Supplementary Methods

Plasmid construction, gene targeting and

Using standard molecular biology and recombineering (ref) techniques a region of the mouse p107 and p130 genes harbouring exon 10 and 2 were isolated, respectively. In the final gene-targeting constructs exon 10 of p107 and exon 2 of p130 were subsequently flanked in surrounding introns with unidirectional loxP sites and a FRT flanked neo selection cassette. The targeting of each construct into mouse 129/SvJES cells was performed as described previously (Tonks et al, 2005). To detect correctly recombined clones EcoRI digested p107 candidates and SacI digested p130 candidates were subject to Southern Blot analysis with a probe that was external to the targeting homologies and would detect a change in the size of a correctly targeted verses a wild type allele. In the case of the p107 candidates the Southern blot analysis detected the distal loxP site unlinked to the selection cassette and was sufficient to determine which clones had recombined the entire targeting construct merely by the conversion of the wild type band to a 6.5kb targeted variant. In contrast, in the case of p130 the initial SacI digested DNA detected the conversion of the 6.8 kb wild type band to a 4.2 kb variant. The clones yielding the 4.2kb band were subsequently subject to PCR with the p130X2RB (5’-tgtgactggctggtattagaac-3’) and p130UL1 (5’-GCGCATCTCCATTTCCTTCAT-3’) primers to determine which possessed the loxP site not linked to the neomycin selection cassette, with the floxed allele yielding a larger band than the wild type allele. PCR conditions employed were as per the Rb1 amplifications describe elsewhere (Tonks et al, 2005). The overall targeting frequency for the p107 and p130 targetings was 1.9% and 0.7%, respectively.

Targeted clones were subsequently grown and injected into C57BL/6J blastocysts, with injected blastocysts being transferred into pseudopregnant hosts. The resultant high percentage chimaeras were mated to C57BL/6J hosts with germline transmission of the targeted allele being initially confirmed by the coat colour and subsequently by Southern blot analysis as described above. As the presence of the PGK-neo cassette is known to manifest hypomorphic alleles in some circumstances by virtue of its strong transcriptional elements and inherent cryptic splice sequences (Meyers et al., 1998), the mice possessing the floxed alleles were crossed with flpe deleter mice [TgN(ACTFLPe)9205Dym] (Rodriguez et al., 2000) to remove the FRT flanked PGK-neo cassette.

Mice, genotyping and tissue analysis

All experimental animals were treated in accordance with the Australian Government National Health and Medical Research Council (NHMRC) guidelines for the care of experimental animals and the work was approved by the QIMR Animal Ethics Committee. The mice were maintained on a 12 h light/dark cycle with ad libitum feed and appropriate environmental enrichment. The details of the derivation of Rb1F2/F2, Tyr::CreER(T2) and p53F2-10/F2-10 mice and their genotyping have been described elsewhere (Bosenberg et al., 2006, Tonks et al., 2005) (Jonkers et al, 2001). The floxed p130 and p107 alleles were detected using the PCR primer pairs that detect the distal loxP site. For p130 these were the p130F (seq) and p130R (seq) and they amplify a (see the instructions that I sent for genotyping) products for the wild type and (wild type plus approximately 58bp- loxP plus 4 6bp cutter sites) floxed allele respectively. The p107-4371F (seq) and p107-4703R (seq) primers were used to genotype for the wild type and floxed p107 allele as they amplified 332 bp and 428 bp products, respectively. The induction of Cre recombinase activity in the melanocytes of Tyr::CreER(T2) mice, 4-OHT (Sigma-Aldrich, St Louis, MO, USA) at a concentration of 25 mg/ml in DMSO (Sigma-Aldrich) was administered topically using a paintbrush to the flank skin and ears of mice daily for two successive days.

Melanocyte culture, PCR of culture DNA and Western blot analysis.

Primary cultures of melanocytes were derived from neonate skins and cultured as described elsewhere (Tonks et al., 2005). To evaluate the percentage of melanocytes that possessed a homozygous deletion of Rb1 in Rb1F2/F2:p107F10/F10:p130F2/F2:Tyr::CreERT2 neonates following topical 4-OHT treatment the melanocyte cultures were initially established from flank skins in 6-well plates (Falcon, Becton Dickinson, Franklin Lakes, NJ, USA). At approximately 70% confluency genomic DNA was harvested from individual melanocytes colonies and analysed by PCR to determine the efficiency of recombination using the Rb5-2 (5’-ggtccttcaatttctagcccatc-3’)/Rbamp2-2 (5’-acctagcctgagagtaggcaac-3’) primer combination reaction and amplification conditions as described elsewhere (Tonks et al, 2005). The remaining cells were replated to obtain established cultures and levels of Rb1, p107 and p130 protein expression was determined from lysates of these melanocytes cultures using an anti-Rb1 antibody (Pharmingen, Becton Dickinson), anti-p107 antibody (Santa Cruz sc-318) and anti-p130 antibodies (Santa Cruz sc-317) as previously described (Schroder et al., 2005).

To perform serum deprivation studies cells from 70% confluent parent cultures were harvested, seeded onto 15 X 90mm diameter tissue culture dishes and grown for 24 hours. After 24 hours cells on one 90mm dish were counted and the melanocytes culture medium was replaced with identical medium except that it lacked any horse or foetal bovine serum. At intervals of one (No Serum day 1-NS1), three (NS3), five (NS5) and seven (NS7) days the cells were harvested from three of the 90mm dishes and counted. On day seven the two remaining 90mm plates were re-fed with normal melanocyte culture medium, allowed them to grow for 48 hours and cells were subsequently counted.

Immunohistochemistry

The evaluation of apoptosis in cultured melanocytes, both controls and those subject to serum withdrawal as described previously, was performed the anti-active caspase 3 active form antibody (Sigma MAB10753), Sigma Sydney, Australia) (Tonks et al, 2005).

Paraffin-embedded sections of mouse skin and tumours were dewaxed, and treated with Dako low pH antigen retrieval solution (Dako Australia Pty. Ltd. North Sydney, Australia) at 125OC for 5 min. Endogenous peroxidase activity was quenched in 3% hydrogen peroxide (H202) for 10 min, and sections were washed and blocked with 10% goat serum.

For proliferation studies we used e used the mouse monoclonal anti-Ki67 (Dako), and secondary detection was via Jackson biotinylated anti-rat (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA) followed by Jackson streptavidin peroxidase conjugate colour developed with Vector red chromagen (Vector laboratories, Burlingame, CA, USA) and counterstained with haemotoxylin.

For pERK signalling we used the cell signalling anti-phospo-ERK antibody (epitope MAPK p44/42 p-Erk 5736S). For staining of melanomas we use anti-S100 SP127 aorimary antibody (Sigma, Sydney, Australia), and a pan-melanoma antibody (containing Abs to MART1, MAGE and pMel17) (Abcam, Melbourne, Australia). For these antibodies we used the Biocare Medical Mach 1 anti-Rabbit HRP polymer secondary antibody system visualised with vector red.

DNA extraction and quantitation.

DNA was extracted from cultured cells using a QIAamp DNA Mini kit (Qiagen) according to the manufacturer's instructions, quantified using a NanoDrop 1000 (ThermoScientific) and quality assessed with a Bioanalyzer DNA 12000 chip (Agilent).

Sanger sequencing

A total of 17 biopsies of melanoma were analyzed for mutations in the specified genes. The following table shows the primers used in this study to examine the hostspots for mutation in these genes. Primer sequences were taken directly from published by van Schanke et al., except for the pair of primers for H-Ras exon 1 which we modified.

van Schanke A, van Venrooij GMCAL, Jongsma MJ, Banus HA, Mullenders LHF, van Kranen HJ, de Gruijl FR. 2006. Induction of nevi and skin tumors in Ink4a/Arf Xpa knockout mice by neonatal, intermittent, or chronic UVB exposures. Cancer Res 66(5): 2608-2615.

Gene / PCR primers (5’-3’) / Amplicon size (bp) / Hotspot / Mutations found
BRAF exon 18 / Forward / TTC CTT TAC TTA CTG CAC CTC AGA / 143 / codon 600: GTG / None
Reverse / AGC GCT GCT CCG GTT CAT AGA TTC CAT CCA AAT AGA TCC AGA
NRAS exon 1 / Forward / ACA GGT TTT TGC TGG TGT GA / 115 / codon 12: GGT
codon 13: GGT / None
Reverse / AGC GCT GCT CCG GTT CAT AGA TTC ATC CAC AAA GTG GTT CTG G
NRAS exon 2 / Forward / CCT TCG CCT GTC CTC ATG TA / 125 / codon 61: CAA / None
Reverse / GGG ACA CCG CTG ATC GTT TAT CCC AGG ATT CTT ACC GAA A
HRAS exon 1 / Forward / TTG GCT AAG TGT GCT TC / 198 / codon 12: GGC
codon 13: GGT / None
Reverse / GCA AAT ACA CAG AGG AAG CC
HRAS exon 2 / Forward / CGT GTT GTT TTG CAG GAC TC / 119 / codon 61: CAG / None
Reverse / ATG TAC TGG TCC CGC ATG G
KRAS exon 1 / Forward / AGG CCT GCT GAA AAT GAC TG / 119 / codon 12: GGT
codon 13: GGC / None
Reverse / AGC GCT GCT CCG GTT CAT AGA TTC GTA TCA TAC TCA TCC AC
KRAS exon 2 / Forward / TTG GAT ATT CTC GAC ACA GCA / 142 / codon 61 : CAA / None
Reverse / AGC GCT GCT CCG GTT CAT AGA TTT TAA ACC CAC CTA TAA TGG TGA A

Exome sequencing and analysis

We harvested MM and liver from the same mouse (9 tumours from 6 mice). Tumours were selected randomly, both males and females used. DNAs were sequenced at Otogenetics Corporation, Atlanta, Georgia, USA. We used an equal number of males and females within each cohort. An Agilent SureSelect Mouse exome kit was used for exome capture from 3 μg of genomic DNA. Sequencing was performed on Illumina HiSeq to produce 100bp paired-end reads with > 60X coverage. Reads were aligned to the mouse reference genome (mm10) using Novoalign. For local realignment around indels, and base quality recalibration, we used GATK software. Only reads with mapping quality of Q30 or greater were considered. Somatic SBSs were called using VarScan and MuTect. Those called using VarScan was further filtered with p>0.05. MuTect variants were filtered to log of (likelihood tumour event is real/likelihood event is sequencing error) (t_lod_fstar) >100. Somatic variants were removed if present in dbSNP. The final set of variants after filtering using both somatic callers were annotated using Annovar and SNPeffect.