NDP52 activates nuclear myosin VI to enhance RNA polymerase II transcription

Myosin VI (MVI) has been found to be overexpressed in ovarian, breast and prostate cancers. Moreover, it has been shown to play a role in regulating cell proliferation and migration, and to interact with RNA Polymerase II (RNAPII). Here, we find that backfolding of MVI regulates its ability to bind DNA and that a putative transcription co-activator NDP52 relieves the auto-inhibition of MVI to enable DNA binding. Additionally, we show that the MVI–NDP52 complex binds RNAPII, which is critical for transcription, and that depletion of NDP52 or MVI reduces steady-state mRNA levels. Lastly, we demonstrate that MVI directly interacts with nuclear receptors to drive expression of target genes, thereby suggesting a link to cell proliferation and migration. Overall, we suggest MVI may function as an auxiliary motor to drive transcription.

(b) Prediction of DNA binding residues using BindN (http://bioinformatics.kstate.edu/bindn/). Bars indicate residues with a predicted confidence greater than 70%. A full list of residues are presented in Supplementary Table 1 Traces were fitted as described in the Methods to yield the rate constants plotted in d.
Association and dissociation rate constants were calculated from linear fits to the data (Supplementary Table 3). Data were averaged from three independent experiments.
(e) Titration of Folch vesicles against 1 M CBD or CBDSiteA. Data were fitted as described in Methods, following background subtraction of the folch autofluorescence. Fitting gives an indication of the affinity with Kd 39 +/-4 mg/ml. Due to the unknown mixed composition, it was not possible to have molar concentrations.
Poor binding by CBDSiteA means data could not be fitted to the model.

Supplementary Figure 3.
(a-b) Intracellular distribution of EGFP-LI-MVI transiently expressed in HeLa cells. The EGFP-LI-MVI localisation was restricted to the cell periphery, being excluded from the nucleus. This is highlighted by the plot of mean intensity of Hoechst (Hst) and EGFP across each image stack. The position is measured relative to the mid-point of the nucleus. The intensity was measured within a region of interest through the centre of the nucleus. Scale bar 10 m.
(c) Intracellular distribution of EGFP-LI-MVI(M1062Q). The mutant, in which the LI helix is destabilised, displayed a similar distribution to the NI isoform. Images were acquired at the mid-point of the nucleus. Scale bar 10 m.

Supplementary Figure 4.
(a) Raw fluorescence intensity data for titration of NMVITAIL against CBD in Fig. 3a. (b) Raw intensity data as shown in (a) but on a smaller scale.
(c) Raw fluorescence intensity data for titration of NMVITAIL against CBD in the presence of DNA (Blue is FITC intensity and Red is AF555 intensity).  (g) Pull-down of CBD (5 M) by His-tagged Motor1-1060 and Motor1-814 both at 10 M +/-1 mM Ca 2+ . P and S represent pellet and supernatant fractions, respectively.
Interaction of the CBD with the motor constructs is shown in the cartoon scheme.
(c) Representative images of transiently expressed NI-, LI-and LI (M1062Q) -EGFP-MVI in HeLa cells stained against DNA (Cyan) and immune-stained against endogenous NDP52 (magenta). Scale bar 10 m. These images were used to assess the significance of the Pearson's coefficient regarding MVI and NDP52 co-localisation.
We have established how the NI and LI (M1062Q) can form interactions with NDP52.
Therefore, to assess this in vivo through imaging, we performed the Pearson's coefficient test (d) with the three MVI constructs. Here a region of interest was drawn around the cells and images were analysed from 5 different stacks. LI MVI shows a significantly lower co-localisation that the NI or mutant **p <0.001.