Evolution of maternal and zygotic mRNA complements in the early Drosophila embryo

Abstract

The earliest stages of animal development are controlled by maternally deposited mRNA transcripts and proteins. Once the zygote is able to transcribe its own genome, maternal transcripts are degraded, in a tightly regulated process known as the maternal to zygotic transition (MZT). While this process has been well-studied within model species, we have little knowledge of how the pools of maternal and zygotic transcripts evolve. To characterize the evolutionary dynamics and functional constraints on early embryonic expression, we created a transcriptomic dataset for 14 Drosophila species spanning over 50 million years of evolution, at developmental stages before and after the MZT, and compared our results with a previously published Aedes aegypti developmental time course. We found deep conservation over 250 million years of a core set of genes transcribed only by the zygote. This select group is highly enriched in transcription factors that play critical roles in early development. However, we also identify a surprisingly high level of change in the transcripts represented at both stages over the phylogeny. While mRNA levels of genes with maternally deposited transcripts are more highly conserved than zygotic genes, those maternal transcripts that are completely degraded at the MZT vary dramatically between species. We also show that hundreds of genes have different isoform usage between the maternal and zygotic genomes. Our work suggests that maternal transcript deposition and early zygotic transcription are remarkably dynamic over evolutionary time, despite the widespread conservation of early developmental processes.

Fig 1. Reproducibility of the data.

(A,B) Spearman rank correlation coefficients are high when FPKM values of embryonic transcriptomes at the same stage are compared. (C) The transcriptome changes dramatically between stage 2 and stage 5.

Fig 2. Evolution of the early transcriptome.

Stage-specific transcriptomes are highly correlated across species. (A) Phylogeny of species in the study. The star represents the focal species, to which the species with filled circles are compared in parts B and C. (B,C) The D. melanogaster (star, 2A) stage 2 and stage 5 transcriptomes are more highly correlated with comparable transcriptomes (dots) from the closely related D. simulans (B) than the distantly related D. virilis (C). (D) When comparing D. melanogaster with the other species, Spearman correlation coefficients decrease over evolutionary time.

Fig 3. Evolutionary transcriptomic divergence across different classes of genes.

Separate correlation plots show interspecific pairwise Spearman correlation coefficients of transcript levels (FPKM) of all genes represented at stage 2 and stage 5 in both species (top row) and genes that are maternal-only (represented at stage 2 and with all transcripts degraded by stage 5) or zygotic-only (represented at stage 5 and not maternally deposited) in both species (bottom row). When comparing species that are closely related (e.g. members of the melanogaster subgroup or obscura group), Spearman coefficients are relatively high for all stage 2, all stage 5 and maternal-only genes, and drop sharply when more distant species are compared. This pattern is less obvious for genes that are zygotic-only in the two species being compared, which show relatively high correlations at large evolutionary distances.

Fig 4. Comparison to Aedes aegypti highlights patterns at longer evolutionary divergence times.

(A) Transcriptomes of each stage cluster together within Drosophila, while the equivalent stages of Aedes aegypti (250 MY diverged) cluster by species rather than by stage. Within Drosophila, clustering does not fully recapitulate the phylogeny. For both stage 2 and stage 5, the obscura group (orange boxes) form an outgroup to those from all other species of an equivalent stage. This is true when all transcripts are examined (this plot) or only autosomal transcripts are examined (S1C Fig). B) Transcript levels (FPKM) for all transcripts at each stage are relatively highly correlated between Drosophila (here D. melanogaster) and Aedes aegypti. C) Genes that have conserved zygotic-only representation in both Drosophila and Aedes are highly enriched in transcription factors, as well as known developmental functions. A small number of these genes have unknown functions in the early embryo.

Fig 5. Changes in gene representation across the phylogeny.

A) The number of gains and losses of gene representation at each stage are shown for all clades with at least three species. A threshold of FPKM = 1 was used throughout. There are more stage 5 changes (499 transcripts) than stage 2 changes (245), when gradual transitions are included. Only transitions between adjacent nodes are shown. Gains in representation are as common as losses, if not more abundant, across both stages. B) Examples of genes with changes in transcript representation at different stages, in the obscura group (first two panels), the melanogaster subgroup (third panel), and with highly variable patterns across stages and species (fourth panel).

Fig 6. Zygotic-only expression evolves rapidly in unannotated genes.

(A) Abundance of genes in different categories (maternal-only, predominantly stage 2, predominantly stage 5, zygotic-only) across species. In addition to the model species D. melanogaster, D. simulans is shown as it has the largest proportion of its transcripts represented at stage 2, whereas D. virilis was chosen as it has one of the largest proportions of zygotic transcripts (D. miranda was equally as zygotically biased). B) The vast majority of unannotated genes, most of which are taxonomically restricted, are zygotic-only, suggesting that zygotic expression evolves more rapidly than maternal deposition. N indicates the number of genes in each category (all genes, unannotated genes) for each species. See S10 Table for data on all species.

Fig 7. Hundreds of genes have stage-specific isoforms for both stage 2 and stage 5.

(A) Genes with separate isoforms for stage 2 and stage 5 (ALT), are found in all 14 Drosophila species. (B) Both the long and short isoforms of cnc are predominantly stage 2, indicated with the red letter M, while the intermediate length isoforms are predominantly stage 5, indicated with the blue letter Z. While total levels of the hdc transcript (C) are similar between sister species D. simulans and D. sechellia, the isoform usage evolves even over this short evolutionary timescale (D).