-
[
International Worm Meeting,
2015]
The evolution of embryogenesis is biased since many variations affect vital processes and lead to nonviable organisms. Species comparisons have enabled to evolutionary developmental biologists to characterize such constraints. For the animal embryo, it has been proposed that a stage in mid-development is conserved throughout the kingdom. As part of a consortium of ten labs, we test this hypothesis by identifying the developmental transcriptomes of ten species, each of a different animal phylum, including Nematodes, Ctenophores, Annelids, Chordates, Platyhelminthes, Tardigrades, Arthropods, Echinoderms, Cnidarians, and Poriferas. We find that animal embryonic development comprises the coupling of two conserved gene expression modules, primarily involved in proliferation and differentiation, where the switch between them occurs at the apparent phylotypic period of each phylum. Genes whose expression is coherent across the examined species are expressed in the first module, while clade specific variations occur in the second module and the transition to it. Our results are consistent with an inverse hour-glass model of animal evolution: gene expression differences across phyla are concentrated at the inter-module switch. Since, expression differences within a phylum appear to be minimal at the phylotypic stage, our work leads us to define a phylum as a collection of species where an hourglass model holds for gene expression differences. All together, we provide a map of embryology upon which ancestral constraints are mapped and distinguished from the stages in which variations are available to the agency of natural selection. [This work is a collaboration among the Arendt (Heidelberg), Rink (Dresden), Blaxter (Edinburgh), Goldstein (North Carolina), Salzberg (Haifa), deLeon (Haifa), Yaniv (Rehovot), Gat (Jerusalem), Degnan (Queensland), Martindale (Florida), and Yanai (Haifa) labs.].
-
[
C. elegans: Development and Gene Expression, EMBL, Heidelberg, Germany,
2010]
Metazoan development is a highly conserved process governed by precise temporal and spatial gene regulation. However, our previous work has provided evidence that a pair of organisms with a very similar mode of development, yet divergent genomes, show evolved transcriptomes. We therefore hypothesized that a detailed comparison of gene expression throughout embryonic development on a phylogeny of nematodes - including 5 Caenorhabditis species - will reveal the functional component of gene expression. Using custom-designed species-specific microarrays containing probes for both a stable set of orthologs as well as quickly evolving paralogs, we examined gene expression levels in precisely staged cohorts of embryos across ten developmental stages. Our preliminary analyses indic ate that gene expression evolution can be understood through the principles of differential constraints across functional categories and developmental stages. Further, the consensus of orthologous gene expression profiles appear to be more correlated with the timings of known functions in C. elegans than the actual C. elegans gene expression profile, suggesting that gene expression profiles across species can be analyzed as gene sequences and queried for the signature of selection through observed patterns of conservations and divergences. We propose that genetic drift of the transcriptome is an important developmental principle with implications to other principles such as robustness and modularity.
-
[
Development & Evolution Meeting,
2008]
Biological systems show remarkably diverse transcriptional responses to new environments, yet it is unclear which component of the response is adaptive and which correlates with the general responsiveness of the genes. To study this problem we examined as a model phenotype the maternal mRNA dowry provided to the C. elegans embryo. We reasoned that a gene's responsiveness would also be manifested across genotypes and that its characterization would enable us to distinguish the adaptive environmental response. Using microarrays, we measured transcript levels of the maternal dowry across five different geographical C. elegans isolates, whose mothers had been raised in five different environments. We found a comparable percentage of genes showing significant variation across strains (9%) as with variation across environments (8%). Moreover, 3% of the genes show variation across both strains and environments -- 4-fold more than expected by chance -- signifying a strong correlation between these two causes of transcriptional variation. Integrating data from global RNAi screens, we asked if the set of genes with variation across strains and/or environment also show differential frequencies of genes essential for development. We found that genes with variation across both strains and environments are 2-fold less likely to be essential than randomly selected genes. In contrast, genes showing an environmental response that is conserved across strains are 40% more likely to be essential. This pattern of variation suggests that a transcriptional response can be parsed according to non-adaptive and adaptive modes by referring to the presence or absence of regulatory variation across genotypes. Experiments are currently underway to test for increased sensitivity of such genes under the responsive-inducing environment. Our results demonstrate that the correlation between transcriptional variation across strains and environments can be used to understand gene expression programs.
-
[
International Worm Meeting,
2005]
We are interested in studying the genetic basis for the remarkably robust network of events that specify the development of a multicellular organism. Genetic analysis has identified a number of master regulators that specify founder cell fate; however, little is known about the constraints that act upon the participating genes throughout development. We hypothesized that a comparison among wild C. elegans isolates with extant Caenorhabditis species would reveal the evolutionary plasticity of the process in molecular detail. Our approach involves comparing, on a gene by gene basis, the effect of two evolutionary time frames: micro-evolution through a comparison of wild isolates, and macro-evolution through a comparison with divergent Caenorhabditis species. To begin we are exploring the magnitude and pattern of transcriptional differences among C. elegans strains to generate a framework for the characterization of the mode of evolution across developmental stages.
-
[
International Worm Meeting,
2007]
Changes in gene expression account for a large fraction of morphological evolution yet the background rate of gene expression divergence in the absence of morphological evolution has not been formally examined. Using species-specific custom whole-genome microarrays, we have produced a gene expression timecourse data that captures dynamic gene expression changes during early embryonic development for C. elegans and C. briggsae. These two species are morphologically near identical yet they last shared a common ancestor ~100 million years ago. We found that the level of transcript abundances across orthologs is well conserved (R=0.74). Strikingly, a strong positive correlation was detected between transcript abundance and conservation of temporal expression profiles, indicating that highly expressed genes are more constrained by natural selection to maintain their expression profile. Divergent transcript abundance was correlated with divergence in the expression profile; however, the majority of expression profile divergences do not result in changes to transcript abundances. What mechanisms might underlie changes to a genes temporal expression profile? We found that chromosomal rearrangements promote expression divergence: genes with different upstream gene between the two species have an increased divergence in expression. Since chromosomal gene neighbors often have similar expression profiles, this finding suggests that expression similarity of gene neighbors may be attributed to leaky-expression of neighboring genes rather than to functional relationships. In conclusion, while a genes pattern of expression is generally assumed to correlate to its function, we find that transcript abundance and gene location influence the level of expression divergence. These results indicate that gene expression comparisons over multiple species and developmental stages may identify the component of gene expression under selection.
-
[
International Worm Meeting,
2009]
Even before specific transcription factors can begin to modulate gene expression, the initiation of transcription is crucially dependent upon the properties of core promoter sequence. While many short motifs such as the initiator motif and the TATA-box are known to occupy this region, a surprising level of variation is present across a genome''s core promoters. We hypothesized the existence of a code of short motifs distinguishing the functional properties of core promoters. Here we characterize 12 regulatory elements located in C. elegans core promoters. We found that these cluster into two mutually exclusive groups of motifs (Core1 and Core2). For example the initiator motif and T-repeats (with a consensus of TTTTTCCTCCATTTTT) from the Core1 group do not co-occur with the TATA-box and purine-repeats ((R)12), of the second group. Invoking gene expression data across embryonic development, we asked whether different core promoter motifs correlate with different expression levels. We found that genes with elements from the first group show significantly elevated expression level. For example genes with the initiator motif exhibit strong expression levels (z-score = 35.7), while genes with motifs from the second group such as the TATA-box show low expression (z-score = -1.7). Our analysis enables us to examine the evolution of motif architecture in the core promoter across different species. Such an analysis of the dynamics of motif composition may lead to an understanding of the tuning of gene expression level by motif combinatorics of the core promoter.
-
[
International Worm Meeting,
2011]
The realization that all nucleated cells in a multicellular organism contain the same genomic information, despite carrying out very specialized functions, is one of the hallmarks of cell biology. In the context of development, it is well-known that cis-regulatory elements (CREs) are crucial in recruiting transcription factors that ensure the precise spatio-temporal regulation of target genes. The DNA sequence of any higher eukaryote can therefore be viewed as a solution to the daunting challenge of arranging thousands of CREs with widely divergent roles in a linear and functionally coherent fashion. Studying the genome from this perspective can serve as a foundation for a broader understanding of the principles of gene regulation in development. To this end, we have integrated the location and magnitude of tens of thousands of genomic binding events from more than 40 C. elegans transcription factors (TF), acting at various stages of development, based on publicly available datasets from the modENCODE consortium. Our approach is unique in that it is fully quantitative and avoids the high false discovery rate associated with many previous studies. For the first time, this allows us to map out the cis-regulatory information content at any given location in the genome. We use this powerful analysis to estimate the extent to which CREs are generally organized into modules that integrate multiple regulatory inputs to produce cell-specific expression. Surprisingly, many promoters exhibit proximal regions with degenerate binding events. We provide evidence that these bindings do not mediate specific regulation and show that clustering genes based only on their specific CREs gives a much better correlation with their corresponding expression profiles, based on transcriptomic data from our lab and others. Our integrative approach provides a framework for understanding how individual sequence elements contribute to developmental expression programs. The analytical power of our approach will further increase with the integration of additional datasets.
-
[
International Worm Meeting,
2011]
Core promoters mediate transcription initiation by the integration of diverse regulatory signals encoded in the proximal promoter and enhancers. It has been suggested that genes under simple regulation may have low-complexity permissive promoters. For these genes, the core promoter may serve as the principal regulatory element; however the mechanism by which this occurs is unclear. We report here a periodic poly-thymine motif, we term T-blocks, enriched in occurrences within core promoter forward strands in C. elegans. An increasing number of Tblocks on either strand is associated with increasing nucleosome eviction. Strikingly, only forward strand T-blocks are correlated with expression levels, whereby genes with ³6 T-blocks have five fold higher expression levels than genes with 3 T-blocks. We further demonstrate that differences in T-block numbers between strains predictably affect expression levels of orthologs. Highly expressed genes and genes in operons tend to have a large number of T-blocks, as well as the previously characterized SL1 motif involved in trans-splicing. The presence of T-blocks thus correlates with low nucleosome occupancy and the precision of a trans-splicing motif, suggesting its role at both the DNA and RNA levels. Collectively our results suggest that core promoters may tune gene expression levels through the occurrences of T-blocks, independently of the spatio-temporal regulation mediated by the proximal promoter.
-
[
International Worm Meeting,
2011]
Embryonic development is broadly characterized by the presence of apparently distinct stages, however their actual distinctiveness and significance remain poorly understood. To address this question, we examined the expression levels of all genes throughout the embryology of five Caenorhabditis species. These nematodes are phylogenetically distant and undergo development at different rates, allowing us to probe which developmental stages exhibit transcriptomes independent of these differences. We report evidence that development is punctuated with non-gradual transcriptomic changes. In particular, the difference in time it takes to arrive at the epithelial morphogenesis (ventral enclosure) stage across the species is independent of the transcriptomic difference among them. Thus we provide evidence for the existence of a developmental milestone that is independent of time. The ventral enclosure stage is enriched with the expression onset of Homeodomain and Zn-finger transcription factors and genes involved in muscle and neuronal functions, suggesting that this specific stage is essential for the progression of embryonic development by integrating diverse signals to orchestrate morphogenesis. In addition, genes whose onset of expression occurs in this stage are enriched with RNAi derived phenotypes such as slow development and late developmental arrest, further supporting the coherence of this stage. Our work provides a molecular definition of a developmental stage in terms of its time independent, non-gradualistic and integrative nature.
-
[
International Worm Meeting,
2011]
Genetic analyses have identified key developmental regulators, yet even well characterized networks may not be complete due to functional redundancies which are not easily identifiable. For a more unbiased and systematic approach, embryonic gene networks have been studied using whole transcriptome methods, however these have lacked the spatial and temporal resolution required for addressing the intricate events occurring in the early embryo. Here we determine the whole-transcriptomes of precisely staged single blastomeres spanning the entire development of a single cell lineage, the E-lineage, encoding the endoderm in C. elegans. We hypothesized that such data would allow us to recover the underlying network components, both known and uncharacterized. Accordingly, we find that early uncommitted blastomeres are extremely similar to one another, in comparison with blastomeres in which cell fate is being specified, suggesting that the transcriptional variation found is related to the specification process. We find that the transcriptome is induced significantly earlier and sharper in the E lineage relative to its sister lineage MS, consistent with E's earlier lineage specification. The known endoderm GATA regulators are highly expressed and E-lineage specific, providing evidence that the lineage gene network is captured by our blastomere transcriptome data. Surprisingly, additional transcription factors, including a dozen members of the nuclear hormone receptor family, are also E-lineage specific, hinting at a more complex network than previously known. In order to understand the functional relevance of the uncharacterized factors, we are repeating our analysis in the C. japonica E-lineage, which may reveal which factors are conserved and consequently play a role in the network. Our results show that given precisely staged single cell transcriptome data, the gene regulatory networks may be determined.