Human Molecular Genetics, 2014, Vol. 23, No. 11 doi:10.1093/hmg/ddu005 Advance Access published on January 12, 2014

2926–2939

PRPF4 mutations cause autosomal dominant retinitis pigmentosa Xue Chen1,{, Yuan Liu1,{, Xunlun Sheng2,{, Pancy O. S. Tam3, Kanxing Zhao4, Xuejuan Chen1, Weining Rong2, Yani Liu2, Xiaoxing Liu1, Xinyuan Pan1, Li Jia Chen3, Qingshun Zhao5, Douglas Vollrath6, Chi Pui Pang3 and Chen Zhao1,∗ 1

Received September 5, 2013; Revised December 4, 2013; Accepted January 8, 2014

Retinitis pigmentosa (RP), a disease characterized by progressive loss of photoreceptors, exhibits significant genetic heterogeneity. Several genes associated with U4/U6 – U5 triple small nuclear ribonucleoprotein (trisnRNP) complex of the spliceosome have been implicated in autosomal dominant RP (adRP). HPrp4, encoded by PRPF4, regulates the stability of U4/U6 di-snRNP, which is essential for continuous splicing. Here, we identified two heterozygous variants in PRPF4, including c.-114_-97del in a simplex RP patient and c.C944T (p.Pro315Leu), which co-segregates with disease phenotype in a family with adRP. Both variants were absent in 400 unrelated controls. The c.-114_-97del, predicted to affect two transcription factor binding sites, was shown to down-regulate the promoter activity of PRPF4 by a luciferase assay, and was associated with a significant reduction of PRPF4 expression in the blood cells of the patient. In fibroblasts from an affected individual with the p.Pro315Leu variant, the expression levels of several tri-snRNP components, including PRPF4 itself, were up-regulated, with altered expression pattern of SC35, a spliceosome marker. The same alterations were also observed in cells over expressing hPrp4Pro315Leu, suggesting that they arose as a compensatory response to a compromised splicing mechanism caused by hPrp4 dysfunction. Further, over expression of hPrp4Pro315Leu, but not hPrp4WT, triggered systemic deformities in wild-type zebrafish embryos with the retina primarily affected, and dramatically augmented death rates in morphant embryos, in which orthologous zebrafish prpf4 gene was silenced. We conclude that mutations of PRPF4 cause RP via haploinsufficiency and dominant-negative effects, and establish PRPF4 as a new U4/U6 – U5 snRNP component associated with adRP.

INTRODUCTION Retinitis pigmentosa [RP (MIM 268000)], the most common form of inherited retinal dystrophies (IRDs), displays a prevalence ranging from 1/3500 to 1/5000 among different populations (1 – 3). Clinically, photoreceptor degeneration and pigment migration are characteristics of RP, and symptoms include night

blindness, a constricted visual field (VF) and sometimes eventual loss of central vision (4). At the cellular level, rod photoreceptors and/or the retinal pigment epithelium (RPE) are affected at the initial stage of the disease, and cone photoreceptors may be involved at a later stage. At the molecular level, RP exhibits significant genetic heterogeneity involving at least 62 diseasecausing genes (RetNet), among which 23 genes have been

∗

To whom correspondence should be addressed. Tel: +86 2568135356; Fax: +86 2568135356; Email: [email protected] These authors contributed equally to the paper.

†

# The Author 2014. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]

Downloaded from http://hmg.oxfordjournals.org/ at University of Birmingham on May 29, 2015

Department of Ophthalmology, The First Affiliated Hospital of Nanjing Medical University and State Key Laboratory of Reproductive Medicine, Nanjing Medical University, Nanjing 210029, China, 2Ningxia Eye Hospital, Ningxia People Hospital, Ningxia 750001, China, 3Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong, Hong Kong, China, 4Tianjin Medical University, Tianjin Eye Hospital, Tianjin Key Laboratory of Ophthalmology and Visual Science, Tianjin 300040, China, 5MOE Key Laboratory of Model Animal for Disease Study, Model Animal Research Center, Nanjing University, Nanjing 210061, China and 6Department of Genetics, Stanford University School of Medicine, Stanford, CA 94305, USA

Human Molecular Genetics, 2014, Vol. 23, No. 11

RESULTS Clinical evaluations of the patients with PRPF4 variants The proband of Family AD01, AD01-II:3 was first referred to ophthalmic examination at the age of 15 for his poor night vision. After that, he developed VF restriction and impaired central vision. At the latest visit when he was 58-year-old, he showed bilateral cataracts and a typical RP fundus, including retinal degeneration with macular involvement, waxy pallor of optic disc, narrowed vasculature and peripheral pigment deposit (Fig. 1C versus E). Optical coherence tomography (OCT) revealed attenuated outer nuclear layer (ONL) and RPE at the maculae with complete loss of outer/inner segments (OS/IS), all of which were consistent with his poor visual acuity (Fig. 1D versus F). Diffused loss of VF and diminished electroretinogram (ERG) responses were also observed (Table 1 and Supplementary Material, Fig. S1). Patient AD01-II:1 and AD01-II:6 developed night blindness at 27 and 24 years of age, respectively. At the time of last examination when they were 66 and 47 years of age, respectively, they showed phenotypes similar to those of the proband, including

poor night vision, poor central visual acuity, severely reduced VF, typical RP fundus appearances and nearly undetectable ERG responses (Table 1 and Supplementary Material, Fig. S1). The individual AD01-III:4 was only 9-year-old at the time of last examination, whereas the ages at onset of disease in this family ranged from 15 to 27. Funduscopy and OCT examinations on AD01-III:4 detected no remarkable findings. ERG test revealed moderate defects including prolonged implicit time of b-wave in scotopic ERG, and decreased amplitudes of a- and b-waves in photopic ERG compared with age-matched controls (Table 1 and Supplementary Material, Fig. S1). We therefore diagnosed the AD01-III:4 as suspected RP. The clinical information of Family AD01 was detailed in Table 1 and Supplementary Material, Figure S1. The Simplex patient, S01, developed night blindness at the age of 20. Constricted VF and decreased VA were noticed by 37 years of age. At a recent visit when he was 53-year-old, he showed severely impaired VA, significantly restricted VF, undetectable ERG responses and a typical RP fundus (Fig. 1G and H and Table 1).

Identification of putative variants in PRPF4 With a targeted NGS approach, 2550 variants were initially identified in Patient AD01-II:3, including 2473 single nucleotide variations (SNVs) and 77 insert/deletions (Indels). The coverage of the targeted region was 99.85% and mean depth was 97.07-fold. Six SNVs and one Indel were retained as potentially pathogenic after initial bioinformatic analyses (Supplementary Material, Table S1). Further mutation validation revealed only one putative pathogenic variant, heterozygous PRPF4 (NM_004697.4) c.C944T, which cosegregated with the disease phenotype in the family (Fig. 1A and 2C). This variant was absent in all single nucleotide polymorphism (SNP) databases and in 400 unrelated controls as demonstrated by Sanger sequencing. We further screened the entire coding sequence of PRPF4 with potential promoter region, 5′ and 3′ untranslated region (UTR) in the proband AD01-II:3 by Sanger sequencing. No other potential pathogenic variant was identified. The c.C944T alteration resulted in a missense amino acids change, proline to leucine, at position 315 (p.Pro315Leu) of hPrp4 (Fig. 2C and D). Multiple orthologous sequence alignment revealed that Pro315 was absolutely conserved among all mammals and most lower species tested, but not in Danio rerio, Saccharomyces cerevisiae or Arabidopsis thaliana (Fig. 2E). Such change may represent a compensatory response during evolution to reserve the function of the protein. The p.Pro315Leu mutation was predicted to be ‘probably damaging’ by PolyPhen-2 (score, 0.999), ‘deleterious’ by PROVEN (score, 28.192), but ‘tolerated’ by SIFT (score, 0.07) and ‘neutral’ by CONDEL. Molecular modeling of hPrp4 (residues 229– 521) was constructed based on the crystal structure of B2J0I0_NOSP7 (Protein Data Bank ID: 2ymu), a WD-40 repeat protein containing a highly repetitive propeller with 14 blades, which exhibits the highest sequence identity (31%) with hPrp4. The constructed model revealed that residue 315 was located in the second blade of the seven repeated blades comprising hPrp4 (Fig. 2D). Taken together, the genetic analyses and structural analysis indicate that PRPF4 p.Pro315Leu could be the pathogenic mutation causing RP in this family.

Downloaded from http://hmg.oxfordjournals.org/ at University of Birmingham on May 29, 2015

implicated in autosomal dominant RP (adRP). Notably, 6 out of the 23 adRP-causing genes are involved in pre-messenger RNA (pre-mRNA) splicing, a fundamental cellular function required by nearly all human genes. The six genes include pre-mRNAprocessing factor 8 [PRPF8 (MIM 607300)] (5), PRPF31 (MIM 606419) (6), PRPF3 (MIM 607301) (7), PIM1-associated protein [RP9 (MIM 607331)] (8), small nuclear ribonucleoprotein 200 kDa [SNRNP200 (MIM 601664)] (9,10) and PRPF6 (MIM 613979) (11). Pre-mRNA splicing is executed by the spliceosome, which functions as a molecular machine to precisely recognize the splice sites of pre-mRNAs and remove the introns accordingly (12). The spliceosome consists of U1, U2, U4/U6 and U5 small nuclear ribonucleoproteins (snRNPs), as well as over a hundred non-snRNP-specific proteins (13). The U4/U6 – U5 tri-snRNP is a dynamic complex, of which structural rearrangements are critical for the assembly and the catalytic activation of the spliceosome (14– 16). Of the six splicing genes associated with adRP, five encode components of the tri-snRNP, highlighting the important role of this complex in disease pathogenesis and raising the question whether other components of the tri-snRNP that function tightly with the five splicing factors could also be correlated with RP. To test this idea, we have previously developed and validated a targeted next-generation sequencing (NGS) approach to screen mutations in 179 IRDs-causing genes and 10 putative genes associated with the U4/U6– U5 tri-snRNP (17). PRPF4 (MIM 607795), a gene encoding the U4/U6 di-snRNP 60 kDa protein (hPrp4), is among the 10 putative genes. HPrp4, a protein compromises 522 amino acids with its C-terminal portion containing seven WD40 repeats, interacts with other crucial components within the di-snRNP to regulate its stability (12,18– 21). Herein, we report two naturally occurring heterozygous mutations in PRPF4, c.C944T (p.Pro315Leu) and c.-114_-97del, which correlate, respectively, with RP in a family with autosomal dominant disease and in a sporadic case. Functional analyses have revealed that the two mutations could lead to RP via dominant-negative and loss-of-function effect, respectively.

2927

2928

Human Molecular Genetics, 2014, Vol. 23, No. 11

We further sequenced PRPF4 for mutations in another 225 index patients affected with RP. A heterozygous 18 bp deletion, c.-114_-97del, in PRPF4 was detected in a Sporadic patient, S01 (Fig. 1B and 2A–B). This deletion was also absent in all SNP database and 400 unrelated controls. We further screened for mutations in all known RP-causing genes in the Patient S01 and his unaffected son by Targeted NGS approach combined with Sanger sequencing for Retinitis Pigmentosa GTPase Regulator (RPGR) gene ORF15 (see Materials and Methods). No potential mutation was identified in any known RP-causing genes in the Patient S01. Ubiquitous expression of PRPF4 Previous study revealed that hPrp4 was widely expressed in multiple human tissues by northern blot and murine neuron cells through in situ hybridization (22). The expression of hPrp4 was also detected in a retinal-derived (NbLib0042 and NbLi0129) and RPE/choroid-derived (NbLi0047) cDNA library (NEIbank). Here, we revaluated the expression pattern of Prpf4 (NM_027297.3) in a panel of murine tissues (Fig. 2F). As expected, the Prpf4 was expressed in all murine tissues, including neural retina and RPE cells.

The c.-114_-97del decreased the expression of PRPF4 The 18 bp deletion, c.-114_-97del, crossing the transcription starting site (TSS) (Fig. 2A), was located in a predicted promoter region of PRPF4 (see Materials and Methods). Using PROMO (v3) on-line program, we found that this deletion could eliminate the binding sites for two transcription factors (TF), c-Ets-1 (T00112) and GR-alpha (T00337) (Fig. 3A). We then hypothesized that this variant would decrease the promoter activity of PRPF4. To test this idea, we conducted a luciferase assay to evaluate the promoter activity of the fragments with and without the deletion. As expected, in both ARPE19 and 293T cells transfected with PromoDEL, the luciferase activities were dramatically decreased by 66 and 75%, respectively, when compared with those transfected with PromoWT (Fig. 3B and C). The results indicate that the c.-114_-97del could significantly downregulate the expression of PRPF4. Therefore, we further determined the expression levels of PRPF4 in the blood cells of Patient S01 and negative controls. Interestingly, we found the expression of PRPF4 was decreased by 92.58% in the Patient S01 compared with four controls (Fig. 3D). As Patient S01 is heterozygous for c-114_-97del, the 92.58% reduction of PRPF4 expression observed in the patient is greater than

Downloaded from http://hmg.oxfordjournals.org/ at University of Birmingham on May 29, 2015

Figure 1. Pedigrees and clinical evaluations of Family AD01 and Sporadic Case S01. (A and B) The pedigree of Family AD01 indicates a likely dominant inheritance of RP over three generations (A), while no other family members are affected within the pedigree of sporadic case S01 (B). The genotypes, age of final examination and disease onset (inside parentheses) are given below the pedigree symbols. Black filled and blank symbols represent affected and unaffected status, respectively. (C and D) Fundus and OCT examination of the right eye of the unaffected member (AD01- III:2) are provided as normal controls. (E and F) Right eye fundus of the proband (AD01-II:3) reveals typical RP degeneration (E), including a waxy optic disc, attenuated retinal vessels and numerous bone spicule-like pigments. The white line indicates the layer for OCT examination (F), which demonstrates macular atrophy showing attenuated ONL and RPE with complete loss of OS and IS (denoted by yellow arrows). (G and H) The fundus of case S01 shows typical RP appearance for both eyes.

0.3 0.3 53/M c.-114_-97del 20 S01

1.0 0.8 / AD01-III:4 c.C944T

10/M

HM/40 cm 0.6 58/M 47/M 15 24 AD01-II:3 c.C944T AD01-II:6 c.C944T

WT, wild-type; F, female; M, male; BCVA, best corrected visual acuity; OD, right eye; OS, left eye; logMAR, logarithm of the minimum angle of resolution; HM, hand moving; NA, not available; MD, macular degeneration; OD, optic disk; AA, artery attenuation; PD, pigment deposits; MOD, moderate; NOR, normal; ERG, electroretinography; VF, visual field.

Mild

Waxy Yes Yes Mild

NOR No No Slightly NOR NOR Reduced Waxy Yes Yes Undetectable ,58 ,58 NOR No No No No

Severe Waxy Yes Yes Severe Waxy Yes Yes Undetectable ,108 ,58 Mild Waxy Yes Yes Mild Waxy Yes Yes Undetectable 158 158

,58 Waxy Yes Yes Undetectable 108 Waxy Yes Yes MOD

+0.75DS/24.0DC × 1808 NA 23.0DS/20.75DC × 1658 20.25DS/20.75DC 20.25DC × 1278 × 158 21.25DS 20.25DS/ 0.75DC × 108 27 AD01-II:1 c.C944T

66/F

0.05

22.0DS/24.5DC × 158 HM/10 cm NA 0.6 23.5DS 0.06

MOD

OD OS OD OS OD

Refractive error BCVA (logMAR)

Age (year) /sex Onset age (year) Genotype Patient ID

Table 1. Clinical features of attainable patients

2929

expected. We therefore expanded screening of potential promoter region (2920 bp to TSS) to search for any potential functional SNP that might regulate the expression of wild-type (WT) allele. However, no putative functional variant was found in the patient or in controls. Pathogenic consequences of p.Pro315Leu in cellular models P.Pro315Leu is associated with increased expression of U4/U6 – U5 tri-snRNP components Based on the genetic analysis and functional predictions of PRPF4 p.Pro315Leu, we further evaluated the functional effects of this missense variant in cellular models. Previous studies showed that silencing of sart3 and prpf31 in zebrafish would affect the expression of other tri-snRNP components (23,24), so we first determined whether this variant would have the same influence. To this end, we performed Q-PCR on cDNA templates generated from cultured fibroblasts derived from Patient AD01-II:3 and the unaffected family member AD01-III:2. Surprisingly, the expression of PRPF4 was up-regulated in the patient fibroblasts by 2.83-fold when compared with that of the control cells (Fig. 4A). We further applied SNaPshot to detect allele-specific mRNA expression of PRPF4 in cultured patient fibroblasts, using c.C944 or c.T944 as a marker. We found both WT allele (C944) and mutant allele (T994) were similarly up-regulated; the ratio between allele C944 and allele T944 in cDNA was 1.027 + 0.195, indicating that the up-regulation of PRPF4 expression was a compensatory response rather than a direct consequence to the mutation. In addition to PRPF4, increased expression levels of other tri-snRNP components were also observed in the patient’s fibroblast cells compared with control cells including PRPF3 (NM_004698.2), PRPF6 (NM_012469.3), PRPF8 (NM_006445.3), PRPF31 (NM_015629.3), EFTUD2 (NM_ 004247.3) and SART1 (NM_005146.4) (Fig. 4A). Altogether, the up-regulation of a number of splicing factors in the patient cells strongly suggested that the p.Pro315Leu variant caused hPrp4 dysfunction, which in turn, triggered a compensatory response of the U4/U6 – U5 tri-snRNP network. P.Pro315Leu is associated with altered expression of non-snRNP components The non-snRNP factors are also necessary for splicing, of which serine/arginine (SR)-rich proteins were considered as markers of splicing speckles (25). We determined whether p.Pro315Leu could affect the pattern of splicing speckles using SC-35, a non-snRNP component, as a marker (25). We performed immunofluorescence and Q-PCR on cells carrying either p.Pro315Leu or WT allele. By confocal microscopy, we found the level of endogenous SC-35 is dramatically increased with a diffused pattern in hPrp4Pro315Leu fibroblasts, whereas in control fibroblasts the SC-35 showed a distinctly speckled pattern (Fig. 4B and C). Consistently, Q-PCR revealed that the expression levels of SRSF1 (encoding SF2/ASF; NM_006924.4), another non-snRNP component and SRSF2 (encoding SC-35; NM_ 003016.4) were significantly increased in hPrp4Pro315Leu fibroblasts compared with WT fibroblasts (Fig. 4A). Similar changes in the levels and distribution of SC-35 were also observed in ARPE19 and 293T cells transfected with hPrp4Pro315Leu, but not in those transfected with hPrp4WT (Fig. 4D–G), indicating

Downloaded from http://hmg.oxfordjournals.org/ at University of Birmingham on May 29, 2015

Fundus appearance OD OS MD OD AA PD MD

OD

AA PD

ERG

VF

OS

Human Molecular Genetics, 2014, Vol. 23, No. 11

2930

Human Molecular Genetics, 2014, Vol. 23, No. 11

that the p.Pro315Leu variant of PRPF4 affected the splicing machinery. Deleterious effect of PRPF4 p.Pro315Leu in zebrafish Over expression of PRPF4c.C944U causes deformations in a large fraction of zebrafish We performed microinjection of PRPF4WT or PRPF4c.C944U mRNA into zebrafish embryos, and embryos injected with PRPF4 antisense RNA (PRPF4Anti) served as negative controls. While the injections of PRPF4WT or PRPF4c.C944U resulted in similar levels of expression of exogenous PRPF4 as demonstrated by Q-PCR, the expression of endogenous prpf4 (NM_199755.1) was increased in the PRPF4c.C944U group

compared with PRPF4WT group (Fig. 5A). Since PRPF4 is a ubiquitously expressed gene involved in a fundamental cellular function, pre-mRNA splicing, we further assessed the systemic effects of over expressed mutant and WT hPrp4 in zebrafish. At 4 day post fertilization (dpf), we found the injection of PRPF4c.C944U caused significant increase in systemic deformation of larvae when compared with injection of PRPF4WT (56 versus 12%, P , 0.001) (Fig. 5B and C). The systemic deformation included malformed brains, short trunks, cardiac edema and curved body axis (data not shown). Injections of PRPF4Anti, PRPF4WT and PRPF4c.C944U resulted in small numbers of embryonic lethality with no statistical difference among the three groups, presumably owing to the toxicity of exogenous mRNA.

Downloaded from http://hmg.oxfordjournals.org/ at University of Birmingham on May 29, 2015

Figure 2. Genetic analyses of variants identified in the PRPF4 Gene, and expression of Prpf4 in murine tissues. (A) PRPF4 gene (red filled box) spanning 17.56 kb on chromosome 9q32 (upper panel) contains 14 exons (middle panel). The two identified heterozygous variants, c.-114_-97del and c.C944T (p.Pro315Leu), were indicated by a red box crossing the green line and a red line in the middle panel, respectively. The two red arrows denote the transcription start site (TSS) and the translation start site (ATG). (B and C) Sequencing chromatograms of the affected member from Family S01 (B) and AD01 (C) showing the c.-114_-97del and c.C944T substitution, respectively (top). WT sequences of the unaffected family members were shown at bottom. (D) The top view illustrates the predicted structure of the WD-40 repeat domain of hPrp4 (residues 229– 521), which comprises seven highly repeated blades. Each blade is indicated by different colors. Residue 315 is located in the second blade. A close view of residue 315 highlighting the WT (proline) and mutated (leucine) amino acids at the position are presented in the two bottom boxes. (E) Orthologous protein sequence alignment of PRPF4 from human (Homo sapiens), chimpanzees (P. troglodytes), dogs (C. lupus), cows (B. taurus), rats (R. norvegicus), chickens (G. gallus), zebrafish (D. rerio), fruit flies (D. melanogaster), roundworms (C. elegans), yeast (S. cerevisiae and S. pombe) and plant (A. thaliana). Conserved residues are shaded. The mutated residue 315 is boxed and indicated. (F) Expression of Prpf4 in multiple murine tissues including lung, kidney, brain, spleen, heart, liver, neural retina and retinal pigmented epithelium (RPE) is shown. A 176 bp PCR product of the murine Prpf4 (top panel) was detected in all tested tissues. PCR products (162 bp) of the murine Gapdh were analyzed in parallel as a loading control.

Human Molecular Genetics, 2014, Vol. 23, No. 11

2931

Over expression of hPrp4Pro315Leu primarily affects zebrafish retina Silencing of prpf4 in zebrafish was previously shown to cause specific retinal phenotypes (24). We therefore investigated whether over expression of PRPF4c.C944U would have similar effects on retina. We randomly selected zebrafish with a normal systemic appearance from both PRPF4WT and PRPF4c.C944U groups at 4 dpf, and characterized their retinal phenotypes. Indeed, while the zebrafish from both groups showed normal ocular size, the structure of photoreceptors and IS/OS were severely interrupted in the zebrafish injected with PRPF4c.C944U as demonstrated by hematoxylin and eosin (H&E) staining. The PRPF4WT zebrafish had no remarkable changes in the retina (Fig. 5D). In addition to morphological changes, immunostaining revealed diminished reactivity of rhodopsin, a characteristic maker of rod photoreceptors, in the zebrafish injected with PRPF4c.C944U (Fig. 5E–J). By contrast, the PRPF4WT zebrafish showed robust expression of rhodopsin in the IS/OS layer, in a pattern similar to that of the uninjected zebrafish. Further, Q-PCR revealed that the expression levels of a number of important retinal transcripts including opn1lw1 (NM_131175.1), gnat2 (NM_131869.2), rs1 (NM_0010034 38.2), snrpg (ENSDART00000150308), prph2b (NM_1315 67.1), gnb3 (NM_213202.1) and rho (NM_131084.1) were decreased in PRPF4c.C944U zebrafish when compared with PRPF4WT zebrafish (Fig. 5K). In contrast, the mRNA levels of several splicing factors including prpf4 (NM_199755.1), prpf3 (NM_205748.1), prpf31 (NM_200504.1), prpf6 (NM_2126 55.1), prpf8 (NM_200976.2), snrnp200 (NM_001123257.1), eftud2 (NM_200508.2) and sart1 (NM_001002673.1) were up-regulated in PRPF4c.C944U zebrafish (Fig. 5A and L), consistent with the aforementioned observations in cellular models.

Altogether, we demonstrated that the mutation p.Pro315Leu had primary pathogenic effects on the retina. Over expression of hPrp4Pro315Leu significantly increased death rates in zebrafish with prpf4 silenced We further determined the effects of over expressed PRPF4WT or PRPF4c.C944U in a previously described zebrafish model, in which the orthologous zebrafish prpf4 was silenced by the injection of translational blocking morpholino oligos (MOs) into embryos (24). We performed co-injection of MO with PRPF4WT or PRPF4c.C944U mRNA, and compared their phenotypes with the sole MO-injected zebrafish. Consistent with a previous report (24), the MO injection in our study resulted in frequencies of 48% deformation and 31% death at 4 dpf, both of which were significantly higher than those of the control MO-injected zebrafish (Fig. 6A and B). As expected, co-injection of PRPF4WT partially rescued the morphant phenotypes as evidenced by decreased frequencies of deformation and death (35 and 19%, respectively). In contrast to PRPF4WT, co-injection of PRPF4c.C944U dramatically increased the frequency of dead morphants to 81% (Fig. 6A and B), suggesting that p.Pro315Leu is a dominant-negative mutation that further suppresses the residual function of zebrafish prpf4 in the morphant.

DISCUSSION Recent genetic and functional studies for RP have revealed the important role of defects in the U4/U6 – U5 tri-snRNP complex in the disease etiology and pathogenic mechanism. PRPF4, a gene encoding a crucial component associated with U4/U6

Downloaded from http://hmg.oxfordjournals.org/ at University of Birmingham on May 29, 2015

Figure 3. Luciferase assay for the promoter activity of the fragment with and without c.-114_-97del in PRPF4. (A) Transcription factor binding sites (TFBS) in the fragment across the mutational site of c.-114_-97del were shown. Two TFBS were eliminated due to the 18 bp deletion, including c-Ets-1 (T00112), GR-alpha (T00337). (B and C) Promoter activities of the fragments inserted are detected by luciferase assay in ARPE19 (B) and 293T cells (C). In both cell lines, promoter activity of PRPF4 was significantly reduced in cells transfected with PromoDEL, when compared with those transfected with PromoWT. Error bars represent SD from technical triplicates. (D) The expression level of PRPF4 in the blood cells of Patient S01 is about 7.42% of that of controls (n ¼ 4). Error bars represent SD from controls. ∗ P , 0.05, ∗∗ P , 0.01 and ∗∗∗ P , 0.001.

2932

Human Molecular Genetics, 2014, Vol. 23, No. 11

(blue), and the green channels represent reactivity of hPrp8 in (B and C) and reactivity of flag, a label for transfected hPrp4 in (D– G). Notably, SC-35 is increased in a diffuse pattern in all cells carrying PRPF4 p.Pro315Leu (C, E and G), whereas in the control cells (B, D and F), they consistently show distinct speckled pattern. Microscopy for all the cells were conducted under same settings, and all the images were automatically generated with no further modification. Scale bar: 20 mm.

Downloaded from http://hmg.oxfordjournals.org/ at University of Birmingham on May 29, 2015

Figure 4. Expression analysis and immunofluorescence for tri-snRNP and non-snRNP components in cells with or without p.Pro315Leu. (A) Expression levels of U4/U6-U5 tri-snRNP and non-snRNP components in fibroblasts with p.Pro315Leu were analyzed by Q-PCR, and plotted relative to the levels of the corresponding genes in the control fibroblasts. Data are represented as mean + SD from technical triplicates. ∗ P , 0.05, ∗∗ P , 0.01 and ∗∗∗ P , 0.001. (B–G) Immunofluorescence for SC-35 (red) in fibroblasts obtained from the unaffected member (B) and patient (C); in 293T cells transfected with PRPF4WT (D) and PRPF4c.C944U (E); and in ARPE19 cells transfected with PRPF4WT (F) and PRPF4c.C944U (G). All cells are counterstained with DAPI

di-snRNP, has not been previously implicated in RP, but has been suggested as a candidate gene for retinal degeneration based on zebrafish model (24). In this study, we report the correlation of RP with two naturally occurring heterozygous variants in PRPF4, both are absent in 400 unrelated controls. Functional analyses of the two alleles in cell culture models and in zebrafish further demonstrate their pathogenic effects. More importantly, we have discovered the particular linkage between the retinal phenotypes and the pathogenic mechanism of the two identified mutations. These findings establish PRPF4 as a new U4/U6 – U5 tri-snRNP component involved in adRP. Five adRP-related splicing factors are embodied in the U4/ U6– U5 tri-snRNP, and can be further divided into two groups: U4/U6-specific proteins including hPrp3 and hPrp31, and U5-specific proteins including hPrp6, hPrp8 and hBrr2 (12). HPrp4, another U4/U6-specific protein, directly binds to and interacts with hPrp3 and CypH to form a stable CypH-hPrp4hPrp3 heterotrimeric complex, which regulates the stability of the tri-snRNP (11,19,20,26,27), thus indicating its crucial role in the mechanism of pre-mRNA splicing. The silencing of endogenous prpf4 in zebrafish has been previously shown to cause retinal degeneration (24), suggesting that the insufficient gene dosage of PRPF4 could cause RP in patients. In this study, the c.-114_-97del was predicted to eliminate binding sites for c-ETS-1 and GR-a, and was further demonstrated to significantly decrease the putative promoter activity of PRPF4 by a luciferase assay, indicating a low expressivity of the mutant allele. However, further investigations of the core promoter region of PRPF4 and the TF binding defects caused by the deletion are required to better illustrate the pathogenic mechanism by which the deletion affects the regulatory network of PRPF4. Consistent with the luciferase assay results, the expression of PRPF4 was decreased by 92.58% in the blood cells of Patient S01 compared with controls. This finding further provides direct evidence for the loss-of-function nature of c-114_-97del, and strongly suggests that the function of PRPF4 gene in the patient is insufficient for proper splicing, which is very likely the cause of RP. Similar mechanism of haploinsufficiency has been illustrated for the mutations in PRPF31 (28–30). The 92.58% reduction of PRPF4 expression in the Patient S01 suggests the existence of either modifier genes or functional SNPs in the unknown regulatory region of PRPF4 that may regulate the expression of WT allele of PRPF4. Similar to this hypothesis, CNOT3 has recently been demonstrated as a modifier of PRPF31, which attribute to the symptomatic–asymptomatic phenomenon observed in family members carrying heterozygous haploinsufficient mutations of PRPF31 (31). Compared to PRPF4 c.-114_-97del, dominant-negative effect is likely the pathogenic mechanism associated with PRPF4 p.Pro315Leu as indicated by the zebrafish study and cellular models. In zebrafish injected with mutant mRNA, symptoms including high rates of deformation and death in

Human Molecular Genetics, 2014, Vol. 23, No. 11

2933

Downloaded from http://hmg.oxfordjournals.org/ at University of Birmingham on May 29, 2015

Figure 5. Deleterious effects of PRPF4 p.Pro315Leu in zebrafish. (A) The expressions of endogenous prpf4 (zf.) and exogenous PRPF4 (H.) in larvae at 2 dpf after injection of 200 pg PRPF4WT or PRPF4c.C944U were determined by Q-PCR and were relative to the expression of prpf4 in uninjected larvae. (B) Morphological changes at 4 dpf after injections of different mRNAs. Significant systemic deformations were observed in larvae injected with PRPF4c.C944U. Most zebrafish injected with PRPF4Anti or PRPF4WT mRNA were morphological normal at the same time. (C) Quantification of normal, deformed and dead zebrafish after injection with PRPF4Anti, PRPF4WT or PRPF4c.C944U from 2 to 4 dpf. n: total number of injected zebrafish from triple experiments. (D) H&E staining of zebrafish eye sections injected with mRNA at 4 dpf. Normal retinal layers (denoted by arrows) were observed in uninjected zebrafish and injected with PRPF4WT. The retinal structures including photoreceptors and IS/OS were severely interrupted in PRPF4c.C944U-injected zebrafish. GCL: ganglion cell layer; IPL: inner plexiform layer; INL: inner nuclear layer; OPL: outer plexiform layer; OP: photoreceptor layer; RPE: retinal pigmented epithelium. Scale bar: 20 mm. (E – J) Retinal frozen sections of the larvae injected with different mRNAs at 4 dpf were immunostained for rhodopsin (green). The uninjected and PRPF4WT mRNA injected zebrafish showed robust expression of rhodopsin in the IS/OS layer (E– H), whereas, the reactivity of rhodopsin was diminished in PRPF4c.C944U-injected larvae (I and J). The boxed areas in (F, H and J) were shown in higher magnification in E, G and I. Scale bar: 20 mm. (K, L) The mRNA expressions of retina-specific transcripts (K) and U4/U6– U5 tri-snRNP components (L) in PRPF4c.C944U mRNA-injected zebrafish were relative to the expressions of corresponding genes in PRPF4WT mRNA-injected larvae. Data in A, C, K, L are represented as mean + SD from technical triplicates. ∗ P , 0.05, ∗∗ P , 0.01 and ∗∗∗ P , 0.001.

zebrafish, diffused morphological changes in H&E staining, and loss of photoreceptors indicated by immunofluorescence were observed, which were not observed in those injected with WT

mRNA, indicating that the mutation functions as an antimorph or a neomorph. Similar symptoms are observed in the prpf4 morphant zebrafish in our study. In the previously described

2934

Human Molecular Genetics, 2014, Vol. 23, No. 11

zebrafish model injected with a sub-lethal dosage of the same prpf4-MO, the retina is also specifically affected with reduced expressions of a panel of retinal characteristic markers, which is similar to the observations made in zebrafish over expressing PRPF4c.C944U (24). This observation indicates that PRPF4 p.Pro315Leu functions as an antimorph instead of a neomorph, which is expected to generate distinct phenotypes compared with a morphant model. Indeed, co-injection of PRPF4c.C944U with MO dramatically increased the death rates of the morphants in our study. As the morphant zebrafish is not a true null allele for prpf4, this observation is presumably caused by the dominantnegative effect of PRPF4c.C944U on the sustained function of orthologous zebrafish prpf4 gene, which shows 80% identity with human PRPF4. In support of our hypothesis, mutations identified in PRPF3 and PRPF8 have been proposed to have dominant-negative effects correlating with severe phenotypes of RP (5,7,32). As PRPF4 and PRPF31 are both U4/ U6-associated proteins crucial for spliceosome assembly, a fundamental process required by all cells, it’s expected that over expression of the two mutant proteins or silencing of the two genes could cause similar systemic anomalies as we have observed and others have previously described (24,33). Interestingly, in all zebrafish models, including the one used in this study, photoreceptors and retinal transcripts were primarily affected and mimicked the phenotypes presented by RP patients. Pro315 is located in the second unit of the seven repeated WD40 motifs, of which analogs in yeast are involved in the interaction with Prp3 (34). Therefore, the p.Pro315Leu could decrease the stability of tri-snRNP by compromising the interaction between hPrp3 and hPrp4, further interrupting the

assembly of spliceosome. Other than the tri-snRNP components, two additional non-snRNP factors, SC-35 and SF2/ASF proteins, which belong to the SR protein family, are also affected by the mutant protein in this study. These two proteins interact directly with each other and present similar protein – protein interaction profiles (35). We have identified that SC-35 and SF2/ASF repeatedly increased in cells carrying hPrp4Pro315Leu, including primary fibroblasts, and transiently transfected 293T cells and RPE cells. These consistent and reproducible findings in various cells clearly reveal that p.Pro315Leu could very likely to cause defects in non-snRNP factors and to further affect the continuous splicing. Interestingly, in fibroblasts carrying p.Pro315Leu, we have observed that WT and mutant alleles were similarly up-regulated by nearly 3-fold, similar to the previous finding that translational blocking of prpf31 in zebrafish would cause increased expression of prpf31 (24). Based on the dominant-negative mechanism, the up-regulated protein encoded by the mutant PRPF4 allele could just counteract the additional WT hPrp4 protein. Additionally, several other core proteins involved in the U4/U6 – U5 tri-snRNP complex were also up-regulated in fibroblasts naturally carrying p.Pro315Leu and in zebrafish over expressed with the mutant protein, mimicking the findings in previous zebrafish models with silencing of prpf31 and sart3, a gene encoding a general splicing factor required by recycling of U4 and U6 snRNPs (23,24). For instance, in zebrafish with sart3 silenced, expression of prpf4 is up-regulated (23). The consistency among these models proves a coordinately compensatory response of overall splicing machinery to the deficiency of a single tri-snRNP factor.

Downloaded from http://hmg.oxfordjournals.org/ at University of Birmingham on May 29, 2015

Figure 6. Phenotypic comparison among larvae with prpf4 silencing, and with additional over expression of PRPF4WT, or PRPF4c.C944U. (A) Systemic morphology of larvae injected with 4 ng Control-MO (control); 4 ng Prpf4-MO (MO); 4 ng Prpf4-MO plus 200 pg PRPF4WT mRNA (MO + PRPF4WT); and 4 ng Prpf4-MO plus 200 pg PRPF4c.C944U mRNA (MO + PRPF4c.C944U). (B) Normal, deformed and dead larvae in four groups with various injections as indicated in (A) were quantified from 2 to 4 dpf. Silencing of prpf4 by injection of MO caused death in 31% of larvae and deformation in 48% of larvae. Compared with MO-injection, injection of MO + PRPF4WT particularly decreased the fraction of deformed and dead larvae, whereas injection of MO + PRPF4c.C944U significantly increased the rate of death (31 versus 81%, P , 0.001). Data are represented as mean + SD from technical triplicates. n: total number of larvae for each type of injection from triple experiments. ∗ P , 0.05, ∗∗ P , 0.01 and ∗∗∗ P , 0.001.

Human Molecular Genetics, 2014, Vol. 23, No. 11

MATERIALS AND METHODS Patients and controls A Han Chinese family (Fig. 1A) affected with adRP (numbered as AD01 hereafter), including three patients, one suspected patient and five asymptomatic members, was recruited in the First Affiliated Hospital of Nanjing Medical University (NJMU). Another 225 RP patients were also included, representing 43 probands in adRP families and 69 sporadic cases collected at the First Affiliated Hospital of Nanjing Medical University, and 113 sporadic cases recruited in the Hong Kong Eye Hospital and the Prince of Wales Hospital in Hong Kong. Peripheral blood samples were collected from all participants for genomic DNA extraction. Genomic DNA samples obtained from 400 unrelated Chinese individuals that were free of ocular diseases served as controls. Skin biopsies were performed on two members from Family AD01 (Patient AD01-II:3 and his asymptomatic son AD01-III:2) for primary culture of fibroblasts. All participants or their legal guardians signed informed written consents before they donated blood samples or skin tissue. All procedures

were prospectively reviewed and approved by the institutional ethical committees in accordance with Declaration of Helsinki. Animals C57BL/6 mice and Tuebingen zebrafish were housed in the Model Animal Research Center (MARC), Nanjing University, conformed to IACUC-approved protocol. Embryo rearing and fish husbandry were maintained at 28.58C with a 14 h light/ 10 h dark cycle. All embryos were produced by natural mating. Animal experiments were approved by the local ethical review board and conformed to the Guide for the Care and Use of laboratory animals. Targeted NGS and mutation screening For Family AD01, we employed a targeted sequence capture with NGS approach to screen 179 genes associated with IRDs and 10 candidate genes involved with U4/U6 – U5 tri-snRNP complex for potential mutations in the proband AD01-II:3. All known genes for adRP (RetNet) were included in the targeted sequence capture microarray, termed as RD-189 array (17). The methods of targeted sequence capture, DNA preparation and NGS were described previously (17). Initial bioinformatic analyses and mutation validation were conducted as previously described (17). Variants found in the normal populations as claimed by the five SNP databases, including dbSNP137, 1000 Genomes Project, HapMap Project, YH database and Exome Variant Server (EVS) were excluded. Intronic variants were discarded with the exception of those could potentially affect splicing sites, and variants do not fit with the inheritance mode of Family AD01 were further removed. We next performed Sanger sequencing to validate the variants, to analyze the intrafamilial cosegregation, and to test prevalence in 400 unrelated controls. Further, all 14 exons of PRPF4 including the 5′ untranslated region (5′ UTR) and 3′ UTR, and the 5′ flanking region of the TSS were screened on the proband AD01-II:3 and another 225 RP cases via Sanger sequencing. A pathogenic variant, c.-114_-97del, found in a simplex RP Patient, S01 (Fig. 1B), was further screened in 400 controls by Sanger sequencing. The standard protocol for Sanger sequencing had been described previously (43). Primer information was detailed in Supplementary Material, Table S3. We also performed targeted NGS using a commercial sequence capture array (Beijing Genome Institute, Shenzhen, China), which is similar to the RD-189 array in design, to screen 316 genes including all known RP-causing genes (detailed in Supplementary Material, Table S2) in the Patient S01 and his unaffected son (Fig. 1B). The bioinformatic analyses of the NGS variants detected in S01 family was performed using the same protocol as described above. As RPGR ORF15 is often not well captured by NGS, we then performed Sanger sequencing to screen for mutations in this gene in Patient S01 and his son using primers as previously described (44,45) and detailed in Supplementary Material, Table S3. In silico analyses and structural modeling We used Vector NTI AdvanceTM 2011 software to compare PRPF4 (NP_004688.2) with orthologues of Pan troglodytes

Downloaded from http://hmg.oxfordjournals.org/ at University of Birmingham on May 29, 2015

Mutations in pre-mRNA splicing genes are hypothesized to cause retina-specific phenotypes via distinct mechanisms. The first hypothesis is that the retina is a metabolically active tissue with continuous renewal of OS that highly demand pre-mRNA splicing of many genes. Therefore, the systemic defects in splicing caused by such mutations, as supported by indirect but convincing data (26,36,37), is only compromising to retina. This hypothesis is particularly relevant to the observation that PRPF3, PRPF8 and PRPF31 are highly expressed in murine retina (27), and retinal transcripts were most severely affected in zebrafish models with reduced expressions of splicing genes (24). The second hypothesis is that some mutations may affect the fidelity mechanisms of splicing, to which the retina might be highly sensitive. Possible examples are the RP-mutations identified in hBrr2 (encoded by SNRNP200), which is required by proofreading of pre-mRNA splicing (9,12,38– 40). Such mutations may primarily impair the fidelity of retinal gene expression leading to misspliced transcripts that are toxic to retina. However, there has been no direct evidence to support the altered splicing of retinal genes in vivo. Aberrant splicing of RHO gene was only observed in the cells over expressed with mini-gene construct (41), and was disapproved by subsequent study (42). Nevertheless, the mechanism by which the RP-mutations in splicing genes cause retinal specific disease is yet to be illustrated, and these mutations may affect some unknown functions specific to retina other than splicing. Understanding of these mechanisms would assist genetic counseling and direct future gene therapy for the patients with pre-mRNA splicing deficiency. In conclusion, PRPF4 is a new RP-associated pre-mRNA splicing gene. Taken together the genetic findings and the phenotypes of the corresponding zebrafish models, the two mutations of PRPF4 identified in this study, like mutations of PRPF31, are evidenced to cause RP via both haploinsufficiency and dominantnegative effect. Screening of PRPF4 in more RP patients for mutation with genotype–phenotype correlations are needed to discover its role in the RP etiology.

2935

2936

Human Molecular Genetics, 2014, Vol. 23, No. 11

(XP_520198.3), Canis lupus (XP_532038.2), Bos taurus (NP_001029502.1), Rattus norvegicus (NP_001100129.1), Gallus gallus (XP_415544.3), D. rerio (NP_956049.1), Drosophila melanogaster (NP_648990.1), Caenorhabditis elegans (NP_492363.1), S. cerevisiae (NP_015504.1), Schizosaccharomyces pombe (NP_592966.1) and A. thaliana (NP_181681.1). Possible functional impact of the substitution was predicted by four online prediction programs PolyPhen-2, SIFT Human Protein DB, PROVEAN (v.1.1.3) and CONDEL, and the transcription factor binding sites around the deletion were predicted by PROMO (v3) with 100% similarity (46,47). Protein modeling of hPrp4WT and hPrp4Pro315Leu was performed by SWISSMODEL online server (48,49). PyMol software (version 1.5) was used to display the predicted structures. Promoter of PRPF4 was predicted by Switchgear Genomics online server.

Reverse-transcriptase polymerase chain reaction (RT-PCR) and quantitative real-time PCR (Q-PCR) were performed to determine the expression of PRPF4 in the blood cells from Patient SD01 and controls. RNA isolation, subsequent cDNA synthesis and PCR were performed as described previously (9,43). Q-PCR was performed using FastStart Universal SYBR Green Master (ROX; Roche, Basel, Switzerland) with the StepOne Plus Realtime PCR System (Applied Biosystems, Darmstadt, Germany) per the manufacturer’s instructions. The same primers for expression study in fibroblasts were used to probe PRPF4 transcript for relative expression analysis according to previously described protocol (50). RT-PCR was also used to evaluate the expression pattern of the Prpf4 in multiple tissues obtained from C57BL/6 mice including lung, kidney, brain, spleen, heart, liver, neural retina, RPE – choroid – sclera eye cup. A 176 bp product of the murine Prpf4 and a 162 bp product of the murine housekeeping gene Gapdh (NM_008084.2) were generated from 100 ng of the synthetic cDNA via PCR using primers as detailed in Supplementary Material, Table S3. In vitro functional analyses of PRPF4 mutation Isolation and culture of skin fibroblasts Skin fibroblasts obtained from the Patient AD01-II:3 and his unaffected son AD01-III:2 were isolated and cultured per a previously published method (51). Briefly, punch biopsies (6 mm in diameter) were obtained from abdominal skin. The tissues were collected in Hank’s balanced salt solution (HBSS), supplemented with penicillin (100 IU/ml), streptomycin (100 mg/ml) and amphotericine B (0.25 mg/ml). Dulbecco’s modified Eagle’s medium/Ham’s F12 nutrient medium 1:1 (DME/F12 medium; Invitrogen, Carlsbad, CA, USA) was used for skin organ cultures. We chose enzymatic digestion for fibroblast primary cultures. The dermis was isolated from epidermis after incubating with 0.25% trypsin for 30 min, and tissue fragments were further removed after incubating with enzyme solution for 3 h at room temperature. The isolated fibroblasts were maintained in DMEM supplemented with 10% fetal bovine serum (FBS; Invitrogen, Carlsbad, CA, USA), penicillin (100 U/ml) and streptomycin (100 g/ml) at 378C, 5% CO2. Complete medium was short for supplemented culture medium in the following passage.

SNaPshot assay for allelic expression of PRPF4 in fibroblasts DNA and cDNA extracted from fibroblasts of AD01-II:3 and AD01-III:2 were amplified with amplification primers (Supplementary Material, Table S3) and purified as described previously (43). A primer extension method (SNaPshot) was then performed to analyze those purified PCR products using extension primers, designed adjacent to the PRPF4 c.C944T site, and labeled with different single fluorescent dideoxyribonucleoside triphosphate (ddNTP) added to the polymorphic site (52). The resulted products were then run on an ABI 3730XL capillary electrophoresis instrument (Applied Biosystems, Carlsbad, CA, USA). Gene Mapper 4.1 software (Applied Biosystems, Carlsbad, CA, USA) was used for data analysis. For each sample, we used peak area ratios to measure the relative amount of C944/T944 for both genomic DNA and cDNA templates. The cDNA ratio between C944 allele and T944 allele were normalized by the corresponding ratio in genomic DNA. Plasmids, cell transfection, and luciferase reporter assay The open reading frame (ORF) sequence of WT PRPF4 was cloned from embryonic kidney (293T) cells and further inserted into pCMV-C-Flag plasmid (Beyotime) with Flag sequence in-frame fused for cell transfection, and pxT7 plasmid (a kind gift from Dr Anming Meng, Tsinghua University) for in vivo transcription. The WT cDNA was obtained from 293T cells and amplified via PCR using primers as detailed in Supplementary Material, Table S3. The primers were designed to match the PRPF4 sequence at their 3′ moiety but carry artificially introduced restriction sites (underlined) for endonucleases at their 5′ end (Supplementary Material, Table S3). The purified PCR amplicons were cloned into the designated plasmid to produce recombinant plasmids AcFlag-PRPF4WT and AcpxT7-PRPF4WT (43). We further introduced the c.C944T mutation into the WT insert by a two-step PCR overlap extension method, then subcloned the mutated insert into the two plasmids to obtain AcFlagPRPF4c.C944T and AcpxT7-PRPF4c.C944T (53,54). In a predicted promoter region of PRPF4, we selected a fragment for luciferase assay, which spans from 2401 to 21 bp upstream to the translation starting site and crossed the TSS of PRPF4 (Fig. 2A). The fragment was synthesized, amplified and ligated into firefly luciferase expression vector pGL3-Basic vector (Promega, Madison, WI, USA) with XhoI and NcoI restriction sites to construct the recombinant plasmid termed as PromoWT. Similarly, the same fragment but with the c.-114_97del was synthesized, amplified and ligated to the same vector to obtain the PromoDEL plasmid. All cDNA inserts as well as junction sequences were confirmed by direct sequencing in both directions. Human RPE (ARPE19) and 293T cells were seeded into 24-well plates using complete DMEM medium at 378C, 5% CO2. Transfections were performed at 40–50% confluence using LipofectamineTM 2000 Transfection reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s protocols.

Downloaded from http://hmg.oxfordjournals.org/ at University of Birmingham on May 29, 2015

Expression analyses of PRPF4

Expression analyses of related components RNA was extracted from primary cultured fibroblasts. Reverse transcription was performed followed by Q-PCR using the same protocol as mentioned above. Relative expression analyses of several snRNP and non-snRNP components were performed using primers as detailed in Supplementary Material, Table S3.

Human Molecular Genetics, 2014, Vol. 23, No. 11

2937

Functional analyses of PRPF4 mutation in zebrafish

Statistics

Microinjections of zebrafish Capped and tailed mRNAs of PRPF4-antisense (PRPF4Anti), PRPF4WT (hPrp4WT) and PRPF4c.C944U (hPrp4Pro315Leu) were generated from the linearized plasmids AcpxT7-PRPF4WT and AcpxT7-PRPF4c.C944T with the mMESSAGE mMACHINE T7 Ultra Kit (Ambion, USA). Free nucleotides were removed to purify the synthesized mRNAs using the RNeasy Kit (Qiagen, Hilden, Germany). Translational blocking MOs for zebrafish prpf4 (Prpf4-MO) and Control-MO were designed as previously described and purchased from Gene Tools, LLC (Philomath, Oregon, USA) (24). Embryos at 1- to 2-cell-stage (0 dpf) were selected for microinjections. For mRNA microinjection, each embryo was injected with 1 nl solution containing 200 pg PRPF4Anti, PRPF4WT or PRPF4c.C944U mRNA. For co-injections of MO and mRNA, zebrafish embryos were separated into four groups and were injected with 4 ng Control-MO, 4 ng Prpf4-MO, 4 ng Prpf4-MO plus 200 pg PRPF4WT mRNA and 4 ng Prpf4-MO plus 200 pg PRPF4c.C944U mRNA, respectively. For all injections described above, embryos died within 24 h post injection were excluded because such death likely resulted from unspecific causes. The percentage of deformation and death of zebrafish embryos was calculated from 2 to 4 dpf.

Comparisons between different groups were made using one-way ANOVA or Student’s T-test, with or without repeated measures, respectively. P , 0.05 was taken as statistically significant. All data were presented as mean + standard deviation (SD). Statistical analysis was conducted using GraphPad Prism version 4.0 (GraphPad Software, San Diego, CA, USA).

Expression analysis, H&E staining and immunostaining of zebrafish At 4 dpf, zebrafish from PRPF4WT-injected group or PRPF4c.C944U-injected group were collected and pooled for expression analysis. RNA isolation and RT-PCR were performed as described above. Q-PCR was carried out to determine the expression of several U4/U6–U5 tri-snRNP components, including prpf3, prpf4, prpf31, prpf6, prpf8, snrnp200, eftud2 and sart1, together with the expressions levels of several retina-specific transcripts, including opn1lw1, gnat2, rs1, opnlsw1, snrpg, prph2b,

SUPPLEMENTARY MATERIAL Supplementary Material is available at HMG online.

ACKNOWLEDGEMENTS We thank all patients and family members for their participation in this study. We also thank Professor Jonathan P. Staley from University of Chicago for his important comments, and Professor Qingjiong Zhang from Zhongshan Ophthalmic Center for his help with mutation screening. We appreciate Liping Guan, Jingjing Jiang and Jingjing Xiao from BGI-Shenzhen for technical support. We are also grateful to Genesky Biotechnologies Inc. (Shanghai, China) for their technical help with the SNaPshot assay. Conflict of Interest Statement. None declared.

FUNDING This work was supported by National Key Basic Research Program of China (2013CB967500 to C.Z.); National Natural Science Foundation of China (81222009 and 81170856 to C.Z., 81260154 to X.S., 81170867 to K.Z.); Ningxia Important Sciences and Technologies Project (2011ZYS175 to X.S.);

Downloaded from http://hmg.oxfordjournals.org/ at University of Birmingham on May 29, 2015

Immunofluorescence and antibodies Fibroblasts were grown on 8-well chamber slides (Millipore, Billerica, MA, USA), fixed, and incubated with designated primary antibodies. Similarly, ARPE19 or 293T cells were harvested at 48 h post transfection, fixed and incubated with designated primary antibodies at 48C overnight. After five washes, the designated fluorescence-conjugated secondary antibodies (Invitrogen, Carlsbad, CA, USA) were incubated with cells for another 1 h at room temperature. The cell nuclei were counterstained by 4′ , 6-diamidino-2-phenylindole (DAPI; Sigma, USA) for 5 min. An Olympus IX70 confocal laser-scanning microscope (Olympus, Tokyo, Japan) was used for image collecting. Antibodies are described in Supplementary Material, Table S4.

gnb3 and rho, using primers as listed in Supplementary Material, Table S3. At 4 dpf, zebrafish with a normal systemic appearance were randomly selected from the PRPF4WT-injected group (n ¼ 14) and PRPF4c.C944U-injected group (n ¼ 12). H&E staining and immunostaining were performed to evaluate their ocular morphology and rod photoreceptors. The fish was fixed in 4% paraformaldehyde (PFA) overnight at 48C, washed three times with phosphate buffered saline (PBS) containing 0.01% Tween 20 (PBST), and dehydrated in 30% sucrose overnight. After that, they were embedded in optimal cutting temperature solution, frozen in liquid nitrogen for 1 min., and sectioned by Leica CM1900 cryostat (Leica, Germany). For H&E staining, the slides were stained with hematoxylin, rinsed with 1% acidic alcohol solution, and counterstained with 0.5% eosin. Alcohol baths for dehydration were then performed followed by the mounting with a resin. Staining was visualized with the Olympus BX-51 microscope (Olympus, Tokyo, Japan). For immunostaining, the slides were dehydrated for 30 min. at 378C, rehydrated with PBST for 5 min., and blocked with PBS supplemented with 10% normal goat serum and 2% BSA for 1 h at room temperature. The slides were then incubated with antizebrafish rhodopsin (Supplementary Material, Table S4) over night at 48C. The same procedures were conducted same as immunofluorescence for human cells.

For luciferase assays, both cells were transfected with 10 ng cytomegalovirus (CMV)–Renilla (Promega, Madison, WI, USA) and 500 ng PromoWT or PromoDEL. Cells were harvested at 48 h post transfection to detect their luciferase activities using the dual luciferase system (Promega, Madison, WI, USA). Firefly luciferase activities were normalized to Renilla luciferase activities, which were taken as internal standard indicators for transfection efficiency. A GloMax-96 luminometer was used to measure the activities of firefly and Renilla luciferases. Experiments were conducted in triplicates with data averaged.

2938

Human Molecular Genetics, 2014, Vol. 23, No. 11

Thousand Youth Talents Program of China (to C.Z.); Jiangsu Outstanding Young Investigator Program (BK2012046 to C.Z.); Jiangsu Province’s Key Provincial Talents Program (RC201149 to C.Z.); the Fundamental Research Funds of the State Key Laboratory of Ophthalmology (to C.Z.); Jiangsu Province’s Scientific Research Innovation Program for Postgraduates (CXZZ13_05 to X.C.); Foundation Fighting Blindness (to D.V.); and Applied Research Program of the Science and Technology Commission Foundation of Tianjin (013111411 to K.Z.).

REFERENCES

Downloaded from http://hmg.oxfordjournals.org/ at University of Birmingham on May 29, 2015

1. Carrillo, J., Martinez, P., Solera, J., Moratilla, C., Gonzalez, A., ManguanGarcia, C., Aymerich, M., Canal, L., Del Campo, M., Dapena, J.L. et al. (2012) High resolution melting analysis for the identification of novel mutations in DKC1 and TERT genes in patients with dyskeratosis congenita. Blood Cells Mol. Dis., 49, 140–146. 2. Chizzolini, M., Galan, A., Milan, E., Sebastiani, A., Costagliola, C. and Parmeggiani, F. (2011) Good epidemiologic practice in retinitis pigmentosa: from phenotyping to biobanking. Curr. Genomics, 12, 260–266. 3. Collin, R.W., Safieh, C., Littink, K.W., Shalev, S.A., Garzozi, H.J., Rizel, L., Abbasi, A.H., Cremers, F.P., den Hollander, A.I., Klevering, B.J. et al. (2010) Mutations in C2ORF71 cause autosomal-recessive retinitis pigmentosa. Am. J. Hum. Genet., 86, 783– 788. 4. Hartong, D.T., Berson, E.L. and Dryja, T.P. (2006) Retinitis pigmentosa. Lancet, 368, 1795–1809. 5. McKie, A.B., McHale, J.C., Keen, T.J., Tarttelin, E.E., Goliath, R., van Lith-Verhoeven, J.J., Greenberg, J., Ramesar, R.S., Hoyng, C.B., Cremers, F.P. et al. (2001) Mutations in the pre-mRNA splicing factor gene PRPC8 in autosomal dominant retinitis pigmentosa (RP13). Hum. Mol. Genet., 10, 1555– 1562. 6. Vithana, E.N., Abu-Safieh, L., Allen, M.J., Carey, A., Papaioannou, M., Chakarova, C., Al-Maghtheh, M., Ebenezer, N.D., Willis, C., Moore, A.T. et al. (2001) A human homolog of yeast pre-mRNA splicing gene, PRP31, underlies autosomal dominant retinitis pigmentosa on chromosome 19q13.4 (RP11). Mol. Cell, 8, 375–381. 7. Chakarova, C.F., Hims, M.M., Bolz, H., Abu-Safieh, L., Patel, R.J., Papaioannou, M.G., Inglehearn, C.F., Keen, T.J., Willis, C., Moore, A.T. et al. (2002) Mutations in HPRP3, a third member of pre-mRNA splicing factor genes, implicated in autosomal dominant retinitis pigmentosa. Hum. Mol. Genet., 11, 87–92. 8. Keen, T.J., Hims, M.M., McKie, A.B., Moore, A.T., Doran, R.M., Mackey, D.A., Mansfield, D.C., Mueller, R.F., Bhattacharya, S.S., Bird, A.C. et al. (2002) Mutations in a protein target of the Pim-1 kinase associated with the RP9 form of autosomal dominant retinitis pigmentosa. Eur. J. Hum. Genet., 10, 245– 249. 9. Zhao, C., Bellur, D.L., Lu, S., Zhao, F., Grassi, M.A., Bowne, S.J., Sullivan, L.S., Daiger, S.P., Chen, L.J., Pang, C.P. et al. (2009) Autosomal-dominant retinitis pigmentosa caused by a mutation in SNRNP200, a gene required for unwinding of U4/U6 snRNAs. Am. J. Hum. Genet., 85, 617–627. 10. Li, N., Mei, H., MacDonald, I.M., Jiao, X. and Hejtmancik, J.F. (2010) Mutations in ASCC3L1 on 2q11.2 are associated with autosomal dominant retinitis pigmentosa in a Chinese family. Invest. Ophthalmol. Vis. Sci., 51, 1036– 1043. 11. Tanackovic, G., Ransijn, A., Ayuso, C., Harper, S., Berson, E.L. and Rivolta, C. (2011) A missense mutation in PRPF6 causes impairment of pre-mRNA splicing and autosomal-dominant retinitis pigmentosa. Am. J. Hum. Genet., 88, 643– 649. 12. Liu, S., Rauhut, R., Vornlocher, H.P. and Luhrmann, R. (2006) The network of protein–protein interactions within the human U4/U6.U5 tri-snRNP. RNA, 12, 1418–1430. 13. Wahl, M.C., Will, C.L. and Luhrmann, R. (2009) The spliceosome: design principles of a dynamic RNP machine. Cell, 136, 701– 718. 14. Cheng, S.C. and Abelson, J. (1987) Spliceosome assembly in yeast. Genes Dev., 1, 1014– 1027. 15. Lamond, A.I., Konarska, M.M., Grabowski, P.J. and Sharp, P.A. (1988) Spliceosome assembly involves the binding and release of U4 small nuclear ribonucleoprotein. Proc. Natl. Acad. Sci. U S A, 85, 411–415.

16. Pikielny, C.W., Rymond, B.C. and Rosbash, M. (1986) Electrophoresis of ribonucleoproteins reveals an ordered assembly pathway of yeast splicing complexes. Nature, 324, 341 –345. 17. Chen, X., Zhao, K., Sheng, X., Li, Y., Gao, X., Zhang, X., Kang, X., Pan, X., Liu, Y., Jiang, C. et al. (2013) Targeted sequencing of 179 genes associated with hereditary retinal dystrophies and 10 candidate genes identifies novel and known mutations in patients with various retinal diseases. Invest. Ophthalmol. Vis. Sci., 54, 2186– 2197. 18. Lauber, J., Plessel, G., Prehn, S., Will, C.L., Fabrizio, P., Groning, K., Lane, W.S. and Luhrmann, R. (1997) The human U4/U6 snRNP contains 60 and 90 kD proteins that are structurally homologous to the yeast splicing factors Prp4p and Prp3p. RNA, 3, 926– 941. 19. Wang, A., Forman-Kay, J., Luo, Y., Luo, M., Chow, Y.H., Plumb, J., Friesen, J.D., Tsui, L.C., Heng, H.H., Woolford, J.L. Jr. et al. (1997) Identification and characterization of human genes encoding Hprp3p and Hprp4p, interacting components of the spliceosome. Hum. Mol. Genet., 6, 2117– 2126. 20. Horowitz, D.S., Kobayashi, R. and Krainer, A.R. (1997) A new cyclophilin and the human homologues of yeast Prp3 and Prp4 form a complex associated with U4/U6 snRNPs. RNA, 3, 1374–1387. 21. Gonzalez-Santos, J.M., Wang, A., Jones, J., Ushida, C., Liu, J. and Hu, J. (2002) Central region of the human splicing factor Hprp3p interacts with Hprp4p. J. Biol. Chem., 277, 23764– 23772. 22. Kojima, T., Zama, T., Wada, K., Onogi, H. and Hagiwara, M. (2001) Cloning of human PRP4 reveals interaction with Clk1. J. Biol. Chem., 276, 32247– 32256. 23. Trede, N.S., Medenbach, J., Damianov, A., Hung, L.H., Weber, G.J., Paw, B.H., Zhou, Y., Hersey, C., Zapata, A., Keefe, M. et al. (2007) Network of coregulated spliceosome components revealed by zebrafish mutant in recycling factor p110. Proc. Natl. Acad. Sci. USA, 104, 6608–6613. 24. Linder, B., Dill, H., Hirmer, A., Brocher, J., Lee, G.P., Mathavan, S., Bolz, H.J., Winkler, C., Laggerbauer, B. and Fischer, U. (2011) Systemic splicing factor deficiency causes tissue-specific defects: a zebrafish model for retinitis pigmentosa. Hum. Mol. Genet., 20, 368–377. 25. Krainer, A.R., Conway, G.C. and Kozak, D. (1990) The essential pre-mRNA splicing factor SF2 influences 5′ splice site selection by activating proximal sites. Cell, 62, 35– 42. 26. Zhang, Z., Lotti, F., Dittmar, K., Younis, I., Wan, L., Kasim, M. and Dreyfuss, G. (2008) SMN deficiency causes tissue-specific perturbations in the repertoire of snRNAs and widespread defects in splicing. Cell, 133, 585– 600. 27. Cao, H., Wu, J., Lam, S., Duan, R., Newnham, C., Molday, R.S., Graziotto, J.J., Pierce, E.A. and Hu, J. (2011) Temporal and tissue specific regulation of RP-associated splicing factor genes PRPF3, PRPF31 and PRPC8— implications in the pathogenesis of RP. PLoS One, 6, e15860. 28. Abu-Safieh, L., Vithana, E.N., Mantel, I., Holder, G.E., Pelosini, L., Bird, A.C. and Bhattacharya, S.S. (2006) A large deletion in the adRP gene PRPF31: evidence that haploinsufficiency is the cause of disease. Mol. Vis., 12, 384– 388. 29. Rio Frio, T., Civic, N., Ransijn, A., Beckmann, J.S. and Rivolta, C. (2008) Two trans-acting eQTLs modulate the penetrance of PRPF31 mutations. Hum. Mol. Genet., 17, 3154– 3165. 30. Saini, S., Robinson, P.N., Singh, J.R. and Vanita, V. (2012) A novel 7 bp deletion in PRPF31 associated with autosomal dominant retinitis pigmentosa with incomplete penetrance in an Indian family. Exp. Eye Res., 104, 82– 88. 31. Venturini, G., Rose, A.M., Shah, A.Z., Bhattacharya, S.S. and Rivolta, C. (2012) CNOT3 is a modifier of PRPF31 mutations in retinitis pigmentosa with incomplete penetrance. PLoS Genet., 8, e1003040. 32. Graziotto, J.J., Inglehearn, C.F., Pack, M.A. and Pierce, E.A. (2008) Decreased levels of the RNA splicing factor Prpf3 in mice and zebrafish do not cause photoreceptor degeneration. Invest. Ophthalmol. Vis. Sci., 49, 3830– 3838. 33. Yin, J., Brocher, J., Fischer, U. and Winkler, C. (2011) Mutant Prpf31 causes pre-mRNA splicing defects and rod photoreceptor cell degeneration in a zebrafish model for retinitis pigmentosa. Mol. Neurodegener., 6, 56. 34. Ayadi, L., Callebaut, I., Saguez, C., Villa, T., Mornon, J.P. and Banroques, J. (1998) Functional and structural characterization of the prp3 binding domain of the yeast prp4 splicing factor. J. Mol. Biol., 284, 673– 687. 35. Zhang, W.J. and Wu, J.Y. (1996) Functional properties of p54, a novel SR protein active in constitutive and alternative splicing. Mol. Cell. Biol., 16, 5400– 5408.

Human Molecular Genetics, 2014, Vol. 23, No. 11

45. Meindl, A., Dry, K., Herrmann, K., Manson, F., Ciccodicola, A., Edgar, A., Carvalho, M.R., Achatz, H., Hellebrand, H., Lennon, A. et al. (1996) A gene (RPGR) with homology to the RCC1 guanine nucleotide exchange factor is mutated in X-linked retinitis pigmentosa (RP3). Nat. Genet., 13, 35– 42. 46. Messeguer, X., Escudero, R., Farre, D., Nunez, O., Martinez, J. and Alba, M.M. (2002) PROMO: detection of known transcription regulatory elements using species-tailored searches. Bioinformatics, 18, 333–334. 47. Farre, D., Roset, R., Huerta, M., Adsuara, J.E., Rosello, L., Alba, M.M. and Messeguer, X. (2003) Identification of patterns in biological sequences at the ALGGEN server: PROMO and MALGEN. Nucleic Acids. Res., 31, 3651– 3653. 48. Arnold, K., Bordoli, L., Kopp, J. and Schwede, T. (2006) The SWISS-MODEL workspace: a web-based environment for protein structure homology modelling. Bioinformatics, 22, 195–201. 49. Kiefer, F., Arnold, K., Kunzli, M., Bordoli, L. and Schwede, T. (2009) The SWISS-MODEL Repository and associated resources. Nucleic Acids. Res., 37, D387–D392. 50. Livak, K.J. and Schmittgen, T.D. (2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) method. Methods, 25, 402– 408. 51. Rittie, L. and Fisher, G.J. (2005) Isolation and culture of skin fibroblasts. Methods Mol. Med., 117, 83– 98. 52. Pinsonneault, J., Nielsen, C.U. and Sadee, W. (2004) Genetic variants of the human H+/dipeptide transporter PEPT2: analysis of haplotype functions. J. Pharmacol. Exp. Ther., 311, 1088– 1096. 53. Ho, S.N., Hunt, H.D., Horton, R.M., Pullen, J.K. and Pease, L.R. (1989) Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene, 77, 51–59. 54. Horton, R.M., Cai, Z.L., Ho, S.N. and Pease, L.R. (1990) Gene splicing by overlap extension: tailor-made genes using the polymerase chain reaction. Biotechniques, 8, 528– 535.

Downloaded from http://hmg.oxfordjournals.org/ at University of Birmingham on May 29, 2015

36. Huranova, M., Ivani, I., Benda, A., Poser, I., Brody, Y., Hof, M., Shav-Tal, Y., Neugebauer, K.M. and Stanek, D. (2010) The differential interaction of snRNPs with pre-mRNA reveals splicing kinetics in living cells. J. Cell. Biol., 191, 75–86. 37. Tanackovic, G., Ransijn, A., Thibault, P., Abou Elela, S., Klinck, R., Berson, E.L., Chabot, B. and Rivolta, C. (2011) PRPF mutations are associated with generalized defects in spliceosome formation and pre-mRNA splicing in patients with retinitis pigmentosa. Hum. Mol. Genet., 20, 2116–2130. 38. Raghunathan, P.L. and Guthrie, C. (1998) RNA unwinding in U4/U6 snRNPs requires ATP hydrolysis and the DEIH-box splicing factor Brr2. Curr. Biol., 8, 847 –855. 39. van Nues, R.W. and Beggs, J.D. (2001) Functional contacts with a range of splicing proteins suggest a central role for Brr2p in the dynamic control of the order of events in spliceosomes of Saccharomyces cerevisiae. Genetics, 157, 1451– 1467. 40. Nguyen, T.H., Li, J., Galej, W.P., Oshikane, H., Newman, A.J. and Nagai, K. (2013) Structural basis of Brr2–Prp8 interactions and implications for U5 snRNP biogenesis and the spliceosome active site. Structure, 21, 910–919. 41. Yuan, L., Kawada, M., Havlioglu, N., Tang, H. and Wu, J.Y. (2005) Mutations in PRPF31 inhibit pre-mRNA splicing of rhodopsin gene and cause apoptosis of retinal cells. J. Neurosci., 25, 748– 757. 42. Wilkie, S.E., Vaclavik, V., Wu, H., Bujakowska, K., Chakarova, C.F., Bhattacharya, S.S., Warren, M.J. and Hunt, D.M. (2008) Disease mechanism for retinitis pigmentosa (RP11) caused by missense mutations in the splicing factor gene PRPF31. Mol. Vis., 14, 683– 690. 43. Zhao, C., Lu, S., Zhou, X., Zhang, X., Zhao, K. and Larsson, C. (2006) A novel locus (RP33) for autosomal dominant retinitis pigmentosa mapping to chromosomal region 2cen-q12.1. Hum. Genet., 119, 617– 623. 44. Demirci, F.Y., Rigatti, B.W., Wen, G., Radak, A.L., Mah, T.S., Baic, C.L., Traboulsi, E.I., Alitalo, T., Ramser, J. and Gorin, M.B. (2002) X-linked cone– rod dystrophy (locus COD1): identification of mutations in RPGR exon ORF15. Am. J. Hum. Genet., 70, 1049– 1053.

2939