The role of interruptions in polyQ in the pathology of SCA1

Abstract

At least nine dominant neurodegenerative diseases are caused by expansion of CAG repeats in coding regions of specific genes that result in abnormal elongation of polyglutamine (polyQ) tracts in the corresponding gene products. When above a threshold that is specific for each disease the expanded polyQ repeats promote protein aggregation, misfolding and neuronal cell death. The length of the polyQ tract inversely correlates with the age at disease onset. It has been observed that interruption of the CAG tract by silent (CAA) or missense (CAT) mutations may strongly modulate the effect of the expansion and delay the onset age. We have carried out an extensive study in which we have complemented DNA sequence determination with cellular and biophysical models. By sequencing cloned normal and expanded SCA1 alleles taken from our cohort of ataxia patients we have determined sequence variations not detected by allele sizing and observed for the first time that repeat instability can occur even in the presence of CAG interruptions. We show that histidine interrupted pathogenic alleles occur with relatively high frequency (11%) and that the age at onset inversely correlates linearly with the longer uninterrupted CAG stretch. This could be reproduced in a cellular model to support the hypothesis of a linear behaviour of polyQ. We clarified by in vitro studies the mechanism by which polyQ interruption slows down aggregation. Our study contributes to the understanding of the role of polyQ interruption in the SCA1 phenotype with regards to age at disease onset, prognosis and transmission.

Figure 1. Correlation between polyQ length expansion and age at disease onset.

Pathogenic allele size as determined via diagnostic sizing (A, n = 35) and by clone sequencing (B, n = 35). Interrupted alleles are indicated on the diagnostic sizing graph by bold crosses for reference only, since these cannot be differentiated by this method. The clone sequencing approach further sub-divided the patient group into those with primarily uninterrupted (crosses, n = 31) and interrupted (open circles – total allele length; filled circles – longest CAG stretch, n = 4) allele repeats; these subsets are not distinguished via conventional sizing. A linear model was used to relate the patient's age at disease onset to their indistinguishable pathogenic allele (A) or uninterrupted pathogenic allele (B) size. The solid lines depict the fit result while the dashed lines show the prediction bounds at the 95% confidence level. Sequencing of cloned patient alleles produces a higher quality fit with tighter prediction bounds. The data used to prepare this figure may be found in Table S3.

Figure 2. Pedigree analysis and the transmission of the pathogenic expanded allele in two families.

Subject numbers (#) correspond to cloned individuals indicated in Table S3, where individual cloned sequences can be found. The mean total repeat size as determined by clone sequencing for each allele is shown, along with the age at disease onset. The first family (A) illustrates the transmission of a pure CAG repeat tract across three generations. The second family (B) shows transmission of an interrupted pathogenic allele. The limited number of clones analysed may have introduced some degree of sampling error.

Figure 3. Interruption of the expanded polyQ stretch by histidines reduces its aggregation in transfected cells.

COS cells were transfected with uninterrupted (A, B and C) or His interrupted polyQ-ataxinCT constructs (D and E). The repeat pattern is shown above each panel. Cells were stained with anti-ataxin-1 antibodies and FITC conjugated secondary antibodies. The % of cells with aggregates (average from at least 3 different experiments) is shown at bottom right hand corner of each panel. Arrows indicate cells with diffused expression of ataxinCT proteins and arrowheads indicate aggregates. (F) Analysis of the data in Table 1 plotting the percentage of aggregates observed in cells expressing uninterrupted polyQ proteins (green triangles, R² = 0.94) and His-interrupted polyQ proteins plotted either against total repeat length (blue squares, R² = 0.40) or against the longest polyQ tract length (orange triangles, R² = 0.97). The regression lines and the R² values were calculated using Microsoft Excel.

Figure 4. Detection of protein aggregates formed in transfected COS cells by slot blot filter assay.

Slot blot filter assay was carried out on the insoluble fraction isolated from transfected COS cells. COS cells were transfected with ataxinCT constructs expressing polyQ repeats with and without His interruptions. Cell pellets obtained after lysis were treated with DNAse I, boiled in 2% SDS and filtered through a cellulose acetate membrane. Aggregated protein retained on the membrane was detected with anti-ataxin-1 antibodies. (A) Proteins retained from uninterrupted (82Q; 54Q) and interrupted (Q30HQ12HQ12HQ12HQ12; Q30HQ11HQ11) polyQ expressing cells. Results were consistent across three separate experiments. (B) Supernatants obtained after cell lysis from each of the above samples were subjected to immunoprecipitation using anti GFP antibodies and the precipitates were subjected to western blot analysis using anti-GFP antibodies. Lane 1, uninterrupted Q82; lane2, interrupted 82Q; lane 3, uninterrupted 54Q; lane 4, interrupted 54Q. Molecular weight markers are indicated on the left.

Figure 5. Comparison of the structural and aggregation properties of interrupted and uninterrupted polyQ.

(A) ¹H NMR spectra of interrupted (upper trace) and uninterrupted (lower trace) of the GST fused polyQ41 and polyQ41H peptides recorded at 600 MHz and 10°C on 50 µM protein samples in 50 mM sodium phosphate at pH 6.9 and 3 mM DTT. (B) Aggregation kinetics followed by NMR collected forpolyQ41 (black) and polyQ41H (red). Continuous and broken lines are used for filtered and non-filtered samples respectively. (C) Aggregation kinetics obtained by ThT fluorescence detection at 25°C in a stirring plate reader. The conventions are the same as in (B). (D) Simultaneous recording of the SLS, plotting the absorbance at 473 nm (black), coupled with ThT fluorescence aggregation kinetics (grey) for polyQ41. (E) The same as in (D) for polyQ41H (in red and orange respectively). The measurements were carried out at 37°C. (F) Plot of the light scattering intensity as a function of temperature for polyQ41 (black) and polyQ41H (red).