Zebrafish survival motor neuron mutants exhibit presynaptic neuromuscular junction defects

Abstract

Spinal muscular atrophy (SMA), a recessive genetic disease, affects lower motoneurons leading to denervation, atrophy, paralysis and in severe cases death. Reduced levels of survival motor neuron (SMN) protein cause SMA. As a first step towards generating a genetic model of SMA in zebrafish, we identified three smn mutations. Two of these alleles, smnY262stop and smnL265stop, were stop mutations that resulted in exon 7 truncation, whereas the third, smnG264D, was a missense mutation corresponding to an amino acid altered in human SMA patients. Smn protein levels were low/undetectable in homozygous mutants consistent with unstable protein products. Homozygous mutants from all three alleles were smaller and survived on the basis of maternal Smn dying during the second week of larval development. Analysis of the neuromuscular system in these mutants revealed a decrease in the synaptic vesicle protein, SV2. However, two other synaptic vesicle proteins, synaptotagmin and synaptophysin were unaffected. To address whether the SV2 decrease was due specifically to Smn in motoneurons, we tested whether expressing human SMN protein exclusively in motoneurons in smn mutants could rescue the phenotype. For this, we generated a transgenic zebrafish line with human SMN driven by the motoneuron-specific zebrafish hb9 promoter and then generated smn mutant lines carrying this transgene. We found that introducing human SMN specifically into motoneurons rescued the SV2 decrease observed in smn mutants. Our analysis indicates the requirement for Smn in motoneurons to maintain SV2 in presynaptic terminals indicating that Smn, either directly or indirectly, plays a role in presynaptic integrity.

Figure 1.

Zebrafish smn mutations. (A) Schematic diagram showing the smn mutations smnY262stop in exon 6, smnG264D, and smnL265stop in exon 7. (B) Clustal W (1.82) alignment of zebrafish Smn exons 6 and 7 (black, vertical line demarcates the boundary) with the human homologue showing mutated amino acids. (C) Schematic diagram showing wild-type and mutant Smn proteins. (D) dCAPS of zebrafish smn mutations. The PCR product of wild-type (lane 1, 4, 7), heterozygous (lanes 2, 5, 8) and homozygous smn mutants (lanes 3, 6, 9). aa, amino acid; wt, wild-type. *= identical amino acid, : = conserved substitution, . = semi-conserved substitution.

Figure 2.

smn mutants are larval lethal. (A) Progeny from incrosses were grown in the same nursery tank and monitored until 20 dpf (smnY262stop, and smnL265stop) or 30 dpf (smnG264D, wild-type). Fish were genotyped by dCAPS and survival of homozygous mutants was plotted (wild-type, green line, n = 143; smnY262stop^−/−, red line, n = 42; smnG264D^−/−, black line, n = 35; smnL265stop^−/−, blue line, n = 33). The average survival for both smnY262stop^−/− and smnL265stop^−/− larvae was 12 dpf with ∼80% dying between 11 and 13 dpf. The average survival for smnG264D^−/− larvae was 17 dpf with ∼80% dying between 15 and 21 dpf. Survival of all mutants was significantly different compared with wild-types (P < 0.0001). (B) Image of live 11 dpf wild-type and smnY262stop^−/− larvae. (C) smnY262stop^−/− (mutant, n=63) were smaller than wild-type (n = 55; P < 0.0001), and smnY262stop^−/+ larvae (n = 114; P < 0.0001) at 11 dpf. Scale bar, 300 µm.

Figure 3.

Smn protein levels in smn mutants. Smn protein levels from (A) wild-type (wt) and (B) smnY262stop^−/− larvae at 3–8 dpf. (C) wt and smnY262stop^−/− at day 9–11 dpf. (D) wt, heterozygous (hets) and smnL265stop^−/− larvae at 11 dpf (E) wt and smnG264D^−/− larvae at 10 and 15 dpf. Each lane was loaded with protein extracts from three larvae. Protein mass markers (kD) are present in A and α-tubulin served as loading control for all blots.

Figure 4.

SV2 is decreased at NMJs in smn mutants. (A) Wild-type (wt, n = 10), (B) smnY262stop^−/− (n = 14), (C) smnL265stop^−/− (n = 8) larvae at 11 dpf. (D) wt (n = 10) and (E) smnG264D^−/− (n = 11) larvae at 16 dpf. Presynaptic regions were labeled with SV2 antibodies (red) and postsynaptic regions were labeled with α-bungarotoxin (α-bgt; green). Merge is an overlay (yellow). (F) The coefficients of co-localization were plotted. All mutants were significantly different compared with wild-types (P < 0.0001). Scale bar, 80 µm.

Figure 5.

Synaptotagmin is not affected in smn mutants. (A) Wild-type (wt, n = 8) and smnY262stop^−/− (n = 10) labeled with synaptotagmin II antibody (synapt, red), and α-bgt (green) at 11 dpf. (B) The coefficients of co-localization were plotted. Mutants were not significantly different from wild-types (P = 0.50). Scale bar, 80 µm.

Figure 6.

Transgenic expression of hSMN in zebrafish motoneurons. (A) Schematic diagram showing the hb9:hSMN transgene construct with hb9:mCherry marker. (B) Lateral views of live Tg(hb9:hSMN;hb9:mCherry) (abbreviated Tg(hb9:hSMN)) zebrafish at 3 dpf. Top, brightfield; middle, fluorescence; bottom, merge. (C) Multiplex RT–PCR of wild-type (wt) and Tg(hb9:hSMN) (Tg) with human and zebrafish smn primers. (D) Kaplan–Meier survival curves of smnY262stop^−/− larvae and smnY262stop^−/−;Tg(hb9:hSMN). The average survival for smnY262stop^−/− with the hb9:SMN transgene was significantly different from those without the transgene (P < 0.0001). Blue line, Tg(hb9:hSMN) (n = 132); Green line, Y262stop^+/−; Tg(hb9:hSMN) (n = 64); black line, Y262stop^−/−; Tg(hb9:hSMN) (n = 43); red line, Y262stop^−/− (n = 42). Scale bar, 250 µm.

Figure 7.

hSMN expression in motoneurons rescues the NMJ defect in smn mutants (A) Immunolabeling of postsynaptic regions (α-bgt, green), and presynaptic regions (SV2, red) of 11 dpf smnY262stop^−/− larvae (n = 14), Tg(hb9:SMN) (n = 10) and smnY262stop^−/−;Tg(hb9:hSMN) larvae (n = 14). (B) The coefficients of co-localization were plotted and the means of each group were calculated. (C) The distribution of log ratio of presynaptic only regions versus postsynaptic only regions. Wild-type zebrafish had a ratio close to zero, indicating the same amount of pre- and postsynaptic regions. The smnY262stop^−/− larvae (red) had a reduced ratio of co-localization when compared with wild-type zebrafish (blue), but increased with addition of the hb9:hSMN transgene (black). For co-localization, smnY262stop^−/−;Tg(hb9:smn) versus wt, P = 0.1; smnY262stop^−/−;Tg(hb9:smn) versus smnY262stop^−/−, P < 0.0001. For pre/post log-ratios, smnY262stop^−/−;Tg(hb9:smn) versus wt, P = 0.482; smnY262stop^−/−;Tg(hb9:smn) versus smnY262stop^−/−, P < 0.0001. For all the tests, the P-values for smnY262stop^−/− versus wt <0.0001. Scale bar, 40 µm.