HPC samples

This study used germline DNA from 462 index cases of early-onset/familial PrCa (HPC) cases, previously recruited [4]. The patients were recruited based on two criteria: early-onset disease, with PrCa diagnosis before the age of 56, and/or familial/hereditary PrCa, with more than one case with PrCa and at least one of them diagnosed before the age of 66. Of the 462 HPC cases, 240 (51.9%) fulfilled the early-onset disease criterion, and 311 (67.3%) fulfilled the family history criterion, with 89 (19.3%) fulfilling both criteria. Demographic and clinicopathological characteristics of all the carriers are listed in Table S1.

Targeted next-generation sequencing (T-NGS)

We used a customized gene panel designed with Agilent SureDesign (Agilent Technologies, Santa Clara, CA, USA) to sequence germline DNA from the index patients of all 462 PrCa cases. The panel covered, among other PrCa genes under investigation, the coding and splicing regions of BUB1B (NM_001211.5). Capture and sequencing, data processing, variant annotation, and prioritization were performed as previously described [22]. Detailed information is described in Supplementary Methods.

Dataset of patients tested for other hereditary cancer syndromes

To gain insights into the possible pan-cancer role of pathogenic/likely pathogenic variants in BUB1B, we searched for BUB1B carriers among 1,416 cancer patients referred for molecular diagnosis of multiple inherited cancer syndromes at the Department of Genetics of IPO Porto, already screened with the TruSight Cancer Panel v.1 (Illumina). Variant annotation and filtering were performed as described for the custom T-NGS panel mentioned above and detailed in Supplementary Methods.

Control samples

To estimate the risk between the carrier status for the c.1171_1173del variant and PrCa development, we used germline DNA from 459 healthy male individuals (mean age 48.3 years; SD ± 10.2 years) as control samples, including 288 blood donors from the Portuguese Oncology Institute of Porto with no personal history of cancer at the time of blood collection and 171 healthy relatives with negative predictive genetic testing (each from independent families). To estimate the global risk for cancer development, germline DNA from 416 healthy females (mean age 48.6 years; SD ± 9.7 years) was also used, which included 243 blood donors and 173 healthy relatives with negative predictive genetic testing.

To assess the basal levels of the PCS trait in our population, we performed PCS on normal lymphocyte metaphase spreads of ten age-matched healthy males aged between 42 and 78 years (mean age 52.6 years; SD ± 10.7 years), specifically recruited for this purpose, under signed informed consent.

KASP genotyping

Kompetitive Allele Specific PCR (KASP) genotyping, with variant-specific KASP probes, was performed according to the manufacturer’s instructions. Assay primers (Metabion, Köln, Germany) summarized in Table S2 were designed using the Primer-BLAST design tool from the National Center for Biotechnology Information (NCBI) [23], and the PCRs were run on a LightCycler 480 Real-Time instrument (Roche Life Sciences, Basel, Switzerland). For data analysis, LightCycler 480 Software 1.5.0 was used.

Haplotype analyses

The T-NGS data were phased using BEAGLE 4.1 [24], and IBD haplotypes were determined using the BEAGLE Refined IBD algorithm [25]. The lengths of the shared haplotype segments were estimated by the distance between the last two shared markers flanking the variants. A similar IBD and haplotype approach was applied to the high-density SNP genotype data from the Portuguese early-onset/familial PrCa sample collection (374 PrCa cases and 180 controls) obtained with the Infinium OncoArray-500 K BeadChip (Illumina) as part of the PRACTICAL consortium, as previously described [26].

Microsatellite haplotype analysis was performed using nine polymorphic microsatellite markers flanking the gene, namely, D15S118, D15S1012, D15S1044, D15S146, D15S214, TR20GT, D15S968, AFM196XB8, and D15S781. A total of seven probands carrying the BUB1B c.1171_1173del variant were genotyped, including all five HPC patients, the two additional carrier patients found among the 1,416 screened for multiple hereditary cancer syndromes (Table 2), and seven unaffected family members. Primers were designed using the Primer-BLAST tool [23] (Table S2) and acquired from Metabion. All markers were assayed by PCR using fluorescently end-labeled primers, and PCR products were run on a 3500 Genetic Analyzer together with the fluorescence-labeled DNA fragment size standard 600-LIZ (Thermo Fisher Scientific, Waltham, MA, USA). Haplotype construction was performed manually based on the genotypes obtained from probands and family members.

Next-generation DNA and RNA sequencing of FFPE tumor samples

To assess the impact of BUB1B germline variants on BUB1B mRNA expression and transcriptomic profiles in the corresponding prostate carcinomas, available prostate tissue samples from carriers were submitted to RNA sequencing. For this purpose, matched tumor and normal RNA were extracted from ~ 5 μm sections of formalin-fixed paraffin-embedded (FFPE) prostate tissues after deparaffinization using xylene and ethanol, according to the recommendations of the High Pure FFPET RNA Isolation Kit (Roche Life Sciences). For each library preparation, 100 ng of total RNA was used as input for the TruSeq™ RNA Exome (Illumina) protocol, according to the manufacturer’s instructions. Libraries were sequenced on an Illumina NextSeq 500 using the NextSeq High Output v2.5 kit using the pair-end run mode with a run setup of 2 × 76 bps. Data was processed for both transcript expression and variant calling, as described in Supplementary Methods.

To look for a somatic “second hit” in BUB1B in the tumors of HPC patients carrying germline variants, DNA was extracted from ~ 5 μm sections of FFPE tumor tissues using the Cobas® DNA Sample Preparation Kit (Roche Life Sciences) after deparaffinization with xylene and ethanol, according to the manufacturer’s recommendations. A T-NGS custom panel covering, among other genes, the coding and splicing regions of BUB1B (NM_001211.5), designed with Agilent SureDesign, was used. Capture, sequencing, data processing, and variant annotation were performed as described for germline DNA.

Generation of prostate in vitro cell models carrying the BUB1B variant c.1171_1173del

The human non-tumorigenic prostate-derived cells RWPE-1 were kindly provided by Prof. Margarida Fardilha (University of Aveiro, Portugal). Wild-type RWPE-1 and derived monoclonal cell populations were cultured in sterile conditions, at 37ºC, 5% CO2, and a humidified atmosphere, in Kerotinocyte Serum-Free (KSF) Medium (Gibco) supplemented with L-glutamine, 0.005 ng/µL Human Recombinant Epidermal Growth Factor (EGF), and 50 ng/µL of Bovine Pituitary Extract (BPE), and 1% penicillin–streptomycin (Gibco).

To induce specific editing of the BUB1B gene in RWPE-1 cells, we used the Alt-R CRISPR/Cas9 System from Integrated DNA Technologies (IDT, Coralville, IA, USA) following the manufacturer’s instructions. Among 153 isolated clones, gene-editing leading to the presence of the recurrent BUB1B variant c.1171_1173del was observed in four independent clones, in which gene-editing also led to the occurrence of a second in-frame variant (c.1133_1156del) in the other allele, predicted to lead to loss of amino acids from Proline 378 to Histidine 385 [p.(Pro378_His385del)]. No heterozygous clones with monoallelic gene-editing for the c.1171_1173del variant were found. Thus, we randomly selected two of the four clonal populations with the same genotype (BubR1Δ391/Δ378−385) – named C1 and C2 – for further phenotypic evaluation. Detailed information is described in Supplementary Methods.

Sanger sequencing

For Sanger sequencing, monoclonal cell populations were lysed directly in the well of a 96-well plate with 10 µL of reaction buffer mixture of the Xpert directXtract Lysis Buffer (GRiSP, Porto, Portugal) after a D-PBS wash (Gibco, Thermo Fisher Scientific). Cell lysates were incubated at 75ºC for 5 min, followed by heat-inactivation at 95ºC for 10 min. The cell lysate was diluted 5 × for PCR amplification of the BUB1B in-frame deletion, as previously described [5]. Specific primers were designed using Primer-BLAST [23] (Table S2) and acquired from Metabion. The PCR products were purified with Exo/SAP Go (GRiSP) and forwarded for sequencing PCR using the BigDye Terminator v3.1 Cycle Sequencing Kit (Thermo Fisher Scientific), according to the manufacturers’ instructions. Samples were run in a 3500 Genetic Analyzer (Thermo Fisher Scientific).

Western blot

To assess steady-state BubR1 expression levels, total protein extracts were obtained from 90% confluent cells. Briefly, adherent cells were washed twice with ice-cold D-PBS and scrapped with ice-cold lysis buffer (50 mM Tris HCl, pH 7.5, 150 mM NaCl, and 0.1% NP-40). After incubation on ice for 15 min, soluble fractions were collected by centrifugation at 13,000 rcf for 30 min.

To obtain mitotic extracts, 90% confluent cells in T75 flasks were synchronized with 0.5 µM nocodazole. Mitotic cells were harvested by shake-off and centrifugation at 1200 rpm for 5 min. Cell pellets were washed once with D-PBS and resuspended in ice-cold lysis buffer (50 mM Tris HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 0.5% NP-40, and 0.5% Triton X-100) supplemented with a cocktail of protease inhibitors (Roche Life Sciences). Proteins (50 µg) were separated in a 10% SDS-PAGE system and transferred to a nitrocellulose membrane using the Trans-Blot Turbo Transfer System (Bio-Rad Laboratories, Hercules, CA, USA). BubR1, Bub3, Cdc20 and β-actin were probed using the following antibodies: rabbit anti-BubR1 (1:1,000; ab70544, Abcam, Cambridge, UK), anti-Bub3 (1:2,000; 27,073–1-AP, Proteintech, Rosemont, IL, USA), mouse anti-cdc20 (1:250; sc-5296, Santa Cruz Biotechnology, Santa Cruz, CA, USA), and anti-β-actin antibody (1:10,000; A1978, Sigma-Aldrich). Anti-rabbit (1:10 000) and anti-mouse (1:2,500) HRP-conjugated secondary antibodies (Sigma-Aldrich and Bio-Rad Laboratories, respectively) were visualized using Clarity Western ECL Substrate (Bio-Rad Laboratories).

Cycloheximide (CHX) chase assay

Cells at 80–90% confluence in T25 flasks were treated with 20ug/mL of cycloheximide (Sigma-Aldrich; C4859) without medium change. At different time points (0 h, 4 h, 8 h and 16 h), total protein extracts were obtained and analyzed by western blot, as described above. Two independent experiments were performed for each clonal cell line paired with WT cells. Images were acquired in the ChemiDoc™ XRS System and data quantified in ImageJ.

Quantitative PCR (qPCR)

To evaluate the expression levels of BUB1B in the gene-edited cell line models and WT cells, total RNA was extracted from each cell population using the RNEasy Mini Kit (QIAGEN, Hilden, Germany), and cDNA was synthesized using the H-minus RevertAid cDNA synthesis kit (Fermentas, part of Thermo Fisher Scientific) with oligo-dT primers, according to the manufacturers’ protocols. Primers and probes (TaqMan) for BUB1B were designed using Primer3 software (v4.1, https://primer3.ut.ee) and acquired from Metabion (Martinsried, Germany) (Table S2). The beta-glucuronidase (GUSB) housekeeping gene was used as an endogenous control for normalization of the expression levels, and the primers/probe mix was acquired as a pre-developed TaqMan Gene Expression Assay from Applied Biosystems (part of Thermo Fisher Scientific). Relative expression levels were obtained by calibrating GUSB normalized BUB1B expression values from each population for the expression levels of the WT control population.

Proliferation and apoptosis assays

Cellular proliferation/viability and apoptosis levels were evaluated with MTT (Sigma-Aldrich) and APOPercentage (Biocolor, Carrickfergus, UK) colorimetric assays, respectively, as previously described [27, 28]. Growth rate (GR) was calculated from the absorbance values obtained with the MTT assay using the formula GR = (T96-T0)/T0. To evaluate Taxol growth inhibition, complete growth medium containing 10 nM Taxol or an equivalent volume of the drug vehicle (DMSO) was added to cells at T0 (h), and the MTT assay was evaluated at T96 (h). For each cell population, the percentage of growth inhibition was calculated using the formula % Inhibition = 100x [1- (T96-T0)Taxol/(T96-T0)Vehicle], and the percentage of apoptosis was obtained by correcting apoptotic cells to the “total cells” (sum of viable and apoptotic cells) at T96 (h). Statistical data was obtained from four independent experiments.

Time-lapse microscopy

Live cell analysis was performed under conditions of mitotic arrest generated by persistent unattached KTs using 0.5 µM nocodazole (Sigma-Aldrich) for 16 h. Mitotic controls and BUB1B-edited cell clones were imaged by phase-contrast microscopy every 20 min for 36 h using an IN Cell Analyzer 2000 microscope (GE Healthcare, CH, IL, USA) at 37 °C in complete KSF medium. Image processing and quantification were performed in ImageJ.

Premature Chromatid Separation (PCS) analysis

Peripheral blood samples were cultured for 72 h in RPMI 1640 medium with GlutaMAX-I (Gibco) supplemented with 20% fetal bovine serum (Gibco) and stimulated with Phytohaemagglutinin-M (Biological Industries, Israel). Colcemid™ (100 ng/mL, Gibco) was added 90 min before cell harvesting by trypsinization and cells were processed for G-banding with Leishman staining, according to standard protocols. The analysis of PCS in patients HPC343 and HPC369 carrying missense variants was not performed due to the absence of biological material.

For analysis of PCS levels in cell lines, cells at 90% confluence in T75 flasks were Colcemid™-synchronized for 16 h before harvesting and G-banding processing.

Automatic capture of metaphases from each case was performed using the microscope slide scanning system GSL-120 (CytoVision version 7.4; Leica Biosystems, Baden-Wurttemberg, Germany). The percentage of metaphases with PCS, corresponding to cells with separate and splayed chromatids with discernible centromeres, involving all or most chromosomes of a metaphase cell, was quantified manually by a certified cytogeneticist.

High-throughput screening of lagging chromosomes

Quantification of lagging chromosomes was performed using 50–500 images of contiguous fields acquired in an IN Cell Analyzer 2000 microscope (GE Healthcare, CH, IL, USA) with a Nikon 40 × /0.95 NA Plan Fluor objective (binning 2 × 2) using a large chip CCD Camera (CoolSNAP K4) with a pixel array of 1,024 × 1,024 (2.7027 pixel/µm resolution), as previously described [29]. All early to late anaphase figures were classified regarding the presence or absence of lagging chromosomes. Any DAPI-positive material between the two chromosome masses, but distinguishably separated from them, was counted as lagging chromosomes. DNA bridges were excluded from this analysis.

Immunofluorescence

RWPE-1 cells were fixed with 4% paraformaldehyde in PBS (Electron Microscopy Sciences) for 10 min and permeabilized with 0.5% Triton X-100 (Sigma-Aldrich) for another 5 min. BubR1, Cdc20 and ACA were immunostained using the following antibodies: rabbit anti-BubR1 1:1,000; ab70544, Abcam, Cambridge, UK, mouse anti-cdc20 (1:100; sc-5296, Santa Cruz Biotechnology, Santa Cruz, CA, USA), and anti-centromere antibodies (ACA, 1:5,000; kind gift from B. Earnshaw, Welcome Trust Centre for Cell Biology, University of Edinburgh, Edinburgh, UK). Alexa Fluor 488, 568 and 647 (1:1000, Themofisher) were used as secondary antibodies, and DNA was counterstained with 1 µg/ml DAPI (Sigma-Aldrich).

Quantification of fluorescence intensity at the kinetochores

The fluorescence intensity signal of BubR1 and Cdc20 was measured directly at KTs in a circular ROI (region of interest) using ImageJ and normalized to the intensity signal of ACA in the same ROI. Background fluorescence was measured outside the ROI and subtracted from each KT. The mean values of all quantified KTs quantified were plotted in a scattered dot plot.

Computational analysis of BubR1 variants

Because there is no experimental structure of the human mitotic checkpoint serine/threonine-protein kinase BubR1 (EC:2.7.11.1) in the Protein Data Bank (http://www.rcsb.org), the molecular structure of the wild-type protein was first constructed by homology modeling using MODELLER software [30] (detailed in Supplementary Methods).

The molecular dynamics simulations were performed to assess the molecular dynamics of the BubR1 proteins in explicit solvent at atomistic resolution. The Amber 18.0 simulation package (parm14SB force field) was used to carry out the optimizations and MD simulations (detailed in Supplementary Methods).

The bioinformatic tools MUpro [31] and I-Mutant 2.0 [32] were used to predict the effect of the currently studied variants (BubR1R120Q, BubR1I147T, BubR1R244C, BubR1Δ391 and BubR1R416Q) and the control variants (BubR1 F175G, BubR1F175L, BubR1E413K) on the structural stability of the protein. Starting from the amino acid sequence data, both web servers use a set of machine learning methods to automatically predict protein stability changes upon single-site mutations. These indicators provide fast, quite accurate (77–80% correct previsions) and quantitative previsions of the effects of the variants.