Editor—Spinal muscular atrophy (SMA) is an autosomal recessive disease resulting from mutations in the telomeric copy of the survival motor neuron gene (SMN1),1-7 which results in reduced expression of the survival of motor neuron protein (SMN).4
7 The SMN protein is ubiquitously expressed but is found at high levels in motor neurons.4
8 The SMN protein associates with Sm proteins,9
10 SIP-1 protein,9
10 and itself.8-11 SMN is found in structures termed gems8 that are associated with coiled bodies in the nucleus. The SMN protein is involved in RNA biogenesis10 and is important for the maturation of a functional snRNP complex that performs splicing.12 The complete loss of SMN is lethal13 whereas the low levels of SMN found in SMA cause loss of the motor neurons.4
7 The mechanism by which the reduction of SMN protein results in the loss of motor neurones is unknown. Some groups have suggested it occurs by apoptosis.14
15 Apoptosis is a conserved, highly regulated mechanism of non-chronic cell death for the removal of surplus, aged, or damaged cells.15 Apoptosis is regulated by the interactions of apoptosis agonists and antagonist with the Bcl-2 protein being one of the key inhibitors of apoptosis.16
Recently Iwahashi et al
17 have suggested a direct interaction between SMN and Bcl-2 using transfected constructs. In an effort to confirm and extend their results, we have attempted to coimmunoprecipitate SMN and Bcl-2 both in a native environment and using transfected cells. The SMN protein and Bcl-2 are expressed in Jurkat cells and in spinal cord. Jurkat cells are a human lymphoblast T cell line that has been previously shown to express Bcl-2 and can be induced to undergo apoptosis.18
Immunoprecipitation experiments were performed using two different methods.17
20 Fig 1A shows a western blot of the immunoprecipitation reactions using the method of Iwahashiet al.17 Jurkat cell lysate was precipitated with anti-SMN (MANSMA2), anti-Bcl-2 (Bcl-2(100)), and anti-dystrophin (MANDYS-1) monoclonal antibodies. The precipitated proteins were western blotted and probed with anti-SMN or anti-Bcl-2 polyclonal antibodies. As shown in fig 1A-1, anti-SMN polyclonal antibody SMN-C3 detects SMN only in the immunoprecipitation using anti-SMN monoclonal antibody MANSMA2 and not in the immunoprecipitation reaction using anti-Bcl-2 monoclonal Bcl-2(100). The reciprocal experiment is shown in fig 1A-2 in which the blot is probed with anti-Bcl-2 polyclonal antibody Bcl-2 (ΔC21). Bcl-2 is detected only in the immunoprecipitation reaction using Bcl-2(100) but not in the sample immunoprecipitated with MANSMA2. The blots were stripped and reprobed with the anti-Bag-1 (fig 1A-3) and anti-Sm antibody Y12 (fig1A-4). Sm proteins and Bag-1 have been previously shown to interact with SMN and Bcl-2 respectively and act as controls for the immunoprecipitation procedure.8
21 The expected results were obtained with Bcl-2 antibodies immunoprecipitating Bag-1 and SMN antibodies precipitating Sm proteins. Similar results were obtained using Jurkat cells undergoing apoptosis after induction with phytohaemagglutinin (data not shown).
(A). Endogenous SMN and Bcl-2 are not coimmunoprecipitated from Jurkat cells. Shown is a western blot of proteins immunoprecipitated with anti-Bcl2 (Bcl-2(100)), anti-SMN (MANSMA2), and control (MANDYS-1)22 monoclonal antibodies. The western blot was serially probed with the antibodies shown below the blot as previously described.7 1 SMN protein (arrow), as detected with anti-SMN chicken polyclonal antibody C3,7is immunoprecipitated with MANSMA2 and not coimmunoprecipitated with Bcl-2. 2 Bcl-2 protein (arrow), as detected with the anti-Bcl-2 polyclonal antibody Bcl-2(ΔC21), is immunoprecipitated with the Bcl-2(100), but is not coimmunoprecipitated with SMN. The 30 kDa band present in all three lanes is non-specific and is probably the IgG light chain, as previously reported.17 3 Bag-1 protein, which is known to associate with Bcl-2, is coimmunoprecipitated only with Bcl-2. This indicates immunoprecipitation conditions were suitable to precipitate associated proteins. 4 SM proteins coprecipitate with SMN as detected with anti-SM polyclonal antibody Y12 as previously reported and are not immunoprecipitated with Bcl-2. This also shows conditions were suitable to precipitate associated proteins. (B) Endogenous SMN and Bcl-2 are not coimmunoprecipitated from non-SMA spinal cord. Shown is a western blot of spinal cord proteins immunoprecipitated with anti-Bcl-2 (Bcl-2(100)) (lane 2), anti-SMN (MANSMA2), and control (MANDYS-1) monoclonal antibodies. The western blot was serially probed with the antibodies shown below the blot. 1 SMN is detected only in the sample precipitated with MANSMA2 and is not coprecipitated with Bcl-2. 2 The reciprocal experiment shows Bcl-2 is precipitated only with Bcl-2(100) and is not coimmunoprecipitated with SMN. (C) Western blot of unprecipitated Jurkat cell lysate and immunoprecipitations probed with anti-Bcl-2 (Bcl-2(ΔC21)) and anti-SMN (SMN C3) antibodies. Methods. Approximately 1.5 × 108 cells from subconfluent cultures were lysed with 1 ml of ice cold lysis buffer.17
20 The lysate was preclarified by incubating the lysate with washed Omnisorb cells (Calbiochem) for 30 minutes on ice. The Omnisorb cells were removed by centrifugation, and the preclarified lysate was divided into 3 aliquots of 300 μl each. Approximately 1 μg of anti-SMN (MANSMA2) was added to the first aliquot, 1 μg of Bcl-2 (100) monoclonal (Santa Cruz Biotechnology, Santa Cruz, CA) was added to the second aliquot, and 1 μg of the anti-dystrophin MANDYS-1 hybridoma supernatant was added to the last aliquot. MANDYS-1 was made from the same myeloma cell line used to make MANSMA2. The antibodies were rocked on ice with the lysate for one hour to overnight. Immune complexes were captured with protein G sepharose (Sigma). Briefly, 40 μl of a 50% suspension of protein G sepharose was washed with 1 ml of lysis buffer. The protein G sepharose was pelleted and the pellet resuspended with the lysate. The lysate was rocked on ice for 30 minutes. The protein G sepharose was washed with 1 ml of lysis buffer five times for five minutes on ice. The residual wash buffer was removed by aspiration and sepharose was boiled in 30 μl of SDS buffer for two minutes. The samples were resolved by SDS-PAGE on 12% gels as described previously.7 Spinal cord collected from a type I SMA and a non-SMA subject was dissected and ground into a fine powder in liquid nitrogen using a mortar and pestle. The powder was dissolved in 1 ml of Empigen BB (1% Empigen BB, 5 mmol/l EDTA, 1 mmol/l PMSF, 5 mmol/l DTT, and 100 mmol/l Tris, pH 8, 10 μmol/l aprotinin, 1 μmol/l leupeptin, 1 μmol/l pepstatin A)20 or the Iwahashi lysis buffer.17 The insoluble material was removed by centrifugation and immunoprecipitations carried out on the supernatant as described above.
In normal spinal cord (fig 1B), SMN-C3 detects a 38 kDa band in the immunoprecipitation reaction using MANSMA2 but not in the immunoprecipitation reactions using Bcl-2(100) or MANDYS-1. Similarly, in the reciprocal experiment, the anti-Bcl-2 rabbit polyclonal antibody Bcl-2 (ΔC21) detects a 29 kDa band only in the sample immunoprecipitated with Bcl-2(100) and not in the other immunoprecipitation reactions.
As immunoprecipitation experiments failed to show a direct interaction of native SMN and Bcl-2 in proliferating and apoptotic Jurkat cells and in spinal cord, we investigated cells transfected with SMN/Bcl-2 expression constructs. We have prepared various SMN expression constructs with or without the HA epitope tag at the amino-terminus. Immunoprecipitations were performed on the transfected cells using an HA epitope monoclonal and the Bcl-2 antibodies described above. The HA antibodies resulted in immunoprecipitation of SMN and not Bcl-2 whereas Bcl-2 antibodies immunoprecipitated Bcl-2 alone (fig 2A). The transfection studies provided no evidence for interaction of SMN and Bcl-2.
(A) Immunoprecipitation of transiently transfected HA tagged SMN and Bcl-2 from COS-7 cells. COS-7 cells, transiently cotransfected with Bcl-2 and HA tagged SMN expression constructs, were immunoprecipitated using rabbit anti-Bcl-2 polyclonal (lanes 1-3) and rabbit anti-HA (lanes 5-7) polyclonal antibody as indicated above the blot. As indicated below the blot, lanes 1 and 7 were probed with anti-HA monoclonal antibody. Lanes 2 and 6 were probed with anti-SMN monoclonal antibody MANSMA2. Lanes 3 and 5 were probed with anti-Bcl-2(100) monoclonal antibody. As seen in lanes 1 and 2, HA-SMN is not coprecipitated when using the anti-Bcl-2 polyclonal antibody. In the reciprocal experiment (lane 5), Bcl-2 is not coimmunoprecipitated with HA-SMN. The middle and lower panels show western blots of the proteins remaining in the supernatants following the immunoprecipitation reactions. These are included to indicate the relative levels of Bcl-2, SMN, and HA-SMN expression in the transfected cells. Additionally, neither SMN nor HA-SMN are depleted from the supernatant incubated with anti-Bcl-2 polyclonal antibody (middle panel), but Bcl-2 is depleted (bottom panel). The reciprocal experiment shows only HA-SMN is depleted by the anti-HA monoclonal antibody (middle panel), but neither SMN nor Bcl-2 (bottom panel) is depleted. (B) Immunofluorescent detection of SMN and HA tagged SIP-1 in transiently cotransfected COS-7 cells. SMN (left panel), indirectly labelled with an FITC conjugated antibody, is located in gems, cytoplasmic particles, and diffusely in the cytoplasm. HA tagged SIP-1 (middle panel), indirectly labelled with TRITC conjugated antibodies, shows a similar distribution pattern. Merging the left and middle images (right panel) clearly shows the association of SMN and SIP1 as previously reported.9 (C) Dual labelling of HA tagged SMN and Bcl-2 detected with anti-HA polyclonal and anti-Bcl-2 monoclonal antibodies. SMN (left panel), indirectly labelled with an FITC conjugated antibody, is localised in gems, cytoplasmic particles, and faint diffuse staining (green indicated by arrows). Bcl-2 (middle panel), indirectly labelled with a TRITC conjugated secondary antibody, is expressed throughout the cytoplasm (red, diffuse fluorescence). In contrast to the SIP-1/SMN merged image above, the merged image (right panel) indicates a lack of colocalisation of SMN and Bcl-2. Methods. COS-7 cells (∼105) were seeded onto coverslips and prepared as described previously.7 Bcl-2(100) and Bcl-2(ΔC21) (Santa Cruz Biotechnology Inc) and anti-HA polyclonal antibody (BabCO) were used at 1/500 dilution. The secondary antibodies, rabbit anti-haemagglutinin antibody conjugated to TRITC (Sigma) and donkey anti-chicken F(ab)2 antibody conjugated to FITC (Jackson Immunoresearch) were used at 1/400 dilution. SIP-1 cDNA was ligated to haemagglutinin (HA) tag and subcloned into the pcDNA3 vector. Bcl-2 was subcloned into pcDNA3 vector. DNA plasmids were prepared for transfection using Endo Free Plasmid Kit (Qiagen). COS-7 cells were transfected with lipofectamine (Life Technologies) according to the manufacturer's instructions. Images were obtained using a Zeiss microscope equipped with an Optronics colour digital camera using FITC and TRITC/DAPI double pass filters. Total magnification of images is 945×.
Dual labelling of COS-7 cells cotransfected with SMN and HA tagged SIP-1 expression constructs shows SMN and SIP-1 are colocalised as previously reported (fig 2B).9 Dual labelling of SMN and Bcl-2 using cotransfection of a HA tagged SMN expression construct and a Bcl-2 expression construct failed to show convincing colocalisation of SMN and Bcl-2 in either COS-7 (fig 2C) or HeLa cells (not shown). As seen in fig 2C, in the cotransfected cell Bcl-2 is expressed abundantly throughout the cytoplasm. Similarly, overexpressed SMN is located in gems, cytoplasmic particles, and throughout the cytoplasm. Even under these conditions of extreme overexpression, there is no clear example of colocalisation, as would be expected if they associate. The apparent association of diffuse staining of SMN and Bcl-2 in the cytoplasm most likely arises from overlapping expression distributions as supported by the immunoprecipitation results.
Although we cannot exclude the possibility that SMN and Bcl-2 interact indirectly with each other in cell death pathways, these results strongly indicate that SMN and Bcl-2 do not directly interact in vivo. It is most likely that the high expression levels in the experiments of Iwahashi et al
17 resulted in artefactual aggregation of Bcl-2 and SMN and that this interaction does not exist in vivo. In order to show clearly a synergistic effect of Bcl-2 and SMN either by direct or indirect mechanisms, it is important to eliminate the possibility that transfection of SMN increases Bcl-2 levels. It appears most likely that the reduction in SMN levels results in inefficient splicing which leads to a reduction or accumulation of a critical product that causes death of the motor neurons.
This work was funded by MDA, Families of SMA, and the Preston Fund, and NIH grant NS 38650.