Beyond zinc fingers: steroid hormone receptors have a novel structural motif for DNA recognition.

The highly conserved DNA-binding domain of the steroid hormone receptors contains two 'zinc finger'-like sequence motifs. The three-dimensional structure in solution has been determined using two-dimensional 1H nuclear magnetic resonance (NMR) spectroscopy and shows that the two 'zinc finger'-like motifs fold to form a single structural domain. The combination of this structural information and mutagenesis data reveals how this family of transcriptional regulators bind to DNA.

Histone deacetylase (HDAC) 1 and 2 complexes regulate both histone acetylation and crotonylation in vivo.

Proteomic analysis of histones has shown that they are subject to a superabundance of acylations, which extend far beyond acetylation, to include: crotonylation, propionylation, butyrylation, malonylation, succinylation, ß-hydroxybutyrylation and 2-hydroxyisobutyrylation. To date, much of the functional data has focussed on histone crotonylation which, similar to acetylation, has been associated with positive gene regulation and is added by the acyltransferase, p300. Although Sirtuins 1-3, along with HDAC3, have been shown to possess decrotonylase activity in vitro, there is relatively little known about the regulation of histone crotonylation in vivo. Here we show that Histone Deacetylase 1 and 2 (HDAC1/2), the catalytic core of numerous co-repressor complexes, are important histone decrotonylase enzymes. A ternary complex of HDAC1/CoREST1/LSD1 is able to hydrolyse both histone H3 Lys18-acetyl (H3K18ac) and H3 Lys18-crotonyl (H3K18cr) peptide substrates. Genetic deletion of HDAC1/2 in ES cells increases global levels of histone crotonylation and causes an 85% reduction in total decrotonylase activity. Furthermore, we mapped H3K18cr in cells using ChIP-seq, with and without HDAC1/2, and observed increased levels of crotonylation, which largely overlaps with H3K18ac in the vicinity of transcriptional start sites. Collectively, our data indicate that HDAC1/2 containing complexes are critical regulators of histone crotonylation in vivo.

Transcriptional control: how nuclear receptors get turned on.

The recently determined structures of the ligand-binding domains from three nuclear receptors show that a receptor undergoes a significant conformational change on ligand binding. It is not yet clear how this structural change results in transcriptional activation.

Signal transduction: fast lane to transcriptional activation.

Two newly reported structures of homodimeric 'STAT' transcription factors bound to DNA reveal at atomic resolution the elegant mechanism through which kinase activity at the cell membrane can be transduced into transcriptional activation within the cell nucleus.

Limbs are moving: where are they going?

The past decade has witnessed many changes in the way in which biologists study vertebrate development. Like curious children, we have progressed from merely watching and playing with our toys to the more exciting activity of taking them apart. This progression is mainly due to the application of a number of new techniques that allow us not only to ablate gene function, but also to induce gene activity inappropriately in time and space. Through the use of these techniques we can now disassemble our 'toys' and begin to understand how the pieces fit together and, thus, we are beginning to understand how the vertebrate embryo develops. Additionally, the analysis and comparison of limb development in diverse species has provided much insight into the evolutionary mechanisms through which changes in developmental pathways have led to the extraordinary diversity of limbs.

The role of water in protein-DNA interactions.

It is becoming well accepted that water plays an important role in both the specificity and affinity of protein-DNA interactions. Recently, a combination of structural, biochemical and thermodynamic techniques has particularly enhanced our understanding of the role of water in complexes between DNA and three different proteins: the trp repressor; the homeodomain; and the glucocorticoid receptor DNA-binding domain.

Human androgen receptor gene ligand-binding-domain mutations leading to disrupted interaction between the N- and C-terminal domains.

Most mutations in the androgen receptor (AR) ligand-binding domain (LBD) disrupt binding of the natural ligands: dihydrotestosterone and testosterone. Some AR LBD mutations do not affect ligand binding but they disrupt androgen-induced interaction of the N-terminal motif FXXLF and C-terminal activation function 2 (AF2). As N-/C-terminal interaction requires binding of agonists that have androgen activity in vivo, it correlates well with the phenotype. To study this further, we searched the Cambridge intersex database for patients with a detected missense mutation in the AR LBD presenting with normal ligand binding. Six mutations (D695N, Y763C, R774H, Q798E, R855H and L907F) were selected and introduced by site-directed mutagenesis into the pSVAR and pM-LBD plasmids. The transactivational potential of the wild-type and mutant androgen receptors (pSVAR) was examined by dual-luciferase assay using pGRE-LUC as a reporter vector. N-/C-terminal interaction was studied by mammalian two-hybrid assay using wild-type and mutated AR LBD (pM-LBD), pVP16-rAR-(5-538) (encoding rat amino-terminal AR) and pCMX-UAS-TK-LUC as a reporter. AR LBD mutations D695N, R774H and L907F presented with minimal transactivational capacity and N-/C-terminal interaction was totally disrupted. Mutations Y763C and R885H had some residual dose-dependent transactivational potential and minimal N-/C-terminal interaction. Q798E presented with good transactivational potential and it showed only mild reduction in N-/C-terminal interaction. With the selected mutations, N-/C-terminal interaction correlated well with AR transactivation and the phenotype. Disrupted N-/C-terminal interaction is capable of providing the mechanism for androgen-insensitivity syndrome in most cases where the mutation in the LBD does not disrupt ligand binding. Furthermore, mutations leading to the disrupted N-/C-terminal interaction can be localized to certain critical regions in the three-dimensional structure of the AR LBD. Our study shows that apart from the previously reported regions, regions just before helix 3, between helices 5 and 6, and at helix 10 are also important for AR N-/C-terminal interaction.

The oestrogen receptor recognizes an imperfectly palindromic response element through an alternative side-chain conformation.

BACKGROUND: Structural studies of protein-DNA complexes have tended to give the impression that DNA recognition requires a unique molecular interface. However, many proteins recognize DNA targets that differ from what is thought to be their ideal target sequence. The steroid hormone receptors illustrate this problem in recognition rather well, since consensus DNA targets are rare. RESULTS: Here we describe the structure, at 2.6 A resolution, of a complex between a dimer of the DNA-binding domain from the human oestrogen receptor (ERDBD) and a non-consensus DNA target site in which there is a single base substitution in one half of the palindromic binding site. This substitution results in a 10-fold increase in the dissociation constant of the ERDBD-DNA complex. Comparison of this structure with a structure containing a consensus DNA-binding site determined previously, shows that recognition of the non-consensus sequence is achieved by the rearrangement of a lysine side chain so as to make an alternative base contact. CONCLUSIONS: This study suggests that proteins adapt to recognize different DNA sequences by rearranging side chains at the protein-DNA interface so as to form alternative patterns of intermolecular contacts.

DNA recognition by the oestrogen receptor: from solution to the crystal.

BACKGROUND: The steroid/nuclear hormone receptors are a large family of conserved ligand-activated transcription factors that regulate gene expression through binding to response elements upstream of their target genes. Most members of this family bind to DNA as homodimers or heterodimers and recognize the sequence, spacing and orientation of the two half-sites of their response elements. The recognition and discrimination of the sequence and arrangements of these half-sites are mediated primarily by a highly conserved DNA-binding domain. RESULTS: Here we describe the DNA-binding properties of the isolated DNA-binding domain of the oestrogen receptor, the ERDBD, and its refined NMR structure. This domain is monomeric in solution, but two molecules bind cooperatively to specific DNA sequences; this cooperativity determines the arrangement of half-sites that is recognized by the ERDBD. The 10 carboxy-terminal residues and a region of 15 residues within the domain are disordered in the solution structure, yet are important for DNA binding. CONCLUSION: The cooperative nature of ERDBD binding to DNA is important. The previously-determined X-ray structure of the ERDBD dimer bound to DNA shows that the 15 internal residues disordered in solution make contact both with DNA and with the corresponding region of the other monomer. These results suggest that these residues become ordered during the process of binding to DNA, forming the dimer interface and thus contributing to the cooperative interaction between monomers.

A dominant negative human peroxisome proliferator-activated receptor (PPAR){alpha} is a constitutive transcriptional corepressor and inhibits signaling through all PPAR isoforms.

Several missense mutations in the ligand-binding domain of human peroxisome proliferator-activated receptor (PPAR)gamma have been described in subjects with dominantly inherited severe insulin resistance associated with partial lipodystrophy, hypertension, and dyslipidemia. These mutant receptors behave as dominant-negative inhibitors of PPARgamma signaling when studied in transfected cells. The extent to which such dominant-negative effects extend to signaling through other coexpressed PPAR isoforms has not been evaluated. To examine these issues further, we have created a PPARalpha mutant harboring twin substitutions, Leu459Ala and Glu462Ala, within the ligand binding domain (PPARalpha(mut)), examined its signaling properties, and compared the effects of dominant-negative PPARalpha and PPARgamma mutants on basal and ligand-induced gene transcription in adipocytes and hepatocytes. PPARalpha(mut) was transcriptionally inactive, repressed basal activity from a PPAR response element-containing promoter, inhibited the coactivator function of cotransfected PPAR-gamma coactivator 1alpha, and strongly inhibited the transcriptional response to cotransfected wild-type receptor. In contrast to PPARgamma, wild-type PPARalpha failed to recruit the transcriptional corepressors NCoR and SMRT. However, PPARalpha(mut) avidly recruited these corepressors in a ligand-dissociable manner. In hepatocytes and adipocytes, both PPARalpha(mut) and the corresponding PPARgamma mutant were capable of inhibiting the expression of genes primarily regulated by PPARalpha, -gamma, or -delta ligands, albeit with some differences in potency. Thus, dominant-negative forms of PPARalpha and PPARgamma are capable of interfering with PPAR signaling in a manner that is not wholly restricted to their cognate target genes. These findings may have implications for the pathogenesis of human syndromes resulting from mutations in this family of transcription factors.

Solution structures of two zinc-finger domains from SWI5 obtained using two-dimensional 1H nuclear magnetic resonance spectroscopy. A zinc-finger structure with a third strand of beta-sheet.

This paper describes the detailed three-dimensional structures of two zinc-finger domains from the yeast transcription factor SWI5, calculated using the results of the n.m.r. experiments described in the accompanying paper. The structure of finger 2 is essentially similar to those previously obtained by others for isolated, synthetic single zinc-finger domains in solution, and for the three zinc-finger peptide Zif268 in its crystalline complex with DNA. The N-terminal half of the sequence forms a two-stranded, irregular beta-sheet containing both of the metal-binding cysteine residues, while the remainder of the structure forms a helix. Approximately the first half of this helix is alpha-helical, whereas the C-terminal portion, including the two metal-binding histidine residues, is 3(10) helical. Four invariant hydrophobic residues form a core to the structure. In contrast to all previously described structures of zinc-finger domains, finger 1 has an additional strand in the beta-sheet, formed by residues N-terminal to the formal start of the finger motif. This additional strand plays a role in stabilising the folded form of finger 1, since a two-finger peptide lacking the N-terminal residues showed folded structure in finger 2 but not in finger 1.

Three novel mutations at serine 314 in the thyroid hormone beta receptor differentially impair ligand binding in the syndrome of resistance to thyroid hormone.

The syndrome of resistance to thyroid hormone is associated with diverse mutations in the ligand-binding domain of the thyroid hormone beta receptor, localizing to three clusters around the hormone binding cavity. Here, we report three novel resistance to thyroid hormone mutations (S314C, S314F, and S314Y), due to different nucleotide substitutions in the same codon, occurring in six separate families. Functional characterization of these mutant receptors showed marked differences in their properties. S314F and S314Y receptor mutants exhibited significant transcriptional impairment in keeping with negligible ligand binding and were potent dominant negative inhibitors of wild-type receptor action. In contrast, the S314C mutant bound ligand with reduced affinity, such that its functional impairment and dominant negative activity manifest at low concentrations of thyroid hormone, but are more reversible at higher T3 concentrations. The degree of functional impairment of mutant receptors in vitro may correlate with the magnitude of thyroid dysfunction in vivo. Modelling these mutations using the crystal structure of thyroid hormone receptor beta shows why ligand binding is perturbed and why the phenylalanine/tyrosine mutations are more deleterious than cysteine.

St John's wort, a herbal antidepressant, activates the steroid X receptor.

St John's wort (SJW), an extract of the medicinal plant Hypericum perforatum, is widely used as a herbal antidepressant. Recently, this agent has been found to adversely affect the metabolism of various coadministered drugs. Steroid X receptor (SXR), an orphan nuclear receptor, induces hepatic cytochrome P450 gene expression in response to diverse endogenous steroids, xenobiotics and drugs. Here, we report that, when coexpressed with SXR, a reporter construct derived from the cytochrome P450 3A promoter is activated by St John's wort. A GAL4-SXR ligand binding domain (LBD) fusion mediates concentration-dependent transactivation by SJW, whereas a mutant GAL4-SXR fusion, containing substitutions in key residues in a transactivation domain, is inactive. SJW recruits steroid receptor coactivator-1 to SXR in a two-hybrid assay and competes with radiolabelled ligand in binding studies, suggesting it interacts directly with the receptor LBD. Of two constituents of SJW, we find that hyperforin, but not hypericin, mediates both transactivation and coactivator recruitment by SXR. Our observations suggest that SXR activation by St John's wort mediates its adverse interaction with drugs metabolised via the CYP 3A pathway. Future development of SJW derivatives lacking SXR activation, may enable its antidepressant and drug-metabolising properties to be dissociated.

All wrapped up.

How is it possible that nine small repeated 'zinc finger' units (each spanning just 3 or 4 base pairs) can protect the whole 50 base pair binding site of TFIIIA and why should such a periodic protein structure give rise to such an asymmetric footprint on DNA? The crystal structure of the first six fingers of TFIIIA bound to 31 base pairs of DNA explains everything: not all zinc fingers act alike.

Protein motifs 5. Zinc fingers.

The term zinc finger was first used to describe a 30-residue, repeated sequence motif found in an unusually abundant Xenopus transcription factor. It was proposed that each motif is folded around a central zinc ion to form an independent minidomain and that adjacent zinc fingers are combined as modules to make up a DNA-binding domain with the modules "gripping" the DNA (hence the term finger). We now know that these proposals were correct and that these DNA-binding motifs are found in many eukaryotic DNA-binding proteins. More recently, crystal structures of three different complexes between zinc finger domains and their target DNA binding sites have revealed a remarkably simple mode of interaction with DNA. The simplicity of the zinc finger structure, and of its interaction with DNA, is a very striking feature of this protein domain. After the discovery of the zinc finger motif, patterns of potential zinc ligands have been found in several other proteins, some of which also bind to DNA. Structural studies of these domains have revealed how zinc can stabilize quite diverse protein architectures. In total, 10 such small zinc-binding domains have been studied structurally. These form a diverse collection, but each in turn has been termed a zinc finger motif-although clearly what they have in common is only their zinc-binding property, which stabilizes an apparently autonomously folded unit.

Towards an understanding of protein-DNA recognition.

Understanding how proteins recognize DNA in a sequence-specific manner is central to our understanding of the regulation of transcription and other cellular processes. In this article we review the principles of DNA recognition that have emerged from the large number of high-resolution crystal structures determined over the last 10 years. The DNA-binding domains of transcription factors exhibit surprisingly diverse protein architectures, yet all achieve a precise complementarity of shape facilitating specific chemical recognition of their particular DNA targets. Although general rules for recognition can be derived, the complex nature of the recognition mechanism precludes a simple recognition code. In particular, it has become evident that the structure and flexibility of DNA and contacts mediated by water molecules contribute to the recognition process. Nevertheless, based on known structures it has proven possible to design proteins with novel recognition specificities. Despite this considerable practical success, the thermodynamic and kinetic properties of protein/DNA recognition remain poorly understood.