Cell culture
The GBM39 neurosphere cell line has been previously described60: it is derived from a patient with glioblastoma undergoing surgery at the Mayo Clinic. The COLO320DM and K562 cell lines were purchased from the American Type Culture Collection (ATCC), and the GM12878 cell line was purchased from the Coriell Institute for Medical Research. The colorectal cancer cell line COLO320DM and the immortalized chronic myelogenous leukaemia cell line K562 were cultured in RPMI 1640 medium with GlutaMAX (Thermo Fisher Scientific, 61870127) supplemented with 10% FBS (Thermo Fisher Scientific, A3840002) and 1% penicillin–streptomycin (Thermo Fisher Scientific, 15140163). GBM39 cells were maintained in DMEM/F12 (Thermo Fisher Scientific, 11320082), B-27 supplement (Thermo Fisher Scientific, 17504044), 1% penicillin–streptomycin, human epidermal growth factor (EGF, 20 ng ml–1; Peprotech, AF-100-15), human fibroblast growth factor (FGF, 20 ng ml–1; Peprotech, AF-100-18B) and heparin (5 µg ml–1; Sigma-Aldrich, H3149). The lymphoblastoid cell line GM12878 was grown in RPMI 1640 with GlutaMAX supplemented with 15% FBS and 1% penicillin–streptomycin. The COLO320DM live-cell imaging line was cultured in DMEM (Corning, 10-013-CV) supplemented with 10% FBS and 1% penicillin–streptomycin–glutamine (Thermo Fisher Scientific, 10378016). GBM39 neurospheres were previously authenticated by the Mischel Laboratory using metaphase DNA-FISH12; other cell lines obtained from the ATCC and Coriell were not authenticated. All cell lines tested negative for mycoplasma contamination.
Analysis of ecDNA hitchhiking in IF–DNA-FISH of anaphase cells
Analysis of ecDNA hitchhiking in IF–DNA-FISH of anaphase cells was performed on raw images used in a previous publication5. Mitotic cells were identified using Aurora kinase B, which marks daughter cell pairs undergoing mitosis, as previously described5,6. Colocalization analysis for ecDNAs with mitotic chromosomes in GBM39 cells (EGFR ecDNA), PC3 cells (ecMYC), SNU16 cells (FGFR2 ecDNA and ecMYC) and COLO320DM cells (ecMYC) described in Fig. 1 was performed using Fiji (v.2.1.0/1.53c)61. Images were split into the FISH colour + DAPI channels, and the signal threshold was manually set to remove background fluorescence. DAPI was used to mark mitotic chromosomes, and FISH signals overlapping with mitotic chromosomes were segmented using watershed segmentation. Colocalization was quantified using the ImageJ-Colocalization Threshold program, and individual and colocalized FISH signals in dividing daughter cells were counted using particle analysis.
Retain-seq
We cloned random genomic sequences into the pUC19 plasmid backbone for the Retain-seq experiments. pUC19 is a simple, small (about 2.7 kb) vector that lacks a mammalian origin of replication and contains few sequences that could be immunogenic or have mammalian promoter or enhancer activity. Therefore, we considered that pUC19 represents an inert and selectively neutral backbone. Consequently, changes in plasmid persistence can be more confidently ascribed to insert sequences as opposed to backbone components under selection. To generate a pool of random genomic sequences, we first fragmented the gDNA of GM12878 cells via transposition with Tn5 transposase, produced as previously described62, in a 50-µl reaction with TD buffer63, 50 ng DNA and 1 µl transposase. The reaction was performed at 37 °C for 5 min, and transposed DNA was purified using a MinElute PCR Purification kit (Qiagen, 28006). GM12878 human B lymphoblastoid cells were selected as the genome of origin owing to their relatively low copy-number variability and the presence of an EBV genome as a positive control; the majority of inserts ranged from 600 to 1,300 bp. The resulting mixture of gDNA fragments was then amplified using 500 nM forward (p5_pUC19_SmaI_20bp) and reverse (p7_pUC19_SmaI_20bp) primers using NEBNext High-Fidelity 2× PCR master mix (NEB, M0541L) followed by gel purification of DNA fragments between 400 bp and 1.5 kb. To insert the mixture of gDNA fragments into a plasmid, the pUC19 vector (Invitrogen) was linearized with SmaI, purified using NucleoSpin Gel and PCR Clean-up (Macherey-Nagel, 740609.250) and the genomic fragments were inserted into the backbone using Gibson assembly (New England Biolabs, NEB). The DNA product was electroporated into Endura Competent Cells (Biosearch Technologies, 60242-2) using a MicroPulser electroporator (Bio-Rad; default bacteria setting) following the manufacturer’s protocol, and the resulting mixed episome library was prepared using a HiSpeed Plasmid Maxi Kit (Qiagen, 12663). The analysis of representation of DNA sequences in this mixed episome library and the retained episomes in transfected cells is described below.
COLO320DM and K562 cells were seeded into a 15 cm dish per biological replicate at a density of 1 × 107 cells in 25 ml of medium. GBM39 cells were seeded into a T75 flask at a density of 5 × 106 cells in 25 ml of medium. Each cell line was incubated overnight. COLO320DM, GBM39 and K562 cells were transfected with 15 µg of an input mixed episome library using Lipofectamine 3000 transfection reagent following the manufacturer’s directions. In brief, 1.5 × 107 GM12878 cells were electroporated with 50 µg input mixed episome library using the Neon Transfection system (Thermo Fisher Scientific, MPK5000). The cells were counted, centrifuged at 300g for 5 min and washed twice with PBS before resuspension in Neon Resuspension buffer to a density of 4.2 × 106 in 70 µl of buffer. The input mixed episome library was also diluted to a density of 14 µg in 70 µl with Neon Resuspension buffer. Next, 70 µl of cell suspension and 70 µl of library were mixed and electroporated according to the manufacturer’s instructions using a 100 µl Neon pipette tip under the following settings: 1,200 V, 20 ms, 3 pulses. Five electroporation reactions were pooled per replicate of GM12878 Retain-seq screens.
Cells were incubated for 2 days before the first subculture to allow recovery from transfection, and then subcultured every 3–4 days afterwards as dictated by the doubling time of each cell line. Once each cell line reached a count of 100–400 million cells per replicate, we collected all but 10 million cells, which were maintained in culture and passaged in the same manner until all subsequent time points had been collected (for a maximum of 3 time points per cell line). Thus, COLO320DM cells were collected at days 7, 14 and 21 after transfection, with a total cell count of approximately 4 × 108 cells at each time point, per replicate. GBM39 cells were collected at days 10, 20 and 30, with total cell counts of approximately 1.5 × 108 per replicate. K562 cells were collected at days 6, 12 and 18, with cell counts of approximately 4.5 × 108 per replicate. GM12878 cells were collected at day 12, with a cell count of approximately 2 × 108.
The output plasmid library was extracted using a HiSpeed Plasmid Maxi kit (Qiagen, 12663) and concentrated to a final volume of 50 µl by isopropanol precipitation. DNA was precipitated with a 1:10 volume of 3 M sodium acetate and 2 volumes of isopropanol, chilled at 4 °C for 10 min and centrifuged at 15,000g for 15 min at 4 °C. The pellet was washed with 500 µl ice-cold 70% ethanol and dissolved in 50 µl Buffer EB (Qiagen, 19086).
To enrich for input mixed episome library inserts, a preliminary PCR amplification (PCR1) of 10 cycles using primers (at 500 nM) annealing to the pUC19 vector (forward: pUC19_SmaI_5prime_fwr; reverse: pUC19_SmaI_3prime_rev) were performed on the concentrated DNA using NEBNext High-Fidelity 2× PCR master mix (NEB, M0541L). Each PCR1 reaction used a maximum of 2 µg concentrated DNA as template, with reactions assembled successively until all concentrated DNA was consumed; all reactions for a given sample were pooled following PCR1 and purified using a NucleoSpin Gel & PCR Clean-up kit (Macherey-Nagel, 740611), resulting in PCR product 1. Owing to variabilities in the insert size and the amount of retained plasmid DNA in the output library, artificial over-representation of fragments caused by PCR overcycling represented a concern for subsequent sequencing. Thus, we used qPCR to identify the cycle before saturation and halted amplification at this point. For qPCR, 50 ng of DNA from PCR product 1, NEBNext High-Fidelity 2× PCR master mix, 500 nM forward and reverse primers (forward: p5_adapter_only; reverse: p7_adapter_only) and 1 µl of 25× SYBR Green I (diluted from 10,000× stock; Thermo Fisher Scientific, S7563) were used in a 50 µl reaction. The SYBR Green signal of amplification products was measured in technical triplicates per reaction using a Lightcycler 480 (Roche) and plotted against the cycle number to identify the PCR cycle before saturation. According to the cycle numbers identified by this qPCR step, we then performed PCR2 by amplifying PCR product 1 (50 ng DNA) using the same primers as for the qPCR with the following number of cycles: 5, 10 and 12 PCR cycles for days 7, 14 and 21, respectively, of the COLO320DM experiment; 5, 11 and 18 PCR cycles for days 10, 20 and 30, respectively, of the GBM39 experiment; 5, 11 and 17 PCR cycles for days 6, 12, and 18, respectively, of the K562 experiment; and 10 PCR cycles for day 12 of the GM12878 experiment. We also collected a day-17 time point from the GM12878 experiment (amplified using 16 PCR cycles) that was specifically used to study retention of the EBV FR element, as this time point was assumed to be more comparable to the second time point in other cell lines. Next, output DNA from this step (PCR product 2) was purified using a MinElute PCR Purification kit (Qiagen, 28006) and then transposed with Tn5 transposase produced as previously described62 in a 50 µl reaction with TD buffer63, 50 ng DNA (PCR product 2) and 1 µl transposase. The reaction was performed at 50 °C for 5 min, and transposed DNA was purified using a MinElute PCR Purification it (Qiagen, 28006). The above PCR steps and transposition were also carried out on the input mixed episome library originally used for cell transfection, but with 25 ng of input mixed episome library for PCR1. According to the cycle numbers identified by this qPCR step, we then amplified PCR product 1 (1 ng DNA) over 9 PCR cycles (PCR2). Finally, the previous PCR steps and transposition were also performed on a dilution series of 10 ng, 1 ng, 0.1 ng, and 0.01 ng of input mixed episome library as PCR1 template DNA to standardize analysis of screen output across varying DNA amounts.
... continue reading