Pharyngula

Posted by: Kristine | March 17, 2006 3:10 PM

Back when I studied chemistry, I used to have nightmares about carbon and carbon chains. Literally. Creationists just don't understand how the unique structure of carbon allows it to readily combine with other elements.

Posted by: Caledonian | March 17, 2006 3:35 PM

To paraphrase Homer Simpson:

"Pfft, facts. You can use facts to prove anything that's even remotely true."

As long as certain important events are shrouded by ignorance and uncertainty, theists can pretend that the explanations they've been peddling for millennia are just as valid as rational speculation. Any attempt to shed light on those mysteries will be perceived as an attack and treated as such.

Posted by: rainbows4dinosaurs | March 17, 2006 4:18 PM

I thought the creationist idea of abiogenesis was "Dirt ---> Man".

That would be really funny if he wasn't hopping all over the country teaching that crap to children.

Posted by: wamba | March 17, 2006 5:31 PM

Is there any evidence that there ever was, or that it is even possible, to have a reverse translatase (reverse transcriptase but protein to RNA) complex.

Posted by: Theodore Price | March 17, 2006 6:03 PM

Kagehi, You're missing something. I'm talking about a very specific process (translation of RNA into protein) working in reverse. It is well known that DNA into RNA can be reversed into RNA into DNA with a trascriptase called reverse transcriptase. In terms of energy, I have no question that sufficient energy could be found somewhere, I am just wondering out-loud if the RNA-protein complex that would be needed to reverse translation could harness enough energy to actually achieve the reaction.

Wamba, I figured there was no evidence (kinda a silly question) but how about reasonable hypotheses on whether something of the sort could work or was ever involved in abiogenesis. I've been trying to find information on this for a long time and have never come up against anything other than exam question type google hits. Just figured this was big chance to ask a crowd that might know.

Posted by: wamba | March 17, 2006 6:50 PM

Wamba, I figured there was no evidence (kinda a silly question) but how about reasonable hypotheses on whether something of the sort could work or was ever involved in abiogenesis.

Energetics is not the problem; you would only have to do it once for each polypeptide, since you could then reproduce the result using standard polynucleotide replication technology.

To make it happen, you would need to build a separate capability to recognize each amino acid, either while it is in the polypeptide chain - or I suppose while you are taking it apart piece by piece. This would be more difficult than the forward direction of tRNAs and tRNA charging enzymes, since those can be done off-site and fit interchangably into the same ribosome. Polypeptide do not have the sort of base-pairing properties that polynucleotides do. The actual ligation of the RNA pieces together is enzymatically simple, it's all the specific recognition; and of course you would have to take care to match up the same genetic code in reverse as you did in the forward system. I cannot at present think of a way the code-matching could be done naturally.

Is that enough BS for one helping?

Posted by: Theodore Price | March 17, 2006 7:27 PM

I've actually managed to finally find some papers that support the hypothesis (albeit they are from the 70s and 80s). I did find one paper that shows evidence for the engineering of an rt tRNA (J theor Biol (2001) 209 181-187). I'm not qualified to comment on the validity of the claim, but it sounds reasonable. Found all this business in pubmed by searching (in quotations) reverse translation. Reading the theory papers from the 70s was great -- its wild to see in writing just how much we've learned in a short period of time.

As for getting the message right in reverse, maybe that could actually be an advantage. Assuming a diversity of codons for the same amino acid (at a prebiotic stage) you could generate an astonishing diversity of RNA codes simply.

Thanks for the helping of BS... I'm listening if you're willing to offer more.

Posted by: tacitus | March 18, 2006 1:37 AM

Then there's the Goo to You version.

I've heard "Goo to Zoo to You" before and, believe it or not, "who's ever heard of a walking, talking rock?"

Posted by: wamba | March 18, 2006 8:25 AM

And then there's this... http://www.miraribiosciences.com/peplica.phtml . Its from a biotech so of course there is no info but it sounds interesting (and if it really works I would use it in the lab -- what a tool!!!)

It can be done with modern machinery: run the protein through peptide sequencing equipment to derive the sequence, use a lookup table to find the corresponding RNA or DNA sequence, and synthesize it on a DNA synthesizer. This would take a lab bench full of equipment; I assume the original question was about accomplishing it at the molecular level.

Assuming a diversity of codons for the same amino acid (at a prebiotic stage) you could generate an astonishing diversity of RNA codes simply.

Posted by: wamba | March 18, 2006 10:00 AM

OK, I read the Nashimoto paper. I see numerous weaknesses in his proposed scheme for reverse translation, and no indication that the author is aware of those weaknesses.

• The author proposes that polypeptides once replicated themselves. There is no evidence for that, and immense practical difficulties to the possibility. Polypeptides lack any property analogous to the double helical base-pairing of polynucleotides, which is what allows polynucleotide replication to work.
  
• The author proposes that reverse translation used to occur. There is no evidence at all.
  
• The author suggests that the development of a useful suite of polypeptides through mutation of the RNA genetic material, translation and selection would have taken "tremendous time", and throws in a line about monkeys producing Hamlet. He blithely ignores the fact that the same criticism applies to his proposal that useful proteins came about through random thermal polymerization of amino acids, and he's lacking any replication method for polypeptides to ratchet him out of it.
  
• Figure 3, in which he's managed to get a couple pieces of RNA to swap a few bases doesn't prove much of anything. It doesn't address the most severe difficulties of his scheme.
  
• The author blithely ignores the independence of the forward and backward genetic codes, an obstacle I have already mentioned.
  
• The forward translation system smuggles in a lot of recognition specificity in that the amino acids are attached to the tRNAs by tRNA charging enzymes. This smuggling is possible because of the base-pairing properties of RNA. I can't come up with any way to acheive such smuggling in the reverse translation system. The author suggests rtRNAs which both recognize the peptide residue and carry the condon information; i.e. they perform the job that in the forward system is done by the tRNA AND the tRNA charging enzyme. This has serious obstacles.
  
• The author expects to have a working reverse translation system working by 2011. I admire but do not share his optimism.
  

Posted by: David Harmon | March 18, 2006 2:37 PM

Some more issues about "anti-transcriptases":

1) RNA and DNA are linear molecules, while proteins fold into all sorts of shapes. The endpoints may not be accessible, and folding can involve guided transformations, clipping off sections or assembling polypeptides from different sources. Information about those changes is not to be found in the sequence of amino acids. Likewise for metal ions, saccharide attachments, etc.

2) IIRC, some amino acids have multiple codings. That means the reverse mapping is ambiguous.

Posted by: Theodore Price | March 18, 2006 6:19 PM

I don't do TOP mass spec myself but I have tried to set up a collaboration to do it and no one seems to think it will work. Maybe I just haven't asked the right or willing person yet, but based on what I do know about i think the chances are slim (although I'd be willing to give it a go for a few months).

In terms of losing post-translational modifications, that would certainly be a downfall, but plenty of applications don't really require that level of information.

5 more years indeed...

Posted by: Torbjorn Larsson | March 18, 2006 7:34 PM

This is one (of many) reasons to hang around Pharyngula - one is often invited to see an area of knowledge sketched in a post, and inform ones naive views a little.

This post perhaps settled a naive discussion about why life has a common ancestor in another thread. If one searches wikipedia for hypercycle one is redirected to the quasispecies model ( http://en.wikipedia.org/wiki/Hypercycle ).

The theory (since modern RNA quasispecies are observed) chooses one quasispecies. It also seems to balance the mutation rates through pushing them close to the error threshold, is that correct?

It also both confirmed and refuted my naive view that abiogenesis seems simpler to look at as a process instead of looking at life specifics. The model of quasispecies specifics is process based. :-) It's always fun then the right choice is a third option.

I have put the real deal Genesis in my reading list too.

Posted by: Xiao Bao Clark | April 28, 2007 8:29 AM

I realise this thread is stupidly old, but (coming from an incredibly naive 1st year biomed student) couldn't the problem of the protein folding be over come by simple denaturation by high or low pH, temperature etc? If that's the case, then it's not hard to imagine a 'reverse-ribosome' binding onto one end and working it's way across, then using the same method reverse transcriptase uses to bring it back to DNA.
In response to the cutting out of parts, I believe this only occurs when the molecule is to be activated? So it would then produce inactive DNA. This is obviously going way out on a limb, but did somebody say 'introns'? :) I realise they get cut out before translation occurs, but what's to say that isn't a process that has undergone development to the point where they do cut out the unnecessary parts? Saying this, if the protein got to the backtranslation complex fresh out of production (ie before any part removals) then what you'd get is a functional DNA gene. It would definately help to explain why the majority of DNA doesn't appear to code for anything.
As for the codon triplet specificity: I could be very very wrong here, but do certain codons appear for an amino acid more than other ones? If this is the case, it could well be that those common ones were the 'ancestor' codons, the original ones the system made knowing(I use this term loosely) full well that DNA replication can fuck up alot, and reducing the effects of base pair substitution(mutation) significantly.

I'm really hoping someone smart will see this and thoroughly discredit everything I've said here. Please inform me :) Thanks.

Posted by: David O'Hagan | July 30, 2007 6:45 PM

I have talked with Mark Martin and Nashimoto about their experimental evidence for reverse translation... and they have none. I tried my own approach to the pRT chemistry using RNA aptamers to recognize the N-termini of peptides... and to sort them based on the N-teminal residue. After sorting I was going to tag them with an artificial codon, then use edmund degradation to remove the n-termini and repeat the process. It was so far out there that I could only find temporary funding for the project. On my Myspace profile I have published the grant application that describes the mechanism and preliminary findings. I believe that with further work it might be possible... but the preliminary results suggest that the specific sorting of peptides based on the recognition of the last three N-terminal residues with RNA aptamer pools is not specific enough for a complex mixture. I tried to affinity purify those peptides with an RNA aptamer population that was selected against peptides of the sequence L-X-X, and this population of RNA aptamers selected the L-X-X-L-X peptides from a random population of peptides (X-X-X-X-X). If anyone else is working on this... I would like to here from them. Dave