(A) Flies expressing F20E Aβ₄₂ (blue line) live significantly longer (median survival 29 ± 1 d, n = 400, p < 0.0001) than flies expressing WT Aβ₄₂ (red line) (median survival 24 ± 1 d, n = 100).
(B) Flies expressing I31E/E22G Aβ₄₂ (blue line) show a dramatic increase in longevity (median survival = 27 ± 1 d, n = 600, p < 0.0001) compared to flies expressing E22G Aβ₄₂ (red line) (median survival = 8 ± 1 d, n = 100).
(C) Flies expressing the F20E Aβ₄₂ peptide (blue squares) have significantly improved locomotor ability (p < 0.001, n = 90 observations per line per time point) compared with flies expressing the WT Aβ₄₂ peptide (red circles).
(D) Flies expressing the I31E/E22G Aβ₄₂ peptide (blue squares) have significantly improved locomotor ability (p < 0.001, n = 90 observations per line per time point) compared to flies expressing E22G Aβ42 (red circles).
(E) Flies expressing E11G Aβ₄₂ (green line) die significantly quicker (median survival 21 ± 1, n = 500, p < 0.0001) than those expressing WT Aβ₄₂ (red line) (median survival 26 ± 1 d, n = 100).
(F) Flies expressing M35F Aβ₄₂ (green line) die significantly more quickly (median survival = 15 ± 1, n = 500, p < 0.0001) than those expressing WT Aβ₄₂ (red line) (median survival 26 ± 1 d, n = 100). Larger values of S_tox indicate higher toxicity

There is a significant correlation between neuronal dysfunction measured both by longevity (A) (S_tox; r = 0.79, p = 0.017) and mobility (B) (M_tox; r = 0.73, p = 0.03) and the rate of aggregation (K_agg) in vitro (measured by thioflavin T fluorescence).

(A) Z_tox predicts more accurately (r = 0.83, p < 0.00001) than Z_agg (Figure 2A) the relative longevity (S_tox) of flies expressing different Aβ₄₂ variants.
(B) Z_tox predicts more accurately (r = 0.75, p = 0.001) than Z_agg (Figure 2B) the relative locomotor ability (M_tox) of flies expressing different Aβ₄₂ variants.
The errors in S_tox measurements (y-axis) are standard errors of the mean arising from the average of the independent lines of flies tested for each variant. The errors in the predictions of protofibril formation propensity (Z_tox) are also shown (x-axis).

(A) There is a significant correlation between the propensities of the Aβ₄₂ variants to aggregate (Z_agg) and the relative survival of the flies (S_tox; r = 0.75, p = 0.001).
(B) There is a similarly significant correlation between the propensities of Aβ₄₂ variants to aggregate (Z_agg) and the locomotor abilities of the flies (M_tox; r = 0.65, p = 0.009).
In both panels the errors in the in vivo measurements (y-axis) are standard errors of the mean arising from the average of the independent lines tested for each variant. The errors in the predictions of aggregation propensity (Z_agg) are also shown (x-axis).

(A) F20E Aβ₄₂ (blue squares) aggregates more slowly than WT Aβ₄₂ (red circles), and both were found to have formed well-defined fibrils at the end point of this assay (Figure S3).
(B) Immunohistochemistry shows extensive Aβ₄₂ deposition (brown staining) in the brain of WT-Aβ₄₂--expressing flies at 20 d of age (arrows).
(C) In contrast, F20E Aβ₄₂ flies show no evidence of Aβ₄₂ deposition at 20 d of age.
(D) Both E22G (red circles) and I31E/E22G (blue squares) Aβ₄₂ aggregate at similar rates as measured by Thioflavin T fluorescence.
(E) Aβ₄₂ immunohistochemistry shows deposition throughout the cortex in the brain of E22G-Aβ₄₂-expressing flies at 8 d of age (arrows). This deposition is also associated with the appearance of vacuoles (asterisks).
(F) Flies expressing I31E/E22G Aβ₄₂ show extensive deposition of Aβ₄₂ throughout their cortex (arrows). In contrast to (E), no evidence of neurodegeneration (vacuolation) is seen.