A role of cellular translation regulation associated with toxic Huntingtin protein

Abstract

Huntington’s disease (HD) is a severe neurodegenerative disorder caused by poly Q repeat expansion in the Huntingtin (Htt) gene. While the Htt amyloid aggregates are known to affect many cellular processes, their role in translation has not been addressed. Here we report that pathogenic Htt expression causes a protein synthesis deficit in cells. We find a functional prion-like protein, the translation regulator Orb2, to be sequestered by Htt aggregates in cells. Co-expression of Orb2 can partially rescue the lethality associated with poly Q expanded Htt. These findings can be relevant for HD as human homologs of Orb2 are also sequestered by pathogenic Htt aggregates. Our work suggests that translation dysfunction is one of the contributors to the pathogenesis of HD and new therapies targeting protein synthesis pathways might help to alleviate disease symptoms.

Electronic supplementary material

The online version of this article (10.1007/s00018-019-03392-y) contains supplementary material, which is available to authorized users.

Introduction

Huntington’s disease (HD) is an autosomal dominant neurodegenerative disorder with no known treatment till date [1]. The symptoms associated with the disease include severe motor defects such as chorea and uncoordinated movements along with cognitive and behavioral problems that are progressive in nature. In the human brain, HD phenotypes are manifested as severe atrophy of the caudate nucleus and putamen regions which together are part of the dorsal striatum. While the adult-onset HD phenotypes start manifesting after 40 years of age, there are also cases of juvenile HD which begin manifesting at very early ages and are more aggressive in nature [2]. HD is caused by the expansion of CAG repeats in Exon1 of the Huntingtin gene on chromosome 4p16 [3]. The CAG triplet codes for Glutamine (Q) and disease-causing expansions can range from approximately 40–200 [4]. Due to polyQ expansion, the expanded Huntingtin (Htt) protein forms misfolded aggregates causing an imbalance in cellular proteostasis and eventually neuronal cell death. Cellular proteostasis is regulated through a careful orchestration of protein synthesis, folding, and degradation [5]. In HD models, while the involvement of protein-folding pathways through chaperone networks and proteasomal degradation pathways are well studied [6–12], the connection to protein synthesis is not very well established. There are several indicators pointing towards such a plausible connection. A network of proteins involved in rRNA processing and ribosome biogenesis were found to be differentially expressed in Yeast expressing pathogenic HttQ103 [13]. Components of ribosome quality control are also involved in Htt aggregate clearance in Yeast [14]. In mouse brains, both non-pathogenic and pathogenic Htt was found to be associated with translating polysomes [15]. Proteomic analysis of Neuro2a cells expressing pathogenic Htt identified proteins involved in RNA processing, ribosome biogenesis, and translation to interact with oligomers of Htt [16]. Further proteomic profiling of the brain from a mouse model of HD also found several RNA binding proteins involved in translational control and ribosome biogenesis as part of the insoluble proteome [17], suggesting the presence of pathogenic Htt might impair the functionality of these proteins. HD brains also show sense and antisense repeat-associated non-ATG (RAN) translation proteins, which are responsible for toxicity [18].

Motivated by these observations here we investigated the status of cellular translation associated with non-pathogenic and pathogenic forms of the Htt protein. We find that pathogenic Htt expression causes a cellular translation deficit in both Drosophila and Yeast models. We further found that two isoforms of Orb2, a translation regulator in Drosophila are sequestered by Htt aggregates. The sequestration probably happens through adsorption like process. Coexpression of Orb2 isoforms can rescue the lethality caused by pathogenic Htt expression in Drosophila. We further noticed that the human homologs of Orb2, the Cytoplasmic polyadenylation binding proteins (hCPEB1-4) also can be sequestered by Htt aggregates, indicating that our observations with Orb2 may be relevant for human HD.

Results

Expression of pathogenic HttQ138 is associated with reduced translation in cells

We expressed RFP tagged constructs of non-pathogenic HttQ15 and pathogenic HttQ138 in S2 cells. The Htt fragment used here codes for the Caspase-6 cleaved N- terminal 588 amino acid protein, which is important for disease progression as mutating this to prevent cleavage by Caspase-6 prevents neurodegeneration [19]. The HttQ15 construct is expressed in a diffused cytoplasmic pattern, whereas the pathogenic Q138 shows punctate aggregates in cells. To investigate translation in these cells we performed polysome analysis. Polysomes are complexes of ribosomes along with the mRNAs they are translating to proteins [20]. For this, cells were incubated with a translation inhibitor Cycloheximide which stalls the translating ribosomes on the mRNAs. Lysates from these cells were then centrifuged on a 5–45% sucrose density gradient which separates complexes based on their sedimentation coefficients, and as there is RNA in these complexes, absorbance measurement at 254 nm can be used to visualize the separation of the 40S, 60S, 80S ribosomes, and polysome fractions. For quantitation purposes, we compared the area under the curve for polysome and 80S fractions and presented the data as a ratio between the two. In comparison to HttQ15, we observed a ~ 24% decrease in this ratio for HttQ138 expressing cells suggesting that there are fewer polysomes or reduced translation associated with the pathogenic construct (Fig. 1a, b).

Fig. 1.

a Polysome profiles of S2 cells expressing HttQ15 and HttQ138. Inset show images of S2 cells expressing RFP tagged HttQ15 and HttQ138. b Quantitation of Polysome/80S ratio. Data is from n = 4 experiments and is represented as a relative fold change for HttQ138 compared to HttQ15. Error bars represent SEM and significance is tested using unpaired one-tailed t test. c Polysome profiles of Yeast expressing HttQ25 and HttQ103. Inset show images of cells expressing GFP tagged HttQ25 and HttQ103. d Quantitation of Polysome/80S ratio from HttQ25 and HttQ103 expressing Yeast. Data from n = 3 experiments represented as a relative fold change of HttQ103 compared to HttQ25. Error bars represent SEM and significance is tested using unpaired one-tailed t-test

We also performed polysome analysis in a Yeast model of HD expressing Exon1 of Huntingtin with 25Q (non-pathogenic) and 103Q repeats (pathogenic) under galactose-inducible promoter [21]. Similar to our observations with S2 cells, here we found a ~ 42.5% decrease in the polysome/80S ratio for pathogenic HttQ103 compared to non-pathogenic HttQ25 (Fig. 1c, d). One common feature in the polysome profiles of both Drosophila S2 cells and Yeast is a higher 80S peak associated with pathogenic Htt expression. These results from two different systems suggest that the expression of pathogenic Htt is associated with decreased polysome/80S ratio in cells. We observed no significant difference in the number of dead cells between the two groups expressing non-pathogenic and pathogenic Htt, at the time point at which the polysome experiments were done with S2 cells and Yeast cells (Supplemental figure 1A, B and C). This suggests the observed deficit in the Polysome/80S ratio is not due to a difference in cellular toxicity.

We next performed the Puromycin incorporation assay to compare levels of newly translated proteins in HttQ15 and HttQ138 cells. Puromycin gets incorporated in translating proteins and blocks further incorporation of amino acids by halting translation by premature chain termination [22]. These cells can then be lysed and puromycylated proteins detected on western blots using anti Puromycin antibody [23, 24]. Here, we observed a ~ 46.6% lesser Puromycin incorporation in HttQ138 cells in comparison to HttQ15 cells (Fig. 2a, b). Together, these observations suggest that pathogenic Htt expression results in translation dysfunction in cells.

Fig. 2.

a Puromycin incorporation assay was performed using anti Puromycin antibody. Western blot shows reduced Puromycin incorporation in HttQ138 cells in comparison to HttQ15 cells. The right panel is the Ponceau-S stained membrane for the same blot. b Quantitation of Puromycin incorporation from HttQ15 and HttQ138 cells shows significantly reduced Puromycin incorporation in HttQ138 cells. Data is from n = 3 experiments and is represented as a relative fold change for HttQ138 compared to HttQ15. Error bars represent SEM and significance is tested using unpaired one-tailed t test. c O-Propargyl-Puromycin (OPP) incorporation assay shows incorporation of OPP in HttQ138 aggregates

In order to visualize nascent protein synthesis in cells, we used an alkyne analog of Puromycin called O-propargyl-puromycin (OPP) which similar to Puromycin gets incorporated in translating proteins and can be detected by a click-based chemical reaction [25]. Surprisingly, upon staining HttQ138 cells with fluorescent OPP we noticed its incorporation into Htt aggregates (Fig. 2c). Based on this observation and a previous study showing Htt binding to its own RNA [26] and proteomics studies which identified several RNA binding proteins and proteins involved in translation as binding partners of Htt [16, 17], we hypothesized that Htt aggregates are sequestering the RNA and translation machinery or its regulators. Due to this some amount of translation, albeit stalled, can occur in the aggregates. An alternate explanation can be that OPP labeled terminated proteins get misfolded and further accumulate in Htt aggregates.

Isoforms of a protein synthesis regulator Orb2 are sequestered by HttQ138 aggregates and rendered non-dynamic

We asked if other RNA binding proteins were also present in these Htt aggregates. Previously Orb2A, an isoform of the translation regulator Orb2 was reported to be sequestered by HttQ128 in Drosophila larval brain [27]. The Htt construct used in the previous report was a much shorter Htt fragment with less number of Q repeats. Orb2 a functional prion-like protein is the homolog of the mammalian Cytoplasmic Polyadenylation Element Binding protein (CPEB), which can de-adenylate or polyadenylate the poly-A tail of its target mRNAs by binding to CPE elements in their 3′UTR [28]. Orb2 is necessary for the maintenance of long term memory [29, 30] and has two isoforms Orb2A and Orb2B, both containing an unstructured low-complexity prion-like domain [30, 31]. On coexpressing GFP tagged Orb2A and Orb2B transgenes along with HttQ138 in the optic lobe neurons in the Drosophila brain, we noticed that both colocalize with HttQ138 aggregates suggesting their sequestration by Htt aggregates (Fig. 3a).

Fig. 3.

a Left panels show Drosophila optic lobes expressing GFP tagged Orb2A, Orb2B and RFP tagged HttQ138 with DAPI in blue color. Right panels show optic lobes coexpressing Orb2A and Orb2B with HttQ138. Upon coexpression, Orb2A and Orb2B are sequestered in HttQ138 aggregates. b Western blot from Drosophila head extracts of HttQ15 and HttQ138 shows increased levels of Orb2B associated with HttQ138. Anti α-Tubulin antibody is used as the loading control. c Quantitation of relative Orb2 levels shows increased Orb2B in HttQ138 expressing Drosophila heads. Data is from n = 4 experiments and is represented as a relative fold change for HttQ138 as compared to HttQ15. Error bars represent SEM and significance is tested using unpaired one-tailed t test

We next asked what happens to endogenous Orb2 levels in the fly heads expressing Htt Q15 and Q138. Of the two isoforms, Orb2B is readily detectable in the fly head extract due to its higher enrichment in comparison to Orb2A [30]. Interestingly, we noticed higher amounts of Orb2B in HttQ138 fly heads compared to HttQ15 (Fig. 3b, c), while there was no significant difference in the mRNA level of Orb2 between them (Supplemental figure 1D). The increased Orb2B protein level might represent a steady-state situation arising from compromised proteostasis in the brain or change in the half-life of the protein due to sequestration by Htt aggregates.

We further asked if endogenous Orb2 interacts with Htt. To address this we performed immunoprecipitations from fly head extract expressing HttQ15 and HttQ138 using an anti Orb2 antibody. On probing the immunoprecipitate with anti-Htt antibody we observed endogenous Orb2 pulls down HttQ138 but not HttQ15 (Fig. 4a). Immunostaining of larval brain dissociated neurons also revealed the presence of endogenous Orb2 in HttQ138 aggregates (Fig. 4b).

Fig. 4.

a Immunoprecipitation from Elav; HttQ15 and Elav; HttQ138 Drosophila heads with anti Orb2 antibody and probed with an anti-Htt antibody shows Orb2 pulls down HttQ138 but not HttQ15. Sup is supernatant/input lane, beads lane represents negative control done with only beads and α-Orb2 IP is the lane for the anti-Orb2 immunoprecipitate. b Representative immunostained image of a cell body from a dissociated larval brain neuron from Elav; HttQ138 larvae stained with anti-Orb2 antibody, showing enrichment of endogenous Orb2 (green) in HttQ138 (red) aggregate

How is Orb2 different in its sequestered state in comparison to a non-sequestered state? To address this, we performed fluorescence recovery after photobleaching (FRAP) experiments on S2 cells coexpressing Orb2 with HttQ138. In S2 cells also we observed the sequestration of both the Orb2 isoforms by Htt aggregates (Fig. 5a). Compared to unsequestered Orb2A and Orb2B, where a decreased but partial recovery for Orb2A and almost complete recovery for Orb2B could be seen, in the sequestered state with HttQ138 there was almost no recovery for both (Fig. 5b, c). This indicates that the dynamic nature and probably function of Orb2 are lost in the sequestered state.

Fig. 5.

a S2 cells expressing Orb2A and Orb2B in left panels showing their expression all over the cell. Panels on the right side of the red line show sequestration of Orb2A and Orb2B upon their coexpression with HttQ138. b, c Fluorescence recovery after photobleaching (FRAP) experiment data shows significantly decreased recovery for sequestered Orb2A and Orb2B in the presence of HttQ138 in comparison to only Orb2A and Orb2B expressing cells. Data from n = 10 experiments were tested using Mann–Whitney U test and the p values are < 0.0001. d Representative image of a fused cell suggests sequestration of Orb2 in HttQ138 aggregates is through adsorption like process

We further asked if the sequestration was due to the coexpression of two sticky proteins inside the cell or through a seeding like mechanism. Towards this, we fused Orb2AGFP expressing cells with HttQ138 cells. In the fused cells, we observed the presence of Orb2 on the peripheral surface of Htt aggregates and not in their core (Fig. 5d). These are probably representatives of starting intermediates in the process of sequestration of Orb2 by Htt aggregates using an adsorption like process.

Coexpression of Orb2 isoforms rescue lethality associated with pathogenic HttQ138

We wanted to know the effect of Orb2 coexpression on the phenotypes associated with Huntington’s disease model in Drosophila. Pan-neuronal expression of HttQ138 using Elav Gal4 causes lethality at pupal stages where only 4.8% of flies emerge. On the coexpression of Orb2A and Orb2B, we found an increase in the survival rate to the extent of 60 and 51% respectively (Fig. 6a). We next performed a loss of function study using an Orb2 RNAi line to test if decreasing Orb2 levels can increase HttQ138 associated lethality. Towards this end, we used a weaker expressing Elav Gal4 line to co-express HttQ15 and HttQ138 along with control Luciferase and Orb2 RNAi lines. Low-level expression of HttQ15 with Orb2 RNAi and control Luciferase RNAi had no effect on the survival of animals. However, upon coexpression with HttQ138 we observed that Orb2 RNAi reduced survival to 50% as compared to 77.8% survival rate with Luciferase RNAi (Fig. 6b). This suggests that the Orb2 knockdown enhances the lethality associated with Htt aggregates. Together the gain and loss of function experiments suggest that Orb2 genetically interacts with the expanded Htt associated pathways and can modulate its toxicity.

Fig. 6.

a Eclosion experiments show that coexpression of Orb2A and Orb2B can rescue the lethality associated with pan-neuronal expression of HttQ138 using Elav Gal4. In comparison to only HttQ138 (n = 608) which showed 4.8% eclosion, coexpression with CD8GFP (n = 26) showed 3.8% eclosion, while coexpression with Orb2A (n = 131) and Orb2B (n = 187) showed 60% and 51% eclosion respectively. b Eclosion experiments performed with a lower expressing Elav Gal4 line show that in the presence of Orb2RNAi coexpression there is decreased survival of HttQ138 expressing animals. For nonpathogenic HttQ15 coexpression with Luciferase (Luc) RNAi (n = 35) or Orb2 RNAi (n = 35) did not show any difference. HttQ138 coexpression with LucRNAi showed 77.8% survival (n = 185) and with Orb2 RNAi showed 50% survival (n = 165). c HttQ138 aggregate numbers in motor neurons in larvae expressing HttQ138 alone (left panel) or together with Orb2A and Orb2B (right panels) do not show any significant difference. d Quantitation of aggregate numbers from HttQ138 alone (n = 3, 11 axons) and HttQ138 with Orb2A (n = 3, 20 axons) or Orb2B from (n = 3, 18 axons) motor neurons do not show any significant difference. Error bars represent SEM and significance tested using unpaired one-tailed t test. e Representative SDD-AGE immunoblot showing no significant difference between the Htt oligomers in HttQ138 and HttQ138 with Orb2A and Orb2B. f Quantitation of Polysome/80S ratio from HttQ138 and HttQ138 + Orb2A GFP and HttQ138 + Orb2B GFP expressing cells shows that coexpression of Orb2A and Orb2B rescues the reduced polysome/80S ratio associated with HttQ138 cells. Data are from n = 3 experiments and represented as a relative fold change for others compared to HttQ15. Error bars represent SEM and significance is tested using unpaired one-tailed t test. g Quantitation of Puromycin incorporation from HttQ138 and HttQ138 + Orb2A GFP and HttQ138 + Orb2B GFP expressing cells shows significantly reduced Puromycin incorporation in HttQ138 cells and rescue with coexpression of Orb2A and Orb2B. Data are from n = 3 experiments and is represented as a relative fold change for others compared to HttQ15. Error bars represent SEM and significance is tested using unpaired one-tailed t test

Coexpression of Orb2 isoforms does not decrease the aggregate load but rescues the translation deficit

As coexpression of Orb2A and Orb2B could rescue the lethality associated with pathogenic HttQ138, we next asked what the mechanism of this rescue might be. As pathogenic Htt forms misfolded toxic aggregates, we first asked if coexpression of Orb2 isoforms decrease the Htt aggregate load in neurons. To test this we imaged the axons coming out of the ventral ganglion in Drosophila larvae. In animals coexpressing Orb2A and Orb2B with pathogenic HttQ138, we observed sequestration of these proteins in the Q138 aggregates (Fig. 6c). This phenotype in the larval axons was similar to what we observed in the optic lobe neurons and S2 cells. On quantitation of the Htt aggregates in axons, we found no significant difference between HttQ138 and the Orb2A or Orb2B rescued animals (Fig. 6d). We also performed semi-denaturing detergent agarose gel electrophoresis assay (SDD-AGE) [32] to detect Htt oligomers. Here also we did not observe any difference in the size distribution of Htt oligomers (Fig. 6e). Both these experiments suggest the mode of rescue by Orb2 coexpression is most likely not due to a decrease in aggregate load.

We next revisited our previous observation that translation is perturbed in HttQ138 cells and asked what happens on coexpression of Orb2 isoforms. On polysome analysis, we observed that coexpression rescues the deficit in the Polysome/80S ratio (Fig. 6f). We also performed a Puromycin incorporation experiment to quantitate protein synthesis and observed that co-expression of Orb2A and Orb2B rescues the reduced translation associated with HttQ138 (Fig. 6g). Overall these experiments suggest that improving the translatory status of the cell might mediate the rescue of the pathogenicity associated with HttQ138.

Human CPEB’s are sequestered by Htt aggregates

Mammals including humans have 4 CPEB genes, CPEB1-4. We asked if human CPEB’s can also be sequestered by pathogenic Htt. We coexpressed GFP tagged hCPEB1-4 along with HttQ138. In all cases, we observed the hCPEBs to be sequestered by Htt aggregates (Fig. 7a). On performing FRAP experiments we noted that similar to Orb2, hCPEB2-4 are rendered non-dynamic by sequestration in Htt aggregates. While there was some recovery of sequestered hCPEB1, it was still significantly lower in comparison to only hCPEB1 (Fig. 7b). This suggests that hCPEBs like their Drosophila homolog can be sequestered by Htt aggregates, which probably makes them non-functional.

Fig. 7.

a S2 cells expressing Human CPEB’s (hCPEB1-4) in the left panel. Right panels show sequestration of Human CPEB’s on coexpression with HttQ138. b Fluorescence recovery after photobleaching (FRAP) experiment data shows significantly decreased recovery for sequestered hCPEB1, 2, 3 and 4 in the presence of HttQ138 in comparison to only hCPEB1-4 expressing cells. Data from n = 9 experiments were tested using Mann–Whitney U test and the p values are < 0.0001

As these experiments were done with overexpression in Drosophila S2 cells, we further attempted to test if our observations with respect to translation and endogenous CPEB’s will correspond in a neuroblastoma origin cell line. In both Neuro2a (mouse neuroblastoma cell line) and SH-SY5Y (human neuroblastoma cell line) cell lines, we found that expression of HttQ138 caused a decrease in puromycin incorporation (Fig. 8a–c) suggesting a reduction in translation. On immunostaining Neuro2a cells expressing HttQ15 and HttQ138 with antibodies against CPEB1 and CPEB3, we observed the presence of endogenous CPEB1 and CPEB3 in the HttQ138 aggregates (Fig. 8d). These observations indicate that our findings with the Drosophila cells and Orb2, a Drosophila translation regulator can be relevant in higher vertebrate origin cells.

Fig. 8.

a Representative image of puromycin incorporation assay in SH-SY5Y cells expressing HttQ15 and HttQ138 shows reduced Puromycin incorporation in HttQ138 cells. The right panel is the Ponceau-S stained membrane for the same blot. b Quantitation of Puromycin incorporation from HttQ15 and HttQ138 cells in SH-SY5Y show significantly reduced Puromycin incorporation in HttQ138 cells. Data is from n = 3 experiments and is represented as relative fold change for HttQ138 compared to HttQ15. Error bars represent SEM and significance is tested using unpaired one-tailed t test. c Quantitation of Puromycin incorporation from HttQ15 and HttQ138 cells in Neuro2A cells show significantly reduced Puromycin incorporation in HttQ138 cells. Data is from n = 3 experiments and is represented as a relative fold change for HttQ138 compared to HttQ15. Error bars represent SEM and significance is tested using an unpaired one-tailed t test. d Immunostaining of Neuro2A cells expressing HttQ15 and HttQ138 (in red) with anti CPEB1 and anti CPEB3 antibodies (in green) showing enrichment of endogenous CPEB1 and CPEB3 in HttQ138 aggregates

Discussion

Previous studies in Alzheimer’s, pathogenic Tau, prion disease and ALS/FTD spectrum disorders associated with Fus, Tdp-43 and C9orf72 expansion have reported translation dysfunction [33–41]. Our study indicates that pathogenic Huntingtin protein can also cause similar dysfunction.

Recently, similar observations on the suppression of protein synthesis by expanded Huntingtin protein in striatal neurons from HD mouse model with 111 CAG repeats are also reported [42]. In these neurons, a 40% decrease in translation compared to control neurons with 7 CAG repeats, was observed using puromycin incorporation assay. Ribosome profiling showed ribosome density to be shifted towards the 5′ end of the mRNAs suggesting a stalled or slow passage of ribosomes during translation. Depletion of expanded Htt could rescue the deficit in protein synthesis and ribosome movement and this deficit was independent of the translation staller Fmr1.

This observation of translation deficit in striatal neurons from HD mouse model complements our study with the N terminal 588 amino acid fragment with 138Q repeats. One reason for the translation dysfunction is possibly through sequestration of translation regulating proteins. Here we find that a translation regulator Orb2 is sequestered by Htt aggregates probably leading to disturbed Orb2 function. Pulldown experiments show endogenous Orb2B interacts with HttQ138 but not with HttQ15. This observation is also analogous to the previous observation that expanded Htt shows stronger affinity towards several ribosomal and RNA binding proteins [42]. The sequestration of Orb2, a prion-like protein by Htt aggregates, resembles the cross-seeding phenomenon seen with other proteins associated with neurodegenerative conditions [43–48]. Orb2 was previously shown to be a modifier of GGGGCC expansion disease model associated with C9orf72 and Ataxin-3 mediated neurodegeneration in Drosophila [49, 50]. Here we find that coexpression of Orb2 rescues Htt associated toxicity. What might be the reason for the rescue of HttQ138 associated toxicity by Orb2? One possibility is that Orb2 sequestration by HttQ138 prevents the HttQ138 aggregates from sequestering other proteins. In the Yeast model for HD, prion-like proteins were observed to rescue the toxicity associated with HttQ103 expression by reducing the sequestration of other prion-like proteins without decreasing the aggregate numbers [51, 52]. Here we notice no significant difference between the Htt aggregates in the axons of only HttQ138 expressing larvae and the Orb2A and B rescued ones. Another possibility is, as several ribosomal proteins and stress-responsive kinases, chaperones, proteins involved in unfolded protein response have putative Orb2 binding elements in its 3′ UTR [53], Orb2 isoform overexpression can somehow modulate their expression to rescue the effects of their downregulation. Overexpression of two chaperones having CPE elements in their 3′ UTR, DnaJ1 and Mrj have been previously reported to rescue Htt associated toxicity in Drosophila [54, 55].

Our findings connecting Orb2 with HD in the Drosophila model might also be relevant in humans as we find that Htt aggregates can sequester hCPEB1-4, and in Neuro2a cells we could detect the presence of endogenous CPEB1 and CPEB3 in the HttQ138 aggregates. In the human genome, almost 20% of the genes can be CPEB1 targets [56, 57]. In a mouse model for HD, CPEB3 and 4 were present in the insoluble proteome in relevant brain regions [17] and CPEB4 specific target mRNAs were found to be enriched in deadenylated transcripts [58]. Patients suffering from HD are also reported to have problems with memory [59, 60]. As Orb2 was identified as a regulator of maintenance of memory, its human homologs might also have similar roles. Translation dysfunction associated with the sequestration of CPEBs might be one of the factors behind memory-associated symptoms in HD patients. Orb2 is probably not the only candidate for Htt sequestration, and towards this, an exhaustive screen for all other RNA binding proteins might be needed in the future to identify other possible targets. The translation regulation pathways can be important targets for improving the disease-associated symptoms in Huntington’s disease.

Materials and methods

Drosophila strains and plasmids

UAS-RFP-HttQ15, UAS-RFP-HttQ138 transgenic lines, and plasmids were kind gifts from Prof. Troy Littleton’s lab [61]. Elav Gal4 (X and 2nd chromosome insertions), UAS-Orb2A-GFP and UAS-Orb2B-GFP lines, Orb2 RNAi lines were kind gifts from Dr. Kausik Si’s lab. Yeast expressing plasmids for HttQ15 and HttQ103 were kind gifts from Prof. Michael Sherman’s lab.

For making mammalian expression plasmids expressing RFP tagged HttQ15 and HttQ138, the protein-coding regions from pUASt plasmids were subcloned into pcDNA3.1 plasmid by digesting the pUASt clones with EcoRI and XbaI enzymes and ligating the released inserts into digested pcDNA3.1 vector using T4 DNA Ligase (NEB).

Cell culture

S2 cells were grown in Schneider’s media supplemented with 10% FBS. Transfections of plasmids were done with Effectene using the manufacturer’s protocol.

Neuro2A and SH-SY5Y cells were obtained from NCCS cell repository and cultured in Minimum Essential Medium Eagle supplemented with 10% FBS and Hanks F12K supplemented with 10% FBS, respectively. For transient transfections, cells were seeded 1 day in advance at an appropriate density. Transfections were done with PEI. For this, 2 μg of plasmid DNA was mixed with PEI and incubated for 30 min. This mixture was then added to cells in an incomplete medium. 6 h post-transfection, the medium was supplemented with 10% FBS until the time for harvesting.

Primary neuron culture

Drosophila larval brains of Elav; HttQ15 and Elav; HttQ138 were dissected in Schneider’s media followed by washing them in Rinaldini’s buffer and digestion using collagenase in Rinaldini’s buffer for 1 h. After removing the enzyme mix, the digested brains were washed three times with Schneider’s media. The digested brains were dissociated using repeated pipetting followed by plating them on Concavalin-coated coverslip bottom dishes and put in a humidified box in 25 °C incubator. After 3 days in culture, the neurons were fixed with 4% Paraformaldehyde and further processed for immunostaining.

Antibodies

Anti-Puromycin antibody was obtained from Kerafast. Anti- α-Tubulin antibody (66031-1) was from Proteintech, Anti CPEB1 (ab73287) and anti CPEB3 (ab10883) antibodies were from Abcam and anti-Huntingtin antibody (Mab2166) was obtained from Chemicon. Secondary antibodies were obtained from CST. Drosophila Orb2 antibody was developed in our lab. Western blots were processed using chemiluminescence and detected with GE AI 600 Imager.

Western blot assay

S2 cells or fly heads were lysed in S2 cell lysis buffer (50 mM Tris–HCl pH 7.8, 150 mM NaCl, 1% Nonidet P-40). Total protein levels of lysates obtained from S2 cells or fly heads were quantified using the Pierce BCA Protein Assay kit (Thermo Scientific) according to manufacturer’s protocol, using a plate reader. Lysates were denatured by boiling in 5 × loading buffer (0.25% Bromophenol blue, 0.5 M DTT, 50% Glycerol, 10% SDS and 0.25 M Tris–Cl (pH 6.8)) and loaded on 10% polyacrylamide gel. Proteins were transferred on PVDF membranes using the transfer apparatus from Bio-rad. Membranes were blocked in TBS, 0.1% Tween, 5% non fat milk. Incubations with primary antibodies were performed overnight at 4  °C. Incubations with secondary anti-mouse- or rabbit horseradish peroxidase conjugated antibodies were carried out for 1 h at room temperature. Protein bands were revealed using chemiluminescence substrate (Clarity TM Western ECL substrate by Bio-rad) and images were acquired using Amersham imager 600.

Growth curve

Htt Q25 and Htt Q103 transformed S. cerevisiae strains were cultured overnight at 30 °C in Glucose containing media. Next day cultures were grown to mid-log phase (OD600 nm = 0.4) in 2% Raffinose containing media. Cells were diluted to an optical density (OD 600 nm = 0.05) in a 96 well plate and were induced with 2% Galactose for induced expression. Growth curves were assessed at 30 °C using a Multimode Reader by collecting absorbance at 600nm every 1 h with uninterrupted shaking.

Trypan blue-based cell viability assay

After 8 h of Galactose induction, HttQ25 and HttQ103 cells were centrifuged and the pellet was resuspended in the residual medium. Cells were mixed with 0.05% Trypan blue dye (1:1) and incubated at room temperature for 2–3 min. Cells were then observed under 20× objective of Nikon Eclipse E600 Fluorescence Microscope on a hemocytometer and the percentage of viable cells was calculated.

Polysome analysis

Polysome analysis was done using the Biocomp gradient station. 5–45% sucrose density gradient in resolving buffer consisting of 140 mM NaCl, 25 mM Tris–Cl (pH:8), 10 mM MgCl2) was poured using Biocomp gradient station. Each gradient was 11 ml in volume. S2 cells expressing the mentioned constructs were incubated with cycloheximide solution (final concentration of 20 μg/ml) for 20 min, before spinning them down at 3500 g for 10 min. The supernatant was removed and cells were immediately lysed with lysis buffer consisting of 300 mM NaCl, 50 mM Tris.Cl (pH:8), 10 mM MgCl₂, 1 mM EGTA, 1% Triton-X100, 0.02% sodium deoxycholate. RNase inhibitor (Rnasin, Promega) was added to the lysate to a final concentration of 1 unit/100 μl. The lysate was spun at 10,000 g at 4 °C for 10 min, and the supernatant was removed to a separate Eppendorf tube, RNA concentration across samples were normalized and equal quantities of RNA were loaded on each gradient. The gradients were loaded on a swing bucket SW41 rotor and were spun at 4 °C at 27,000 g speed for 3 h. The gradients were then unloaded from the ultracentrifuge and were fractionated with simultaneous monitoring of OD at 254 nm using a Biocomp fractionator station. For quantitation purposes, the ratio of area under the curve of polysome/80S was measured.

Puromycin incorporation assay

S2 cells were transfected with RFP-Htt Q15, RFP-Htt Q138, and Orb2A-GFP or Orb2B-GFP as mentioned in the figures. 48 h post-transfection cells were treated with 100ug/ml Puromycin (Sigma) for 30 min and sorted using BD FACSAria™ III Standard Sorter. Sorted cells were then lysed in S2 lysis buffer (150 mM NaCl, 10 mM Tris pH 7.5 and 0.1% NP40) spun at 10000 g for 10 min. The protein concentration of the lysates was measured using the Bradford Assay and an equal amount of protein was used to perform western blot analysis. Membranes were probed with anti-Puromycin antibody from Kerafast (EQ 0001) and goat anti-mouse secondary antibody tagged with HRP from CST.

Immunostaining

Post-fixation cells were washed with 0.1% PBST and blocked for 2 h with blocking buffer (5% NGS in PBST). Cells were then incubated with primary antibodies diluted in the blocking buffer for 2 h followed by washes with PBST buffer. Secondary antibodies conjugated with fluorophores (Alexa 488 or 555) were diluted in blocking buffer and were added to these cells and incubated for 1 h. The secondary antibody solution was removed and cells were washed with PBST. Post immunostaining these were imaged on a Nikon A1R confocal microscope.

OPP staining

OPP staining was done using Click-iT plus OPP Alexa-647 protein synthesis kit as per manufacturer’s protocol. Cells were incubated at 25 °C for 30 min after adding 1:400 dilution of the Click-iT OPP reagent in Schneiders media. These cells were further washed three times with PBS and then fixed with 4% formaldehyde solution and then permeabilized with 0.5% Triton X-100 in PBS for 15 min. For the Click-iT reaction, cells were incubated in the dark at room temperature in the Click-iT reaction cocktail. After 30 min, samples were washed with Click-iT reaction rinse buffer and PBS, followed by imaging.

Cell fusion assay

Cell fusion assays were done by mixing cells expressing RFP tagged HttQ138 with Orb2A GFP expressing cells. These cells were also coexpressing the C. elegans fusogenic protein EFF1. Post mixing, fused cells were identified manually under a microscope and imaged within 4 h.

FRAP assay

FRAP experiments were performed using a 100× oil immersion lens in a Nikon Eclipse Ti-based A1R confocal system. The analysis was done using Fiji and easyfrap (https://easyfrap.vmnet.upatras.gr/) [62, 63]. Mann–Whitney test was used for statistical analysis of FRAP data as the data was not normally distributed.

FACS analysis

S2 cells were transfected with RFP tagged HttQ15 or HttQ138 using Effectene reagent (Qiagen) as per manufacturer-recommended protocol. 48 h post-transfection, cells were harvested and washed with PBS. Cells were then treated with 7AAD to a final concentration of 4 μg/ml for 20 min and then subjected to FACS analysis in BD FACS ARIA III Standard instrument using PE and PerCP/Cy5.5 emission channels. Using FACS Diva software the histogram plots were derived and 7AAD and RFP positive cell percentages were calculated. Triplicate data was converted into % live cells and plotted using Graph Pad Prism.

qPCR analysis

Heads of Elav; HttQ15 and Elav; HttQ138 expressing flies were dissected out and crushed in Trizol reagent. RNA was isolated using the manufacturer’s recommended protocol. Total RNA was reverse transcribed using Primescript 1st Strand cDNA synthesis kit (Clontech). qPCR was performed with Power SYBR green PCR master mix (Thermo Scientific) in QuantStudio 6 Flex Real-Time PCR system. For data analysis, the average cycle thresholds from four independent biological replicate samples were calculated and normalized to housekeeping control gene GAPDH. Normalization was performed using HttQ15 as a normalization control using the formula: [2^(Ct control − Ct target)]. Student’s t test was used to compare gene expression between two different groups and graphs were plotted using Graph Pad Prism 7. P < 0.05 was considered statistically significant.

SDD-AGE

S2 cells (approximately 1 × 10⁷ for each sample) were pelleted down at 3000 g for 1 min. The cell pellet was washed with PBS and further lysed using lysis buffer while rotating, for 1 h. Lysed cells were spun at 10,000 g for 10 min. The supernatant was transferred to a new tube, 4X SDS loading dye was added and loaded onto a 1.5% Agarose gel containing 0.1% SDS. After running the gel it was overnight transferred to a nitrocellulose membrane and further probed with an anti-Htt antibody.

Immunoprecipitation

30 fly heads from Elav; HttQ15 and Elav; HttQ138 were taken in a tube, snap frozen and lysed using pestle in 200 μl of S2 cell lysis buffer. Tubes were kept on ice for 15 min and then spun at 10,000 g at 4 °C for 10 min. The supernatant was collected in new tubes. Out of 200 μl one-tenth was kept aside as input (20 μl) and the remaining 180 μl was divided equally into two parts. One part was incubated with anti-Orb2 antibody while on rotation based mixing for 1.5 h. Protein A beads (blocked with 2.5% BSA) were then added to lysate containing antibody and also to the lysate without antibody, and kept on rotation based mixing for 1.5 h. After removing the lysates the beads were washed with 500 μl of S2 lysis buffer three times after every 10 min keeping on rotation for washing. After third wash, 30 μl 0.2 M glycine (pH 4) was added to the beads for elution, the eluate was neutralized with 1 M Tris pH 9. To this mix 5× running buffer was added and sample was run in 10%SDS PAGE. Blot was probed with anti-Htt antibody (MAB 2166, Millipore, Ms) 1:2000 dilution. Immunoprecipitation experiments were repeated three times.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

HJ started this project and identified Orb2 isoforms as rescuers of Htt pathogenicity. VG performed the S2 cell polysome and puromycin incorporation experiments. MS¹ performed S2 cell imaging and FRAP experiments. MS² did the Yeast polysome experiments. AR did the Orb2 rescue and Orb2 RNAi experiments. MD performed the Orb2 level quantitation, OPP staining, Htt clonings, and immunoprecipitation experiments. RC and AD performed the Yeast lethality and growth curve experiments. TB and AM conceived and designed the experiments. AM wrote the manuscript.

Compliance with ethical standards

Conflict of interest

Authors declare no competing interests.