letters
© 2014 Nature America, Inc. All rights reserved.
TDP2 protects transcription from abortive topoisomerase activity and is required for normal neural function Fernando Gómez-Herreros1,12, Janneke H M Schuurs-Hoeijmakers2,3,12, Mark McCormack4,12, Marie T Greally5, Stuart Rulten1, Rocío Romero-Granados6, Timothy J Counihan7, Elijah Chaila8, Judith Conroy9, Sean Ennis9, Norman Delanty4,8, Felipe Cortés-Ledesma6, Arjan P M de Brouwer2,3, Gianpiero L Cavalleri4, Sherif F El-Khamisy10,11, Bert B A de Vries2,3 & Keith W Caldecott1 Topoisomerase II (TOP2) removes torsional stress from DNA and facilitates gene transcription by introducing transient DNA double-strand breaks (DSBs). Such DSBs are normally rejoined by TOP2 but on occasion can become abortive and remain unsealed. Here we identify homozygous mutations in the TDP2 gene encoding tyrosyl DNA phosphodiesterase-2, an enzyme that repairs ‘abortive’ TOP2-induced DSBs, in individuals with intellectual disability, seizures and ataxia. We show that cells from affected individuals are hypersensitive to TOP2-induced DSBs and that loss of TDP2 inhibits TOP2-dependent gene transcription in cultured human cells and in mouse post-mitotic neurons following abortive TOP2 activity. Notably, TDP2 is also required for normal levels of many gene transcripts in developing mouse brain, including numerous gene transcripts associated with neurological function and/or disease, and for normal interneuron density in mouse cerebellum. Collectively, these data implicate chromosome breakage by TOP2 as an endogenous threat to gene transcription and to normal neuronal development and maintenance. Clinical evaluation of a consanguineous Irish family identified three brothers with intellectual disability, epilepsy and various levels of ataxia (Fig. 1a,b). The affected brothers (denoted IV-9, IV-14 and IV-16) developed seizures at ~2 months, 12 years and ~6 months of age, respectively. Electroencephalography conducted on two of the brothers (IV-9 at age 23 years and IV-16 at age 11 years) showed pronounced epileptiform activity, consistent with a diagnosis of epileptic encephalopathy. The seizures are frequent and are refractory to antiepileptic drugs. Ataxia appears to be progressive, with all three brothers learning to walk at a normal age (10–12 months) but now
requiring support when walking, with the two oldest (IV-9 and IV-14) largely wheelchair bound. Dense SNP genotyping was conducted in ten individuals from generations III and IV of this family (Fig. 1a). Extensive homozygosity was apparent, and analysis identified a 9.08-Mb region of homozygosity (chr. 6: 20,334,647–29,417,748; NCBI Build 36) common only to the three affected siblings (Supplementary Fig. 1). Two different centers (Dublin, for individuals IV-9 and IV-14, and Nijmegen, The Netherlands, for individual IV-9) independently performed exome sequencing, and candidate mutations were selected under the assumption of an autosomal recessive disease model (Online Methods and Supplementary Tables 1 and 2). After combining results, three variants in two genes remained as potentially causal: a missense and a putative splice-site mutation in TDP2, encoding an enzyme involved in DNA repair1, and a missense mutation in ZNF193, encoding a zincfinger protein of unknown function (Supplementary Table 3). The TDP2 missense variant (c.919T>C; p.Ile307Val) is most likely benign because it is predicted not to affect protein function by both SNP&Go and PolyPhen-2 (Supplementary Table 3)2,3 and because the affected isoleucine residue is not conserved in several other vertebrate species. In contrast, the putative splice-donor mutation (c.425+1G>A) in TDP2, which we have confirmed by Sanger sequencing (Fig. 1c and Supplementary Fig. 2a,b), is predicted to result in either retention of intron 3, skipping of exon 3 or the use of a cryptic splice-donor site upstream of the mutation (Supplementary Fig. 2c). Each of these splicing errors would result in the insertion of a premature stop codon within the N-terminal half of the encoded protein, thereby deleting conserved domains critical for TDP2 activity (Supplementary Fig. 3). Consistent with this hypothesis, mRNA analyses in lymphoblastoid cells prepared from the affected brothers (IV-9, IV-14 and IV-16) showed that mutant TDP2 mRNA is both truncated and subject
1Genome
Damage and Stability Centre, School of Biological Sciences, University of Sussex, Sussex, UK. 2Department of Human Genetics, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands. 3Department of Cognitive Neurosciences, Donders Institute for Brain Cognition and Behaviour, Radboud University Medical Centre, Nijmegen, The Netherlands. 4Molecular and Cellular Therapeutics, The Royal College of Surgeons in Ireland, Dublin, Ireland. 5National Centre for Medical Genetics, Our Lady’s Children’s Hospital, Crumlin, Dublin, Ireland. 6Centro Andaluz de Biología Molecular y Medicina Regenerativa (CABIMER), Departamento de Genética, CSIC (Centro Superior de Investigaciones Científicas)–Universidad de Sevilla, Sevilla, Spain. 7Department of Neurology, University Hospital Galway, Galway, Ireland. 8Division of Neurology, Beaumont Hospital, Dublin, Ireland. 9School of Medicine and Medical Science, University College Dublin, Dublin, Ireland. 10Kreb’s Institute, Department of Molecular Biology and Biotechnology, University of Sheffield, Sheffield, UK. 11Center of Genomics, Helmy Institute, Zewail City of Science and Technology, Giza, Egypt. 12These authors contributed equally to this work. Correspondence should be addressed to K.W.C. (k.w.caldecott@ sussex.ac.uk), B.B.A.d.V. (
[email protected]), S.F.E.-K. (
[email protected]) or G.L.C. (
[email protected]). Received 1 September 2013; accepted 28 February 2014; published online 23 March 2014; doi:10.1038/ng.2929
Nature Genetics ADVANCE ONLINE PUBLICATION
letters
*
Y-P P
TDP1
Y-P P
IV
-2
4
Actin
-9
-1 IV
N
R
6
b
3′-TDP activity
IV
P
-1
*
TDP2
Actin Y-P P
IV
P
TDP1
5′-TDP activity
TOP2α TDP1
Y-P P
II
3′-TDP activity
TOP2β TDP1
+/–
III 1
IV 1
b
+/+
+/– +/+
2
3
4
2
+/– 5
6
7
8
–/– 9
+/– 10
–/– 11
12 13 14
+/+ 15
–/– 16
V
Proband (age)
IV-9 (32 years) IV-14 (26 years) IV-16 (23 years)
Seizures/epilepsy (age at onset)
Yes ( 0.05). (b) Recruitment of RNAP II at the KLK3 promoter in the absence Organ morphology of gene induction, 8 h after gene induction with 100 nM DHT and after gene induction with 100 µM etoposide present for two additional hours. Data TOP2β-dependent Organismal genes development are presented as the amount of DNA precipitated relative to the recovery of a non-transcribed region of chromosome 5 and are the mean (±s.e.m.) of three independent experiments. (c) Top, representative images of EU pulse labeling in Tdp2+/+ and Tdp2∆1–3 astrocytes and Tdp2+/∆1–3 and Tdp2∆1–3 granule neurons before and after 2-h treatment with 0.5 mM etoposide and after a further 90-min recovery period in drug-free medium. Bottom, transcription was quantified using SimplePCI 6.0. Data are the mean (±s.e.m.) of three independent experiments (two-tailed t test, **P < 0.005). (d) mRNA levels of the indicated developmentally regulated TOP2β-dependent genes (Cacna2d1, Kcnd2, Syt1) and controls (Actb, Tubb3, Gapdh) in E16.5 brain from wild-type and Tdp2∆1–3 mice. mRNA levels were quantified by qRT-PCR as in a but were normalized to Gapdh levels instead of to Actb levels. Data are the mean (±s.e.m.) for three independent littermate pairs. Statistically significant differences are indicated (two-tailed t test, *P < 0.05, **P < 0.01). NS, not statistically significant (P > 0.05). (e) Physiological processes that are significantly over-represented in terms of altered gene transcription in embryonic brain from Tdp2∆1–3 mice. The percentage of affected genes (n = 165) that fall into each category is plotted. Bars are sorted by the P-value range for the relevant genes: (top to bottom) P = 1.5 × 10−2 to 5.9 × 10−13 (53 genes); P = 1.5 × 10−2 to 3.4 × 10−7 (28 genes); P = 8.0 × 10−5 (5 genes); P = 1.5 × 10−2 to 8.2 × 10−5 (20 genes); P = 1.5 × 10−2 to 8.2 × 10−5 (20 genes). Data are from three independent littermate pairs.
d
© 2014 Nature America, Inc. All rights reserved.
Granule neurons
Tdp2∆1–3 EU DAPI
Tdp2+/+ EU DAPI
EU incorporation (%)
**
*** * ***
c
No DHT DHT DHT + etoposide
EU incorporation (%)
4
b
Etoposide No Etoposide + recovery etoposide
6
Scrambled, no DHT Scrambled, DHT Scrambled, DHT + etoposide TDP2 siRNA, no DHT TDP2 siRNA, DHT TDP2 siRNA, *** DHT + etoposide
Relative immunoprecipitation of KLK3 DNA
mRNA levels (AU)
a8
defects in DSB repair in mice and has previously been associated with seizures and ataxia18. To explain how loss of TDP2 might affect neural function, we reasoned that abortive TOP2 activity might be sufficiently frequent to disrupt gene transcription in the absence of TDP2-dependent DNA repair. Such disruption could occur because TOP2 activity is required for the transcription of many genes, including numerous genes associated with normal neurological development and/or function 8,9,19–23. To test this hypothesis, we first employed a model cell system in which TOP2-dependent transcription of androgen receptor (AR)-responsive genes could be induced by dihydrotestosterone (DHT)24. This model system has previously been used to demonstrate induction of sitespecific TOP2-induced DSBs at the TMPRSS2 locus, a region commonly associated with TOP2-induced chromosome translocations. As reported previously24, incubation of wild-type LNCaP prostate cancer cells with DHT induced expression of the AR-responsive genes TMPRSS2, KLK2 and KLK3 but did not induce the expression of three control genes that lack AR-responsive elements (Fig. 4a, compare blue and yellow bars, and Supplementary Fig. 8a). Strikingly, transcriptional induction of the AR-responsive genes was greatly reduced in LNCaP cells in which TDP2 was depleted by small interfering RNA (siRNA), suggesting that TDP2 is required for the normal induction of these genes, even at endogenous levels of abortive TOP2 activity (Fig. 4a,
compare yellow bars and striped yellow bars, and Supplementary Fig. 9). In agreement with this finding, whereas coincubation with etoposide to increase abortive TOP2 activity reduced DHT-induced transcription of the AR-responsive genes by up to 50% in wild-type cells (Fig. 4a, compare yellow and red bars, and Supplementary Fig. 8b), it almost ablated the induction of these genes in TDP2-depleted cells (Fig. 4a, compare striped yellow and striped red bars). Intriguingly, promoter occupancy by RNA polymerase II (RNAP II) at TMPRSS2 and KLK3 after induction with DHT was greatly reduced in wild-type LNCaP cells by treatment with etoposide, as measured by chromatin immunoprecipitation, suggesting that abortive TOP2 activity might reduce gene expression in part by reducing RNAP II recruitment or retention (Fig. 4b and Supplementary Fig. 8c). Consistent with this possibility, RNAP II occupancy at the KLK3 promoter in TDP2-depleted cells was greatly reduced even in the absence of etoposide treatment, mirroring the impact of TDP2 depletion on KLK3 expression (Fig. 4b). Collectively, these data suggest that TDP2 is required for normal levels of androgen-stimulated gene transcription, even at physiological levels of abortive TOP2 activity. Next, we extended our experiments to examine the levels of gene transcription in post-mitotic neural cells. First, we examined whether loss of TDP2 affected the recovery after treatment with etoposide of global nuclear gene transcription in quiescent neural astrocytes aDVANCE ONLINE PUBLICATION Nature Genetics
© 2014 Nature America, Inc. All rights reserved.
letters and granule neurons from mice at postnatal days 4–6 (P4–P6) by measuring the incorporation of 5′-ethynyl uridine (EU) into nascent RNA by pulse labeling. Indeed, whereas etoposide suppressed transcription to a similar extent in wild-type and Tdp2∆1–3 astrocytes or neurons, the recovery of transcription after the removal of etoposide was much slower in the Tdp2∆1–3 cells (Fig. 4c). To examine whether TDP2 also influenced neural gene expression with physiological levels of abortive TOP2 activity (in the absence of etoposide), we used quantitative RT-PCR (qRT-PCR) to measure the expression of three genes previously demonstrated to be transcribed in a TOP2βdependent manner in embryonic day 16.5 (E16.5) brain20. Notably, two of these genes (Kcnd2 and Syt1) exhibited significantly reduced expression levels (decreased by ~50%) in Tdp2∆1–3 developing brain, confirming that TDP2 is required for the normal neural expression of TOP2-dependent genes, even at physiological levels of abortive TOP2 activity (Fig. 4d). Finally, we compared genome-wide gene expression in wildtype and Tdp2∆1–3 developing brain using Affymetrix gene chips. Strikingly, 165 genes exhibited expression that was altered by 1.5-fold or more (P < 0.05) in Tdp2∆1–3 brain, with ~70% of these genes showing reduced expression (Supplementary Table 4). Of the physiological processes that were significantly over-represented by altered gene expression, nervous system development and function was the most affected, accounting for >30% of the deregulated genes (Fig. 4e and Supplementary Table 4). Moreover, >50 of the deregulated genes are associated with neurological disease in humans, including many that are linked to seizures/epilepsy, cognitive impairment and ataxia. Of particular note was the reduced expression of multiple γ-aminobutyric acid (GABA) and glutamate receptors. Also of note, the average size of the downregulated genes in Tdp2∆1–3 brain was ~150 kb, compared to an average for all expressed genes in the array data of ~50 kb. This finding is consistent with a recent report that etoposide-induced abortive TOP2β activity preferentially affects the expression of very large human genes19. Surprisingly, of the genes that were upregulated in Tdp2∆1–3 brain, most were involved in the cell cycle or cell division, suggesting an increase in cell proliferation. The reason for this effect is currently unknown but could reflect delayed development and/or reduced expression of GABA and/or glutamate receptors, as both GABA and glutamate signaling can suppress cell proliferation25,26. In summary, we have identified inactivating mutations in TDP2 in individuals with intellectual disability, seizures and ataxia, and we show that TDP2 protects gene transcription from endogenous abortive TOP2 activity, including the transcription of many genes involved in neurological development or function. Collectively, these data identify TDP2-dependent DNA break repair as a critical guardian of gene expression and highlight abortive TOP2 activity as a threat to normal neuronal development and maintenance. URLs. Sequence Variant Analyzer (SVA) program, http://www. svaproject.org/; Ensembl Variant Effect Predictor (VEP), http://www. ensembl.org/info/docs/variation/vep/index.html; Online Mendelian Inheritance in Man (OMIM), http://www.omim.org/; Primer3 program, http://bioinfo.ut.ee/primer3-0.4.0/. Methods Methods and any associated references are available in the online version of the paper. Accession codes. The gene expression microarray data discussed in this publication have been deposited in the NCBI Gene Expression Nature Genetics ADVANCE ONLINE PUBLICATION
Omnibus (GEO)27 and are accessible through GEO series accession GSE54628. Note: Any Supplementary Information and Source Data files are available in the online version of the paper. Acknowledgments We thank the patients and their family for their cooperation in this research project, N. Sabry for the provision of blood from the Egyptian patient, J.H.J. Hoeijmakers for critical evaluation of the manuscript, T. van Moorsel and L. Ju for technical assistance and D. Huylebroeck for the provision of unpublished behavioral data on Tdp2∆1–3 mice. Microarray analysis was conducted by M. Hubank and colleagues at University College London (UCL) Genomics, UCL Institute of Child Health. We also thank D.B. Goldstein, E.L. Heinzen and the Duke Center for Human Genome Variation Genetic Analysis Facility, and we thank the following individuals associated with the Carol Woods and Crosdaile Retirement Communities, the MURDOCK Study Community Registry and Biorepository, and the Washington University Neuromuscular Genetics Project: J. McEvoy, A. Need, J. Silver, M. Silver; E.T. Cirulli, V. Dixon, D.K. Attix, O. Chiba-Falek, K. Schmader, S. McDonald, H.K. White, M. Yanamadala, C. Depondt, S. Sisodiya, W.B. Gallentine, A.M. Husain, M.A. Mikati, R.A. Radtke, S.R. Sinha, J. Hoover-Fong, N.L. Sobreira, D. Valle, D. Daskalakis, W.L. Lowe, V. Shashi, K. Schoch, D.H. Murdock, S.M. Palmer, Z. Farfel, D.D. Lancet, E. Pras, A. Holden, E. Behr, A. Poduri; P. Lugar, D. Marchuk, S. Kerns, H. Oster, R. Gbadegesin, M. Winn, E.J. Holtzman, Y.-H. Jiang, R. Brown, S.H. Appel, E. Simpson, S. Halton, L. Lay, R. Bedlack, K. Grace. This work was funded in the Caldecott laboratory by the Medical Research Council (MRC; MR/J006750/1 and G0901606/1) and Cancer Research UK (C6563/A16771), in the Cortes-Ledesma laboratory by the Spanish government (SAF2010-21017, RYC-2009-03928 and JAE-Doc 2010-011) and European Union (PERG07-2010268466), in the El-Khamisy laboratory by the Wellcome Trust (fellowship 085284 and grant 091043) and the Lister Institute of Preventative Medicine (fellowship), and in part by the Netherlands Organization for Health Research and Development (ZonMW; VIDI grant 917-86-319 to B.B.A.d.V.), the GENCODYS project (EU-7th-2010-241995 to B.B.A.d.V. and A.P.M.d.B.), a Brainwave–Irish Epilepsy Association/Medical Research Charities Group of Ireland/Health Research Board award (2009/001) and a Health Research Board of Ireland Translational Research Scholars award. Control samples were funded by National Institute for Mental Health (NIMH) awards (RC2MH089915, K01MH098126, R01MH099216 and R01MH097971), the Epi4K Gene Discovery in Epilepsy study (National Institute for Neurological Disorders and Stroke (NINDS) U01NS077303), the Epilepsy Genome/ Phenome Project (EPGP; NINDS U01NS053998), the Center for HIV/AIDS Vaccine Immunology (CHAVI) study (National Institute of Allergy and Infectious Diseases (NIAID) UO1AIO67854), the Ellison Medical Foundation New Scholar award (AG-NS-0441-08), SAIC-Frederick, Inc. (M11-074) and Biogen Idec, Inc. AUTHOR CONTRIBUTIONS K.W.C. devised and coordinated the project. K.W.C. and F.G.-H. designed and interpreted the biochemical, cell biology and mouse experiments and wrote the manuscript. F.G.-H. conducted all biochemical and cell biology experiments. S.R. analyzed mouse interneurons, and F.C.-L. and R.R.-G. measured TDP2 activity in mouse brain tissue. J.H.M.S.-H., A.P.M.d.B. and B.B.A.d.V. conducted and interpreted exome sequencing under the supervision of B.B.A.d.V. and identified the human splice-site mutation in Nijmegen. M.M., J.C., S.E. and G.L.C. conducted and interpreted genome-wide association study and homozygosity mapping in Ireland under the supervision of G.L.C. and identified the TDP2 splice-site mutation by exome sequencing in collaboration with the Duke Center for Human Genome Variation. E.C., N.D. and T.J.C. recruited and phenotyped patients in Ireland. M.T.G. consulted, phenotyped and liaised with patients and their families in Ireland. S.F.E.-K. identified and coordinated the analysis of the TDP2 patient in Egypt. COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. Reprints and permissions information is available online at http://www.nature.com/ reprints/index.html. 1. Cortes Ledesma, F., El-Khamisy, S.F., Zuma, M.C., Osborn, K. & Caldecott, K.W. A human 5′-tyrosyl DNA phosphodiesterase that repairs topoisomerase-mediated DNA damage. Nature 461, 674–678 (2009). 2. Calabrese, R., Capriotti, E., Fariselli, P., Martelli, P.L. & Casadio, R. Functional annotations improve the predictive score of human disease-related mutations in proteins. Hum. Mutat. 30, 1237–1244 (2009).
© 2014 Nature America, Inc. All rights reserved.
letters 3. Adzhubei, I.A. et al. A method and server for predicting damaging missense mutations. Nat. Methods 7, 248–249 (2010). 4. Rodrigues-Lima, F., Josephs, M., Katan, M. & Cassinat, B. Sequence analysis identifies TTRAP, a protein that associates with CD40 and TNF receptor–associated factors, as a member of a superfamily of divalent cation-dependent phosphodiesterases. Biochem. Biophys. Res. Commun. 285, 1274–1279 (2001). 5. Pype, S. et al. TTRAP, a novel protein that associates with CD40, tumor necrosis factor (TNF) receptor-75 and TNF receptor–associated factors (TRAFs), and that inhibits nuclear factor-κB activation. J. Biol. Chem. 275, 18586–18593 (2000). 6. Gómez-Herreros, F. et al. TDP2-dependent non-homologous end-joining protects against topoisomerase II–induced DNA breaks and genome instability in cells and in vivo. PLoS Genet. 9, e1003226 (2013). 7. Zeng, Z. et al. TDP2 promotes repair of topoisomerase I–mediated DNA damage in the absence of TDP1. Nucleic Acids Res. 40, 8371–8380 (2012). 8. Nitiss, J.L. DNA topoisomerase II and its growing repertoire of biological functions. Nat. Rev. Cancer 9, 327–337 (2009). 9. Vos, S.M., Tretter, E.M., Schmidt, B.H. & Berger, J.M. All tangled up: how cells direct, manage and exploit topoisomerase function. Nat. Rev. Mol. Cell Biol. 12, 827–841 (2011). 10. Nitiss, J.L. Targeting DNA topoisomerase II in cancer chemotherapy. Nat. Rev. Cancer 9, 338–350 (2009). 11. Pommier, Y., Leo, E., Zhang, H. & Marchand, C. DNA topoisomerases and their poisoning by anticancer and antibacterial drugs. Chem. Biol. 17, 421–433 (2010). 12. Schellenberg, M.J. et al. Mechanism of repair of 5′-topoisomerase II–DNA adducts by mammalian tyrosyl-DNA phosphodiesterase 2. Nat. Struct. Mol. Biol. 19, 1363–1371 (2012). 13. Shi, K. et al. Structural basis for recognition of 5′-phosphotyrosine adducts by Tdp2. Nat. Struct. Mol. Biol. 19, 1372–1377 (2012). 14. Esguerra, C.V. et al. Ttrap is an essential modulator of Smad3-dependent Nodal signaling during zebrafish gastrulation and left-right axis determination. Development 134, 4381–4393 (2007).
15. Takashima, H. et al. Mutation of TDP1, encoding a topoisomerase I–dependent DNA damage repair enzyme, in spinocerebellar ataxia with axonal neuropathy. Nat. Genet. 32, 267–272 (2002). 16. Katyal, S. et al. TDP1 facilitates chromosomal single-strand break repair in neurons and is neuroprotective in vivo. EMBO J. 26, 4720–4731 (2007). 17. Hirano, R. et al. Spinocerebellar ataxia with axonal neuropathy: consequence of a Tdp1 recessive neomorphic mutation? EMBO J. 26, 4732–4743 (2007). 18. Lee, Y. et al. The genesis of cerebellar interneurons and the prevention of neural DNA damage require XRCC1. Nat. Neurosci. 12, 973–980 (2009). 19. King, I.F. et al. Topoisomerases facilitate transcription of long genes linked to autism. Nature 501, 58–62 (2013). 20. Lyu, Y.L. et al. Role of topoisomerase IIβ in the expression of developmentally regulated genes. Mol. Cell. Biol. 26, 7929–7941 (2006). 21. Ju, B.-G. et al. A topoisomerase IIβ–mediated dsDNA break required for regulated transcription. Science 312, 1798–1802 (2006). 22. Thakurela, S. et al. Gene regulation and priming by topoisomerase IIα in embryonic stem cells. Nat. Commun. 4, 2478 (2013). 23. Tiwari, V.K. et al. Target genes of topoisomerase IIβ regulate neuronal survival and are defined by their chromatin state. Proc. Natl. Acad. Sci. USA 109, E934–E943 (2012). 24. Haffner, M.C. et al. Androgen-induced TOP2B-mediated double-strand breaks and prostate cancer gene rearrangements. Nat. Genet. 42, 668–675 (2010). 25. LoTurco, J.J., Owens, D.F., Heath, M.J., Davis, M.B. & Kriegstein, A.R. GABA and glutamate depolarize cortical progenitor cells and inhibit DNA synthesis. Neuron 15, 1287–1298 (1995). 26. Liu, X., Wang, Q., Haydar, T.F. & Bordey, A. Nonsynaptic GABA signaling in postnatal subventricular zone controls proliferation of GFAP-expressing progenitors. Nat. Neurosci. 8, 1179–1187 (2005). 27. Edgar, R., Domrachev, M. & Lash, A.E. Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res. 30, 207–210 (2002).
aDVANCE ONLINE PUBLICATION Nature Genetics
ONLINE METHODS
© 2014 Nature America, Inc. All rights reserved.
Affected family. Members of the Irish pedigree were initially recruited through the outreach genetics clinic of the National Centre for Medical Genetics at University Hospital Galway (UHG) and, later, through Beaumont Hospital, Dublin. Consent for research was obtained from the mother, in accordance with the ethics protocol at the National Centre for Medical Genetics (NCMG), from the Beaumont Hospital research ethics committee (protocol number 05/56) and from the regional ethics committee in Nijmegen (Commissie Mensgebonden Onderzoek Regio Arnhem-Nijmegen; protocol 2007/263). Written informed consent and blood samples were obtained from all ten participating family members. The proband and two of his affected brothers underwent neurological examinations as part of routine clinical care at Beaumont hospital. All participants without epilepsy or neurological dysfunction were interviewed by a neurologist in relation to their medical history. Previous copy number variant analysis by SNP array, karyotyping and fragile X screening had not led to an etiological diagnosis. Clinical assessment of the affected individual is presented as a Supplementary Note. Homozygosity mapping. Genome-wide marker data were obtained for nine family members (III-2, IV-2, IV-3, IV-4, IV-8, IV-9, IV-14, IV-15 and IV-16) on the Human1Mv1 DNA Analysis BeadChips (Illumina) using a scanning platform at the Health Sciences Centre genotyping facility, University College Dublin. An unaffected male sibling (IV-10) recruited at a later date was genotyped using a Human610-Quad BeadChip (Illumina) at the Duke University Centre for Human Genome Variation. In total, there were 571,241 overlapping genetic markers on the Human1Mv1 and Human610-Quad BeadChips. All quality control measures were performed using PLINK v1.07 (ref. 28). All subjects had a genotyping success rate of >95%. Markers that failed genotyping in more than 5% of individuals were removed. After quality control, the >500,000 common markers on the 2 genotyping platforms were analyzed for the detection of long runs of homozygosity using PLINK v1.07. We set a minimum size of 1 Mb and then 10 Mb for the sliding window to determine runs of consecutive homozygosity. The segments showing runs of homozygosity for each subject were aligned, and segments that were uniquely homozygous in the affected male siblings were identified and subsequently termed the ‘candidate causal region’. We used the UCSC Genome Browser table reporting tool to identify all RefSeq genes and transcripts within the candidate causal region. Exome sequencing. Exome sequencing in the Genomic Analysis Facility at the Duke Center for Human Genome Variation was performed on DNA isolated from the blood of the two eldest affected brothers (IV-9 and IV-14). Samples were enriched with the Agilent SureSelect Human All Exon (38Mb) capture array and sequenced on an Illumina Genome Analyzer IIx machine. This kit captures >37 Mb of the genome (approximately 1.22%) and includes NCBI Consensus CDS database (CCDS) regions, >700 human microRNAs from the Sanger v13 database and >300 additional human noncoding RNAs such as snoRNAs (small nucleolar RNAs) and scaRNAs (small cajal body– specific RNAs). Briefly, 1 µg of genomic DNA was fragmented by nebulization, the fragmented DNA was repaired, an adenine dNTP was ligated to the 3′ end, Illumina adaptors were ligated to the fragments, and the sample was size selected, aiming for a product of 350–400 bp. The size-selected product was PCR amplified, and the final product was validated using the Agilent Bioanalyzer. Samples were then amplified on the flow cell and sequenced using the Genome Analyzer IIx, following the Illumina-supplied protocols. The majority of sequence runs were paired end with 75-base reads29. Variant calling was performed with SAMtools software30. The Sequence Variant Analyzer (SVA) program (see URLs) was used for annotating variants identified from the sequence data31. This software provides each variant with a genomic context (nonsynonymous or splice-site coding, gene name, transcript, associated Gene Ontology term and other relevant transcript information). Each sample was checked for X-chromosome heterozygosity to ensure a correct sex match with the expected pedigree information, and genotype data were checked against exome calls to ensure fidelity during sequencing. High-confidence variant calls for 30,636 common and rare variants were obtained from wholeexome sequencing in the 2 affected siblings. Exome sequencing in the Department of Human Genetics in Nijmegen was performed on 3 µg of genomic DNA isolated from the blood of
doi:10.1038/ng.2929
individual IV-9. The exome was captured with an ABI SOLiD optimized SureSelect 50Mb human exome kit (Agilent) representing exonic sequences for ~21,000 genes (including >99% of genes from CCDS, version September 2009, and >95% of RefSeq genes and transcripts, version June 2010, as specified by the company). The manufacturer’s instructions (version 1.5) for enrichment were followed with a minor modification, which was the reduction of the number of post-hybridization ligation-mediated PCR cycles from 12 to 9. To allow for multiplexing of several libraries before sequencing, post-hybridization sample barcodes were used (Agilent) that were compatible with SOLiD sequencing technology. The enriched exome library was pooled with three other libraries in equimolar amounts, based on a combined library concentration of 0.7 pM. Subsequently, the obtained pool was used for emulsion PCR and bead preparation with the EZbead system, following the manufacturer’s instructions (version 05/2010; Life Technologies). A full sequencing slide was used for the four pooled libraries on a SOLiD 4 System (Life Technologies), thereby anticipating that all four samples would be represented by ~25% of the total beads sequenced on the slide. Color space reads were mapped to the hg19 reference genome with SOLiD Bioscope software version 1.3. Selecting candidate variants. To identify candidate variants of interest at the Duke Center for Human Genome Variation, filtering criteria including singlenucleotide variant quality, consensus score and a requirement of ≥3 reads supporting the variant were applied as standard in SVA 31. The 2 sequenced affected siblings (IV-9 and IV-14) were compared to exome data from 128 neurologically normal controls made available for use by E. Heinzen (Duke Center for Human Genome Variation). Controls were European-American subjects enrolled in the Duke Center for Human Genome Variation studies and consented for use through Duke Institutional Review Board–approved protocols. These controls were initially used to identify and filter out variants that were considered common (>5% minor allele frequency, MAF) and thus very unlikely to be disease causing under our presumed recessive model. For a further estimation of allele frequencies in a much larger population, we accessed data on 5,400 individuals from the Exome Variant Server database, maintained by the University of Washington National Heart, Lung, and Blood Institute (NHLBI) Grand Opportunity (GO) Exome Sequencing Project (ESP). The Ensembl Variant Effect Predictor (VEP) online tool (see URLs) was used to identify and predict the effect of amino acid changes on a protein’s viability as well as the impact of individual insertions and deletions. VEP uses the PolyPhen version 2 (Polymorphism Phenotyping)3 algorithm to provide, for nonsynonymous variants, a classification of impact (with categories being ‘benign’, ‘possibly damaging’ or ‘probably damaging’). To obtain a comprehensive set of candidate disease-causing polymorphisms within the candidate causal region, filters were applied to reduce the shared exome variants as follows. Initially, the shared variants were reduced to the 9-Mb candidate causal region obtained from homozygosity mapping. We removed variants with MAF of >1%. Where the EVS database reported zero or unavailable genotype data for a variant, the coverage at this location was checked on the EVS website. If a site was covered sufficiently (>10×), then the variant was considered novel (0% EVS MAF), whereas, if the site was not covered, the variant frequency could not be established from EVS and was obtained from the initial 128 neurologically normal controls. The remaining rare candidate variants were filtered by function, and only nonsynonymous, stop-gain, stop-loss and essential splice-site variants were kept for further consideration. From the Ensembl VEP report, PolyPhen scores were available to prioritize the most damaging variants (Supplementary Table 4). At the Department of Human Genetics in Nijmegen, exome-wide sequence variations, including indel variations, were selected using quality settings that required the presence of at least two unique variant reads as well as requiring the variation to be present in at least 20% of all reads. Non-genic, intronic and synonymous sequence variants (but not canonical splice-site variants) were excluded as well as alleles found in >1% of the population on the basis of dbSNPv134 and the local variant database consisting of data derived from 672 exome experiments performed in Nijmegen. Under the hypothesis of a recessive inheritance model, all rare variants were selected in genes on the autosomes present either in a presumed homozygous state (>80% of variation reads) or in a presumed compound heterozygous state in genes containing at least two rare variants (>20% of variation reads). Because the three investigated
Nature Genetics
© 2014 Nature America, Inc. All rights reserved.
individuals were all males from a branch of the pedigree with only affected male siblings, we also investigated possible hemizygous sequence variants (>80% of variation reads) on the X chromosome. Special attention was given to all rare variants, regardless of the percentage of variation reads, in genes that have intellectual disability described as a phenotypic feature in their OMIM clinical synopsis or OMIM clinical features (see URLs). Raw sequence reads of each variant were manually checked by use of the Integrative Genomics Viewer32. Candidate indel variations with a percentage of variant reads below 80% for the homozygous candidates and below 20% for the heterozygous candidates after manual inspection of the sequence reads were excluded from further analyses. Exome copy number analysis was performed by use of cn.MOPs 33. The sample was analyzed with a reference set of 30 exomes for comparison. The threshold for the detection of homozygous deletions was set at a z score of A mutation was observed in one individual in a heterozygous state. That individual is of European-American descent and has a diagnosis of refractory epilepsy. Analysis of TDP2 expression in human lymphoblastoid cells and tissue by qRT-PCR. RNA was isolated from case and control lymphoblastoid cells using the NucleoSpin RNA II kit (Macherey-Nagel) according to the manufacturer’s protocols. For first-strand synthesis, RNA was isolated from the appropriate cells (NucleoSpin RNA II kit), and 0.5 µg of total RNA was reverse transcribed using the iScript cDNA synthesis kit (Bio-Rad), according to the manufacturer’s instructions. The resulting cDNA was purified using the NucleoSpin extract II kit, and quantification by quantitative PCR was conducted in duplicate on the equivalent of 7.8 ng of total RNA from the first-strand synthesis. SYBR Green–based real-time quantitative PCR (qRT-PCR) expression analysis was performed on a 7500 Fast Real-Time PCR System (Applied Biosystems) using Power SYBR Green PCR Master Mix (Applied Biosystems) according to the manufacturer’s instructions. The TDP2 primers for qRT-PCR were designed by the Primer3 program34 (see URLs) and are described in Supplementary Table 5. The resulting PCR products encompassed the boundary between exons 5 and 6. GUSB was used as a reference gene35. Differences in TDP2 mRNA expression between case and control samples were calculated by the comparative Ct or 2∆∆Ct method36,37. To inhibit nonsense-mediated decay, cells were mock treated or treated with 100 µg/ml cycloheximide for 4 h before being snap frozen until used for RNA isolation. For TDP2 expression profiling in human tissue, total RNA from different human adult and fetal tissues was purchased from Stratagene Europe. All fetal tissues were from 20- or 21-week-old embryos, except for cochlear RNA, which was from 8-week-old embryo. We reverse transcribed 5 µg of total RNA from each tissue into cDNA as described above. qRT-PCR was performed in duplicate on the equivalent of 12.5 ng of total RNA input, and GUSB and PPIB were used as reference genes. Differences in expression of a gene of interest between two samples were calculated as described above. Analysis of TOP2/TDP2-dependent gene expression in LNCaP cells and mouse brain by qRT-PCR. RNA was extracted from pelleted LNCaP cells (obtained from the American Type Culture Collection (ATCC) and confirmed to be mycoplasma free) using the RNeasy kit (Qiagen) with an additional DNase step. Total RNA (1 µg) and oligo(dT) (0.16 µg; Ambion) were heated at 70 °C for 5 min, chilled on ice and reverse transcribed for 2 h at 42 °C. The cDNA was treated with RNase at 37 °C for 30 min and purified using a
Nature Genetics
PCR purification kit (Qiagen). Aliquots of 2.5 µl were employed in qRT-PCR (25-µl total volume). The expression data for each gene of interest were first normalized to the data for ACTB in the same experimental condition and then to the relevant gene of interest in the scrambled and uninduced control (i.e., not treated with siRNA and not induced with DHT). Primers are listed in Supplementary Table 6. Mouse embryos were collected by cesarean section at E16.5. Whole brains (olfactory bulb, forebrain, midbrain, hindbrain and brain stem) were homogenized in TRIzol (Sigma), chloroform extracted, isopropanol precipitated and purified by RNeasy kit. RNA from embryos of the appropriate genotypes (three pairs of wild-type and Tdp2∆1–3 mice; note that the absence of observable phenotypic differences between genotypes prevented selection bias) were processed as follows. Before cDNA synthesis, samples were treated with DNase I and further purified by RNeasy kit. cDNA was generated as described above and employed where indicated in qRT-PCR experiments to measure the expression of the indicated control and TOP2βdependent genes. For microarray analysis, conducted by UCL Genomics, UCL Institute of Child Health, total RNA was labeled with the Affymetrix GeneChip WT Sense Target Labeling and Control Reagents kit and hybridized to GeneChip Mouse Gene 1.0 ST arrays in an Affymetrix hybridization oven following the standard Affymetrix protocol. Arrays were washed and labeled using the Affymetrix hybridization wash and stain kit in a Fluidics station 450 before scanning on a 7G scanner. Microarray quality was assessed using the Affymetrix GeneChip Expression Console, and CEL files were transferred to Genespring (Agilent) for analysis of differential expression. The data (obtained from three independent pairs of wild-type and Tdp2∆1–3 brains) were normalized using RMA (Robust Multi-Array Average), and genes with expression changing by 1.5-fold between Tdp2+/+ and Tdp2∆1–3 were selected. Further pathway analysis was performed using Ingenuity Pathway Analysis (Qiagen), in which 156 genes were mapped. Mouse experiments were conducted under Home Office licence number PPL 70/7007. Cerebellar interneuron density Nissl staining of mouse brain sections. Mice were killed at 10 weeks, and brains were fixed in 4% neutral-buffered formalin overnight and rinsed in PBS. Sagittal sections were blinded and stained with Nissl, and interneurons in the molecular layer (5,625-µm2 region) of the cerebellum were imaged and counted by light microscopy. Cells and cell culture. Human B lymphocytes were immortalized by transformation with Epstein-Barr virus (EBV) according to established procedures38. EBV-transformed lymphoblastoid cells from cases and controls were grown at 37 °C and 7.5% CO2 in RPMI 1640 medium (Gibco) containing 10% FCS (Sigma), a 10 U/µl penicillin/10 µg/µl streptomycin (Gibco) solution at 1% and 1% GlutaMAX (Gibco). Cells were maintained at 37 °C and 5% CO2 in RPMI supplemented with penicillin, streptomycin and 10% FCS. LNCaP cells were propagated in RPMI medium supplemented with 10% FBS. For depletion of androgens, cells were washed with serum-free medium three times for 30 min and incubated in medium containing 5% charcoal-stripped FBS for at least 48 h before use. Primary mouse astrocytes were cultured at 37 °C in a humidified, low-oxygen (3%) incubator in DMEM F-12 Ham’s nutrient mixture F-12 (Gibco) supplemented with 10% FBS, 2 mM l-glutamine, 100 U/ml penicillin, 100 µg/ml streptomycin and 1× MEM nonessential amino acids (Gibco) with or without 20 ng/ml epidermal growth factor (EGF; Sigma). Astrocytes were isolated by the dissection of the cerebral cortex from wild-type and mutant littermate pups (P5). The cortex was mechanically homogenized, trypsinized and dissociated by passage through a 10-ml pipette. The cell suspension was transferred to a fresh T-25 flask for 1 h to allow fibroblasts to attach, and the supernatant was transferred to a fresh T-25 flask for 48 h to allow astrocytes to attach. Astrocyte cultures were incubated until confluency was achieved, with the medium changed every 48 h. For post-mitotic astrocytes, cells were trypsinized, counted and seeded at 2.5 × 104 cells/cm2 onto coverslips and cultured in the presence of EGF until confluent. Cells were then incubated for a further 72 h in the absence of EGF. Absence of fibroblast contamination was confirmed by immunofluorescence with antibody to GFAP (Sigma, G3893) and post-mitotic arrest by EdU (Invitrogen) pulse labeling. Granule cerebellar neurons were cultured at 37 °C and 7.5% CO2 (3%) in Neurobasal-A medium (Gibco), a 10 U/µl penicillin/10 µg/µl streptomycin (Gibco) solution to 1%,
doi:10.1038/ng.2929
1% GlutaMAX (Gibco), 250 µM KCl and 1× B27 supplement (Gibco). Cerebellar granule neurons were isolated using the Worthington Papain Dissociation System as previously described39. Briefly, three to four cerebellums were isolated from P4–P6 littermates and, after the removal of meninges and choroid plexus, mechanically dissociated in cold HBSS-glucose and digested in a Papain-DNase solution. After the removal of large pieces and membrane fragments, cells were filtered through a 70-µm nylon mesh and centrifuged in a 35–60% Percoll gradient. Granule cells at the Percoll interphase were isolated, washed once with HBSS-glucose and preplated twice for 20 min to remove non-neural cells. Finally, cells were plated at high density in coverslips coated with poly-D-lysine. The medium was changed 24 h after plating and every 48 h until use. All experiments were performed between 4 and 10 d after isolation. The purity of the neuron culture was confirmed by immunofluorescence with antibody to p27 (Abcam, ab3928).
© 2014 Nature America, Inc. All rights reserved.
Cell survival and growth curves. Lymphoblastoid cells were seeded at 2 × 105 cells/ml in medium containing DMSO vehicle, etoposide or MMS at the indicated concentrations and counted at 48-h intervals thereafter. Cell death was quantified as the fold increase in dead cells in lymphoblast cultures at the indicated time points and was calculated by flow cytometry after differential cell staining with Sytox, as recommended by the manufacturer (Invitrogen). Tyrosyl DNA phosphodiesterase activity. [32P]-labeled single- or doublestranded oligonucleotide substrates harboring a single 3′-phosphotyrosyl or 5′-phosphotyrosyl terminus were generated as previously described1. Reactions contained 0.5 µM oligonucleotide substrate, 40–400 µM unlabeled competitor single-stranded oligonucleotide and the indicated amount of protein extract in a total volume of 6 µl of reaction buffer (50 mM Tris-HCl, pH 7.5, 50 mM KCl, 1 mM MgCl2, 1 mM DTT, 100 µg/ml BSA). Reactions were conducted at 37 °C for 2 h unless indicated otherwise, and reaction products were resolved by denaturing PAGE and analyzed by phosphorimaging. Protein extracts from brain and total blood were prepared by mild sonication in 4 volumes of lysis buffer (40 mM Tris-HCl, pH 7.5, 100 mM NaCl, 0.1% Tween-20, 1 mM DTT) supplemented with 1 mM PMSF (phenylmethanesulfonylfluoride) and protease inhibitor cocktail (Sigma) and clarification by centrifugation for 10 min (16,500g at 4 °C). White blood cells were isolated from the mononuclear cell layer of a 1:1:0.6 blood:PBS:Histopaque-1077 gradient and lysed in 100 µl of lysis buffer for every 0.5 × 106 cells. Antibodies, pulse labeling and immunofluorescence. Cells were treated as indicated and fixed for 10 min in 4% paraformaldehyde in PBS, permeabilized for 2 min in 0.2% Triton X-100 in PBS, blocked for 30 min in 5% BSA in PBS and incubated with the indicated primary antibody for 1–3 h in 1% BSA in PBS. Primary antibodies for immunoblotting and/or immuno fluorescence were to γH2AX (Millipore, 05-636; 1:1,000 dilution), cyclin A (Santa Cruz Biotechnology, sc-751; 1:500 dilution), CENPF (Sigma, ab5; 1:500 dilution), GFAP (Sigma; 1:500 dilution), TDP1 (Abcam, ab4166; 1:1,000 dilution), TDP2 (ref. 40), TOP2α (Santa Cruz Biotechnology, sc-5348; 1:1,000 dilution), TOP2β (Santa Cruz Biotechnology, sc-13059; 1:1,000 dilution), XRCC1 (Mab, 33-2-5; 1:100 dilution of hybridoma supernatant) and actin (Sigma, a-4600; 1:2,500 dilution). Cells were then rinsed (3 × 5 min) in 0.1% Tween-20 in PBS and incubated for 30 min with the appropriate AlexaFluor-conjugated secondary antibody (1:1,000 dilution in 1% BSA in PBS). Cells were counterstained with DAPI (4′,6-diamidino-2-phenylindole; Sigma) and mounted in Vectashield (Vector Labs). γH2AX foci were counted (double blind) in 40 cells from each experimental sample. To identify cells in G1, cells were immunostained with antibody to CENPF or cyclin A, and cells negative for these factors were scored. Lymphoblastoid cells were rinsed once in PBS, resuspended in 0.03 M sodium citrate and cytospun for 3 min at 800g. For pulse labeling with EU, cells were incubated with 1 mM EU (Invitrogen) for 45 min (astrocytes) or 30 min (granule cerebellar neurons) and fixed, and coverslips were incubated for 30 min at room temperature with 1 µM AlexaFluor-conjugated azide (Invitrogen) in 100 mM Tris-HCl, pH 8.5, 1 mM CuSO4 and 100 mM ascorbic acid. Mean fluorescence was quantified by Simple PCI software from n ≥ 30 cells for each experiment.
doi:10.1038/ng.2929
Transfection of LNCaP cells. LNCaP cells were transfected with non-targeting scrambled siRNA or TDP2 siRNA (targeting 5′-GUACAGCCCAGAUGUGAUA-3′) using Oligofectamine (Invitrogen). Cells were transfected twice over two consecutive days and were used for experiments 48 h after the second transfection. Before the second transfection, cells were washed in serum-free medium for 2 h and subjected to androgen deprivation as described above. For short hairpin RNA (shRNA), LNCaP cells were transfected (Amaxa Kit R, Lonza) with a plasmid vector encoding G418r (pCD2E) and either empty pSUPER or pSUPER-TDP2 (ref. 1) and selected in G418 for 5 d before androgen deprivation. Chromatin immunoprecipitation. Cells were cross-linked with 1% formaldehyde at 37 °C for 10 min, and cross-linking was stopped by the addition of glycine to 125 mM for 5 min at room temperature. Cells were recovered by scraping, rinsed in PBS and pelleted. Cell pellets were resuspended in 0.3 ml of lysis buffer (1% SDS, 10 mM EDTA, 50 mM Tris-HCl, pH 8, 2 mM DTT, 50 µg/ml PMSF, 1× protease inhibitor cocktail (Sigma, P8340)) and sonicated in a Bioruptor at maximum intensity for 30 min in 30-s cycles (30-s off, 30-s on) followed by clarification by centrifugation. Supernatants were collected, 1/30 volume was reserved as whole-cell extract, and the remainder was diluted in 1% Triton X-100, 200 mM NaCl, 10 mM Tris-HCl, pH 8, 50 µg/ml PMSF and 1× protease inhibitor cocktail. Diluted extract was precleared with magnetic beads (Dynabeads, Invitrogen) presaturated with sheared salmon sperm DNA for 1 h at 4 °C. RNA polymerase II immunoprecipitation was performed overnight at 4 °C with 8WG16 antibody (Santa Cruz Biotechnology). After immunoprecipitation with 1 µg of antibody, protein G magnetic beads (presaturated with sheared salmon sperm DNA) were added, and incubation was continued for another 2 h. Beads were washed in NaCl buffer (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris-HCl, pH 8, 500 mM NaCl) followed by LiCl buffer (1% deoxycholate, 1% NP-40, 1 mM EDTA, 10 mM Tris-HCl, pH 8, 250 mM LiCl). Beads were then washed two times with TE buffer, and bound material was eluted with 2% SDS in 1× TE at 65 °C for 20 min. Eluates were heated at 65 °C for >6 h to reverse cross-links and treated with 100 µg of proteinase K at 50 °C for 1 h. DNA fragments were purified (QIAquick PCR purification kit, Qiagen) and eluted in 100 µl of milliQ water. For PCR, 2.5 µl of immunoprecipitated material or whole-cell extract (1:100 dilution) was employed per 25-µl reaction. Primers are listed in Supplementary Table 6.
28. Purcell, S. et al. PLINK: a tool set for whole-genome association and populationbased linkage analyses. Am. J. Hum. Genet. 81, 559–575 (2007). 29. Pelak, K. et al. The characterization of twenty sequenced human genomes. PLoS Genet. 6, e1001111 (2010). 30. Li, H. et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics 25, 2078–2079 (2009). 31. Ge, D. et al. SVA: software for annotating and visualizing sequenced human genomes. Bioinformatics 27, 1998–2000 (2011). 32. Robinson, J.T. et al. Integrative genomics viewer. Nat. Biotechnol. 29, 24–26 (2011). 33. Klambauer, G. et al. cn.MOPS: mixture of Poissons for discovering copy number variations in next-generation sequencing data with a low false discovery rate. Nucleic Acids Res. 40, e69 (2012). 34. Rozen, S. & Skaletsky, H. Primer3 on the WWW for general users and for biologist programmers. Methods Mol. Biol. 132, 365–386 (2000). 35. de Brouwer, A.P.M., van Bokhoven, H. & Kremer, H. Comparison of 12 reference genes for normalization of gene expression levels in Epstein-Barr virus–transformed lymphoblastoid cell lines and fibroblasts. Mol. Diagn. Ther. 10, 197–204 (2006). 36. Livak, K.J. & Schmittgen, T.D. Analysis of relative gene expression data using realtime quantitative PCR and the 2(−∆∆CT) method. Methods 25, 402–408 (2001). 37. Pfaffl, M.W. A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res. 29, e45 (2001). 38. Wall, F.E., Henkel, R.D., Stern, M.P., Jenson, H.B. & Moyer, M.P. An efficient method for routine Epstein-Barr virus immortalization of human B lymphocytes. In Vitro Cell. Dev. Biol. Anim. 31, 156–159 (1995). 39. Lee, H.-Y., Greene, L.A., Mason, C.A. & Manzini, M.C. Isolation and culture of post-natal mouse cerebellar granule neuron progenitor cells and neurons. J. Vis. Exp. 23, e990 (2009). 40. Thomson, G. et al. Generation of assays and antibodies to facilitate the study of human 5′-tyrosyl DNA phosphodiesterase. Anal. Biochem. 436, 145–150 (2013).
Nature Genetics