The silencing protein is a Salmonella histone-like protein called H-NS. The group was searching for direct targets of this known repressor when they noticed that the H-NS binding sites were GC poor (∼47%) compared with the rest of the chromosome (∼52%). A GC-poor foreign gene that the group recombined into Salmonella was also repressed by H-NS.
Most bacteriophage and other bacteria are lower in GC content than Salmonella and its relatives, so invading DNA is an obvious target for H-NS. “It's like a primitive immune system,” says Fang. “Reduce their expression, and the foreign genes can be tolerated.”
Useful newcomers might eventually be expressed, however, via mutations that increase their GC content or through the evolution of antisilencers. Many of Salmonella's foreigner-derived virulence genes, for instance, are shut off by H-NS but can be reactivated when needed by a transcription factor called SlyA.
Bacteriophage, of course, also evolved means to get around this defense system. Some encode their own H-NS antagonists, whereas others maintain GC-neutral genomes.
Other GC-rich bacteria have related DNA-binding proteins that are possible analogues of H-NS. Bugs with AT-rich genomes might in turn have GC-binding repressors. If universal, this immune strategy would explain why each bacterial species maintains its distinctive GC/AT ratio.
H-NS seems to recognize short stretches of DNA, although how the protein reads GC content within just a couple hundred base pairs is not clear. Perhaps it recognizes the intrinsically curved or partially melted structure of AT-rich sequences.
Others have found that tandemly bound copies of H-NS form multimers. This probably blocks transcription by compacting the DNA in that vicinity. By locking many helical turns in place, it would also restrict changes in DNA superhelicity, thus explaining the known repressive effect H-NS has on heat- and salt-induced responses.