Darrel Falk Badly Mischaracterizes RNA World Experiments. . .and Stephen Meyer
On April 16, BioLogos posted the first critical review of Return of the God Hypothesis. The review, written by biologist Darrel Falk, offered no criticism of the book’s new design or cosmological arguments based on physics, astronomy, and cosmology, but instead critiqued only those portions of the book where Meyer reprised or updated his previous arguments for intelligent design based on biology. Much of Falk’s review repeated criticisms that Falk had made in reviews of Meyer’s previous books — critiques that Meyer and others have already refuted in detail (here, here, here, and here).
Falk did advance one new (for him) objection to Meyer’s critical analysis of chemical evolutionary theory — a critique that Meyer makes as part of his overall positive case for intelligent design as the best explanation for the origin of the information necessary to produce the first living cell. In his review, Falk claimed that recent “RNA world” simulation experiments challenge Meyer’s claim that origin-of-life researchers have been unable to produce RNA molecules capable of copying more than 10 percent of themselves. As he asserts:
…in discussing the hypothetical RNA world and the origin of life, Meyer writes, “To date, scientists have been able to design RNA catalysts that will copy only about 10% of themselves” (p. 280). He then references a paper from 2001. However, the field has progressed quite well in the past twenty years.
To support this challenge, Falk cites an experiment from 2014 that he claims did succeed in demonstrating that RNA molecules can replicate themselves in their entirety. Here’s what Falk claims:
For example, in 2014, Robertson and Joyce reported a similar system with a tweak which resulted in 100 percent effectiveness. They summarized their results with these words: “Each parental enzyme can give rise to thousands of copies per hour, and each of these copies in turn can do the same, all the while transmitting molecular information across the generations.”
Falk then reiterates his claim that Meyer misled his readers by not mentioning this allegedly significant newer result. Says Falk of the Joyce and Robertson experiment: “Although they changed the enzymatic reaction a bit, this comes very close to being what Meyer tells the reader has not been done.”
So what is the story here? Did Meyer, who wrote his Cambridge PhD thesis on origin-of-life biology and who has written an acclaimed book on the subject, mislead his readers? Have origin-of-life researchers actually produced a genuinely self-replicating RNA molecule capable of replicating itself in its entirety “with 100% effectiveness.” Have they done so under realistic or plausible pre-biotic conditions without extensive investigator interference — something that Meyer also insists has not been done? If so, was Meyer ignorant of these allegedly dramatic advances in the field?
The answer to all these questions is a resounding no. In fact, Meyer advanced a perceptive and decisive critique of an earlier, nearly identical version of the experiment that Falk cites — an experiment performed in 2009 by Joyce and Lincoln. Yet Falk, in faulting Meyer for failing to report allegedly new experimental evidence in support of the RNA world hypothesis, completely failed to mention that Meyer already addressed Joyce’s supposed demonstration of RNA self-replication. That was in a defense of his book Signature in the Cell in an exchange in the (London) Times Literary Supplement (No. 5572, p. 6) in 2010 with another critic who made the same claim that Falk now makes. Falk also failed to mention that Meyer critiqued the same claim that Falk is now making about Joyce’s experiment on page 309 of Return of the God Hypothesis — the book that Falk claims to have read as the basis for his review.
Gerald Joyce’s Experiments
So what about Joyce’s experiments? Did they show that RNA molecules can self-replicate more than 10 percent of themselves under plausible prebiotic conditions and without intelligent intervention — the specific claim that Meyer disputes.
No, they did not. Instead, here’s what Joyce and Robertson, and earlier Joyce and Lincoln, actually did.
In these experiments, Gerald Joyce and his colleagues demonstrated that a specifically designed RNA enzyme (or “ribozyme”) that they designated as E could link together two partial strands or halves of another RNA molecule (which they called the RNA substrates A’ and B’). The resulting new RNA enzyme (designated E’) could then join together two parts of the original ribozyme (RNA substrates A and B). The longer strands fused together by this process (that is, ribozymes E and E’) could then repeatedly fuse together the two halves of the opposite ribozyme if (1) a continuous supply of the two halves (either A’ and B’ or A and B) were provided in ample amounts to the experiment and if (2) critical protein enzymes were also introduced into the experiment at specific times.
Here is a figure that depicts the entire process.
Though Falk, and the researchers themselves, give the impression that these experiments produced a self-replicating system that simulates “self-sustaining Darwinian evolution,” they in fact did no such thing. Nor did they produce an RNA molecule that could copy more than 10 percent of itself or, still less, one that could reproduce itself with “100% effectiveness” and do so under plausible prebiotic conditions.
Ligation, not Polymerization or Replication
In the first place, Joyce and colleagues did not produce a genuinely self-replicating molecule. As envisioned by RNA World proponents, the emergence of a self-replicating RNA molecule is the crucial step in the emergence of the first life on earth since only after the emergence of such a self-replicating molecule would something like natural selection and random mutation begin to occur.
Moreover, in the RNA world scenario a self-replicating RNA molecule would emerge only after (1) the chemical subunits of RNA formed on the early earth and then (2) those subunits linked together in specific ways to form an RNA molecule capable of producing copies (and near copies) of itself. RNA world researchers envision such self-replication occurring as the result of a ribozyme (specifically an RNA replicase) using a complementary copy of itself as a template to produce another copy of the original strand from free-floating RNA subunits (in particular, activated RNA nucleotides).
Nevertheless, as Meyer has repeatedly noted, the molecules in Joyce’s experiment do not demonstrate the capability for such template-directed self-replication — a capability that RNA world advocates envision as crucial to the process of life originating from RNA molecules. Such self-replication necessarily requires the ribozyme to function as a polymerase — in other words, the ribozymes need to have the ability to link many nucleotide bases together to form long RNA chains. The ribozymes in the Joyce experiments do not perform this action. Instead, they catalyze (ligate) a single linkage between two ends of two pre-made, pre-sequenced halves or sections of RNA — sections that, once linked, will become a separate RNA chain that folds into a ribozyme. Thus, the RNA enzymes in Joyce’s experiments function as simple ligases rather than polymerases or replicases.
Moreover, as noted, although Falk portrays Meyer as being unaware of Joyce’s work demonstrating the ability of ribozymes to perform extensive feats of self-replication, Meyer had already critiqued these experiments showing that they lacked this capability and did so again in Return of the God Hypothesis. As he stated (on p. 309):
The “self-replicating” RNA molecules in this experiment did not copy a template of genetic information from free-standing nucleotides as protein machines (called polymerases) do in actual cells. Instead, in the experiment, a presynthesized specifically sequenced RNA molecule merely catalyzed a single chemical bond, fusing together two other presynthesized partial RNA chains. Their version of “self-replication,” therefore, amounted to nothing more than joining two sequence-specific premade halves together.
This limitation underscores why Meyer has correctly emphasized that simulations of RNA self-replication have failed to produce molecules capable of producing more than 10 percent of themselves. In Joyce’s experiments the single linkages performed by his RNA ligases provide far less than 10 percent of the total number of linkages in the resulting RNA strands (each of which include more than 60 such linkages between nucleotide bases). Indeed, Joyce himself has acknowledged that his experiment merely demonstrates the capacity of RNA molecules to perform ligation not polymerization and, thus, not genuine self-replication. As he noted, his use of “a directed evolution strategy required selecting for the ability to catalyze a simple ligation reaction, rather than replication itself.”
Physical Limitations of RNA Self-Replication
There is an important and more general biochemical consideration that also supports Meyer’s critique. Nucleotide sequences in RNA enzymes must fold into precise three-dimensional structuresbefore they can function as catalysts. But RNA enzymes fold into such structures as a result of complementary bases pairing. When they do so, much of the RNA molecule ceases to present an exposed single strand that can function as an RNA template. Consequently, RNA replicases will almost inevitably encounter a physical limit in the extent to which they can copy themselves — or at least they would encounter such a limit in any realistic prebiotic environment or simulation experiment. Once ribozymes have folded in such a way so as to allow them to function as a replicases or polymerases (or as ribozymes generally) they will also no longer be able to serve as a template that could allow them to be copied by other versions of themselves.
Moreover, because of the strong stereochemical affinities between complementary bases in RNA strands, free-standing single-stranded RNA molecules quickly fold up on themselves to form complex three-dimensional structures with features such as hairpin loops. These facts render extremely implausible all scenarios involving RNA replicases copying themselves in their entirety using free-standing complements of themselves as templates since the templates will quickly fold upon themselves as well. Attwater and his colleagues detail this problem:
However, even the most highly-evolved RPRs [RNA polymerase ribozymes] are substantially impeded by template secondary structures. Such structures are ubiquitous in larger, functional RNAs (including the RPRs themselves) and generally indispensable for function. The strong inhibitory role of this central feature of RNA leads to an antagonism between the degree to which an RNA sequence is able to fold into a defined three-dimensional structure to encode function (such as catalysis) and the ease with which it can be replicated.
Protein-Enzyme Mediated Replication, Not RNA Self-Replication
So, in light of all this, how did Joyce and his colleagues produce many complete copies of their original ribozymes E and E’? It turns out the production of the copies of the RNA enzymes in their experiment depends — not on the ability of the RNA molecules to copy themselves — but instead on complex protein enzymes derived from living cells. Specifically, to make more copies of the RNA enzymes Joyce and colleagues employed the reverse transcription polymerase chain reaction (RT-PCR) procedure that requires using two complex protein enzymes — a reverse transcriptase and a DNA polymerase — as well as other molecular tools such as primers. Indeed, in order to make more copies of the most efficient ribozymes (rather than making complementary RNA strands with the opposite bases at each site) this procedure requires turning RNA into DNA and then reconstituting RNA from DNA. But that procedure necessarily employs an RNA reverse transcriptase, as mentioned, and an RNA polymerase — both of which are derived from living bacteria. As Meyer has told me, “Joyce and his team did not produce a self-replicating RNA molecule. Instead, they intelligently designed a system of protein-enzyme mediated replication.” Since these proteins had to be extracted from already living cells, Meyer also commented that “these experiments lead to the paradoxical conclusion that simulating a crucial step toward the origin of the first life from non-living RNA molecules requires the use of protein enzymes derived from already living cells.”
There is another reason that these experiments do not demonstrate the capacity of the RNA molecules in the experiment to self-organize or self-replicate. Every crucial step depended upon external guidance — often in the form of inputs of functional sequence-specific information — from highly intelligent chemists, in particular, Gerald Joyce and his colleagues.
Consider first that Joyce intelligently designed the larger ribozymes designated E and E’ that could link each other’s halves together. To build a precursor ribozyme in an original 2001 experiment, Joyce started with a random crop of 100 trillion RNA molecules with many different nucleotide base sequences. He then repeatedly applied chemical screens to select out those few RNAs that could perform ligation and performed it best (Rogers and Joyce 2001).
Next, he selectively altered the base sequences in particular regions of these RNAs to enhance their ability to link together the halves of a duplicate strand. For example, he wanted the ribozymes to be able to bind strongly enough to the complementary base pairs on the substrate molecules (i.e., A and B) and yet not to bind so strongly as to prevent the larger ribozyme from breaking away once the two RNAs halves had linked together. Thus, Joyce not only used his intelligence to select molecules that could perform the function that he wanted from a random crop, he also optimized the function of these ribozymes through modifying carefully chosen regions (Paul and Joyce 2002).
Joyce then altered the original RNA enzyme (which he called T) in order produce two new ribozymes (which he called E and E’) that would have the ability to link the two halves of each of these new enzymes together — where E would link together A’ and B’ to form E’, and E’ would link together A and B to form E. By his own admission, he used what he characterized as a “rational design” approach to create this mutually interdependent, cross-catalyzing system. He specifically arranged the RNA base sequences in the “paired regions” of the two enzymes so that they would bind by complementary base pairing to the substrates. In addition, the regions near the ends of the break between the two halves of E and E’ had to be engineered to ensure that a ribozyme-mediated linkage could occur (Lincoln and Joyce 2009).
All this implies that Joyce necessarily had to design the pre-made, sequence-specific halves (i.e., both A and B and A’ and B’) that his ribozymes would join together. Indeed, the break point between the two halves needed to be at just the right location in order to ensure that ligation would occur. As mentioned, the arrangement of the nucleotide bases on the pre-made halves needed to be precise so that they would bind to their opposite base on the ribozyme by complementary base pairing. Meeting these specifications required Joyce’s repeated, active, and intelligent intervention in his experiment.
Once Joyce had designed this cross-catalyzing system, he used “directed evolution” in an attempt to improve the efficiency of the ligase ribozymes. His team started by altering specific positions in the original ribozymes to generate numerous variants of E and E’ in the 2009 study and to generate numerous variants of E in the 2014 version of the experiment. They then isolated the variants that demonstrated the most efficient substrate-joining (ligase) function and differentially reproduced those. The 2014 study also tested for the variants’ ability to link their own half-strands together as well as the half strands of the opposite ribozyme.
Clearly, this process also required extensive investigator guidance and intelligent design. For example, Joyce and his colleagues employed advanced laboratory techniques to generate trillions of variants of the original enzyme(s) and trillions of copies of the substrates. They then executed equally advanced procedures such as “selecting the reacted, biotinylated products by capturing them on a streptavidin-agarose resin” to tag and capture the variants that most efficiently joined substrates (Robertson and Joyce 2014). One cannot overstate the implausibility of comparable processes occurring outside of an advanced laboratory setting staffed with highly trained and intelligent technicians, let alone on the pre-biotic earth, presumably devoid of any source of intelligent guidance (here, here).
Indeed, as Meyer argued in Return of the God Hypothesis (p. 310),
…whenever chemists set up or interfere in a reaction sequence — or whenever they otherwise apply constraints to a chemical system — to ensure one outcome and preclude others, they effectively input information into that system. In so doing, they inadvertently simulate, if anything, the need for intelligent design to generate biologically relevant chemistry and information.
Moreover, Meyer specifically applied this critique to Gerald Joyce’s ribozyme engineering experiments in his discussion of them in in RGH (p. 309). As he notes:
Lincoln and Joyce themselves intelligently arranged the base sequences in these RNA chains. They generated the sequence-specific functional information that made even this limited form of replication possible. Thus, the experiment not only demonstrated that even a limited capacity for RNA self-replication depends upon information-rich RNA molecules; it also lent additional support to the hypothesis that intelligent design is the only known means by which functional information arises.
A Final Thought on Falk
One final observation about Falk’s review must be offered. As much as I appreciate his interest in Meyer’s book, and his willingness to engage the arguments that Meyer makes, there is something oddly tone-deaf about an attempt at engagement that critiques a book for its ignorance of a new scientific result when the book itself has already described and critiqued the result in question. I am quite willing to give Professor Falk the benefit of the doubt and to assume that he merely overlooked the passages in Meyer’s work where Meyer discussed and critiqued the very experiment (albeit in only a slightly different and earlier form) that Falk cites. But I would suggest that Falk now acknowledge, as a matter of scholarly integrity, that he mischaracterized Meyer’s argument. He should retract that portion of his review that left his own (Falk’s) readers with the misimpression that Meyer was ignorant of developments of great consequence in origin of life simulation experiments when instead, as I have shown, nothing could be further from the truth. Meyer was not ignorant of these experimental results. Nor did they significantly advance our understanding of how RNA might have acquired a capacity for genuine self-replication, let alone how it might have done so under plausible pre-biotic conditions on the early earth.