© 2000 Oxford University Press
Human Molecular Genetics, 2000, Vol. 9, No. 19 2869–2877
The exon 2b region of the spinal muscular atrophy protein, SMN, is involved in self-association and SIP1 binding Philip J. Young1, Nguyen thi Man1, Christian L. Lorson2, Thanh T. Le4, Elliot J. Androphy2,3, Arthur H.M. Burghes4 and Glenn E. Morris1,+ 1MRIC
Biochemistry Group, North East Wales Institute, Mold Road, Wrexham LL11 2AW, UK, of Dermatology and 3Department of Molecular Biology and Microbiology, New England Medical Center and Tufts University School of Medicine, Boston, MA 02111, USA and 4Department of Medical Biochemistry, Ohio State University, Columbus, OH 43210, USA 2Department
Received 3 August 2000; Revised and Accepted 27 September 2000
Spinal muscular atrophy (SMA) is caused by mutations in the SMN (survival of motor neurons) gene and there is a correlation between disease severity and levels of functional SMN protein. Studies of structure–function relationships in SMN protein may lead to a better understanding of SMA pathogenesis. Self-association of the spinal muscular atrophy protein, SMN, is important for its function in RNA splicing. Biomolecular interaction analysis core analysis now shows that SMN self-association occurs via SMN regions encoded by exons 2b and 6, that exon 2b encodes a binding site for SMN-interacting protein-1 and that interaction occurs between exon 2- and 4encoded regions within the SMN monomer. The presence of two separate self-association sites suggests a novel mechanism by which linear oligomers or closed rings might be formed from SMN monomers. INTRODUCTION Childhood spinal muscular atrophy (SMA) is an autosomal recessive disorder characterized by the loss of α motor neurons from the lower spinal cord (1). The disease-determining SMN (survival of motor neurons) gene encodes a 40 kDa protein which is ubiquitously expressed in mammalian cells. There are two inverted copies of the SMN gene on chromosome 5q13: telomeric SMN (SMN-1) and centromeric SMN (SMN-2) (2). The two genes differ by 11 nucleotides, none of which alter the coding sequence (2,3). Both genes produce a full-length transcript and an alternatively spliced SMN isoform lacking exon 5. SMN-2 produces two additional splicing products lacking exon 7 or exons 5 and 7. Whereas SMN-1 produces mainly the fulllength protein, the main SMN-2 product lacks exon 7 because of a single nucleotide difference in SMN-2 that affects splicing (3–5). In SMA, the SMN-1 gene is deleted or inactive and the variable severity of the disease correlates with production of full-length SMN from the SMN-2 gene (2,6,7). +To
The SMN protein is found both in the cytoplasm and in prominent nuclear dot-like structures termed gemini of coiled bodies, or ‘gems’, because of their close association with coiled bodies (8). Within the cytoplasm, SMN co-localizes with SMN-interacting protein-1 (SIP1 or gemin2), U1 and U5 small nuclear RNAs (snRNAs) and Sm core proteins (including B, D1-3 and E) (9). SMN also interacts directly with RNA (10,11) and fibrillarin (8). It undergoes self-oligomerization (12) and an essential role in small nuclear ribonucleoprotein (snRNP) biosynthesis has been suggested for the SMN–SIP1 complex (13), since SIP1 has sequence homology with Brr1, a yeast protein essential for snRNP synthesis and maturation (14). The SIP1 interaction site has been mapped to a conserved N-terminal sequence, SMN13–44 (9). The N-terminal sequences SMN1–91 (10) or SMN1–76 (11) bind RNA. In cells over-expressing SMN lacking amino acids 1–27, large SMN aggregates accumulate within the cytoplasm and the nucleus (15). The cytoplasmic structures contain SMN, SIP1 and snRNPs lacking the trimethylguanosine cap, leading to the suggestion that the SMN/SIP1 complex may be involved in methylation of immature snRNPs and their subsequent transport back into the nucleus (15). Inside the nucleus, SMN may be involved in recycling/regeneration of spliceosomal U snRNPs (15) and may influence pre-mRNA splicing by modulating the Sm core protein composition of U snRNPs (16). A self-oligomerization site has been mapped to a conserved C-terminal domain (SMN240–267) encoded by exon 6 (12). This site overlaps SMN258–279, a region that contains a tyrosine/ glycine-rich motif conserved in many RNA-binding proteins (17). The majority of point mutations identified in SMA patients are present within the tyrosine/glycine region (Y272C, T274I and G279V), the adjacent self-oligomerization site (S262I) or the tudor domain encoded by exon 3 (E134K), suggesting a functional importance for these sites (2,17–21). It has been suggested that SMN may act as a scaffold for snRNP assembly (22) and Sm proteins themselves form oligomeric ring structures (23,24). Removal of exon 7 (SMN280–294) also impairs self-association (12,20), interaction with Sm core
whom correspondence should be addressed. Tel: +44 1978 293 082; Fax: 44 1978 290 008; Email:
[email protected]
2870 Human Molecular Genetics, 2000, Vol. 9, No. 19
Table 1. SMN isoforms and fragments used for BIA studies Exon fragment
Amino acid nos
Vector
Molecular weight including fusion tag (kDa)
1
1–27
pGEX3x
29
1–2
1–90
pGEX3x
36
2
28–91
pGEX3x
33
2a
28–62
pGEX3x
30
2b
52–91
pGEX3x
31
3
91–158
pGEX3x
34
4
159–209
pGEX3x
32
6
242–279
pGEX3x
30
6–7
242–294
pRSETc
7
SMN iso-5
1–209, 242–294
pRSETc
34
SMN iso-7
1–279
pRSETc
36
SMN-FL
1–294
pET32c
55
SIP1
1–280
pET32c
53
proteins (20) and gem formation (25), suggesting an important role for the last 15 amino acids of SMN. In the present study, we have used real-time biomolecular interaction analysis (BIA) to characterize protein interactions within the SMN monomer, between SMN monomers in dimers/oligomers and between SMN and SIP1. We have used the same approach to characterize surface epitopes on SMN recognized by a panel of 21 monoclonal antibodies (mAbs) essential for the BIA study. Two different sites for selfassociation were identified in the SMN monomer, suggesting a possible mechanism for forming linear SMN oligomers or oligomeric ring structures. RESULTS Monoclonal antibody production and epitope mapping New panels of 16 mAbs against SMN (MANSMA6–21) and 4 against SIP1 (MANSIP1A–D) were produced by the hybridoma method and screened for recognition of SMN or SIP1 using western blots of HeLa total protein extracts. The 16 anti-SMN mAbs described here and five described previously (MANSMA1–5) (26) all recognized a 40 kDa SMN band on total protein extract of HeLa cells whereas all four anti-SIP1 mAbs recognized a 38 kDa SIP1 band (data not shown). All mAbs against SMN and SIP1 were of the IgG1 subclass and they all recognized ‘gems’ in HeLa cell nuclei (26). To map the SMN mAb panel, SMN fragments corresponding to exons 1, 2, 2a, 2b, 3, 4 and 6/7 produced as GST fusion proteins (5,10,12) and an SMN isoform lacking exon 5 were used. Table 1 shows the sizes and amino acid numbers of these fragments. mAb binding was assessed by western blotting (Fig. 1a–c) and BIAcore (Table 2; Fig. 2). Figure 2 illustrates the BIAcore method used for both epitope mapping and subsequent SMN interaction studies [the large, but transient, increases in resonance units (RU) are refractive index artefacts
that occur commonly in BIA and are mainly due, in this case, to small amounts of urea in the sample]. Each mAb was captured onto a rabbit anti-mouse Ig sensor chip and was screened with all the SMN subfragments. Table 2 shows that 18 mAbs recognized the exon 2 protein (Figs 1a and 2), 2 mAbs (MANSMA4 and 5) recognized the exon 4 protein (Fig. 1b) and MANSMA3 failed to recognize the SMN isoform (Fig. 1c), indicating exon 5 specificity. None of the mAbs recognized exon 1 or exon 3 fragments (Table 2). Figure 2 shows that the exon 2a and 2b fragments do not bind the exon 2-specific mAb MANSMA1, even though they overlap by 11 amino acids. This was found to be the case for all of the exon 2-specific mAbs (Table 2), suggesting that the exon 2 epitope is created by protein folding. In all BIAcore experiments, it is essential to perform controls for non-specific binding. The most important of these is to omit the proposed binding partner from the sequence of additions to the sensor chip. In the present case, no SMN fragment binding occurred when the binding partner, the mAb, was omitted (Table 2). This form of control was performed in all the following experiments, though it is not shown in every case. Intramolecular interactions within SMN Figure 3 shows that exon 2a and 2b proteins will bind sequentially to exon 4 protein and in doing so acquire the ability to bind the exon 2-specific mAb, MANSMA1. This suggests that exon 4 protein is acting as a scaffold on which the exon 2a and 2b fragments can fold together to reform the exon 2 epitope. Table 3 shows direct interactions of the captured exon 4 protein with exon 2 protein and vice versa. The exon 2–exon 4 interaction on the BIAcore was confirmed by immunoprecipitation, using Dynabeads with covalently attached rat antimouse IgG1 antibody. MANSMA1 mAb against exon 2 immunoprecipitated exon 4 in the presence of exon 2, but not in the absence of exon 2 (Fig. 4). Exon 4 protein in the immunoprecipitate was detected using MANSMA4 on a western blot. In theory, the exon 4–exon 2 interaction could be
Human Molecular Genetics, 2000, Vol. 9, No. 19 2871
Figure 1. Monoclonal antibody mapping by western blotting. (a) The recombinant 35 kDa SMN exon 2-encoded fragment was run as a strip on SDS– PAGE and then probed with MANSMA1–21 using a miniblotter. MANSMA3, -4 and -5 failed to identify the recombinant protein. The 18 remaining mAbs recognized the 35 kDa SMN fragment band. (b) The recombinant SMN exon 4-encoded fragment was run as a strip on SDS–PAGE and then probed with MANSMA1–21. Only MANSMA-4 and -5 recognized the 33 kDa band. (c) The recombinant 34 kDa isoform lacking exon 5 was run as a strip on SDS–PAGE. Using a miniblotter, the western blot was probed as follows: lane 1, MANSMA2; lane 2, MANSMA4; lane 3, MANSMA5; lane 4, MANSMA1; lane 5, MANSMA3. All mAbs except MANSMA3 recognize the 34 kDa band.
either intermolecular between two SMN monomers or intramolecular within the same SMN monomer. However, we shall now show that exon 4 protein does not bind to SMN monomer, supporting an intramolecular interaction. SMN self-association Although exon 2b protein will not bind to MANSMA1 mAb it will bind to exon 2 protein captured onto the sensor chip by MANSMA1 (Fig. 5; Table 3). The interaction is 2b-specific, since exon 2a, 3 or 6 proteins will not bind (Fig. 5; Table 3). This suggests a 2b–2b interaction. To determine whether such an intermolecular interaction can occur in full-length SMN, SMN was attached to a Cm5 sensor chip by amino-coupling. It was then made monomeric by passing 0.1% SDS over the sensor chip, followed by 8 M urea in 10 mM HCl. Monoclonal antibodies against exon 2 (MANSMA1–2), exon 4 (MANSMA4–5) and exon 5 (MANSMA3) all bound to the captured monomer. This showed that SMN was still present with epitopes intact (data not shown). This monomeric SMN was then exposed to SMN exons 1–2, 2, 2a, 2b, 4 or 6, as well as full-length SMN (Table 3). Only fragments containing exon 2b were captured by monomeric SMN (Table 3). This supports
the view that the 2b–2b interaction is a self-association site, whereas the exon 2–4 interaction is intramolecular. For these studies, stock solutions in 8 M urea were diluted 1:100 in HEPES-buffered saline (HBS) and incubated at room temperature for 1 h to allow renaturation and refolding before application to the sensor chip. Failure of renatured full-length SMN and the exon 6 fragment to bind monomeric SMN suggests that they are already self-associated in solution. To test this hypothesis, full-length SMN and the exon 6 fragment were incubated in 8 M urea for 1 h at 37°C to ensure complete separation of monomers. The samples were then diluted 1:100 in HBS and injected into the BIAcore within 30 s to minimize self-association. Table 3 shows that monomeric SMN and exon 6 protein produced in this way will bind to monomeric SMN on the chip. Exon 2b and 6 proteins bind to SMN monomer on the chip independently of each other and they reduce subsequent binding of SMN monomer. When both exon 2b and 6 proteins are attached to SMN monomer on the chip, SMN can no longer bind, suggesting that there are only these two selfassociation sites within SMN (Table 3). To explain why urea treatment is required for SMN and exon 6 protein, but not for exon 2b protein, we suggest that the 2b–2b interaction is weaker, enabling dissociation and exchange whereas exon 6– exon 6 interaction is essentially irreversible. Earlier reports have shown that the isoform lacking exon 7 displays a decreased ability to self-associate (12,20). This was confirmed by BIA comparison of full-length SMN and the SMN isoform lacking exon 7 in their binding to the monomeric SMN chip (Table 3). Both the full-length and isoform proteins were incubated in 8 M urea and injected over the monomeric chip immediately after dilution to 10 µg/ml. Pre-incubation of the SMN monomer on the chip with exon 2b inhibited binding of the isoform lacking exon 7 much more effectively than preincubation with exon 6 (Table 3). Consistent with the earlier reports, the absence of the last 15 amino acids encoded by exon 7 appears to reduce self-association via exon 6, leaving the isoform more dependent on the alternative exon 2b self-association site. SMN–SIP1 protein interactions SIP1 was captured onto a BIAcore sensor chip via covalently bound anti-mouse IgG and MANSIP1A and was shown to bind full-length SMN (Table 3). The experiment was also performed in reverse by capturing SMN using MANSMA1 and demonstrating subsequent SIP1 binding (Table 3). Monomeric full-length SMN bound directly to the sensor chip also interacts with SIP1 (Table 3), indicating that SMN dimerization is not essential for SIP1 binding. To map the SIP1 interaction domain further, captured SIP1 was exposed to SMN fragments encoded by exons 1–2, 2, 2a, 2b, 3, 4 and 6. Table 3 shows that SIP1 could bind exon 2b protein but not exon 2a protein. Only SMN fragments containing exon 2b sequence were captured by SIP1 (Table 3). The experiments were repeated in reverse using exon 4 to capture exon 2a and/or 2b (because capture by mAb is not possible). SIP1 was captured by exon 4/2b and exon 4/2a/2b, but not bl for SIP1 interaction (Table 3). SIP1 binding by the exon 2-encoded fragment was similar in the presence or absence of the exon 4-encoded fragment (Table 3). As a control experiment, SIP-1 on the biosensor chip was probed with exon 2b protein samples that had been pre-incubated with or without SIP-1 bound to solid
2872 Human Molecular Genetics, 2000, Vol. 9, No. 19
Table 2. Epitope mapping by BIA mAb
RU values
no.a
1
1–2
2
2a
2b
3
4
6
6–7
Iso-5
C
12
16
14
11
13
17
16
3
0
21
1
17
251
232
12
15
16
15
2
1
378
2
16
232
218
19
7
15
13
2
1
327
3
15
17
13
11
13
20
18
4
1
28
4
14
16
12
17
12
14
167
2
0
267
5
16
16
13
19
10
18
189
1
1
298
6
17
267
251
19
10
17
16
3
1
383
7
12
242
241
11
8
17
14
6
0
342
8
12
274
249
15
11
13
9
8
1
378
9
16
174
144
10
5
8
6
5
0
251
10
18
218
192
19
16
18
5
4
1
259
11
12
271
261
18
11
19
17
8
1
362
12
13
284
258
10
10
18
16
4
2
384
13
14
18
17
18
9
4
8
1
1
211
14
13
298
264
16
11
12
9
2
1
352
15
13
276
271
18
10
10
7
2
1
310
16
16
245
216
15
9
11
9
3
0
341
17
9
221
198
11
6
12
9
4
1
243
18
11
234
209
11
7
13
10
5
1
277
19
13
267
217
17
10
11
9
4
0
275
20
17
276
221
11
9
15
14
2
0
312
21
11
167
139
7
6
8
9
3
0
233
22
11
201
177
11
8
12
10
5
1
377
The anti-SMN mAbs were diluted 1:100 in HBS. SMN fragments were diluted to 1 µg/ml in HBS. The binding studies were repeated three times and the RU values shown are the mean values. aCorresponds to the MANSMA number. MANDYS1 mAb against dystrophin (C) was used as a control for non-specific binding. SMN isoform lacking exon 5 and full-length SMN are abbreviated to iso-5 and FL, respectively. Binding values >3× background were taken as significant and are shown in bold type (most are >10× background).
phase (via His-tag to Ni-NTA beads). Pre-incubation with solid phase SIP-1 to deplete exon 2b reduced the biosensor response RU by >85%, confirming an authentic interaction between exon 2b and SIP-1 (see Materials and Methods: Competitive inhibition). DISCUSSION BIA studies have identified a SIP1-binding site within SMN52–91 encoded by exon 2b (Table 3). The BIA studies also show that SMN self-association is not required for SIP1 binding. Although exon 1 (SMN1–28) and exon 2a (SMN28– 64) fragments did not bind SIP1 in the present study, previous studies (9) have shown a second binding site at the junction between these two fragments (SMN13–44). The relative contribution of these two sites to SIP1 binding remains to be determined. An interaction between regions encoded by SMN exon 2 and exon 4 has also been identified (Table 3; Figs 3 and 4). This is an interaction within the SMN monomer and is not involved in
oligomerization (Table 3). The SMN exon 4 protein also interacts with both the exon 2a and 2b proteins to reform a major antigenic determinant recognized by 18 of 21 anti-SMN mAbs (Fig. 3; Table 2). This suggests that exon 4 protein is acting as a scaffold on which the exon 2a and 2b fragments can fold together to reform the exon 2 epitope (Fig. 3). The conformational nature of this epitope is shown by lack of mAb binding to exons 2a and 2b separately (Fig. 2; Table 2). The alternative possibility that the mAbs bind to the exon 2a–2b junction is unlikely because the two SMN fragments overlap by 11 amino acids at this point (Table 1). mAb binding to exon 2 protein did not inhibit either SIP1 binding or exon 4 protein binding (Table 3), so simultaneous binding of more than one protein to this region of SMN is clearly possible. The most interesting observation to emerge from BIA studies is the identification of a novel site for self-association of SMN encoded by exon 2b. BIA also detected a previously established self-association site encoded by exon 6 (12). BIA also showed that exon 7 enhances the ability of SMN to selfassociate through the exon 6-encoded site (Table 3). This
Human Molecular Genetics, 2000, Vol. 9, No. 19 2873
Figure 2. BIA sensorgrams showing the exon 2 specificity of MANSMA1. Although MANSMA1 captures the exon 2-encoded fragment, it fails to bind the fragments encoded by exons 2a and 2b independently. A 56 kDa recombinant fragment of neuronal apoptosis inhibitory protein (NAIP) was used as a control for nonspecific binding. MANSMA1 was captured in both experiments using an immobilized rabbit anti-mouse Ig.
Figure 3. Binding of exons 2a and 2b to exon 4 re-assembles a native epitope that is recognized by MANSMA1. The exon 4 fragment was captured using MANSMA4. MANSMA4 was diluted 1:10 in HBS and captured on a Cm5 chip with covalently bound rabbit anti-mouse IgG. Purified SMN fragments encoded by exons 2a, 2b and 4 are shown in the first inset on a Coomassie Blue-stained 10% polyacrylamide gel. The control experiment (second inset) shows that MANSMA1 binding is neither to exon 2a alone nor non-specific. A 56 kDa recombinant fragment of neuronal apoptosis inhibitory protein (NAIP) was used as an additional control for non-specific binding.
agrees with other evidence for an important role for the 15 C-terminal amino acids encoded by exon 7 (12,20,25). Both sites presumably form dimers/oligomers in solution, as well as on the the sensor chip, so the exon 2b–2b interaction must be weak enough to allow exchange as the probe solution passes over the chip. The exon 6–6 interaction appears to be stronger since urea treatment to separate monomers is required immediately before passing the probe solution over the chip. The two sites are independently sufficient for self-association and blocking both sites on monomeric SMN prevents interaction with more SMN monomer (Table 3), suggesting that
these are the only two self-association sites. In the hypothetical model shown in Figure 6a, the two self-association sites are portrayed as forming a dimer. This is the simplest hypothesis conistent with our own data. However, the presence of the two sites means that more complex associations are theoretically possible, either in vivo or in vitro, so we have avoided using the terms ‘dimer’ and ‘dimerization’ in favor of ‘dimer/oligomer’ and ‘self-association’. Pellizzoni et al. (20) have shown that SMN is capable of forming larger oligomers, so SMN selfassociation may be more complex than the simple dimeric model presented in Figure 6a. For larger oligomers to form, a
2874 Human Molecular Genetics, 2000, Vol. 9, No. 19
Table 3. BIA of SMN exon and SMN/SIP1 interactions Capturing protein
C
1–2
2
2a
2b
3
4
6 (monomer)
SMN-FL (monomer) SMN∆7 (monomer) SIP1
24
–
–
7
167
13
301
–
–
Exon interations 1–2
–
131
2
21
–
–
4
140
13
295
–
–
–
127
4
15
204
190
111
156
13
–
2
12
7
20
–
–
–
–
–
–
–
–
–
–
167
Monomeric and dimeric SMN interactions SMN-FL (d) SMN-FL (m)
16
211
189
14
176
14
21
110
410
190
115
SMN-FL (m)/6(m)
14
–
–
–
137
–
–
–
117
103
108
SMN-FL (m)/2b
17
–
–
–
–
–
–
87
102
17
106
SMN-FL (m)/6(m)/2b
21
–
–
–
–
–
–
–
3
5
110
21
211
205
16
197
21
14
2
184
–
–
SMN/SIP1 interactions SIP1 4
16
204
190
109
89
13
–
2
12
–
20
4/1–2
14
–
–
–
–
11
–
2
–
–
140
4/2
19
–
–
–
–
9
–
1
–
–
143
4/2a
21
–
–
–
86
17
–
1
–
–
39
4/2b
15
–
–
103
–
10
–
3
–
–
136
4/2a/2b
17
–
–
–
–
13
–
2
–
148
SMN∆7, SMN isoform lacking exon 7; monomer, sample unfolded with 8 M urea and diluted immediately before use to minimize re-association; –, not determined. BSA was used as an additional control for non-specific binding (C). RU values are the averages of triplicate determinations and the range was ± 20% of the mean. RU binding values >3× background were taken as significant and are shown in bold type.
protein must contain at least two independent self-association sites (20). The identification of the additional exon 2b-encoded self-association site suggests a working model by which dimers linked via a strong exon 6-encoded interaction may be further associated into tetramers, hexamers by exon 2bencoded interactions (Fig. 6b). Two different Sm-binding sites have been identified on the SMN protein (9,18). The YG motif encoded by exon 6 has been shown to interact with the arginine–glycine-rich C-terminal tail of the larger core proteins (D1–3 and B). The smaller core proteins (E, F, G) that contain the conserved Sm1 and Sm2 motifs, but not the C-terminal tail, fail to interact with the exon 6-encoded site (22). In contrast, all seven subclasses of Sm core proteins bind to the tudor domains encoded by exon 3 of SMN (18). This suggests that the tudor domain interacts with either the Sm1 or the Sm2 conserved motifs. The Sm2 motif has been shown to mediate Sm core protein self-association (24). It has been suggested that the complete Sm core complex forms a ring structure, containing seven different Sm proteins, which interacts with the U snRNA through a basic central hole (23,24). It is of particular interest to note that SMN oligomers could form ring structures that would present exposed tudor domains (Fig. 6b) capable of interacting directly with each of the Sm core protein components of the proposed ring-structured Sm core complex (18,20,23,24). However, the model in Figure 6b can only produce even-numbered rings, whereas the proposed Sm core ring is seven-membered (23,24). Further studies on the composition of the Sm core complex and the functional relationship between
the exon 3- and 6-encoded Sm-binding sites are needed before a more precise model can be produced. Missense mutations causing SMA have been found in both the exons 6 (17,19,21) and 3 (18), confirming the importance of these regions for SMN function in motor neurons. A missense mutation at the SMN N-terminus (A2G) (27) also causes SMA and the model in Figure 6 may have to be modified when a function for this region of SMN is identified. In future BIA work, it would clearly be of great interest to study the effects of these missense mutations on self-association and to extend the studies to gemin 3 and 4, Sm core proteins and even nucleic acid interactions. The SMN region encoded by exon 2a binds to single stranded DNA and RNA homopolymers, with exons 2b and 3 modulating binding affinity (10,11). MATERIALS AND METHODS Cloning and expression of fusion proteins SIP1 and SMN (full-length and isoforms) cDNAs were cloned into pET32 expression plasmids for expression in Escherichia coli BL21(DE3). Transformed bacteria were induced with 1 mM isopropyl-thio-β-D-galactoside (IPTG) for 16 h at 37°C. Expressed fusion protein was purified from inclusion bodies by sequential extraction with increasing concentrations (2, 4, 6 and 8 M) of urea in phosphate-buffered saline (PBS), followed by 8 M urea with 10 mM 2-mercaptoethanol (2-ME). The extract with 8 M urea/10 mM 2-ME was further purified by gel
Human Molecular Genetics, 2000, Vol. 9, No. 19 2875
negative results due to interference by the GST moiety can never be entirely ruled out. In the present study, however, all major conclusions are based on positive binding data. Production of antibodies
Figure 4. Immunoprecipitation analysis of the interaction between exon 2- and 4-encoded proteins. Fifty microlitres of Dynabeads were washed three times with 200 µl of wash buffer (1× PBS/4% BSA) and incubated with 100 µl of MANSMA1 for 1 h at 4°C. The beads were pelleted using a magnetic particle concentrator (MPC; Dynal) and washed three times with 200 µl of wash buffer. The beads were then incubated in 200 µl of the SMN exon 2 protein (1 mg/ml in HBS) for 1 h at 4°C. The beads were pelleted with the MPC, washed three times with 200 µl of wash buffer and incubated with the SMN exon 4 protein (10 µg/ml) for 1 h at 4°C. The beads were pelleted again, washed five times with 200 µl of wash buffer and extracted in 100 µl of 2× SDS sample buffer. The extracted samples were analyzed by western blot using MANSMA4. A Dynabeads–MANSMA1–SMN exon 2 protein complex was used to immunoprecipitate SMN exon 4 protein (lane 1). SMN exon 4 protein exposed to a Dynabead–MANSMA1 complex was used a negative control (lane 2) and the exon 4 protein input is shown (lane 3). Markers are 50 kDa (IgG H chain), 32 kDa (exon 4) and 25 kDa (IgG L chain).
filtration on Sephadex G-15 (Pharmacia, St Albans, UK) to remove the 2-ME before His-tag affinity chromatography (Novagen, Madison, WI). SMN exon subfragments (Table 1) were cloned into pGEX-3X for expression as GST fusion proteins, as described previously (10,12). Although GST fusion proteins are widely used in ‘pulldown’ experiments without problems, the possibility of false
Monoclonal antibodies were produced by immunization of BALB/c mice and fusion of spleen cells with Sp2/0 myeloma cells as described elsewhere (28). Both the sera and the hybridoma culture supernatants were screened by ELISA, western blot (40 kDa band on HeLa total protein extract) and immunofluorescence microscopy (nuclear gems in COS-7 and HeLa cell lines). Hybridoma cell lines were cloned to homogeneity by limiting dilution. Ig subclass was determined using an isotyping kit (Serotec, Oxford, UK). SDS–PAGE and western blotting SDS–PAGE and western blotting were carried out as described elsewhere (29). The respective protein bands were visualized following development with a biotin–avidin detection system for mouse mAbs (Vectastain; Vector Laboratories, Burlingame, CA) and diaminobenzidine or with peroxidase-conjugated rabbit anti-mouse Ig (Dako, Ely, UK) and a chemiluminescent system (SuperSignal; Pierce, Rockford, IL). BIA The BIAcore biosensor detects changes in total mass at the surface of a sensor chip by measuring variations of the critical angle needed to produce total internal refraction. The critical angle is dependent on the refractive index of the medium that light has to pass through. An increase in mass bound to the chip increases the refractive index and this increases the critical angle. The change in critical angle is proportional to the amount of bound protein, and is expressed as resonance units (RU). The shift in RU is plotted against time and is displayed
Figure 5. The self-association site encoded by exon 2b. The exon 2 fragment was captured with MANSMA1 and probed with fragments encoded by exons 3, 6, 2a and 2b. Only the fragment encoded by exon 2b interacted with captured exon 2 fragment. MANSMA1 was diluted 1:10 in HBS and captured on a Cm5 chip with covalently bound rabbit-anti-mouse IgG.
2876 Human Molecular Genetics, 2000, Vol. 9, No. 19
Figure 6. Schematic model of SMN as (a) a dimer and (b) an oligomer. (a) Both the exon 2b- and exon 6-encoded self-association sites are involved in dimerization in this model. (b) In the proposed oligomer model, the exon 6-encoded self-association site would produce initial dimerization and the exon 2b-encoded site would allow further oligomerization. All structural interactions between SMN fragments identified by BIA are shown as bars.
as a sensorgram. A research grade Cm5 sensor chip with rabbit anti-mouse IgG covalently attached by amino-coupling was prepared according to the manufacturer’s instructions (BIAcore AB, Stevenage, UK). A BIAcore-X apparatus was used with an operating flow rate of 5 µl/min. All injected volumes were 10 µl. Concentrations of stock recombinant proteins were determined by Coomassie Brilliant Blue staining after SDS–PAGE. Stocks were diluted in HBS to 10 µg/ml unless otherwise stated. This reduced the urea concentration, allowing protein renaturation. SMN was also bound directly to a Cm5 chip. Stock SMN solutions in 8 M urea/10 mM 2-ME were subjected to gel filtration on a calibrated Sephadex G-15 column in coupling buffer to remove all traces of urea and 2-ME in preparation for amino coupling to the chip. SMN eluted in the excluded volume was coupled immediately to the chip. After coupling, the chip was pulsed with HBS containing 0.1% SDS followed by 10 mM HCl/8 M urea to remove all non-covalently bound protein and leave monomeric SMN on the chip. Successful coupling was confirmed by binding of previously described SMN-specific mAbs, MANSMA1–3 (26). Controls for non-specific binding were always performed on the same sensor chip, rather than a parallel chip or flow-cell, since deterioration of a chip may cause increased non-specific binding. Immunoprecipitation Magnetic Dynabeads with covalently attached rat anti-mouse Ig antibodies were used according to the manufacturer’s instructions (Dynal, Oslo, Norway). Immunocomplexes obtained with MANSMA1 (exon 2) were removed from the beads with SDS–PAGE sample buffer and analyzed by western blotting with MANSMA4 (exon 4).
Competitive inhibition SIP-1 was attached to Ni-NTA beads (Qiagen, Crawley, UK) via its His-tag. The beads were incubated for 1 h at 4°C with 200 µl of exon 2b protein (10 µg/ml in HBS). The supernatant was compared on the BIAcore with exon 2b incubated with control beads (no SIP1 attached). SIP1 was captured onto the sensor chip via rabbit anti-mouse Ig and an mAb against its thioredoxin-tag. ACKNOWLEDGEMENTS This work was supported by grants from the Muscular Dystrophy Association (USA) and Families of SMA (USA). REFERENCES 1. Munsat, T.L. and Davies, K.E. (1992) Workshop report: international SMA collaboration. Neuromusc. Disord., 2, 423–428. 2. Lefebvre, S., Burglen, L., Reboullet, S., Clermont, O., Burlet, P., Viollet, L., Benichou, B., Cruaud, C., Millasseau, P., Zevianni, M. et al. (1995) Identification and characterization of a spinal muscular atrophy-determining gene. Cell, 80, 155–165. 3. Monani, U.R., Lorson, C.L., Parsons, D.W., Prior, T.W., Androphy, E.J., Burghes, A.H.M and McPherson, J.D. (1999) A single nucleotide difference that alters splicing patterns distinguishes the SMA gene SMN1 from the copy gene SMN2. Hum. Mol. Genet., 8, 1177–1183. 4. Gennarelli, M., Lacarelli, M., Capon, F., Pizzuti, A., Merlini, L., Angelini, C., Novelli, G. and Dallapiccola, J. (1995) Survival of motor neuron gene transcripts analysis in muscles from spinal muscular atrophy. Biochem. Biophys. Res. Commun., 213, 342–348. 5. Lorson, C.L., Hahnen, E., Androphy, E.J. and Wirth, B. (1999) A single nucleotide in the SMN gene regulates splicing and is responsible for spinal muscular atrophy. Proc. Natl Acad. Sci. USA, 96, 6307–6311. 6. Monani, U., Sendtner, M., Coovert, D., Parsons, D., Andreassi, C., Le, T.T., Jablonka, S., Schrank, B., Rossol, W., Prior, T.W. et al. (2000) The human centomeric survival of motor neuron gene (SMN2) rescues embryonic lethality in SMN –/– mice and results in a mouse with spinal muscular atrophy. Hum. Mol. Genet., 9, 333–339. 7. Jablonka, S., Schrank, B., Kralewski, M., Rossoll, W. and Sendtner, M. (2000) Reduced survival of motor neuron (SMN) gene dose in mice leads
Human Molecular Genetics, 2000, Vol. 9, No. 19 2877
8. 9. 10.
11.
12.
13. 14. 15.
16.
17.
18.
19.
to motor neuron degradation: an animal model for spinal muscular atrophy type III. Hum. Mol. Genet., 9, 341–346. Liu, Q. and Dreyfuss, G. (1996) A novel nuclear structure containing the survival of motor neuron protein. EMBO J., 15, 3555–3565. Liu, Q., Fischer, U., Wang, F. and Dreyfuss, G. (1997) The spinal muscular atrophy disease gene product, SMN, and its associated protein SIP1 are in complex with spliceosomal snRNP proteins. Cell, 90, 1013–1021. Lorson, C.L. and Androphy, E.J. (1998) The domain encoded by exon 2 of the survival of motor neuron protein mediates nucleic acid binding. Hum. Mol. Genet., 7, 1269–1275. Bertrandy, S., Burlet, P., Clermont, O., Huber, C., Fondrat, C., ThierryMieg, D., Munnich, A. and Lefebvre, S. (1999) The RNA binding properties of SMN: deletion analysis of the zebrafish orthologue defines domains conserved in evolution. Hum. Mol. Genet., 8, 775–782. Lorson, C.L., Strasswimmer, J., Jun-Mei, Y., Baleja, J.D., Hahnen, E., Wirth, B., Le, T.T., Burghes, A.H.M. and Androphy, E.J. (1998) SMN oligomerization defects correlates with spinal muscular atrophy severity. Nature Genet., 19, 63–66. Fischer, U., Liu, Q. and Dreyfuss, D. (1997) The SMN-SIP1 complex has an essential role in spliceosomal snRNP biosynthesis. Cell, 90, 1023–1029. Nobel, S.M. and Guthrie, C. (1996) Identification of novel genes required for yeast pre-mRNA splicing by means of cold-sensitive mutations. Genetics, 143, 67–80. Pellizzoni, L., Kataoka, N., Charroux, B. and Dreyfuss, G. (1998) A novel function for SMN, the spinal muscular atrophy disease gene product, in pre-mRNA splicing. Cell, 95, 615–624. Meister, G., Buhler, D., Laggernauer, B., Zobawa, M., Lottspeich, F. and Fischer, U. (2000) Characterization of a nuclear 20S complex containing the survival of motor neurons (SMN) protein and a specific subset of spliceosomal Sm proteins. Hum. Mol. Genet., 9, 1977–1986. Talbot, K., Ponting, C.P., Theodosiou, A.M., Rodrigues, N.R., Surtees, R., Mountford, R. and Davies, K.E. (1997) Missense mutation clustering in the survival of motor neuron gene: a role for a conserved tyrosine and glycine rich region of the protein in RNA metabolism? Hum. Mol. Genet., 6, 497–500. Buhler, D., Raker, V., Luhrmann, R. and Fischer, U. (1999) Essential role for the tudor domain of SMN in spliceosomal U snRNP assembly: implications for spinal muscular atrophy. Hum. Mol. Genet., 8, 2351–2357. Clermont, O., Burlet, P., Cruad, C., Bertrandy, S., Melki, J., Munnich, A. and Lefebvre, S. (1997) Mutation analysis of the SMN gene in undeleted SMA patients. Am. J. Hum. Genet., 61, A329.
20. Pellizzoni, L., Charroux, B. and Dreyfuss, G. (1999) SMN mutants of spinal muscular atrophy patients are defective in binding to snRNP proteins. Proc. Natl Acad. Sci. USA, 96, 11167–11172. 21. Hahnen, E., Schonling, J., Rudnik-Schoneborn, S., Raschke, H., Zerres, K. and Wirth, B. (1997) Missense mutations in exon 6 of the survival of motor neuron gene in patients with spinal muscular atrophy (SMA). Hum. Mol. Genet., 6, 821–825. 22. Friesen,W.J. and Dreyfuss, G. (2000) Specific sequences of the Sm and Sm-like (Lsm) proteins mediate their interaction with the spinal muscular atrophy disease gene product (SMN). J. Biol. Chem., 275, 26370–26375. 23. Plessel, G., Luhrmann, R. and Kastner, B. (1997) Electron microscopy of assembly intermediates of the snRNP core: morphological similarities between the RNA-free (E.F.G) protein heteromer and the intact snRNP core. J. Mol. Biol., 265, 87–94. 24. Kambach, C., Walke, S., Young, R., Avis, J.M., de la Fortelle, E., Raker, V.A., Luhrmann, R., Li, J. and Nagai, K. (1999) Crystal structures of two Sm protein complexes and their implications for the assembly of the spliceosomal snRNPs. Cell, 96, 375–387. 25. Frugier, T., Tiziano, F.D., Cifuentes-Diaz, C., Miniou, P., Roblot, N., Dierich, A., Le Meur, M. and Melki, J. (2000) Nuclear targeting defect of SMN lacking the C-terminus in a mouse model of spinal muscular atrophy. Hum. Mol. Genet., 9, 849–858. 26. Young, P.J., Le, T.T., Nguyen thi Man, Burghes, A.H.M. and Morris, G.E. (2000) The relationship between SMN, the spinal muscular atrophy protein, and nuclear coiled bodies in differentiated tissues and cultured cell lines. Exp. Cell Res., 256, 365–374. 27. Wirth, B. (2000) An update of the mutation spectrum of the survival of motor neuron (SMN1) gene in autosomal recessive spinal muscular atrophy (SMA). Hum. Mutat., 15, 228–257. 28. Nguyen thi Man and Morris, G.E. (1996) Production of panels of monoclonal antibodies by the hybridoma method. In Morris, G.E. (ed.), Epitope Mapping Protocols. Humana Press, Totowa, NJ. pp. 377–389. 29. Nguyen thi Man, Ellis, J.M., Love, D.R., Davies, K.E., Gatter, K.C., Dickson, G. and Morris, G.E. (1991) Localization of the DMDL gene encoded by dystrophin-related protein using a panel of nineteen monoclonal antibodies: presence of neuromuscular junctions in the sarcolemma of dystrophic skeletal muscle, in vascular and other smooth muscle and in proliferating brain cell lines. J. Cell Biol., 115, 1695–1700.
2878 Human Molecular Genetics, 2000, Vol. 9, No. 19