Nascent mutant Huntingtin exon 1 chains do not stall on ribosomes during translation but aggregates do recruit machinery involved in ribosome quality control

Mutations that cause Huntington’s Disease involve a polyglutamine (polyQ) sequence expansion beyond 35 repeats in exon 1 of Huntingtin. Intracellular inclusion bodies of mutant Huntingtin protein are a key feature of Huntington’s disease brain pathology. We previously showed that in cell culture the formation of inclusions involved the assembly of disordered structures of mHtt exon 1 fragments (Httex1) and they were enriched with translational machinery when first formed. We hypothesized that nascent mutant Httex1 chains co-aggregate during translation by phase separation into liquid-like disordered aggregates and then convert to more rigid, amyloid structures. Here we further examined the mechanisms of inclusion assembly in a human epithelial kidney (AD293) cell culture model and examined whether ribosome quality control machinery previously implicated in stalled ribosomes were involved. We found mHttex1 did not appear to stall translation of its own nascent chain and there was no recruitment of RNA into inclusions. However, proteins involved in translation or ribosome quality control were co-recruited into the inclusions (Ltn1 and Rack1) compared to a protein not anticipated to be involved (NACAD). Furthermore, we observed co-aggregation with other proteins previously identified in inclusions, including Upf-1 and chaperone-like proteins Sgta and Hspb1, which also suppressed aggregation at high co-expression levels. The newly formed inclusions contained immobile mHttex1 molecules which points to the disordered aggregates being mechanically rigid prior to amyloid formation.


36
The transgenic expression of the Htt exon 1 fragment (Httex1) in polyQ-expanded form is sufficient to 37 produce a HD-like pathology in rodent and primate models, which is suggestive of these fragments 38 mediating proteotoxicity [5][6][7]. The mechanism of toxicity remains to be unequivocally determined but it 39 is thought to involve two distinct components; soluble and inclusion states of Httex1 [8]. Soluble states, 40 which may include monomeric or small nanometer-sized oligomers of mutant Httex1 cause oxidative 41 and mitochondrial stress and increase the risk of apoptosis in cell culture models of disease [9-12]. We 42 previously suggested that the toxicity of the soluble forms of mutant Httex1 may involve a quality 43 control feedback mechanism during translation involving stalled Httex1 nascent chains, which when 44 unresolved triggers apoptosis [8]. Once inclusions form survival times are improved in cell culture 45 models of disease, leading to a hypothesis that inclusion formation alleviates toxicity by sequestering 46 the soluble toxic forms away from harm (reviewed in [13]). However, rather than returning the cell to a normal state of homeostasis, cells in culture with inclusions are metabolically quiescent and die at a primers and ligation into the P2A T2A Htt25Q Stall Reporter via the NotI and BamHI restriction sites.  168 RESULTS

169
We previously postulated the presence of a translation-related quality control mechanism that clears 170 aggregating or misfolding proteins emerging from the ribosome in cells lacking inclusions. Prior data has 171 shown that polyglutamine (polyQ)-expanded Httex1 is more efficiently degraded than the wild-type 172 counterpart, which is consistent with an elevated clearance mechanism at this step  (Fig 1A). Each construct is encoded in frame without stop 182 codons however the test sequence is flanked by viral P2A sequences, which causes the ribosome to skip 183 the formation of a peptide bond but otherwise continue translation elongation uninterrupted [22].

184
Complete translation of the cassette from one ribosome will generate three independent proteins (GFP,   185 test protein, and mCherry). However, should the ribosome stall during synthesis (such as through the 186 previously established poly-lysine (20K) sequence used here as a control [21]), mCherry is produced at 187 lower stoichiometries than the GFP. In our hands, we noted that there was a small fraction of protein 188 synthesized reading through the P2A sequences (Fig 1B). Of particular note was the appearance of SDS-189 insoluble material in the stacking gel for the mutant polyQ-length form of Httex1 (97Q), which is 190 indicative of SDS-insoluble mHtt products that arises from aggregation [3] (Fig 1B) (Fig 1C). into nuclear RNA pools and to a lesser extent the cytoplasm (Fig 2A-B). The levels of EU inside the 229 inclusions was lower than the surrounding cytosol but was higher than background levels of 230 fluorescence determined by control cells labelled with Alexa 647 in the absence of EU (Fig 2C). This  staining, yet both Ltn1 and Rack1 were enriched in the outer layer of the inclusions suggesting a specific 264 enrichment of the RQC machinery to the aggregates (Fig 3A and B). The over-expression of these 265 proteins did not appear to influence the formation of inclusions of Httex1(97Q) (Fig 3C). Sgta, Upf1) and HBRi inclusions (Snu13, Rpl18). Hspb1 (Hsp27), which is a small heat shock protein 285 involved in chaperone activity, appeared to enrich in the outer layer of Httex1(97Q) inclusions (Fig 4A).

286
Furthermore, at high expression levels, Hspb1 lowered the potential of Httex1 to form inclusions which 287 suggests it plays a role in mitigating (or reversing) aggregation (Fig 4A). Similarly, Sgta, which is a co-288 chaperone involved in the Bag6 system and ERAD also enriched to the inner and outer layer of Httex1 289 and had a more potent effect on suppressing aggregation of Httex1(97Q) at high co-expression levels 290 (Fig 4B). Snu13, which is involved in pre-mRNA splicing, also was enriched in the inner and outer layers 291 of the Httex1(97Q) inclusions and could suppress the aggregation of Httex1(97Q) (Fig 4C). Upf1 also 292 enriched in the inner and outer edges of the inclusions but did not appear to affect the aggregation 293 process of Httex1 (Fig 4D). Previously we found RPL18 as the most enriched protein in Httex1 inclusions 294 by proteomics [8]. Halo-tagged Rpl18 was mostly present in the nucleus as punctate structures but 295 small levels were seen in the cytosol (Fig 4E). Endogenous Rpl18 is anticipated to mostly reside in the  (Fig 4E). However, the strong nuclear localization of the Halo-tagged Rpl18 suggests it might 300 not be properly forming complexes with the ribosomes under these conditions. 301 both HBRi and PBRi inclusions did not recover over a period of 20 mins, which is consistent with a non-314 liquid aggregation state (Fig 5A-B). As a further probe for porosity we immunostained cells with a GFP 315 antibody to test whether the antibody was able to penetrate the inclusions formed by TC9-Httex1(97Q) 316 fused to GFP derivative Cerulean. The antibody formed a tight ring around the inclusions of both PBRi 317 and HBRi, which suggested both inclusion states formed impenetrable aggregates (Fig 5C) and there was 318 no statistical difference in penetration distance of the antibody staining between the HBRi and PBRi 319 suggesting that inclusions form a dense, and immobile core structure quickly after formation (Fig 5D).

354
Of the proteins that were co-aggregated Sgta, Snu13 and Upf1 were mildly enriched inside the inclusion 355 whereas Rack1, Ltn1 and Hspb1 were only or more extensively enriched on the outside edge of the 356 inclusion. Previously it was shown that Hspb1 can form molecular condensates, which raises the 357 possibility of a mixed phase separation process with polyQ that may explain some of the co-aggregation 358 mechanism [35,36]. Nonetheless, the overexpression of the proteins involved in ribosome quality 359 control did not alter the aggregation propensity. This result is more consistent with them playing non 360 rate-limiting roles if they are involved in aggregation or clearance, or indeed acting as bystanders that 361 do not play a critical role in mediating inclusion formation but are co-aggregated.

362
In conclusion, our data suggests that nascent chains of mutant Httex1 emergent from the ribosome are 363 unlikely to stall and therefore unlikely to drive inclusion formation as stalled entities. However, given 364 that we did see some ribosome-associated proteins co-aggregating as well as other proteins we 365 previously identified as enriched in inclusions by proteomics, it remains possible that newly synthesized 366 nascent Httex1 contributes to the aggregation process substoichiometrically by nucleating further 367 association of post translated pools of Httex1. Alternatively, it is possible that aggregation of mHttex1 368 can co-aggregate with endogenous Htt that is engaging in a physiological function of regulating 369 ribosome translocation rates, and thereby drawing translation machinery into the inclusions in trans.

370
Both contexts are consistent with other reports of pre-existing pools of Httex1 monomer and small 371 oligomers being quickly absconded into the inclusion once they form [37].