Homeodomain determinants of major groove recognition

Dynamics and recognition of homeodomain containing protein-DNA complex of IRX4

Iroquois Homeobox 4 (IRX4) belongs to a family of homeobox TFs having roles in embryogenesis, cell specification, and organ development. Recently, large scale genome-wide association studies and epigenetic studies have highlighted the role of IRX4 and its associated variants in prostate cancer. No studies have investigated and characterized the structural aspect of the IRX4 homeodomain and its potential to bind to DNA. The current study uses sequence analysis, homology modeling, and molecular dynamics simulations to explore IRX4 homeodomain-DNA recognition mechanisms and the role of somatic mutations affecting these interactions. Using publicly available databases, gene expression of IRX4 was found in different tissues, including prostate, heart, skin, vagina, and the protein expression was found in cancer cell lines (HCT166, HEK293), B cells, ascitic fluid, and brain. Sequence conservation of the homeodomain shed light on the importance of N- and C-terminal residues involved in DNA binding. The specificity of IRX4 homodimer bound to consensus human DNA sequence was confirmed by molecular dynamics simulations, representing the role of conserved amino acids including R145, A194, N195, S190, R198, and R199 in binding to DNA. Additional N-terminal residues like T144 and G143 were also found to have specific interactions highlighting the importance of N-terminus of the homeodomain in DNA recognition. Additionally, the effects of somatic mutations, including the conserved Arginine (R145, R198, and R199) residues on DNA binding elucidated the importance of these residues in stabilizing the protein-DNA complex. Secondary structure and hydrogen bonding analysis showed the roles of specific residues (R145, T191, A194, N195, R198, and R199) in maintaining the homogeneity of the structure and its interaction with DNA. The differences in relative binding free energies of all the mutants shed light on the structural modularity of this protein and the dynamics behind protein-DNA interaction. We also have predicted that the C-terminal sequence of the IRX4 homeodomain could act as a potential cell-penetrating peptide, emphasizing the role these small peptides could play in targeting homeobox TFs.

Phylogenetic and mutational analyses of human LEUTX, a homeobox gene implicated in embryogenesis

Recently, human PAIRED-LIKE homeobox transcription factor (TF) genes were discovered whose expression is limited to the period of embryo genome activation up to the 8-cell stage. One of these TFs is LEUTX, but its importance for human embryogenesis is still subject to debate. We confirmed that human LEUTX acts as a TAATCC-targeting transcriptional activator, like other K50-type PAIRED-LIKE TFs. Phylogenetic comparisons revealed that Leutx proteins are conserved across Placentalia and comprise two conserved domains, the homeodomain, and a Leutx-specific domain containing putative transcriptional activation motifs (9aaTAD). Examination of human genotype resources revealed 116 allelic variants in LEUTX. Twenty-four variants potentially affect function, but they occur only heterozygously at low frequency. One variant affects a DNA-specificity determining residue, mutationally reachable by a one-base transition. In vitro and in silico experiments showed that this LEUTX mutation (alanine to valine at position 54 in the homeodomain) results in a transactivational loss-of-function to a minimal TAATCC-containing promoter and a 36 bp motif enriched in genes involved in embryo genome activation. A compensatory change in residue 47 restores function. The results support the notion that human LEUTX functions as a transcriptional activator important for human embryogenesis.

Lysine Side-Chain Dynamics in the Binding Site of Homeodomain/DNA Complexes As Observed by NMR Relaxation Experiments and Molecular Dynamics Simulations

An important but poorly characterized contribution to the thermodynamics of protein-DNA interactions is the loss of entropy that occurs from restricting the conformational freedom of amino acid side chains. The effect of restricting the flexibility of several side chains at a protein-DNA interface may be comparable in many cases to the other factors that determine the binding thermodynamics and may, therefore, play a key role in dictating the binding affinity and/or specificity. Because the entropic contributions, including the presence and influence of side-chain dynamics, are especially difficult to estimate based on structural information, it is important to pursue experimental and theoretical studies that can provide direct information regarding these issues. We report on studies of a model system, the homeodomain/DNA complex, focusing on the Lys50 class of homeodomains where a key lysine residue in position 50 was shown previously to be critical for binding site specificity. NMR methodology was employed for determining the dynamics of lysine side-chain amino groups via 15N relaxation measurements in the Lys50-class homeodomains from the Drosophila protein Bicoid and the human protein Pitx2. In the case of Pitx2, complexes with both a consensus and a nonconsensus DNA binding site were examined. NMR-derived order parameters indicated moderate to substantial conformational freedom for the lysine NH3+ group in the complexes studied. To complement the experimental NMR measurements, molecular dynamics simulations were performed for the consensus complexes to gain further, detailed insights regarding the dynamics of the Lys50 side chain and other important residues in the protein-DNA interface.

The combination of sequence-specific and nonspecific DNA-binding modes of transcription factor SATB1

Transcription factor SATB1 (special AT-rich sequence binding protein 1) contains multiple DNA-binding domains (DBDs), i.e. two CUT-domain repeats (CUTr1 and CUTr2 from the N-terminus) and a homeodomain, and binds to the matrix attachment region (MAR) of DNA. Although CUTr1 and the homeodomain, but not CUTr2, are known to contribute to DNA binding, different research groups have not reached a consensus on which DBD is responsible for recognition of the target sequence in MAR, 5'-TAATA-3'. Here, we used isothermal titration calorimetry to demonstrate that CUTr1 has binding specificity to this motif, whereas the homeodomain shows affinity for a variety of DNAs without specificity. In line with nonspecific DNA-binding properties of the homeodomain, a mutation of the invariant Asn at position 51 of the homeodomain (typically in contact with the A base in a sequence-specific binding mode) did not affect the binding affinity significantly. The NMR analyses and computational modeling of the homeodomain, however, revealed the tertiary structure and DNA-binding mode that are typical of homeodomains capable of sequence-specific binding. We believe that the lack of highly conserved basic residues in the helix relevant to the base recognition loosens its fitting into the DNA groove and impairs the specific binding. The two DBDs, when fused in tandem, showed strong binding to DNA containing the 5'-TAATA-3' motif with an affinity constant >10(8) M(-1) and retained nonspecific binding activity. The combination of the sequence-specific and nonspecific DNA-binding modes of SATB1 should be advantageous in a search for target loci during transcriptional regulation.

Transcription of follicle-stimulating hormone subunit genes is modulated by porcine LIM homeobox transcription factors, LHX2 and LHX3

The LIM-homeobox transcription factors LHX2 and LHX3s (LHX3a and LHX3b) are thought to be involved in regulating the pituitary glycoprotein hormone subunit genes Cga and Fshβ. These two factors show considerable differences in their amino acid sequences for DNA binding and protein-protein interactions and in their vital function in pituitary development. Hence, we compared the DNA binding properties and transcriptional activities of Cga and Fshβ between LHX2 and LHX3s. A gel mobility shift assay for approximately 1.1 kb upstream of Cga and 2.0 kb upstream of Fshβ varied in binding profiles between LHX2 and LHX3s. DNase I footprinting revealed DNA binding sites in 8 regions of the Cga promoter for LHX2 and LHX3s with small differences in the binding range and strength. In the Fshβ promoter, 14 binding sites were identified for LHX2 and LHX3, respectively. There were alternative binding sites to either gene in addition to similar differences observed in the Cga promoter. The transcriptional activities of LHX2 and LHX3s according to a reporter assay showed cell-type dependent activity with repression in the pituitary gonadotrope lineage LβT2 cells and stimulation in Chinese hamster ovary lineage CHO cells. Reactivity of LHX2 and LHX3s was observed in all regions, and differences were observed in the 5'-upstream region of Fshβ. However, immunohistochemistry showed that LHX2 resides in a small number of gonadotropes in contrast to LHX3. Thus, LHX3 mainly controls Cga and Fshβ expression.

OCT4/SOX2-independent Nanog autorepression modulates heterogeneous Nanog gene expression in mouse ES cells

NANOG, OCT4 and SOX2 form the core network of transcription factors supporting embryonic stem (ES) cell self-renewal. While OCT4 and SOX2 expression is relatively uniform, ES cells fluctuate between states of high NANOG expression possessing high self-renewal efficiency, and low NANOG expression exhibiting increased differentiation propensity. NANOG, OCT4 and SOX2 are currently considered to activate transcription of each of the three genes, an architecture that cannot readily account for NANOG heterogeneity. Here, we examine the architecture of the Nanog-centred network using inducible NANOG gain- and loss-of-function approaches. Rather than activating itself, Nanog activity is autorepressive and OCT4/SOX2-independent. Moreover, the influence of Nanog on Oct4 and Sox2 expression is minimal. Using Nanog:GFP reporters, we show that Nanog autorepression is a major regulator of Nanog transcription switching. We conclude that the architecture of the pluripotency gene regulatory network encodes the capacity to generate reversible states of Nanog transcription via a Nanog-centred autorepressive loop. Therefore, cellular variability in self-renewal efficiency is an emergent property of the pluripotency gene regulatory network.

Structural effect of the L16Q, K50E, and R53P mutations on homeodomain of pituitary homeobox protein 2

The transcription factor pituitary homeobox protein 2 (PITX2) is involved in genetic control of development. Mutations in PITX2, most in the homeodomain, cause the autosomal-dominant disorder Rieger syndrome. The mutants L16Q, K50E and R53P destabilize the structure and disrupt DNA-binding activity. The biological functions of these mutants have been characterized but not the structural basis behind the loss of DNA-binding activity. We performed multiple molecular dynamics simulations at 37°C to investigate the structural and dynamic effects of the 3 PITX2 homeodomain mutants. Compared with the wild type (WT), the L16Q mutant induces a kink in the α3 helix, which is stabilized by the hydrogen bond of Q21-R59. The disruption in backbone hydrogen bonds of V47-N51 and W48-R52 leads to a kink formation in the α3 helix of K50E. The R53P mutant alters the relative orientation of helices, which is apparently stabilized by the formation of new hydrogen bonds of T38-Q11, T38-Q12, T38-R2, N39-R2, L40-Q1, L40-R2, and T41-Q4. The hydrophobic core residues F8, L13, L40 and V45 change their positions in all mutants to break the hydrophobic core. Thus, changes in helical orientations and hydrophobic core cause rearrangement of the DNA-binding surface and disrupt DNA-binding activity in the mutants. The structural and molecular dynamics properties of 3 PITX2 homeodomain mutants differ from those of the WT, especially in formation of a kink in the recognition helix, change in the packing of helices and disruption of the hydrophobic core. This structural basis for the loss of DNA-binding activity for these polymorphisms may help in understanding the effect of mutations on other homeodomains with other diseases.

Exploring the DNA-recognition potential of homeodomains

The recognition potential of most families of DNA-binding domains (DBDs) remains relatively unexplored. Homeodomains (HDs), like many other families of DBDs, display limited diversity in their preferred recognition sequences. To explore the recognition potential of HDs, we utilized a bacterial selection system to isolate HD variants, from a randomized library, that are compatible with each of the 64 possible 3' triplet sites (i.e., TAANNN). The majority of these selections yielded sets of HDs with overrepresented residues at specific recognition positions, implying the selection of specific binders. The DNA-binding specificity of 151 representative HD variants was subsequently characterized, identifying HDs that preferentially recognize 44 of these target sites. Many of these variants contain novel combinations of specificity determinants that are uncommon or absent in extant HDs. These novel determinants, when grafted into different HD backbones, produce a corresponding alteration in specificity. This information was used to create more explicit HD recognition models, which can inform the prediction of transcriptional regulatory networks for extant HDs or the engineering of HDs with novel DNA-recognition potential. The diversity of recovered HD recognition sequences raises important questions about the fitness barrier that restricts the evolution of alternate recognition modalities in natural systems.

Molecular cloning of LIM homeodomain transcription factor Lhx2 as a transcription factor of porcine follicle-stimulating hormone beta subunit (FSHβ) gene

We cloned the LIM-homeodomain protein LHX2 as a transcription factor for the porcine follicle-stimulating hormone β subunit gene (Fshβ) by the Yeast One-Hybrid Cloning System using the upstream region of -852/-746 bases (b) from the transcription start site, called Fd2, as a bait sequence. The reporter assay in LβT2 and CHO cells revealed the presence of an LHX2-responsive region other than Fd2. A potential LHX2 binding sequence was confirmed as AATTAAT containing a consensus homeodomain binding core sequence AATT by Systematic Evolution of Ligands by Exponential Enrichment analysis. DNase I footprinting demonstrated three AATTAAT sequences located at regions -835/-829, -818/-812 and -806/-800 b in the Fd2 region and 12 binding sites in the distal and proximal regions mostly containing an AATT-core sequence. RT-PCR analysis of Lhx2 expression during porcine fetal and postnatal pituitary development showed a gradual increase from fetal day (f) 40 to postnatal day (p) 8 followed by a slight decrease to p230, suggesting that LHX2 may play its role largely in the late fetal and postnatal periods. The analyses of Lhx2 expression in pituitary tumor-derived cell lines showed their expressions in cell lines including αT31, LβT2 and others. Since LHX2 was previously identified as a transcription factor for Cga and the in vitro experiments in the present study suggested that LHX2 regulated the expression of Fshβ, it is possible that LHX2 controls the synthesis of FSH at the transcription level.

Oct1 regulates trophoblast development during early mouse embryogenesis

Oct1 (Pou2f1) is a transcription factor of the POU-homeodomain family that is unique in being ubiquitously expressed in both embryonic and adult mouse tissues. Although its expression profile suggests a crucial role in multiple regions of the developing organism, the only essential function demonstrated so far has been the regulation of cellular response to oxidative and metabolic stress. Here, we describe a loss-of-function mouse model for Oct1 that causes early embryonic lethality, with Oct1-null embryos failing to develop beyond the early streak stage. Molecular and morphological analyses of Oct1 mutant embryos revealed a failure in the establishment of a normal maternal-embryonic interface due to reduced extra-embryonic ectoderm formation and lack of the ectoplacental cone. Oct1(-/-) blastocysts display proper segregation of trophectoderm and inner cell mass lineages. However, Oct1 loss is not compatible with trophoblast stem cell derivation. Importantly, the early gastrulation defect caused by Oct1 disruption can be rescued in a tetraploid complementation assay. Oct1 is therefore primarily required for the maintenance and differentiation of the trophoblast stem cell compartment during early post-implantation development. We present evidence that Cdx2, which is expressed at high levels in trophoblast stem cells, is a direct transcriptional target of Oct1. Our data also suggest that Oct1 is required in the embryo proper from late gastrulation stages onwards.

Mutational analysis of the preferential binding of human topoisomerase I to supercoiled DNA

Human topoisomerase I binds DNA in a topology-dependent fashion with a strong preference for supercoiled DNAs of either sign over relaxed circular DNA. One hypothesis to account for this preference is that a second DNA-binding site exists on the enzyme that mediates an association with the nodes present in supercoiled DNA. The failure of the enzyme to dimerize, even in the presence of DNA, appears to rule out the hypothesis that two binding sites are generated by dimerization of the protein. A series of mutant protein constructs was generated to test the hypotheses that the homeodomain-like core subdomain II (residues 233-319) provides a second DNA-binding site, or that the linker or basic residues in core subdomain III are involved in the preferential binding to supercoiled DNAs. When putative DNA contact points within core subdomain II were altered or the domain was removed altogether, there was no effect on the ability of the enzyme to recognize supercoiled DNA, as measured by both a gel shift assay and a competition binding assay. However, the preference for supercoils was noticeably reduced for a form of the enzyme lacking the coiled-coil linker region or when pairs of lysines were changed to glutamic acids in core subdomain III. The results obtained implicate the linker and solvent-exposed basic residues in core subdomain III in the preferential binding of human topoisomerase I to supercoiled DNA.

Dimeric PROP1 binding to diverse palindromic TAAT sequences promotes its transcriptional activity

Mutations in the Prop1 gene are responsible for murine Ames dwarfism and human combined pituitary hormone deficiency with hypogonadism. Recently, we reported that PROP1 is a possible transcription factor for gonadotropin subunit genes through plural cis-acting sites composed of AT-rich sequences containing a TAAT motif which differs from its consensus binding sequence known as PRDQ9 (TAATTGAATTA). This study aimed to verify the binding specificity and sequence of PROP1 by applying the method of SELEX (Systematic Evolution of Ligands by EXponential enrichment), EMSA (electrophoretic mobility shift assay) and transient transfection assay. SELEX, after 5, 7 and 9 generations of selection using a random sequence library, showed that nucleotides containing one or two TAAT motifs were accumulated and accounted for 98.5% at the 9th generation. Aligned sequences and EMSA demonstrated that PROP1 binds preferentially to 11 nucleotides composed of an inverted TAAT motif separated by 3 nucleotides with variation in the half site of palindromic TAAT motifs and with preferential requirement of T at the nucleotide number 5 immediately 3' to a TAAT motif. Transient transfection assay demonstrated first that dimeric binding of PROP1 to an inverted TAAT motif and its cognates resulted in transcriptional activation, whereas monomeric binding of PROP1 to a single TAAT motif and an inverted ATTA motif did not mediate activation. Thus, this study demonstrated that dimeric binding of PROP1 is able to recognize diverse palindromic TAAT sequences separated by 3 nucleotides and to exhibit its transcriptional activity.

Regulation of porcine pituitary glycoprotein hormone alpha subunit gene with LIM-homeobox transcription factor Lhx3

The aim of this study was to characterize the promoter activity of the porcine pituitary glycoprotein hormone common alpha gene (Cga) promoter (-1059/+12) and the role of LIM homeodomain transcription factor Lhx3. A transfection assay using Chinese hamster ovary (CHO) cells showed that the -1059/-101 region of the Cga promoter definitely responds to Lhx3 and that the -1059/-240 region exhibits a high basal transcriptional level in a pituitary-derived cell line, LbetaT2. A DNA binding and DNase I footprinting assay demonstrated that Lhx3 has seven binding sites in the -1059/+12 region of Cga, including a pituitary glycoprotein hormone basal element (PGBE) known as a LIM homeodomain factor-binding site. A transfection assay of the sequence of Lhx3-binding sites fused with minimal promoter vector confirmed their Lhx3-dependent stimulations in LbetaT2 cells. RT-PCR analysis of porcine pituitary ontogeny demonstrated that porcine Lhx3 showed striking changes of expression in both sexes during the fetal period but a stable high level of expression after birth. Thus, the porcine Cga promoter is regulated by Lhx3 through seven sites in the distal and proximal regions.

Function of HNF1 in the pathogenesis of diabetes

The importance of hepatocyte nuclear factors (HNFs), as well as other transcription factors in β-cell development and function, was underlined by the characterization of human mutations causing maturity-onset diabetes of the young (MODY). HNF1A and HNF1B mutations lead to MODY forms 3 and 5, respectively. Thus, transcriptional control is an essential mechanism underlying the precise metabolic control exerted by β-cells in regulating insulin release. The diabetes phenotype of MODY3 (HNF1α) and the phenotypes of MODY5 (HNF1β), which can also include renal disease and genitourinary malformations, as well as neonatal diabetes and pancreatic agenesis, have now been described. However, detailed molecular pathology remains elusive. The large array of dominant-negative and deletion mutations, and the lack of structure-phenotype relationships for most mutations, have not helped us to formulate a mechanistic understanding. Further molecular studies of HNF1 actions and gene regulation are anticipated to provide useful insights into β-cell biology and potential therapeutic tools.

Proneural bHLH and Brn Proteins Coregulate a Neurogenic Program through Cooperative Binding to a Conserved DNA Motif

Proneural proteins play a central role in vertebrate neurogenesis, but little is known of the genes that they regulate and of the factors that interact with proneural proteins to activate a neurogenic program. Here, we demonstrate that the proneural protein Mash1 and the POU proteins Brn1 and Brn2 interact on the promoter of the Notch ligand Delta1 and synergistically activate Delta1 transcription, a key step in neurogenesis. Overexpression experiments in vivo indicate that Brn2, like Mash1, regulates additional aspects of neurogenesis, including the division of progenitors and the differentiation and migration of neurons. We identify by in silico screening a number of additional candidate target genes, which are recognized by Mash1 and Brn proteins through a DNA-binding motif similar to that found in the Delta1 gene and present a broad range of activities. We thus propose that Mash1 synergizes with Brn factors to regulate multiple steps of neurogenesis.

Characterization of NOBOX DNA Binding Specificity and Its Regulation of Gdf9 and Pou5f1 Promoters

Nobox (newborn ovary homeobox gene) deficiency disrupts early folliculogenesis and the expression of oocyte-specific genes in mice. Here, we identified several cis-acting sites, TAATTG, TAGTTG, and TAATTA as NOBOX DNA binding elements (NBEs) using a library of randomly generated oligonucleotides by cyclic amplification of sequence target assay and mutation analyses. We show that NOBOX preferentially binds to the NOBOX binding elements with high affinity. In addition, we found that promoter regions of mouse Pou5f1 and Gdf9 contain one (-426) and three NOBOX binding elements (-786, -967, and -1259), respectively. NOBOX binds to these putative NOBOX binding elements with high affinity and augmented transcriptional activity of luciferase reporter driven by mouse Pou5f1 and Gdf9 promoters containing the NOBOX binding elements. In chromatin immunoprecipitation assays, DNA sequences from Pou5f1 and Gdf9 promoters co-precipitated with anti-NOBOX antibody. These results suggest that NOBOX directly regulates the transcription of Pou5f1 and Gdf9 in oocytes during early folliculogenesis.

Molecular analysis of a human PAX6 homeobox mutant

Pax6 controls eye, pancreas and brain morphogenesis. In humans, heterozygous PAX6 mutations cause aniridia and various other congenital eye abnormalities. Most frequent PAX6 missense mutations are located in the paired domain (PD), while very few missense mutations have been identified in the homeodomain (HD). In the present report, we describe a molecular analysis of the human PAX6 R242T missense mutation, which is located in the second helix of the HD. It was identified in a male child with partial aniridia in the left eye, presenting as a pseudo-coloboma. Gel-retardation assays revealed that the mutant HD binds DNA as well as the wild-type HD. In addition, the mutation does not modify the DNA-binding properties of the PD. Cell transfection assays indicated that the steady-state levels of the full length mutant protein are higher than those of the wild-type one. In cotransfection assays a PAX6 responsive promoter is activated to a higher extent by the mutant protein than by the wild-type protein. In vitro limited proteolysis assays indicated that the presence of the mutation reduces the sensitivity to trypsin digestion. Thus, we suggest that the R242T human phenotype could be due to abnormal increase of PAX6 protein, in keeping with the reported sensitivity of the eye phenotype to increased PAX6 dosage.

Use of altered-specificity binding Oct-4 suggests an absence of pluripotent cell-specific cofactor usage

Oct-4 is a POU domain transcription factor that is critical for maintaining pluripotency and for stem cell renewal. Previous studies suggest that transcription regulation by Oct-4 at particular enhancers requires the input of a postulated E1A-like cofactor that is specific to pluripotent cells. However, such studies have been limited to the use of enhancer elements that bind other POU-protein family members in addition to Oct-4, thus preventing a ‘clean’ assessment of any Oct-4:cofactor relationships. Other attempts to study Oct-4 functionality in a more ‘stand-alone’ situation target Oct-4 transactivation domains to DNA using heterologous binding domains, a methodology which is known to generate artificial data. To circumvent these issues, an altered-specificity binding Oct-4 (Oct-4RR) and accompanying binding site, which binds Oct-4RR only, were generated. This strategy has previously been shown to maintain Oct-1:cofactor interactions that are highly binding-site and protein/binding conformation specific. This system therefore allows a stand-alone study of Oct-4 function in pluripotent versus differentiated cells, without interference from endogenous POU factors and with minimal deviation from bound wild-type protein characteristics. Subsequently, it was demonstrated that Oct-4RR and the highly transactive regions of its N-terminus determined here, and its C-terminus, have the same transactivation profile in pluripotent and differentiated cells, thus providing strong evidence against the existence of such a pluripotent cell-specific Oct-4 cofactor.

Homeodomain revisited: a lesson from disease-causing mutations

The homeodomain is a highly conserved DNA-binding motif that is found in numerous transcription factors throughout a large variety of species from yeast to humans. These gene-specific transcription factors play critical roles in development and adult homeostasis, and therefore, any germline mutations associated with these proteins can lead to a number of congenital abnormalities. Although much has been revealed concerning the molecular architecture and the mechanism of homeodomain-DNA interactions, the study of disease-causing mutations can further provide us with instructive information as to the role of particular residues in a conserved mode of action. In this paper, I have compiled the homeodomain missense mutations found in various human diseases and re-examined the functional role of the mutational "hot spot" residues in light of the structures obtained from crystallography. These findings should be useful in understanding the essential components of the homeodomain and in attempts to design agonist or antagonists to modulate their activity and to reverse the effects caused by the mutations.

Transcriptional activity of the paired-like homeodomain proteins CHX10 and VSX1

CHX10 and VSX1 are homeodomain (HD) proteins essential for normal retinal development. CHX10 is required first for retinal progenitor cell proliferation and later for bipolar cell differentiation, whereas VSX1 is important in the terminal differentiation of a subset of bipolar cells. Elucidating the transcriptional activity of CHX10 and VSX1 is required to understand how these factors control retinal development. We show that CHX10 and Vsx1 can function as transcriptional repressors. When tethered to a promoter by a heterologous LexA DNA-binding domain or its HD, CHX10 repressed multiple classes of activators in different immortalized cell lines. CHX10 blocked TATA-containing and TATA-less promoters, repressed at a distance, and inhibited a complex enhancer positioned upstream or downstream of the reporter gene, whereas retinoblastoma protein (RB) inhibited the downstream enhancer only. Interestingly, CHX10 mildly potentiated a subset of activators in chick neuronal cultures. Thus, CHX10 is both a versatile repressor and a context-specific weak activator. The CHX10 HD and CVC domains were sufficient for DNA binding and repression. VSX1 contains closely related homeo and CVC domains and, like CHX10, also repressed transcription. A VSX1 HD mutation, R166W, that impairs DNA binding and causes keratoconus in humans, hindered repressor function. Therefore, CHX10 and VSX1 may control retinal bipolar cell specification or differentiation by repressing genes required for the development of other cell types.

Tat-dependent repression of human immunodeficiency virus type 1 long terminal repeat promoter activity by fusion of cellular transcription factors

Transcription initiation from HIV-1 long terminal repeat (LTR) promoter requires the virally encoded transactivator, Tat, and several cellular co-factors to accomplish the Tat-dependent processive transcription elongation. Individual cellular transcription activators, LBP-1b and Oct-1, on the other hand, have been shown to inhibit LTR promoter activities probably via competitive binding against TFIID to the TATA-box in LTR promoter. To explore the genetic interference strategies against the viral replication, we took advantage of the existence of the bipartite DNA binding domains and the repression domains of LBP-1b and Oct-1 factors to generate a chimeric transcription repressor. Our results indicated that the fusion protein of LBP-1b and Oct-1 exhibited higher DNA binding affinity to the viral promoter than the individual factors, and little interference with the host cell gene expression due to its anticipated rare cognate DNA sites in the host cell genome. Moreover, the chimera exerted increased Tat-dependent repression of transcription initiation at the LTR promoter both in vitro and in vivo compared to LBP-1b, Oct-1 or combination of LBP-1b and Oct-1. These results might provide the lead in generating a therapeutic reagent useful to suppress HIV-1 replication.

Characterization of Nkx3.2 DNA binding specificity and its requirement for somitic chondrogenesis

We have previously shown that Nkx3.2, a member of the NK class of homeoproteins, functions as a transcriptional repressor to promote somitic chondrogenesis. However, it has not been addressed whether Nkx3.2 can bind to DNA in a sequence-specific manner and whether DNA binding by Nkx3.2 is required for its biological activity. In this work, we employed a DNA binding site selection assay, which identified TAAGTG as a high affinity Nkx3.2 binding sequence. Sequence-specific binding of Nkx3.2 to the TAAGTG motif in vitro was confirmed by electrophoretic mobility shift assays, and mutagenesis of this sequence revealed that HRAGTG (where H represents A, C, or T, and R represents A or G) comprises the consensus DNA binding site for Nkx3.2. Consistent with these findings, the expression of a reporter gene containing reiterated Nkx3.2 binding sites was repressed in vivo by Nkx3.2 co-expression. In addition, we have generated a DNA nonbinding point mutant of Nkx3.2 (Nkx3.2-N200Q), which contains an asparagine to glutamine missense mutation in the homeodomain. Interestingly, despite being defective in DNA binding, Nkx3.2-N200Q still retains its intrinsic transcriptional repressor function. Finally, we demonstrate that unlike wild-type Nkx3.2, Nkx3.2-N200Q is unable to activate the chondrocyte differentiation program in somitic mesoderm, indicating that DNA binding by Nkx3.2 is critical for this factor to induce somitic chondrogenesis.

Missense mutations of human homeoboxes: A review

The homeodomain (encoded by the homeobox) is the DNA-binding domain of a large variety of transcriptional regulators involved in controlling cell fate decisions and development. Mutations of homeobox-containing genes cause several diseases in humans. A variety of missense mutations giving rise to human diseases have been described. These mutations are an excellent model to better understand homeodomain molecular functions. To this end, homeobox missense mutations giving rise to human diseases are reviewed. Seventy-four independent homeobox mutations have been observed in 17 different genes. In the same genes, 30 missense mutations outside the homeobox have been observed, indicating that the homeodomain is more easily affected by single amino acids changes than the rest of the protein. Most missense mutations have dominant effects. Several data indicate that dominance is mostly due to haploinsufficiency. Among proteins having the homeodomain as the only DNA-binding domain, three "hot spot" regions can be delineated: 1) at codon encoding for Arg5; 2) at codon encoding for Arg31; and 3) at codons encoding for amino acids of recognition helix. In the latter, mutations at codons encoding for Arg residues at positions 52 and 53 are prevalent. In the recognition helix, Arg residues at positions 52 and 53 establish contacts with phosphates in the DNA backbone. Missense mutations of amino acids that contribute to sequence discrimination (such as those at positions 50 and 54) are present only in a minority of cases. Similar data have been obtained when missense mutations of proteins possessing an additional DNA-binding domain have been analyzed. The only exception is observed in the POU1F1 (PIT1) homeodomain, in which Arg58 is a "hot spot" for mutations, but is not involved in DNA recognition.

E2F1-mediated transcriptional inhibition of the plasminogen activator inhibitor type 1 gene

Gene expression of the plasminogen activation system is cell-cycle dependent. Previously, we showed that ectopic expression of E2F1 repressed the plasminogen activator inhibitor type 1 (PAI-1) promoter in a manner dependent on the presence of DNA-binding and transactivation domains of E2F1 but independent of binding to pocket-binding proteins, suggesting a novel mechanism for E2F-mediated negative gene regulation [Koziczak, M., Krek, W. & Nagamine, Y. (2000) Mol. Cell. Biol. 20, 2014-2022]. However, it remains to be seen whether endogenous E2F can exert a similar effect. We report here that down-regulation of PAI-1 gene expression correlates with an increase in endogenous E2F activity. When cells were treated with a cdk2/4-specific inhibitor, which maintains E2F in an inactive state, the decline of serum-induced PAI-1 mRNA levels was suppressed. In mutant U2OS cells expressing a temperature-sensitive retinoblastoma protein (pRB), a shift to a permissive temperature induced PAI-1 mRNA expression. In U2OS cells stably expressing an E2F1-estrogen receptor chimeric protein that could be activated by tamoxifen, PAI-1 gene transcription was markedly reduced by tamoxifen even in the presence of cycloheximide. These results all indicate that endogenous E2F can directly repress the PAI-1 gene. DNase I hypersensitive-site analysis of the PAI-1 promoter suggested the involvement of conformation changes in chromatin structure of the PAI-1 promoter. 5' deletion analysis of the PAI-1 promoter showed that multiple sites were responsible for the E2F negative regulation, some of which were promoter dependent. Interestingly, one of these sites is a p53-binding element.

Optimal Oct-2 affinity for an extended DNA site and the effect of GST fusion on site preference

The regulator of immunoglobulin expression Oct-2 and the related widely expressed transcription factor Oct-1 have been shown to interact with DNA sequences containing an "octamer" motif, ATGC(A/T)AAT. To better understand Oct-2 function we have used random oligonucleotide selection and competition assays to define the optimal recognition site for this protein. The selected site contains an extended sequence that is remarkably similar to octamer-heptamer sequences found in immunoglobulin heavy-chain regulatory sequences, and the affinity of Oct-2 for this site is at least 50-fold greater than for sites containing the octamer motif alone. Fusion to glutathione S-transferase, a widely used model for protein-DNA and protein-protein interaction, does not alter the optimal Oct-2 recognition site, but inhibits Oct-2 POU-domain dimerization, slows the dissociation rate of the GST-Oct-2/DNA complex, and increases the relative importance of the heptamer domain for Oct-2 binding. These data advance our ability to identify in vivo targets of POU-factor regulation and also suggest that GST-fusion proteins should be used with caution in DNA-binding studies.

Geometric analysis and comparison of protein-DNA interfaces: why is there no simple code for recognition?

Structural studies of protein-DNA complexes have shown that there are many distinct families of DNA-binding proteins, and have shown that there is no simple "code" describing side-chain/base interactions. However, systematic analysis and comparison of protein-DNA complexes has been complicated by the diversity of observed contacts, the sheer number of complexes currently available and the absence of any consistent method of comparison that retains detailed structural information about the protein-DNA interface. To address these problems, we have developed geometric methods for characterizing the local structural environment in which particular side-chain/base interactions are observed. In particular, we develop methods for analyzing and comparing spatial relationships at the protein-DNA interface. Our method involves attaching local coordinate systems to the DNA bases and to the C(alpha) atoms of the peptide backbone (these are relatively rigid structural units). We use these tools to consider how the position and orientation of the polypeptide backbone (with respect to the DNA) helps to determine what contacts are possible at any given position in a protein-DNA complex. Here, we focus on base contacts that are made in the major groove, and we use spatial relationships in analyzing: (i) the observed patterns of side-chain/base interactions; (ii) observed helix docking orientations; (iii) family/subfamily relationships among DNA-binding proteins; and (iv) broader questions about evolution, altered specificity mutants and the limits for the design of new DNA-binding proteins. Our analysis, which highlights differences in spatial relationships in different complexes and at different positions in a complex, helps explain why there is no simple, general code for protein-DNA recognition.

Transcriptional regulation of the murine 3' IgH enhancer by OCT-2

Targeted disruption of either of the B cell-specific transcription factors Oct-2 or OCA-B/BOB-1/OBF-1 dramatically affects B cell terminal differentiation. The 3' enhancer of immunoglobulin heavy chain (IgH) locus is important for transcription of the locus in terminal plasma cells. Allele-specific suppression of mutant Oct-2 binding sites in this enhancer by a variant Oct-2 protein revealed that in a mature B cell line this enhancer was specifically dependent upon Oct-2, as contrasted to the closely related Oct-1 transcription factor. Phosphorylation of the Oct-2 protein was important for this activation and was synergistic for coactivation by the OCA-B factor. These results indicate that Oct-2 and OCA-B interact with the 3' enhancer in regulation of the IgH locus during B cell activation.

Determinants of the DNA-binding specificity of the Avian homeodomain protein, AKR

AKR (Avian Knotted-Related) was the first example of a vertebrate homeodomain protein with a highly divergent Ile residue at position 50 of the DNA-recognition helix. The protein was cloned from a liver cDNA expression library of a day-9 chick embryo by virtue of its ability to bind to the F' site in the proximal promoter of the avian apoVLDLII gene. Expression of the apoVLDLII gene is completely estrogen dependent, and mutation or deletion of the F' site decreases estrogen inducibility 5- to 10-fold. Subsequent data indicated that AKR is capable of repressing the hormone responsiveness of the apoVLDLII promoter, specifically through binding to F'. Involvement of the F' site in the hormone-dependent activation of apoVLDLII gene expression, as well as AKR-mediated repression, strongly suggests that both positive and negative regulatory factors interact with this site. Although several mammalian proteins have now been isolated whose homeodomains share many of the structural features of AKR, including the Ile at position 50, little is known of their functions in vivo or the identities of the genes they regulate. Consequently, the elements through which they exert their effects and the structural determinants of their binding specificities remain largely uncharacterized. In this study, we defined the sequence specificity of binding by AKR using polymerase chain reaction-assisted optimal site selection and determined the affinity with which the protein binds to both the optimized site and the F' site. Additionally, we generated a three-dimensional model of the AKR homeodomain binding to its optimized site and probed the validity of the model by examining the consequences of mutating amino acid residues in recognition helix 3 and the N-terminal arm on the binding specificity of the homeodomain. Finally, we present evidence that the F' site itself may act as an estrogen response element (ERE) when in the vicinity of imperfect or canonical EREs and that AKR can repress hormone inducibility mediated via this site.

Highly cooperative homodimerization is a conserved property of neural POU proteins

POU-domain proteins have been shown to play important roles in the development of the nervous, endocrine, and immune systems. However, the distinctive DNA recognition properties of the six major POU subclasses have not been well defined. Here, we have used random oligonucleotide selection and competitive binding assays to determine the optimal DNA recognition elements for the POU-III and POU-VI protein classes, represented by Brn-2 and Brn-5, respectively. The optimal Brn-5 consensus binding sequence GCATAA(T/A)TTAT strongly resembles that previously determined for the POU-IV (Brn-3) class, whereas Brn-2 exhibits highest affinity for non-octamer sites of the form ATG(A/C)AT(A/T)0-2ATTNAT and for octamer sites that contain a full associated heptamer sequence. Brn-2, Brn-3.0, and their invertebrate homologues all exhibit highly cooperative homodimerization on the Brn-2 consensus sequence, demonstrating that cooperative dimerization is a general property of these neural POU proteins. However, modified sites to which Brn-2 binds only as a monomer mediate the transcriptional effects of Brn-2 better than the consensus sequence, demonstrating that dimerization on these sites diminishes the transactivation ability of the protein. Together with the findings of our prior studies these data greatly facilitate the identification of functional POU recognition elements in the regulatory regions of neural genes.

Engrailed homeodomain-DNA complex at 2.2 A resolution: a detailed view of the interface and comparison with other engrailed structures

We report the 2.2 A resolution structure of the Drosophila engrailed homeodomain bound to its optimal DNA site. The original 2.8 A resolution structure of this complex provided the first detailed three-dimensional view of how homeodomains recognize DNA, and has served as the basis for biochemical studies, structural studies and molecular modeling. Our refined structure confirms the principal conclusions of the original structure, but provides important new details about the recognition interface. Biochemical and NMR studies of other homeodomains had led to the notion that Gln50 was an especially important determinant of specificity. However, our refined structure shows that this side-chain makes no direct hydrogen bonds to the DNA. The structure does reveal an extensive network of ordered water molecules which mediate contacts to several bases and phosphates (including contacts from Gln50), and our model provides a basis for detailed comparison with the structure of an engrailed Q50K altered-specificity variant. Comparing our structure with the crystal structure of the free protein confirms that the N and C termini of the homeodomain become ordered upon DNA-binding. However, we also find that several key DNA contact residues in the recognition helix have the same conformation in the free and bound protein, and that several water molecules also are "preorganized" to contact the DNA. Our structure helps provide a more complete basis for the detailed analysis of homeodomain-DNA interactions.

Comparison of X-ray and NMR structures for the Antennapedia homeodomain-DNA complex

Homeodomains are one of the key families of eukaryotic DNA-binding motifs and provide an important model system for studying protein-DNA interactions. We have crystallized the Antennapedia homeodomain-DNA complex and solved this structure at 2.4 A resolution. NMR and molecular dynamics studies had implied that this homeodomain achieves specificity through an ensemble of rapidly fluctuating DNA contacts. The crystal structure is in agreement with the underlying NMR data, but our structure reveals a well-defined set of contacts and also reveals the locations and roles of water molecules at the protein-DNA interface. The synthesis of X-ray and NMR studies provides a unified, general model for homeodomain-DNA interactions.

Using altered specificity Oct-1 and Oct-2 mutants to analyze the regulation of immunoglobulin gene transcription

Oct-1 and Oct-2 represent the prototypical example of related transcription factors with identical DNA recognition properties. They bind functionally critical octamer elements found in diverse regulatory sequences. It has not been possible to determine directly if these factors are functionally redundant or selective when interacting with different regulatory sequences in the appropriate cell type. An equivalent pair of altered DNA-binding specificity mutants of Oct-1 and Oct-2 are used to examine their function from varied regulatory contexts in B cells. These factors function as redundant activators of immunoglobulin (Ig) gene promoters (Vkappa and VH) and a histone H2B promoter. The structural basis of redundancy resides in the highly conserved DNA-binding POU domain, because this domain of either protein can activate transcription from both Ig and H2B promoters. We find that the nature of a distal enhancer dictates the relative potency of Oct-1 versus Oct-2 bound to a promoter. Oct-1 preferentially stimulates transcription from a VH or Vkappa promoter in combination with enhancers from the IgH or Igkappa locus, respectively. In this context, the more potent action of Oct-1 is dependent on a region external to the POU domain. These results suggest that Oct-1 may be the critical regulator of Ig gene transcription during B cell development and provide an explanation for selective Ig isotype expression defects in Oct-2 and OCA-B null mice.

Engrailed (Gln50-->Lys) homeodomain-DNA complex at 1.9 A resolution: structural basis for enhanced affinity and altered specificity

BACKGROUND

The homeodomain is one of the key DNA-binding motifs used in eukaryotic gene regulation, and homeodomain proteins play critical roles in development. The residue at position 50 of many homeodomains appears to determine the differential DNA-binding specificity, helping to distinguish among binding sites of the form TAATNN. However, the precise role(s) of residue 50 in the differential recognition of alternative sites has not been clear. None of the previously determined structures of homeodomain-DNA complexes has shown evidence for a stable hydrogen bond between residue 50 and a base, and there has been much discussion, based in part on NMR studies, about the potential importance of water-mediated contacts. This study was initiated to help clarify some of these issues.

RESULTS

The crystal structure of a complex containing the engrailed Gln50-->Lys variant (QK50) with its optimal binding site TAATCC (versus TAATTA for the wild-type protein) has been determined at 1.9 A resolution. The overall structure of the QK50 variant is very similar to that of the wild-type complex, but the sidechain of Lys50 projects directly into the major groove and makes several hydrogen bonds to the O6 and N7 atoms of the guanines at base pairs 5 and 6. Lys50 also makes an additional water-mediated contact with the guanine at base pair 5 and has an alternative conformation that allows a hydrogen bond with the O4 of the thymine at base pair 4.

CONCLUSIONS

The structural context provided by the folding and docking of the engrailed homeodomain allows Lys50 to make remarkably favorable contacts with the guanines at base pairs 5 and 6 of the binding site. Although many different residues occur at position 50 in different homeodomains, and although numerous position 50 variants have been constructed, the most striking examples of altered specificity usually involve introducing or removing a lysine sidechain from position 50. This high-resolution structure also confirms the critical role of Asn51 in homeodomain-DNA recognition and further clarifies the roles of water molecules near residues 50 and 51.

Conservative Val47 residue of POU homeodomain: role in DNA recognition

Conservative Val47 residue, located in the third recognition helix of the Oct-2 POU domain, was alternately substituted with other 19 amino acids. Affinity and specificity of interaction with oct-site ATGCAAANGA and homeo-specific site ATAANGA were determined for all mutants. The wild type protein (with Val47) has maximal affinity and specificity in POU domain interaction with octamer sequence. However, V47I mutant showed stronger interaction with homeo-specific site. The highest specificity of interaction with homeo-site was recorded for V47S mutant. We conclude that only Val47 provides sequence-specific high-affinity binding of POU proteins with octamer targets other than the homeo-specific site. It is shown also that damages caused by point mutations may be at least partially compensated by participation in the oct-site recognition of both POUh and POUs domains.

Cysteine 50 of the POU H domain determines the range of targets recognized by POU proteins

The best target of POU proteins (Oct-1, Oct-2) is an octamer sequence ATGCAAAT. POU proteins also recognize, with weaker affinity, the TAAT-like targets of another group of regulatory factors, the homeoproteins. Up to now, it has not been known why Cys50 of the POUHdomain is absolutely conserved in contrast to that in homeoproteins. To assess the importance of Cys50 in determining the binding specificity of POU proteins, all possible amino acids were substituted for Cys at position 50, and the resulting mutants were tested with probes containing octamer (ATGCAAATNN) or homeospecific binding sites. Only the wild-type POU was shown to adequately discriminate between the octamer and homeospecific sites, and the protein affinity was only slightly affected by the nucleotide sequence flanking the octamer at the 3'-end. Any amino acid substitution at position 50 resulted in the mutant protein binding efficiently both to the octamer and the TAAT-like sequences. Moreover, in this case the 3'-flanking sequences influenced the binding to a much greater extent.

Electrostatics and hydration at the homeodomain-DNA interface: chemical probes of an interfacial water cavity

Electrostatics and hydration of a homeodomain-DNA complex are dissected by chemical modification. Selective neutralization of phosphate charges by methylphosphonate substitution demonstrates the differential importance of short- and long-range electrostatic interactions. Whereas the footprint of direct contacts is in accord with crystal structures, interference is also observed at non-contacted sites. Such sites adjoin a novel interfacial water cavity in the major groove. Non-contacted phosphodiester groups in the cavity are proposed to contribute to long-range ordering of an extended protein-water-DNA interface. Use of isolated S(p) and R(p) methylphosphonate diastereomers demonstrates that interference at this extended interface is stereoselective and charge-independent. Attenuation of protein binding presumably reflects groove-specific reorganization of bound water. Surprisingly, such attenuation can exceed that due to neutralization of a direct phosphate-side-chain salt bridge. These results support the hypothesis that hydration of an interfacial cavity functions as a non-covalent extension of the DNA surface. Stereo-specific interrogation of bound water by chemical synthesis provides a general method to assess the coupling between solvation and DNA recognition.

Physical basis of a protein-DNA recognition code

Can a stereochemical recognition code explain sequence-specific protein-nucleic acid interactions? Whereas a code that is generally applicable to DNA-binding proteins of all known structural families is unattainable, the indications are that a code can describe at least some of the interactions of classical zinc fingers with DNA. The crystal structures of related zinc finger-DNA complexes reveal a remarkable mode of interaction that sets the framework for this code, and recent biochemical studies have elucidated the intermolecular contacts (contingent on this framework) that result in specificity.

In the TTF-1 homeodomain the contribution of several amino acids to DNA recognition depends on the bound sequence

The thyroid transcription factor-1 homeodomain (TTF-1HD) shows a peculiar DNA binding specificity, preferentially recognizing sequences containing the 5'-CAAG-3' core motif. Most other homeodomains instead recognize sites containing the 5'-TAAT-3' core motif. Here, we show that TTF-1HD efficiently recognizes another sequence, called D1, devoid of the 5'-CAAG-3' core motif. Different experimental approaches indicate that TTF-1HD contacts the D1 sequence in a manner which is different to that used to interact with sequences containing the 5'-CAAG-3' core motif. The binding activities that mutants of TTF-1HD display with the D1 sequence or with the sequence containing the 5'-CAAG-3' core motif indicate that the role of several DNA-contacting amino acids is different. In particular, during recognition of the D1 sequence, backbone-interacting amino acids not relevant in binding to sequences containing the 5'-CAAG-3' core motif play an important role. In the TTF-1HD, therefore, the contribution of several amino acids to DNA recognition depends on the bound sequence. These data indicate that although a common bonding network exists in all of the HD/DNA complexes, peculiarities important for DNA recognition may occur in single cases.

Conservation and diversification in homeodomain-DNA interactions: a comparative genetic analysis

Nearly all metazoan homeodomains (HDs) possess DNA binding targets that are related by the presence of a TAAT sequence. We use an in vitro genetic DNA binding site selection assay to refine our understanding of the amino acid determinants for the recognition of the TAAT site. Superimposed upon the conserved ability of metazoan HDs to recognize a TAAT core is a difference in their preference for the bases that lie immediately 3' to it. Amino acid position 50 of the HD has been shown to discriminate among these base pairs, and structural studies have suggested that water-mediated hydrogen bonds and van der Waals contacts underlie for this ability. Here, we show that each of six amino acids tested at position 50 can confer a distinct DNA binding specificity.

A novel homeobox protein which recognizes a TGT core and functionally interferes with a retinoid-responsive motif

We describe here a novel homeobox gene, denoted TGIF (5"TG3' interacting factor), which belongs to an expanding TALE (three amino acid loop extension) superclass of atypical homeodomains. The TGIF homeodomain binds to a previously characterized retinoid X receptor (RXR) responsive element from the cellular retinol-binding protein II promoter (CRBPII-RXRE), which contains an unusual DNA target for a homeobox. The interactions of both the homeprotein TGIF and receptor RXR alpha with the CREBPII-RXRE DNA motif occur on overlapping areas and generate a mutually exclusive binding in vitro. Transient cellular transfections demonstrate that TGIF inhibits the 9-cis-retinoic acid-dependent RXR alpha transcription activation of the retinoic acid responsive element. TGIF transcripts were detected in a restricted number of tissues. The canonical binding site of TGIF is conserved and is an integral part of several responsive elements which are organized like the CRBPII-RXRE. Hence, a novel auxiliary factor to the steroid receptor superfamily may participate in the transmission of nuclear signals during development and in the adult, as illustrated by the down-modulation of the RXR alpha activities.

Drosophila engrailed-1,10-phenanthroline chimeras as probes of homeodomain-DNA complexes

We have converted the Drosophila engrailed homeodomain into a sequence-specific nuclease by linking the protein to the chemical nuclease 1,10-phenanthroline-copper (OP-Cu). Unique cysteines were introduced at six positions into the homeodomain by site-directed mutagenesis for the covalent attachment of OP-Cu. The varied DNA-binding affinity and specificity of these mutants and the DNA cleavage pattern of their OP-Cu derivatives allowed us to assess the crystal structure of the engrailed homeodomain-DNA complex. We have also achieved site-specific double-stranded DNA scission with one of the homeodomain mutants, E28C, which has the potential of being used to identify engrailed binding sites in the genome. Because the homeodomain is so well conserved among members of the homeodomain-containing protein family, other homeodomain proteins can be converted into nucleases by attaching OP-Cu at position 28 of their homeodomains.

Analysis of homeodomain function by structure-based design of a transcription factor

The homeodomain is a 60-amino acid module which mediates critical protein-DNA and protein-protein interactions for a large family of regulatory proteins. We have used structure-based design to analyze the ability of the Oct-1 homeodomain to nucleate an enhancer complex. The Oct-1 protein regulates herpes simplex virus (HSV) gene expression by participating in the formation of a multiprotein complex (C1 complex) which regulates alpha (immediate early) genes. We recently described the design of ZFHD1, a chimeric transcription factor containing zinc fingers 1 and 2 of Zif268, a four-residue linker, and the Oct-1 homeodomain. In the presence of alpha-transinduction factor and C1 factor, ZFHD1 efficiently nucleates formation of the C1 complex in vitro and specifically activates gene expression in vivo. The sequence specificity of ZFHD1 recruits C1 complex formation to an enhancer element which is not efficiently recognized by Oct-1. ZFHD1 function depends on the recognition of the Oct-1 homeodomain surface. These results prove that the Oct-1 homeodomain mediates all the protein-protein interactions that are required to efficiently recruit alpha-transinduction factor and C1 factor into a C1 complex. The structure-based design of transcription factors should provide valuable tools for dissecting the interactions of DNA-bound domains in other regulatory circuits.

Structure-based design of transcription factors

Computer modeling suggested that transcription factors with novel sequence specificities could be designed by combining known DNA binding domains. This structure-based strategy was tested by construction of a fusion protein, ZFHD1, that contained zinc fingers 1 and 2 from Zif268, a short polypeptide linker, and the homeodomain from Oct-1. The fusion protein bound optimally to a sequence containing adjacent homeodomain (TAATTA) and zinc finger (NGGGNG) subsites. When fused to an activation domain, ZFHD1 regulated promoter activity in vivo in a sequence-specific manner. Analysis of known protein-DNA complexes suggests that many other DNA binding proteins could be designed in a similar fashion.