|Cancer||Cell death||Cell cycle||Cytoskeleton||Exo/endocytosis||Differentiation||Division||Organelles||Signalling||Stem cells||Trafficking|
The composition of the polyglutamine-containing proteins influences their co-aggregation properties
Daniel Bąk*† and Michał Milewski*1
*Laboratory of Cell Biology, Department of Medical Genetics, Institute of Mother and Child, Kasprzaka 17a, 01211 Warsaw, Poland, and †Postgraduate School of Molecular Medicine, wirki i Wigury 61, 02091 Warsaw, Poland
The sequestration of crucial cellular proteins into insoluble aggregates formed by the polypeptides containing expanded polyglutamine tracts has been proposed to be the key mechanism responsible for the abnormal cell functioning in the so-called polyglutamine diseases. To evaluate to what extent the ability of polyglutamine sequences to recruit other proteins into the intracellular aggregates depends on the composition of the aggregating peptide, we analysed the co-aggregation properties of the N-terminal fragment of huntingtin fused with unrelated non-aggregating and/or self-aggregating peptides. We show that the ability of the mutated N-terminal huntingtin fragment to sequester non-related proteins can be significantly increased by fusion with the non-aggregating reporter protein [GFP (green fluorescence protein)]. By contrast, fusion with the self-aggregating C-terminal fragment of the CFTR (cystic fibrosis transmembrane conductance regulator) dramatically reduces the sequestration of related non-fused huntingtin fragments. We also demonstrate that the co-aggregation of different non-fused N-terminal huntingtin fragments depends on their length, with long fragments of the wild-type huntingtin not only excluded from the nuclear inclusions, but also very inefficiently sequestered into the cytoplasmic aggregates formed by the short fragments of mutant protein. Additionally, our results suggest that atypical intracellular aggregation patterns, which include unusual distribution and/or morphology of protein aggregates, are associated with altered ability of accumulating proteins to co-aggregate with other peptides.
Key words: co-aggregation, huntingtin, Huntington's disease, polyglutamine, protein aggregation
Abbreviations: CFTR, cystic fibrosis transmembrane conductance regulator, DAPI, 4, 6-diamidino-2-phenylindole, GFP, green fluorescence protein, HA, haemagglutinin
1To whom correspondence should be addressed (email firstname.lastname@example.org).
The polyglutamine diseases (or polyglutaminopathies) are a group of hereditary disorders, characterized by abnormal elongation (expansion) of the unstable (CAG)n trinucleotide repeats, encoding polyglutamine (poly-Q) tracts in expressed proteins. Proteins with expanded polyglutamine tracts have been implicated in the pathogenesis of nine neurodegenerative diseases, including HD (Huntington's disease), different types of spinocerebellar ataxia (SCA1-3, 6, 7, 17), DRPLA (dentatorubropallidoluysian atrophy) and SBMA (spinobulbar muscular atrophy) [reviewed in Shao and Diamond (2007)]. Since all these mutated proteins show increased propensity to form insoluble intranuclear and/or cytoplasmic aggregates, such abnormal protein aggregation has been commonly associated with the pathophysiology of all polyglutamine disorders.
Several different mechanisms have been, so far, proposed to explain the relationship between the abnormal aggregation of proteins with expanded poly-Q tracts and the pathophysiology of particular polyglutamine diseases. For example, the aggregates have been assumed to form abnormal structures within neurites that block the intracellular transport between the cell body and the cellular periphery, thus obstructing the delivery of crucial cellular factors to the synapses (Lee et al., 2004). Another model of cellular pathophysiology points to the ability of the aggregation intermediates, namely, small hydrophobic protein oligomers, to form non-specific channels within the cellular membranes, especially within the mitochondrial membrane, which could potentially contribute to the mitochondrial dysfunction and subsequent cell death (Hirakura et al., 2000). However, during the last decade, most attention has been given to the so-called sequestration hypothesis that connects abnormal protein aggregation with increased sequestration of crucial cellular proteins, including important transcription factors, frequently containing poly-Q tracts (Stenoien et al., 1999; Preisinger et al., 1999; Suhr et al., 2001). Notably, as the sequestration of proteins into the insoluble aggregates assumes some specificity of interactions between the aggregating polypeptides and sequestered molecules, this model may allow for certain differences between the co-aggregation pathways in particular polyglutamine diseases, even when assuming that the poly-Q tracts, which are common to all mutated polyglutamine proteins, play a central role in these presumably pathogenic interactions.
Although protein sequestration seems to be essential for the polyglutamine-associated pathology, little is known whether and to what extent amino acid sequences other than the poly-Q tracts can contribute to the interactions with sequestered proteins. Many experimental in vitro models of the polyglutamine-induced aggregation use synthetic poly-Q peptides only, thus evading the question of how accompanying sequences could modulate this process, although it has been demonstrated that sequences flanking the polyglutamine stretches can influence the self-aggregation propensities of the poly-Q-containing proteins (Nozaki et al., 2001). Moreover, fusions with unrelated reporter proteins are frequently used to study protein self- and co-aggregation in vivo. Since many studies have shown that modifying the peptide's sequence by fusing it with the reporter protein may influence its ability to self-aggregate (Stenoien et al., 1999; Thomas and Maule, 2000; Milewski et al., 2002; Link et al., 2006; Bulone et al., 2006; Bąk et al., 2007, Rich and Varadaraj, 2007), it is conceivable that such fusion may also affect the ability of fused proteins to specifically co-aggregate with other proteins, although no previous reports have investigated this issue.
Huntingtin, a poly-Q-containing protein mutated in Huntington's disease, is a frequently studied model of the poly-Q-induced protein aggregation (Bates, 2003). The huntingtin aggregates, formed by the N-terminal protein fragments containing expanded poly-Q tracts (Cooper et al., 1998; Martindale et al., 1998; Sieradzan et al., 1999), are able to sequester not only the wild-type huntingtin fragments (Huang et al., 1998; Kazantsev et al., 1999; Busch et al., 2003) but also many other polypeptides, including the poly-Q-containing transcription factors (Preisinger et al., 1999; Steffan et al., 2000; Schaffar et al., 2004), which presumably contributes to the pathogenesis of the disease.
Among the models of non-poly-Q-mediated protein aggregation, the model based on the intracellular aggregation of the C-terminal fragment of the CFTR (cystic fibrosis transmembrane conductance regulator) shows remarkable dependence on the presence of a short amino acid sequence (Milewski et al., 2002; Bąk et al., 2007). When fused to GFP (green fluorescence protein), this CFTR-derived fragment leads to the formation of typical insoluble electron-dense protein aggregates, while the non-fused peptide exhibits a very unique accumulation pattern with numerous small aggregates dispersed through the cytoplasm and associated with mitochondria (Milewski et al., 2002). Intriguingly, both these aggregation forms can be abolished by a deletion of a short nine-amino-acid sequence, called the ag region.
To evaluate a potential influence of the protein composition on the co-aggregation of polyglutamine-containing peptides, we tested the outcome of fusing the N-terminal huntingtin fragments with non-related peptides/proteins, showing either minimal tendency to form intracellular inclusions or a strong propensity to self-aggregate. The GFP, commonly used as a reporter molecule, was used as a non-aggregating fusion partner, while the CFTR-derived fragment served as an example of a non-related aggregation-prone peptide. Additionally, we tested whether the natural context of the poly-Q sequence in huntingtin-derived fragments of different lengths will significantly affect their co-aggregation properties.
2. Materials and methods
2.1. Plasmid constructions
The eukaryotic expression plasmids, encoding the N-terminal huntingtin fragment containing either the normal-range (pN63–Q18–Myc–His) or expanded (pN63–Q75–Myc–His) poly-Q tracts were a gift from C.A. Ross (Johns Hopkins University, Baltimore, MD, U.S.A.) (Cooper et al., 1998). The site-directed mutagenesis system Transformer (BD Biosciences) was used to introduce a codon for the conserved lysine residue K6 (according to the UniProt reference sequence, accession no. P42858) that was missing in both constructs, as was revealed by the DNA sequencing analysis, and to introduce a new NotI restriction site, designed for the subsequent subcloning of the huntingtin-encoding DNA fragments. The Transformer mutagenesis system, in conjunction with the subcloning of the XhoI–MssI DNA fragment, was used to further modify these huntingtin-encoding plasmids, designated as pH64m and pH64wt, by replacing the c-Myc tag with the HA (haemagglutinin) epitope sequence (thus obtaining the pH64m–HA and pH64wt–HA plasmids). The sequences of mutagenic primers and selection primers used to create all of the above modifications are available upon request.
The plasmids encoding the longer (amino acids 1–588 of the full-length reference sequence) wild-type (Q17) and mutant (Q146) huntingtin fragments, tagged N-terminally with the 3xFlag epitope, were a gift from D.C. Rubinstein (Cambridge Institute for Medical Research, Cambridge, U.K.) and were previously described (Luo et al., 2005). The 3xFlag epitope has been shown to not affect the localization and aggregation of these protein constructs (Luo et al., 2005).
The construction of the pRK5–SK-derived expression plasmids, encoding the GFP, the C-terminal fragment of CFTR (amino acids 1370–1480) tagged with the HA epitope (HA-C), its Δag version (HA-CΔag) or the C-terminal CFTR fragments fused to GFP (GFP-C and GFP-CΔag), was described elsewhere (Milewski et al., 2001). The NotI and XbaI restriction sites in these plasmids were used to insert the pH64wt- or pH64m-derived NotI–XbaI DNA fragments, containing the huntingtin-encoding sequence, thus leading to the construction of a series of huntingtin-containing fusion proteins, all expressed from the same vector, carrying the CMV (cytomegalovirus) promoter. The coding sequences of all of the above plasmids have been confirmed by DNA sequencing prior to the transfection procedure.
2.2. Cell cultures
Mouse neuronal hippocampus-derived HT-22 cells, kindly provided by K. Domańska-Janik (Mossakowski Medical Research Centre, Warsaw, Poland), were grown in DMEM (Dulbecco's modified Eagle's medium; Gibco) supplemented with 10% fetal bovine serum (Gibco). IB3-1 human bronchial epithelium cells, derived from a cystic fibrosis patient (Zeitlin et al., 1991), were cultured in LHC-8 medium (BioSource) containing 5% fetal bovine serum (Gibco). Both cell lines were cultured in polystyrene flasks in 5% CO2-balanced air at 37°C.
2.3. Transient transfections
Cells were seeded on to collagen-coated glass coverslips (placed in six-well plates) and grown in appropriate standard medium for 20–30 h. After reaching 40–60% confluence, the IB3-1 cells were transiently transfected with the Lipofectin reagent (Gibco), according to the manufacturer's instructions. Since this protocol was not efficient in the case of HT-22 cells, another transfection reagent (ExGen500, Fermentas) was applied to transiently transfect this cell line. In all co-transfection experiments, a 1:1 DNA ratio was used for plasmids encoding different polypeptides. At 24 h after transfection, cells were fixed in 4% paraformaldehyde for 20 min at room temperature and, if not immunostained, mounted in SlowFade (Invitrogen) containing 0.1 mg/ml DAPI (4′,6-diamidino-2-phenylindole; Sigma).
2.4. Western blot analysis
To assess the protein level in transiently transfected IB3-1 cells and to confirm the structural integrity of expressed proteins, the transfected cells were lysed overnight in RIPA buffer (Sigma), and the total cell lysates were separated by SDS/PAGE (12%). The separated proteins were electrophoretically transferred to PVDF membranes (Amersham Pharmacia Biotech) and probed with monoclonal anti-GFP (BD Biosciences), anti-HA (HA.11, Covance), anti-Flag (SIGMA) or anti-huntingtin (MAB5492, Chemicon International) antibodies. The ECL Plus system (Amersham Pharmacia Biotech) was used for the subsequent chemiluminescence detection step, according to the manufacturer's instructions.
2.5. Fluorescence immunostaining
The immunostaining protocol for the paraformaldehyde-fixed IB3-1 and HT-22 cells was essentially as described in the work of Milewski et al. (2002), with the monoclonal anti-HA antibody (12CA5) from Roche Molecular Biochemicals, the polyclonal anti-c-Myc antibody from Santa Cruz Biotechnology and the monoclonal or polyclonal anti-Flag antibodies from Sigma used as primary antibodies. FITC- or Cy3-conjugated anti-mouse and anti-rabbit immunoglobulins (Sigma) were used as secondary antibodies.
2.6. Fluorescence microscopy and statistical analysis
An IX71 fluorescence microscope (Olympus) was used to analyse the protein distribution in transfected (GFP-expressing and/or immunostained) cells. For the self-aggregation analysis, the presence of protein aggregates in at least 100 transfected cells was examined in duplicate in two separate experiments. The self-aggregation level was calculated as the percentage of transfected cells showing the aggregation of the overexpressed protein. In the co-aggregation experiments, the co-aggregation level was calculated as the percentage of cells with at least one common (double signal) aggregate among at least 100 cells (analysed in duplicate in two separate experiments) co-expressing both tested proteins and showing the presence of at least one aggregate (common or separate). Data were given as means±S.D. Standard Student's t test was used to estimate the statistical significance (assumed at P<0.05) of observed differences. Images were prepared for publication using the Cell-F (Olympus) and Acrobat 5.0 (Adobe) software.
3.1. The aggregation of N-terminal huntingtin fragments can be enhanced by fusion with other peptides
Before studying the effect of altered peptide composition on the co-aggregation properties of the N-terminal huntingtin fragments, we first analysed to what extent the intended modifications of the huntingtin-derived peptides would affect their propensity to self-aggregate. As a basic construct, we used the N-terminal fragment of huntingtin, corresponding to the first 64 amino acids (H64) of the reference protein sequence and containing either the normal-range (Q18; H64wt) or expanded (Q75; H64m) polyglutamine tract. These two huntingtin fragments were fused in frame with the GFP and/or the C-terminal fragment of the CFTR, either containing or not the nine-amino-acid ag region that enhances the aggregation of the CFTR-derived peptides (Milewski et al., 2002) (Figure 1A).
Human epithelial IB3-1 and mouse neuronal HT-22 cells were transiently transfected with plasmids expressing the above constructs and analysed for the presence of intracellular aggregates. By contrast with the non-aggregating wild-type N-terminal huntingtin fragment (H64wt), showing only diffused cytoplasmic localization in all transfected cells (Figure 1B), its expanded counterpart (H64m) formed apparent cytoplasmic and/or nuclear aggregates (Figure 1C) in 15.8±1.3% of transfected IB3-1 cells, thus confirming the previously reported relationship between the presence of expanded poly-Q tract and the ability of the N-terminal huntingtin fragments to self-aggregate. However, the aggregation level of the expanded huntingtin fragment was significantly increased following its fusion with the non-aggregating GFP reporter protein. This fusion protein (H64m-G) showed self-aggregation in more than half (53.5±3.5%, P<0.0001) of transfected IB3-1 cells. Moreover, in contrast to the non-fused H64m construct, no intranuclear inclusions were formed, with all aggregates showing apparent cytoplasmic, mostly perinuclear, distribution (Figure 1E), clearly resembling the aggresomes that have been previously reported for huntingtin (Waelter et al., 2001) and other aggregating polypeptides [for a review, see Olzmann et al. (2008)]. Surprisingly, fusion with GFP altered also the intracellular accumulation pattern of the non-aggregating H64wt construct. A small proportion (7.7±0.3%, P<0.0001) of IB3-1 cells expressing this fusion protein (H64wt-G) showed the presence of atypical amorphous cytoplasmic bodies (Figure 1D). Closer examination of this particular accumulation pattern revealed that it likely represents an increased concentration of the protein around the perinuclear vacuoles (data not shown), with similar accumulation form of huntingtin previously noted by others (Kim et al., 1999). Although these structures did not resemble the compact bodies formed by the fused or non-fused huntingtin fragments with expanded poly-Q tracts, this result suggested that fusion with GFP cannot only increase the aggregation level of the aggregation-prone proteins but may also induce some accumulation when fused to the otherwise non-aggregating peptides.
Even stronger increase of the aggregation level was observed when the mutant or wild-type N-terminal huntingtin fragments were fused with the self-aggregating CFTR-derived peptide, containing the last 111 amino acids of CFTR. The vast majority of IB3-1 cells expressing either the H64m-C (75.3±5.2%) or H64wt-C (78.0±7.7%) fusion proteins showed the presence of intracellular aggregates (Figures 1G and 1F, respectively), and their distribution pattern resembled that of aggregates formed by the CFTR C-terminal peptide alone (Milewski et al., 2002) (HA-C, Figure 1I), indicating that the CFTR-derived component determined the behaviour of such fusion proteins, even in the presence of the expanded poly-Q tract. However, the subsequent removal of the aggregation-promoting ag region from the CFTR-derived sequence lead to significant decrease of the aggregation level only when the normal range, but not the expanded poly-Q tract, was present. The corresponding self-aggregation levels decreased to 71.6±2.4% (H64m-CΔag, P = 0.25) and 46.1±5.1% (H64wt-CΔag, P = 0.0004). Additionally, the morphology of most aggregates formed by the ag region-deleted fusion protein H64m-CΔag resembled the morphology of aggregates formed by the mutated huntingtin fragment alone (Figure 1H), which together suggested that, in the absence of the ag region, the expanded poly-Q tract contributes more significantly to the aggregation of the fusion protein. Importantly, the Western blot analysis revealed that different GFP- or CFTR-fused huntingtin fragments were expressed at a comparable level in transfected IB3-1 cells (data not shown), indicating that the observed differences in aggregation propensity cannot be attributed to different expression levels or drastic alterations in stability of expressed proteins.
The relative self-aggregation levels for all of the above constructs were similar in both cell lines tested (IB3-1 and HT-22), although the aggregation was always stronger in the IB3-1 cells (Figure 1J), which most likely corresponded to higher transfection rates in this cell line (data not shown). Therefore, all further co-aggregation experiments were performed in IB3-1 cells to assure high aggregation levels.
3.2. Fusion with GFP induces the co-aggregation of huntingtin with non-related peptides
Since fusion with GFP increased the self-aggregation level of both mutated and non-mutated huntingtin, we tested whether fusion with this reporter protein will also affect the ability of mutated huntingtin fragments to sequester their non-mutated counterparts and other related and non-related proteins. Co-expression of two mutated (Q75) N-terminal huntingtin fragments tagged with different epitopes (HA and c-Myc, respectively) revealed that both these peptides always form common aggregates (100% co-aggregation) in transfected IB3-1 cells (Figure 2A). Similarly, the co-expression of the HA-tagged mutated huntingtin fragment (H64m-HA) with its non-mutant c-Myc-tagged counterpart (H64wt) resulted in formation of aggregates that always contained both peptides (Figure 2B), suggesting that the mutant N-terminal huntingtin fragments are able to sequester the wild-type peptides of similar composition.
The fusion with GFP did not significantly alter the ability of mutated huntingtin to co-aggregate with its non-fused mutated (Figure 2C) or non-mutated (Figure 2D) counterpart. In both cases, the co-aggregation level was very high, reaching 98.9±0.6% and 90.7±3.6%, respectively. However, by contrast with the non-fused H64m construct, the H64m-G fusion protein was able to form common aggregates with the non-related CFTR-derived C-terminal peptides, tagged with the HA epitope (HA-C). Among the cells expressing both constructs (H64m-G and HA-C) and showing protein aggregation of any kind, 83.8±6.4% contained large cytoplasmic spherical aggregates composed of both co-expressed proteins (Figure 2E), whereas the remaining cells (16.2%) contained aggregates formed by HA-C only. Importantly, this co-aggregation process did not require the presence of an expanded poly-Q tract in the huntingtin–GFP fusion protein. Similar common aggregates were also observed, although less frequently (co-aggregation level 53.9±8.9%), when HA-C was co-expressed with the non-mutated huntingtin–GFP fusion protein (H64wt-G, Figure 2F). In both cases, the morphology of common inclusions was typical for the aggregates formed by the corresponding huntingtin–GFP fusions expressed alone. However, these large common inclusions were almost always accompanied by additional smaller aggregates composed of the HA-C peptide only that showed subcellular distribution typical for the accumulations of this particular construct. This indicated that the co-aggregation of both proteins is enabled only when a specific aggregation pathway, leading to the formation of large spherical aggregates, is initiated. Additionally, the GFP component of the above fusion proteins seems to play a key role in this co-aggregation process, as similar spherical aggregates formed by non-fused huntingtin fragments (H64m) did not sequester the HA-C peptide (data not shown) (Milewski et al., 2002).
3.3. The co-aggregation of different huntingtin-derived peptides can be disturbed by fusion with the self-aggregating CFTR C-terminus
To further explore the relationship between the composition of aggregating peptides and their ability to co-aggregate, we examined the co-aggregation properties of huntingtin fragments fused to the self-aggregating CFTR-derived C-terminal peptide. In contrast with the mutated N-terminal huntingtin fragment alone (H64m), its fusion with the CFTR C-terminus (H64m-C) was unable to efficiently sequester not only the wild-type (H64wt, co-aggregation level 7.0±6.1%), but even the expanded (H64m, co-aggregation level 7.0±4.8%) huntingtin fragment (Figures 3A and 3B, respectively), thus indicating that such fusion may partially prevent the co-aggregation with related non-fused huntingtin fragments. Similarly, the corresponding fusion containing the normal-range poly-Q tract (H64wt-C) showed a very low level of co-aggregation (<1%) with the non-fused H64m or H64wt constructs (Figures 3C and 3D, respectively), demonstrating that the normal huntingtin fragment cannot be efficiently sequestered into the aggregates formed by the H64m when fused to the C-terminus of CFTR.
To test whether additional introduction of GFP into the huntingtin–CFTR fusions will influence the inability of these fusion proteins to sequester the non-fused huntingtin fragments, we created the H64m-G-C construct. This fusion protein, showing strong self-aggregation propensity (76.7±1.9% in transfected IB3-1 cells), was able to sequester the non-aggregating H64wt construct (Figure 3E) and form common aggregates with the self-aggregating H64m peptide (Figure 3F). This strongly confirms our previous results and again indicates that the presence of the GFP reporter protein may significantly affect not only the self-aggregation level, but also the co-aggregation level of fused proteins, including those containing the expanded poly-Q tracts.
3.4. Atypical intracellular aggregation patterns can limit the ability of proteins to co-aggregate
Since fusion with the self-aggregating C-terminal portion of CFTR significantly reduced the ability of the N-terminal-mutated huntingtin fragment to sequester the non-fused huntingtin-derived peptides, we further tested whether the co-aggregation of different fusion proteins containing both huntingtin- and CFTR-derived peptides will correlate with the presence of specific amino acid sequences. Two panels of different huntingtin–CFTR fusions, either containing or not the GFP reporter protein, were created. Different pairs of GFP and non-GFP fusions were transiently expressed in IB3-1 cells, and the corresponding co-aggregation levels were calculated based on the co-localization analysis. As shown in Table 1, each pair of peptides was able to form common aggregates, which could be expected from the pairs that always included a peptide fused to GFP, a molecule that was previously demonstrated to induce non-specific co-aggregation when fused to huntingtin-derived peptides. These common aggregates, usually observed as round protein accumulations, were always localized outside the nucleus, as shown on an exemplary image (Supplementary Figure S1A at http://www.cellbiolint.
Table 1 Co-aggregation propensity of different huntingtin–CFTR fusion proteins
aNon-GFP fusion constructs containing the intact ag region within the CFTR C-terminus and showing relatively low (<80%) co-aggregation levels.
aNon-GFP fusion constructs containing the intact ag region within the CFTR C-terminus and showing relatively low (<80%) co-aggregation levels.
org/cbi/034/cbi0340933add.htm). However, such common protein accumulations were missing in a small fraction of co-transfected cells, where separate aggregates of two co-expressed proteins were seen only, suggesting a certain level of aggregation specificity. Importantly, increased specificity of protein aggregation, as estimated based on the relatively low co-aggregation levels (<80%), was observed when one of the co-expressed constructs was the non-GFP fusion protein that contained the intact ag region. As has been shown above, such fusions, although containing the expanded (Q75) or wild-type (Q18) N-terminal huntingtin fragment, were still able to demonstrate the accumulation pattern typical for the CFTR-derived C-terminal peptides. In co-transfected cells, this distinctive accumulation pattern frequently coexisted with round cytoplasmic aggregates formed by the co-expressed GFP-fused huntingtin–CFTR hybrids (as shown in Supplementary Figure S1B).
Together, these results imply that certain characteristic aggregation patterns that are invoked by the presence of specific amino acid sequences may correlate with limited ability of aggregating proteins to sequester other peptides, even those showing significant sequence similarity and/or high propensity to self-aggregate. Consequently, one can conclude that the sequence modifications that influence the distribution of huntingtin-derived peptides may affect their self- and co-aggregation propensities.
3.5. The length of the N-terminal huntingtin fragments determines their co-aggregation properties
Since the results presented above suggested a possible relationship between the subcellular distribution/localization of huntingtin-derived peptides and their ability to co-aggregate with other related peptides, we tested whether the length of the huntingtin N-terminal fragments will affect their ability to co-aggregate. Previous reports have shown that long N-terminal fragments of mutant huntingtin, containing more than ∼200 amino acids of the reference sequence, form only cytoplasmic aggregates, whereas their shorter counterparts are able to generate both cytoplasmic and nuclear inclusions (Hackam et al., 1998; Martindale et al., 1998). Therefore, we investigated whether short (H64) and long (H588) N-terminal fragments of mutant or wild-type huntingtin will co-aggregate when simultaneously overexpressed in transfected IB3-1 cells.
When expressed alone, the wild-type version of H588 (H588wt) showed diffused cytoplasmic distribution (Figure 4A), while its mutant form (H588m) frequently accumulated in cytoplasmic aggregates (Figure 4B), seen in 17.1±2.2% of transfected cells. Importantly, the co-expression of H588m with its shorter counterpart H64m lead to the co-localization of both peptides in similar cytoplasmic aggregates (Figure 4C), with the co-aggregation level reaching 100%. Also, the non-mutant variant of H64 (H64wt), previously shown to be sequestered by H64m (Figure 2B), was similarly sequestered into the cytoplasmic aggregates formed by H588m (Figure 4D). By contrast, the longer fragment of non-mutant huntingtin (H588wt) was not efficiently sequestered into the cytoplasmic inclusions formed by H64m, showing instead diffuse cytoplasmic distribution, although a slightly marked presence of H588wt in the H64m aggregates was occasionally seen (data not shown). This suggested that the sequestration of the wild-type huntingtin fragments may depend on the length of those fragments.
Noticeably, both the wild-type (H588wt) and mutant (H588m) variants of the long N-terminal huntingtin fragment were not incorporated into the intranuclear aggregates formed by H64m (Figures 5A and 5B, respectively), although the short fragment of the wild-type huntingtin (H64wt) was readily sequestered into these nuclear inclusions (Figure 5C). This supported our previous conclusion that the subcellular localization of aggregates may determine the spectrum of co-aggregating proteins. Additionally, the above results clearly indicated that the short fragments of mutant huntingtin were unable to recruit the long fragments of the wild-type protein into the nucleus, thus suggesting that the intranuclear sequestration of relatively long huntingtin fragments is not likely to occur in vivo.
This study demonstrates that the ability of different poly-Q-containing proteins to co-localize in common intracellular inclusions can be influenced by the presence of specific amino acid sequences, accompanying the polyglutamine tract. Two important consequences of this finding are apparent. First, it draws our attention to the possible relationship between the amino acid composition of different pathogenic proteins with expanded poly-Q tracts and their ability to sequester other proteins. Since the specificity of co-aggregation may determine the appearance of disease-specific symptoms in different polyglutamine disorders, future investigations need to focus on identifying specific amino acid sequences within particular poly-Q-containing proteins that could potentially contribute to this process. This should be important even when assuming that the aggregated proteins are not the toxic species of poly-Q polypeptides, as has been recently suggested by many researchers [for review, see Truant et al. (2008)], since the composition of aggregates may reflect the potentially toxic processes that precede the formation of insoluble protein inclusions.
Secondly, our results indicate that fusion of a reporter protein to the poly-Q-containing peptides may drastically alter their co-aggregation properties, which calls for a very careful designing of protein constructs used in future studies on polyglutamine-associated protein sequestration. In this context, using the GFP reporter protein should be especially avoided, as we demonstrated that this commonly used reporter molecule may not only increase the aggregation level of fused aggregation-prone peptides but also induce the intracellular accumulation of the otherwise non-aggregating peptides. Even more importantly, fusion with GFP reduced the specificity of the polyglutamine-associated co-aggregation process. This clearly indicates that using GFP fusions may lead to obtaining false-positive results in sequestration analysis. It has been previously suggested that the ability of GFP to dimerize may facilitate the self-aggregation of different GFP-fused peptides (Milewski et al., 2002). Therefore, the GFP variants that are unable to dimerize (Snapp et al., 2003; Zacharias et al., 2002) should be used to investigate whether the dimerization of GFP is indeed responsible for the increased aggregation level of the GFP-fused proteins.
As demonstrated in our work, specific amino acid sequences attached to the poly-Q-containing huntingtin fragment may either increase or drastically reduce the ability to sequester related huntingtin-derived peptides. By contrast with the sequestration-promoting GFP fusions, similar fusions with the CFTR-derived aggregation-prone peptide lead to the inhibition of the poly-Q-associated co-aggregation process. In addition, we have shown that the length of the N-terminal huntingtin fragments is critical for their co-aggregation properties. Thus, it seems likely that certain huntingtin-derived sequences may differentially contribute to the co-aggregation process, and it would be of special interest to identify the non-poly-Q huntingtin-derived amino acid sequences that could increase or decrease the sequestration rate of specific huntingtin-binding partners. Our results strongly indicate that the huntingtin fragment located downstream of the poly-Q tract, more specifically the region encompassing amino acids 65–588 of the full-length protein, can substantially interfere with the sequestration process. Importantly, this portion of huntingtin is almost entirely enclosed within the pathogenesis-related N-terminal fragment that is produced by the caspase-6 cleavage (Graham et al., 2006). It contains many characteristic amino acid motifs, including the entire proline-rich domain (Dehay and Bertolotti, 2006), the palmitoylated cystein residue at position 214 (Yanai et al., 2006), phosphorylated serines at positions 421 (Zala et al., 2008) and 434 (Luo et al., 2005) and the first three of the ten initially reported huntingtin's HEAT repeats (Andrade and Bork, 1995). Intriguingly, many of these motifs are suspected of regulating the intracellular transport of huntingtin (Takano and Gusella, 2002; Qin et al., 2004; Yanai et al., 2006; Zala et al., 2008). Future studies should explore the potential involvement of those amino acids in protein–protein interactions mediating the co-aggregation of the mutated huntingtin with other proteins, including the wild-type huntingtin.
It is commonly believed that the short N-terminal fragments of mutant huntingtin are responsible for the formation of the intranuclear huntingtin aggregates (DiFiglia, 2002; Lunkes et al., 2002), usually associated with increased toxicity (Gutekunst et al., 1999; Peters et al., 1999; Yang et al., 2002; Schilling et al., 2004). Additionally, we have now demonstrated that such short fragments of the mutant protein are unable to recruit the long huntingtin fragments into the nucleus. Although it has been suggested that small huntingtin fragments are passively transported through the nuclear pores (Hackam et al., 1999; Trushina et al., 2003), there are indications that certain amino acid sequences in huntingtin may facilitate the nuclear import and export (Atwal et al., 2007; Truant et al., 2007). Interestingly, it is known that the amino acid sequences contributing to the nuclear localization of aggregates may influence the protein's ability to co-aggregate with other nuclear polypeptides, as has been demonstrated for ataxin-3, a poly-Q protein mutated in spinocerebellar ataxia type 3 (Chai et al., 2001). Therefore, future studies should determine whether specific N-terminal regions of huntingtin, other than the aggregation-inducing poly-Q tract, are required for the huntingtin-mediated sequestration of specific nuclear proteins, including different transcription factors.
Although several previous studies have demonstrated that the aggregates formed by the mutant huntingtin fragments are able to sequester the relatively short fragments of the wild-type protein (Huang et al., 1998; Busch et al., 2003), the presence of the full-length wild-type huntingtin in such aggregates, although occasionally reported (Martindale et al., 1998), remains a controversial issue, as other studies failed to detect the whole-length protein in aggregates (Huang et al., 1998; Sieradzan et al., 1999). The results of the present study indicate that long fragments of the wild-type huntingtin, apart from being excluded from the nuclear inclusions, are also not efficiently sequestered into the cytoplasmic aggregates formed by the short fragments of mutant huntingtin. This suggests that long fragments of the wild-type huntingtin, including possibly the whole-length protein, have limited ability to co-aggregate with the short fragments of mutant huntingtin, although further studies are required to definitely answer this important question. In this context, one should take into account the very recent finding that the nucleus and not the cytoplasm seems to be the primary site of the caspase-6-mediated production of the relatively long N-terminal huntingtin fragments, encompassing amino acids 1–586 (Warby et al., 2008).
The data presented here indicate that atypical aggregation patterns are usually associated with an altered spectrum of co-aggregating peptides. However, aside from the diverse subcellular distributions of the co-expressed proteins, other factors, including dissimilar physical properties and different kinetics of aggregation, are also likely to interfere with the co-aggregation process. Hence, the biochemical analysis of all observed accumulation forms should be undertaken to test whether particular biochemical parameters of the aggregation process are responsible for the observed differences.
Daniel Bąk created all new DNA constructs. He performed all transfection experiments and most of the microscopy analysis. Michał Milewski designed all experiments and contributed to the microscopy analysis. He also wrote the manuscript. Both authors contributed equally to the analysis of the obtained data.
This work was supported by grants from the
Atwal, RS, Xia, J, Pinchev, D, Taylor, J, Epand, RM and Truant, R (2007) Huntingtin has a membrane association signal that can modulate huntingtin aggregation, nuclear entry and toxicity. Hum Mol Genet 16, 2600-15
Bulone, D, Masino, L, Thomas, DJ, San Biagio, PL and Pastore, A (2006) The interplay between polyQ and protein context delays aggregation by forming a reservoir of protofibrils. PLoS ONE 1, (1), e111
Busch, A, Engemann, S, Lurz, R, Okazawa, H, Lehrach, H and Wanker, EE (2003) Mutant huntingtin promotes the fibrillogenesis of wild-type huntingtin: a potential mechanism for loss of huntingtin function in Huntington's disease. J Biol Chem 278, 41452-61
Chai, Y, Wu, L, Griffin, JD and Paulson, HL (2001) The role of protein composition in specifying nuclear inclusion formation in polyglutamine disease. J Biol Chem 276, 44889-97
Cooper, JK, Schilling, G, Peters, MF, Herring, WJ, Sharp, AH and Kaminsky, Z (1998) Truncated N-terminal fragments of huntingtin with expanded glutamine repeats form nuclear and cytoplasmic aggregates in cell culture. Hum Mol Genet 7, 783-90
Graham, RK, Deng, Y, Slow, EJ, Haigh, B, Bissada, N and Lu, G (2006) Cleavage at the caspase-6 site is required for neuronal dysfunction and degeneration due to mutant huntingtin. Cell 125, 1179-91
Gutekunst, CA, Li, SH, Yi, H, Mulroy, JS, Kuemmerle, S and Jones, R (1999) Nuclear and neuropil aggregates in Huntington's disease: relationship to neuropathology. J Neurosci 19, 2522-34
Hackam, AS, Singaraja, R, Wellington, CL, Metzler, M, McCutcheon, K and Zhang, T (1998) The influence of huntingtin protein size on nuclear localization and cellular toxicity. J Cell Biol 141, 1097-105
Hackam, AS, Singaraja, R, Zhang, T, Gan, L and Hayden, MR (1999) In vitro evidence for both the nucleus and cytoplasm as subcellular sites of pathogenesis in Huntington's disease. Hum Mol Genet 8, 25-33
Hirakura, Y, Azimov, R, Azimova, R and Kagan, BL (2000) Polyglutamine-induced ion channels: a possible mechanism for the neurotoxicity of Huntington and other CAG repeat diseases. J Neurosci Res 60, 490-4
Huang, CC, Faber, PW, Persichetti, F, Mittal, V, Vonsattel, JP and MacDonald, ME (1998) Amyloid formation by mutant huntingtin: threshold, progressivity and recruitment of normal polyglutamine proteins. Somat Cell Mol Genet 24, 217-33
Kazantsev, A, Preisinger, E, Dranovsky, A, Goldgaber, D and Housman, D (1999) Insoluble detergent-resistant aggregates form between pathological and nonpathological lengths of polyglutamine in mammalian cells. Proc Natl Acad Sci USA 96, 11404-9
Kim, M, Lee, HS, Laforet, G, McIntyre, C, Martin, EJ and Chang, P (1999) Mutant huntingtin expression in clonal striatal cells: dissociation of inclusion formation and neuronal survival by caspase inhibition. J Neurosci 19, 964-73
Lee, WC, Yoshihara, M and Littleton, JT (2004) Cytoplasmic aggregates trap polyglutamine-containing proteins and block axonal transport in a Drosophila model of Huntington's disease. Proc Natl Acad Sci USA 101, 3224-9
Link, CD, Fonte, V, Hiester, B, Yerg, J, Ferguson, J and Csontos, S (2006) Conversion of green fluorescent protein into a toxic, aggregation-prone protein by C-terminal addition of a short peptide. J Biol Chem 281, 1808-16
Lunkes, A, Lindenberg, KS, Ben Haiem, L, Weber, C, Devys, D and Landwehrmeyer, GB (2002) Proteases acting on mutant huntingtin generate cleaved products that differentially build up cytoplasmic and nuclear inclusions. Mol Cell 10, 259-69
Luo, S, Vacher, C, Davies, JE and Rubinsztein, DC (2005) Cdk5 phosphorylation of huntingtin reduces its cleavage by caspases: implications for mutant huntingtin toxicity. J Cell Biol 169, 647-56
Martindale, D, Hackam, A, Wieczorek, A, Ellerby, L, Wellington, C and McCutcheon, K (1998) Length of huntingtin and its polyglutamine tract influences localization and frequency of intracellular aggregates. Nat Genet 18, 150-4
Milewski, MI, Mickle, JE, Forrest, JK, Stafford, DM, Moyer, BD and Cheng, J (2001) A PDZ-binding motif is essential but not sufficient to localize the C terminus of CFTR to the apical membrane. J Cell Sci 114, 719-26
Milewski, MI, Mickle, JE, Forrest, JK, Stanton, BA and Cutting, GR (2002) Aggregation of misfolded proteins can be a selective process dependent upon peptide composition. J Biol Chem 277, 34462-70
Nozaki, K, Onodera, O, Takano, H and Tsuji, S (2001) Amino acid sequences flanking polyglutamine stretches influence their potential for aggregate formation. Neuroreport 12, 3357-64
Preisinger, E, Jordan, BM, Kazantsev, A and Housman, D (1999) Evidence for a recruitment and sequestration mechanism in Huntington's disease. Philos Trans R Soc London B Biol Sci 354, 1029-34
Qin, ZH, Wang, Y, Sapp, E, Cuiffo, B, Wanker, E and Hayden, MR (2004) Huntingtin bodies sequester vesicle-associated proteins by a polyproline-dependent interaction. J Neurosci 24, 269-81
Schaffar, G, Breuer, P, Boteva, R, Behrends, C, Tzvetkov, N and Strippel, N (2004) Cellular toxicity of polyglutamine expansion proteins: mechanism of transcription factor deactivation. Mol Cell 15, 95-105
Schilling, G, Savonenko, AV, Klevytska, A, Morton, JL, Tucker, SM and Poirier, M (2004) Nuclear-targeting of mutant huntingtin fragments produces Huntington's disease-like phenotypes in transgenic mice. Hum Mol Genet 13, 1599-610
Sieradzan, KA, Mechan, AO, Jones, L, Wanker, EE, Nukina, N and Mann, DM (1999) Huntington's disease intranuclear inclusions contain truncated, ubiquitinated huntingtin protein. Exp Neurol 156, 92-9
Snapp, EL, Hegde, RS, Francolini, M, Lombardo, F, Colombo, S and Pedrazzini, E (2003) Formation of stacked ER cisternae by low affinity protein interactions. J Cell Biol 163, 257-69
Steffan, JS, Kazantsev, A, Spasic-Boskovic, O, Greenwald, M, Zhu, YZ and Gohler, H (2000) The Huntington's disease protein interacts with p53 and CREB-binding protein and represses transcription. Proc Natl Acad Sci USA 97, 6763-8
Stenoien, DL, Cummings, CJ, Adams, HP, Mancini, MG, Patel, K and DeMartino, GN (1999) Polyglutamine-expanded androgen receptors form aggregates that sequester heat shock proteins, proteasome components and SRC-1, and are suppressed by the HDJ-2 chaperone. Hum Mol Genet 8, 731-41
Suhr, ST, Senut, MC, Whitelegge, JP, Faull, KF, Cuizon, DB and Gage, FH (2001) Identities of sequestered proteins in aggregates from cells with induced polyglutamine expression. J Cell Biol 153, 283-94
Takano, H and Gusella, JF (2002) The predominantly HEAT-like motif structure of huntingtin and its association and coincident nuclear entry with dorsal, an NF-kB/Rel/dorsal family transcription factor. BMC Neurosci 3, 15
Thomas, CL and Maule, AJ (2000) Limitations on the use of fused green fluorescent protein to investigate structure–function relationships for the cauliflower mosaic virus movement protein. J Gen Virol 81, 1851-5
Truant, R, Atwal, RS, Desmond, C, Munsie, L and Tran, T (2008) Huntington's disease: revisiting the aggregation hypothesis in polyglutamine neurodegenerative diseases. FEBS J 275, 4252-62
Trushina, E, Heldebrant, MP, Perez-Terzic, CM, Bortolon, R, Kovtun, IV and Badger, JD (2003) Microtubule destabilization and nuclear entry are sequential steps leading to toxicity in Huntington's disease. Proc Natl Acad Sci USA 100, 12171-6
Waelter, S, Boeddrich, A, Lurz, R, Scherzinger, E, Lueder, G and Lehrach, H (2001) Accumulation of mutant huntingtin fragments in aggresome-like inclusion bodies as a result of insufficient protein degradation. Mol Biol Cell 12, 1393-407
Warby, SC, Doty, CN, Graham, RK, Carroll, JB, Yang, YZ and Singaraja, RR (2008) Activated caspase-6 and caspase-6-cleaved fragments of huntingtin specifically colocalize in the nucleus. Hum Mol Genet 17, 2390-404
Yanai, A, Huang, K, Kang, R, Singaraja, RR, Arstikaitis, P and Gan, L (2006) Palmitoylation of huntingtin by HIP14 is essential for its trafficking and function. Nat Neurosci 9, 824-31
Zala, D, Colin, E, Rangone, H, Liot, G, Humbert, S and Saudou, F (2008) Phosphorylation of mutant huntingtin at S421 restores anterograde and retrograde transport in neurons. Hum Mol Genet 17, 3837-46
Zeitlin, PL, Lu, L, Rhim, J, Cutting, G, Stetten, G and Kieffer, KA (1991) A cystic fibrosis bronchial epithelial cell line: immortalization by adeno-12-SV40 infection. Am J Respir Cell Mol Biol 4, 313-9
Received 8 December 2009/6 May 2010; accepted 1 June 2010
Published as Cell Biology International Immediate Publication 1 June 2010, doi:10.1042/CBI20090474
© The Author(s) Journal compilation © 2010 Portland Press Limited
ISSN Print: 1065-6995
ISSN Electronic: 1095-8355
Published by Portland Press Limited on behalf of the International Federation for Cell Biology (IFCB)
Figure 2 Co-aggregation between the GFP-fused and non-fused huntingtin fragments in transfected IB3-1 cells
Figure 3 Lack of co-aggregation between the CFTR-fused and non-fused huntingtin fragments in transfected IB3-1 cells
Figure 4 Co-aggregation of different short (H64) and long (H588) fragments of huntingtin in the cytoplasm of transfected IB3-1 cells