A great deal has been learned in recent years about transcriptional control in eukaryotes. Genes encoding some 1400 transcription factors can be recognized in the human genome. We now have a glimpse of how the astronomical number of possible combinations of these transcription factors can generate the complexity of gene control required to produce organisms as remarkable as those we see around us. But very much remains to be understood. Although we now have some understanding of what processes turn a gene on and off, we have very little understanding of how the frequency of transcription is controlled in order to provide a cell with the appropriate amounts of its various proteins. In a red blood cell precursor, for example, the globin genes are transcribed at a far greater rate than the genes encoding the enzymes of intermediary metabolism (the so-called housekeeping genes). How are the vast differences in the frequency of transcription initiation at various genes achieved? What happens to the multiple interactions between activation domains, co-activator complexes, general transcription factors, and RNA polymerase II when the polymerase initiates transcription and transcribes away from the promoter region? Do they completely dissociate at promoters that are transcribed infrequently, so that the combination of multiple factors required for transcription must be reassembled anew for each round of transcription? Do complexes of activators with their multiple interacting co-activators remain assembled at promoters from which reinitiation takes place at a high rate, so that the entire assembly does not have to be reconstructed each time a polymerase initiates transcription?

Much remains to be learned about the structure of chromatin and how that structure influences transcription. What additional components besides HP1 and methylated histone H3 lysine 9 are required to direct certain regions of chromatin to form heterochromatin, in which transcription is repressed? Precisely how is the structure of chromatin changed by activators and repressors, and how do these changes promote or inhibit transcription? Once chromatin-remodeling complexes and histone acetylase complexes become associated with a promoter region, how do they remain associated with it? Current models suggest that certain subunits of these complexes associate with modified histone tails so that the combination of binding to a specific histone tail modification plus modification of neighboring histone tails in the same way results in retention of the modifying complex at an activated promoter region. In some cases, this type of assembly mechanism causes the complexes to spread along the length of a chromatin fiber. What controls when such complexes spread and how far they will spread?

Single activation domains have been discovered to interact with several co-activator complexes. Are these interactions transient, so that the same activation domain can interact with several co-activators sequentially? Is a specific order of co-activator interaction required? How does the interaction of activation domains with Mediator stimulate transcription? Do these interactions simply stimulate the assembly of a preinitiation complex, or do they also influence the rate at which RNA polymerase II initiates transcription from an assembled preinitiation complex?

Transcriptional activation must be a highly cooperative process because genes expressed in a specific type of cell are expressed only when the complete set of activators that control that gene are expressed and activated. As mentioned earlier, some of the transcription factors that control expression of the TTR gene in the liver are also expressed in intestinal and kidney cells. Yet the TTR gene is not expressed in these other tissues, since its transcription requires two additional transcription factors expressed only in the liver. What mechanisms account for this highly cooperative action of transcription factors that is so critical to cell-type-specific gene expression?

The discovery that long noncoding RNAs can repress transcription of specific target genes has stimulated tremendous interest. Do they always repress transcription by targeting Polycomb complexes? Can long noncoding RNAs also activate transcription of specific target genes? How are they targeted to specific genes? Do the roughly 1600 long noncoding RNAs that are conserved among mammals all function to regulate transcription of specific target genes, adding to the complexity of transcriptional control by sequence-specific DNA-binding proteins? Research to address these questions will be an exciting area of investigation in the coming years.

A thorough understanding of normal development and of abnormal processes associated with disease will require answers to these and many related questions. As further principles of transcriptional control are discovered, applications of that knowledge are likely. This understanding may allow fine control of the expression of genes introduced by gene therapy vectors as they are developed. Detailed understanding of the molecular interactions that regulate transcription may provide new targets for the development of therapeutic drugs that inhibit or stimulate the expression of specific genes. A more complete understanding of the mechanisms of transcriptional control may allow improved engineering of crops with desirable characteristics. Certainly, further advances in the area of transcriptional control will help to satisfy our desire to understand how complex organisms such as ourselves develop and function.