Fig 2: DYRK2 levels correlate with HSF1 nuclear levels and associates with poor prognosis and tumour recurrence in tissue from TNBC patients.A Chi-Square test correlations between nuclear/cytoplasmic HSF1, and nuclear/cytoplasmic DYRK2 in triple-negative breast cancer. Relationship between either cytoplasmic DYRK2 levels (left panel) or nuclear DYRK2 levels (right panel) and cancer-specific survival in (B) patients with triple-negative invasive ductal breast cancer, and (C) patients with triple-negative and AR-negative invasive ductal breast cancer. D Relationship between nuclear DYRK2 levels in tumour cells and time to recurrence in patients with triple negative (left panel) or triple negative and AR-negative (right panel) invasive ductal breast cancer. See also Fig. S6.

Fig 3: DYRK2 affects the expression levels of the HSF1 target gene HSP70.A Control (WT), DYRK2-KO or HSF1-KO MDA-MB-468 cells were incubated at 37 or 42 °C for 1 h. The mRNA levels for HSP70 (HSPA1A) were quantified using real-time PCR. The data were normalised using β-actin as an internal control. Data represent means ± SD (n = 3) and are expressed relative to the control sample levels at 37 °C. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, B Control (WT), DYRK2-KO or HSF1-KO MDA-MB-468 cells were incubated at 37 or at 42 °C for 1 h followed by another hour at 37 °C. Cells were lysed and the levels of the indicated proteins were analysed by western blot. The corresponding quantifications of HSP70 protein levels are shown in Supplementary Fig. S4B. C Control (WT) and DYRK2-KO MDA-MB-468 cells were incubated at 37 °C, at 42 °C for 5, 15, 30 or 60 min, or at 42 °C for 60 min plus recovery time at 37 °C (up to 180 min). After that, mRNA levels for HSP70 (HSPA1A) were quantified using real-time PCR. The data were normalised using β-actin as an internal control. Data represent means ± SD (n = 3) and are expressed relative to the control samples at 37 °C. D DYRK2-KO MDA-MB-468 infected with virus encoding for either empty vector (Ctrl), DYRK2-WT (R. WT) or the DYRK2 HSF1-interaction deficient mutant (R. BR1 + 2) were incubated at 37 °C or at 42 °C for 60 min. After that, mRNA levels for HSP70 (HSPA1A) were quantified using real-time PCR. The data were normalised using β-actin as an internal control. Data represent means ± SD (n = 3) and are expressed relative to the control samples at 37 °C. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. E DYRK2-KO MDA-MB-468 infected with virus encoding for either empty vector (Ctrl), DYRK2-WT (R. WT) or the DYRK2 HSF1-interaction deficient mutant (R. BR1 + 2) were incubated at 37 °C or at 42 °C for either 60 or 120 min. Cells were lysed and the levels of the indicated proteins were analysed by western blot. The corresponding quantifications of HSP70 protein levels are shown in Supplementary Fig. S4D. See also Fig. S4.

Fig 4: DYRK2 phosphorylates HSF1.A 293T cells were transiently transfected to express GFP-tagged DYRK2 wild-type (WT) or a kinase dead (KD) version. After 48 h, cells were lysed and the levels of endogenous HSF1 and phospho-HSF1 were analysed as indicated. B 293T cells were transiently transfected with a GFP-tagged DYRK2 analogue sensitive (AS) version. After 48 h, cells were treated for a further 3 h with increasing concentrations of three different PP1 inhibitors as indicated. Cells were lysed and the levels of endogenous HSF1 and phospho-HSF1 were analysed by western blot. C Upper panel, purified recombinant His-HSF1 (1 µg) was incubated in kinase buffer with increasing concentrations of recombinant GST-DYRK2 or GST-DYRK2 kinase dead (KD) at 30 °C for 30 min. Lower panel, purified recombinant His-HSF1 (1 µg) was incubated with either 20 ng of GST-DYRK2 WT or KD at 30 °C for various times as indicated. The reactions were terminated by the addition of SDS gel loading buffer, the proteins were resolved by SDS-PAGE, and the levels of phosphorylated HSF1 were analysed. D MDA-MB-468 cells were treated with either DMSO, the p38 inhibitor SB202190 (10 µM), the mTOR inhibitor rapamycin (30 nM) or harmine (10 µM). After 1 h, cells were incubated at 42 °C. After 1 h, cells were lysed in SDS buffer and the levels of the indicated proteins were analysed. Upper panel is a representative western blot and the bottom panel shows the quantification of the ratio between the phospho-HSF1 and total HSF1 levels. Data represent means ± SD (n = 3) and are expressed relative to the DMSO treated samples. See also Fig. S1.

Fig 5: DYRK2 promotes HSF1 nuclear stability.A 293T cells were transfected with either empty vector or GFP-DYRK2. After 48 h, cells were incubated at 37 or 42 °C for 5 min. Nuclear and cytoplasmic fractions were analysed by western blot for the levels of the indicated proteins. The corresponding quantifications of HSF1 protein levels are shown in Supplementary Fig. S3A. B Control (WT) and CRISPR-mediated DYRK2-KO MDA-MB-468 cells were incubated at 37 or at 42 °C for 5 min. Nuclear and cytoplasmic fractions were analysed by western blot for the levels of the indicated proteins. The corresponding quantifications of HSF1 protein levels are shown in Supplementary Fig. S3B. C MDA-MB-468 cells were transfected with either siControl or siDYRK2 (pool) or individual siDYRK2 (1–4). After 48 h, cells were incubated at 37 or at 42 °C for 5 min. Nuclear fractions were analysed by western blot for the levels of the indicated proteins. The corresponding quantifications of HSF1 protein levels are shown in Supplementary Fig. S3E. D DYRK2-KO MDA-MB-468 cells transduced with virus encoding for either the empty vector (Ctrl), DYRK2-WT (R. WT) or the DYRK2 HSF1-interaction deficient mutant (R. BR1 + 2) were incubated at 37 or at 42 °C for 5 min. Nuclear fractions were analysed by western blot for the levels of the indicated proteins. The corresponding quantifications of HSF1 protein levels are shown in Supplementary Fig. S3F. E 293T cells were transfected with either Flag-tagged HSF1 or with Flag-tagged HSF1 mutant were S320 and S326 were mutated to alanine (S326AS320A) in combination with either empty plasmid or GFP-DYRK2 as indicated. After 48 h, cells were lysed and nuclear and cytosolic fractions were analysed for the levels of the indicated proteins by western blotting. The corresponding quantifications of HSF1 protein levels are shown in Supplementary Fig. S3G. F Control (WT) and CRISPR-mediated DYRK2-KO MDA-MB-468 cells were incubated with either DMSO or MG132 (10 µM). After 1 h, cells were incubated at 37 or at 42 °C for 5 min. Nuclear and cytoplasmic fractions were analysed by western blot for the levels of the indicated proteins. The corresponding quantifications of HSF1 protein levels are shown in Supplementary Fig. S3K. G Control (WT) and CRISPR-mediated DYRK2-KO MDA-MB-468 cells were incubated with either DMSO (0) or Cycloheximide (10 µM) for the indicated times. Nuclear fractions were analysed by western blot for the levels of HSF1. The corresponding quantifications of HSF1 protein levels are shown in Supplementary Fig. S3L. See also Fig. S3.