Active transport of the survival motor neuron protein and the role of exon-7 in cytoplasmic localization

Abstract

Spinal muscular atrophy (SMA) is a neurodegenerative disease caused by deletion and/or mutation of the survival motor neuron protein Gene (SMN1) that results in the expression of a truncated protein lacking the C terminal exon-7. Whereas SMN has been shown to be an important component of diverse ribonucleoprotein (RNP) complexes, its function in neurons is unknown. We hypothesize that the active transport of SMN may be important for neurite outgrowth and that disruption of exon-7 could impair its normal intracellular trafficking. SMN was localized in granules that were associated with cytoskeletal filament systems and distributed throughout neurites and growth cones. Live cell imaging of enhanced green fluorescent protein (EGFP)-SMN granules revealed rapid, bidirectional and cytoskeletal-dependent movements. Exon-7 was necessary for localization of SMN into the cytoplasm but was not sufficient for granule formation and transport. A cytoplasmic targeting signal within exon-7 was identified that could completely redistribute the nuclear protein D-box binding factor 1 into the cytoplasm. Neurons transfected with SMN lacking exon-7 had significantly shorter neurites, a defect that could be rescued by redirecting the exon-7 deletion mutant into neurites by a targeting sequence from growth-associated protein-43. These findings provide the first demonstration of cytoskeletal-based active transport of SMN in neuronal processes and the function of exon-7 in cytoplasmic localization. Such observations provide motivation to investigate possible transport defects or inefficiency of SMN associated RNPs in motor neuron axons in SMA.

Figure 7.

Fusion of SMN exon-7 to the nuclear protein, DBF1, targets it to the cytoplasm. Left panels, EGFP fusion protein signal in nucleus (arrowhead) and cytoplasm (arrows). Right panels, Location of nuclear boundary (arrowheads) using DAPI (red). A, An EGFP-DBF1 transfected neuron presented an exclusive nuclear localization of its EGFP fusion protein (arrowhead). B, DAPI stained nucleus (red). C, In contrast, an EGFP-DBF1/Ex7 transfected neuron showed fluorescence signal in the cytoplasm (arrow) beyond the (D) DAPI stained nucleus (red). E, Proximal segment of exon-7 showed clear nucleus (arrowhead) with two foci and no evidence of EGFP signal accumulation. Cytoplasmic signal (arrows) in the perinuclear region was observed beyond the (F) DAPI stained nuclear border (red). G, A similar pattern was observed when the middle segment of SMN exon-7 was fused to EGFP-DBF1. H, DAPI (red). I, In contrast, the distal segment of the exon-7 was present mostly in the nucleus (arrowhead) with some signal in the perinuclear region (arrow). J, DAPI (red). K, Quantitative analysis of subcellular localization showed the percent of transfected cells with only nuclear signal, signal in both the nucleus and cytoplasm (uniform), and only cytoplasmic signal. L, Schematic of the three subregions of SMN exon-7 (P, proximal; M, middle; D, distal) that were fused to EGFP-DBF1.

Figure 1.

Western blot analysis of SMN expression from rat spinal cord and chick forebrain at different embryonic stages. A, In rat spinal cord, SMN expression was prominent at embryonic stages and declined developmentally. B, In chick forebrain, SMN was also highly expressed at embryonic stages and declined after birth. A, B, Density ratio of SMN to β-actin is shown in the top panel. SMN was detected with a monoclonal antibody and showed a single band at 38 kDa. β-actin was detected with a monoclonal antibody and showed a single band at 42 kDa, blotted on the same membrane as an internal control.

Figure 2.

Localization and microtubule association of SMN granules in neuronal processes and growth cones. A, Rat spinal cord neurons were cultured for 3 d and fixed for double label immunofluorescence staining with a monoclonal antibody to SMN (red) and a rabbit polyclonal antibody to tau (green), as an axonal marker. SMN formed granules that localized to the axonal process (arrows) and its growth cone (arrowhead). B, Cultured chick forebrain neurons stained with the monoclonal antibody to SMN showed SMN granules (red) within minor neurites (arrows) and growth cones (arrowhead). C, Transfection of EGFP-SMN in chick forebrain neurons similarly revealed granules within processes. Microtubules were detected with a monoclonal antibody and Cy3-labeled anti-mouse antibody (red). Z-series stacks were acquired with a cooled CCD camera, and images were deconvolved (Hazebuster; Vaytek). Higher magnification of two regions (a, b) suggest association of EGFP-SMN granules (yellow, arrows) and microtubules. D, Immunogold detection of SMN (arrows) at the EM levels shows proximity to microtubules (arrowheads) within neuronal processes. E, Rat spinal cord neuron showed colocalization of SMN (green) with ribosomal RNA (red). Ribosomal RNA was detected by biotin-labeled oligonucleotide probes and Cy3-conjugated streptavidin. SMN was stained with a monoclonal antibody and Cy5-conjugated anti-mouse antibody. Higher magnification region (c) indicates SMN colocalization with ribosomal RNA (yellow, arrows).

Figure 3.

EGFP-SMN granules exhibit rapid, bidirectional movements in processes that are dependent on either microtubules or microfilaments. A, Small (S) granules frequently exhibited directed movements, both anterograde and retrograde (long white arrows denote trajectories of several small granules). Granules often moved beyond branch points into secondary neurites (see video 1, available at www.jneurosci.org). Large (L) granules often displayed either oscillatory or bidirectional movements over short distances (short black arrows). B-E, Histograms showing frame by frame analysis of granule velocities. Anterograde movements are positive, and retrograde movements are negative. B, Example of an anterograde trajectory of a small granule. C, Example of small granule with a predominantly retrograde trajectory. D, Example of granule showing bidirectional movements. E, Example of a larger granule that showed predominantly nondirected, oscillatory movements. F, Directed movements of small granules were dependent on microtubules and microfilaments. G, Large granules showing nondirected movements, both oscillatory and bidirectional displacements over short distances, were dependent on actin filaments. ^*p < 0.01 when compared with untreated control.

Figure 4.

Association of EGFP-SMN with the actin cytoskeleton in fibroblasts. A, EGFP-SMN was distributed predominantly in the cytoplasm and signal often accumulated at the leading edge of cultured fibroblast (arrow). One or two dots were also observed in the nucleus. B, Corresponding DIC image. C, EGFP-SMN granules were retained after Triton X-100 extraction in buffer (arrow). D, DIC image. E, Retention of EGFP-SMN on the cytoskeleton after colchicine treatment and Triton X-100 extraction (arrows). F, This procedure did not disrupt actin filaments, although microtubules were depolymerized (data not shown). G, In contrast, cytochalasin-D and Triton X-100 extraction released EGFP-SMN, and little signal was apparent. H, Note the disruption of most actin filaments after cytochalasin-D treatment. I, Quantitative analysis of the mean cytoplasmic fluorescence intensity in Triton X-100 extracted cells demonstrated a 70% reduction after cytochalasin-D treatment and a 26% reduction after colchicine treatment.

Figure 5.

EGFP-SMN accumulates in the nucleus after deletion of exon-7. SMN constructs, fused to EGFP, were transfected into cultured chick forebrain neurons to investigate the role of specific exons in subcellular distribution. Left panels, Imaged at 0.05 sec exposure times to show nuclear and perinuclear signal. Right panels, Imaged at 0.5 sec exposure time to show fluorescence signal in processes. Nuclei were stained with DAPI (red). A, Transfection of full-length EGFP-SMN revealed numerous granules in the perinuclear region. In the nucleus, one or two foci were often observed. B, In the same cell at longer exposure, granules were apparent in the processes. C, Exon-5 truncation (EGFP-SMNΔEx5) demonstrated a similar pattern to the full-length protein. SMN granules were observed in the perinuclear region with little signal in the nucleus. D, In the same cell, SMN granules were frequent in the processes. E, F, By contrast, deletion of exon-7 resulted in exclusive nuclear localization (EGFP-SMNΔEx7) with no signal in the perinuclear region (E) or processes (F). G, H, Deletions of both exon-5 and 7 (EGFP-SMNΔEx5&7) also showed an exclusive nuclear localization.

Figure 6.

Nuclear aggregates in the SMN exon-7 deletion mutant were not associated with coilin. A, An EGFP-SMN transfected neuron showed a single nuclear foci (arrowhead). In the cytoplasm, EGFP-SMN granules were observed in processes (arrows). B, A coiled body (red) was detected with an anti-coilin antibody in the nucleus. C, EGFP-SMN foci colocalized to a coiled body in an overlay image (yellow, arrowhead). D, DAPI stained nucleus (blue). E, By contrast, an EGFP-SMNΔEx7 transfected neuron showed nuclear aggregation with several foci (arrows). (see also video 2, available at www.jneurosci.org). F, Coiled bodies (red) were detected in the nucleus. G, Overlay of images indicated two foci of EGFP-SMNΔEx7 association with coiled bodies. H, DAPI stained nucleus of the same neuron (blue).

Figure 8.

Redistribution of nuclear EGFP-SMNΔEx7 by cytoplasmic targeting sequences. A, Fusion of the QNQKE sequence to EGFP-DBF1 (EGFP-DBF1/tif) revealed signal in the cell body (arrows), but not in a granular pattern in neurites (arrow). No signal was observed within distal neurites. B, Fusion of the QNQKE sequence to EGFP-SMNΔEx7 (EGFP-SMNΔEx7/tif) resulted in the presence of granules within the cell body and processes (arrows). C, Fusion of the membrane targeting sequence of GAP-43 to EGFP-SMNΔEx7 (mem/EGFP-SMNΔEx7) resulted in its redistribution into the cytoplasm and localization of granules in neurites (arrows; see video 3 for demonstration of its active transport in axons and video 4 for example of granule transport from the cell body into proximal neurite, both available at www.jneurosci.org). Only weak signal from the mem/EGFP-SMNΔEx7 was observed in the nucleus. D, The GAP-43 targeting sequence also promoted EGFP-DBF1 localization in the cytoplasm (arrow). A-D, DAPI stained nuclei (red), approximate location of nuclear border (arrowhead).

Figure 9.

Decreased neurite length after overexpression of SMNΔEx7 was rescued by fusion of GAP-43 targeting sequence. The average length of the measured neurites is indicated in the histogram. Transfection of the SMNΔEx7 showed a 25% decrease in neurite length in comparison with full-length SMN. Addition of a GAP-43 membrane-targeting sequence to SMNΔEx7 did not show a reduction of the neurite length compared with wild-type SMN. ^**p < 0.01; ^*p < 0.05 (Student's t test).