hatch with typical external gills, but have an extremely short larval period. Unfortunately, there don't seem to be any photos of Amphiuma (or for that matter, Cryptobranchus) larvae out there.

I've always found these guys interesting. Or at least weird looking. And now I've learned they're going to take over the world? Oh, no!

In that first photo, did the feature that looks like eyeglasses on the specimen develop in lieu of the eyelids it lacks? Also in that picture the forelimb looks pretty large with an especially large extended digit. And it appears to have five digits on the forelimb. Isn't that unusual?

Mike's brief comment (# 4) about the vertebral structures has just led me to realise that amphiumas have vertebral laminae. I'm looking right now at figures of amphiuma vertebrae, and on the lateral side of the centrum there are structures that - if seen in a saurischian dinosaur - would be identified as anterior and posterior centrodiapophyseal laminae. There are also structures on the dorsal surfaces of the postzygapophyses that look like laminae. None of this is unprecedented, as it's already been said in the saurischian literature that some plethodontids have laminae (Aneides, I think).

OT, but I just read that the carcass of a 12-foot Giant Oarfish was found on the coast of Sweden earlier this week. Apparently it hasn't been sighted since 1879.

You said that the occipital condyles occur on either side of the foramen magnum -- is this typical of salamanders in general, or just the Amphiumas? I had always learned that double occipital condyles were a mammalian characteristic (damn mammal-centric biology classes, I guess).

So amphiumid genomes are ~75 billion base pairs? That is absolutely enormous. Given that the human genome was recently found to be 95% transcribed (ENCODE project) despite only ~1-2% identified as protein coding regions, I wonder what the situation is for amphiumids. That is an awful lot of (putatively) useless mRNA to be floating around (although it's not all transcribed in the same cell), unless the amphiumid genome is not transcribed to the same extent, or unless the RNA is actually functional in some way; although of course, 3 billion bp of RNA is a lot too. The question is, even if most of the transcripts have some function, if the human genome needs 3.2 billion bp worth of extravagant transcription to work, how do amphiumids find use for 25 times as much?

They are reported to bite readily and nastily and are sometimes referred to as to the only amphibians within their range that can pose any sort of physical (as opposed to chemical) danger to humans...

True, too true. The only scar that I bear from a wound inflicted by an animal is the small crescent-shaped mark on one index finger - still nice and clear some thirty years after a little amphiuma got over-excited when I was feeding it. Blood all over the place, as I recall.

Forgot to add - I believe that sirens can give you a nasty nip as well. And one of my field assistants was bitten sufficiently hard to draw blood by her pet tiger salamander (another hand-feeding incident). Nothing to match the carapace-cracking ferocity of the amphiuma, though.

Amphiumas (together with Agkistrodons and Dendrobatidae) have always been of interest to me. When I finally do save enough money to return for graduate school, I would probably have to choose these as my preferred subject of study. As far as being obscure, I'm not entirely sure I would agree. Most of the crawfish (crayfish) farmers here in Louisiana are quite familiar with them. When I was working for a seafood company, we would frequently encounter them in our traps as they attempt to eat the trapped Procambarus. I don't recall them being particularly aggressive, though.

How does internal fertilization work for amphiumas? Do males have a phallus?

Hurray for Amphiumas and hurray for the shout out to Bruce Means!

Brad Moon's photo makes them look rather playful. The third photo seems indistinguishable from synthesized image; probably they would make excellent animated characters.

Sebastian: They don't need one, they are one.

Haha, walked right into that one.

I should have followed the links in the article closer though:

Breeding occurs in the spring, males possessing enlarged testes and swollen cloaca between January and April. Copulation among amphiuma has been observed only once by Baker et al. (1947) during July in Tennessee, and was described as an eight-day event where the male chose between two females and sperm transfer occured via cloacal apposition.

8 day event? Whoa...

The skull reminds me of the lysorophians (occiput excepted). So do the mud burrows...

They have internal gills even as adults.

Internal? I knew they retain an open gill slit, but are there really gills behind them!?!

Is there any analogue to the infamous postbranchial lamina on the shoulder girdle?

Paleoamphiuma (why did they go for this spelling, and not Palaeoamphiuma?)

Perhaps out of ignorance? Or perhaps it's a pun on "Eocene", just like how Paleothyris could be a pun on Eothyris (though the paper doesn't say any such thing, and R. L. Carroll isn't the joking type).

BTW, it should have been "Palaeamphiuma" -- Greek (and Latin) don't like too large vowel clusters. It should also, while I am at it, have been "Amphiumatidae".

A. antica

A. anticum, and should have been "antiquum".

Salamandroidea, Salamandriformes or Diadectosalamandroidei (I have yet to work out whether these names properly overlap, as the definitions given for each are somewhat different)

Diadectosalamandroidei, like almost all names by Frost et al. (2006), lacks a definition altogether. Only a vague "concept" is given. There's no way to tell if it's supposed to be a crown-group, a total group, or anything in between; I actually ranted about this in Appendix 6 of MarjanoviÄ & Laurin (2007). :-þ

As I said at the start of this article, there really isn't much on amphiumas in the literature at all, and most herpetology books only mention them in passing. Hopefully this little review can act as a useful synthesis.

It does!!!

And it appears to have five digits on the forelimb.

How can you see such a thing in such a tiny photo?

I had always learned that double occipital condyles were a mammalian characteristic

Mammalian, and lissamphibian; also found in some lepospondyls (some "microsaurs" and most "nectrideans") and in many temnospondyls (all stereospondyls for a start). Intermediate morphologies are found in other lepo- and temnospondyls.

In the 1860s, before those fossils and the fossils of close mammalian relatives were discovered, some people thought mammals were not derived from "reptiles" but directly from amphibians because of the double occipital condyle.

being sister grouped with Plethodontids,do they present direct development or they hatch as larvae...old fashioned ?

If I remember correctly, the current hypothesis on why amphibians have such enormous amounts of DNA is based on their being cold-blooded and not making any arrangements for incubating their eggs; the extra DNA includes proteins for just about every conceivable environment so that they'll always develop more or less normally. Birds, mammals and some reptiles keep a stable environment, and so require far less DNA to encode all the developmental proteins they need.

the occipital condyles (which are normally paired convexities on either side of the foramen magnum) extend posteriorly as short stalks, allowing the whole skull to be rotated much further dorsally (and ventrally) than is normally the case (a similar system evolved convergently in plagiosauroid temnospondyls)

And in uropeltid snakes, though it's usually discussed as part of their burrowing adaptation rather than affecting gape (thought of as relatively narrow). The condyle's single in snakes though.

The proportions of the mandible in the skull shown above are unlike any living snakes but similar to Eocene marine fossils Archaeophis turkmenicus and Palaeophis (the latter shown here.

I don't think anything much has been writtern on amphiuma vocalization. The CFZ once had a lage tree toed amphiuma that would periodicaly (about once per month)thrash about violently and neigh like a horse! The other sound it made was asort of high pitched scream like that of a woman. The latter was not unlike the noise made by some tortoises whilst mateing.

If I remember correctly, the current hypothesis on why amphibians have such enormous amounts of DNA is based on their being cold-blooded and not making any arrangements for incubating their eggs; the extra DNA includes proteins for just about every conceivable environment so that they'll always develop more or less normally. Birds, mammals and some reptiles keep a stable environment, and so require far less DNA to encode all the developmental proteins they need.

That doesn't work as an explanation, because most of the genome is trash. Over half of the human genome consists of rotting retrovirus corpses at all stages of decay, and most of the rest is repetitive sequences (runaway copy errors). And of the rest of the rest, well over half consists of pseudogenes (corpses of genes of our own at all stages of decay).

And besides, all amphiumas live in subtropical environments. It's not like huge genomes were restricted to things like the alpine salamander.

Salamanders in general, and amphiumas in particular, obviously have a lot more junk DNA than even we do.

Interestingly, the amount of DNA in a nucleus is correlated to the size of the cell. 2 of 3 Amphiuma species are unusually large. Have they got unusually large cells...?

being sister grouped with Plethodontids,do they present direct development or they hatch as larvae...old fashioned ?

The post says they have a larval period, but an unusually short one.

But then, their metamorphosis is incomplete, persisting gill slits and all.

Nice new banner, BTW.

That doesn't work as an explanation, because most of the genome is trash. Over half of the human genome consists of rotting retrovirus corpses at all stages of decay, and most of the rest is repetitive sequences (runaway copy errors). And of the rest of the rest, well over half consists of pseudogenes (corpses of genes of our own at all stages of decay).

I don't know if the previous poster is correct about genome size tendencies in amphibians and mammals, although birds I think typically have smaller genomes (which appears to correlate with cell size; Organ et al. 2007). However I especially object that you could then make a further supposition about the ecological/life history role genome size may play, when we don't have a handle on what most of the genome is doing, regardless of what function such activity may have, and what role size may play in this function.

But, to respond to the quoted paragraph, the bulk of the genome is indeed dead retroviruses and mobile elements, pseudogenes, and to a smaller extent repetive DNA (including Alu as a ME/retrovirus). I don't know if you were trying to address this point, but just because these aren't included in the tiny proportion of recognized protein-coding sequences (genes) doesn't mean they're trash- just because they may have originated as errors or junk or extraneous, doesn't mean they remained that way. They may have a regulatory role, including providing proper spacing in the genome or providing binding sites, they may be transcribed regulators, or they may even be translated and be unrecognized protein-coding genes. At the least, the point I was musing on is the fact that even if they ARE trash, nearly all of this 'trash' becomes RNA at some point. Weird.

Organ et al. 2007:
nature.com/nature/journal/v446/n7132/abs/nature05621.html
ENCODE Consortium:
nature.com/nature/journal/v447/n7146/full/nature05874.html

Nice new banner, BTW.

Oh yes, noticed the new logo as well. Ground hornbill, babirusa, and some sort of frog. Nice touch, really.

Also, just another point to no one in particular, while cell size (which appears to be a relatively weak correlation overall; perhaps more importantly, long-term effective population size; see Lynch & Conery 2003) may provide a restrictive or permissive environment for genome size expansion, that doesn't mean that cell size evolved to allow greater genome size/complexity.

Lynch & Conery 2003
sciencemag.org/cgi/content/abstract/sci;302/5649/1401

Richard (comment 23): are you serious? If yes, this is certainly the first I've heard of this (though I am aware that some air-breathing fishes - lungfish, for example - can make loud noises when inhaling, and mudpuppies and others are said to do similar things on occasion). Is anyone else aware of vocalising in amphiumas... some salamanders make squeaks or squeals when molested, and sirens are said to make a clicking noise (I reckon this is non-vocal and to do with their beaks), but I think that's about it.

Um, to clarify my own comment:

that doesn't mean that cell size evolved to allow greater genome size/complexity

because there is some advantage to larger genome size. There, I think that clarifies it.

birds I think typically have smaller genomes (which appears to correlate with cell size; Organ et al. 2007)

Yes. Smaller cells mean a higher surface/volume ratio for each cell, and that allows faster metabolism; this way, the amount of filler in the genome does end up being an adaptation.

we don't have a handle on what most of the genome is doing

We absolutely do have a handle on what most of the genome is doing: it's doing nothing. Decaying retrovirus corpses are useless, and microsatellites even more so.

You've got some 20,000 genes of your own, and 34,000 damaged retrovirus genes -- broken copies of gag, pol, and env over and over again.

just because they may have originated as errors or junk or extraneous, doesn't mean they remained that way. They may have a regulatory role, including providing proper spacing in the genome or providing binding sites, they may be transcribed regulators,

How would that work? How does this hypothesis pass the onion test?

or they may even be translated and be unrecognized protein-coding genes.

How would that work?

At the least, the point I was musing on is the fact that even if they ARE trash, nearly all of this 'trash' becomes RNA at some point. Weird.

Yes. Transcription is horribly inefficient. Our cells are awash in useless RNA.

Stupid Design.

Is anyone else aware of vocalising in amphiumas...

I think I read about it once, but there was even less information than here.

Yes. Smaller cells mean a higher surface/volume ratio for each cell, and that allows faster metabolism; this way, the amount of filler in the genome does end up being an adaptation.

As I tried to say, there isnt likely selective force for genome size expansion itself per se, even "to make room for more genes", all those uses arise after the fact; but anyways, I don't understand what you are saying in the above quote.

We absolutely do have a handle on what most of the genome is doing: it's doing nothing. Decaying retrovirus corpses are useless, and microsatellites even more so.

Actually, we don't. Read the refs.

You've got some 20,000 genes of your own, and 34,000 damaged retrovirus genes -- broken copies of gag, pol, and env over and over again.

Again, read the refs. Also, this science journalism piece may be germane:
nytimes.com/2008/11/11/science/11gene.html

or they may even be translated and be unrecognized protein-coding genes.

How would that work?

Current gene-recognition algorithms may have gene definitions that do not include all types of "genes" extant in the genome. In fact this is a hypothesis based on recent work involving proteomic data comparisons to genome sequence.

just because they may have originated as errors or junk or extraneous, doesn't mean they remained that way. They may have a regulatory role, including providing proper spacing in the genome or providing binding sites, they may be transcribed regulators,

How would that work? How does this hypothesis pass the onion test?

Read Lynch and Conery 2003 and also the ENCODE Consortium (2008) papers to start with, I would suggest.

At the least, the point I was musing on is the fact that even if they ARE trash, nearly all of this 'trash' becomes RNA at some point. Weird.

Yes. Transcription is horribly inefficient. Our cells are awash in useless RNA.

Stupid Design.

Well, I think the conclusions from recent work like the ENCODE project is that the human genome is transcribed to a previously unguessed-at extent. We don't know how much of this is alternative splice forms of genes, regulatory RNA, structural RNA (?), completely useless RNA, etc etc. You are papering over a lot of (interesting and important) biology just by saying "transcription is inefficient and cells make a lot off useless RNA". As I have said, the latter half of that statement may be inaccurate, to what extent is unknown.

Whoopsies, didn't close my blockquotes properly, but hopefully it is clear enough

In case it's not, this shouldn't be blockquoted:

Current gene-recognition algorithms may have gene definitions that do not include all types of "genes" extant in the genome. In fact this is a hypothesis based on recent work involving proteomic data comparisons to genome sequence.

For those who haven't clicked on the link, this is the "onion test", apparently:
"Whatever your proposed function, ask yourself this question: Can I explain why an onion needs about five times more non-coding DNA for this function than a human?"

To that I would say, that is basically my question. If indeed this newly discovered pervasive transcription indicates unknown functionality, how do genomes of such massive size difference find use for all that extra DNA?

That actually may be a good argument against "extra" DNA having any use at all, even if it's transcribed. But we don't know to what extent etc etc.

Of course, I am not saying that massive differences in genome size are explained by organisms needing it, just that they may use it once additional real estate appears. An example of this is gene duplication followed by subfunctionalization.

The CFZ's current amphiumas are both 3 toed, the one that (I'm told as it was before my time) made noises was a two toed species.

I've yet to hear them make a sound, however there are a few things about these fellows behaviour I can add that I don't think has been mentioned elsewhere.

Despite what most of the literature says, my experience with A. tridactylum is that they are quite docile and will only act aggressively on very rare occasions, their main means of defence is not to bite but to flick their bodies in an almost whip like manner. In late spring to early summer they may attack other members of their species in competition for space in captivity and presumably in the wild as well as it would follow there may be competition for the best pools for mating and feeding at this time.

As far as the legs go I've often seen them use these as if they were little hooks to hold their body in position as they start to twist themselves round things like wood.

For those who haven't clicked on the link, this is the "onion test", apparently:

"Whatever your proposed function, ask yourself this question: Can I explain why an onion needs about five times more non-coding DNA for this function than a human?"

This is only half of it!

Here's the other half, which follows only two paragraphs later:

"Further, if you think perhaps onions are somehow special, consider that members of the genus Allium range in genome size from 7 pg to 31.5 pg. So why can A. altyncolicum make do with one fifth as much regulation, structural maintenance, protection against mutagens, or [insert preferred universal function] as A. ursinum?"

Brackets and link in the original. Homo sapiens has 3.5 picograms of DNA as the haploid genome size, Allium cepa (the kitchen onion) has 17.

The post mentions that there are far smaller genomes than ours, and indeed Takifugu rubripes has only 390 Mb compared to our 3000. Yet, there are at least as many genes in it as in ours (as had to be expected -- first, there aren't any big differences in complexity among vertebrates; second, Takifugu is an actinopterygian, which means its ancestors have gone through a whole-genome duplication that our ancestors missed out on (for instance, they have about 7 sets of Hox genes, we have 4).

This means that if junk DNA serves a spacer function or something along these lines, almost none of it is needed.

I don't have access to Nature or Science, so I read the NYT article instead (it is, after all, by Carl Zimmer!). It mentions that 1.2 % of our genome are protein-coding, and 4 % in total show signs of having undergone selection. Evidently, all the recently discovered regulatory RNA is included in those 4 %, and so are all promoters, enhancers, silencers, terminators and so on.

The article further mentions that some of the viral stuff does have a function. Yes, for instance there's a viral promoter involved in pregnancy, and I've read of two or three similar cases. But such cases are so rare they only confirm the rule. How do I know they really are rare and there aren't lots more awaiting discovery? Because only 4 % of the human genome are under selection (and most of that appears to be regulatory RNA).

It follows from this that most RNA is useless and that transcription is horribly inefficient.

BTW, I don't understand why it was ever proposed that junk DNA could serve as a trap for mutations. There isn't a fixed number of DNA polymerase mistakes per genome, there's a fixed number per amount of DNA. Adding more junk won't decrease the number of mutations in the selected part of the genome at all -- and indeed, the fugu hasn't gone extinct.

So why is all the junk still there? Why haven't we lost most of it, like Takifugu has? Because it's much easier to gain than to get rid of. There are, of course, no enzymes that recognize useless DNA and cut it out. DNA only gets lost by random deletion (as a mistake during replication or repair), and there's simply no selection for that, except when there's strong selection for small cell size.

BTW, isn't each and every one of the retrovirus corpses preceded by a promoter? I bet many of these still work. Promoters don't need to have a very precise sequence to work; most mutations will only affect how active they are, but won't completely switch them off.

Many (or all?) integrated retrovirus genomes have a promoter at each end, and both work in the same direction, so that both the virus genome and the next gene are transcribed. That's how some viruses cause cancer. I suppose it's also a way to cause the transcription of junk DNA that isn't viral in origin.

For those who haven't clicked on the link, this is the "onion test", apparently:

"Whatever your proposed function, ask yourself this question: Can I explain why an onion needs about five times more non-coding DNA for this function than a human?"

This is only half of it!

Here's the other half, which follows only two paragraphs later:
"Further, if you think perhaps onions are somehow special, consider that members of the genus Allium range in genome size from 7 pg to 31.5 pg. So why can A. altyncolicum make do with one fifth as much regulation, structural maintenance, protection against mutagens, or [insert preferred universal function] as A. ursinum?"

Brackets and link in the original. Homo sapiens has 3.5 picograms of DNA as the haploid genome size, Allium cepa (the kitchen onion) has 17.

That does drive home the large variability in genome size even in closely related organisms, but it doesn't bring up anything I was arguing against, if you read my post.

This means that if junk DNA serves a spacer function or something along these lines, almost none of it is needed.

Again, I am not saying it is like a B vitamin for the genome. More like sequence evolution keeps introducing more DNA, and the genome's functional components work around this, and because so much of gene regulation depends on relative location, order, spacing, etc., whatever random new configuration is in place becomes important (essential?) to correct genome operation. That is one idea.

It mentions that 1.2 % of our genome are protein-coding, and 4 % in total show signs of having undergone selection. Evidently, all the recently discovered regulatory RNA is included in those 4 %, and so are all promoters, enhancers, silencers, terminators and so on.

Well, since we currently cannot identify a large proportion of peptides to genome sequences, since we do not have perfect ways of recognizing all the genome elements you listed, I don't think you can draw a line and say "all the rest is useless junk", even if most of it is, and close the book. I would also note that Ka/Ks ratio tests for selection, if that is what they used, are also not perfect, especially if sequences are under weak selection, although I'm not sure whether this is relevant. However the 4% selection number is a good point- any additional functionality (in the traditional sense) that we can find, it probably won't depart significantly from this number. Although the whole spacing thing may still come into play, etc.
ALso, it's one thing to say gene expression is "horribly inefficient" and perhaps 96% of the genome is 'junk', and quite another to count the ways it is amazingly inefficient. The genome is 95% transcribed? Perhaps most of those unknown peptides (the large majority, according to one paper I read) are useless proteins? Pardon me if I find wonder in the fact that evolution is so much more than just-so stories of natural selection.
Also, I think we don't have a good handle on how this potentially massive useless-ome interacts with the genome, transcriptome, proteome, etc., if it does so at all. That is sort of why I am wondering about what

BTW, I don't understand why it was ever proposed that junk DNA could serve as a trap for mutations. There isn't a fixed number of DNA polymerase mistakes per genome, there's a fixed number per amount of DNA. Adding more junk won't decrease the number of mutations in the selected part of the genome at all -- and indeed, the fugu hasn't gone extinct.

One example: suppose you have an intron (a type of "junk" DNA) emplaced in a gene. That intron must maintain the proper splice/excision sites, or else it potentially takes out the gene product. Plus, you must keep it clear of erroneous translation start/stop codons etc. Introns might also be target sites for yet more junk DNA like retroviruses and mobile elements. More chances for things to go wrong if you have more introns.
However, all of these marginal increases in genomic hazard are opposed by weak selection at best, and in higher animals and plants low effective population size means selection is too weak relative to drift to purge this noncoding DNA from the genome. Thus, the splice sites and so on are maintained, instead of just getting rid of the whole intron, if I understand the science correctly.

However, the drift/selection ratio as determined by long-term effective population size only determines central tendencies, and doesn't say necessarily why fugu in particular has an order of magnitude less DNA than humans. Perhaps there are additional enzymes/molecular biology processes in some animals that happened to evolve- a retroviral defense at one point?- just as we humans do not have telomerase, which serves a seemingly pretty important genome maintenance function in other organisms.

Cell size, again, may be a rare instance where genome size itself is selected upon in higher eukaryotes, although since it is only a correlation we don't know the direction of causation. Do organisms evolve small genomes by chance, which then allows smaller cell size? Do shrinking cells put downward pressure on genome size? If so, how is this mediated?

Still haven't figured out the blockquotes apparently. This should be blockquoted:
"Whatever your proposed function, ask yourself this question: Can I explain why an onion needs about five times more non-coding DNA for this function than a human?"

This is only half of it!

Here's the other half, which follows only two paragraphs later:
"Further, if you think perhaps onions are somehow special, consider that members of the genus Allium range in genome size from 7 pg to 31.5 pg. So why can A. altyncolicum make do with one fifth as much regulation, structural maintenance, protection against mutagens, or [insert preferred universal function] as A. ursinum?"

Brackets and link in the original. Homo sapiens has 3.5 picograms of DNA as the haploid genome size, Allium cepa (the kitchen onion) has 17.

its identification was doubted by Gardner (2003) and it might be a sirenid

I briefly wondered how on Earth a comparatively itzy amphibian could possibly be confused with a sirenian. There should be a law or something against reusing the same roots for distantly related groups.

Still haven't figured out the blockquotes apparently.

Whether the blockquote tag is terminated after the first quoted paragraph seems to depend on your browser, bizarrely enough. IE8 works, and Safari (at least for Mac) also works...

so much of gene regulation depends on relative location, order, spacing, etc.

Well, no. DNA can and does bend around protein complexes. Enhancers and silencers are often far away from the gene they control, by hundreds or thousands of base pairs, and moving them around changes little or nothing. Their number is much more important.

Well, since we currently cannot identify a large proportion of peptides to genome sequences,

You mean there are lots of peptides for which we don't know the genes???

since we do not have perfect ways of recognizing all the genome elements you listed, I don't think you can draw a line and say "all the rest is useless junk", even if most of it is, and close the book. I would also note that Ka/Ks ratio tests for selection, if that is what they used, are also not perfect, especially if sequences are under weak selection, although I'm not sure whether this is relevant.

Suppose the tests are off by a factor of four. That would still mean 84 % of the human genome are not under selection.

Introns might also be target sites for yet more junk DNA like retroviruses and mobile elements.

AFAIK all such elements insert almost perfectly at random...

just as we humans do not have telomerase, which serves a seemingly pretty important genome maintenance function in other organisms.

Oh nonono. We do have telomerase, and in our germline it's active. The trick is that switching telomerase off is a protection against cancer: by the time a single cell would have grown into a tumor, all its descendants are dead.

And indeed, every malignant tumor contains a mutation in telomerase regulation and has telomerase production and activity turned on.

Do organisms evolve small genomes by chance, which then allows smaller cell size? Do shrinking cells put downward pressure on genome size? If so, how is this mediated?

I'd say there's selection on faster metabolism. This puts those with smaller cells at an advantage, and smaller cells are caused by smaller genomes. Thus, selection for faster metabolism is selection for DNA deletions. At the same time, the usual selection against deletion of useful sequences stays upright, so you end up with a population with less junk DNA.

Well, no. DNA can and does bend around protein complexes. Enhancers and silencers are often far away from the gene they control, by hundreds or thousands of base pairs, and moving them around changes little or nothing. Their number is much more important.

Actually, spacing, which will strongly affect the torsion stress that the DNA is under during gene regulation, does matter. DNA torsion is operated on by a range of proteins including the well known chromatins, all of which play a key role in gene expression. Or, it simply may be that things must be the right distance apart for two different loci to be bound by the same complex. Spacing and location can change a lot in gene expression, right down to e.g. the structure of transcription initiation sites.

Let me just say that a good way to look at the genome is that it is neither precisely designed, nor is it perfectly chaotic. It is disordered, yes, but it is sort of like my workdesk at times: a mess, but an ordered mess where I know where everything is (more or less). Don't mess up those drifts! (Although that sort of trivializes the much much greater amount of information carried in the ordered disorder of the genome (including in its extant structure))

You mean there are lots of peptides for which we don't know the genes???

Yes, when looking at MS/MS spectra of peptides the vast majority are unidentified.

AFAIK all such elements insert almost perfectly at random...

Nope, that is definitely not true, some, in fact many are quite specific. But you shouldn't take my poor summary as the last word- there is a lot I am probably leaving out. A good macro overview is Lynch (2006) The origins of eukaryotic gene structure.

Suppose the tests are off by a factor of four. That would still mean 84 % of the human genome are not under selection.

If the test is not sensitive enough, then for all we know 100% of the genome could be under selection, strong and weak. But again one would have to look at what sort of algorithms were implemented to get that 4% number- they may have sophisticated models that are able to detect a range of selective signatures to arrive at that number.

Oh nonono. We do have telomerase, and in our germline it's active.

I wasn't sure about that when I wrote it, probably should have checked. But it was a poor example for a point that probably didn't need one. Imagine one, I guess.

I'd say there's selection on faster metabolism.

Hm, a good idea. Has anyone else thought of this?

smaller cells are caused by smaller genomes

as discussed in Kozlowski et al. 2003, pnas.org/content/100/24/14080.full (and elsewhere). Apparently genome size exerts a significant instantaneous control over cell size in higher eukaryotes, which in turn controls basal metabolic rate. In which case, the suggestion

Do organisms evolve small genomes by chance, which then allows smaller cell size?

[which then allows faster metabolism] would appear to be the winner.

Spacing and location can change a lot in gene expression, right down to e.g. the structure of transcription initiation sites.

That, yes -- but the location of enhancers and silencers matters a lot less, and how much space there is between genes doesn't seem to matter at all.

I remember reading about an experiment where hundreds of megabases of presumed junk DNA were removed from mice; the mice grew up normally and didn't show any abnormal phenotype.

Although that sort of trivializes the much much greater amount of information carried in the ordered disorder of the genome (including in its extant structure)

Like what?

Yes, when looking at MS/MS spectra of peptides the vast majority are unidentified.

And you're sure they can't result from alternative splicing of known genes? The Zimmer article you linked to says alternative splicing is a lot more common than used to be thought.

If the test is not sensitive enough, then for all we know 100% of the genome could be under selection, strong and weak.

Come on, that's special pleading. :-)

But it was a poor example for a point that probably didn't need one. Imagine one, I guess.

Ockham's Razor doesn't like it when I imagine things like that.

I'd say there's selection on faster metabolism.
Hm, a good idea. Has anyone else thought of this?

Of course. Look for publications by Chris Organ.

Apparently genome size exerts a significant instantaneous control over cell size

Yes. For example, extra-large cells in plants (like epidermal hairs) are polyploid. Arabidopsis has 32-ploid single-celled branched hairs on its leaves.

higher eukaryotes

What do you mean? I hope you don't mean "everything except yeast", which is the usage I've encountered...

I don't really have anything to add to the discussion, I just wanted to say how awesome it is to have one of my specimens featured in your article. TetZoo rocks!

-boneman_81

Thanks Kyler, I'm glad you're ok with me using your photo: very much appreciated!

I remember reading about an experiment where hundreds of megabases of presumed junk DNA were removed from mice; the mice grew up normally and didn't show any abnormal phenotype.

So you have an example. Here are some counterexamples.
neuromuscular.wustl.edu/mother/dnarep.htm

Like what?

Please see above.

And you're sure they can't result from alternative splicing of known genes? The Zimmer article you linked to says alternative splicing is a lot more common than used to be thought.

Well, that is an area of research. However, a recent study using the latest in peptide-to-genome mapping was only able to make a small dent in the number of unidentified peptides. Some were novel splice variants. Some appear to be novel peptides, which may be unrecognized by current annotation algorithms. By the way, if they are alternative splice variants, that generally means that intron regions- noncoding DNA- become parts of exons.

Come on, that's special pleading. :-)

No it's not, it's just making an observation.

Ockham's Razor doesn't like it when I imagine things like that

. It's called a hypothetical. For some solid examples of molecular biological innovation relating to periods of genome expansion, enumerated in this article are various cellular retroviral defenses, which have periodically been outmaneuvered, resulting in genome expansion: pnas.org/content/101/29/10496.full

Of course. Look for publications by Chris Organ.

I was being facetious. (BTW, Organ is the lead author of one of the refs I gave)

What do you mean? I hope you don't mean "everything except yeast", which is the usage I've encountered...

I mean plants and animals. A common operational definition of multicellularity and its various associated properties, although perhaps seaweeds should be included.