Here is one reason I don’t think life as we know it is the result of ordinary processes.
From Wiki:
https://en.wikipedia.org/wiki/Stop_codon
In the genetic code, a stop codon (or termination codon) is a nucleotide triplet within messenger RNA that signals a termination of translation.[1] Proteins are based on polypeptides, which are unique sequences of amino acids. Most codons in messenger RNA (from DNA) correspond to the addition of an amino acid to a growing polypeptide chain, which may ultimately become a protein. Stop codons signal the termination of this process by binding release factors, which cause the ribosomal subunits to disassociate, releasing the amino acid chain. While start codons need nearby sequences or initiation factors to start translation, a stop codon alone is sufficient to initiate termination.
Now what happens when there is no stop codon?
A nonstop mutation is a point mutation that occurs within a stop codon. Nonstop mutations cause the continued translation of an mRNA strand into an untranslated region. Most polypeptides resulting from a gene with a nonstop mutation are nonfunctional due to their extreme length.
….
Nonstop mutations have been linked with several congenital diseases including congenital adrenal hyperplasia,[15] variable anterior segment dysgenesis,[16] and mitochondrial neurogastrointestinal encephalomyopathy.
In other words, it would be bad juju if there are no means of reading of DNA and recognizing where one gene ends and the other begins. In fact, without stop codons, it looks like a DNA-RNA-protein-based life on Earth would be dead.
One could postulate a DNA-RNA-protein-based life that had an alternate stopping mechanism that eventually evolved a stop codon. But that just moves the problem elsewhere in as much as a DNA translation system that contains multiple genes needs a gene delimiting mechanism. A stopping mechanism needs proteins to implement it, but without a stopping mechanism to implement proteins, there is no stopping mechanism. We have then a chicken and egg paradox.
One could postulate proteins arose by a method outside of DNA translation and somehow recruited DNAs and RNAs and then defied all probability and somehow figured out how to code the next generation of proteins using DNAs that just happen to be coding the proteins like the ones that miraculously recruited them. At some point, such scenarios are so out of the ordinary they are not distinguishable from miracles.
Some will argue Darwinian evolution in the origin of life. That’s problematic for at least two reasons.
1. even most evolutionists don’t view the origin of life and origin of the protein translation cycle as part of typical evolutionary theory. So in that sense of the word “evolution”, stop codons didn’t evolve.
2. something dead can’t evolve by Darwinian mechanisms, and if this is an origin of life scenario we’re dealing with dead pools of chemicals.
So Darwinian evolution isn’t a solution. Chemistry isn’t a solution. Physical laws aren’t a solution. Probability isn’t a solution. In fact what we know of physical process would work against evolution of stop codons, not for it. Hence stop codons and life that depend on stop codons did not arise out of ordinary physical processes.
“POOF” Is your claim however. God did it all at once, like magic. So why criticize another’s “POOF” (as you see it) when you actually believe exactly the same (as you see it).
I see what you mean, I don’t really see what the next step is supposed to be that you could somehow “couple” to the synthesis of amino acids in this way which would make protein synthesis exergonic. Presumably the authors have not thought about that either or they would have mentioned something about it.
I should be noted, though, that in most experiments that produce amino acids, some of them polymerize into small (though heterochical) peptides some times up to 6 or 8 amino acids in length.
I think much of the “issue” seen with peptides first is misplaced, because it is often the case that proponents of peptides-first scenarios are confused with protein-first proponents (which I agree is absurd) and further, it is also mistakenly thought that the peptides produced non-translationally need to be sequence-specific and homochiral (and big), to be functional and/or replicated.
This is of course a very speculative area, but there have been some interesting publications on the subject of what kinds of functions short peptides could have before coding, and how they would be made and “pseudo-replicated”. See for example:
Functional Capabilities of the Earliest Peptides and the Emergence of Life.
Predicting the conformations of peptides and proteins in early evolution.
And I recently ran into this (very speculative) but interesting attempt and trying to explain a coevolutionary (RNA+amino acids/peptides) model of the translation system from the bottom-up:
How Amino Acids and Peptides Shaped the RNA World.
I should be noted… = It should be noted…
I’m not that noteworthy 😛
Rumraket,
Yep, interesting. The noncatalytic, and sequence independent catalytic, roles of peptides are certainly worth a look. I do feel, though, that biochemistry almost predisposes people to try and find a role for peptides: they are such ubiquitous actors today. But I’m not sure why peptides as opposed to (say) metal ions would hold the key. After all, catalysis is often performed by metals now, deep in the reaction centre.
Transition and replication seem to me to be the main stumbling blocks. I may be exercising undue haste in hoping to get replication going asap, but it unlocks the power of NS, and gives a unit in whose service all subsequent tuning is done.
I would disagree with this (from the 1st link)
The clincher for me is the intimate involvement of RNA monomers in biochemical pathways. It seems one hell of a coincidence that the energetic moiety of choice has a planar structure sticking out of it at just the right angle for stacking, with just the right edge groups for complementary hydrogen bonding in such a pair of stacks oriented antiparallel. One can of course hypothesise different cofactors in a precursor metabolism which subsequently adapted to RNA monomers when they became available, but we do need to get homochiral cyclic ribose and complementary bases going at some point. I find it persuasive that this may need to happen first, before replication (and hence, before what I would call Life).
I hope I don’t appear to be assuming homochiral alpha peptides. Any peptide of any isomer composition, and linking through any of the -NH2/-COOH-bearing carbon atoms, can potentially be functional. The H where a side chain ‘should’ be, or the side chain where an H ‘should’ be, merely introduce a conformational kink in the peptide chain, likewise beta and gamma linkage. They produce something different, but little more different than substitution of any old same-chirality acid with a different side chain.
The problem is functional selective advantage doesn’t decide how chemicals will behave before the supposed advantage appears. You keep invoking selection for non-existent features before they appear. Differential reproductive success doesn’t happen for non-existent features. The enucleated cells illustrates the problem that differential reproductive success won’t happen for features that don’t exist even if would be lead to differential reproductive success after it merges.
You keep saying you don’t invoke enucleated cells. Ok, a cell totally missing DNA. Why will differential reproductive success select for DNA when DNA does not exist in the system. How does Darwinian Selection chemically select for non existent chemical features?
lol
Heretics and apostates need punishment.
Why else would IDiots call evolution–the one successful model actually based upon the evidence–religion?*
Glen Davidson
*OK, there are other reasons, but the punishment idea crops up repeatedly.
Because IDist unconsciously have such contempt for religion that the worst insult they can think of is to call competing ideas religious.
Thanks for answering.
Then isn’t it fair at the least to say “we don’t know” or “we believe” we have no idea of what the scientific expectation is in general environments, and we have no idea what specific environments will make the expectation higher based on first principles.
Not exceptional for the entire known universe. But a qualification, that is if it is predicted to happen “once per 1000 galaxies” from first principles.
Thank you for answering. But if you’ll indulge me one more question, and there is no right or wrong answer, “Supposing you were able to calculate from first principles the probability of life, is there any probability (or deviation from expectation) that would be remote enough to convince you life was a miracle?”
Ok, I’ll give my answer. Koonin used the much maligned UPB that Dembski uses. At a personal level, that’s good enough for me. I suppose there is no right answer for what number it should be, but I wouldn’t be betting my soul on an estimate of 1 out of 1000 galaxies given that’s just a guess to begin with.
Thanks by the way for your comments. Sorry I have to disagree, I appreciate the time you’ve taken to write.
That’s not the point of bringing up the enucleated cell. The point was to illustrate DNA should not evolve in a system merely because it will confer differential reproductive success after DNA transcription and translation are added to the system. Selection can’t select for non-existent features.
One can’t exercise options that don’t exist.
If life was a miracle, how much more unlikely is a intelligent designers existence?
Enzymes are promiscous, usually. It is actually relatively rare for enzymes to only catalyze one single specific reaction. In this sense, the evolution of DNA is the biochemical synthesis of DNA by modification of RNA precursors, which really just requires a switch in the specificity of certain enzymes. In all likelihood, the capacity to make DNA from the RNA precursors already existed in some enzymes, but at very low activity levels.
I did not read it that way. It was selection via alternate universes as means for explaining the apparent selection of a functioning system out of see of non-functioning systems.
FWIW, he even referred to the much maligned UPB that Bill Dembski uses of 10^150.
Multiverse, and multiworlds are not the only interpretations of supposed physical laws, an Ultimate MIND is also a valid possibility as pointed out in the 2005 nature paper I linked to. If one is willing to argue multiverses from physics, it is no less outrageous to argue for the ULTIMATE MIND which can be inferred from various interpretations of quantum mechanics.
And if one argues multiverses, then there is one universe where the Creationists are right, and the Darwinists are wrong. How do you know which universe we’re in. 🙂
Go on then. Infer away.
Right about what?
By the way, it only took one: https://en.wikipedia.org/wiki/Ribonucleotide_reductase
stcordova,
Do I hell as like! Selection does not cause the 2′ oxygen to be removed; a mutation does. Selection determines its future course in the population.
The main thing enucleated cells illustrate is your willful determination to locate the wrong end of any available stick and grasp at it like a drowning man.
You seem to think I am thick enough to propose that it would.
I am not proposing cells which do not possess hereditary material. The hereditary material is RNA (some viruses still use it). It becomes DNA through the biochemically trivial tasks of removal of the 2′ oxygen from ribose and methylation of uridine during biosynthesis. RNA replicators can be subject to ‘Darwinian’ selection in exactly the same way as DNA replicators are.
stcordova,
Of course we don’t know. People are trying to find out. You are standing on the sidelines doing some hokey calculations in order to tell them they are wasting their time.
There is no probability too low to make me think something is impossible, except for zero. You act like you know what the probability is, when you have no more idea than I, and base it upon some dubious understanding of the issues.
It only has to happen once to explain present observations – and present observers. It only has to be probable enough to happen. All points in a distribution must be visited by repeat runs, including the tails. But you have to know the distribution to assess where you are in it – how ‘unlikely’ the event.
stcordova,
Not necessarily. A universe has a temporal dimension as well as a spatial one, and some things could still be impossible in an infinity of them due to blockage of pathways. An infinity of chess games would not throw up every distribution of the pieces on an 8×8 matrix – only the ones that can be accessed by a series of legal moves.
On the specific matter of Creation, I’d say either all universes were created or none were. Seems a bit iffy to have some spring up unaided and others from the local god’s fevered brow.
If supposing there is an RNA replicator and the RNAs are the hereditary material, and the hereitable RNAs become DNAs, why will the replicator keep functioning?
I’m assuming the model you suggest is some sort of polymerized RNA strand. Are you saying we’ll have a polymerized strand with both RNA and DNA or DNA only? How many bases are we talking about. Will RNAs then keep copying the DNA templates now? If they are now DNA templates, they won’t, as far as I know have the same ability as RNA to copy themselves.
So suppose the RNAs convert to DNAs as you suggest. It would seem the catalyzing ability would be lost now that RNA replicator converted itself to a pile of DNA.
So you could then propose a gradual scenario instead. But how then does one convert the RNA enzyme to DNA enzymes (if such a thing is even possible) to do the same task? If along the way the species dies, there is no further chemical evolution.
Sal should take a look at Korzybski.
The models of genomes are not genomes. Information is not molecules.
Logical manipulation of abstractions are not binding on reality.
It is true that scientists “assume” natural causes. This assumption did not pop fully formed from the head of some Ur Atheist. It arose gradually over hundreds of years, derived from the continuing success of the scientific program. It’s a useful and productive assumption. It will continue to gain prestige and social power because it works.
So Biochemists will continue prodding and poking at models of OOL. Because that activity is productive. Occasionally it produces something commercially useful, which is why basic research is funded.
And Sal will grow old and die without having done anything useful at all.
That is not necessarily true, RNA-Ribozymes capable of polymerizing RNA, DNA and even mixed-monomer RNA and DNA sequences are known:
DNA polymerization catalysed by a group II intron RNA in vitro.
An (admittedly very speculative) scenario would go something like this: In an RNA/Protein world, possibly with a RNA-based RNA-replicase, an enzyme ala the RNR(Ribonucleotide-Reducatse) arises with a low activity (or an already existing enzyme has trace RNR activity and is duplicated).
Meaning it only makes few DNA monomers, which are subsequently used in replication when they are made. The immediate selective advantage of this is that some selfish genetic elements made of RNA (think parasitic transposons or RNA viruses) stop working, or are just reduced in activity, combined with the greater strand stability conferred by reduced hydrolysis rates of DNA containing polymers compared to pure RNA polymers. (This shit has actual computational support, simulations have shown there would, in fact, be a significant selective advantage to a gradual incorporation of DNA against selfish genetic elements).
But due to the low rate of DNA monomer synthesis, they are not frequent enough to also break the function of the RNA-replicase, meaning RNA replication can still proceed and other cellular functions remain intact as long as initially the rate of DNA incorporation remains low and only gradually increases over many generations. In this sense there is dual slope for selection to climb, on the one hand increasing the frequency of DNA monomers will further reduce the activity of RNA-based selfish genetic elements and increase genome stability, but in turn also reduce the functional rate of the RNA-polymerase, which is in turn forced to adapt to an increasing frequency of DNA monomers. A long slow climb from RNA to DNA is mediated by this process. Gene duplications along the way ensure multiple copies of the original RNA polymerase exist, which remain RNA-polymerases operating eventually on DNA genomes to produce mRNA (which right from the begining was required by the Ribosome).
Now, to be sure, I made this up. It is not something you should believe happened, I don’t. I simply don’t know and that is where my own beliefs stop. But you’re here arguing for a design origin because, presumably, you can’t see any other way. But the fact is, you don’t know either. You’re effectively here arguing from ignorance.
There’s an endless supply of ammunition available.
I know I sound like I’m being personal in attacking Sal, but it’s not at all personal. My animosity is directed to everyone who argues by quote mining and cherry picking and straw manning. Toward people like Doug Axe who prove that birds can’t evolve from squirrels and generalize this to all of evolution.
Here’s a clue, Sal. Modern things did not evolve from other modern things. No multi-celled entity evolved from modern bacteria. So just stop shitting in the punchbowl.
If you want respect, show us you are worthy of respect by demonstrating you can argue the case for evolution. When you have demonstrated some mastery of current theory you can begin picking it apart.
Once again – enucleated cells have no genes in them. Evolution needs genes, which are then copied to produce new, variant versions of the genes. This doesn’t happen in a cell without genes. It is not a test of how DNA could evolve, because nothing at all evolves in this scenario. It proves no point, other than removing all the genes from an organism is a good way to make sure it doesn’t evolve anymore.
No one is proposing RNA enzymes got converted to DNA enzymes. Whether RNA genes were being replicated by RNA enzymes or protein enzymes at that point, the genes would be converted to DNA, not the enzymes. Transition to genes with some DNA and some RNA is perfectly feasible – there are polymerases that read both RNA and DNA, and a few which write both DNA and RNA. A few mutations or condition changes can even get existing polymerases to read and write with XNAs with entirely different back bones than DNA or RNA. As for selective advantages – even a little bit of DNA will decrease degradation rates and decrease melting temps. If life is competing for ribonucleosides, converting them to deoxyribonucleosides is a good way to monopolize the resource.
For the record, many functional RNAs can retain function when partially made out of DNA, and one ribozyme has even been converted by evolution from fully RNA to fully DNA. But again, this isn’t necessary for the origin of DNA genes.
Thanks!
stcordova,
Why would it stop? Base recognition is by hydrogen bonding to the edge of the nucleotide bases. These bases present identical profiles in DNA and RNA. Neither the 2′ -OH nor the methyl group in thymine interfere with this.There are, in nature, both DNA-dependent RNA polymerases and RNA-dependent DNA polymerases.
I’m starting with random RNA strands of indeterminate length. They cannot replicate, but hybridisation of complementary strands preferentially stabilises their components from a more complex mixture. This would tend towards homochirality (with equal chance, initially, of separate D-ribose and L-ribose pairings), and would favour pairable bases over others.
Dunno. Not many, initially. Random nucleic acid chains tend to cyclise at very short lengths (fewer than 10). But a cyclic genome is exactly what prokaryotes (mostly) have. Cyclisation is due to flexion of the strand – the 3′ end bends round and grabs its own 5′ tail, like Kekule’s serpents. Single strands are much more flexible than double.
Flexion also causes hydrolysis of the links in the chain. This provides a possible mechanism of chain growth – two short cyclic chains break, and instead of recyclising to the same chain, the 3′ and 5′ ends of the two shorter chains swap. Voila! A longer cyclic chain.
I don’t see why not. It’s a bit harder to separate DNA strands because of the tighter packing, but it’s only a question of degree, not an absolute barrier.
No. A given sequence would fold differently if made of DNA vs RNA, but there would never need to be DNA catalysts. An organism that used RNA as both its genetic material and its catalysts would make multiple copies of those RNA catalysts from the genetic RNA. Following partition of the 2 roles, the RNA catalysts would be templated from DNA instead. No change in catalytic ability, just a change in the ‘master copy’s font.
Just to be clear, are these polymerases proteins or some RNA that does the same job. The literature uses the term “RNA polymerase” but I take that to mean a protein that takes DNAs and makes RNAs.
The reason I asked it seems you think the RNA to DNA transition would likely need protein based polymerases to make the RNA to DNA gene transition possible.
If anyone wants to weigh in, when will helicases emerge or when will something that does their function be needed:
https://en.wikipedia.org/wiki/Helicase
stcordova,
They are proteins. But their presence demonstrates the fundamental feasibility of polymerisation in all 4 of the possible combinations of template and new strand.
eta – interestingly, the process of xNA polymerisation is essentially the same for a proto-replicase and a proto-transcriptase. This limits the number of things that have to be ‘invented’.
Thank you and David for your highly informative comments.
So it seems then the transition from RNA to DNA is not so big a step since RNAs and DNAs are chemically similar.
Is it a fair statement then, if we had an RNA world, the next major transition would be from RNA replicators to RNA+protein replicators?
For what it’s worth, would an RNA gene require helicase? Presumably since mRNAs are single stranded, with the RNA gene (I’m presuming it looks like an mRNA) becoming a DNA gene, will the DNA gene be single stranded? When will the double helix emerge?
Thanks in advance. This is one of the most informative discussions I’ve been a part of. Since it is OOL, some of the ideas are not exactly textbook biology, so thank you for taking the time to respond.
Sal, you do realize that OOL is speculative. If we knew how it happened we could reverse engineer it in the lab.
I suspect you will take whatever you have learned back to AIG or some similar site and come back with reasons why this or that would not work. Of course we already know that existing models don’t work. that goes along with not knowing exactly how it happened.
A test of character is whether you stop at not knowing and declare that goddidit, or whether you try to advance human knowledge.
stcordova,
As soon as you have double strands, you need a means of separating them, both for replication and generation of ribozyme or mRNA transcripts – in order to ‘see’ the bases, the hydrogen bonded groups need to be unbonded. You might also look at topoisomerases, if you’re looking to bolster your ‘impossible’ hand.
As I have said, when ignorance is your weapon of choice, ammunition is not scarce.
stcordova,
As I have said, it is not clear in which order DNA and protein would have come. There is no particular reason for either to have preceded the other.
In my model, before everything else – but that is personal speculation based on my perceived need to get homochirality and base complementarity off the ground before anything else can take place.
As an aside, I’d recommend disabusing yourself of the common notion that RNA is single stranded while DNA is double stranded. Both can exist in either form.
Thank you. Didn’t know that. I guess double stranded RNA is not too common in biology today???
Meant to ask you about that too thanks.
Double-helix early. I like that. Though I would guess base complementarity is important, given we’re talking a scaled down system, I actually was not acquainted with the reason why this is needed. Not something I looked into since I assumed if life had it, it was important, but didn’t bother asking why.
From here:
http://www.ncbi.nlm.nih.gov/books/NBK26821/
the S and S’ strand make DNA facilitates DNA copying. Would it stand to reason a pre-DNA RNA genome would also be complementary.
I recognize that is your model, but whether right or wrong, it seems to point the the value of double-helix existing early.
I think calling it a model is giving it too much credit.
I feel like I fit right in!
Mung,
Yeah, thanks for your substantive critique.
Visiting the Sandwalk blog, Larry recommend an article on “metabolism first” by James Trefil. It was interesting.
http://www.americanscientist.org/issues/feature/2009/3/the-origin-of-life/5
PS
Dr. Trefil was my former professor in Quantum Mechanics in 2004. I was a little surprised to see the article since I didn’t think Dr. Trefil delved into OOL.
Dr. Trefil campaigned against ID, but he was always friendly to me even though he eventually knew I was on the opposing side. At one of his anti-ID talks, I stood up in the Q&A just to point out to him 3 creationists that got biology PhDs from his school and that one of his students (me) was a creationist. He introduced me to his audience and went out his way to tell them I got an A in his class. 🙂
It is not, though there are still some viruses that use dsRNA genomes. Many structural RNAs natively fold back on themselves and form doubles strands via so-called ‘hairpin loops’, and will even form a double-helix in the base-pairing part. Case in point, tRNA.
The double-helix is not something that had to evolve (or be designed into) RNA or DNA, the strands will natively wind around each other in that structure if the strands are basepairing complementarily. It is also pretty much sequence independent, a pure G strand will form a double helix as long as the antiparallel strand is the complementary base (C).
Nice conflation of ID and creationism 😛
stcordova,
More so than you’d think. RNA viruses are the only ones that appear to use it as genetic material. The significance of that depends on where viruses came from, which is somewhat obscure. It seems, on grounds of parsimony, that LUCA would be expected to be a dsDNA organism (if one buys the concept of a LUCA, that is …).
But locally double stranded RNA occurs whenever ribozymes or tRNA fold, or one RNA binds to another, so it is pretty vital in all cells. Hairpin loops are antiparallel pairings of complementary bases – if you look along the axis, one strand goes 3′-5′ and the other 5′-3′ according to the convention for denoting backbone linkage. It’s somewhat incidental that it turns out, when you follow the loop, that it’s the same strand.
Further, any RNA which binds to another RNA is doing so through antiparallel base pairing and forming a local stretch of dsRNA. This happens in regulation, in splicing, in protein synthesis.
You also get many stretches of ‘hybrid’ double stranded RNA-DNA (I use scare quotes because, confusingly, RNA-RNA and DNA-DNA binding is also called hybridisation). This is how RNA probes work, which are a powerful means of targetting a specific sequence, likewise PCR primers.
Basically, complementary nucleic acid sequences have a very high affinity for each other. This is only possible in homochiral strands with complementary bases – there are many possible configurations and exotic bases which would have less physical affinity. But I see homochirality and base affinity as vital precursors of replication and ribozyme generation, hence my tending, in my
apparently-not-a-model, to use hybridisation itself as a mechanism for partial purification of homochiral and complementary RNA components before replication or functional activity.I consider transition to be a tough problem – I don’t see how one can get complementary nucleic acid strands evolving after a different kind of replicator has got going, nor how important biochemical pathways involving molecules which aren’t RNA monomers can be retuned to use those monomers later. RNA monomers are embedded deep within genetics and within cell energetics. They are dual-purpose, so I address (or evade, if one prefers) the issue by seeing complementarity as fundamental to the whole enterprise.
Complementarity is at the heart of biology. It is the reason why organisms can increase – the very fact you can take a double strand and make a copy of each is the reason why there is a – ahem – ‘Darwinian’ competition. You get a net doubling, in ideal conditions, and hence exponential growth.
It is also the means by which you can make huge numbers of copies of functional molecules from the same strand – either RNAs, or proteins. It’s the reason tRNAs bind to mRNA. It’s one of the reasons functional RNAs fold. It’s the reason nucleic acid promoters and repressors bind to their sites, and the reason genes can be silenced by antisense transcripts. It’s the reason recombinational mechanisms can work – separate two pairs of near-homologous strands and one can ‘invade’ the other instead of simply re-annealing, and hence LGT, homologous repair and sex. But you don’t get any of this if your sugars aren’t homochiral or your bases not complementary.
stcordova,
Topoisomerases are cool. They perform a neat trick of cleaving one or two strands and then either rotating around the remaining strand (single break) or passing an entire double strand through the gap (double break) before re-ligating. This helps sort out the pile of knitting that is a lengthy xNA molecule.
The latter is a ‘type ll’ isomerase, and interestingly, the initiating protein in sexual recombination double-strand breaks shares clear homology with an archaeal Type ll topoisomerase. It performs the double-strand breakage, but lacks the re-ligation activity, so the cell’s repair pathways kick in to resolve the gap, with far-reaching consequences. The gap is resolved by strand invasion from the homologue, and can result in crossover or non-crossover products with equal probability. So: no topoisomerase, no sex.
On looking on the issue of topoisomerase, if the genome is small enough, I guess one could by without it for an early protocell????
The following considerations were pretty much inspired by this video of replication.
https://www.youtube.com/watch?v=27TxKoFU2Nw
Would not DNA Gyrases (or some facsimile) be more important than topoisomerase?
Do we need something like beta clamps and clamp loader (or some facsimile)?
How about Primase (or some facsimile)?
If we have a double helix wouldn’t we need something like DNA Polymerase 3 (or some facsimile) for the Okazaki fragments?
Is there any way that a double helix (be it RNA or DNA) can be read and replicated without the lagging strand being duplicated as in the video?
I respect that these questions are conjecture. I like your double helix early model because it at least agrees with my “every thing first” model. 🙂
If I could convey where I was headed, there are certain interdependent relationships that need all parts to make a working Rube Goldberg machine.
Most certainly a task might conceptually be done with fewer parts, but there are contexts where it just makes sense for stuff to be there all at once before it becomes operational, such as this example:
http://www.evolutionnews.org/2012/03/illustrating_ir057831.html
In the engineering world, if needed, we may provide temporary supports, scaffolds, bricolage, etc. in the construction process until the machine is switched on, hence we circumvent the need for everything to be there all at once in the construction process. But once it’s operational, there is a core that has to be there in place at the same time.
In the early-life world, there problem is how everything doesn’t have to be there all at once in the abiogenetic and/or early evolution process — bricolage and construction scaffolds have to build themselves, and the chemical evolutionary scenario attempts to explain how the stages can collectively act as factories for the eventual product.
In engineering experience, the factory that makes something is different than the product. In the biological world, the factories are identical to the final product. The “everything first” model at least works in that sense if one will allow for a POOF. The gradual chemical evolution model has to describe a plausible factory in the stages of chemical evolution.
This just in: The exact history of OOL is STILL unknown.
stcordova,
You don’t need a single-function enzyme like modern topoisomerases. What you actually need, though, depends on whether the DNA is single- or double-stranded, and whether looped or linear. Topological issues depend on whether there are free ends that can spin, and on what kind of replication is performed – the modern ‘Okazaki’ format, both strands at once, or a primitive version generating single strands independently which subsequently anneal.
I favour looped, and I favour double stranded, but these are really just prejudices based on a couple of observations.
It appears to be a fundamental requirement that polymerases need both something to ‘bite on’ and a region of overshoot. Replicating a linear genome is therefore tricky. You can’t replicate the first or the last few bases. Prokaryotes circumvent this by (usually) having circular DNA. The DNA then provides the overshoot bases in both directions. Eukaryotes can have linear genomes because of telomerase, which adds a few sacrificial bases to the ends, but one can’t suppose telomerase to arise before replication.
Another factor is that cyclisation is a common terminator of nonenzymatic polymerisation, so it occurs ‘naturally’ – albeit making ‘genomes’ of a mere single-figure number of bases. Such genomes can grow, however, by breakage and re-ligation between separate loops.
And finally, looping provides a hedge against mechanical damage. Damage a linear strand and the components float off. Break a loop and you need a second break before all is lost. You have a window of opportunity to re-ligate.
My prejudice favouring double strands arises from the need to establish homochirality and complementarity early, and the fact that single strands can only synthesise a complement, not the original strand. Indeed, having synthesised a complement, it will likely anneal to the template, so double strands are pretty much inevitable, one way or another.
But both of these lead me to topological problems, because I have paired strands which do not have free end rotation, and semiconservative replication of which leads to a tangled mess of parent and daughter strands.
So, pretty early on, I would need an active mechanism of break formation, and an active mechanism of ligation. These would not, however, need to be done by the one enzyme.
One would not need a modern-style system, where there is continuous replication in one direction and mini-replication in short bursts with repeated ligation in the other. This, indeed, is asking a lot of probabilistic resources to arise all at once! No, a single polymerase adding bases to the 3′ end of a growing strand (and hence moving in the 5′ ->3′ direction of the template strand) could in principle replicate both strands, because both strands have a 5′->3′ direction. Set off in one direction to get a single-strand copy of the first, which is free to escape (barring topological linkage) because it has no complement. Then do a turn in the opposite direction against the second strand to get a single-strand copy of that. Since these two single strands are complementary (assuming some minimum fidelity) they will hybridise. Voila! a replicated genome. Set two such enzymes off at the same time in opposite directions to speed things up. This provides a second means by which genomes can lengthen too – on a circular genome, overshoot means that you have replicated a few additional bases, so if your polymerases aren’t very good at stopping, the genome will grow, subject to a converse pressure limiting wastefulness.
Heh! In one respect … !
My favourite opener to an ID paragraph, along with “as a programmer” …!
I appreciate this, but there is a fundamental difference to me between the stuff of engineering and programming, and the world of atoms and molecules: your components are not inert. A living organism is a complex cascade of energy surgically applied at very precise places. Atoms react very readily when passed close to each other. If there is a minimal configuration, and it requires Design to work out what it is, it also requires the suspension of known physics in order to build it. A favourite analogy of mine (stop me if you’ve heard this one) is trying to build a jumbo in mid-flight, from parts some of which react explosively with oxygen, others shrivel up to get away from water, others which will clamp magnetically to anything they pass near and require a shedload of force to disengage. All the while aviation fuel is spilling out, while you try and clamp the hoses …
I don’t think everything has to be there at once, but I do think that RNA needs to have its properties fixed early on. And that means, I think, before replication, which is why I favour hybridisation as a means of stabilising molecules. Once you have stable double strands, it is an obvious benefit if, instead of ultimately succumbing to entropic degradation, you can make a copy first – which complementarity allows you to do. Then, things start to get really interesting.
So, a few not particularly interconnected points –
1. circular and double-stranded DNA/RNA: I understand the appeal of both properties, but a circular, double-stranded nucleic acid couldn’t be smaller than 150bp, and probably closer to 500bp. Double-stranded nucleic acids are fairly stiff as polymers go, with persistence lengths of ~50 nm, which translates into 100s of bp. So I could see short dsRNA being around early, and I could see short circular ssRNA around early, but short dsRNA wouldn’t have been an option.
I think as soon as you have nucleic acid replication, double strandedness will be the default option, since nucleic acid replication has to pass through a double-stranded intermediate in the first place. Circular genomes have advantages and disadvantages, and linear genomes aren’t that challenging to replicate – most viruses manage it through a variety of methods, without any telomere problems.
2. Ancillary replication enzymes (helicases, topoisomerases, etc) – all are useful, none are necessary. As Allan mentions, topoisomerases are only useful in contexts where the ends of the genome are either absent (because its circular) or not free to rotate. Even without them, if enough strain builds up, any nick that occurs in the genome with also manage to remove excess supercoiling. Its messy and less controlled, but it should still work. The same thing holds for helicases – they help separate strands, but many polymerases are able to displace complementary RNA/DNA strands without any other enzymes. Strand displacement falls out of polymerization pretty readily – the energy released by synthesizing the new strand is normally enough to force off the old strand.
3. To add to what Allan said – In engineering and especially in programming, every function needs to be explicitly set up, and the default in the absence of this set up is to not do anything. In biology, overlapping activities and unexpected interactions are the norm. You don’t need a helicase function to get strand displacement, it falls out of polymerization, you don’t need to explicitly code for stopping during translation or transcription, termination is going to happen regardless. Often times specific enzymes or functions are present not to make something happen that won’t otherwise, but rather to regulate an event that will occur spontaneously, for the sake of efficiency or some other goal.
david,
Granted, I may be ‘joining the dots’ a little hastily between ssRNA cyclisation at less than 10bp and circular prokaryote genomes. But something has to make ribose homochiral, bases complementary and 3′ linkage the norm in order that complementary pairing and templating be possible. A single strand has no constraints. It is only complementary pairing that determines the particular properties from which everything else flows.
Single strands, because they are so flexible, are susceptible to 2′ -OH attack (or 3′ -OH, in a pre-constrained polymer of mixed linkage). Cyclisation has some appeal in helping minimise the catastrophic consequences of a single break. But dsRNA, by stiffening the chain, both reduces the incidence of such attack, and pins the components in place to permit re-ligation much more readily.
It is not necessary for the entire strand to be dsRNA. If a circular ssRNA polymer is long enough, shorter local regions may hybridise with free short random complementary ssRNA. This stabilises those regions, and favours homochiral, complementary, 3′-5′ regions over others, even if the entire ‘genome’ is yet too short to permit both double stranding and circularity. The ssRNA region will be more susceptible to damage and excision by paired breaks.
All that said, is not essential that the genome remain circular during this growth phase. I just see it as an added form of stabilisation initially, which may (or may not) make polymerisation easier later.
But how do you replicate a region with no complement?
Viruses usually parasitise on the host machinery, so I’m not sure they provide a guide to primitive mechanism, unless a polymerase can start and just ‘fall off’ a bare-ended viral genome without missing anything.
Agreed, and ultimately these things need to be resolved by experiments (which, of course, won’t prove what did happen in the past, but lay out the range of what’s possible). The idea that stabilization by base pairing solves these things is not implausible, though there are problems for each – enantiomeric cross-inhibition (L-RNA monomers blocking D-RNA replication and vice versa) is a big problem, so unless stabilization of homochiral strands feeds back into the monomer supply, this will be an issue. Whatever prebioitic process might produce nucleobases could supply a set more prone to base pairing or less, which could pose a problem or not; even a few ambiguous bases could be tolerated as a higher error rate or heterogenous RNA, and eliminated only once RNA life became more complex and could only tolerate lower error rates. A similar principle could be at work for 3-5 vs 2-5 linked RNA – ribozymes evolved as 3-5 chains can tolerate a very high fraction of 2-5 linkages.
But we could argue about these things all day… My own view is that we should keep an open mind and realize that very little has been explored out there. The origins field is a hard one to be in, there’s not much funding and very few “safe” projects. But more so than bright new theories, I’d very much like to see a much larger bedrock of experimental systems that look at how simple we can make life, and how life-like we can make chemical systems.
Right, given how fast viruses need to evolve (given the arms race with hosts) and are able to evolve (given small genomes), there’s not much hope that they’ve preserved sequences from deep time, although you never know what manages to stick around (I’ve always liked Patrick Forterre’s theory for the viral origin of DNA). My main point in referring to viruses is that there do seem to be many strategies to avoid the telomere problem. First, reading flush to the 5′-end of a template isn’t a problem for most polymerases. Many read all the way to the end and even add a few extra nucleotides after than. The 3′-end is more of a problem, particularly if the polymerase initiates without a primer. But many viral replication cycles will involve priming off of an RNA or even a protein, and cover the complete 5′ ends of their genome this way.
PS, I should add that many viruses don’t use host-cell polymerases, but encode their own. Most of these virally replicated genomes are linear, and to my knowledge most don’t have a specific telomere problem.