Proteomic analysis of wild-type and mutant huntingtin-associated proteins in mouse brains identifies unique interactions and involvement in protein synthesis

Abstract

Huntington disease is a neurodegenerative disorder caused by a CAG repeat amplification in the gene huntingtin (HTT) that is reflected by a polyglutamine expansion in the Htt protein. Nearly 20 years of research have uncovered roles for Htt in a wide range of cellular processes, and many of these discoveries stemmed from the identification of Htt-interacting proteins. However, no study has employed an impartial and comprehensive strategy to identify proteins that differentially associate with full-length wild-type and mutant Htt in brain tissue, the most relevant sample source to the disease condition. We analyzed Htt affinity-purified complexes from wild-type and HTT mutant juvenile mouse brain from two different biochemical fractions by tandem mass spectrometry. We compared variations in protein spectral counts relative to Htt to identify those proteins that are the most significantly contrasted between wild-type and mutant Htt purifications. Previously unreported Htt interactions with Myo5a, Prkra (PACT), Gnb2l1 (RACK1), Rps6, and Syt2 were confirmed by Western blot analysis. Gene Ontology analysis of these and other Htt-associated proteins revealed a statistically significant enrichment for proteins involved in translation among other categories. Furthermore, Htt co-sedimentation with polysomes in cytoplasmic mouse brain extracts is dependent upon the presence of intact ribosomes. Finally, wild-type or mutant Htt overexpression inhibits cap-dependent translation of a reporter mRNA in an in vitro system. Cumulatively, these data support a new role for Htt in translation and provide impetus for further study into the link between protein synthesis and Huntington disease pathogenesis.

FIGURE 1.

Strategy to identify wild-type (7Q) and mutant (140Q) Htt-interacting proteins from mouse brain. a, a flow chart shows centrifugation-based subcellular fractionation of mouse brain lysates. The red-boxed fractions were used for FLAG purifications. Western blot analyses below show the distribution of a nuclear protein (YY1), ribosomal protein (Rps6), cytoskeletal protein (β-tubulin), and Htt. The numbers indicate the amount of each fraction blotted. b, shown is a representative Western blot (WB) and silver-stained gels of 7Q and 140Q S2 (cytoplasmic) and P2 (membranous) purifications. The arrowheads and arrows indicate 7Q and 140Q Htt protein isoforms, respectively. IP, immunoprecipitates. c, the bar graph shows the number of spectral counts for Htt observed in each trial, for each genotype, under each condition. d, shown is a flow chart depicting the data analysis pipeline of the 24 total FLAG affinity purifications. This paradigm was followed for both S2 and P2 purifications. e, shown is a graph illustrating the relationship between Pearson correlation values (r) and co-variation of candidate Htt-interacting proteins. Proteins with positive r values tend to co-vary in abundance with respect to Htt in each trial. f, the graph shows the distribution of r values (y axis) and the log₂ enrichment of average spectral counts (x axis, 7Q/control) of reproducibly identified control proteins (black balls) and 7Q proteins (green balls). The pink-shaded area highlights the proteins considered to be candidate 7Q-interacting proteins.

FIGURE 2.

Functional categories of 7Q and 140Q interacting proteins. a and b, shown is a Gene Ontology analysis of S2 (a) and P2 (b) 7Q- and 140Q-associated proteins. The x axis labels individual molecular function or pathway categories as annotated by the PANTHER data base. The left y axis marks the enrichment (observed/expected), and right y axis marks the −log₁₀ of the p value. Enrichment values are presented as bar graphs (7Q = white, 140Q = gray), and p values are presented as balls. ns, not significant (p > 0.05). c, a schematic shows different types of shared interactions between 7Q and 140Q. Balls represent individual protein categories (black = S2 identification, white = P2 identification). The red boxes group interaction types, which are labeled to the left of the schematic and given letter notations (A, B, C, D, and E within protein balls). d, the bar graph illustrates how 7Q and 140Q candidate interacting proteins were distributed between 7Q and 140Q purifications with letter notations summarizing shared interaction distributions as diagrammed in c. e, procedures were the same as in a and b except for the proteins grouped as D or E. N-Ach-R, nicotinic acetylcholine receptor.

FIGURE 3.

Comparison and confirmation of 7Q- and 140Q-associated proteins. a and b, volcano plots compare the log₂-fold difference (x axis) and −Log₁₀ p values (y axis) of commonly observed S2 (a) and P2 (b) proteins. The different shaded balls signify in which genotype a protein passed all the stringency criteria. In b, Myo5a was a low confidence candidate interacting protein in both 7Q and 140Q purifications (positive r value; -fold change <1.5 over control). Arrows highlight the identities of specific proteins. The dotted lines show regions of particular significance (raw -fold difference >1.5, p value <0.05). c and d, Western blot analysis for the indicated proteins from S2 (c) and P2 (d) FLAG purifications is shown. e, Western blot analysis of GFP and Htt in GFP immunoprecipitations from Neuro-2a cells transfected with the indicated plasmids is shown. f, Western blot analysis of Htt and FLAG in IgG or Htt immunoprecipitations (IP) from Neuro-2a cells transfected with FLAG-RbFox1 is shown. Equal amounts of protein lysate were used in each condition. g, Western blot analysis of IgG and Htt immunoprecipitations from non-FLAG-tagged wild-type (+/+) and homozygous mutant (140Q/140Q) whole mouse brain lysates probed with antibodies to Htt, Myo5a, and Rps6 is shown. Both wild-type and mutant Htt specifically co-immunoprecipitated with Myo5a and Rps6. Substantial amounts of Htt remained in extracts after immunoprecipitation with anti-Htt antibodies (unbound lanes).

FIGURE 4.

Htt associates with ribosomes. a, shown are polysome profiles of cytoplasmic wild-type (7Q/7Q) and mutant (140Q/+) mouse brain extracts with (black traces) or without (red traces) EDTA pretreatment. The x axis shows the position in the gradient (left = top), and the y axis shows the 254-nm absorbance. The numbers below the traces indicate sample fraction in Western blot analyses shown below. The colors of the boxes surrounding the Western blot images correspond to the traces (red = untreated, black = EDTA treatment). b, fluorescent micrographs show the localization of Htt (green) with respect to the stress granule marker protein Tia1 (red) in wild-type (STHdh^7Q/7Q) and mutant (STHdh^111Q111Q) arsenite-stressed striatal precursor cells. Cells were treated with 0.5 mm arsenite in complete media for 1 h at 33 °C before fixation. c, overexpression of full-length wild-type (Htt17Q) or mutant Htt (Htt75Q) in HeLa cells inhibits the translation capacity of extracts prepared from these cells. Extracts derived from untransfected (Control) and GFP-, Htt17Q-, and Htt75Q-transfected cells were programmed with in vitro transcribed 5′m7G-capped and polyadenylated firefly luciferase mRNA, and translation was assessed by luminescence after 1 h at 37 °C (unmixed extracts, black bars with S.D.). Translational activity could be partially restored by mixing transfected cell extracts 1:1 with control extract (1:1 mixed extract, gray bars with S.D.). The control 1:1 mixed extract was diluted 1:1 with lysis buffer alone. The Western blots confirm overexpression of the indicated proteins and equal protein content in extracts. The lack of phospho-Eif2α signal in Htt17Q and Htt75Q extracts indicates that translation was not inhibited as part of a general stress response mechanism. Arsenite-treated (Tx) HeLa lysate serves as a positive control for the antibody.