CONTRIBUTIONS TO CODON-ANTICODON RECOGNITION

JACQUES NINIO

 

( included in the web site http://www.lps.ens.fr/~ ninio )


return to Ninio's home page

 

TOPICS DISCUSSED HERE:

 

- Biographical notes
- The missing triplet hypothesis.
- The logics of transfer RNA modification
- Mutual adaptation of organelle and cytoplasmic translation
- The evolution of the genetic code
- Comments on codon usage

 

SEE IN OTHER SECTIONS

- The origin of the genetic code (in the origins of life section)
- Non- Watson-Crick pairs in RNA and DNA (in sections on Bioinformatics and on RNA structure)
- Errors in protein synthesis, and their control (in Section on Accuracy, and its control)
- wobble pairs in prebiotic replication (in section on the origins of life)

++++++++++++++++++++++++++++++++++++++++++++++++++++++++

 

BIOGRAPHICAL NOTES

 

In June 1965, at a time I was making erroneous speculations on the origin of the genetic code, my thesis supervisor, Vittorio Luzzati handed to me a scientific memo (note 1) by Francis Crick, entitled "Codon-anticodon pairing. The wobble hypothesis". Vittorio said "Here is the good theory". I was fascinated by this text. It presented a general theoretical framework, then used a deductive, abstract reasoning to narrow down a wide class of theoretical possibilities into a single, precise hypothesis. Subsequently, the deductive argumentation was only briefly summarized in the final paper (Journal of Molecular Biology, volume 19, pages 548-555; 1966). On the other hand, I was also somewhat dissatisfied with the hypothesis, in relation to the origin of the genetic code. Degeneracy, in the framework of the wobble hypothesis, is a consequence of inevitable structural constraints. It seemed more natural to me to consider that the structures of ribosomes and tRNAs evolved to optimize translation, while preserving as much as possible some early coding relationships.

In May 1967, while listening to a talk by my friend Pierre Claverie, I had an illumination, which led me to the "missing triplet hypothesis" and from there to my subsequent carreer in the accuracy of molecular processes. Pierre Claverie's talk was about quantum-mechanical calculations, in which he studied the stacking energy of every doublet of Watson-Crick pairs, sliding one pair over the other, and looking for the geometry giving the minimum energy. This geometry was found to be dependent upon which nucleotide pair was on top of which, a result which can be reappraised as an anticipation of modern ideas on sequence-dependent local conformations in DNA (see for instance, Edwin G. Trifonov's work). Claverie's calculations were made with the limitations of the time: the nucleotide pairs were interacting in vacuum, and the sugar-phosphate backbones were not taken into account.

I saw immediately that if Claverie was basically right, the wobble hypothesis could be wrong - or at least, the structural basis of the wobble hypothesis could be wrong. The wobble hypothesis made a sharp distinction between nucleotide pairings in the first two positions of the codon-anticodon associations, assumed to occur in a completely rigid geometry, and nucleotide pairings in the third position, assumed to occur in a looser geometry. Claverie's results, on the other hand, were suggesting that a single, rigid geometrical arrangement (as was assumed by the wobble hypothesis for the first two positions) could not be simultaneously optimal for all the complementary associations between codon and anticodon doublets. So, upon binding, a confomational change to a more optimal structure could be expected, and it beacme conceivable, energetically, that some non-complementary associations were not too disfavoured, compared to strictly complementary ones. Thus, I considered that non-complementary nucleotide pairings could occur in all three positions of the codon-anticodon associations. How could this be compatible with the experimental observation that translation was practically error-free on the first two codon positions? I saw rapidly how the energy-levels could be manipulated to achieve this goal. The idea will be explained in the section below on the missing triplet hypothesis. In a nutshell, and reasoning in terms of energy levels, you can use the degeneracy of the genetic code in the third position to remove potential errors on the first two positions. Actually, I had in mind an even more radical view, according to which third position degeneracy was there, in large meausre, as a way to remove the potential ambiguities in the first two positions.

After Claverie's talk, I raised a question about local geometry in RNA structures, which reflected my statistical observations on succession rules in RNA stems. This was the first time that my name appeared printed in a scientific publication [1]. The next day, I explained my idea on codon-anticodon recognition to Vittorio Luzzati, and he reacted immediately, saying that my theory could not be correct, for it were, Crick would have thought of it before me. The theory remained three years in my drawers. I was terrified at the idea that someone else might hit at the same idea, but it turned out to be an unfounded apprehension. Several experimental papers on codon-recognition patterns were published, all claiming perfect agreement with the wobble hypothesis, but it seemed to me that my alternative hypothesis was not ruled out by the new pieces of evidence. In September 1969, I started a one year period of military duties in the French army, as an officer in the transmission corps. In Verdun, far from the laboratory, I did a lot of writing: the first draft of my thesis, the manuscript on RNA topology, later submitted to "Biochimie" (see section on mathematical biology), and I started working again on "the missing triplet hypothesis". In parallel, Pierre Claverie made quantum mechanical calculations on triplet-triplet interactions in vacuum, using the RNA, not the DNA geometry, which suggested possible ambiguities in codon-anticodon interactions.

After my return to civil life, I finalized the manuscript, which was completed by a long appendix, written by Pierre Claverie. The manuscript was submitted to the Journal of Molecular Biology. This was, considering the time, and my situation, an act of boldness. I was 27 years old, the laboratory boss was completely skeptical of the work, and the Journal of Molecular Biology was reputed for having the highest quality requirements in molecular biology. Publishing there was becoming a member of the scientific elite (Note 2). The manuscript was channelled by François Gros to the Journal of Molecular Biology, and received at their office on February 18th, 1970. Then, I waited and waited and waited. My boss said, with compassion, that this was probably a bad sign, a sign that the sollicited reviewers were withdrawing from the task one after the other.

Seven months later, I was still without news of the manuscript. François Gros wrote a letter of inquiry, and at last (September 25th) I received from the JMB office a letter saying that the second reviewer's report had not yet been received, and that the manuscript could be accepted after being amended according to the first reviewer's report. The first reviewer was, almost certainly, Francis Crick himself (Note 3), and he appeared, in his report, far more open to the ideas expressed in the MS, than most of the people who worked in the field, the next 35 years. Actually, he never published on the subject again, as far as I know. The paper was accepted on November 19th, 1970 [2].

In May 71, I was still in Gif-sur-Yvette, and I made the very bold step of writing to J.N. Davidson, then co-editor in chief of the series "Progress in Nucleic Acid Research", proposing to him to write a review chapter on codon-anticodon recognition. He accepted at once, and I started writing my review. In retrospect, the time elapsed between the JMB paper and this review was too short, and the review does not show any serious progress with respect to the JMB paper, on the topic of which codons are recognized by which anticodons. I expressed though with more vigour the notion that we had to think about the mutual interactions between two families of molecules. This was reflected in the title "Recognition in nucleic acids and the anticodon families". The manuscript was completed in April 1972. What was interesting in this review, was not the core, but everything which revolved around it. Thus, I introduced the kinetic ideas on ribosome accuracy in the first pages, and interpreted Gorini's ribosomal mutants in terms of kinetics. I included my ideas on the evolution of the genetic code, and on the evolution of transfer RNA 3D structure. At the end of the review, I stated very briefly how the notion of recognition as a matching between families of molecules could be extended to various systems, in particular the recognition between tRNAs and aminoacyl tRNA synthetases (Note 4) and protein-protein interactions within the cell (Note 5).

After the death of Norman Davidson, the review was handled by the other editor, Waldo Cohn, and it came back to me with pages and pages of objections from a referee. The referee wrote as though he was as an acknowledged authority in the genetic code, but his report was totally idiotic, so his identity was not hard to guess (Note 6). I answered point by point, and made a number of clarifications in the text. The revised manuscript was published without further problem [3].

In December 1971, I happened to attend a meeting on tRNA, organized in Strasbourg by the "Société de Chimie Biologique". This meeting made a lasting impression on most of the participants, due to an incident between Brian Clark and I. The French organizers had, very generously, allowed me to give a 10 minutes talk on Friday 10th, on anticodon families. Brian F.C. Clark, from Cambridge gave the next day a talk on protein synthesis, and spoke of the wobble hypothesis without mentioning alternatives. So, I stood up at the end of his talk and pointed out - in terms I have completely forgotten - that scientific truth in molecular biology was not an exclusive property of the school of Cambridge. It was, in the context of the time, absolute sacrilege. Brian Clark replied that if I wished to have my work published in top journals, I had better tone down my criticisms. The audience took this exchange laughingly. I think that Hans Zachau was chairing the session, and he concluded in a convivial manner. This is a story of the past. Recently, I met Brian Clark in a FEBS meeting, and had pleasure exchanging a few words with him. Even more recently, reading my old documents for the purpose of this scientific biography, I realized that the transfer RNA preparation with which I was able to do my thesis work came in fact from his own hands (see section on RNA structure). So, I here solemnly bury the hatchet.

At the Strasbourg meeting, I became acquainted with Frank Unger and Georg Högenauer. Both were working in Sandoz Forschungsintitut in Vienna, and were carrying experiments of oligonucleotde bindings to the anticodon loops of transfer RNAs. I saw a possibility to test my ideas in their lab. If I was right, one could detect the binding of trinucleotides which formed non-copmlementary base-pairs to the anticodon, in the first or second positions of the codon-anticodon association. I was invited to their laboratory and spent ten days there in July. They made binding experiments with the oligonucleotides GAA and UGAA and several tRNA's, including a half-tRNA molecule, specific for aspartic acid, offered by Jean Gangloff from Strasbourg. In fact the experiments probed the binding potential of the GTUC loops of tRNA's, and the anticodon loops were to be studied later, but there was no follow up.

In September 1972, I was already at the Salk Institue in Orgel's laboratory. I was keeping an eye on the tRNAs sequencing work. To my disapointment, John Carbon had sequenced a glycine tRNA, with an unmodified CCC anticodon, which, I had predicted, should not be used. However, I was distracted from this topic by my involvement in the development of the kinetic theory of accuracy ([4,5] and see Section on the kinetic theory of accuracy), the analysis of Gorini's ribosomal mutants, the next nucleotide effect, kinetic amplification, and all that. Remember that I had been living with the missing triplet hypothesis since 1967, and was therefore ready for new adventures.

However, the topic of codon-anticodon recognition entered by the back door, through a problem raised by Leslie Orgel. He considered the situation of a cell hosting two different translation apparatuses, one in the cytoplasm, and one in the mitochondria. What could happen? There would be, on one hand, a trend to limit the number of mitochondrial genes, and transfer as many functions as possible to the nucleus, and on the other hand, an accuracy problem arising from the mixing of components from two different translation apparatuses. I worked a bit on the subject , and reached the conclusion that the mitochondria would use tRNAs having more extended reading patterns than the nuclear ones. Upon my return in France, I was invited to a congress on nucleocytoplasmic interactions (see section below on mitochondrial codes), gave a talk on the subject which was subsequently published in a book [6]. I was aware of the fact that, pushing my ideas one step forward, one could anticipate that, perhaps, the mitochondrial code could be different from the cytoplasmic one. This appeared in an allusive way at the end of the article, but I did not state it clearly enough. Later, when it was shown that the mitochondrial codes differed from the cytoplasmic one, I regretted my timidity. Still later, when it became clear that not all cytoplasmic codes were identical, I ceased reproaching myself the timidity, because reality was clearly beyond my expectations. On the other hand, my past work on the co-evolution of two translation apparatuses led me naturally to consider that the evolution of the cytoplasmic codes, observed in paramecia and a few other organisms might have been driven by some analogous problem. I pointed out that if a paramecium swallows a bacterium, then there could be some mixing of the translation apparatuses. But instead of peaceful coexistence, we would rather have a genomic war, each translation apparatus evolving to silence the other. From there, an evolution of the host's termination codons became a very likely possibility (see [7] and section below on the evolution of the genetic code).

In December 1977, I received a letter from Samir Kumar Mitra. He was working in Ulf Lagerkvist's laboratory in Göteborg, and they were finding extensive misreading at the third position of the codon-anticodon asociation. Ultimately, this led Lagerkvist to propose his "two out of three hypothesis". Mitra complained about the way his work had been treated. As always, I was ready to protect the weak and fight against the strong, so I took position in his favour. Soon afterwards, Mitra started organizing a round table on codon-anticodon recogntion at the next FEBS meeting in Dresden, in (East) Germany. He asked me to be a co-organizer, and we prepared and run the session together. After the session, we were asked to write a paper for the proceedings of the meeting, and it had to be given to the organizers before the end of the meeting. So we wrote the paper together in the conference hall, and it appeared without modification [8].

Ten years later, I attended a Cold Spring Harbor Meeting on "The evolution of catalytic functions". I gave a talk on "Kinetic devices in protein synthesis, DNA replication, and mismatch repair". It was published as a review in the CSHSQB series [9]. In the last section, I reviewed codon-anticodon recogntion, in the light of the latest results on in vivo reading patterns, tRNA modifications, and missense suppressor tRNAs. I made the link between the missing triplet hypothesis, and the kinetic treatments of accuracy. Indeed, in Table 1 of this article, page 644, I show step by step how a first position ambiguity may be erased and absorbed by a third position degeneracy. This makes the hypothesis complete, and fully compatible with modern thinking, and I suggest to the reader of these pages to study carefully this table 1 in order to appreciate the full power of the hypothesis. After this farewell paper, I wrote one more review on the code, which reflects my participation to two meetings, a tRNA meeting in Umea (Sweden), in 1987, and a meeting on the origins of life in Prague, in 1990 [10].

In a wider perspective, being interested both in the genetic code and in accuracy problems in general, I maintained a longstanding interest for quantifying the incidence of all "abnormal" nucleotide pairs. Thus, I showed, experimentally in a study on prebiotic replication with oligomers, that U could be incorporated non-enzymatically opposite to a G, when it was part of a UG dimer ([11], and see section on the origins of life). More importantly, I was able to derive, using arguments of RNA secondary structure predictions, sets of energy values describing the energetics of U.G pairs, and those of all the "odd pairs" (i.e., other than A.U, G.C, G.U) - see [12, 13] and Section on Bioinformatics.

++++++++++++++++++++++++++++++++++++++++++++++++++++++

THE MISSING TRIPLET HYPOTHESIS

Let us consider an imaginary situation, in which the genetic code would use, say, 64 codons, and all the 64 unmodified nucleotide triplets as anticodons, each one being carried by a specific transfer RNA. Let us visualize, not the codon-anticodon associations, but the energy levels of all the possible 64x64 codon-anticodon associations. We expect that there will be some overlap between the levels for cognate and non-cognate interactions. Actually, what is important is to examine, for each codon separately, the set of the 64 energy levels corresponding to the 64 tRNAs. Let "A" be one such codon. Among the non-cognate interactions, some are too weak to generate an appreaciable level of misreading. So let us focus on the presumably few interactions which could contribute to translation errors. Let us take the strongest of these interactions. So, there is a transfer RNA X, who reads normally a certain codon "B" in accordance with the genetic code, and misreads the codon "A". What can be done about that?

One solution is simply not to use this tRNA ! This is feasible if the codon "B" can be read, in the absence of this tRNA X, by another tRNA, Y, which reads the codon through a non complementary interaction at the third codon position.

So, the potential ambiguous binding between codon A and tRNA X disappears, thanks to the use of the wobble binding between codon B and tRNA Y.

Another solution, is to modify the anticodon of tRNA X, when this is possible, in a way which weakens its binding potential to both codons A and B. While misreading of A would decrease sharply, the residual binding capacity with respect to B could be sufficient to sustain cognate reading with a reasonable probability.

With the help of modern kinetic concepts the argument can be cast into a practical example which one can follow step by step (see Table 1 in [9]).

I see no reason to withdraw this hypothesis and find that the attitude of the molecular biologist's community towards the wobble hypothesis is extremely strange, by the standards of other scientific disciplines.

As I wrote in the CSHSQB review [9],

" The wobble hypothesis (Crick, 1966) is a historical accident. It would perhaps not have seen the light of day, had the first sequenced tRNA been a tRNA-Val1 from E. coli, rather than a tRNA-ala from yeast. The tRNA-Val1 anticodon VAC where V is uridine-5-oxyacetic acid probably reads the codons GUU, GUA, GUG, whereas the wobble hypothesis would have suggested an anticodon IAC reading the codons GUU, GUC, and GUA".

I have never seen codon-anticodon recognition explained as it should, first by putting on the table the basic experimental findings, then discussing their implications. The usual exposition begins with the wobble hypotyhesis plus a few very selected facts. Then, eventually, anything which does not fit is either not mentioned at all, or given as further complications.

It must be stressed that codon reading patterns have almost never been determined in vivo, and that most in vitro experiments lack the sensitivity that could have been achieved through in vivo genetic manipulations. The biology students have to accept at the same time that there are universal reading rules in the code, and that there are plenty of practical exceptions, that there is a structural folding principle for 7-membered loops in RNAs, explaining the wobble and making it inevitable in an evolutionary context, and that each anticodon loop is shaped for optimal reading, using sequence constraints on both sides of the anticodon, and chemical modifications of the anticodon itself. They must repeat obediently that some special device, at the 3rd position of the codon-anticodon association makes the wobble pairs possible, and at the same time, they must say that the wobble G.U pair is universally present in RNA secondary structures. All this is again a sign of scientific immaturity.

It would be much more honest to state that unconstrained codon-anticodon interactions would lead to massive decoding errors and most of these are prevented by using two tricks in combination:

(1) Modifying chemically some nucleotides in the anticodon loops in tRNAs, thereby changing the binding efficiencies, and making non-cognate interactions very weak.

(2) Using a restricted ( < 64) number of anticodons. This is feasible, thanks to the degeneracy of the code. tRNAs with anticodons which are not strictly needed and which might provoke misreading are simply not synthesized.

It seems to me that everyone in the field can agree with statements (1) and (2). In the missing triplet hypothesis, I add two statements, which are not intuitive to many people:

(1') A chemical modification does not need to make the correct codon-anticodon binding more efficient. In many cases, it is enough to shift the energy levels of the correct and the incorrect association in the same direction, by the same amount.

(2') The existence of a legal degeneracy somewhere, makes it possible to do without some anticodons which generate an illegal ambiguity. Then the observed 3rd position degeneracies may be related to the existence of ambiguities at other positions.

Again, how this might work, quantitatively, has been presented in [9], Table 1.

++++++++++++++++++++++++++++++++++++++++++++++++

THE LOGICS OF tRNA MODIFICATIONS

Many nucleotides are modified post-transcriptionally in tRNAs. In a narrow-minded perspective, one is tempted to justify every individual modification. Experiments designed to prove the essential character of this or that modification have in general failed. Modifying enzymes work on a wide subset of tRNA species. They need not be very specific. As I probably wrote somewhere, such modifications need not improve every individual tRNA of the subset. It is just important, for the cell, that the effect of modifying all the tRNAs of the subset be, on the whole, advantageous. Thus, a same kind of modification may have an effect in one direction for a given tRNA, and in the opposite direction for another tRNA. Evolution proceeds by implementing modifications which strike at many tRNAs at the same time, and affect them in various ways, provided that each new modification brings about a global improvement.

+++++++++++++++++++++++++++++++++++++++++++++++++

MUTUAL ADAPTATION OF ORGANELLE AND CYTOPLASMIC TRANSLATIONS

From "Molecular approaches to Evolution", Princeton edition [14], chapter 15:

"Initially we have two translation apparatuses functioning in parallel, one in the host-cell's cytoplasm and the other in the ancestor of the organelle. If there is communication between the organelle and the host-cell's cytoplams and the components of the translation apparatuses (tRNAs and others) differ from one compartment to the other, errors will be produced, just as when tRNAs and activating enzymes from two different species are mixed. Worse yet, some codons may correspond to different amino acids in the two apparatuses. The later evolution of the two apparatuses will obey two contradictory tendencies: to reduce the number of genes which do the same job and to maintain precise translation in the two apparatuses. Can we deduce from the foregoing any rules about the characteristics of translation components in the two compartments ? Hre are some of the early guesses I made, to solve this problem submitted to me by Orgel.

"Initially, the organelle membrane is slightly permeable: communication with cytoplasm is possible but autonomy is preserved. How can communication be maintained while avoiding errors due to partial mixing between the two translation apparatuses? Two readily applicable solutions may be glimpsed here. Activating enzymes apt to make mistakes would be confined to their cellular compartment following mutations which would give them an affinity for the membranes of their compartment or for other proteins of the same apparatus. Gross aggregates would have difficulty in crossing the organelle envelope. Enzymes could "stamp' molecules synthesized in one compartment by adding on methyl groups, phosphate groups, or others which would make them suspect beyond their own borders where they would be degraded by cleaning enzymes. Or one can envisage a slow evolution of each component of one of the apparatuses to minimize interactions with components of the other. Cross-relations between tRNAs and elongation factors or activating enzymes would thus be quite different from those observed when one compares translation apparatuses of two species, whether they are closely or distantly related. As the translation apparatuses adjust to one another, it becomes possible to suppress parts which are doubly employed. It would be difficult to get rid of the activating enzymes of serine, leucine and arginine which have the delicate task of charging four or five different tRNAs. Besides, an organelle tRNA can be replaced by one from the central apparatus on the condition that all codons continue to be read. It will be easier to replace a tRNA which reads codons C1 and C2 by one which reads C1, C2 and C3 than the reverse. Therefore we can foresee that in the organelle there wil be a tendency for tRNAs which read a single codon to disappear. The 'coverage' of the codons must be particularly large in organelles. Activating enzymes do form aggregates in the cytoplasm and the codon coverage is large indeed in organelles. Unforeseen was the fact that the genetic codes would differ. Perhaps they were different at the start. But it is also tempting to speculate that in some cases the error-level in translation of some codons was so high that a change in the meaning of these codons could not have made the situation worse, and was therefore tolerable."

++++++++++++++++++++++++++++++++++++++++++++++++

THE EVOLUTION OF THE GENETIC CODE

From [9]:

" Consider an in vivo situation where there is some mixing of the components of two translation apparatuses due to endosymbiosis (the mitochondria), predation (the protozean ciliates) or infection (the mycoplasms). Then, the translation of one particular codon may be so ambiguous that the assignment of this codon could evolve. This might be the key to the observed divergence in the genetic code"

For a more detailed presentation, with discussion of specific pathways for the genetic code's evolution, see [7].

++++++++++++++++++++++++++++++++++++++++++++++++

COMMENTS ON CODON USAGE

From "Molecular approaches to evolution", Princeton edition [14], chapter 9:

" (---) Presumably, codon usage is mainly adjusted to the needs of the translation machinery"

"As postulated by Garel then shown by him in the extreme case of silk biosynthesis, the cell synthesizes preferentially the tRNAs that read the most frequent codons. To optimize translation fully, the cell would make a messenger RNA structured enough to resist nucleases but loose enough to let ribosomes creep in. It would use a successsion of codons that slow down translation at some places to let the polypeptide chain fold properly, and successions that allow full speed elsewhere. It would try to reduce errors by keeping error-prone codons scarce and remote from error-enhancing contexts. Above all, it would choose the codons so as to avoid having the unfinished protein fall off from the ribosome (especially when close to the end). Single out any of the factors just listed, or any other plausible one, and you get one of the present or future theories of why particular cells use particular codons."

To this, I have two new elements to bring into the picture:

First, multiple mutations occur much more frequently than people think. I have explained why this is so in [15, 16]. So, even though single codon changes might not generate seizable selective advantage, multiple codon changes do occur with a significant probability, and they may generate conspicuous selective effects.

Second, the mutational spectrum of DNA replication is itself subject to changes that can be brought about by single mutational events. Therefore, codon usage may evolve in particular directions thanks to changes in the directional mutation pressure. In this case, the codon usage may not be fully optimal, because it is the general direction of the mutation pressure which is beneficial, and not the individual codon changes produced.

 

++++++++++++++++++++++++++++++++++++++++++++++++++++

+++++++++++++++++++++

NOTES

++++++++++++++++++++++++++

++++++++++++++++++++++++++++++++++++++++++++++++++++

Note 1. In 1965, we received regularly in the laboratories scientfiic memos, written before formal submission to scientific journals. The memos of our series had the generic title "information exchange group N° 7: Nucleic acids and the genetic code".

+++++++++++++++++++++++++++++++++++++++++++++++++++++

Note 2. This attitude may seem decadent to-day, because in biolgy, scienttific quality has ceased to be a guiding criterion. Articles are evaluated and accepted or rejected, first on the basis of geographic origin, second on the basis of their short-term "selling potential" (expected contribution to the impact factor), and rarely on the basis of scientific truth.

+++++++++++++++++++++++++++++++++++++++++++++++++++++

Note 3. The report began with the sentence:

" This paper presents interesting ideas concerning the genetic code.The proposed "missing anticodon" hypothesis would, if correct, shed new light on the mechanism of codon recognition. The article could be significantly improved by the clarification of certain points."

There were 9 points, of the kind a careful reviewer would make. Two of them, points 3 and 7, are particularly interesting in the light of modern knowledge.

"Point 3. Only one tRNA (tRNA-Val from yeast) has a GU base-pair in a stretch of only three base-pairs. The various base-paired regions of tRNA may interact to stabilize any weak pairings; it is incorrect to state that: "GU pairs can be incorporated into segments of double-helical RNA without any special device." (p. 6)

This criticism illustrates the change of attitude from Crick's wobble hypothesis (a G.U pair needs something special) from my own standing which was to try to explain why G.U pairs were NOT seen at the first two positions of the codon-anticodon association.

"Point 7. It is stated that a weakening of a G.C pair makes the genetic code less ambiguous (p. 11). Unless inosine affects the diagonal interactions of stacking, the difference in association energy between UGC/ACG and UGC/CCG would be similar to the difference between UGC/ACI and UGC/CGI. The energy calculations should be extended to include inosine".

Here, the reviewer, was reasoning in terms of energy differences, as I did in the manuscript. In the revised manuscript, I went around the objection, by stating (p. 70, section 6) that the anticodon responsible for the ambiguity should be modified, but not its correct competitor. To-day, I would have answered differently, using kinetic principles instead of free-energy differences arguments.

Concerning Crick's intimate conviction, I find the following passage in a letter which I sent to Waldo Cohn, dated July, 27th, 1972,

"As far as I know, the refereeing of my hypothesis for the Journal of Molecular Biology was signed by Crick himself. Crick seems to be less confident in the absolute reliability of the wobble hypothesis than some of his followers are. "I am not a wobblist myself" is one of his expressions (Lecture given at a meeting of the British Biophysical Society, December 1966; private talk, Cambridge, March 1968)."

 

+++++++++++++++++++++++++++++++++++++++++++++++++++

Note 4

"The set of molecules used in any particular cell form a coherent family, with quite unpredictable relationships between some of their features. For instance, within one cell, the set of the aminoacyl synthetases is adapted to the recognition of the set of the tRNA species of that cell. When one takes tRNA from one organism, and enzymes from another, mischarging occurs (reference to Dudock et al., 1971). Through these experiments, some investigators are attempting to delineate a "recognition site" for the syntetase on the tRNA. We confidently predict that such a reductionist hope will lead to inconsistent results."

++++++++++++++++++++++++++++++++++++++++++++++++++

Note 5.

The penultimate paragraph of this review:

"The information content of a protein must include the compatibility of that protein with the various processes that take place in the cell. Part of the information can be described in the form of negative instructions: do not go accross the membrane, do not bind to the active site of that protein, etc. The ability of an enzyme to perturb the functioning of another can be suppressed if it forms an oligomeric structure in which a portion of the monomer surface is now buried in the quaternary structure."

++++++++++++++++++++++++++++++++++++++++++++++++++

Note 6

From the reviewer's report on the "Progress in Nucleic Acid Research" manuscript.

(to be written later)

++++++++++++++++++++++++++++++++++++++++++++++++++++

++++++++++++++++++++

REFERENCES

+++++++++++++++++++++++

++++++++++++++++++++++++++++++++++++++++++++++++++++

 

[1] Claverie, P. (1968) Intéractions moléculaires et structure hélicoïdale des acides nucléiques. Journal de Chimie 65, 57-65. My question is page 65.

[2] Ninio, J. (1971) Codon-anticodon recognition : the missing triplet hypothesis. J. Mol. Biol. 56, 63-74. .Appendix by Claverie, P.: Calculation of interaction energy between triplets in the RNA 11 configuration. Pages 75-82.

[3] Ninio, J. (1973) Recognition in nucleic acids and the anticodon families. Progress in Nucleic Acids Res. Mol. Biol. 13, 301-337.

[4] Ninio, J. (1974) A semi-quantitative treatment of missense and nonsense suppression in the strA and ram ribosomal mutants of Escherichia coli . Evaluation of some molecular parameters of translation in vivo. J. Mol. Biol. 84, 297-313

[5] Ninio, J. (1975) Kinetic amplification of enzyme discrimination. Biochimie 57, 587-595.

[6] Ninio, J. (1975) Considerations on the problem of the joint evolution of two different translation apparatuses within the same cell. In Molecular biology of nucleocytoplasmic relationships (Puiseux-Dao, C., ed.) pp. 31-39, Elsevier, Amsterdam.

[7] Ninio, J. (1986) Divergence in the genetic code. Biochemical Systematics and Ecology 14, 455-457.

[8] Mitra, S.K. & Ninio, J. (1979) Recognition between codon and anticodon. The limits of our knowledge. In: "FEBS Federation of European Biochemical Societies 12th meeting Dresden 1978". Vol. 51: "Gene functions" (S. Rosenthal et al., eds), Pergamon Press, Oxford, pp. 437-444.

[9] Ninio, J. (1987) Kinetic devices in protein synthesis, DNA replication and mismatch repair. Cold Spring Harb. Symp. Quant. Biol. 52, 639-646.

[10] Ninio, J. (1990) The revised genetic code. Origins of Life and Evolution of the Biosphere 20, 167-171.

[11] Ninio, J. & Orgel, L.E. (1978) Heteropolynucleotides as templates for non-enzymatic polymerizations. J. Mol. Evol. 12, 91-99.

[12] Ninio, J. (1979) Prediction of pairing schemes in RNA molecules. Loop contributions and energy of wobble and non-wobble pairs. Biochimie 61, 1133-1150.

[13] Papanicolaou, C., Gouy, M. & Ninio, J. (1984) An energy model that predicts the correct folding of both the tRNA and the 5S RNA molecules. Nucleic Acids Res. 12, 31-44.

[14] Ninio, J. (1979) Approches moléculaires de l'évolution. Masson, Paris. (1982) Molecular approaches to evolution. Pitman, London. (1983) idem, Princeton University Press, Princeton, (1984) Japanese translation, Kinokuniya Press, Tokyo.

[15] Ninio, J. (1991) Transient mutators: a semiquantitative analysis of the influence of translation and transcription errors on mutation rates. Genetics 129, 957-962.

[16] Ninio, J. (1996) Gene conversion as a focusing mechanism for correlated mutations: a hypothesis. Molecular and General Genetics 251, 503-508.