Wild-type BICD2 interactome

(A). Schematic of human BICD2, indicating the three coiled-coil domains and sites of interaction with dynein/dynactin, kinesin-1, and cargo.

(B). Model of BICD2 binding to dynein/dynactin in a cargo dependent manner.

(C, C’). A volcano plot indicating proteins that were enriched with BICD2-mTrbo in comparison to RFP-mTrbo. The dashed line along the x-axis indicates a fold enrichment of 2, whereas the dashed line along the y-axis indicates a p value of 0.05. Specific interacting proteins were considered those that were enriched at least two-fold with BICD2-mTrbo and had a p value of at least 0.05. These candidates are shown in the zoomed in image in C’.

(D). Known BICD2 interacting proteins, components of the dynein motor, and components of the HOPS complex that were specifically enriched with BICD2-mTrbo are indicated.

(E). A cellular component GO analysis of the BICD2 interactome.

BICD2 interacts with components of the HOPS complex in vivo.

(A). A co-immunoprecipitation experiment was performed with HEK cells expressing the indicated constructs. The lysates were incubated with V5 trap beads to precipitate the tagged proteins. The co-precipitating proteins were analyzed using western blotting with the indicated antibodies. VPS41 specifically co-precipitates with BICD2_wt. Minimal binding was observed with RFP-mTrbo or a BICD2 construct lacking the cargo binding domain.

(B). A similar co-precipitation experiment was set up as in panel A. The co-precipitating proteins were analyzed by blotting using antibodies against VPS16, VPS18 and the V5 epitope. VPS16 and VPS18 co-precipitate specifically with BICD2_wt-mTrbo.

(C). A schematic of the BICD2 constructs used in the binding experiment.

(D). Cells expressing either RFP-mTrbo, BICD2_delCC3-mTrbo (lacking the cargo binding domain), BICD2_wt-mTrbo (full length BICD2) or BICD2_CC3-mTrbo (just the cargo binding domain) were used to examine the interaction with VPS41. Lysates were incubated with streptavidin beads to precipitate biotinylated proteins. The precipitated proteins were analyzed by blotting using the indicated antibodies. Although RANBP2 interacts with the isolated cargo binding domain of BICD2, VPS41 does not.

Role of BICD2 in localization of GFP-VPS41 and LAMP1 vesicles.

(A-C). HeLa cells were transfected with either a control-siRNA (A), an siRNA targeting dynein heavy chain (B), or an siRNA targeting BICD2 (C). Two days after the siRNA transfection, the cells were transfected with a plasmid encoding GFP-VPS41. The next day, the cells were fixed and processed for immunofluorescence using an antibody against GFP. The cells were counterstained with DAPI (cyan) and Phalloidin (grey). Depletion of DHC results in an outward spreading of GFP-VPS41 vesicles, whereas depletion of BICD2 results in more perinuclear clustered vesicles.

(D, E). The distance of GPF-VPS41 vesicles relative to the nucleus was determined and plotted. Vesicles present within 10 microns of the nucleus are shown in D, and those present at a distance greater than 10 microns are shown in panel E.

(F-H). HeLa cells were transfected with the indicated siRNAs. Three days later, the cells were fixed and processed for immunofluorescence using an antibody against LAMP1. As with GFP-VPS41 vesicles, depletion of DHC resulted in peripheral vesicles, whereas depletion of BICD2 resulted in perinuclear clustering of LAMP1 vesicles.

(I-K). HeLa cells were transfected with either a control siRNA (I, J) or an siRNA targeting KIF5B

(K). Two days later the cells were transfected with a plasmid encoding either GFP (I) or BICD2-mNeon (J, K). The cells were fixed on day 4 and processed for immunofluorescence using an antibody against LAMP1. The cells were counterstained with DAPI. Over-expression of BICD2 results in the peripheral spreading of LAMP1 vesicles. This phenotype was reversed upon knocking down KIF5B.

(L-M). The distance of LAMP1 vesicles relative to the nucleus was determined and plotted. Vesicles present within 10 microns of the nucleus are shown in L, and those present at a distance greater than 10 microns are shown in panel M.

The signal for GFP-VPS41 and LAMP1 is displayed using the “red hot” LUT in FIJI. The scale bar is 20 microns. A one-way ANOVA was used for the quantifications shown in panel D, E, L and M with the values compared to the mean of BICD2_wt. ns = not significant, *, p ≤ 0.05, ***, p<0.001.

BICD2 cargo binding domain mutants hyperactivate dynein.

(A). A co-immunoprecipitation experiment was performed using HEK cells expressing the indicated constructs. The tagged proteins were purified using V5 trap beads and the co-precipitating proteins were analyzed by western blotting with the indicated antibodies. A greater amount of DIC and DCTN1 co-purified with mutant BICD2 compared to the wild-type protein.

(B-E). HeLa cells expressing BICD2_wt (B), BICD2_N188T (C), BICD2_R694C (D) or BICD2_R747C (E) were fixed and processed for immunofluorescence using antibodies against V5 (cyan) and pericentrin (magenta). Merged images are also shown. All three mutants displayed a centrosomal localization pattern to varying degrees. The scale bar is 20 microns.

(F). The centrosomal enrichment of BICD2_wt or mutant was quantified.

(G). Schematic of the peroxisome tethering assay.

(H). The average distance of peroxisomes to the nucleus was calculated on a cell-by-cell basis. In comparison to wild-type BICD2, all three mutants showed increased clustering of peroxisomes close to the nucleus. A one-way ANOVA was used for the quantifications shown in panels F and H with the values compared to the mean of BICD2_wt. n = not significant, **, p<0.01, ***, p<0.001, ****, p<0001.

Localization of BICD2 wild-type and mutants in neurons.

(A-D). E18 rat hippocampal neurons were transfected with the indicated constructs. Two days after transfection, the cells were fixed and processed for immunofluorescence using a V5 antibody. The axon outline for cells expressing BICD2_R694C and BICD2_R747C are indicated. Signal for wild-type BICD2 could be detected in the cell body and axon. A similar phenotype was noted for BICD2_N188T. By contrast, BICD2_R694C and BICD2_R747C displayed reduced axonal signal. The scale bar is 100 microns. The signal for BICD2 is displayed using the “red hot” LUT in FIJI.

(E). Quantification of the cell body enrichment of BICD2 wild-type and mutant.

(F). The axon length of neurons expressing either wild-type or mutant alleles of BICD2 was quantified. Expression of BICD2 mutants correlated with shorter axonal lengths. A one-way ANOVA was used for the quantifications shown in panels E and F and the values were compared to the mean of BICD2_wt. n = not significant, **, p<0.01, ***, p<0.001, ****, p<0.0001.

BICD2 mutations are associated with altered interactomes.

(A-C). Volcano plots comparing the interactome of BICD2_wt vs BICD2_N188T (A), vs BICD2_R694C (B) and vs BICD2_R747C (C). Interacting proteins that show at least a twofold change in comparison to BICD2_wt and have a p value of at least 0.05 are indicated in the shaded boxes. Red boxes indicate proteins that display a greater interaction with BICD2 mutants vs the wild-type, whereas blue boxes indicate proteins that display a lower interaction vs the wild-type protein.

(D) The proteomics results were validated by repeating the experiment and analyzing the bound fractions using the indicated antibodies. Streptavidin beads were used to purify the biotinylated proteins.

(E-G). Quantification of binding of BICD2_wt and mutants with RANBP2 (E), VPS41 (F), and CSPP1 (G). The level of binding for the mutants was compared to BICD2_wt. Consistent with the proteomics results, BICD2_R747C displayed reduced binding to RANBP2 and VPS41. All three mutants bound CSPP1 at a greater level than the wild-type. A one-way ANOVA was used for this analysis. ns = not significant, *, p ≤ 0.05, **, p<0.01, ***, p<0.001, ****, p<0.0001.

BICD2_R747C is associated with a gain of function interaction with GRAMD1A.

(A). Lysates from cells expressing BICD2_wt and mutants were incubated with streptavidin beads to purify biotinylated proteins. Bound proteins were eluted and analyzed by blotting using the indicated antibodies. BICD2_R747C interacted with substantially more GRAMD1A than either the control, BICD2_wt, or the other mutants.

(B). Quantification of binding of BICD2_wt and mutants with GRAMD1A. The level of binding for the mutants was normalized to BICD2_wt. A one-way ANOVA was used for this analysis. ns = not significant, ***, p<0.001.

(C). Cos7 cells were co-transfected with constructs expressing either BICD2_wt or mutant (magenta) along with a plasmid expressing GRAMD1A-mScarlet3 (cyan). Except for cells expressing BICD2_R747C, GRAMD1A was localized to the ER. By contrast, GRAMD1A was highly enriched at the centrosome in cells expressing BICD2_R747C.

(D). Quantification of the co-localization between BICD2_wt and mutants with GRAMD1A. A one-way ANOVA was used for this analysis and the values were compared to the mean of BICD2_wt. ns = not significant, ****, p<0.0001.