A, Post-debris supernatants and total membranes prepared from striata of young (10 – 14 weeks old) WT and HD mice as described in Methods were analyzed by SDS-PAGE and Western blot analysis with indicated antibodies. Shown are a series of blot analyses. B, Quantitative data for A. Films were scanned and the band density was measured using NIH ImageJ. The density of calnexin was used for normalization of loading variations. Data are represented as ratio of Rab11 on HD striatal membranes to Rab11 on WT striatal membranes (n=6 mice for each of WT and HD, Mean±SD, Student t-test: * p<0.01). C, RabGDI-mediated extraction of Rab11 was performed as described in Methods. In this assay we used 20μg HD membranes and 60μg WT membranes to keep the same level of Rab11GDP to be retrieved by RabGDI. The extracted Rab11 (released) in complex with GST-RabGDI on glutathione and the un-extracted Rab11 (unreleased) in the supernatant was precipitated with chloroform/methanol and was boiled in SDS-PAGE sample buffer. Samples were analyzed by SDS-PAGE and Western blot with anti-Rab11 and anti-GDI antibodies. Shown are blot analyses from one of the three experiments. D, Densitometry data obtained in C were used to calculate the percentage of extracted Rab11 by GST-RabGDI by dividing the extracted amount from the total input (released + unreleased). Average percentage of released Rab11 by GST-RabGDI was plotted (n=3, Mean±SD, Student t-test: * p<0.001).

A, Primary neurons on coverslips were cultured in PBS for 1hr. After two washes in cold PBS and incubation in cold PBS on ice for 15min, primary neurons on coverslips were incubated with 5μg ml^-1 Alexa568-transferrin on ice for 30min. After two washes in cold PBS, coverslips were incubated in pre-warmed culture medium containing 10μg ml^-1 non-labeled transferrin at 37°C for 2min or 15min and fixed for analysis. Shown are confocal images. Nuclear stain with Hoechst 33258 is in blue. Alexa568-transferrin is in red. Fewer neurons appear in the HD cultures than in the WT cultures. Arrowheads show neurons with strong Alexa568-transferrin signals. Boxed regions are shown at higher magnification at lower left of each panel. Arrows identify endosomes or transport intermediates labeled with Alexa568-transferrin. Scale bar: 20 μm. B, Signal intensity for Alexa568-transferrin in soma of neurons and the corresponding background levels were measured using NIH ImageJ. Data are represented as mean signal intensity of Alexa568-transferrin in the soma region per neuron (Mean±SD, Student t-test; 2min chase, n=42 neurons for each of WT and HD^140Q/140Q, n/s: no significance; 15min chase, n=42 neurons for each of WT and HD^140Q/140Q, * p<0.001).

Left panel depicts the two cycles of Rab11 that are required for its proper function (normal). Rab11 associates and dissociates with membranes and alternates between GDP-bound inactive and GTP-bound active states. In the cytoplasm, GDP-bound Rab11 partners with Rab-GDP dissociation inhibitor (GDI). To fulfill functions, Rab11-GDP is recruited onto endosomal membranes for activation by a guanine nucleotide exchange factor (GEF) and conversion into GTP-bound form. After fulfilling the function by recruiting a cohort of effectors, Rab11-GTP is inactivated by a GTPase-activating protein (GAP) and returns back to GDP-bound form that is extracted by RabGDI. Right panel shows a proposed model for Rab11 dysfunction in HD. The key event causing neurodegeneration is a deficient nucleotide conversion from Rab11GDP to Rab11GTP or deficient activation of Rab11 (indicated by long dashed arrow), which occurs on endosomal membranes. Another problem in HD endosomal membranes is the accumulation of inactive GDP-bound form of Rab11 (enlarged oval) caused in part by reduced extraction of RabGDP by RabGDI (indicated by short dashed arrow). The role of the accumulation of Rab11GDP to neuronal dysfunction is unclear.

A and B, 200μg of striatal total membranes prepared from freshly sacrificed WT and homozygous HD mice (see Methods) were used as the source of GEFs for catalyzing [³H]GDP release from [³H]GDP-GST-Rab11 (A) or [³H]GDP-6×His-Rab5 (B) for indicated times. Data were represented as the percentage of [³H]GDP released from GST-Rab11 or 6×His-Rab5 after incubation with striatal total membranes (n=3 for each time point in both A and B, Mean±SD, 2-tailed Student t-test: * p<0.01). C, Rab11-enriched membrane fractions (fraction-7 for HD and fractions-6 and 7 for WT) were prepared as described in Methods. 50μg of HD and WT mouse brain Rab11-enriched membranes were used to catalyze [³H]GDP release from GST-Rab11 for indicated times. Data are represented as the percentage of [³H]GDP released from GST-Rab11 after incubation with Rab11-enriched membrane fractions (n=3 for each time point, Mean±SD, 2-tailed Student t-test: * p<0.05).

A, AAV expressing EGFP or HA-tagged Rab11S25N were unilaterally injected into the striatum and overlying cortex. Beam walking was tested at indicated dates post-injection as described in Materials and Methods. The time for mice to cross the beam was measured pre- and post-injection of virus. Data were graphed as the average increase in time at each post-injection period compared to pre-injection time (n=6, Mean±SD, Student t-test: * p<0.01). B, Shown are representative vibratome cut brain sections at the level of the striatum from two of the mice reported in (A). Brain slices were stained with anti-HA antibody or anti-GFP antibody using the immunoperoxidase method. Arrow points to the side injected with AAV-HARab11S25N or AAV-EGFP. C, Expression of HA Rab11S25N led to neuropathological changes. Shown are representative examples of neurons expressing HA-tagged Rab11S25N in the striatum and cortex that appear shrunken (arrowhead) or have a retracted proximal dendrite (arrow). Scale bar: 5μm.

Glutamate-induced cell death in primary cortical neurons. Primary WT and HD cortical neurons were infected with lentivirus or left alone for seven days and challenged with 0.2mM glutamate for 20min. After washout of glutamate, primary neurons were cultured in fresh media for 24hr. Cell death was determined by trypan blue uptake. Data in bar graph are represented as the percentage of trypan blue positive neurons from four independent preparations of neurons (n=462 for WT and 285 HD neurons for no-virus; n= 403 for WT and 349 for HD neurons infected with virus-EGFP; n= 432 for WT and 409 for HD neurons infected with virus-Rab11Q70L; Chi square: # p<0.0001; * p<0.000001).