Biochimica et Biophysica Acta 1842 (2014) 326–337

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta

journal homepage: www.elsevier.com/locate/bbadis

Review The biophysical and biochemical properties of the autoimmune regulator (AIRE) ☆

Roberto Perniola a,⁎, Giovanna Musco b a Department of Pediatrics – Neonatal Intensive Care, V. Fazzi Regional Hospital, Piazza F. Muratore, I-73100, Lecce, Italy b Biomolecular NMR Laboratory, Center of Translational Genomics and Bioinformatics, Dulbecco Telethon Institute at San Raffaele Scientific Institute, Via Olgettina 58, I-20132, Milan, Italy article info abstract

Article history: AIRE (for autoimmune regulator) is a multidomain protein that performs a fundamental function in the Received 27 July 2013 and possibly in the secondary lymphoid organs: the regulation, especially in the sense of activation, of the process Received in revised form 11 November 2013 of in cell lines deputed to the presentation of self- to the maturing T lymphocytes. The Accepted 18 November 2013 of the elements bearing T-cell receptors with critical affinity for the exhibited self-antigens prevents Available online 22 November 2013 the escape of autoreactive clones and represents a simple and efficient mechanism of deletional self-tolerance. However, AIRE action relies on an articulated complex of biophysical and biochemical properties, in most cases Keywords: Autoimmune polyendocrinopathies attributable to single subspecialized domains. Here a thorough review of the matter is presented, with a Self-tolerance privileged look at the pathogenic changes of AIRE that interfere with such properties and lead to the impairment Transcription factors in its chief function. Protein conformation © 2013 The Authors. Published by Elsevier B.V. All rights reserved. Post-translational protein processing

1. Introduction polyglandular syndrome type 1 (APS1), also known as autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy (APECED), The autoimmune regulator (AIRE) is encoded which is characterized by autoimmune damage of the endocrine by the homonymous gene, AIRE, situated in the region q22.3 of chro- glands and various ectodermal dystrophies. A chronic or recurrent mosome 21 [1,2].PathogenicAIRE variants cause the autoimmune mucocutaneous candidiasis completes the clinical picture [3,4].The homologous murine gene, Aire (printed in lower case to avoid confu- sion), is located in the 10, and shows 77% nucleotide Abbreviations: aa, amino acid(s); AIRE, autoimmune regulator; APECED, autoimmune – polyendocrinopathy-candidiasis-ectodermal dystrophy; APS1, autoimmune polyglandular correspondence with human AIRE [5 7]. syndrome type 1; CARD, caspase-recruitment domain; CBP, CREB-binding protein; ChIP, Since the delineation of its sequence, it has been evident that AIRE chromatin immunoprecipitation; CIITA, MHCII transactivator; coIP, coimmunoprecipitation; has a multidomain structure typical of the able to bind to chro- DAXX, death-associated protein 6; Deaf1, deformed epidermal autoregulatory factor 1; matin and to regulate the process of gene transcription [8–10]. Its orga- fl EMSA, electrophoretic mobility shift assay; Gfp, green uorescent protein; GST, nization resembles that of the 100-, 110-, and 140-kDa members of the glutathione-S-transferase; HEK293, human-embryonic-kidney epithelial-cell line 293; HSR, homogeneously-staining region; ING2, inhibitor of growth 2; MHCII, speckled-protein family (Sp100, Sp110 and Sp140, respectively), which class-II major histocompatibility complex; MM/PBSA, Molecular Mechanics/ have been identified in the nuclear bodies (NBs) and carry out the above Poisson–Boltzmann Surface Area; MS, mass spectrometry; mTEC, medullary thymic functions [11–14]. Starting from the amino terminus, AIRE comprises epithelial cell; NB, nuclear body; NLS, nuclear localization signal; NMR, nuclear a homogeneously-staining region (HSR) or caspase-recruitment do- magnetic resonance; NucP41/75, nuclear phosphoprotein 41/75; PGE, promiscuous main (CARD) [15], a nuclear localization signal (NLS) [16],aSAND ; PHD, plant homeodomain; PIAS1, protein inhibitor of activated STAT1;PK,proteinkinase;PML,promyelocytic leukemia; PRR, proline-rich region; (for Sp100, AIRE, nuclear phosphoprotein 41/75 or NucP41/75, and de- P-TEFb, positive transcription elongation factor b; PTM, post-translational modifi- formed epidermal autoregulatory factor 1 or Deaf1) domain likely to cation; RING, really-interesting new gene; RNA-PolII, RNA-polymerase II; SAND, mediate DNA binding [17], and two plant-homeodomain (PHD) fingers Sp100, AIRE, NucP41/75 and Deaf1; Sp100, 100-kDa speckled protein; Sp110, [18]. In addition, four LXXLL (where L stays for leucine) motifs are dis- 110-kDa speckled protein; Sp140, 140-kDa speckled protein; STAT, signal trans- ducer and activator of transcription; TsAg, tissue-specificantigen tributed among the domains [19]. ☆ This is an open-access article distributed under the terms of the Creative Commons AIRE expression reaches its highest level in a subset of medullary Attribution-NonCommercial-No Derivative Works License, which permits non-commercial thymic epithelial cells (mTECs), one of the cell types that are part of use, distribution, and reproduction in any medium, provided the original author and source the thymic stroma [20,21]. Not casually, mTECs are involved in the are credited. mechanisms of , where the term “central” refers to ⁎ Corresponding author. Tel.: +39 0832 661 359; fax: +39 0832 661 719. E-mail addresses: [email protected] (R. Perniola), [email protected] the primary lymphoid organs [22,23]. An apparently simple mechanism (G. Musco). useful to this end is the apoptosis (negative selection) of the clones of

0925-4439/$ – see front matter © 2013 The Authors. Published by Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.bbadis.2013.11.020 R. Perniola, G. Musco / Biochimica et Biophysica Acta 1842 (2014) 326–337 327 (the maturing T lymphocytes) bearing T-cell receptors with of the intermediate filaments or microtubules: the staining is not due to critical affinity for the single self-antigens [24].But,toestablishacon- saturation of the nuclear import, and colocalizes AIRE with vimentin tact with the counterpart, a large number of these self-antigens, mostly and α-tubulin, but does not with actin [59–61]. Similarly to what was tissue-specific antigens (TsAgs), need to be synthesized in the thymus. demonstrated for PML- and Sp100-related NBs [62], the number and This phenomenon, termed promiscuous gene expression (PGE), dis- size of the nuclear speckles in each cell, and the percentage of cells covered in the second part of the eighties [25–27],andbetterdefined in displaying fluorescence in the cytoplasm were thought to differ accord- the following decade [28–30], is just restricted, both in humans and ing to the modulation of the involved structures during the cell cycle mice, to mTECs [31,32], albeit it is presumable that PGE is replicated [60]. in, or extended by contiguity to other -presenting cells (of epi- To confirm this hypothesis, HeLa cells were transfected with a fusion thelial and bone-marrow lineage) of the thymus and secondary lym- construct, in which AIRE was placed downstream of, and in frame with phoid organs [33,34]. As shown in the murine thymus, PGE correlates the gene encoding the green fluorescent protein (Gfp): unexpectedly, with Aire expression and the exposition, on mTEC surface, of the cluster two-thirds of Gfp-AIRE+ cells displayed only a homogeneous and faint of differentiation CD80 [35].Inthissense,bothcellularkineticsandsur- fluorescence of the nucleoplasm, without apparent nuclear speckles. face phenotype of the murine Aire+ mTECs would indicate that AIRE Time-lapse analysis cleared that the presence of the nuclear speckles marks terminally-differentiated elements [36,37]. The participat- is dependent on the cell cycle, as they disappear before the mitosis ing in PGE colocalize in chromosomal clusters [32,38], and the pattern of and are promptly reconstituted after the cell division. In addition, it PGE in single mTECs suggests a probabilistic mechanism of gene was observed that, after chromatin digestion and removal, a significant targeting [39,40]. amount of Gfp-AIRE associated with the nuclear matrix, the insoluble These features may be read taking in mind an alternative model of subcomponent of the nuclear skeleton that directs the functional sub- AIRE action: this theory moves from the evidence that both thymic ru- units of the chromatin [63]. diments and the normal thymus exhibit structures that are reminiscent of various epithelia and contain the related TsAgs [41–43]. The presence 3. The basic properties of AIRE: homomerization and DNA-binding of several transcription factors in the mitotically active mTECs, which is representative of a condition of cellular multipotentiality, and the very AIRE homology with nuclear proteins involved in the regulation of low frequency of expression of each TsAg-encoding gene in murine the gene transcription was easily confirmed by fusion constructs Aire+ mTECs would mean that small clusters of mTECs are programmed encoding AIRE bound to heterologous DNA-binding domains: in the re- to reproduce the cellular and histological arrangement of the extra- porter assays, wild-type AIRE displayed strong activation of the gene thymic epithelia [44,45], and that AIRE action is addressed to target transcription, while defective AIRE variants had various degrees of im- genes whose expression is relevant, in transit elements, to the final pairment [64–67]. stage of multiple mTEC differentiations [46–48]. The acceptance of this Concomitantly, of great importance was the delineation of AIRE theory implies that some in vitro performances of AIRE (in particular properties propedeutic to the chief function: firstofall,itwas the enhancement of the transcription of various TsAg-encoding genes) demonstrated, by pull-down assay utilizing glutathione-S-transferase would be influenced by the experimental settings and environmental (GST)-AIRE and 35S-labeled AIRE, that AIRE is able to homomerize. conditions of the transfected cells. This phenomenon was confirmed by mammalian and yeast two- Whatever its biological role may be [49,50], AIRE regulates the gene hybrid assays, albeit the gene transcription was not enhanced in the transcription (gene transactivation and transrepression) thanks to an yeast cells [65]. Accordingly, Kumar et al. [68] processed AIRE under ox- articulated complex of biophysical and biochemical properties held by idation and refolding conditions, and observed two bands of 110 and its subspecialized domains. Here a comprehensive recapitulation of 220 kDa, corresponding to AIRE homodimers and homotetramers, re- the pertaining studies is reported, with a preferential look at the distur- spectively, on native polyacrylamide gel electrophoresis; the denatur- bances in AIRE function caused by pathogenic changes occurring in its ation of the refolded product yielded a 56-kDa band (AIRE monomers). sequence. A remarkable difficulty encountered in the draft of the pres- Similarly, Western blotting of murine Aire showed positive bands at ent review was represented by the differences in the nomenclature of molecular weights compatible with a homomeric pattern. In addition, gene and protein variants utilized in the referenced articles. To reach coming from the observation that several serines and threonines (and the best accuracy and to avoid any ambiguity, we followed the related possibly Tyr394) of AIRE are phosphorylated in vivo, it was demonstrated recommendations in use [51–54]. that the corresponding in vitro process, catalyzed by the protein kinases A and C (PKA and PKC, respectively), favored AIRE homomerization. 2. The subcellular localization of AIRE In the same study the authors used an electrophoretic mobility shift assay (EMSA) to test the interaction between AIRE and 32P-labeled oli- It is widely accepted that the subset of AIRE+ mTECs is too exiguous gonucleotide sequences known to be recognized by PHD fingers and to represent an adequate substrate for the best definition of the in vivo leucin-zipper DNA-binding motifs: they found that AIRE homodimers and in vitro properties of the protein. Although there has been some suc- and homotetramers produced a strong shift with EGR-consensus cess in establishing either cultures of murine Aire-expressing mTECs motif, which contains guanine doublets spaced by seven nucleotides, [55] or, based on the common organotypic model of thymic medulla while AIRE monomers were inert. They then screened G2N7G2 and ran- and skin [56,57], cocultures of murine mTECs and dermal fibroblasts dom N26 libraries to identify further oligonucleotide sequences prefer- resulting in three-dimensional surrogates [58], most information entially targeted by AIRE: both screenings revealed that AIRE has the comes from the study of cell lines, such as human cervical carcinoma highest affinity for one TTATTA motif (T sequence) and two repeats HeLa, human-embryonic-kidney epithelial-cell line 293 (HEK293 cells), of ATTGGTTA motif (GG sequence). Finally, four oligonucleotide se- monocytes U937, COS1 and CV1 cells from the African green monkey, quences (G, GG, TG and TGG) were engaged to measure AIRE affinity and murine and human fibroblasts, transfected with AIRE (or Aire) by one-to-one EMSAs with 32P-labeled and cold competitors, and the cDNA: almost invariably, AIRE localizes into small speckles uniformly strongest binding was directed to TGG sequences [68]. distributed in the nucleoplasm, with exception of the nucleoli; this pat- In the following studies, the same research group first tested the af- tern resembles, but does not coincide with that of other proteins finity of thirteen AIRE fragments, which consisted of single or coupled known to be part of NBs, such as the promyelocytic-leukemia (PML) pro- contiguous domains, for T, GG and TGG sequences: the results cleared tein and the members of Sp family [59–61]. that several fragments of the protein retain a strong DNA-binding ability In addition, AIRE is visualized in the cytoplasm of a variable number [69]. Later, chromatin immunoprecipitation (ChIP) was used to con- of transfected cells, where it forms a scaffold-like structure reminiscent struct a library of DNA sequences that are found in the murine thymus 328 R. Perniola, G. Musco / Biochimica et Biophysica Acta 1842 (2014) 326–337 and are bound in vivo by Aire. The promoters of various genes that en- polar amino acids define the core of the structure, whose integrity is code TsAgs targeted by in Aire-deficient (Aire–/–)mice needed for the properties of AIRE [74]. CARDs are common to several were included, and in most cases the selected regions contained the pro-apoptotic proteins: not surprisingly, murine Aire+ mTECs have a above oligonucleotide sequences [70]. brief life span, followed by rapid turnover [37]. A recent study shows In any case, it should be kept in mind that DNA-binding of AIRE may that an Aire truncated chain comprising only HSR and NLS is sufficient simply represent a non-specific way of action, and that the property of to induce the apoptosis of murine mTECs, and that in HEK293 cells regulating the gene transcription should be rewritten on the basis of transfected with AIRE cDNA there is nuclear accumulation of the its participation in transcriptional machineries. glycolytic enzyme glyceraldehyde-3-phosphate dehydrogenase, which reflects the levels of nitric-oxide-induced cellular stress and apoptosis 4. Homogeneously-staining region or caspase-recruitment domain [75]. The amino terminus of AIRE was thought to perform additional func- The domain located in the amino terminus of AIRE, and consisting of tions: first, it may include a nuclear export signal, as suggested by the the amino acids (aa) 1–100, was originally named HSR and a three- dual localization of the protein. Accordingly to this hypothesis, the var- dimensional model of it predicted a four-α-helical bundle arrangement, iant p.Gly208* showed the same subcellular distribution of the wild- with the hydrophobic sides pointing to the interior of the structure [65]. type protein, while p.Met1_Lys83del was almost exclusively restricted HSR denomination, which was borrowed from Sp100 [71] and in itself to the nucleus. Moreover, after exposition to a specific inhibitor of the does not imply any structural and functional meaning, is still in use, al- nuclear export such as leptomycin B, the cells showing nuclear localiza- beit, as indicated, following studies have cleared its real nature [15]. tion of the protein approximately doubled, while the percentage of HSR is involved in the process of AIRE homomerization: in two- them showing only cytoplasmic staining decreased consensually. Inter- hybrid assays, strong homomeric interaction with the wild-type protein estingly, the exposition to leptomycin B produced the same response in was guaranteed by AIRE truncated chains consisting of aa 1–256, 1–293, the cells transfected with cDNA of AIRE deletion construct encoding 1–348 [65] and 1–524 [72] and corresponding to the variants p.Arg257* p.Gly208* [66]. (commonly observed in the Northern-European APS1 patients), Lastly, it was hypothesized that HSR would contain a consensus p.Lys294*, p.Gln349* and p.Phe525*, respectively, while it was slightly motif for sumoylation and interact with the enzyme ubiquitin- weaker using the shorter variant p.Arg139*, which is typical of the conjugating 9, which represents the E2 component of the above process Sardinian APS1 patients and retains only aa 1–138 [65]. Interestingly, [76], but the related tests gave negative results [77]. 35S-labeled variant p.Gly208* (aa 1–207) was able to bind itself when translated as GST fusion construct. In contrast, 35S-labeled AIRE 5. Nuclear localization signal variants lacking HSR, such as p.Met1_Lys83del, p.Met1_Arg174del and p.Met1_Leu291del (aa 84–545, 175–545 and 292–545, respectively), Contiguously to HSR, AIRE contains a bipartite NLS, which consists of did not show any interaction [65]. Also the variants containing missense two basic modules (aa 110–114 and 131–133). Molecules containing changes within HSR were unable to homomerize, albeit a certain degree monopartite and bipartite NLSs are transported into the nucleus, of interaction with wild-type AIRE [65] and p.Phe525* [72] was through nuclear pores, by adapters and receptors belonging to the fam- conserved. ily of karyopherins. After NLSs are recognized by specific importins α, Presumably linked to the impairment in homomerization, immuno- another molecule, named importin β, binds to this complex to mediate fluorescence of cell lines transfected with cDNA of deletion constructs its effective translocation [78]. encoding various AIRE fragments showed that the integrity of HSR is AIRE variant p.[Met1_Phe100del; Ala142*] (aa 101–141), as well as also required for a proper compartmentalization [66,73]. However, not larger amino-terminal fragments, can be efficaciously transported into all missense changes involving this domain have a significant impact: the nucleus [66]. In this sense, Ilmarinen et al. [79] showed that AIRE Halonen et al. [67] grouped them into those causing a more (group is recognized by the members of the importin-α family, mainly 1A) or less (group 1B) severe alteration of the α-helices and hydropho- importins α1, α3andα5, and that the success of the operation requires bic core, and those (group 2) predicted to alter only the α-helical sur- the intactness of aa 131–133 (LysArgLys, or KRK). Conversely, the fact face. AIRE variants of group 1A induced a diffuse, unorganized staining that the subcellular distribution of AIRE variants containing missense of the nucleus and cytoplasm, albeit the distribution of the fluorescence changes at Arg113 and Lys114 was not altered, led to the conclusion resembled roughly that of the wild-type protein. Those belonging to that the nuclear import of AIRE is provided by a monopartite, rather group 2, which includes the variant p.Tyr85Cys (commonly observed than bipartite, NLS. in the Iranian-Jewish APS1 patients), localized mainly within the nucle- Later, comparative analysis of AIRE domains not only strengthened us, where they formed irregular dots of various sizes and numbers. Last- the hypothesis that aa 110–114 represent the effective module of a ly, the distribution and appearance of group-1B variant were similar to bipartite NLS (being Arg110 and Lys111 able to neutralize the loss of those of wild-type AIRE. positive charges at aa positions 113 and 114), but also identified aa Importantly, missense changes within HSR decrease consensually 159–167 as a further NLS [80]. AIRE efficiency in the activation of the gene transcription [65,67]:anex- ception might be represented by the variants containing the most 6. SAND domain amino-terminal amino-acid substitutions, and, on the other extremity, by p.Tyr85Cys [64,67], albeit the latter would confer to the protein SAND domain (aa 189–280) is characterized by several amino-acid a precocious decay [73]. On the other hand, AIRE fragments lacking basic groups, which suggest the existence of negatively-charged ligands, this domain, such as the aforementioned variants p.Met1_Lys83del, such as the phosphate groups of DNA [17]. p.Met1_Arg174del and p.Met1_Leu291del, are still able to promote AIRE variants lacking SAND domain aggregate into coarse foci most- the gene transcription, and some discrepancy emerging from connected ly localized around the nuclear membrane [61,64,66,73].Ontheother studies may simply reflect the difference in the sensitivity of the report- hand, not all missense changes generated in this domain alter signifi- er assays utilized [65,66]. cantly the subcellular distribution of the protein [73].Gly228belongs More recently, powerful software programs for the recognition of to SAND domain: it is important to remember that AIRE variant distant polypeptide homologs by sequence alignment and structural determining the amino-acid substitution Gly228Trp is thought to model comparison have revealed a previously undiscovered CARD in cause APS1 in a dominant fashion [81]. As expected, such missense the amino terminus of AIRE: the domain is predicted to fold into a six- change impairs roughly AIRE functions: when a two-hybrid assay was α-helical bundle with Greek-key topology. Highly-conserved non- performed in mammalian cells cotransfected with AIRE cDNAs encoding R. Perniola, G. Musco / Biochimica et Biophysica Acta 1842 (2014) 326–337 329 p.Gly228Trp and the wild-type protein, homomerization, subcellular 7. Plant homeodomain 1 distribution and the property of promoting the gene transcription were significantly disturbed. The impairment was less obvious when AIRE contains two PHD fingers, named PHD1 (aa 299–340) and the cohabitation involved either an AIRE variant having within SAND PHD2 (aa 434–475), which are separated by a proline-rich region domain two missense changes other than Gly228Trp, or the naturally- (PRR, with nineteen prolines among aa 350–430). PHD fingers are occurring variant p.Pro252Leu [82]. Due to the observation that cysteine-rich domains featured by a four-cysteine, one-histidine, Gly228 is conserved across a large number of species, similar evidences three-cysteine (Cys4HisCys3, or C4HC3) motif that coordinates two were reproduced in a murine model of disease [83]. zinc ions in an interleaved fashion [18]. The amino-terminal extremities The solution structure of SAND domain of Sp100, detected by nucle- of AIRE PHDs should be represented, according to Pitkänen et al. [65],by ar magnetic resonance (NMR) spectroscopy, displays a five-stranded Cys299 and Cys434, respectively: actually, fundamental interactions of anti-parallel β-sheet and four α-helices packing against one side of both domains seem to be performed by more amino-terminal amino the β-sheet. Since highly-conserved non-polar amino acids stabilize acids, so their boundaries should be formally revised. the two regions and form the core of the structure, a similar conforma- In general, PHDs are able to read (or cause) the epigenetic marks of tion is expected for all SAND domains [84]. However, expression in the chromatin, in particular the post-translational modifications (PTMs) Escherichia coli and purification of SAND domain of AIRE results in an of the histones, mostly, but not exclusively, the methylation of the unfolded structure, indicating that additional intra- or intermolecular amino acids situated in the tail of histone H3 (henceforth indicated as forces are required to stabilize its folding (G. Musco, unpublished results). H3). The combination of these processes determines a “code” that con- According to Halonen et al. [67], DNA-binding site of AIRE SAND ditions the downstream events and ultimately regulates, both in per- should be identified in the positively-charged α-helical surface centered missive and repressive sense, the gene transcription. Importantly, around an AsnLysAlaArg (NKAR) module (aa 244–247), which in AIRE PHD1 belongs to the subfamily of PHDs that recognize the integrity some manner differs from the canonical LysAspTrpLys (KDWK) and of the above amino acids, underlining in this way that the unmodified LysAsnTrpLys (KNWK) DNA-binding modules of SAND domains. On tail of H3 represents a distinct epigenetic mark, rather than a transient the other hand, in the study of Purohit et al. [69], AIRE variants condition of post-translational default [88,89].Incontrast,aninterac- p.Asp198*, p.[Met1_Ser99del; Asp198*], p.[Met1_Ser99del; Ala281*] tion between H3 and PHD2 is improbable, due to the biophysical prop- and p.[Met1_Val188del; Ala281*] (aa 1–197, 100–197, 100–280 and erties of the latter, as delineated in the next section. 189–280, respectively), showed strong affinity for TGG sequences and The deletion of one or both PHDs do not disturb significantly the as- were the only fragments able to bind to T sequences, while the variant sociation of AIRE with the intermediate filaments and microtubules, p.[Met1_Ser196del; Ala300*] (aa 197–299) failed to bind to any oligo- provided that the amino-terminal domains of the protein are intact, nucleotide sequence: so it is possible that a module based on aa but the formation of the nuclear speckles is compromised, and the nu- 189–196 (GlnArgAlaValAlaMetSerSer, or QRAVAMSS) encloses the ef- cleus displays either large inclusions [61] or a faint, diffuse staining fective DNA-binding site of SAND domain. [64,66,73]. The consequences of single or combined missense changes Actually, the functional role of this domain has not yet been de- within PHDs are not univocal: some authors described an unorganized termined. A negative consequence of the disruption of SAND do- nuclear staining [64], while in other instances a subcellular distribution main may be the missed interaction with p63, a member of indistinguishable from that of wild-type AIRE was observed [66,67]. family of transcription factors, which is implied in the expression Again, Gaetani et al. [90] reported that two AIRE variants, p.Asp297Ala of the genes encoding the antigens of the class-II major histocom- and p.Val301Met, having missense changes within PHD1, form the nu- patibility complex (MHCII): once again, AIRE variant p.Gly228Trp clear speckles, while a variant having a similar change within PHD2, was unable to establish a heteromeric complex with p63, as p.Cys446Gly, is barely detectable both in the cytoplasm and nucleus. shown by coimmunoprecipitation (coIP) and molecular modeling In any case, the most significant impairment due to AIRE changes in- analysis [85]. volving PHDs regards the regulation of the gene transcription: as indi- Based on the homology with the protein Ultrapetala 1, which func- cated, an unmodified carboxyl terminus confers an efficiency in the tions as a trithorax-group member in maintaining the expression state above property comparable to that of the wild-type protein [66],butei- of developmental genes in Arabidopsis thaliana, it was hypothesized ther deletion or disruption of PHDs attenuate drastically such process that the proteins containing SAND domains, AIRE included, would [64–67,90–92]. To this regard, some authors delineated a different be part of transcriptional machineries, with a crucial role: to bind to rank filled by AIRE PHDs: missense changes within PHD1 would DNA, and at the same time to offer an anchorage to other proteins, determine a limited loss of efficiency [64,67,91,92],whilemissense such as those having PHD fingers, deputed to bind to the histones, and changes and frameshifts involving PHD2, either encoded by naturally- those functioning as coactivators or belonging to RNA-polymerase-II occurring AIRE variants or strategically designed to reproduce into (RNA-PolII) complexes involved in the initiation of the gene transcrip- PHD2 amino-acid substitutions corresponding to those disrupting tion and its elongation [86]. PHD1, abolish nearly completely the process [64,91,92]. However, Independently from the spectrum of roles and interactions attribut- there are studies in which such differences are less obvious or inconsis- ed, SAND domain, together with NLSs, seems to represent the privileged tent [66,90]. site of AIRE acetylation: lysates of stable AIRE-expressing cell lines test- PHD1 solution structure was determined by NMR spectroscopy on ed by anti-acetyl-lysine antibodies showed that AIRE is strongly AIRE variant p.[Met1_His292del; Pro355*] (aa 293–354): the domain acetylated following the transfection with a cDNA encoding a pro- folds into a globular structure made of three extended loops (named tein, p300, which has an acetyl-transferase domain and is paralog L1, L2 and L3), with a short two-stranded (aa 308–311 and 316–319, re- of CREB-binding protein (CBP). Mass spectrometry (MS) identified spectively β1- and β2-strand) anti-parallel β-sheet and a carboxyl- the location of 12 acetylated lysines, distributed among NLSs (at aa terminal α-helix (aa 338–342) representing the only elements of the positions 114, 120, 131, 159, 164 and 165) and SAND domain (at aa secondary structure. A well-defined tertiary structure includes aa positions 221, 222, 243, 245, 253 and 259). Interestingly, increasing 295–345, while Gln293, Lys294 and aa 346–354 are disordered [93]. amounts of p300 stabilized AIRE, but did not enhance the gene tran- The folding of PHD1 is zinc-dependent, and each of the two zinc ions, scription. Moreover, missense changes mimicking the structural which are situated at the extremities of the β-sheet, is coordinated by modifications induced by acetylation of Lys243 and Lys253 modified four amino acids in a cross-braced scheme. The first zinc-binding site the set of genes activated by AIRE, indicating that the process of is formed by the thiol groups of Cys299, Cys302 and Cys322, and by acetylation may exert a role in the selection of AIRE-dependent His319 through Nδ1 of the imidazole ring. The second zinc-binding genes [87]. site is coordinated by Cys311, Cys314, Cys337 and Cys340. The domain 330 R. Perniola, G. Musco / Biochimica et Biophysica Acta 1842 (2014) 326–337 is stabilized by hydrophobic interactions between the side-chain non- Analogously, the monomethylation, and above all the symmetric polar groups of Leu308, Ile309, Phe318, Leu323, Leu327 and Trp335. and asymmetric dimethylation of Arg2, which rests about 3.0 Å away Lastly, it is important to underline that the surface of PHD1 is negatively from the pocket rim, represent a sterical obstacle to the interaction; charged, suggesting a role in protein interactions, more than in nucleic- also the phosphorylation of Thr3 has unfavorable electrostatic and acid binding [93,94]. sterical effects on the stability of the complex [97,98]. Lastly, Chignola In agreement with such hypothesis, it was observed that AIRE inter- et al. [97] suggested a further stabilizing role for Arg8 and Lys9, as acts selectively with H3, as detected by pull-down assays performed both trimethylation and acetylation of Lys9 increase significantly the with various GST-AIRE fusion constructs [95]. Peptide microarrays con- constant of dissociation between H3 and PHD1. Moreover, calculations taining bare and modified histones, and probed with murine GST-Aire performed with the Molecular Mechanics/Poisson–Boltzmann Surface including PHD1, PHD2, or both, confirmed these findings [96].The Area (MM/PBSA) method show that Arg8 and Lys9 may favorably and interaction is effective also in vivo,asdemonstratedbytheuseofnative specifically contribute to the binding free energy of the complex, with or reconstituted nucleosomes [95,96]. PHD1 is necessary and sufficient respect to PHDs of other proteins included in the same subfamily [99]. to interact with H3, provided that its zinc-dependent folding is con- This last study also highlighted the presence, in unbound PHD1, of a served: for example, amino-acid substitutions such as Cys302Pro flapping movement involving L3 loop and the edge of the β1-strand. Im- and Cys311Tyr impair the interaction [95],asconfirmed by the intro- portantly, this intrinsic “breathing” of the domain, which could be rele- duction of the corresponding missense changes into murine Aire vant to its function, was blocked in an open conformation upon binding [96]. Similarly, a conserved H3 tail (presumably not beyond ten to the unmethylated tail of H3. Moreover, the recognition of H3 by PHD1 amino-terminal amino acids) is necessary and sufficient for the bind- strongly influenced the network of several dynamically-related amino ing of PHD1. Various basic amino acids contribute to the constitution acids, changing the conformation of the domain. Consequently, it was of a positively-charged surface that perfectly complements that of hypothesized that such events may influence the biochemical proper- the partner [95,96]. ties of the protein in its complex, and ultimately have an impact on The solution structure of PHD1 bound to peptides consisting of, or the interaction between AIRE and chromatin [99]. including an unmethylated H3 tail has given new insights into the inter- In another experimental setting, Garske et al. [100] screened, by on- molecular forces that determine the partnership [97,98]. PHD1 hosts a bead colorimetric assay, a peptide library consisting of about 5000 PTM preformed platform on which H3 creates a third β-sheet, anti-parallel combinations involving H3 tail, and found that AIRE PHD1 tolerates - to aa 306–308 (Fig. 1A). The complex is stabilized by hydrogen bonds atively, when compared to PHDs of other transcription factors, the between backbone amide and carbonyl groups of both chains, from monomethylation of Lys4, provided that Arg2 is unmethylated. Quite Arg2 to Thr6 of H3, and from Gly306 to Cys310 of PHD1, while, on the unexpectedly, the specificity of the histone “code” was significantly extremity of H3 tail, Ala1 is engaged in two hydrogen bonds with the increased by PTMs of Thr3 and Thr6, as PHD1 demonstrated really intol- backbone carbonyl groups of Pro331 and Gly333 [95,97,98].Inany erant to the phosphorylation of both threonines. Suggestively, it is suffi- case, the force really stapling PHD1 to H3 is given by the electrostatic in- cient to replace AIRE PHD1 with PHD of another transcription factor teractions between side-chain basic groups of the latter, and negatively- such as the inhibitor of growth 2 (ING2), which acts mainly as repressor charged groups of the former, mainly aspartic acids. In fact, the side- of the gene transcription and recognizes Lys4 trimethylation, to modify chain groups of Arg2 and Lys4 localize into two adjacent surface grooves AIRE attitude towards histone PTMs, and to switch the set of AIRE- of PHD1, separated by Asn295: the guanidine group of Arg2 forms elec- dependent genes [101]. trostatic interactions with the backbone carbonyl groups of Cys311 and The ultimate mechanism by which the interaction between H3 and Asp312, and the side-chain carboxyl group of the latter. In the other PHD1 regulates the gene transcription is poorly understood: it was pocket, the ε-amino group of Lys4 establishes a network of polar con- demonstrated that, in HEK293 cells, AIRE-dependent genes have a low tacts with both backbone and side-chain carbonyl groups of Asn295, level of expression and that, following AIRE transfection, their pro- and above all with the side-chain carboxyl group of Asp297 [97,98] moters are prone to acquire marks of chromatin activation. However, (Fig. 1A and B). The narrow cages entrapping Arg2 and Lys4 do not the disruption of PHD1 abrogates the transcription of only a part of the allow PTMs of the tail of H3: as a matter of fact, the monomethylation above genes [102], and increasing levels of a H3-specific demethylase of Lys4 increases the constant of dissociation between the two does not enlarge their number [103]. partners by two orders of magnitude, while the dimethylation and The hypothesis that the promoters of AIRE-dependent and AIRE- trimethylation abolish the interaction [95,97]. independent genes are simply differentiated by chromatin marks is

Fig. 1. (A) Ribbon representation of the plant homeodomain 1 (PHD1) of the autoimmune regulator (AIRE) protein (in red), bound to the tail of the histone H3 (in blue). The latter forms a third β-sheet anti-parallel to amino acids 306–308 of PHD1. The side-chain ε-amino group of Lys4 (histone H3) is entrapped in a narrow cage, where it establishes polar bridges with the backbone and side-chain carbonyl groups of Asn295, and above all with the side-chain carboxyl group of Asp297. Post-translational modifications of Lys4 impair significantly (monomethylation) or abolish (dimethylation and trimethylation) the interaction. The zinc ions are shown as spheres. (B) Electrostatic surface plot of PHD1 in complex with the tail of the histone H3. Red and blue colors indicate negatively- and positively-charged areas of the surface, respectively. R. Perniola, G. Musco / Biochimica et Biophysica Acta 1842 (2014) 326–337 331 also called in question by the ontogeny of PGE in the casein gene locus: construct, and tumble independently in solution, supporting the hy- it was seen that DNA demethylation of the region surrounding one of pothesis that the two domains have evolved towards different func- the clustered genes (casein b) precedes its transcription in murine im- tions. In agreement with previous data excluding a direct involvement mature mTECs, and that the expression involves the other genes com- of PHD2 in AIRE binding to H3 tail [95,96], pull-down assay of H3 and prised in the locus, spreading in either direction. In strict correlation GST-AIRE variant p.Cys446Gly confirms that PHD2 is dispensable for with the degree of gene transcription, the chromatin becomes such interaction, at least in vitro [90]. However, as indicated, the struc- decondensed, thus offering novel points of anchorage to the transcrip- tural integrity of PHD2 appears to be fundamental for AIRE property of tional machinery, and acquires marks of activation upon terminal regulating the gene transcription, as unfolding or the deletion of the en- mTEC differentiation. Overall, these data suggest that AIRE property of tire domain leads to a failure in the activation of AIRE-dependent genes regulating the gene transcription could not reflect a mechanism merely [90,107]. This result has been recently confirmed by deletion of PHD2 based on the recognition of the unmethylated tail of H3 [104]. from murine Aire [108]. Other functions were attributed to PHD1: due to the structural It should not be thought that the functional domains of AIRE end similarity to the protein encoded by the really-interesting new gene with PHDs, as Meloni et al. [92] demonstrated clearly that the thirty (RING) [105], the domain was thought to confer to AIRE the function amino acids positioned at the carboxyl terminus of AIRE are required of ubiquitin-ligase enzyme, the E3 component of the process of for the induction of the gene transcription. In the same study, the ubiquitination [91]. Moreover, it was thought that AIRE itself would un- short fragment represented by AIRE variant p.Met1_Asp515del (aa dergo ubiquitination [63]. However, structure calculations revealed that 516–545) was sufficient to exert such function. the relationship between PHD of RING protein and AIRE PHD1 is ex- A comprehensive recapitulation of all protein variants utilized in the tremely low. At the same time, biophysical and biochemical assays did studies of the biophysical and biochemical properties of AIRE is offered not detect any direct interaction between PHD1 and its putative cognate in Table 1. E2 [93]. Again, it was suggested that, similarly to PHD of ING2, AIRE PHD1 acts as nuclear phosphoinositide [106], but once again 9. LXXLL motifs later studies did not confirm this hypothesis [93]. AIRE contains four LXXLL motifs (aa 7–11, 63–67, 414–418 and 516–520). LXXLL motifs are found in protein, either coactivators or co- 8. Plant homeodomain 2 repressors, that bind to nuclear receptors and may consequently influ- ence the process of gene transcription [19]. In this sense, it has been The solution structure of PHD2, as visualized by NMR spectroscopy demonstrated that AIRE potentiates the transcriptional activity of the – on AIRE variant p.[Met1_Gly424del; Gly486*] (aa 425 485), reveals glucocorticoids by interacting with the glucocorticoid-receptor α,and that also this domain adopts the typical globular shape based on three the sequence involved in this function would be the second (from the β extended loops, with a short two-stranded anti-parallel -sheet (aa amino terminus of the protein) LXXLL motif [109]. – – 444 445 and 452 453, respectively), followed by a carboxyl-terminal The fourth LXXLL motif lies in the thirty carboxyl-terminal amino α – – -helix (aa 473 476) (Fig. 2A). Only aa 433 476 are included in a acids of AIRE: similarly to the other missense changes involving this re- fi well-de ned tertiary structure, while the amino acids positioned more gion, the substitution of one or all three leucins (AIRE variants amino- and carboxyl-terminally are disordered. The two zinc ions are p.Leu516Ala and p.[Leu516Ala; Leu519_Leu520delinsAlaAla], respec- coordinated in a cross-braced scheme by Cys434, Cys437, His454 and tively) leads to a substantial loss of efficiency in the promotion of the Cys457, and by Cys446, Cys449, Cys472 and Cys475, respectively. The gene transcription [92]. stabilization is given by hydrophobic interactions between the side- chain non-polar groups of Val443, Leu444, Ala452, Phe453, Trp455, 10. Phylogenesis of AIRE domains Phe459 and Leu470 [90]. Despite the structural resemblance with PHD1, PHD2 displays a pos- By sequence alignment, Saltis et al. [80] performed a comparative itive electrostatic surface that makes it unsuitable for interactions with and phylogenetic analysis of Aire from animals of various gnathostome the tail of H3 (Fig. 2B). Among other things, NMR experiments show classes, such as an amphibian (the Western clawed frog, or Xenopus that AIRE PHDs do not interact in the context of a double-domain tropicalis), a bird (the Leghorn chicken, or Gallus gallus)andabony fish (the zebrafish, or Danio rerio), in relation to human AIRE. Mamma- lian species, such as mouse and opossum, were included as well. The study revealed that the domains associated with the typical cytoplasmic pattern and nuclear localization, as well as PHD1, are evolutionarily con- served, while other domains, in particular PHD2, are divergent (and in some cases absent). An intriguing observation was reported on Pro252: AIRE variant p.Pro252Leu was found, in a genetic condition of compound heterozy- gosity,inanApulianAPS1girl[110–112], but just a leucine occupies the corresponding amino-acid position in wild-type Aire from Pan troglodytes and other mammals such as mouse, rat and dog. It was hy- pothesized that the phenomenon may be attributable to neutralizing in- teractions between potentially damaging and compensatory alleles able to avoid deleterious consequences on the protein properties [113].

11. AIRE partners

Studies on the biophysical and biochemical properties of AIRE have Fig. 2. (A) Ribbon representation of the plant homeodomain 2 (PHD2) of the autoimmune been paralleled by the delineation of its functional partners. The first regulator (AIRE) protein (in red). The zinc ions are shown as spheres. (B) Electrostatic sur- one to be considered was CBP, a common coactivator of the gene tran- face plot of PHD2. Red and blue colors indicate negatively- and positively-charged areas of the surface, respectively. As clearly displayed, PHD2 exhibits a preponderantly positive scription containing three cysteine- and histidine-rich segments: it surface, which induces a repulsive interaction with the tail of the histone H3. was seen that in vitro and in vivo AIRE binds specifically to two of 332 R. Perniola, G. Musco / Biochimica et Biophysica Acta 1842 (2014) 326–337 these domains, named CH1 (aa 344–451) and CH3 (aa 1680–1890), and transcription given by the contemporary introduction of AIRE and CBP its efficiency in the regulation of the gene transcription depends on this overcame largely the sum of the activities of each of the two proteins, physical interaction [65]. when singularly assayed; the cooperation was null when AIRE variants Later, the same research group performed a series of reporter assays p.Leu28Pro, p.Arg257*, p.[Cys302Pro; Cys437Pro] and p.Cys437Pro re- in mammalian cells to demonstrate that the enhancement of the gene placed the wild-type protein. Moreover, although both AIRE and CBP

Table 1 AIRE (for autoimmune regulator) variants (with the related studies) used in the delineation of the biophysical and biochemical properties of the protein. Note that the variants are listed according to the position of the amino-acid change, from the amino terminus to the carboxyl terminus of AIRE. Shaded-in variants are encoded by naturally-occurring gene variants.

Authors [Refs.] Tested properties Challenged AIRE variants

Björses et al. [64], Subcellular localization, p.Arg15Leu, p.Thr16Met, p.Ala21Val, p.Leu28Pro, p.Leu29Pro, p.Trp78Arg, p.Val80Leu, p.Lys83Glu,

Halonen et al. [67] Homomer and transactivation p.Tyr85Cys, p.Tyr90Cys , p.Leu93Arg , p.Gly228Trp, p.Arg257* , p.Cys311Tyr , p.Pro326Gln, p.Leu397Cysfs*83

HSR/CARD structured p.Ile96*, p.[Arg15Leu; Ile96*], p.[Thr16Met; Ile96*], p.[Ala21Val; Ile96*], p.[Leu28Pro; Ile96*], p.[Leu29Pro;

Ile96*], p.[Trp78Arg; Ile96*], p.[Val80Leu; Ile96*], p.[Lys83Glu; Ile96*], p.[Tyr85Cys; Ile96*], p.[Tyr90Cys;

Ile96*], p.[Leu93Arg; Ile96*]

SAND-domain structured p.[Met1_Ser187del; Gln290*], p.[Met1_Ser187del; Gly228Trp; Gln290*]

participation in protein complexesp.Ala21Val, p.Leu28Pro, p.Leu29Pro, p.Trp78Arg, p.Tyr85Cys , p.Leu93Arg , p.Gly228Trp, p.Arg257* ,

p.Cys311Tyr , p.Pro326Gln, p.Leu397Cysfs*83

Pitkänen et al. Subcellular localization p.Met1_Lys83del, p.[Met1_Phe100del; Ala142*], p.Met1_Arg174del, p.Met1_Leu291del, p.Leu28Pro,

[65,66,114] p.[Leu28Pro; Lys83Glu], p.Lys83Glu , p.Gly208*, p.Arg257* , p.Lys294*, p.Cys302Pro, p.[Cys302Pro;

Cys437Pro], p.Gln349*, p.Cys437Pro

HSR/CARD structured p.Ile96*

Homomerization p.Met1_Lys83del, p.Met1_Arg174del, p.Met1_Leu291del, p.Leu28Pro, p.[Leu28Pro; Lys83Glu], p.Arg139* ,

p.Gly208*, p. Arg257*, p.Lys294*, p.Cys302Pro, p.Gln349*, p.Cys437Pro

Nuclear export p.Met1_Lys83del, p.Gly208*

Transactivation p.Met1_Lys83del, p.Met1_Arg174del, p.Met1_Leu291del, p.Leu28Pro, p.[Leu28Pro; Lys83Glu], p.Lys83Glu ,

p.Arg139* , p.Ser217*, p. Arg257*, p.Lys294*, p.Cys302Pro, p.[Cys302Pro; Cys437Pro], p.Gln349*, p.Cys437Pro

Contransactivation (CBP) p.Leu28Pro, p.Arg257* , p.[Cys302Pro; Cys437Pro], p.Cys437Pro

Purohit et al. [69] Binding to DNA p.[Met1_Ser99del; Asp198*], p.[Met1_Ser99del; Ala281*], p.[Met1_Val188del; Ala281*], p.[Met1_Ser196del;

Ala300*], p.[Met1_Ser196del; Arg356*], p.[Met1_Val279del; Gly431*], p.[Met1_Glu298del; Gly431*],

p.Met1_Gln349del, p.Met1_Cys419del, p.Met1_Arg433del, p.Met1_Ser474del, p.Pro101*, p.Asp198*

Meloni et al. [72] Homomerization p.Arg15Leu , p.Thr16Met, p.Val22_Asp23del, p.Phe77Ser, p.Trp78Arg, p.Phe525*

Ferguson et al. [74] Subcellular localization p.Met1_Phe100del, p.Leu28Pro, p.Leu29Pro, p.Lys83Glu

HSR/CARD structured p.Pro101*, p.[Leu28Pro; Pro101*], p.[Leu29Pro; Pro101*], p.[Lys83Glu; Pro101*]

Transactivation and p.Leu28Pro, p.Leu29Pro, p.Lys83Glu

cotransact. (CBP)

Ilmarinen et al. [79] Subcellular localization p. Arg113Ala, p.Arg113_Lys114delinsAlaAla, p.Lys114Glu, p.Lys131Glu, p.Arg132Ala, p.Lys133Ala

Binding to importins α p.Arg113_Lys114delinsAlaAla, p.Lys131Glu, p.Lys133Glu,p.Ala140*

Ilmarinen et al. [118] Binding to PIAS1 p.Ala140*

Meloni et al. [119] Binding to DAXX p.[Met1_Leu11del; Ser518*], p.[Met1_Leu67del; Ser518*], p.Pro101*, p.Pro162*, p.Leu414*, p.Ser518*,

p.[Trp78Arg; Ser518*]

Rinderle et al. [61] Subcellular localization p.Leu210Glnfs*18, p.Glu307Glyfs*4

Bottomley et al. [84] SAND-domain structured p.[Met1_Gly180del; Ala281*]

Ramsey et al. [73]a Subcellular localization and post- p.[Met1_Arg174del; Cys299*], p.[Met1_Arg174del; Gln346*], p.[Met1_Arg174del; Cys434*],

transfection decay p.Met1_Arg174del, p.Met1_Leu285del, p.Tyr85Cys , p.Lys221Ala, p.Lys222Ala, p.Lys222Glu, p.Glu237Ala,

p.Lys253Ala, p.Lys253Glu, p.Arg257Ala, p.Cys299*, p.Gln346*

Ilmarinen et al. [82] Subcellular localization, p.Leu28Pro, p.Tyr85Cys , p.Gly228Trp, p.[Lys243Ala; Arg247Ala], p.Pro252Leu

Homomer and transactivation

Participation in protein complexesp.[Lys243Ala; Arg247Ala ], p.Pro252Leu

Tonooka et al. [85]b SAND-domain structured p.[Met1_Ala193del; Pro280*], p. [Met1_Ala193del; Gly228Trp; Pro280*]

Binding to p63 and p.Gly228Trp

contransactivation

Saare et al. [87] Subcellular localization and p.Lys243Arg, p.Lys243Gln, p.[Lys243Arg; Lys245Arg], p.[Lys243Gln; Lys245Gln], p.[Lys243Arg; Lys253Arg],

transactivation p.[Lys243Gln; Lys253Gln], p.Lys245Arg, p.Lys245Gln, p.Lys253Arg, p.Lys253Gln

(continued on next page) R. Perniola, G. Musco / Biochimica et Biophysica Acta 1842 (2014) 326–337 333

Table 1 (continued)

Authors [Refs.] Tested property Challenged AIRE variants Akiyoshi et al. [63] Subcellular localization p.Cys299Ala, p.Cys434Ala

Uchida et al. [91] PHDs structured p.[Met1_Asp297del; Thr344*], p.[Met1_Ala432del; Gly477*]

Transactivation p.[Met1_Leu291del; Gln342*], p.[Met1_Leu291del; Arg433*], p.Met1_Leu291del, p.[Met1_Leu341del;

Arg433*], p.Met1_Leu341del, p.Met1_Ala432del, p.Cys311Tyr , p.Pro326Gln, p.Cys434Ala

Ubiquitination p.[Met1_Val279del; Glu375*], p.[Met1_Leu414del; Ala510*], p.Cys299Ala, p.[Cys299Ala; Cys434Ala],

p.Cys311Tyr , p.Pro326Gln, p.Cys434Ala Gozani et al. [106] Binding to phosphoinositides p.[Met1_Cys299del; Arg371*]

Bottomley et al. [93] PHDs structure p.[Met1_His292del; Pro355*], p.[Met1_His292del; Val301Met; Pro355*], p.[Met1_His292del; Cys311Tyr;

Pro355*], p.[Met1_His292del; Pro326Gln; Pro355*], p.[Met1_His292del; Pro326Leu; Pro355*],

p.[Met1_Pro422del; Gly486*]d, p.[Met1_Gln425del; Gly486*]

Binding to phosphoinositides p.[Met1_His292del; Pro355*]

Ubiquitination p.[Met1_Val279del; Glu375*], p.[Met1_His292del; Pro355*], p.[Met1_His292del; Cys311Tyr; Pro355*],

p.[Met1_His292del; Glu375*]

Liiv et al. [120] Binding to DNA-PK p.[Met1_His292del; Pro355*]

Phosphorylation p.[Met1_Arg174del; Cys299*], p.[Met1_Leu177del; Pro483*], p.Thr68Ala, p.Val80Leu , p.Ser156Ala,

p.Arg257* , p.[Thr68Ala; Arg257*], p.[Ser156Ala; Arg257*], p.Lys294*, p.Gln349*

Transactivation p.Thr68Ala, p.Ser156Ala

Org et al. [95] PHDs structure p.[Met1_Pro289del; Pro350*], p.[Met1_Pro289del; Asp297Ala; Pro350*], p.[Met1_Pro289del; Asp312Ala;

Pro350*], p.[Met1_His292del; Glu347*]d, p.[Met1_Asn427del; Ala482*]d

Binding to histones p.[Met1_Leu177del; Pro483*], p.[Met1_Pro289del; Pro350*], p.[Met1_Pro289del; Asp297Ala; Pro350*],

p.[Met1_Pro289del; Val301Met; Pro350*], p.[Met1_Pro289del; Cys302Pro; Pro350*], p.[Met1_Pro289del;

Cys311Tyr; Pro350*], p.[Met1_Pro289del; Asp312Ala; Pro350*], p.[Met1_Pro289del; Pro483*],

p.[Met1_His292del; Pro355*], p.[Me t1_Gly 424del; Pro483*], p.[Met1_Gln425del; Gly486*], p.Arg139* ,

p.Lys294*, p.Cys302Pro, p.Cys311Tyr , p.Gln349*

Transactivation p.Asp297Ala, p.Asp312Ala

Chignola et al. [97], PHDs structure and binding to p.[Met1_His292del; Pro355*], p.[Met1_His292del; Asn295Ala; Pro355*], p.[Met1_His292del; Asp297Ala;

Spiliotopoulos et al. [99] histones Pro355*], p.[Met1_His292del; Glu298Ala; Pro355*], p.[Met1_His292del; Arg303Pro; Pro355*],

p.[Met1_His292del; Asp304Ala; Pro355*], p.[Met1_His292del; Glu307Ala; Pro355*]

Chakravarty et al. [98] PHDs structure and binding to p.[Met1_Met184del; Lys294*], p.[Met1_Gln293del; Val348*], p.[Met1_Gln293del; Val301Met; Val348*],

histones p.[Met1_Gln293del; Cys311Tyr; Val348*], p.[Met1_Gln293del; Pro326Gln; Val348*], p.[Met1_Gln293del;

Pro326Leu; Val348*], p.[Met1_Leu428del; Pro483*]

Garske et al. [100] Binding to histones p.[Met1_His292del; Pro355*], p.[Met1_Gln425del; Gly486*]

Žumer et al. [101]c Binding to histones p.[Met1_His292del; Glu354*], p.[Met1_His292del; Glu296Ala; Glu354*], p.[Met1_His292del; Asp297Ala;

Glu354*]

Transactivation p.Lys243Ala, p.Lys245Ala, p.[Lys243Ala; Lys245Ala; Arg247Ala], p.Asp297Ala

Gaetani et al. [90] Subcell. localization, transact. and p.Asp297Ala, p.Val301Met, p.Cys446Gly

participation in protein complexes

PHDs structure p.[Met1_Gln293del; Glu354*]d, p.[Met1_Lys294del; Gly486*], p.[Met1_Gly424del; Gly486*],

p.[Met1_Gly424del; Cys446Gly; Gly486*], p.[Met1_Asn427del; Glu485*]d

Binding to histones p.Cys446Gly

Fierabracci et al. [107] PHDs structured p.[Met1_Gly431del; Pro481*]

Meloni et al. [92] Transactivation p.Met1_Asp515del, p.Cys311Tyr , p.[Cys311Tyr; Cys446Tyr], p.Cys446Tyr, p.Leu516*, p.Leu516Ala,

p.[Leu516Ala; Leu519_Leu520delinsAlaAla], p.Trp531*, p.Pro539Leu , p.[Pro539Leu; Pro544Leu]

Žumer et al. [116] Subcellular localization p. Ala505Profs*16 , p.Ala510*

Transactivation p.[Met1_Gly424del; Ala505Profs*16], p.Met1_Gly424del, p.[Met1_Val487del; Ala505Profs*16],

p.Met1_Val487del, p.Ala505Profs*16

Cotransactivation (P-TEFb) p.Met1_Val487del, p.Ala505Profs*16

HSR, homogeneously-staining region; CARD, caspase-recruitment domain; SAND, Sp100, AIRE, NucP41/75 and Deaf1; CBP, CREB-binding protein; PIAS1, protein inhibitor of activated STAT 1; DAXX, death-associated protein 6; PHDs, plant homeodomains; DNA-PK, DNA-dependent protein kinase; P-TEFb, positive transcription elongation factor b. aRamsey et al. [73] utilized also two AIRE variants lacking SAND domain and the amino acids encoded by the exon 7 of AIRE gene, respectively, but the exact spanning of the deletions was not reported. bTonooka et al. [85] utilized also six AIRE variants represented by protein fragments whose amino-acid spanning was not detailed. cŽumer et al. [101] utilized also an AIRE variant in which PHD1 was replaced by PHD of the inhibitor of growth 2 (ING2), but the amino-acid sequence of the insertion was not reported. dSoftware-based calculations and comparisons by sequence alignment and homology modeling. 334 R. Perniola, G. Musco / Biochimica et Biophysica Acta 1842 (2014) 326–337 are known to form nuclear inclusions, it was observed that the of the related genes [122]. This work produced an excellent basis for a colocalization occurs in a minority of AIRE+ mTECs [114]. CBP contrib- comprehensive delineation of the molecular interactions that involve utes to AIRE acetylation, but to a lesser extent than its paralog p300 [87]. AIRE, and overall determine its function [123,124]. In subsequent studies, AIRE was seen to interact with various pro- In this vein of research, Giraud et al. [125] disclosed the terms of the teins, both activators and repressors of the transcriptional machinery, interaction between murine Aire and RNA-PolII: Aire absence does not such as: condition the gene transcription in mTECs, but blocks its elongation briefly after starting. When Aire is introduced, mapping of Aire and i) The positive transcription elongation factor b (P-TEFb), which RNA-PolII by ChIP shows that Aire binds to RNA-PolII-rich transcription- removes the block of elongation of the gene transcription al start sites, suggesting that, irrespective of the downstream events, produced by counteracting elements [115]. To this regard, it Aire acts in this context by unleashing stalled RNA-PolII. was observed that the interaction between AIRE and P-TEFb is It is noticeable that Aire lacking PHD2 still collaborates with proteins compromised by the introduction of a frameshift in the carboxyl of the pre-mRNA processing group, thus suggesting that other domains terminus of the former (AIRE variant p.Ala505Profs*16), which (with specific reference to PHD1) are deputed to the selection of Aire- makes the complex inactive [116,117]. dependent genes. In contrast, coIP with members of the transcriptional ii) The protein inhibitor of activated STAT 1 (PIAS1; STAT stays for pathway, such as RNA-PolII, is fatally compromised [108]. signal transducer and activator of transcription), which belongs to PIAS family of coregulators targeting several transcription fac- tors involved mainly in immunological pathways [118].Inthis 12. Conclusions study, the interaction between AIRE and PIAS1 was assayed on the promoter of the gene [118]. The bulk of data above reported leads to a dual conclusion: on one iii) p63, a nuclear phosphoprotein which performs the control of cell hand, we have now remarkable knowledge of the organization of AIRE proliferation and apoptosis. As indicated, the interaction would and its murine homolog, their domains and the related subspecializations. rely on SAND domain of AIRE, and the amino-acid substitution Within certain limits, this permits to predict, in front of a deletion or dis- Gly228Trp would damage such process [85]. In the same study, ruption involving a determined region of it, the corresponding impair- AIRE-p63 interaction determined, in mTECs, the level of tran- ment in AIRE properties. Of course, it should be clearly stated that this is scription of the gene encoding the MHCII transactivator (CIITA), true exclusively for the in vitro performances of AIRE, as both human which is essential for the transcriptional activity of the MHCII APS1 and the animal models of disease do not contemplate a genotype- promoter [85]. phenotype correlation, and the clinical picture seems to be rather modu- iv) The death-associated protein 6 (DAXX), which binds to the lated by the input to the autoimmunity conferred by species-specific, ge- amino terminus of AIRE (NLS included), and necessitates other netic (on geo-ethnic, familial and individual basis) and environmental interacting factors, such as the histone deacetylase, to repress backgrounds [126–132]. AIRE action [119]. On the other hand, no enough information is currently available to v) Finally, an important partner of AIRE is the heterotrimeric com- depict the intimate mechanisms of the chief function of AIRE, the regu- plex of DNA-dependent PK (DNA-PK), which consists of two lation of the gene transcription. Most recent studies move into three 70-kDa and 86-kDa regulatory subunits (Ku70 and Ku80, respec- main directions: mapping the genes, encoding either TsAgs or factors tively), and a 465-kDa catalytic subunit (DNA-PKcs) [120].Asin- able to promote a more advanced stage of mTEC differentiation, dicated, it had been reported that AIRE phosphorylation by PKA whose transcription is regulated by AIRE; unveiling the details of the in- and PKC leads to its homomerization in vitro [68].Later,pull- teraction between AIRE and chromatin; and placing AIRE in a more down assays and coIP, followed by MS, showed that DNA-PK in- complex scenario, that of the transcriptional machineries. teracts with AIRE; specifically, DNA-PK phosphorylates AIRE at The variable expression of AIRE-dependent genes in the single mTECs Thr68 and Ser156. AIRE variants containing missense changes in- suggests that the way of AIRE action is not univocal, but either relies on volving these amino acids, such as p.Thr68Ala and p.Ser156Ala, chromatin accessibility and physical availability of the transcriptional do not undergo phosphorylation and show a reduced efficiency machinery, or complies with the pre-constituted program of differentia- in the activation of the gene transcription [120].However,a tion of each mTEC cluster. Future studies are needed to refine our under- major role attributed to DNA-PK in this field would be related standing of AIRE properties, and at the same time to disclose the details to the ability to recruit AIRE to its dependent genes with an un- of AIRE-related events in the preservation of the self-tolerance. known mechanism [101,117].

Interestingly, Tao et al. [121] used a mathematical approach to ana- Acknowledgements lyze an Affymetrix dataset mapping AIRE-dependent genes and their tendency to colocalize in non-casually distributed regions of the chro- We thank Dr K. Žumer, University of Helsinki, and Dr T. Ilmarinen, Uni- mosomes. On this basis, the authors suggested that AIRE optimizes spa- versity of Tampere, Finland; Prof B. Kyewski and Dr S. Pinto, German Can- tially its function by targeting grouped genes and obtaining a physical cer Research Center, Heidelberg, Germany; Prof C.D. Surh and Dr C. facilitation to the recruitment of multiple transcriptional components. Ramsey, The Scripps Research Institute, La Jolla, CA, USA; and Prof M.-M. Later, Abramson et al. [122] identified AIRE-associated proteins by Zhou and Dr S. Chakravarty, Icahn School of Medicine at Mount Sinai, coIP and MS analysis; the first steps were followed by quantitative mod- New York, NY, USA, for their gentle and precious collaboration. ulation of the transcription of each candidate, to assay the respective rel- evance to the expression of AIRE-dependent genes. Remarkably, the authors utilized, in addition to HEK293 cells, also two cell lines derived References from the human and murine thymic epithelia. Besides to proteins in- [1] K. Nagamine, P. Peterson, H.S. Scott, J. Kudoh, S. Minoshima, M. Heino, K.J.E. Krohn, volved in the nuclear transport (thus deputed to shuttle AIRE into and M.D. Lalioti, P.E. Mullis, S.E. Antonarakis, K. Kawasaki, S. Asakawa, F. Ito, N. Shimizu, – out of the nucleus), and other ones belonging to chromatin or binding Positional cloning of the APECED gene, Nat. Genet. 17 (1997) 393 398. [2] Finnish-German APECED Consortium, An , APECED, caused by to it, AIRE associated with proteins that are involved in post-initiation mutations in a novel gene featuring two PHD-type zinc-finger domains, Nat. Genet. events of the RNA-PolII-mediated gene transcription, with particular ref- 17 (1997) 399–403. erence to those that participate in a putative complex including DNA-PK. [3] C. Betterle, N.A. Greggio, M. Volpato, Autoimmune polyglandular syndrome type 1, J. Clin. Endocrinol. Metab. 83 (1998) 1049–1055. Finally, another consistent association was seen with proteins able to [4] J. Perheentupa, Autoimmune polyendocrinopathy-candidiasis-ectodermal dystro- process pre-mRNAs, and consequently to condition the expression level phy, J. Clin. Endocrinol. Metab. 91 (2006) 2843–2850. R. Perniola, G. Musco / Biochimica et Biophysica Acta 1842 (2014) 326–337 335

[5] L. Mittaz, C. Rossier, M. Heino, P. Peterson, K.J.E. Krohn, A. Gos, M.A. Morris, J. [37] D. Gray, J. Abramson, C. Benoist, D. Mathis, Proliferative arrest and rapid turnover Kudoh, N. Shimizu, S.E. Antonarakis, H.S. Scott, Isolation and characterization of of thymic epithelial cells expressing Aire, J. Exp. Med. 204 (2007) 2521–2528. the mouse Aire gene, Biochem. Biophys. Res. Commun. 255 (1999) 483–490. [38] J.B. Johnnidis, E.S. Venanzi, D.J. Taxman, J.P.-Y. Ting, C.O. Benoist, D.J. Mathis, Chro- [6] K. Blechschmidt, M. Schweiger, K. Wertz, R. Poulson, H.-M. Christensen, A. mosomal clustering of genes controlled by the aire transcription factor, Proc. Natl. Rosenthal, H. Lehrach, M.-L. Yaspo, The mouse Aire gene: comparative genomic se- Acad. Sci. U.S.A. 102 (2005) 7233–7238. quencing, gene organization, and expression, Genome Res. 9 (1999) 158–166. [39] J. Derbinski, S. Pinto, S. Rösch, K. Hexel, B. Kyewski, Promiscuous gene expression [7] C.-Y. Wang, J.-D. Shi, A. Davoodi-Semiromi, J.-X. She, Cloning of Aire,themouseho- patterns in single medullary thymic epithelial cells argue for a stochastic mecha- mologue of the autoimmune regulator (AIRE) gene responsible for autoimmune nism, Proc. Natl. Acad. Sci. U.S.A. 105 (2008) 657–662. polyglandular syndrome type 1 (APS1), Genomics 55 (1999) 322–326. [40] J. Villaseñor, W. Besse, C. Benoist, D. Mathis, Ectopic expression of peripheral-tissue [8] P. Björses, J. Aaltonen, N. Horelli-Kuitunen, M.-L. Yaspo, L. Peltonen, Gene defect antigens in the thymic epithelium: probabilistic, monoallelic, misinitiated, Proc. behind APECED: a new clue to autoimmunity, Hum. Mol. Genet. 7 (1998) Natl. Acad. Sci. U.S.A. 105 (2008) 15854–15859. 1547–1553. [41] A.G. Farr, J.L. Dooley, M. Erickson, Organization of thymic medullary epithelial het- [9] P. Peterson, K. Nagamine, H. Scott, M. Heino, J. Kudoh, N. Shimizu, S.E. Antonarakis, erogeneity: implications for mechanisms of epithelial differentiation, Immunol. K.J.E. Krohn, APECED: a monogenic autoimmune disease providing new clues to Rev. 189 (2002) 20–27. self-tolerance, Immunol. Today 19 (1998) 384–386. [42] J. Dooley, M. Erickson, H. Roelink, A.G. Farr, Nude thymic rudiment lacking func- [10] J. Aaltonen, P. Björses, Cloning of the APECED gene provides new insight into tional resembles respiratory epithelium, Dev. Dyn. 233 (2005) 1605–1612. human autoimmunity, Ann. Med. 31 (1999) 111–116. [43] J. Dooley, M. Erickson, A.G. Farr, An organized medullary epithelial structure in the [11] C. Szostecki, H.H. Guldner, H.J. Netter, H. Will, Isolation and characterization of cDNA normal thymus expresses molecules of respiratory epithelium and resembles the encoding a human nuclear antigen predominantly recognized by autoantibodies epithelial thymic rudiment of nude mice, J. Immunol. 175 (2005) 4331–4337. from patients with primary biliary cirrhosis, J. Immunol. 145 (1990) 4338–4347. [44] G.O. Gillard, A.G. Farr, Features of medullary thymic epithelium implicate postnatal [12] D.B. Bloch, S.M. de la Monte, P. Guigaouri, A. Filippov, K.D. Bloch, Identification and development in maintaining epithelial heterogeneity and tissue-restricted antigen characterization of a leukocyte-specific component of the nuclear body, J. Biol. expression, J. Immunol. 176 (2006) 5815–5824. Chem. 271 (1996) 29198–29204. [45] G.O. Gillard, J. Dooley, M. Erickson, L. Peltonen, A.G. Farr, Aire-dependent alter- [13] D.B. Bloch, J.-D. Chiche, D. Orth, S.M. de la Monte, A. Rosenzweig, K.D. Bloch, ations in medullary thymic epithelium indicate a role for Aire in thymic epithelial Structural and functional heterogeneity of nuclear bodies, Mol. Cell. Biol. 19 differentiation, J. Immunol. 178 (2007) 3007–3015. (1999) 4423–4430. [46] J. Dooley, M. Erickson, A.G. Farr, Alterations of the medullary epithelial compart- [14] D.B. Bloch, A. Nakajima, T. Gulick, J.-D. Chiche, D. Orth, S.M. de la Monte, K.D. Bloch, ment in the Aire-deficient thymus: implications for programs of thymic epithelial Sp110 localizes to the PML-Sp100 nuclear body and may function as a nuclear hor- differentiation, J. Immunol. 181 (2008) 5225–5232. mone receptor transcriptional coactivator, Mol. Cell. Biol. 20 (2000) 6138–6146. [47] M. Yano, N. Kuroda, H. Han, M. Meguro-Horike, Y. Nishikawa, H. Kiyonari, K. [15] K. Hofmann, P. Bucher, J. Tschopp, The CARD domain: a new apoptotic signalling Maemura, Y. Yanagawa, K. Obata, S. Takahashi, T. Ikawa, R. Satoh, H. Kawamoto, motif, Trends Biochem. Sci. 22 (1997) 155–156. Y. Mouri, M. Matsumoto, Aire controls the differentiation program of thymic epi- [16] C. Dingwall, R.A. Laskey, Nuclear targeting sequences—a consensus? Trends thelial cells in the medulla for the establishment of self-tolerance, J. Exp. Med. Biochem. Sci. 16 (1991) 478–481. 205 (2008) 2827–2838. [17] T.J. Gibson, C. Ramu, C. Gemünd, R. Aasland, The APECED polyglandular autoim- [48] Y. Nishikawa, F. Hirota, M. Yano, H. Kitajima, J. Miyazaki, H. Kawamoto, Y. Mouri, mune syndrome protein, AIRE-1, contains the SAND domain and is probably a M. Matsumoto, Biphasic Aire expression in early embryos and in medullary thymic transcription factor, [letter] Trends Biochem. Sci. 23 (1998) 242–244. epithelial cells before end-stage terminal differentiation, J. Exp. Med. 207 (2010) [18] R. Aasland, T.J. Gibson, A.F. Stewart, The PHD finger: implications for 963–971. chromatin-mediated transcriptional regulation, Trends Biochem. Sci. 20 (1995) [49] D. Danso-Abeam, S. Humblet-Baron, J. Dooley, A. Liston, Models of Aire-dependent 56–59. gene regulation for thymic negative selection, Front. Immunol. 2 (2011) e14. [19] M.J. Plevin, M.M. Mills, M. Ikura, The LxxLL motif: a multifunctional binding se- [50] M. Matsumoto, Y. Nishikawa, H. Nishijima, J. Morimoto, M. Matsumoto, Y. Mouri, quence in transcriptional regulation, Trends Biochem. Sci. 30 (2005) 66–69. Which model better fits the role of Aire in the establishment of self-tolerance: [20] S.A. Eldershaw, D.M. Sansom, P. Narendran, Expression and function of the autoim- the transcription model or the maturation model? Front. Immunol. 4 (2013) e210. mune regulator (Aire) gene in non-thymic tissue, Clin. Exp. Immunol. 163 (2011) [51] S.E. Antonarakis, Nomenclature Working Group, Recommendations for a nomen- 296–308. clature system for human gene mutations, Hum. Mutat. 11 (1998) 1–3. [21] R. Perniola, Expression of the autoimmune regulator gene and its relevance to the [52] J.T. den Dunnen, S.E. Antonarakis, Mutation nomenclature extensions and sug- mechanisms of central and peripheral tolerance, Clin. Dev. Immunol. 2012 (2012) gestions to describe complex mutations: a discussion, Hum. Mutat. 15 (2000) e207403. 7–12. [22] D. Mathis, C. Benoist, Back to central tolerance, Immunity 20 (2004) 509–516. [53] J.T. den Dunnen, S.E. Antonarakis, Nomenclature for the description of human se- [23] B. Kyewski, L. Klein, A central role for central tolerance, Annu. Rev. Immunol. 24 quence variations, Hum. Genet. 109 (2001) 121–124. (2006) 571–606. [54] P.E.M. Taschner, J.T. den Dunnen, Describing structural changes by extending HGVS [24] L. Klein, M. Hinterberger, G. Wirnsberger, B. Kyewski, in the sequence variation nomenclature, Hum. Mutat. 32 (2011) 507–511. thymus for positive selection and central tolerance induction, Nat. Rev. Immunol. [55] Y. Yamaguchi, A. Takayanagi, J. Chen, K. Sakai, J. Kudoh, N. Shimizu, Mouse 9(2009)833–844. lines expressing “Aire” and peripheral tissue-specific [25] V. Geenen, J.-J. Legros, P. Franchimont, M. Baudrihaye, M.-P. Defresne, J. Boniver, antigens reproduce in vitro negative selection of T cells, Exp. Cell Res. 317 The neuroendocrine thymus: coexistence of oxytocin and neurophysin in the (2011) 2019–2030. human thymus, Science 232 (1986) 508–511. [56] R.A. Clark, K. Yamanaka, M. Bai, R. Dowgiert, T.S. Kupper, Human skin cells support [26] V.K.M. Han, A.J. D'Ercole, P.K. Lund, Cellular localization of somatomedin thymus-independent development, J. Clin. Invest. 115 (2005) 3239–3249. (insulin-like growth factor) messenger RNA in the human fetus, Science 236 [57] P. Bonfanti, S. Claudinot, A.W. Amici, A. Farley, C.C. Blackburn, Y. Barrandon, Micro- (1987) 193–197. environmental reprogramming of thymic epithelial cells to skin multipotent stem [27] P.J. Fuller, K. Verity, Somatostatin gene expression in the thymus gland, J. Immunol. cells, [letter] Nature 466 (2010) 978–982. 143 (1989) 1015–1017. [58] S. Pinto, K. Schmidt, S. Egle, H.-J. Stark, P. Boukamp, B. Kyewski, An organotypic [28] D. Hanahan, Peripheral-antigen-expressing cells in thymic medulla: factors in coculture model supporting proliferation and differentiation of medullary thymic self-tolerance and autoimmunity, Curr. Opin. Immunol. 10 (1998) 656–662. epithelial cells and promiscuous gene expression, J. Immunol. 190 (2013) [29] M. Sospedra, X. Ferrer-Francesch, O. Domínguez, M. Juan, M. Foz-Sala, R. 1085–1093. Pujol-Borrell, Transcription of a broad range of self-antigens in human thymus sug- [59] M. Heino, P. Peterson, J. Kudoh, K. Nagamine, A. Lagerstedt, V. Ovod, A. Ranki, I. gests a role for central mechanisms in tolerance toward peripheral antigens, Rantala, M. Nieminen, J. Tuukkanen, H.S. Scott, S.E. Antonarakis, N. Shimizu, K. J. Immunol. 161 (1998) 5918–5929. Krohn, Autoimmune regulator is expressed in the cells regulating immune toler- [30] O. Werdelin, U. Cordes, T. Jensen, Aberrant expression of tissue-specificproteins ance in thymus medulla, Biochem. Biophys. Res. Commun. 257 (1999) 821–825. in the thymus: a hypothesis for the development of central tolerance, Scand. [60] P. Björses, M. Pelto-Huikko, J. Kaukonen, J. Aaltonen, L. Peltonen, I. Ulmanen, Local- J. Immunol. 47 (1998) 95–100. ization of the APECED protein in distinct nuclear structures, Hum. Mol. Genet. 8 [31] J. Derbinski, A. Schulte, B. Kyewski, L. Klein, Promiscuous gene expression in medullary (1999) 259–266. thymic epithelial cells mirrors the peripheral self, Nat. Immunol. 2 (2001) 1032–1039. [61] C. Rinderle, H.-M. Christensen, S. Schweiger, H. Lehrach, M.-L. Yaspo, AIRE encodes [32] J. Gotter, B. Brors, M. Hergenhahn, B. Kyewski, Medullary epithelial cells of the a nuclear protein co-localizing with cytoskeletal filaments: altered sub-cellular dis- human thymus express a highly diverse selection of tissue-specificgenes tribution of mutants lacking the PHD zinc fingers, Hum. Mol. Genet. 8 (1999) colocalized in chromosomal clusters, J. Exp. Med. 199 (2004) 155–166. 277–290. [33] L. Klein, M. Hinterberger, J. von Rohrscheidt, M. Aichinger, Autonomous versus [62] T. Sternsdorf, T. Grötzinger, K. Jensen, H. Will, Nuclear dots: actors on many stages, dendritic cell-dependent contributions of medullary thymic epithelial cells to cen- Immunobiology 198 (1997) 307–331. tral tolerance, Trends Immunol. 32 (2011) 188–193. [63] H. Akiyoshi, S. Hatakeyama, J. Pitkänen, Y. Mouri, V. Doucas, J. Kudoh, K. Tsurugaya, [34] A.L. Fletcher, D. Malhotra, S.J. Turley, Lymph node stroma broaden the peripheral D. Uchida, A. Matsushima, K. Oshikawa, K.I. Nakayama, N. Shimizu, P. Peterson, M. tolerance paradigm, Trends Immunol. 32 (2011) 12–18. Matsumoto, Subcellular expression of autoimmune regulator is organized in a spa- [35] J. Derbinski, J. Gäbler, B. Brors, S. Tierling, S. Jonnakuty, M. Hergenhahn, L. Peltonen, tiotemporal manner, J. Biol. Chem. 279 (2004) 33984–33991. J. Walter, B. Kyewski, Promiscuous gene expression in thymic epithelial cells is reg- [64] P. Björses, M. Halonen, J.J. Palvimo, M. Kolmer, J. Aaltonen, P. Ellonen, J. ulated at multiple levels, J. Exp. Med. 202 (2005) 33–45. Perheentupa, I. Ulmanen, L. Peltonen, Mutations in the AIRE gene: effects [36] D.H.D. Gray, N. Seach, T. Ueno, M.K. Milton, A. Liston, A.M. Lew, C.C. Goodnow, R.L. on subcellular location and transactivation function of the autoimmune Boyd, Developmental kinetics, turnover, and stimulatory capacity of thymic epi- polyendocrinopathy-candidiasis-ectodermal dystrophy protein, Am. J. Hum. thelial cells, Blood 108 (2006) 3777–3785. Genet. 66 (2000) 378–392. 336 R. Perniola, G. Musco / Biochimica et Biophysica Acta 1842 (2014) 326–337

[65] J. Pitkänen, V. Doucas, T. Sternsdorf, T. Nakajima, S. Aratani, K. Jensen, H. Will, [91] D. Uchida, S. Hatakeyama, A. Matsushima, H. Han, S. Ishido, H. Hotta, J. Kudoh, N. P. Vähämurto, J. Ollila, M. Vihinen, H.S. Scott, S.E. Antonarakis, J. Kudoh, N. Shimizu, V. Doucas, K.I. Nakayama, N. Kuroda, M. Matsumoto, AIRE functions as Shimizu, K. Krohn, P. Peterson, The autoimmune regulator protein has tran- an E3 ubiquitin ligase, J. Exp. Med. 199 (2004) 167–172. scriptional transactivating properties and interacts with the common coacti- [92] A. Meloni, F. Incani, D. Corda, A. Cao, M.C. Rosatelli, Role of PHD fingers and vator CREB-binding protein, J. Biol. Chem. 275 (2000) 16802–16809. COOH-terminal 30 amino acids in AIRE transactivation activity, Mol. Immunol. 45 [66] J. Pitkänen, P. Vähämurto, K. Krohn, P. Peterson, Subcellular localization of the au- (2008) 805–809. toimmune regulator protein. Characterization of nuclear targeting and transcrip- [93] M.J. Bottomley, G. Stier, D. Pennacchini, G. Legube, B. Simon, A. Akhtar, M. Sattler, tional activation domain, J. Biol. Chem. 276 (2001) 19597–19602. G. Musco, NMR structure of the first PHD finger of autoimmune regulator protein [67] M. Halonen, H. Kangas, T. Rüppell, T. Ilmarinen, J. Ollila, M. Kolmer, M. Vihinen, J. (AIRE1). Insights into autoimmune polyendocrinopathy-candidiasis-ectodermal Palvimo, J. Saarela, I. Ulmanen, P. Eskelin, APECED-causing mutations in AIRE re- dystrophy (APECED) disease, J. Biol. Chem. 280 (2005) 11505–11512. veal the functional domains of the protein, Hum. Mutat. 23 (2004) 245–257. [94] G. Musco, P. Peterson, PHD finger of autoimmune regulator. An epigenetic link be- [68] P.G. Kumar, M. Laloraya, C.-Y. Wang, Q.-G. Ruan, A. Davoodi-Semiromi, K.-J. Kao, tween the histone modifications and tissue-specific antigen expression in thymus, J.-X. She, The autoimmune regulator (AIRE) is a DNA-binding protein, J. Biol. Epigenetics 3 (2008) 310–314. Chem. 276 (2001) 41357–41364. [95] T. Org, F. Chignola, C. Hetényi, M. Gaetani, A. Rebane, I. Liiv, U. Maran, L. Mollica, [69] S. Purohit, P.G. Kumar, M. Laloraya, J.-X. She, Mapping DNA-binding domains of M.J. Bottomley, G. Musco, P. Peterson, The autoimmune regulator PHD finger the autoimmune regulator protein, Biochem. Biophys. Res. Commun. 327 (2005) binds to non-methylated histone H3K4 to activate gene expression, EMBO Rep. 9 939–944. (2008) 370–376. [70] Q.-G. Ruan, K. Tung, D. Eisenman, Y. Setiady, S. Eckenrode, B. Yi, S. Purohit, W.-P. [96] A.S.Koh,A.J.Kuo,S.Y.Park,P.Cheung,J.Abramson,D.Bua,D.Carney,S.E. Zheng, Y. Zhang, L. Peltonen, J.-X. She, The autoimmune regulator directly controls Shoelson, O. Gozani, R.E. Kingston, C. Benoist, D. Mathis, Aire employs a the expression of genes critical for thymic epithelial function, J. Immunol. 178 histone-binding module to mediate immunological tolerance, linking chroma- (2007) 7173–7180. tinregulationwithorgan-specificautoimmunity,Proc.Natl.Acad.Sci.U.S.A. [71] T. Sternsdorf, K. Jensen, B. Reich, H. Will, The nuclear dot protein Sp100, char- 105 (2008) 15878–15883. acterization of domains necessary for dimerization, subcellular localization, [97] F. Chignola, M. Gaetani, A. Rebane, T. Org, L. Mollica, C. Zucchelli, A. Spitaleri, V. and modification by small ubiquitin-like modifiers, J. Biol. Chem. 274 (1999) Mannella, P. Peterson, G. Musco, The solution structure of the first PHD finger of 12555–12566. autoimmune regulator in complex with non-modified histone H3 tail reveals the [72] A. Meloni, E. Fiorillo, D. Corda, R. Perniola, A. Cao, M.C. Rosatelli, Two novel muta- antagonistic role of H3R2 methylation, Nucleic Acids Res. 37 (2009) 2951–2961. tions of the AIRE protein affecting its homodimerization properties, Hum. Mutat. [98] S. Chakravarty, L. Zeng, M.-M. Zhou, Structure and site-specific recognition of his- 25 (2005) 319. tone H3 by the PHD finger of human autoimmune regulator, Structure 17 (2009) [73] C. Ramsey, A. Bukrinsky, L. Peltonen, Systematic mutagenesis of the functional do- 670–679. mains of AIRE reveals their role in intracellular targeting, Hum. Mol. Genet. 11 [99] D. Spiliotopoulos, A. Spitaleri, G. Musco, Exploring PHD fingers and H3K4me0 in- (2002) 3299–3308. teractions with molecular dynamics simulations and binding free energy calcula- [74] B.J. Ferguson, C. Alexander, S.W. Rossi, I. Liiv, A. Rebane, C.L. Worth, J. Wong, M. tions: AIRE-PHD1, a comparative study, PLoS ONE 7 (2012) e46902. Laan, P. Peterson, E.J. Jenkinson, G. Anderson, H.S. Scott, A. Cooke, T. Rich, AIRE's [100] A.L. Garske, S.S. Oliver, E.K. Wagner, C.A. Musselman, G. LeRoy, B.A. Garcia, T.G. CARD revealed, a new structure for central tolerance provokes transcriptional plas- Kutateladze, J.M. Denu, Combinatorial profiling of chromatin binding modules re- ticity, J. Biol. Chem. 283 (2008) 1723–1731. veals multisite discrimination, Nat. Chem. Biol. 6 (2010) 283–290. [75] I. Liiv, U. Haljasorg, K. Kisand, J. Maslovskaja, M. Laan, P. Peterson, AIRE-induced ap- [101] K. Žumer, A.K. Low, H. Jiang, K. Saksela, B.M. Peterlin, Unmodified histone H3K4 optosis is associated with nuclear translocation of stress sensor protein GAPDH, and DNA-dependent protein kinase recruit autoimmune regulator to target Biochem. Biophys. Res. Commun. 423 (2012) 32–37. genes, Mol. Cell. Biol. 32 (2012) 1354–1362. [76] K.A. Wilkinson, J.M. Henley, Mechanisms, regulation and consequences of protein [102] T. Org, A. Rebane, K. Kisand, M. Laan, U. Haljasorg, R. Andreson, P. Peterson, AIRE SUMOylation, Biochem. J. 428 (2010) 133–145. activated tissue specific genes have histone modifications associated with inactive [77] M.C. Rosatelli, D. Corda, E. Fiorillo, A. Cao, A. Meloni, Interacting proteins of Autoim- chromatin, Hum. Mol. Genet. 18 (2009) 4699–4710. mune Regulator (AIRE), [abstract] The American Society of Human Genetics – 56th [103] A.S. Koh, R.E. Kingston, C. Benoist, D. Mathis, Global relevance of Aire binding to Annual Meeting, New Orleans (LO), October 9–13, 2006, The American Society of hypomethylated lysine-4 of histone-3, Proc. Natl. Acad. Sci. U.S.A. 107 (2010) Human Genetics, Bethesda (MD), 2006, p. e1120, (No Eds. Listed). 13016–13021. [78] L.F. Pemberton, B.M. Paschal, Mechanisms of receptor-mediated nuclear import [104] L.-O. Tykocinski, A. Sinemus, E. Rezavandy, Y. Weiland, D. Baddeley, C. Cremer, S. and nuclear export, Traffic 6 (2005) 187–198. Sonntag, K. Willecke, J. Derbinski, B. Kyewski, Epigenetic regulation of promiscuous [79] T. Ilmarinen, K. Melén, H. Kangas, I. Julkunen, I. Ulmanen, P. Eskelin, The gene expression in thymic medullary epithelial cells, Proc. Natl. Acad. Sci. U.S.A. monopartite nuclear localization signal of autoimmune regulator mediates its nu- 107 (2010) 19426–19431. clear import and interaction with multiple importin α molecules, FEBS J. 273 [105] M.B. Metzger, V.A. Hristova, A.M. Weissman, HECT and RING finger families of E3 (2006) 315–324. ubiquitin ligases at a glance, J. Cell Sci. 125 (2012) 531–537. [80] M. Saltis, M.F. Criscitiello, Y. Ohta, M. Keefe, N.S. Trede, R. Goitsuka, M.F. Flajnik, [106] O. Gozani, P. Karuman, D.R. Jones, D. Ivanov, J. Cha, A.A. Lugovskoy, C.L. Baird, H. Zhu, Evolutionarily conserved and divergent regions of the Autoimmune Regulator S.J. Field, S.L. Lessnick, J. Villasenor, B. Mehrotra, J. Chen, V.R. Rao, J.S. Brugge, C.G. (Aire) gene: a comparative analysis, Immunogenetics 60 (2008) 105–114. Ferguson,B.Payrastre,D.G.Myszka,L.C.Cantley,G.Wagner,N.Divecha,G.D. [81] F. Cetani, G. Barbesino, S. Borsari, E. Pardi, L. Cianferotti, A. Pinchera, C. Marcocci, A Prestwich, J. Yuan, The PHD finger of the chromatin-associated protein ING2 novel mutation of the autoimmune regulator gene in an Italian kindred with auto- functions as a nuclear phosphoinositide receptor, Cell 114 (2003) 99–111. immune polyendocrinopathy-candidiasis-ectodermal dystrophy, acting in a domi- [107] A. Fierabracci, C. Bizzarri, A. Palma, A. Milillo, E. Bellacchio, M. Cappa, A novel nant fashion and strongly cosegregating with hypothyroid autoimmune heterozygous mutation of the AIRE gene in a patient with autoimmune thyroiditis, J. Clin. Endocrinol. Metab. 86 (2001) 4747–4752. polyendocrinopathy-candidiasis-ectodermal dystrophy syndrome (APECED), [82] T. Ilmarinen, P. Eskelin, M. Halonen, T. Rüppell, R. Kilpikari, G. Duran Torres, Gene 511 (2012) 113–117. H. Kangas, I. Ulmanen, Functional analysis of SAND mutations in AIRE sup- [108] S. Yang, K. Bansal, J. Lopes, C. Benoist, D. Mathis, Aire's plant homeodomain(PHD)-2 ports dominant inheritance of the G228W mutation, Hum. Mutat. 26 (2005) is critical for induction of immunological tolerance, Proc. Natl. Acad. Sci. U.S.A. 110 322–331. (2013) 1833–1838. [83] M.A. Su, K. Giang, K. Žumer, H. Jiang, I. Oven, J.L. Rinn, J.J. DeVoss, K.P.A. Johannes, [109] A. Tiulpakov, G.P. Chrousos, T. Kino, The autoimmune regulator protein (AIRE) in- W. Lu, J. Gardner, A. Chang, P. Bubulya, H.Y. Chang, B.M. Peterlin, M.S. Anderson, creases transcriptional activity of the α (GRα) via direct in- Mechanisms of an autoimmunity syndrome in mice caused by a dominant muta- teraction and this effect is dependent on the integrity of one of its LXXLL motifs, tion in Aire, J. Clin. Invest. 118 (2008) 1712–1726. [abstract] Horm. Res. 62 (Suppl. 2) (2004) 26–27. [84] M.J. Bottomley, M.W. Collard, J.I. Huggenvik, Z. Liu, T.J. Gibson, M. Sattler, The SAND [110] R. Perniola, A. Falorni, M.G. Clemente, F. Forini, E. Accogli, G. Lobreglio, Organ- domain structure defines a novel DNA-binding fold in transcriptional regulation, specific and non-organ-specific autoantibodies in children and young adults Nat. Struct. Biol. 8 (2001) 626–633. with autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy (APECED), [85] A. Tonooka, T. Kubo, S. Ichimiya, Y. Tamura, T. Ilmarinen, I. Ulmanen, S. Kimura, S. Eur. J. Endocrinol. 143 (2000) 497–503. Yokoyama, Y. Takano, T. Kikuchi, N. Sato, Wild-type AIRE cooperates with p63 in [111] A. Meloni, R. Perniola, V. Faà, E. Corvaglia, A. Cao, M.C. Rosatelli, Delineation of HLA class II expression of medullary thymic stromal cells, Biochem. Biophys. Res. molecular defects in the AIRE gene in autoimmune polyendocrinopathy- Commun. 379 (2009) 765–770. candidiasis-ectodermal dystrophy patients from Southern Italy, J. Clin. Endocrinol. [86] C.C. Carles, J.C. Fletcher, Missing links between histones and RNA Pol II arising from Metab. 87 (2002) 841–846. SAND? Epigenetics 5 (2010) 381–385. [112] R. Perniola, O. Filograna, G. Greco, V. Pellegrino, High prevalence of thyroid autoim- [87] M. Saare, A. Rebane, B. Rajashekar, J. Vilo, P. Peterson, Autoimmune regulator munity in Apulian patients with autoimmune polyglandular syndrome type 1, [let- is acetylated by transcription coactivator CBP/p300, Exp. Cell Res. 318 (2012) ter] Thyroid 18 (2008) 1027–1029. 1767–1778. [113] L. Azevedo, G. Suriano, B. van Asch, R.M. Harding, A. Amorim, Epistatic interactions: [88] C.A. Musselman, T.G. Kutateladze, PHD fingers. Epigenetic effectors and potential how strong in disease and evolution? Trends Genet. 22 (2006) 581–585. drug targets, Mol. Interv. 9 (2009) 314–323. [114] J. Pitkänen, A. Rebane, J. Rowell, A. Murumägi, P. Ströbel, K. Möll, M. Saare, J. [89] C.A. Musselman, T.G. Kutateladze, Handpicking epigenetic marks with PHD fingers, Heikkilä, V. Doucas, A. Marx, P. Peterson, Cooperative activation of transcription Nucleic Acids Res. 39 (2011) 9061–9071. by autoimmune regulator AIRE and CBP, Biochem. Biophys. Res. Commun. 333 [90] M. Gaetani, V. Matafora, M. Saare, D. Spiliotopoulos, L. Mollica, G. Quilici, F. (2005) 944–953. Chignola, V. Mannella, C. Zucchelli, P. Peterson, A. Bachi, G. Musco, AIRE-PHD [115] I. Oven, N. Brdičková, J. Kohoutec, T. Vaupotič, M. Narat, B.M. Peterlin, AIRE recruits fingers are structural hubs to maintain the integrity of chromatin-associated inter- P-TEFb for transcriptional elongation of target genes in medullary thymic epithelial actome, Nucleic Acids Res. 40 (2012) 11756–11768. cells, Mol. Cell. Biol. 27 (2007) 8815–8823. R. Perniola, G. Musco / Biochimica et Biophysica Acta 1842 (2014) 326–337 337

[116] K. Žumer, A. Plemenitaš, K. Saksela, B.M. Peterlin, Patient mutation in AIRE disrupts [126] C. Ramsey, O. Winqvist, L. Puhakka, M. Halonen, A. Moro, O. Kämpe, P. Eskelin, M. P-TEFb binding and target gene transcription, Nucleic Acids Res. 39 (2011) Pelto-Huikko, L. Peltonen, Aire deficient mice develop multiple features of APECED 7908–7919. phenotype and show altered immune response, Hum. Mol. Genet. 11 (2002) [117] K. Žumer, K. Saksela, B.M. Peterlin, The mechanism of tissue-restricted antigen 397–409. gene expression by AIRE, J. Immunol. 190 (2013) 2479–2482. [127] M.S. Anderson, E.S. Venanzi, L. Klein, Z. Chen, S.P. Berzins, S.J. Turley, H. von [118] T. Ilmarinen, H. Kangas, T. Kytömaa, P. Eskelin, J. Saharinen, J.-S. Seeler, K. Boehmer, R. Bronson, A. Dierich, C. Benoist, D. Mathis, Projection of an immunolog- Tanhuanpää, F.Y.-L. Chan, R.M. Slattery, K. Alakurtti, J.J. Palvimo, I. Ulmanen, Func- ical self shadow within the thymus by the Aire protein, Science 298 (2002) tional interaction of AIRE with PIAS1 in transcriptional regulation, Mol. Immunol. 1395–1401. 45 (2008) 1847–1862. [128] E. Kekäläinen, A. Miettinen, T.P. Arstila, Does the deficiency of Aire in mice really [119] A. Meloni, E. Fiorillo, D. Corda, F. Incani, M.L. Serra, A. Contini, A. Cao, M.C. Rosatelli, resemble human APECED?, [letter] Nat. Rev. Immunol. 7 (2007) 1. DAXX is a new AIRE-interacting protein, J. Biol. Chem. 285 (2010) 13012–13021. [129] F.-X. Hubert, S.A. Kinkel, P.E. Crewther, P.Z.F. Cannon, K.E. Webster, M. Link, R. Uibo, [120] I. Liiv, A. Rebane, T. Org, M. Saare, J. Maslovskaja, K. Kisand, E. Juronen, L. Valmu, M.K. O'Bryan, A. Meager, S.P. Forehan, G.K. Smyth, L. Mittaz, S.E. Antonarakis, P. M.J. Bottomley, N. Kalkkinen, P. Peterson, DNA-PK contributes to the phosphoryla- Peterson, W.R. Heath, H.S. Scott, Aire-deficient C57BL/6 mice mimicking the com- tion of AIRE: importance in transcriptional activity, Biochim. Biophys. Acta 1783 mon human 13- deletion mutant present with only a mild autoimmune (2008) 74–83. phenotype, J. Immunol. 182 (2009) 3902–3918. [121] Y. Tao, R. Kupfer, B.J. Stewart, C. Williams-Skipp, C.K. Crowell, D.D. Patel, S. Sain, R.I. [130] M. Halonen, P. Eskelin, A.-G. Myhre, J. Perheentupa, E.S. Husebye, O. Kämpe, F. Scheinman, AIRE recruits multiple transcriptional components to specificgenomic Rorsman, L. Peltonen, I. Ulmanen, J. Partanen, AIRE mutations and human leuko- regions through tethering to nuclear matrix, Mol. Immunol. 43 (2006) 335–345. cyte antigen genotypes as determinants of the autoimmune polyendocrinopathy- [122] J. Abramson, M. Giraud, C. Benoist, D. Mathis, Aire's partners in the molecular con- candidiasis-ectodermal dystrophy phenotype, J. Clin. Endocrinol. Metab. 87 trol of immunological tolerance, Cell 140 (2010) 123–135. (2002) 2568–2574. [123] B. Kyewski, P. Peterson, Aire, master of many trades, Cell 140 (2010) 24–26. [131] W. Jiang, M.S. Anderson, R. Bronson, D. Mathis, C. Benoist, Modifier loci condition [124] R. Towns, M. Pietropaolo, Aire's partnership: an answer for many questions and autoimmunity provoked by Aire deficiency, J. Exp. Med. 202 (2005) 805–815. new questions in search of answers, Pediatr. Diabetes 11 (2010) 85–87. [132] L. De Martino, D. Capalbo, N. Improda, F. D'Elia, R. Di Mase, R. D'Assante, I. [125] M. Giraud, H. Yoshida, J. Abramson, P.B. Rahl, R.A. Young, D. Mathis, C. Benoist, Aire D'Acunzo, C. Pignata, M. Salerno, APECED: a paradigm of complex interactions unleashes stalled RNA polymerase to induce ectopic gene expression in thymic ep- between genetic background and susceptibility factors, Front. Immunol. 4 (2013) ithelial cells, Proc. Natl. Acad. Sci. U.S.A. 109 (2012) 535–540. e331.