Mutations in NSUN2 cause autosomal-recessive intellectual disability

Abstract

With a prevalence between 1 and 3%, hereditary forms of intellectual disability (ID) are among the most important problems in health care. Particularly, autosomal-recessive forms of the disorder have a very heterogeneous molecular basis, and genes with an increased number of disease-causing mutations are not common. Here, we report on three different mutations (two nonsense mutations, c.679C>T [p.Gln227(∗)] and c.1114C>T [p.Gln372(∗)], as well as one splicing mutation, g.6622224A>C [p.Ile179Argfs(∗)192]) that cause a loss of the tRNA-methyltransferase-encoding NSUN2 main transcript in homozygotes. We identified the mutations by sequencing exons and exon-intron boundaries within the genomic region where the linkage intervals of three independent consanguineous families of Iranian and Kurdish origin overlapped with the previously described MRT5 locus. In order to gain further evidence concerning the effect of a loss of NSUN2 on memory and learning, we constructed a Drosophila model by deleting the NSUN2 ortholog, CG6133, and investigated the mutants by using molecular and behavioral approaches. When the Drosophila melanogaster NSUN2 ortholog was deleted, severe short-term-memory (STM) deficits were observed; STM could be rescued by re-expression of the wild-type protein in the nervous system. The humans homozygous for NSUN2 mutations showed an overlapping phenotype consisting of moderate to severe ID and facial dysmorphism (which includes a long face, characteristic eyebrows, a long nose, and a small chin), suggesting that mutations in this gene might even induce a syndromic form of ID. Moreover, our observations from the Drosophila model point toward an evolutionarily conserved role of RNA methylation in normal cognitive development.

Figure 1

Family Pedigrees and Linkage Intervals Filled symbols indicate affected individuals (A–C). Bars alongside the idiogram of chromosome 5 shown in (D) represent the overlapping linkage intervals. The enlarged picture detail shows the flanking markers of the linkage intervals from the different families, distinguished by shades of gray (dark, G-013; medium, M-192; and light, M-314).

Figure 2

Mutations (A) Sequence chromatograms of heterozygous carriers (“HET”) and individuals homozygous (“MUT”) for c.679C>T (Gln227^∗) (NM_017755.5) are compared to the normal sequence (“WT”). (B) RT-PCR results and (C) qPCR results with primers specific to exons 3 and 6 (indicated by arrowheads in the schematic on top) show loss of NSUN2 transcription in an individual homozygous for Gln227^∗ (“P”) and in controls (“C1”–“C4”). Relative expression was determined by comparison to GAPDH expression levels. The presence of RNA was controlled for with the use of primers specific to the X-chromosomal HUWE1 transcript. Size marker (“M”): HyperLadder IV (Bioline.) Error bars represent the standard deviation (n = 3). (D) Sequence chromatograms of individuals heterozygous (“HET”) and homozygous (“MUT”) for c.1114C>T (Gln372^∗) (NM_017755.5) are compared to the normal sequence (“WT”). (E) RT-PCR results and (F) qPCR results with primers specific to exons 7–9 (indicated by arrows in the schematic on top) show loss of NSUN2 transcription in an individual homozygous for Gln372^∗ (“P”) and in controls (“C1” – “C3”). Relative expression was determined by comparison to GAPDH expression levels. The presence of RNA was controlled for with the use of primers specific to the X-chromosomal HUWE1 transcript. Size marker (“M”): HyperLadder IV (Bioline). Error bars represent the standard deviation (n = 3). (G) Sequence chromatograms of individuals heterozygous (“HET”) and homozygous (“MUT”) for g.6622224A>C (NC_000005.9) (hg19) are compared to the normal sequence (“WT”). (H) Sequence chromatograms of the normal junction between exons 6 and 7 (“WT”) and the junction between exons 5 and 7 (“MUT”) caused by mutation g.6622224A>C (NC_000005.9) (hg19). The resulting protein sequences are indicated; the frameshift in the mutant leads to a downstream stop codon starting at nucleotide 71 of exon 7 (p.Ile179Argfs^∗192). (I) RT-PCR results with primers specific to the junctions of exons 2 and 3 and exons 8 and 9 (positions are indicated in the schematic representations of NSUN2 transcripts 1 and 2 presented on top) and commercially available RNA from fetal and adult brains. RNA prepared from lymphoblasts (“LCL”) of affected individuals (“P”) and controls (“C1” –“C5”) show loss of the main transcript (Trscpt 1) in lymphoblasts (“LCL”) of both affected individuals. The presence of RNA was controlled for with the use of primers specific to the X-chromosomal HUWE1 transcript (the right panel shows the results from one affected individual and controls). Size marker (“M”): HyperLadder IV (Bioline).

Figure 3

Mutants of the Drosophila Ortholog of NSUN2 Show an STM-Specific Defect (A) A phylogenetic reconstruction by the neighbor-joining method is based on the amino acid sequences of Drosophila NSUN2 and mouse and human methyltransferases of the NOL1/NOP2/Sun domain family. The scale bar indicates phylogenetic distance, and the tree has been rooted with the midpoint method. Note that the tree shows a distinct clade that clusters the Drosophila, human, and mouse NSUN2 (red box). (B) Human NSUN2 and the Drosophila ortholog (dNsun2) show a high degree of similarity. A region of about 300 aa in the center of the protein shares 74% sequence similarity and 59% identical amino acid (for a complete alignment, see Figure S3). The locations of the two nonsense mutations are indicated. Amino acid positions bordering the region of highest similarity are given, and blue boxes indicate SUN domains. (C) Genomic location of dNsun2 on the X chromosome at 4A5–4A6. dNsun2-deficient animals were constructed with Drosophila lines carrying transposon-mediated flippase recognition target sites that neighbored dNsun2 (black boxes indicate dNsun2 exons, and gray boxes indicate neighboring loci). The obtained deficiency (black line with arrowheads) was confirmed by PCR. In addition to the entire dNsun2, the 5′ dgt4 was also excised. (D) Behavioral tests for STM performance demonstrate a requirement for dNsun2. In contrast to a performance index (PI) of 65.30 ± 3.58 (n = 11) for wild-type flies (w¹¹¹⁸ males), the mutant flies only show a PI of 20.71 ± 3.15 (n = 11). Expressing the dNsun2⁺ cDNA or the driver line alone into the mutant background does not rescue the phenotype (26.51 ± 5.93, n = 7 and 24.5 ± 2.97, n = 6, respectively). The impaired STM performance in dNsun2^ex1 flies is only rescued by pan-neuronal (elav-GAL4) re-expression of the dNsun2⁺ cDNA (72.64 ± 5.55, n = 16). All data represent means ± SEM. Asterisks indicate p < 0.005.