FRA2A is a CGG repeat expansion associated with silencing of AFF3

Abstract

Folate-sensitive fragile sites (FSFS) are a rare cytogenetically visible subset of dynamic mutations. Of the eight molecularly characterized FSFS, four are associated with intellectual disability (ID). Cytogenetic expression results from CGG tri-nucleotide-repeat expansion mutation associated with local CpG hypermethylation and transcriptional silencing. The best studied is the FRAXA site in the FMR1 gene, where large expansions cause fragile X syndrome, the most common inherited ID syndrome. Here we studied three families with FRA2A expression at 2q11 associated with a wide spectrum of neurodevelopmental phenotypes. We identified a polymorphic CGG repeat in a conserved, brain-active alternative promoter of the AFF3 gene, an autosomal homolog of the X-linked AFF2/FMR2 gene: Expansion of the AFF2 CGG repeat causes FRAXE ID. We found that FRA2A-expressing individuals have mosaic expansions of the AFF3 CGG repeat in the range of several hundred repeat units. Moreover, bisulfite sequencing and pyrosequencing both suggest AFF3 promoter hypermethylation. cSNP-analysis demonstrates monoallelic expression of the AFF3 gene in FRA2A carriers thus predicting that FRA2A expression results in functional haploinsufficiency for AFF3 at least in a subset of tissues. By whole-mount in situ hybridization the mouse AFF3 ortholog shows strong regional expression in the developing brain, somites and limb buds in 9.5-12.5dpc mouse embryos. Our data suggest that there may be an association between FRA2A and a delay in the acquisition of motor and language skills in the families studied here. However, additional cases are required to firmly establish a causal relationship.

Figure 1. Description of FRA2A Families A–C.

Females are represented by circles, males by squares. The percentage of cells showing FRA2A expression are indicated on the bottom right-hand side of the symbol. As discussed in the text the 4% FRA2A expression seen in individual AI.1 represents a false positive. The case number used to indicate the FRA2A carriers in the text is given on the top left-hand side of the symbol. N within a symbol indicated individuals expression of the fragile site was not examined or observed. The associated phenotypes in individuals AII.3, AII.4, BII.1 and CII.1 are detailed in the boxed symbol key. The proband in each family is arrowed.

Figure 2. FISH analysis of the FRA2A fragile site.

A. The BAC-clone 549H5 (labelled green) spans the fragile site FRA2A. B. FISH analysis using the 10 kb L10K (labelled green; chr2: 100721983–100733233; hg19) and 18 kb L18K (labelled red; chr2: 100700447–100718834; hg19) PCR-generated probes, targeted to map either side of FRA2A. The additional telomeric FISH signal on the red channel (red arrow chr2: 110520380–110538822 and chr2:111347822–111366260; hg19) is the result of a 24 kb low copy repeat (24 kb LCR pink text) encompassing L18K. C. A schematic representation of the position of the LCR.

Figure 3. Allele frequencies of the FRA2A associated CGG repeat in a population of 100 control individuals.

PCR amplification of the repeat and subsequent sequencing in 200 control chromosomes revealed that it is highly polymorphic with a length ranging from 3 to 20 copies. The most commonly found allele contains eight repeated units.

Figure 4. Southern blot analysis of the AFF3 CGG repeat in all available members of the three families and two unrelated control individuals (C).

DNA restriction fragments obtained from blood samples of all available members of family A (lanes 1–4), B (lanes 5–6) and C (lanes 7–9) where blotted with a specific 32P-labeled probe (chr2:100088460–100089451; hg19) after digestion with HindIII. Lanes 10 and 11 contain DNA restriction fragments from two unrelated control samples (represented as diamonds) after digestion with HindIII. In addition to a normal size allele fragment of 4.4 kb, individuals CII.1, CI.2, BII.1, BI.2, AII.4, AII.3 and AIII.1 show additional fragments of a larger size, indicating repeat expansion. These expanded fragments were not present in the controls and individuals CI.1 and AI.1. A 1 kb length marker is presented at the left of the figure.

Figure 5. AFF3 protein-domain and gene structure.

A. diagrammatic representation of the AFF3 protein, showing the N-terminal homology domain (NHD), C-terminal homology domain (CTHD) and two predicted nuclear localisation signals. The 18 amino acids encoded by the alternatively spliced exon 4 are represented as an “insertion” at position 18 of isoform 1. The 6-bp in-frame deletion we identified in exon 14 of subjects AI.1 and AII.4, removing two amino acids (position 619 and 621 respectively) is indicated in red. This deletion was predicted to be benign. B. Genomic structure of AFF3 with the alternately used spliced exon 4 shown in red with asterisk. AFF3 exons 2, 3 and the alternate first exon are not separately resolvable at this resolution but shown in detail in panel C. Transcription left to right is shown in blue, right to left in brown. CAGE tag defined transcription start sites are shown aligned with annotated gene structure (blue forward strand transcription, brown reverse strand transcription shown with negative counts). Y-axis values show average tags per million (TPM) from CAGE libraries from the indicated tissue groups. C. Finer details of the AFF3 TSS regions are shown along with histograms of human fetal brain RNA-seq read coverage; three replicates are coloured separately. Splicing of the intron between the alternate first exon and exon 3 (blue asterisk) was supported by 9 independent RNA-seq reads and found in all three replicates. The CGG repeat and abnormally methylated region (AMR) are shown in red and green respectively, The major transcription start sites are shown as black arrowed lines. There is no supporting evidence for the RefSeq TSS represented by the pink arrowed line in our data. D. Alignment of human CGG repeat region and associated TSS with the orthologous mouse region. Nucleotides are color coded (A = green, G = yellow, C = blue, T = red, alignment gaps are grey). Orange histograms show the predicted G-quadruplex forming potential of the human and mouse sequences. Outer histograms show CAGE tag 5′ ends at single nucleotide resolution in both human (top) and mouse (bottom). TPM counts shown are the average from brain derived CAGE libraries in each species and represent the precise location of transcription initiations.

Figure 6. Site and stage specific expression of Aff3 in mouse embryos.

A. photomicrograph of a lateral view of a 12.5Aff3. The only area of strong expression at this stage is in the hand plate. B. show a false coloured but unthresholded image of the brightfield optical projection tomography (OPT) scan of a 11.5 dpc embryo with digital sagittal and coronal sections from the same embryo shown in C and D. Strong expression is seen in the somites, upper limb bud, the fusing primary palate and the diencephalon and prosomere1 regions of the developing brain.

Figure 7. Analysis of rs4851214 which maps to coding exon 14 of AFF3 using paired genomic DNA and cDNA templates from the unaffected carrier mother BI.2 and the affected carrier son BII.1 from family B.

A: Both BI.2 and BII.1 are heterozygous for SNP rs4851214 in genomic DNA as T and C peaks are visible at this site. B: Analysis of cDNA from subject BII.1, shows that only the C allele could be detected in this patient, indicating monoallelic expression, while in cDNA of subject BI.2 the heterozygous T/C signal was observed, indicating transcription of both alleles.