Structural variants create new topological-associated domains and ectopic retinal enhancer-gene contact in dominant retinitis pigmentosa

ARTICLE S uc u al Va ian s C ea e New Topological-Associa ed Domains and Ec opic Re inal Enhance -Gene Con ac in Dominan Re ini is Pigmen osa Suzanne E. de B uijn, 1,2,26 Alessia Fio en ino, 3,4,5,26 Daniele O a iani, 3 S ephanie Fanucchi, 6 Ui a ´S. Melo, 7,8 Julio C. Co al-Se ano, 3 Timo Mulde s, 2,9 Michalis Geo giou, 3,10 Ca lo Ri ol a, 11,12,13 Nikolas Pon ikos, 3,4,5 Ga in A no, 3,4,5,10 Lisa Robe s, 14 Jacquie G eenbe g, 14 Sil ia Albe , 1 Ch is ian Gilissen, 1 Ma co Aben, 1 Geo ge Rebello, 14 Simon Mead, 15 F. Lucy Raymond, 16,17 Jo di Co ominas, 1 Clai e E.L. Smi h, 18 Hannie K eme , 1,2,19 Susan Downes, 4,5,20 G aeme C. Black, 4,5,21 And ew R. Webs e , 3,4,5,10 Ch is F. Inglehea n, 4,5,18 L. Ingebo gh an den Bo n, 22 Robe K. Koenekoop, 23 Michel Michaelides, 3,4,5,10 Raj S. Ramesa , 14 Ca el B. Hoyng, 2,9 S e an Mundlos, 7,8 Musa M. Mhlanga, 1,6,24,25 F ans P.M. C eme s, 1,2 Michael E. Chee ham, 3,4,5 Susanne Roosing, 1,2,27, *and Alison J. Ha dcas le 3,4,5,27 Summa y The cause o au osomal-dominan e ini is pigmen osa (adRP), which leads o loss o ision and blindness, was in es iga ed in amilies lacking a molecula diagnosis. A e ined locus o adRP on Ch 17q22 (RP17) was delinea ed h ough geno yping and genome sequencing, leading o he iden i ica ion o s uc u al a ian s (SVs) ha seg ega e wi h disease. Eigh di e en complex SVs we e cha - ac e ized in 22 adRP-a ec ed amilies wi h >300 a ec ed indi iduals. All RP17 SVs had b eakpoin s wi hin a genomic egion spanning YPEL2 o LINC01476. To in es iga e he mechanism o disease, we ep og ammed ib oblas s om a ec ed indi iduals and con ols in o induced plu ipo en s em cells (iPSCs) and di e en ia ed hem in o pho o ecep o p ecu so cells (PPCs) o e inal o ganoids (ROs). Hi-C was pe o med on ROs, and di e en ial exp ession o egional genes and a e inal enhance RNA a his locus was assessed by qPCR. The epigene ic landscape o he egion, and Hi-C RO da a, showed ha YPEL2 si s wi hin i s own opologically associa ing domain (TAD), ich in enhance s wi h binding si es o e inal ansc ip ion ac o s. The Hi-C map o RP17 ROs e ealed c ea ion o a neo-TAD wi h ec opic con ac s be ween GDPD1 and e inal enhance s, and modeling o all RP17 SVs was consis en wi h neo-TADs leading o ec opic e inal-speci ic enhance -GDPD1 accessibili y. qPCR con i med inc eased exp ession o GDPD1 and inc eased exp ession o he e inal enhance ha en e s he neo-TAD. Al e ed TAD s uc u e esul ing in inc eased e inal exp ession o GDPD1 is he likely con e gen mechanism o disease, consis en wi h a dominan gain o unc ion. Ou s udy highligh s he impo ance o SVs as a genomic mecha- nism in unsol ed Mendelian diseases. In oduc ion Despi e ecen ad ances in nex -gene a ion sequencing, app oxima ely 30%–40% o indi iduals wi h inhe i ed e inal diseases (IRDs) lack a molecula diagnosis. This is p obably due o a combina ion o a e no el disease genes, which equi e la ge coho s o alida ion, and p e iously in ac able mu a ion classes, such as in onic a ian s, s uc u al a ian s (SVs), and a ian s in egula o y e- gions. 1,2 1 Depa men o Human Gene ics, Radboud Uni e si y Medical Cen e , Nijmegen, 6500 HB, he Ne he lands; 2 Donde s Ins i u e o B ain Cogni ion and Beha iou , Radboud Uni e si y Medical Cen e , Nijmegen, 6500 HB, he Ne he lands; 3 UCL Ins i u e o Oph halmology, London, EC1V 9EL, UK; 4 UK In- he i ed Re inal Disease Conso ium; 5 Genomics England Clinical In e p e a ion Pa ne ship; 6 Gene Exp ession and Biophysics G oup, Di ision o Chem- ical, Sys ems and Syn he ic Biology, Depa men o In eg a i e Biomedical Science, Ins i u e o In ec ious Disease & Molecula Medicine, Facul y o Heal h Sciences, Uni e si y o Cape Town, Cape Town, 7935, Sou h A ica; 7 Max Planck Ins i u e o Molecula Gene ics, RG De elopmen & Disease, Be lin, 14195, Ge many; 8 Ins i u e o Medical and Human Gene ics, Cha i e ´– Uni e si a ¨ smedizin, Be lin, 10117, Ge many; 9 Depa men o Oph halmology, Radboud Uni e si y Medical Cen e , Nijmegen, 6500 HB, he Ne he lands; 10 Moo ields Eye Hospi al, London, EC1V 2PD, UK; 11 Depa men o Gene ics and Genome Biology, Uni e si y o Leices e , Leices e , LE1 7RH, UK; 12 Clinical Resea ch Cen e , Ins i u e o Molecula and Clinical Oph halmology Basel (IOB), Basel, 4031, Swi ze land; 13 Depa men o Oph halmology, Uni e si y Hospi al Basel, Basel, 4001, Swi ze land; 14 Uni e si y o Cape Town/MRC Genomic and P ecision Medicine Resea ch Uni , Di ision o Human Gene ics, Depa men o Pa hology, Ins i u e o In ec ious Disease and Molecula Med- icine, Facul y o Heal h Sciences, Uni e si y o Cape Town, Cape Town, 7935, Sou h A ica; 15 MRC P ion Uni a UCL, UCL Ins i u e o P ion Disease, Lon- don, W1W 7FF, UK; 16 NIHR BioResou ce, Camb idge Uni e si y Hospi als, Camb idge, CB2 0QQ, UK; 17 Depa men o Medical Gene ics, Camb idge Ins i- u e o Medical Resea ch, Uni e si y o Camb idge, Camb idge, CB2 OXY, UK; 18 Di ision o Molecula Medicine, Leeds Ins i u e o Medical Resea ch, Uni e si y o Leeds, Leeds, LS2 9JT, UK; 19 Depa men o O o hinola yngology, Radboud Uni e si y Medical Cen e , Nijmegen, 6500 HB, he Ne he lands; 20 Ox o d Eye Hospi al, Ox o d Uni e si y Hospi als NHS T us and Nu ield Labo a o y o Oph halmology, Uni e si y o Ox o d, Ox o d, OX3 9DU, UK; 21 Manches e Cen e o Genomic Medicine, S . Ma y’s Hospi al, Manches e , M13 9WL, UK; 22 The Ro e dam Eye Hospi al, Ro e dam, 3011 BH, he Ne he lands; 23 Depa men o Paedia ic Su ge y, Human Gene ics and Oph halmology, McGill Uni e si y, Mon e ´al, QC H4A 3J1, Canada; 24 Gene Exp es- sion and Biophysics Uni , Ins i u o de Medicina Molecula , Faculdade de Medicina Uni e sidade de Lisboa, Lisbon, 1649-028, Po ugal; 25 Epigenomics & Single Cell Biophysics G oup, Radboud Ins i u e o Molecula Li e Sciences (RIMLS), Radboud Uni e si y, Nijmegen, 6525 GA, he Ne he lands 26 These au ho s con ibu ed equally 27 These au ho s con ibu ed equally *Co espondence: susanne. oosi[email p o ec ed] h ps://doi.o g/10.1016/j.ajhg.2020.09.002. 802 The Ame ican Jou nal o Human Gene ics 107, 802–814, No embe 5, 2020 Ó2020 The Au ho s. This is an open access a icle unde he CC BY license (h p://c ea i ecommons.o g/licenses/by/4.0/). The mos common o m o IRD is e ini is pigmen osa (RP [MIM: 268000]), which is gene ically he e ogeneous, wi h a p e alence o 1 in 4,000. 3 RP is de ined as a e inal degene a ion ha p ima ily a ec s od pho o e- cep o s, esul ing in nigh blindness and p og essi e loss o pe iphe al ision, o en p og essing in o he cen al e ina and a ec ing cone pho o ecep o s, leading o se e e isual impai men o blindness (see ‘‘Nonsynd omic Re ini is Pigmen osa O e iew’’ in Web Resou ces). Au osomal-dominan RP (adRP) accoun s o 25%–40% o cases, depending on he popula ion s udied, and has been associa ed wi h mu a ions in 30 genes, including CA4 (MIM: 114760) on Ch 17q23.1 (RP17 [MIM: 600852])(see ‘‘Re Ne ’’ in Web Resou ces). 4,5 Following ini ial publica ions de ining his locus 6,7 a a ian inCA4 wasimplica edas hecause o adRP in amilies o Sou h A ican o igin, howe e pa hogenici y o he epo ed a ian has been ques- ioned because i has a popula ion equency o 4% in heal hy con ols in no he n Sweden. 8–10 Subsequen ly epo ed CA4 a ian s in indi iduals wi h RP we e iden- i ied by a ge ed Sange sequencing and do no ully exclude a ian s in o he genes as a cause o disease (Table S1). We in es iga ed he cause o adRP in unsol ed amilies, including he i s pedig ee (GC1, e e ed o as UK1) d awn up a Moo ields Eye Hospi al o e 35 yea s ago and he o iginal Du ch amily (W97-079, e e ed o as NL1) ha showed linkage o he RP17 locus bu lacked a mu a ion in CA4. 7 He e, we epo iden i ica ion and cha ac e iza ion o complex SVs on Ch 17q22, h ough whole-genome sequencing (WGS), as he genomic cause o adRP a he RP17 locus in a la ge numbe o amilies, including he amilies o Sou h A ican o igin. To explo e a con e gen mechanism o disease, we in es iga ed he e ec o RP17 SVs on h ee-dimensional (3D) ch oma in o ganiza ion ha esul s in he compa - men aliza ion o he genome in o opologically associ- a ing domains (TADs) and he epigene ic landscape o he egion. TADs a e ch oma in domains wi hin he genome ha acili a e enhance p omo e con ac s wi hin he nuclea 3D space. 11 Dis up ion o TAD s uc u es can lead o loss o ch omosomal con ac be ween egula o y egions and hei a ge genes o he o ma ion o no el ac i e domains wi h ec opic con ac s occu ing be ween egula o y egions and a new a ge gene, esul ing in pa hogenic al e a ions in gene exp ession. 12–15 We demons a e ha al e ed TAD s uc u e a he RP17 locus leads o ec opic e inal enhance -gene in e ac ions, consis en wi h a dominan gain o unc ion. Ou s udy highligh s he pa hogenici y o SVs ha al e 3D ch oma in o ganiza- ion and gene exp ession by ea anging TAD s uc u es and he need o e isi a e Mendelian diseases o which genes and a ian s ha e no been subs an ia ed in o he coho s. Ma e ial and Me hods S udy Coho The s udy was app o ed by he medical e hics commi ee o he E asmusMC Ro e dam, Radboudumc Nijmegen, and Moo ields Eye Hospi al and was pe o med in acco dance wi h he p inciples o he Wo ld Medical Associa ion Decla a ion o Helsinki. In o med consen was ob ained om all pa icipan s o hei legal ep esen a i es. Gene ic Analyses We pe o med SNP geno yping o index amilies NL1 and UK1 o de ine and e ine he RP17 locus. Genomic DNA om a ec ed indi iduals and hei amily membe s was analyzed by whole- exome sequencing (WES) and WGS. Sequence da a was aligned o he Human Re e ence Genome build hg19. Va ian s we e p io i ized on he basis o a mino allele equency (MAF) % 0.0001 in gnomAD. SVs we e called wi h ExomeDep h, Man a S uc u al Va ian Calle , Can as Copy Numbe Va ian Calle , and Con ol-FREEC. De ails o geno yping, sequencing, and analysis pipelines a e p o ided in he Supplemen al Ma e ial and Me hods. Cha ac e iza ion and Valida ion o S uc u al Va ian s SV b eakpoin junc ions we e PCR ampli ied and alida ed wi h Sange sequencing. P ime sequences and coo dina es a e lis ed in Table S2. SV b eakpoin egions we e assessed o he p esence o mic ohomology and epe i i e elemen s. To alida e a ipli- ca ed egion o UK-SV6, we pe o med quan i a i e eal- ime PCR (qPCR) on genomic DNA om a ec ed indi iduals om am- ily UK13 and una ec ed con ols (Supplemen al Ma e ial and Me hods). Clinical Analysis A ailable clinical no es o cases o he pedig ees iden i ied a Rad- boudumc, Moo ields Eye Hospi al, Uni e si y o Cape Town, and McGill Uni e si y Heal h Cen e we e e iewed, as well as de ailed e inal imaging, undus au o luo escence, and op ical cohe ence omog aphy. Age o onse is de ined as he age a which symp oms we e i s expe ienced. In e oga ion o he Genomic Region We in e oga ed ch oma in and genome egula ion da ase s o explo e he epigenomic landscape o he egion. A ailable da ase s we e ob ained and analyzed ia he UCSC genome b owse (de ails o da ase s used a e p o ided in Supplemen al Ma e ial and Me hods). Rep og amming Fib oblas s in o iPSCs and Di e en ia ion in o Pho o ecep o P ogeni o Cells and 3D Re inal O ganoids Fib oblas s we e cul u ed om skin biopsies o wo indi iduals wi h NL-SV1, one indi idual wi h UK-SV2, and i e anonymous con ol indi iduals. Cell lines we e ep og ammed in o induced plu ipo en s em cells (iPSCs) and di e en ia ed in o pho o e- cep o p ogeni o cells (PPCs) ollowing he p e iously desc ibed 60 day p o ocol (Supplemen al Ma e ial and Me hods). 16,17 3D e inal o ganoids (ROs) we e di e en ia ed o UK-SV2 and con ols, as p e iously desc ibed (Supplemen al Ma e ial and Me hods). 18 The Ame ican Jou nal o Human Gene ics 107, 802–814, No embe 5, 2020 803 P epa a ion o Low Inpu Hi-C Lib a ies (Low-C) Hi-C was pe o med on UK-SV2 and con ol 3D ROs ia a low inpu p o ocol (Low-C) wi h ew modi ica ions (Supplemen al Ma e ial and Me hods). 19 Two lib a ies pe sample we e sequenced o 200 million agmen s in a 100 bp pai ed-end un on a No a- Seq 6000 (Illumina). Pai ed-end sequencing da a was p ocessed ia Juice 20 and he Hi-C maps we e c ea ed wi h a bin size wi h 10 kb esolu ion. Fu he in o ma ion abou he bioin o ma ics pipeline is de ailed in Melo e al., 2020. 21 Exp ession Analysis o Genes and Enhance RNA wi hin he RP17 Locus To assess exp ession o genes, we pe o med qPCR o di e en hu- man issues, including e ina (Table S3 and Supplemen al Ma e ial and Me hods). Single-cell RNA sequencing da a o human 22 and p ima e 23 e inal cell ypes we e ob ained and isualized ia he B oad Ins i u e Single Cell Po al (Supplemen al Ma e ial and Me hods). cDNA was syn hesized om o al RNA ex ac ed om PPCs, ROs, and ib oblas s. Di e en ial exp ession o genes implica ed in he SVs, and con ol housekeeping and e inal p ogeni o genes, was assessed by qPCR (Table S3 and Supplemen al Ma e ial and Me hods). We designed p ime s o he enhance egion con- aining mul iple e inal ansc ip ion ac o binding si es impli- ca ed in all SVs o analyze a ge ed enhance RNA exp ession by qPCR (Table S3 and Supplemen al Ma e ial and Me hods). Resul s Re inemen o he RP17 Locus in Two Un ela ed adRP- A ec ed Families The a ec ed haplo ype o a Du ch adRP-a ec ed amily (NL1) (Figu e 1A) was p e iously mapped o a 7.18 Mb egion spanning he RP17 locus on ch omosome 17. 7 The RP17 locus was e ined o a 5.16 Mb in e al by SNP haplo yping in an ex ended pedig ee (Figu e 1D and Supplemen al In o ma ion). No a e coding o splice si e he e ozygous a ian s (MAF %0.0001) sha ed be ween a ec ed indi iduals we e ound h ough WES. Subse- quen ly, WGS was pe o med, and simila ly, no a e candida e coding, splice si e, in onic, o in e genic he e o- zygous single-nucleo ide a ian s we e iden i ied (Table S4 and Supplemen al In o ma ion). In pa allel, WES and WGS we e pe o med o a ec ed indi iduals om a gene ically unexplained UK adRP- a ec ed amily (UK1) (Figu e 1B). This also ailed o iden- i y a a e causa i e a ian ; howe e , a disease-associa ed haplo ype on ch omosome 17 was iden i ied (Figu e 1E, Table S5, and Supplemen al In o ma ion). In e oga ion o unsol ed IRD sequence da a gene a ed h ough he UK IRDC, UCL-Ex, NIHR-Bio esou ce, and Genomics England iden i ied o he adRP p obands ha sha ed he same haplo ype o Ch 17 SNVs and es ablished his as a ounde haplo ype in ele en addi ional UK adRP-a ec ed amilies (Figu e 1C). The adRP locus was e ined o a 4.4 Mb in e al on Ch 17q22 (Figu e 1E). This genomic in e al o e laps he p e iously desc ibed RP17 locus in amilies o Du ch and Sou h A ican o igin (Figu e 1F). A missense a ian in CA4 [c.40C>T (p.A g14T p); GenBank: NM_000717.4] was p e iously desc ibed as he cause o adRP a he RP17 locus in amilies o Sou h A ican o igin. 8 No a e coding, in onic, o ups eam a ian s in CA4 we e iden i ied in he Du ch and UK amilies. Iden i ica ion o S uc u al Va ian s wi hin he RP17 Locus Nex , we analyzed genome and exome da a o copy num- be a ian s and SVs (Supplemen al In o ma ion). In amily NL1, WGS e ealed a 226 kb duplica ion wi hin he RP17 locus: ch 17: 57,291,905_57,518,137dup (NL-SV1). This SV in ol es wo duplica ed genes (GDPD1 [MIM: 616317] and YPEL2 [MIM: 609723]), an in agenic mic oRNA (MIR4729), and pa ial duplica ion o SMG8 (MIM: 613175) and he long non-coding RNA LINC01476. The duplica ion c ea es a b eakpoin junc ion (ch 17: g.57,518,137–57,291,905) speci ic o he mu a ed allele in NL1 (Figu es 2A, 2B, and S1), which was used o con i m seg ega ion o he SV wi h he adRP pheno ype in his am- ily. No o e lapping SVs in he RP17 locus we e obse ed in WES o ~7,500 indi iduals wi hou e inal disease gene - a ed in-house a he Depa men o Human Gene ics, Radboudumc. Fo he wel e UK RP17 ounde haplo ype am- ilies, WGS e ealed a duplica ed in e sion: ch 17: 57,456,098–57,468,960delins57,275,839_57,559,114in (UK-SV2) (Figu e 2B). The SV was cha ac e ized, and b eakpoin junc ions we e alida ed (Figu e S1 and Sup- plemen al In o ma ion). This SV in ol ed ou coding genes (PRR11 [MIM: 615920], SMG8,GDPD1,and YPEL2) and wo non-coding RNA genes (MIR4729 and LINC01476)(Figu e 2B). UK-SV2 seg ega ed wi h adRP in all amilies o which DNA was a ailable o analysis. UK-SV2 was absen in WGS con ol genome da a gene - a ed o 58,000 UK indi iduals (Genomics England). Di e en S uc u al Va ian s wi hin he RP17 Locus in Mul iple adRP-A ec ed Families These da a p omp ed us o in es iga e whe he SVs we e p esen in he wo o iginal Sou h A ican amilies (SA1 and SA2) ha we e linked o he RP17 locus (Figu e S2A). 6,8 In addi ion, a Canadian adRP-a ec ed am- ily (CA1) was also mapped o he RP17 locus (unpublished da a, Figu e S2B). WGS was pe o med o a ec ed indi id- uals om hese amilies, and in e sion duplica ion e en s we e iden i ied in all samples analyzed (Figu e 2C). In SA1 and SA2, an iden ical SV, SA-SV3, was e ealed, sugges ing his is a ounde a ian in his popula ion. SA-SV3 was also ound by b eakpoin PCR in wo addi ional amilies o Sou h A ican o igin (SA3 and SA4), con i ming he ounde e ec (Figu e S2A). In he Canadian amily, a di e en in e sion duplica ion e en was iden i ied, CA- SV4. SA-SV3 and CA-SV4 b eakpoin s we e cha ac e ized and alida ed (Figu e S1), and seg ega ion o he SVs wi h he adRP pheno ype was con i med. 804 The Ame ican Jou nal o Human Gene ics 107, 802–814, No embe 5, 2020 Ou da a sugges ed ha SVs a he RP17 locus a e an impo an cause o adRP. The e o e, WGS and WES da a o gene ically unexplained adRP-a ec ed amilies we e analyzed o SVs wi hin his locus. In ou un ela ed am- ilies o Du ch o UK o igin, ou addi ional unique com- plex SVs we e disco e ed (Figu es 2D and S2C). Fo Figu e 1. Mapping o he RP17 Locus in Two Un ela ed Families (A) Pedig ee o Du ch NL1 amily. (B) Pedig ee o UK1 amily. (C) Pedig ees o addi ional UK amilies wi h he ounde haplo ype on Ch 17q. WGS o WES was pe o med in indi iduals highligh ed in blue o ed, espec i ely. (D) SNP haplo yping esul s o NL1. The e ined RP17 locus ( s8078110– s9910672) is sha ed by all a ec ed indi iduals (n ¼35) and no p esen in una ec ed indi iduals (n ¼28, only indi iduals wi h ecombina ion close o o e ining he c i ical egion a e depic ed) wi h a maximum LOD sco e o 15.0. The ho izon al numbe s ep esen he numbe o indi iduals wi h his haplo ype. (E) UK ounde haplo ype e ining he RP17 locus in UK amilies. Rep esen a i e haplo ypes om se e al un ela ed amilies a e shown wi h a ec ed (a ) indi iduals compa ed o an una ec ed (una ) indi idual. Black lines and a ows indica e ecombina ion e en s. Sha ed haplo ype in indi iduals is shaded ed. (F) O e lap o e ined RP17 loci in UK, NL, and p e iously desc ibed SA amilies. 8 The Ame ican Jou nal o Human Gene ics 107, 802–814, No embe 5, 2020 805 PRR11 SMG8 GDPD1 YPEL2 LINC01476 NL-SV1 AB SMG8 GDPD1 YPEL2 LINC01476 BC PRR11 SMG8 GDPD1 YPEL2 UK-SV2 DE SMG8GDPD1YPEL2LINC01476 PRR11 GF E YPEL2 LINC01476 GH PRR11 SMG8 GDPD1 YPEL2 LINC01476 SA-SV3 IJ KL PRR11 SMG8 GDPD1 JM LINC01476 LINC01476 MNL CA-SV4 PRR11 OP LINC01476 YPEL2 GDPD1 SMG8 PRR11 RQP LINC01476 RS PRR11 SMG8 GDPD1 YPEL2 LINC01476 NL-SV5 TU PRR11 SMG8 GDPD1 YPEL2 U LINC01476 V PRR11 SMG8 GDPD1 YPEL2 LINC01476 UK-SV6 WX GDPD1 YPEL2 LINC01476 X GDPD1 YPEL2 X LINC01476 Y UK-SV7 UK-SV8 GDPD1 YPEL2 Z AA GDPD1YPEL2LINC01476DHX40 CLTC CLTC AC AB AA AC AD DHX40 LINC01476 YPEL2 B C D MIR4729 MIR4729 MIR4729 MIR4729 MIR4729 MIR4729 MIR4729 MIR4729 MIR4729MIR4729 MIR4729 MIR4729 APRR11 SMG8 GDPD1 YPEL2 LINC01476 WT MIR4729 CLTCDHX40 PRR11 SMG8 PRR11SMG8 MIR4729 M PRR11 SMG8 GDPD1 AE AF SMG8GDPD1YPEL2LINC01476 AH AG AF PRR11 YPEL2 LINC01476 AH AIMIR4729 MIR4729 E PRR11 SMG8 GDPD1 YPEL2 LINC01476 MIR4729 NL-SV1 UK-SV2 SA-SV3 CA-SV4 NL-SV5 DHX40 CLTC UK-SV6 UK-SV7 UK-SV8 Figu e 2. O e iew o S uc u al Va ian s wi hin he RP17 Locus in adRP-A ec ed Families B eakpoin s a e indica ed wi h dashed lines. Blue segmen s ep esen duplica ed o iplica ed egions, whe eas in e sions a e high- ligh ed in pu ple. (A) Wild- ype (WT) ch omosomal o ganiza ion. (B) S uc u al a ian s iden i ied in NL1 (NL-SV1) and UK ounde haplo ype amilies (UK-SV2). (C) S uc u al a ian s iden i ied in adRP-a ec ed amilies ha we e p e iously linked o he RP17 locus; SA-SV3 8 and CA-SV4 (unpub- lished da a). (legend con inued on nex page) 806 The Ame ican Jou nal o Human Gene ics 107, 802–814, No embe 5, 2020 indi iduals ha had only unde gone WES, we pe o med WGS o de e mine he b eakpoin junc ions and iden i y po en ial in e sions o o he SVs. In all amilies, b eak- poin s we e alida ed, and seg ega ion analysis was pe - o med whe e possible. T iplica ion o UK-SV6 was con i med by qPCR in amily UK13 (Figu e S3 and Supple- men al In o ma ion). De ails o all SVs iden i ied in his s udy a e shown in Table S6,Figu e 2, and Figu e S4, and an o e iew o SV- speci ic b eakpoin junc ions is shown in Figu e S1. All RP17 SVs sha e a common duplica ed (o iplica ed) e- gion o 11.5 kb and ha bo unique b eakpoin s dis up ing he genomic egion spanning YPEL2 o LINC01476 (ch 17: 57,499,214–57,510,765) (Figu e S4). We analyzed all b eak- poin junc ion sequences o in es iga e he po en ial mechanism(s) ha c ea ed RP17 SVs. No single mecha- nism could accoun o he RP17 SVs because a combina- ion o (mic o)homology-media ed epai and non-homol- ogous end joining e en s we e iden i ied (Table S7 and S8, Figu e S5, and Supplemen al In o ma ion). Consis en Au osomal-Dominan Re ini is Pigmen osa Pheno ype o RP17-A ec ed Families The SVs iden i ied we e ully pene an in all amilies. A ailable clinical da a a e p esen ed in Table S9. Twen y- ou a ec ed indi iduals om se en een pedig ees we e e alua ed. The e is signi ican co ela ion o pheno ype ac oss all geno ypes wi h ela i ely mild disease, dec eased isual acui y, isual ield cons ic ion, nyc alopia, and slow p og ession consis en wi h adRP. Many a ec ed indi id- uals ha e p ese ed cen al isual unc ion and acui y un il he 6 h –7 h decade. Fo eal spa ing and cys oid macula edema we e a common inding in indi iduals wi h UK- SV2. On he basis o a small numbe o a ec ed indi iduals (n ¼2), UK-SV6 (wi h a iplica ed SV) may be associa ed wi h an ea lie age o onse and mo e se e e pheno ype (Figu e S6). Topologically Associa ing Domain S uc u e and Epigene ic Landscape o he RP17 Genomic Region Allo heRP17SVslead odis up iono hegenomic e- gion spanning YPEL2 o LINC01476 (Figu e 2E). SVs ha in e e e wi h genome s uc u e can ha e dis inc e - ec s on gene egula ion depending on he ype and ex en o he SV and landscape o he genomic egion. 15 TADs a e sepa a ed by bounda ies, egions o low ch o- ma in in e ac ion ha insula e he egula o y ac i i ies o neighbo ing TADs. The ansc ip ion ac o CTCF (CCTC-binding ac o ) ypically binds in hese egions whe e i plays a pi o al ole in he main enance o bound- a ies. SVs can cause loss o unc ion by disconnec ing en- hance s om hei a ge genes; howe e , dis up ion o TAD s uc u es and bounda ies can also exe a gain-o - unc ion e ec . Dele ions, o example, can lead o he using o wo p e iously sepa a ed TADs (TAD- usion), in- e sions can esul in he exchange o egula o y ma e ial be ween TADs (TAD-shu ling), whe eas duplica ions can gi e ise o he gene a ion o no el domains, so-called neo-TADs. 12,13 In each case, SVs esul in he gene a ion o ec opic con ac s o enhance s wi h he p omo e s o no el a ge genes esul ing in abe an gene ac i a ion. The human limb mal o ma ions caused by SVs ha al e he CTCF-associa ed bounda y o he WNT6/IHH/ EPHA4/PAX3 locus a e a p ominen example. The SVs esul in ec opic in e ac ions be ween EPHA4 (MIM: 602188) limb enhance s and he neighbo ing de elop- men al genes ha a e no mally insula ed, d i ing ec opic exp ession in he limb. 14 Simila ly, he dele ion o a CTCF si e loca ed be ween he Xis (MIM: 314670) and Tsix (MIM: 300181) TADs on he X ch omosome esul ed in a no el domain by usion o he adjacen TADs ( used- TAD). 24 As a consequence, p e iously insula ed en- hance s ac i a ed genes in he adjacen TAD, leading o he dys egula ion o hese genes. Hi-C da a we e no a ailable o human e ina, and he e o e, we gene a ed Hi-C maps o con ol human 3D ROs o ob ain maps o he ch oma in o ganiza ion o ou egion o in e es . Hi-C e ealed a s uc u ed domain con aining YPEL2 (YPEL2 TAD) lanked by less s uc u ed neighbo ing domains (Figu e 3A). CTCF binding is p e- sen on bo h bounda ies (Figu e 3B) suppo ing he TAD s uc u e a his locus. CTCF ChIA-PET da a highligh ed in e ac ions be ween he CTCF binding si es a he 50 and he 30bounda y o he YPEL2 TAD (Figu e S7B). Assay o ansposase-accessible ch oma in using seqenc- ing (ATAC-seq) da a om human e ina show ha he ch oma in in he YPEL2 TAD is accessible, and H3K27Ac ChIP-seq da a e ealed ha he e a e se e al ac i e en- hance s loca ed wi hin he YPEL2 TAD ha a e expec ed o d i e YPEL2 exp ession in he e ina (Figu e 3B). 25 Impo an ly, he YPEL2 TAD ha bo s wo egions o ac i e enhance s wi h binding si es o ansc ip ion ac o (TFs) known o be equi ed o pho o ecep o unc ion, including NRL, CRX, and OTX2 (Figu e 3B). NRL is a TF ha is p e e en ially exp essed in od pho o ecep o s. These TF binding si es co ela ed wi h H3K27Ac and ATAC-seq peaks in e ina. The published GeneHance da a- se shows ha hese egula o y elemen s ha e in e ac ions wi h he YPEL2 p omo e (Figu e S7C). 26 Collec i ely, hese analyses e ealed ha YPEL2 is loca ed wi hin an ac i e compa men ha con ains e inal-speci ic enhance s (Figu e 3C). (D) S uc u al a ian s ound in a coho o unsol ed adRP-a ec ed amilies; NL-SV5, UK-SV6, UK-SV7, and UK-SV8. Le e s A–AI depic he genomic in e als o each SV used o analyze and anno a e SV b eakpoin s. (E) O e iew o all SV b eakpoin s iden i ied in he RP17 locus. An o e lapping genomic egion ha is duplica ed o iplica ed in all SVs was iden i ied (ch 17: 57,499,214–57,510,765) and is highligh ed by a ligh blue e ical ba . The size o DHX40 is educed, and CLTC is pa ially shown o he pu pose o his igu e. The Ame ican Jou nal o Human Gene ics 107, 802–814, No embe 5, 2020 807 Exp ession o YPEL2 and GDPD1 Exp ession o YPEL2 and GDPD1 was assessed by qPCR in mul iple heal hy human issues, including e ina (Figu e S8). YPEL2 is ubiqui ously exp essed in he is- sues s udied, including e ina, wi h highes ela i e exp ession in b ain. Single-cell e ina RNA-seq da ase s e ealed YPEL2 is exp essed a highe le els in od pho o- ecep o cells, which is he p ima y cell ype a ec ed in e ini is pigmen osa, compa ed o cone pho o ecep o s (Figu e S8). 23 GDPD1 is de ec ed a low exp ession in all issue ypes bu has highe exp ession in es is and he b ain. These da a suppo he hypo hesis ha YPEL2 exp ession is egula ed by e inal enhance s wi hin he YPEL2 TAD. RP17 SVs C ea e New Topologically Associa ing Domains and Ec opic Enhance -Gene In e ac ions Using he wild- ype e inal o ganoid Hi-C map, we modeled he TAD bounda ies, CTCF si e o ien a ion, and e inal TF binding si e posi ions o each unique RP17 SV LINC01476 55 _ 0 _ 200 _ 0 _ 0 _ -5 _ 55 _ 0 _ 55 _ 0 _ 55 _ 0 _ 55 _ 0 _ 55 _ 0 _ 10.0 20.0 0.0 CTCF Re ina-speci ic exp ession H3K27ac ATAC-seq NRL CRX OTX2 Re inal o ganoid Hi-C mapA BHuman e ina ChIP-seq and RNA-seq da a C LINC01476 B B YPEL2 TAD B Re inal enhance Bounda y SMG8 GDPD1 Scale ch 17: 200 kb hg19 57 , 300 , 000 57 , 400 , 000 57 , 500 , 000 57 , 600 , 000 57 , 700 , 000 SKA2 PRR11 S MG8 GDPD1 YPEL2 MIR4729 DHX40 CLTC PTRH2 VMP1 Figu e 3. YPEL2 is Loca ed wi hin a S uc u ed Ac i e Compa men ha Con ains Re inal-Speci ic Enhance s (A) The TAD landscape o he genomic egion dis up ed by he RP17 SVs. Hi-C map o con ol e inal o ganoids e ealed a s uc u ed domain con aining YPEL2. (B) YPEL2 TAD bounda ies co espond wi h CTCF si es iden i ied in human e ina. Analysis o RNA-seq and assay o ansposase- accessible ch oma in using seqencing (ATAC-seq) da a ac oss he YPEL2 egion shows YPEL2 e inal exp ession and an accessible ch o- ma in con igu a ion. Analysis o H3K27Ac ChIP-seq da a in he same egion e ealed se e al ac i e enhance s loca ed wi hin he YPEL2 TAD, which a e en iched o e inal ansc ip ion ac o binding si es, including NRL, CRX, and OTX2. 25 These enhance s we e loca ed 50 o he CTCF bounda y si e wi hin LINC01476. (C) Schema ic ep esen a ion o he YPEL2 TAD s uc u e. 808 The Ame ican Jou nal o Human Gene ics 107, 802–814, No embe 5, 2020 Figu e 4. RP17 SVs C ea e No el Domains (neo-TADs) and Hype -ac i a ion o Re inal Enhance s (A) Schema ic modeling o he genome a chi ec u e spanning he RP17 egion using Hi-C maps. The wild- ype Hi-C map de i ed om neu onal issue shows a TAD wi h CTCF bounda ies con aining YPEL2 and e inal enhance s, lanked by uns uc u ed domains. TAD models o NL-SV1 and UK-SV2 (do ed e ical lines ep esen SV b eakpoin s) p edic he o ma ion o neo-TADs and ec opic in e ac- ions o he e inal enhance wi h GDPD1. (B) Hi-C pe o med on e inal o ganoids (ROs) de i ed om con ol ( op) and RP17 UK-SV2 indi iduals (bo om) (10 kb esolu ion; aw coun map). The ch oma in o ganiza ion in con ol ROs shows he YPEL2 TAD (indica ed by dashed lines). Two no el domains (neo- TAD 1 and 2) a e isible in he UK-SV2 ROs, and neo-TAD 2 allows ec opic e inal enhance con ac s o GDPD1 and SMG8. The dashed ci cle indica es he s ong ch oma in con ac be ween e inal enhance s and he GDPD1 p omo e . (C) qPCR e ealed signi ican ly up egula ed e inal enhance RNA exp ession in UK-SV2 ROs compa ed o con ols (n ¼3, mean 5s an- da d e o o he mean, **p %0.01). The Ame ican Jou nal o Human Gene ics 107, 802–814, No embe 5, 2020 809 Wild ype GDPD1SMG8 YPEL2 LN 01476 >< NL-SV1 GDPD1SMG8 YPEL2 01476 GDPD1 SMG8 LN 01476 GDPD1 YPEL2 LN YPEL2 SMG8GDPD1 Bounda y CTCF o ien a ion B Enhance in e ac ion > UK-SV2 Re inal enhance YPEL2 YPEL2 GDPD1 > < > >< < A B D GDPD1 CTL NL-SV1 0.0 0.5 1.0 1.5 2.0 Fold changes SMG8 CTL NL-SV1 0.0 0.5 1.0 1.5 2.0 Fold changes YPEL2 CTL NL-SV1 0.0 0.5 1.0 1.5 2.0 Fold changes GDPD1 SMG8 GDPD1 CTL UK-SV2 0.0 1.0 2.0 3.0 Fold changes SMG8 CTL UK-SV2 0.0 1.0 2.0 3.0 Fold changes YPEL2 CTL UK-SV2 0.0 1.0 2.0 3.0 Fold changes C E ** ** ** **** Re inal o ganoid exp ession Pho o ecep o p ecu so s exp ession Neo-TAD Neo-TAD No el TAD domain Gene ansc ip ion BB BB B B B B B B B B B B B BBBB BBBB B B B B YPEL2 TAD (legend on nex page) 810 The Ame ican Jou nal o Human Gene ics 107, 802–814, No embe 5, 2020