Impaired XK recycling for importing manganese underlies striatal vulnerability in Huntington's disease

Abstract

Mutant huntingtin, which causes Huntington's disease (HD), is ubiquitously expressed but induces preferential loss of striatal neurons by unclear mechanisms. Rab11 dysfunction mediates homeostatic disturbance of HD neurons. Here, we report that Rab11 dysfunction also underscores the striatal vulnerability in HD. We profiled the proteome of Rab11-positive endosomes of HD-vulnerable striatal cells to look for protein(s) linking Rab11 dysfunction to striatal vulnerability in HD and found XK, which triggers the selective death of striatal neurons in McLeod syndrome. XK was trafficked together with Rab11 and was diminished on the surface of immortalized HD striatal cells and striatal neurons in HD mouse brains. We found that XK participated in transporting manganese, an essential trace metal depleted in HD brains. Introducing dominantly active Rab11 into HD striatal cells improved XK dynamics and increased manganese accumulation in an XK-dependent manner. Our study suggests that impaired Rab11-based recycling of XK onto cell surfaces for importing manganese is a driver of striatal dysfunction in Huntington's disease.

Figure 1.

Proteomic analysis of Rab11 endosomes uncovers XK. (A) Schematic representation of two-step isolation of Rab11 endosomes. STHdhQ7/Q7 cells were transfected with pcDNA_3.1-6xHis-Rab11 to label endosomes. Homogenates were prepared using a ball-bearing homogenizer and centrifuged to remove nuclei and then incubated with Nickel resins. Eluates from Nickel resins were overlaid on a sucrose gradient and centrifuged to obtain membrane-bound His-Rab11 in pellets (endosomes). (B) Western blot analysis of eluates from nickel resins and ultracentrifuge pellets (membrane-bound His-Rab11) with indicated antibodies. Shown are blot analyses from one of the four individual endosomal preparations. The open arrowhead indicates a protein likely to be an isoform of XK, as signals for this protein were very low at the basal state and appeared to increase when different amounts of XK-expressing plasmids were transfected (Fig. S1). Star symbols indicate signals for exogenous and endogenous Rab11, which were still persistent during the incubation of the blots with anti-EGFP and secondary antibodies. (C) Membrane-bound His-Rab11 in ultracentrifugation pellets were resuspended in glutaraldehyde and dropped onto grids for electron microscopic analysis. Images in both panels showed profiles of tubulovesicular clusters, which are reminiscent of recycling endosomes in cells. Shown are electron microscopic images from one of four individual endosomal preparations.(D) Examples of known Rab11 interactors and/or known cargo proteins identified in the proteomic analysis of isolated Rab11 endosomes are shown. Shown are the proteins identified in all four endosomal preparations. Source data are available for this figure: SourceData F1.

Figure S1.

Characterization of anti-XK antibodies and the XK-EGFP reporter. (A) Western blot analysis of lysates of cells transfected with the indicated amounts of plasmids expressing WT XK. (B) Densitometry of blot analyses in A showed that signals for the protein band of ∼43 kD identified with an arrow increased in a plasmid dose-dependent manner. Signals for the protein band of ∼52 kD indicated by an open arrowhead were very weak at basal state, but also appeared to increase in a plasmid dose-dependent manner. Signals for the protein bands of 90 and 100 kD marked by a star symbol remained constant regardless of the amount of plasmids transfected, suggesting that they are crossreactive. (C) Western blot analysis of cytosol and total membranes prepared from STHdhQ7/Q7 cells. The protein bands of 43 and 52 kD, respectively, were present in total membranes, whereas the protein bands of 90 and 100 kD were cytosolic, further supporting their crossreactivity to XK antibodies. (D) Schematic representation of the XK-EGFP reporter. (E) STHdhQ7/Q7 cells were transfected with pcDNA3.1-XK for 16 h and further cultured in the presence or absence of β-cycloheximide (CYX) for 5 h. Cells were collected for preparing homogenates by passing through a 25-gauge needle. Postnuclear supernatants were overlaid on a discontinuous Nycodenz gradient and centrifuged as in Materials and methods. The same volume of each fraction was analyzed by Western blot. Protein bands identified by fixed arrowheads were detected by other antibodies left over in the XK antibody solutions which were reused. XK-EGFP was expressed at the expected size. The open arrowhead indicated the 52 kD isoform of XK. Source data are available for this figure: SourceData FS1.

Figure 2.

Rab11 regulates XK trafficking. (A) Colocalization of XK-EGFP with mCherry-Rab11 in striatal cells. SThdhQ7/Q7 cells were transfected with plasmids expressing XK-EGFP and with plasmids expressing dsRed-Rab4, dsRed-Rab5, dsRed-Vps35, or mCherry-Rab11. After treatment with β-cycloheximide, cells were processed for fluorescence microscopy. Shown are representative confocal images, which were captured individually for each channel and merged. Yellow structures indicate where the colocalization of XK-EGFP with mCherry-fused endosomal marker proteins took place. Scale bars: 10 μm (merge) and 2 μm (inset). (B) Pearson’s coefficient of colocalization. Digital images were analyzed with the JACoP plugin of the NIH ImageJ Fiji. Each symbol represents one cell. Data are mean ± SD. (C) Western blot of biotinylated proteins and postnuclear supernatants of SThdhQ7/Q7 cells transfected with plasmids expressing dNrab11 or dArab11 or empty vectors. Shown are blot analyses from one of the three individual experiments. (D) Densitometry for blot analyses in C. Data are mean ± SD. One-way ANOVA and post hoc Tukey’s analysis: F_(2,6) = 44.26, P < 0.001; Tukey’s test * P < 0.01, # P < 0.001. (E) Effects of dominant active and inactive mutants of Rab11 on the subcellular distribution of XK-EGFP. STHdhQ7/Q7 cells were transfected with plasmids expressing XK-EGFP along with plasmids expressing mCherry-fused dArab11 or dNrab11. Confocal images showed that XK-EGFP and mCherry-dArab11 colocalized at small vesicular structures distributed throughout the cytoplasm and clustered at perinuclear regions, whereas the colocalization of XK-EGFP and mCherry-dNrab11 occurred at large tubulovesicular structures. Scale bars: 20 μm (merge) and 3 μm (inset). Source data are available for this figure: SourceData F2.

Figure 3.

Dynamics of Rab11 endosomes in HD striatal cells declines and is enhanced upon expression of dominantly active Rab11. Time-lapse cell imaging of SThdhQ7/Q7 cells expressing XK-EGFP and mCherry-Rab11, SThdhQ111/Q111 expressing XK-EGFP and mCherry-Rab11, and SThdhQ111/Q111 expressing XK-EGFP and mCherry-dArab11. Prior to imaging and during the whole period of imaging, cells were treated with β-cycloheximide to ascertain the detection of XK-EGFP in the endocytic pathway. A series of six consecutive frames were chosen from Videos 1, 3, and 5, respectively, to highlight cotrafficking of XK-EGFP with mCherry-Rab11 or with mCherry-dArab11 in motile vesicles and the dynamics of large tubulovesicular structures labeled with XK-EGFP and mCherry-Rab11 or with XK-EGFP and mCherry-dArab11. Boxed regions were enlarged and shown below the corresponding frame. Arrows in insets trace motile structures containing both XK-EGFP and mCherry-Rab11/dArab11, whereas arrowheads point to structures changing in their size. Dashed circles indicate motile structures containing both XK-EGFP and mCherry-Rab11/dArab11 disappearing in the following images, and dashed polygons identify those appearing in the following images. Enlarged dashed contours indicate dynamic changes in the morphology of large vesiculotubular structures, likely reflecting events of vesicle fusion and budding. Scale bars in the last frame of each of SThdhQ7/Q7, SThdhQ111/Q111, and SThdhQ111/Q111 + dArab11: 10 μm (upper), 2 μm (middle), and 0.5 μm (lower). Source data are available for this figure: SourceData F3.

Figure S2.

Another set of cells to illustrate that the dynamics of Rab11 endosomes in HD striatal cells declines and is enhanced upon the expression of dominantly active Rab11, as in Fig. 3. Boxed regions were enlarged and shown below the corresponding frame. Arrows in Insets trace motile structures containing both XK-EGFP and mCherry-Rab11/dArab11, whereas arrowheads point to structures changing in their size. Dashed circles indicate motile structures containing both XK-EGFP and mCherry-Rab11/dArab11 disappearing in the following images, and dashed polygons identify those appearing in the following images. Enlarged dashed contours indicate dynamic changes in the morphology of large tubulovesicular structures, likely reflecting events of vesicle fusion and budding. Scale bars in the last frame of each of SThdhQ7/Q7, SThdhQ111/Q111, and SThdhQ111/Q111 + dArab11: 10 μm (upper), 2 μm (middle), and 0.5 μm (lower).

Figure 4.

Defective trafficking of XK onto cell surfaces in HD cells. (A) Confocal images of SThdhQ7/Q7 and SThdhQ111/Q111 cells transiently transfected with plasmids expressing XK-pHluorin showed that pHluorin signals occurred at cell surfaces and intracellular compartments. Arrows point to cells with XK-pHluorin signals concentrated in the perinuclear regions. (B) Densitometry of pHluorin signals on the sharp edges of each of SThdhQ7/Q7 and SThdhQ111/Q111 cells showed a decline in the expression of XK-pHluorin on the surface of HD striatal cells. (C) Western blot analysis of biotinylated proteins and postnuclear supernatants of SThdhQ7/Q7 and SThdhQ111/Q111 cells with indicated antibodies. Shown are blot analyses from one of the three experiments. (D and E) Densitometry of signals of XK in postnuclear supernatants (total, D) and biotinylated XK (cell surfaces, E) for blot analyses in (C). (F) Images of XK staining in a brain section of WT and HD140Q/140Q mice. A series of consecutive coronal brain sections cut through the striatum were processed for labeling with antibodies for XK, with detergents omitted in all solutions to ensure the integrity of the plasma membrane. Images were captured from the nucleus accumubens of each brain section with the same settings. (G and H) Digital images from three brain sections for each of 3 WT and 3 HD mice were analyzed with the NIH ImageJ Fiji software to measure the cross-sectional area (G) and signal intensity (H) of striatal neurons immunoreactive to the XK antibody. Arrowheads in F indicate examples of neurons applied for densitometry and for measuring their cross-section areas. Scale bars: 20 μm (A), 100 μm (F), and 5 μm (inset in F). Data are mean ± SD. Each symbol in B, G, and H represents one cell and in D and E represents one experiment. Two-tailed Student’s t test. Source data are available for this figure: SourceData F4.

Figure S3.

XK is expressed normally in the striatum but is diminished on the cell surface of striatal neurons in CAG140 knock-in mice. (A and B) Postnuclear striatal and cortical supernatants of WT (N = 7) and HD (N = 6) mice were analyzed by Western blot (A) followed by densitometry (B) to examine expression levels of XK. The age of the mice was 10 mo. (C–E) A series of three consecutive coronal brain sections cut through the striatum of WT and HD140Q/140Q mice were processed for labeling with antibodies for XK with the same procedures as in Fig. 4. Immunolabeled XK molecules were detected by the avidin–biotin peroxidase method. (C) Images of immunolabeled XK in one brain section of one animal for WT and HD. Boxed regions were enlarged and shown below the corresponding photograph. (D and E) Digital images captured from the striatum of four brain sections for each genotype were analyzed with the NIH ImageJ/Fiji software to measure the cross-sectional areas (D) and signal intensity (E) of striatal neurons immunoreactive to the XK antibody. Scale bars: 150 μm. Intensities of XK, as well as GAPDH (A), immunoreactive signals were measured with the NIH ImageJ/Fiji software. Data are mean ± SD. Each symbol represents one animal (A) and one cell (D and E), respectively. Two-tailed Student’s t test was done for comparison. Source data are available for this figure: SourceData FS3.

Figure S4.

Analysis of the feasibility for XK to act as a manganese transporter. (A) Metal ion interaction analysis predicted 15 sites to which a Mn²⁺ ion could bind in XK. Arrows point to two predicted sites exposed to external environments or the pore, and their corresponding residues were also displayed. (B) The simulated trajectory of the Mn²⁺ movement through the pore formed by the helices of XK. At the beginning of the simulation, the Mn²⁺ ions were clustered in the same place. At 10 ns, Mn²⁺ binding to the predicted site (230t-232V-233L) triggered the pore to change from a compact closed conformation to an open basket shape, which facilitated the inward movement of the Mn²⁺ ion. Once the channel was in the open state, the Mn²⁺ inward movement might not be driven by the electrostatic interaction with the binding sites but be driven by a concentration gradient. The ribbon structure was rotated at 30, 40, and 50 ns, respectively, to illustrate that the Mn²⁺ ion moved through the pore. Arrows in the ribbon structures point to the Mn²⁺ ion located inside the pore during the simulation.

Figure S5.

Effects of altered levels of XK on manganese accumulation in HD striatal cells. (A) Deficiency of Mn in STHdhQ111/Q111 cells. At basal and Mn-triggered states, STHdhQ111/Q111 cells enriched less Mn than STHdhQ7/Q7 cells, suggesting that HD striatal cells have a deficit in taking up Mn. (B and C) Ectopic expression of XK significantly improved STHdhQ111/Q111 cells to accumulate Mn. (D and E) However, XK-siRNA treatment had little effect on Mn accumulation in STHdhQ111/Q111 cells, though XK was efficiently lowered down upon transfection with XK-specific siRNA. Blot analyses in B and D were from one of three experiments. Data in C and E are mean ± SD. One-way ANOVA and post hoc Tukey’s analysis: F_(3,8) = 1,083, P < 0.0001 (A); F_(3,8) = 55.8, P < 0.001 (C); F_(3,8) = 2.495, P = 0.134 (E); Tukey’s test: # P < 0.0001. Source data are available for this figure: SourceData FS5.

Figure 5.

XK is involved in transporting manganese. (A–D) Mn accumulation in SThdhQ7/Q7 cells was elevated and reduced upon ectopic expression (A and B) and knockdown (C and D) of XK, respectively. SThdhQ7/Q7 cells were transfected with pcDNA3.1-XK or empty vector (Mock) or XK-specific or Scramble siRNA. After exposure to Mn, cells were harvested for verifying changes in expression levels of XK by Western blot analysis (A and C) and for measuring the contents of cellular Mn (B and D). Blot analyses shown in A and C were obtained from one of the three experiments. (E and F) Expression of dArab11 in SThdhQ111/Q111 cells promotes Mn accumulation in an XK-dependent manner. SThdhQ111/Q111 cells were transfected with plasmids expressing dsRed-dArab11 along with XK-specific or Scramble siRNA. After Mn exposure, cells were collected for Western blot analysis (E) and measurement of cellular Mn (F). Data in B, D, and F are mean ± SD. One-way ANOVA and post hoc Tukey’s analysis: F_(3,8) = 619.5, P < 0.0001 (B), F_(3,8) = 81.47, P < 0.0001 (D), F_(5,12) = 144, P < 0.0001 (F); Tukey’s test: # P < 0.0001. Source data are available for this figure: SourceData F5.

Figure 6.

A model of selective degeneration of striatal neurons in HD. The idea that Rab11 activation is achieved by a protein complex composed of HTT, Kalirin, and TRAPPII was based on our recent studies (Ke et al., 2020; McClory et al., 2018; Wang et al., 2020). Mutant HTT compromises the Kalirin–TRAPPII complex in activating Rab11 on endosomal membranes (1), thereby impeding XK recycling back onto neuronal surfaces (2) and reducing XK-mediated import of Mn into neurons (3). Constant deficiency of Mn along with the chronic decline of functions of other Rab11-regulated trafficking proteins render the dysfunction of striatal neurons (4) and eventually leads to degeneration of striatal neurons and atrophy of the striatum (5).