RIZ1 — a suspected protein methyltransferase — has long been thought to possess tumour-suppressor activity: its gene maps to a region that is frequently deleted in human cancers, and reintroducing the gene into tumours causes apoptosis and growth suppression in animal tumour models. Definitive proof has, however, not been available until now. Steele-Perkins et al., in the September issue of Genes & Development, show that mice deleted for Riz1 develop a range of tumours, and that several human tumours contain mutations in RIZ1. RIZ1 is therefore the founding member of a new class of tumour suppressors.

RIZ1 was initially isolated as an RB-binding protein and a transcriptional co-activator. It possesses two important motifs — an RB-binding motif and a nuclear hormone receptor binding motif — in addition to the PR domain, which places it within the PR subfamily of protein methyltransferases. RIZ1's substrates are a complete mystery, but the gene is deleted in many human cancers, and mRNA expression is also decreased or lost in many more, probably by epigenetic silencing. So is RIZ1 a tumour suppressor?

Riz1−/− mice developed normally, but had a shorter lifespan than Riz1+/− and Riz1+/+ mice, owing to an increase in tumour burden. In fact, 80% of Riz1−/− mice had tumours at 20 months (compared with 30% of Riz1+/+ mice), and almost half of these had tumours in several organs. The predominant tumour type in Riz1−/− mice is malignant B-cell lymphoma, with a histology similar to diffuse large B-cell lymphoma (DLBL). Mutation of Riz1 in a Trp53 +/− background also increased tumour formation, compared with Riz1−/− and Trp53+/− mice, indicating a cooperative effect on tumorigenesis and confirming that Riz1 acts as a tumour suppressor in mice.

So, does RIZ1 also participate in tumorigenesis in humans? Single-strand conformation polymorphism analysis of RIZ1 in 35 human DLBL tumours revealed that ten of these had an abnormal band, corresponding to a single base substitution, C-terminal to the PR domain. Not one of the other 432 tumours tested had a mutation in this region.

The authors also found other mutations in this region of RIZ1, either within or C-terminal to the PR domain. Using its ability to stimulate oestrogen-receptor function in activating target gene promoters as a functional assay, the authors showed that none of the RIZ1 mutations discussed were active.

RIZ1 therefore represents the first in a new class of tumour suppressors — the identification of many more will surely follow. Discovery of RIZ1's substrates should elucidate how this class acts.