Mutant copper-zinc superoxide dismutase associated with amyotrophic lateral sclerosis binds to adenine/uridine-rich stability elements in the vascular endothelial growth factor 3'-untranslated region

Abstract

Vascular endothelial growth factor (VEGF) is a neurotrophic factor essential for maintenance of motor neurons. Loss of this factor produces a phenotype similar to amyotrophic lateral sclerosis (ALS). We recently showed that ALS-producing mutations of Cu/Zn-superoxide dismutase (SOD1) disrupt post-transcriptional regulation of VEGF mRNA, leading to significant loss of expression [Lu et al., J. Neurosci.27 (2007), 7929]. Mutant SOD1 was present in the ribonucleoprotein complex associated with adenine/uridine-rich elements (ARE) of the VEGF 3'-untranslated region (UTR). Here, we show by electrophoretic mobility shift assay that mutant SOD1 bound directly to the VEGF 3'-UTR with a predilection for AREs similar to the RNA stabilizer HuR. SOD1 mutants A4V and G37R showed higher affinity for the ARE than L38V or G93A. Wild-type SOD1 bound very weakly with an apparent K(d) 11- to 72-fold higher than mutant forms. Mutant SOD1 showed an additional lower shift with VEGF ARE that was accentuated in the metal-free state. A similar pattern of binding was observed with AREs of tumor necrosis factor-alpha and interleukin-8, except only a single shift predominated. Using an ELISA-based assay, we demonstrated that mutant SOD1 competes with HuR and neuronal HuC for VEGF 3'-UTR binding. To define potential RNA-binding domains, we truncated G37R, G93A and wild-type SOD1 and found that peptides from the N-terminal portion of the protein that included amino acids 32-49 could recapitulate the binding pattern of full-length protein. Thus, the strong RNA-binding affinity conferred by ALS-associated mutations of SOD1 may contribute to the post-transcriptional dysregulation of VEGF mRNA.

Fig. 1

SOD1 binds to the AU-rich elements of VEGF 3′-UTR. (a) Schematic showing region of the VEGF 3′-UTR analyzed. Probes are shown below with nucleotide limits shown in parentheses. FL, full-length and T, truncation. (b) Western blot of WT and G93A SOD1 proteins probed with an anti-SOD1 antibody. (c) EMSA with SOD1 and HuR proteins using FL probe with and without an excess of unlabeled VEGF 3′-UTR probe or control (Ctl) RNA. (d) EMSA of FL and truncated probes with G93A SOD1. EMSA experiments were repeated at least three times with similar results; -, probe alone; +, probe and protein; *, free probe.

Fig. 2

G93A SOD1 binds to other class II AREs. Upper panels show schematic of riboprobes derived from IL-8 and TNF-α 3′-UTRs. Lower panels show EMSA gels with these probes using different recombinant proteins. Assays were repeated five times with similar results. P, probe only.

Fig. 3

Ribopolymer competition for VEGF 3′-UTR binding. Recombinant G93A SOD1 and HuR were incubated with 2.5 fmol of labeled VEGF 3′-UTR in the presence of excess ribopolymers as shown. The reactions were then analyzed by native gel electrophoresis. Results are representative of three independent experiments; P, probe only; *, free probe.

Fig. 4

Other ALS-associated SOD1 mutants bind to VEGF, IL-8, and TNF-α 3′-UTRs. (a) The upper panel is a Coomassie-stained gel showing WT and mutant SOD1 proteins tagged to GST. The lower panel is a Western blot of the same proteins using an anti-SOD1 antibody. (b) EMSA analysis of mutant and WT SOD1 proteins with the VEGF 3′-UTR riboprobe; P, probe only; M+, metallated; M-, unmetallated. (c) Same as (b) except TNF-α and IL-8 riboprobes were used. (d) Ribopolymer competition of G37R SOD1 for VEGF 3′-UTR binding as described in Fig. 3. Panels (b-d) are representative of three to seven independent experiments.

Fig. 5

Comparison of binding affinities among SOD1 and HuR proteins. With limiting quantities of VEGF 3′-UTR probe (0.25 fmol), each protein was analyzed by EMSA over a range of concentrations as shown. Bound and free probe bands were quantitated by densitometry and expressed as a ratio of bound/(bound + free). Curves were calculated as described in Materials and methods. Apparent K_d are shown under each curve. Data points are the mean ± SEM of three independent experiments.

Fig. 6

Mutant SOD1 competes with HuR for binding to VEGF 3′-UTR probe. An ELISA-based RNA-binding assay was used to assess competition between SOD1 and HuR (top panel) or HuC (bottom panel) (see Materials and methods and King 2000). With HuR or HuC affixed to the ELISA well, the binding reaction was initiated by addition of a biotinylated VEGF 3′-UTR riboprobe and SOD1 protein. VEGF 3′-UTR binding was assessed by colorimetry. Results were expressed as a percentage of the value when no competitor was added. All data represent the mean ± SD of four independent measurements. HuR: ***p < 0.001; HuC: ***p < 0.001; **p < 0.01 and *p < 0.05. At 20× for HuC, ***A4V and L38V, **G37R, and *G93A.

Fig. 7

Localization of RNA-binding activity in SOD1. (a) Schematic of human SOD1 and truncated peptides used in the analysis of RNA binding (Deng et al. 1993). Numbers in parentheses refer to the amino acids included in each truncation. ▲, dimer contact sites; *, metal binding sites; S, residue involved in disulfide bridge formation. Peptides included mutant or WT sequence as indicated in parentheses. (b) Coomassie stain of peptides fused to GST. (c) EMSA analysis of peptides using VEGF, IL-8, and TNF-α 3′-UTR riboprobes as indicated. This assay was repeated three times with similar results.