Health Articles

 

Pii: s0959-440x(99)80007-4

sb9103.qxd 12/16/1999 9:19 AM Page 48 Combinatorial gene regulation by eukaryotic transcription factors
Lin Chen
Recent structure determinations of high order complexes of eukaryotic transcription factors bound to DNA have revealedthat residues from their DNA-binding domains are involved in MCM1 MADS box dimer protein–protein interactions between distinct factors.
Protein–protein interactions between transactivation domains and coactivators have also been characterized in a number of recently determined structures. These studies support thecombinatorial mechanism of transcription regulation ineukaryotic cells and multicellular organisms.
Department of Chemistry and Biochemistry, University of Colorado at MATα2 homeodomain Modeled half-site Boulder, Boulder, CO 80309-0215, USA;e-mail: [email protected] Current Opinion in Structural Biology 1999, 9:48–55
Elsevier Science Ltd ISSN 0959-440X activation function-2 activating protein-1 CREB-binding protein cAMP-response element CRE-binding protein Modeled half-site GA-binding protein general transcription factor Current Opinion in Structural Biology CREB-binding domain of CBP The structure of MCM1 (residues 1–100) and MATα2 (residues ligand-binding domain 103–189) bound to DNA. On a natural a-specific operator, the MCM1
nuclear factor of activated T cells MADS box DNA-binding site is flanked on both sides by a MATα2 site.
nuclear factor κB The MADS box dimer and the left hand MATα2 homeodomain bound phosphorylated kinase-inducible domain to DNA is the crystallographically observed complex [9••]. The PPAR-γ peroxisome proliferator-activated receptor-γ
MATα2–DNA complex on the right hand (light shading and thin stick Rel homology region DNA drawing) was modeled manually to represent a steroid receptor coactivator MATα2–MCM1–MATα2 complex on an a-specific operator. Each
signal transducer and activator of transcription subunit is shaded differently. Note that the β hairpin of MATα2 extends TBP-associated factor the central β sheet of MCM1, which forms the major protein–protein TATA-box-binding protein interaction interface. The β-strand interaction is probably conserved in the modeled right-hand half-site, but the rest of the linker and the DNA transcription factor conformation may be different with different DNA spacing. (a) View
herpes simplex virus protein-16 perpendicular to the MCM1 dimer dyad axis. (b) View from the top,
with a rotation 90° from (a). Note the conserved minor groove DNA
binding by the MATα2 homeodomain ‘arm'. All figures drawn using Regulated eukaryotic gene transcription involves the MOLSCRIPT [52].
assembly of an initiation complex at the core promoterregion and regulatory complexes at promoter-enhancer/operator regions [1]. The core promoter complex here because of the limited space. This review will focus of RNA polymerase II contains multiple protein factors on recent structural studies of high order regulatory com- (referred to as general transcription factors, GTFs), includ- plexes [9••–11••] and activation domain–coactivator ing the TATA-box-binding protein (TBP) and its complexes [12••–15••]. associated factors (TAFs). The structures of a number ofbinary and ternary GTF–DNA complexes have been The activation of eukaryotic genes in vivo often requires determined in previous years [2–4]. More GTF complexes the coordinated binding of multiple transcription factors will probably appear soon. Structural studies of eukaryotic to the promoter-enhancer region; many of these factors transcription factors continue to reveal novel DNA-bind- are regulated by distinct signal transduction pathways. In ing motifs and variants of known motifs [5•,6•,7,8•]. Many several cases, it has been shown that the binding of mul- of these studies from the past year will not be discussed tiple transcription factors to a specific promoter-enhancer sb9103.qxd 12/16/1999 9:19 AM Page 49 Combinatorial gene regulation by eukaryotic transcription factors Chen 49
region is cooperative and requires a unique composition and spatial arrangement of transcription-factor-bindingsites [16•,17,18]. The assembly of these enhancer com- plexes (also referred to as the enhanceosome [19]) is facilitated by protein–protein interactions betweenDNA-bound factors and protein-induced DNA bending.
Two features of enhanceosome assembly are especially important for combinatorial transcription regulation: com-binations of multiple transcription factors generate diverse patterns of regulation [20,21] and highly coopera-tive binding ensures the specificity of transcriptional The enhanceosome may serve as a template for the assembly of the core promoter complex by either direct orcoactivator-mediated interactions. Most eukaryotic tran-scription factors contain one or more transactivationdomain involved in interactions with downstream coacti-vators and GTFs [23]. These transactivation domains areusually rich in proline and glutamine, or acidic aminoacids, and appear to be very flexible before binding totheir respective targets. This flexibility is probably the main reason that the structural characterization of thesedomains lagged behind that of the DNA-binding Current Opinion in Structural Biology domains. The recent identification of various targets of The structure of NFAT1 (residues 399–678), Fos (residues 138–200) transactivation domains, including TAFs, cAMP- and Jun (residues 254–315) bound to DNA [10••]. Multiple loops (loops response element (CRE) binding protein (CREB) 1, 2 and 3 are shown in this orientation) presented by the NFAT N- binding proteins (CBPs/p300) and steroid receptor coac- terminal RHR (RHR-N) contact Fos and Jun. Each component (Fos, Junand the two NFAT domains [RHR-N and RHR-C]) is shaded differently. tivators (SRCs), however, has allowed the structuralcharacterization of several transactivation domains incomplex with their respective target proteins [12••–15••].
bonding of mainchain atoms and hydrophobic packing of Structures of eukaryotic transcription factor
sidechains. β-strand-mediated interactions have also been complexes bound to DNA
observed between TFIIA and TBP in TFIIA–TBP–DNA In Saccharomyces cerevisiae, three transcription factors, the complexes [3,4] and may be involved in a wide range of homeodomain proteins MATα2 and MATa1, and the physiological and pathological protein–protein interactions. MADS box protein MCM1, play crucial roles in cell-typespecification. Various combinations of these three factors Combinatorial control of transcription activation in high- and additional factors are involved in gene activation and er eukaryotic cells has been characterized biochemically repression in different yeast cell types, providing the best for several promoter-enhancers, including that of the example of the combinatorial control of eukaryotic gene TCR α gene [18], the IFN-β gene [16•] and the inter- regulation [24]. These transcription factors bind various leukin (IL)-2 gene [27]. A crystallographic study of this DNA sites cooperatively to form higher order complexes last was reported recently [10••]. The nuclear factor of that have distinct regulatory functions. The crystal struc- activated T cells (NFAT) and members of the activating ture of a MATα2–MATa1–DNA ternary complex revealed protein-1 (AP-1) transcription factor family (including that cooperative DNA binding is induced by a C-terminal Fos and Jun) bind cooperatively to their target DNA sites amphipathic α helix of MATα2 bound to a hydrophobic in the promoter-enhancer and participate in the tran- groove on the MATa1 homeodomain [25]. In another com- scriptional regulation of IL-2 and other immune response plex, that of MATα2–MCM1–MATα2–DNA, an genes [28]. Although NFAT is activated by calcium sig- N-terminal flexible linker of MATα2 has been suggested nals through calcineurin, the AP-1 transcription factors by biochemical studies to interact with the DNA-binding are induced by agents that activate protein kinase C. The domain (referred to as the MADS box) of MCM1 to medi- DNA-binding domain of NFAT is distantly related to ate cooperative DNA binding [26]. Indeed, the recent that of the Rel family of transcription factors, including crystal structure of a DNA-bound ternary complex of the nuclear factor-KB (NF-KB) proteins. The DNA-bind- MCM1 and MATα2 [9••] shows that part of this flexible ing domains of AP-1 proteins consist of the basic region, linker forms a β hairpin that binds to the outer strand of the leucine zipper (bZIP) motif. The DNA-binding domains central β sheet of the MCM1 MADS box (Figure 1). The of NFAT, Fos and Jun are necessary and sufficient for binding interactions involve parallel β-strand hydrogen cooperative DNA binding. The recently determined sb9103.qxd 12/16/1999 9:19 AM Page 50 Protein–nucleic acid interactions
enhances GABPα DNA binding both by displacing theC-terminal α helix from its inhibitory orientation and bydirectly stabilizing the DNA-binding residues ofGABPα. GABPβ is also thought to mediate the higher order assembly of a GABPβ/α complex on DNA (αβ–αβ dimer) through its C-terminal leucine zipper motif. Complexes of transactivation domains and
their target proteins
The complex structure of a CREB activation domain (the
phosphorylated kinase-inducible domain, pKID)
bound to its target (the KIX domain) on CBP has been
determined by NMR spectroscopy [12••]. Complex inter-actions between the herpes simplex virus protein-16(VP16) activation domain and human TAF31 were alsocharacterized by an NMR study [13••]. More recently, thecrystal structure of the ligand-binding domain (LBD) of anuclear receptor, the peroxisome proliferator-activatedreceptor-γ (PPAR-γ), which contains the activation func-tion-2 (AF-2) motif, was determined in a ternary complex with SRC-1 and the anti-diabetic ligand rosiglitazone Current Opinion in Structural Biology [14••,15••]. CBP, TAFs and SRC represent the major tar-gets of transcription activators bound to upstream Complex of GABPβ (residues 1–157) and GABPα (residues enhancers. In both the pKID–KIX and the VP16–TAF 311–430) bound to DNA [11••]. Each subunit is shaded differently complexes, the transactivation domain presents an amphi- (GABPβ dark and GABPα light). Note that the GABPβ ankyrin repeatsbind to the GABPα Ets domain and the C-terminal α helix using the pathic α helix to a hydrophobic surface on their respective extended loop tips. The ankyrin repeats of GABPβ do not contact the target. These helices are flexible and largely unstructured DNA in this structure.
before binding. This flexibility may be important for thebinding of the same region in different conformations toother regulatory proteins. In the pKID–KIX complex, a crystal structure of this complex [10••] shows phosphoserine at one end of the amphipathic helix plays protein–protein interactions between NFAT and Fos–Jun an important regulatory role, conveying the phosphoryla- that involve multiple patches of surface residues on the tion signal through a hydrogen bond interaction. In the three proteins and that are facilitated by the bending of ternary complex of PPAR-γ LBD, rosiglitazone and SRC- the DNA and the Fos α helix (Figure 2). Previous bio- 1, the coactivator SRC-1 displays an amphipathic helix to chemical studies had suggested that many of these a surface formed by AF-2 and other structural elements interacting residues were important for complex assem- from the nuclear receptor LBD. SRC-1 binding is ligand bly [29–31]. In the ternary complex, the NFAT dependent, providing another means of conveying a signal DNA-binding domain seems to have moved closer to to the transcription machinery. Fos–Jun from its orientation in the binary NFAT–DNAcomplex [32•]. The interaction surfaces are mostly Implications for the mechanisms of eukaryotic
hydrophilic, with a small hydrophobic center. transcription regulation
Together with biological studies on transcriptional synergy
DNA binding by a transcription factor can also be mod- or ‘cross talk', what do the above complex structures tell us ulated by interactions with a protein that does not bind about the combinatorial gene regulation mechanisms in DNA itself. One example is the GA-binding protein eukaryotic cells? (GABP) that binds DNA with a conserved GA sequencemotif, whose core DNA-binding complex structure was Members of a transcription factor family may be
recently determined [11••] (Figure 3). The two subunit differentiated in higher order transcription
GABP complex is involved in the transcriptional regula- tion of mitochondrial protein genes and some viral In the higher order complex structures discussed above, the genes. GABPα belongs to the Ets family of transcription residues of a DNA-binding domain that do not contact the factors, which have a ‘winged helix-turn-helix' DNA- DNA mediate protein–protein interactions. Transcription fac- binding motif. GABPβ is a large protein that contains a tors from the same family usually have highly homologous leucine zipper-like motif and a number of ankyrin DNA-binding surfaces, but different surface residues outside repeats, but it does not bind DNA directly. GABPβ binds the DNA-binding region. There is some biological evidence to the GABPα Ets motif and a C-terminal α helix using that these variable residues may be important determinants the extended loop tips of its ankyrin repeats. GABPβ of each family member's distinct in vivo functions [33,34].
sb9103.qxd 12/16/1999 9:19 AM Page 51 Combinatorial gene regulation by eukaryotic transcription factors Chen 51
A sequence comparison of NFAT (NFATs 1–4) and Fos (c- Fos, FosB, Fra-1, Fra-2) family members indicated thatdifferent NFAT–Fos pairs should have different interaction interfaces [10••]. A similar specificity is also seen in the MATα2–MATa1–DNA complex, in which the MATα2 C- terminal α helix specifically binds to the MATa1homeodomain, but not its own homeodomain [25]. The assembly of specific transcription factor complexes
contributes to the specificity of transcriptional control
For each individual transcription factor in the ternary
complexes discussed above, the homeodomain–DNA,
MADS box–DNA and Ets–DNA interactions are all simi- lar to the interactions seen in their respective binaryprotein–DNA complexes. This is true even when thebinding site deviates significantly from the consensussequence, as does DNA binding by Fos–Jun in theNFAT–Fos–Jun complex. The recognition of nonconsen- sus sites by a transcription factor often involves only local changes or adjustments of DNA-binding residues. Suchapparent ‘relaxation' of specificity seems very common inin vivo DNA binding by eukaryotic transcription factors, especially as part of larger complexes. This may allow the assembly of enhancer complexes on various compositesites while maintaining the specificity by cooperative pro- tein–protein interactions. In higher order complexes, protein–protein interactionsadd specificity to the combined DNA-binding specificityof each component. In the MATα2–MATa1–DNA andMATα2–MCM1–DNA complexes, protein–protein inter-actions are mediated by structural modules that aretethered to DNA-binding domains through a peptide link-er. These linkers, though flexible in the crystal structures (as evident from high B factors), still seem to impose speci-ficity on complex formation by restricting the arrangementof each component's binding site [35]. In contrast, the Current Opinion in Structural Biology interactions between NFAT and Fos–Jun are directlymediated by residues of their respective DNA-binding Comparison of NFAT in the (a) NFAT–Fos–Jun–DNA complex and (b)
domains and are continuous with the DNA-binding sur- NF-κB p50 in the p50 dimer–DNA complex. The N-terminal DNA- faces, similar to a nuclear receptor heterodimer–DNA binding domains of NFAT and p50 are similarly anchored on the DNA,but their C-terminal domains are in very different orientations. The C- complex [36]. Thus, the whole complex has a continuous terminal domain of NFAT is similar to that of p50 (root mean square DNA-binding groove, explaining why the spacing deviation of 0.84 Å for 65 β strand Cα atoms, and most of the between the NFAT and AP-1 DNA-binding sites is highly dimerization residues at the abe sheet are conserved in NFAT). Significant conformational changes of the protein and region of DNA-bound bZIP proteins is usually flexible, DNA are observed in the above higher order complexes which may favor interactions with other proteins on and they may play important roles in the specificity DNA, although it is not clear if members of the Fos fam- (‘indirect read out') and diversity of higher order complex ily have different flexibility in this region or not. DNA assembly. In the MATα2–MCM1–DNA complex, the conformational changes in the NFAT–Fos–Jun complex MATα2 linker contacting MCM1 shows flexibility by may explain why the DNA spacer between the NFAT being able to assume either a β strand or an α-helical con- and AP-1 sites shows a strong preference for AT-rich formation. Such flexibility may be important for MATα2 binding to a MCM1 dimer on both sides, with a two orthree base pair spacing on a natural a-specific operator Another level of specificity that is achieved through higher [9••]. The Fos α helix in the NFAT–Fos–Jun–DNA com- order complex assembly is the specific orientation of het- plex has a significant bend in its fork region. The fork erodimeric DNA-binding proteins on DNA. Fos and Jun sb9103.qxd 12/16/1999 9:19 AM Page 52 Protein–nucleic acid interactions
Minor groove contact Current Opinion in Structural Biology A comparison of the immunoglobulin DNA-binding domains of (a) p53,
groove, whereas a β-strand loop is inserted in the adjacent minor (b) NFAT, (c) STAT, (d) NF-κB and (e) the T domain. The β barrel is
groove. The orientation of p53 on DNA is significantly different from oriented similarly, with the ABE sheet shown in front. The conserved ab that of the others. NF-κB does not contact the DNA minor groove. A loop bound in the DNA major groove is indicated. Note that in p53, C-terminal α helix of the T domain binds deeply in the minor groove, NFAT and STAT, the C-terminal α helix is positioned in the major opposite the groove contacted by p53, NFAT and STAT.
have almost identical DNA-binding surfaces. Their bZIP without binding to DNA. As seen in the crystal structures, heterodimer was found to bind DNA in two orientations in the protein–protein interaction interfaces observed in the binary (Fos–Jun)–DNA complexes [37,38], but adopted these complexes are either limited or highly hydrophilic.
a unique orientation in the ternary NFAT–Fos–Jun–DNA This feature is also observed in other complexes, including complex [37]. As seen in the ternary crystal structure [10••], those of the signal transducer and activator of transcription the NFAT–Fos and NFAT–Jun interaction interfaces are (STAT) proteins. Crystal structures of two STAT DNA- different, leading to orientation specificity on the DNA.
binding complexes and a STAT N-terminal domain have Such a specific orientation of heterodimeric transcription been determined recently [39••,40••,41•]. DNA-bound factors on DNA, either through their own asymmetric DNA STAT dimers are suggested to form higher order complex- binding or through interactions with partner proteins, may es through the conserved N-terminal protein interaction have functional importance. domain that probably only dimerizes after DNA binding.
The potential dimer interface of this STAT N-terminal Higher order complexes form upon DNA binding
domain is highly hydrophilic, as seen in the crystallograph- Unlike many tightly associated transcription factor dimers, ic dimer [41•]. Complex formation between NFAT and most of the ternary complexes discussed above do not form Fos–Jun, and between MATα2 and MATa1 is also DNA- sb9103.qxd 12/16/1999 9:19 AM Page 53 Combinatorial gene regulation by eukaryotic transcription factors Chen 53
binding-dependent. This DNA-dependent complex for- recent structure determinations of NFAT, STATs and the mation may enable the combinatorial use of transcription T-domain [47•] reveal that these proteins, together with factors [25].
p53 [48] and NF-κB [43,44], form a superfamily that usesthe Ig module as a scaffold for presenting various sec- Combinatorial diversity is achieved at multiple levels
ondary structural elements for DNA recognition In addition to complex formation between distinct and (Figure 5). Whether the above proteins are evolutionarily related transcription factors, the same DNA-binding related or not, however, is unclear at present. The recent domain of a given transcription factor may adopt different structure determination of a DNA-binding complex of conformations in different promoter contexts, while Skn-1 illustrated another way of generating diverse DNA- maintaining its specific DNA-binding interactions [42].
binding functions, combining various DNA-binding NFAT may be one example. On composite NFAT and motifs and scaffolds to form a novel DNA-binding domain AP-1 DNA-binding sites, such as those found in the IL-2 [49•]. This combination of DNA-binding modules is relat- promoter-enhancer, NFAT binds the DNA cooperatively ed to but is different from those seen in other in complex with AP-1 transcription factors. On some DNA-binding proteins, including zinc-finger proteins and other DNA sites that have two NFAT DNA-binding sites the POU domain proteins.
arranged with dyad symmetry and appropriate spacing,NFAT binds the DNA cooperatively as a dimer. The lat- Transcriptional coactivators play important roles in the
ter mode of DNA binding by NFAT may be similar to combinatorial control of transcription
that of the Rel NF-KB transcription factors, which are Considering the complexes between transactivation involved in a wide range of transcription regulation in the domains and coactivators discussed above and similar immune system and in viral gene expression. The DNA- studies on p53 [50], it seems that the binding of transacti- binding sites of the NFAT dimer resemble that of the vation domains to their target proteins generally involves NF-KB proteins (the Kb DNA site), raising the possibili- an induced amphipathic helix. This mode of protein–pro- ty that NFAT may regulate gene transcription through tein interaction is also seen in the MATα2–MATa1–DNA DNA sites (KB or KB-like DNA sites) that are also respon- complex. Amphipathic helices have one or more φxxφφ sive towards NF-KB regulation. The structures of the motifs, where φ represents a hydrophobic residue. In the DNA-binding domains of NFAT and NF-KB are also case of pKID, VP16 and other transactivation domains remarkably similar, both consisting of two Ig modules. In (p53 and p65), the motif shows a preference for pheny- the NFAT–Fos–Jun-DNA complex, each of NFAT's two lalanine and tyrosine at the first φ, imposing some Ig modules, Rel homology regions (RHRs) N and C, specificity on coactivator selection. In SRC-1 and other resembles the corresponding parts of Rel NF-κB p50, but nuclear receptor coactivators, the motif is more restricted the relative orientation of the two Ig modules in NFAT is to LXXLL (L represents leucine) [51]. The specificity of significantly different from that seen in p50–DNA com- coactivator binding also involves hydrogen bonding and plexes [43,44]. The C-terminal Ig module in p50 electrostatic interactions. mediates dimer formation, whereas NFAT is a monomerin solution [37,45] (Figure 4). Strikingly, residues Thus, interactions between transactivation domains and coac- involved in p50 dimerization are largely conserved in the tivators are mediated by small, relatively independent NFAT C-terminal Ig module, suggesting that NFAT can structural modules, each making limited but specific contacts.
probably dimerize upon binding to DNA using the same There are three LXXLL motifs in SRC-1, two of which can surface. There is evidence that NFAT may bind κB-like interact with a nuclear receptor dimer. The presence of a DNA sites as a dimer under certain physiological situa- third LXXLL may imply that these motifs can be used tions [28], probably by adopting a conformation that is either combinatorially with different nuclear receptor dimers very much like a typical NF-κB dimer (Figure 4).
or simultaneously in forming higher order complexes.
Crystals of a NFAT dimer bound cooperatively to a κB Similarly, coactivators such as CBP and p300 also contain DNA fragment have been obtained and the structure separated binding sites for the transactivation domains of var- determination is underway (L Chen, A Breier, B Tasic, ious transcription factors. These general features of SC Harrison, unpublished data). Alternatively, the transactivation domain–coactivator interactions are consistent exposed NFAT C-terminal dimer interface can mediate with the combinatorial gene regulation mechanism. higher order complex formation between multiple copiesof DNA-bound NFAT–Fos–Jun complexes on an enhancer [46], similar to the higher order assembly of In the past year, we have seen more structural characteri- DNA-bound GABPβ/α and STAT dimers (see above). zations of the protein–protein interactions involved in thefunction of eukaryotic transcription factors. These include The Ig DNA-binding domain is also found in a novel class the interactions between DNA-bound transcription fac- of transcription regulators, referred to as the T-domain tors and interactions between transactivation domains and proteins. The T-domain proteins are the products of the their respective targets. These structural studies support so-called T-box gene, which plays important roles in tran- the combinatorial transcription regulation mechanism in scription regulation in embryonic development. The eukaryotic cells.
sb9103.qxd 12/16/1999 9:20 AM Page 54 Protein–nucleic acid interactions
Structural studies of eukaryotic transcription factors with 12. Radhakrishnan I, Perez AG, Parker D, Dyson HJ, Montminy MR, Wright PE: Solution structure of the KIX domain of CBP bound to
higher and higher complexity continue. These structures the transactivation domain of CREB: a model for
will allow the analysis of the molecular details of activator–coactivator interactions. Cell 1997, 91:741-752.
enhancer complexes (enhanceosome), the core promoter The first detailed structural characterization of the interactions between anactivation domain (CREB) and a coactivator (CBP).
complex and the interactions between them. These mol- 13. Uesugi M, Nyanguile O, Lu H, Levine AJ, Verdine GL: Induced α helix
ecular pictures will provide an important framework for in the VP16 activation domain upon binding to a human TAF.
studying and understanding the in vivo mechanisms of Science 1997, 277:1310-1313.
The first structural characterization of the interactions between VP16 and eukaryotic gene regulation. TAF31, using NMR spectroscopy and biochemical methods.
14. Westin S, Kurokawa R, Nolte RT, Wisely GB, McInerney EM, Rose DW, Milburn MV, Rosenfeld MG, Glass CK: Interactions
controlling the assembly of nuclear-receptor heterodimers and
The author is grateful to Ernest Fraenkel, Rachelle Gaudet, Susanne co-activators. Nature 1998, 395:199-202.
Swalley and Marc Jacobs for critically reading this review and for their many See annotation to [15••].
helpful comments. The author would like to thank Stephen C Harrison fora wonderful postdoctoral experience in his laboratory. A postdoctoral 15. Nolte RT, Wisely GB, Westin S, Cobb JE, Lambert MH, Kurokawa R, fellowship from the Medical Foundation is acknowledged.
Rosenfeld MG, Willson TM, Glass CK, Milburn MV: Ligand binding
and coactivator assembly of the peroxisome proliferator-activated
receptor-γ. Nature 1998, 395:137-143.
References and recommended reading
This paper, together with [14••], describes the first crystallographic study of anuclear receptor ligand-binding domain and its coactivator in a ternary complex.
Papers of particular interest, published within the annual period of review,have been highlighted as: 16. Kim TK, Maniatis T: The mechanism of transcriptional synergy of
an in vitro assembled interferon-β enhanceosome. Mol Cell 1997,
• of special interest •• of outstanding interest A detailed biochemical study on enhanceosome assembly, demonstrating a Tjian R, Maniatis T: Transcriptional activation: a complex puzzle
direct correlation between cooperative DNA binding and transcriptional syn- with few easy pieces. Cell 1994, 77:5-8.
ergy in the context of an assembled enhancer.
Tan S, Richmond TJ: Eukaryotic transcription factors. Curr Opin
Garrity PA, Chen D, Rothenberg EV, Wold BJ: Interleukin-2
Struct Biol 1998, 8:41-48.
transcription is regulated in vivo at the level of coordinated
binding of both constitutive and regulated factors. Mol Cell Biol
Tan S, Hunziker Y, Sargent DF, Richmond TJ: Crystal structure of a
yeast TFIIA/TBP/DNA complex. Nature 1996, 381:127-151.
18. Grosschedl R: Higher-order nucleoprotein complexes in
Geiger JH, Hahn S, Lee S, Sigler PB: Crystal structure of the yeast
transcription: analogies with site-specific recombination.
TFIIA/TBP/DNA complex. Science 1996, 272:830-836.
Curr Opin Cell Biol 1995, 7:362-370.
Shi Y, Wang YF, Jayaraman L, Yang H, Massague J, Pavletich NP: 19. Thanos D, Maniatis T: Virus induction of human IFNβ gene
Crystal structure of a Smad MH1 domain bound to DNA:
expression requires the assembly of an enhanceosome. Cell
insights on DNA binding in TGF-β signaling. Cell 1998,
This paper describes the first crystal structure of the DNA-binding domain 20. Yamamoto KR, Pearce D, Thomas J, Miner JN: Combinatorial
from a Smad family protein in complex with DNA, showing a novel regulation at a mammalian composite response element. In
DNA-binding mode involving a β hairpin bound in the major groove.
Transcriptional Regulation. Edited by McKnight SL, Yamamoto KR.
New York: Cold Spring Harbor Laboratory Press; Escalante CR, Yie J, Thanos D, Aggarwal AK: Structure of IRF-1 with
bound DNA reveals determinants of interferon regulation. Nature
1998, 391:103-106.
21. Johnson A: A combinatorial regulatory circuit in budding yeast. In
The first crystal structure determination of a DNA-binding domain from the Transcriptional Regulation, vol 2. Edited by McKnight SL, Yamamoto interferon regulatory factor (IRF) family of proteins in complex with DNA, KR. New York: Cold Spring Harbor Laboratory Press; showing a variant of the ‘helix-turn-helix' DNA-binding motif.
Chen FE, Huang DB, Chen YQ, Ghosh G: Crystal structure of
22. Hochschild A, Ptashne M: Cooperative binding of lambda
p50/p65 heterodimer of transcription factor NF-κB bound to DNA.
repressors to sites separated by integral turns of the DNA helix.
Nature 1998, 391:410-413.
Cell 1986, 44:681-687.
Parraga A, Bellsolell L, Ferre-D'Amare AR, Burley SK: Co-crystal
23. Ptashne M, Gann A: Transcriptional activation by recruitment.
structure of sterol regulatory element binding protein 1a at 2.3 Å
Nature 1997, 386:569-577.
resolution. Structure 1998, 6:661-672.
A recent example showing that binding to a nonconsensus DNA site by a tran- 24. Johnson AD: Molecular mechanisms of cell-type determination in
scription factor involves only localized changes at the protein–DNA interface.
budding yeast. Curr Opin Genet Dev 1995, 5:552-558.
Tan S, Richmond TJ: Crystal structure of the yeast MATα2/MCM1/
25. Li T, Stark MR, Johnson AD, Wolberger C: Crystal structure of the
DNA ternary complex. Nature 1998, 391:660-666.
MATa1/MAT α2 homeodomain heterodimer bound to DNA.
A higher order eukaryotic transcription factor complex bound to DNA, demon- Science 1995, 270:262-269.
strating the importance of protein–protein interactions in the assembly.
26. Vershon AK, Johnson AD: A short, disordered protein region
10. Chen L, Glover JN, Hogan PG, Rao A, Harrison SC: Structure of the
mediates interactions between the homeodomain of the yeast α2
DNA-binding domains from NFAT, Fos and Jun bound specifically
protein and the MCM1 protein. Cell 1993, 72:105-112.
to DNA. Nature 1998, 392:42-48.
Crabtree GR, Clipstone NA: Signal transmission between the
A higher order eukaryotic transcription factor complex bound to DNA, demon- plasma membrane and nucleus of T lymphocytes. Annu Rev
strating the importance of protein–protein interactions in the assembly.
Biochem 1994, 63:1045-1083.
11. Batchelor AH, Piper DE, de la Brousse FC, McKnight SL, 28. Rao A, Luo C, Hogan PG: Transcription factors of the NFAT family:
Wolberger C: The structure of GABPα/β: an ETS domain-ankyrin
regulation and function. Annu Rev Immunol 1997, 15:707-747.
repeat heterodimer bound to DNA. Science 1998,
279:1037-1041.
29. Erlanson DA, Chytil M, Verdine GL: The leucine zipper domain
The structure determined shows a different kind of higher order transcription controls the orientation of AP-1 in the NFAT–AP-1–DNA complex.
factor complex, as compared with [9••] and [10••], in that one component Chem Biol 1996, 3:981-991.
does not bind DNA directly; however, the protein–protein interactions areequally important for the assembly and the enhanced DNA binding. It is also 30. Peterson BR, Sun LJ, Verdine GL: A critical arginine residue
one of the first two structures to show the detailed protein–protein interac- mediates cooperativity in the contact interface between
tions mediated by ankyrin repeats that play important roles in a wide range transcription factors NFAT and AP-1. Proc Natl Acad Sci USA
of physiologically important protein complexes. sb9103.qxd 12/16/1999 9:20 AM Page 55 Combinatorial gene regulation by eukaryotic transcription factors Chen 55
31. Sun LJ, Peterson BR, Verdine GL: Dual role of the nuclear factor of
for the DNA-binding-dependent, higher order complex assembly of tran- activated T cells insert region in DNA recognition and cooperative
scription factors.
contacts to activator protein 1. Proc Natl Acad Sci USA 1997,
94:4919-4924.
42. Lefstin JA, Yamamoto KR: Allosteric effects of DNA on
transcriptional regulators. Nature 1998, 392:885-888.
32. Zhou P, Sun LJ, Dotsch V, Wagner G, Verdine GL: Solution structure
of the core NFATC1/DNA complex. Cell 1998, 92:687-696.
43. Ghosh G, van Duyne G, Ghosh S, Sigler PB: Structure of NF-κB
An NMR study of a binary NFAT–DNA complex, showing the different orien- p50 homodimer bound to a κB site. Nature 1995, 373:303-310.
tation of NFAT on DNA compared with [10••].
44. Muller CW, Rey FA, Sodeoka M, Verdine GL, Harrison SC: Structure
of the NF-κB p50 homodimer bound to DNA. Nature
33. Diamond MI, Miner JN, Yoshinaga SK, Yamamoto KR: Transcription
factor interactions: selectors of positive or negative regulation
from a single DNA element. Science 1990, 249:1266-1272.
45. Hoey T, Sun YL, Williamson K, Xu X: Isolation of two new members
of the NF-AT gene family and functional characterization of the
34. Davis RL, Cheng PF, Lassar AB, Weintraub H: The MyoD DNA
NF-AT proteins. Immunity 1995, 2:461-472.
binding domain contains a recognition code for muscle-specific
gene activation. Cell 1990, 60:733-746.
46. Rooney JW, Sun YL, Glimcher LH, Hoey T: Novel NFAT sites that
mediate activation of the interleukin-2 promoter in response to
35. Jin Y, Mead J, Li T, Wolberger C, Vershon AK: Altered DNA
T-cell receptor stimulation. Mol Cell Biol 1995, 15:6299-6310.
recognition and bending by insertions in the α2 tail of the yeast
α1/α2 homeodomain heterodimer. Science 1995, 270:290-293.
Muller CW, Herrmann BG: Crystallographic structure of the T
domain-DNA complex of the Brachyury transcription factor. Nature
36. Rastinejad F, Perlmann T, Evans RM, Sigler PB: Structural
determinants of nuclear receptor assembly on DNA direct
The structure not only expands the Ig DNA-binding protein family, but it also repeats. Nature 1995, 375:203-211.
reveals unusual minor groove DNA binding by an α helix. This observation Chen L, Oakley MG, Glover JN, Jain J, Dervan PB, Hogan PG, Rao A, is important for understanding the general mechanisms of DNA recognition Verdine GL: Only one of the two DNA-bound orientations of AP-1
by proteins.
found in solution cooperates with NFATp. Curr Biol 1995,
48. Cho Y, Gorina S, Jeffrey PD, Pavletich NP: Crystal structure of a p53
tumor suppressor-DNA complex: understanding tumorigenic
mutations.
38. Glover JN, Harrison SC: Crystal structure of the heterodimeric
Science 1994, 265:346-355.
bZIP transcription factor c-Fos-c-Jun bound to DNA. Nature 1995,
49. Rupert PB, Daughdrill GW, Bowerman B, Matthews BW: A new
DNA-binding motif in the Skn-1 binding domain-DNA complex.
Nat Struct Biol 1998, 5:484-491.
39. Chen X, Vinkemeier U, Zhao Y, Jeruzalmi D, Darnell JJ, Kuriyan J: This structure shows a novel DNA-binding domain that can be regarded as Crystal structure of a tyrosine phosphorylated STAT-1 dimer
a hybridization of parts of other known DNA-binding motifs, illustrating a bound to DNA. Cell 1998, 93:827-839.
unique way of varying DNA-binding motifs in order to generate new DNA- See annotation to [40••].
binding functions by nature.
40. Becker S, Groner B, Muller CW: Three-dimensional structure of the
50. Kussie PH, Gorina S, Marechal V, Elenbaas B, Moreau J, Levine AJ, Stat3β homodimer bound to DNA. Nature 1998, 394:145-151.
Pavletich NP: Structure of the MDM2 oncoprotein bound to the
The structures described in this paper and [39••], the first for the STAT fam- p53 tumor suppressor transactivation domain. Science 1996,
ily, show the unique features of the STAT family proteins, whose multiple functional domains are highly integrated at the structural level.
51. Heery DM, Kalkhoven E, Hoare S, Parker MG: A signature motif in
41. Vinkemeier U, Moarefi I, Darnell J Jr, Kuriyan J: Structure of the
transcriptional co-activators mediates binding to nuclear
amino-terminal protein interaction domain of STAT-4. Science
receptors. Nature 1997, 387:733-736.
This is the first crystallographic study of potential protein-binding surfaces 52. Kraulis PJ: MOLSCRIPT: a program to produce both detailed and
from the conserved STAT N-terminal domain. The structural features of the schematic plots of protein structures. J Appl Crystallogr 1991,
dimer interface observed in the crystal structure have general implications

Source: http://thacher.biz/ChenLab/Selected_papers/COSB_LinChen_review99.pdf