Despite considerable efforts to sequence hypermutated
cancers such as
melanoma, distinguishing
cancer-driving genes from thousands of recurrently mutated genes remains a significant challenge. To circumvent the problematic background mutation rates and identify new
melanoma driver genes, we carried out a low-copy piggyBac transposon mutagenesis screen in mice. We induced eleven
melanomas with mutation burdens that were 100-fold lower relative to human
melanomas. Thirty-eight implicated genes, including two known drivers of human
melanoma, were classified into three groups based on high, low, or background-level mutation frequencies in human
melanomas, and we further explored the functional significance of genes in each group. For two genes overlooked by prevailing discovery methods, we found that loss of
membrane associated guanylate kinase, WW and PDZ domain containing 2 and
protein tyrosine phosphatase, receptor type, O cooperated with the v-raf murine
sarcoma viral oncogene homolog B (BRAF) recurrent V600E mutation to promote cellular transformation. Moreover, for infrequently mutated genes often disregarded by current methods, we discovered recurrent
mitogen-activated protein kinase kinase kinase 1 (Map3k1)-activating insertions in our screen, mirroring recurrent MAP3K1 up-regulation in human
melanomas. Aberrant expression of Map3k1 enabled
growth factor-autonomous proliferation and drove BRAF-independent ERK signaling, thus shedding light on alternative means of activating this prominent signaling pathway in
melanoma. In summary, our study contributes several previously undescribed genes involved in
melanoma and establishes an important proof-of-principle for the utility of the low-copy transposon mutagenesis approach for identifying
cancer-driving genes, especially those masked by hypermutation.