Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2017 Apr 19;94(2):322-336.e5.
doi: 10.1016/j.neuron.2017.03.026. Epub 2017 Apr 6.

Non-nuclear Pool of Splicing Factor SFPQ Regulates Axonal Transcripts Required for Normal Motor Development

Swapna Thomas-Jinu  1 , Patricia M Gordon  1 , Triona Fielding  1 , Richard Taylor  1 , Bradley N Smith  2 , Victoria Snowden  1 , Eric Blanc  1 , Caroline Vance  2 , Simon Topp  2 , Chun-Hao Wong  2 , Holger Bielen  1 , Kelly L Williams  3 , Emily P McCann  3 , Garth A Nicholson  4 , Alejandro Pan-Vazquez  1 , Archa H Fox  5 , Charles S Bond  6 , William S Talbot  7 , Ian P Blair  3 , Christopher E Shaw  2 , Corinne Houart  8
Affiliations

Non-nuclear Pool of Splicing Factor SFPQ Regulates Axonal Transcripts Required for Normal Motor Development

Swapna Thomas-Jinu et al. Neuron. .

Erratum in

Abstract

Recent progress revealed the complexity of RNA processing and its association to human disorders. Here, we unveil a new facet of this complexity. Complete loss of function of the ubiquitous splicing factor SFPQ affects zebrafish motoneuron differentiation cell autonomously. In addition to its nuclear localization, the protein unexpectedly localizes to motor axons. The cytosolic version of SFPQ abolishes motor axonal defects, rescuing key transcripts, and restores motility in the paralyzed sfpq null mutants, indicating a non-nuclear processing role in motor axons. Novel variants affecting the conserved coiled-coil domain, so far exclusively found in fALS exomes, specifically affect the ability of SFPQ to localize in axons. They broadly rescue morphology and motility in the zebrafish mutant, but alter motor axon morphology, demonstrating functional requirement for axonal SFPQ. Altogether, we uncover the axonal function of the splicing factor SFPQ in motor development and highlight the importance of the coiled-coil domain in this process. VIDEO ABSTRACT.

Keywords: PSF; RNA processing; RNA-binding protein; SFPQ; amyotrophic lateral sclerosis; axonogenesis; central nervous system; motor neurons; neurodegeneration; neurodevelopment.

PubMed Disclaimer

Figures

Figure 1
Figure 1
SFPQ Is Required for Brain Boundaries Dorsal (A–D and I–L) and lateral (E–H) views of 32 hpf zebrafish brain with anterior to the top (A and B) or left (C–L). (A and B) Immunostaining of sfpq; Tg(dlx4.6 GFP) embryos. Anti-acetylated tubulin staining (red) reveals asymmetrical folds in the midbrain (white arrowheads) and thickening of the isthmic organizer (IsO) in coma mutant (B; n = 8) compared to its wild-type sibling (A; n = 24). GFP staining (green) reveals disorganized neuronal distribution in the cerebellum. (C and D) Dorsal views, anterior to the top, of sibling (C) and mutant (D) sfpq;Tg(βactin: HRAS gfp) embryos showing failure of morphological thinning of the isthmus (white arrows in D, n = 9/9) in the homozygous mutant. (E and F) Expression of pax2.1 in the MHB, greatly reduced in all sfpq mutants (F; n = 10) compared to siblings (E; n = 30) at 36 hpf. (G and H) Expression of boundary marker, rfng, in siblings (G; n = 21) and mutant (H; n = 7, reduced or absent) at 24 ss. (I and J) Expression of zash1a in the hindbrain at 15 ss, in siblings (I; n = 19) and mutant (J; n = 6). (K and L) HuC staining at 32 hpf, in siblings (K; n = 17) and mutant (L; n = 5). Scale bar, 100 μm. (M) Schematic of the sfpq gene and the zebrafish and human mutations described in this report and, below, the sequence altered in the zebrafish coma mutant. PQ, proline (P) glutamine (Q) rich; RRM, RNA recognition motif; NOPS, NONA/paraspeckle domain; NLS, nuclear localization signals.
Figure 2
Figure 2
Axonogenesis Is Affected in sfpq−/− Embryos Lateral (A, B, and G–P) and dorsal (C–F) views, anterior to the left, of zebrafish brain at 32 hpf (A and B) and spinal cord at 24–56 hpf. (A and B) sfpq;Tg(dlx4.6:gfp) embryos. Lack of supra-optic commissure (white arrow) and posterior commissure (white arrowhead) in coma (n = 8/32) is revealed by acetylated tubulin staining. (C and D) Dorsal view, anterior to the left of acetylated tubulin staining showing hindbrain disorganized axonal tracks in sfpq mutant (D; green arrowhead, n = 8) compared to siblings (C; n = 26). (E and F) Dorsal view, anterior to the left of GFP+ motor neurons in sfpq; Tg(isl1:gfp) siblings (E) and mutant (F), showing cranial motor neuronal clusters lacking axonal projections in the mutant (red arrowhead, n = 7). (G–N) Lateral view, anterior to the left, of confocal live imaging, showing temporal defect in axonogenesis in the majority of spinal motor neurons in a mutant (H, J, L, and N; n = 14) compared to a sibling (G, I, K, and M; n = 42) in the Tg(mnx1: gfp) background. (O and P) Lateral view, anterior to the left, of SV2 antibody staining of siblings (O) and sfpq null mutant (P), showing pre-synaptic protein in the few axons formed in sfpq mutants (n = 9). Scale bar, 100 μm.
Figure 3
Figure 3
Sfpq Is Required Cell Autonomously for Motor Axon Development Homotopic transplantation at 70%–80% epiboly of sfpq−/−; Tg(mnx1:GFP) and sibling Tg(mnx1:GFP) ventral spinal cord progenitors into wild-type or sfpq−/− hosts. (A and B) Schematic of the transplants from donor (A) to host (B) embryo. (C, D, and G–L) Lateral view, anterior to the left (C, D, and G–J) or transverse (K and L) of 30 hpf zebrafish transplanted trunks, showing the transplanted mnx1+ neurons in green and all axons in red (acetylated tubulin staining). (E and F) Lateral view, anterior left of transplanted wild-type MNs (green) in early larvae (3 dpf) in wild-type (E) or sfpq−/− (F) hosts. (M) Quantification of VeLD interneurons in transplanted clones at 30 hpf (value is number of interneurons/total number of transplanted neurons, number of embryos in box). (N) Quantification of axonal lengths of transplanted neurons at 30 hpf. Scale bar, 100 μm.
Figure 4
Figure 4
Sfpq Loss of Function Affects a Specific Set of Transcripts (A) Scattered plot of the expression level of the total probe sets in the array. The 38,489 dots represent total individual probes for genes or expressed sequence tags (ESTs). The gene expression level is expressed in relative arbitrary units of fluorescence intensity on a logarithmic scale. Genes whose expression level is identical between the two groups lie on the diagonal line and the non-identical ones lie outside of the two lines parallel to the central diagonal purple line. The dots positioned above the upper diagonal line represent probe sets that are upregulated and the dots appearing below represent probe sets that are downregulated in the mutants as compared to siblings. Black dots represent those 571 probes with expression level at least 2-fold significantly different (p < 0.05, moderated t test) between sfpq+/−;+/+; (Tg:mnx1 gfp) and sfpq−/−;(Tg:mnx1 gfp) embryos. A high correlation (R2 = 0.98) was observed for the linear regression performed on the gene set. (B) Validation of microarray results by qRT-PCR performed on randomly picked downregulated transcripts. The gene expression data from microarray and qPCR exhibited high correlation (Spearman’s rho test; r(6) = 1, p = 0.01). Statistical comparison was performed using paired t test with Bonferroni correction for multiple comparisons. Fold change is expressed as the ratio between the linear expression levels of sfpq+/−;−/−; (Tg: mnx1 gfp) and sfpq−/−;(Tg:mnx1 gfp) samples and negative value represents downregulation of gene expression. (C) GO analysis of the differentially expressed genes. (D) Analysis of number of splice variants in whole transcriptome and SFPQ-dependent transcriptome. The SFPQ-dependent genes are substantially enriched for high numbers of splice variants (one-sided Wilcoxon-Mann-Whitney rank-sum test with continuity correction, p = 2.51 × 10−26).
Figure 5
Figure 5
SFPQ Protein Is Found in Axons, with Unprocessed RNA Targets and Spliceosome Component U1 snRNP70 Transverse single confocal sections of 24 hpf (A, N, and O) and 48 hpf (B–M) forebrain (A and I–K) or spinal cord dorsal to the top (G, H, L, and M) or to the left (B–D), and lateral view of head, anterior to the left (E and F). (A–F) SFPQ protein expression shown by antibody staining (A–D) or GFP-tagged ectopically expressed protein (E and F). Identity of red, green, and blue staining is labeled on bottom right corner of each picture. Arrow shows axonal localization. (G–K) Localization of first alternative introns (in red) of SFPQ-dependent transcripts, bcas3 (G and H), and gnao1a (I–K). Axons are stained by acetylated tubulin antibody in green. (L and M) High-resolution imaging of bcas3 intron detection (L; red) in motor axons stained with acetylated tubulin antibody (M; green). (N and O) Double antibody staining of U1 snRNP70 proteins (N) with acetylated tubulin (O; merged channels) in spinal axons. Arrows are pointing to motor axons. Scale bar, 25 μm.
Figure 6
Figure 6
Non-nuclear Sfpq Rescues Loss of Transcripts and Axonal Phenotype in the coma Mutant Lateral view of 30 hpf zebrafish (A–D) and spinal cord (G–L), anterior to the left. (A–D) General morphology of sfpq−/− (A); sfpq+/− siblings (B); sfpq−/− injected with GFP-hsfpqΔNLS (C) and sfpq+/− injected with GFP-hsfpqΔNLS (D). Insets showing localization of the human GFP-tagged protein in early ss embryos, CAAX-Cherry in red. (E) Whole-mount detection of SFPQ (red) in a 28 hpf sfpq−/− mutant, injected at one-cell stage with pmnx1:GFP and pCS2:hsfpqΔNLS DNA, expressing these mosaically. Arrow shows motor axon. Note the negative nuclei (arrowhead). (F) Measurement of ventral motor axon length in siblings and homozygous mutant uninjected or injected with RNA (DNLS) or DNA (MOSAIC DNLS) coding for hSFPQΔNLS. For the controls and RNA injections, the ventral motor axons were measured in five embryos over five-somite length upstream of the cloaca. For mosaic DNA-injected sample, measures were done only in seven homozygous mutants, in the same trunk area for the rare SFPQ+ neurons. (G–L) Lateral view, anterior to the left of spinal cords, showing rescue of the axonal defect both in the anterior (G and J) and posterior (H and K) trunk of sfpq−/−;Tg (mnx1:GFP) by injection of the ΔNLS human SFPQ. Sibling axons are unaffected by the injection (I and L). (M) qPCR results for six out of ten DES transcripts from 32 hpf siblings and sfpq−/− mutant embryos uninjected or injected with human ΔNLS sfpq RNA. Bars show expression fold changes compared to the expression level of the transcript in uninjected sibling embryos taken as reference. The four transcripts not plotted did not show any improvement after ΔNLS rescue.
Figure 7
Figure 7
Human Mutations Affect SFPQ Localization and Motor Development in the Zebrafish (A–K) Lateral (A–F) and transverse (G–K, N, and O) views of the spinal cord in sfpq−/− mutant (A, B, and H), siblings (C and G), and homozygous mutant rescued by wild-type human Sfpq (D and I) or L534 (E, J, and O) or N533 (F and K) mutant human Sfpq. (L) Distribution of SFPQ variants in all exome-sequenced human individuals. In red, the position of the variants only found in fALS patients. (M) Schematic of the wild-type hSFPQ protein and the seven variants cloned and tested in double-blind rescue experiments in the Tg(mnx1:GFP) background. S8N (c.23G>A) is present in 4 SALS and 24 ExAC samples, showing a modest enrichment in SALS (p = 0.029, two-tailed Fisher’s test). P51L (c.152C>T) is private to a single SALS patient and absent from all controls. M469V (c.1405A>G) is in a single SALS patient and twice in ExAC (p = 0.071). A515V (c.1544C>T) and A515T (c.1543G>A) are in two and one ExAC normal individuals, respectively, and are located in the coiled-coil domain close to the two FALS variants we identified. (N and O) Close-up of the motor axon and its environment, stained for SFPQ in red and GFP in green in the sfpq−/−; Tg(mnx1:GFP) rescued by injection of the human (N and N′) or L-mutated form (O and O′). Arrowheads show localization of SFPQ in motor axons. (P) Quantification of the α-SFPQ red fluorescent signal in motor axons on confocal stacks, for five pairs of motor axons per 48 hpf embryo stained with SFPQ antibody. Measurement was done in no less than 30 (maximum, 46) embryos per variant injected. Injection is done in sfpq+/−;Tg(mnx1:GFP) incross progeny. The embryos showing poor signal intensity were genotyped and were all homozygous mutants. (Q) Quantification of the length of ventral motor axons. Measures were made for five segments per 48 hpf embryo (five somites anterior to the cloaca) on confocal stacks, using FIJI Single Neurite Tracer in no less than 28 (maximum, 41) embryos per variant injected. Injection is done in sfpq+/−;Tg(mnx1:GFP) incross progeny. The embryos showing shorter axon length were genotyped and were all homozygous mutants. Asterisks indicate highly significant reduction in axon lengths compared to wild-type (pairwise ANOVA, ∗∗∗p < 0.0001). The two significantly different variants show same difference when comparing pairwise to the other variants. All other variants are not significantly different from wild-type. (R) Quantification of ectopic ventral motor branching. Measures were made for five segments per 48 hpf embryo (five somites anterior to the cloaca) on confocal stacks using FIJI Single Neurite tracer in no less than 28 (maximum, 41) embryos per variant injected. Injection is done in sfpq+/−;Tg(mnx1:GFP) incross progeny. The embryos showing excessive branching were genotyped and were all homozygous mutants. Asterisks indicate significance (pairwise ANOVA, ∗∗∗p < 0.0001 and ∗∗p < 0.001) compared to wild-type. The two significantly different variants show same difference when comparing pairwise to the other variants. All other variants are not significantly different from wild-type.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Allende M.L., Weinberg E.S. The expression pattern of two zebrafish achaete-scute homolog (ash) genes is altered in the embryonic brain of the cyclops mutant. Dev. Biol. 1994;166:509–530. - PubMed
    1. Amoyel M., Cheng Y.C., Jiang Y.J., Wilkinson D.G. Wnt1 regulates neurogenesis and mediates lateral inhibition of boundary cell specification in the zebrafish hindbrain. Development. 2005;132:775–785. - PubMed
    1. Barry G., Briggs J.A., Vanichkina D.P., Poth E.M., Beveridge N.J., Ratnu V.S., Nayler S.P., Nones K., Hu J., Bredy T.W. The long non-coding RNA Gomafu is acutely regulated in response to neuronal activation and involved in schizophrenia-associated alternative splicing. Mol. Psychiatry. 2014;19:486–494. - PubMed
    1. Bell T.J., Miyashiro K.Y., Sul J.Y., Buckley P.T., Lee M.T., McCullough R., Jochems J., Kim J., Cantor C.R., Parsons T.D., Eberwine J.H. Intron retention facilitates splice variant diversity in calcium-activated big potassium channel populations. Proc. Natl. Acad. Sci. USA. 2010;107:21152–21157. - PMC - PubMed
    1. Brand M., Granato M., Nuesslein-Volhard C. Keeping and raising zebrafish. In: Nuesslein-Volhard C., Dahm R., editors. Zebrafish, a Practical Approach. Oxford University Press; 2002. pp. 7–37.

Supplementary concepts

LinkOut - more resources