A loss of function allele for murine Staufen1 leads to impairment of dendritic Staufen1-RNP delivery and dendritic spine morphogenesis

Abstract

The dsRNA-binding protein Staufen was the first RNA-binding protein proven to play a role in RNA localization in Drosophila. A mammalian homolog, Staufen1 (Stau1), has been implicated in dendritic RNA localization in neurons, translational control, and mRNA decay. However, the precise mechanisms by which it fulfills these specific roles are only partially understood. To determine its physiological functions, the murine Stau1 gene was disrupted by homologous recombination. Homozygous stau1(tm1Apa) mutant mice express a truncated Stau1 protein lacking the functional RNA-binding domain 3. The level of the truncated protein is significantly reduced. Cultured hippocampal neurons derived from stau1(tm1Apa) homozygous mice display deficits in dendritic delivery of Stau1-EYFP and beta-actin mRNA-containing ribonucleoprotein particles (RNPs). Furthermore, these neurons have a significantly reduced dendritic tree and develop fewer synapses. Homozygous stau1(tm1Apa) mutant mice are viable and show no obvious deficits in development, fertility, health, overall brain morphology, and a variety of behavioral assays, e.g., hippocampus-dependent learning. However, we did detect deficits in locomotor activity. Our data suggest that Stau1 is crucial for synapse development in vitro but not critical for normal behavioral function.

Fig. 1.

Truncated stau1 lacking dsRBD3 leads to reduced Stau1 protein levels and an impairment of RNA-binding capacity. (A) Targeting strategy for disrupting the Stau1 gene in mouse. WT genomic locus (Top) shows the location of the dsRBD3, the sites for homologous recombination with the targeting vector (IRESβgeopA, Middle) and the final targeted, mutated Stau1 locus (Bottom). Arrowheads indicate the primers used for genotyping. 5′ probe, indicates the probe used for Southern blotting. (B) 3′ PCR (Top) and 5′ Southern blot (Bottom) analysis of WT (+/+), heterozygous (+/-), and mutant (−/−) Stau1 mice. For Southern blotting, genomic DNA was EcoRI-digested and probed with a PCR fragment (nt 325–568 of the Stau1 locus, 5′ probe). The DNA from stau1^tm1Apa homozygous mice (−/−) shows a 1.5 kb deletion (from 10 to 8.5 kb) due to the presence of an additional EcoRI site in the recombinant locus. For verification of 3′ targeting (Top, 3′ PCR), primers stauR5 and neo236 were used, generating a 2.9kbp PCR product in the correctly targeted clones. m, marker. (C) PCR genotyping indicating the two resulting PCR fragments for either WT (+/+) or stau1^tm1Apa homozygous mice (−/−). M, marker. (D) Western blot of WT (+/+) and stau1^tm1Apa (−/−) forebrain extracts probed with an immunopurified polyclonal anti-Stau1 antibody (Upper). In mutant brain, a faint band of lower molecular weight is visible (asterisk). The lower panel shows a blot probed with the preimmune serum (PIS) of the Stau1 rabbit. (E) Western blot of WT (+/+) and stau1^tm1Apa (−/−) forebrain extracts probed with an immunopurified polyclonal anti-Stau2 antibody or a monoclonal anti-tubulin antibody as internal loading control. Expression levels of the three Stau2 isoforms (62, 59, and 52 kDa, respectively) are unchanged. (F) Detection of full-length Stau1 and Stau2 mRNAs isolated from brains of WT (+/+), heterozygous (+/-), or stau1^tm1Apa homozygous mice (−/−) by RT-PCR. In contrast to WT (arrowhead), a shortened Stau1 transcript is found in stau1^tm1Apa homozygous mice caused by an in-frame deletion of 207 nts in RBD3 (arrow). As internal control, the dsRBD5 of Stau2 was amplified in parallel. The scheme below shows skipping of exon 5 in stau1^tm1Apa homozygous mice (−/−) leading to a deletion of 69 aa within the dsRBD3 (amino acids in italics, LQ is added in ΔStau1 due to alternative splicing). (G) ΔStau1 lacking RBD3 does not bind RNA in vitro. A Northwestern assay was performed on HeLa cell extracts transfected with WT (+/+) or ΔStau1 (−/−) and probed with an immunopurified polyclonal anti-Stau1 antibody (WB). The asterisk marks endogenous Stau1 protein in HeLa. Proteins were renatured on the filter and exposed to a synthetic ³²P-labeled poly rI/rC probe. Arrowhead indicates unspecific RNA binding to a protein present in all extracts. UT, untransfected control.

Fig. 2.

Hippocampal neurons lacking WTStau1 show impaired RNP transport to distal dendrites. Neurons derived from WT mice are displayed on the left (A, C, E, G, and K) and those derived from stau1^tm1Apa homozygous mice on the right (B, D, F, H, and L). (A and B) stau1^tm1Apa homozygous mice do not express detectable amounts of Stau1. Immunocytochemistry on hippocampal slices using a polyclonal anti-Stau1 (αStau1) antibody shows that Stau1 is expressed in both cell bodies and dendrites (arrows) of WT mice but not of stau1^tm1Apa homozygous mice. (C and D) Cultured hippocampal neurons derived from stau1^tm1Apa homozygous mice do not express Stau1. Immunocytochemistry was performed on hippocampal neurons derived from WT or stau1^tm1Apa homozygous mice using a polyclonal anti-Stau1 antibody. Inset shows an enlargement of the cell body in D with little nuclear staining. Nuclei were visualized with DAPI. (E and F). WTStau1-EYFP forms particles that are transported into distal dendrites of hippocampal neurons derived from WT or stau1^tm1Apa homozygous mice. (G and H) ΔStau1-EYFP needs WTStau1 to form RNPs that are transported into distal dendrites. When WTStau1 is lacking, the ΔStau1-EYFP expression pattern is diffuse with few detectable particles in neuronal dendrites. (I and J) Quantification of the results obtained in E, F, G, and H (letters superimposed on each bar indicate the panel of the representative neuron from each condition). (I) When mutant ΔStau1-EYFP is transfected in neurons derived from stau1^tm1Apa homozygous mice, it results in a failure to form ΔStau1-containing RNPs (81.7 ± 6% SE, P < 0.001). In cases where ΔStau1-containing RNPs do form (I, 18.3 ± 6%, P < 0.001), (J) they are less prominent in distal dendrites as for WTStau1-GFP (45.2 ± 7.2 μm SE, versus 102.9 ± 8.7 μm SE from the cell body, P < 0.001). White bars represent neurons derived from WTStau1 mice; gray bars neurons from stau1^tm1Apa homozygous mice. White bars indicate neurons transfected with WTStau1-EYFP whereas dotted bars indicate neurons transfected with ΔStau1-EYFP. Neurons were fixed 24 h after transfection and their expression pattern was quantified. The maximum distance a WT or ΔStau1-EYFP particle could be detected in all puncta-expressing neurons was measured. (K and L) Detection of β-actin mRNA by FISH. Neurons derived from stau1^tm1Apa homozygous mice (L) show a reduction in the number of distal dendritic β-actin mRNPs compared to their WT counterparts (K). Lower exposure insets of the cell bodies reveal similar expression levels of β-actin mRNA in the soma. (M and N) Quantification of β-actin (M) and polyadenylated (N) mRNA particles. Proximal dendrites were considered as the first 25 μm, distal dendrites as the subsequent 25 μm of dendrite length. Black bars, WT neurons; gray bars, stau1^tm1Apa homozygous neurons. (M) Quantification reveals a significant reduction of β-actin mRNA in distal dendrites of stau1^tm1Apa homozygous mice (P < 0.001). (N) Neurons were probed with oligo(dT) to detect polyadenylated mRNA and quantified as described. Significant reductions of mRNA particles in dendrites of neurons lacking Stau1 were detected both proximally (*, P < 0.05) and distally (**, P < 0.01). (Scale bars: 10 μm.)

Fig. 3.

Neurons lacking WTStau1 show impaired dendritic outgrowth and dendritic spine formation. WT mouse neurons are displayed on the left (A, E, and G) and those from stau1^tm1Apa homozygous mice are displayed on the right (B, F, and H). (A, B, C, and D) 7 DIV neurons derived from stau1^tm1Apa homozygous mice display impaired dendritic outgrowth and branching. (C) Sholl quantification of neurons lacking WTStau1 demonstrates significantly (P < 0.001) less complexity at distances greater than 21 μm from the center of the cell soma. 14 μm from the cell soma, no significant differences were detected. Quantification of primary dendritic outgrowth as well as primary and secondary branch points indicates primary dendrite number is unaffected. The stau1^tm1Apa homozygous mice demonstrate significantly fewer primary branch points (P < 0.01) and significantly fewer secondary branch points (P < 0.001) as compared to WT dendrites. (E, F, J, and K) 18 DIV mature neurons lacking Stau1 show alterations in dendritic spine morphology and density. Two representative neurons from WT (E) and stau1^tm1Apa homozygous mice (F) are shown with one overview (green) and one enlargement (gray scale) of a dendrite from another neuron. (J and K) Quantification reveals that mature neurons from stau1^tm1Apa homozygous mice form significantly fewer protrusions (P < 0.001, ± SE) than WT neurons (J) and that these protrusions are significantly longer in length as compared to WT protrusions (K; P < 0.001, ± SE). (G, H, and I) Neurons lacking WTStau1 have significantly fewer synapses (P < 0.001) and less presynaptic input. (G and H) Synapses were visualized via transfection with PSD95-GFP (green) followed by immunocytochemistry for the presynaptic protein synapsin (red). Enlarged insets show that these two markers are in close proximity indicating functional synapses. High magnifications below show dendrites from another neuron expressing PSD95-GFP. (I) Quantification of F and G reveals a significant reduction in synapse numbers (P < 0.001, ± SE). (Scale bars: 10 μm.)

Fig. 4.

Mice lacking WTStau1 exhibit intact hippocampus-dependent learning but decreased locomotor activity in the open field test. (A) In the Morris water maze, homozygous stau1^tm1Apa mice (n = 17) and WT littermates (n = 20) were trained over multiple days to find a hidden platform at an initial location (acquisition, A1–6) and at a new location on the opposite side of the pool (reversal, R1–3). Escape latency decreased across training trials and did not differ significantly between genotypes. (B) No significant difference between genotypes in preference for the quadrant of the water maze (Ra, right adjacent; La, Left adjacent; Op, opposite) that previously contained the hidden platform (T, target) during probe trials (P1–2). (C) No significant difference in freezing behavior between genotypes. In fear conditioning, homozygous stau1^tm1Apa (n = 13) mice and WT littermates (n = 17) learned to associate an unconditioned stimulus (US), a foot shock, with an auditory conditioned stimulus (CS) and the conditioning chamber itself (Context). (D) In the open field test, homozygous stau1^tm1Apa (n = 19, 9 males and 10 females) and WT littermates (n = 21, 12 males and 9 females) were placed in an automated open field apparatus for 30 min and activity was recorded. Female mice showed less activity overall and homozygous stau1^tm1Apa mice of both sexes exhibited significantly reduced activity than WT littermates. *, P < 0.05; **, P < 0.01.