ArticlePDF Available

Evolution of the Genetic Code. The Ribosome-Oriented Model

Authors:

Abstract

There are currently three major theories on the origin and evolution of the genetic code: the stereochemical theory, the coevolution theory, and the error-minimization theory. The first two assume that the genetic code originated respectively from chemical affinities and from metabolic relationships between codons and amino acids. The error-minimization theory maintains that in primitive systems the apparatus of protein synthesis was extremely prone to errors, and postulates that the genetic code evolved in order to minimize the deleterious effects of the translation errors. This article describes a fourth theory which starts from the hypothesis that the ancestral genetic code was ambiguous and proposes that its evolution took place with a mechanism that systematically reduced its ambiguity and eventually removed it altogether. This proposal is distinct from the stereochemical and the coevolution theories because they do not contemplate any ambiguity in the genetic code, and it is distinct from the error-minimization theory because ambiguity-reduction is fundamentally different from error-minimization. The concept of ambiguity-reduction has been repeatedly mentioned in the scientific literature, but so far it has remained only an abstract possibility because no model has been proposed for its mechanism. Such a model is described in the present article and may be the first step in a new approach to the study of the evolution of the genetic code.
ORIGINAL ARTICLE
Evolution of the Genetic Code: The Ribosome-Oriented Model
Marcello Barbieri
1
Received: 24 August 2015 / Accepted: 2 September 2015 / Published online: 7 October 2015
Konrad Lorenz Institute for Evolution and Cognition Research 2015
Abstract There are currently three major theories on the
origin and evolution of the genetic code: the stereochem-
ical theory, the coevolution theory, and the error-mini-
mization theory. The first two assume that the genetic code
originated respectively from chemical affinities and from
metabolic relationships between codons and amino acids.
The error-minimization theory maintains that in primitive
systems the apparatus of protein synthesis was extremely
prone to errors, and postulates that the genetic code
evolved in order to minimize the deleterious effects of the
translation errors. This article describes a fourth theory
which starts from the hypothesis that the ancestral genetic
code was ambiguous and proposes that its evolution took
place with a mechanism that systematically reduced its
ambiguity and eventually removed it altogether. This pro-
posal is distinct from the stereochemical and the coevolu-
tion theories because they do not contemplate any
ambiguity in the genetic code, and it is distinct from the
error-minimization theory because ambiguity-reduction is
fundamentally different from error-minimization. The
concept of ambiguity-reduction has been repeatedly men-
tioned in the scientific literature, but so far it has remained
only an abstract possibility because no model has been
proposed for its mechanism. Such a model is described in
the present article and may be the first step in a new
approach to the study of the evolution of the genetic code.
Keywords Code ambiguity Code evolution
Codepoiesis Genetic code Ribosomal proteins
Ribosomes
The Apparatus of Protein Synthesis
In 1946, Jean Brachet argued that ‘‘DNA is primarily
confined to the nucleus, while RNA is mainly found in the
cytoplasm, and protein synthesis is associated with RNA.’’
More precisely, Brachet (1944,1946) argued that protein
synthesis takes place on heavy cytoplasmic particles that he
called ribonucleoprotein granules.
Shortly afterwards, Boivin and Vendrely (1947) carried
the scheme of Brachet to its logical conclusion and pro-
posed that ‘‘DNA makes RNA makes Proteins.’’ This 1947
version of the central dogma of molecular biology was
largely ignored, but the idea was born and the evidence
started accumulating.
In 1952, Alexander Dounce made another revolutionary
proposal. He suggested that there must be a code of cor-
respondence between nucleic acids and proteins, and since
there are 20 amino acids but only four nucleotides, he
proposed that each amino acid is coded by a group of three
nucleotides (a codon). On top of that, Dounce (1952,1953)
proposed that the attachment of the 20 canonical amino
acids to nucleotides is promoted by 20 specific catalysts
that he called activating enzymes. Since their discovery,
these molecules have been called aminoacyl-tRNA-syn-
thetases, or, more briefly, synthetases.
Between 1946 and 1952, in short, the basic concepts of
molecular biology had already been formulated but were
largely ignored. The turning point came in 1953, when
James Watson and Francis Crick proposed the model of the
double helix, the idea that in one swift stroke illuminated
&Marcello Barbieri
brr@unife.it
1
Department of Morphology and Embryology, University of
Ferrara, Ferrara, Italy
123
Biol Theory (2015) 10:301–310
DOI 10.1007/s13752-015-0225-z
the structure of DNA and suggested that hereditary infor-
mation is carried by linear sequences of nucleotides
(Watson and Crick 1953).
In 1957 Francis Crick argued there must be intermediary
molecules between nucleotides and amino acids, molecules
that are necessarily made of RNA because they must be able to
recognize a codon by a complementary triplet of nucleotides
that he called anticodon.Crick(1957) called these molecules
adaptors, but in that same year they were discovered by
Hoagland et al. (1957) and became known as transfer RNAs.
A year later, Crick (1958) re-proposed the central
dogma, and this time the idea was immediately accepted:
DNA makes RNA (transcription) and RNA makes proteins
(translation). In the same year, Roberts (1958) gave the
name ribosomes to the molecular machines that make
proteins (the ribonucleoprotein granules described by
Brachet), and the conclusion that the ribosome is the
decoder of genetic information acquired the status of an
experimental fact.
Ribosomes account for more than 80 % of the total
RNA of a cell, and in the 1950s it was taken for granted
that the information of the genes is transported by riboso-
mal RNAs, but things turned out very differently. In 1961,
Franc¸ois Jacob and Jacques Monod proved that the carriers
of genetic information are a completely different family of
RNA molecules that they called messenger RNAs (Jacob
and Monod 1961). Later on, it was discovered that the
scanning of the messenger RNAs requires a whole new
battery of enzymes that were called initiation,elongation,
and termination factors (Nomura et al. 1974).
The apparatus of protein synthesis, in conclusion, con-
sists of ribosomal RNAs, messenger RNAs, transfer RNAs,
ribosomal proteins, aminoacyl-synthetases, scanning fac-
tors, and amino acids. It is a huge supramolecular system
made of more than 120 different types of molecules.
What is most extraordinary, however, is that the rules of
the genetic code are virtually identical in all living crea-
tures (Hinegardner and Engelberg 1963; Woese et al.
1964). A few exceptions do exist but they are very minor
changes and occur in an infinitesimal number of organisms.
The genetic code, in other words, is virtually universal, and
this means that it has been transmitted to all forms of life
by a population of primitive systems that has become
known as the common ancestor. Ever since this discovery,
the origin of the genetic code has become one of the
greatest problems of biology.
Characteristics of the Code
A messenger RNA is scanned by a ribosome in groups of
three nucleotides, called codons, and every codon is rec-
ognized by the anticodon of a transfer RNA (tRNA) that
carries one of the 20 canonical amino acids. The total
number of codons that can be obtained with four nucleo-
tides (A, U, C, G) is 64 (4
3
), and the rules of correspon-
dence between 64 codons and 20 amino acids represent,
collectively, the genetic code.
In 1961, Nirenberg and Matthaei announced that an
artificial messenger containing only uracil (U) is translated
into a protein that contains only phenylalanine (Phe), which
means that the codon UUU codes for phenylalanine
(Niremberg and Matthaei 1961). They had deciphered the
first code word of the genetic code. Other code words were
identified with artificial messengers made of nucleotides
arranged in various orders (Speyer et al. 1963). It was
found, for example, that a sequence of alternating uracil
and cytosine (UCUCUCUC) codes for a polypeptide
made of alternating serine and leucine (Ser-Leu-Ser-
Leu), thus proving that UCU is a codon for serine and
CUC is a codon for leucine (Niremberg and Leder 1964;
Nishimura et al. 1965).
Various other techniques were designed to unlock the
meaning of the other codons, and by 1966 the genetic code
was completely deciphered (Khorana et al. 1966; Nirem-
berg et al. 1966).
It turned out that 61 codons code for amino acids, and
that one of them (usually AUG) is also used as a start
signal, whereas the remaining three (UAA, UAG, and
UGA) are termination signals. Between 61 codons and 20
amino acids there is necessarily a many-to-one corre-
spondence, and this is expressed by saying that the genetic
code is degenerate (or redundant). More precisely, some
amino acids are specified by six codons, some by four,
others by two, and only two amino acids are coded by a
single codon. The genetic code is therefore redundant but
not ambiguous because any one of the 61 codons codes for
one and only one amino acid.
In principle, this implies that every cell contains 61
different types of tRNAs, one for each codon, but in
practice the actual number is about 40 per cell. The best
explanation for this surprising fact was proposed by Crick
(1966) with what has become known as the wobble
hypothesis. Crick pointed out that the three nucleotides of
an anticodon stick out like fingers from the surface of their
tRNA and this allows them to oscillate, or wobble. The
result is that a nucleotide in an anticodon (especially in
third position) can form a temporary bond not only with its
complementary nucleotide but also with another one with
which it has a partial similarity. A uracyl in third position,
for example, can form a bond not only with adenine but
also with guanine, and in this case its tRNA can associate
the same amino acid to two distinct codons. This means
that only one tRNA, rather than two, is sufficient when two
codons specify the same amino acid, and this is indeed
what happens in various cases. It has been found, for
302 M. Barbieri
123
example, that the codons that end with adenine (XYA) and
those that end with guanine (XYG) code for the same
amino acid, and the same is true for the codons that end
with uracyl (XYU) and with cytosine (XYC). Lisine, for
example, is codified by AAA and AAG, whereas tyrosine is
codified by UCU and UAC.
There are, in short, regularities in the genetic code that
allow a cell to carry far less than 61 different types of
tRNAs, and that have, as we will see, important biological
effects. This implies that the modern code has been the
result of an evolutionary process, but first we must address
a preliminary question: what was the starting point of that
process?
The Ancestral Genetic Code
The ribosomal RNAs are among the most conserved
molecules in evolution (Woese 1987,2000) and this means
that they appeared very early on the primitive Earth. It is
also known that they contain regions that have the ability to
form peptide bonds (Nitta et al. 1998), and this means that
some primitive ribosomal RNAs could stick amino acids
together at random and produce statistical proteins. These
proteins did not have biological specificity but could still
be useful because the RNAs can barely work on their own.
They need amino acids and peptides to maintain
stable conformations, and their functions are greatly
enhanced by the attachment of small proteins (Orgel 1973).
This is why an apparatus of protein synthesis started
evolving from pieces of ribosomal RNAs, possibly stabi-
lized by random polypeptides.
The next step in the evolution of this apparatus was the
acquisition of transfer RNAs, molecules that have the
ability to carry amino acids to the site of protein synthesis.
All modern transfer RNAs are small molecules (75–90
nucleotides long) with a basic cloverleaf structure that has
been highly conserved in evolution, which strongly sug-
gests that they descended from a common ancestor. The
contribution of these molecules to protein synthesis, on the
other hand, was greatly enhanced by a third type of RNAs,
because at the site of synthesis it is necessary that the
amino acids be kept in place for a long enough time to
allow the formation of a peptide bond (Wolf and Koonin
2007;Fox2010). This means that the transfer RNAs
required temporary anchoring sites, and in primitive sys-
tems these were provided by anchoring RNAs, the ances-
tors of the messenger RNAs (Osawa 1995).
The combination of ribosomal RNAs, transfer RNAs, and
anchoring RNAs gave origin to an apparatus of protein syn-
thesis where the transfer RNAs were automatically creating a
bridge, or a mapping, between codons and amino acids, and
any such mapping is, by definition, a genetic code.
We realize in this way that the genetic code appeared on
Earth when transfer RNAs and anchoring RNAs joined the
ribosomal RNAs and became an integral part of the
apparatus of protein synthesis. Here, this first code is
referred to as the ancestral genetic code.
But what type of code was it? We know that the modern
code is nonambiguous because any one of its codons codes
for one and only one amino acid, but what about the
ancestral code? It has been underlined that that code was
almost certainly ambiguous because at such an early stage
nothing could prevent a codon from coding for two or more
amino acids (Fitch and Upper 1987; Osawa 1995). This
means that a sequence of codons was translated sometimes
into one protein and at other times into a different protein,
and the apparatus was inevitably producing statistical
proteins.
It is a fact, on the other hand, that a fully nonambiguous
genetic code did appear on the primitive Earth, and this
means that the ambiguity of the ancestral code was steadily
reduced until it reached a point in which any codon could
code for one and only one amino acid. When that happened
the first nonambiguous code came into existence, a code
that here is referred to as the ancient genetic code.
The transition from ancestral to ancient genetic code
transformed the early systems based on statistical proteins
into the first systems that were producing specific proteins,
but how did it happen?
Koonin and Novozhilov (2009) have shown that today
there are three major theories on the origin and the evolution
of the genetic code—the error-minimization theory, the
stereochemical theory, and the coevolution theory—and we
need therefore to examine the mechanisms that they propose.
The Error-Minimization Theory
The translation of a sequence of nucleotides into a
sequence of amino acids is subject to a variety of errors that
can be studied in vitro. In particular, they have been studied
by experiments where protein synthesis takes place in
different environmental conditions and with a variety of
artificial messengers. The most used messenger is the poly-
U molecule where all codons are UUU and all amino acids
should be phenylalanine (Phe). In this case, the appearance
of any other amino acid can only be the result of a trans-
lation error, and it is possible therefore to make a statistical
study of these errors. The overwhelming result of these
studies is that the error rate in the third position of a codon
is about 100 times greater than that in the first position,
which in turn is about ten times greater than the error rate
in the second position. The third position, in other words, is
the most error prone, whereas the second position is the
most stable.
Evolution of the Genetic Code: The Ribosome-Oriented Model 303
123
On the basis of these results, in 1965 Carl Woese pro-
posed a theory on code evolution that consisted in two
main concepts:
(1) The first is the idea that in ancestral systems ‘‘the
translation mechanism was a far more rudimentary
thing than at present, in particular far more prone to
make translation errors. We shall assume that
errors in translation were extreme, to such an extent
that the probability of translating correctly any given
messenger-RNA was essentially zero. From this
concept of error-ridden translation in the primitive
cell it follows that the proteins produced by any
given gene will have to be statistical proteins’
(Woese 1965, p. 1548).
(2) The second concept is the idea that the codons of the
ancestral code were ‘‘readjusted’’ or ‘‘reallocated’’ in
order to minimize the effects of the translation
errors. More precisely, the ancestral code evolved in
such a way that the codon resulting from a transla-
tion mistake would code either for the same amino
acid or for an amino acid with very similar chemical
properties (Woese 1965).
This has become known as the ‘‘error-minimization
theory,’’ but it must be underlined that what is minimized are
not the translation errors but their biological effects. Woese
pointed out that the translation apparatus did improve its
performance in the course of evolution, and became less and
less prone to errors, but that was an altogether different
process, and one that necessarily took place at a later stage.
More precisely, Woese argued that it would be tautolog-
ical to say that the primitive cells evolved a more efficient
translation apparatus by learning to translate more effi-
ciently. ‘‘The way out of this paradox is that although unable
to reduce the translation error rate, the primitive cell can do
something tantamount to this by adjusting the code so that the
‘effect’ of the translation errors is lessened’’ (1965, p. 1549).
Hence the idea that the primitive cells had first to evolve
a genetic code that could minimize the effects of the
translation errors, and only after that could they start
improving the translation apparatus. Woese proposed in
this way a theory where ‘the evolution of the genetic
code starts with a primitive cell possessing random,
ambiguous codon assignments, and a very error-ridden
translation process, and it was the ‘necessity’ of minimiz-
ing the effects of translation errors that led to the highly
ordered code that we observe today’’ (1965, p. 1550).
Woese acknowledged that a similar theory was proposed
by Sonneborn (1965) with the idea that the genetic code
evolved in order to minimize the lethal effects of ordinary
mutations, and the concept of ‘‘error minimization’’ was
extended to all errors deriving from translation mistakes
and from mutations.
At a later stage, Woese proposed that horizontal gene
transfer is another powerful mechanism of code evolution
and suggested that it was most probably that mechanism
that was responsible for the near universality of the modern
genetic code (Woese 2002; Vetsigian et al. 2006).
The Stereochemical Theory
In 1954, George Gamow proposed the stereochemical
hypothesis, the idea that the amino acids fit with a lock-
and-key mechanism into ‘‘holes’’ formed by four nucleo-
tides, and that it is the three-dimensional shape of each hole
that determines which amino acid binds to which quartet of
nucleotides. According to this diamond code (Gamow
1954) the rules of the genetic code are the result of
chemical affinities between codons and amino acids and
are therefore determined by chemistry.
Gamow’s model was quickly abandoned when it became
clear that in protein synthesis there are no direct contacts
between codons and amino acids, but the idea of stereo-
chemical interactions between them was not discarded and has
been re-proposed ever since in many different forms.
Pelc and Weldon (1966) suggested that there is a
stereochemical complementarity between amino acids and
their codons; Dunnill (1966) argued that the anticodon loop
of the transfer RNA forms a molecular pocket in which the
amino acid can be trapped; Melcher (1974) proposed that
amino acids have a stereochemical correlation with their
anticodons; and Shimizu (1982) maintained that a com-
plementary relationship exists between amino acids and
groups of four nucleotides in the transfer RNAs.
It must be underlined that the interactions between
nucleotides and amino acids are beyond dispute. The
negative charges of the nucleotide phosphates attract the
positive charges of the basic amino acids and this gives
origin to countless interactions between them. The crucial
point is that these interactions take place between any
nucleotide and any amino acid and in no way account for
the specificity of the coding rules. The stereochemical
theory is the idea that in addition to the standard chemical
interactions there are also specific affinities between
codons and amino acids, and the history of this theory has
been but a long journey in search of such alleged affinities.
A systematic study on chemical affinities was conducted
by Saxinger et al. (1971) by filtering nucleotides on gels
containing amino acids and by measuring the quantities of
amino acids that were selectively retaining triplets of
nucleotides. The results were disappointing and indicated
that there are at best very weak specific interactions
between amino acids and codons.
The stereochemical theory was revived again when
Yarus discovered that there is a selective interaction
304 M. Barbieri
123
between the amino acid arginine and a nucleotide that is
present in all four codons that code for arginine (Yarus
1988,1998). Later on Yarus and colleagues extended the
research to eight amino acids and reported other positive
correlations (Yarus et al. 2005), but in these cases the
evidence was much weaker and arginine remained an iso-
lated exception.
Another difficulty for the stereochemical theory is its
potential conflict with the error-minimization mechanism.
If it is true that codons can be reallocated to different
amino acids in order to minimize the effects of translation
errors, one is bound to conclude that there are no specific
chemical affinities between them. Yarus replied to this
objection with the argument that only some coding rules
are determined by stereochemistry whereas others are free
to change and allow the system to mitigate the effects of
mutations and translation errors (Yarus et al. 2005).
After decades of research, in conclusion, there still is no
real evidence in favor of the stereochemical theory, and
what keeps it alive is the possibility that stereochemical
interactions between codons and amino acids might have
been important at some early stages of evolution (Koonin
and Novozhilov 2009).
The Coevolution Theory
The origin of the genetic code is still a mystery, but we do
have some interesting clues. The first is that the number of
amino acids changed during the early evolution of the
code. This is because only less than half of the 20 canon-
ical amino acids can be synthesized from inorganic mole-
cules and for this reason are referred to as ‘‘primary’’ (or
‘precursor’’) amino acids. The others are always synthe-
sized taking primary amino acids as starting points and are
referred to as ‘‘secondary’’ (or ‘‘product’’) amino acids.
The crucial point is that only primary amino acids are
produced in laboratory experiments that simulate prebiotic
conditions, which strongly suggests that less than ten pri-
mary amino acids appeared on the primitive Earth (Wong
and Bronskill 1979). This conclusion is supported by the
discovery that the amino acids that are missing in labora-
tory syntheses are also missing from meteorites (Higgs and
Pudritz 2007). The implication is that the evolution of the
genetic code started with less than ten amino acids and
went all the way up until it reached the canonical set of 20
that has been strongly conserved ever since.
The second important clue on the genetic code is that the
codons assigned to the secondary amino acids differ very
little from the codons of their precursors amino acids. The
best explanation of this pattern, so far, is the theory proposed
by Jeffrey Wong, the so-called coevolution theory of genetic
code and amino acid biosynthesis (Wong 1975,1981).
Wong proposed that the secondary amino acids received
their codons from the primary amino acids that served as
precursors in their biosynthesis. This theory predicts that
‘the codons of precursor-product amino acids should be
contiguous, i.e., separated by only a single base change’’
(1975, p. 1909), and in many cases this is what is actually
observed.
Wong concluded that the first genetic code that appeared
on Earth was codifying only primary amino acids, and all
codons were assigned to them. Later on, during the evo-
lution of the genetic code, the secondary amino acids
steadily increased in number by new biosynthetic pathways
and received their codons from the primary amino acids
that served as precursors in their biosynthesis.
The coevolution theory proposed by Wong starts from a
genetic code where all codons are assigned to less than ten
primary amino acids, but does not say how this first code
came into existence. This issue was taken on by Di Giulio
(2008) who proposed an ‘‘extension’’ of the original theory
that applies the same mechanism also to the primary amino
acids of the first genetic code.
The key idea of the coevolution theory, in both the
original and the extended form, is the hypothesis that the
codon of any secondary amino acid comes from the
codons assigned to its precursor amino acid because in the
early stages of evolution the synthesis of amino acids was
taking place on transfer RNAs, and there was therefore a
metabolic continuity between RNAs and amino acids. In
this case, the relationships between codons and amino
acids are no longer due to chemical affinities, as in the
stereochemical theory, but continue to be deterministic
relationships because they are dictated by metabolic
reactions.
The problem with the coevolution theory is that the
synthesis of amino acids could have taken place on RNAs
in the RNA-world but this is certainly no longer true in the
protein world. When the amino acids ceased to be syn-
thesized on RNAs, the ancient metabolic connections
between codons and amino acids disappeared and with
them disappeared the rules of the ancestral genetic code, so
how did these rules reemerge in the protein world? The
coevolution theory, in other words, might account for the
origin of the genetic code in the RNA world, but leaves us
with the huge problem of understanding how the same
coding rules reappeared in the protein world.
Arbitrariness
The stereochemical theory and the coevolution theory
assume that the genetic code originated respectively from
chemical affinities and from metabolic relationships
between codons and amino acids, and in both cases it
Evolution of the Genetic Code: The Ribosome-Oriented Model 305
123
would not be a real code because its rules would not have
the arbitrariness that characterizes all true codes.
The crucial point, here, is the relationships that exist
between the recognition of the amino acids, performed by
the synthetates (aminoacyl-tRNA-synthetases) and the
recognition of the codons performed by the anticodons of
the transfer RNAs. If the synthetases could recognize both
the amino acids and the anticodons, they would establish
direct connections between them and the rules of the
genetic code would be deterministic, but this is precisely
what the evidence has ruled out. Experiments have shown
that the recognition of the amino acids is independent
from the recognition of the anticodons because in many
cases the synthetases have no access to the anticodons
(Schimmel 1987; Schimmel et al. 1993). On top of that, it
has been shown that the links between codons and amino
acids could have been made in countless different ways.
Hou and Schimmel (1988), for example, managed to
introduce two extra nucleotides in a tRNA without
changing its anticodon, and found that that the resulting
tRNA was recognized by a different synthetase and was
carrying therefore a different amino acid. They had
changed one of the rules of the genetic code, a result that
achieved in vitro what a few microorganisms have
achieved in vivo in the course of evolution (Jukes and
Osawa 1990,1993).
The lesson that comes from these experiments is that the
rules of the genetic code are the result of interactions
between synthetases and tRNAs that can be modified vir-
tually at will by adding or subtracting a few molecules.
This means that the number of adaptors between codons
and amino acids is potentially unlimited, and only the
selection of a fixed number of them can ensure a specific
correspondence. It also means that the rules of the genetic
code are not dictated by any form of chemical necessity,
and in this sense they are arbitrary.
It is worth mentioning, at this point, a fairly widespread
argument according to which the rules of the genetic code
cannot be arbitrary because they have been optimized in
the course of evolution. In reality, the two things are not
incompatible. The rules of the Morse code, for example,
have been optimized by associating the most frequent let-
ters of the alphabet with the shortest combinations of dots
and dashes, and yet they continue to be arbitrary rules. An
optimization of the coding rules, in other words, simply
means that they are not random, and is perfectly compat-
ible with their arbitrariness.
We reach in this way the conclusion, first expressed by
Monod (1970), that the genetic code is chemically arbitrary
because its rules are not dictated by necessity. The genetic
code, in short, is a real code and this makes the problem of its
origin all the more challenging and interesting.
The Missing Theory in Code Evolution
If we admit (1) that the ancestral genetic code was
ambiguous and (2) that the ancestral apparatus of protein
synthesis was error prone, we conclude that the primitive
systems contained two different types of statistical pro-
teins: those produced by code ambiguity and those pro-
duced by translation errors. This in turn implies that the
evolution of the genetic code took place by two distinct
mechanisms, one that reduced code ambiguity and one that
minimized the effects of translation errors. We should have
therefore two distinct theories on code evolution, but what
we have had up to now is only the error-minimization
framework. So far, the ambiguity-reduction framework has
remained an abstract possibility and can rightly be regarded
as the missing theory in code evolution.
One may be tempted to suggest that ambiguity reduction
can be included in the error-minimization category, but this
is not the case. Error minimization implies that some
associations between codons and amino acids are normal
and others are the result of errors, whereas in code ambi-
guity all associations between transfer RNAs and amino
acids are equally ‘‘normal.’’ A solution of the error-mini-
mization problem, in other words, is in no way a solution of
the ambiguity-reduction problem. A genetic code can
continue to be ambiguous even when the translation
apparatus has become completely error free.
In his 1965 paper, Woese underlined that the ancestral
genetic code was necessarily ambiguous, but then he
concentrated on error minimization only and did not
mention a separate mechanism for the reduction in code
ambiguity. Some forty years later, Woese and collaborators
underlined again that codon ambiguity has nothing to do
with errors, and said so in no uncertain terms: ‘‘Ambiguity
is therefore not the same thing as error’’ (Vetsigian et al.
2006, p. 10696). That seminal paper, however, was dedi-
cated to the role of horizontal gene transfer in the evolution
of the genetic code, and the reduction in ambiguity was
mentioned only as something that necessarily took place
but whose mechanism is still unknown.
The ambiguity-reduction concept has remained in this
way a missing theory, a project for the future, and it is high
time therefore that we try to address it. More precisely, that
we try to figure out the mechanism that brought it about. To
this purpose, it may be useful to recall the ‘‘little parable’
of the hotel keys proposed by Nino (1982).
In any hotel there are two types of keys: the familiar
keys that open individual doors and the passkey that opens
all doors. At first, one may be forgiven for thinking that the
passkey is the most complex of all, whereas the truth is
precisely the other way round. The passkey is the simplest
one because what is complex in a key is not the ability to
306 M. Barbieri
123
open a door (that can easily be done with a screwdriver) but
the ability to open one door and not all the others.This
suggests an interesting parallel with the evolution of the
genetic code.
An ancestral ribozyme might have been able to attach
amino acids to all transfer RNAs—like a passkey that
opens all doors—thus giving origin to a completely
ambiguous genetic code. A reduction in the ambiguity of
this ancestral code would have been achieved by evolving
ribozymes that became capable of attaching fewer and
fewer amino acids to any transfer RNA, until the point was
reached when a ribozyme could deliver to any transfer
RNA one and only one amino acid.
The problem is: why did this happen? What was the
driving force that fueled a systematic reduction of ambi-
guity in the ancestral genetic code?
The Ribosome-Oriented Model
Ribosomes consist of a small and a large subunit that are
made of ribosomal.
RNAs and ribosomal proteins. The ribosomal-RNAs
account for more than half of the huge molecular weight of
the ribosomes (over 2 million), and the rest is accounted for
by more than 50 different ribosomal proteins of low
molecular weight, each present in one or a few copies
(Nomura et al. 1974). These proteins are the descendants of
the statistical proteins of the ancestral ribosomes, but how
did they evolve? Today we have at least two important
clues about this major transition.
The first comes from the fact that there was an evolu-
tionary advantage in increasing the total number of ribo-
somal proteins irrespective of their individual
characteristics. The reason comes from a general principle
in engineering that Burks (1970) expressed in this way:
‘there exists a direct correlation between the size of an
automaton—as measured roughly by number of compo-
nents—and the accuracy of its function.’’ In our case, this
principle means that increasing the number of ribosomal
proteins was making the ribosomes more heavy, more
resistant to thermal noise, and therefore more reliable in
protein synthesis.
The second clue comes from the fact that ribosomes are
formed by self-assembly from their components, and it has
been possible to discover the contribution of individual
ribosomal proteins by studying what happens when ribo-
somes are reassembled without any one of them in turn.
These experiments have shown that the ribosomal proteins
fall into three major groups: some are necessary for func-
tion, others are required for self-assembly, and those of the
third group have a stimulating effect but are fundamentally
disposable (Kurland 1970;Fox2010).
The evolution of the ribosomal proteins was therefore a
process that gave origin first to three great families and
then to an increasing number of subfamilies. These sub-
families, on the other hand, could be transmitted to future
generations only if the ambiguity of the genetic code was
lower than the statistical differences between them. The
ambiguity of the code, in other words, was the limiting
factor that determined how many subfamilies of statistical
ribosomal proteins could reappear in the descendants.
Which means that only by decreasing the ambiguity of the
ancestral genetic code was it possible to promote the
evolution of the ribosomal proteins—an evolution that
increased their number, that diversified their functions, and
that favored the reappearance of increasingly similar types
of ribosomal proteins in the descendants because this made
it easier for ribosomes to self-assemble in every new
generation.
It will be noticed that the evolution of the ribosomal
proteins was not about this or that protein or this or that
protein function. It involved all statistical proteins at the
same time. It was not about individual features but about
collective relationships. It was an evolution that went on
until the ambiguity of the genetic code was completely
removed and biological specificity came into existence.
It is possible, in other words, that the evolution of the
ancestral genetic code was ribosome oriented. But do we
have any evidence of this? Luckily we do. If the evolution
of the ancestral genetic codes was in function of the ribo-
somal proteins, we should find that these proteins were the
first specific proteins that appeared on the primitive Earth,
and the evidence does seem to support this conclusion.
The last common ancestor gave origin to the cells of the
three primary kingdoms (Archaea, Bacteria, and Eukarya),
and it has been shown that most ribosomal proteins are
present in all kingdoms (Woese 2002; Fox 2010). This
means that the primary kingdoms received from the last
common ancestor not only a universal genetic code, but
also a set of universal ribosomal proteins. Which in turn
means that the evolution of these proteins had already
taken place when the last common ancestor came into
being. The molecular trees, on the other hand, do not reveal
the existence of older proteins, and this strongly suggests
that the ribosomal proteins were indeed the first specific
proteins that appeared on the primitive Earth.
The Modern Genetic Code
The modern genetic code is a mapping between 64 codons
carried by transfer RNAs and 20 amino acids carried by 20
aminoacyl-tRNA-synthetases, each of which attaches one
amino acid to one or more tRNAs. The synthetases are
specific proteins that can be produced only by an apparatus
Evolution of the Genetic Code: The Ribosome-Oriented Model 307
123
that already has a genetic code, and this gives us a classic
chicken-and-egg paradox: how could the genetic code
come into existence if its rules are implemented by proteins
that can be made only when the code already exists?
A possible solution to this paradox is that the modern
apparatus of protein synthesis was preceded by an ancient
apparatus where the amino acids were attached to the
transfer RNAs not by proteins but by RNAs (Maizels and
Weiner 1987). The modern genetic code, in other words,
was preceded by an ancient genetic code based on RNA
synthetases that were later replaced by protein synthetases.
Such a replacement, on the other hand, was bound to have
major biological consequences. Proteins can mimic RNAs
but only up to a point, and replacing the RNA synthetases
with protein synthetases could well have modified the rules
of the ancient genetic code. But did that actually happen?
Evidence in support of the idea that the ancient genetic
code was repeatedly modified has come from computer
studies that suggest that the modern genetic code performs
better than most of its many potential alternatives (Haig
and Hurst 1991; Freeland and Hurst 1998; Bollenbach et al.
2007). Gilis et al. (2001) have shown that the modern code
is optimal with respect to the stabilization of protein
structure; Itzkovitz and Alon (2007) have argued that the
modern code is nearly optimal for the acquisition of
additional information into genetic sequences, whereas
Drummond and Wilke (2008) have suggested that the
modern code is ideally suited to favor the process of pro-
tein folding. It must also be reported, however, that some
authors have warned against reaching overoptimistic con-
clusions about this issue. Novozhilov et al. (2007) have
pointed out that there are 10
84
possible codes and that a
huge number of them are more robust than the modern
code. Their computer simulations revealed that the genetic
code did go through processes of optimization but appar-
ently went only halfway up the optimality ladder.
Altogether, the computer simulation data leave little
doubt that the genetic code was, at least partially, opti-
mized, but it is unlikely that the optimization process was
conducted on countless proteins. A more realistic scenario
is that it was optimized in a group of 50 or so ribosomal
proteins, and once it was optimized for them it was adopted
without further changes for all other proteins.
It is possible, in conclusion, that the step-by-step intro-
duction of protein synthetases in the apparatus of protein
synthesis provided the means for optimizing its perfor-
mance until the point was reached when the accuracy of
protein synthesis became so high as to be virtually error
free.
This suggests that the two historical evolutions of the
genetic code were both ribosome oriented: the evolution
from ancestral to ancient genetic code was driven by the
ribosomal proteins, whereas the evolution from ancient to
modern genetic code was driven by the synthetase proteins
(Barbieri 2015).
The Conservation of the Genetic Code
The modern genetic code appeared on Earth before the first
cells of the three primary kingdoms and has been highly
conserved ever since (Woese 2000,2002). Today, this
extraordinary process of conservation is usually explained
by saying that the genetic code is a set of constraints
(Pattee 2001) and that physical constraints cannot be
changed, an idea that appears to explain why the genetic
code has been ‘‘frozen’’ since the origin of life.
The conclusion that the genetic code is a set of con-
straints is formally correct because a code is indeed a set of
rules that impose limitations on a virtually unlimited
number of possibilities. It must be underlined, however,
that the rules of the genetic code are biological constraints,
not physical ones.
They are biologically generated rules and in no way can
be assimilated to physical constraints. This is because the
genes of the genetic code are constantly subject, like all
other genes, to mutation and neutral drift. They are in a
continuous state of flux and the fact that they have been
highly conserved in evolution means that there is a bio-
logical mechanism that actively and continuously restores
their original structure. The conservation of the genetic
code, in other words, is not the passive result of physical
constraints. It can only be the result of an active biological
mechanism that is continuously at work, a mechanism that
has been referred to as codepoiesis (Barbieri 2012).
The concept of autopoiesis, or self-production, describes
the ability of living systems to produce their own compo-
nents and eventually to generate copies of themselves.
Before the genetic code, however, specific proteins did not
exist, and the ancestral systems were producing descen-
dants that were inevitably different from themselves.
Autopoiesis, in short, did not exist before the first cells, so
it was not the mechanism that gave origin to them.
The ancestral apparatus of protein synthesis was
engaged in the process of evolving coding rules and was
therefore a code-generating system. After the origin of the
genetic code the situation completely changed, and the
system in question became a code-conservation system.
Another part of the living systems, however, maintained
the potential to evolve other coding rules and behaved as a
new code-generating, or code-exploring, system. In the
early Eukarya, for example, the cells had a code-conser-
vation part for the genetic code, but also a code-exploring
part for the splicing code. The evolution of the first cells, in
other words, was based on two complementary processes:
one was the generation of new organic codes and the other
308 M. Barbieri
123
was the conservation of the existing ones. Taken together,
these two processes are the two sides of codepoiesis.
The ancestral systems, in conclusion, were not
autopoietic systems but they had to be codepoietic sys-
tems. And all cells that came after them were not always
engaged in autopoiesis but were inevitably engaged in
codepoiesis. This is the great message of the conserva-
tion of the genetic code: what is always and necessarily
present in all living systems is codepoiesis, not
autopoiesis.
Conclusions
The history of the genetic code can be divided into four
great phases: (1) the origin of the ancestral code, (2) the
evolution from ancestral to ancient code, (3) the evolution
from ancient to modern code, and (4) the conservation of
the modern code. The present article takes as a starting
point the hypothesis that the ancestral code was ambiguous
and proposes that the second of those four major transi-
tions—the evolution from ancestral to ancient genetic
code—took place with a mechanism that gradually reduced
the ambiguity of the ancestral code and eventually removed
it completely, a mechanism that here is described, at least
in first approximation, by the ribosome-oriented model.
The ambiguity-reduction theory and the ribosome-ori-
ented model have implications also for the other evolu-
tionary phases of the genetic code. On the third phase—the
evolution from ancient to modern code—the implication is
that the synthetase proteins had a driving role similar to
that of the ribosomal proteins in the second phase. About
the fourth phase—the conservation of the modern genetic
code—it is argued that the underlying mechanism is
probably far more complex than that of a frozen accident.
As for the first phase—the origin of the ancestral code—the
implication is that the result of that process was an am-
biguous code.
The ribosome-oriented model describes a mechanism
that accounts for a steady reduction of ambiguity in the
evolution of the genetic code, a process that ended with the
origin of biological specificity, the very hallmark of life as
we know it. The model is likely to require further devel-
opments, but the important point is that ambiguity reduc-
tion is no longer an abstract possibility. Perhaps the most
significant implication of this article is the fact that the
ribosome-oriented model describes a mechanistic approach
to the problem of the origin of coding (and therefore of
meaning) in a population of primitive systems based on
statistical proteins.
Acknowledgments I am indebted to two anonymous referees whose
comments greatly improved the initial version of this manuscript.
References
Barbieri M (2012) Codepoiesis: the deep logic of life. Biosemiotics
5:297–299
Barbieri M (2015) Code biology: a new science of life. Springer,
Dordrecht
Boivin A, Vendrely R (1947) Sur le ro
ˆle possible deux acides nucleic
dans la cellule vivant. Experientia 3:32–34
Bollenbach T, Vetsigian K, Kishony R (2007) Evolution and
multilevel optimization of the genetic code. Genome Res
17:401–404
Brachet J (1944) Embriologie chimique. Masson et Cie, Paris
Brachet J (1946) Nucleic acids in the cell and the embryo. Symp Soc
Exp Biol 1(213–215):222
Burks AW (1970) Essays on cellular automata. University of Illinois
Press, Urbana
Crick FHC (1957) The structure of nucleic acids and their role in
protein synthesis. Biochem Soc Symp 14:25–26
Crick FHC (1958) On protein synthesis. Symp Soc Exp Biol
12:138–163
Crick FHC (1966) Codon-anticodon pairing: the wobble hypothesis.
J Mol Biol 19:548–555
Di Giulio M (2008) An extension of the coevolution theory of the
origin of the genetic code. Biol Direct 3:1–37
Dounce AL (1952) Duplicating mechanism for peptide chain and
nucleic acid synthesis. Enzymologia 15:251–258
Dounce AL (1953) Nucleic acid template hypothesis. Nature 172:541
Drummond DA, Wilke CO (2008) Mistranslation-induced protein
misfolding as a dominant constraint on coding-sequence evolu-
tion. Cell 134:341–352
Dunnill P (1966) Triplet nucleotide-amino-acid pairing; a stereo-
chemical basis for the division between protein and non-protein
amino-acids. Nature 210:1267–1268
Fitch WM, Upper K (1987) The phylogeny of tRNA sequences
provides evidence for ambiguity reduction in the origin of the
genetic code. Cold Spring Harb Symp Quant Biol 52:759–767
Fox GE (2010) Origin and evolution of the ribosome. Cold Spring
Harb Perspect Biol 2:a003483
Freeland SJ, Hurst LD (1998) The genetic code is one in a million.
J Mol Evol 47:238–248
Gamow G (1954) Possible relation between deoxyribonucleic acid
and protein structures. Nature 173:318
Gilis D, Massar S, Cerf NJ, Rooman M (2001) Optimality of the
genetic code with respect to protein stability and amino-acid
frequencies. Genome Biol 2:41–49
Haig D, Hurst LD (1991) A quantitative measure of error minimiza-
tion in the genetic code. J Mol Evol 33:412–417
Higgs PG, Pudritz RE (2007) From protoplanetary disks to prebiotic
amino acids and the origin of the genetic code. In: Pudritz RE,
Higgs PG, Stone J (eds) Planetary systems and the origins of life,
vol 3., Cambridge series in astrobiology. Cambridge University
Press, Cambridge, pp 62–88
Hinegardner RT, Engelberg J (1963) Rationale for a universal genetic
code. Science 142:1083–1085
Hoagland MB, Zamecnik PC, Stephenson ML (1957) Intermediate
reactions in protein biosynthesis. Biochem Biophys Acta
24:215–216
Hou Y-M, Schimmel P (1988) A simple structural feature is a major
determinant of the identity of a transfer RNA. Nature
333:140–145
Itzkovitz S, Alon U (2007) The genetic code is nearly optimal for
allowing additional information within protein-coding
sequences. Genome Res 17:405–412
Jacob F, Monod J (1961) Genetic regulatory mechanisms in the
synthesis of proteins. J Mol Biol 3:318–356
Evolution of the Genetic Code: The Ribosome-Oriented Model 309
123
Jukes TH, Osawa S (1990) The genetic code in mitochondria and
chloroplasts. Experientia 46:1149–1157
Jukes TH, Osawa S (1993) Evolutionary changes in the genetic code.
Comp Biochem Physiol 106 B:489–494
Khorana HG, Bu
¨chi H, Ghosh H et al (1966) Polynucleotide synthesis
and the genetic code. Cold Spring Harb Symp Quant Biol
31:39–49
Koonin EV, Novozhilov AS (2009) Origin and evolution of the
genetic code: the universal enigma. IUBMB Life 61(2):99–111
Kurland CG (1970) Ribosome structure and function emergent.
Science 169:1171–1177
Maizels N, Weiner AM (1987) Peptide-specific ribosomes, genomic
tags and the origin of the genetic code. Cold Spring Harb Symp
Quant Biol 52:743–757
Melcher G (1974) Stereospecificity and the genetic code. J Mol Evol
3:121–141
Monod J (1970) Le Hasard et la Nece
´ssite
´. Editions du Seuil, Paris.
English edition: Monod J (1971) Chance and necessity (trans:
Wainhouse A). Knopf, New York
Ninio J (1982) Molecular approaches to evolution. Pitman Books,
London
Niremberg M, Leder P (1964) RNA codewords and protein synthesis.
Science 145:1399–1407
Niremberg M, Matthaei H (1961) The dependence of cell-free protein
synthesis in E. coli upon naturally occurring or synthetic
polyribonucleotides. Proc Natl Acad Sci USA 47:1588–1602
Niremberg M, Caskey T, Marshal R et al (1966) The RNA code and
protein synthesis. Cold Spring Harb Symp Quant Biol 31:11–24
Nishimura S, Jones DS, Khorana HG (1965) The in vitro synthesis of
a co-polypeptide containing two amino acids in alternating
sequence dependent upon a DNA-like polymer containing two
nucleotides in alternating sequence. J Mol Biol 13:302–324
Nitta I, Kamada Y, Noda H et al (1998) Reconstitution of peptide
bond formation. Science 281:666–669
Nomura M, Tissie
`res A, Lengyel P (1974) Ribosomes, Cold Spring
Harbor monograph series. Cold Spring Harbor Laboratory Press,
Cold Spring Harbor
Novozhilov AS, Wolf YI, Koonin EV (2007) Evolution of the genetic
code: partial optimization of a random code for robustness to
translation error in a rugged fitness landscape. Biol Direct 2:24
Orgel LE (1973) The origins of life. Wiley, New York
Osawa S (1995) Evolution of the genetic code. Oxford University
Press, New York
Pattee HH (2001) The physics of symbols: bridging the epistemic cut.
BioSystems 60:5–21
Pelc SR, Weldon MGE (1966) Stereochemical relationship between
coding triplets and amino-acids. Nature 209:868–870
Roberts RB (1958) Microsomal particles and protein synthesis.
Pergamon Press, Washington
Saxinger WC, Ponnamperuma C, Woese CR (1971) Evidence for the
interaction of nucleotides with immobilized amino-acids and its
significance for the origin of the genetic code. Nat New Biol
234:172–174
Schimmel P (1987) Aminoacyl tRNA synthetases: general scheme of
structure-function relationship in the polypeptides and recogni-
tion of tRNAs. Ann Rev Biochem 56:125–158
Schimmel P, Giege
´R, Moras D, Yokoyama S (1993) An operational
RNA code for amino acids and possible relationship to genetic
code. Proc Natl Acad Sci USA 90:8763–8768
Shimizu M (1982) Molecular basis for the genetic code. J Mol Evol
18:297–303
Sonneborn TM (1965) Degeneracy of the genetic code: extent, nature,
and genetic implications. In: Bryson V, Vogel HJ (eds) Evolving
genes and proteins. Academic Press, New York, pp 377–397
Speyer J, Lengyel P, Basilio C et al (1963) Synthetic polynucleotides
and the amino acid code. Cold Spring Harb Symp Quant Biol
28:559–567
Vetsigian K, Woese C, Goldenfeld N (2006) Collective evolution and
the genetic code. Proc Natl Acad Sci USA 103:10696–10701
Watson JD, Crick FHC (1953) Molecular structure of nucleic acids. A
structure for deoxyribose nucleic acid. Nature 171:737–738.
Genetical implications of the structure of deoxyribose nucleic
acid. Nature 171:964-967
Woese CR (1965) Order in the genetic code. Proc Natl Acad Sci USA
54:71–75
Woese CR (1987) Bacterial evolution. Microbiol Rev 51:221–271
Woese CR (2000) Interpreting the universal phylogenetic tree. Proc
Natl Acad Sci USA 97:8392–8396
Woese CR (2002) On the evolution of cells. Proc Natl Acad Sci USA
99:8742–8747
Woese CR, Hinegardner RT, Engelberg J (1964) Universality in the
genetic code. Science 144:1030–1031
Wolf YI, Koonin EV (2007) On the origin of the translation system
and the genetic code in the RNA world by means of natural
selection, exaptation, and subfunctionalization. Biol Direct 2:14
Wong JT (1975) A co-evolution theory of the genetic code. Proc Natl
Acad Sci USA 72:1909–1912
Wong JT (1981) Coevolution of genetic code and amino acid
biosynthesis. Trends Biochem Sci 6:33–36
Wong JT, Bronskill PM (1979) Inadequacy of prebiotic synthesis as
origin of proteinous amino acids. J Mol Evol 13:115–125
Yarus M (1988) A specific amino acid binding site composed of
RNA. Science 240:1751–1758
Yarus M (1998) Amino acids as RNA ligands: a direct-RNA-template
theory for the code’s origin. J Mol Evol 47:109–117
Yarus M, Caporaso JG, Knight R (2005) Origins of the genetic code:
the escaped triplet theory. Ann Rev Biochem 74:179–198
310 M. Barbieri
123
... Neither is it indexical reference, for the action of the codons does not depend upon association by contiguity. Employing Barbieri (2015b), we may describe-RNA transcripts of-codons and amino acids, as two independent worlds, which are connected by transfer-RNAs functioning as adaptors. The mapping between codon and amino acid is by a nearly universal, context independent (i.e., fixed) code, named the genetic code. ...
... The mapping between codon and amino acid is by a nearly universal, context independent (i.e., fixed) code, named the genetic code. The signs involved are called symbols, indicating that they are associated with an object by an entirely arbitrary link (Barbieri, 2015b;cf. Pattee, 1972;Keller, 2009;Emmeche, 2011), not by similarity or a physical link. ...
... The way in which arbitrariness in the mapping between codon and amino acid originated, was explicitly discussed by Barbieri (2015a). After a phase with merely "statistical"-i.e. ...
Article
Full-text available
In this article two questions are discussed with regard to semiosis in protein biosynthesis for nano engines. (1) What kind of semiosis is involved in the construction of these proteins? and (2) How can we explain the semiotic process observed? With regard to the first issue we draw attention to comparisons between semiosis in protein biosynthesis and human natural language. The notion of normativity appears to be of great importance for both. A comparison also demonstrates differences. Nevertheless, because of the normative symbolic information processing in it, we suggest to employ the term symbolic reference (employed in linguistics as a distinguishing feature of human language) to indicate the semiotic processes in protein biosynthesis. With regard to explaining semiosis in protein synthesis we compare different approaches. We conclude that a Kantian approach should be preferred. In such an approach strengths of the mechanistic and organicist approaches can be combined, and the observed symbolic information processing acknowledged.
... The structure and properties of the standard genetic (SGC) code have intrigued scientists since the first ascriptions of codons to amino acids were discovered in the sixties of the last century [1,2]. From that time, many hypotheses concerning the origin and evolution of SGC have been proposed (see for review: [3][4][5][6][7][8]). However, it is still unclear which factors played a decisive role in the process of the genetic code emergence and evolution. ...
... This assumption appears to be a good starting point for further investigations of the SGC evolution. It was proposed that these codes evolved together from a set of ambiguous codon assignments towards a coding system with a low level of ambiguity to reduce an initial high translational noise [5,[12][13][14]. During the genetic code evolution, amino acids were gradually added to the code, which was profitable because it increased the diversity of synthesized proteins [6,[15][16][17]. ...
... This assumption seems reasonable because the genetic code at very early stages of evolution was likely a mixture of different reading systems for the set of encoded items and characterized by ambiguity in the coding of amino acids until the whole system became more precise [5,12]. The genetic code could have evolved from a highly ambiguous and non-homogeneous system to an unambiguous system with nearly identical reading mechanisms for each encoded label. ...
Article
Full-text available
The standard genetic code (SGC) is a set of rules according to which 64 codons are assigned to 20 canonical amino acids and stop coding signal. As a consequence, the SGC is redundant because there is a greater number of codons than the number of encoded labels. This redundancy implies the existence of codons that encode the same genetic information. The size and organization of such synonymous codon blocks are important characteristics of the SGC structure whose evolution is still unclear. Therefore, we studied possible evolutionary mechanisms of the codon block structure. We conducted computer simulations assuming that coding systems at early stages of the SGC evolution were sets of ambiguous codon assignments with high entropy. We included three types of reading systems characterized by different inaccuracy and pattern of codon recognition. In contrast to the previous study, we allowed for evolution of the reading systems and their competition. The simulations performed under minimization of translational errors and reduction of coding ambiguity produced the coding system resistant to these errors. The reading system similar to that present in the SGC dominated the others very quickly. The survived system was also characterized by low entropy and possessed properties similar to that in the SGC. Our simulation show that the unambiguous SGC could emerged from a code with a lower level of ambiguity and the number of tRNAs increased during the evolution.
... The early genetic code most likely characterized by a high translational noise, which was further reduced during its evolution (Fitch and Upper, 1987;Barbieri, 2015;Błażej et al., 2019b). Such a state remained likely a long time because the last universal common ancestor of three domains of life, bacteria, archaea and eukaryotes was still a progenote, with not fully developed translational apparatus (Di Giulio, 2001;Giulio, 2014;Di Giulio, 2020a,b). ...
... One of them assumes that newly added amino acids captured codons due to duplications of genes coding for tRNAs and aminoacyl-tRNA synthetases (Cavalcanti et al., 2000(Cavalcanti et al., , 2004Massey, 2015Massey, , 2016Koonin, 2017;Koonin and Novozhilov, 2017). The products of the duplicated genes most likely recognized initially groups of similar codons and there existed an ambiguity in the coding of amino acids until the whole system became more precise (Fitch and Upper, 1987;Barbieri, 2015). ...
Article
It is assumed that at the early stage of cell evolution its translation machinery was characterized by high noise, i.e. ambiguous assignment of codons to amino acids in the genetic code, which initially encoded only few amino acids. Next, during its evolution new amino acids were added to this code. Taking into account this facts, we investigated theoretical models of genetic code’s structure, which evolved from a set of ambiguous codons assignments into a coding system with a low level of uncertainty. We considered three types of translational inaccuracies assuming a different number of fixed codon positions. We applied a modified version of evolutionary algorithm for finding the genetic codes that the most effectively reduced the initial uncertainty in the assignment of codons to encoded labels, i.e. amino acids and a stop translation signal. We examined codes with the number of labels from four to 22. Our results indicated that the quality of genetic code structure is strongly dependent on the number of encoded labels as well as the type of translational mechanism. The more strict assignments of codon to the labels was preferred by the codes encoding more number of labels. The results showed that a smaller degeneracy of codes evolved from a more tolerant coding with the stepwise addition of coded amino acids to the genetic code. The distribution of codon groups in the standard genetic code corresponds well to the translation model assuming two fixed codon positions, whereas the six-codon groups can be relics form previous stages of evolution when the code characterized by a greater uncertainty.
... The very high optimization of the partition energy of amino acids, in particular on the columns of the genetic code ( Tabs. 3 and 4), is in agreement with the hypothesis of Woese (1965Woese ( , 1967) that one of the main selective pressures that would have structured the genetic code would have been the lowering of translational noise. Other forms of reduction of ambiguity hypothesized to be present in some stages of the origin of the genetic code (Fitch and Upper, 1987;Ardell and Sella, 2001;Barbieri, 2015), would also be compatible with the very high optimization found . CC-BY-ND 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. ...
Preprint
Full-text available
We used the Moran’s I index of global spatial autocorrelation with the aim of studying the distribution of the physicochemical or biological properties of amino acids within the genetic code table. First, using this index we are able to identify the amino acid property - among the 530 analyzed - that best correlates with the organization of the genetic code in the set of amino acid permutation codes. Considering, then, a model suggested by the coevolution theory of the genetic code origin - which in addition to the biosynthetic relationships between amino acids took into account also their physicochemical properties - we investigated the level of optimization achieved by these properties either on the entire genetic code table, or only on its columns or only on its rows. Specifically, we estimated the optimization achieved in the restricted set of amino acid permutation codes subject to the constraints derived from the biosynthetic classes of amino acids, in which we identify the most optimized amino acid property among all those present in the database. Unlike what has been claimed in the literature, it would appear that it was not the polarity of amino acids that structured the genetic code, but that it could have been their partition energy instead. In actual fact, it would seem to reach an optimization level of about 96% on the whole table of the genetic code and 98% on its columns. Given that this result has been obtained for amino acid permutation codes subject to biosynthetic constraints, that is to say, for a model of the genetic code consistent with the coevolution theory, we should consider the following conclusions reasonable. (i) The coevolution theory might be corroborated by these observations because the model used referred to the biosynthetic relationships between amino acids, which are suggested by this theory as having been fundamental in structuring the genetic code. (ii) The very high optimization on the columns of the genetic code would not only be compatible but would further corroborate the coevolution theory because this suggests that, as the genetic code was structured along its rows by the biosynthetic relationships of amino acids, on its columns strong selective pressure might have been put in place to minimize, for example, the deleterious effects of translation errors. (iii) The finding that partition energy could be the most optimized property of amino acids in the genetic code would in turn be consistent with one of the main predictions of the coevolution theory. In other words, since the partition energy is reflective of the protein structure and therefore of the enzymatic catalysis, the latter might really have been the main selective pressure that would have promoted the origin of the genetic code. Indeed, we observe that the β-strands show an optimization percentage of 94.45%, so it is possible to hypothesize that they might have become the object of selection during the origin of the genetic code, conditioning the choice of biosynthetic relationships between amino acids. (iv) The finding that the polarity of amino acids is less optimized than their partition energy in the genetic code table might be interpreted against the physicochemical theories of the origin of the genetic code because these would suggest, for example, that a very high optimization of the polarity of amino acids in the code could be an expression of interactions between amino acids and codons or anticodons, which would have promoted their origin. This might now become less sustainable, given the very high optimization that is instead observed in favor of partition energy but not polarity. Finally, (v) the very high optimization of the partition energy of amino acids would seem to make a neutral origin of the ability of the genetic code to buffer, for example, the deleterious effects of translation errors very unlikely. Indeed, an optimization of about 100% would seem that it might not have been achieved by a simple neutral process, but this ability should probably have been generated instead by the intervention of natural selection. In actual fact, we show that the neutral hypothesis of the origin of error minimization has been falsified for the model analyzed here. Therefore, we will discuss our observations within the theories proposed to explain the origin of the organization of the genetic code, reaching the conclusion that the coevolution theory is the most strongly corroborated theory.
... Hence, the very beginning of the genetic code formation should be considered in the light of a certain ambiguity, which was gradually by increasing some regularities and became "frozen" [9]. For a review of these approaches to the life origin, the very early evolution and meaning of the genetic code, see papers [1,2,3] on code biology. Anyhow, the GC was present already in the LUCA and has evolved to nowadays into about 30 versions, differing only in the few assignments of codons to amino acids (or stop signal). ...
Chapter
Full-text available
The genetic code (GC) plays a central role in all living organisms. From a mathematical point of view, the GC is a map from a set of 64 elements (which are codons) onto a set of 21 elements (which are 20 amino acids and 1 stop signal). The GC is the result of evolution, experimentally deciphered as early as the mid-1960s, but its satisfactory theoretical understanding does not yet exist. There are many papers on the GC modeling, its origin and evolution, but also many unsolved issues. In this contribution we provide a brief overview of genetic code modeling, highlighting the p-adic approach, which can describe many properties of the GC. Our primary mathematical tool is a p-adic distance, which simply and adequately describes similarities within the GC. We also point how one could apply this mathematical method to other sequences with a bioinformatic content.
Chapter
The genetic code defines the mapping between codons and amino acids, but its origin has remained elusive for the last 50 years. Of the 64 possible codons, 61 designate amino acids, indicating that many amino acids are encoded by more than 1 codon. The remaining three codons are used as stop signals. In the genetic code, the digital information in mRNA is translated into analog information in proteins. Aminoacyl-tRNA, the encoder, and decoder of mRNA, determines the correspondence between a codon and an amino acid. The genetic code is virtually universal in all life and was used by the last universal common ancestor (LUCA) four billion years ago. Most researchers believe that the code developed gradually, depending on the availability of amino acids in the prebiotic environment. From the beginning, the first code used triplet codons. There are two prevailing theories on why the genetic code might have originated: stereochemical and coevolutionary. The stereochemical theory suggests that the origin of the genetic code must lie in the stereochemical interactions between anticodons or codons and amino acids. The coevolutionary theory postulates that the code structure coevolved with the amino acid biosynthesis pathways: the 10 primary or primitive amino acids available in the prebiotic conditions that gave rise to the remaining 10 amino acids derived from the first set. We discuss these two leading views on the origin of the genetic code and conclude that the code’s organization is the outcome of the coevolution of genes and the genetic code. In our view, the stereochemical interactions between codons and amino acids are possible in some early stages of evolution. However, after decades of research, there is still no real evidence in favor of the stereochemical theory. The coevolutionary theory asserts two generations of canonical amino acids, depending on whether they were readily available in the prebiotic environment (Phase 1) or biosynthetically produced (Phase 2). Each phase contains 10 amino acids. With the availability of 20 amino acids, the genetic code expanded and stabilized in the universal code. The coevolutionary theory has been modified and extended through time. For example, Phase 1 amino acids are subdivided into two stages based on the evolution of the genetic code: the first stage (the GADV hypothesis) comprises four primitive amino acids—Gly (G), Ala (A), Asp (D), and Val (V) —for the GNC code. The second stage consists of 10 (4+6) cumulative amino acids (Phase 1) that gave rise to the SNS code. The incremental third stage (Phase 1 + Phase 2) includes all 20 amino acids that correlate with the universal genetic (UG) code. We have integrated the coevolutionary theory and identified two distinct coevolution phenomena: (1) the coevolution of the translation machine and the genetic code and (2) the coevolution of genes and the genetic code. The genetic code developed in three stages with the refinement of the translation machine: (1) the GNC code by the pre-tRNA/pre-mRNA machine; (2) the SNS code by the tRNA/aaRS machine, and (3) the universal genetic code by the tRNA/aaRS/ribosome machine. The emergence of translation machines began the Darwinian evolution, an interplay between information and its supporting structure. We also show the coevolution of the genes with three stages of the genetic code: (1) pre-mRNA and the GNC code; (2) short-chain mRNA and the SNS code; and (3) long-chain mRNA and the universal code. Encoding mRNA by aa-tRNA created the genetic code as a memory bank (origin), but decoding of mRNA processed the genetic code for protein synthesis (translation).
Article
We used the Moran's I index of global spatial autocorrelation with the aim of studying the distribution of the physicochemical or biological properties of amino acids within the genetic code table. First, using this index we are able to identify the amino acid property - among the 530 analyzed - that best correlates with the organization of the genetic code in the set of amino acid permutation codes. Considering, then, a model suggested by the coevolution theory of the genetic code origin - which in addition to the biosynthetic relationships between amino acids took into account also their physicochemical properties - we investigated the level of optimization achieved by these properties either on the entire genetic code table, or only on its columns or only on its rows. Specifically, we estimated the optimization achieved in the restricted set of amino acid permutation codes subject to the constraints derived from the biosynthetic classes of amino acids, in which we identify the most optimized amino acid property among all those present in the database. Unlike what has been claimed in the literature, it would appear that it was not the polarity of amino acids that structured the genetic code, but that it could have been their partition energy instead. In actual fact, it would seem to reach an optimization level of about 96% on the whole table of the genetic code and 98% on its columns. Given that this result has been obtained for amino acid permutation codes subject to biosynthetic constraints, that is to say, for a model of the genetic code consistent with the coevolution theory, we should consider the following conclusions reasonable. (i) The coevolution theory might be corroborated by these observations because the model used referred to the biosynthetic relationships between amino acids, which are suggested by this theory as having been fundamental in structuring the genetic code. (ii) The very high optimization on the columns of the genetic code would not only be compatible but would further corroborate the coevolution theory because this suggests that, as the genetic code was structured along its rows by the biosynthetic relationships of amino acids, on its columns strong selective pressure might have been put in place to minimize, for example, the deleterious effects of translation errors. (iii) The finding that partition energy could be the most optimized property of amino acids in the genetic code would in turn be consistent with one of the main predictions of the coevolution theory. Since the partition energy is reflective of the protein structure and therefore of the enzymatic catalysis, the latter might really have been the main selective pressure that would have promoted the origin of the genetic code. Indeed, we observe that the β-strands show an optimization percentage of 95.45%; so it is possible to hypothesize that they might have become the object of selection during the origin of the genetic code, conditioning the choice of biosynthetic relationships between amino acids. (iv) The finding that the polarity of amino acids is less optimized than their partition energy in the genetic code table might be interpreted against the physicochemical theories of the origin of the genetic code because these would suggest, for example, that a very high optimization of the polarity of amino acids in the code could be an expression of interactions between amino acids and codons or anticodons, which would have promoted its origin. This might now become less sustainable, given the very high optimization that is instead observed in favor of the partition energy but not polarity. Finally, (v) the very high optimization of the partition energy of amino acids would seem to make a neutral origin of error minimization, i.e. of the ability of the genetic code to buffer, for example, the deleterious effects of translation errors, very unlikely. Indeed, an optimization of about 100% would seem that it might not have been achieved by a simple neutral process, but this ability should probably have been generated instead by the intervention of natural selection. In actual fact, we show that the neutral theory of the origin of error minimization has been falsified for the model analyzed here. Therefore, we will discuss our observations within the theories proposed to explain the origin of the organization of the genetic code, reaching the conclusion that the coevolution theory is the most strongly corroborated theory.
Article
The modern genetic code reveals numerous traces of specific relationships between the early codons which, together with its internal asymmetries, suggest a sequential appearance of the nucleobases in primitive RNA molecules. Keeping the hypothesis of triplet pairings between primitive RNA molecules at the origin of the code, this work systematically examines complete codon-anticodon interaction matrices assuming distinct pairing options at each position of the triplet duplexes. Application of these principles suggests that a 27-codon precursor having a reasonable coding capacity for short peptide synthesis could have started with primitive RNA molecules able to form two distinct pairs with different free energies between a single purine and two pyrimidines (such as G with C and U). Conservation of the same pairing options at positions 1 and 2 of codons at the arrival of a second purine with distinct pairing preferences (such as A) generated a 64-codon intermediate code made of interrelated pairs or groups of codons (designated here as intricacy). The numerous traces of this hypothetical scheme that are visible in the standard and variant forms of the modern code demonstrate without ambiguity that the ancestral codon-anticodon duplexes required high energetic pairings at their central position (Watson-Crick) but tolerated less energetic pairings at the first codon position (G • U type). Combined with the sequential appearance of the nucleobases, the predicted codon intricacy allows a stepwise reconstruction of the evolution of the coding repertoire, by simple a posteriori comparison to the modern code. This reconstruction reveals a remarkable internal coherence in terms of amino acids and tRNA synthetases recruitment. The code started with a group of amino acids (Ala, Gly, Pro, Ser and Thr) that are now all activated by class II tRNA synthetases before reaching an intermediate period during which up to 14 distinct amino acids could be encoded by a full set of intricated codons. The perfect coincidence between the last 6 amino acids predicted in this reconstruction and the speculated action of the arrival of free atmospheric oxygen on proteins is spectacular, and suggests that the code has only reached its present form after the great oxidation event.
Article
Here I use the rationale assuming that if of a certain trait that exerts its function in some aspect of the genetic code or, more generally, in protein synthesis, it is possible to identify the evolutionary stage of its origin then it would imply that this evolutionary moment would be characterized by a high translational noise because this trait would originate for the first time during that evolutionary stage. That is to say, if this trait had a non-marginal role in the realization of the genetic code, or in protein synthesis, then the origin of this trait would imply that, more generally, it was the genetic code itself that was still originating. But if the genetic code were still originating - at that precise evolutionary stage - then this would imply that there was a high translational noise which in turn would imply that it was in the presence of a protocell, i.e. a progenote that was by definition characterized by high translational noise. I apply this rationale to the mechanism of modification of the base 34 of the anticodon of an isoleucine tRNA that leads to the reading of AUA and AUG codons in archaea, bacteria and eukaryotes. The phylogenetic distribution of this mechanism in these phyletic lineages indicates that this mechanism originated only after the evolutionary stage of the last universal common ancestor (LUCA), namely, during the formation of cellular domains, i.e., at the stage of ancestors of these main phyletic lineages. Furthermore, given that this mechanism of modification of the base 34 of the anticodon of the isoleucine tRNA would result to emerge at a stage of the origin of the genetic code - despite in its terminal phases - then all this would imply that the ancestors of bacteria, archaea and eukaryotes were progenotes. If so, all the more so, the LUCA would also be a progenote since it preceded these ancestors temporally. A consequence of all this reasoning might be that since these three ancestors were of the progenotes that were different from each other, if at least one of them had evolved into at least two real and different cells - basically different from each other - then the number of cellular domains would not be three but it would be greater than three.
Article
The function of the glutaminyl-tRNA synthetase and Glu-tRNAGln amidotransferase might be related to the origin of the genetic code because, for example, glutaminyl-tRNA synthetase catalyses the fundamental reaction that makes the genetic code. If the evolutionary stage of the origin of these two enzymes could be unambiguously identified, then the genetic code should still have been originating at that particular evolutionary stage because the fundamental reaction that makes the code itself was still evidently evolving. This would result in that particular evolutionary moment being attributed to the evolutionary stage of the progenote because it would have a relationship between the genotype and the phenotype not yet fully realized because the genetic code was precisely still originating. I then analyzed the distribution of the glutaminyl-tRNA synthetase and Glu-tRNAGln aminodotrasferase in the main phyletic lineages. Since in some cases the origin of these two enzymes can be related to the evolutionary stages of ancestors of archaea and eukaryotes, this would indicate these ancestors as progenotes because at that evolutionary moment the genetic code was evidently still evolving, thus realizing the definition of progenote. The conclusion that the ancestor of archaea and that of eukaryotes were progenotes would imply that even the last universal common ancestor (LUCA) was a progenote because it appeared, on the tree of life, temporally before these ancestors.
Chapter
Several major breakthroughs have helped contribute to the emerging field of astrobiology. Focusing on these developments, this fascinating book explores some of the most important problems in this field. It examines how planetary systems formed, and how water and the biomolecules necessary for life were produced. It then focuses on how life may have originated and evolved on Earth. Building on these two themes, the final section takes the reader on a search for life elsewhere in the Solar System. It presents the latest results of missions to Mars and Titan, and explores the possibilities of life in the ice-covered ocean of Europa. This interdisciplinary book is an enjoyable overview of this exciting field for students and researchers in astrophysics, planetary science, geosciences, biochemistry, and evolutionary biology. Colour versions of some of the figures are available at www.cambridge.org/9780521875486.
Article
Part I of this paper deals with the amount of degeneracy in the genetic code or, looked at in reverse, the amount of nonsense. Section A considers the experimental evidence and suggests (a) that technical difficulties have prevented assigning meaning to the still unassigned triplets and (b) that the validity of the earlier evidences for nonsense triplets may be questioned in the light of recent discoveries. Complete degeneracy of the code and total absence of nonsense have not yet been excluded. Section B comes to the conclusion, on the basis of general evolutionary considerations, that natural selection would be expected to establish and preserve a completely degenerate code. Section C points out that different nondegenerate codes differ greatly in the builtin frequency of nonsense mutations by single base substitutions.