Probably, shorter codons lead to faster synthesis, smaller molecules and thus faster diffusion, and thus faster reproduction overall. Also, there's a correlation between cell size and genome size; smaller cells have a higher surface-to-volume ratio and therefore faster metabolism, not to mention faster reproduction again.

However, it could just be a historical accident: the first successful replicator happened to have three-base codons, and so do all its descendants... maybe longer codons weren't selected against, but simply never evolved in the first place. Who knows.

(Two-base codons are of course too short for 20 amino acids. But then, why do we have 20... or 22 or 18...)

I don't think there is any "reason" why most things in biology turned out the way they did. Surely there were not dozens of different origins of cellular life on earth using various genetic codes etc. and the one with a 3-base codon code won out. All evidence points to life on earth all sharing a single common ancestor. In contrast, there are some good "reasons" why computer code begins with binary data, divided into 8-bit bytes, etc... Likewise, most human spoken languages have vowels and consonants, but there are some languages with click sounds and other alternatives. There are several ways of creating written symbols to represent the spoken languages.

Given four DNA (and RNA) bases, and 20 amino acids, the three letter code makes a lot of sense. But there is no real reason life could not have used 200 or 600 different amino acids instead of 20, and no reason to have 4 bases instead of 2 or 20. If there were 6 or 10 bases and 200 amino acids a 5-base codon system may have been worked out.

Within the coding scheme that did evolve on earth, there are a lot of interesting questions. For example the 64 codons are not equally divided with each of the 20 amino acid having 3 codons, plus 4 extras for stop codons. Why do we have 6 codons for some amino acids and only 1 or two for some others? Some bacteria have close to 80% G+C genomes (20% A+T) and others are very A+T rich and C+C poor, which presents challenges for maintaining protein sequence conservation. In evolutionary studies, the third codon position is often considered to be "silent" and have very low selection pressure, but in many cases there is indeed quite a bit of selection on the "silent" sites.

The scientific basis for the 3-base (triplet) codon is well documented. Why 20 genetically-encoded amino acids and not 60 or 80? could also be a question. This is biological and not chemical or physical science. We know in part and only attempt an elucidation of weird mechanisms. Only attempting an explanation of nature's why's and how's. A sublime academic exercise. Three, four five six.base codons all presented and explored. A three-base codon won. Because nature has taken a decision and it has made it's choice. A three-base codon it is and not five. Just natural selection.

Clicks are just one type of consonant; all languages that use clicks also have good-sized inventories of other consonants and vowels. But of course the point stands – despite a bunch of constraints, the sound systems of different languages do vary a lot.

Hanna Schumacher

First I want to comment on your question and its starting itself. Its not naturally possible for tRNA to have four or five anticodon, so your question is wrong. The paper you are mentioning it itself says about the non-natural amino acid. As they are not natural one can engineer two, four, five or whatever number of anticodon in tRNA, which should not be the topic for the discussion. Moreover, if they do not exist naturally, the question about how and why they are prevented/not used in transation, does not arise.

Further according to the triplet codes, the number of codons and amino acids with the degeneracy of codes seems to be logical. Now if you want to argue why three, this is different question from why not more than three.

To answer this question, we need to see the chemistry of biomolecules. At the origin of life it was more of a chemistry than biology. There might be several DNA bases too which could have been present in DNA but did not because may be of the thermodynamics and stability of the molecules only ATGC could have made it to form meaningful bond and code for information. The molecules which were not stable culd not have been used to code (as we can see for several other DNA bases which are not stable and thus do not present universally in normal DNA). Other molecules derived from DNA like RNA (all types), also follows the principals of the chemical stability and functional property. The structure of tRNA is such that it has three bases which can serve as binding site between mRNA and tRNA (naturally) and thus surves as anticodon.

By the way, the codon theory was experimentally proven and have been analyzed from bottom to top approach (amino acid/tRNA to DNA code).

Why don't we use a set of three- and five-base-codons together?

>> I suppose its not us to decide. It is possible that there might be some tRNA which were having more or less anticodons but they might have resulted in something which is not stable and thus those mutants were extinct. Again, I would like to empasise on the chemical bonding and stability of tRNA molecules. the secondary structure of tRNA () and its ability to carry specific amina acid might have been the best combination in the process of origination of life and evolution. We could also argue why tRNA only carry one amino acid molecules, if it would carry two amino acid, polypeptide building could have been faster, isn't it?

I don't agree with David Marjanović 's answer regarding three base codon could give faster reproduction. according to this hypothesis, the rate of reproduction must be positively co-related to the codon usage. If Faster reproduction can be due to three codon (lets say some bacteria or cyanobacteria), why not there are more codons for the organism which has slower reproduction for example archaea?

"Moreover, if they do not exist naturally, the question about how and why they are prevented/not used in transation, does not arise."

The question, as I understand it, is why they don't exist in nature.

"If Faster reproduction can be due to three codon (lets say some bacteria or cyanobacteria), why not there are more codons for the organism which has slower reproduction for example archaea?"

Why would there be? It could just be inherited from an ancestor that reproduced faster than its competition.

David Marjanović If you read further down, I mentioned the possibility why they do not exist in nature. Moreover, what and which ancestor? What do you propose the single common ancestor in tree of life? And where are the ancestor which had comparatively slow reproduction? Do you want to suggest that they extinct just because they had slower reproduction? For me it does not sounds logical. If you can propose some evidence for that, I would like to read further.

Its natural biological phenomenon.

"What do you propose the single common ancestor in tree of life? And where are the ancestor which had comparatively slow reproduction? Do you want to suggest that they extinct just because they had slower reproduction?"

In short, yes. If you have different organisms in the same environment and all else is equal, except that one kind reproduces faster than the others, then that kind will be overrepresented in future generations. If the probability of death is equal for all, then the fastest reproducers will sooner or later end up as the only "species".

David Marjanović Now its becoming like an argument over the rate of reproduction. From my side I would like to put a full stop on it until you provide some evidence of the rate of reproduction related to codon length. Its not a question about reproduction rate anyway.

As others have said, I see this as a question of why a universal three-codon translational code, rather than a two-codon or four-codon (or larger) code, or a set of non-universal codes, is what we observe. The evidence we have to work with is the universal system that did evolve, presumably once and by out-competing all competitors that evolved at any time before or since. So what do the characteristics of this system suggest?

The observed system is a redundant, 64-codon code that supports two control signals (start and stop) and 20 amino acids _directly encoded_ as codons. The only codons that do not cause an amino acid to be added are the three stop codons, which are by and large functionally redundant. The facts that stop codons don't cause an amino acid to be added to the chain, while the start codon does, appears to be due to mechanism: the start codon provides a c-terminal onto which the next amino acid can be attached, while the stop codon causes release rather than addition of a terminator without an available c-terminus. The initiator codon starts the amino acid chain with an amino acid (typically Methionine, but sometimes Valine) which is often cleaved off after it (the initiator amino acid) emerges from the ribosome, which can be interpreted as a "reason" for why there is only one universal start codon: the identity of the initiator amino acid isn't particularly important, any amino acid would do.

One might be tempted to say that there ARE only two control operations (start and stop) encoded in a protein-coding sequence, however as someone who has studied the mechanisms of splicing, I can say this is false. There are a variety of other control operations encoded in protein coding sequence: not just splicing donor and acceptor sequences, but also exonic splicing enhancers, translational pause sites, regulated frameshift sites, and (very likely) RNA-binding sites which regulate how the RNA is exported from the nucleus (in eukaryotes) and localized within the cell (in eukaryotes, and possibly in prokaryotes as well).

Since the code is redundant, but only contains these two kinds of control signals embodied _as codons_, I think it is clear that a redundant, three-base codon system provides enough coding potential for both (a) the biochemical diversity required of the encoded amino acid chains, and (b) the initiation and termination of the amino acid chain itself, and (c) the required information for proper post-transcriptional regulation of the transcribed pre-mRNA. Suppose that the coding potential was NOT sufficient for encoding these three kinds of information: then we would not observe the system that has evolved, we would necessarily observe a more complex code, perhaps with four codon positions!

By the same argument, this is precisely why we do not see a one-position or two-position codon system: they simply do not provide enough coding potential to encode all of the required signals.

It is important to recognize that it is the redundancy of the three-position system which permits additional regulatory signals to overlap with the amino acid code; without such redundancy, any post-transcriptional information would necessarily either (a) require a particular encoded amino acid sequence, or (b) be encoded outside the protein-coding region.

We can therefore conclude that a two-position amino acid code does not provide enough coding potential to support the (full) information required to produce a broad enough repertoire of proteins, including all control information required to be present in the transcript itself, but that a three-position amino acid code does. I think this is as close as we can possibly come to answering "why" the amino acid code that evolved is a three-position system.

"until you provide some evidence of the rate of reproduction related to codon length"

Longer codons mean longer genomes, and longer genomes take longer to replicate.

"Its not a question about reproduction rate anyway."

It could be, because it could be a question about natural selection.

Follow

For 20 amino acids this is the shortest possible length. A codon consisting of a single base could only code for 4 amino acids, a length of two bases for 16 (4x4), and of three bases for 64 (4x4x4). Given that tRNAs have to interact via their anticodons with the mRNA, we have an upper limit for the codon length. An A-helical turn consists of 8 (or 9, I don't remember) bases, rendering a codon-anticodon interaction sterically impossible beyond that length. Furthermore, large and unstructured loop structures destabilise an RNA, so that also gives an upper bound for the codon/anticodon length.

Ali, thanks for pointing out than an RNA alpha-helix has a strongly constrained length, that is a substantive argument for why tRNAs may not have evolved to recognize longer codons.

My answer (that the information we have about codon structure is limited to the observations we can make now) is similar to the first part of your answer: that the number of amino acids that can be encoded is a function of codon length, in that both imply the (circular) argument that we must need more than 14 amino acids (plus a start and a stop, making 16) because we observe three nucleotide codons rather than two nucleotide codons. It would be nice to have a noncircular argumentf for why the minimum number of distinct amino acids is more than 14, but that is beyond my ability to construct.