Menu

Close

Ancient genomic regulatory blocks are a major source for gene deserts in vertebrates after whole genome duplications

María Touceda-Suárez, Elizabeth M. Kita, Rafael D. Acemel, Panos N. Firbas, Marta S. Magri, Silvia Naranjo, Juan J. Tena, Jose Luis Gómez-Skarmeta, Ignacio Maeso, Manuel Irimia

Preprint posted on September 25, 2019 https://www.biorxiv.org/content/10.1101/776369v1

On the origin of gene deserts: large intergenic regions in vertebrates likely originated from exon erosion.

Selected by Jesus Victorino

Categories: evolutionary biology

Background & Summary

Whole Genome Duplication (WGD) events, which might have resulted from erroneous chromosome segregation during meiotic or zygotic cell division, have occurred in many lineages such as yeasts or plants [1]. Having all genes duplicated at once is likely a source of rapid evolutionary change: from the incredible amount of raw material a set of new features could be acquired through specialization and subfunctionalization of the duplicated paralogs. Such potential source of variation was “inserted” into the vertebrate genome after the two rounds of WGD that took place more than 450 million years ago [2].

However, the extra copies of each gene (ohnologs) produced after a WGD event often disappear due to the lack of providing a selective advantage to the organism, making it hard to trace evolutionary paths back, and explaining why there are not four paralogs of each gene in humans. Microsyntenic associations are formed between a developmental transcription factor (trans-dev, as they are referred to in the preprint) and another gene (bystander) when the regulatory elements of the trans-dev gene are found in the introns of bystander genes; such microsyntenic associations are known as genomic regulatory blocks (GRBs). In this work, Touceda-Suárez et al. investigated how GRBs have evolved in vertebrates in the context of trans-dev genes.

The authors studied the dismantling of the GRB microsyntenic associations after the two rounds of WGD and found that the non-trans-dev gene disappeared while the intronic regulatory elements were kept. This “exon erosion” likely resulted in large intergenic regions full of non-coding regulators, and the authors postulate that this is an important mechanism of gene desert formation (Figure 1).

 

Figure 1.- Schematic representation of the formation of a gene desert. An ancient Genomic Regulatory Block (GRB) present in a common ancestor is kept in microsynteny after Whole Genome Duplication (WGD) at one of the loci, while at the other loci the bystander gene is lost in the duplicated region, giving rise to a gene desert in the vicinities of the trans-dev.

 

Key findings

– Most microsyntenic associations (GRB pairs) common to chordates are present in humans. 131 out of the 156 syntenic gene pairs studied (84%) in 116 putative GRBs were present in the human genome in at least one copy.

– The majority of these GRB pairs are conserved in single copy. Despite the two rounds of WGD, most GRBs analyzed (present in humans) only kept the microsyntenic association between the trans-dev and the bystander gene in one place in the genome, even when there were more ohnologs maintained.

– The dismantling of the extra copies of GRB pairs involve preferential loss of the non-trans-dev gene. In most cases, only the trans-dev gene has been kept in more than one copy (70.1%).

– Exon erosion of the non-trans-dev gene is likely the mechanism of gene loss. In most of the GRBs analyzed, the absence of synteny in more than one trans-dev ohnolog could have been produced by genomic rearrangement or by the loss of one of the genes. However, the longer the intergenic region, the higher the number of putative regulatory elemements (as identified by ATAC-seq peaks). The number of putative regulatory elements increases to reach as many as there would be from the trans-dev gene to the next one after the bystander, which would have not been the case if the synteny had been lost by a genomic rearrangement event. For instance, the ohnologs isl2a and isl1 in zebrafish both maintain a highly conserved non-coding region located within an intron of the scaper gene in the first case, and within the gene desert originated in the second scenario.

– Large intergenic regions (gene deserts) arise upon the loss of large non-trans-dev genes. Trans-dev genes were enriched for large intergenic regions and, after dismantling of the bystander gene, the size of these intergenic regions would increase. The size of these larger intergenic regions reached approximately the distance between the trans-dev genes and the following gene after the bystanders in those cases where the bystander gene had been kept.

 

Why I liked this preprint

The regulation of gene expression in metazoans is coordinated to a large extent by distal enhancers, which are highly prevalent in vertebrates. Genes encoding developmental transcription factors usually lie near large intergenic regions in humans. These regions are commonly known as “gene deserts” and are an important source of regulatory elements. However, the evolutionary paths that these genomic blocks might have followed all the way through the vertebrate lineage until they became large intergenic regions remained largely unknown.

This research led by M. Irimia, J.L. Gómez-Skarmeta and I. Maeso shows a very likely mechanism by which these enigmatic genomic regions might have appeared in evolved vertebrates. A beautiful answer to an interesting question, which is the main reason why I liked this preprint.

 

Questions to the authors

– What has been the criteria used by the authors to define a gene as a transcription factor involved in development?

– What do the authors think could be the main reason behind the preferential loss of the bystander gene? Is it known what’s the proportion of trans-dev (retained in multiple copies) that acquired new functions or specialization?

– The authors show examples of trans-dev and bystander pairs in synteny, while contrasting it with non-trans-dev genes likely sharing bidirectional promoters. Did the authors explore the frequency of trans-dev and bystander genes sharing a bidirectional promoter? Would this GRB pairs be expected to undergo the same exon erosion phenomena?

– Do the authors think that having large intergenic regions -sort of specialized in trans-dev gene regulation- confer any evolutionary advantage over having regulatory elements at the intronic regions of a bystander?

 

References

1 https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.0030314

2 https://www.nature.com/articles/ng.3526#f1

 

Tags: bystander, evolutionary biology, gene deserts, intergenic regions, trans-dev, vertebrates

Posted on: 14th October 2019 , updated on: 5th November 2019

doi: https://doi.org/10.1242/prelights.14600

Read preprint (No Ratings Yet)




  • Author's response

    Manuel Irimia shared

    – What have been the criteria used by the authors to define a gene as a transcription factor involved in development?

     

    We defined trans-dev genes as “transcription factors involved in the regulation of developmental processes”, using Gene Ontology terms. An important consideration is that we decided to focus on transcription factors only because they are easier to define in terms of their constituent protein domains (i.e. sequence-specific DNA binding domains) than other developmental genes, reducing the fraction of false positives. This of course has the problem that many developmental genes (e.g. morphogens or alternative splicing regulators) that also have complex distal cis-regulation are left outside of the definition, but we preferred to be restrictive and have a high-confident set.

     

    – What do the authors think it could be the main reason behind the preferential loss of the bystander gene? Is it known the proportion of trans-dev (retained in multiple copies) that acquired new functions or specialization?

     

    This a very interesting question, and many researchers have investigated why certain genes are retained or lost more frequently than others after WGDs, so many different explanations and hypotheses have been put forward.

    With respect to the first question, we could say that, strictly speaking, it is not so much that bystanders are lost preferentially, but that trans-dev genes are retained more often after WGDs. One explanation for this maybe that their larger regulatory landscapes make their duplicates more likely to accumulate mutations in their regulatory regions in a reciprocal manner, so the chances that two or more duplicates become non-redundant and essential are higher than in bystander genes (the classic duplication-degeneration-complementation, DDC, model). Another, non-mutually exclusive reason could be that the acquisition of novel or specialised functions is particularly advantageous in the case of developmental regulators, since they could contribute to the evolution and co-option of gene regulatory networks, providing the raw material for the appearance of novel cell types and morphological novelties. This connects with the second question, for which we do not have a definite answer. While there are multiple cases of trans-dev ohnologs with clearly non-identical functions (e.g. Pax4/6, Srrm2/3/4 or Shh/Ihh/Dhh), the exact fraction is hard to define.

     

    – The authors show examples of trans-dev and bystander pairs in synteny while contrasting with non-trans-dev genes likely sharing bidirectional promoters. Did the authors explore the frequency of trans-dev and bystander genes sharing a bidirectional promoter? Would this GRB pairs be expected to undergo the same exon erosion phenomena?

     

    We did not specifically search for examples of this particular scenario, but this an interesting point that emphasizes some of the differences between trans-dev and bystander genes. The expression patterns of trans-dev genes are vastly different from those of their associated bystander genes and this is because only the core promoters of the trans-dev genes are able to respond to the multiple long-range regulatory elements present in the GRB. Thus, trans-dev and bystander genes normally have different core promoters types. This would in principle rule out the possibility of trans-dev and bystander genes sharing a common bidirectional promoter, since this would probably lead to a highly coordinated expression of both genes (and would probably be incompatible with avoiding interactions between distal cis-regulatory elements and the bystander gene).

     

    – Do the authors think that having large intergenic regions -sort of specialized in trans-dev gene regulation- confer any evolutionary advantage over having regulatory elements at the intronic regions of a bystander?

     

    We have frequently asked ourselves this question, which is probably not easy to test. If we take into account that, as mentioned previously, trans-dev and bystander genes are expressed very differently, we could speculate different reasons why having enhancers within a bystander gene could be disadvantageous. For instance, as a consequence of the activation or silencing of the bystander expression, regulatory elements located in the introns of the bystander could be placed in local chromatin environments that are very different from the rest of the regulatory landscape of the trans-dev gene. Also, there could be interferences between the transcriptional machineries of the two genes. Finally, since most enhancers are also transcribed, this could affect bystander expression and splicing, especially in the case of antisense transcription.

    All these are interesting possibilities that would be worth pursuing in future studies.

    Have your say

    Your email address will not be published. Required fields are marked *

    This site uses Akismet to reduce spam. Learn how your comment data is processed.

    Sign up to customise the site to your preferences and to receive alerts

    Register here
    Close