|Home | About | Journals | Submit | Contact Us | Français|
Brain malformations are individually rare but collectively common causes of developmental disabilities1–3. Many forms occur sporadically and have reduced reproductive fitness, pointing towards a causative role for de novo mutations4,5. Here we report our studies of Baraitser-Winter syndrome, a well-defined syndrome characterized by distinct craniofacial features, ocular colobomata and a neuronal migration defect6,7. By using whole-exome sequencing in three proband-parent trios, we identified de novo missense changes in the cytoplasmic actin genes ACTB and ACTG1 in one and two probands, respectively. Sequencing of both genes in fifteen additional patients revealed disease-causing mutations in all probands, including two recurrent de novo mutations (ACTB p.Arg196His and ACTG1 p.Ser155Phe). Our results confirm that trio-based exome sequencing is a powerful approach to discover the genes causing sporadic developmental disorders, emphasize the overlapping roles of cytoplasmic actins in development, and suggest that Baraitser-Winter syndrome is the predominant phenotype associated with mutations of these two genes.
Whole-exome sequencing has recently been used to detect disease-causing mutations underlying rare sporadic syndromes in just a few unrelated individuals8–10. However, this becomes more difficult in the presence of any locus heterogeneity. For disorders hypothesized to result from de novo mutations, a potentially more powerful alternative approach involves exome sequencing in parent-child trios, which should rapidly narrow the number of candidate mutations to a handful in view of the rarity of de novo events in protein coding sequences11,12. This strategy has been successfully applied to identify de novo point mutations in intellectual disability4, autism13, and schizophrenia14,15.
Baraitser-Winter syndrome is a rare but well-defined developmental disorder recognized by the combination of congenital ptosis, high-arched eyebrows, hypertelorism, ocular colobomata and a brain malformation consisting of anterior predominant lissencephaly. Other typical features include postnatal short stature and microcephaly, intellectual disability, seizures and hearing loss6,7,16–18. Neither familial recurrence nor consanguinity have been observed in any families including the 18 reported here (Supplementary note), and chromosome microarrays have not detected any pathogenic copy number variants (CNVs). We therefore hypothesized that the genetic basis of Baraitser-Winter syndrome was likely to result from de novo point mutations and performed whole-exome sequencing in three probands (patient LP98-083 is shown in Fig. 1) and their unaffected parents using two different platforms. Exome capture and sequencing were performed using NimbleGen solution-based capture and Illumina sequencing for trio 1 (LP98-083), and Agilent SureSelect target enrichment system with ABI SOLiD sequencing for trios 2 (58248) and 3 (58431) (Online Methods). Both platforms generated a coverage of at least ten reads for more than 85% of the targeted exome and we identified 22,591 to 29,685 genetic variants per proband (Table 1). As previously described4,13, we filtered variants to systematically identify candidate de novo events in each proband. Given the severe phenotype, we focused on protein-altering and splice-site variants absent from other exome datasets available locally or from the dbSNP or 1000 Genomes Project databases19. Similar to previous studies4,13–15, we identified two to six candidate de novo mutations per proband (Table 1 and Supplementary Table 1), all of which were tested by Sanger sequencing.
The two sequencing and analysis platforms converged to identify de novo missense changes in the cytoplasmic γ-actin gene ACTG1 (NM_001614.2) in two probands and in the β-actin gene ACTB (NM_001101.2) in the third (Supplementary Fig. 1). We used Sanger sequencing to screen the coding sequence of both genes in 15 additional patients and detected pathogenic mutations in all. Altogether, we found 10 and 8 mutations in ACTB and ACTG1, respectively (Table 2). These proved to be de novo in all 11 subjects with parental DNA available, and 6 of 7 remaining mutations were identical to mutations shown to be de novo in the first group. Strikingly, 8 patients (44%) carried a mutation disrupting Arg196 of β-actin, including 7 with the same nucleotide change (c.587G>A, p.Arg196His). These recurrent transitions occurred at a CpG dinucleotide, which is known to be susceptible to deamination of methylcytosines20, thus suggesting a possible mechanism for the high frequency of this mutation. Another 3 (17%) had a mutation disrupting Ser155 of cytoplasmic γ-actin.
Several control datasets were analyzed to validate our findings and assess the number of coding variants in ACTB and ACTG1. These included whole-exome sequence data from 244 unrelated healthy individuals of European descent (Online Methods and Supplementary Table 2) and genome-wide sequence data from 629 and ~2,500 individuals available from the 1000 Genomes Pilot Project19 and NHLBI Exome Sequencing Project (ESP) websites (Supplementary Tables 3 and 4). None of the mutations identified in Baraitser-Winter syndrome patients were present in this large dataset. Further, no CNVs encompassing ACTB or ACTG1 exons were observed in 2,349 control individuals (Online Methods). The ESP dataset showed a striking ratio of synonymous to non-synonymous substitutions of 79:1 for the covered portion of ACTB and ACTG1 combined. On a genome-wide scale19, this ratio is estimated at ~1:1, thus indicating strong selection against non-synonymous mutations in both genes. Surprisingly, one of the ACTG1 de novo mutations (p.Ala135Val) was recorded in dbSNP (rs11549190, Table 2). This variant was originally identified, but not validated, from large-scale sequencing of expressed sequence tags21. Given its absence in the aforementioned datasets, we concluded that this report was likely an error in dbSNP.
Actins are a family of essential cytoskeletal proteins implicated in nearly all cellular processes22–25. Among the six human genes encoding actins, only ACTB and ACTG1 are ubiquitously expressed; the remaining four are expressed primarily in muscle. The cytoplasmic β- and γ-actins are highly conserved throughout evolution and nearly identical to each other, differing by only four amino acids26 (Supplementary Fig. 2). Despite their striking homology and presence in all cell types27, β- and γ-actins have been proposed to play at least partially distinct physiological functions based on their relative spatial and temporal enrichment in neurons28 and other cell types29,30, and on different phenotypes observed in homozygous mouse mutants. Actb−/− is embryonically lethal, while Actg1−/− shows reduced viability with some surviving to adulthood25,31,32. Previously reported heterozygous ACTB and ACTG1 mutations in humans also suggest a more severe phenotype for ACTB mutations33–35 (phenotype reviewed in Supplementary note and Supplementary Table 5). In contrast, our findings suggest that both isoforms have significant overlapping but non-redundant functions during human development, as illustrated by the indistinguishable clinical presentation of Baraitser-Winter syndrome patients carrying mutations in ACTB and ACTG1 (Table 3 and Supplementary Table 6).
Protein levels and morphology of lymphoblastoid cell lines from six affected individuals were examined, representative examples of which appear in Fig. 2 and and3.3. No change in protein levels of β- and γ-actin were seen between these two cell lines (Fig. 2) or four additional mutant lines and controls (Supplementary Fig. 3), similar to one previously reported ACTB mutation34. Both mutant lines derived from patients carrying one or the other of the two ACTB and ACTG1 recurrent mutations (p.Arg196His and p.Ser155Phe, Fig. 3) contained greatly increased F-actin content and multiple anomalous F-actin rich filopodia-like protrusions compared to control cells, resulting in an increased cell perimeter (Fig. 3b). The stability of F-actin in these cells was probed using latrunculin A, which binds actin monomers thereby preventing their incorporation into growing filaments36. Both mutant lines showed altered sensitivity to latrunculin A, albeit with different outcomes. We observed an increased resistance to latrunculin A in β-actin p.Arg196His mutant cells, which is consistent with the previously reported p.Arg183Trp mutation34, whereas the γ-actin p.Ser155Phe mutant displayed increased sensitivity to the treatment (Fig. 3c). Immunofluorescent staining of F-actin in lymphoblastoid cell lines from four additional patients showed patterns of cytoskeletal changes and abnormal accumulation of F-actin that were reproducible between independent cell lines carrying the same mutation (p.Arg196His in LR04-173 and LR09-079; p.Ser155Phe in LP98-096 and LR04-298, Fig. 3 and Supplementary Fig. 4). Consistent with our initial analysis of p.Arg196His and p.Ser155Phe mutant lines, all mutants examined presented striking yet distinct patterns of abnormal F-actin organization, suggesting that morphologies are mutation specific. Altogether, our data indicate that mutations associated with Baraitser-Winter syndrome result in increased F-actin content and altered F-actin dynamics in response to latrunculin A treatment. These effects should impact cell morphology, motility, and other actin-related functions.
Several observations support a dominant-negative or gain-of-function mechanism for the disease-causing mutations. Firstly, none of our subjects had deletions or protein-truncating mutations, which are an indication of haploinsufficiency. Further, 11 of 18 mutations (61%) disrupt the same two amino-acids (ACTB Arg196 and ACTG1 Ser155). Although these could be hypermutable sites, mutation clustering suggests a gain-of-function effect, as observed in Noonan37, FGFR-related craniosynostosis38 and Proteus syndromes39. Second, patients with complete deletion or duplication of ACTB do not have Baraitser-Winter syndrome. To assess the latter, we searched the databases of two large clinical laboratories performing chromosome microarrays, Signature Genomic Laboratories and the Department of Human Genetics, Radboud University Nijmegen Medical Centre, and ascertained four children: two with a deletion and one with a duplication of 7p22 that included ACTB, and one with a deletion of 17q25.3 that included ACTG1. All lacked the major and most severe features of Baraitser-Winter syndrome, which are congenital ptosis and high-arched eyebrows, neuronal migration malformation and colobomata (Supplementary note and Supplementary Fig. 5). Finally, the unchanged β- and γ-actin protein levels coupled with increased F-actin content observed in patient-derived lymphoblastoid cell lines also argue against haploinsufficiency.
In conclusion, our results show that Baraitser-Winter syndrome is a clinically well-defined syndrome caused by mutations in the cytoplasmic actin genes ACTB and ACTG1 and provide additional evidence that actin plays a central role in the pathogenesis of lissencephaly40,41. Our data also show that even in a scenario of limited genetic heterogeneity, whole-exome sequencing of a small number of trios combined with validation in additional patients is a powerful approach to identify mutations underlying well-defined sporadic disorders, especially those with low reproductive fitness. Because mutations of either ACTB or ACTG1 cause an indistinguishable phenotype, we hypothesize that mutations observed in Baraitser-Winter syndrome affect developmental functions shared by cytoplasmic β- and γ-actins. Functional comparison of mutations identified in Baraitser-Winter syndrome and other actinopathies should help elucidate both the distinct and overlapping functions of cytoplasmic β- and γ-actin isoforms. We conclude that Baraitser-Winter syndrome comprises the severe end of the predominant phenotype spectrum associated with mutations of both ACTB and ACTG1, a spectrum that extends from Baraitser-Winter syndrome to non-syndromic hearing loss.
Written, informed consent was obtained from all subjects prior to enrollment in the study. Experiments on human subjects were approved by institutional review boards at all participating institutions. The study included 18 unrelated patients; the diagnosis of Baraitser-Winter syndrome was based on previously reported clinical features6,7,17,18, and review of MRI brain scans when available. Genomic DNA was extracted from whole-blood or saliva using standard methods. When parental DNA was available, paternity-maternity testing was performed by genotyping a panel of polymorphic short tandem repeats.
Five micrograms of whole-blood genomic DNA from the affected proband and his parents were sent to the University of Washington Genome Sciences Genomic Resource Center for exome capture and sequencing. Whole-exome sequence capture was performed using the SeqCap EZ Exome Library v2.0 liquid phase sequence capture kit (Roche). Libraries were then sequenced on an Illumina Genome Analyzer IIx according to the manufacturer’s recommendations for paired-end 76-bp reads. On average, we generated 81 million total reads (68 million mapped reads) and 5.2 gigabases (Gb) of mappable sequence per individual. We aligned reads to the human reference genome (hg19) with the Burrows-Wheeler Aligner (BWA)42 and removed potential duplicate paired-end reads. The Genome Analysis Toolkit (GATK)43,44 was used for base quality score recalibration and indel realignment as well as for single nucleotide variant and indel discovery and genotyping using standard hard filtering parameters43. Variants with quality scores < 30, allele balance > 0.75, sequencing depth < 4, quality/depth ratio < 5.0, length of homopolymer run > 5.0, and strand bias > −0.10 were flagged and excluded from subsequent analyses. Coverage was assessed with the GATK Depth of Coverage tool by ignoring reads with mapping quality < 20 and bases with base quality < 30. 92% of the primary target was covered ≥ 4 times in all three individuals.
Massively parallel sequencing of genomic DNA from two patient-parent trios was performed at the Radboud University Nijmegen Medical Center by using the ABI SOLiD™ 4 platform (Life Technologies). Enrichment of exonic sequences was achieved by using the 50 Mb human SureSelect set (Agilent), which targets ~21,000 human genes. On average, we obtained > 120 million mappable sequencing reads and 4.6 Gb of mappable sequence data per individual after multiplex sequencing, using 4 exomes per sequencing slide. Colour space reads were mapped to the hg19 reference genome with the SOLiD bioscope software version 1.3, which utilizes an iterative mapping approach. On average, 81% of bases originated from the targeted exome, resulting in a mean coverage of 88-fold (median 66-fold). 88% of the targeted exons were covered more than 10 times. Single nucleotides variants were subsequently called by the DiBayes algorithm using high stringency calling settings and small insertions and deletions were detected using the SOLiD Small Indel Tool.
The 22,591 variants identified in the patient were annotated with SeattleSeq SNP annotation (http://snp.gs.washington.edu/SeattleSeqAnnotation131/). We focused on protein-altering variants (missense, nonsense, splice-site variants, and coding indels) absent from dbSNP (build 132), 1000 Genomes Project data19, and 101 other exomes. From this list of rare variants, we identified potential de novo events using the following parameters: presence of ≥ 2 variant reads in the proband, sequencing depth at the variant position ≥ 4 in both parents, and parental allelic balance > 0.90. Evolutionary conservation at the nucleotide level and impact of amino acid substitutions were assessed using the Genomic Evolutionary Rate Profiling (GERP)45 and Grantham matrix scores46, respectively. Candidate de novo events were then inspected using IGV47 and validated in all three individuals by standard PCR reactions using custom primers designed with Primer3 (http://frodo.wi.mit.edu/primer3/) and Sanger sequencing.
A total of 28,497 and 29,685 variants were identified in probands of trio 2 (58248) and trio 3 (58431), respectively. All variants were annotated using an in-house annotation pipeline built in Nijmegen, as previously reported4,8,9,48. These variants were filtered and analyzed for potential de novo occurrence using an automated prioritization scheme4 involving an automated check of all private non-synonymous variants identified in the patients for occurrence in the respective parental bam files. Candidate de novo events were then verified using conventional Sanger sequencing.
For trio 1, variants were compared against 101 other exomes sequenced at the University of Washington, including 88 exomes from the NIEHS SNPs project (http://snp.gs.washington.edu/niehsExome/), and 13 exomes sequenced in the same batch as the trio and consisting of healthy individuals and patients with unrelated disorders. For trios 2 and 3, we used 332 other exomes sequenced at the Radboud University Nijmegen Medical Center to remove systematic artifacts and low-frequency variants.
We also analyzed the coding exons (+/− 2 bp) of ACTB and ACTG1 in exome sequence data from 244 unrelated healthy individuals of European descent, which consist of parents of children with sporadic autism. These samples were sequenced at the University of Washington Genome Sciences using SeqCap EZ Exome Library v2.0 (Roche) and the Illumina sequencing platform. For all control samples combined, 89% of the bases of ACTB and ACTG1 were covered ≥ 8x. The average per-base coverage was 23x and 31x for ACTB and ACTG1, respectively. The first coding exon of ACTB had a low read depth, with an average coverage of 4x; all other exons had sufficient coverage to call variants in the majority of samples. Variants were filtered to 8x coverage and consensus or variant quality of 30 using SAMtools49, as previously described13.
Genomic DNA samples from 2,349 control individuals (including 1,953 of European ancestry) were genotyped using the InfiniumII HumanHap610 Quad BeadChip array (Illumina) at the Center for Applied Genomics at Children’s Hospital of Philadelphia. Control subjects consisted of 0–18 years of age healthy children that were primarily recruited from the Philadelphia region through the Hospital’s Health Care Network. Copy number variations were identified using the PennCNV algorithm50 and further filtered using a threshold of 10 consecutive single nucleotide variants, a length of 30 kilobase pairs, and a PennCNV confidence score of 10.
We amplified the coding and flanking intronic regions of the ACTG1 (NM_001614.2) and ACTB (NM_001101.2) genes in genomic DNA from 18 probands. PCR products were sequenced at the University of Washington High-Throughput Genomics Unit or the Radboud University Nijmegen Medical Centre. All mutations were confirmed by reamplification of the fragment and resequencing of the proband and his or her available relatives. Primer sequences and PCR conditions are available on request.
Epstein-Barr virus immortalized lymphoblastoid cell lines were established from peripheral blood samples of six affected individuals (LR04-173, LP98-096, LR09-079, LP98-085, LR04-298 and LR06-241) and two control individuals (09.1359 and 07.0841) using standard procedures. Cells were collected from suspension culture by centrifugation at 1,000 g for 10 min. Cell pellets were rinsed once with PBS and lysed in M-PER lysis buffer (Pierce) supplemented with protease and phosphatase inhibitor cocktail (Sigma). Protein concentration of the lysates cleared of insoluble cell debris were determined using 660 nm Protein Assay reagent (Pierce) and more concentrated samples were diluted to the concentration of the least concentrated sample with the lysis buffer. For SDS-PAGE equal volumes of concentration-adjusted lysates were mixed with LDS electrophoresis loading buffer (Life Technologies), then reduced and denatured for 10 min at +70°C. For 1D electrophoresis equal volumes of reduced denatured samples were separated on two 4–12% polyacrylamide gels (Life Technologies). One of the gels was processed for Coomassie staining (Bio-Rad), as per manufacturer’s recommendations. The other gel was transferred onto 0.2 μm nitrocellulose membrane (Pall) and processed for Western Blotting. Primary antibodies for Western Blotting were used at the following dilutions: mouse anti-ACTB (clone AC-15, Sigma, 1:32,000), mouse anti-ACTG1 (clone 2–220.127.116.11, Sigma, 1:16,000), and goat anti-GAPDH (Santa-Cruz biotechnology, 1:1,000). Appropriate secondary HRP-conjugated antibodies (Bio-Rad) were used at 1:40,000. Proteins were detected by incubation with chemiluminescent substrate SuperSignal Pico (Pierce), as per manufacturer’s protocol. Appropriate secondary HRP-conjugated antibodies (Bio-Rad) were used at 1:40,000. Proteins were detected by incubation with chemiluminescent substrate SuperSignal Pico (Pierce), as per manufacturer’s protocol.
Lymphoblastoid cells grown in suspension were placed on fibronectin (Millipore, 100 μg/ml) and laminin (Invitrogen, 25 μg/ml) coated glass cover slips for 2 hr. After cells attached, experimental cultures were treated with Latrunculin A (Cayman Chemical Company, 0.6 μM) for 2 hr. Immediately after treatment, cells were fixed with 0.25% glutaraldehyde and permeabilized with 0.1% triton x100. Mouse anti-α-tubulin primary antibody (Sigma, 1:5,000) was applied for overnight at +4°C. Secondary AlexaFluor-conjugated antibody (Life Technologies, 1:1,000) along with AlexaFluor-conjugated phalloidin to visualize F-actin (Life Technologies, 1:100) were applied for 1 hr at room temperature. Cover glasses were mounted in ProLong anti-fade media (Life Technologies) and visualized with 100x oil objective on inverted microscope (Zeiss) fitted with spinning disc confocal scanner (Perkin-Elmer). All imaging analysis was performed using ImageJ software as follows. Confocal stacks were projected into a single plane (Z-project, Maximal Intensity). Images were thresholded, perimeter was measured as object perimeter and fluorescence intensity measured as a mean gray value.
We would like to thank the families for their contribution to this study. We thank all members of the Northwest Genomics Center and Genomic Disorders Group Nijmegen, as well as personnel from the Microarray Facility and Sequencing Facility Nijmegen for excellent technical assistance. This work was supported by the US National Institutes of Health (grants 1R01-NS058721 to W.B.D. and 2PO1-NS048120 to M.E.R.), the Netherlands Organization for Health Research and Development (ZonMW grants 917-66-363 and 911-08-025 to J.A.V.), the EU-funded TECHGENE project (Health-F5-2009-223143 to J.A.V.) and the AnEUploidy project (LSHG-CT-2006-37627 to A.H., B.W.M.v.B., H.G.B. and J.A.V.). J.B.R. is supported by a Banting Postdoctoral Fellowship from the Canadian Institutes of Health Research. T.R. is supported by an Australian NHMRC post-doctoral fellowship. We would like to thank the Simons Foundation Autism Research Initiative (SFARI) that provided control exome data (grant 191889), the NIEHS Environmental Genome Project for providing support for this project (contract No.HHSN273200800010C), and the NHLBI GO Exome Sequencing Project and its ongoing studies which produced and provided exome variant calls for comparison: the Lung GO Sequencing Project (HL-102923), the WHI Sequencing Project (HL-102924), the Broad GO Sequencing Project (HL-102925), the Seattle GO Sequencing Project (HL-102926) and the Heart GO Sequencing Project (HL-103010).
The whole-exome sequencing data from individuals of trio 1 have been deposited in dbGaP under accession number (pending).
Author ContributionsD.T.P., W.B.D., A.V. and H.G.B. designed the study. J.B.R., B.W.M.v.B. and A.H. designed and performed the genetics experiments. S.S.K. performed the experiments in lymphoblastoid cell lines. J.B.R., A.H., B.J.O. and C.G. performed the bioinformatics experiments. B.J.O. analyzed the control exome datasets. S.G. contributed to the genetics experiments. C.T.S. and S.L.C. prepared DNA samples and lymphoblastoid cell lines. O.A.A.R., J.F.A., N.C., V.D.G., A.E.F., J.P.F., K.W.G., M.K., T.K., G.M.S.M., M.J.M.N., C.M.A.v.R.A., T.R., B.B.A.d.V., M.M., V.M.S., A.V., H.G.B., D.T.P. and W.B.D. recruited and evaluated the study subjects. J.A.R. analyzed the Signature Genomic Laboratories dataset. J.S. supervised B.J.O., J.A.V. supervised A.H., C.G. and S.G., H.G.B. supervised B.W.M.v.B., M.E.R. supervised and evaluated data with S.S.K. and wrote sections dealing with lymphoblastoid cell lines, and W.B.D. supervised J.B.R.. J.B.R. and W.B.D. wrote the paper.
Competing financial interests
The authors report no competing financial interests.