PRPF31 gene organisation and protein structure

PRPF31 is a 16 kb gene composed of 14 exons, located at chromosome 19q13.4 (figure 1). PRPF31 encodes a 499 amino acid, 61 kDa protein, referred to as PRP31, PRPF31, hPrp31 and, in older literature, 61K protein. The protein contains a NOP domain: a highly conserved multihelical domain generated by amino acid residues 215–333. The NOP domain mediates protein–protein interactions during the splicing cycle via HAT repeats 1–6, HAT repeats 7*–13 and the TPR-repeat domains.2 3 Furthermore, the protein has a coiled-coil domain, formed by two non-contiguous helices (residues 85–120; 181–215) and a nuclear localisation signal (residues 351–364).4

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Genomic architecture of PRPF31 showing exon-intron arrangement and untranslated regions.

The protein is ubiquitously expressed. Additionally, distant homologues of 61K protein include the nucleolar proteins NOP56 and NOP58 from the box C/D snoRNPs.3 A central domain of protein 61K, from amino acids 93–328, shares homology with human NOP proteins. It has been proposed that as these proteins are components of RNP complexes, the conserved central domain of 61K may be involved in RNA binding.4

Biological functions of PRPF31 protein

PRPF31 plays an essential role in spliceosomal function, the spliceosome being a ribozyme responsible for pre-mRNA splicing. Most eukaryotic pre-mRNA contains segments of non-coding intervening sequences (introns) that must be removed by splicing. The remaining coding sequences, exons, are then ligated together to form translatable mRNAs. Differential inclusion and exclusion of exons during this process increase the number of unique proteins that can be expressed from a single gene.

The current model of spliceosome function involves a dynamic, stepwise process of assembly, activation and catalysis (figure 2).5 6 After each spliceosome cycle, the components are disassembled and recycled for future use.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Simplified splicing cycle showing main rearrangements of the small nuclear ribonucleoproteins and interactions of key splicing factors. Boxes show other splicing factors involved in each stage. ss, splice site.

The first stage of the splicing cycle is spliceosome assembly. Initially, the U1 snRNP base pairs with the 5′ splice site (ss) of an intron and U2 snRNP associates with the branch site close to the 3′ ss. This defines the boundaries of the intron and results in the formation of complex A, also known as the pre-spliceosome. Meanwhile, U4, U6 and U5 snRNPs preassemble to form the U4/U6-U5 tri-snRNP. The tri-snRNP is the central unit in the spliceosome and is necessary for spliceosome activation. The successful integration of tri-snRNP with complex A forms precatalytic B complex.

The second stage of the splicing cycle is activation. Dissociation of the U1 from the 5′ ss and U4 snRNP from the precatalytic B complex activates the spliceosome and gives rise to the catalytic B (B^cat) complex. Dissociation of U4 from U6 allows U6 snRNP to refold and create the active site necessary for splicing catalysis.

The catalytic splicing step of spliceosome activity involves two transesterification reactions. A DEAH box containing RNA helicase, Prp2, catalyses the first splice at the 5′ ss and formation of a lariat structure.7 This reaction yields the catalytic C complex, which in turn catalyses the splicing at the 3′ ss. Finally, the two free exons are ligated, with the 3′-hydroxyl group of the 5′ exon nucleophilic attacking the 3′ ss. The intron is released and the remaining snRNPs disassemble. The free snRNPs are trafficked into Cajal bodies in nucleus to be recycled for additional rounds of splicing.8

PRPF31 plays an essential role in the splicing cycle. PRPF31 acts as a bridging protein between the U4/U6 di-snRNP and the U5 snRNP, allowing formation and stabilisation of the tri-snRNP.9 This is facilitated by prior binding of U4-15.5K protein to the 5′ stem-internal loop-stem structural motif of the U4 snRNP.10 11 PRPF31 subsequently links the U4/U6 di-snRNP with the U5 snRNP, forming the 25S U4/U6.U5 tri-snRNP. PRPF31 is phosphorylated during spliceosome assembly, preceding formation of complex B—this phosphorylation significantly adds increased stability of the tri-snRNP complex.12 PRPF31 also interacts with other splicing factors (figure 3). Critically, PRPF31 interacts with the U5-associated protein, PRPF6, during the process of tri-snRNP formation.6 13 PRPF31 also interacts directly with PRPF3 and PRPF4, and indirectly with PRPF8 and SNRNP200.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Interactions of PRPF31 with other splicing factors. Solid lines indicate strong interactions, while dashed lines indicate weaker interactions (data on interactions and strength from STRING interaction network; http://version10.string-db.org/).

Mutations in PRPF31 cause defective splicing. There is an approximately 50% decrease in the quantitative assembly of tri-snRNPs and the correctly assembled tri-snRNPs are less efficient.14 Furthermore, PRPF31 knockdown by RNA interference leads to inhibition of tri‐snRNP formation, and stable U5 mono-snRNPs and U4/U6 di-snRNPs accumulate within the Cajal bodies.15

PRPF31 mutations in RP

Mutations in PRPF31 have been implicated as a major cause of adRP. RP is the umbrella term for one type of retinal dystrophy, which is characterised by degeneration of the retinal rod photoreceptor cells, followed by the cone cells. Clinically, the loss of rod function leads to nyctalopia (night blindness) and progressive constriction of the visual fields (tunnel vision). Later in the disease process, cone cell death causes loss of central vision—eventually leading to severe visual impairment in many cases. RP can be inherited by all modes of inheritance; autosomal dominant accounts for 30%–40% of cases, autosomal recessive for 50%–60% and X-linked recessive for 5%–15%. Non-Mendelian inheritance, such as digenic and mitochondrial inheritance, has also been reported but these represent only a small proportion of cases.16 adRP is the second most common inheritance pattern of RP, accounting for 30%–40% of cases. The disease is typified by genetic heterogeneity and, to date, 27 genes and one locus have been identified as causative of adRP.17

A major adRP locus was identified at chromosome 19q13.4 and designated RP11; subsequently, PRPF31 was identified as the causative gene underlying this locus.18 19 Mutations in PRPF31 account for 5%–10% of adRP cases, and mutations in the gene have been observed in European and Asian populations.20–22 There is wide heterogeneity of mutations, with all varieties of mutations having been identified, including non-sense, missense, deletions, insertions and ss mutations, all leading to loss of function of PRPF31. A consistent feature of PRPF31-associated adRP is the unusual pattern of inheritance: affected families show non-penetrance, with some mutation carriers being visually impaired, whereas other family members harbouring an identical mutation are normally sighted.

It is thought that non-penetrance is due to a phenomenon termed ‘variant haploinsufficiency’.23 In the normal population, there are differentially expressed PRPF31 alleles, both higher expressivity alleles and lower expressivity alleles. If a mutant allele and a higher expressivity wild-type allele are inherited, residual protein level is sufficient for normal retinal function. If, however, a mutant allele and a lower expressivity allele are coinherited, PRPF31 level falls beneath the threshold required for normal retinal function and clinically manifest disease results.

In the general population, expression of PRPF31 follows a continuous distribution, with an approximately fivefold change between the lowest and the highest expression levels.24 Several studies have confirmed that the expression of wild-type PRPF31 is significantly higher in asymptomatic, as compared with symptomatic, RP11 mutation carriers.25–27 Given the association of gene expression levels and disease phenotype, research has focused on identifying genetic factors that influence gene expression of the wild-type PRPF31 gene. Critically, linkage analysis of asymptomatic–symptomatic sibling pairs in families affected by PRPF31-adRP has demonstrated that non-penetrance is strongly linked to chromosome 19q13, in a 1 megabase interval between microsatellite markers D19S572 and D19S926—which contains the wild-type PRPF31 gene.28 29 It was proposed, therefore, that there are cis-acting factors in close proximity to PRPF31, which act to alter expression of the gene.

One gene that lies in close proximity to PRPF31 is CNOT3, and it has been demonstrated that the expression of CNOT3 was inversely proportional to the expression of PRPF31.30 It was also demonstrated that CNOT3 protein could bind the PRPF31 core promoter and that suppression of CNOT3 by siRNA leads to a significant increase in PRPF31 expression and it was, therefore, concluded that CNOT3 is a negative regulator of PRPF31 expression.30 CNOT3 protein is a component of the Ccr4-Not complex, an evolutionary conserved multimeric structure involved in modulation of gene expression.31 32 The apparent paradox between the strong cis-acting linkage data and the trans-acting nature of modulation by CNOT3 has been explained by the theory of ‘linked trans-acting epistasis,’ where the trans-acting factor is not meiotically separated from the target due to close genetic distance33

A second factor thought to affect the expression of PRPF31 is copy number variation (CNV) of MSR1 repeat elements. A cluster of MSR1 repeats lies approximately 600 bp upstream to the PRPF31 transcription start site; with three copies or four copies of MSR1 being normal variants in the European population.34 It was demonstrated that CNV of MSR1 significantly altered luciferase reporter activity, and that the higher expressivity four-copy allele was significantly under-represented in a cohort of symptomatic PRPF31 mutation carriers.34 Finally, a number of trans-acting factors might also affect gene expression; in particular, an eQTL on chromosome 14 was identified that might play a role and this region contained several transcription factors and genes with a role in retinal biology which were considered interesting candidates.35