Fig 2: Exogenous KMO expression and localisation in HEK293T cells. (A) Schematic representation of KMO ICC constructs: flKMO-RFP and tKMO-RFP. Grey boxes show putative transmembrane domains (TM). tKMO has a deletion of amino acids from position 368 to 380 that resembles the deletion in KMO isoform 2, but is also truncated at its C-terminus, with 67 aa missing from the protein. (B) Immunoblotting of KMO constructs using anti-KMO antibody (10698-1-AP). (C,D) HEK293T cells were transfected with the flKMO-RFP (C), Left panel)) or tKMO-RFP ((D), Left panel) constructs and fixed 24 h after transfection; ((C,D) middle panels): immunolabelling for the mitochondrial protein HtrA2 using anti-HtrA2 antibody (AF1458) (Alexa Fluor 488); ((C,D) right panels): merge of the RFP and anti HtrA2 signal. Nuclei were stained with Hoechst 33342. Scale bar = 8 µm. flKMO-RFP depicts mitochondrial localisation (punctate structures), whereas the tKMO-RFP signal is diffuse throughout the cell. The squares on the images indicate the selected areas for co-localisation analysis (an enlarged view of the selected area is shown on the side of each merge panel). (E,F) Co-localisation analysis of optical z-sections from eight deconvolved confocal images, using JACoP in ImageJ. (E) Pearson’s coefficient shows a significant difference in the mitochondrial co-localisation of flKMO-RFP and tKMO-RFP. (F) Mander’s coefficient correlation. M1 represents the red signal overlapping the green signal, while M2 indicates the green signal overlapping the red signal. **** (p < 0.0001), for unpaired t-test. ns = not significant. Data are expressed as mean ± SEM (n = 8).

Fig 3: Cellular localisation of BiFC complexes in live HEK293T cells, using confocal microscopy. Cells were seeded on ibiTreat dishes and co-transfected with BiFC constructs for 48 h. (A–C) illustrate the HTT signal using the BiFC system (BiFC signal = left image, and internal control RFP signal = right image) in each panel. (A) Cells were transfected with 19Q-VN, 25Q-VC and RFP, the BiFC signal is cytosolic and slightly punctate. (B) Cells were transfected with 97Q pair and RFP; the BiFC signal is generally cytosolic, with HTT inclusions present. (C) Cells were transfected with flKMO-CC, 97Q-VN and RFP; the BiFC signal is mainly mitochondrial as suggested by the dotted appearance of the signal. Scale bar = 8 µm. (D–F) Localisation of KMO-BiFC complexes. Left panels show BiFC signal of the following pairs: (D) flKMO-CC and HTT19Q-VN, (E) flKMO-CC and HTT46Q-VN, and (F) flKMO-CC and HTT97Q-VN. Second column of panels (D–F): mitochondria stained with MitoTracker Red CMXRox (M-7512). Third column of panels (D–F): merge of the BiFC signal and the MitoTracker signal. Scale bar = 8 µm. The BiFC signal in panels (D–F) exhibits dotted structures that co-localise with the MitoTracker signal, as seen in the merge images in the right panels of (D–F). This confirms the mitochondrial localisation of all the BiFC complexes of flKMO-CC with different polyQ lengths of HTT-VN.

Fig 4: Electron micrographs of dual immunogold labelling in transfected HEK293T cells for 48 h. Cells were co-transfected and co-probed with anti-KMO antibody (10698-1-AP) and anti-HTT (mEM48) antibody (MAB5374), followed by 30 and 15 nm gold conjugate secondary antibodies, respectively. (A,C) Overview of dual labelling of flKMO-CC and 19Q-VN, respectively. Scale bar = 1 µm. (B,D) Zoomed view of the regions indicated by the black box in (A) and (C), respectively. Mitochondria are all intensely labelled with HTT (15 nm particles), and some particles are seen in the cytoplasm. flKMO-CC labelling is seen on the outer membrane of some HTT-labelled mitochondria (30 nm particles). Scale bar = 1 µm.

Fig 5: Interaction of flKMO-CC and HTT-VN in HEK293T cells. (A) Cells were transfected for 48 h with 0.16 µg of each plasmid and 0.08 µg of RFP. Fluorescence intensities were analysed using ScanR analysis software. Mean green intensity of the fluorescence complementation signal of three independent experiments shows a clear interaction between flKMO-CC and HTT-VN. The histogram shows a significant reduction in the fluorescence complementation as the polyQ length increases, which is significantly different from the background (background = flKMO-CC + VN-backbone); positive control = DJ-1-GN + DJ-1-CC from [34]. **** p < 0.0001, for one-way ANOVA, followed by Tukey’s multiple comparison tests. Data are expressed as mean ± SEM. The number of analysed cells ranged from 15,000 to 18,500 cells per condition. (B) A representative immunoblot of the lysates from one BiFC experiment shows the expression levels of flKMO-CC and the soluble fraction of HTT-VN proteins, using anti-GFP antibody (ab6556). Positive control = DJ-1-GN + DJ-1-CC. (C) Filter trap of cells lysates expressing only HTT-VN to reveal polyQ dependent protein aggregation, using anti-GFP antibody (ab6556; 1:10,000); each lysate was blotted in duplicate. (D) Activity of BiFC KMO constructs after expression in HEK293T cells: truncation leads to complete loss of activity, but flKMO-CC activity is maintained when co-expressed with BiFC-VN constructs. **** p < 0.0001 and ns = not significant for one-way ANOVA, followed by Tukey’s multiple comparison tests. Data are expressed as mean ± SEM. (E) HEK293T cells were transfected with untagged flKMO and either MYC alone, 1–90 amino acid HTT 23Q-MYC or 145Q-MYC constructs. Upon crosslinking, HTT constructs were pulled down by using MYC-Trap and revealed with anti-KMO (Proteintech, 1:1000). A strong interaction is observed between flKMO and 1–90 HTT-Q23-MYC whereas flKMO and 1–90 HTT-Q145-MYC displays a weaker interaction. (F) HEK293T cells were transfected with a construct expressing flKMO-RFP, which was pulled down with the RFP-Trap system and revealed with anti-HTT (4C8) antibody (MAB2166; 1:1000). An interaction between flKMO-RFP and endogenous HTT was detected. TCL = total cell lysate, FT = flow through, W = wash and IP = immunoprecipitation.