Evolution of DNA - First Feedback
With a Roscoe in the puddle, the number of Sofias would increase dramatically, limited only by Roscoe's speed of replication and the quantity of raw materials.
Of course, Roscoe was not a perfect replicator, and it would make many defective copies of Sofia (which in turn would create various odd Freds when they were transcribed by a good Fred).
At some point, Roscoe would have accidentally produced an incorrect copy of Sofia with an extremely interesting property-- this 'bad' Sofia would just happen to create a Roscoe, when transcribed by a Fred.
We'll call it a Sorrel (short for Sequence of Roscoe Replicating ELements).
Once Sorrel came on the scene, there were two chains and two short proteins in the neighborhood, and they had a brand new synergy. Roscoes made more Sofias and Sorrels, Freds made more Freds and Roscoes, and then their children continued to make more of the same. A positive feedback loop!
From an evolutionary point of view, we are definitely getting somewhere with this! It's the first chemical system that is able to completely self-replicate, at least under extremely ideal conditions. All it takes is a Fred, a Sofia and a Sorrel in the right puddle, and the loop will run until all raw materials are turned into our four interesting large molecules.
In the presence of some catalytic or natural process that keep producing more of the same two amino acids and the same two chain molecules, Fred and Roscoe will keep on flooding the entire region with Freds, Roscoes, Sofias and Sorrels more or less permanently.
How long would it have taken for Sorrel to appear?
As with Roscoe, it probably wasn't long. The difference between Fred and Roscoe was probably just in one or two amino acids at the elbow end. That means that Sorrel would have been only one or two chain molecules different from Sofia. So it would only take a small number of glitches to accidentally produce a Sorrel from a Sofia.
The jump from protein replication gene to chain replication gene may have only taken hours, days or years, given the high percentage of bad copies that our inefficient Fred would have made. Practically an instant, in geological time.
We talked earlier about alternative versions of Fred. Now that we're familiar with Roscoe and Sorrel, there is another possible pathway to consider.
It's possible that Fred may have consisted of more than one small polypeptide, rather than one moderately large one. In that case, it would have developed via a slightly more complicated path.
Consider the original Fred puddle, but with two shorter half-Fred proteins-- we'll call them knee-Fred (which can match to chain molecules) and elbow-Fred (which can line up amino acids). We'll also throw in an elbow-Sofia which happens to create an elbow-Fred when it is transcribed by a Fred.
The sequence might go as follows:
1. When elbow-Fred, knee-Fred and elbow-Sofia got together, the temporary Fred would read the elbow-Sofia, and create a bunch of elbow-Freds.
2. A mutant elbow-Fred might be produced which would act like the elbow part of a Roscoe. It would match with the knee-Fred and start duplicating elbow-Roscoes.
3. A bad copy of elbow-Sofia might turn out to be a knee-Sofia, which would produce knee-Freds. Now the puddle would be filled with many copies of both Fred halves.
4. A bad copy of knee-Sofia might be a knee-Sorrel, which would produce knee-Roscoes. Now the puddle would be filled with many copies of both Roscoe halves, and positive feedback would start to occur.
5. Eventually the two half-Sofias might merge into a full Sofia, and the two half-Sorrels might merge into a full Sorrel. That would bring the system to the full Fred/Roscoe/Sofia/Sorrel system.
The advantage of the 'half Fred' system is that it uses much smaller chains and polypeptides (perhaps as small as 10 molecules in length). Those smaller molecules would have formed much more frequently from catalysis or templates. For example, in a two-amino-acid puddle, there are only 210 or about 1,000 possible permutations of 10-molecule polypeptides. So the chance of seeing two half-Freds and a half-Sofia might be considerably greater than waiting around for 20-molecule chains and polypeptides to appear.
The disadvantage of the 'half Fred' system is that it requires more 'puddle evolution' at the very beginning, before it can start replicating in quantity.
Evolution of Sofia and Sorrel
You might think OK, now we're all set, we have everything we need for positive feedback, and we'll soon have an entire ocean full of Freds, Roscoes, Sofias and Sorrels. And then they'll get better and better, and evolve into life.
Well, sorry, but it's just not that simple. We still have many tedious steps ahead, before anything like that can happen. There are two basic problems.
First of all, the Fred/Roscoe system is still very dependent on an exact mix of chemical raw materials, and there were probably only a few extremely rare puddles with conditions they needed for their chemical wizardry.
Even worse, the four-molecule system itself is still too simple to evolve successfully via plain old Darwinian selection. They are just free-floating chemicals, and they lack most of the properties of a true organism.
Let's take a closer look at these problems.
The biggest stumbling block for Fred and Roscoe is a simple matter of ingredients-- they need to have a high local concentration of just two amino acids for Fred to work, and high concentrations of just two chain molecules for Roscoe to work.
Since Fred is so primitive about amino acid selection, having other, similar compounds around can easily 'poison' its transcription results, and create non-functional polypeptides. Fred would pop a similar-but-wrong molecule into place, and the result would be something other than a Fred.
Likewise, Roscoe is easily poisoned by any molecules that are similar to its original chain molecules.
Fred and Roscoe could have arisen in the just right puddle complex where there was an extremely high concentration of just the right ingredients. But outside of the local range of puddles, there would be too many other compounds which would sabotage the process, and result in totally unsuitable results.
So for the moment, our little self-replicating quartet can only work when in a rare puddle with just the right concentrations of raw materials-- formed from a cluster of supercatalysts, or by some natural process.
Even with quadrillions of puddles in the world, there would have been only a very small number that were hospitable enough for Fred and Roscoe. It's possible that there was only one solitary puddle complex on the entire Earth that could support their lifestyle.
Another problem is that we are still in a chemical world of free-floating molecules, and we really don't have true organisms yet. Because Fred and Roscoe are still loose in the soup, they will not have an easy time competing against their own bad copies (and it's too much to believe that the first Fred and Roscoe would start right out reading and duplicating chains with 100% accuracy).
With no true genetic inheritance, yet, natural selection doesn't work as directly as it does with modern, living organisms.
For example, in the puddle where Fred and Roscoe first appeared, there would be a few good Freds and a few good Roscoes, but there would also be many bad copies. Likewise there would be good Sofias and Sorrels in a mix of many bad copies. In that kind of genetic chaos, there's no reason that Roscoe would have preferentially replicated the real Sorrels and Sofias, as opposed to the bad ones. Likewise Fred would have connected with plenty of bad chains, and then produced a host of bad proteins with no enzymatic action.
Natural selection works effectively for living organisms because there is a direct link between each individual's fitness, and its genetic material. If an organism does well, then its genes do well, automatically. But this simple, four-molecule prebiotic system has no direct linkage between the genetic inheritance (Sofia and Sorrel) and its expression (Fred and Roscoe). Or to use more technical language, the genotype and phenotype were not linked.
Fortunately, a certain amount of natural selection and evolution could happen with such simple molecules-- but only if there were large numbers of isolated populations where fairly simple, chemical-based selection could take place. Let's take a look at that now.