Survival motor neuron function in motor axons is independent of functions required for small nuclear ribonucleoprotein biogenesis

Abstract

Spinal muscular atrophy (SMA) is a motor neuron degenerative disease caused by low levels of the survival motor neuron (SMN) protein and is linked to mutations or loss of SMN1 and retention of SMN2. How low levels of SMN cause SMA is unclear. SMN functions in small nuclear ribonucleoprotein (snRNP) biogenesis, but recent studies indicate that SMN may also function in axons. We showed previously that decreasing Smn levels in zebrafish using morpholinos (MO) results in motor axon defects. To determine how Smn functions in motor axon outgrowth, we coinjected smn MO with various human SMN RNAs and assayed the effect on motor axons. Wild-type SMN rescues motor axon defects caused by Smn reduction in zebrafish. Consistent with these defects playing a role in SMA, SMN lacking exon 7, the predominant form from the SMN2 gene, and human SMA mutations do not rescue defective motor axons. Moreover, the severity of the motor axon defects correlates with decreased longevity. We also show that a conserved region in SMN exon 7, QNQKE, is critical for motor axon outgrowth. To address the function of SMN important for motor axon outgrowth, we determined the ability of different SMN forms to oligomerization and bind Sm protein, functions required for snRNP biogenesis. We identified mutations that failed to rescue motor axon defects but retained snRNP function. Thus, we have dissociated the snRNP function of SMN from its function in motor axons. These data indicate that SMN has a novel function in motor axons that is relevant to SMA and is independent of snRNP biosynthesis.

Figure 1.

Classification of motor axon defects caused by Smn reduction. Motor axons were scored in Tg[gata2:gfp] embryos at 2 dpf. When compared with larvae in which motor axonal pathways follow the stereotypical pattern (A), severe fish display multiple truncations with (B) or without branching. Moderate fish show a more chimeric display of defects ranging from multiple nerves innervating the neighboring myotome (C) to a single truncation (D).

Figure 2.

Analysis of the Smn protein levels in the morphant classification groups. Smn Western blot analysis using protein from 2 dpf embryos scored as severe (lane 2), moderate (lane 3), mild (lane 4), and no defects (lane 5) compared with wild type (lane 1). β-Actin is used as a loading control.

Figure 3.

Diagrams of the human RNAs used to analyze SMN function in motor axons. SMN exons (ex) and the name of the constructs are denoted. Asterisks indicate the location of single nucleotide changes in the SMN sequence.

Figure 4.

Full-length hSMN, but not SMN lacking exon 7, rescues motor axon defects caused by reduction of Smn. Tg[gata2:gfp] zebrafish injected with smn MO and scored at 2 dpf (n = 407 fish, 8140 nerves) using criteria in Tables 1 and 2 resulted in a distribution of fish with motor nerve defects. Coinjection of smn MO with full-length hSMN RNA was able to partially rescue the nerve defects (p < 0.001; n = 110 fish, 2200 nerves), whereas coinjection with hSMNΔ7 did not (p = 0.97; n = 192 fish, 3840 nerves). Control (ct) MO (n = 86 fish, 1720 nerves) and hSMN RNA alone (n = 104 fish, 2080 nerves) are shown as controls. For details on the statistical analysis for this and similar figures, see Materials and Methods.

Figure 5.

RNAs encoding human SMA mutations fail to rescue motor axon defects caused by the reduction of Smn. Motor nerves in G279V RNA coinjected fish (p = 0.77; n = 117 fish, 2340 nerves) and Y272C RNA coinjected fish (p = 0.26; n = 89 fish, 1780 nerves) were not statistically different from smn morphant fish.

Figure 6.

Addition of the VDQNQKE motif to the SMNΔ7 sequence rescues motor axon defects caused by the reduction of Smn. A, Alignment of exons 6 and 7 from different species. Arrow denotes the beginning of exon 7 (G279 in human). The QNQKE motif (red) has a high degree of evolutionary conservation most notably the first glutamine (Q282) and the glutamic acid (E286). B, Motor nerves in hSMNΔ7-VDQNQKE RNA (n = 117 fish, 2340 nerves) coinjected fish were significantly different when compared with motor nerves in smn morphant fish (p < 0.001) and equivalent to the rescue seen with the full-length hSMN RNA (p = 0.54).

Figure 7.

Amino acids added to SMNΔ7 enable transport into axons, but only exon 7-specific sequences rescue motor axon defects. Fluorescently tagged DNA constructs were transfected into cultured chick cortical neurons, and the distribution of the tagged protein was analyzed. Arrows indicate the presence of SMN protein in the axonal projections. Like wild-type SMN protein (A), SMNΔ7-VDQNQKE (B) and SMNΔ7 read-through (C) proteins are transported out of the nucleus and into the axons. D, SMNΔ7-VDQNQKE rescued motor axon defects in smn morphants (see Fig. 6). This was not the case with SMNΔ7 read-through (n = 85 fish, 1700 nerves), which was not statistically different from SMNΔ7 (p = 0.11) or smn MO alone (p = 0.08).

Figure 8.

Evaluation of the QNQKE motif. Substitution of an alanine for the glutamic acid (E286A) followed by coinjection with smn MO rescues as well as hSMN RNA (p = 0.10; n = 152 fish, 3040 nerves). The same change at the first glutamine residue (Q282A; n = 106 fish, 2120 nerves) coinjected with smn MO leads to defects that are statistically worse than hSMN RNA coinjections (see Fig. 4) (p < 0.001).

Figure 9.

Analysis of SMN binding properties. The Q282A mutation maintains the ability to self-associate (A), bind to Sm proteins (B), and bind to wild-type SMN (C) and does so in a manner that is similar (within 2 SDs) to full-length SMN (FL-SMN). This is unlike the other three proteins tested: hSMNΔ7 (Δ7), hSMNΔ7-VDQNQKE (QNQKE), and hSMNΔ7 read-through (RT) that showed minimal ability to perform these functions. The human mutation A111G can self-associate and has moderate Sm and SMN binding. The Sm protein used in this assay is the neuronal form (SmN), which is most similar to SmB/B⁻.

Figure 10.

Sm core assembly does not lead to motor axon rescue. RNA encoding SMA mutation A111G, which retains the ability to undergo Sm core assembly, does not rescue motor axon defects caused by low Smn levels. n = 92 fish, 1840 nerves; p = 0.12 versus smn MO alone; p < 0.001 versus smn MO plus hSMN.

Figure 11.

Severity of motor axon defects correlates with decreased survival. A, smn morphant fish were scored (severe, moderate, mild, and no defects), placed in individual housing tubes, and monitored up to 30 dpf. Severe fish (n = 27) died more abruptly, and to a greater extent, than the other groups (p = 0.006 vs moderates; p < 0.001 vs all other groups). Moderate (n = 52) and mild (n = 42) fish also showed a significant difference when compared with the control MO (n = 24) injected fish (p = 0.006 and 0.05, respectively). The fish with no observable differences (n = 20) were not significantly different from the mild category (p = 0.619) or the control MO group (p = 0.236). The control MO and uninjected (n = 48) fish had no difference in their survival (p = 0.308). B, To evaluate the effect of SMN forms on the survival of the fish, RNA was coinjected with smn MO and survival was monitored. Compared with the smn MO-injected fish, the p values are as follows for the coinjections: hSMN, p < 0.001, n = 93; hSMNΔ7, p = 0.152, n = 79; hSMNΔ7-VDQNQKE, p < 0.001, n = 76; Q282A, p = 0.224, n = 106; control MO, p < 0.001, n = 39; uninjected, p < 0.001, n = 48. The solid dots at the end of each plotted line are representative of fish that were still alive at the end of the survival assay. The difference in color of the last dot and the line is an artifact of the SPSS software.