4-Billion-Year-Old Fossil Protein Resurrected

First time accepted submitter Zoë Mintz writes "Researchers have 'resurrected' a 4-billion-year-old Precambrian protein and found they resembled those that existed when life began, proving that protein structures have the ability to remain constant over extended periods of time."

Comments

Re:Alive

[GigaplexNZ]

Nowhere in the title, summary or article was the word "alive" mentioned.

Sorta

[dnadoc]

They took present-day versions of the protein in living organisms, used a computer to interpolate a hypothetical common ancestor, then 'found' sequence homology - but people already knew the sequence was highly conserved, it's evident in modern organisms. There were no "fossils" involved. And conserved sequences make for poor molecular clocks, so who knows if it was 4 billion years ago.

Re:Sorta

[Samantha Wright]

The key is that because thioredoxins are found in all of the kingdoms, and are so conserved, the authors are assuming this is what the thioredoxin in the LUCA looked like. Even if the molecular clock isn't accurate over this one protein because of masked mutations, the number's most likely sound. (To one significant figure, anyway, since the LUCA is held to be 3.5 to 3.8 gya.)

Re:Sorta

[c0lo]

Better source

A team of scientists from Columbia University, Georgia Institute of Technology and the University of Granada in Spain have successfully reconstructed active enzymes from four-billion-year-old extinct organisms.
[...]
In their study, published in the journal Nature Structural & Molecular Biology, the researchers used vast amounts of genetic data to computationally reconstruct the genes of extinct species, a technique known as ancestral sequence reconstruction. The researchers then went a step further and synthesized the proteins encoded by these genes. They focused their efforts on a specific protein, thioredoxin, which is a vital enzyme found in all living cells.

Re:Sorta

[bzipitidoo]

(LUCA = Last Universal Common Ancestor.)

I read a lot of conflicting info about the early Earth. Is the end of the Hadean Eon and beginning of the Archean supposed to be when life began? Or is some other event supposed to divide the two eons, like perhaps the emergence of conditions hospitable to life? I've read that it is 3.8 gya or 4.0 gya. Why not say 3.9+/-0.1 gya? Obviously 4 was picked for being a very round number, but settling on a single number however round seems a bad idea. Makes it sound like we're more certain of those dates than we really are.

Much of our knowledge is sketchy and speculative. No one really says whether the first life forms might be considered bacteria, or archaea. The archaean domain is still new to science. Was only in the late 1970s that archaea were recognized as being different enough to qualify as a separate domain and not part of the domain of bacteria. Then there are fun ideas like the RNA world hypothesis. There's the idea that life could have started and died several times before achieving permanence. Panspermia is another notion.

Re:Sorta

[Samantha Wright]

The LUCA dates range from 3.5 to 4 gya, so it's even broader than that. Different estimates come from different sources and with different precision, though, so it's not quite right to give a single symmetrical error measurement. I'd personally vote for saying 3.8 +0.2/-0.3 gya. In the case of this article, however, they chose 4 exactly because of their molecular clock predictions.

The article doesn't clarify between the Archean and Hadean periods, however, and it's probably bad to equate the LUCA with the beginning of life because we have pretty strong evidence that the LUCA was already a very well-developed organism, with a complete central dogma, hundreds of enzymes, and a preference for potassium ions over sodium ones. Wikipedia cites several science journalism pieces that argue for a Palaeoarchaean LUCA.

As for what the LUCA actually looked like, I would say somewhere between Archaea and Bacteria, but defying both categories. Archaeans have a number of later innovations that definitely disqualify them from being good representatives, since they can do sophisticated chromatin modelling (folding DNA to make gene transcription more efficient) and have a unique membrane composition (which I personally like to imagine may be evidence of multiple abiogenesis events, but that's a bit of an uninformed theory.) Bacteria, on the other hand, are known to have a tendency towards simplifying their genomes. If anything the bias seems to be toward Bacteria as the root; no one has recently proposed that Archaeans pre-date Bacteria.

Re:Sorta

[ColdWetDog]

Mitochondria are thought to be a (relatively late) Major Event - the genetic and biosynthetic pathways were already fairly advanced.

The major idea behind a LUCA is that you can 'work the clock back' with DNA sequencing to something around 4 billion years ago AND that all life-as-we-know-it relies on a fairly specific set of chemical reactions. It's pretty clear that modern organisms are a mismash of A+C+B+x - nature hates to throw genetic material away. It's also perfectly reasonable to assume that life started / stopped multiple times and that bits of earlier life were indeed incorporated into later critters. When you start talking about that, you get into some terribly annoying semantic arguments (perfect for Slashdot).

The Holy Grail would evidence of organisms using wildly different chemistry (like incorporating arsenic into the DNA backbone instead of Phosphorus) or some molecule that transferred genetic information without DNA or RNA entirely.

This is one strong reason why we should get our respective asses towards Mars. It offers the closest laboratory for finding off-world life. What that looks like (if it exists) is going to be one of the biggest scientific discoveries ever.

So, for life on earth, it's pretty clear that there is one LUCA - something started us along the pathway to RNA and DNA based life (I'm personally a proponent of RNA World). And TFA implies that that this enzyme was present fairly early on. But something further back had to set the stage for the ability of an organism to fold amino acids into proteins, etc. TFA doesn't even begin to discuss what the chemistry would have been - that''s another story - but by the time Thioredox was on the scene, something was making RNA and there was something that looked like a ribosome - pretty advanced functions. How they got there is pretty much hand waving.

Re:Alive

[Black Parrot]

It is implied by the usage of resurrected.

Actually, we use "resurrected" for lots of non-living things, e.g. a plan.

Re:Alive

[Anonymous Coward]

No it isn't. I'm a scientist and we use the word resurrected too. We are talking about molecular resurrection, not whole organism resurrection. The scientific community re-purposes common words to mean different specific scientific things all the time. We use the word resurrection to mean that we made an ancient protein in the lab, and the protein still has it's original function. There is a big difference between just making a predicted ancestral protein in the lab, and having it actually work the same way it used to. The protein needs to fold correctly and be in the correct environment.

For more details of the previous use of this word, google "Ribosomal Paleontology and Resurrection".

Re:"resembled those that existed when life began"

[MightyMartian]

I can't sort out why. The statistical tools they used seem little different from how the proto-languages of major language families are reconstructed. In both cases you look at the genomic unit (in molecular biology that is genes and the proteins they encode, in comparative linguistics it's words, or more specifically cognates). In either case you cannot state with absolute certainty that the proposed progenitor unit (gene or cognate) has been reconstructed absolutely, but you can say with a reasonably high probability that you're pretty close.

Re:"resembled those that existed when life began"

[Samantha Wright]

With protein sequence evolution it's a little more controlled: the modifications occur more-or-less randomly, and there are almost no cases where a letter (residue) is replaced throughout the entire vocabulary (proteome) due to phonological shifts. As a result, if you have enough datapoints to work from, like with the thioredoxins, it's simply a matter of picking the version most commonly agreed upon by all of the branches. In that sense, it's more like textual criticism than historical linguistics, particularly since you can also use the requirement of "it has to be well-formed language" (i.e. a working protein) to weed out obviously bad combinations of changes.

For some reason, that bewilders a lot of reasonably scientifically-minded people.

Re:Does anyone, and i mean ANYONE, question the ag

[Samantha Wright]

It's faintly possible that an absolutely essential component of cellular function suddenly worked its way into the genomes of every single organism on Earth one Tuesday afternoon, and that despite every indication of all copies being descendant from a single master source, they were simply made to look that way after the fact, and that the last universal common ancestor got along just merrily without it, despite it being much more logical that this one particular protein happened to be there alongside all the other ancient essential proteins we know and cherish... but that would require an incredibly petty and childish divine being, or one with terrible planning skills. Possibly the divine being that buries dinosaur bones to test the faith of His followers.

So, no; not really. Why do you ask?

Re:"resembled those that existed when life began"

[MightyMartian]

I didn't mean to say they were identical, but still, at least in the Indo-European languages (and I'm sure it can be found in other language families) there are some pretty highly conserved cognates, like pHtér (father). In most cases throughout the various Indo-European families one can trace pretty predictable sound changes to explain why pHtér became pater in the Italic languages. pitár in Sanskrit and father/fadar in the Germanic languages. Yes, there's a good deal more horizontal transfer in languages, and indeed in some cases words will disappear from some members of the family, but in general, the core vocabulary of the proto-language is pretty highly conserved in its descendants. Even in English, with its vast importation from the Norman invasion onward of Romance and Greek words, the core vocabularly remains Germanic, and the sound shifts from the Proto-Germanic thru West Germanic thru to Modern English tend to follow regular rules. It's actually kind of cool, because even where you have a word that was adopted from another language, you can usually determine when roughly it happened by the way in which it was or was not effected by the sound changes going at the time.

Re:"resembled those that existed when life began"

[Samantha Wright]

Save your cynicism for other headlines. The amount of potential variability in this situation is genuinely minuscule, and it is inappropriate to worry about the predicted structure being 'out and out "wrong".' Claiming knowledge of the original structure is not that much of an exaggeration; the researchers demonstrated that none of the differences in the various versions of thioredoxin had a significant impact on its shape.

Re:"resembled those that existed when life began"

[Samantha Wright]

...sorry, I've read the article a little more closely and I made a couple of factual errors in my other reply to you and the one before it. There were small structural changes, and the primary purpose of this paper was to investigate ways of detecting them. Convergent evolution (every copy changes at once after a split) does occur in protein structure and sequences, primarily due to large-scale environmental changes.

The paper's primary contribution is that they stepped back gradually, rather than doing a bulk sequence alignment (what they called a "vertical" approach rather than a "horizontal" one) and found that to maintain function, certain shifts had to occur. (The details of which are rather boring.) Rather encouragingly, they found that, by the time they'd stepped back all the way to the beginning, the changes the protein experienced meant that it would perform optimally in a chemical environment much like the one archaeology has shown us was ubiquitous in the Precambrian era. Not only does this support the idea that their results are very close to being correct, it also tells us that the LUCA probably had a fair amount of time to evolve its thioredoxin to that environment.

Re:"resembled those that existed when life began"

[khallow]

For some reason, that bewilders a lot of reasonably scientifically-minded people.

Just because you can come up with one or more best fit versions from what currently survives, doesn't mean that they resemble the original source of the evolutionary pattern.

Re:"resembled those that existed when life began"

[Samantha Wright]

To be fair, the IB Times article doesn't do a very good job of explaining the lengths the researchers went to in order to avoid that. Here's a summary:

• - The researchers were acutely aware of the exact problem you described, and spent several paragraphs on it in their background section.
• - Their way of getting around it was to ensure consistency at various time-steps, which revealed a gradual change in the shape of the protein overall that wouldn't have been evident if they'd just used a multiple sequence alignment and gone with the maximum parsimony route.
• - The models they ended up generating are extremely high-precision and based on a ton of data.
• - The final results perform optimally in an environment similar to ones from 4 gya according to archaeological evidence. They didn't try to force that, which means they must've gotten it very close to right.

In the future, here's how to read scientific news stories (at least molecular biology ones):

• - If the article was published in Nature or Science, it's a really big deal. Any shortcomings you can see are probably failures on the part of the journalist. I think PNAS is in this category, too, but as a bioinformatician I don't know quite enough to comment.
• - If the article was published in Cell, it's also almost certainly really serious, although you should note that the authors have no scruples because they're publishing in an Elsevier journal.
• - If the journal name sounds like it's just a description of the field with no organization attached ("Journal of Microbiology") then it's probably from a fake journal. (But not necessarily. Tread with caution. In particular, one-word journal titles don't fall in this category.) Any exciting claims made by the journalist are probably exaggerations.
• - If the article came from PLOS ONE (and only that journal; there are lots of good PLOS journals) it was terrible or boring science and couldn't cut it anywhere else. Anything exciting in PLOS ONE is probably a typo.
• - Any other journal (e.g. Oxford Bioinformatics) is fairly reputable but not infallible. It's unlikely that the journalist understands more than half the article.
• - If the source is a university press release, it's complete crap. Unless you're a potential donor, in which case you should eat it up.

Re: Alive

[Anonymous Coward]

Ever heard of weather being hellish, or the flight being a torture, a meal being an orgasm in the mouth, a person being an asspain or buttmad? Lauguage is strange thing, and as they say: time flies like an arrow, fruit flies like a banana.

Re:Alive

[fellip_nectar]

Our chief weapon is surprise...surprise and fear...fear and surprise.... Our two weapons are fear and surprise...and ruthless efficiency.... Our *three* weapons are fear, surprise, and ruthless efficiency...and an almost fanatical devotion to the Pope.... Our *four*...no... *Amongst* our weapons.... Amongst our weaponry...are such elements as fear, surprise.... I'll come in again.

Re:Alive

[Anonymous Coward]

They traced back the mutations of every thioredoxin variation to a common ancestor 4bn years ago. If you have three close species: A, B, C. The three share a variation of a protein which is exactly the same at nucleotid level except for one site, lets say: A: CGCGTA, B: CGTGTA, C: CGCGTA. You know, because of the rest of the genome, that A and B had a common ancestor 2 million years ago, and that common ancestor had a common ancestor with C 3 million years ago. Chances are that the original protein was CGCGTA. In this case, the reconstructed protein is the same as the A and C proteins, but given enough species you can use this kind of reconstruction techniques to figure out how the ancient version of a specific protein looked like.