Gene Gains and Losses Shape Arthropod Evolution, Genomic Survey Finds

Genome sequencing has become cheap and powerful enough that scientists can now compare the complete genetic blueprints of hundreds of arthropods, from fruit flies to horseshoe crabs, and ask a question that was once out of reach: how do gains and losses of genes drive the evolution of whole groups of animals? A new review in Development Genes and Evolution by Alys M. Cheatle Jarvela and Judith R. Wexler argues that the answer is reshaping how biologists think about macroevolution, the large-scale changes that separate major lineages over millions of years. The authors survey recent literature on diverse genomic variants and their impacts on phenotypic evolution, focusing mostly on arthropods because the phylum combines extraordinary morphological diversity with an explosion of evolutionary genetics research.

The technological shift is dramatic. As of 2009, the only sequenced arthropod genomes were about a dozen Drosophila species, the honeybee, the silkworm, the flour beetle, and two disease-vector mosquitoes, representing just four orders of holometabolous insects. Thanks to consortia such as i5k, the Earth Biogenome Project, and the Darwin Tree of Life, researchers can now access annotated genome assemblies for more than 490 species spanning over 40 orders through the International Nucleotide Sequence Database Collaboration. Long-read sequencing has been central to this progress. A 2021 meta-analysis concluded that insect genomes assembled with long reads are 48 times more contiguous than those assembled without them, which means fewer missing and fragmented genes and a stronger ability to compare the makeup of distantly related genomes.

But the review is emphatic that finding an interesting variant is only the first step. Identifying a gene gain, loss, or structural change in a genome does not by itself establish that the change caused an evolutionary shift. Demonstrating a causal link between genome variants and phenotypes remains labor intensive, especially in non-model species. The current gold standard is to knock out the candidate region with CRISPR technology, yet in a non-model organism that requires first establishing a protocol for delivering CRISPR-Cas9 to early embryos or the germline, then rearing the mutant animals. Two recent shortcuts have helped. ReMOT Control, described in 2018, fuses Cas9-gRNA complexes to yolk protein precursors that are taken up by ovaries after injection into female insects. DIPA-CRISPR, reported in 2022, skips the yolk-protein tagging step entirely; in the German cockroach, where it was first developed, knockout rates in the first generation reached 20 percent. Even then, mutant individuals are typically crossed with wild-type animals for at least one generation to remove non-specific lesions, work that can take several months to a year depending on generation time.

Gene duplication, an idea famously advanced by Susumu Ohno in 1970, remains a rich source of novelty. Ohno proposed that after a gene is duplicated, the new copy can become raw material for the evolution of a new structure or function. Later work by Force and colleagues added the concept of subfunctionalization, in which two redundant copies divide an ancestral role between them, often through changes in expression domains. Recent research shows what this can look like in its earliest stages. In winged insects, the Hox gene zen is required for forming extraembryonic membranes, structures that provide structural support and desiccation resistance to eggs and helped the group radiate into a hugely successful terrestrial lineage. Within the flour beetle genus Tribolium, zen duplicated into two paralogs, zen1 and zen2, which regulate sequential steps in membrane development through non-overlapping target gene regulation. A negative feedback loop prevents co-expression, a mechanism the authors call iterative subfunctionalization. Crucially, the mechanism behind the paralogs’ distinct functions only came to light with an updated genome assembly and gene set for differential expression analysis, plus new genomic data from related species.

Whole gene families prone to massive duplication pose their own challenges. Chemosensory proteins, which mediate an insect’s perception of the environment, often expand in lineage-specific ways, as with odorant receptors in corbiculate bees, ionotropic receptors in dipterans, and gustatory receptors in the German cockroach. Assaying a single odorant receptor typically requires expressing it in a controlled genetic background with no endogenous receptors, often by inserting it into a Drosophila neuron and recording from that sensillum. One study illustrates the pipeline from genome to function: researchers investigating two recently diverged, sympatric orchid bee species that collect different floral perfume compounds localized selective sweeps to genome regions containing odorant receptor loci, then cloned one receptor with an elevated number of non-synonymous substitutions and expressed it transgenically in a functionally empty Drosophila neuron for single sensillum recording. The inference that duplications of odorant receptors allowed species-specific floral preferences, assortative mating, and eventually reproductive isolation took substantial time and work. Toxins and venoms show a similar tendency to duplicate independently across metazoan taxa, a phenomenon termed convergent amplification, and dissecting their function often means synthesizing candidate genes in bacteria and testing them on animal victims or cell cultures.

Whole genome duplication is harder to detect, particularly when the event is ancient. A 2018 paper claimed evidence for 18 instances of whole genome duplication in insects, but a follow-up analysis found no evidence of macrosynteny supporting whole genome duplication in Lepidoptera, underscoring that synteny conservation remains the gold standard for confirming such events. Several duplications have been confidently identified in arthropods: one in Arachnopulmonata, the group containing spiders and scorpions, and three in the lineage leading to horseshoe crabs. In spiders, duplicated genes are more likely to be expressed in silk-producing glands than in other tissues, suggesting the duplication might have contributed to the evolution of silk production, though the review notes this remains a correlation rather than a demonstrated causal link. In the common house spider, Hox gene paralogs show spatially and temporally divergent expression patterns, and a similar non-overlapping pattern appears in scorpion embryos. Different gene families fare differently after duplication: Wnt family duplicates appear preferentially lost in spiders, and in horseshoe crabs, one arm of duplicated microRNAs is generally more conserved and more highly expressed than the alternate arm. Given the low number of and long divergence times between these events in arthropods, the authors caution that overarching conclusions remain difficult.

Gene loss, long assumed to be a neutral process of paring down redundancy, is emerging as a potentially adaptive force. The i5k Arthropod Genome Pilot Project detected increased rates of gene loss in the lineages leading to Diptera and Lepidoptera, in the lineage leading to Blattodea, and in leafcutter ants, none of which are characterized by dependence on hosts or regressive evolution. One striking example comes from Drosophila: after analyzing odorant and gustatory receptors in five genomes, researchers found that Drosophila sechellia and Drosophila erecta, both specialists on novel host fruits, have preferentially lost gustatory receptors responding to bitter compounds. That loss may have allowed these species to specialize on plants repulsive to other Drosophila lineages; D. sechellia’s host plant is toxic to all other Drosophila species, and detecting bitter compounds helps other drosophilids identify it as toxic. The authors verified true gene loss by detecting corresponding pseudogenes rather than relying on a failure to find hits by BLAST. In another pathway, loss of the photosensitive cryptochrome cry1 gene has been documented in honeybee and flour beetle genomes, with more recent sampling upholding a convergent loss in Coleoptera and Hymenoptera while the gene is otherwise well conserved across insect orders. Little is known about how circadian rhythms work in the species that lost cry1, only that they must function differently from the well-studied Drosophila system.

Gene loss can even strike early developmental transcription factors. The Pax gene paired controls segment patterning in many diverse insect species, yet it is absent from mosquito genomes and transcriptomes despite improvements in the number and quality of available genomic resources. Mosquitoes nonetheless retain a typical insect body plan with three thoracic and ten abdominal segments. Recent work shows that gooseberry, a paralog of paired conserved in mosquito genomes, has expanded its expression into paired’s typical spatiotemporal domain, and analysis of a gooseberry loss-of-function mutant generated by CRISPR-Cas9 shows that gooseberry has taken on paired’s functionality. This reversal of subfunctionalization, coined synfunctionalization, may provide stability but could also introduce incompatibilities that drive species divergence; in Drosophila, gooseberry can rescue segmentation defects in paired null embryos but cannot rescue male fertility defects. Similar losses may remain undetected in poorly sampled taxa, as illustrated by the discovery that crickets in the genus Gryllus have lost short gastrulation/chordin, a gene used throughout Bilateria for dorsal-ventral patterning, a finding made possible only recently because orthopteran genomes are large, rich in transposable elements, and high in GC content.

Horizontal gene transfer adds another layer. Aphids synthesize their own carotenoids, molecules almost every other metazoan must obtain from food, and their carotenoid biosynthesis genes are most closely related to fungal orthologs, implying transfer from a fungus to an ancestral aphid lineage. Different alleles produce red and green morphs with different susceptibility to predators and parasitoids, so the acquired genes have shaped multiple aspects of aphid ecology. By one estimate, 40 percent of arthropods are infected with Wolbachia bacteria, which have repeatedly hijacked sex determination systems to promote female development; in the isopod Armadillidium vulgare, transfer of Wolbachia genes to a previously autosomal chromosome generated an entirely new sex chromosome. Lineage-specific or emergent genes, which lack clear orthology outside a narrow taxon, may make up 10 to 20 percent of a genome. Data from the i5k project found 147 emergent gene families in the last common ancestor of all insects, 227 in Hymenoptera, 205 in Coleoptera, 156 in Diptera, and a staggering 1038 in Lepidoptera. A novel gene called panish, found only in chironomid midges, rapidly developed an essential role in anterior-posterior patterning: knockdown embryos develop a second abdomen in place of a head. And in water striders of the genus Rhagovelia, two closely related emergent genes control development of a fan-like leg structure that lets these insects inhabit fast-moving streams inaccessible to most semiaquatic insects; knocking down the pair by RNAi reduced fan growth and left the insects unable to row effectively in moving water.

The review’s conclusion is a sober counterweight to the excitement of the genomic era. Supergenes, sets of linked genes controlling complex polymorphisms, illustrate both the promise and the limits: long-read sequencing finally allowed researchers to dissect the genetic architecture of butterfly wing mimicry, fire ant social polymorphism, and stick insect pigmentation, yet functional validation still demands painstaking lab or fieldwork. Genome annotation remains a non-trivial hurdle, database continuity is fragile, and even FlyBase struggles to secure funding. The bottleneck to understanding the relative importance of different genomic variants, the authors write, will for the foreseeable future remain laboratory work, limited by investigators’ ability to design experiments that test hypotheses generated from genomic data. Correlating variants with phenotypes has become increasingly trivial; proving causation has not.

Subject of Research: How advances in genome sequencing reveal the role of gene gain, loss, duplication and horizontal transfer in arthropod macroevolution.

Article Title: Advances in genome sequencing reveal changes in gene content that contribute to arthropod macroevolution

Article References: Cheatle Jarvela, A. M., & Wexler, J. R. (2023). Advances in genome sequencing reveal changes in gene content that contribute to arthropod macroevolution. Development Genes and Evolution, 233(2), 59-76. https://doi.org/10.1007/s00427-023-00712-y

Image Credits: AI Generated

DOI: 10.1007/s00427-023-00712-y

Keywords: genome sequencing, arthropods, gene duplication, gene loss, horizontal gene transfer, whole genome duplication, evo-devo, supergenes, lineage-specific genes, CRISPR, long-read sequencing, macroevolution