The transcriptional program regulated by the tumor suppressor p53 was analy
sed using oligonucleotide microarrays. A human lung cancer cell line that e
xpresses the temperature sensitive murine p53 was utilized to quantitate mR
NA levels of various genes at different time points after shifting the temp
erature to 32 degreesC. Inhibition of protein synthesis by cycloheximide (C
HX) was used to distinguish between primary and secondary target genes regu
lated by p53. In the absence of CHX, 259 and 125 genes were up or down-regu
lated respectively; only 38 and 24 of these genes were up and down-regulate
d by p53 also in the presence of CHX and are considered primary targets in
this cell line. Cluster analysis of these data using the super paramagnetic
clustering (SPC) algorithm demonstrate that the primary genes can be disti
nguished as a single cluster among a large pool of p53 regulated genes. Thi
s procedure identified additional genes that co-cluster with the primary ta
rgets and can also be classified as such genes, In addition to cell cycle (
e.g. p21, TGF-beta, Cyclin E) and apoptosis (e.g, Fas, Bak, IAP) related ge
nes, the primary targets of p53 include genes involved in many aspects of c
ell function, including cell adhesion (e.g, Thymosin, Smoothelin), signalin
g (e.g. H-Ras, Diacylglycerol kinase), transcription (e.g. ATF3, LISCH7), n
euronal growth (e.g. Ninjurin, NSCL2) and DNA repair (e.g. BTG2, DDB2), The
results suggest that p53 activates concerted opposing signals and exerts i
ts effect through a diverse network of transcriptional changes that collect
ively alter the cell phenotype in response to stress.