Testis-Derived Microrna Profiles of African Clawed Frogs (Xenopus) and Their Sterile Hybrids ⁎ Pawel Michalak , John H

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Available online at www.sciencedirect.com

Genomics 91 (2008) 158–164 www.elsevier.com/locate/ygeno

Testis-derived microRNA profiles of African clawed frogs (Xenopus) and their sterile hybrids ⁎ Pawel Michalak , John H. Malone 1

Department of Biology, University of Texas Arlington, Box 19498, Arlington, TX 76019, USA Received 5 October 2007; accepted 29 October 2007

Abstract

Gene regulation was long predicted to play a vital role in speciation and species divergence. Only recently with the advent of new technologies, however, has it been possible to address the question of the relative contributions of different mechanisms of gene expression to the evolution of phenotypic diversity. Here we broaden the question and ask whether microRNAs, a large class of small regulatory RNAs, play a role in reproductive isolation between species by contributing to hybrid male sterility. MicroRNAs from the testes of clawed frogs (Xenopus) were extracted and the expression profiles of sterile hybrids were compared with males of a parental species. Hybrid testes were largely microRNA- depleted relative to those of nonhybrids, and this pattern was validated with quantitative RT-PCR. A number of candidate differential microRNAs from this study have previously been described as testis-specific in the mouse, suggesting that microRNA structural conservation may be associated with functional retention. © 2007 Elsevier Inc. All rights reserved.

Keywords: Xenopus; MicroRNA; Reproductive isolation; Hybrid sterility

Comparative genomics provides new tools for helping to predictions related to the genetics of reproductive isolation. In understand how phenotypic differences between species are particular, the Dobzhansky–Muller model does not easily encoded in their genomes and what genetic processes underlie accommodate such widespread phenomena as maternal effects, the formation of new species (speciation). It has long been transposable element mobilization, genomic imprinting, and recognized that speciation often results in sterility and non- other epigenetic changes despite their importance to reproduc- viability of species hybrids but only recently has it been pos- tive isolation [4–7]. sible to identify genes causing incompatibilities on the hybrid The merging of two distinct genomes typically sets in motion genetic background and associated hybrid defects [1]. The extensive modifications of the genome and transcriptome, evolution of reproductive isolation between species has been creating cascades of novel expression patterns [8–10], regula- formalized as the so-called Dobzhansky–Muller model of tory interactions, and new phenotypic variation [11,12].To incompatible epistatic interactions between genes that have complete this picture, we examined the contribution of micro- diverged in different gene pools [2,3]. Despite its attractiveness, RNAs to hybrid transcriptome anomalies and male sterility. this broad model has obvious limitations for generating specific MicroRNAs (miRNAs) are a large class of small (∼22 nucleo- tides long) noncoding ribonucleic acids that function as fundamental transcriptional and posttranscriptional regulators in plants and animals [13]. ⁎ Corresponding author. Fax: +1 817 272 2855. There are several reasons to expect that miRNAs and other E-mail address: [email protected] (P. Michalak). small RNAs have a critical role to play in creating postzygotic 1 Present address: Laboratory of Cellular and Developmental Biology, National Institute of Diabetes and Digestive and Kidney Diseases, National reproductive isolation. First, it is becoming increasingly clear Institutes of Health, Department of Health and Human Services, Building 50, that there are barely any cellular or developmental functions Room 3341, Bethesda, MD 20892, USA. unaffected by miRNAs. For example, it has been estimated that

pathways are conserved between taxa as distant as Xenopus and the mouse. An analysis of the miRNA putative targets shows that a diversity of gene activities could be affected and that there was substantial cross-targeting between the top candidate miRNAs, at least at the level of the target Gene Ontology categories. Although X. laevis and X. muelleri belong to two relatively distant clades that diverged more than 30 million years ago, these two species still hybridize in nature and produce viable hybrids [23–27]. Notably, hybrid females are fertile and hybrid males are sterile, despite the fact that the females are hetero- gametic (ZW) and males are homogametic (ZZ), a pattern opposite to that predicted by Haldane’s rule [27]. These two species along with other Xenopus taxa provide an exciting system for investigations into the genetic background of reproductive isolation in general and exceptions from Haldane’s rule in particular. Fig. 1. Hierarchical clustering visualizing miRNA overrepresentation (red color) in the parental X. laevis relative to hybrids. Only differentially expressed (q value b0.2) miRNAs were included. Results

Differential expression from miRCURY locked nucleic acid (LNA) microRNA arrays known miRNAs make up N1% of the human gene content and regulate N30% of all protein-coding genes [14,15]. Second, There was a global miRNA underexpression in hybrids there are abundant testis-specific miRNAs [16,17], repeat- compared with parental X. laevis, as reflected by the extent of associated small interfering RNAs [18], and Piwi-interacting red color in Fig. 1. Both Cyber-T [30] and Edge [31] produced RNAs involved with spermato-and spermiogenesis [19–21], identical orders of the top candidate miRNAs differentially which all provide candidate causative agents of hybrid male expressed between the parental X. laevis and the F1 hybrids sterility. Finally, given that miRNAs along with other transcripts (Supplementary Table 1). Three miRNAs remained signifi- are abundant in maternal deposits [22], it is tempting to cantly different at the 0.05 α level after a Bonferroni correction speculate that species-specific miRNA deposits may account for (0.05/1458) was applied to the Cyber-T results: xtr-miR-202⁎, cases of reproductive isolation with maternal effects, when hsa-miR-494, and dre-miR-202⁎, all underexpressed in hybrids reciprocal crosses between the same two species produce dif- (Table 1). The Edge-generated q values for the top 3 and the ferent patterns of hybrid defects. next 5 best candidates (according to their p values) were all African clawed frogs (Xenopus) provide a novel system equal to 0.125. The fact that 2 of the best candidate miRNAs, for exploring the importance of miRNAs to the problem of xtr-miR-202⁎ from X. tropicalis and dre-miR-202⁎ from Danio reproductive isolation through hybrid male sterility. There are rerio, have identical sequences is interpreted as an indication of N20 Xenopus species, and experiments with over 100 reciprocal high array reproducibility. Of the top 10 candidate miRNAs, 6 crosses between selected pairs of species have revealed that were previously known in Xenopus, 3 were known exclusively hybrid males and rarely females are sterile [23–26]. Certain in mammals (miR-494, mmu-miR-489, and mmu-miR-155), reciprocal crosses are more affected or not possible, suggesting a and 1 is from miRPlus. These mammalian miRNA sequences contribution of maternal effects. Additionally, there is evidence did not produce significant hits when blasted against the that hybridization in Xenopus is accompanied by a transcriptome X. tropicalis genome (X. laevis and X. muelleri genomes are shock and preferential silencing of one of the parental ge- unavailable). nomes [27], thus suggesting the activity of ample transcriptional repressors. As spatial and temporal expression patterns of miRNAs can Table 1 provide clues about their possible functions, tissue specificity of Top five candidate miRNAs differentially expressed between hybrid and nonhybrid testes miRNA signaling is currently an area of active research. For example, miRNA profiling identified pancreas islet-specific miRNA Fold change Edge p value Cyber-T p value ⁎ − 6⁎⁎,a miRNA involved in insulin secretion [28] as well as brain- xtr-miR-202 2.491 0.0004 9.776 hsa-miR-494 2.436 0.0008 1.921− 5⁎⁎ specific miRNA regulating the size of dendritic spines— − dre-miR-202⁎ 2.392 0.0012 3.206 5⁎⁎,a postsynaptic sites of excitatory synaptic transmission [29]. Here xtr-miR-125b 2.179 0.0016 7.321− 5a we report significant changes to miRNA levels between testes miRPlus 17927 2.117 0.0020 0.0025 of sterile Xenopus hybrids (X. laevis×X. muelleri) and testes of Fold change is a mean ratio between X. laevis and hybrid expression values. one of the parental species (X. laevis). The detected repertoire of aConfirmed with quantitative RT-PCR. miRNAs additionally suggests that testis-specific miRNA ⁎⁎Significant at pb0.05 after Bonferroni correction. 160 P. Michalak, J.H. Malone / Genomics 91 (2008) 158–164

Quantitative real-time PCR analysis sented among the miR-202⁎ targets, including 6 various ace- tyltransferase activities. In miR-125b putative targets, 47 BPGO, As the false discovery rate estimated as q values was 11 CCGO, and 23 MFGO categories were overrepresented. A relatively high, differential expression of miR-202⁎, miR-494, similar analysis of miR-100 showed that 19 BPGO, 8 CCGO, miR-125b, and miR-100 was independently tested using larger and 15 MFGO target categories were overrepresented. There RNA sampling and quantitative RT-PCR. Amplification pat- was a substantial overlap (24%) between overrepresented BPGO terns of miR-202⁎, miR-125b, and miR-100 confirmed results categories of all three miRNA targets, including chromatin of the microarray analysis and showed consistent miRNA assembly, cellular protein metabolism, protein localization, and underexpression in hybrids relative to X. laevis (Table 2). Ex- metabolism. A 19% overlap between overrepresented CCGO pression of miR-494 was low compared to the other three (including intracellular organelle part) and no clear overlap miRNAs and no statistically significant differential expression between MFGO categories for these combined miRNA targets between X. laevis and hybrids was observed. Although were observed. reproducibility between the consecutive RT reactions was remarkably high, their statistical effect was significant in all Discussion four assays (Table 2). Interaction factors from the one-way repeated-measures ANOVA were not significant ( pN0.05). Here we report the identification of a number of miRNAs depleted in testes of sterile hybrids relative to fertile parentals Predicted targets of candidate miRNAs in Xenopus. Previously, a number of X. laevis miRNAs were isolated from various developmental stages ranging from MicroRNAs positively verified with qRT-PCR were ana- oocytes to tadpoles [32,33]. Although miRBAse (release 9.1) lyzed with respect to their putative targets with miRanda provides a total of 177 predicted miRNAs for X. tropicalis, (Table 3). A total of 23 Gene Ontology (GO) categories related no data on their spatial and temporal distribution have been to Biological Processes (BPGO) were overrepresented among available. putative targets of miR-202⁎ (Supplementary Table 2). These Sequences of a large proportion of miRNAs are highly categories included eight chemotaxis-, six protein processing-, conserved among taxa and many are expressed in a tissue- and two behavior-related processes. Only 2 Cell Component specific manner [34]. Although this does not necessarily imply GO (CCGO) categories were overrepresented in miR-202⁎ spatial and temporal conservation of their expression, such targets: small ribosomal subunit and cell surface. There were overlap in expression between various taxa has been demon- 12 Molecular Function GO (MFGO) categories overrepre- strated [35,36]. For example, miR-10 regulates Hox genes in both Drosophila and vertebrates [37,38]. Hence, one may ask whether the miRNAs that we observed to be differentially expressed between hybrid and nonhybrid Xenopus males match the list of testis-specific miRNAs isolated from the Table 2 mouse [16,17]. One of such testis-specific miRNAs, miR-122a, ANOVA of TaqMan RT-PCR results (L - X. laevis, H - hybrids) for example, is known to target a reporter mRNA containing sequences from the 3′-UTR of the transition protein 2 (Tnp2) L mean H mean ANOVA MS MS Fdf(1, 8) p value CT ratio CT ratio effects effect error mRNA [17]. Four species of transition proteins, including miR-202⁎ Tnp2, facilitate chromatin transformation from the nucleosome 0.5834 0.6032 Type (1) 0.0065 0.0006 11.3723 0.0100 structure to the nucleoprotamine structure during spermatid Replicates (2) 0.0014 0.0001 9.5812 0.0148 differentiation in mammals. All genes involved in mammalian − (1)×(2) 6.68 5 0.0001 0.4675 0.5134 spermiogenesis, such as Tnp2, are subject to translational re- interaction pression for up to 7 days [39], and small RNAs are logical miR-494 candidate silencers in this process. 1.1480 1.1420 Type (1) 0.0002 0.0015 0.1175 0.7406 Remarkably, one of our top three candidate miRNAs, miR- Replicates (2) 0.0012 0.0005 2.5307 0.1503 100, has been characterized as testis-specific in mice [17].Its − (1)×(2) 1.74 5 0.0005 0.0365 0.8531 targets and a potential role in spermatogenesis/spermiogenesis interaction are unknown, but the miRanda-predicted best candidate target miR-125b of miR-100 in X. tropicalis is cereblon (CRBN; Table 2), in 0.6844 0.7438 Type (1) 0.0176 0.0007 25.7364 0.0010 mammals known to be enriched in brain and testes [40]. Replicates (2) 0.0251 0.0003 92.0373 1.16×10− 5 CRBN encodes an ATP-dependent Lon protease that selec- (1)×(2) 0.0002 0.0002 0.6183 0.4543 tively degrades short-lived polypeptides and regulates mito- interaction chondrial replication and transcription [41].Inbrain,CRBNis 2+ miR-100 typically associated with the large conductance Ca -acti- + 0.7089 0.7544 Type (1) 0.0104 0.0008 12.3913 0.0078 vated K (BK(Ca)) channels and mutations in the CRBN gene Replicates (2) 0.0084 0.0002 52.7265 8.71×10− 5 result in long-term potentiation and learning dysfunctions − (1)×(2) 6.91 5 0.0002 0.4335 0.5288 leading to mental retardation. The complementarity between interaction miR-100 and its putative targets, including CRBN, is not P. Michalak, J.H. Malone / Genomics 91 (2008) 158–164 161

Table 3 MiRanda-predicted top candidate targets of miR-202⁎, miR-125b, and xtr-miR-100 Gene Description Score Energy p value Alignment xtr-miR-202⁎ BC076989.1 Hypothetical protein 20.2392 −14.06 0.0009

ENSXETT00000039632 cAMP-dependent protein kinase type Iα 20.142 −19.5 10− 6 regulatory subunit (tissue-specific extinguisher 1) (TSE1) ENSXETT00000058115 Sacsin 18.8403 −22.04 0.0035 ap3s2 Hypothetical protein LOC496606 18.5336 −18.46 0.0047

Egg005f01.1 Zinc finger, AN1-type domain 2A 18.5336 −12.62 0.0047

xtr-miR-125b hs3st1 Heparan sulfate (glucosamine) 18.0788 −27.24 0.0023 3-O-sulfotransferase

JARID1A Jumonji/ARID domain-containing protein 17.6511 −34.21 0.0042 1A (retinoblastoma-binding protein 2) (RBBP-2) pecr Pecr-prov protein (fragment) 17.5634 −22.72 0.0004

ENSXETT00000013476 β-1,3-Galactosyltransferase 5 17.442 −27.59 3.0− 5

xtr-miR-100 CRBN_XENTR Protein cereblon 18.0964 −25.93 0.0018

UCK2 Uridine–cytidine kinase 2 17.0573 −21.18 0.0003

TNeu066d14.1 Hypothetical protein 17.0555 −21.39 0.0052

GLP2R Glucagon-like peptide 2 receptor 16.9971 −23.87 0.0055 precursor

CBLN2_HUMAN Cerebellin-2 precursor 16.6989 −27.64 0.0018

perfect but this seems to be typical of most animal miRNAs, miR-30b, miR-15b, miR-7c, and miR-467 (Supplementary which, unlike small interfering RNAs (siRNAs) and plant Table 1). This finding suggests a considerable conservation of miRNAs, are characterized by an imperfect match to their male germ-line miRNA pathways in vertebrates and necessi- mRNA targets [42,43]. The GO categorization of the putative tates further explorations into the roles of miRNA in male miR-100 targets in Xenopus suggests enrichment in genes reproduction and hybrid sterility. There is no evidence, how- related to neurophysiological processes, sensory perception of ever, that the other top 2 candidate miRNAs, miR-202⁎ and light stimulus, exocytosis, chromatin and nucleosome assem- miR-125b, are testis-specific in Xenopus or other vertebrates. bly, and protein localization. Lee and colleagues [44] have used a siRNA against the loop In addition to miR-100, 7 other miRNAs from the Yu and region of the mirR-125b precursor to down-regulate selectively colleagues [17] list of the 29 mouse testis-specific miRNAs the mature miRNA in human cell lines. The depletion of miR- were also present in our list: miR-16, miR-143, miR-21, 125b resulted in a significant proliferation defect in PC-3 162 P. Michalak, J.H. Malone / Genomics 91 (2008) 158–164

(prostate cancer) cells. The miRanda-predicted targets of mirR- transcription factors, miRNAs, signaling pathways, and epige- 125b in Xenopus are enriched in transferase and hydrolase/ netic regulators are wired together into multicomponent protease activities and associated with such biological process- regulatory networks. es as proteolysis, protein complex, chromatin and nucleosome assembly, DNA packaging, RNA export, and localization. There was a substantial overlap between the target GO categories of the three miRNAs, with the largest between miR-100 and miR- 125b (19%) and smallest between miR-100 and miR-202⁎ (2%). This result suggests the miRNA cross-targeting of such functions as chromatin assembly, nucleosome assembly, and cellular protein metabolism. In contrast to many hybrid anomalies in the expression of translated genes, differential miRNAs were consistently under- represented and virtually no significant overrepresentations were observed in hybrids relative to one of the parental species. This may reflect an overall miRNA underexpression due to hybrid genetic incompatibilities between upstream components of the miRNA pathway. In animal cells, miRNAs are transcribed from genes as primary miRNAs by RNA polymerase II, then cleaved by Drosha, an RNase III enzyme, into hairpin-like precursor miRNA, and eventually transported into the cytoplasm with the help of Exportin [34]. In the cytoplasm, pre-miRNAs are cut into double-stranded RNA duplexes by Dicer, another RNase III enzyme, processed into mature miRNAs, and incorporated into RISC with other components for target recognition. Dysfunctions at any of these levels are expected to result in a global underexpression of miRNAs, which in turn may lead to overexpression of their target genes on the hybrid genomic background. Future studies will need to address whether this pattern of miRNA underexpression in hybrid males also holds in relation to the other parental species, X. muelleri, and how that compares to parental and hybrid females. Could this prediction of uniform upregulation of the target genes due to misregulation of the upstream components in the miRNA pathway explain the pattern of other gene misexpressions in hybrids between X. laevis and X. muelleri? Given the lack of direct experimental data on the true miRNA targets in these two Xenopus species, all answers to this question will necessarily be speculative at this moment. Nevertheless, Malone and colleagues [27] have observed that 65% of all hybrid gene misexpressions relative to the parental X. laevis were in fact overexpressions. Among all misexpressed genes, only 6 of 142 were predicted target genes of Xenopus miR-100, miR-125b, or miR-202⁎, and all 6 but 1 (a miR-100 target) were overexpressed in hybrids. This tendency for enrichment in upregulations among putative miRNA targets was also apparent in a comparison of hybrid males with the other parental species, X. muelleri: 44% of all misexpressions (n=3995) were represented by overexpressions in hybrids, and 51% of the misexpressed putative miRNA targets (n=180) were upregu- lated (χ2 test, p=0.085). As there are examples of negative feedback loops that involve multiple miRNAs and transcription factors [45–47],it is likely that the prediction of uniform upregulation of miRNA targets due to miRNA misexpression is an oversimplification. A Fig. 2. Hy5 vs Hy3 scatterplots for the spike-in controls. These diagrams show comprehensive understanding of the genetic basis of hybrid that two different dyes used in spike-in controls produced similar fluorescence defects and speciation will require further insights into how signal. The black trend line represents x=y. P. Michalak, J.H. Malone / Genomics 91 (2008) 158–164 163

Materials and methods the target miRNA and the endogenous control were calculated and arcsin- transformed, and one-way repeated-measures ANOVA was used to test dif- RNA extraction ferences between male types (X. laevis and hybrids) and determine the effects of technical replications. A total of 5 μg/sample of RNA was extracted from mature testis in X. laevis (n=3) and sterile hybrids of X. laevis×X. muelleri (n=3) using the Ambion miRNA target prediction MirVana miRNA isolation kit and the manufacturer’s protocols. Adults were euthanized with MS-222 and tissue was immediately extracted and homo- To identify potential binding genomic sites for candidate miRNA, the genized in lysis/binding buffer. Samples of RNA were checked for quality using miRanda algorithm provided by a Welcome Trust Sanger Institute server was a Nanodrop ND 1000 spectrophotometer and the Agilent 2100 bioanalyzer. The used (http://microrna.sanger.ac.uk/targets). The miRanda uses dynamic pro- origin of the animals and the patterns of hybrid sterility have been described gramming alignment to identify highly complementary sites, which are scored elsewhere [27]. between 0 and 100, where 0 represents no complementarity and 100 complete complementarity. The algorithm incorporates a weighted scoring system that miRCURY LNA microRNA arrays rewards complementarity at the 5′ end of the miRNA and the Vienna RNA folding routines to estimate the target site thermodynamic stability. Statistical The Exiqon (Vedbaek, Denmark) miRCURY LNA microRNA arrays and overrepresentations of the target GO categories were examined with GoMiner the service to process the samples were used. This technology is based on the use [50]. of LNA-modified oligonucleotides that exhibit improved thermostability in nucleic acid duplexes [48]. The arrays consist of control probes, mismatch probes, and 1458 capture probes interrogating all known miRNAs as annotated Acknowledgments in miRBase release 8.1 [49], including 7 cloned X. laevis [32] and 177 predicted X. tropicalis miRNAs, as well as 146 miRPlus miRNA sequences unavailable in Comments made by two anonymous reviewers greatly miRBase. Thirty control capture probes include 10 spike-in, 8 negative control, improved the quality of the manuscript. Funding was provided and 12 small nuclear RNA-interrogating probes. The arrays are manufactured on by a SPRING Research and Nanotechnology Transfer Program a standard size glass slide and organized as eight subarrays in four replicates each with a spot size of 120 μm. For labeling, the miRCURY LNA kit was used grant to P.M., a National Science Foundation Dissertation according to the Exiqon protocols. The six biologically independent samples Improvement grant (DEB-0508882), and a Texas Academy of were labeled with Hy3 and Hy5 dyes and competitively hybridized on three Science Student Research grant to J.H.M. arrays, each contrasting hybrid with X. laevis samples. Dyes were alternated between sample types on one of the arrays. The spike-ins were labeled as normal Appendix A. Supplementary data miRNA types and hybridized to capture probes specifically designed for these RNAs. Supplementary data associated with this article can be found, Microarray data processing in the online version, at doi:10.1016/j.ygeno.2007.10.013.

Slides were scanned using a Genepix 4000B laser scanner (Axon Instruments) and microarray images were preprocessed using the Genepix Pro References 4.0 software. Data were normalized using the background subtraction and the global lowess (locally weighted scatterplot smoothing) regression algorithm. [1] H.A. Orr, J.P. Masly, N. Phadnis, Speciation in Drosophila: from pheno- Spike-in controls (10 different RNA oligonucleotides) were used to assess the types to molecules, J. Heredity 98 (2007) 103–110. overall performance of hybridizations. Signals from the Hy3 and Hy5 channels [2] T. Dobzhansky, Genetics and the Origin of Species, Columbia Univ. Press, 2 were highly correlated (R range between 0.972 and 0.992), indicating highly New York, 1937. reproducible labeling and hybridization across all slides (Fig. 2). The normalized [3] H.J. Muller, Bearing of the Drosophila work on systematics, in: J. Huxley expression values were log2 transformed and implemented in Cyber-T [30] and (Ed.), The New Systematics, Clarendon Press, Oxford, 1940, pp. 185–268. Edge [31] for tests of differential gene expression. To control the false discovery [4] A. Salmon, M.L. Ainouche, J.F. Wendel, Genetic and epigenetic conse- rate, a Bonferroni correction was applied to the p values generated by Cyber-T, quences of recent hybridization and polyploidy in Spartina (Poaceae), and q values were estimated using Edge. Hierarchical clustering with Euclidean Mol. Ecol. 14 (2005) 1163–1175. distances and heuristic ordering in Edge was used to visualize the global pattern [5] C. Josefsson, B. Dilkes, L. Comai, Parent-dependent loss of gene silencing of miRNA expression in hybrids relative to X. laevis. during interspecies hybridization, Curr. Biol. 16 (2006) 1322–1328. [6] M.C. Ungerer, S.C. Strakosh, Y. Zhen, Genome expansion in three hybrid Quantitative real-time PCR validation sunflower species is associated with retrotransposon proliferation, Curr. Biol. 16 (2006) R872–R873. Differentially expressed miRNAs were subjected to quantitative RT-PCR [7] J.B. Nasrallah, P. Liu, S. Sherman-Broyles, R. Schmidt, Nasrallah, validation using TaqMan MicroRNA Assays from Applied Biosystems (ABI Epigenetics mechanisms for breakdown of self-incompatibility in inter- Product Nos. 4373219, 4373148, 4373019, 4373160, 4373274). As there ap- specific hybrids, Genetics 175 (2007) 1965–1973. pears to be no consensus regarding endogenous control selection for the miRNA [8] P. Michalak, M.A.F. Noor, Genome-wide patterns of expression in relative quantification yet, we chose to use hsa-miR-382, which according to our Drosophila pure-species and hybrid males, Mol. Biol. Evol. 20 (2003) microarray data was highly expressed at very similar levels in all samples. A total 1070–1076. of 10 ng/sample of RNA was extracted as described above from testes of 10 [9] J.M. Ranz, K. Namgyal, G. Gibson, D.L. Hartl, Anomalies in the X. laevis and 10 hybrid males. cDNA was produced using the ABI TaqMan expression profile of interspecific hybrids of Drosophila melanogaster and MicroRNA Reverse Transcription Kit (Product No. 4366596) and the man- Drosophila simulans, Genome Res. 14 (2004) 373–379. ufacturer’s protocol. To test for genomic DNA and reaction contamination, two [10] M.J. Hegarty, G.L. Barker, I.D. Wilson, R.J. Abbott, K.J. Edwards, S.J. types of negative controls were used for PCR: reverse transcriptase-omitted Hiscock, Transcriptome shock after interspecific hybridization in Senecio products and blanks (DEPC-treated water). No amplification was observed in any is ameliorated by genome duplication, Curr. Biol. 16 (2006) 1652–1659. of the negative controls. The ABI 7300 RT-PCR system and the instrument [11] N.C. Riddle, J.A. Birchler, Effects of reunited diverged regulatory default cycling conditions were used. Each RNA sample was assayed twice for hierarchies in allopolyploids and species hybrids, Trends Genet. 19 every miRNA in separate RT reactions. The threshold cycle (CT) ratios between (2003) 597–600. 164 P. Michalak, J.H. Malone / Genomics 91 (2008) 158–164

[12] K.L. Adams, J.F. Wendel, Novel patterns of gene expression in polyploid [32] T. Watanabe, et al., Stage-specific expression of microRNAs during Xe- plants, Trends Genet. 21 (2005) 539–543. nopus development, FEBS Lett. 579 (2005) 318–324. [13] V. Ambros, MicroRNAs: tiny regulators with great potential, Cell 107 [33] J.F. Chen, et al., The role of microRNA-1 and microRNA-133 in skeletal (2001) 823–826. muscle proliferation and differentiation, Nat. Genet. 38 (2006) 228–233. [14] X. Xie, et al., Systematic discovery of regulatory motifs in human [34] K. Chen, N. Rajewsky, The evolution of gene regulation by transcription promoters and 3′ UTRs by comparison of several mammals, Nature 434 factors and microRNAs, Nat. Rev., Genet. 8 (2007) 93–103. (2005) 338–345. [35] L.F. Sempere, S. Freemantle, I. Pitha-Rowe, E. Moss, E. Dmitrovsky, V. [15] O.A. Kent, J.T. Mendell, A small piece in the cancer puzzle: microRNAs Ambros, Expression profiling of mammalian microRNAs uncovers a as tumor suppressors and oncogenes, Oncogene 25 (2006) 6188–6196. subset of brain-expressed microRNAs with possible roles in murine and [16] M. Lagos-Quintana, R. Rauhut, A. Yalcin, J. Meyer, W. Lendeckel, T. human neuronal differentiation, Genome Biol. 5 (2004) R13. Tuschl, Identification of tissue-specific microRNAs from mouse, Curr. [36] B. Ason, et al., Differences in vertebrate microRNA expression, Proc. Natl. Biol. 12 (2002) 735–739. Acad. Sci. U. S. A. 103 (2006) 14385–14389. [17] Z. Yu, T. Raabe, N.B. Hecht, MicroRNA Mirn122a reduces expression of [37] L.F. Sempere, et al., The phylogenetic distribution of metazoan microRNAs: the posttranscriptionally regulated germ cell transition protein 2 (Tnp2) insights into evolutionary complexity and constraint, J. Exp. Zool. B Mol. messenger RNA (mRNA) by mRNA cleavage, Biol. Reprod. 73 (2005) Dev. Evol. 306 (2006) 575–588. 427–433. [38] D. Lemons, W. McGinnis, Genomic evolution of Hox gene clusters, [18] A.A. Aravin, M.S. Klenov, V.V. Vagin, F. Bantignies, G. Cavalli, V.A. Science 313 (2006) 1918–1922. Gvozdev, Dissection of a natural RNA silencing process in the Drosophila [39] K.C. Kleene, Patterns, mechanisms, and functions of translation regulation melanogaster germ line, Mol. Cell. Biol. 24 (2004) 6742–6750. in mammalian spermatogenic cells, Cytogenet. Genome Res. 103 (2003) [19] A. Girard, R. Sachidanandam, G.J. Hannon, M.A. Carmell, A germline- 217–224. specific class of small RNAs binds mammalian Piwi proteins, Nature 442 [40] W. Xin, N. Xiaohua, C. Peilin, C. Xin, S. Yaqiong, W. Qihan, Primary (2006) 199–202. function analysis of human mental retardation related gene CRBN, Mol. [20] S.T. Grivna, B. Pyhtila, H. Lin, MIWI associates with translational Biol. Rep. (in press) (Epub ahead of print). machinery and PIWI-interacting RNAs (piRNAs) in regulating sperma- [41] J.J. Higgins, J. Pucilowska, R.Q. Lombardi, J.P. Rooney, A mutation in a togenesis, Proc. Natl. Acad. Sci. U. S. A. 103 (2006) 13415–13420. novel ATP-dependent Lon protease gene in a kindred with mild mental [21] N.C. Lau, et al., Characterization of the piRNA complex from rat testes, retardation, Neurology 63 (2004) 1927–1931. Science 313 (2006) 363–367. [42] P.D. Zamore, T. Tuschl, P.A. Sharp, D.P. Bartel, RNAi: double-stranded [22] F. Tang, et al., Maternal microRNAs are essential for mouse zygotic RNA directs the ATP-dependent cleavage of mRNA at 21 to 23 nucleotide development, Genes Dev. 21 (2007) 644–648. intervals, Cell 101 (2000) 25–33. [23] M.D. Picker, Hybridization and habitat selection in Xenopus gilli and [43] M.A. Valencia-Sanchez, J. Liu, G.J. Hannon, R. Parker, Control of Xenopus laevis in the southwestern Cape Province, Copeia 1985 (1985) translation and mRNA degradation by miRNAs and siRNAs, Genes Dev. 574–580. 20 (2006) 515–524. [24] H.R. Kobel, Reproductive capacity of experimental Xenopus gilli×X. l. [44] Y.S. Lee, H.K. Kim, S. Chung, K.S. Kim, A. Dutta, Depletion of human laevis hybrids, in: R.C. Tinsely, H.R. Kobel (Eds.), The Biology of Xe- micro-RNA miR-125b reveals that it is critical for the proliferation of nopus, Clarendon Press, Oxford, 1996, pp. 73–80. differentiated cells but not for the down-regulation of putative targets [25] H.R. Kobel, C. Loumont, R.C. Tinsley, The extant species, in: R.C. Tinsely, during differentiation, J. Biol. Chem. 280 (2005) 16635–16641. H.R. Kobel (Eds.), The Biology of Xenopus, Clarendon Press, Oxford, [45] R.J. Johnston Jr., S. Chang, J.F. Etchberger, C.O. Ortiz, O. Hobert, 1996, pp. 9–33. MicroRNAs acting in a double-negative feedback loop to control a neuronal [26] W.J. Fischer, W.A. Koch, A. Elepfandt, Sympatry and hybridization cell fate decision, Proc. Natl. Acad. Sci. U. S. A. 102 (2005) 12449–12454. between the clawed frogs Xenopus laevis laevis and Xenopus muelleri [46] A.S. Yoo, I. Greenwald, LIN-12/Notch activation leads to microRNA- (Pipidae), J. Zool. London 252 (2000) 99–107. mediated down-regulation of Vav in C. elegans, Science 310 (2005) [27] J.H. Malone, T.H. Chrzanowski, P. Michalak, Sterility and gene expression 1330–1333. in hybrid males of Xenopus laevis and X. muelleri, PLoS One 2 (2007) e781. [47] F. Fazi, A minicircuitry comprised of microRNA-223 and transcription [28] M.N. Poy, et al., A pancreatic islet-specific microRNA regulates insulin factors NFI-A and C/EBPalpha regulates human granulopoiesis, Cell 123 secretion, Nature 432 (2004) 226–230. (2005) 819–831. [29] G.M. Schratt, et al., A brain-specific microRNA regulates dendritic spine [48] M. Castoldi, et al., A sensitive array for microRNA expression profiling development, Nature 439 (2006) 283–289. (miChip) based on locked nucleic acids (LNA), RNA 12 (2006) 913–920. [30] P. Baldi, A.D. Long, A Bayesian framework for the analysis of microarray [49] S. Griffiths-Jones, R.J. Grocock, S. Van Dongen, A. Bateman, A.J. expression data: regularized ttest and statistical inferences of gene changes, Enright, miRBase: microRNA sequences, targets and gene nomenclature, Bioinformatics 17 (2001) 509–519. Nucleic Acids Res. 34 (2006) D140–D144. [31] J.T. Leek, E. Monsen, A.R. Dabney, J.D. Storey, EDGE: extraction and [50] B.R. Zeeberg, et al., GoMiner: a resource for biological interpretation of analysis of differential gene expression, Bioinformatics 22 (2006) 507–508. genomic and proteomic data, Genome Biol. 4 (2003) R28.