Acetylation of nuclear receptors in health and disease: an update

Lysine acetylation is a common reversible post‐translational modification of proteins that plays a key role in regulating gene expression. Nuclear receptors (NRs) include ligand‐inducible transcription factors and orphan receptors for which the ligand is undetermined, which together regulate the expression of genes involved in development, metabolism, homeostasis, reproduction and human diseases including cancer. Since the original finding that the ERα, AR and HNF4 are acetylated, we now understand that the vast majority of NRs are acetylated and that this modification has profound effects on NR function. Acetylation sites are often conserved and involve both ordered and disordered regions of NRs. The acetylated residues function as part of an intramolecular signalling platform intersecting phosphorylation, methylation and other modifications. Acetylation of NR has been shown to impact recruitment into chromatin, co‐repressor and coactivator complex formation, sensitivity and specificity of regulation by ligand and ligand antagonists, DNA binding, subcellular distribution and transcriptional activity. A growing body of evidence in mice indicates a vital role for NR acetylation in metabolism. Additionally, mutations of the NR acetylation site occur in human disease. This review focuses on the role of NR acetylation in coordinating signalling in normal physiology and disease.


Introduction
Histone acetylation requires the transfer of an acetyl group from acetyl-CoA to the e-amino group of lysine side chains within the substrate.Families of histone acetyltransferase (HAT) proteins have been identified with distinct substrate specificities.In addition to modifying histones, acetylases and deacetylases regulate lysine modifications of diverse proteins, including structural proteins, transcription factors, metabolic enzymes, components of the cell cycle [1] and nuclear receptors (NRs).High-resolution mass spectrometry identified 1750 acetylated proteins, and now over 6000 proteins are thought to be acetylated at lysine residues, suggesting the regulatory function of lysine acetylation is comparable with other major post-translational modifications [2].Transcription factor acetylation may alter DNA binding, protein-protein interactions, subcellular location, and responses to diverse signalling cascades.Acetylation may govern sequential modifications of substrates by phosphorylation, ubiquitination, methylation and other modifications.
A growing body of literature supports a key role for direct NR acetylation in governing NR function (Fig. 1).NR regulates the expression of genes governing diverse essential functions including development, metabolism, homeostasis, reproduction and tumorigenesis [3].Evolutionary relatedness and the capacity to bind ligand has been used to characterize NR subfamilies (Fig. 2).The type 1, 'classical' or steroid receptors include receptors for androgen receptor (AR), oestrogen (ERa), glucocorticoid receptor (GR), mineralocorticoid (MR).and progestin (PR).The type II receptors include those for 9-cis-retinoic acid (RXR), all-transretinoic acid (RAR), thyroid hormone (TR), and vitamin D3 receptor (VDR).A third class, referred to as 'orphan receptors' based on the lack of a known ligand at the time of discovery, has since been associated with endogenous ligands.
The structure of an NR typically comprises an amino-terminal AF-1 (A/B) domain, the DNA-binding domain (DBD) (C domain), the hinge region (D domain) and a carboxyl-terminal ligand-binding domain (LBD) (E domain).The E domain often participates in subcellular localization, is essential for binding to a ligand and interacting with heat shock proteins [3].The E domain may also contain a second activation function (AF-2).In the absence of ligand, the type I receptors and several orphan receptors are transcriptional repressors.Similarly, the type II receptors RARs and TRs bind to co-repressor (SMRT and N-CoR) [4][5][6][7] in the absence of hormone.We now know a broad array of NR-binding proteins include Alien [8] and others [9] fine-tune NR function.
The conformational changes induced by ligands binding to NR results in disengagement of corepressor protein complexes, with associated HDAC activity and the recruitment of transcriptional coactivators with associated HAT activity [10,11].In the absence of ligand, NR helix 12 projects away from the LBD and with the addition of ligand, undergoes~180°r otation to interface helices 3-5, creating a surface for coactivator recruitment [12][13][14][15].Thus, NR reside as an inactive complex which, on the addition of ligand alters the NR structure, dissociating heat shock proteins, translocating to the nucleus, dimerizing, binding to DNA in chromatin, often at sites of local DNA damage with subsequent recruitment of DNA topology regulatory proteins (TOPIIA), DNA repair complexes and interaction with a basal transcription apparatus and associated recruitment of coactivators.

Cointegrators govern the acetylation status of nuclear receptor complexes and associated histones
The removal of acetyl groups from lysine residues of these substrates is conducted by either NADdependent or NAD-independent histone deacetylases.NR signalling engages cyclical recruitment of HAT complexes wherein direct acetylation of both cointegrators and NR occurs, thereby providing additional checkpoints for ligand-regulated gene transcription [16].NRs directly contact the basal transcription apparatus (reviewed in Ref. [3]) with interactions between the AR and TFIIH, ERa with TFIIB, RAR and TFIIH, RXR with TBP and VDR with TFIIB described.

Acetylation of nuclear receptors governs hormone signalling
The first classical NR shown to be directly acetylated was the AR [59,60] (Fig. 1).The mechanisms governing AR acetylation in vitro and in vivo have been characterized by a number of different laboratories [59][60][61][62][63][64].
On binding of androgens to the AR, conformational changes occur resulting in HSP90 dissociation, nuclear translocation and the induction of AR target gene transcription [64].Dihydrotestosterone (DHT) and other ligands, such as bombesin, induce AR acetylation [62].Mutation of the AR K632A/K633A acetylation site abolished DHT-dependent AR transactivation in cultured cells.Acetylation-deficient AR mutants were selectively defective in DHT-induced transactivation of androgen-responsive reporter genes and coactivation by SRC-1, Ubc9, TIP60, and p300 [59,60].In the original studies, p300 and p300/cAMP-response element-binding protein (P/CAF) acetylated the AR at a 630 KLKK 633 in the AR hinge region, representing an RXKK motif that is conserved across species and among NRs (Fig. 2).The highly conserved lysine-rich motif resides carboxyl-terminal to the DBD.Tip60 acetylated the AR which was necessary for Tip60induced transcription [61].AR acetylation by p300, demonstrated by 14 C-acetate-labeling in cultured cells required the same 630 KLKK 633 motif that had been identified in vitro.Compared with the AR wild-type, the AR acetylation site mutant showed 10-fold enhanced N-CoR binding.p300 binding to the AR was augmented by AR acetylation in vitro.A charged residue (AR(K630R)) substitution reduced p300 binding and enhanced co-repressor binding, whereas AR acetylation mimics with neutral polar substitutions, AR(K630Q) and AR(K630T), enhanced p300 binding and reduced N-CoR/HDAC/Smad3 co-repressor binding.Sirtuin-1 (SIRT1) inhibited AR-dependent gene expression by deacetylating the AR at (K630/633) and restrained prostate cancer cell growth in mice [65].Additional sites of AR acetylation (K618) were identified as targets of Arrest-defect-1 protein, ARD1, (Naa10).ARD1 complexed with AR and Hsp90 induced AR acetylation (K618) and subsequent Hsp90 dissociation, nuclear translocation, and the induction of AR gene expression [64].Conversely, ARD1 silencing resulted in decreased AR-targeted geneinteractions and gene expression.Although p300-mediated acetylation of AR lysine residues (KLKK 633 ) decreases the total amount of AR, it did not impact AR acetylation via ARD1 [63].
The AR acetylation motif functions as a pivotal node governing diverse signalling pathways.AR recruitment into chromatin in response to AR phosphorylation [66] is determined by the state of AR acetylation, as are interactions with co-repressors, including BRCA1, SIRT1 [65] and the NCoR-like repressor DACH1 [33].The AR acetylation site also serves as a substrate for methylation by the histone methyltransferase enzyme SET9 [67,68], and functions as a tethering site for lnc-RNAs that are overexpressed in therapy-resistant prostate cancer.PRNCR1 (Prostate-Cancer-Associated Non-Coding RNA 1; PCAT8) and PCGEM1 (PCGEM1 Prostate-Specific Transcript) bind the AR enhancing ligand-dependent and ligand-independent AR-mediated gene expression and prostate cancer cell proliferation [69].PRNCR1 bound via the acetylated AR and sequentially recruited DOT1L and lncRNA PCGEM1 to the AR amino terminus, which in turn was methylated by DOT1L.Binding of PRNCR1 to the C-terminally acetylated AR on enhancers and its association with DOT1L were required for recruitment of lncRNA, PCGEM1, to the DOT1L-mediated methylated AR.
The ERa is acetylated in vivo [70].ERa acetylation restrains ligand sensitivity.p300, but not P/CAF, directly acetylated the ERa hinge/LBD (lysine residues 302, 303).Either polar or charged residue substitution of these acetylated residues enhanced ERa hormone sensitivity indicating ERa acetylation restrains ligand sensitivity [70].MAPK signalling to the ERa was unaffected by ERa acetylation mimics indicating dissociable signalling pathways to the ERa.Acetylation of the ERa at K302, and 303 and acetylation of the p160 coactivator s participate in subsequent transcriptional attenuation.Mutation of the ERa K302/303 acetylation site affects regulation by ligand, tumour suppressors and therapeutic responses.The ERa acetylation sites are within a hot spot of mutually exclusive posttranslational modifications (phosphorylation, ubiquitination, acetylation), located within an intrinsically disordered region (IDR) [76].Phosphorylation at the S305 site prevented K303 acetylation [77].Modifications within this region (K302, K303) affects acetylation by p300, methylation by SET7/9 [78], phosphorylation by PKA or PAK1 [70,79] and a receptor acetylation-ubiquitination balance mediated by BRCA1 [80].The K303R is adjacent to a protein kinase A (PKA) phosphorylation site.Acetylation of K303 ERa was blocked by a substitution of serine residue 305 to aspartic acid, designed to mimic constitutive phosphorylation [81].
An additional ERa acetylation site (K266, 268) was subsequently reported [82].and acetylation at lysine 266 and 268 enhanced transactivation and DNA binding, whereas acetylation of lysine 302 and 303 suppressed transactivation.Unbiased mass spectrometry analysis identified ERa acetylation at lysine residues 302, and 303 [83] suggesting the ERa acetylation site may be cell-type dependent.The ERa lysine residues K299, 302 and 303 bind to glutamate residues in calmodulin via salt bridges to prevent ubiquitination thereby inhibiting ERa degradation [84].
Transcriptional repression of ERa can occur through HDAC complex recruitment by BRCA1 [30,31,85].BRCA1 binds HDAC complexes [85].BRCA1 restrained estradiol-induced ERa activity [30], BRCA1 repression of ERa involves a direct physical interaction between ERa and BRCA1 [86].BRCA1 represses expression of ERa via p300-dependent andindependent mechanisms [86].ERa mutations of the K302/303 motifs are resistant to BRCA1 inhibition [80].The level of K302/K303 acetylated ERa was reduced by BRCA1 expression and increased by BRCA1 knockdown without affect K266/268R or K266/268Q acetylation.Germ line BRCA1 mutations from patients with breast cancer have reduced ERa transcriptional inhibition [30,31,87] leading to the proposal that enhanced ERa activity in such patients may contribute to the pathogenesis of early-stage breast cancer [30,31].
The progesterone receptors (PRs) undergo acetylation, phosphorylation and ubiquitination [88].The PR is acetylated at the hinge region consensus sequence, KXKK (amino acids 638-641), [89].P300 also enhanced PR acetylation at K183 (Lys-183) [90].Analysis of the PR hinge region acetylation site showed that both acetylation mimics (K-Q or K-T) and an acetylation-deficient (K-A) mutant conveyed delayed phosphorylation and nuclear entry.PR acetylation altered the magnitude and kinetics of progesterone transcriptional response [89].
The RXR receptors (RXRa, RXRb and RXRc) respond to retinoids or vitamin A derivatives including 9-cis retinoic acid (9-cRA), and other endogenous ligands.RXRa is acetylated by p300 on lysine 145, facilitating its DNA binding and transcriptional activity.RXRc is also acetylated by p300.RXRa acetylation (K145) increased the transcriptional activity in cultured cells by promoting receptor binding to DNA.RXRa acetylation by p300 was attenuated by the orphan nuclear receptor (ONR) TR3 [97].The binding of RXRa to TR3 was increased by 9-cRA, which attenuated p300-induced cell proliferation.
The GR (NR3C1) and MR (NR3C2) are responsive to glucocorticoids including corticosterone or cortisol and mineralocorticoids such as aldosterone, respectively.Both receptors are ligand-dependent transcription factors typically found in the cytoplasm, but on ligand binding, are transported into the nucleus.The cytosolic interaction with HSP90 maintains MR and GR in a ligand-binding conformation and controls nucleo-cytoplasmatic shuttling GR (K 494 and K 495 ) [98] and MR (K 677 ) share the hinge region KXKK acetylation motif [99] which, when mutated, attenuates acetylation and promotes transcriptional activity [100,101].Inhibition of class 1 HDACs promoted MR acetylation, attenuated recruitment of MR and RNA polymerase II to the promoters of target genes and reduced expression of MR target genes [100] manifesting as decreased transcriptional activity [102].CBP or p300 increased MR acetylation, decreased recruitment of MR and Pol II to specific hormone response elements and decreased expression of MR target genes in HEK cells [102,103] without affecting sub-cellular localization.
CLOCK acetylates the nuclear glucocorticoid-bound pool of GR, attenuating binding of the receptor to GR response elements [101].
Knockdown or inhibition of HDAC3 increased MR acetylation, reduced expression of MR target genes in response to mineralocorticoids [100].HDAC3 interaction with MR reduces its transcriptional activity [100], and HDAC4 was identified as a necessary scaffold for assembly of the MR-HDAC3 complex.Inhibition of class I HDACs promoted MR acetylation and abrogating hypertension in spontaneously hypertensive rats [107] HDAC2 [98] and SIRT1 [108] govern GR deacetylation, which is required for GR binding to NFjB with consequent repression of inflammation [109].Increasing acetylation of residue K295 of HSP90 by inhibiting HDAC6 weakened HSP90 interaction with the MR [110] and GR [111] and supported nuclear import of both receptors.While HSP90-K 295 acetylation does not alter MR or GR expression it differentially affects transactivation promoting GR activation but not that of MR [110,111].
Nuclear receptor acetylation of in thermogenesis and metabolism (Fig. 3) Peroxisome proliferator-activated receptor gamma (PPARc) is activated by synthetic ligands including the anti-diabetic thiazolidinediones (TZDs) and natural ligands such prostaglandins (15d-PGJ2) [112].PPARc governs cellular proliferation and autophagy, adipocyte differentiation, lipogenesis, inflammation and insulin sensitivity.PPARc has a cell type-specific role in tumorigenesis, typified by gene deletion studies that demonstrated endogenous PPARc1 promoting ErbB2induced mammary tumour growth [113].
PPARc lysine acetylation by p300 and deacetylation by SIRT1 have been described in the context of cellular senescence, with the abundance of acetylated PPARc increasing with cell cycle number.SIRT1 is itself transcriptionally regulated by PPARc, inferring a feedback loop wherein cellular senescence is accompanied by increased acetylation of PPARc, which in turn regulates the transcriptional activation of SIRT1 [114].Prior studies had shown HDACs are associated with PPARc in chromatin [115].PPARc/HDAC complexes are recruited to genes governing lipogenesis, which is facilitated by cyclin D1, a regulatory subunit of a holoenzyme that phosphorylates the retinoblastoma protein (pRB).Chromatin immunoprecipitation assay demonstrated that cyclin D1 augmented recruitment of the histone methyltransferase SUV39H1 together with histone deacetylases HDAC1 and HDAC3 to the PPAR response element of the lipoprotein lipase promoter.Cyclin D1 decreased acetylation of total histone H3 and specifically histone H3 lysine 9. Cyclin D1 bound HDAC in vivo and is preferentially physically associated with HDAC1, HDAC2, HDAC3 and HDAC5 [115].
PPARc isoforms (PPARc1 and PPARc2) are generated by alternative splicing and differential promoter utilization.The murine Pparc2 encodes an additional N-terminal 30 amino acids (28 amino acids in human PPARc).PPARc1 is expressed at low levels in many tissues.PPARc2 is expressed at high levels in adipose tissue.Five acetylated lysine residues have been identified in the adipose-tissue restricted PPAR2 (K98, K107, K218, K268 and K293).Acetylation at two sites (K268ac and K293ac) was reduced by the TZD agonist, rosiglitazone or by SIRT1 deacetylase activation [116].SIRT1-mediated Lys268 and Lys293 deacetylation led to recruitment of PRDM16 (the brown adipogenic activator PR domain containing 16) to PPARc2 [116].K268 and K293 acetylation enhanced binding to NCoR, whereas K293 deacetylation enhanced PPARc2 binding to PRDM16.Deletion in breast cancer-1 (Dbc1) promoted 'browning' of white adipose tissue, also by deacetylating PPARc on Lys268 and Lys293 [116].Consistent with these findings, mice with constitutively deacetylated PPARc mutations of K268R/K293R, (2KR) demonstrated augmented brown remodelling of white adipose tissues [117].
Lysine acetylation of PPARc has been mapped, employing techniques that had been developed for mass spectrometric identification of potential modification sites on histones, where traditional tryptic digestion approaches are hampered by the abundance of basic amino acid residues in fragments.Nine lysyl residues in Pparc1 (including the lysine residues corresponding to K218 and K268 on PPARc2), of which K154 and K155 (K184 and K185 in PPARc2) were further characterized [112].Quantification of Pparc1 K154/155 acetylation showed that endogenous Pparc1 acetylation levels are ~1%.Qiang et al. [116] had made a similar observation with PPARc2, where expression of Cbp was needed to reveal acetylation of at K268 and K293 in vitro.The PPARc1 K154 and K155 residues identified in mass spectrometry were assessed further.A PPARc1 peptide encoding acetylated K154 and K155 PPARc1 served as a substrate for SIRT1 deacetylation and in vivo labeling studies showed the PPARc mutant K154/155R conveyed dramatically decreased of [ 3 H] acetyl-CoA incorporation.K154/K155A and K154/K155Q mutants both showed severely diminished lipogenic potential compared with the WT protein.Collectively these studies illustrate that PPARc is acetylated at multiple residues which may participate in diverse biological processes.
FXR, the nuclear bile acid receptor farnesoid X receptor, which is highly expressed in the liver and small intestine, is an important regulator of lipid and glucose metabolism [118].FXR is acetylated by p300 and deacetylated by SIRT1 at K217 [119,120].p300 activity depends on acetyl-CoA as the acetyl donor and SIRT1 activity is regulated by cofactor NAD + , thus linking metabolism and FXR acetylation.p300 regulates FXR transactivation in conjunction with acetylation of the receptor and local histones at the target gene promoters [121].FXR acetylation (K217, K157) increased FXR stability and inhibited heterodimerization with RXRa, with a consequent reduction in DNA binding and transactivation activity [119].Down-regulation of SIRT1 in mice infected with adenovirus-siSIRT1 increased FXR acetylation in liver and SIRT1 activation by resveratrol reduced FXR acetylation in-vivo, consistent with the in-vitro findings.Findings in ob/ob and dietary mouse models of metabolic disease were consistent with FXR acetylation/deacetylation participating in SIRT1 regulation of the metabolic state [119].SIRT1 agonists experiments and analysis from Sirt1 À/À mice revealed metabolic targets [122].Studies using acetylation-mimic and acetylation-defective K217 mutant mice have assisted in the understanding of FXR acetylation function overall SIRT1 action [123], but this is an area warranting further inquiry.
The LXR family (LXRa and LXRb isotypes) [124,125] share ~75% sequence homology in the DBD and LBD [126].SIRT1 deacetylation (LXRa, K432; LXRb, K433) promotes ubiquitination, proteasomal degradation and transcriptional activation [127].Sirt1 deficient mice have defective reverse cholesterol transport, in part due to reduced expression of the LXR target gene ABCA1 resulting in decreased HDL and thereby increased hepatic and testicular cholesterol levels.
Hepatocyte nuclear factor 4a (HNF4a), which governs the development of the liver, kidney and intestine, participates in normal metabolism and is mutated in maturity-onset diabetes of youth.Acetylation of HNF4a enhances its nuclear retention and DNA binding affinity [129], whereas SIRT1 deacetylation inactivates HNF4a [130].
Steroidogenic factor-1 (SF-1), which is necessary for the expression of steroidogenic pathway components and development of steroidogenic tissues, is acetylated by p300 in vivo [131] at the hinge region (KQQK (AA 102-106)).GCN5 also acetylates SF-1 in vitro stimulating its transcriptional activity.K34, K38 and K72 in the C terminal zinc finger motifs appears to be the targets of GCN5.Inhibiting deacetylation with trichostatin A (TSA) stabilized SF-1 and induced its nuclear export, but at the same time increased SF-1-mediated transactivation [132].
REV-ERB governs the mammalian circadian rhythm, via the clock genes, and regulates metabolic pathways.Tip60 acetylates Rev-erbb at the RXKK motif, with consequences that include relieving Reverbb-mediated repression of apolipoproteinCIII (apoC-III) expression [133].HDAC1 is recruited to the apoC-III promoter by Rev-erbb, antagonizing Tip60.REV-ERB repressor activity is induced by heme, the intracellular synthesis of which is regulated by the NR coactivator PGC-1a, in turn mediating glucose and fatty acid metabolism.Acetylation of REV-ERBb thereby integrates the circadian rhythm and energy metabolism [134].

Mitogenic signalling cascades deploy nuclear receptor acetylation in growth control
Induction of mitogenic signalling by receptor tyrosine kinases regulates HAT and HDAC subcellular location and enzyme activity.Acetylation of the AR [59] is a key determinant of AR phosphorylation [66].The 630T AR acetylation site mutant was defective in MEKK1-induced apoptosis [60].AR acetylation site mutants transactivated more efficiently than AR wildtype in response to activation of the p42/p44 MAPK pathway but failed to respond to the PKA signalling pathway [66].AR wild-type, but not the AR acetylation site mutants, responded to PKA inhibition with increased reporter activity and increased association at an androgen response element in chromatin immunoprecipitation assays.AR acetylation site mutations reduced TSA responsiveness also reduced ligandinduced phosphorylation of the AR.Point mutation of a subset of AR phosphorylation sites reduced TSA responsiveness and trans-activation by HATs.MAPK (ERK1/2) inhibition with PD98059 blocked T4 induced TR acetylation of.In contrast, the ERa acetylation site does not affect MAPKK induction [70].

Nuclear receptor acetylation and human disease
Fatty liver disease and non-alcoholic steatohepatitis (NASH) As noted above, dysregulation of NR acetylation function in vivo alters cellular metabolism.In this regard PPARc, LXR and FXR are each acetylated and participate in lipid metabolism.FXR agonists are being developed for the treatment of nonalcoholic fatty liver disease [150,151].As noted above, FXR (K157 and K217) is deacetylated by SIRT1 and acetylated by p300.
FXR acetylation reduces FXR-RXRa heterodimerization leading with a consequent reduction in expression of FXR target genes.FXR acetylation is increased in obese individuals, in mice chronically fed a western style diet and ob/ob mice [119,152].These observations support a link between NR acetylation and metabolic disease [153] and provide a basis for therapeutic targeting in insulin resistance, dyslipidemia [154,155] and fatty liver disease [156].

Cancer
Mutation of NR acetylation sites has been identified in human tumours, including the ERa [76] and the AR (K630T).The K303R missense mutation in ERa (ESR1) is very common, with over 200 instances recorded in COSMIC [76].This cancer-associated mutation confers resistance to the aromatase inhibitor, anastrozole, when expressed in MCF-7 breast cancer cells [81,157].The AR acetylation mimics, including the prostate cancer associated K630T somatic mutation, promoted cell survival and a dramatic increase in the growth of prostate cancer cells in soft agar and in nude mice, likely mediated via augmented transcription of growth control target genes [158].Enhanced transactivation occurred notwithstanding delayed nuclear import of K630T mutant AR [159].

Kennedy's disease
Kennedy's disease also known as X-linked spinal and bulbar muscular atrophy is an adult-onset form of motor neuron disease associated frequently with androgen insensitivity.AR gene mutations associated with the disease encode increased length of a polymorphic tandem CAG repeat in the coding region [160], leading to polyglutamine (polyQ)-expanded AR, which is thought to mediate disease pathogenesis in a DHTdependent manner.SIRT1 deacetylates the AR at lysines 630/632/633 protecting against polyQ-expanded AR toxicity.Pharmacologic inhibition of the hyperacetylated polyQ-expanded AR reduced mutant AR aggregation, offering a mechanistic link between AR acetylation status and disease pathogenesis.Inhibition of polyQ-expanded AR lysines 630/632/633 acetylation decreased AR aggregation and consequent toxicity in motor neurons [161].

Intrinsically disordered regions and acetylation sites
A significant part of NR is comprised of IDRs [162,163].IDRs do not adopt a well-defined conformation in isolation even under physiological conditions but carry out important functions that are often regulated by PTMs [164].Using a recent approach based on the structure prediction of Alphafold2 [165], nearly 40% of residues of NRs are predicted to be disordered (Fig. 4, Table 1).In a number of cases, there is direct experimental evidence for the disordered state of NRs (Fig. 4).IDRs are predominantly in the N-terminal region and the hinge region (Fig. 4, orange and red boxes).In the case of ERa and HNF4a, the Cterminal regions are also disordered (Fig. 4, human).At least parts of the hinge region were shown to be disordered for ERa [166], ERR-2 [167], PGR [168], AR [169] or RXRa [170].There are no or very few IDRs predicted for the DBDs and LBDs, in agreement with their largely structured characteristics.The coincidence of experimentally verified acetylation sites, coincident with IDR (Fig. 4, Red boxes), raises the question of the potential functional relationship between these characteristics.
The important regulatory functions of IDRs in NRs are regulated by a multitude of different types of PTMs, including acetylation.PTMs, including phosphorylation or ubiquitination have a strong preference for IDRs [164].Acetylation sites are enriched in the disordered hinge region (Table 1, observed 30 vs. expected 10.1) but depleted in the disordered Nterminal region and enriched in the ordered DBD (Table 1, observed 15 vs. expected 6.8).While the disordered hinge region shows very little overall sequence conservation, a small motif containing two and three lysines can be identified in the majority of NRs.A second site corresponding to the acetylation sites of K299, K302 in ERa is more specific and is only present in two other members of the oestrogen family and in members in the nr1 family.In contrast, the K303 site mutated in breast cancer is specific to ERa.
Intrinsically disordered regions play crucial roles in the function of NRs and are key to their functional versatility.The hinge region has important regulatory roles and can increase the affinity and specificity of DNA recognition [167].The hinge region contains additional interaction sites with regulatory functions such as the nuclear localization site in AR [159], or a calmodulin binding site in ERa [171].The disordered N-terminal region is the main transactivation domain providing interaction surfaces for coregulatory proteins such as the TATA box-binding protein (TBP) or RAP74 subunit of the general transcription factor TFIIF.The N-terminal domain (NTD) region also plays a crucial role in the allosteric communication between the DBD and ligand binding site [172].Steroid receptors, including ERa [173], AR [174], PGR [175] and GR [176] contain long NTD that are highly dynamic with little or no secondary structural elements.In addition, the NTDs of RXRa and PPARc were also shown to be disordered by a combination of techniques [170].While it adopts a folded structure, the LBD is highly flexible, which is important to bind diverse ligands.It can also contain some disordered segments, as indicated by residues with missing electron density observed in known structures of this domain [162].The NR-box which mediates interactions between NR and its coregulator is typically located within disordered regions [177,178].The disordered N-terminal region is the main transactivation domain providing interaction surfaces for coregulatory proteins such as the TBP or RAP74 subunit of the general transcription factor TFIIF.The specific binding to these partners induces a transition to a more ordered structural state [173,174].In the context of full length proteins, DNA binding can influence the properties of this region [179].The NTD region also plays a crucial role in the allosteric communication between the DBD and ligand binding site [172].Alternative initiation sites described for multiple NRs create different isoforms by changing the length of the N-terminal region [180].
Due to intrinsic disorder, these alterations enable another level of regulation of allosteric communication without disrupting the overall structural features.The hinge regions can be directly involved in DNA binding and undergoes a disorder-to-order transition on binding specific DNA elements.In the case of RXR, the transition was observed only in the homodimeric  [188] and are indicated with boxes.Experiment disorder information was collected from Mobidb [189] and Disprot [190] and are indicated by darker red.Predicted disorder is indicated as lighter red and are based on low (< 70%) pLDDT scores of Alpahafold2 [165].Lysine residues are indicated as dark blue (exposed) and light blue (buried) ticks.Orange lines indicate known acetylation sites.The sequence alignment for the DBD was generated with MAFFT [191].To detect conserved motifs within disordered hinge region, which shows high evolutionary variability, this region was aligned with PHI-BLAST [192] using the motif K-X(0,2)-K-K(0,1).Lysine residues are highlighted by blue background in columns containing acetylated lysines, with the modified sites indicated by red.Areas of coincident acetylated lysine residues confirmed experimentally, with intrinsically disordered domains are shown with a red open box.complex, while the same segment remained unfolded in heterodimeric complexes [163].A similar folding on binding interactions can be important for NR that bind DNA as monomers, such as ERR, to increase the affinity and specificity of DNA recognition [167].The hinge region contains additional interaction sites with regulatory functions such as the nuclear localization site in AR [159], or a calmodulin binding site in ERa [171].The important regulatory and switching functions of IDRs of NRs are regulated by a multitude of different types of PTMs, including acetylation.
Acetylation in NRs are enriched in the disordered hinge region (Table 1, observed 30 vs. expected 10.1) as well as the ordered DBD (Table 1, observed 15 vs. expected 6.8).While the DBD generally shows high sequence conservation, acetylation sites can follow different evolutionary patterns (Fig. 4).In addition to highly conserved lysines which are present in nearly all of the NRs, some lysines are specific to certain families, or smaller subfamilies.Interestingly, the conserved position corresponding to the known acetylation sites K38 of SF1 is mutated to arginine in the nr1 family.While the Lys and Arg both occur frequently in DNA binding, they behave very differently in terms of PTMs, hence their mutation could indicate an altered modification strategy.Additional hot-pots for acetylation are located within the disordered hinge region.While this region shows very little overall sequence conservation, a small motif containing two and three lysines can be identified in the majority of NRs.A second site corresponding to the acetylation sites of K299, K302 in ERa is more specific and is only present in two other members of the oestrogen family and in members in the nr1 family.In contrast, the K303 site mutated in breast cancer is specific to ERa.Overall, as many of the acetylation sites are conserved in multiple NR sequences, the number of acetylation sites is expected to be even higher than has been verified so far.

Unanswered questions and future directions
Collectively, these studies have made appreciable progress toward understanding the biology, disease significance and therapeutic opportunities that reside in lysine acetylation of NR.The challenges ahead are summarized in three broad questions.Firstly, what is the structural basis by which the acetylated motif may interact with such a diverse array of partner proteins, in the context of histones?Secondly, how might one predict the outcome of engaging such diverse signalling pathways?Thirdly, with the goal of designing therapeutics to these molecular targets, what is the in vivo biological significance of NR acetylation in each of its many contexts?

Perspectives and summary
Prior reviews have described the profound effects of NR acetylation on NR functions, including ligand sensitivity, ligand specificity, subcellular localization, association with particular protein complexes and responses to mitogenic signalling pathways [181][182][183].Over time, the diversity of potential interactions between histone modifying proteins, NRs and the function of NR acetylation has been revealed.The questions that remain, though, center on understanding the circumstances under which acetylation of NR comes about, how acetylation governs interaction with chromatin and other proteins and how acetylation affects the response to other signalling pathways.Firstly, as the resolution of proteomic analysis has improved, it has become apparent that acetylation of NR and NR co-integrators is common, rivalling the incidence of phosphorylation.Secondly, the networks of interaction between particular HATs and NR, at least in cultured cells appear to be relatively promiscuous.For example, TIP60 interacts with a broad array of NR (AR, ERa, PR, GR, PPARc, REV-ERBB and PXR [184]).Other proteins within the TIP60 complex (RUVBL1, RUVBL2, p400, TRRAP, BRD8, EPC1, DMAP1, MRG15, ING3, BAF53a, GAS41, Eaf6, PARP-1 and MDT1) expand the repertoire of interactions to additional NR (FXR, LXR).Such findings raise the question how these multiprotein complexes Table 1.Disorder and acetylation sites within different modules in nuclear receptor (NR).The table shows the percentage of disordered regions according predicted based on Alphafold pLDDT scores [165] within the different functional modules.The minimum and maximum values among the different NR are also indicated.The number of acetylation sites within each module was compared with the expected number of sites assuming an even distribution of PTMs along the sequence.Statistically significant differences between observed and expected number acetylation sites based on Fisher's exact test are indicated by asterisks.
Percentage of disordered residues (min-max) contribute specificity to signal transduction?Thirdly, the sites of NR that mediate acetylation, appear to serve as nodes of interaction between diverse signalling pathways, beyond the kinases, including the noncoding genome.Such findings raise the question of how these diverse signalling pathways are coordinated through a common signalling nodes?Structural modelling has predicted that some NR acetylation sites reside within intrinsically disordered domains (IDDs), allowing development of plausible mechanistic models that might explain how diverse post-translational modifications can occur at these sites.IDRs participate in liquid-liquid phase separation, a process that underlies the creation of nonmembrane bound intracellular structures, with functional consequences that include proximity-facilitated biochemical reactions, protein sequestration and structural organization [185].Nucleosomes, for example, exist as separated liquid phase structures operating in a dynamic, condensed fashion [186].These interactions can carry a significant element of specificity, likely arising in an ensemble-encoded, 'emergent' way [187].
Potential acetylation sites are prevalent in NR and IDRs facilitate liquid-liquid phase separation, suggesting a locus of regulation [185].Given the importance of these acetylated motifs in governing aberrant growth, including prostate cancer [158], their IDD contexts warrant examination for the opportunities of targeted therapy.

Fig. 1 .
Fig. 1.Timeline identifying acetylated nuclear receptors (NRs).NRs that have been documented as acetylated in vitro and or in vivo are shown by the year of publication.

Fig. 2 .
Fig. 2. Phylogenetic tree of nuclear receptor family.Nuclear receptors shown to be acetylated are shown highlighted in yellow.Figure reproduced in part from [70] with modifications.

Fig. 3 .
Fig. 3. Function of nuclear receptor (NR) governed by acetylation.NRs (top left panel) have been shown to undergo acetylation by acetyltransferases.Acetylated NRs are deacetylated by HDACs (bottom-right panel).Acetylation-mediated regulation of NRs includes transactivation, subcellular localization, DNA binding, stability and degradation, ligand binding and cofactor binding (bottom-left panel).

Fig. 4 .
Fig. 4. Sequence alignment for acetylation sites in nuclear receptor (NR) and disordered domains.Domain assignments for the DBD and the ligand-binding domain are based on Pfam[188] and are indicated with boxes.Experiment disorder information was collected from Mobidb[189] and Disprot[190] and are indicated by darker red.Predicted disorder is indicated as lighter red and are based on low (< 70%) pLDDT scores of Alpahafold2[165].Lysine residues are indicated as dark blue (exposed) and light blue (buried) ticks.Orange lines indicate known acetylation sites.The sequence alignment for the DBD was generated with MAFFT[191].To detect conserved motifs within disordered hinge region, which shows high evolutionary variability, this region was aligned with PHI-BLAST[192] using the motif K-X(0,2)-K-K(0,1).Lysine residues are highlighted by blue background in columns containing acetylated lysines, with the modified sites indicated by red.Areas of coincident acetylated lysine residues confirmed experimentally, with intrinsically disordered domains are shown with a red open box.