Families 1, 2 and 3 have three, two and one members with cardiac angiosarcomas (CAS), respectively. CAS and other tumours are shown (colour code below the Fig.). Exomes of blue-squared CAS cases were sequenced. One member of family 4 presented with breast angiosarcoma. Only family members of whom DNA samples were available are numbered (1–19) in bold and italics. Age of onset (CAS cases) and age of sample (mutation carriers) is shown in brackets. Age of onset of other tumours is shown in the . ‘*' indicates immortalized lymphocyte cell lines from family 1: II-1 and II-4 (64- and 65-years old, respectively) and III-15 and III-16 (34- and 35-years old, respectively). ‘**' indicates paraffin-embedded tumour tissue available. Black arrows show the probandus for each family.

(a) Amino acid conservation across representative phylogeny of POT1 orthologues (tet, Tetraodon nigroviridis; xen, Xenopus laevis; gal, Gallus gallus; hum, Homo sapiens; mou, Mus musculus). The mouse POT1a gene sequence was used in the alignment. Position p.117 is shown in blue; mutations described in previous studies are shown in yellow. Triangles indicate positions with putative conformational changes (red) and loss of TPP1-binding site (green) due to the p.R117C mutation. ‘*' indicates amino acid (aa) conserved in all POT1 orthologues. (b) Heat map representation shows the tolerance to independent aa substitutions (y-axis) for each position of the protein (x-axis). Dark red indicates the highest score for a deleterious effect (score 100); white indicates a small effect; green indicates a neutral effect/no effect (score −100); and black represents the corresponding wild-type residue. Deleteriousness effect score is shown for highlighted positions. (c) Putative tertiary structure. p.152 and p.266 residues change PACC score value driving a putative protein conformation change. Left: homology-based three-dimensional model of human POT1 (Uniprot, Q9NUX5). Right: structural impact of the p.R117C mutation using the same algorithm from Protein Model Portal (Uniprot, PSI_SBKB). Black triangle shows the loop where p.117 is located. Red triangles show the principal detected structural changes (p.152 and p.266). Blue arrows show OB1 and OB2 domains.

(a) Quantification of telomere-bound TRF1 (left panel) and POT1 (right panel) protein levels in immortalized LCLs corresponding to p.R117C carriers and non-carriers. Two independent experiments with replicated samples were performed. a.u.f., arbitrary units of fluorescence. (b) Representative images of TRF1 (red) and POT1 (green) double immunofluorescence in wild-type and p.R117C carriers. White arrows indicate colocalization of both proteins (yellow spots). Scale bar, 5 μm. (c) Quantification of telomeric DNA bound to POT1, TRF1 and TRF2 by ChIP analysis. IgG was used as negative control. Results were normalized to input chromatin. Black bars, wild-type; grey bars, p.R117C carriers. Two independent experiments from each genotype were performed. Lower panel: representative ChIP dot-blot is shown. (d) Left panel: western blot analysis of in vitro-translated FLAG-TPP1, FLAG-TPP1^OBD, MYC-POT1, MYC-POT1^R117C and MYC-POT1^ΔOB1. Right Panel: p.R117C substitution decreased POT1 binding capacity to TPP1. Co-immunoprecipitation assays of the in vitro-translated proteins. FLAG-TPP1 was pulled down with MYC antibody and revealed by FLAG antibody. FLAG-TPP1 lacking the TIN2 and POT1 binding domain, FLAG-TPP1OBD, and a POT1 mutant lacking its OB1 domain were used as controls. Two exposures are showed. (e) p.R117C substitution decreased POT1 binding capacity to telomeric ssDNA. Electrophoretic mobility shift assay of [³²P]-labelled oligonucleotide (5′-TTAGGG-3′)₇ in the presence of the indicated in vitro-translated POT1 proteins. Data from two independent experiments are shown in d,e. (f) Upper panel: quantification of multitelomeric signal (MTS) events per metaphase in primary lymphocytes by telomeric FISH (n=2 in triplicate). Lower panel: example of MTS (white arrows). Red fluorescence shows telomere signals. Scale bar, 5 μm. (g) Quantification of cells positive for γH2AX (left) and per cent of cells with >5 telomeric induced foci (TIFs) (right) in immortalized LCLs (n=2 in triplicate). Lower panel: representative images of γH2AX and TRF1 immunofluorescence. White arrows show examples of TIFs (yellow spots) with anti-TRF1 (green fluorescence) and anti-γH2AX (red fluorescence). Scale bar, 5 μm. Values are expressed as mean+s.e. The two-tailed student's unpaired t-test was used for the statistical analysis, NS, not significant. DAPI (blue) was used for DNA labelling.

(a) Telomere length (TL) analysis by qPCR (t/s ratio) for family 1 members (n=29). (b) TL analysis by qFISH of primary immortalized LCLs corresponding to non-carriers (II-1 and III-16) and p.R117C carriers (II-4 and III-15). (c) Percentage of telomeres shorter than 3 kb of primary immortalized LCLs corresponding to non-carriers (II-1 and III-16) and p.R117C carriers (II-4 and III-15). (d) TRAP Tel/IC ratio values for telomerase activity calculated from primary lymphocytes of different members of family 1 (n=7). In b–d two independent experiments with samples in triplicate were performed. (e) Left panel: number of telomere sister chromatid exchange (T-SCE) events/metaphase in primary lymphocytes by CO-FISH (n=2 individuals per genotype in triplicate). Right panels: representative CO-FISH images showing the leading (green) and lagging (red) telomere strands. T-SCEs are indicated with arrows. DAPI (blue) was used for DNA labelling. Below a magnified merge image is shown. Scale bar, 1 μm. (f) Upper panel: per cent of cells positive for ultra-bright spots (ubs) at telomeres by FISH in three different paraffin-embedded cardiac tumour (T) and normal (N) tissue samples carrying the mutation from members of family 2 (F2) and 3 (F3). Lower panel: examples of large red spots corresponding to positive signals (white arrows). DAPI (blue) was used for DNA labelling. Scale bar, 10 μm. (g) TL adjusted for age of wt and p.R117C carriers of all members of families 1, 2 and 3 and the 3 CAS tumours. DNA from CAS samples was extracted from paraffin-embedded tissues. Values are expressed as mean+s.e. and the two-tailed student's unpaired t-test was used for the statistical analysis, NS, not significant.

(a) Representative images of MYC (green) and TRF1 (red) double immunofluorescence in Hela cells with endogenous expression of POT1 protein infected with retroviral empty vector, MYC-POT1, MYC-POT1^R117C and MYC-POT1^ΔOB1. DAPI (blue) was used to DNA labelling. Scale bar, 5 μm. (b) Quantification of MYC-POT1 and TRF1 colocalization foci (left panel) and of nuclear TRF1 foci intensity (right panel) (a.u.f., arbitrary units of fluorescence) in Hela cells infected with retroviral empty vector, MYC-POT1, MYC-POT1^R117C and MYC-POT1^ΔOB1. (c,d) Quantification of nuclear telomere length per metaphase (a.u.f) (c) and of multitelomeric signal (MTS) events/metaphase (d) in Hela cells infected with retroviral empty vector, MYC-POT1, MYC-POT1^R117C and MYC-POT1^ΔOB1. (e) Per cent of γH2AX-positive cells in Hela cells infected with retroviral empty vector, MYC-POT1, MYC-POT1^R117C and MYC-POT1^ΔOB1. (f) Representative qFISH images of metaphase spreads. Examples of MTS are shown in the magnified insets (arrow). Red fluorescence shows telomere signal. DAPI (blue) was used for DNA labelling. Scale bar, 5 μm. Two independent infections were performed. Values are expressed as mean+s.e. The two-tailed student's unpaired t-test was used for the statistical analysis, NS, not significant.