Mechanisms by which a DNA-binding protein could function as a repressor and inhibit transcription (see Milestone 2) were established by the early 1970s. For example, a repressor could bind to DNA at a region that overlapped with a promoter, thereby preventing the RNA polymerase holoenzyme from binding. By contrast, mechanisms through which a DNA-binding protein could act as an activator and stimulate transcription (see Milestone 4) were not immediately obvious, and were not defined until the 1980s and 1990s.
By the beginning of the 1980s, Mark Ptashne and co-workers had shown that the bacteriophage lambda repressor not only repressed transcription at the lambda PR and PL promoters, but also activated transcription at the lambda PRM promoter. In 1982 and 1983, the Ptashne laboratory isolated mutants of the lambda repressor that had lost the ability to activate transcription, but retained the ability to repress transcription, and therefore were presumably still able to bind to DNA. The mutations all occurred in a small surface-exposed region located on the part of the lambda repressor that was expected to be closest to RNA polymerase when the two were bound, side-by-side, at the PRM promoter. Ptashne proposed that activation required a protein–protein interaction between the lambda repressor and RNA polymerase, and postulated that the mutations disrupted this interaction. In 1994, using suppression genetics, Susskind and co-workers obtained compelling support for the proposed interaction and identified its target as a surface in the 
-subunit of RNA polymerase holoenzyme.
Analogous results were obtained for transcription activation at the Escherichia coli lac promoter by catabolite gene-activator protein (CAP; also known as the cyclic-AMP-receptor protein or CRP). In 1993, Richard Ebright and co-workers isolated mutants of CAP defective in activation at lac but not in DNA binding, and showed that they mapped to a small surface patch located on the part of CAP expected to be closest to RNA polymerase when the two were bound at lac. Then, in 1994, using protein–protein photocrosslinking, the Ebright laboratory showed that this surface patch contacts RNA polymerase in the CAP–RNA-polymerase–lac complex, and identified the target of the contact as a surface within the carboxy-terminal domain of the 
-subunit of RNA polymerase.
Taken together, the results with lambda and lac led to the simple 'velcro' model, in which activation requires small surface patches on the activator (the 'activating region') and the transcription machinery (the 'activation target'), and involves direct protein–protein interaction between the two. Activation proceeds through a 'recruitment' mechanism, in which 'adhesive' interactions between the activator and the general transcription machinery facilitate assembly of a stable and catalytically competent transcription-initiation complex.
Subsequent work has defined the protein–protein interactions in the lambda and lac systems in atomic detail. Although transcriptional activation in other systems can be more complex, the model derived from the lambda and lac systems has been shown to apply to many, perhaps most, examples of activation, both in prokaryotes and in eukaryotes.
