Epigenetics and transcription regulation during eukaryotic diversification: the saga of TFIID

Abstract

The basal transcription factor TFIID is central for RNA polymerase II-dependent transcription. Human TFIID is endowed with chromatin reader and DNA-binding domains and protein interaction surfaces. Fourteen TFIID TATA-binding protein (TBP)-associated factor (TAF) subunits assemble into the holocomplex, which shares subunits with the Spt-Ada-Gcn5-acetyltransferase (SAGA) coactivator. Here, we discuss the structural and functional evolution of TFIID and its divergence from SAGA. Our orthologous tree and domain analyses reveal dynamic gains and losses of epigenetic readers, plant-specific functions of TAF1 and TAF4, the HEAT2-like repeat in TAF2, and, importantly, the pre-LECA origin of TFIID and SAGA. TFIID evolution exemplifies the dynamic plasticity in transcription complexes in the eukaryotic lineage.

Keywords: SAGA; TFIID; basal transcription; phylogenetic analyses.

Figure 1.

Structural variation between human (h) and yeast (y) TFIID and SAGA complexes. Shared TAFs between TFIID and SAGA may reflect a common ancestral origin for the two complexes (here “ancestor?”). Reduction of shared TAFs between TFIID and SAGA in human versus yeast Saccharomyces cerevisiae as well as loss of epigenetic domains in S. cerevisiae (e.g., TAF1 BrDs [bromodomains] and TAF3 PHD) indicate divergence in TFIID and SAGA adaptation to transcriptional requirements across different eukaryotic branches (Matangkasombut et al. 2000; Gangloff et al. 2001a; Spedale et al. 2012). Unique and shared subunits as well as epigenetic reader domains are color-coded as indicated.

Figure 2.

Inferred evolutionary history of TAF1 and TAF2. (A) TAF1 is duplicated in the Old World monkeys. BrD is gained in the ancestor of metazoa and lost in dikarya. Streptophyta acquired a ubiquitin-like domain. (B) TAF2 contains previously unrecognized HEAT2-like repeats. Various BrDs were acquired early in fungal evolution and subsequently lost late in fungi. Duplications are represented as red arrows; gradient domains are not predicted in all species of that respective (super)group.

Figure 3.

Inferred evolutionary history of TAF3, TAF8, and SPT7. (A) TAF3 arises from a duplication of a shared ancestor of TAF8 in opisthokonta. TAF3 acquired a PHD, which is secondarily lost in late fungi. (B,C) SPT7 duplicated either in the ancestor of the amoebozoa (B) or pre-LECA (C), implying differential loss. Metazoan SPT7 lost its BrD. Duplications are represented as red arrows.

Figure 4.

Evolutionary history of the relative invariable TFIID subunits. (A) TAF5 duplicated in the ancestor of animals and contains seven WD40 repeats. (B) TAF6 duplicated in the ancestor of animals. TAF5 and TAF6 paralogs subfunctionalized to either SAGA or TFIID. (C) TAF9 duplicated in placentalia but did not subfunctionalize to SAGA. Duplications are represented as red arrows; gradient domains are not predicted in all species of that respective (super)group.

Figure 5.

Inferred evolutionary history of TAF4/Ada1 and the TAF12 HF partner. (A) TAF4 duplicated in the ancestor of vertebrates through a WGD. Afterward, an additional small-scale duplication took place, named TAF4x, which is lost in tetrapoda. The RST domain is acquired in the ancestor of streptophyta, while the NHR1 domain is acquired in animals specifically. (B) TAF12 duplicated in the angiosperm through a WGD. (C) TAF4 and Ada1 emerged through a pre-LECA duplication and subfunctionalized to either SAGA or TFIID. WGD events are represented as blue arrows.

Figure 6.

Inferred evolutionary history of TAF11/TAF13/SPT3. (A) SPT3 is the ancestral protein that gave rise to TAF11 and TAF13 through a duplication followed by a gene fission. (B) TAF11 and TAF13 are ancestral and gave rise to SPT3 through independent duplications followed by gene fusion. WGD events are shown in blue arrows.

Figure 7.

Model of TFIID and SAGA evolutionary divergence from pre-LECA until fungal and metazoan ancestors. In a pre-LECA, the ancestral repertoire (green) of TFIID and SAGA was completely shared. Through duplication and subfunctionalization of the resulting paralogs, the complexes diverged to share fewer subunits throughout eukaryotic evolution (pink and gray). Metazoan TFIID acquired several lineage-specific paralogs (e.g., TAF1L, TAF4B, TAF4x, TAF7L, and TAF9B). Epigenetic domains are differentially gained and lost in metazoan and fungal TFIID and SAGA: Metazoan TFIID acquired epigenetic domains (double BrDs in TAF1 and a PHD in TAF3), while metazoan SAGA lost BrD in SUPT7L (retained in fungal SAGA); in contrast, fungal TFIID gradually lost the TAF3 PHD and carries only one BrD in TAF1 (in some late fungi, the BrDs are completely lost). Additionally, fungal TAF2 displays dynamic gains and losses of numerous BrDs, in contrast to metazoan TAF2. Unique and shared subunits as well as dynamics in epigenetic reader domains are color-coded as indicated.