DNAunion: Here are some attempts to explain why the "self-replicating" peptide mentioned in both articles (what others have referred to as the "Ghadiri Ligase", which I will follows) is not a True Self-Replicator
First, I will use an analogy that employs the letters of the English alphabet and a short sentence in order to demonstrate why the Ghadiri ligase is not a true self-replicator.
For this analogy, I will equate the 23-character sentence METHINKSITISLIKEAWEASEL with the 32-amino-acid Ghadiri ligase. Each of the letters represents an amino acid residue along the length of the GL (my abbreviation for the Ghadiri ligase) where each of the individual letters is covalently bonded to its nearest neighbor(s) on the same strand (analogous to the same physical sentence, when there are two). The covalent bonds between the units will be represented with dashes (-) between the letters (M-E-T-H-I-N-K-S-I-T-I-S-L-I-K-E-A-W-E-A-S-E-L).
A true self-replicator can extract its individual building blocks (monomers/letters) one at a time from its surroundings (a pool of monomers/letters) and construct a functional copy of itself using itself as a template for the sequencing of the units, followed by release of the copy. In order to allow them to separate from each but to not decompose, they should not be covalently bonded together but both the template and the copy should be covalently bonded internally. Note that the letters cant simply line up according to the templates sequence and be done with it; they also have to be covalently linked to their nearest neighbors in the growing copy after being non-covalently attached to the template. Forming this bond between units within the same strand requires either a catalyst or the pre-activation of each of the building blocks. And since we are looking for a true self-replicator, the sequence itself should be performing the function, whether it is catalyzing the bond directly or pre-activating incoming monomers. The process we will look at (the less involved of the two) involves two basic steps for each monomer added: first, the correct monomer is chosen from the stocked pool of monomers and lines up along the template, then the template sequence covalently bonds the new monomer to the elongating string.
M-E-T-H-I-N-K-S-I-T-I-S-L-I-K-E-A-W-E-A-S-E-L
M (correct monomer lines up non-covalently with template)
M-E-T-H-I-N-K-S-I-T-I-S-L-I-K-E-A-W-E-A-S-E-L
M E (correct monomer lines up non-covalently with template)
M-E (template sequence covalently bonds new monomer to growing string)
M-E-T-H-I-N-K-S-I-T-I-S-L-I-K-E-A-W-E-A-S-E-L
M-E T (correct monomer lines up non-covalently with template)
M-E-T (template sequence covalently bonds new monomer to growing string)
M-E-T-H-I-N-K-S-I-T-I-S-L-I-K-E-A-W-E-A-S-E-L
M-E-T H (correct monomer lines up non-covalently with template)
M-E-T-H (template sequence covalently bonds new monomer to growing string)
M-E-T-H-I-N-K-S-I-T-I-S-L-I-K-E-A-W-E-A-S-E-L
M-E-T-H I (correct monomer lines up non-covalently with template)
M-E-T-H-I (template sequence covalently bonds new monomer to growing string)
M-E-T-H-I-N-K-S-I-T-I-S-L-I-K-E-A-W-E-A-S-E-L
M-E-T-H-I N (correct monomer lines up non-covalently with template)
M-E-T-H-I-N (template sequence covalently bonds new monomer to growing string)
M-E-T-H-I-N-K-S-I-T-I-S-L-I-K-E-A-W-E-A-S-E-L
M-E-T-H-I-N K (correct monomer lines up non-covalently with template)
M-E-T-H-I-N-K (template sequence covalently bonds new monomer to growing string)
M-E-T-H-I-N-K-S-I-T-I-S-L-I-K-E-A-W-E-A-S-E-L
M-E-T-H-I-N-K S (correct monomer lines up non-covalently with template)
M-E-T-H-I-N-K-S (template sequence covalently bonds new monomer to growing string)
M-E-T-H-I-N-K-S-I-T-I-S-L-I-K-E-A-W-E-A-S-E-L
M-E-T-H-I-N-K-S I (correct monomer lines up non-covalently with template)
M-E-T-H-I-N-K-S-I (template sequence covalently bonds new monomer to growing string)
[next 26 steps omitted to save space]
M-E-T-H-I-N-K-S-I-T-I-S-L-I-K-E-A-W-E-A-S-E-L
M-E-T-H-I-N-K-S-I-T-I-S-L-I-K-E-A-W-E-A-S-E L (monomer lines up with template)
M-E-T-H-I-N-K-S-I-T-I-S-L-I-K-E-A-W-E-A-S-E-L (final monomer covalently bonded to others)
So how does the actual Ghadiri ligase measure up? Not very well. Using the same analogy, here is how the GL functions.
The first PREEXISTING half of the sequence -- M-E-T-H-I-N-K-S-I-T, which for some unknown reason just happens to be floating around nearby, already covalently linked together -- lines up with template.
M-E-T-H-I-N-K-S-I-T-I-S-L-I-K-E-A-W-E-A-S-E-L
M-E-T-H-I-N-K-S-I-T
The second PREEXISTING half of the sequence -- I-S-L-I-K-E-A-W-E-A-S-E-L, which is also just floating around nearby for some unknown reason, already covalently linked together -- lines up with template.
M-E-T-H-I-N-K-S-I-T-I-S-L-I-K-E-A-W-E-A-S-E-L
M-E-T-H-I-N-K-S-I-T I-S-L-I-K-E-A-W-E-A-S-E-L
The two halves are covalently bonded together BUT NOT BY ANY EXTRA ACTION PERFORMED BY THE TEMPLATE SEQUENCE ITSELF, BUT BY THE SEPARATE TWO HALVES THEMSELVES, BECAUSE ONE OF THEM WAS PRE-ACTIVATED.
M-E-T-H-I-N-K-S-I-T-I-S-L-I-K-E-A-W-E-A-S-E-L
M-E-T-H-I-N-K-S-I-T-I-S-L-I-K-E-A-W-E-A-S-E-L
How is it that the needed halves just happen to be floating around? Because the researchers intentionally synthesize those exact two sequences, preactivate the copies of one of the sequences, and then supply both for reaction.
This analogy points out some conceptual reasons why the Ghadiri ligase is not a true self-replicator: it absolutely requires (1) the correct 15- and 17-aa sequences already be available in the surroundings, (2) both halves to already be held together by covalent bonds, and (3) one of the two halves to already be activated. The Ghadiri ligages is powerless to recreate itself from the individual building blocks that make it up.
Lets try looking at it with a second analogy.
What do we know that truly self-replicates? The most obvious answer is, life. To make sure that we are not throwing in excess complexity, what is the simplest form of life known? A bacterium (or to be more precise, the bacterium Mycoplasma genitalium). Lets double check
does a bacterium self-replicate? Yes. Okay, how?
In very simple terms, a bacterium takes in simple, raw materials from its surroundings and then the bacterium uses those simpler precursors to build extra copies of its own constituents such as DNA, proteins, mRNA, etc. and then divides to form two bacteria, each like the original.
Now, is this similar to the way the GL self-replicates? No, not at all. If the GL were a bacterium, it would require two preexisting halves of another bacterium that would then simply line up with it and join together to form a whole bacterium. In the real world, a bacterium is given only simple raw materials such as inorganic substances and sugars, that it uses to build a complete copy of itself from scratch; whereas in the GL world the bacterium has to be handed EVERYTHING already setup, for free, and just joins the two preexisting halve together.
Okay, lets try looking at this from an informational point of view.
A true self-replicating protein would not need any help by being supplied large amounts of very specific, external information; the needed information for self-replication would be contained in the self-replicator itself. But for the GL ligase, it requires being handed approximately 130 bits of information*. How much is that? Suppose a random process were used to select a single number between 1 and 429,000,000,000,000,000,000,000,000,000,000,000,000,000, and that you know nothing other than the range and that the integer was chosen randomly, with each integer in that range being just a likely to have been selected as any other. Now, handed 130 bits of information, you'd be able to correctly pick that one integer in one try. And even if the rough calculation is too high by 10 orders of magnitude, or even 20 orders of magnitude, the fact still remains that the GL requires a huge amount of preexisting information to be handed to it for free in order for it to self-replicate.
However you slice it, the GL is not a true self-replicator.
So is the GL a catalyst? Yes it accelerates the rate of the two halves joining without itself being altered in the process. It has been shown that even in the absence of the GL, the preexisting, pre-activated 15-aa and 17-aa fragments will bond together to form the full 32-aa GL. In the presence of the full GL, this rate of combination of the halves to form the full template is increased, and after doing so, the original template is ready to align another set of two halves so that they too will bond together. What the GL does is to orient two preexisting, pre-activated, specific sequences in the correct manner so that they can interact more readily (of course the probability of two halves finding each other and being properly oriented in order to link up is much greater when they are aligned linearly in tandem on a template than when colliding randomly in a solution). So yes, the GL is a true catalyst.
So does the term autocatalytic fit the GL? Yes it is a catalyst whose product is itself.
Is the GL a true self-replicator? No.
* The information calculation was done very quickly in back-of-the-napkin fashion. The complete peptide is 32 amino acids long, and the rule is that living cells use 20 amino acids in the production of peptides. Since we are not looking at preexisting life in order to evolve the correct sequence, and since intelligent direction is not involved either, the sequence of monomers would be generated by undirected, basically random forces. So, using that assumption that each amino acid is just as likely as any other to be incorporated at position X, we have 20 equally likely possibilities for each of 32 positions. That gives a total of 20^32, or about 4.29 x 10^41, possible unique sequences. On a piece of paper, in just four lines of code, I calculated that 4.29 x 10^41 is about 2^130 (the first number is actually almost 1,000 times greater than the second, but I was trying to avoid doing a complicated calculation and also didnt want to overestimate). With 2^130 equally likely possibilities, you need, on average, 130 bits of information to find the one correct outcome. Also, note that I mention above in the actual discussion that the infomation calculation's being too high by even 20 orders of magnitude would not alleviate the problem with the GL ligase. That is, if a more detailed calculation came out to 10^21, it would still require about 70 bits of information, which is still probably enough to win a state lottery twice in a row.
********************************
PS: Edited the day after it was posted to correct the analogy explanation of how much 130 bits of information is (I realized the error just as soon as I left school last night, so couldn't fix it until now).
********************************