6.4 Synthetic Yeast Promoters 119Hybrid promoters containing heterologous TFBSs need a heterologous tran-
scription factor. To ensure orthogonality, this transcription factor should not
have any other target than the heterologous promoter itself. The modular struc-
ture of natural transcription factors suggests that it is possible to combine differ-
ent protein domains to obtain new factors.
A protein able to bind DNA can stimulate transcription when fused to an acti-
vation domain. The first heterologous transcription factor tested in yeast con-
tained the bacterial DNA‐binding protein LexA fused to an activation domain
and triggered transcription of promoters containing LexA TFBSs [34, 108, 109].
Transcription activators containing the bacterial DNA‐binding protein tetR are
regulated by tetracycline [97, 110], which prevents the binding of tetR to the
cognate TFBSs [111–113]; therefore, the expression of the target gene can be
modulated by adjusting the concentration of this chemical in the culture medium.
A reverse tetR mutant, which binds its TFBS upon addition of tetracycline, is also
available. However, transcription activators containing reverse tetR have a rela-
tively strong basal activity in the absence of tetracycline [114].
Hybrid promoters containing artificial TFBSs require the construction of
artificial DNA‐binding domains. Zinc fingers and transcriptional activator‐like
effectors (TALEs) are short peptidic modules binding to specific and short DNA
sequences. Protein engineering has diversified these modules, and libraries of
protein moieties recognizing virtually all DNA sequences of three to four nucleo-
tides have been constructed. By fusing several zinc fingers or TALE modules,
it is possible to obtain arrays that specifically bind longer DNA sequences
[89, 115, 116]. Artificial transcription activators are obtained by fusing these
DNA‐ binding domains to activation domains [100, 106].
In general, any mechanism able to target an activation domain to the DNA can
be used to stimulate transcription initiation. In the clustered regularly inter-
spaced palindromic repeats (CRISPR)‐derived system, the target DNA sequences
are identified via RNA‐mediated interactions, instead of binding of a protein. In
this system, the DNA‐binding activity consists of a single‐guide RNA (sgRNA),
which targets specifically the DNA region to be regulated, and the catalytically
inactive version of the protein Cas9 (dCas9), which binds specifically the sgRNA.
By fusing dCas9 to an activation domain, a transcription activator is obtained
[117, 118].
The activation domain stimulates transcription initiation by establishing
protein–protein interactions with coactivators and components of the transcrip-
tional machinery [119, 120]. While DNA‐binding domains have well‐defined
conserved architectures (reviewed in [12]), activation domains do not share
common structures, except for a marked acidity [35, 121]. They usually consist of
multiple unstructured acidic patches; each acidic patch triggers transcription
initiation when fused to a DNA‐binding domain [122]. Any peptide stretch
displaying such properties can be used to activate transcription [123].
The DNA‐binding and activation activities do not need to reside on a unique
protein but can be physically separated on two different molecules. An interac-
tion between these two is sufficient for transcription initiation. This is exploited
in the yeast two‐hybrid assay [124, 125]. This principle was also used to construct
a light switchable system. Here, the DNA‐binding domain and the activation