Dilysine motifs in exon 2b of SMN protein mediate binding to the COPI vesicle protein α-COP and neurite outgrowth in a cell culture model of spinal muscular atrophy

NSC-34 cells as a model of spinal muscular atrophy

To model SMA in vitro, we chose the NSC-34 cell line. These immortalized cells were derived from a hybrid of neuroblastoma and motor neuron-enriched embryonic mouse spinal cord cells and display many characteristics of motor neurons (16). We used the pSuperior system to create stably transfected NSC34 cells expressing the TetR protein. The highest expressing clonal line was selected and transfected with pSuperior.puro containing a short-hairpin RNA (shRNA) directed against murine SMN. Selection followed by clonal isolation resulted in NSC-34 clone #56, which produces robust knockdown of SMN protein to ∼10–20% of normal levels within 48 h (Fig. 1A). These cells maintain viability under SMN knockdown for up to 14 days, but demonstrate a marked reduction in their proliferation. The impact of SMN depletion on the ability of these cells to form neuronal processes in culture was then assessed. Cultures were maintained in Dulbecco's modified Eagle's medium (DMEM) with 10% FBS for 48 h in the presence or absence of doxycycline to induce SMN knockdown. Following depletion of SMN protein, the cultures were plated on poly-D-lysine coated glass coverslips in DMEM/F12 with 1% FBS to induce neuronal differentiation. After 3 days of culture in neuronal differentiation media, the cells were fixed and stained with α-tubulin antibodies to visualize processes. Neurite length and number were measured for control and SMN-depleted cultures. While control cells produce lengthy neurites and mostly bipolar or pyramidal cellular morphologies, SMN-depleted cultures produced significantly shorter neurites (P < 0.01) similar to the recent reports of poor axonal outgrowth in SMN-depleted motor neuron derived from human embryonic stem cells (17) and adopted a more stellate morphology resulting in a statistically significant shift in the number of neurites per cell (P < 0.05) (Fig. 1B–E). These changes persisted as far out as 120 h in culture. Interestingly, SMN-depleted cells also display what appear to be prominent tubulin-immunopositive microtubule organizing centers, which is consistent with previous reports of altered tubulin dynamics in SMN-depleted cells (18).

To determine whether the neurite outgrowth phenotype in NSC-34 clone #56 was SMN-dependent, two experiments were carried out. First, cells from the parent NSC-34 clone expressing only the TetR protein were cultured in the presence and absence of doxycycline. No knockdown of SMN protein was detectable following doxycycline treatment. When cells were grown in neuronal differentiation media, no alteration in neurite length or number was detectable in doxycycline-treated cultures (data not shown). Moreover, we were able to rescue the defects in both neurite outgrowth and number by transiently transfecting the cultures with eGFP-tagged human SMN (a generous gift from Philip J Young) on the day of neuronal induction. The eGFP-SMN is mainly cytoplasmic and punctate, and fluorescence extends throughout the length of the cell body and neurites. Transfection with human SMN fully rescues neurite length and significantly shifted the distribution of neurites per cell toward that of control cultures (Fig. 2). These experiments combine to show that a decreased neurite length and altered morphology in NSC-34 clone #56 upon doxycycline treatment is a result of SMN depletion. Interestingly, eGFP-SMN bearing the E134 K mutation, which is associated with severe disease presentation (3), did not rescue neurite outgrowth in these cells, while there was modest rescue seen with a construct containing the A2G mutation (Fig. 3), which is associated with a milder SMA phenotype (19). Similar rescue of neurite number was seen with the A2G mutant and the wild-type SMN, but no significant rescue of neurite number was seen in cells transfected with the E139 K mutant as determined by Chi-squared analysis. Western blot analysis from cultures transfected with wild-type, A2G or E139 K eGFP-SMN shows that expression levels were comparable and the eGFP-SMN constructs are not affected by the doxycycline-induced knockdown of murine SMN (Supplementary Material, Fig. S2).

The SMN protein contains a number of domains that are known to mediate its subcellular localization and its ability to interact with various binding partners (20,21). To determine which regions of SMN are important for its ability to rescue neurite outgrowth in our NSC34 model, we used a series of eGFP-tagged human SMN constructs lacking exons 1 through 6. SMN lacking exon 7 was very unstable as has been previously reported (22), and could not be expressed in these cells at detectable levels for the required 3 days in cultures and thus was removed from the analysis. As seen in Figure 4A and quantified in Figure 4G, SMN lacking exon 1 (SMN Δ1) adopted a punctate cytoplasmic distribution, extending throughout the cell body and neurites. This construct was able to rescue neurite outgrowth and was indistinguishable from full-length eGFP-tagged SMN. Similar results were seen with constructs lacking exons 4 or 5. Although SMN lacking exon 5 adopted a much more diffuse subcellular distribution, both SMN Δ4 and SMN Δ5 rescued neurite outgrowth in doxycycline-treated cultures. In contrast, SMN lacking exons 2b, 3 or 6 displayed diffuse nuclear and cytoplasmic distribution and failed to restore neurite outgrowth or normal morphology in doxycycline-treated cultures (Fig. 4B, D, F and G).

We were particularly interested in the failure of SMN Δ2b to rescue neurite outgrowth in SMN-depleted NSC34 clone #56 cultures. Our previous in vitro binding experiments demonstrated that SMN exon 2b was sufficient for binding of α-COP (10). In support of the role of a-COP in neurite outgrowth, over-expression human α-COP co-transfected with eGFP in either control or SMN-depleted NSC34 clone #56 cultures resulted in a significant rescue of neurite length, similar to the rescue achieved with transfection of human SMN (Fig. 5).

As shown in Figure 4, SMN lacking exon 2b fails to rescue neurite outgrowth. Several proteins that bind to α-COP contain characteristic lysine-rich motifs (K–K or K×K) (23) and exon 2b contains these lysine motifs at amino acids 65–x–67, 71–x–73, 76–77-x-79 and 82–83 (Fig. 5H). To determine whether these putative a-COP binding residues affect the ability of SMN to rescue neurite outgrowth, we used two incomplete deletions of exon 2b. The first, SMN Δ2b-1, lacks only the first eight amino acids of exon 2b (ALKNGDIC) and retains all four potential α-COP binding lysine motifs. As shown in Figure 6A, eGFP-hSMN Δ2b-1 adopts a cytoplasmic punctate pattern similar to full-length SMN and is able to rescue neurite outgrowth in SMN-depleted cultures (Fig. 6C). The second construct, SMN Δ2b-2 contained a larger deletion within exon 2b that removes lysines 65–76 but retains lysines 82–83. This construct demonstrated a diffuse nuclear and cytoplasmic staining pattern and failed to neurite outgrowth (Fig. 6B and C). This indicates that the region of exon 2b containing the putative α-COP binding residues is critical for the ability to rescue neurite outgrowth.

To examine the ability of these lysine residues to bind a-COP protein in vitro, we used GST-tagged fragments of SMN exon 2. The first was a fusion of exon 2a with the first 20 amino acids of exon 2b (SMN 2a+). This contained lysine 65, 67 and 71 but lacked the lysines from 76–82. We compared the behavior of this 2a+ fragment with GST-tagged exon 2b containing all four lysines. Flag-tagged human α-COP was transcribed and translated in vitro, and the resulting reticulocyte lysate was incubated with GST-tagged SMN 2b or SMN 2a+ immobilized on glutathione beads. Blotting bound proteins with an anti-FLAG antibody showed that exon 2b successfully bound α-COP but exon 2a+ that lacks lysines 76–82 did not (Fig. 6D). These experiments together demonstrate that lysines 76, 77 and 79 appear to be crucial for both binding of α-COP to SMN and for SMN-dependent neurite outgrowth.

We used site-directed mutagenesis to mutate each of the putative a-COP binding dibasic motifs in eGFP-tagged SMN. Alanine mutations were introduced into GST-tagged exon 2b, although in this case, only lysines 76 and 82 were altered. In cultured SMN-depleted cells, eGFP-tagged SMN K65A and K71A were indistinguishable from wild-type proteins, adopting a cytoplasmic punctate appearance with fluorescence throughout the length of the cell body and neuritic processes, and significant rescue of neurite length compared with non-transfected neighbors (Fig. 7A and B). In contrast, SMN K76A and SMN K82A predominantly appear to diffuse, evenly distributed throughout the nucleus and the cytoplasm, and fail to rescue neurite outgrowth in SMN-depleted cultures (Fig. 7C and D). When these mutations were introduced into GST-tagged SMN exon 2b, they destroyed the binding of in vitro translated FLAG-a-COP (Fig. 7F). We infer that dilysine motifs 76–77 and 82–83 mediate α-COP binding to SMN exon 2b and are important for rescue of neurite outgrowth in SMN-depleted NSC34 cells. When transfected in HeLa cells, these proteins co-immunoprecipitate at equal levels with Gemin2 when compared to wild type SMN (Supplementary Material, Fig. S1a and b). When co-transfected into NSC-34 cells along with HA-tagged human SMN, all mutants equally co-immunoprecipitated with antibodies against HA (Supplementary Material, Fig. S1d), implying that despite being within the exon 2b self-association domain (21), these retain the ability to oligomerize. Finally, when expressed in HeLa cells, the eGFP-SMN mutants all co-localize with Gemin2-positive puncta (Supplementary Material, Fig. S1c). Taken together, these data demonstrate that we have disturbed the COPI binding in exon 2b without significantly altering other SMN protein–protein interactions.

Generation of SMN knockdown NSC-34 cells

To establish an inducible system for knocking down the expression of mouse SMN gene, we started with generating motor neuron-like cells NCS-34 that stably expresses Tet repressor protein (tet-R). For that purpose, NSC-34 cells were transfected with pcDNA6/TR. Individual clones were isolated after 4 weeks of selection with 10 mg/ml of blasticidin S HCl (Invitrogen). To identify a clone with the highest level of tet-R we used a reporter construct, pcDNA4/TO-luciferase. In this construct Firefly luciferase is expressed from a CMV promoter that contains tet-R binding sites. More than 100 of randomly selected individual clones were screened for their ability to repress luciferase expression. Briefly, each clone was transiently transfected with pcDNA4/TO-luciferase and allowed to grow in the absence or presence of 1 mg/ml doxycycline (Sigma). Approximately 30 h after transfection cells were lysed in Luciferase lysis buffer (Promega), and assayed using a Firefly Luciferase assay kit (Promega). NSC-34 clone #4 (NSC-34–4) that showed the highest difference (∼10 times) in luciferase amount when grown with and without doxycycline was subsequently used to establish the inducible shRNA expression system. To express shRNAs of interest, we used a pSUPERIOR.puro plasmid (Oligoengine). In this plasmid, shRNA is expressed from the H1 promoter that contains the tet-R binding site. The shRNA GGAGAAUGAAAGUCAAGUU that targets the junction of exon3/exon4 was cloned into a pSUPERIOR.puro vector following manufacturer's recommendations. Also, shRNA that targets Photinus pyralis luciferase mRNA sequence CGUACGCGGAAUACUUCG was used as a negative control. The silencing efficiency of the shRNA was first confirmed using a psiCHECK-moSMN construct, in which mouse SMN cDNA was fused to the Renilla luciferase gene using XhoI and NotI restriction sites in the psiCHECK-2 plasmid (Promega). Briefly, NSC34–4 clone was co-transfected with psiCHECK-moSMN (25 ng) and the pSUPERIOR.puro-shRNA (175 ng) in a 24-well plate, and luciferase activity was measured 30 and 50 h later using a Dual-Glo Luciferase assay (Promega). The shRNA caused >90% decrease in luciferase activity when compared with the control shRNA as well as mock-transfected cells. The pSUPERIOR.puro-shRNA construct was transfected into NSC-34–4 cells and individual clones were isolated after 4 weeks of selection with 2.5 mg/ml of puromycin. To test for SMN knockdown, >30 randomly selected individual clones were tested by RT–PCR for mRNA levels after being grown for 3 days in the absence or presence of 2 mg/ml doxycycline. Clones with a significant decrease in SMN mRNA levels upon doxycycline addition in growth medium were then used in western blot analysis. Clone NSC-34–4 #56 that expresses shRNA was selected for subsequent experiments.

Neurite outgrowth assays

Prior to induction of neurite outgrowth, cells were maintained in either vehicle or doxycycline-treated (2 mg/ml) growth medium (DMEM, 10% FBS) for 48 h to reduce the levels of SMN. Following induction of knockdown, the cells were plated on PDL-coated 18 mm glass coverslips in Neurite outgrowth media (DMEM/F12, 1% FBS) at a density of 20 000 cells/ml. eGFP-SMN and alpha-COP expression vectors (1 µg DNA per well) were introduced at the time of plating by reverse transfection with Lipofectamine 2000 for 6 h to yield a transfection efficiency of 5–10%. Following transfection, the media was replaced with either vehicle or doxycycline-treated neurite outgrowth media and replaced every 48 h. Following 72 h of neurite extension, cells were fixed in 4% paraformaldehyde. For immunohistochemistry, fixed cells were incubated in blocking buffer (PBS + 5% normal goat serum and 1% Triton-X 100). Neurites were stained with a monoclonal antibody against α-tubulin (1:10 000) followed by goat anti-mouse AlexaFluor 598 fluorescent secondary. Images were captured at ×20 magnification using the QuickPro software, and the neurite length was determined using the measure function. Neurites were measured from at least 50 cells from two separate culture dates. Statistically significant changes in neurite length were determined by Student's t-test. Significant shifts in the number of neurites per cell relative to control cultures were determined by Chi-squared analysis. All error bars represent the standard error of the mean.

Site-directed mutagenesis

Mutations were introduced into previously described eGFP-SMN and GST-SMN exon 2b constructs using the QuickChange system. The following primer pairs were generated to introduce the K65A, K71A, K76A and K82A mutations into these vectors: K65AF 5′-GTGAAACTTCGGGTGCACCAAAAACCACACC-3′, K65AR 5′-GGTGTGGTTTTTGGTGCACCCGAAGTTTCAC-3′, K71AF 5′-GTAAACCAAAAACCACACCTGCAAGAAAACCTGCTAAGAAG -3′, K71AR 5′-CTTCTTAGCAGGTTTTCTTGCAGGTGTGGTTTTTGGTTTAC-3′, K76AF 5′-CACCTAAAAGAAAACCTGCTGCCGCGAATAAAAGCCAAAAGAAG-3′, K76AR 5′-CTTCTTTTGGCTTTTATTCGCGGCAGCAGGTTTTCTTTTAGGTG-3′, K82AF 5′-CTAAGAAGAATAAAAGCCAACAGCAGAATACTGCAGCTTCCTTAC-3′, K82AR 5′-GTAAGGAAGCTGCAGTATTCTGCTGTTGGCTTTTATTCTTCTTAG-3′.

Expression of GST-fusion proteins

GST-SMN exon 2a+ and GST-SMN exon 2b fusions were cloned into pGex4T-3 as previously described. The constructs were transformed into DH5a cells, and a single colony was isolated and grown to OD600. Cells were collected and lysed following 3 h of IPTG induction, and GST fusions were bound to High Performance Glutathione Sepharose (GE Healthcare) for 1 h at 4°C. Protein expression was confirmed by sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) of 20 µl of beads in 2× sodium dodecyl sulphate (SDS) sample buffer on a 4–15% Tris–Glycine gel followed by visualization with Coomassie blue stain. The GST-SMN fragments run at ∼32 and 31 kD, respectively. For α-COP-binding studies, FLAG-tagged human α-COP was cloned into the pTNT vector as previously described and expressed from the T7 promoter using the TNT T7 Quick Coupled Transcription/Translation System (Promega). Immobilized GST–SMN fusions were incubated with FLAG-α-COP protein for 4 h at 4°C and washed five times. Bound proteins were removed in 2× SDS sample buffer and separated by SDS–PAGE on a 4–15% Tris–glycine gradient gel. Proteins were transferred to PVDF membrane and visualized by western blotting with an anti-FLAG antibody (M2, 1:2500) or antibodies against SMN exon 2 (MANSMA 1).