Participation of the TBP‐associated factors (TAFs) in cell differentiation

The understanding of the mechanisms that regulate gene expression to establish differentiation programs and determine cell lineages, is one of the major challenges in Developmental Biology. Besides the participation of tissue‐specific transcription factors and epigenetic processes, the role of general transcription factors has been ignored. Only in recent years, there have been scarce studies that address this issue. Here, we review the studies on the biological activity of some TATA‐box binding protein (TBP)‐associated factors (TAFs) during the proliferation of stem/progenitor cells and their involvement in cell differentiation. Particularly, the accumulated evidence suggests that TAF4, TAF4b, TAF7L, TAF8, TAF9, and TAF10, among others, participate in nervous system development, adipogenesis, myogenesis, and epidermal differentiation; while TAF1, TAF7, TAF15 may be involved in the regulation of stem cell proliferative abilities and cell cycle progression. On the other hand, evidence suggests that TBP variants such as TBPL1 and TBPL2 might be regulating some developmental processes such as germ cell maturation and differentiation, myogenesis, or ventral specification during development. Our analysis shows that it is necessary to study in greater depth the biological function of these factors and its participation in the assembly of specific transcription complexes that contribute to the differential gene expression that gives rise to the great diversity of cell types existing in an organism. The understanding of TAFs' regulation might lead to the development of new therapies for patients which suffer from mutations, alterations, and dysregulation of these essential elements of the transcriptional machinery.

A major issue in Developmental Biology is to understand the mechanisms that regulate gene expression for determining cell lineages and the onset of terminal differentiation.These processes consist in a spatiotemporal control that depends upon a collection of processes that involve the participation of (i) transcription factors and their interaction with housekeeping and specific promoters and enhancers (Russo et al., 2018), (ii) chromatin remodeling complexes (Chen et al., 2023), (iii) the epigenetic machinery for DNA and histone modifications (Joseph & Young, 2023;Peng et al., 2023), and (iv) posttranscriptional regulation by noncoding RNAs (Hunkler et al., 2022;Van Wijk et al., 2022).Together, they allow the expression of different phenotypes during embryogenesis, the development of specialized cell types, and the establishment of stem cell reservoirs in mature individuals.
In this review, we will focus on the role of the general transcription factors (GTFs) that constitute the TFIID complex, as recent work in the field has revealed their involvement in the regulation of cell differentiation through changes in TFIID composition to promote the expression of specific phenotypes.

| TRANSCRIPTION FACTORS AND THE REGULATION OF CELL DIFFERENTIATION
The search for a mechanism to explain tissue-specific gene expression began with the study of transcriptional regulation in bacteria.Such studies demonstrated the existence of transcriptional units that respond to specific metabolites and lead to their activation or repression (Jacob & Monod, 1961).After being extended from the lac operon to other systems, it was a conditio sine qua non the existence of gene networks to regulate cell functional states.However, at that time, the intermediate mechanisms were still obscure.Several authors proposed theoretical models to explain eukaryotic cell regulation based on sequence-specific binding RNAs as activators and histones playing a role as repressors (Britten & Davidson, 1969).
Nevertheless, the isolation of bacterial repressors and the demonstration of their specific DNA-binding activities showed that neutral or slightly acidic proteins act as transcription factors (Gilbert & Müller-Hill, 1966;Maniatis & Ptashne, 1973;Pirrotta & Ptashne, 1969;Ptashne, 1967aPtashne, , 1967b;;Riggs et al., 1968), discarding RNAs as candidates to fulfill the role of repressors of gene expression.Nowadays, it is well-known that eukaryotic RNA polymerases cannot recognize and bind to their corresponding gene promoters by themselves.They require the participation of GTFs that recognize certain elements within promoters and recruit the correct RNA polymerase to assemble the preinitiation complex (PIC).The analysis of transcriptional regulation of oncogenic viral sequences and of those mutations that alter development in animal models led to realize that gene expression involves both activators and repressors.Thus, while activators stimulate transcription by interacting with GTFs to facilitate the assembly of the transcription machinery on the promoter, repressors bind to specific sequences near the transcription start site (TSS) to block either the interaction of the RNA polymerase or GTFs with the promoter through protein-protein interactions, or alternatively, they can bind to GTFs to abolish their function in the PIC assembly.
Among the earliest suspected activators/repressors in eukaryotic cells was the SV40 large T antigen, whose binding to DNA activated viral replication, while simultaneously repressed early transcription (Myers et al., 1981;Tegtmeyer, 1972;Tjian, 1978).Likewise, the transcription factor Sp1 emerged as a promoter-specific protein involved in transcriptional activation by RNA polymerase at the SV40 early promoter (Dynan & Tjian, 1983).In the case of animal systems, the study of mutations associated with developmental alterations in Drosophila melanogaster led to the identification of genes such as Krüppel, which acts as a repressor and is essential for the normal embryonic segmentation; its mutation leads to the reversion in polarity at the posterior abdomen and modifies the number of body segments (Wieschaus et al., 1984).
Once it was demonstrated that eukaryotic gene expression was regulated by specific DNA-binding proteins, a plethora of transcription factors with activation or repression activities was described (de Mendoza et al., 2013).Such transcription factors regulate most cellular activities, from the response to external stimuli to cell growth, migration, and differentiation.To date, it is recognized that transcription factors do not necessarily bind to DNA; instead, multiple molecules function by contacting other transcription factors with DNA binding activity or being associated with RNA polymerase to enhance its activity as occurs for the GTFs (Verrijzer, 1996) or for CBP (CREB binding protein) (Akinsiku et al., 2021).
There are two main classes of transcription factors based on their role in gene expression regulation: (i) GTFs, defined as necessary for the basal transcriptional activity of RNA polymerase, and that interact with gene promoters for the formation of the PIC (reviewed in Chen & Pugh, 2021); (ii) transcription factors involved in the control of differential gene expression as a result of their specific binding to DNA sequences located either at the core promoter (CP), at proximal or distal sequences from the TSS, or enhancers (Patient, 1990;Lobe, 1992;reviewed in Polyanovsky & Stepchenko, 1990).
As a result, the analysis of the functions of transcription factors led to the discovery of proteins associated with embryo development and the programming into specialized cell phenotypes.Such is the case of homeotic genes, which encode transcription factors that directly control the pattern of body formation and the development of structures (Gehring & Hiromi, 1986).Other examples are transcription factors functioning as master genes that regulate the determination of specific cell lineages as described for MyoD in myogenesis (Davis et al., 1987;Lassar et al., 1986;Weintraub et al., 1989); NeuroD/BETA2 in neuronal development from ectoderm (Lee et al., 1995); or Skn-1 in keratinocyte differentiation (Cabral et al., 2003), among others.However, few transcription factors promote by themselves the development of cell lineages, as occurs with MyoD or Neuro D. To be more precise, the analysis of tissue-specific gene regulation has shown that regulation of developmental programs requires combinatorial and cooperative interactions among different transcription factors; such interactions lead to establishing complex transcriptional networks implicated in development and differentiation (Bondos, 2006;Bondos & Tan, 2001).An example of these combinatorial interactions is found during lens morphogenesis, which involves a transcriptional network of at least 20 transcription factors interacting sequentially to culminate in the expression of the lensdifferentiation genes (Ogino et al., 2012).
It is important to mention that the role of a transcription factor as a master regulator of a developmental process does not exclude its participation in other processes.This may be exemplified with Pax6, which is a typical case of a transcription factor with multiple activities during development; it is essential for the overall central nervous system and eye development and contributes to the formation of the olfactory system and pancreas (Ashery-Padan & Gruss, 2001;Lopez-Mascaraque et al., 1998;St-Onge et al., 1997;Walther & Gruss, 1991).Its functionality varies depending on the tissue type, the timing of its expression, and its isoforms; these elements determine its interaction with other transcription factors and the transcriptional network to be activated.Thus, during early eye development, Pax6 is essential for expressing Math5, Ngn2, and Mash1 to maintain neural fate in retinal precursors (Marquardt et al., 2001).On the other hand, Pax 6 depends upon the expression of FoxC1 to promote the differentiation of limbal precursors into corneal epithelial cells; the knock-down of the latter promotes an alternative differentiation into epidermal keratinocytes (Li et al., 2021).
These examples illustrate the tremendous complexity of transcriptional regulation.Nevertheless, until now, most attention in the field has forgotten the participation of GTFs in regulating gene expression associated with development and differentiation.In the following paragraphs, we will address the role of the TATA-box binding protein (TBP)-associated factors (TAFs) and their variants when they assemble into the TFIID complex, as well as the participation of TBP variants to constitute alternative transcriptional complexes, to regulate cell differentiation and tissue-specific gene expression.

| Transcription in eukaryotes
Basal gene expression depends on transcription factors that help the RNA polymerases to bind the promoters of those genes that are transcriptionally active.Such factors designated as basal or GTFs are essential for RNA polymerase activities; in their absence, polymerases remain inactive despite strong promoters.
There are three RNA polymerases in eukaryotes (Table 1).RNA polymerase I (RNAP I) which transcribes the 47S pre-rRNA coding the 28S, 18S, and 5.8S ribosomal RNA (rRNA) genes, sequences that are tandemly arranged as a single unit that is repeated between 200 and 600 times along the human genome (Drygin et al., 2010;Girbig et al., 2022).Transcription by RNAP I is tightly linked to demanding cell growth, division, and proliferative metabolism in coordination with the transcriptional activity of RNAP III.At the same time, RNAP III is involved in the transcription of short noncoding RNAs, including tRNAs, the 5S rRNA, U6 snRNA, 7SK RNA, 7SL RNA, Y RNA, MRP RNA, H1 RNA, and the Vault RNA.These small RNAs participate in diverse functions such as translation, elongation by RNAP II, protein translocation to the endoplasmic reticulum, RNA splicing, tRNA precursors processing, replication of mitochondrial DNA, pre-rRNA processing; or through the association with the telomerase-associated reverse transcriptase (TERT) they also downregulate their own levels, as occurs for the MRP RNA (Dergai & Hernandez, 2019;Dieci et al., 2013;Mattijssen et al., 2010;White, 2011).Even though most of the promoters of short interspersed nuclear elements (SINEs) are occupied by the RNAP III, their transcriptional regulation is scarcely known (Dergai & Hernandez, 2019;Dieci et al., 2013;White, 2011).

| Promoters for RNA polymerase I
The function of a promoter is to allow the sequential binding of basal transcription factors to recruit an RNAP for the assembly of the transcription PIC through distinctive elements found in each promoter, which are recognized by specialized GTFs.Consequently, ribosomal promoters comprise the Upstream Control Element (UCE) and the CP, which in turn, also contains the ribosomal gene TSS or ribosomal initiator (rInr) (Girbig et al., 2022).Promoter Selector Factor 1 (SL1) binds to CP in concert with a dimer of the Upstream Binding Transcription Factor (UBTF or UBF) which contains multiple highmobility group boxes to recognize the UCE to promote DNA bending, leading to the recruitment of RNAP I (Drygin et al., 2010).Another form of RNAP I, the RNAP Ia, also possesses catalytic activity but cannot bind to a promoter.SL1 is a protein complex built by TBP and the TAFs TAFIA, TAFIB, TAFIC, and TAFID (Drygin et al., 2010;Tremblay et al., 2022).Then, RRN3 (TIF-IA) binds to the complex, and RNAP I (RNAP Ib) is recruited to the promoter through interactions of UBTF and SL1 with the DNA-directed RNA polymerase I subunit RPA49 (POLR1E/PAF53).UBTF is also involved in chromatin decondensation at ribosomal gene loci by competition with the linker histone H1 (Sanij et al., 2008).
Other components that assemble into the PIC include the protein kinase CK2, the PCAF and the G9a chromatin remodelers, as well as nuclear actin, myosin IC (NM1), Topoisomerases I and IIα, Ku70(XRCC6 or X-Ray Repair Cross Complementing 6), PCNA, TFIIH and the Chimeric ERCC6-PGBD3 protein (CSB/ERCC6 or ERCC Excision Repair 6).

| Promoters for RNA polymerase III
The RNA polymerase III works onto three types of promoters.Then, TFIIIB, a three-subunit protein complex composed of TBP, BDP1, and the TFIIB-related factor BRF1 is mobilized to the promoter, allowing the recruitment of RNAP III to establish the transcription PIC.Vertebrates also possess the BRF2 factor, which can replace BRF1 in TFIIIB (Dieci et al., 2013;Graczyk et al., 1861;Schramm & Hernandez, 2002;White, 2011).On the other hand, type II promoters in all tRNA genes have boxes A and B, which TFIIIC, a six-protein subunit complex, recognizes, allowing the recruitment of T A B L E 1 Target genes transcribed by the eukaryotic RNA polymerases and their corresponding GTFs involved in core promoter recognition and RNA polymerases recruitment for PICs assembly.RNAP III • tRNAs Note: TBP is bolded to remark on its participation in transcription by the three RNA polymerases.
TFIIIB and RNAP III.The third promoter type, such as that of the U6 snRNA coding promoter gene involved in splicing, contains the TATA-box and the proximal specific element (PSE) located upstream of the TSS, although others lack PSE.TFIIIB binds the TATA-box, and PSE is recognized by the snRNA activating protein complex (SNAPc) (Schramm & Hernandez, 2002;White, 2011).For the recruitment of RNAP II to protein-coding gene promoters, a sequential addition of specific GTFs is needed, followed by the Mediator complex and RNAP II (Allen & Taatjes, 2015;Burley & Roeder, 1996;Roeder, 2019).PIC assembly begins with the recognition and binding of TFIID to the CP, followed by the sequential recruitment of TFIIA, TFIIB, the TFIIF-RNAP II complex, TFIIE, and TFIIH (Roeder, 2019) (Figure 1).TFIID is a protein complex formed by the TBP and 13 to 14 TAFs (Antonova et al., 2019;Burley & Roeder, 1996;Hahn, 2004;Patel et al., 2020) (Figure 2).TBP recognizes a specific sequence, the TATA-box, and binds to this motif assisted by TAFs.

| RNA polymerase II uses different promoters
Approximately 20% of the yeast genes contain the TATA-box (Patel et al., 2020).In humans, 24% of the genes have a TATA-boxlike sequence; however, only 2.4% of the human genes have the Transcription by RNA polymerase II requires assembly of a preinitiation complex (PIC).PIC assembly begins with the recognition and binding of TFIID to the core promoter.TFIID is a protein complex formed by the TATA-box binding protein (TBP) and 13 to 14 TBP-associated factors (TAFs) which contribute to transcription of most genes.After TFIID is bound to the gene promoter, assembly of PIC is followed by the sequential recruitment of TFIIA, TFIIB, the TFIIF-RNAP II complex, TFIIE, and TFIIH.
In the opposite direction, the protein-coding gene promoters can also drive the synthesis of small noncoding RNAs whose TSSs are generally located less than 250 nucleotides upstream of the TSS for the protein-coding region.These promoter upstream transcripts or PROMPTs, also known as upstream antisense RNAs (uaRNAs), are unstable.Some PROMPTs can transcribe long noncoding polyadenylated RNAs.The machinery synthesizing these uaRNAs also involves the GTFs and RNAP II recruited by transcription factors bound to the nucleosome-depleted region between these two TSS types (Core et al., 2014;Haberle & Stark, 2018).GTFs and RNAP II are also recruited by transcription factors bound to enhancers for bidirectional production of the small noncoding RNAs known as enhancer RNAs or eRNAs, which have a short half-life (Haberle & Stark, 2018).
Other RNAs transcribed by RNAP II correspond to lncRNAs, RNAs longer than 200 nucleotides that do not encode proteins.They are often modified at their 5' ends through nitrogen methylation at position 7 of guanosine (m7G), splicing, and polyadenylation at their 3'ends; modifications that contribute to their stability (Statello et al., 2021).Those nonpolyadenylated lncRNAs are quickly hydrolyzed by the eukaryotic exosome complex (Nojima & Proudfoot, 2022), although they can also be stabilized by forming secondary structures (Bridges et al., 2021).The Functional Annotation of the Mammalian Genome (FANTOM5) identified a collection of 27,919 human lncRNAs (Hon et al., 2017), while NONCODEV6 database contains a human data set of 96,411 lncRNA genes and 173,112 lncRNA transcripts (Zhao et al., 2021).These lncRNAs include those produced in intergenic regions (lincRNAs), as well as eRNAs, PROPMTs, and antisense RNAs (Gil & Ulitsky, 2020;Nojima & Proudfoot, 2022;Statello et al., 2021).Other types of lncRNAs are the circular RNAs (cirRNAS), which correspond to lncRNAs generated during splicing and covalently linked between their 5' and 3' ends.Recently, it was found that lncRNAs play roles in several biological processes, including differentiation, cell cycle, metabolism, and diseases.These processes might be modulated either by proteins, epigenetic modifications, transcriptional regulation, as well by RNAs stability, translation, posttranslational modifications, and through the interaction with signaling receptors (Bridges et al., 2021;Kopp & Mendell, 2018).
2.5 | The TBP-associated factor IID (TFIID) TFIID is built by TBP and 13 to 14 TAFs in humans and yeasts, respectively (Kolesnikova et al., 2018;Nogales et al., 2017;Sanders & Weil, 2000).However, metazoans possess different TFIID complexes to assure tissue-specific transcription (Müller et al., 2010).In the following paragraphs, we will examine the evidence of the contribution of different TAFs in the regulation of gene expression associated with cell differentiation, maintenance of stem cell reservoirs, and survival.

| CELL DIFFERENTIATION AND THE ROLE OF TAFS IN THE REGULATION OF GENE EXPRESSION
GTFs play a crucial role in the recruitment of RNA polymerase II to the promoters of genes that are being expressed, and therefore, it was thought that they were only involved in basal transcription.
However, several studies provided evidence showing the interaction of TFIID with other proteins bound to sequences located upstream of the promoter region (Horikoshi et al., 1988(Horikoshi et al., , 1991;;Pugh & Tjian, 1990;Smale et al., 1990).Further investigation demonstrated that TFIID was composed of at least eight subunits in addition to TBP (Dynlacht et al., 1991;Tanese et al., 1991;Timmers et al., 1992).We currently know that TFIID is a collection of different complexes whose composition depends on the existence of alternative forms of TBP or members of the TBP family (Mishal & Luna-Arias, 2022;Ohbayashi et al., 2001;Persengiev et al., 2003;Pugh & Tjian, 1991) and tissuespecific TAFs that might lead to cell-type-or stage-specific gene expression (Figure 3).These variations in TFIID composition can explain the selectivity for specific genes that play critical roles in development and cell differentiation, as well as their activation by master transcription factors through protein-protein interactions with TAF variants, and the recruitment of chromatin remodelers or molecules involved in epigenetic control at these gene loci (reviewed in Levine et al., 2014).

| TAF4 and TAF4b
The earliest evidence supporting the participation of specific TAFs in differentiation-linked gene expression derives from experiments that F I G U R E 2 Schematic representation of the canonical TFIID complex.This scheme shows the trilobular structure of canonical TFIID, which comprises the TATA-box binding protein (TBP) and the number of different TBP-associated factors (TAFs) enlisted at the right column.
described the presence of a high molecular weight polypeptide in terminally differentiated B lymphocytes but not in neuroblastoma, glioblastoma, and other cell types (Dikstein, Ruppert, et al., 1996), as well in the granulosa cells of the ovarian follicle (Freiman et al., 2001).
Such factor, named TAFII105 (currently TAF4b), interacted specifically with TAFII250 (TAF1) and seemed to be regulated in a cell typespecific manner, leading to propose its possible role as a coactivator or target for a subset of activators responsible for transcription of genes expressed in B-cells (Dikstein, Zhou, et al., 1996).
In contrast with the ubiquitously expressed TAF4 (TAFII130), TAF4b seems to be a cell type-specific subunit of TFIID (Bahat et al., 2013); moreover, some reports suggest that their activity might be antagonistic to TAF4 activity in embryonic stem cells and liver (Alpern et al., 2014;Bahat et al., 2013).TAF4b expression maintains the stemness and proliferative abilities of mouse embryonic stem cells (mESC); its knockdown facilitates their differentiation after stimulation with retinoic acid (Bahat et al., 2013;Gura et al., 2020).
F I G U R E 3 Tissue-specific gene transcription is mediated by a variety of transcriptional complexes.Canonical human TFIID comprises TBP and 13 to 14 TAFs, which participate in transcription of 20%-30% of yeast and human genes (a).However, recent studies have shown that, besides tissue-specific transcription factors, several components of the core transcription machinery determine cell type-specific gene expression.It was detected the lack of a canonical TFIID complex in human embryonic stem cells (hESCs).Transcription of class I genes such as SOX2 and KLF4 is mediated by a TFIID complex composed of TBP, TAF5, and TAF3 (b); class II genes such as OCT4 and NANOG are transcribed by a TFIID complex which contains TBP, TAF2, TAF3, TAF5, TAF6, TAF7, and TAF11 (d) (Goodrich & Tjian, 2010).In other cases, as occurs during germ cell differentiation, hematopoiesis, and myogenesis, transcription of specific genes is dependent of a TFIID constituted by TBPL2 and TAF3 (c) (Suzuki et al., 2012).Alternatively, the expression of some TATA box-less genes associated to specific proliferative and/or differentiation stages may depend upon a complex constituted by TBPL1 and TAFIIA (e), as occurs for p21 waf1 expression (Suzuki et al., 2012).TAF, TBP-associated factor; TBP, TATA-box binding protein.
The role of TAF4 in cell differentiation is clear both during nervous system development and in epidermal keratinocytes.In the mouse embryo, TAF4 shows low levels at gestational stages E13 and E15 to rise at E17.In postnatal stage P1, its highest levels are found at the cortex, hippocampus, striatum, and at ventral nuclei of the thalamus (Metsis et al., 2001).Then, TAF4 shows its maximal expression in the adult brain at the cortex, cerebellum, and hippocampal formation, particularly enriched in cerebellar Purkinje cells (Metsis et al., 2001).Similarly, 48 h after the induction of teratocarcinoma PCC7 cells into the neuronal lineage, TAF4 expression increased notably, suggesting its possible function as a regulator of neuron differentiation.This result is supported by TAF4 overexpression in nondifferentiated PCC7 cells, which induce the expression of neural differentiation markers such as the brainderived neurotrophic factor (BDNF), the Neurofilament light chain (NF-L) and GAP-43, among others (Metsis et al., 2001).Supporting these results, it was reported that in null-TAF4 mESCs, neuronal differentiation is defective, besides the alteration of other tissues (Langer et al., 2016).
In epidermal keratinocytes, TAF4 inactivation/knockdown has enormous implications for the expression of terminal differentiation leading to an increase in cell proliferation as indicated by the augment in the number of Ki67 positive cells and the upregulation of molecular markers associated with a hyperproliferative phenotype such as amphiregulin or the cytokeratin KRT16 (Fadloun et al., 2007;Kazantseva & Palm, 2014).Such alterations also involve the downregulation of keratinization markers like transglutaminase I, loricrin, filaggrin, and tight junction and cell adhesion components (Fadloun et al., 2007), as well as the upregulation of trichohyalin.Together, these changes suggest that TAF4 inactivation/knockdown promotes the expression of a phenotype like the one expressed by keratinocytes at the hair follicle (DasGupta et al., 2002;Lyle et al., 1999;Purba et al., 2014).Such characteristics explain the early postnatal death associated with an altered skin barrier (Fadloun et al., 2007).
Similarly, it has been demonstrated that during embryo development, TAF4 is essential for liver maturation and the post-natal activation of the hepatocyte gene expression program (Alpern et al., 2014).Its inactivation induces alterations in hepatic parenchyma organization, which is accompanied by a reduction in the number of proliferating hepatocytes in the developing liver, and the downregulation of enzymes involved in glucose and lipid metabolism as well as the disruption of the hepatic epithelium and the blood-bile barrier (Alpern et al., 2014).Such effects lead to an early post-natal death of TAF4-deficient animals (Alpern et al., 2014).

| TAF15
Among the TAFs associated with cell differentiation and the control of tissue-specific gene expression, TAF15 has been most extensively studied because it is also involved in neurodegenerative disorders and cancer (Harrison & Shorter, 2017;Sankar & Lessnick, 2011).
TAF15 is a protein that belongs to the FET family of DNA/RNA binding proteins (FUS, fused in sarcoma/EWS, Ewing sarcoma protein/TAF15) (Kovar, 2011), and it is considered a noncanonical TAF since it is not essential for TFIID complex assembly (Andersson et al., 2008;Tora, 2002).TAF15 knockdown with siRNAs inhibits cell proliferation in cancer cell lines by decreasing the expression of cell cycle-associated proteins (Ballarino et al., 2013).
Also, during mouse nervous system development, TAF15 seems to maintain the same expression levels, although it shows a differentially regulated expression, depending on the cerebral regions analyzed (Svetoni et al., 2017).In such a sense, it was described that TAF15 functionally regulates neuronal transcripts in neurons differentiated from mouse and human ESCs.Among these transcripts, the NMDA receptor subunits, potassium voltage-gated channels, neurexins 1 and 3, neuroligin 1, and protocadherin-9 stand out; and the use of siRNAs to knock down TAF15 expression induced a 75% decrease in the expression of the corresponding mRNAs and proteins (Ibrahim et al., 2013).On the other hand, TAF15 is differentially expressed in various murine tissues, suggesting its possible role in tissue-specific gene regulation, showing high levels in the testis, thymus, lung, liver, and spleen, but low levels in the brain, cerebellum, and muscle (Melot et al., 2001).
Moreover, it has acetyltransferase activity on H3 and H4 histones (Mizzen et al., 1996) and on TFIIEb and TFIIF (Imhof et al., 1997), besides being able to mediate ubiquitination of H1 histone (Crane-Robinson, 1999;Pham & Sauer, 2000).Based on such a variety of biochemical activities and its association with diseases, it has been proposed that TAF1 is a major transcription factor that acts as a transcriptional coactivator as well as a repressor, being essential for development and regulating processes such as cell cycle (Hilton & Wang, 2003;Li et al., 2004;Zeng et al., 2022), cell proliferation, and viability (Wassarman et al., 2000).
Four TAF1 isoforms were described in Drosophila, generated by alternative splicing.They show differential expressions both in tissues with different proliferative capacities (i.e., testes, ovaries, and salivary glands), and after DNA damage (Katzenberger et al., 2006).These results suggested that TAF1 isoforms are implicated in initiating gene-specific transcription.Evidence supporting this proposal comes from studies showing that twi and sna activation during Drosophila development indirectly depend on TAF1 (Pham et al., 1999).Such effect seems to be mediated by TAFII110 (human homolog TAF4) and TAFII60 (human homolog TAF6), which in turn are the target of Dorsal and Twist, driving the development of mesoderm-determining genes in the fly (Pham et al., 1999).
Likewise, it was reported that during the spermatocyte growth phase in Drosophila, TAF1 and testis-specific TAFs (tTAFs) follow a similar expression pattern (Metcalf & Wassarman, 2007).This phase, characterized by its high transcriptional activity and TAF1 expression, is restricted to transcriptionally active premeiotic cells, consistent with its role as a component of TFIID.Such behavior led to suggest that TAF1 and tTAFs-in a developmentally regulated event-become components of a distinctive tTFIID in which TAF4 and TAF5 are replaced by their tTAF paralogs Nht and Can for regulation of transcription in spermatocytes (Chen et al., 2005;Metcalf & Wassarman, 2007).
TAF1 is also involved in early brain development and alterations associated with nervous system pathologies.It was described that TAF1 is associated with a gene network that regulates early brain development (Choi et al., 2016;Hurst et al., 2018).Research has identified eight TAF1 splice variants expressed in the human fetal brain (Herzfeld et al., 2013), whose involvement in brain development and function is unclear.However, mutations of this factor are involved in pathologies such as dystonia (Newman et al., 2012), parkinsonism (Herzfeld et al., 2007;Zeng et al., 2022), intellectual disability (Hurst et al., 2018), and other neurological alterations (O'Rawe et al., 2015).Moreover, TAF1 overexpression is found in gliomas interacting with the NOTCH1 signaling pathway (O'Rawe et al., 2015).This effect depends on the overexpression of the lncRNA FOXD2-AS1, which induces TAF1 overexpression and NOTCH1 pathway activation, leading to repress glioma stem cell differentiation and stimulating the expression of the stem cell markers Oct4, Sox2, Nanog, Nestin and CD133) (Wang et al., 2022).

| TAF7 and TAF7L
TAF7 is an intron-less gene that encodes a GTF which, together with TAF2, directly interacts with TAF1 to nucleate the assembly of the TFIID complex (Chen et al., 2021), functioning as a checkpoint regulator that suppresses transcription initiation until PIC is assembled; thereafter, it is released from TAF1 allowing the transcription initiation (Gegonne et al., 2006).Accumulated evidence suggests that TAF7 is required for cell proliferation, while TGFβ arrest of mammary epithelial cell proliferation depends on its degradation (Nakagawa et al., 2018), and its deletion in mouse embryonic fibroblast cell lines impairs proliferation (Nakagawa et al., 2018).In contrast, TAF7 is unnecessary for cell differentiation, as suggested by several laboratories (Gegonne et al., 2012(Gegonne et al., , 2013)).
Contrary to TAF7, its paralogue TAF7L is important in cell differentiation processes.Initially, TAF7L was considered a spermatogenesis-specific component of TFIID since it was coimmunoprecipitated together with TBP from spermatocyte and haploid cell extracts (Pointud et al., 2003).In this case, TAF7L participates by interacting with TRF2 to regulate spermatogenesis and metabolism; TAF7L-KO mice become infertile, with an abnormal sperm structure, low sperm count, and reduced motility (Pointud et al., 2003;Zhou, Grubisic, et al., 2013).These results led to the suggestion that TAF7L may function as a repressor or a positive regulator in the haploid spermatocyte.
Nevertheless, more recent work has revealed that TAF7L carries out an essential function in the fate determination of mesenchymal cells.Specifically, TAF7L is an important regulator of stem cell differentiation into adipocytes.The first results suggesting its participation in adipogenesis derived from observations showing that during C3H10T½ and 3T3-L1 cells differentiation, TAF7L underwent upregulation, in contrast with most of the TFIID canonical subunits (Zhou, Kaplan, et al., 2013).Moreover, such an increase seemed to be adipose-specific, since TAF7L was downregulated during myogenesis (Zhou, Kaplan, et al., 2013).Further experiments using shTAF7L cells inhibited adipose conversion, while ectopic expression of TAF7L in depleted cells restored their adipogenic abilities (Zhou, Kaplan, et al., 2013).Similar results were observed when TAF7L was knocked out in mice, impairing adipose tissue development.Furthermore, ectopic expression of TAF7L in C2C12 myoblasts promoted an increase in the expression of the transcription factors PPARγ and C/EBPα, associated with adipose differentiation (Zhou, Kaplan, et al., 2013).Other experiments suggested that TAF7L might be playing a common role in the differentiation of both white (WAT) and brown (BAT) adipose tissues (Zhou et al., 2014).
Information on the role of TAF9 in cell differentiation is scarce.
Earlier studies suggested that TAF9 (TAFII32a) is highly expressed in rodents' bone marrow, fibroblasts, and megakaryocytes (Thompson & Ravid, 1999).Since this factor reduced its expression in differentiated cells, it was proposed that TAF9 is mainly involved in the proliferation of poorly differentiated cells (Thompson & Ravid, 1999).In contrast, TAF9b is associated with neuronal differentiation.While it is found at low levels in mESCs, after in vitro differentiation promoted by the combined effect of retinoic acid and the SMO receptor agonist SAG, TAF9b expression increased about 10-fold (Herrera et al., 2014).
These results were also found in newborn spinal cord neurons, suggesting that TAF9b is induced during neuronal differentiation.
Such proposal was supported by experiments that showed that TAF9b knocking-down blocks the development of a neuronal phenotype as demonstrated by the downregulation of the neuronal-specific markers Lhx3, Lhx4, Isl1, Mnx1, and Tubb3 (Herrera et al., 2014).Interestingly, in motor neurons, TAF9b seems to be mainly associated with the PCAF complex instead of the canonical TFIID complex, supporting its role as a tissue-specific transcriptional regulator (Herrera et al., 2014).

| TAF10
TAF10, previously known as TAFII30, was cloned and characterized after its discovery as a protein required for transcriptional activation of the estrogen receptor AF-2 (Jacq et al., 1994).TAF10 is also part of the PCAF histone acetylase complex, the TATA-binding proteinfree TAF complex (TFTC), and the STAGA transcription coactivator-HAT complex (Martinez et al., 2001;Zhao et al., 2008).Moreover, TAF10 has been associated with human dilated cardiomyopathy (Li et al., 2018) and Systemic Lupus Erythematosus (Peng et al., 2022), among other diseases.
The possible participation of TAF10 in the regulation of cell proliferation and differentiation is concluded from experiments showing that TAF10 disruption through a Cre recombinase-LoxP strategy, stopped proliferation of undifferentiated F9 cells; once TA10 expression was restored using a Dox-inducible TAFII30 expression system, cells grew up to confluence (Metzger, 1999).In these experiments, cell growth arrest was accompanied by a decrease in the expression of cyclin E and a reduction in pRb phosphorylation (Metzger, 1999).In addition, knocking out of TAF10 in F9 cells impaired their differentiation into parietal endoderm by treatment with Retinoic Acid and cAMP, but not into primitive endoderm as determined by quantification of thrombomodulin (Metzger, 1999).
In further experiments, this group also disrupted TAF10 in C57BL/6 mice, showing that this factor is essential for development since null animals were not viable 5-6 days after postcoitus, and demonstrating that lacking TAF10, the TFIID complex barely possesses TBP (Mohan et al., 2003).Analysis of TAF10 expression along normal mouse development shows variable levels of the corresponding mRNA in a tissue-dependent manner.So, high TAF10 mRNA levels were observed mainly in epithelia and ectodermalderived tissues.In contrast, the urogenital ridge or heart did not express detectable TAF10 (Mohan et al., 2003), suggesting its possible participation in tissue-specific gene expression.
Given the evidence that revealed the role of TAF10 in developing epithelia in mice embryos, Indra et al. studied the effect of TAF10 ablation in fetal and adult epidermal keratinocytes.Because of the lack of TAF10, newborn mice did not survive for more than 24 h, suggesting an impaired epidermal barrier, a result which was confirmed by whole mount permeability assays and by measurement of transepidermal water loss (Indra et al., 2005).These observations correlated with a decrease in the number of both corneodesmosomes and lamellar bodies, and a size reduction of the epidermal cornified layer (Indra et al., 2005).In contrast, TAF10 knockout in adult epidermal keratinocytes did not promote effects on cell proliferation and differentiation (Indra et al., 2005).

| TAF3, TAF6, and TAF8
Given the scarce evidence on the participation of TAF3, TAF6, and TAF8 in cell differentiation and tissue-specific gene expression, this section will discuss important data supporting their possible biological function.
TAF3 is a core component of TFIID which is necessary for myogenesis.It was demonstrated that muscle differentiation implies a pronounced reduction in the expression of TAF1, TAF4, and TBP both at mRNA and protein levels, which is also accompanied by a reduction in the expression of TAF9 and TAF10 while levels of other GTFs remain unaltered (Deato & Tjian, 2007).These observations in C2C12 cells suggested that the canonical TFIID complex is lost during myogenesis, raising the need for an alternative complex to activate gene transcription.Further confirmation derives from primary myoblasts.Together, the studies in C2C12 cells and primary myoblasts demonstrated the existence of a switch of the core transcriptional complex involving the activity of TAF3 and the presence of a protein partner that could function as a TBP substitute.
After some experiments which showed similar expression patterns and the interaction of TAF3 and TBPL2 (also known as TBP2 or TRF3), authors concluded that TAF3/TBPL2 constitutes a complex that substitutes the conventional TBP/TFIID (Deato & Tjian, 2007;Deato et al., 2008).Such proposal was supported by using RNAi to knock down either TAF3 or TBPL2, which in both cases inhibited myoblast differentiation as determined by quantification of myogenic regulators such as myogenin and Myf5 (Deato & Tjian, 2007), and whose promoters are regulated by a MyoD mediated activation of a TAF3/TBPL2 complex (Deato et al., 2008).
However, opposed to these results, RT-qPCR and Western blot studies demonstrated that C2C12 cells and muscle stem satellite cells (MuSCs) express the TBP gene but not the TBPL2 gene, and that the differentiation of myoblasts into myotubes requires TBP assembled into the TFIID complex (Malecova et al., 2016).Moreover, they also demonstrated that a TBPL2 -/-null mutant mouse has unaffected the muscle regeneration process and MuSCs-derived from this model, also showed differentiation into myotubes, supporting that TBP is essential for differentiation of myoblast into myotubes.
On the other hand, TAF6 is required for a variety of developmental processes, as suggested by alterations such as the frameshift mutation c.1052delT or the missense mutation c.323T>C (Ile108Thr), which causes in humans, intellectual disability, nystagmus, left ventricular hypertrophy, short stature, hypermetropia, and other changes.These mutations and alterations of other genes lead to phenotypes known as the Alazami-Yuan syndrome (Lin et al., 2022;Tuc et al., 2020).
In Drosophila, TAF6 is required for cell fate specification in the eye.Its mutation causes atypical eye development with an abnormal number of photoreceptor cells, and loss of function results in lethality (Aoyagi & Wassarman, 2001).Also, this transcription factor is necessary for imaginal development, spermatogenesis, and oogenesis (Aoyagi & Wassarman, 2001;Hiller et al., 2001).
Finally, TAF8 has been linked to neurodegenerative disorders characterized by hypomyelination, a thin corpus callosum, and brain atrophy (Wong et al., 2022).Such alterations are associated with heterozygosity for several gene variants, including the intronic TAF8 variant GRCh38:chr6:g.42050590A>G and the exonic TAF8 variant GRCh38:chr6:g.42057513G > A (Wong et al., 2022).It has been reported that TAF8 is an integral subunit of the functionally active TFIID, essential for integrating TAF10, an action mediated by its histone-fold domain (Guermah et al., 2003).
In vitro, during adipose conversion of 3T3-L1 cells, TAF8 expression is increased explicitly from practically imperceptible levels (Guermah et al., 2003), in contrast with no increases during C2C12 myoblast differentiation.TAF8 interacts directly with C/EBPα and PPARγ promoters in this case.However, it does not influence C/EBPβ expression (Guermah et al., 2003), suggesting that in contrast with TAF7L, which might be part of the regulatory machinery that switches on adipose tissue development (Zhou, Kaplan, et al., 2013;Zhou et al., 2014), TAF8 could only be involved in the regulation of adipose phenotype expression.

| TBP family members in development and differentiation
TBP has a central role in global gene expression of protein-coding genes.
TBPL1 (TRF2), a metazoan TBP paralog sharing 40% identity with TBP, plays an essential role in transcription in Zebrafish, Xenopus, and Drosophila embryogenesis (Müller et al., 2001;Veenstra et al., 2000), it is essential for mice spermiogenesis, but it has no role in mouse embryogenesis (Martianov et al., 2001(Martianov et al., , 2016;;Mishal & Luna-Arias, 2022;Zhang et al., 2001).TBPL1 associates with the GTF IIA subunit 1 like (GTF2A1L) and remains in the cytoplasm as a complex with several heat shock proteins in mice.When TBL1 is recruited to active gene promoters in haploid cells, it drives their expression associated with TAF7L, thus establishing specific complexes to turn on some genes.However, TBPL1 is not essential for cell viability (Martianov et al., 2016).
Another member of the TBP family is TBPL2 (TRF3), also known as TBP2, because of the 95% similarity with the TBP core domain (reviewed in Mishal & Luna-Arias, 2022).This protein has been found located in loops of the lampbrush chromosome at meiosis I of Xenopus oocytes, which are transcriptionally active (Akhtar & Veenstra, 2009).
TBPL2 also mediates transcription by RNAP II in oocytes and during their meiotic maturation, but its role in embryonic development is still controversial.TBPL2 works in zebrafish differentiation pathways and ventral specification during frog development (Bártfai et al., 2004;Jacobi et al., 2007;Persengiev et al., 2003).It has also been involved in the zebrafish hematopoiesis initiation (Hart et al., 2009;Persengiev et al., 2003).Due to its high similarity with TBP, TBPL2 can replace it during transcription in Xenopus oocytes (Jallow et al., 2004).
TBPL2 only expresses in mice oocytes and might regulate the differentiation of female germ cells, but not embryonic development (Gazdag et al., 2007(Gazdag et al., , 2009)).TBPL2 substitutes TBP during oocyte growth but does not replace TBP in TFIID; instead, it is associated with TFIIA and TFIIB (Tora & Vincent, 2021;Yu et al., 2020).
Interestingly, TBPL2 has a robust selectivity for TATA-like CPs.
Furthermore, oocytes lacking the two copies of the TBPL2 gene showed downregulation of 1802 genes, spotlighting TBPL2 as an emerging player in gene expression and regulation of oocyte growth (Tora & Vincent, 2021;Yu et al., 2020).
The complexity of the machinery involved in transcription regulation during embryonic cell differentiation was recently evidenced using murine embryonic stem cells (mESc) with strong depletion of TBP levels.These cells did not show alterations in their global gene expression pattern dependent on RNAP II transcription, suggesting no dependency on TBP or TBPL1.In contrast, the RNAP III transcription was severely affected by the absence of TBP (Kwan et al., 2023).Thus, there is a need to identify all protein complexes involved in development and differentiation to understand these processes and spotlight the key players involved.

| TAFs and disease
Just as the differential expression of TAFs is involved in the control of cell differentiation and development or in the maintenance of "stemness" (see above), numerous reports show the participation of these transcription factors in disease and cancer development.In this regard, one of the first reports was the description of the chromosomal region known as Ewing Sarcoma Breakpoint region 1 (EWSR) (Delattre et al., 1992), which encodes a multifunctional protein with the same name, and which belongs to the FET family of proteins that includes Fused in Sarcoma (FUS) and TAF15 (Andersson et al., 2008;Bertolotti et al., 1996).Currently, it is known that the 5' regions of the human FET genes are the result of chromosomal translocations leading to gene fusions to various transcription factor coding genes in multiple human malignancies, being considered the cause of cancer development in those patients that express these fused genes (Law, 2006;Riggi et al., 2007), as occurs in acute lymphoblastic leukemia (Kim et al., 2016;Nyquist et al., 2011), and in Ewing Sarcoma which is an aggressive cancer of bone and soft tissue (Picard et al., 2022).
Likewise, it has been reported that TAF12, in association with Nfyc and Rad54l, is required to initiate severe hyperplasia and choroid plexus carcinomas in mouse brains (Tong et al., 2015); as well as the alterations of TAF1 and TAF1L lead to tumorigenesis in gastric and colorectal cancers (Oh et al. 2015(Oh et al. , 2017)).In addition, high-grade serous ovarian cancer is associated either with amplifications, gain in the number of gene copies, and mRNA upregulation of TAF2, TAF4, and TAF4B or with deletions and downregulation of TAF9 (Ribeiro et al., 2014).
On the other hand, there exists another group of diseases that have been recently associated with transcriptional dysregulation.
Given the enormous importance of mutations and alterations that lead to dysregulation of the expression of different TAFs in cancer and other diseases, it is necessary to understand how these transcription factors are regulated.Understanding their activity as specific transcriptional regulators, as well as their association with signaling pathways, will allow the development of treatments for such diseases.Several research groups have proposed the possible use of antibodies with therapeutic and diagnostic use, as suggested for a human antibody with affinity for TAF15, that inhibits neoplastic cell motility and cell adhesion (Schatz et al., 2010).
Others attempted a pharmacological approach by developing new compounds that, at nanomolar concentrations, inhibit growth of Ewing's sarcoma tumor cells (Esfandiari Nazzaro et al., 2021).Thus, in the next future, research should focus on TAFs as novel targets for therapeutic intervention.

| CONCLUSION
Cumulative pieces of evidence obtained in the last decades indicate that TAFs in the TFIID complex are essential for the assembly of the transcription PIC, as well as for the communication with those transcription factors bound at the proximal promoters or enhancers through protein-protein interactions and for the targeting of diverse signaling pathways and epigenetic marks that drive and modulate gene expression at specific moments of the cell cycle.This complexity increases with the involvement of TAF and TBP variants, as well as other members of the TBP family such as TBPL1 and TBPL2, which might suggest a repertory of TFIID complexes specialized in the expression of gene sets necessary for the proper establishment of cell lineages and specialized phenotypes (Akhtar & Veenstra, 2011), as shown for human embryonic stem cells (Maston et al., 2012), and for muscle development (Deato & Tjian, 2007;Deato et al., 2008) (Figure 3).Therefore, the selective expression and use of TAFs and TBP variants needs to be explored to characterize those mechanisms which participate in gene regulation during development and differentiation, which by now, are not-so-well studied.
Promoters type I and II contain an internal control region (ICR) located downstream of the TSS.In type I promoters ICR is conserved among different species and is localized within the transcribed region of the 5S rRNA gene.It contains Boxes A and C separated by an intermediate element (IE).Box C is recognized and bound by the basal transcription factor GTF3A (TFIIIA), while Boxes A and IE are involved in the TFIIIA-dependent recruitment of GTF3C (TFIIIC).
Metazoan promoters utilized by RNA polymerase II are classified into two main types based on the number of TSS, CpG content, epigenetic tags, presence of nucleosomes, and their covalent modifications (Bhuiyan & Timmers, 2019; Haberle & Stark, 2018; Lenhard et al., 2012).Type I promoters require a precise TSS, low CpG content, imprecisely positioned nucleosomes, and frequently harbor the TATA-box and Inr element.The nucleosomes in this region contain histones with the H3K4me3, H3K4me2, and H3K27ac epigenetic marks involved in active chromatin.These promoters regulate the expression of terminal differentiation-specific genes in adult tissues; therefore, they are known as adult promoters.Type II promoters are called broad promoters because they are TATA-less and possess several TSS.They have a high CpG content and are considered ubiquitous because they are expressed in all cell types and regulate the expression of housekeeping genes.They also have a nucleosome-depleted region flanked by nucleosomes containing the H3K4me3, K3K4me2, and H3K27ac characteristic of active chromatin.A third type of promoter has been recognized, corresponding to developmental genes involved in patterning and morphogenesis.They do not contain a TATA-box or an Inr motif.Instead, they have large CpG islands encompassing the promoter and some regions of the structural gene, and they are tagged with the H3K4me3 and H3K27ac marks in the region that contains the TSS, although they might possess the repressive epigenetic mark in H3K27me3 (Haberle & Stark, 2018).
TATA-box consensus sequence TATAWAWR, localized at position −31 to −23 relative to the TSS, where W can be an A or a T, and R = G or C(Haberle & Stark, 2018;Yang et al., 2007).Other elements found in human promoters are: (i) the Inr motif found in nearly 46% of the genes and localized at positions −3 to +3; (ii) the TFIIB recognition element (BRE), which can be found upstream (BREu) and downstream (BREd) at positions −38 to −32, and −23 to −17, respectively; (iii) the Downstream Promoter Element (DPE), first described in Drosophila and that could be found in humans; (iv) the downstream core elements (DCE) I, II and III found at +6 to +11, +16 to +21, and +30 to . TAF1 possesses various biochemical activities.It contains N-terminal and C-terminal Ser/Thr kinase domains, which can catalyze protein autophosphorylation, or phosphorylation of other transcription factors, as occurs with p53