InteractiveFly: GeneBrief

Deformed epidermal autoregulatory factor-1: Biological Overview | References

Gene name - Deformed epidermal autoregulatory factor-1

Synonyms -

Cytological map position - 76C6-76D1

Function - transcription factor

Keywords - regulation of Deformed, Immune response, embryonic development

Symbol - Deaf1

FlyBase ID: FBgn0013799

Genetic map position - 3L: 19,811,280..19,822,623 [+]

Classification - SAND domain and zf-MYND

Cellular location - nuclear

NCBI links: Precomputed BLAST | EntrezGene

Immunity genes are activated in the Drosophila fat body by Rel and GATA transcription factors. Evidence that an additional regulatory factor, Deformed epidermal autoregulatory factor-1 (DEAF-1), also contributes to the immune response and is specifically important for the induction of two genes encoding antimicrobial peptides, Metchnikowin (Mtk) and Drosomycin (Drs). The systematic mutagenesis of a minimal Mtk 5' enhancer identified a sequence motif essential for both a response to LPS preparations in S2 cells and activation in the larval fat body in response to bacterial infection. Using affinity chromatography coupled to multidimensional protein identification technology (MudPIT), DEAF-1 was identified as a candidate regulator. DEAF-1 activates the expression of Mtk and Drs promoter-luciferase fusion genes in S2 cells. SELEX assays and footprinting data indicate that DEAF-1 binds to and activates Mtk and Drs regulatory DNAs via a TTCGGBT motif. The insertion of this motif into the Diptericin (Dpt) regulatory region confers DEAF-1 responsiveness to this normally DEAF-1-independent enhancer. The coexpression of DEAF-1 with Dorsal, Dif, and Relish results in the synergistic activation of transcription. It is proposed that DEAF-1 is a regulator of Drosophila immunity (Reed, 2008).

Transcriptional regulation of Drosophila antimicrobial genes depends on Rel and GATA transcription factors. Many immunity genes contain tightly linked Rel- and GATA-binding sites in promoter-proximal regions. GATA sites are important for establishing responses in distinct tissues such as the fat body and midgut. Serpent (dGATAb) is thought to regulate antimicrobial gene expression in the fat body, whereas dGATAe activates such genes in the midgut in response to ingested microbes. In contrast, Dorsal, Dif, and Relish, the NF-kappaB homologues in flies, shuttle between the cytoplasmic and nuclear compartments, acting as 'on/off switches' for induction). Additional factors, such as HOX and POU domain proteins, bind to distal enhancer elements and maintain constitutive domains of gene activity. A regulatory element (R1) also has been described within the CecA1 enhancer, although the factor that interacts with this motif is unknown (Reed, 2008).

Deformed epidermal autoregulatory factor-1 (DEAF-1) is a transcription factor that was originally shown to bind the autoregulatory enhancer of the Deformed (Dfd) Hox gene, which is activated in embryonic head segments of Drosophila (Gross, 1996). DEAF-1 recognizes several TTCG motifs within the portion of the Dfd autoregulatory region termed 'module E.' In addition, DEAF-1 binds several similar motifs within a Dfd response element (DRE) from the 1.28 gene that enhances maxillary gene expression during embryogenesis (Pederson, 2000). The DEAF-1 binding elements identified in these studies are reportedly not required for enhancer activity however (Reed, 2008).

The 576-aa DEAF-1 protein possesses two conserved domains, SAND and MYND. The 113-aa SAND domain (named for SP100, AIRE-1, NucP41/75, and DEAF-1) (Gibson, 1998) is responsible for DNA binding via a highly conserved KDWK peptide motif (Christensen, 1999; Surdo, 2003; Bottomley, 2001). The 32-aa MYND domain (for myeloid, Nervy, and Deaf-1) contains non-DNA-binding zinc fingers that are thought to mediate protein-protein interactions (Gross, 1996). DEAF-1 is maternally expressed, and the encoded protein is broadly distributed throughout the early embryo. It exhibits augmented expression in the CNS after stage 14 (Gross, 1996). Zygotic mutants develop to pupal stages, but do not eclose, whereas maternal mutants display severe defects in early embryonic patterning (Veraksa, 2002). Overexpression of DEAF-1 by using a maternal driver inhibits germ-band retraction and causes defects in dorsal closure, whereas overexpression at later stages causes cell death (Reed, 2008).

In vertebrates, the closest relatives of DEAF-1 are nuclear DEAF-1-related factor (NUDR) and Suppressin (SPN) (Huggenvik, 1998; LeBoeuf, 1998). Both factors are expressed in a wide variety of tissue types. NUDR functions to either activate (Michelson, 1999) transcription depending on its context, and it binds sequences bearing TTCGGG or TTTCCG motifs (Huggenvik, 1998). SPN does not have a characterized role in transcription. It was originally identified as a protein secreted by the bovine pituitary gland that, when added to tissue culture media, inhibits splenocyte proliferation (LeBoeuf, 1990) and stimulates IFN-gamma production (Carr, 1990) in leukocytes (Reed, 2008).

Studies in Drosophila have identified a 208-bp proximal enhancer that regulates the expression of the Mtk gene. This enhancer directs high levels of transcription in the fat bodies of infected larvae and also is induced by LPS preparations in S2 cells. These regulatory activities depend on a cluster of Rel- and GATA-binding sites. This study presents evidence that an additional sequence motif (E8) contributes to Mtk activation. Enhancer DNA affinity chromatography assays and proteomic analysis identified DEAF-1 as a protein that interacts with the E8 motif. DEAF-1 binds to the consensus sequence TTCGGBT, which is contained within the E8 region of the Mtk enhancer. Additional DEAF-1 consensus motifs are found in the regulatory regions of other immunity genes, such as Drosomycin (Drs). Evidence is presented that DEAF-1 works synergistically with Dorsal, Dif, and Relish to induce gene expression in response to LPS. It is proposed that DEAF-1 is an essential component of the immune response in Drosophila (Reed, 2008).

To identify regulatory motifs within the minimal 208-bp Mtk enhancer, the nucleotide sequences of 11 regions (E1-E11) flanking the previously identified Rel- and GATA-binding sites were scrambled. Several of these scrambled elements (SEs) were found to alter the activities of Mtk-Luciferase (Mtk-Luc) promoter-reporter fusion genes in transient transfection assays with S2 cells in the presence of LPS. It should be emphasized that this assay does not measure a response to LPS, but rather to Gram-negative peptidoglycans that commonly occur in commercial preparations of LPS. Gram-negative peptidoglycans have been shown to signal through the Imd and, to a lesser degree, the Toll pathway, so this assay probably reflects both types of signaling cascades. Fusion genes bearing scrambled sequences in region 4, 5, or 6 are roughly twice as active as the native enhancer, suggesting a disruption of potential repressor elements. A 50% reduction in induction is seen for the fusion gene containing scrambled sequences in region 9, suggesting the loss of a weak activator element. Most notably, there is a severe 21-fold decrease in the induced expression of the Mtk-Luc fusion gene containing scrambled sequences in region 8 (SE8). There also is a 3-fold reduction in the constitutive activities of this fusion gene. Thus, region 8 appears to contain an essential activator element (Reed, 2008).

To test the activities of the E8 sequence in vivo, transgenic larvae carrying an Mtk SE8-LacZ transgene were examined. Upon septic injury with a mixture of Escherichia coli and Micrococcus luteus, the wild-type Mtk-lacZ transgene drives intense lacZ expression in the fat body. Mutation of E8 essentially abolishes reporter gene expression in three of four independent lines and allows only a weak response in the fourth. The wild-type Mtk-lacZ fusion gene is constitutively active in the posterior proventriculus of most larvae and in the anterior midgut of ?20% of the larvae. Upon ingestion of Erwinia carotovora, there is at least a doubling in the number of larvae that exhibit expression in the anterior midgut. In contrast, the Mtk SE8-LacZ transgene completely lacks both constitutive and induced activity throughout the midgut in three of four transgenic lines, with a weak response in the fourth line (Reed, 2008).

A proteomics approach was used to identify proteins that bind E8. The entire Mtk regulatory domain (WT and SE8) was biotinylated and coupled to magnetic Dynal beads. An EcoRI restriction site was included proximal to the biotin moiety. Dynal bead-DNA complexes were incubated with nuclear extracts from S2 cells that had been treated with LPS. The resulting nucleoprotein complexes were washed extensively and eluted with a brief EcoRI digest. Eluted proteins were then subjected to Multidimensional Protein Identification Technology (MudPIT) analysis (Reed, 2008).

MudPIT identified several candidate proteins that were uniquely present in the eluate from the native Mtk regulatory sequence and not from the SE8 mutant enhancer sequence. One of these, DEAF-1, was particularly interesting because it recognizes a sequence motif, TTCG (Gross, 1996; Pederson, 2000), which resembles the E8 sequence (TCATTCGGC). This led to a pursuit of the role of DEAF-1 in regulating Mtk expression (Reed, 2008).

To test whether DEAF-1 recognizes Mtk regulatory sequences, gel shift assays were performed with increasing amounts of recombinant DEAF-1 protein and radiolabeled E8 and SE8 oligonucleotides. The DEAF-1 protein binds to E8, but not the SE8 scrambled sequence. Competition assays were done by incubating DEAF-1 with the radiolabeled E8 sequence, followed by the addition of an excess of unlabeled E8 or SE8 oligonucleotides. A 10-fold excess of unlabeled E8 removes DEAF-1 from the radioactive probe, whereas the same amount of the SE8 oligonucleotide only weakly disrupts binding (Reed, 2008).

The previously published DEAF-1-binding site, TTCG, is based on footprint assays using the Dfd autoregulatory enhancer and the 1.28 gene enhancer. This analysis was extended by performing systematic evolution of ligands by exponential (SELEX) enrichment experiments. Recombinant His-tagged DEAF-1 protein was incubated with a random library of radiolabeled oligonucleotides. Protein-DNA complexes were gel-purified and PCR-amplified, and the selected DNA was subjected to two additional rounds of binding and amplification. The DNAs were sequenced and aligned with the published footprint data to generate a position-weighted matrix. The broadest consensus sequence using this approach is TTCGGBT. The SELEX data show a weaker preference for cytosine at position 3 than the previous footprinting data and a stronger selection for guanine at position 5. This finding may reflect differences in the binding of DEAF-1 to individual sites, compared with the clustered sites seen in the Dfd enhancer (Reed, 2008).

The strongest selection by DEAF-1 occurs at positions 2-5 (TCGG). The importance of each position was verified by performing gel shift competition assays. DEAF-1-E8 complexes were incubated with unlabeled oligonucleotides bearing mutations at each position along the consensus binding sequence. Oligonucleotides bearing mutations at position 1, 6, or 7 (TTCGGBT) successfully competed with radiolabeled E8 for DEAF-1 binding, suggesting that they contain an intact core-binding sequence. In contrast, mutations at positions 2-5 (TTCGGBT) greatly impaired competitive binding of the modified oligonucleotides. Hence, strong DEAF-1-binding sites appear to contain a TCGG core sequence (Reed, 2008).

Transient transfection assays were done to investigate the ability of DEAF-1 to activate transcription in S2 cells. DEAF-1 was expressed in S2 cells by placing the DEAF-1 coding sequence under the control of the actin promoter (pMA6-DEAF-1). This DEAF-1 expression vector was cotransfected with an Mtk-Luc reporter construct in S2 cells. The fusion gene is normally induced 12-fold upon addition of LPS to the culture medium. The addition of pMA6-DEAF-1 causes a 5-fold increase in the basal expression of the Mtk-Luc reporter gene and a 28-fold increase upon addition of LPS. In contrast, an Mtk-Luc reporter construct bearing the scrambled E8 sequence (Mtk SE8-Luc) did not respond to expression of DEAF-1 (Reed, 2008).

Other immunity genes were surveyed for sequences that conformed to the DEAF-1-binding consensus. The 746-bp 5' enhancer of Drs contains five potential DEAF-1-binding sites (E8.1-E8.5). Each site binds DEAF-1 with a different affinity; sites E8.3 and E8.4 bind particularly well. The five recognition sequences were scrambled in the context of an otherwise normal Drs enhancer and analyzed in S2 cells by using a Drs-Luc reporter construct. A Drs-Luc reporter containing wild-type sequences is induced ~2-fold upon LPS addition. Cotransfection with the pMA6-DEAF-1 expression vector causes an additional 12-fold increase in Drs-Luc reporter gene expression in the absence of LPS and a 17-fold increase with LPS. Mutations in individual binding sites reduce both basal and activated transcription. Mutation of site 8.3 causes the most dramatic reduction in activity (Reed, 2008).

The 201-bp Diptericin enhancer mediates strong expression in the fat bodies of infected larvae, but is only weakly induced in S2 cells (~2-fold). The enhancer likely lacks DEAF-1-binding sites based on DNA sequence analysis, and a Dpt-Luc reporter construct responds only weakly to transfected DEAF-1 in S2 cells. To determine whether insertion of a DEAF-1 consensus binding site can confer an ability to respond to DEAF-1 expression, a single, optimal DEAF-1 site was placed 10 bp upstream of the endogenous GATA motif, either in the forward or reverse orientation. Both constructs exhibit a significant increase in luciferase expression when coexpressed with a DEAF-1 expression construct, compared with the unmodified reporter fusion gene. In contrast, insertion of the scrambled E8 (SE8) sequence led to only a modest increase in reporter gene expression (Reed, 2008).

Mutation of all three Rel sites or all three GATA sites within the Mtk enhancer causes a complete loss of induced activity in S2 cells. Thus, Rel and GATA factors function synergistically to activate immunity gene expression. The presence of DEAF-1-binding sites near the Rel and GATA sites suggests that it may cooperate with these factors during the mounting of an immune response. To test this, pMA6-DEAF-1 were cotransfected with expression vectors for Dorsal, Dif, and Relish. Both Mtk-Luc and Drs-Luc fusion genes were used as reporters to monitor the combinatorial activities of these transcription factors (Reed, 2008).

The Mtk-Luc reporter gene is induced 12-fold with LPS. Addition of pMA6-DEAF-1 raised the basal activity 5-fold and boosts the response to added LPS ~30-fold. Separate transfections with individual Dorsal, Dif, and Relish expression vectors also result in substantial activation (up to 67-fold for Dorsal, 166-fold for Dif, and 350-fold for Relish with LPS). Cotransfection of pMA6-DEAF-1 with each Rel factor results in some degree of synergy. The level of synergy is calculated by dividing the activity of two factors working together by the sum of their individual activities. The average fold synergy with DEAF-1 on the Mtk-Luc reporter is 1.6-fold with Dorsal, 2.2-fold with Dif, and 1.7-fold with Relish (Reed, 2008).

Similar results were obtained by using the Drs-Luc reporter, which is induced an average of 2.8-fold upon addition of LPS in S2 cells. The addition of pMA6-DEAF-1 raises basal activity 6-fold and LPS induction 10-fold. Separate transfections of either Dorsal or Dif dramatically activate this reporter (up to 60-fold with LPS), whereas Relish appears to be a much weaker activator (7-fold with LPS). Cotransfection of pMA6-DEAF-1 with each Rel factor results in an average of 3-fold synergy with Dif and 1.9-fold synergy with Dorsal. Only additive effects (average 1.2-fold synergy) were obtained with Relish. Altogether, these results suggest that DEAF-1 differentially augments the activities of different Rel factors during the induction of immunity genes. Particularly strong synergy is seen between DEAF-1 and Dif (Reed, 2008).

In summary, DEAF-1 is an essential component of the immune response in both Drosophila larvae and S2 cells. It appears to augment the synergistic activities of Rel and GATA transcription factors during the immune response. In so doing, it provides a signal amplification mechanism so that certain innate immunity genes, such as Mtk, can be transcribed at particularly high levels while using the same signaling pathways as other less highly expressed immunity genes. The critical E8 motif, TTCGGCT, is highly conserved among the Mtk enhancers of divergent Drosophilids and is closely linked to a paired set of Rel and GATA sites. It is therefore conceivable that DEAF-1 facilitates the binding or transcriptional efficacy of Rel and GATA factors at linked sites (Reed, 2008).

The requirement for DEAF-1 in the regulation of Drs, but not Dpt, hints at a possible function for DEAF-1 in Toll signaling. Cotransfection experiments using Mtk-Luc and Drs-Luc reporter genes demonstrate that DEAF-1 synergizes with Dorsal and especially Dif, two effectors of Toll signaling. Only weak cooperation occurs between DEAF-1 and Relish, a target of the Imd pathway. Microarray studies of flies mutant for Toll and Imd pathway components have comprehensively identified groups of genes coregulated by each pathway. Interestingly, several genes that require Toll signaling for regulation, such as Cactus, IM1, and Dif, contain one perfect and several near-perfect consensus DEAF-1-binding sites within 1 kb of the transcription start sites. Future studies should determine whether DEAF-1 is a constitutive component of immune tissues like the GATA factors or is regulated in response to infection by Toll signaling as seen for the Rel factors (Reed, 2008).

DEAF-1 function is essential for the early embryonic development of Drosophila

The Drosophila protein DEAF-1 is a sequence-specific DNA binding protein that was isolated as a putative cofactor of the Hox protein Deformed (Dfd). This study analyzed the effects of loss or gain of DEAF-1 function on Drosophila development. Maternal/zygotic mutations of DEAF-1 largely result in early embryonic arrest prior to the expression of zygotic segmentation genes, although a few embryos develop into larvae with segmentation defects of variable severity. Overexpression of DEAF-1 protein in embryos can induce defects in migration/closure of the dorsal epidermis, and overexpression in adult primordia can strongly disrupt the development of eye or wing. The DEAF-1 protein associates with many discrete sites on polytene chromosomes, suggesting that DEAF-1 is a rather general regulator of gene expression (Veraksa, 2002).

The DEAF-1 protein was identified as protein that bound to DNA sites containing TTCG nucleotides in an autoregulatory enhancer that is transcriptionally activated by the Drosophila Hox protein Deformed (Dfd). Initially, DEAF-1 was proposed to be a cofactor that assisted in the maxillary-specific activation of this small Dfd response element (Gross, 1996). However, further mutagenesis of the module E Dfd response element, as well as a Dfd response element from the Drosophila 1.28 gene, has indicated that the TTCG DEAF-1 binding motifs are not required for the activation of either element (Li, 1999; Pederson, 2000; Veraksa, 2002 and references therein).

In Drosophila embryos, DEAF-1 protein is ubiquitously expressed and appears to be constitutively localized in nuclei. The DEAF-1 protein contains two conserved domains, SAND and MYND, which are present in several transcription factors. The SAND domain (for SP100, AIRE-1, NucP41/75 and DEAF-1) was originally named "KDWK" after a conserved core of amino acid residues in DEAF-1 and other proteins (Gross, 1995). NUDR (nuclear DEAF-1-related factor, also known as Suppressin or mDEAF-1) is an apparent mammalian ortholog of DEAF-1 and is the only non-Drosophila protein containing both the SAND and the carboxy-terminal MYND domain. NUDR was characterized as a nuclear DNA binding protein that also preferentially binds DNA sites containing TTCG motifs, with a preference for clustered TTCG sequences (Veraksa, 2002).

Experiments on NUDR suggest that it is required to activate the proenkephalin promoter, but in a manner that does not involve DNA binding. In contrast, NUDR protein, as well as NUDR DNA binding sites, are required for the repression of the hnRNP A2/B1 promoter in tissue culture cells (Veraksa, 2002).

Unpaired copies of the SAND and MYND domains are also found in other proteins. A divergent SAND domain is present in the SP100 family of proteins, which are localized to subnuclear structures (PML bodies), and are thought to play a role in the etiology of acute promyelocytic leukemia. Different SP100 isoforms associate with chromatin components, and when bound to a promoter, behave as transcriptional activators or repressors. Other well-studied SAND domain proteins include GMEB-1 and GMEB-2 (glucocorticoid modulatory element binding factors). Their heterooligomeric complex was shown to bind variably spaced PuCGPy motifs. In cell transfection assays, GMEB-1 and GMEB-2 have been shown to function as transcriptional activators, repressors, or modulators. The three-dimensional structure of a SAND domain has been recently determined (Bottomley, 2001). The authors proposed that the SAND domain represents a novel DNA binding module characteristic for chromatin-dependent transcriptional regulation (Veraksa, 2002).

The MYND domain (myeloid, Nervy, and DEAF-1) is present in a large group of proteins that includes RP-8 (PDCD2), Nervy, and several ESTs and predicted proteins from Drosophila, mammals, C. elegans, yeast, and plants. The MYND domain consists of a cluster of cysteine and histidine residues, arranged with an invariant spacing to form a potential zinc-binding motif (Gross, 1995). Mutating conserved cysteine residues in the DEAF-1 MYND domain does not abolish DNA binding, which suggests that the MYND domain might be involved in protein-protein interactions (Gross, 1995). Indeed, the MYND domain of ETO/MTG8 interacts directly with the N-CoR and SMRT corepressors. Aberrant recruitment of corepressor complexes and inappropriate transcriptional repression is believed to be a general mechanism of leukemogenesis caused by the t(8;21) translocations that fuse ETO with the acute myelogenous leukemia 1 (AML1) protein. Recently, ETO was shown to be a corepressor recruited by the promyelocytic leukemia zinc finger (PLZF) protein. A divergent MYND domain present in the adenovirus E1A binding protein BS69 was also shown to interact with N-CoR and mediate transcriptional repression. The current evidence suggests that the MYND motif in mammalian proteins constitutes a protein-protein interaction domain that functions as a corepressor-recruiting interface (Veraksa, 2002).

Despite all of the molecular and biochemical information concerning DEAF-1 and related proteins, the lack of conventional genetic studies has hampered the understanding of this family of transcription factors. This study identified and characterized mutations in the Drosophila DEAF-1 gene. DEAF-1 maternal function is essential for early embryonic development; most maternal mutant embryos arrest before reaching the stage at which the zygotic segmentation genes are activated (Veraksa, 2002).

Two point mutations were isolated in the DEAF-1 gene; maternal and zygotic reduction of DEAF-1 function results in many embryos with arrested development during embryogenesis. In those embryos that develop to late embryogenesis and secrete cuticule, there are segmental defects of variable severity. In salivary cells, DEAF-1 protein localizes to about 200 sites on the polytene chromosomes. The loss of function phenotypes and in vivo chromosomal localization pattern are consistent with the idea that DEAF-1 functions in the regulation many different genes, especially at the earliest stages of Drosophila development, and perhaps also has an important role in the regulation of maternal genes whose products are deposited in the oocyte (Veraksa, 2002).

The DEAF-1S10B mutation eliminates all of the MYND domain and a predicted coiled coil structure. MYND domains in mammalian DEAF-1 homologs have been shown to interact with corepressors of transcription. The S10B mutation may compromise the repressive function of DEAF-1 by abolishing such interaction (Veraksa, 2002).

Another allele, DEAF-1k3, is a missense mutation that changes a cysteine residue to tyrosine. This residue corresponds to the carboxy terminus of the 2 helix in the SAND domain (Bottomley, 2001). In other DEAF-1 homologs, only glycine or alanine are encountered at the same position (see alignment in Bottomley, 2001). These amino acids and cysteine have small side chains, and introduction of a bulky hydrophobic residue of tyrosine may impair the folding of the SAND domain in this mutant or prevent crucial intermolecular interactions. Without any other point mutations available, it is difficult to assess whether DEAF-1k3 and DEAF-1S10B represent true null alleles. The variability of DEAF-1 mutant phenotypes suggests that these alleles are hypomorphic (Veraksa, 2002).

However, variable phenotypes resulting from loss of maternal protein have been reported for the TrxG genes kismet, osa, and the histone deacetylase Rpd3 (Veraksa, 2002).

As a complement to the loss of function studies reported in this study, an inducible DEAF-1 transgene was generated using a DEAF-1 cDNA downstream of UAS sequences. This construct was tested with a variety of imaginal disc specific GAL4 drivers. In all cases, overexpression of DEAF-1 resulted in partial or complete loss of tissues in which the transgene was activated. These abnormalities were likely to be a consequence of apoptosis, as revealed by increased uptake of acridine orange by cells overexpressing DEAF-1 (Veraksa, 2002).

Regulation by homeoproteins: A comparison of Deformed-responsive elements

Homeotic genes of Drosophila encode transcription factors that specify segment identity by activating the appropriate set of target genes required to produce segment-specific characteristics. Advances in understanding target gene selection have been hampered by the lack of genes known to be directly regulated by the HOM-C proteins. Evidence is presented that the gene 1.28, coding for a glycine-rich domain/Proline-rich domain protein, is likely to be a direct target of Deformed in the maxillary segment. A 664-bp Deformed Response Element (1.28 DRE) has been identified that directs maxillary-specific expression of a reporter gene in transgenic embryos. The 1.28 DRE contains in vitro binding sites for Deformed and DEAF-1. The Deformed binding sites do not have the consensus sequence for cooperative binding with the cofactor Extradenticle, and no cooperative binding to these sites is detected, though an independent role for Extradenticle cannot be ruled out. Removing the four Deformed binding sites renders the 1.28 DRE inactive in vivo, demonstrating that these sites are necessary for activation of this enhancer element, and supporting the proposition that 1.28 is activated by Deformed. Comparisons of the 1.28 DRE with other known Deformed-responsive enhancers indicate that there are multiple ways to construct Deformed Response Elements (Pederson, 2000).

To function as a maxillary enhancer the 120 bp module E requires at least one sequence in addition to the Deformed and Exd binding sites. This sequence is found in an imperfect inverted repeat. Site-directed mutagenesis of this imperfect inverted repeat abolishes module E enhancer function. Though it seemed noteworthy that a similar imperfect inverted repeat sequence is located within the 1.28 DRE, that sequence is not required for 1.28 DRE enhancer function, and deletion of this repeat has no consequence on expression of the endogenous 1.28 gene. Attaching the module E inverted repeat sequence to the Deformed binding portion of the 1.28 DRE does increase activity of the 1.28 DRE, indicating that this sequence can function in a heterologous enhancer. The idea is favored that the module E inverted repeat region contains a binding site or sites for other unknown factors, and that these factors act to enhance maxillary-specific expression. Such a site is likely to be within the inverted repeat sequence of module E. It has been proposed that factors bind to sequences GGC and AAAGC of the module E repeat. This sequence is not present in the 1.28 DRE, suggesting again that regulation through these two enhancers uses different mechanisms (Pederson, 2000 and references therein).

The DEAF-1 protein was initially hypothesized to be an activator involved in Deformed autoregulation because it binds tightly to the inverted repeat region of module E. However, accumulating evidence indicates that this may not be the case. The DEAF-1 binding site is located in region 6 of module E. Though this region is necessary for maxillary enhancer function, eliminating the DEAF-1 binding does not alter the ability of this fragment to be a maxillary enhancer. The DEAF-1 binding region of the 1.28 DRE does not enhance maxillary expression of either the 1.28 DRE or the module E Deformed binding sites, and furthermore, in both cases this region suppresses the weak, endogenous activity often observed for the pHZ-white reporter alone. The DEAF-1 binding region appears to, at least under some circumstances, act as a negative element. DEAF-1 perhaps does play a role in expression, as a repressor (Pederson, 2000).

DEAF-1, a novel protein that binds an essential region in a Deformed response element

A 120 bp homeotic response element that is regulated specifically by Deformed in Drosophila embryos contains a single binding site for Deformed protein. However, a 24 bp sub-element containing this site does not constitute a Deformed response element. Specific activation requires a second region in the 120 bp element, which presumably contains one or more binding sites for Deformed cofactors. A novel protein was isolated from Drosophila nuclear extracts that binds specifically to a site in this second region. This protein, called DEAF-1 (Deformed epidermal autoregulatory factor-1), contains three conserved domains. One of these includes a cysteine repeat motif that is similar to a motif found in Drosophila Nervy and the human MTG8 proto-oncoprotein, and another matches a region of Drosophila Trithorax. Mutations in the response element designed to improve binding to DEAF-1 in vitro resulted in increased embryonic expression. Conversely, small mutations designed to diminish binding to DEAF-1 resulted in reduced expression of the element. Thus, DEAF-1 is likely to contribute to the functional activity, and perhaps to the homeotic specificity, of this response element. Consistent with this hypothesis, DEAF-1 binding sites were discovered in other Deformed response elements (Gross, 1996; full text of article).


Search PubMed for articles about Drosophila DEAF-1

Bottomley, M. J., et al. (2001). The SAND domain structure defines a novel DNA-binding fold in transcriptional regulation. Nat. Struct. Biol. 8(7): 626-33. PubMed ID: 11427895

Carr, D. J., Blalock, J. E., Green, M. M. and LeBoeuf, R. D. (1990). Immunomodulatory characteristics of a novel antiproliferative protein, suppressin. J. Neuroimmunol 30: 179-187. PubMed ID: 2121798

Christensen, J., Cotmore, S. F. and Tattersall, P. (1999). Two new members of the emerging KDWK family of combinatorial transcription modulators bind as a heterodimer to flexibly spaced PuCGPy half-sites. Mol. Cell. Biol. 19(11): 7741-50. PubMed ID: 10523663

Gibson, T. J., Ramu, C., Gem√ľnd, C. and Aasland, R. (1998). The APECED polyglandular autoimmune syndrome protein, AIRE-1, contains the SAND domain and is probably a transcription factor. Trends Biochem Sci 23: 242-244. PubMed ID: 9697411

Gross, C. T. and McGinnis, W. (1996). DEAF-1, a novel protein that binds an essential region in a Deformed response element. EMBO J. 15(8): 1961-70. PubMed ID: 8617243

Huggenvik, J. I., Michelson, R. J., Collard, M. W., Ziemba, A. J., Gurley, P. and Mowen, K. A. (1998). Characterization of a nuclear deformed epidermal autoregulatory factor-1 (DEAF-1)-related (NUDR) transcriptional regulator protein. Mol. Endocrinol. 12(10): 1619-39. PubMed ID: 9773984

LeBoeuf, R. D., Burns, J. N., Bost, K. L. and Blalock, J. E. (1990). Isolation, purification, and partial characterization of suppressin, a novel inhibitor of cell proliferation. J. Biol. Chem. 265: 158-165. PubMed ID: 2294102

LeBoeuf, R. D., et al. (1998). Molecular cloning, sequence analysis, expression, and tissue distribution of suppressin, a novel suppressor of cell cycle entry. J. Biol. Chem. 273(1): 361-8. PubMed ID: 9417089

Michelson, R. J., et al. (1999). Nuclear DEAF-1 related (NUDR) protein contains a novel DNA binding domain and represses transcription of the heterogeneous nuclear ribonucleoprotein A2/B1 promoter. J. Biol. Chem. 274: 30510-30519. PubMed ID: 10521432

Pederson, J. A., et al. (2000). Regulation by homeoproteins: A comparison of Deformed-responsive elements. Genetics 156: 677-686. PubMed ID: 11014815

Reed, D. E., et al. (2008). DEAF-1 regulates immunity gene expression in Drosophila. Proc. Natl. Acad. Sci. 105(24): 8351-6. PubMed ID: 18550807

Surdo, P. L., Bottomley, M. J., Sattler, M. and Scheffzek, K. (2003). Crystal structure and nuclear magnetic resonance analyses of the SAND domain from glucocorticoid modulatory element binding protein-1 reveals deoxyribonucleic acid and zinc binding regions. Mol. Endocrinol. 17: 1283-1295. PubMed ID: 12702733

Veraksa, A., Kennison, J. and McGinnis, W. (2002). DEAF-1 function is essential for the early embryonic development of Drosophila. Genesis 33(2): 67-76. PubMed ID: 12112874

Biological Overview

date revised: 4 September 2007

Home page: The Interactive Fly © 2008 Thomas Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.