Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

History: The Canonical Model for the Regulation of RAR-Target Gene Expression

The Functional Domains of RARs

Like most nuclear receptors (NRs) [49–51], RARs and RXRs exhibit a well-defined domain organization consisting of a central conserved DNA binding domain (DBD) linked to a variable N-terminal domain (NTD), and a C-terminal Ligand-Binding Domain (LBD) [10, 28, 73, 108] (Fig. 6.1a).

Fig. 6.1
figure 1

Structure of RARs and of their DNA binding sites. a RARs depict a domain organization with an unstructured N-terminal domain (NTD), and two well structured domains: a central DNA binding domain (DBD) and a C-terminal ligand-binding domain (LBD). The phosphorylation sites located in the NTD and the LBD are shown. b Structural changes induced upon RA binding. The crystal structures of the unliganded RXRα and liganded RARγ LBDs are shown with the binding domains for corepressors, coactivators and cyclin H. Helices are represented as ribbons and labelled from H1 to H12. Adapted from Protein Data Bank 1lbd and 2lbd. c The retinoid response elements (RAREs) are composed of a direct repeat of the motif 5’- Pu G (G/T) TCA spaced by 0 (DR0), 1 (DR1), 2 (DR2), 5 (DR5) or 8 (DR8) base pairs. DR8 comprise three half sites with DR2 and DR0 spacing. Some RARE-associated genes are shown. d Binding of the RAR/RXR heterodimers to DR1 and DR5 RAREs (adapted from [104])

Full size image

The structures of RAR and RXR LBDs are rather similar. This domain in each protein is composed of 12 conserved alpha helices and a beta-turn, separated by loops and folded into a three-layered, and parallel helical sandwich [20, 28, 103], with helices H4, H5, H8, H9 and H11 sandwiched between H1, H2 and H3 on one side, and H6, H7 and H10 on the other (Fig. 6.1a, b). In contrast, the C-terminal helix, H12, is more flexible and adopts conformations that differ from one RAR to the other. The conformation pf H12 also changes after RA binding.

The primary feature of the LBD is its functional complexity. It contains the Ligand-Binding Pocket (LBP) [19], the heterodimerization surface [21], and interaction surfaces involved in the binding of multiple coregulators (Fig. 6.1b). A well-described hydrophobic surface, generated by H3 and H4, is involved in the binding of corepressors/coactivators [57, 111] (see also Chap. 3 in this volume). The LBD also contains a recently described docking site for cyclin H, a subunit of the cyclin-dependent activating (CAK) sub complex of the general transcription factor, TFIIH, that is formed by loop L8-9 and the first amino acids of H9 [15] (Fig. 6.1b).

The DBD is composed of two zinc-nucleated modules and two alpha-helices [137, 138], which contribute to a second dimerization interface and define the contacts for specific DNA sequences, named RA response elements (RAREs). Classically, RAREs are composed of two direct repeats of a core hexameric motif (A/G) G (G/T) TCA separated by 1, 2 or 5 nucleotides and referred as DR1, DR2 and DR5 [8, 10, 48] (Fig. 6.1c). However, recent genome wide chromatin immunoprecipitation coupled with deep sequencing (ChIP-seq) technology has allowed identification of new RAR binding loci [85, 91] and revealed that RARs can occupy a larger repertoire of sites with an unexpected diversity in the spacing and the topology of the DNA binding elements, including DR0, DR8 and IR0 (inverted repeats) elements (Fig. 6.1c). Recent structural studies have also indicated that the architecture of DNA-bound heterodimers is dictated by the DNA sequence (Fig. 6.1d) [22, 104] (see also Chap. 2 in this volume).

In contrast to the DBD and the LBD, the NTD is not conserved between RARs and RXRs and even between the different subtypes and isoforms. As yet, high-resolution, three-dimensional structures of this region have not been produced [108]. Several biochemical and structural studies coupled to structure prediction algorithms suggest that the NTDs of RARs and RXRs, as well as of any member of the NR family, are naturally disordered [74, 131]. An interesting feature of these NTDs is that they contain phosphorylation sites [105], which are conserved between RARs (Fig. 6.1a) [112]. Moreover, they contain proline-rich motifs (PRMs) which are well known to bind proteins with Src-homology-3 (SH3) or tryptophan-tryptophan (WW) domains. Phosphorylation prevents or favors these interactions [7].

RAR-Mediated Gene Expression

RAR/RXR heterodimers control transcription via several distinct mechanisms, including both repression and activation. According to the canonical model, the transcriptional regulation of RA-target genes relies not only on the binding of RAR/RXR heterodimers to specific RAREs, but also on corepressors that dissociate and coactivators that associate with the LBD upon ligand-induced conformational changes [57, 81, 95, 133]. At the molecular level, the discrimination between corepressors and coactivators is governed by the ligand-induced, orientation of H12, which contributes in a critical manner to the generation or removal of cofactor interaction surfaces.

Repression of Transcription in the Absence of Ligand

In the absence of ligand and in a context of chromatin where the nucleosomes do not impede binding to RAREs, the RARα subtype is a strong repressor of target gene expression (Fig. 6.2a) [34]. In this unliganded state, H12 adopts an open conformation that unmasks a hydrophobic groove generated by H3 and H4 [76] (see also Chap. 3 in this volume). This interface specifically binds an LxxI/HIxxxI/L motif in the extended alpha helix box of the corepressors, NCoR or SMRT [98]. NCoR and SMRT are genetic paralogs with multiple protein variants, but SMRT is the favored corepressor for RARs [90]. According to recent studies, SMRT is recruited by the heterodimer only through the RAR partner [76] (see also Chap. 3 in this volume).

Fig. 6.2
figure 2

The classical model of activation of RA-target genes. a Repression in the absence of ligand. b Corepressors and coactivators exchange after ligand binding. c Recruitment of the transcription machinery and initiation of transcription. d End of the RA signal upon recruitment of non-conventional coativators such as RIP140, associated to large complexes with chromatin repressing activity. The end of the RA signal occurs also through the degradation of RARs by the ubiquitin proteasome system

Full size image

SMRT does not have intrinsic enzymatic activity, but serves as an adaptor to recruit other high molecular weight complexes that are endowed with histone deacetylase activity (HDACs) [95]. These complexes deacetylate lysine residues in the N-terminal tails of histones and maintain chromatin in a condensed and repressed state over the target promoter [34, 111] (Fig. 6.2a). The corepressor complexes also contain other components such as transducer β-like proteins (TBL1 and TBLR1), which serve as adaptors regulating corepressor assembly and function [94].

In contrast to RARα, the RARγ and RARβ subtypes poorly interact with corepressors [39, 60, 102], most probably due to the fact that, in these receptors, H12 interacts with H3 even in the absence of ligand, thus occluding the corepressor docking site.

Initiation of Transcription in Response to the Ligand: A Process Governed by the LBD via the Exchange of Coregulators

According to the canonical model, ligand binding to RARs must be understood in terms of structural features (Fig. 6.1b). When entering the cavity of the RARa LBP, the ligand induces a β-strand-to-α-helix secondary structure switch [76], which induces repositioning of H11 relative to H10 and a concomitant swinging of H12 inward to pack against H3 and H4 in a mouse trap model that locks RA in the LBP. Consequently, corepressors are released and a new hydrophobic cleft is formed between H3, H4 and H12 [28], with a charge clamp between a conserved glutamate residue in H12 and a lysine in H3. This charge clamp specifically grips the ends of the helix specified by the LxxLL motif of the p160 subfamily of steroid receptor coactivators, SRC-1, SRC-2 and SRC-3 [81, 96]. The p160 coactivators have an intrinsic histone acetyl transferase (HAT) activity, and according to recent structural studies, only one coactivator molecule is recruited by the heterodimer through the RAR partner [22, 92].

When recruited to the LBD of the liganded RAR, the p160 coactivators initiate a dynamic, ordered and coordinated recruitment of other proteins with HAT activity [p300/CBP (CREB binding protein) and p/CAF (p300/CBP-associated factor)] or with histone methyl transferase (HMT) activities, such as Coactivator-associated arginine methyl-transferase 1 (CARM1) or Protein-arginine methyl transferase 1 (PRMT1) [57, 81, 111, 133] (Fig. 6.2b). Acetylation and methylation weaken histone DNA contacts and create marks forming an «histone code», which coordinates the recruitment of additional HATs or HMTs to histones for further chromatin decompaction. This code also orchestrates the recruitment of chromatin remodelers, which use the energy of ATP-hydrolysis to reposition nucleosomes by sliding them in cis or displacing them in trans, allowing the formation of nucleosome-free or nucleosome-spaced regions at the promoter.

The p160 coregulators also recruit large complexes which contain modules with other enzymatic activities, such as histone de-ubiquitinases [139] and histone lysine methyl transferases [42, 80]. Lysine methylation creates marks for the binding of other enzymes that erase repressive marks. According to a recent study, this step requires Poly (ADP-Ribose) glycohydrolase (PARG) which is corecruited at RAR-dependent promoters in response to RA [78]. Thus, due to their varied composition, the coactivator complexes provide an elegant mechanism to reorganize chromatin by writing activating histone marks, erasing repressive marks, and remodeling nucleosomes.

All of these steps pave the way for recruitment of the transcription machinery to the promoter region (Fig. 6.2c) [34, 111] including RNA Polymerase II, General Transcription Factors, and a specific subunit (DRIP205/TRAP220) of the multisubunit Mediator complex. According to recent studies, the transcription machinery assembles sequentially with nucleotide excision repair (NER) factors, in order to achieve optimal histone modifications, and thus, efficient RNA synthesis [77].

According to recent chromatin conformation capture technologies, RARs that bind to different enhancer elements of a gene can form loops [25]. An emerging view is that RARs, similar to other TFs, promote the formation of long range chromatin loops, bridging genomic loci located even on different chromosomes, thus creating hot spots of transcription [115].

Turn “OFF” of Transcription

After the “ON” switch, transcription of the RA target genes has to be terminated. Several scenarios have been proposed, but it is still unclear whether termination activities are gene or cell-specific. One possibility is that liganded RARs recruit unconventional coregulators with LxxLL motifs which, in contrast to classical p160 SRCs, inhibit, rather than activate, the transcriptional activity of RARs (Fig. 6.2d). These coregulators include the receptor interacting protein of 140 kDa (RIP140/NRIP1) [62], the preferentially expressed antigen in melanoma (PRAME) [38], and the transcription Intermediary factor-1 alpha (TIF1α/Trim24) [75]. The mechanism of TIF1α-mediated repression has not been elucidated yet [67], but the repressive activities of RIP140 and PRAME have been attributed to the recruitment of HDACs and PcG proteins, respectively [38, 132].

According to another scenario, an efficient way to limit RAR function and/or to signal the end of the transcriptional process would be the degradation of RARs and RXRs by the ubiquitin proteasome system [126] (Fig. 6.2d). Supporting such a hypothesis, RARs have been shown to be ubiquitinated and degraded by the proteasome through the recruitment of TRIP1/SUG-1, which is a subunit of the 19S regulatory sub complex of the proteasome with an ATPase activity [41, 52].

Development of the Field: RARs also Have Non-canonical Extra-Nuclear Effects, Which Are Integrated in the Nucleus

RA Activates Kinase Signaling Pathways via a Pool of RARs Present in Membrane Lipid Rafts

It is now appreciated that RARs have additional extra-nuclear and non-transcriptional effects that activate kinase signaling pathways [4]. Studies from several laboratories have demonstrated that RA rapidly (within minutes after RA addition) and transiently activates several kinase cascades. RA activates p38 mitogen activated kinase (MAPK) in fibroblasts, mouse embryo carcinoma cells, mammary breast tumor cells, and leukemia cells [5, 25, 52, 100]. RA activates the p42/p44 MAPKs (also called Erks) in neurons, Sertoli cells, and embryonic stem cells [30, 33, 59, 87, 123].

Since the RA-induced activation of the MAPK pathways occurs after the activation of upstream cascades involving RhoGTPases [5, 33, 100], PI3 kinase and/or protein kinase B (PKB)/Akt [31, 87, 93] in the cytosol, it has been suggested that the RA-induced cytosolic activities must involve an atypical, non-genomic event similar to that described for steroid NRs [82, 99]. In line with this concept, RARα proteins have recently been found in lipid rafts isolated from the membranes of several cell types [87, 100] (Fig. 6.3). Moreover, the activation of p38MAPK has been shown to involve the interaction of RARα present in lipid rafts with Gαq proteins [100] (Fig. 6.3). However, the activation of Erks by RA did not involve Gαq proteins (Piskunov et al., unpublished results), but rather PI3K [31, 87, 93] or the Src kinase [33] (Fig. 6.3). Thus depending on the cell type, the extra nuclear effects of RA appear to involve different mechanisms and kinase cascades.

Fig. 6.3
figure 3

Activation of kinase cascades by an extra nuclear pool of RARs. A subpopulation of RARα is present in membrane lipid rafts. Depending on the cell type, in response to RA, this pool of RARα can either interact with Gαq proteins or with PI3K to activate the p38 or p42/p44 MAPK pathways respectively. In other cell types, RARγ can also activate the p42/p44MAPK pathway via Src. Then the activated MAPKs translocate into the nucleus where they phosphorylate and activate MSK1

Full size image

Once activated by RARα, p38MAPK and Erks translocate to the nucleus where they phosphorylate MSK1 (Fig. 6.3) [99]. Finally MAPKs and MSK1 phosphorylate several nuclear factors involved in the expression of RA-target genes, including RARs themselves and their coregulators.

In the Nucleus, RARs Are Rapidly Phosphorylated by a Cascade of Kinases

A few years after the cloning of RARs, it emerged that these receptors are also phosphoproteins [45, 106, 107, 109, 110]. However, at that time, the analysis of RAR phosphorylation was a challenging task because of its highly dynamic nature and also because of the low ratio of phosphorylated versus non-phosphorylated RARs that are found in vivo [71]. The early studies of RAR phosphorylation required radioactive material and large amounts of recombinant NRs overexpressed in cultured cells or of bacterially expressed NRs purified and phosphorylated in vitro with different kinases. Though technically limited, these studies resulted in the identification of a number of phosphorylation sites in RARs and RXRs [1, 79, 105, 106, 110, 121]. Two main RAR phosphorylation sites were identified in intrinsically disordered regions; one in the loop between helices 9 and 10 of the LBD (S369 in RARα) and the other one in the NTD (S77 in RARα) [25] (Fig. 6.1a). The serine located in the LBD belongs to an Arginine-Lysine-rich motif and can be phosphorylated by several kinases such as the cyclic AMP dependent Protein Kinase (PKA) [110, 114] or MSK1 [25, 112]. In contrast, the serine located in the NTD belongs to a proline-rich motif and is phosphorylated by cdk7/cyclin H [106, 130], a kinase that belongs to the CAK subcomplex of the general transcription factor TFIIH. Most importantly, the correct positioning of the cdk7 kinase and thereby the efficiency of the NTD phosphorylation by cdk7 relies on the docking of cyclin H at a specific site in the LBD located in L8-9 and the N-terminal part of H9 [15] (Fig. 6.1b).

More recently, the emergence of new methods for enrichment of phosphopeptide samples, as well as the availability of phosphospecific antibodies, made it possible to analyze endogenous RAR phosphorylation. These studies revealed that the serines located in the LBD and in the NTD are both rapidly phosphorylated in response to RA via a cascade of kinases (Fig. 6.4). First, RA-activated MSK1 phosphorylates RARα at the serine located in the LBD [25]. Then, phosphorylation of this residue promotes phosphorylation of the serine located in the NTD through subtle conformational changes [25, 43]. Molecular dynamic simulations of the RARα LBD showed that phosphorylation of S369 (located in loop L9-10) leads to changes in the structural dynamics of the cyclin H binding site (composed of loop L8-9) situated at a 30 Å distance. This change in dynamics has been correlated with an increase in cyclin H binding and phosphorylation of the NTD at S77 by cdk7 [29, 112]. Thus, the coordinated phosphorylation of RARα results from a coordinated cascade that can be explained by changes in the structural features of the molecule.

Fig. 6.4
figure 4

RARs are phosphorylated by a cascade of cascades. Activated MSK1 phosphorylates RARs at a serine residue located in the LBD (loop L9-10). Phosphorylation of this residue induces conformational changes in loop L8-9, which promote the binding of the cyclin H subunit of the CAK subcomplex of TFIIH. Consequently the cdk7 kinase can phosphorylate the serine residue located in the NTD. In the case of RARγ, phosphorylation of the NTD induces the dissociation of vinexin β. Finally the phosphorylated RAR can be recruited to response elements located in the promoters of target genes

Full size image

This phosphorylation cascade has been described in cells that respond to RA via the activation of p38MAPK [25]. Whether it also occurs in cells that respond via the activation of Erks requires further investigations. Remarkably, the two phosphorylation sites are conserved between the mammalian RAR subtypes (α, β and γ) [112] (Fig. 6.1a) and the RARγ subtype is also phosphorylated at the same residues [9, 70, 110] through a similar cascade (our unpublished results). It is interesting to note that the serine residue located in the NTD has been conserved during evolution of chordates, indicating that the phosphorylation of this residue is likely important for RARs activity. In contrast, the serine residue located in the LBD is not present in non-mammalian RARs, suggesting that in other vertebrates, the phosphorylation cascade described above does not occur. Consequently, in other vertebrates, the phosphorylation of the NTD would be controlled by different regulatory circuits [112].

Consequences of RARs Phosphorylation at the NTD: Dissociation of Coregulators and Degradation by the Proteasome

The NTD is an intrinsically disordered region [74, 131], but the serine residue of this domain belongs to a PRM (Fig. 6.1a), which can form polyproline helices and bind proteins with SH3 or WW domains [65, 84, 136]. Recently, our laboratory identified vinexinβ as a new binding partner for the RARγ PRM [17]. Vinexinβ is an adaptor protein characterized by the presence of three SH3 domains, the third C-terminal one interacting with the PRM of RARγ. Recently, the combination of nuclear magnetic resonance (NMR), Circular dichroïsm, small angle X-ray scattering (SAXS) and molecular dynamics simulations revealed that phosphorylation of the serine residue located in the PRM of RARγ changes the global hydrodynamic behavior of the polyproline helix and decreases the propensity of the PRM to bind SH3 domains (Kieffer et al., unpublished results). Consequently, vinexinβ dissociates from RARγ [70] (Fig. 6.4).

In addition, at the end of the transcriptional process, phosphorylation of the N-terminal serine residue of RARγ and of an additional one located at position -2 has been shown to promote the ubiquitination of the receptor and its subsequent degradation by the proteasome [52, 68]. This is a typical example of interplay between different posttranslational modifications [118].

The N-terminal PRM of RARα has been shown to bind the proline isomerase, Pin1, in a phospho-dependent manner [23, 53]. Pin1 is a WW domain-containing protein that is well known to induce cis-trans isomerization of proline residues that follow phosphorylated serines, and in so doing, to create new specific recognition sites for other interacting factors [135]. Pin1 interaction has been correlated with the degradation of RARα by the proteasome and the inhibition of RARα activity [53]. However the mechanism of the Pin1-mediated degradation of RARα remains to be defined.

Not only RARs but also Several Other Proteins Integrate MAPK Signaling and Become Phosphorylated in Response to RA

RA-activated MAPKs and MSK1 also phosphorylate other factors involved in RA target gene transcription, their phosphorylation favoring programs induced by the ligand. MSK1 is recruited at RAR-target promoters where it phosphorylates histone H3 tails at serines S10 and S28 (Fig. 6.5a). H3S10 phosphorylation has been correlated with the recruitment of HATs and the SWI/SNF ATP-dependent chromatin-remodeling complex [25, 99], while H3S28 phosphorylation induces the displacement of PcG complexes that maintain chromatin in a repressive state [47].

Fig. 6.5
figure 5

Working models for the role of phosphorylations in the activation of RAR-target genes. When RARs already occupy RAREs in the absence of ligand, the RA-activated MAPKs phosphorylate components of the corepressor complexes such as SMRT and TBLR1. (b) Phosphorylation promotes their dissociation and their degradation by the proteasome, thus facilitating their exchange for coactivators. Then the coactivators such as SRC3 become also phosphorylated (c). Subsequently, they dissociate from RARs and are degraded, allowing the recruitment of other coregulators. Histones are also phosphorylated (a) and their phosphorylation induces the recruitment of HATs and remodeling complexes. Altogether these events cooperate to decompact chromatin at the promoters and pave the way for the recruitment of the transcription machinery (g). When RAREs are not occupied, phosphorylation of histones (a) and RARs (d) cooperate for the recruitment of RARs at their response elements (e). Next, liganded and DNA bound RARs recruit coactivator complexes, which decompact further chromatin. As above, coactivators phosphorylation (f) leads to their dissociation and degradation by the proteasome, thus facilitating the dynamics of coregulators exchange and the recruitment of the transcription machinery (g). Finally, RARs are degraded by the ubiquitin-proteasome system

Full size image

In addition, MSK1 and the upstream kinases, p38MAPK and Erks, phosphorylate several other actors in RA signaling, including RXRs, corepressors and coactivators. The important point is that phosphorylations alter protein structure, protein-protein interactions and protein activity, thus constituting an important cellular integration mechanism.

As an example, RA-activated p38MAPK rapidly phosphorylates rapidly RXRα at three residues located in the NTD [24, 128] by an as yet, unknown mechanism. MAPKs also phosphorylate components of corepressor complexes, such as SMRT and TBLR1 (Fig. 6.5b). Phosphorylation of SMRT induces its release from RARα [64] and disrupts its interaction with HDACs and other proteins in the corepressor complexes [129]. Consequently, the architecture, composition, and function of the corepressor complexes are disrupted. A current model of TBLR1phosphorylation is that this adaptor mediates recruitment of the ubiquitin proteasome system to ubiquitinate and degrade NCoR, SMRT and HDACs [94, 97]. It has been proposed that the phosphorylation-dependent dissociation and degradation of components of the corepressor complexes mediates the exchange of corepressors with coactivators.

Coactivators are also phosphorylated in response to RA [54]. The p160 coactivator, SRC-3, is phosphorylated by p38MAPK at a serine residue located in the vicinity of the RAR binding domain. Phosphorylation of this residue results in dissociation of SRC3 from RARα. It also marks SRC-3 for ubiquitination and degradation by the proteasome [40] (Fig. 6.5c). This phosphorylation-ubiquitination-degradation process facilitates the dynamics of RARs-mediated transcription by allowing other coregulators to bind. The other components of the coactivator complexes, such as p300/CBP, can be also phosphorylated by several kinases in response to several signaling pathways [94, 96, 97], but whether they are phosphorylated in response to RA requires further investigation. Overall, coregulators respond to several signals that fine-tune their functional interactions with RARs and thus, their ability to modulate RAR transcriptional activity.

Current State of the Field

The RA-Induced Kinases and RAR Recruitment to DNA

Recent ChIP-seq profiles have confirmed that the occupancy of many RAREs is increased in response to RA [25, 70, 85, 89], and that there is a significant correlation between transcription activation and the binding of RAR/RXR heterodimers to DNA [89]. Though the mechanism of RAR/RXR recruitment to DNA in response to RA is still ill-defined, one cannot exclude a role for phosphorylation processes.

In the absence of RA, many RAREs are inaccessible due to a compact epigenetic landscape of chromatin. This implies that RAR binding requires an initial rapid modification of the chromatin environment in order to alleviate compaction and make the RAREs accessible. Among the candidates for chromatin reorganization, there are the RA-activated kinases. Indeed, RA-activated MSK1 is rapidly recruited to RAR-target genes promoters and phosphorylates histones H3 at serines S10 and S28 [25, 47]. Once phosphorylated, these serine residues are marks that induce the recruitment of remodeling complexes and the displacement of repressive complexes [47, 55, 56].

Another possibility is that the phosphorylation state of RARs themselves control their recruitment to DNA. To validate such a hypothesis, RNA-seq and ChIP experiments were performed with mutant mouse embryonic stem cell (mESC) lines expressing RAR phosphomutants in a RAR null background. Such experiments highlighted direct target genes whose expression is controlled by phosphorylation of the N-terminal serine residue of RARγ [2]. These studies also revealed that only RAREs with specific spacings recruit the phosphorylated form of RARγ in response to RA [2].

Then, the question was how phosphorylation of the N-terminal serine could promote recruitment of RARγ to specific RAREs in response to RA. It must be noted that this serine belongs to a PRM that is located in the vicinity of the DBD and which interacts with vinexin β, a repressor of RARγ-mediated transcription (see discussion, above). Remarkably, vinexin β bound to the nonphosphorylated form of RARγ prevents DNA binding, while vinexin β dissociation upon RARγ phosphorylation allows the binding of the receptor to DNA [70] (Fig. 6.4). Thus phosphorylation of the NTD that occurs in response to RA would promote DNA binding via the dissociation of proteins that occlude the DBD.

A Working Model for the Role of Phosphorylation in the Activation of RAR-Target Genes

RA-induced phosphorylation processes play an important role in the expression of RAR-target genes through modulating RAR recruitment to DNA, the sequential recruitment of the different classes of coregulators, and also the stability of the target proteins, thus constituting an important cellular integration mechanism. Two models can be proposed for the role of phosphorylation in the activation of RAR target genes. One for RARs already bound to DNA and another for RARs that are recruited to RAREs in response to RA (Fig. 6.5).

When RARs are constitutively bound to their DNA targets, ligand binding is the crucial molecular event that switches transcription from repression to activation via coregulator exchanges and chromatin reorganization. However it is now evident that RA-activated kinases provide additional layers of regulation for this switch through a phosphorylation code examplified by the phosphorylation of the corepressors and coactivators that promotes their dissociation from RARs and their degradation by the proteasome (Fig. 6.5b, c). Such a code facilitates the exchange of coregulator complexes and the reorganization of the epigenetic landscape. The kinases also phosphorylate histones, introducing additional marks for the recruitment of activating complexes and dissociation of repressive cofactors (Fig. 6.5a).

Not all RAREs are occupied in the absence of RA due to a compact genetic landscape. In this case, it is hypothesized that RA-activated kinases promote recruitment of RARs to DNA by first, phosphorylating histones, a process that alleviates chromatin compaction (Fig. 6.5a). The RA-activated kinases also phosphorylate RARs, but this process controls the recruitment of the receptors only to a subset of RAREs with a specific spacing (Fig. 6.5d, e). Additional approaches are required to investigate why phosphorylation controls RARs recruitment only to certain RAREs and not to the others. When RARs are recruited to DNA, it is evident that phosphorylation of the different coactivators also facilitates the dynamics of their association-dissociation for further chromatin decompaction and recruitment of transcription machinery (Fig. 6.5f, g). Finally, the phosphorylation of RARs signals their degradation by the ubiquitin proteasome system in order to stop the transcriptional process [16] (Fig. 6.5g).

Relevance: In vivo Relevance of the Cross Talk Between the Nuclear and Extra-Nuclear Effects of RARs

Embryonic Development and cell differentiation

Several genetic approaches performed in animals demonstrate that RARs and RXRs are the conductors of RA signaling during development [86, 113]. However, without further development of appropriate technology, current animal models cannot be used to study the role of RAR and RXR phosphorylation due to the complexity of the processes and signaling pathways. Instead, cell differentiation models have provided interesting tools to study the influence of RAR phosphorylation. These have included mouse embryo carcinoma cells (F9 cell line) which markedly resemble embryonic cells from the blastocyst and differentiate into primitive, parietal or visceral endoderm-like cells [18] after RA addition. They also included mouse embryonic stem cells that are pluripotent cells which self-renew indefinitely and have the propensity to differentiate in vitro into a larger variety of cell types [58, 134], such as neurons in response to RA [13].

Experiments from our laboratory revealed that differentiation of F9 cells into primitive endoderm [125] and ES cells into neurons [2] involves the RARγ2 subtype. Most interestingly, the generation of stable rescue cell lines expressing RARγ2 phosphomutants in a RARγ null background indicated that phosphorylation of the RARγ2 NTD is critically required for the RA-induced differentiation of these cells [2, 125]. Moreover, recent genome wide RNA-seq analysis experiments highlighted a subset of genes belonging to early phosphoRAR-regulated gene programs that are critical for triggering the effects of RA [2, 3]. These data suggest an important role for RAR phosphorylation in RA signaling, and pave the way for further investigations during embryonic and tissue development.

Cancer and Diseases

Available evidence suggests that the integrity of signaling pathways is required for the proper activity of RARs. Consequently, one can speculate that deregulation of the “kinome” would have deleterious downstream effects. Accordingly, in Xeroderma Pigmentosum patients, who are characterized by mutations affecting subunits of the general transcription factor, TFIIH, cdk7 does not efficiently phosphorylate RARα. This deficient phosphorylation has characteristic downstream consequences on the expression of RAR target genes [66] and has been correlated at least in part, to clinical abnormalities observed in patients.

In addition, in several cancers characterized by amplified or deregulated cytosolic kinase cascades [14], ending at Akt or MAPKs [119, 127] RARα has been shown to be aberrantly phosphorylated [121, 122]. Moreover, the RA-induced activation of the MAPK pathway is abrogated [100]. Subsequently, RARα is degraded and/or its transcriptional activity is suppressed. Similarly, in hepatocellular carcinoma, RXRα is aberrantly phosphorylated in its LBD, with characteristic inhibition of its transcriptional activity [88]. Thus, one can postulate that aberrant kinome signaling and RAR/RXR phosphorylation and activity may correlate with tumoral growth and/or RA resistance [37].

There is an increasingly body of evidence indicating that RA signaling plays an important role in brain function, such as synaptic plasticity and learning and memory via transcriptional effects [72], as well as through non genomic effects involving a pool of RARα present in dendrites that becomes phosphorylated in response to RA [6, 30, 101]. An interesting observation is that RA signaling is also involved in the pathophysiology of Alzheimer’s disease [72]. Alzheimer’s disease is complex, but its striking and increasingly important characteristic is the aberrant expression and activity of several protein kinases [35, 61]. Therefore, one can suggest that in this disease, aberrant phosphorylation of RARα could have consequences on both its transcriptional activity and its extra nuclear effects on synaptic plasticity.

In conclusion, RA signaling and RAR phosphorylation represent potentially exploitable pathways for devising novel therapies in several diseases, including the Alzheimer disease [120].

Relevance in the Biology of Other Nuclear Receptors

According to the classical model, all-trans RA is channeled to RARs in the nucleus via the cellular RA-binding protein CRABPII, which is a small cytosolic protein belonging to the family of intracellular lipid binding proteins (iLBP) [26, 32, 36]. Such a process markedly facilitates formation of the liganded receptor. However, recent studies revealed that in certain CRABPII-deficient, cell types, such as brain, adipose tissue, skeletal muscle and skin, RA binds another fatty acid-binding protein of the iLBP family, FABP5. The interesting point is that upon RA binding, FABP5 does not deliver the ligand to RARs, but to the Peroxisome Proliferator-Activated Receptor β/δ (PPARβ/δ), another nuclear receptor, resulting in the regulation of genes that are not direct RAR targets [11, 116, 117]. Consequently, new functions of RA in the regulation of energy homeostasis and insulin responses were revealed [12]. PPARs are also known to be phosphoproteins [27] and to have nongenomic effects [83], but the β/δ subtype is the least studied in terms of phosphorylation. Whether it becomes phosphorylated in response to RA requires further investigation.

Some in vitro studies suggest that RA signaling could be mediated by other nuclear factors such as the Retinoic acid receptor related Orphan Receptor beta (RORβ) [124], the Chicken Ovalbumin Upstream Promoter Transcription Factor (COUP-TFII) [69] or the Testicular Receptors (TR2/4) [140], which are also able to integrate several signaling pathways [46, 63]. However, the in vivo relevance of such observations remains to be determined as well as whether these receptors can be also phosphorylated in response to RA.

Future Directions: What Is Still Left to do

The regulation of RAR-target gene activation by RA is controlled not only by simple on/off conformational switches of RARs, but also by kinase signaling pathways. Importantly, these signaling pathways target several actors in retinoid regulatory processes through phosphorylations that fine-tune the RA response via rapid changes in chromatin organization, RAR dynamics, coregulator interactions, and structural and functional shifts in protein-DNA interactions. The future challenges are to connect these data directly with new highly sensitive, real-time or large-scale technologies in order to get novel, critical information about the influence of phosphorylations on the regulation of RARs and RXRs activity. The last-generation dual linear ion trap mass spectrometers coupled with the Orbitrap technology should allow the identification of new phosphorylation sites in endogenous RARs, RXRs and their coregulators, and should provide information about their regulation by RA. Biophysical approaches, such as NMR, coupled to molecular dynamics simulations are other promising tools to investigate how phosphorylations fine-tune the structure of RARs and RXRs to control their recruitment to RAREs with specific spacings, but not to others with different spacings and/or sequences.

Finally, the recent TALEN (Transcription Activator-like Effector Nucleases) or CRISPR (Clustered Regularly Interspaced Short Palindromic Repeats)-based technologies [44] should make possible the generation of point mutations at phosphorylation sites in vivo, providing a more powerful tool than the classical re-expression of a mutant in a null background. These tools coupled to RNA-seq, ChIP-seq and quantitative proteomics should provide interesting information about downstream gene expression and changes in protein complexes induced by RAR phosphorylation. Large-scale and quantitative phosphorylation screens of RARs, combined with other large-scale data sets, should pave the way to breakthroughs in disease-related research. In conclusion, further insights into the effects of RA will likely continue to reveal new targets and mechanisms that will help explain their pleiotropic effects and how these features might be manipulated in the treatment of metabolic disorders.