Homozygous Nonsense Mutations in KIAA1279 Are Associated with Malformations of the Central and Enteric Nervous Systems

Abstract

We identified, by homozygosity mapping, a novel locus on 10q21.3-q22.1 for Goldberg-Shprintzen syndrome (GOSHS) in a consanguineous Moroccan family. Phenotypic features of GOSHS in this inbred family included microcephaly and mental retardation, which are both central nervous system defects, as well as Hirschsprung disease, an enteric nervous system defect. Furthermore, since bilateral generalized polymicogyria was diagnosed in all patients in this family, this feature might also be considered a key feature of the syndrome. We demonstrate that homozygous nonsense mutations in KIAA1279 at 10q22.1, encoding a protein with two tetratrico peptide repeats, underlie this syndromic form of Hirschsprung disease and generalized polymicrogyria, establishing the importance of KIAA1279 in both enteric and central nervous system development.

Goldberg-Shprintzen syndrome (GOSHS) (Goldberg and Shprintzen 1981) is a disorder characterized by microcephaly, mental retardation, facial dysmorphism, and Hirschsprung disease (HSCR). GOSHS shares the same accession number (MIM 235730) in the Online Mendelian Inheritance in Man (OMIM) database as the comparable disorder Mowat-Wilson syndrome (MWS) (Mowat et al. 1998). Genetically, however, they are different. MWS occurs as a de novo dominant syndrome, whereas GOSHS is most likely inherited as an autosomal recessive trait, a hypothesis based on the occurrence of the disease in consanguineous pedigrees and in affected siblings with unaffected parents (Hurst et al. 1988; Brooks et al. 1999). MWS is associated with de novo mutations in ZFHX1B, located at 2q22 (Amiel et al. 2001; Wakamatsu et al. 2001; Zweier et al. 2002), whereas, for GOSHS, no causative gene has yet been identified. Also, some clinical differences exist. In patients with MWS, neurological abnormalities such as epilepsy (in a high percentage of patients), agenesis of the corpus callosum (in 35% of patients), and cortical malformations (in a minority of patients) have been reported (Amiel et al. 2001; Zweier et al. 2002; Silengo et al. 2004). In the few patients with GOSHS discussed in published reports, neurological symptoms have not been studied in detail, and the cause of the microcephaly and mental retardation remains obscure (Goldberg and Shprintzen 1981; Hurst et al. 1988).

Here, we describe a family with HSCR (an anomaly of the enteric nervous system that is of neural crest origin and is histologically characterized by the absence of ganglion cells in the myenteric and submucosal plexuses) (Okamoto and Ueda 1967) as a variable feature and bilateral generalized polymicrogyria (PMG) (a developmental malformation of the cerebral cortex, characterized by an enlarged number of smaller convolutions or gyri and disruption of the normal six-layered cerebral cortical structure) (Friede 1989) as a constant feature. The established diagnosis for this family was GOSHS, although bilateral generalized PMG (BGP) has not been reported as part of the syndrome before.

To unravel the genetic basis of GOSHS, we performed a complete genome scan and homozygosity mapping in a large consanguineous Moroccan family reported elsewhere (Brooks et al. 1999). The segregation of the disease in the pedigree is consistent with an autosomal recessive mode of inheritance, with several consanguineous loops (fig. 1), and the family therefore was considered suitable for homozygosity mapping (Lander and Botstein 1987). Informed consent was obtained from the parents of all subjects. In this family, five patients had the cardinal signs characterizing GOSHS (one boy lacked HSCR). One child with long-segment HSCR died in the neonatal period; his DNA was unavailable for our study. When the youngest patient was diagnosed, brain magnetic resonance imaging (MRI) revealed BGP. MRI scans of all living patients also revealed BGP (fig. 2). Gait disturbance, speech defect, excessive drooling, and pseudobulbar signs indicate diffuse cortical dysfunction including the perisylvian areas.

Figure 1.

Pedigree structure and haplotypes of the Moroccan family with GOSHS. Patients are represented as blackened symbols. The shared region of homozygosity (red) in all affected cases is delimited by markers CGR166 and CGR170. Six of the 11 markers typed for refining the candidate region between marker D10S1652 and D10S537 are shown.

Figure 2.

Representative MRIs of patient VI-1 (panels A and B) and patient VI-10 (panels C and D), showing BGP. A, Axial T2-weighted MRI of patient VI-1 at age 5 mo, showing frontoparietal and occipital bilateral PMG. B, Coronal FLAIR image of the same patient showing perisylvian PMG (arrow). C, T1-weighted axial image of patient VI-10 at age 14 years, showing BGP with distribution similar to that seen in patient VI-1. D, Parasagittal T1-weighted image of patient VI-10, showing PMG of the right peri-insular temporal lobe.

Genomic DNA was isolated from peripheral leukocytes by use of the method described by Miller et al. (1988). By haplotyping, FISH, and sequence analyses, we excluded the ZFHX1B gene as a cause of GOSHS in this family (data not shown). For the systematic genome scan, 381 markers (STRPs) from the ABI Prism Linkage Mapping Set MD-10 (version 2.5), with an average spacing of 10 cM, were tested. DNA amplification was performed using 25 ng genomic DNA in 7.5-μl PCRs containing 1× PCR Gold Buffer, 2.5 mM MgCl₂, 10 μM primer pair mix, and 0.4 U AmpliTaq Gold DNA polymerase (Applied Biosystems). Amplification conditions were 10 min at 95°C, followed by 35 cycles of 30 s at 95°C, 30 s at 55°C, and 1 min 30 s at 72°C; amplification was ended by a final extension for 5 min at 72°C. PCR products were pooled in panels and were loaded on an ABI 3100 automated sequencer. Data were analyzed using GeneMapper software (version 2.1) (Applied Biosystems).

After this initial screening, a region on 10q21.3-q22.1 was further investigated, since one of the four affected individuals was homozygous by descent for two consecutive markers, D10S1652 and D10S537, whereas the other three patients were homozygous for one of these two markers. No other regions of homozygosity for consecutive markers were identified. We saturated the region between markers D10S1652 and D10S537 (12.3 cM) with 11 additional markers; they either were obtained from the Marshfield genetic map (Center for Medical Genetics Web site) or were newly developed. Marker order and genetic distances were obtained from the Marshfield genetic map (Center for Medical Genetics Web site) and the Celera physical map. All patients were homozygous for 7 of these 11 markers. A recombination event between loci CGR166 and CGR167 in patient VI-7 defined the centromeric boundary of the genetic interval. The telomeric border was defined by recombination events between markers D10S1665 and CGR170 in patients VI-1 and VI-3 (fig. 1). These recombination events reduced the candidate region to a minimum of 2.8 Mb and a maximum of 3.8 Mb.

Parametric two-point and multipoint linkage analyses were performed using the MLINK program from the LINKAGE (version 5.1) software package (Lathrop and Lalouel 1984) and the SimWalk2 program (version 2.9) (Sobel et al. 2002). LOD scores were calculated, under the assumption that the disease phenotype is a rare autosomal recessive disorder with 100% penetrance, with a gene frequency of 1:5,000. No phenocopies were allowed, and equal allele frequencies were used because of the limited number of available independent family members. A significant two-point LOD score was obtained (max LOD score 3.29 and θ=0 for marker CGR167) (table 1), which increased to a maximum multipoint LOD score of 5.9 between markers CGR167 and D10S1647 (fig. 3).

Table 1.

Results of Two-Point LOD Score Analysis, Performed for Markers Used in Fine Mapping

	LOD at θ =
Markera	0	.01	.05	.1	.2	.3	.4
D10S196	−13.81	−3.96	−1.52	−.65	−.08	.02	0
D10S1652	−6.96	−2.39	−1.85	−1.36	−.68	−.35	−.15
D10S1743	−4.97	1	1.39	1.31	.9	.46	.14
CGR166	−3.7	.58	1.03	1.02	.73	.4	.13
CGR167	3.29b	3.21	2.9	2.51	1.73	1	.4
D10S210	2.59	2.53	2.26	1.92	1.26	.67	.24
D10S1678	1.49	1.45	1.29	1.09	.72	.4	.16
D10S1647	1.84	1.79	1.58	1.33	.86	.47	.19
D10S1672	2.88	2.81	2.52	2.15	1.42	.76	.25
D10S1665	2.77	2.7	2.43	2.08	1.39	.75	.25
CGR172	1.36	1.32	1.17	.97	.62	.32	.11
CGR170	−.28	1.59	1.97	1.88	1.4	.83	.33
D10S676	−.28	1.25	1.64	1.56	1.1	.59	.18
D10S537	−.11	1.45	1.82	1.71	1.2	.65	.21

^aMarker order follows that of the Celera physical map.

^bMax LOD score.

Figure 3.

Multipoint LOD score analysis across and beyond the region of homozygosity, flanked by markers D10S1652 and CGR170, with a peak LOD score of 5.9.

The maximum 3.8-Mb linked region of shared homozygosity (defined by CGR166 and CGR170) contained a total of 25 known genes and 5 genes encoding hypothetical proteins, in accordance with the National Center for Biotechnology Information (NCBI) build 34 of the human genome and the Ensembl Genome Browser (table 2). We first sequenced seven functional candidate genes on the basis of their putative involvement in neurogenesis (CTNNA3, LRRTM3, ATOH7, DDX50, DDX21, NET-7, and NEUROG3), either using cDNA obtained by RT-PCR from a lymphoblastoid cell line of patient VI-10 and from a control or using genomic DNA from both. Mutations were not identified. We then performed a systematic mutational screening of all transcripts from the candidate region. PCR products from either genomic DNA (exons including flanking intronic sequences) or reverse-transcribed products (ORFs) were amplified and purified (ExoSapIT [USB]). PCR primers, which were designed using standard software (Primer3), and conditions are available on request. Bidirectional sequencing was performed using BigDye Terminator chemistry on an ABI 3100 sequencer (Applied Biosystems). With the use of the software package SeqScape (version 2.1) (Applied Biosystems), sequences were aligned and compared with consensus sequences.

Table 2.

Gene and Marker Positions and Details

Gene or Marker	Celera Position(Mb)	Additional Details
D10S196	…	GenBank accession number Z16598
D10S1652	…	GenBank accession number Z52339
D10S1743	60.5	GenBank accession number Z53951 (located at BAC AC020641)
CTNNA3	60.95–62.73	5′→3′: catenin (cadherin-associated protein), alpha 3
CGR166	61.2	GenBank accession number BV212297 (located at BAC AC27675)
LRRTM3	61.9–62.1	3′→5′: leucine-rich repeat transmembrane neuronal 3
CGR167	61.63–62.1	GenBank accession number BV212296 (located at BAC AC084303)
LOC340888	…	3′→5′: similar to aldo-keto reductase family 1, member B10; aldose reductase-like 1 pseuodogene
JDP1	62.83–62.87	5′→3′: J domain–containing protein 1
LOC389974	…	3′→5′: similar to 60S ribosomal protein L12
SIRT1	62.91–62.95	3′→5′: sirtuin (silent mating type information regulation 2 homolog) 1 (S. cerevisiae)
DKFZP564G092	…	5′→3′: DKFZP564G092 protein, HERC4
MYPN	63.14–63.24	3′→5′: myopalladin
ATOH7	63.26–63.26	5′→3′: atonal homolog 7 (Drosophila)
LOC389975	…	3′→5′: similar to keratin 19; keratin, type I cytoskeletal 19; keratin, type I, 40-kd; pseudogene
MAWBP	63.31–63.36	5′→3′: MAWD-binding protein
D10S210	63.32	GenBank accession number Z16813 (located at BAC AC016395)
HNRPH3	…	3′→5′: heterogeneous nuclear ribonucleoprotein H3 (2H9)
RUFY2	63.37–63.44	5′→3′: RUN and FYVE domain–containing 2
DNA2L	63.44–63.50	5′→3′: DNA2 DNA replication helicase 2-like (S. cerevisiae)
SLC25A16	63.51–63.56	5′→3′: solute carrier family 25
CXXC6	63.68–63.72	3′→5′: CXXC finger 6
CCAR1	63.75–63.8	3′→5′: cell division cycle and apoptosis regulator 1
C10orf24	63.92–63.93	3′→5′: chromosome 10 ORF 24: ORF with an as-yet-unknown function
D10S1678	63.8–64	GenBank accession number Z52660 (located at BAC AL359844)
DDX50	63.93–63.98	3′→5′: DEAD (Asp-Glu-Ala-Asp) box polypeptide 50
DDX21	63.99–63.02	3′→5′: DEAD (Asp-Glu-Ala-Asp) box polypeptide 21
KIAA1279	64.02–64.05	3′→5′: KIAA1279
LOC389976	…	3′→5′: similar to actin 3 (Drosophila) (fragments) pseudogene
PRG1	64.12–64.14	3′→5′: proteoglycan 1, secretory granule
VPS26	64.1–64.2	3′→5′: vacuolar protein–sorting 26 (S. cerevisiae)
SUPV3L1	64.22–64.24	3′→5′: suppressor of var1, 3-like 1 (S. cerevisiae)
FLJ22761	64.2–64.3	3′→5′: hypothetical protein FLJ22761
D10S1647	64.20–64.38	GenBank accession number Z52188 (located at BAC AL596223)
HK1	64.3–64.43	3′→5′: hexokinase 1
TACR2	64.44–64.45	5′→3′: tachykinin receptor 2
NET-7	64.48–64.54	3′→5′: transmembrane 4 superfamily member tetraspan NET-7
NEUROG3	64.6–64.6	5′→3′: neurogenin 3
C10orf35	…	3′→5′: chromosome 10 ORF 35
LOC389977	…	3′→5′: similar to ribosomal protein L5; 60S ribosomal protein L5 pseudogene
COL13A1	64.8–64.9	3′→5′: collagen, type XIII, alpha 1
LOC387686	…	5′→3′: LOC399779
CGR170	65	GenBank accession number BV212295 (located at BAC AC024601)
H2AFY2	65.0–65.1	3′→5′: H2A histone family, member Y2
AMID	65.15–65.18	5′→3′: apoptosis-inducing factor–homologous mitochondrion-associated inducer of death
D10S537	…	5′→3′

Analysis of 98% of all coding sequences from the region revealed one likely disease-causing homozygous nucleotide substitution, 303C→T, in exon 1 of the KIAA1279 gene (fig. 4A). This transversion leads, at the amino acid level, to the replacement of an arginine with a stop codon (R90X), resulting in a shortened protein of 89 aa. The 303C→T mutation showed complete segregation with the disease in the family. One hundred ethnically matched control chromosomes (of North African origin) were screened for the R90X mutation, and none of the controls carried this mutation. Subsequently, one additional family of British Pakistani ancestry, with a phenotype similar to that found in the Moroccan family, was screened for mutations in this gene. This family had four affected individuals with multiple inbreeding loops. Clinical characteristics were consistent with GOSHS, and CT brain scans of this family were described in a previous study (Hurst et al. 1988) and were thought to represent an abnormality of neuronal migration. In the affected individuals, a homozygous G→T transversion was identified at nt 285, resulting, at the amino acid level, in the replacement of a glutamic acid with a stop codon (E84X) (fig. 4B). Two different homozygous nonsense mutations in KIAA1279 in two independent families with GOSHS strongly indicate that KIAA1279 is the gene responsible for GOSHS.

Figure 4.

Mutations of KIAA1279 associated with GOSHS. A, Electropherogram of the mutation identified in the Moroccan family. The 303C→T mutation causes the replacement of an arginine with a stop codon (R90X) in exon 1. B, Electropherogram of the mutation identified in the Pakistani family. A homozygous G→T transversion at nt 285 results in the replacement of a glutamic acid with a stop codon (E84X) in the patients.

KIAA1279 consists of seven exons spanning 28 kb of genomic DNA. The transcript encodes a protein of 621 aa. We identified orthologs of human KIAA1279 in fruit fly, frog, rat, mouse, bee, chicken, and Japanese puffer fish, using BLAST comparison (NCBI Web site). There is strong sequence conservation during evolution; the protein sequence of human KIAA1279 shares 89% amino acid identity with mouse product, 88% with rat product, and 29% with fruit fly product. Multitissue northern blots (both adult and fetal) (fig. 5) showed ubiquitous expression of the 2.4-kb KIAA1279 mRNA without a specific pattern, with expression lacking only in blood leukocytes (fig. 5A). In addition, northern blot hybridization identified KIAA1279 mRNA molecules in different parts of the adult CNS—namely, cerebellum, cortex, medulla, spinal cord, occipital lobe, frontal lobe, temporal lobe, and putamen (fig. 5B). This widespread KIAA1279 expression is consistent with data available in gene expression databases such as the Gene Expression Omnibus, the Human Unidentified Gene-Encoded (HUGE) Protein Database, and the University of California–Santa Cruz (UCSC) Genome Browser. The function of KIAA1279 is unknown. Neither the gene nor its derived protein shows any significant sequence similarity to known human cDNA or protein sequences. Possible aspects of the molecular function of KIAA1279 may be gleaned from its primary sequence. We predicted protein motifs and domains using the program SMART (Simple Modular Architecture Research Tool). The predicted protein contains two tetratrico peptide repeats (TPRs), which are structural motifs consisting of 34 amino acid residues, in exon 1/2 and exon 4, respectively. Because of these repeats, KIAA1279 may be considered a member of the TPR protein family (D’Andrea and Regan 2003). The most basic function of TPR motifs is to mediate protein-protein interactions. Proteins with TPR motifs are involved in a variety of biological processes, such as cell-cycle regulation, transcriptional control, mitochondrial and peroxisomal protein transport, neurogenesis, and protein folding. Several diseases have been described as being caused by mutations in genes encoding proteins containing TPR domains—for example, Leber congenital amaurosis (AIPL1) (Sohocki et al. 2000) and Charcot-Marie-Tooth type 4C neuropathy (KIAA1985) (Senderek et al. 2003).

Figure 5.

Expression of KIAA1279 in human adult and fetal tissues. A, Multitissue northern blots, incubated with a KIAA1279 probe, showing ubiquitous expression without a specific pattern (note the lack of expression in blood leukocytes). B, Northern blots from adult brain tissues, incubated with a KIAA1279 probe. C, Human fetal multitissue northern blot, showing widespread expression.

Since mutations in KIAA1279 are associated with HSCR and PMG, its protein product may play a pivotal role in both peripheral and central nervous system development. PMG is traditionally classified as a result of a disruption such as fetal hypoxic ischemic damage, infection, or exposure to toxic drugs in the second trimester (13–21 wk) (Barth and van der Harten 1985; Barth 2003). However, the identification of KIAA1279 as the GOSHS gene underlines the importance of genetic factors in the etiology of this neuronal migration disorder of the cortex, as was already proven (Mitchell et al. 2003; Piao et al. 2004) or suggested (Ciardo et al. 2001; Chang et al. 2004) for other forms of PMG. Under the assumption that the main function of KIAA1279 is to bind other proteins, these target peptides might be (known) HSCR- or PMG-associated susceptibility factors. More likely, however, these targets might play a more general role in neuronal development, since a malformation of cortical organization is a constant feature found in the Moroccan family, whereas HSCR is a variable feature. Elucidation of KIAA1279 function and, more important, studies of KIAA1279 expression and protein interactions may provide new insight into the molecular basis of PMG and HSCR.