Deletion of TOP3β, a component of FMRP-containing mRNPs, contributes to neurodevelopmental disorders

Abstract

Implicating particular genes in the generation of complex brain and behavior phenotypes requires multiple lines of evidence. The rarity of most high-impact genetic variants typically precludes the possibility of accruing statistical evidence that they are associated with a given trait. We found that the enrichment of a rare chromosome 22q11.22 deletion in a recently expanded Northern Finnish sub-isolate enabled the detection of association between TOP3B and both schizophrenia and cognitive impairment. Biochemical analysis of TOP3β revealed that this topoisomerase was a component of cytosolic messenger ribonucleoproteins (mRNPs) and was catalytically active on RNA. The recruitment of TOP3β to mRNPs was independent of RNA cis-elements and was coupled to the co-recruitment of FMRP, the disease gene product in fragile X mental retardation syndrome. Our results indicate a previously unknown role for TOP3β in mRNA metabolism and suggest that it is involved in neurodevelopmental disorders.

Figure 1. The frequency of neurodevelopmental disorders in Finland varies by region.

This variation corresponds to the migration history of the Finnish population, as exemplified by schizophrenia prevalence^{, ,} (Fig. 1A) and by the percentage of the population with a disability pension due to intellectual disability (Official Statistics of Finland, the Social Insurance Institution of Finland, 2011) (Fig. 1B). Black arrows depict the radiation of migration, beginning in the 16^th century, from the early settlement region to the late settlement region (the border of the two regions is outlined in light grey). Red arrows show the location of sub-isolate municipalities within the Northeastern late settlement region. Gradients in prevalence of neurodevelopmental disorders, across different municipalities, are shown by the intensity of the blue shading.

Figure 2. The 22q11.22 deletion covers about 240 kb, is present in homozygous form in individuals diagnosed with schizophrenia and/or cognitive deficits, and results in dose-dependent reduction in TOP3β.

(a). A schematic representation of 22q11.2 (17 −22.0 Mb) depicts the distinct regions containing the deletions implicated in VCFS and the distal 22q11.22 deletion syndrome; the latter region includes the 240 kb deletion described in this paper. Orange bars depict known disease-related genes (based on OMIM) as well as genes relevant for the current study. Grey bars represent low copy repeats associated with the deletions in this region, while blue bars show the most common syndrome-related deletionsand the 240 kb 22q11.22 deletion reported here. The SNPs defining the breakpoint are marked above the deletion. The breakpoint region between the SNPs is marked with dashed lines. The deletion is flanked by two complementary low- copy repeats (grey bars). The genes located on the deletion region are presented below the deletion, which incorporates the full extent of TOP3β and IGLV2-14. Genomic positions are according to hg build 36. (b). Four individuals with homozygous 22q11.22 deletions, all of whom display schizophrenia and/or intellectual impairment, are members of three pedigrees from Northern Finland. The pedigrees, which include two documented consanguineous matings, are presumed to descend from a common ancestor (dashed line) based upon the common 22q11.22 haplotype observed among all of the homozygous deletion carriers. (c) TOP3β mRNA levels differ among non-carriers, heterozygous carriers and homozygous carriers of the 22q11.22 deletion. The expression level of TOP3β in non-carriers (269.99, 95%-CI = 249.18 - 290.79) was twice that of heterozygotes (127.61, 95%-CI = 103.92- 151.30), while homozygous deletion carriers had no detectable transcript.

Figure 3. TOP3β is part of a cytosolic, mRNP-associated protein complex that contains TDRD3 and FMRP.

(a) Immunoprecipitation of FLAG/HA-TOP3β from stably transfected HEK293 cells led to the co-purification of TDRD3 and FMRP in an RNAse-insensitive manner. In contrast, RNAse abolished the association of TOP3β with the mRNP proteins PABPC1 and MAGOH. (b) The TTF complex is present in neuronal cells. Immunoprecipitations from mouse brain lysates using antibodies directed against TDRD3 (lane 3) or TOP3β (lane 4). Precipitated TTF components were immunodetected as indicated. Specificity was controlled by an immunoprecipitation with pre-immune serum (lane 2). (c) Immunostaining of untreated (CTR; upper panel) and arsenite treated (+ARS; lower panel) HeLa cells with an anti-TOP3β antibody (green); nuclei were counterstained with DAPI (blue). TOP3β was localized predominantly in the cytosol where it showed a diffuse staining pattern. Arsenite treatment led to its accumulation in cytosolic foci that resemble stress granules (see also Fig. S3). Scale bars represent 50 μm. (d) TOP3β binds mature mRNAs. Cells expressing FLAG/HA-TOP3β were lysed in the presence of SDS (lane 1, input) and polyadenylated RNA was purified by affinity to oligo(dT) cellulose (lane2). This led to the co-purification of TOP3β and the known RNA binding proteins FMRP and PABPC1, while the control protein GAPDH was not found in the eluate. Treatment of the lysate with RNAse A prior to oligo(dT) affinity purification (lane 3) diminished binding of TOP3β, FMRP and PABPC1. (e) TOP3β is in direct contact with RNA. Control cells (lane 1) or FLAG/HA-TOP3β-expressing cells (lanes 2-6) were crosslinked using UV light (lanes 1-5) or left untreated (lane 6). FLAG/HA-TOP3β was immunoprurified under stringent conditions and crosslinked RNAs were treated with increasing amounts of RNAse T1 prior to 5′-labeling with [γ³²P]-ATP. Middle panel: Autoradiography of [32P]-labeled RNA crosslinked to FLAG/HA-TOP3β. Lower panel: Western blot control of immunoprecipitated FLAG/HA-TOP3β. All images are representative of at least 3 independent experiments; full-length blots and gels are presented in Figure S8.

Figure 4. TOP3β catalyzes RNA transesterification.

(a) Schematic of the reaction mechanism. TOP3β binds to a substrate oligonucleotide and a cleavage-religation equilibrium is established, leading to the transient formation of a 5′-fragment and a covalent complex with a tyrosyl-5′-phosphodiester bond between the 3′-fragment and the enzyme. Addition of SDS to the reaction intermediate leads to a release of the cleaved 5′-fragment and the covalent complex. (b) Coomassie stain of purified recombinant wildtype TOP3β and an active site mutant (Y336F). (c) TOP3β is active on DNA oligonucleotides and can be competed off by addition of RNA. A single stranded, 5′-[³²P] labeled DNA oligo was incubated with TOP3β, the reaction was stopped with SDS and cleavage products were analyzed by denaturing polyacrylamide gel electrophoresis and autoradiography. Cleavage was observed for wildtype TOP3β (wt; lanes 1-3) but not for an active site mutant (lane 4). Addition of cold competitor RNA resulted in a strong inhibition of the reaction (lanes 2 and 3). (d) TOP3β cleaves RNA. A 5′ [³²P] labeled RNA oligo was incubated with TOP3β either in Y336F mutant (lane 2) or wildtype (lane 6) form and cleavage products were analyzed as in (b). To control for impurities in the recombinant protein preparations that might lead to unspecific RNA fragmentation, wildtype and Y336F mutant protein were titrated against each other (lanes 3-5). (e) TOP3β forms a covalent tyrosyl-5′-phosphodiester bond with RNA. Cleavage reactions using the indicated amounts of TOP3β were performed like in (c) except that 3′-[³²P] labeled RNA was used. Formation of the TOP3β-RNA covalent complex was monitored by denaturing SDS-PAGE and autoradiography (upper panel). As a control, TOP3β protein was detected by Western blot (lower panel). (f) The TOP3β-RNA covalent complex is present in cellular extracts. Proteins bound to polyadenylated RNAs were affinity-purified by oligo(dT) cellulose as described in figure 4e, except that RNAse A treatment was performed not in the extract but after elution of from the column. This revealed high molecular weight species migrating above FLAG/HA-TOP3β, which were recognized by the anti-FLAG antibody. These were enriched by oligo(dT) purification and sensitive to RNAse treatment, indicating for covalent FLAG/HA-TOP3β-RNA complexes. SDS-PAGE of input (lane 1) and eluates either untreated (lane2) or treated with RNAse (lane 3) was analyzed by and Western blotting. Proteins of interest were immunodetected using the indicated antibodies. All images are representative of at least 3 independent experiments; full-length blots and gels are presented in Figure S8.

Figure 5. The TTF complex is present on early mRNPs that undergo the pioneer round of translation.

(a) Schematic of the ribosome salt wash (RSW) used to purify proteins associated with the translational machinery. Fractions used in (b) are underlined. (b) RSW of HeLa cells analyzed by SDS-PAGE with subsequent Coomassie staining (left panel) and Western blot against the indicated marker proteins (right panel). Immunodetection shows that endogenous TDRD3 and TOP3β co-purify with translation initiation factor eIF2γ in the high salt supernatant (S500), indicating that they are indirectly associated with ribosomes via translating mRNPs.(c) Polysome gradient analysis of extracts from a stable cell line expressing FLAG/HA-TDRD3. RNA profiling (upper panel) and Western blot analysis of gradient fractions (lower panel) revealed that the TTF-complex components co-sediment with polyribosomes in an RNAse sensitive manner (d) TTF-bound mRNPs show characteristics of early mRNPs that have not yet undergone steady state translation. Extracts of control cells (lane 1) or FLAG/HA-TDRD3 expressing cells (lane 2) were subjected to anti-FLAG IP to purify TTF-bound mRNPs either without (lanes 3 and 5) or with (lanes 4 and 6) RNAse pretreatment. The TTF complex was associated with ribosomal RNAs and the nuclear cap binding proteins CBP80 and CBP20 but not eIF4E in an RNAse sensitive manner. All images are representative of at least 3 independent experiments; full-length blots and gels are presented in Figures S8 and S9.

Figure 6. Formation of the TTF complex is essential for the co-recruitment of TOP3β and FMRP into mRNPs.

(a) Schematic of the TDRD3 mutants that were used to analyze the biogenesis of TTF-containing mRNPs. (b) Recognition of aDMA by the TDRD3 Tudor domain is required for efficient TTF complex formation in vivo. The indicated mutants were immunoprecipitated and co-precipitated proteins were analyzed by Western blotting. Note that the residual FMRP signal is RNAse sensitive for TDRD3ΔFIM but not for TDRD3_TDRmut, indicating that in the latter case the TTF complex was formed, albeit with an efficiency lower to that of wildtype TDRD3 (see Fig. S5c)..(c) Tudor-mutant TDRD3 can still bind mRNPs. HeLa cell extracts (input; lane 1) were incubated with purified recombinant GST-TDRD3-6xHis either in wildtype (lane 2), Tudor-mutant (lane 3) or EBM-mutant (lane 4) form. GST-6xHis was used as a control (lane 5). Upper panel: Coomassie-stained bait proteins. Lower panels: Western Blot detection of FMRP and MAGOH. (d) Recruitment of TOP3β to mRNPs requires binding of TDRD3 to the EJC. Analysis of extracts from a cell line carrying a deletion of TDRD3 (ΔTDRD3) by sucrose gradient centrifugation revealed that the association of TOP3β with polysomal mRNPs was disrupted. Polysomal migration of TOP3β was restored by transfection of wildtype, but not EBM-mutant TDRD3. Lower panel: Quantification of polysomal TOP3β from three independent experiments; error bars: SD. (e) TDRD3 is essential for the co-recruitment of FMRP to TOP3β-containing mRNPs. TDRD3-negative cells were co-transfected with GFP-TOP3β and either a control (FLAG/HA) or FLAG/HA-TDRD3 in ΔFIM-mutant or wildtype form. GFP-TOP3β-containing mRNPs were immunoprecipitated and analyzed for co-precipitated proteins by Western blotting with the indicated antibodies. While co-transfection of ΔFIM-mutant TDRD3 enabled the formation of mRNPs containing GFP-TOP3β (lane7), a co-recruitment of FMRP was observed only when wildtype TDRD3 was transfected (lane8). All images are representative of at least 3 independent experiments; full-length blots and gels are presented in Figure S9.