The Fleagle Factor

Abstract

You've written your paper, presenting your work as best you can and believing that it's not only important but clear. Then you're sent the reviewers' comments. Thinking arrogantly that they know more about your work than you do, and speaking from behind the cloak of anonymity, they tell you how to write your paper in order for it to be “acceptable” for the journal and to achieve the influence it deserves. If they disagree with you, or have had other ideas themselves, they naturally judge your paper unworthy of being seen in public. They don't want you to get much credit, and point out that they said the same thing first and should have been cited. And, disguised behind “helpful” comments, they try to force you to write the paper they would have written instead. You grumble, make your co-authors reluctantly do their best to cover all bets with perfunctory citations and unwarranted caveats to please the reviewers, and you send back the manuscript. You have to write an accompanying letter, as lengthy as the paper, telling the editor how you've kowtowed to the reviewers' every whim. With the perfunctory agreeable conclusion, you say you hope your gem is now satisfactory. But no, it's not over yet! Even if the paper is accepted, the proofs come back with yet more corrections made by the editor, the copy editor, and who knows who else. It may be frustrating, but we all understand that the editor has the final say. We can refer to this as the Fleagle Factor, after the editor of the journal you're now reading, who has made it a leading journal and a pacesetter in the field (Fig. 1). But editorial oversight is not just about people working in evolutionary anthropology. It is, instead, a surprising fact of biology. External forces pass judgment on the quality of an organism's work, and an editor has the final say. But before we can pass judgment on nature as editor, we need to consider a few basics, for readers who may not be well-versed in them. The regime begins. Left. The first Evolutionary Anthropology cover, 1992. Right. John Fleagle, Editor (well, in younger days(. Source Wikimedia. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.] DNA is a molecule that contains much of the information whereby diversity in living organisms is specified. DNA is a string of molecules called nucleotide bases, of which there are four, A, C, G, and T (the first letters of their chemical names). Remarkably, from a chemical point of view these can be strung together like a charm bracelet, in any order and any length. The order of these bases, established by evolution, is what carries biological information. DNA is a double chain of two such strings, paired one-by-one. For chemical reasons, As are always paired with Ts and Cs with Gs. This base-pairing allows DNA to be copied into two identical descendant copies. When a cell divides, each cell gets a full double-stranded copy because when the two strands are separated in a cell, there are mechanisms to match each element with its natural pair. Furthermore, for many of its important functions, DNA itself is a carrier of information that is transcribed by being copied, also using complementary base-pairing, into a correspondingly ordered chain, called RNA, which is only single stranded. RNA copies of DNA sequences work in two major ways. First, they can have their own function. They can fold up upon themselves, using the same kind of base-pairing that keeps DNA double-stranded; thus, a C in one place in an RNA molecule can pair with a G somewhere else along its chain. This gives the RNA molecule a three-dimensional structure that allows it to interact with other molecules in the cell. There are many such functional forms of RNA. Although it is debated, at least some of these RNA functions probably evolved before there even was any DNA to code for it. Second, a form of RNA called messenger RNA (mRNA), doesn't use its nucleotide sequence to form a physical shape, but instead uses it to specify the sequence order of amino acids that will form a protein. As with folded-up RNA, proteins are amino acid strings that, because of their properties, fold up; their folded shape is what has biological function. Again, it is the order of nucleotides in mRNA that specifies the order of amino acids. The usual evolutionary idea, well established and clearly true, is that mutations that change nucleotides in DNA can affect function by changing the order of amino acids that are coded for. Hence, mutations change the protein's function by altering what, or how effectively, the coded protein interacts with in the cell. If the interaction is not very good, the organism that bears the mutation may not succeed in life and not reproduce as a result, and the mutation disappears from the population. If the mutational change is salubrious, its bearers may reproduce like rabbits, proliferating the change. The screening process for such changes is what we call evolution by natural selection (that's what made rabbits). For several decades, we used DNA sequence data to search for functional effects because we knew how to read the code: From a DNA sequence, we could see how an RNA molecule would fold up by identifying complementary base pairs along its strand, or what amino acids it would code for in a given protein. The underlying idea was that DNA is a code, the fundamental information source of life, which was directly transcribed into function. Nature “edited” the functional result, but only through the evolutionary judgments of natural selection. Not so long ago, something quite unexpected was discovered. The amino acid sequence of some proteins turned out not to be what we read off from the DNA. One or more specific amino acids in the protein, or its mRNA code, differed from what we inferred from the DNA code itself. The deviations from the code were replicable, not due to laboratory error, and could even be conserved among species. This was very mysterious. Among the clearest and most thoroughly studied examples are instances in which an A in DNA is replaced by another nucleotide, called inosine (I), which is not normally part of the DNA-RNA system. This occurs by way of what is known chemically as a deaminase reaction (Fig. 2). An I in an RNA string can alter its function because while A pairs with T (RNA actually uses a nucleotide denoted U rather than T), I pairs with C, as if it were G. RNA editing means that an mRNA molecule edited in this way could then code for a different amino acid, which can harm the function of the coded protein; alternatively, an altered functional RNA molecule may not fold up into the same three-dimensional shape that it would with the original A in its sequence (Fig. 3). This is called RNA editing—the genomic Fleagle Factor. A deaminase reaction altering an A to an I, which will now act as if it were a C, and pairs with G. The NH2 group is replaced by an oxygen atom and the N also changes to a NH. How an edited RNA can have new function. Top, the 3 by 3 amino acid code is changed by an A→I RNA edit (3-letter amino acid abbreviations shown under each coding triplet(. Bottom, RNA is transcribed as a single-stranded copy of the “coding” strand of DNA, but can fold up upon itself by complementary base-pairing, C with G and A with U (which, in DNA, is a T). I doesn't normally occur in normal DNA and RNA, but acts as if it were a G and pairs with C, changing the folded shape of the RNA and thus also its function. An unusual behavior in the humble fruit fly was found to be caused by mutation in a gene called adenosine deaminase acting on RNA (ADAR), which codes for a protein that normally brings about deaminase reactions.1, 2 The normal ADAR protein recognizes specific places in a folded-up RNA where the target A is exposed and converts it to an I. But in the mutant, it is the unedited mRNA that codes for an aberrant protein in the fly's brain. This protein affect the fly's behavior, leading to lack of coordination, seizures, and other anomalies.1, 2 This and other genes carry out their normal function only when they have been edited, meaning, in turn, that the ADAR gene has evolved to be a normal, indeed necessary part of our genomes. RNA editing is not a fluke. A mutant ADAR gene is like the Fleagle Factor, a normal editorial function, being asleep at the editorial desk (John, the real Fleagle, never is). These genetically based functions are long established in evolution, so they have a long history of functional importance. There is one ADAR gene in flies, but two evolutionarily related ADAR genes (ADAR1 and ADAR2) in mammals. In fact, ADARs also exist in invertebrates. A major source of evidence for ADAR is obtained by bioinformatics, or computer-based searches for erasures and cross-outs. The DNA sequence for a given gene is compared to RNA that has been extracted from cells that are using that gene. In the case of editing of the ADAR type, one finds As in the DNA that are replaced by Is in the RNA. When the same discrepancy is found enough times to rule out sequencing errors, one begins to believe that it's real. In cases like those of the fly, the same sites are not edited in ADAR-deficient individuals. Variation was necessary for the editing system to have arisen and evolved in the past. One might expect that, like everything else in life (except death and taxes), editing is variable and probabilistic. That is indeed the case. One systematic large-scale study found 239 genes in which sites were edited in all of seven tested tissues: cerebellum, frontal lobe, corpus callosum, diencephalon, small intestine, kidney, and adrenal gland.3 Stress was on neural tissue because prior work like that with strangely behaving flies suggested that RNA editing was important in such cells. Subsequent work has confirmed that high neural editing activity level in mammals. The same study found 330 additional genes in which editing was not complete. Only a fraction of copies of the mRNAs for the gene were edited. A sense of the reliability of our inferences about these results is reinforced by the fact that there is a statistical bias against such editing in sites within the gene where we have reason to suspect that the editing change may have a harmful effect, such as in the most functionally critical amino acids in a protein.3 This is what would be expected on evolutionary grounds. There is so much interest in RNA editing that an entire online database has been built to serve the needs of research into RNA editing in animals as well as plants and even in simpler forms of life.4 RNA editing is not a newcomer function or an erroneous function, but has been evolved and elaborated upon over hundreds of millions of years. Other genes code for proteins that form a complex with ADAR protein, a structure called an “editosome,” which actually makes the change. This and other RNA editing systems must have been molded by an evolutionary history that involved some sorts of adaptive natural selection. But why? Deaminase reactions are found in ba cteria where they seem to be involved in basic nucleotide processing. Other roles have been suggested as possible, biologically more primitive reactions (see Wikipedia, RNAediting). Subsequently, as with many kinds of systems that are strange to us, editing could have been co-opted for some sort of immune function. Perhaps ADAR editing took on a role to alter the proteins coded by incoming bacteria or viruses so the pathogen's genes don't work right inside the host. It could have been a generic, though not very fine-tuned mechanism in the sense that somewhere in any invader's genes would be something that RNA editing would detect and inactivate out of its normal function, making the infection harmless. Of course for this to work, the host had to be able to survive its own patrolling, so it wouldn't have done in its own RNA. At the same time that the host evolved a way to knock the invader's genes out of their normal function, its own genes may have evolved so that they needed to be edited into their normal function. Further evidence that this is a system for such purposes, in the sense that legs evolved for walking, is that many independently evolved systems of RNA editing, unrelated to the system described here, have been discovered in animals, different systems in plants, and similar functions in bacteria. Various ideas have been suggested for the evolution of editing, but the issues are complex and there are no simple answers. Variable editing potentially enables subtle dose effects of a given gene in cells, if the edited fraction of a given type of mRNA is dependent on the state of the organism at any particular time (such as being asleep or excited(. Gene-expression dose effects have many precedents via other mechanisms and again blur the degree of determinism between the DNA code and its effects on an organism's traits. Indeed, some authors have suggested that these systems may initially have arisen by chance, without any adaptive reasons.5, 6 Once integrated into the genome, such complexity may have been unlikely to be removed by chance, gradually ratcheting ever greater complexity into cells over time.7 In that case, searching for explanations based on natural selection may be at least partly in vain. Over the last few decades, there has been a rather steady stream of discoveries of new functions in DNA. In addition to the long-known function of serving as a code for protein, some bits of DNA sequence are used to package chromosomes so they can fit into the nucleus; some protect chromosomes from damage or separate them when cells divide. Still other sequence elements are recognized by specific proteins that grab onto the DNA at those points to regulate gene to be transcribed into messenger RNA to regulate the gene's use in the cell. The sequence itself leads directly to these functions. Thus, in principle at least, we should be able to learn how to infer the function from sequence data alone. The importance of RNA editing in the overall evolutionary scope is a matter of debate. Surprise or skepticism expressed when such systems are discovered reflects the theory that DNA directly codes for RNA function. How the editing layer of causal complexity arose is almost pure speculation. But that DNA editing is here to stay is not speculative, and shows that our usual views of DNA evolution are at least incomplete. Instead, we find yet another layer of causal complexity between the genome's DNA sequence and the functions that result from its use. Rather than a direct line of communication from DNA to protein, genetic mechanisms work by being filtered through a maze of connections something like cotton wool. If this unexpected indirectness of function changes our understanding of genes and genomes, it doesn't cause any real change in our theories of evolution, so long as we realize that reproduction and natural selection are directly affected by the net result—the organism's traits—that only indirectly reflect the individual's genes. In many ways, the complexity itself becomes a subject of interest. Why would organisms not be as absolutely simple as they can get away with? One possible explanation is that if each function were too starkly dependent on single genes, organisms would be too vulnerable in the face of challenges presented by their environment. As fibers of causal complexity are built into our genomes, they provide more pathways to success and more protective redundancy or buffering against individual genes that may go awry. This doesn't undermine the idea of DNA as cells' main information repository. The system is, after all, inherited in all of its reticulated pathways, since they ultimately reside in the DNA sequence, which ultimately codes for the protein and other RNA codes including those related to the editing mechanism itself. Eventually, change in DNA that arises by mutation of one sort or another either affects traits or doesn't. One of the most powerful aspects of the evolutionary legacy is that it concerns the net result rather than the internal guts of how it occurs. The basic ideas about evolution have been robust enough to withstand a century of discovery of previously unsuspected mechanisms related to how inheritance works, including surviving Darwin's own mistaken ideas about that. The evolutionary legacy also provides an overarching method for approaching biological data: by viewing function and variation alike in terms of gradual descent from common ancestry. For example, DNA sequence comparisons among individuals and species can help identify parts of the genome containing important function without it being necessary for one to know what the function may be or how it works. Evolutionary theory tells us that most of the time important function is conserved because carriers of harmful mutations don't reproduce well, the result being low variability within or conservation among species for the parts of the genome that are involved. The same kind of reasoning applies regardless of how any function is manifest: either it proliferates (for whatever reason) or it doesn't. When screened by natural selection, what changes in frequency is the assemblage of things that make the final organism, regardless of whether they do so directly or indirectly. RNA editing adds layers to mechanisms, but succeeds or not within the same basic processes of evolutionary change. The power of this theory, and the method it provides us, were clear long before the detailed kinds of genetic knowledge that we have today were available, or even suspected. So, even though one can quibble with the details, the core ideas are basically as valid now as they were then.8 The real lesson from the Fleagle Factor in life is, however, profound. It is that RNA editing is but one of many kinds of indirect causation that basically cannot be inferred from DNA sequence alone. The various sequence elements responsible for RNA editing (and some other kinds of DNA function) are scattered across the genome. Only when they work together through the web of intermediaries does the function take place. Since each element can vary from individual to individual, there are more possible variants scattered in more places in the genome, for current informatics-only strategies to try to identify. If such function can't be inferred from DNA sequence alone, that means that many traits can't be predicted from DNA sequence alone. Indeed, here is the real lesson for evolutionary biology, because if we can't predict from DNA to function in a clear-cut way, natural selection can't “see” specific genes when it screens their net result. Even if evolution itself isn't challenged by the cotton-wool nature of life, it is a major challenge in modern evolution and genetics to develop a better understanding of how all this works. With the Fleagle Factor in journals, you can sometimes argue and persuade the editor to let your ideas stand as you want them. At other times, you find, to your surprise, that when a paper you've submitted is finally published, there are many changes that you weren't aware of. Editorial staffers at journals use their proverbial blue pencils to tinker with texts in many ways, not all of them easy to spot even when you look over proofs. But the result is layers of editing between the submitted draft and the final product. Embarrassing goofs—mutational changes that should have been removed by editorial selection—can arise. Like the Fleagle Factor, DNA submits its initial script, which is then checked and revised. I am not impartial enough to say whether nature's editorial judgment is as excellent as this journal's Fleagle Factor. But as in nature, journal readers see only the edited end result, not the primary text. And like a journal editor, nature always has the final say. I welcome comments on this column: kenweiss@psu.edu. I co-author a blog on relevant topics at EcoDevoEvo.blogspot.com. I thank Anne Buchanan and John Fleagle for critically reading this manuscript. This column is written with financial assistance from funds provided to Penn State Evan Pugh professors.