Science v Intelligent Design: Wolf and Koonin on the origin of the translation system

flunked.jpgIn a recent paper, authors Yuri Wolf and Eugene Koonin present a hypothesis on the origin of the translation system and show how science proceeds. I am encouraging ID proponents to present a similarly detailed explanation based on the foundational concepts of Intelligent Design.

The authors explain how the problem is neither trivial and yet can be resolved:

The origin of the translation system is, arguably, the central and the hardest problem in the study of the origin of life, and one of the hardest in all evolutionary biology. The problem has a clear catch-22 aspect: high translation fidelity hardly can be achieved without a complex, highly evolved set of RNAs and proteins but an elaborate protein machinery could not evolve without an accurate translation system. The origin of the genetic code and whether it evolved on the basis of a stereochemical correspondence between amino acids and their cognate codons (or anticodons), through selectional optimization of the code vocabulary, as a “frozen accident” or via a combination of all these routes is another wide open problem despite extensive theoretical and experimental studies. Here we combine the results of comparative genomics of translation system components, data on interaction of amino acids with their cognate codons and anticodons, and data on catalytic activities of ribozymes to develop conceptual models for the origins of the translation system and the genetic code.

In other words, any hypothesis needs to address the catch-22 scenario as well as incorporate the scientific knowledge gathered so far.

In their explanations, these scientists use a main guiding principle

Our main guide in constructing the models is the Darwinian Continuity Principle whereby a scenario for the evolution of a complex system must consist of plausible elementary steps, each conferring a distinct advantage on the evolving ensemble of genetic elements. Evolution of the translation system is envisaged to occur in a compartmentalized ensemble of replicating, co-selected RNA segments, i.e., in a RNA World containing ribozymes with versatile activities. Since evolution has no foresight, the translation system could not evolve in the RNA World as the result of selection for protein synthesis and must have been a by-product of evolution drive by selection for another function, i.e., the translation system evolved via the exaptation route. It is proposed that the evolutionary process that eventually led to the emergence of translation started with the selection for ribozymes binding abiogenic amino acids that stimulated ribozyme-catalyzed reactions.

In other words, the scenario needs to be plausible in the sense that each step confers a distinct advantage, and since evolution has not foresight, the translation system must have evolved via an exaptation route.

In addition to presenting various plausible scenarios, the authors also acknowledge our level of ignorance (or in ID speak ‘complexity’) and provide us with experimental tests that can help resolve several aspects of their scenario.

The authors explain how the origin of ‘complex biological systems’ is not only a classical topic in evolutionary science but also a principal focus of attack by “anti-darwinists of all ilk”.

The origin of complex biological systems is a classical topic in evolutionary biology and, probably, the principal object of attacks of anti-darwinists of all ilk, including the notorious Intelligent Design movement. The gist of the criticisms is that many biological systems are not just complex but “irreducibly complex” and, as such, could never evolve via the Darwinian mechanism of gradual, stepwise adaptive change because intermediate stages of evolution would have no selective value and so could not be fixed. Darwin himself was perfectly aware of the problem and its dimensions and addressed it in one of the most famous passages of the Origin, the one on the evolution of the vertebrate eye [1]. The solution offered by Darwin and developed ever since in numerous works of evolutionary biology was straightforward in principle and extremely ingenious when it came to details.

So what challenge is facing the evolutionary biologist?

When an evolutionary biologist strives to explain the origin of a truly novel system that is seen only in its elaborately complex state and, at face value, appears to be irreducibly complex, the task is much harder.

The scenario needs to be based on the logical premise that “evolution has no foresight”

Because evolution has no foresight, no system can evolve in anticipation of becoming useful once the requisite level of complexity is attained. Instead, the evolving system must have a selectable function(s) distinct from the modern one, a possibility recognized by Darwin [1] and emphasized by Gould in the concept of exaptation, that is, reassignment of function in the course of evolution [11,12].

The steps need to be “mangeable”

In either case, the general Darwinian principle applies: evolution must proceed via consecutive, manageable steps, each one associated with a demonstrable increase in fitness. Darwin did not use a specific term for this crucial tenet of evolutionary biology; we will call it the Continuity Principle, following the recent insightful discussion of this issue by Penny [8]. The developments in the 150 years since Darwin taught us to be more flexible about this principle than he was. It is no longer prudent to demand, as Darwin did, that all evolutionary changes are “infinitesimal”; some genome modifications may have had a substantial one time effect on fitness, e.g., those that involve horizontal gene transfer, gene loss, or genome rearrangement [13].

and the steps need not be selective since evolutionary science has shown how neutral or even slightly deleterious steps can still become fixed in the genome.

Furthermore, it cannot be demanded that every change is selectively advantageous because neutral or even slightly deleterious mutations can be fixed by drift, especially, in small populations [9,14]. Nevertheless, these newly discovered factors of evolution, however important by themselves, are but modifications of the Continuity Principle – evolution of complex systems still needs to be deconstructed into successive steps and explained in a Darwinian way.

The authors explain that the chicken and the egg problem of the “Central Dogma” can be resolved via an RNA-alone step, also known as the RNA world. This conclusion is founded on many scientific findings

The unique property of RNA that makes it a credible, indeed, apparently, the best candidate for the central role in the primordial replicating system is its ability to combine informational and catalytic functions. This notion has been greatly boosted by the study of ribozymes (RNA enzymes), which was pioneered by Cech and coworkers’ discovery, in 1982, of the autocatalytic cleavage of the Tetrahymena rRNA intron [38], and by the demonstration, in 1983, by Altman and coworkers, that RNAse P is a ribozyme [39]. Since the time of these seminal discoveries, the study of ribozymes has evolved into a vast, expanding research area (at the time of this writing, March 1, 2007, the keyword ‘ribozyme’ retrieves 4883 documents from the PubMed database; for recent reviews, see [40-43]).

Of course, the RNA world scenario is further strengthened by the fact that the RNA world can still be found in living organisms

It is often noted that the RNA World is not just a concept supported by the catalytic prowess of ribozymes: while overshadowed by the multitude of proteins with catalytic and structural functions, the RNA World still lurks within modern life forms

The authors conclude

To recapitulate, three independent lines of evidence converge in support of a major role of RNA, and in particular, RNA catalysis at the earliest stages of life’s history, and are compatible with the reality of a complex, ancient RNA world that was first postulated by Woese, Crick, and Orgel on purely logical grounds. First, comparative analysis of the protein components of the translation machinery and their homologs involved in other functions strongly suggests that extensive diversification of the protein world took place at the time when the translation system was comprised, primarily, of RNA. Second, several classes of ribozymes operate within modern cells, and their properties are compatible with the notion that they are relicts of the ancient RNA world. Third, while limited in scope and, obviously, inferior in catalytic activity compared to protein enzymes [41], ribozymes have been shown or, more to the point, evolved to catalyze a remarkable variety of reactions including those that are central to the evolution of translation (Table 1).

The authors then continue, in good scientific fashion, to play the ‘devil’s advocate’ and explore common objections and weaknesses of this scenario. While ID proponents typically perceive this step as a weakness, science has come to rely on it as a strength.

The authors then address the origin of the genetic code, which is strongly linked with the origin of the translation system

To understand how translation might have emerged, the nature and origin of the codon assignments in the universal genetic code are crucial.

Typically there are three proposed scenarios

  1. Stereo-chemistry

  2. Selection

  3. Historical contingency: also known as ‘frozen accident’

Although the authors seem to prefer the ‘frozen accident’, one of the reviewers, the well known genetic code researcher, Bob Knight, observes that the novelty of their scenario does not depend on the exact route through which the genetic code evolved. In another posting I will attempt to discuss the various scenarios and how they stack up, needless to say, science has uncovered plausible scenarios and supporting evidence, even though the origin of the genetic code is likely a historical event hundreds if not thousands of million of years in our past.

How can they do this? By using plausible scenarios based on chemistry, physics and an understanding of plausible prebiotic atmospheres.

(See for instance Signature of a Primitive Genetic Code in Ancient Protein Lineages, in J Mol Evol. 2007 Oct 6)

The genetic code is the syntactic foundation underlying the structure and function of every protein in the history of the biological world. Its highly ordered degenerate complexity suggests an incremental evolution, the result of a combination of selective, mechanistic, and random processes. These evolutionary processes are still poorly understood and remain an open question in the study of early life on Earth. We perform a compositional analysis of ribosomal proteins and ATPase subunits in bacterial and archaeal lineages, using conserved positions that came and remained under purifying selection before and up to the most recent common ancestor. An observable shift in amino acid usage at these conserved positions likely provides an untapped window into the history of protein sequence space, allowing events of genetic code expansion to be identified. We identify Cys, Glu, Phe, Ile, Lys, Val, Trp, and Tyr as recent additions to the genetic code, with Asn, Gln, Gly, and Leu among the more ancient. Our observations are consistent with a scenario in which genetic code expansion primarily favored amino acids that promoted an increase in polypeptide size and functionality. We propose that this expansion would have been critical in the takeover of many RNA-mediated processes, as well as the addition of novel biological functions inaccessible to an RNA-based physiology, such as crossing lipid membranes. Thus, expansion of the genetic code likely set the stage for the transition from RNA-based to protein-based life.

The authors then continue to propose their scenario as well as experiments to test the scenario. In their discussion, the authors clarify their guiding principles, the speculative nature of their claims, and yet they also point out how their claims follow from simple guiding principles of science. Their proposed explanation is based on constraints

Thus, our main incentive with the present analysis was to deconstruct the formidable problem of the emergence of translation into a series of plausible and manageable steps, in accordance with the Continuity Principle. We believe that, in doing so, we achieved a somewhat greater level of detail and coherence than any of the previous models we are aware of. Importantly, in constructing this model, we were both constrained and driven by: i) comparative-genomic data, ii) experimental data on amino-acid-codon recognition, iii) experimental data on the diverse catalytic activities of ribozymes.

The authors then repeat their guiding principles

Discussion and conclusion

The status of the model: incentives and constraints The scenarios for the origin of the translation system and the genetic code outlined here are both sketchy and highly speculative. Why, then, bother building such conceptual, qualitative models at all? The justification for this kind of theorizing can be succinctly put in the short phrase: we have to get from there to here. There being the early, cooling earth with no complex organic molecules, and here being a minimally complex genetic system with modern-type translation, transcription, and replication machineries, a system that would be subject to biological evolution much like modern organisms. The replication and transcription problems are, at least, logically relatively straightforward, even if hard from the chemical point of view, inasmuch as no new principles, beyond base complementarity, and enzymatic catalysis need to be invented. Thus, plausible, even if conflicting, accounts of the emergence of these systems have been derived from comparative-genomic data and evolutionary reasoning [70,140-144]. There is, however, a crucial snag about these models: they all rely on a pre-existing translation system. And the origin of the translation system is far from being a trivial matter. The main difficulty is not even its complexity per se but the necessity to invent a new principle, that of the genetic code, the correspondence between the a priori unconnected sequences of nucleotides and amino acids. It might not be much of an exaggeration to note that, at least, at first glance, the origin of the translation system evokes the scary specter of irreducible complexity.

Thus, our main incentive with the present analysis was to deconstruct the formidable problem of the emergence of translation into a series of plausible and manageable steps, in accordance with the Continuity Principle. We believe that, in doing so, we achieved a somewhat greater level of detail and coherence than any of the previous models we are aware of. Importantly, in constructing this model, we were both constrained and driven by: i) comparative-genomic data, ii) experimental data on amino-acid-codon recognition, iii) experimental data on the diverse catalytic activities of ribozymes.

Comparative-genomic analysis indicates that an elaborate translation system, comparable to the modern one in terms of fidelity and efficiency, has evolved within the RNA world. Indeed, extensive diversification of many protein folds occurred before the advent of some of the essential components of the modern translation system, such as aaRS and translation factors. Before the emergence of these dedicated proteins, the translation system must have been a machine comprised primarily, if not exclusively, of RNA. The only conceivable alternative, that the primordial translation system employed a different, currently, extinct complement of essential protein factors, inevitably leads to infinite regression. Thus, it seems to be a virtually inevitable conclusion that the ancient, RNA-only translation system was comparable in efficiency to the modern one. This might seem paradoxical and even not credible at a superficial glance. However, a quick reflection suggests that: i) the skeleton of the modern translation system actually consists of RNA, with the proteins being elaborations, however numerous and important, and ii) logically, it hardly could have been otherwise: indeed, in order to switch to a new type of constituents (proteins), biological systems needed the means to produce them accurately. It is conceivable and, indeed, likely that peptides produced by the first, RNA-based proto-translation systems provided positive feedback leading to hypercycle formation (Figs. 4, 6). However, this primitive version of translation must have been quite sloppy and hardly could master production of anything beyond relatively short peptides. Evolution of the (nearly) complete set of tRNAs was a pre-requisite for achieving the fidelity required to kick off protein evolution in earnest.

In our description of the model, the alternative scenarios based on CRM, ARM, and FAM are considered on equal footing. As discussed above, the currently available data are too ambiguous to conclude which of these models for the origin of coding is most likely. However, it should be noted that, important as they are in terms of the actual physico-chemical underpinning of the code, the differences between CRM, ARM, and FAM do not translate into major modifications of the evolutionary scenario. Indeed, the central principles remains the same, i.e., specific recognition of amino acids by proto-tRNAs such that an amino acid is paired with the cognate anticodon with sufficient reliability.

Lasting principles and ephemeral details The models presented here were deliberately constructed at the level of considerable detail -at the risk of getting many, perhaps, most aspects wrong – in order to provide a proof of principle, i.e., to illustrate a plausible sequence of selectively advantageous steps along the path from the RNA world to the modern-type translation system. This being said, there seem to be several underlying principles that are likely to stand regardless of further developments. We briefly recapitulate these:

  1. Evolution having no foresight, selection for translation per se is not feasible.

Translation must have evolved as a by-product of selection for some other function, i.e., via the exaptation route.

  1. Given that the essence of translation is the intimate link between RNA and proteins, it seems most likely that, in some form, this connection existed from the very beginning of the evolutionary path from the RNA World to translation. Thus, the proposed starting point, i.e., stimulation of ribozymes by amino acids and peptides seems to be a strong, almost, logically required, candidate for this role (see also [145]).

  2. Synthesis of peptides directly on an RNA template is stereochemically unfeasible. Hence adaptors must have been part of the primordial translation system from the start. Accordingly, from the very onset of translation, adaptors have been key to the establishment of the genetic code. These ancestral adaptors, although, in all likelihood, smaller and simpler than modern tRNAs, must have been endowed with catalytic capacities lacking in the latter, i.e., they would have to catalyze specific self-aminoacylation with the cognate amino acids.

  3. The primordial translation system was dominated by RNA although peptides might facilitate its functioning. However, the fidelity of this primordial, (nearly) RNA-only translation system must have been comparable to that of modern translation systems, considering that extensive protein evolution took place prior to the diversification of the proteins that are essential for the modern translation.

Problems and testability The current scenario for the evolution of translation in the RNA World faces formidable difficulties because, although the ribozyme catalysis of the elementary reactions required for translation has been demonstrated experimentally (Table 1), the required complex RNA-mediated functions have not. The crux of the problem seems to lie in the postulated catalytic adaptors that would have to possess a notable spectrum of capabilities including, in addition to the apparently feasible specific recognition of amino acids and self-aminoacylation, the ordered binding to the progenitor of the large subunit (RL), and at a subsequent stage, recognition of a specific region in the progenitor of the small subunit (RS). With regard to RL and RS themselves, ribozyme stimulation by amino acids and peptides has been demonstrated but, beyond that, the postulated properties of these molecules remain hypothetical. It seems that a focused experimental effort aimed at the construction/selection of ribozymes with the properties of the postulated T RNAs, in particular, their postulated interaction with other, more complex ribozymes, could provide crucial evidence in support of this or a similar scenario for the evolution of translation.

Although the individual ribozyme-catalyzed reactions involved in the postulated scheme are feasible, the succession of multiple evolutionary steps that appear to be required for the emergence of translation might be legitimately viewed as far fetched, particularly, considering the inevitably inefficient ribozyme-mediated replication that must have been prevalent in the RNA World. Be as it may, this is, at present, our best effort to develop a conceptual model for the origin of translation. Elsewhere, one of us (EVK) examines a radical alternative [17].

As far as experimental verification

It seems that a focused experimental effort aimed at the construction/selection of ribozymes with the properties of the postulated T RNAs, in particular, their postulated interaction with other, more complex ribozymes, could provide crucial evidence in support of this or a similar scenario for the evolution of translation.

And that my friends is how science works. Since I do not want to pretend to be speaking for Intelligent Design, I invite ID proponents to share with us their guiding principles, and show how their scenario explains the origin of the translation system better, given the constraints and known data.

Read the full article:

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 2007; 2: 14.