A Homozygous Splice Mutation in the HSF4 Gene Is Associated with an Autosomal Recessive Congenital Cataract

Abstract

purpose. To map the locus and identify the gene causing autosomal recessive congenital cataracts in a large consanguineous Tunisian family.

methods. DNA was extracted from blood samples from a large Tunisian family with an autosomal recessive, congenital, total white cataract. A genome-wide scan was performed with microsatellite markers. All exons and the splice sites of the HSF4 gene were sequenced in all members of the Tunisian family and in control individuals. RT-PCR was used to detect different transcripts of the HSF4 gene in the human lens. The transcripts were cloned in a TA cloning vector and sequenced.

results. Two-point linkage analyses showed linkage to markers on 16q22 with a maximum lod score of 17.78 at θ = 0.01 with D16S3043. Haplotype analysis refined the critical region to a 1.8-cM (4.8-Mb) interval, flanked by D16S3031 and D16S3095. This region contains HSF4, some mutations of which cause the autosomal dominant Marner cataract. Sequencing of HSF4 showed a homozygous mutation in the 5′ splice site of intron 12 (c.1327+4A→G), which causes the skipping of exon 12. A more detailed study of the transcripts resulting from alternative splicing of the HSF4 gene in the lens is also reported, showing the major transcript HSF4b.

conclusions. This is the first report describing association of an autosomal recessive cataract with the HSF4 locus on 16q21-q22.1 and the first description of HSF4 splice variants in the lens showing that HSF4b is the major transcript.

Congenital cataracts are a significant cause of visual disease, being responsible for approximately one third of blindness in infants. Without prompt treatment, cataracts may interfere with sharp imaging on the retina, resulting in failure to develop normal retinal–cortical synaptic connections and placing children at risk of irreversible visual loss. Approximately 50% of congenital cataracts are inherited. ^{1 2} Congenital cataracts, isolated or associated with other abnormalities, have been mapped to 26 genetic loci. Eighteen genes have been identified at these loci, most of which belong to five groups. The major cataract genes encode (1) the crystallin families: CRYAA (online Mendelian Inheritance in Man [OMIM] 123580), CRYBA3/A1 (OMIM 123610), CRYBB2 (OMIM 123620), CRYGC (OMIM 123680), CRYGD (OMIM 123690), CRYAB (OMIM 123590), and CRYBB1 (OMIM 600929); (2) the gap junctional channel proteins: GJA3 (OMIM 121015) and GJA8 (OMIM 600897); (3) the major intrinsic protein of the lens AQP0 (OMIM 154050); (4) the intermediate filament (IF) family of proteins ³ beaded-filament structural protein 2 BFSP2 (OMIM 603212); and (5) the regulators of gene transcription: PITX3 (OMIM 602669), PAX6 (OMIM 607108), FOXE3 (OMIM 601094), EYA1 (OMIM 601653), MAF, ^{4 5} and the more recently identified HSF4 (OMIM 602438). 

The most common mode of inheritance for nonsyndromic cataracts is autosomal dominant, but four loci for autosomal recessive have been reported, including the CRYAA gene on 21q22.3 ⁶ ; the LIM2 gene on 19q13.4 ⁷ ; 9q13-q22 ⁸ ; and a locus on the short arm of chromosome 3. ⁹ Recently a new gene, HSF4, which caused autosomal dominant congenital cataract in a Chinese family and in the Marner-bearing families and has been mapped to 16q21-q22.1, ^{10 11} has been identified. ¹² Four different missense mutations have been reported in sporadic and familial cases, all occurring in the highly conserved DNA-binding domain. The cataract phenotypes described vary in age of onset and the severity and morphology in different families and even within the same family. 

HSF4 belongs to the family of heat shock transcription factors that regulate the expression of heat shock proteins (Hsps) in response to different stresses, such as oxidants, heavy metals, elevated temperature, and bacterial and viral infections. Trimerization of HSFs is mediated by the hydrophobic heptad repeats (HR-A/B) characteristic of helical coiled-coil structures. ¹³ Suppression of HSF trimerization is likely to be mediated by another region of hydrophobic heptad repeats (HR-C). Because they lack HR-C, the HSF4 proteins are multimeric. ^{14 15} Two isoforms, HSF4a and HSF4b, resulting from two alternative splice sites for exons 8 and 9 and leading to two different amino acid sequences, have been reported for HSF4. HSF4a actively represses transcription of other heat shock factor genes by binding directly to the heat shock element (HSE). HSF4b, which contains 30 additional amino acids, acts as an activator of transcription. It has been demonstrated that the additional 30 amino acids are responsible for this activity. ^{16 17} Herein, we describe a splice mutation in intron 12 of the HSF4 gene that is associated with autosomal recessive congenital cataracts. We further investigated the alternative splicing leading to various isoforms of HSF4 in the eye lens. 

Materials and Methods

DNA samples were collected from a consanguineous Tunisian family (30016, Fig. 1 ) of 63 individuals, of whom 22 were affected. Informed consent was obtained from each individual studied. This study was approved by the ethics review board of Charles Nicolle Hospital, Tunis, and the National Eye Institute’s Institutional Review Board and is consistent with the provisions of the Declaration of Helsinki. 

Phenotype

Patients were examined with a slit lamp before and after dilatation. Visual acuity was assessed in all patients. All patients with congenital cataract showed total and bilateral cataract (Fig. 2) associated with nystagmus but without any other ocular anomaly. In some cases, the cataract was regressive; one eye had undergone surgery. Ocular examinations of the parents of these patients revealed clear lenses. 

Linkage Analysis

DNA was extracted directly from blood by standard phenol-chloroform protocols. ¹⁸ A genome-wide scan was performed with 382 fluorescently labeled microsatellite markers (Linkage Mapping Set MD-10, Applied Biosystems, Inc., [ABI] Foster City, CA). Multiplexed PCR was performed as described. ¹⁹ Two-point linkage analyses were performed with the FASTLINK implementation of the MLINK program of the LINKAGE program package. ^{20 21} For screening, equal allele frequencies were used for all markers. For fine mapping the allele frequencies were calculated using 30 Tunisian control individuals. 

Mutation Analysis

For screening genes in the linked interval, coding exons and their splice sites were amplified from genomic DNA of two affected patients and one unaffected individual. Reference genomic sequences are available from GenBank (accession no. NT_010498). PCR amplification of 13 exons using primers listed in Table 1 was performed in 20-μL reactions. Exons 1, 2, 3, 4, 5, and 6 were amplified with Taq polymerase (LATaq; Takara Bio, Inc., Shiga, Japan). Exons 7, 8, 9, 10, 11, 12, and 13, were amplified with 2.5 IU Taq (Taq Gold; ABI), 10 × 2 μL buffer (Taq Gold; ABI), 1.9 mM MgCl₂, 0.25 mM dNTP, 0.25 μM primers, and 100 ng genomic DNA. PCR cycling consisted of an initial 10-minute denaturation step of 95°C for 5minutes; 32 cycles of 94°C for 40 seconds, 55°C for 30 seconds, and 72°C for 30 seconds; and a final elongation step at 72°C for 5 minutes. PCR products were purified with either of two kits (PCR purification kit; Qiagen, Valencia, CA; or a QuickStep2 PCR purification kit; Edge Biosystems, Gaithersburg, MD). The PCR template was sequenced using dye-termination chemistry (Big Dye Terminator, ver. 1; ABI) in a final reaction of 10 μL. Products of the sequencing reactions were purified with a gel-filtration system (Performa DTR System; Edge Biosystem) and run on a gene analyzer (model 3100; ABI). Sequences were aligned on computer (Seqman; DNAStar, Inc., Madison, WI). The same approach was used to sequence exon 12 and the flanking intronic sequences of the HSF4 gene from DNA from all family members and 175 control individuals including 87 Tunisians, 41 Barbadians, and 47 Indians. DNA from Tunisian individuals was collected from the genetics department at Charles Nicolle Hospital, and the remaining control DNA samples were available through studies performed at the National Eye Institute and collaborating institutions. All subjects gave their consent for anonymous use of their DNA samples. 

Splice Site Analysis

Computer analysis of the strength of splice sites of genomic sequences spanning relevant nucleotides was performed with the Delila package. ²² These programs use an information theory–based model of donor and acceptor splice sites to assign a bit value as a representation of the strength of interaction between the splice site and the spliceosome for both the wild-type and the mutant splice sites. 

Expression of HSF4 in the Human Lens and in Lymphoblastoid Cell Lines

Total RNA was extracted from cadaveric human lenses (Trizol; Invitrogen, Carlsbad, CA), in accordance with the manufacturer’s protocol. First-strand synthesis was performed with oligo dT primers and an RT-PCR system (Thermoscipt; Invitrogen). PCR reactions were performed in 20-μL reactions. Exons 10 through 13 were amplified using 2.5 IU polymerase (Taq Gold; ABI), 10 × 2 μL buffer (Taq Gold; ABI) 1.9 mM MgCl₂, 0.25 mM dNTP, 0.25 μM primers, and 1 μL cDNA. After a 2-minute denaturation step at 95°C, 40 cycles consisting of 94°C for 40 seconds, 60°C for 30 seconds, 72°C for 30 seconds, and a final elongation 72°C for 5 minutes were performed. A similar approach was used to characterize of the effect of the intron 12 c.1327+4A→G mutation on mRNA splicing in human lymphoblastoid lines isolated from patients. 

Amplification of exons 1 through 11 in human lens cDNA (Table 2) was performed (LATaq; Takara Bio, Inc.) using conditions described by the manufacturer. The PCR cycle consisted of an initial step of denaturation at 95°C for 1 minute, followed by 40 cycles of 94°C for 40 seconds, 62°C for 30 seconds, and 72°C for 2 minutes and a final elongation 72°C for 10 minutes. The PCR product was subcloned in a TA cloning vector (PCR2.1; Invitrogen) and then sequenced. 

Results

Identification of a Homozygous Mutation in HSF4

Evidence of significant linkage was first observed with D16S503 and D16S512, which gave maximum lod scores of 9.42 at θ = 0.03 and 5.94 at θ = 0.05, respectively (Table 3) . The maximum two-point lod score was 17.78 at θ = 0.01 with D16S3043. Homozygosity mapping narrowed the linked region to 1.8 cM between D16S3095 (individual 157) and D16S3031 (individual 124; Fig. 1 ). This critical interval contains HSF4, some mutations of which have recently been shown to cause autosomal dominant Marner cataracts. ¹²  

Sequencing of the coding region of the HSF4 gene showed no mutations or polymorphisms in affected family members. However, a homozygous A→G change at position 4 of the donor site consensus sequence of intron 12 (c.1327+4A→G; GenBank AB029348; Fig. 3 ) was identified in the two affected individuals and then confirmed in all remaining affected individuals belonging to family 30016. This change was not seen in 175 control individuals comprising 350 chromosomes from 87 Tunisians, 41 Barbadians, and 47 Indians. Information analysis using the program Delila ²² showed a decrease from 8.0 bits in this donor splice site to 5.4 bits in the mutant. Although this would usually correspond to a leaky mutation, the incomplete splicing shown by the native site (described later) may increase the susceptibility of this donor splice site to even a relatively subtle mutation. 

RT-PCR of lymphoblast cell lines mRNA from unaffected individuals using primers 4231fw and 4232rv (Table 1) , which amplify sequences transcribed from exons 10 through 13, showed two bands of 309 and 379 bp (Fig. 4A) . The lower band lacks exon 12, which was confirmed by direct sequencing after agarose gel purification. Affected individuals showed a single band of 309 bp, corresponding to the transcript lacking exon 12, also confirmed by sequencing. Translation of HSF4b mRNA lacking exon 12 is predicted to result in a protein consisting of the first 419 amino acids preceding exon 12, followed by a frameshift resulting in 28 novel amino acids followed by a premature stop codon, so that it is missing the C-terminal 74 amino acids. 

Characterization of HSF4 Transcripts in the Lens

RT-PCR amplification of HSF4 in the human lens (Fig. 4B) using primers 4369fw and 4372rv (Table 2) , shows, besides the two already identified transcripts HSF4a and HSF4b, four additional transcripts—HSF4c, HSF4d, HSF4g, and HSF4h—and three apparently aberrantly spliced mRNAs (Fig. 5) . Of 14 clones sequenced, 6 were HSF4b, confirming that HSF4b is the dominant product in the lens, as has been described in other tissues. All mRNAs contain exons 1 through 3, which encode the HSF4 DNA-binding domain. 

HSF4a and HSF4b differ at exons 8 and 9. Exon 8a is 139 bp, and exon 8b has lost 14 bp in the 5′ end and is 125 bp, whereas exon 9a is 124 bp and exon 9b has an additional 104 bp at the 5′ end and is 228 bp. Exons 4 and 5, which code for the hypervariable region (HRA/B) are missing from HSF4c mRNA, which has exons 8b and 9b. HSF4d has one additional exon compared with HSF4a, exon 7b which is 100 bp in size and located at the 3′ end of exon 7, whereas exon 8d is 96 bp, missing 43 bp of the 5′ sequence of exon 8a. HSF4g mRNA lacks 24 bp in the downstream part of exon 5 and all exons 6 and 7, but contains 49 bp of the 3′ sequence of exon 7b, as well as exons 8d and 9a. This mRNA encodes a protein of 401 amino acids. HSF4h mRNA is similar to HSF4d, except that it contains exons 9a and 8h which is 115 bp, missing 10 bp of the 5′ sequence of exon 8a, and so encodes a protein with 523 amino acids. The mRNAs for HSF4e, -f, and -i have early stop codons in exons 8, 7, and 8, respectively. Thus, these mRNAs probably represent aberrant splicing. HSF4 belongs to the HSF family, which has an amino-terminal helix-turn-helix DNA-binding domain, the most conserved functional domain of HSFs. 

Discussion

This is the first report of an association of an autosomal recessive congenital cataract with the HSF4 gene, which has been reported to be responsible for autosomal dominant cataracts, including the Marner cataract. The association of the HSF4 gene with two different modes of Mendelian inheritance for congenital cataract can be explained by the location and the severity of the mutations in both cases. In the Marner cataract, the reported mutations are missense changes located in exons 1 and 3 of the HSF4 gene. These exons correspond to the HSF DNA-binding domain and possibly cause a dominant negative effect with interference of normal and mutant proteins. Alternatively, these aberrant HSF4 proteins may activate novel genes. 

In the autosomal recessive cataract reported herein, the mutation is located in the 5′ splice-site of intron 12, causing skipping of exon 12 of the HSF4 mRNA and creating a frame shift. This may relate to the severe splicing error seen experimentally relative to that predicted by the Delila program. Donor splice sites show a mean information content of 7.92 ± 0.09 bits, and acceptor sequences show a mean of 9.35 ± 0.12 bits. Usually, sites with less than 2.4 bits of information are not spliced, whereas sites with content less than the normal information content but more than 2.4 bits result in “leaky” or partial splicing. ²² The exon 12 splice site may be unusually susceptible to sequence changes, because of its relatively low starting information content, consistent with the leaky splicing seen in the normal gene. 

This mutation is predicted to cause a premature stop codon, resulting in an inactivation of the protein translation or the production of an abnormal protein, with a complete loss of function of the aberrant HSF4 protein in affected homozygotes. This causes a severe phenotype autosomal recessive with congenital cataracts. This contrasts with the Marner and Chinese families with HSF4 mutations, in which the cataracts appear after 15 months of age. 

In the human, HSF4 is widely expressed, especially in the heart, brain, skeletal muscle, lung, and pancreas. ^{16 17} Alternative splicing is a common feature in HSFs, and tissue-specific splicing patterns have been reported for HSF1 and HSF2. ^{23 24}  

In summary, we have shown that a splice mutation in HSF4 is associated with autosomal recessive total cataract and have described complex and variable splicing of the HSF4 mRNA in the human lens. It will be interesting to study the expression of HSF4 isoforms during normal development in the lens to assess whether there is a predominant isoform, as there is for HSF2. HSF4 is only the second gene after CRYAA in which mutations have been reported to cause both autosomal dominant and recessive congenital cataract in humans. 

Databases

The following databases were used in the course of the study:

•  
  Delila Program, Laboratory of Computational and Experimental Biology, National Cancer Institute, National Institutes of Health (Bethesda, MD): http://www.lecb.ncifcrf.gov/∼toms/
•  
  GenBank: HSF4a NM_001538 (human), HSF4b AB029348 (human): http://www.ncbi.nlm.nih.gov/Genbank; provided in the public domain by the National Center for Biotechnology Information (NCBI; Bethesda, MD)
•  
  LINKAGE: provided in the public domain by the Human Genome Mapping Project Resources Center (Cambridge, UK): http:www.hgmp.mrc.ac.uk/
•  
  Online Mendelian Inheritance in Man (OMIM): http://www.ncbi.nlm.nih.gov/Omim/Genethon/; http://www.genethon.fr/ provided in the public domain by NCBI (Bethesda, MD); or by NCBI through Gènèthon, hosted by the French Association against Myopathies (Evry, France)

 Submitted for publication December 17, 2003; revised March 3, 1004; accepted March 8, 2004.

 Disclosure: N. Smaoui, None; O. Beltaief, None; S. BenHamed, None; R. M’Rad, None; F. Maazoul, None; A. Ouertani, None; H. Chaabouni, None; J.F. Hejtmancik, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Pedigree of family 30016. Another version is available online at www.iovs.org/cgi/content/full/45/8/2716/DC1.

Figure 2.

 

Photograph of a patient with a congenital white total cataract.

Table 1.

 

PCR Primers for Amplification of HSF4, 13 Coding Exons, and the Splice Sites

Exon	Primer Name	Primer Sequence	Size of PCR Product (bp)
Exon 1–6	HSF4_1-6Fw	5′-CCTCCACTCCACTCCACTCCACAC-3′	2083
	HSF4_1–6Rv	5′-TAGCGCAGGGGGAGGAATGAGGAT-3′
Exon 7–8	HSF4_7-8Fw	5′-GGTTCCCTCCCTCCCCTCTC-3′	766
	HSF4_7-8Rv	5′-CTCCCTCTACCCCTACAGCCATCT-3′
Exon 9	HSF4_9Fw	5′-GGCTGTAGGGGTAGAGGGAGAAGT-3′	405
	HSF4_9Rv	5′-ATGCTGGGGCCAAGAAATCA-3′
Exon 10–11	HSF4_10-11Fw	5′-GGGAACTTAATGCGGGGTGGACA-3′	456
	HSF4_10-11Rv	5′-GAGGGGGACAGCTGGGAAATCA-3′
Exon 12–13	HSF4_12-13Fw	5′-TGGGTCGGGGAGGGCAGAG-3′	811
	HSF4_12-13Rv	5′-GCAGGGATAGTCGGGGTAGTGGAG-3′

Table 2.

 

PCR Primers Used to Amplify the cDNA of the HSF4 Gene

Exon	Primer Name	Primer Sequence	HSFb Expected Size (bp)
Exon 1–11	4369fw	5′-TTCCTCGGCAAGCTATGGGCGCTGGTGGGG-3′
	4372rv	5′-ATCAAGGACAGCTCCATGTCCAAGTC-3′	1199
Exon 10–13	4231fw	5′-TGCCTCCGATGCTGCTTC-3′
	4232rv	5′-GTTCCAAGCCCCTTCAGA-3′	379

 fw, forward; rv, reverse.

Table 3.

 

Two-point Lod Scores for Family 30016 with Locus 16q21-q22.1

cM (16qcen)	Markers	0	0.01	0.05	0.1	0.2	0.3	0.4	Z max	θ
81.9	D16S503	−∞	9.16	9.18	8.09	5.46	3.01	1.10	9.42	0.03
83.1	D16S3043	16.43	17.78	16.78	14.92	10.80	6.47	2.60	17.78	0.01
83.7	D16S3050	11.60	13.00	12.21	10.65	7.30	4.13	1.60	13.00	0.01
84.4	D16S3019	12.86	12.54	11.26	9.65	6.48	3.63	1.40	12.86	0.00
84.4	D16S3067	12.98	12.63	11.25	9.53	6.23	3.34	1.20	12.98	0.00
85.5	D16S3096	13.60	13.31	12.10	10.51	7.22	4.06	1.50	13.6	0.00
85.5	D16S496	16.61	16.26	14.79	12.86	8.90	5.12	2.00	16.61	0.00
85.5	D16S3095	5.36	5.21	4.60	3.88	2.56	1.46	0.60	5.36	0.00
88.0	D16S3066	−∞	14.4	13.50	11.78	8.04	4.45	1.60	14.4	0.01
88.1	D16S512	−∞	4.96	5.94	5.48	3.82	2.14	0.80	5.94	0.05
90.2	D16S515	−∞	2.22	5.22	5.53	4.43	2.82	1.30	5.56	0.09

Figure 3.

 

Mutation in the 5′ splice site of intron 12 (arrow) homozygote substitution of A→G at the fourth nucleotide of the splice site (c.1327+4A→G).

Figure 4.

 

(A) Expression of HSF4 in human lymphocytes. Lane 1: affected individual. Only one band (309 bp) is present from which exon 12 (70 bp) is absent. Lane 2: normal individual. The brighter top band corresponded to the amplification of exons 10-11-12 and 13. The primers (4231fw and 4232rv) used for this RT-PCR are listed in Table 2 . (B) Lane 1: expression of HSF4 in human lens. At least three bands were observed, although the 974-bp band dominated. This PCR product was subcloned in a TA cloning vector, and nine different mRNAs were detected with the primers 4369fw and 4372rv (Table 2) , which amplify the cDNA from exon 1 to 11.

Figure 5.

 

Transcripts of the HSF4 gene in the lens. First row: exon number; second row: exon size in base pairs; remaining rows: the different HSF4 mRNAs expressed in the lens. The schematic indicates the domains of the HSF4 protein and the position of mutations reported to cause the autosomal dominant and recessive cataracts.

The authors thank the members of the cataract family for their participation in this study and Tom Schneider, Peter Rogan, and Jim Ellis for advice and discussions concerning the Delila splice site analysis.