N-terminal mutant huntingtin associates with mitochondria and impairs mitochondrial trafficking

Abstract

Huntington's disease (HD) is caused by polyglutamine (polyQ) expansion in huntingtin (htt), a large (350 kDa) protein that localizes predominantly to the cytoplasm. Proteolytic cleavage of mutant htt yields polyQ-containing N-terminal fragments that are prone to misfolding and aggregation. Disease progression in HD transgenic models correlates with age-related accumulation of soluble and aggregated forms of N-terminal mutant htt fragments, suggesting that multiple forms of mutant htt are involved in the selective neurodegeneration in HD. Although mitochondrial dysfunction is implicated in the pathogenesis of HD, it remains unclear which forms of cytoplasmic mutant htt associate with mitochondria to affect their function. Here we demonstrate that specific N-terminal mutant htt fragments associate with mitochondria in Hdh(CAG)150 knock-in mouse brain and that this association increases with age. The interaction between soluble N-terminal mutant htt and mitochondria interferes with the in vitro association of microtubule-based transport proteins with mitochondria. Mutant htt reduces the distribution and transport rate of mitochondria in the processes of cultured neuronal cells. Reduced ATP level was also found in the synaptosomal fraction isolated from Hdh(CAG)150 knock-in mouse brain. These findings suggest that specific N-terminal mutant htt fragments, before the formation of aggregates, can impair mitochondrial function directly and that this interaction may be a novel target for therapeutic strategies in HD.

Figure 1.

Mutant htt associates with mitochondria in HD knock-in mice. A, EM48 immunogold labeling of the cerebral cortex of HD knock-in brain tissue at 9 months of age showing the association of mutant htt with mitochondria (M; arrows). B, EM48 immunogold labeling of the cerebral cortex of wild-type mice at 12 month of age. N, Nucleus. Scale bars, 500 nm.

Figure 2.

Specific N-terminal mutant htt fragments are enriched in brain mitochondria from HD knock-in mice. A, Forebrains from wild-type (7Q/7Q) or HD (150Q/7Q) knock-in mice were fractioned to purify mitochondria on a Percoll-based gradient. Whole-cell (Whole), cytosolic (Cyto), or mitochondrial (Mito) fractions (20 μg/each) were immunoblotted for various subcellular markers to confirm the purity of the isolated mitochondrial fraction. All mitochondrial proteins (30 and 70 kDa SDH, MnSOD, and AAT) showed significant enrichment in mitochondrial fractions. The blots were also probed with antibodies for plasma membrane (Na/K-ATPase), cytosolic (α-tubulin, GAPDH, G6PD, and NeuN), nuclear (NeuN), ER (BiP), endosomal (EEA1), or synaptosomal (synaptophysin) proteins. B, Subsequent blotting for mutant htt and full-length htt demonstrated a preferential association of N-terminal mutant-htt fragments with brain mitochondria. Note that the 1C2 antibody only reacts with mutant htt, whereas the 2166 antibody detects the full-length forms of both wild-type and mutant htt (bottom). C, Comparison of the sizes of N-terminal htt fragments in mitochondria isolated from HD knock-in mice with mutant htt of known lengths.

Figure 3.

N-terminal fragments of wild-type htt associate more readily with mitochondria than full-length htt. A, B, Whole-cell and cytosolic (A) and mitochondrial (B) fractions from wild-type (7Q/7Q) and homozygous HD knock-in (150Q/150Q) mice at the ages of 3–25 months were blotted with 1C2 to specifically detect mutant htt followed by antibodies to either α-tubulin or the mitochondrial protein MnSOD. A nonspecific band (n.s) appears in both 7Q/7Q and 150Q/150Q samples. Longer exposure of blots containing mitochondrial protein revealed an age-dependent accumulation of specific intermediate-sized N-terminal mutant htt fragments on brain mitochondria (bracket). C, Western blotting of whole-cell (W), cytoplasmic (C), and mitochondrial (M) fractions of transfected HEK293 cells expressing full-length htt (23Q-FL and 120Q-FL) or N-terminal htt fragments (23Q-508 and 120Q-508). The blots were probed with EM48 for htt (top) and antibodies to α-tubulin (middle) and the mitochondrial protein MnSOD (bottom).

Figure 4.

Mutant htt affects the in vitro association of trafficking proteins with mitochondria. A, Cytosolic lysates from 6-month-old 7Q/7Q and 150Q/150Q mouse forebrains were incubated with Percoll-purified mitochondria from wild-type mouse brains (top panel). Mitochondria were pelleted, washed repeatedly, and blotted with antibodies to trafficking proteins (P150, KLC, KHC, and HAP1). The mitochondrial protein (SDH) was also detected as a loading control. B, Relative binding of trafficking proteins to mitochondria in the presence of mutant versus wild-type htt was quantified. Blots probed with 1C2 (A, bottom) show that endogenously expressed N-terminal mutant htt fragments associate with brain mitochondria in this in vitro assay. Preferential binding of intermediate-length mutant htt fragments (arrow) is also confirmed in this assay.

Figure 5.

Soluble N-terminal mutant htt impairs mitochondrial trafficking in neurons. A, B, Representative images of DsRed2Mito and GFP-htt expression in striatal neurons transiently transfected with GFP-23Q-208 (A) or GFP-130Q-208 (B). Only neurons showing diffuse htt without visible aggregates were selected for analysis of mitochondrial transport. C, D, White framed rectangles highlight the region presented in the time-series panels below (C, D) and in the full 5 min sequences (supplemental Movies 1, 2, available at www.jneurosci.org as supplemental material). Both anterogradely (white arrowheads) and retrogradely (yellow arrows) moving mitochondria tended to travel longer distances over shorter periods of time in neurons expressing wild-type N-terminal htt fragments compared with those in neurons expressing soluble mutant htt N-terminal fragment. Scale bars, 5 μm.

Figure 6.

Mutant htt reduces the velocity of moving mitochondria in the processes of cultured striatal neurons. A, B, The velocity of moving mitochondria in the processes of cultured striatal neurons that were transfected with GFP-htt-208 (A) or N-terminal htt-508 (B). C, The velocity of moving mitochondria in the processes of cultured striatal neurons of wild-type (7Q/7Q), heterozygous (150Q/7Q), and homozygous (150Q/150Q) HD knock-in mice. *p < 0.05 for 150Q/150Q versus 7Q/7Q or 150Q/7Q. Complete analysis of trafficking parameters is included in supplemental Table 1 (available at www.jneurosci.org as supplemental material). D, 1C2 immunoblotting analysis revealed that higher levels of N-terminal mutant htt fragments (inner bracket) are present in cultured 150Q/150Q neurons than in 150Q/7Q neurons. A nonspecific band (n.s.) in 7Q/7Q cultures indicates equivalent protein loading for each sample.

Figure 7.

Mutant N-terminal htt fragments decrease the distribution of mitochondria in the processes of cultured neurons. Fluorescent images of cultured rat brain striatal neurons that expressed transfected htt (23Q-508 or 120Q-508, green) and DsRed2Mito (red). Arrowheads indicate neuronal processes. Quantitative assessment of the number of DsRed-labeled mitochondria per 10 μm and their occupancy (percentage) in the measured processes is described in the text.

Figure 8.

HD mouse brains show degenerated mitochondria in nerve terminals and decreased ATP level in synaptosomal fraction. A, EM48 immunogold labeling of the cerebral cortex of HD knock-in mice at the age of 12 months. Single arrows indicate synapses. Double arrows indicate degenerated, mitochondria-like organelles in axonal (left) and dendritic (right) terminals. Mutant htt is labeled by dark immunogold particles. Scale bars, 250 nm. B, ATP levels of cytoplasmic and synaptosomal fractions isolated from the brain cortex of wild-type (7Q/7Q) and homozygous HD knock-in (150Q/150Q) mice at 4 and 14 months of age. *p < 0.05 compared with 7Q/7Q.