Highlight: Tiny Bacterial Genome Opens a Huge Mystery: AT Mutational Bias in Hodgkinia

Abstract

The US southwest is home to a remarkable bug. Cicadas, a common sap-feeding insect, are most known for their ear-splitting mating songs, which seem at times to dominate the air waves. However, inside specialized cells in their bodies, cicadas host less known but perhaps more remarkable bugs—of the bacterial variety. A cicada's bacterial symbionts provide the insect with the essential amino acids it is not able to get from its diet. One of those bacteria, Hodgkinia cicadicola, owns a very special genome. It is remarkable for its tiny size (the second smallest known), its rich guanine and cytosine (GC) content, and its potential to turn the study of genetics on its head. Hodgkinia's ability to retain a high GC content, as reported in the latest issue of Genome Biology and Evolution (Van Leuven and McCutcheon 2011) poses a paradox, given geneticists' current understanding of genome dynamics. It suggests that selection may be effective in a smaller population than generally believed. If true, it would require a general revamping of the models used in population genetics and phylogenetics. The data set has a backstory. It was collected as part of a systematic genome sequencing of Alphaproteobacteria, the symbionts in this group of sap-feeding insects, using high-throughput genomics. Primary author John McCutcheon, of the University of Montana, Missoula, collected cicadas from the environs of Tuscon, Arizona. Entomologists assured him the cicadas he found would be a single species, important since cicadas and their symbionts co-speciate. He pooled the DNA from all samples, but in the end, a look at the resulting sequence indicated that two bacterial species were present. It was impossible to tease apart which fragment of genome was from which species, McCutcheon says. He had both uncovered a new species of cicada in the area and rendered his results useless, at least at first. “Sort of naively, I saved all of the experiments,” McCutcheon says. Going back and collecting samples from ten individual cicadas of one species, he sequenced them again and published the genome of one Hodgkinia species. “One of the coolest and weirdest things about this genome was that it had, at the time, the smallest bacterial genome known, but it also has a high GC content.” Bacterial genomes vary hugely in GC content: from 13% at the spectrum's low end up to 75%. For decades, researchers have sought to understand why this spread should be so great. Many assumed that the errors in DNA replication are biased toward different directions in different bacteria. In 2010, though, two papers published in PLoS Genetics challenged this assumption. They showed that mutations seem to be universally biased toward adenine (A) and thymine (T), meaning that any point mutation is more likely to change a GC pair to AT than AT to GC (Hershberg and Petrov 2010; Hildebrand et al. 2010). Bacteria tend to slough off huge amounts of their genome when they go from free-living organisms to symbiotic ones. Given the AT mutational bias, the genome in time should become increasingly AT rich. Almost all empirical data support this hypothesis. In fact, the two bacterial genomes with the highest AT content known to date are small-genome insect nutritional endosymbionts. Two forces are thought to drive this shift. As bacteria become symbionts, genes used for DNA repair and recombination often end up on the cutting block. Mutations are therefore less likely to be patched up. Also, there is very little genetic mixing among endosymbionts due to small effective population sizes and less recombination. Deleterious mutations are more likely to spread because natural selection is thought to be very weak. Curiously, the two known exceptions are the two genomes that should be most strongly shifting toward AT. Hodgkinia cicadicola and Tremblaya princeps have the smallest reported genomes. Yet both have retained a high GC content. (The Hodgkinia examined in this study stems from a group in which most free-living members have large high GC genomes.) The 2010 PLoS Genetics papers caused a stir in the community. “They certainly got a lot of people thinking, myself included,” McCutcheon says. “But both [papers] say, ‘Hodgkinia is pretty weird, it's an outlier. It would be interesting to look at its mutational pattern.’ I thought, ‘I agree!’” It occurred to McCutcheon he could do exactly that. “We had the complete genome from one population. And also, because of this mistake I made, we also had a closely related outgroup genome. So we could unambiguously find the direction of mutation.” Hodgkinia, somewhat mysteriously, shares in this universal bias toward AT mutations. Yet the two forces that are hypothesized to select for increased GC content in other bacterial genomes should not be at play here. Natural selection, as currently understood, should be too weak. Biased gene conversion requires recombination. But these bacteria are not believed to recombine. Natural selection is currently understood in light of the neutral theory of molecular evolution. This theory, developing for approximately the past 40 years, states that most fixed point mutations are neutral, causing no change in functionality. “This assumption allows for nice mathematical properties that facilitate the use of the models we use in population genetics and in phylogenetic studies,” says Eduardo Rocha, a member of Microbial Evolutionary Genomics group, at France's Pasteur Institute. These models implicitly assume that most fixed mutations are neutral, not affected by natural selection. But the only explanation that would seem to explain why Hodgkinia retains a high GC content is that selection is able to detect and quickly purge all As and Ts generated by mutation. Rocha co-authored a 2010 PLoS Genetics Perspective summarizing the score of hypothesis explaining the mutational patterns in bacteria. “And none really works.” McCutcheon's analysis is atypical, Rocha says, since the sequenced DNA came from a population not a strain of cultured bacteria. “But basically I think they’re right, that selection is keeping GC contents high.” “If every single mutation has been imprinted by natural selection,” says Rocha, “It means we have to review all the models we’re using to analyze DNA sequences.”