Studies of the role of ubiquitination in the interaction of ubiquilin with the loop and carboxyl terminal regions of presenilin-2

Abstract

Ubiquilin was originally identified as a presenilin-interacting protein. We previously reported that ubiquilin interacts with both the loop and carboxyl terminus of presenilin proteins and that the ubiquitin-associated (UBA) domain of ubiquilin, which binds poly ubiquitin chains, is important for mediating this interaction. In the present study, we examined whether ubiquitination of presenilin-2 (PS2) is required for interaction with ubiquilin-1 by mutating lysine residues that may be targets for ubiquitination in the presenilin loop and carboxyl terminus regions. Mutation of two lysine residues in the PS2-loop region suggested that ubiquitination is not required for interaction with ubiquilin-1 and may, in fact, even negatively regulate the interaction. Similarly, we found that ubiquitination of the PS2 carboxyl terminus (PS2-C-terminus) is not required for interaction with ubiquilin-1, although our results suggest that it could play some role. Instead, we found that the mutation of either one of the two lysine residues in the carboxyl terminus of PS2 or the proline residues in the highly conserved PALP motif in this region results in destabilization of the mutant PS2 polypeptides because of increased degradation by the proteasome. Furthermore, by GST-pull-down assays we found that the mutant polypeptides were unable to bind ubiquilin, suggesting that loss of ubiquilin interaction leads to destabilization of presenilin polypeptides. Paradoxically, however, knockdown of ubiquilin expression by RNA interference did not alter the rate of turnover of PS2 proteins in cells. Instead, we found that PS2 synthesis was reduced, and PS2 fragment production was increased, suggesting that ubiquilin expression modulates biogenesis and endoproteolysis of presenilin proteins.

Figure 1. Ubiquilin interaction with PS2-loop and PS2-C-terminus proteins

β̃galactosidase yeast two-hybrid liquid culture interaction of ubiquilin-1 as prey and PS2-loop (A) or PS2-C constructs (B) as bait. The PS2-loop constructs used were PS2-loop wt, PS2-loop K271R, PS2-loop K306R, or PS2-loop K271RK306R. The PS2-C constructs used were PS2-C wt, PS2-C K410R, PS2-C K411R, or PS2-C K410RK411R. The β̃galactosidase units were normalized to pEG202, the empty bait vector. Each assay was performed in triplicate. Error bars indicate standard deviation and “**” indicates p<0.01.

Figure 1. Ubiquilin interaction with PS2-loop and PS2-C-terminus proteins

β̃galactosidase yeast two-hybrid liquid culture interaction of ubiquilin-1 as prey and PS2-loop (A) or PS2-C constructs (B) as bait. The PS2-loop constructs used were PS2-loop wt, PS2-loop K271R, PS2-loop K306R, or PS2-loop K271RK306R. The PS2-C constructs used were PS2-C wt, PS2-C K410R, PS2-C K411R, or PS2-C K410RK411R. The β̃galactosidase units were normalized to pEG202, the empty bait vector. Each assay was performed in triplicate. Error bars indicate standard deviation and “**” indicates p<0.01.

Figure 2. Mutation of lysine residues in the PS2-loop and PS2-C-terminus reduces ubiquitination of the proteins in yeast

LexA-PS2-loop (A) and LexA-PS2-C-terminus (B) fusion proteins were immunoprecipitated from certain of the yeast cultures described in Figure 1 using a anti-LexA antibody. An equivalent fraction of the immunoprecipitates and a 10% fraction of total yeast lysates were separated by SDS-PAGE, transferred to nitrocellulose membranes, after which they were probed with anti-ubiquitin (upper blot) or anti-LexA antibodies (lower blot). Note that the anti-LexA blots confirmed that the LexA-PS2-loop and LexA-PS2-C fusion proteins were indeed immunoprecipitated in all the reactions. Asterisk indicates cross-reaction of the IgG heavy chain used in the immunoprecipitations with the anti-ubiquitin antibody.

Figure 3. Mutations of proline residues in the PALP motif of the PS2 C-terminus disrupts interaction of ubiquilin with the PS2-C-terminus

β̃galactosidase yeast two-hybrid liquid culture interaction of ubiquilin-1 with PS2-C- terminus constructs containing mutations in the PALP motif. Yeast were co-transformed with ubiquilin-1 as prey and PS2-C-terminus wt, PS2-C P414L or PS2-C P417S as bait. β-galactosidase liquid assays were used to test for interaction. All experimental values were normalized to yeast co-transformed with ubiquilin-1 prey and pEG202. Error bars indicate standard deviation, and “**” indicates p<0.01. Each assay was performed in triplicate. Mutation of the first and fourth proline within the highly conserved PALP motif leads to a 95% and 80% decrease, respectively, in ubiquilin-1 interaction with the PS2-C-terminus.

Figure 4. Mutation of the PS2-C-terminus leads to destablilization of the protein in yeast

A, Equivalent amounts of protein lysate isolated from yeast transformed with certain of the PS2-loop or PS2-C-terminus baits used in the β-galactosidase liquid assays described in Figures 1 and 3 were immunoblotted with either a anti-LexA antibody (upper panels) or with an anti-actin antibody (lower panel). Note that all of the mutations in the PS2-C-terminus lead to destablization of the expressed protein in yeast. B, Equivalent amounts of protein lysate from yeast transformed with the PS2-C-terminus wt or -P414L mutant construct and treated with MG132 in DMSO or mock-treated with DMSO alone were immunoblotted with either a anti-LexA antibody (upper panels) or with an anti-actin antibody (lower panel). Note that proteasome inhibition by MG132 leads to increased accumulation of the PS2-terminus proteins. Similar results were observed for LexA-PS2-C-terminus-K410R, -K411R, and -P417S mutants (data not shown).

Figure 5. Evidence that mutation of the PS2 C-terminus leads to disruption of interaction with ubiquilin in vitro

PS2-C constructs were in vitro transcribed and translated in the presence of ³⁵S-methionine using cDNA templates encoding LexA-PS2-C-teminus -wt, -P414L, -P417S, -K410R, -K411R, or -K410RK411R proteins. A portion of each of the ³⁵S-labeled reactions was incubated with 2.5μg of purified recombinant ubiquilin-1-GST (GST-Ubqln-1) protein in vitro, after which the ubiquilin-1-GST protein was affinity purified from the mixtures using glutathione-agarose beads. The beads from these pulldown assays were resuspended in protein sample buffer and association of the different ³⁵S-labeled PS2-C-terminus proteins with GST-ubiquilin-1 was analyzed by Coomassie blue staining and autoradiography after SDS-PAGE. A, Coomassie blue (lower gel) and GST-ubiquilin-1 pulldown reactions (upper gel) with LexA-PS2-C PALP mutants. B, GST-ubiquilin-1 pulldown reactions with LexA-PS2-C K→R mutants. C, quantification of the ³⁵S-labeled PS-2 proteins that were pulled down in the experiments shown in A and B. Note that all of the PS2-C-terminus mutations lead to reduced interaction with GST-ubiquilin-1.

Figure 5. Evidence that mutation of the PS2 C-terminus leads to disruption of interaction with ubiquilin in vitro

PS2-C constructs were in vitro transcribed and translated in the presence of ³⁵S-methionine using cDNA templates encoding LexA-PS2-C-teminus -wt, -P414L, -P417S, -K410R, -K411R, or -K410RK411R proteins. A portion of each of the ³⁵S-labeled reactions was incubated with 2.5μg of purified recombinant ubiquilin-1-GST (GST-Ubqln-1) protein in vitro, after which the ubiquilin-1-GST protein was affinity purified from the mixtures using glutathione-agarose beads. The beads from these pulldown assays were resuspended in protein sample buffer and association of the different ³⁵S-labeled PS2-C-terminus proteins with GST-ubiquilin-1 was analyzed by Coomassie blue staining and autoradiography after SDS-PAGE. A, Coomassie blue (lower gel) and GST-ubiquilin-1 pulldown reactions (upper gel) with LexA-PS2-C PALP mutants. B, GST-ubiquilin-1 pulldown reactions with LexA-PS2-C K→R mutants. C, quantification of the ³⁵S-labeled PS-2 proteins that were pulled down in the experiments shown in A and B. Note that all of the PS2-C-terminus mutations lead to reduced interaction with GST-ubiquilin-1.

Figure 6. Reduction of ubiquilin expression by RNA interference is associated with a reduction in steady-state accumulation of full-length PS2 protein without affecting the turnover of the protein

PS2-inducible cultures were transfected with SMARTpool siRNAs directed against ubiquilin-1, ubiquilin-1 and ubiquilin-2, or nonsense siRNA. 20h following transfection, the cultures were treated with PonasteroneA to induce PS2 expression. After an additional 24 h the cultures were treated with cycloheximide to inhibit new protein synthesis, and lysates were collected at 2h intervals, as indicated in the figure. A, Western blots showing immunoreactivity with anti-PS2 N-terminus 096 (upper panel), anti-ubiquilin (middle panel), and anti-actin (lower panel) antibodies. B, Quantification of PS2 protein levels of the experiment shown in A, after normalization for actin loading. C, Western blot showing expression of the PS2-NTF using the anti-PS2 (093) antibody in a experiment similar to that described in A. Note that in cells transfected with ubiquilin 1 and 2 siRNA, PS2 NTFs accumulate to higher levels than in cells transfected with nonsense siRNA.

Figure 6. Reduction of ubiquilin expression by RNA interference is associated with a reduction in steady-state accumulation of full-length PS2 protein without affecting the turnover of the protein

PS2-inducible cultures were transfected with SMARTpool siRNAs directed against ubiquilin-1, ubiquilin-1 and ubiquilin-2, or nonsense siRNA. 20h following transfection, the cultures were treated with PonasteroneA to induce PS2 expression. After an additional 24 h the cultures were treated with cycloheximide to inhibit new protein synthesis, and lysates were collected at 2h intervals, as indicated in the figure. A, Western blots showing immunoreactivity with anti-PS2 N-terminus 096 (upper panel), anti-ubiquilin (middle panel), and anti-actin (lower panel) antibodies. B, Quantification of PS2 protein levels of the experiment shown in A, after normalization for actin loading. C, Western blot showing expression of the PS2-NTF using the anti-PS2 (093) antibody in a experiment similar to that described in A. Note that in cells transfected with ubiquilin 1 and 2 siRNA, PS2 NTFs accumulate to higher levels than in cells transfected with nonsense siRNA.

Figure 6. Reduction of ubiquilin expression by RNA interference is associated with a reduction in steady-state accumulation of full-length PS2 protein without affecting the turnover of the protein

PS2-inducible cultures were transfected with SMARTpool siRNAs directed against ubiquilin-1, ubiquilin-1 and ubiquilin-2, or nonsense siRNA. 20h following transfection, the cultures were treated with PonasteroneA to induce PS2 expression. After an additional 24 h the cultures were treated with cycloheximide to inhibit new protein synthesis, and lysates were collected at 2h intervals, as indicated in the figure. A, Western blots showing immunoreactivity with anti-PS2 N-terminus 096 (upper panel), anti-ubiquilin (middle panel), and anti-actin (lower panel) antibodies. B, Quantification of PS2 protein levels of the experiment shown in A, after normalization for actin loading. C, Western blot showing expression of the PS2-NTF using the anti-PS2 (093) antibody in a experiment similar to that described in A. Note that in cells transfected with ubiquilin 1 and 2 siRNA, PS2 NTFs accumulate to higher levels than in cells transfected with nonsense siRNA.

Figure 7. Reduction of ubiquilin expression by RNA interference leads to decreased synthesis of PS2 proteins but does not alter the turnover of PS2 proteins in neuronal SH-SY5Y cells

SH-SY5Y cell cultures were transfected with SMARTpool siRNAs directed against ubiquilin-1, ubiquilin-1 and ubiquilin-2, or nonsense siRNA. 48h following transfection, the cultures were metabolically labeled with ³⁵S-methionine for 1h. After the labeling period, the cultures were washed and incubated in non-radioactive medium and protein lysates were collected at 2 h intervals as indicated in the figure. A, Autoradiogram showing full-length PS2. B, Quantification of PS2 bands in A. Note that ubiquilin knockdown results in decreased synthesis of PS2 but has little affect on the turnover of the protein.

Figure 7. Reduction of ubiquilin expression by RNA interference leads to decreased synthesis of PS2 proteins but does not alter the turnover of PS2 proteins in neuronal SH-SY5Y cells

SH-SY5Y cell cultures were transfected with SMARTpool siRNAs directed against ubiquilin-1, ubiquilin-1 and ubiquilin-2, or nonsense siRNA. 48h following transfection, the cultures were metabolically labeled with ³⁵S-methionine for 1h. After the labeling period, the cultures were washed and incubated in non-radioactive medium and protein lysates were collected at 2 h intervals as indicated in the figure. A, Autoradiogram showing full-length PS2. B, Quantification of PS2 bands in A. Note that ubiquilin knockdown results in decreased synthesis of PS2 but has little affect on the turnover of the protein.