Such is the mix-and-match nature of cell biology that many different modes of regulation come up time and again. Take, for example, the addition or removal of an acetyl group, which is a well-known means of regulating histones and turning gene expression on or off. But acetylation can also be used to control the activity of p53, and in the 16 November issue of Nature Wei Gu and colleagues describe one way in which this process might be regulated.

Acetylation by p300 stimulates the ability of p53 to bind DNA in a sequence-specific manner. Because p53 acetylation is enhanced when cells are treated with deacetylase inhibitors, there is likely to be a balance between the two processes, with the deacetylases fine-tuning the levels of acetylated p53.

If this is the case, reasoned Gu and colleagues, p53 would be expected to interact physically with a deacetylase complex. They used a glutathione S-transferase (GST)–p53 affinity column to test this hypothesis and, among the eluted proteins, they pulled down the human histone deacetylase HDAC1. The two proteins do not interact directly, however, as purified recombinant HDAC1 failed to bind p53.

So what mediates the interaction? Gu and co-workers purified complexes containing HDAC1, then ran these fractions down the GST–p53 column. Only one protein remained bound after dissociation of the HDAC1 complex, and the authors named this PID (for 'p53 target protein in the deacetylase complexes'). They then showed, by immunoprecipitation, that the interaction between p53 and PID is specific and direct both in vitro and in vivo.

Sequence analysis revealed that PID is homologous to human MTA1, a protein found in the so-called NuRD complexes involved in nucleosome remodelling and histone deacetylation. Confirmation that the p53/PID/HDAC1 complex, too, is a NuRD complex came with the identification of other members, among their number a chromatin-remodelling protein.

The final test was to see whether PID is involved in deacetylation in vivo. As expected, expression of PID inhibited p53 acetylation in an HDAC1-dependent manner. It also affected the cellular functions of p53; for example, treatment with PID blocked p53-mediated growth inhibition, expression of endogenous p21 and apoptosis. It seems, then, that Gu and colleagues have identified a protein that tips the balance towards the deacetylation — and, hence, inactivation — of p53, with obvious implications for its role in human cancer.