SUMO-2 and PIAS1 modulate insoluble mutant huntingtin protein accumulation

Abstract

A key feature in Huntington disease (HD) is the accumulation of mutant Huntingtin (HTT) protein, which may be regulated by posttranslational modifications. Here, we define the primary sites of SUMO modification in the amino-terminal domain of HTT, show modification downstream of this domain, and demonstrate that HTT is modified by the stress-inducible SUMO-2. A systematic study of E3 SUMO ligases demonstrates that PIAS1 is an E3 SUMO ligase for both HTT SUMO-1 and SUMO-2 modification and that reduction of dPIAS in a mutant HTT Drosophila model is protective. SUMO-2 modification regulates accumulation of insoluble HTT in HeLa cells in a manner that mimics proteasome inhibition and can be modulated by overexpression and acute knockdown of PIAS1. Finally, the accumulation of SUMO-2-modified proteins in the insoluble fraction of HD postmortem striata implicates SUMO-2 modification in the age-related pathogenic accumulation of mutant HTT and other cellular proteins that occurs during HD progression.

Figure 1. K6 and K9 Are Primary Sites for HTTex1p In Vitro SUMOylation

(A) Schematic illustration of the SUMOylation pathway. SUMO is expressed as a precursor protein (SUMO-X) that is processed by SUMO-specific proteases (SENPs) to expose a C-terminal diglycine motif (-GG). Enzymatic reactions are similar to ubiquitination and include activation by the SUMO E1-activating enzyme (SAE1/UBA2), to the SUMO E2-conjugating enzyme (Ubc9), and transfer of SUMO to target lysines with or without the assistance of the SUMO E3-ligating enzymes. (B) Time course of SUMO-1 modification of His-tagged HTTex1p (25Q)-purified proteins (WT, K6R, K9R, K6,9R, K6,9,15R, and S13,16D) was performed in vitro. SUMOylation was visualized using anti-His antibody. WT HTTex1p is SUMOylated within 16 min, K6R or K9R mutations delay SUMOylation, and combined mutations (K6,9R or K6,9,15R) greatly reduce SUMOylation. Mutations that mimic phosphorylation (S13,16D) alter kinetics of SUMO-1 modification with SUMO modification observed beginning by 4 min. See also Figure S1.

Figure 2. SUMO-1 and SUMO-2 Are Differentially Expressed in HD Mice versus Control

(A) qRT-PCR analysis of SUMO-modifying proteins and enzymes in cortex and striata of WT mice at 12 weeks. Relative expression for all the SUMO enzymes is normalized to mouse β-actin. (B and C) qRT-PCR of SUMO mRNAs from 12-week-old WT and R6/2 mouse cortex (B) and striatum (C). SUMO enzyme mRNAs are differentially expressed in R6/2 versus control with statistically significant increases in SENP1 (p = 0.01), SENP3 (p = 0.02), and SUMO-1 (p = 0.003) in cortex and SENP1 (p = 0.02), SENP6 (p = 0.04), PIAS3 (p = 0.007), SUMO-1 (p = 0.02), and SUMO-2 (p = 0.02) in striatum. Samples were analyzed in quadruplicate and normalized to mouse β-actin. Data are shown as R6/2 expression relative to WT levels set at 1 for each enzyme with ± SD (n = 4). *p < 0.05. n.s., not significant. (D and E) Western blot analysis of SUMO-2 in 14-week R6/2 cortex (D) and striatum (E) versus aged-matched controls. SUMO-2 is upregulated in R6/2 striatum versus control (p = 0.026; n = 4). Protein is normalized to α-tubulin and quantitated using ImageJ. Note that only the 12-week time point is shown in (B) and (C). Please see 4- and 8-week time points in Figures S2A and S2B.

Figure 3. HTT Is Modified by Both SUMO-1 and SUMO-2

(A) HeLa cells transfected with His-tagged HTTex1p (46Q) or 46QP-K6,9,15R (3R) along with GFP-SUMO-1 or GFP-SUMO-2, lysed under denaturing conditions, and nickel purified (Ni-NTA). Unmodified HTT-46Q is indicated by the arrow and SUMO-modified HTT by the boxed region. The lysine mutant (3R) serves as a negative control. Ni-NTA represents nickel-purified His-tagged HTT, and WC TCA represents 10% of the whole-cell lysate expression of HTT and myc-actin (transfection control). HTT is modified by SUMO-1 (left) and SUMO-2 (right). (B) SUMO isopeptidases (SENP1, SENP2, SENP3, SENP5, and SENP6) modulate SUMO-modified HTT when overexpressed together with HTT (46QP-H4 or 3R) and SUMO-1 (GFP-SUMO-1). SENP1, SENP2, and SENP6 decrease HTT SUMOylation. Graph depicts quantitation of western blot using the Odyssey Infrared Imaging System (LI-COR) to calculate the ratio of HTT purified versus the HTT modified by SUMO multiplied by 100. (C) Titration of SUMO-1. Denaturing nickel purification of HTTex1p (46QP-H4) following transfection with decreasing amounts of SUMO-1 reduces the amount of SUMO-modified HTT to undetectable levels. The Ni-NTA blot displayed in the gray scale shows purified HTTex1 and SUMO-modified HTTex1 using HTT antibody. Note that 0.5 μg of SUMO-1 (¼ the amount of SUMO-1 cDNA ) was used for identifying the SUMO-1 E3 ligase for HTTex1p. Graph depicts quantitation of western blot using the Odyssey Infrared Imaging System to calculate the ratio of HTT purified versus the HTT modified by SUMO multiplied by 100. Note that all experiments were performed in triplicate, and a representative figure is shown. See also Figure S3A.

Figure 4. PIAS1 Is a SUMO E3 Ligase for HTT

(A) Under limiting SUMO conditions (1/4 SUMO-1, lanes 3–9), PIAS1 increases HTT-SUMO modification above 1/4 SUMO alone. Purified HTT (arrow) and SUMO-HTT (boxed region) were detected using anti-HTT. Graph depicts quantitation of the Ni-NTA western blot using the Odyssey Infrared Imaging Software (LI-COR) to calculate the ratio of HTT purified versus the HTT modified by SUMO multiplied by 100. (B) Western analysis of overexpression of HTTex1 (46QP-H4 or 3R), SUMO-2 (GFP-SUMO-2), and all the PIAS proteins (PIAS1, PIASxα, PIASxβ, PIAS3, and PIASy). Under nonlimiting SUMO-2 conditions, PIAS1 enhances SUMO modification of HTT. WC TCA shows overall myc-actin (transfection control) and HTT levels. Graph quantitating the ratio of HTT purified versus the HTT modified using the Odyssey (LICOR). (C) Left panel is the autoradiography results of a GST pull-down assay showing that radiolabeled human PIAS1 interacts with HTTex1p. Right panel is a phosphorimager analysis of GST pull-downs, performed in triplicate, showing the percentage of 35S-labeled PIAS protein that bound the GST proteins: GST alone, unexpanded HTT with and without the proline-rich region (20QP and 20Q), expanded HTT with and without the proline-rich region (51QP and 51Q), and the proline-rich region alone (Pro). Error bars were calculated as an estimate of SE = STDEV(n)/SQRT(n-1).

Figure 5. A 586 aa Fragment of HTT Is SUMO Modified Downstream of Exon 1

(A) SUMOplot (Abgent) predicts that HTT is SUMOylated downstream of HTTex1p lysines. SUMOplot analysis predicts two high-probability SUMO sites (red) and three low-probability SUMO sites (blue) identified downstream of HTT exon 1. Greater than 13 overlapping noncovalent SIMs are observed within this fragment (yellow). (B) Longer HTT polypeptides are modified by SUMO-1. Western blot overexpressing HTTex1p (46QP-H4 or 3R), unexpanded HTT-586 fragment (25Q-586 or 25Q-3R-586 aa), and expanded HTT-586 fragment (137Q-586 or 137Q-3R-586 aa) with SUMO-1 (GFP-SUMO-1). Cell lysates were subjected to HTT immunoprecipitation (IP) using HTT antibody. Western analysis performed with the Odyssey (LI-COR) allows detection of HTT (data not shown) and SUMO simultaneously and shows that all forms of HTT are SUMO-1 modified using the anti-GFP antibody. HTTex1p is covalently SUMO-1 modified (lane 3), and the modification disappears when Lys are mutated to Arg (3R) (lane 4). Both unexpanded (lane 4 and 5) and expanded HTT-586 fragments (lane 6 and 7) are covalently SUMO-1 modified in both the presence and absence of the three Lys in the N-terminal region of HTT (3R). Free SUMO-1 is indicated with the arrow, and SUMO-modified HTT is indicated by the boxes. Inset on the right, from a replicate experiment, shows comigration of expanded HTT (anti-HTT) and SUMO-1 (anti-GFP) displayed in the gray scale and in color when the two antibodies are merged (HTT in red, SUMO-1 in green, and yellow when colocalizing). (C) Bait plasmids (HTT-586-25Q or HTT-586-73Q aa) were transformed into the L40ccua MATa yeast strain. Yeast clones encoding bait proteins were individually mated against a matrix of MATα yeast clones encoding 16,888 prey proteins (with Gal4 activation domain fusions) using pipetting and spotting robots. Diploid yeasts were spotted onto SDIV (-Leu-Trp-Ura-His) agar plates for selection of PPIs as well as nylon membranes placed on SDIV agar plates for β-galactosidase assays. After 5–6 days of incubation at 30ºC, digitized images of the agar plates and nylon membranes were assessed for growth and β-galactosidase activity using the software Visual Grid (GPC Biotech). (D) Overexpression of PIAS1 alone, with unexpanded HTT constructs (25Q-586 aa) plus SUMO-1 (GFP-SUMO-1), shows that PIAS1 binds both full-length HTT and HTT-586 fragment (25Q-586 aa). WT HTT was used in these experiments to preclude confounding aggregation effects. Western analysis detection was performed using Odyssey and is displayed in the gray scale but is shown in color on the merge (HTT is red, and PIAS1 is green). WB, western blot. (E) HeLa cells overexpressing either expanded 586 aa-HTT or the phosphomimetic (S13,16D = DD) with SUMO-2 plus and minus PIAS1. HTT was purified by IP using hydrazide beads (Bioclone) crosslinked to HTT (Enzo) antibody and subjected to western analysis using anti-HTT (MAB5490). Arrows indicate SUMO-2-modified HTT. (F) IP of HTT with hydrazide-linked beads shows that both unexpanded and expanded HTT 586 aa phosphomimetics (S13, 16D-586 aa) are modified by SUMO-2. Arrows indicate SUMO-conjugated HTT. Note that all experiments including the Y2H assay were done in triplicate; representative experiments are shown. Arrows indicate SUMO-2-modified HTT.

Figure 6. SUMO-2 Causes Mutant HTT to Accumulate

(A) Western analysis of whole-cell lysates from HeLa cells transfected with His-SUMO-1 or SUMO-2 and/or 97Q-HTT exon 1 and treated with 5 μM MG132 for 18 hr. Lysates were separated using differential centrifugation into a detergent-soluble fraction (SOLUBLE) with 1% Triton X-100 and a detergent-insoluble fraction (INSOLUBLE) with 4% SDS. Western blot probed with anti-HTT shows full-length endogenous HTT in the SOLUBLE fraction (upper arrow) and 97Q-HTTex1 (lower arrow) (left panel). In the INSOLUBLE fraction, HTT HMW species are indicated by the bracket and asterisks (right panel). (B) MG132 and SUMO-2 cause mutant HTT to accumulate as HMW species (bracke and asterisk). Western blot showing IP with HTT antibody crosslinked beads from the detergent-insoluble fraction probed with the anti-HTT antibody. (C) Mutant HTT (97Q-Httex1) fibrils are detected with anti-HTT in the insoluble fraction with treatment of MG132 or addition of exogenous SUMO-2. (D) Western blot with increasing concentrations of SUMO-2 detected with anti-His antibody (left panel). Middle panel is the same western blot probed with anti-HTT showing soluble forms of HTT. Right panel presents western blot from detergent-insoluble fraction with monomeric HTT (97Q) at 55 kDa, and the asterisk (*) indicates the HMW species. Note that all experiments were performed in triplicate; representative figures are shown.

Figure 7. SUMO-2 Proteins Accumulate in HD Brain

(A) Western blot analysis of HeLa cells over-expressing exogenous PIAS1 in the presence of mutant HTT (97Q) when separated into detergent-soluble and detergent-insoluble fractions. No difference is detected in monomeric HTT (top panel, Soluble), but HMW HTT levels increase with PIAS1 overexpression. Anti-PIAS1 antibody (Invitrogen) was used to detect PIAS1. (B) Acute knockdown of PIAS1 decreases HMW HTT species in the detergent-insoluble fraction. PIAS1 knockdown is detected in detergent-soluble and -insoluble fractions using anti-PIAS1 antibody. (C) Drosophila melanogaster expressing mutant HTTex1p (93Q) in a reduced Su(var)2-10/dPIAS genetic background exhibits statistically significantly reduced photoreceptor neuron degeneration (left panel, p = 0.033) when comparing dPIAS/+ to +/+ flies and increased overall survival (right panel, p = 0.047) when comparing dPIAS1/+ to +/+ flies. Significance was measured by Student’s t test. (D) HMW SUMO-2 accumulates in postmortem HD striata. Western blot analysis of the insoluble fraction from three control and three HD postmortem striata as described (Experimental Procedures).