Pre-Metazoan Origin of Animal Mirnas
Total Page:16
File Type:pdf, Size:1020Kb
bioRxiv preprint doi: https://doi.org/10.1101/076190; this version posted October 1, 2016. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Bråte et al., 20 09 2016 – preprint copy - BioRxiv Pre-metazoan origin of animal miRNAs Jon Bråtea,1, Ralf S. Neumanna,b,1, Bastian Frommc, Arthur A. B. Haraldsena, Paul E. Grinia, Kamran Shalchian-Tabrizia,2 a Centre for Epigenetics, Development and Evolution (CEDE) and Centre for Integrative Microbial Evolution (CIME), Section for Genetics and Evolutionary Biology (EVOGENE), University of Oslo, Norway. b Institute of Clinical Medicine, Department of Immunology, University of Oslo, Norway. c Department of Tumor Biology, Institute for Cancer Research, Norwegian Radium Hospital, Oslo University Hospital, Norway. 1 Contributed equally 2 Corresponding author: [email protected] Abstract microRNAs (miRNAs) are integrated parts of the developmental toolkit in animals. The evolutionary history and origins of animal miRNAs is however unclear, and it is not known when they evolved and on how many occasions. We have therefore investigated the presence of miRNAs and the necessary miRNA biogenesis machinery in a large group of unicellular relatives of animals, Ichthyosporea. By small RNA sequencing we find evidence for at least four genes in the genus Sphaeroforma that satisfy the criteria for the annotation of animal miRNA genes. Three of these miRNAs are conserved and expressed across sphaeroformid species. Furthermore, we identify homologues of the animal miRNA biogenesis genes across a wide range of ichthyosporeans, including Drosha and Pasha which make up the animal specific Microprocessor complex. Taken together we report the first evidence for bona fide miRNA genes and the presence of the miRNA-processing pathway in unicellular Choanozoa, implying that the origin of animal miRNAs and the Microprocessor complex predates multicellular animals. Keywords: miRNA, animal evolution, Sphaeroforma, Ichthyosporea, Holozoa Introduction induced silencing complex (RISC), while the other strand (the microRNAs (miRNAs) are short (20-26 nucleotides) RNA molecules passenger, or star strand) is degraded. Gene regulation takes place via that are involved in gene regulation (Fromm, 2016). They are found in a the RISC and the central component, the Argonaute (AGO) protein that wide variety of organisms including plants, viruses and animals but binds the mature miRNA. Target messenger RNAs (mRNAs) are current evidence does not support a common evolutionary history identified by basepairing between the loaded miRNAs and the mRNA, (Tarver et al., 2012; Cerutti and Casas-Mollano, 2006; Shabalina and and translation of the mRNA is either fully blocked or down regulated. Koonin, 2016). In animals, miRNAs are essential components of the In animals, miRNA genes have been found even in the early gene regulatory machinery and highly important for processes like cell diverging animal lineages such as sponges (Grimson et al., 2008; differentiation and development (Mello and Conte, 2004). Furthermore, Robinson et al., 2013; Liew et al., 2016; Wheeler et al., 2009). In fact, because of the increase of distinct miRNA families during animal only two animal lineages are known to be missing miRNAs, the evolution (combined with lineage specific expansions and rare losses), ctenophores and Placozoa (Maxwell et al. 2012; Grimson et al. 2008; miRNAs have been linked to the evolution of morphological Moroz et al. 2014). However, due to the lack of homologous miRNA complexity in animals (Berezikov, 2011; Wheeler et al., 2009; Kosik, sequences in the basal branching animals, it has been hypothesized that 2010; Fromm et al., 2013; Erwin et al., 2011; Philippe et al., 2011). miRNA genes in animals arose several times independently (Tarver et Mainly based on the absence of miRNAs in single-celled ancestors, it al., 2012). Although, except from one study sequencing the small RNA has even been proposed that miRNAs were one of the key factors that content of the choanoflagellate Monosiga brevicollis (Grimson et al. allowed multicellular animals to evolve (Grimson et al., 2008; Wheeler 2008), almost nothing is known about the presence of miRNA genes, or et al., 2009; Mattick, 2004). the biogenesis machinery, in the unicellular relatives of animals. Generally, the primary transcripts of miRNA genes (pri- Here, we therefore investigate whether miRNA genes miRNA) can be very long RNA molecules that contain hairpin coevolved with animal multicellularity or if these genes were present structures (Axtell et al. 2011; Ameres & Zamore 2013; Chang et al. already in the unicellular ancestors of animals, the choanozoans. We 2015). In animals, the so-called Microprocessor complex (Han et al., address this question by deep sequencing of smallRNAs from species 2006), comprising the RNase III enzyme Drosha and the double- belonging to Ichthyosporea; a basal group at the split between animals stranded RNA (dsRNA) -binding protein Pasha (known as DGCR8 in and fungi, and search for the animal miRNA biogenesis machinery in humans), subsequently excises the hairpin structures (creating the pre- all unicellular relatives of animals. Our results show the existence of six miRNA). In contrast, plants do not possess Drosha or Pasha proteins; miRNA genes in Sphaeroforma that fulfill all criteria for the annotation here, several copies of the nuclear RNase III protein Dicer fulfills the of metazoan miRNAs (Fromm et al., 2015). Three of the miRNAs are role of the Microprocessor. In both animals and plants, the loop region highly conserved and expressed across three species, which strongly of the hairpin pre-miRNA is subsequently removed by Dicer proteins, indicate that the genes have important and similar function. In addition, albeit dicing takes place in the cytoplasm and nucleus, respectively, we show that the genomes of several species of Ichthyosporea, covering resulting in a 20-26 nt dsRNA molecule. Typically, only one of these the genera Sphaeroforma, Creolimax, Pirum, Abeoforma and strands (known as the mature strand) is then loaded into the RNA- 1 bioRxiv preprint doi: https://doi.org/10.1101/076190; this version posted October 1, 2016. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Bråte et al., 20 09 2016 – preprint copy - BioRxiv Amoebidium contain homologs of the animal Pasha and Drosha genes additional miRNAs were found in S. sirkka. Because there is no genome as well as Dicer, Exportin 5 and AGO. available for S. napiecek we could not run the miRNA prediction tools Altogether we present an evolutionary scenario where the as for the other species. But we could identify five of the S. arctica pre- common ancestor of Choanozoa and animals (Holozoan ancestor) miRNAs among the expressed mRNAs (even though the pri-miRNA is already possessed the animal miRNA biogenesis pathway, including transcribed by RNA polymerase II, and hence is poly-adenylated, it Drosha and Pasha, as well as bona fide miRNAs. This implies that both might not be present among the mRNAs because of rapid processing by the animal miRNAs and the Microprocessor pre-date the origin of Drosha and Dicer). These pre-miRNAs also had small RNA reads animals and that the last common ancestor of animals likely used mapping to the same locations as the mature and star strands in S. miRNAs for gene regulation. arctica, except for the star region of Sar-Mir-Nov-3 and 5. The pre- miRNAs were highly conserved between the three species with most Results mutations in the loop region which does not affect the gene targeting. But there were also a few minor differences in the mature and star Sphaeroforma species have bona fide miRNA genes strands, most notably in Sar-Mir-Nov-2. None of the novel miRNAs had any detectable homologs outside of Ichthyosporea. Initially, our small RNA sequencing and bioinformatics pipeline The miRNA genes are either located in intergenic regions or identified six miRNA genes in Sphaeroforma arctica (Figure 1 and in the introns and UTRs of protein-coding genes (Figure S1). In S. Supplementary File 1). Five of the miRNAs fulfilled all of the recently arctica Sar-Mir-Nov-1, 3, 5 and 6 are encoded in intergenic regions and proposed criteria for the annotation of miRNA genes (Fromm et al., transcribed as > 1kb unspliced pri-miRNA fragments with a well- 2015); at least two 20-26 nt RNA strands are expressed from each of the defined TSS, except for Sar-Mir-Nov-3 which is lacking a TSS but this two arms that are derived from a hairpin precursor with 2-nt offsets is probably an artifact of the genome assembly. Interestingly, the region (note that the 1nt Drosha-offset of Sar-Mir-Nov-3 and the 1nt Dicer- covering the Sar-Mir-Nov-5 pri-miRNA is transcribed also from the offset of Sar-Mir-Nov-5 are both likely due to the low number of reads opposite strand which could represent a potential novel long non-coding in general and correspondingly lower read-counts of the star sequence). RNA with binding capacity to the miRNA (an miRNA sponge) as no Further, the reads on both arms show no or very little variation of their protein-coding gene is annotated in the region. Sar-Mir-Nov-2 and Sar- 5’ start positions (5’ homogeneity) and we observe at least 16 Mir-Nov-4 are encoded within protein coding genes, in the intron and complementary nucleotides between the two arms. We include here Sar- the 3’-UTR respectively (Figure S1). These miRNAs are not transcribed Mir-Nov-6, although the star strand is not expressed and thus Dicer or from a separate TSS, but instead co-transcribed with the protein-coding Drosha offsets cannot be assessed because it shows high expression, a gene and subsequently processed into the mature miRNAs.