Protein Interactions

During the development of multicellular organisms, gene expression must be tightly regulated, both spatially and temporally. One set of transcription factors that are important in animal development is encoded by the homeotic (Hox) genes, which govern the choice between alternative developmental pathways along the anterior-posterior axis. Hox proteins, such as Drosophila Ultrabithorax, have low DNA-binding specificity by themselves but gain affinity and specificity when they bind together with the homeoprotein Extradenticle (or Pbx1 in mammals). To understand the structural basis of Hox-Extradenticle pairing, the crystal structure of an Ultrabithorax-Extradenticle-DNA complex at 2.4 A resolution was determined, using the minimal polypeptides that form a cooperative heterodimer. The Ultrabithorax and Extradenticle homeodomains bind opposite faces of the DNA, with their DNA-recognition helices almost touching each other. However, most of the cooperative interactions arise from the YPWM amino-acid motif of Ultrabithorax, located amino-terminally to its homeodomain, which forms a reverse turn and inserts into a hydrophobic pocket on the Extradenticle homeodomain surface. Together, these protein-DNA and protein-protein interactions define the general principles by which homeotic proteins interact with Extradenticle (or Pbx1) to affect development along the anterior-posterior axis of animals (Passner, 1999).

In the embryonic midgut both Decapentaplegic (Dpp) and Wingless (Wg) signaling pathways control the subcellular localization of Extradenticle protein. Exd protein is predominantly nuclear in endoderm cells close to the Dpp-and Wg-secreting cells of the visceral mesoderm, but is found in the cytoplasm in more distant endoderm cells. Both dpp and wg are required for the nuclear localization of Exd in the endoderm; ectopic expression of dpp and wg expands the domain of nuclear Exd (Mann, 1996). The requirement of homothorax for Exd's nuclear localization is apparent in many embryonic tissues, including the ectoderm, visceral mesoderm, and endoderm. This requirement is observed in cells where the signaling molecules Wg and Dpp contribute to Exd's nuclear translocation (e.g., the endoderm). It is suggested that Wg and Dpp both regulate expression of hth in the domains in which Exd nuclear function is required (Rieckhof, 1997).

In all cells where hth is expressed, Exd is localized to nuclei. Conversely, in most cells (but not all), where Exd is nuclear, hth is expressed. For example, during embryogenesis, Exd is cytoplasmic in the labial segment and in the limb primordia cells that express the gene Distal-less. For both of these cell types, hth is not expressed. Meis1/Hth also specifically binds to Exd with high affinity in vitro. Conditional expression of Meis1 in cultured Drosophila cells shifts Exd's subcellular localization within an hour. Hth can induce Exd's nuclear localization even when Asn-51 of the Hth homeodomain (implicated in DNA binding of other homeodomain proteins) has been mutated to Ala. These data suggest a novel and evolutionarily conserved mechanism for regulating HOX activity in which a direct protein-protein interaction between Exd and Hth results in Exd's nuclear translocation (Rieckhof, 1997).

Nuclear localization of the Extradenticle (EXD) and PBX1 proteins is regionally restricted during Drosophila and mammalian development. The subcellular localization of EXD, PBX, and their partners Homothorax (HTH) and PREP1, have been studied in different cell contexts. HTH and PREP1 are cytoplasmic and require association with EXD/PBX for nuclear localization. EXD and PBX1 are nuclear in murine fibroblasts but not in Drosophila Schneider cells, in which the proteins are actively exported to the cytoplasm. Coexpression of EXD/PBX with HTH/PREP1 causes nuclear localization of their heterodimers in both cell contexts. It is proposed that heterodimerization with HTH/PREP induces nuclear translocation of EXD and PBX1 in specific cell contexts by blocking their nuclear export (Berthelsen, 1999).

The homeobox gene extradenticle (exd) acts as a cofactor of Hox function in both Drosophila and vertebrates. It has been shown that the distribution of the Exd protein is developmentally regulated at the post-translational level; in the regions where exd is not functional, Exd is present only in the cell cytoplasm, whereas it accumulates in the nuclei of cells requiring exd function. Maternal EXD mRNA lasts for a few hours and is undetectable by stage 9 of embryogenesis. Protein produced by maternal RNA is stable. Both maternally and zygotically derived protein translocates into nuclei, suggesting that the proteins translocated from both mRNAs are functional and also suggesting that regulation of protein distribution is not dependent on transcriptional control (Azpiazu, 1998).

The subcellular localization of Exd is regulated by the BX-C genes; to varying degrees, BX-C genes can prevent or reduce nuclear translocation of Exd. Embryos homozygous for the deficiency Df(3R)P9, lacking the entire Bithorax complex, contain Exd at high levels in the nuclei of epidermal cells of both thoracic and abdominal segments. Only in the Keilin's organs of each segment does Exd remain cytoplasmic. This rise in the level of nuclear Exd in the abdominal segments (in comparison with the wild-type distribution) already indicates an involvement of the BX-C genes with the subcellular distribution of the product. In contrast, the thorax-determining genes Sex combs reduced (Scr) and Antp do not appear to affect Exd localization, as Scr Antp homozygous embryos exhibit normal distribution of Exd. To discriminate the roles of individual BX-C genes in the nuclear translocation of Exd, the distribution of Exd protein was examined in two more BX-C mutant combinations: Ubx- abd-A+ Abd-B+ and Ubx- abd-A- Abd-B+. Embryos homozygous for the first combination, defective only for Ubx function, show an increased level of nuclear Exd in the first abdominal segment as compared to wild type. In more posterior abdominal segments the levels and distributions of Exd are normal. In the second combination, lacking Ubx and abd-A functions, Exd is detected at high levels in the nuclei of the abdominal segments A1 to A4. In Df(3R)P9 (Ubx abd-A Abd-B) embryos, Exd nuclear localization extends to A8: all these results indicate that each BX-C gene is capable of preventing or reducing the nuclear translocation of Exd (Azpiazu, 1998).

The inhibition of Exd nuclear transport by overexpression of BX-C genes causes exd-like phenotypes. This was shown by inducing ectopic Ubx expression with several Gal4 lines during embryonic development and examining the segmental transformations produced. In the presence of normal exd function, Ubx specifies the pattern of the first abdominal (Al) segment. In contrast, in embryos lacking exd function, Ubx specifies a pattern resembling a more posterior segment, of A3-A5 type. Under conditions in which levels of Gal4 activity are high, larvae develop all segments anterior to A2 with an A3-A5 pattern. This segment pattern closely resembles that found in the A1 segment of zygotic exd larvae and is the same overall pattern observed after heat shock-inducing Ubx expression in zygotic exd embryos. These observations indicate that high levels of Ubx protein are able to produce an exd-like phenotype, in good agreement with the observed negative effect of BX-C genes on the nuclear translocation of Exd (Azpiazu, 1998).

During the evolution of insects from a millipede-like ancestor, the Hox genes are thought to have promoted the diversification of originally identical body structures. In Drosophila, antennae and legs are homologous structures that differ from each other as a result of the Hox gene Antennapedia (Antp), which promotes leg identities by repressing unknown antennal-determining genes. Four lines of evidence are presented that identify extradenticle (exd) and homothorax (hth) as antennal-determining genes. (1) Removing the function of eitherexd or hth (which is required for the nuclear localization of Exd protein), transforms the antenna into leg; such transformations occur without activation of Antp. (2) In most antennal cells, hth is expressed and Exd is nuclear, whereas both are restricted to proximal cells of the leg. (3) Antp is a repressor of hth. (4) Ectopic expression of Meis1, a murine hth homolog, can trigger antennal development elsewhere in the fly. Taken together, these data indicate that hth is an antennal selector gene, and that Antp promotes leg development by repressing hth, consequently preventing the nuclear transport of Exd (Casares, 1998).

homothorax (hth) is a Drosophila member of the Meis family of homeobox genes. hth function is required for the nuclear localization of the Hox cofactor Extradenticle (Exd). There is also a post-transcriptional control of Hth by exd: exd activity is required for the apparent stability of the Hth protein. To determine whether the lack of Hth in exd- clones is a result of transcriptional or post-transcriptional regulation, the expression of an hth enhancer trap was examined. In contrast to Hth protein, lacZ expression from the hth enhancer trap is maintained and in many cases upregulated in exd- clones. This suggests that the loss of Hth protein occurs post-transcriptionally, perhaps by protein degradation. Thus, the activities of hth and exd are intimately associated with one another: removing hth function results in cytoplasmic and presumably non-functional Exd, and removing exd function results in the loss of detectable Hth protein (Abu-Shaar, 1998).

Now that it is clear that hth determines antennal fate, it is worthwhile reconsidering the transformation to leg that is produced by hth or exd mutant cells in the antenna. This is the same phenotype seen with dominant Antp mutants, but the leg develops without the activity of Antp, Scr or Ubx. It follows that a leg can be generated without Hox activity, suggesting that the leg pathway is the ground state for ventral appendages. Thus the ground pattern for both larvae and adults is thoracic. Nor does Antp "select" for a specific leg pathway -- it simply represses hth in the leg primordia, thereby blocking antennal development and allowing the development of legs by default. This supports the idea that Antp promotes a ground (mesothoracic) pattern by repressing cephalic genes. This basal pattern is modified by Scr toward prothoracic (first leg) or by Ubx toward metathoracic (third leg) in their respective primordia. The downregulation of hth by Antp explains the phenotype of the dominant Antp mutants is due to homothorax repression. It also explains the ability of other Hox genes such as Ubx, abdominal-A, and Abdominal-B to induce the transformation of antennae into legs. These genes prevent the nuclear translocation of Exd (most likely through hth repression), so the antennal to leg transformations are probable nonspecific and caused by a property that is common to Antp and other Hox proteins (Morata, 1998 and Casares, 1998).

Hox proteins are transcription factors that assign positional identities along the body axis of animal embryos. Different Hox proteins have similar DNA-binding functions in vitro and require cofactors to achieve their biological functions. Cofactors can function by enhancement of the DNA-binding specificity of Hox proteins, as has been shown for Extradenticle (Exd). Three results support a novel mechanism for Hox cofactor function: regulation of transcriptional activation function. (1) Evidence is provided that the Hox protein Deformed (Dfd) can interact with simple DNA-binding sites in Drosophila embryos in the absence of Exd, but this binding is not sufficient for transcriptional activation of reporter genes. (2) Either Dfd or a Dfd-VP16 hybrid (VP16 is a transcriptional activation domain) mediate much stronger activation in embryos on a Dfd-Exd composite site than on a simple Dfd-binding site, even though the two sites possess similar Dfd-binding affinities. This suggests that Exd is required to release the transcriptional activation function of Dfd independent of Exd enhancement of Dfd-binding affinity on the composite site. (3) Transfection assays confirm that Dfd possesses an activation domain, which is suppressed in a manner dependent on the presence of the homeodomain. The regulation of Hox transcriptional activation functions may underlie the different functional specificities of proteins belonging to this developmental patterning family (Li, 1999).

The neutral state of Dfd on simple binding sites indicates that additional regulatory steps and regulatory sequences are required for Dfd to activate gene expression. To test the hypothesis that Dfd binding per se is inherently neutral in embryos, a test was performed to see whether high levels of Dfd or Dfd-VP16 proteins could activate transcription through simple Dfd recognition sites. In vitro, a DNA sequence consisting of two tandem copies of the simple Deformed binding site (D site or 2×D), is bound by Dfd with high affinity but not detectably bound by Exd. The affinity of Dfd protein for the 2×D-site is not enhanced by the inclusion of Exd protein (Li, 1999).

A test was performed of the embryonic function a varient of the D site reporter construct. This varient contains two tandem copies of a core sequence, to which Dfd and Exd bind together (2×ED2 sites). In vitro, the 2×ED2 site is bound weakly by Dfd protein alone, but is not bound detectably by Exd alone. Binding of Dfd to the 2×ED2 site is enhanced in the presence of Exd as shown by the formation of an abundant complex that contains Dfd, Exd and 2×ED2. The affinity of the Dfd-Exd heterodimer for the 2×ED2 site is approximately the same as the affinity of the Dfd protein alone for the 2×D site. Although the 2×D site and the 2×ED2 site have very similar in vitro affinity for Dfd in the presence of Exd, the 2×ED2 site is much more responsive than the 2×D site to either Dfd or Dfd-VP16 proteins in embryos. This strongly suggests that Exd is required to release the transcriptional activation function of Dfd in a way that is independent of the Exd enhancement of Dfd binding affinity on the 2×ED2 site. At present, the most widely accepted models propose Exd as a cofactor that has its effect on Hox specificity by acting to increase the binding affinity of different Hox proteins to different composite binding sites. The results presented here indicate that Exd has other regulatory effects on Hox proteins that may play a role in the diversification of function within the Hox family (Li, 1999).

Dfd protein contains an autonomous activation domain that is functional in transfection assays when separated from the C-terminal half of the protein. On tandem repeats of simple Dfd-binding sites, the function of the Dfd transcription activation domain is suppressed both in cultured cells and in embryos. In embryos, this suppression can be partially relieved by the addition of Exd-binding sites to simple Dfd-binding sites. This is apparently due to the function of the Exd protein, since exd genetic function is required for the relief of the suppression of Dfd activation function on 2×ED2 sites. In cultured cells, the suppression of Dfd activation function can be conferred by the homeodomain regions from either Dfd or Ubx. Since no evidence is found that there is a direct intramolecular interaction between the Dfd homeodomain and its transcriptional activation region, a model is proposed that invokes a masking factor that suppresses the function of the activation domain by contacting the homeodomain region. In addition, it is speculated that Exd may be required to alleviate the suppressive effect of the proposed masking factor by competing for overlapping protein-protein interaction sites on the homeodomain (Li, 1999).

During embryogenesis, in contrast with leg development, Antp selects for a specific developmental pathway. Loss-of-function mutations and experiments to induce ectopic expression show that Antp determines the larval mesothoracic pattern -- a function that is clearly distinct from the other Hox genes. Why legs should be different is not clear, but different Hox genes have similar effects on appendages, possibly because these appendages have no hth activity, without which the Hox genes lack specificity (Morata, 1998 and references).

extradenticle interacts genetically in parallel with Ultrabithorax, abdominal-A, and perhaps other HOM genes. There is a selective interaction of EXD with UBX and ABD-A proteins but not with an Antennapedia protein. Strong interaction with UBX requires only the Ultrabithorax homeodomain and a 15-residue N-terminal extension that includes Tyr-Pro-Trp-Met (YPWM), a tetrapeptide motif found near the homeodomain in most HOM proteins and their mammalian Hox counterparts. The size and sequence of the region between the YPWM element and the homeodomain differ among UBX isoforms. This variable region appears to influence the interaction detected in the assay (Johnson, 1995).

The homeodomain proteins encoded by the Drosophila extradenticle gene and its mammalian homologues, the pbx genes, contribute to HOX specificity by cooperatively binding to DNA with HOX proteins. For example, the HOX protein labial cooperatively binds with extradenticle protein to a 20-bp oligonucleotide identified in the 5' region of the mouse Hoxb-1 (the mammalian homolog of Drosophila labial) gene that is sufficient to direct a labial-like autoregulated expression pattern in Drosophila embryos. Labial and Extradenticle, their binding sites separated by only 4 bp, bind DNA as a heterodimer in a head-to-tail orientation. Mutations in base pairs predicted to contact the HOX N-terminal arm results in a change in HOX preference in the heterodimer, from Labial to Ultrabithorax. These results demonstrate that Extradenticle prefers to bind cooperatively with different HOX proteins depending on subtle differences in the heterodimer binding site (Chan, 1996b).

The homeotic proteins encoded by the genes of the Drosophila HOM and the vertebrate HOX complexes do not bind divergent DNA sequences with a high selectivity. In vitro, HOM (HOX) specificity can be increased by the formation of heterodimers with Extradenticle (Exd) or PBX homeodomain proteins. A single essential Labial (Lab)/Exd-binding site has been identified in a Decapentaplegic (Dpp)-responsive enhancer of the homeotic gene lab, which drives expression in the developing midgut. Lab and Exd bind cooperatively to the site in vitro, and the expression of the enhancer in vivo requires exd and lab function. In addition, point mutations in either the Exd or the Lab subsite compromise enhancer function, strongly suggesting that Exd and Lab bind to this site in vivo. Interestingly, the activity of the enhancer is only significantly stimulated by Dpp signaling upon the binding of Lab and Exd. Thus, the enhancer appears to integrate positional information via the homeotic gene lab, and spatiotemporal information via Dpp signaling; only when these inputs act in concert in an endodermal cell is the enhancer fully active. These results illustrate how a tissue-specific response to Dpp can be generated through synergistic effects on an enhancer carrying both Dpp- and HOX-responsive sequences (Grieder, 1997).

How does Dpp signaling affect the expression of the lab550 enhancer? The enhancer has been reported as a Dpp response element based on the fact that all of its activity in the midgut requires dpp function, but very little of it is due to lab function; in addition, it drives expression one or two cells more posterior than the endogenous lab gene, thus apparently displaying lab-independent expression. Although these conclusions somewhat contrast with those presented in the current paper, further analysis has suggested that several cyclic AMP response elements (CREs) in the enhancer mediate the Dpp responsiveness, directly or indirectly. More recently, in vitro binding sites for Drosophila Mad (a member of the SMAD protein family) and thought to act as co-activators of transcription in the Dpp signaling pathway, have also been identified in the lab enhancer. In addition, there are numerous binding sites for Schnurri, a putative transcription factor that is required for endodermal cells to respond to Dpp (Grieder, 1997 and references).

Despite the presence of all these putative target sequences of the Dpp signaling pathway, the current paper demonstrates that the expression of the lab enhancer strongly depends on the presence of a 48 bp region that harbors an Exd/Lab site; in exd and lab mutant embryos, expression of the element is strongly reduced. Since it recently has been shown that the translocation of Exd from the cytoplasm to the nucleus in the endoderm is controlled by Dpp and Wingless (Wg) , and since the endogenous lab gene is strongly induced by Dpp, it is likely that part of the Dpp regulation of the lab enhancer enters through the Exd/Lab site. This is consistent with the current findings and with the previous observation that in the absence of lab function, dpp-mediated induction is hardly working or working with reduced efficiency (Grieder, 1997 and references).

What is the role of the additional Dpp response elements in the lab enhancer? Clearly, the activity of the enhancer also depends on the presence of CRE sites; similar to observations concerning the Lab/Exd site, mutations in the four CRE sites result in weaker expression of the enhancer. This suggests that on the full-length enhancer, the additional (direct or indirect) Dpp input via the CRE and other sites is required in concert with the input through the Lab/Exd site, and that the absence of either input impairs expression of the enhancer. Two elements appear to act synergistically with respect to their response to Dpp; only on the addition of a weak Dpp response element to the Lab/Exd-containing element is a cis-acting element generated which is strongly Dpp inducible (and Lab/Exd dependent). It is proposed that the strong Dpp responsiveness of the enhancer is limited, through synergistic interactions, to those cells in which the Lab/Exd site is occupied. The function of this dual requirement might be to insure that the enhancer activates transcription exclusively in cells located in the central portion of the midgut endoderm where low levels of Lab in the tip of the two endoderm primordia coincide with Dpp, secreted from the visceral mesoderm parasegment 7 (Grieder, 1997).

Extradenticle protein raises the DNA binding specificity of UBX and ABD-A but not that of ABD-B. It also modulates the DNA binding activity of Engrailed to a different target site. While a region N-terminal of the EXD homeodomain is required for UBX and ABD-A cooperativity, Engrailed requires a domain C-terminal of the EXD homeobox (van Dijk, 1994).

The Ultrabithorax and Antennapedia genes of Drosophila encode homeodomain proteins that have very similar DNA binding specificities in vitro but specify the development of different segmental patterns in vivo. Cooperative interactions occur between UBX protein and EXD that selectively increase the affinity of UBX (but not ANTP) for a particular DNA target. UBX and EXD bind to neighboring sites on target DNA and interact directly to stabilize the DNA-bound form of UBX (Chan, 1994). EXD nevertheless modulates Antennapedia function (Rauskolb 1994).

A highly conserved region of the Engrailed protein, EH2, N-terminal of the homeodomain is required for interaction with Extradenticle and vertebrate homologues of EXD designated PBX1, PBX2 and PBX3. Two tryptophan residues present in the Drosophila and murine Engrailed EH2 domain are required for cooperativity with EXD and PBX. A related motif, present in Hox proteins, called the hexapeptide, is necessary for Hox interaction with PBX proteins. The EH2 domain is distinct from the hexapeptide present in Hox proteins with respect to the amount of conserved residues, but both contain conserved tryptophan residues and the length of the linker region separating the Pbx interaction motif from the homeodomain in both Hox and Engrailed proteins is important for cooperativity. Nevertheless the N-terminal flanking regions and homeodomains of En and Hox proteins cannot be interchanged, consistent with the idea that the Pbx interaction domain in Hox and Engrailed proteins have evolved with their associated homeodomains (Peltenburg, 1996).

Outside of their homeodomains, HOX proteins have no sequence similarities except for two short peptides: 'MXSYF', at their N-termini, and the hexapeptide (also called the 'YPWM' motif or pentapeptide) N-terminal to their homeodomains. A 20 bp human Hoxb-1 promoter oligonucleotide is sufficient to direct an expression pattern in the fly that is very similar to endogenous labial expression. This expression pattern requires lab and extradenticle. Labial proteins with mutations in the hexapeptide bind DNA in the absence of EXD and have an increased ability to activate transription in vivo. Proteolysis experiments suggest that EXD can induce a conformational change in LAB. Apparently, LAB hexapeptide inhibits LAB function by inhibiting DNA binding, and an EXD-induced conformational change in LAB relieves this inhibition, promoting highly specific interactions with biologically relevant binding sites (Chan, 1996a).

The Distal-less gene is known for its role in proximodistal patterning of Drosophila limbs. However, Distal-less has a second critical function during Drosophila limb development, that of distinguishing the antenna from the leg. The antenna-specifying activity of Distal-less is genetically separable from the proximodistal (PD) patterning function because certain Distal-less allelic combinations exhibit antenna-to-leg transformations without proximodistal truncations. Distal-less has been shown to act in parallel with homothorax (a previously identified antennal selector gene) to induce antennal differentiation. While mutations in either Distal-less or homothorax cause antenna-to-leg transformations, neither gene is required for the others expression, and both genes are required for antennal expression of spalt. Coexpression of Distal-less and homothorax activates ectopic spalt expression and can induce the formation of ectopic antennae at novel locations in the body, including the head, the legs, the wings and the genital disc derivatives. Ectopic expression of homothorax alone is insufficient to induce antennal differentiation from most limb fields, including those of the wing. Distal-less therefore is required for more than induction of a proximodistal axis upon which homothorax superimposes antennal identity. hth encodes a TALE-class homeodomain protein required for the nuclear localization of a PBC-class homeodomain protein encoded by extradenticle. Based on their genetic and biochemical properties, it is proposed that Homothorax and Extradenticle may serve as antenna-specific cofactors for Distal-less (Dong, 2000).

Could Dll form a functional complex with Hth and Exd in the antenna? Given that Dll and Hth cooperate to regulate antennal differentiation, it is of interest to elucidate the molecular basis of this synergy. Exd and its vertebrate counterpart, Pbx, are known cofactors for a variety of homeodomain proteins, including Labial, Engrailed and Ultrabithorax. Hth is required for retention of Exd in the nucleus and may form part of the functional Exd/Hox complex. Vertebrate homologs of Hth, the Meis and Prep proteins, have been shown to form trimeric complexes with Hox and Pbx proteins. Several lines of evidence now support the idea that Exd and Hth are cofactors for the Dll homeodomain protein in the developing Drosophila antenna. These include: (1) the similar antenna-to-leg transformation phenotypes of Dll, hth and exd mutants; (2) the known physical interactions of Exd and Hth with other homeodomain proteins; (3) the fact that Dll and hth function in parallel to regulate antennal development, and (4) the fact that ectopically expressing Hth can mimic loss of Dll function in the antenna. Testing whether Dll, Hth and Exd interact physically and whether such a complex activates antennal enhancers will be important steps toward understanding limb development and tissue-specific gene regulation (Dong, 2000).

The gene homothorax is required for the nuclear import of Extradenticle, The functions of exd/hth and of the Hh/Wg/Dpp pathway are mutually antagonistic: exd blocks the response of Hh/Wg/Dpp target genes such as optomotor-blind and dachshund; high levels of Wg and Dpp eliminate exd function by repressing hth. This repression is mediated by the activity of Dll and dac. One prerequisite for appendage development is the inactivation of the exd/hth genes (Azpiazu, 2000 and references therein).

The main role of hth is to regulate exd function. The loss of hth activity during adult patterning results in changes in segmental identity and morphogenetic alterations that appear to be similar or identical to those produced by eliminating exd. Thus hth and exd can be considered to perform the same developmental function. In the wing disc, hth and exd are only required in the wing hinge region and, in their absence, the cells proliferate but form aberrant patterns indicating that hth/exd function is involved in specifying the wing hinge region. The experiments inducing ectopic hth expression suggest that it has a role in controlling growth, for hth is able to prevent the formation of the wing pouch. It is also consistent with the observation that hth mutant clones in the hinge may reach very large size. The finding that hth suppresses wg activity in the DV border may be related with the repression of growth, a process with which wg has been shown to be involved. The lack of effect of hth on dpp expression emphasizes the independence of the AP wing axis from hth/exd function. One aspect that is not fully understood is the effect of hth on the proximodistal pattern, which has also been observed on the leg disc and on the chicken limb. In the experiments described here, the presence of the Hth product influences the reading of proximodistal signals by the cells towards differentiating more proximal patterns. It is not known which factors are responsible for the proximodistal pattern in the wing, but since hth prevents wg response to Notch, it is possible that a Wg response element or some other Notch response gene may be involved in patterning (Aspiazu, 2000).

Altogether, the results presented here suggest the subdivision of the non-thoracic part of the wing disc into two major domains: the wing hinge, where hth is expressed and Exd is functional (nuclear), and the wing pouch where hth is not expressed, and Exd is cytoplasmic and therefore inactive. By homology with the leg disc, the latter would be the genuine appendage part of the disc. These two regions are formed by two antagonistic genetic systems: in the hinge, the high levels of hth, inherited from the embryo and probably maintained by wg, tsh and maybe other regulators, prevent wg response to Notch signaling, which is necessary for the development of the wing pouch. In the wing pouch, the activities of the Wg and Dpp pathways suppress hth so that Notch may induce wg activity and the appendage is formed. In addition to its role in preventing excessive proliferation, hth may also contribute, together with tsh, wg and nub, to the partition of the wing hinge into two regions that correspond to the outer and inner rings of hth expression. The outer ring domain expresses tsh, wg and hth; has nuclear Exd and does not express vg and nub. The inner domain expresses wg, nub and hth, has nuclear Exd and does not express tsh. The individual role of these genes is not yet established, but it is possible that they function in some combinatorial manner (Aspiazu, 2000).

Drosophila Homothorax (Hth) and Extradenticle (Exd) are two homeoproteins required in a number of developmental processes. Exd can function as a cofactor to Hox proteins. Its nuclear localization is dependent on Hth. Evidence is presented of in vivo physical interaction between Hth and Exd, mediated primarily through an evolutionarily conserved MH domain in Hth. This interaction is essential for the mutual stabilization of both proteins, for Exd nuclear localization, and for the cooperative DNA binding of the Exd-Hth heterodimer. Some in vivo functions require both Exd and Hth in the nucleus, suggesting that the Exd-Hth complex may function as a transcriptional regulator (Jaw, 2000).

To assess the ability of Hth in inducing Exd nuclear localization, the Hth mutant constructs were expressed using the UAS-GAL4 system in two different cell types: the eye field of larval eye disc, and the larval salivary gland. Two Hth deletion constructs were tested: deltaMH deletes residues 31-312, which include the conserved MH domain (residues 91-219), and deltaHD deletes residues 299-459, which include the HD (residues 368-428). deltaHD is still capable of inducing Exd nuclear localization, indicating that the region deleted is not required for this function. deltaMH has no ability to induce Exd nuclear localization, suggesting that the MH domain is required for Exd nuclear localization. A GST pull-down assay shows that the region deleted in deltaMH is required for the physical interaction with Exd, and the region deleted in deltaHD is not essential for the interaction. In salivary gland, both deltaMH and deltaHD proteins are located in the nucleus. The observation that deltaMH is itself nuclear indicates that Hth can enter the nucleus independent of Exd. Since only the 30 N-terminal residues (1-30) and the 28 C-terminal residues (460-487) are shared between deltaMH and deltaHD, it is possible that a nucleus localization signal (NLS) is located in one of these two regions. However, no sequence in these two regions fits the canonical NLS motif. Therefore it is likely that Hth has two independent NLS, one located within residues 91-219 and another in residues 368-428. The two best match of NLS motif are indeed found in these two regions: KRDK (residues 91-94, at the N-terminal end of MH) and KKNQKKR (residues 363-369, at the N-terminal end of HD). The eye field deltaMH, although still retaining the putative NLS in HD, is primarily cytoplasmic. The different distribution in different cells suggests that the remaining NLS function in deltaMH is weak and is influenced by other factors (Jaw, 2000).

Ectopic expression of the full length Hth (driven by dpp-GAL4) causes several major phenotypes in adult flies: eyes are absent or very small, the arista of the antennae are missing, the third antennal segments are occasionally duplicated, and the distal leg segments are deleted, malformed, and occasionally bifurcated. Whether the mutant Hth constructs can cause any of these phenotypes in transgenic flies was examined. deltaMH causes no effect on antenna, eye and leg morphology. Since deltaMH fails to induce Exd nuclear localization, it appears that nExd is required to affect antenna, eye and leg development. Although deltaHD can induce Exd nuclear localization, it caused no effect on antenna and eye development, suggesting that the Hth HD is required to affect eye and antenna development and nExd alone is not sufficient. deltaHD caused leg defects similar to those induced by full length Hth, indicating that the Hth HD is not required to affect leg development. When a NLS-Exd construct is expressed, the FLAG-tagged NLS-Exd is located in the nucleus in the absence of Hth. When NLS-Exd expression is driven by the dpp-GAL4, the eyes and antennae are not affected, but the femur and tibia leg segments are deformed. These results confirm that nExd alone is not sufficient to affect eye and antenna development, while part of its effect on leg development does not require Hth. The leg phenotypes caused by deltaHD and NLS-Exd are not the same: deltaHD affects the distal segments and NLS-Exd affects the medial segments. Even when induced at 29 degrees C, NLS-Exd did not affect the distal segments. The difference in phenotype suggests that some effect on leg development also requires the contribution from Hth, but does not require its HD. It is concluded that some functions require both nExd and nHth, while some require only nExd (Jaw, 2000).

The hexapeptide and linker regions of the AbdA hox protein regulate its activating and repressive functions

The Hox family transcription factors control diversified morphogenesis during development and evolution. They function in concert with Pbc cofactor proteins. Pbc proteins bind the Hox hexapeptide (HX) motif and are thereby thought to confer DNA binding specificity. The mutation of the AbdA HX motif as reported here does not alter its binding site selection but does modify its transregulatory properties in a gene-specific manner in vivo. A short, evolutionarily conserved motif, PFER, in the homeodomain-HX linker region acts together with the HX to control an AbdA activation/repression switch. These in vivo data thus reveal functions not previously anticipated from in vitro analyses for the hexapeptide motif in the regulation of Hox activity (Merabet, 2003).

Hox proteins share a helix-turn-helix DNA binding motif, the homeodomain (HD), and, consequently, recognize very similar TAAT core sequences; this fact contrasts with the highly specific biological functions carried out by Hox proteins during development. It is now well established that Hox proteins gain specificity by physically interacting with Pbc class cofactors. Association with Pbc proteins increases the DNA binding specificity of Hox proteins: Hox/Pbc complexes recognize a larger motif, TGATNNATNN, where the identity of the central NN nucleotides depends on the particular Hox protein involved (Merabet, 2003).

Hox/Pbc interactions are now well characterized in vitro, both in biochemical and structural terms. Pbc proteins belong to the TALE (three amino acid loop extension) class of atypical HD-containing proteins that is characterized by a three amino acid insertion between helices 1 and 2 of the HD. These residues participate in the constitution of a hydrophobic pocket that mediates interaction with Hox proteins, through a short evolutionary conserved sequence, the hexapeptide (HX), lying upstream of the HD in all but the Abdominal-B class of Hox proteins. Structurally, the HX folds into a classical type I reverse turn and is connected to the HD by a short sequence commonly termed the linker region. The variable length and disordered structure of the linker region suggest that it has a passive role in connecting the HX to the HD (Merabet, 2003).

In vitro, the HX promotes the formation of Hox/Pbc complexes with heightened DNA binding affinity and specificity, suggesting that this domain critically contributes to the selection of Hox target genes during development. The role of the HX in vivo has, however, been poorly investigated so far. One study has addressed the point by analyzing in Drosophila the effect of the HX-mutated Labial (Lab) protein on the regulation of a heterologous mouse Hoxb1 enhancer, 3Xrpt3 (Popperl, 1995). The authors concluded that the recruitment of Extradenticle (Exd) by the HX neutralizes an inhibitory effect of the HX on Lab DNA binding. This might, however, be a very specialized function of the HX, since Lab is very peculiar in the sense that, unlike most Hox proteins, it does not bind DNA on its own (Merabet, 2003 and references therein).

A detailed analysis has been carried out of the HX function in the regulation of bona fide target genes during development. To address the contribution of the HX and the linker region, two variants were generated, AbdA(HXm) and AbdA(PFERm), where the YPWM motif and the PFER sequence were mutated into AAAA. PFER lies in the middle of the AbdA linker region, which has been fairly well conserved in the insect phylum. The HX is not involved in controlling DNA binding and target gene selection, nor is it necessary for Exd recruitment, but, rather, it controls transregulatory functions of the Hox protein Abdominal-A (AbdA). An evolutionarily conserved motif in the linker region interferes with the HX, to control a repression/activation switch in AbdA (Merabet, 2003).

Extensive in vitro analyses have demonstrated that the HX is responsible for the interaction with Pbc proteins, leading to the view that this motif imparts Hox DNA binding specificity and therefore assists Hox proteins in the selection of appropriate target genes. In vivo data challenge this view in several ways. (1) The unaltered capacity of AbdA(HXm) to induce A2-like identities in the thorax and to form dimeric complexes on DNA with Exd shows that the HX is not the only motif of AbdA that is able to recruit Exd. A similar situation has been shown to occur in Ubx, indicating that other residues in Hox proteins can compensate for the lack of the HX in mediating Hox/Exd interactions. (2) Mutation of the HX does not affect binding site selection by AbdA, as shown by the ability of the mutant protein to bind target sequences from Dll and dpp in vitro, and to control dpp promoter elements in vivo. Accordingly, the HX mutation does not alter target gene selection (in this case, wg and dpp in the VM) in vivo. (3) The fact that the HX mutation modifies AbdA function in the regulation of dpp, which does not depend on Exd, implies that the HX should interact with additional proteins that remain to be identified. These data thus endow the HX with unexpected functions; this does not preclude that the HX could, however, play a role in target selection in other developmental contexts. The PFER motif within the linker region was found to fulfill an important regulatory function; this was also unexpected, considering the variable length and disordered structure of this region (Merabet, 2003).

The regulation of dpp by AbdA in the VM is mediated by the dpp674 enhancer, which contains seven binding sites for AbdA. Sites 1-4 in dpp419 (the 3' portion of dpp674) mediate repression by AbdA, while sites 5-7 in dpp265 (the 5' portion of dpp674) mediate activation. Interestingly, dpp265 reveals an activating potential of AbdA on dpp transcription that is masked by the prevalence of repression over activation in the regulation of dpp674 or dpp. Exd acts in a Hox-independent manner to repress dpp in the anterior VM. Anterior expression of dpp induced by AbdA(HXm) could therefore result from an interference with the repressive function of Exd, rather than from a direct effect on dpp transcription. However, while dpp265 is not derepressed anteriorly in exd- or hth-deficient animals and, therefore, does not contain the sequences mediating repression by Exd, it is activated by AbdA(HXm). Thus, Exd and AbdA(HXm) act on different regulatory sequences to respectively repress or activate dpp in the anterior VM, which makes it unlikely that activation by AbdA(HXm) results from an interference with the Hox-independent repressive function of Exd. Considering that the HX mutation affects neither DNA binding nor target site recognition in vitro and in vivo, it is proposed that AbdA(HXm), as does AbdA, controls dpp transcription directly (Merabet, 2003).

Engrailed cooperates with Extradenticle and Homothorax to repress target genes in Drosophila

Engrailed is a key transcriptional regulator in the nervous system and in the maintenance of developmental boundaries in Drosophila, and its vertebrate homologs regulate brain and limb development. The functions of both of the Hox cofactors Extradenticle and Homothorax play essential roles in repression by Engrailed. Mutations that remove either of them abrogate the ability of Engrailed to repress its target genes in embryos, both cofactors interact directly with Engrailed, and both stimulate repression by Engrailed in cultured cells. A model is suggested in which Engrailed, Extradenticle and Homothorax function as a complex to repress Engrailed target genes. These studies expand the functional requirements for Extradenticle and Homothorax beyond the Hox proteins to a larger family of non-Hox homeodomain proteins (Kobayashi, 2003).

As a first step in determining the mechanisms whereby exd and hth contribute to repression by En in vivo, the possibility of direct interaction was examined. En can bind co-operatively with Exd in vitro to artificial DNA sites. Whether a direct En-Exd interaction could also occur in other contexts was examined using yeast two-hybrid and in vitro assays. Whether En could interact similarly with Hth was also examined. En appears to interact robustly with Exd in the yeast two-hybrid system, because the signal strength observed with both isolated colonies and colony streaks is consistently higher than that seen with some positive controls, including the functionally important interaction between En and Groucho. This signal was also comparable to that seen with Exd and the mouse homolog of Hth, Meis1. En also gives a somewhat weaker, but apparently specific, signal in combination with either Hth or Meis1 (Kobayashi, 2003).

In vitro, En also interacts specifically with both Exd and Meis1. En fused with GST effectively pulls down either Exd or Meis1. Meis1 was used in these studies because of the high level of non-specific interaction observed with in-vitro-translated Hth, perhaps owing to the heterologous nature of the translation system. In this system, it is unlikely that the interactions are due to co-operative binding to DNA, and these results are interpreted to mean that these interactions can occur in solution. Furthermore, Meis1 appears to interact more strongly with En in the presence of Exd, suggesting that the three proteins form a co-complex (Kobayashi, 2003).

In cultured Drosophila cells, Exd and Hth cooperate with En to repress transcription. Using a co-operative binding site for Exd and En to construct an En-responsive target gene, it was found that both Exd and Hth are required for full repression activity. When a mutation is introduced into an Exd consensus binding sequence that eliminates co-operative binding, co-operative repression is largely eliminated, whereas mutating the En consensus binding sequence eliminates repression. This, along with the fact that RNA interference directed against Exd mRNA also largely eliminates co-operative repression, suggests that a complex containing Exd and En is responsible for the co-operative repression caused by coexpression of Hth and En (Exd is constitutively expressed in these cells). Because Hth regulates the nuclear localization of Exd, it can allow Exd-En repression complexes to form in the nucleus. In addition, the observed molecular interactions suggest that the fully active repression complex might include all three proteins (Kobayashi, 2003).

Exd cooperates with En to repress target genes and to pattern embryos. Loss of exd function has been shown to result in a loss of en expression at later embryonic stages. Because en function is required to maintain its own expression, the loss of en expression could be a downstream effect of a loss of en function, or it could be due to some other consequence of the lack of exd. This ambiguity concerning the role of exd in en function led to an investigation of whether the activities of ectopically expressed En are dependent on exd function. En was ectopically expressed in two ways: from a heat-shock promoter and using a patterned Gal4 'driver' transgene. An advantage of the former approach is that one can often distinguish between immediate and secondary downstream effects based on how rapidly they occur following heat induction. Advantages of the second approach include having normal and altered expression in parts of the same embryo, providing a rigorous internal control. Both of these approaches led to similar conclusions, that exd function is important for the repression by En of its direct target gene slp, that wg also shows a strong dependence on exd function for its repression by En and that the ability of En to alter the pattern of embryonic cuticles is sensitive to the gene dosage of exd. Further, in each set of experiments, the observed dependence of repression on exd was accompanied by a residual repression activity when exd function was removed both maternally and zygotically. This residual exd-independent repression activity might be due to the ability of En to bind to target sites independently of exd but with a reduced affinity, or it could be accounted for by the existence of two classes of binding sites, one exd dependent and the other exd independent. This possibility is paralleled by the relationship of Exd with Ubx, which has been shown to function either co-operatively with Exd or alone on multiple binding sites in target genes. Alternatively, exd might be exerting an indirect effect on repression by En. However, because Exd forms complexes with En in yeast and in vitro, and because it appears to facilitate repression by En directly in cultured cells, it seems likely that the dependence of En on exd function in vivo is due at least in part to the direct action of En-Exd complexes. Confirmation of this model will require the analysis of specific regulatory sites, which have not yet been identified, in target genes such as slp. If this model is correct then these results suggest that the repression activity of Exd-En complexes might come exclusively from En repression domains, because Exd has been shown to act as a cofactor in the activation of target genes in vivo in conjunction with Hox proteins (Kobayashi, 2003).

The effects of eliminating exd function on repression by En appear to be different in the abdomen and the more-anterior regions: En is less dependent on exd in the abdomen (parasegments 6-12). One possible explanation is that hth can provide the observed exd-independent activity. However, in exd mutants, Hth levels are reduced, probably because Hth protein is less stable without Exd. Nevertheless, these data are consistent with the possibility that, on their own, either Hth or Exd might provide partial cofactor activity, whereas both together might be required for full activity. The latter possibility is suggested by the observation that maximal repression activity in S2 cells requires all three gene products (Kobayashi, 2003).

An additional possibility to account for the residual exd- and hth-independent repression activity of En in the abdomen is that other cofactors assist En in binding to its target genes in the abdomen. If there are other cofactors at work, it is likely that their activity (or expression) is dependent, either directly or indirectly, on the Hox genes Ubx and abd-A, because these genes are responsible for all known aspects of differential segment identity in this region of the embryo (Kobayashi, 2003).

It is noteworthy that the difference in the dependence of En on exd in the abdomen versus the thorax is seen only after stage 9, when the levels of Hth, and the consequent nuclear concentration of Exd, have declined in the abdomen. Thus, the dependence of En on exd parallels the nuclear concentration of Exd, and might reflect an evolutionary adaptation to the changing levels of Exd in different regions of the embryo (Kobayashi, 2003).

Hth has been shown to act in part through its facilitation of the nuclear localization of Exd, and strong hth and exd mutants have very similar phenotypes. Although Hth can also interact with En independently of Exd, transfection assays in cultured cells suggest that Hth might depend entirely on Exd for its ability to increase repression by En, at least from artificial En-Exd co-operative binding sites. Because Hth forms complexes with En in these cells, in addition to increasing its repression activity, a simple model is that maximal repression activity is due to complexes containing En, Exd and Hth. However, the possibility cannot be ruled out that Hth acts solely by making Exd available to interact with En on target sites, through its ability to bring Exd into the nucleus (Kobayashi, 2003).

Whether the repression activity of ectopically expressed En in vivo is dependent on hth function was tested using assays similar to those used for exd. In each case, a close similarity was observed to results with exd mutants. En activity shows a strong dependence on hth function, although residual activity remains in hth mutants. In addition, En activity shows a sensitivity to the hth gene dose. All of these results are consistent with the effects of Hth being exerted through its effect on Exd nuclear localization, provided that the nuclear targeting of Exd is necessary for its ability to function with En. However, Hth might also increase the effectiveness of En repression directly, by forming complexes with En and/or as part of En-Exd complexes. A detailed analysis of a number of in vivo target sites will be necessary to distinguish among these possibilities (Kobayashi, 2003).

Exd and Hth are essential to the correct regulation of target genes by the homeodomain proteins of the Hox clusters. However, their functional interactions have not previously been shown to extend beyond the highly restricted subset of homeodomain proteins that are found within the Hox clusters (the Antp, Abd-B and Labial classes). The identification of functional interactions with En suggests that exd and hth might provide functional specificity in conjunction with other non-Hox-class homeodomain proteins (Kobayashi, 2003).

The identification of slp as a direct target gene of En has implications for the mechanism by which En helps to maintain the activity of its own and other genes, including hedgehog, within its domains of expression, the posterior compartments. The slp locus produces two closely related, coordinately regulated gene products (Slp1 and Slp2), which have essentially indistinguishable functions. They are forkhead-domain transcription factors that repress en expression, and both contain a conserved motif (homology region II) that is similar to the Groucho-binding domain of En. Slp1 has also been shown to bind the Groucho co-repressor in vitro, suggesting that it is a repressor and therefore that its action on the en gene is likely to be direct. Thus, the mechanism of en autoregulation, as well as the ability of En to activate other target genes, is likely to be due, at least in part, to an indirect effect of repression of slp expression. In addition, En might activate target genes indirectly by repressing other repressors that are also normally excluded from its expression domain, such as Odd-skipped and the repressor form of Cubitus interruptus (Kobayashi, 2003).

Although there have been previous suggestions that Exd and Hth might participate in active repression as well as activation complexes, most of the well-characterized direct Exd-Hth-Hox target genes are activated in an exd- or hth-dependent fashion. In fact, these observations raised the question of whether Exd and Hth might be dedicated to gene activation. Recently, Hth and Exd have been shown to act directly with Ubx to repress the Hox target gene Distalless in the Drosophila abdomen. The partnership with En in repression further argues that these cofactors can increase the target site discrimination of homeodomain proteins without restricting the resulting transcriptional activity to activation alone. Based on these results, it is suggested that Hth and Exd increase the target-site discrimination of several classes of homeodomain proteins and that they do so without defining the transcriptional activity of the resulting protein complex (Kobayashi, 2003).

In the absence of MEIS family proteins, two mechanisms are known to restrict the PBX family of homeodomain (HD) transcription factors to the cytoplasm. (1) PBX is actively exported from the nucleus via a CRM1-dependent pathway. (2) Nuclear localization signals (NLSs) within the PBX HD are masked by intramolecular contacts. In a screen to identify additional proteins directing PBX subcellular localization, a fragment of murine nonmuscle myosin II heavy chain B (NMHCB) was identified. The interaction of NMHCB with PBX was verified by coimmunoprecipitation; immunofluorescence staining revealed colocalization of NMHCB with cytoplasmic PBX in the mouse embryo distal limb bud. The interaction domain in PBX maps to a conserved PBC-B region harboring a potential coiled-coil structure. In support of the cytoplasmic retention function, the NMHCB fragment competes with MEIS1A to redirect PBX, and the fly PBX homolog EXD, to the cytoplasm of mammalian and insect cells. Interestingly, MEIS1A also localizes to the cytoplasm in the presence of the NMHCB fragment. These activities are largely independent of nuclear export. The subcellular localization of EXD is deregulated in Drosophila zipper mutants that are depleted of nonmuscle myosin heavy chain. This study reveals a novel and evolutionarily conserved mechanism controlling the subcellular distribution of PBX and EXD proteins (Huang, 2003).

Recognition of distinct target sites by a unique Labial/Extradenticle/Homothorax complex

Hox genes encode evolutionarily conserved transcriptional regulators, which define regional identities along the anteroposterior axis of multicellular animals. In Drosophila, Hox proteins bind to target DNA sequences in association with the Extradenticle (Exd) and Homothorax (Hth) co-factors. The current model of Hox-binding selectivity proposes that the nucleotide sequence identity defines the Hox protein engaged in the trimeric complex, implying that distinct Hox/Exd/Hth complexes select different binding sites and that a given Hox/Exd/Hth complex recognizes a consensus DNA sequence. The regulation of a newly identified Lab target gene does not rely on the previously established consensus Lab/Exd/Hth-binding site, but on a strongly divergent sequence. Thus Lab, and most probably other Hox proteins, selects different DNA sequences in regulating downstream target genes. These observations have implications with regard to the current model of Hox-binding selectivity (Ebner, 2005).

Understanding how Hox proteins trigger diversified morphogenesis requires the identification of the mechanisms underlying appropriate target gene selection as well as appropriate target gene regulation, which relies on controlling Hox transregulatory properties. At present, most studies have focused on how Hox proteins cooperate with two classes of co-factors, Exd/Pbx and Hth/Meis, to reach DNA-binding selectivity. Although not valid for the regulation of all Hox target genes, the Hox-binding selectivity model is a useful conceptual framework for understanding how Hox proteins, which as monomers display similar DNA-binding properties, reach specificity in target site recognition by interacting with a single co-factor, Exd. This model implies that distinct Hox/Exd complexes select different binding sites for three reasons, each of which has been well documented: (1) in vitro studies have shown that the prototypical TGAT[NN]ATNN Hox/Exd site recruits Lab or Ubx, depending on the identity of the two central NN nucleotides: GG selects Lab/Exd, while TT or TA recruits a Ubx/Exd complex; (2) the Distalless regulatory element that mediates repression by Ubx contains a Hox/Exd site where the two central nucleotides are TT; (3) switching the identity of these two central nucleotides from GG to TA, within the context of repeat3, leads to the recruitment of a Dfd/Exd complex instead of Lab/Exd, and to transformation of the Lab-responsive enhancer into a Dfd-responsive enhancer, as revealed by an in vivo test. Similar DNA binding preferences were also observed with the vertebrate Hox and Pbx homologues (Ebner, 2005 and references therein).

An in silico approach based on the Hox DNA-binding selectivity model was used to find novel Lab target genes. Although the approach identified 40 putative target sequences for the Lab/Exd/Hth complex, expression analysis of half of them only identified a single novel Lab target, CG11339. This suggests that sequences mediating Lab regulatory function in vivo are insufficiently well defined, which is further supported by the finding that the regulation of CG11339 does not rely on the consensus Lab/Exd/Hth-binding site used for the in silico approach, but on a strongly divergent sequence. These results have implications both with regard to the mode of Lab DNA-binding and more generally to the Hox-binding selectivity model (Ebner, 2005).

Previous work proposed that Lab is very peculiar among all other Hox proteins, in the sense that it does not bind DNA as a monomer, but does so in association with the co-factor Exd. Mutation of the hexapeptide (HX), a short motif upstreaam of the homodomain, confers to Lab the capacity to bind DNA in the absence of Exd. Accordingly, it was proposed that the HX exerts an inhibitory effect on Lab DNA binding, which is neutralized when interaction occurs with Exd. This conclusion was reached by studying the DNA-binding properties of Lab on the mouse repeat3 enhancer. The current study observed that this conclusion does not hold on another target sequence, the EVIII enhancer of CG11339, indicating that the previous conclusion could reflect a specialisation of Lab activity with regard to its autoregulation, rather than a general feature that distinguishes the mode of Lab DNA binding from that of other Hox proteins (Ebner, 2005).

The Hox-binding selectivity model also implies that a given Hox/Exd complex should recognize a consensus nucleotide sequence in downstream target genes; owing to the lack of well characterised Hox target sequences, this still remains to be experimentally validated. The sequence responsible for Lab-mediated regulation of CG11339 is TGAT[CA]ATTA, which diverges from the TGAT[GG]ATTG site mediating lab autoregulation, at the two central positions that are predicted to define the choice of the Hox protein recruited with Exd. The fact that Lab can recognize target sequence differing at the central NN nucleotide is also observed upon mutation of these nucleotides from GG to TA in the lab550 autoregulatory enhancer. Thus, Lab can form a complex with Exd and activates transcription in vivo on at least three sequences that differ with regard to the identity of the central NN nucleotides: GG in repeat3, TA in the mutated lab enhancer and CA in CG11339 (Ebner, 2005).

Since altering the GG identity of the central NN nucleotides in repeat3 to TA or TT alleviates Lab/Exd complex assembling, the readout of the nucleotide identity at the central NN positions most probably depends upon neighbouring nucleotides that are different in repeat3, lab48/95 and CG11339. Examination of the three sites shows that the Exd half sites are conserved, while the Hox half site differs at the most 3' end. In support for a role of nucleotides lying in the Hox half site in the readout of the identity of the central NN nucleotides, it was found that loss of Lab/Exd complex assembly following mutations at the 3' end of the Hox half site can be reversed by modifying the two central positions. This compensatory effect might result from subtle changes in contacting helix 3 of the HD, which in turn might modify the sequence requirement at the central NN position for efficient Lab/Exd recruitment. The importance of the Hox half site 3' end sequences is further supported by the observation that Scr and Dfd both bind in vitro and act in vivo on a prototypical Hox/Exd site that shares a TA at the central NN position, but differs in the identity of nucleotides at the 3' end of the Hox half site: GA for Dfd and CT for Scr (Ebner, 2005).

Variability in the sequence and spacing of the Hth-binding site might also influence the choice of the Hox protein that will preferentially form a complex with Exd and Hth. In any case, this study clearly shows that one Hox/Exd complex can recognize divergent sequences in two different regulated target genes. Although the two central nucleotides play a crucial role in assembling a specific Hox/Exd complex, added complexity to the Hox-binding selectivity model needs to be considered, and the nature of these two base pairs will not necessarily predict which Hox protein will selectively bind with the co-factor Exd (Ebner, 2005).

Finally, the data might also open perspectives on the mechanisms underlying the establishment of complex and distinct transcriptional patterns downstream of Hox genes. Hox transcription factors are usually expressed in broad domains, yet downstream target genes are often activated or repressed only in part of the Hox expression domain. It has previously been shown that regulatory regions of downstream target genes integrate signalling inputs, which provides additional positional information to restrict downstream target gene activation. These observations highlight the importance of the environment of the Hox/Exd-binding sequence in mediating transcriptionally distinct outputs. This study shows that Lab responsive enhancers that bear Lab/Exd-binding sites drive distinct expression patterns, both with regard to spatial and temporal characteristics. It suggests that in addition to environmental cues, the identity of the Hox/Exd sites might also be instructive (Ebner, 2005).

Ubx structure and interaction with Exd

While testing the functions of deletion mutants in the Hox protein Ultrabithorax (Ubx), it was found that the embryonic repression function of Ubx on Distal-less transcription in limb primordia is highly concentration dependent. The steep sigmoidal relationship between in vivo Ubx concentration and Distal-less repression is dependent on the Ubx YPWM motif. This suggests that Ubx cooperatively assembles a multi-protein repression complex on Distal-less regulatory DNA with the YPWM motif as a key protein-protein interface in this complex. Deletion mutants also provide evidence for a transcriptional activation domain in the N-terminal 19 amino acids of Ubx. This proposed activation domain contains a variant of the SSYF motif that is found at the N termini of many Hox proteins, and is conserved in the activation domain of another Hox protein, Sex combs reduced. These results suggest that the N-terminal region containing the SSYF motif has been conserved in many Hox proteins for its role in transcriptional activation (Tour, 2005).

The deletion of the Ubx YPWM region has little detectable effect on the transcriptional activation of the dpp and tsh genes. Because exd genetic function is required for normal levels of dpp and tsh activation in Ubx-expressing cells, this result is difficult to reconcile with a simple model in which the YPWM motif is required for Exd recruitment to activation target sites in dpp and tsh enhancers. However, it is consistent with studies that tested the effect of YPWM mutations on the activation abilities of the Labial and Abd-A Hox proteins in embryos. A YPWM to AAAA mutant of Labial is a more potent activator than wild-type Labial protein of a sequence derived from the Hoxb1 autoregulatory region, whereas a YPWM-to-AAAA mutant of Abd-A converted this protein from a repressor into an activator of dpp transcription. In addition, this YPWM mutation has no effect on the activation function of Abd-A on wingless. The ability of Labial and Abd-A YPWM mutants to retain their transactivation functions is correlated with their ability to bind Exd in vitro in a YPWM-independent fashion. The YPWM-independent interactions between Hox proteins and Exd can be mediated by Hox homeodomains and the C-terminal regions (Tour, 2005).

Since the Ubx-responsive elements from dpp and tsh loci possess a mixture of Ubx monomer and Ubx-Exd heterodimer-binding sites, possible reasons for the ability of the Ubx YMPM deletion mutant to activate these downstream target genes are: (1) Hox activation of target genes often involves a mixture of Exd-dependent and Exd-independent functions; (2) removal of the YPWM motif does not completely abolish Exd-Ubx binding interactions, and (3) the YPWM apparently serves other functions besides binding Exd in the context of developing embryos (Tour, 2005).

Distinct functions of homeodomain-containing and homeodomain-less isoforms encoded by homothorax

The homothorax (hth) gene of Drosophila is required for executing Hox functions, for head development, and for forming the proximodistal (PD) axis of the appendages. Alternative splicing of hth generates two types of protein isoforms, one that contains a DNA-binding homeodomain (HthFL) and one that does not contain a homeodomain (HDless). Both types of Hth isoforms include the evolutionarily conserved HM domain, which mediates a direct interaction with Extradenticle (Exd), another homeodomain protein. Although both HthFL and HDless isoforms of Hth can induce the nuclear localization of Exd, they carry out distinct sets of functions during development. Surprisingly, many of hth’s functions, including PD patterning and most Hox-related activities, can be executed by the HDless isoforms. In contrast, antennal development shows an absolute dependency on the HthFL isoform. Thus, alternative splicing of hth results in the generation of multiple transcription factors that execute unique functions in vivo. It is further demonstrated that the mouse ortholog of hth, Meis1, also encodes a HDless isoform, suggesting that homeodomain-less variants of this gene family are evolutionarily ancient (Noro, 2006: full text of article).

hth includes 16 annotated exons distributed over >100 kb of genomic DNA. All functionally characterized isoforms of hth include both the HM domain, encoded by exons 2–6, and the HD, encoded by exons 11–13. In addition, hth encodes at least two additional alternatively spliced variants that have an intact HM domain but no HD. Both alternatively spliced mRNAs code for two almost identical HM-containing proteins that are largely derived from the first six coding exons. Both of these HDless isoforms have an additional 24 amino acids at their C termini encoded by alternate exons. One of these variants (the 7' isoform) uses an alternative exon 7 (exon 7'). Sequence comparisons between D. melanogaster and Anopheles gambiae hth genes enabled identification of a second hth splice variant that is also missing the HD. This isoform (the 6' isoform) is generated when the splice site at the 3' end of exon 6 is not used, generating an extended ORF. The existence of both 6' and 7' isoforms in vivo was confirmed by sequencing ESTs and performing RT–PCR on mRNA isolated from embryonic and larval tissues. The presence of 6' and 7' isoforms raised the possibility that HDless variants of Hth might carry out distinct functions, suggesting a functional diversification of the hth gene that depends on alternative splicing (Noro, 2006).

This study addresses the functional relevance of alternatively spliced isoforms of Hth, a transcription factor involved in a wide variety of developmental programs that are critical for the construction of the D. melanogaster body plan. Analysis of hth100-1 mutant tissues during both embryonic and larval stages have demonstrated a strict requirement for the HD in a surprisingly small subset of developmental functions, such as the instruction of antennal identity and the correct patterning of the wing hinge. In contrast, partial loss of function of HDless forms, resulting either from siRNA injection against the 6' and 7' isoforms or from the ectopic expression of HthFL, suggest that these forms carry out crucial functions in vivo. Intriguingly, the data further suggest that HthFL is apparently unable to substitute for at least a subset of HDless functions. This idea rests primarily on the observation that 6' + 7' siRNA-injected embyros exhibit hth loss-of-function phenotypes yet still express HthFL. However, the possibility that the injected siRNAs might have off-target effects cannot be excluded, even though the specificity of the observed phenotypes suggests that it is unlikely. In future experiments, it may be possible to more definitively test this idea by generating hth mutant alleles that are unable to express the 6' and 7' isoforms (Noro, 2006).

In contrast, the generally weak phenotypes observed in hth100-1 embryos and adults support the hypothesis that HDless/Exd dimers work as bona fide transcription factors that are essential for the correct regulation of many hth-dependent functions. In some respects, these findings are reminiscent of reports showing that an artificially truncated and HDless version of the segmentation protein Fushi tarazu (Ftz) retains many of the activities of full-length Ftz. These earlier findings provide additional support to the idea that HDless forms of some homeoproteins retain biological activity, probably due to their ability to assemble stable protein complexes in vivo. What is unique to the current results is that hth normally expresses HDless isoforms and that there is a division of labor between HDless and HthFL isoforms. This is best exemplified by the finding that HthFL isoforms are essential for antennal development but largely dispensable for proximal leg development. Based on these observations, it is suggested that distinct Hth/Exd dimers may bind to partially overlapping sets of target genes in vivo, and that the presence or absence of the DNA-binding HD expands the range of target genes that Hth/Exd can select and regulate (Noro, 2006).

Hth is composed of two conserved modules: the HM domain that mediates an interaction with Exd and the DNA-binding HD. Since the 6' and 7' isoforms do not have a HD, they are unable to directly interact with DNA. However, the presence of the HM domain allows them to complex with Exd, whose HD can mediate DNA binding, as demonstrated by the formation of cooperative HDless/Exd/Hox complexes on the fkh[250] and lab48/95 elements. Consistent with these results, Meis has also been shown to form trimeric complexes with Pbx and Hox without binding directly to the DNA (Noro, 2006).

The absence of the Hth HD has several implications for the transcriptional properties of HDless/Exd complexes. (1) It is likely that HDless/Exd and HthFL/Exd complexes have distinct DNA-binding specificities because the latter complex contains two HDs, while the former contacts DNA exclusively through Exd's HD. It is imagined that the two types of complexes regulate partly overlapping sets of target genes by decoding different cis-regulatory architectures, possibly in the same cells. For example, the HthFL isoform appears to be unable to carry out some hth functions since the 6' + 7' siRNA-injected embryos exhibit hth loss-of-function phenotypes. This observation suggests that the presence of the HD might be incompatible with a subset of the cis-regulatory architectures that bind HDless/Exd. (2) HDs can also be protein interactions motifs, raising the possibility that the absence of the Hth HD from HDless/Exd could influence its ability to contact other transcription factors, coactivators, and/or corepressors. The Exd and Pbx TALE HD mediate direct interactions with Hox factors, and with the HD-containing transcription factor Engrailed (En). The HD of Hth, which is also of the TALE family, is also likely to interact with other transcription factors, including Hox proteins. Thus, through alternative splicing, the modular architecture of Hth is exploited to produce unique transcription factor complexes that are likely to have distinct protein and DNA-binding properties (Noro, 2006).

Given that HthFL and HDless isoforms have some unique functions during development, it is tempting to suggest some generalizations about which functions require the Hth HD and which do not require this domain. Insect body plans are made up of repeated units that develop into diverse body parts in the adult due to the activity of selector genes, transcription factors that instruct morphological identities by regulating unique sets of target genes. Legs and antennae in Drosophila represent an example of serially homologous appendages that develop from a leg-like ground-state in response to different selector activities: Hox factors select for legs while Hth/Exd select for antenna. The demonstration that a hth100-1 mutant antenna is completely transformed toward a ground-state leg-like appendage demonstrates that the antennal selector function of Hth is absolutely dependent on its HD (Noro, 2006).

In contrast to its antennal selector role, the data suggest that the Hth HD is largely dispensable for at least some of the Hox-cofactor functions of Hth/Exd. This surprising conclusion is based in part on the cuticle phenotypes of hth100-1 and 6' + 7' siRNA-injected larvae. Specifically, hth100-1 larvae show no or very weak transformations of segmental identity, whereas 6' + 7' siRNA-injected larvae show clear posterior-directed transformations. Consistently, hth100-1 mutant embryos still express two directly activated Hox/Exd/Hth targets, fkh[250] and lab550. Repression of Distalless (Dll), which also requires direct Hox/Exd/Hth input, also occurs normally in hth100-1 mutant embryos. Thus, from these diverse observations it is concluded that the Hth HD is largely dispensable for the Hox-cofactor function of Hth/Exd. However, it is noted that there are exceptions to this generalization. Although activation of lab550 does not require the Hth HD, activation of a weakened derivative of this enhancer, lab48/95, does require the Hth HD. Similarly, mutation of the Hth-binding site in the Dll repressor element, DllR, results in weak abdominal derepression. Taken together, these data suggest that the transcription factor complexes binding to the lab and Dll regulatory elements contain the Hth HD, but that its presence is only required when the activity of these elements is compromised or weakened (Noro, 2006).

A third well-characterized function of Hth/Exd is its role in the establishment of the PD axis in both ventral (legs) and dorsal (wings and halteres) appendages. The experiments suggest that the Hth HD is not required for PD axis formation or for specifying proximal identities in the legs. In the wing, the Hth HD is also apparently dispensable for forming a correct PD axis (in particular, repression of wg at the DV boundary) but is partially required for specifying proximal (hinge) fates. Notably, both functions in which the Hth HD is largely dispensable (PD axis formation and Hox cofactor activity) appear to be evolutionarily ancient. Like Hth/Exd, Meis/Pbx are Hox cofactors and are also instrumental for establishing the PD axis of the vertebrate limb. In contrast, the antennal-specifying activity of Hth/Exd, which requires the Hth HD, is not known to have a vertebrate correlate. Thus, it is tempting to speculate that, in Drosophila, the Hth HD is more essential for executing evolutionarily recent, invertebrate-specific Hth functions and plays a less crucial, supplemental role in evolutionarily ancient Hth/Meis activities. Consistent with the idea that the HDless activities of Hth are ancient is the identification of an analogous HDless isoform made by Meis1 in Mus musculus, which underscores the functional relevance of HDless isoforms for the fulfillment of Hth/Meis-dependent functions during both invertebrate and vertebrate development. Interestingly, Prep2, another vertebrate gene related to hth, also appears to encode both HD-containing and HDless isoforms. Although the functions of these isoforms are not known, the results suggest that there may be a similar division of labor of HD-containing and HDless isoforms encoded by the Meis1 and Prep2 genes of vertebrates (Noro, 2006).

In summary, these results strongly support the idea that alternative splicing of Hth and its vertebrate orthologs is an evolutionarily conserved mechanism to expand the architectural diversity of Hth/Exd and Meis/Pbx transcriptional complexes. It is proposed that by excluding or including the HD of Hth, Hth/Exd complexes acquire distinct DNA-binding and protein interaction properties, which allow them to regulate different sets of target genes and execute unique developmental programs in vivo (Noro, 2006).

A unique Extradenticle recruitment mode in the Drosophila Hox protein Ultrabithorax

Hox transcription factors are essential for shaping body morphology in development and evolution. The control of Hox protein activity in part arises from interaction with the PBC class of partners, pre-B cell transcription factor (Pbx) proteins in vertebrates and Extradenticle (Exd) in Drosophila. Characterized interactions occur through a single mode, involving a short hexapeptide motif in the Hox protein. This apparent uniqueness in Hox-PBC interaction provides little mechanistic insight in how the same cofactors endow Hox proteins with specific and diverse activities. This study identified in the Drosophila Ultrabithorax (Ubx) protein a short motif responsible for an alternative mode of Exd recruitment. Together with previous reports, this finding highlights that the Hox protein Ubx has multiple ways to interact with the Exd cofactor and suggests that flexibility in Hox-PBC contacts contributes to specify and diversify Hox protein function (Merabet, 2007).

The current view of Hox-PBC interactions is that they all occur through a single mode, involving a short hexapeptide (HX) motif. The importance of the Hox HX motif in mediating interaction with PBC proteins is extensively supported by its requirement in in vitro interaction assays and by crystallographic studies that showed that the HX provides most if not all major contacts. In contrast, in vivo functional support for a role of the HX in mediating interaction with PBC proteins is still limited, mainly because effects of HX mutations during development have only been assessed for two vertebrate Hox proteins, Hoxa-1 and Hoxb-8, and for three Drosophila proteins, Labial (Lab), Ubx, and Abdominal-A (AbdA) (Merabet, 2007 and references therein).

Mutation of the HX in Hoxb-8 results in dominant phenotypes, which are at present difficult to interpret with regard to Pbx recruitment. Hoxa-1 HX mutation mimics Hoxa-1 loss of function, including defects in the hindbrain that could relate to loss of Pbx recruitment because inactivation of Pbx2 and Pbx4 in the zebrafish affects hindbrain patterning. However, addressing in vertebrates whether phenotypes resulting from HX mutations are consequences of defects in Pbx interaction will require examination of combinations of Hox-1 paralogous and Pbx gene mutations. In Drosophila, mutation of the HX in Lab, the only representative of Hox-1 class genes, results in an hyperactive protein when assayed for its potential to activate transcription through an evolutionarily conserved Hoxb-1 autoregulatory element. This hyperactivity results from the loss of an inhibitory action of the HX on Lab DNA binding. In this context, it was proposed that HX-mediated recruitment of Extradenticle (Exd) acts to mask the DNA-binding inhibitory activity of the HX motif (Merabet, 2007 and references therein).

Although the Hoxa-1 and Lab studies support, yet not exclusively, a role of the HX in mediating recruitment of PBC class proteins during development, work on Drosophila Ubx and AbdA has provided evidences for HX-independent mode of Exd recruitment. Regarding Ubx, a truncated protein lacking N-terminal sequences (including the HX) was shown in vitro to retain Exd recruitment potential and to interact weakly with Exd in yeast two-hybrid assays. More specifically, mutation of the HX does not affect the capacity to recruit Exd on a Hox/Exd consensus target sequence in vitro and to repress in an Exd-dependent manner the limb-promoting gene Distalless (Dll), which has served as a paradigm to study Hox-Exd interactions. Concerning AbdA, the HX-deficient protein was shown to recruit Exd on the Dll regulatory element that mediates Dll repression, consistent with its retained ability to repress Dll (Merabet, 2007).

Thus, Hox-PBC interactions are not limited to HX-mediated interactions, highlighting that another Hox protein motif, yet to be identified, may also assume this function. The Ubx C terminus [sequences downstream of the homeodomain (HD), UC], important for Ubx segment identity functions and shown to increase the ability of the Ubx HD to associate with Exd in yeast two-hybrid assays, harbors an 8-aa peptide previously termed UbdA as well as a QA repression domain responsible for changes in Ubx activity. Although evolutionarily conserved, the precise function of the UbdA motif, only present in Ubx and AbdA proteins but absent from any other Drosophila Hox protein, is not known. This work reports on the function of the UbdA motif in the context of the Ubx protein. Because this motif is specific to Ubx and AbdA, which share the HX-independent mode of Exd recruitment, the analysis focused on the possible implication of this motif in mediating Exd recruitment (Merabet, 2007).

The results strongly support that the UbdA motif mediates Exd recruitment by the Ubx protein. This finding is first established by the requirement of the motif for Exd recruitment in the process of Dll regulation: mutation of the motif impairs the capacity of Ubx to mediate interaction with Exd on Dll regulatory sequences in vitro, which correlates with the reduced ability of UbxUbdA to perform Exd-dependent repression of Dll in vivo. Evidence is provided that mutation of the UbdA motif does not result in a globally defective protein: the UbdA mutated protein still binds DNA with appropriate affinities as a monomer, still represses the wing promoting genes dSRF, sal, and vg and still activates the dpp target gene in the visceral mesoderm. Thus, mutating the UbdA motif selectively affects a subset of Ubx functions. Importantly, the conclusion that the UbdA motif mediates Exd recruitment is also supported by the demonstration that the motif provides de novo Exd recruitment potential to a Hox protein that has been rendered deficient for this function. This finding is shown both in vitro by the potential of the motif to confer Exd recruitment to Antp on a Hox/Exd consensus sequence and on a cis-regulatory sequence of the Antp/Exd target gene tsh, and in vivo by its potential to restore Exd-dependent activation of tsh. Complexity in Hox-Exd Interactions (Merabet, 2007).

The Ubx protein provides a so far unique situation wherein two identified protein motifs within the same Hox protein have the potential to perform the recruitment of the Exd cofactor, which raises the question of whether these two motifs are effectively used for Exd recruitment by Ubx. Previous work has shown that an HX-deficient Ubx protein was altered in its segment identity specification: whereas Ubx specifies A1 segment identity, the mutated form specifies A2-like identity. Interestingly, in a context deficient for zygotic Exd contribution, Ubx also specifies A2-like identity, suggesting that the HX motif is required for Exd-dependent A1 specification. These observations support that within this context, the HX is the motif used to perform Exd recruitment, although definitive support awaits characterization of Ubx-Exd interaction on a Ubx downstream target gene involved in segment identity specification. Considering the finding that the UbdA motif mediates Exd recruitment in the process of Dll regulation, it is proposed that depending on the developmental context, i.e., on the target gene regulated, Ubx uses different protein motifs for Exd recruitment. The contextual (gene-specific) use of the HX and UbdA protein motifs introduce a first level of complexity in Ubx-Exd interactions (Merabet, 2007).

A second level of complexity in the Ubx-Exd relationship is illustrated by the regulation of the dpp target gene. In this case, it was found that neither the HX nor the UbdA motif was required for Exd-dependent activation by Ubx. The possibility that these two motifs were acting in a redundant way was excluded by the observation that a Ubx protein mutant for both motifs still activates dpp. Thus, other protein motifs, yet to be identified, could confer an additional mode of Exd interaction, further increasing the diversity by which Ubx could contact the Exd cofactor. Alternatively, the dispensability of the HX and UbdA motifs for dpp activation may also suggest that Ubx/Exd contacts are not required. The latter hypothesis is supported by the existence in dpp regulatory regions of Exd-binding sites that are not closely associated to Hox-binding sequences and by the previous observation that Exd can improve Ubx monomer binding to dpp regulatory sequences in a manner that does not require the formation of a Ubx-Exd-DNA tripartite complex. In any case, the regulation of dpp suggests further complexity in the Ubx-Exd relationship, which, by extension, highlights that the functional interplay of Hox-PBC proteins is likely to be more diverse than the current view (Merabet, 2007).

Although previous studies showed that HX-deficient Hox proteins retain the capacity to interact with Exd and to mediate Exd-dependent functions, motifs responsible for alternative modes of interaction were not identified. This work identifies a so far unique HX-alternative mode of PBC recruitment, introducing the notion of flexibility in Hox-PBC contacts. This interaction mode was not anticipated from previous crystallographic studies because the truncated Ubx protein used was lacking the UbdA motif. Given the divergence of the primary sequences of the HX and UbdA motifs, their distinct location in the protein, and the absence of functional redundancy, the UbdA- and HX-mediated interaction modes are likely to be structurally distinct (Merabet, 2007).

These findings have also implications with regard to Hox protein diversity and specificity. Flexibility in Hox-PBC interactions allows addressing of the issue of diversity from a mechanistic point of view: depending on the motif involved in the interaction, which likely relies on the target sequence, the Hox-PBC complex may adopt different conformations, which in turn set structural bases for distinct activities. This process therefore provides cues to explain how diversity can be generated through qualitatively distinct interaction modes involving the same protein partners. Furthermore, because the UbdA motif is only found in Ubx and AbdA, it likely endows these two proteins with a specific Exd interaction mode. This mode may serve to distinguish Ubx and AbdA from other Drosophila Hox proteins, therefore providing basis for Hox protein specificity. Finally, this study questions whether additional HX-independent modes of PBC interaction exist. It was reported previously that the HX-deficient Lab protein retains Exd interaction potential and in vivo Exd-dependent activity. As Lab does not bear a UbdA motif, it supports further flexibility in Hox-PBC interaction. Addressing the issue of diversity in Hox-PBC interaction thus appears as a necessary step to understand the mechanisms underlying Hox protein activity in development and evolution (Merabet, 2007).

Dissecting the functional specificities of two Hox proteins

Hox proteins frequently select and regulate their specific target genes with the help of cofactors like Extradenticle (Exd) and Homothorax (Hth). For the Drosophila Hox protein Sex combs reduced (Scr), Exd has been shown to position a normally unstructured portion of Scr so that two basic amino acid side chains can insert into the minor groove of an Scr-specific DNA-binding site. This study provides evidence that another Drosophila Hox protein, Deformed (Dfd), uses a very similar mechanism to achieve specificity in vivo, thus generalizing this mechanism. Furthermore, it was shown that subtle differences in the way Dfd and Scr recognize their specific binding sites, in conjunction with non-DNA-binding domains, influence whether the target gene is transcriptionally activated or repressed. These results suggest that the interaction between these DNA-binding proteins and the DNA-binding site determines the architecture of the Hox-cofactor-DNA ternary complex, which in turn determines whether the complex recruits coactivators or corepressors (Joshi, 2010).

Previous work on Scr's ability to specifically regulate its target gene, fkh, revealed that the N-terminal arm of its homeodomain and preceding linker region are positioned in such a manner as to allow the insertion of two basic side chains into the minor groove of the target DNA, fkh250 (Joshi, 2007). Importantly, the correct positioning of these residues depends on an interaction between Scr's YPWM motif and the cofactor Exd. This study shows that an analogous mechanism is required for Dfd to bind productively to a Hox-Exd-binding site in the EAE element and to activate EAE-lacZ in vivo. Specifically, it was found that Dfd's YPWM motif is required for cooperative binding to EAE's site I in vitro, and for executing Dfd-specific functions in vivo. Like Scr, Dfd has the same two basic residues -- a histidine (likely to be protonated when bound to DNA) and an arginine -- at the equivalent positions relative to its YPWM motif and homeodomain. Moreover, these residues are also required for Dfd to execute its specific functions in vivo. Thus, the activation of fkh by Scr and the activation of Dfd by Dfd appear to use analogous mechanisms, whereby linker and N-terminal arm residues are used to bind paralog-specific binding sites in an Exd-dependent manner (Joshi, 2010).

The YPWM-to-YPAA mutation severely impaired Dfd's ability to carry out its specific in vivo functions, such as activation of EAE-lacZ and production of cirri. Thus, the YPWM motif of Dfd is critical for Dfd function in vivo. This situation contrasts with other apparently more complex scenarios. For example, mutation of the YPWM motif of the Hox protein Ultrabithorax (Ubx) did not significantly impair some of its in vivo functions. In this case, it appears that other sequence motifs, in particular a domain C-terminal to the Ubx homeodomain, are important for Ubx to carry out its specific functions in vivo. These Ubx sequences also appear to help recruit Exd to DNA, and therefore may be used for binding site selection in conjunction with YPWM at a subset of Ubx target-binding sites. Interestingly, a sequence motif immediately C-terminal to Dfd's homeodomain also plays a role in in vivo specificity, although its impact on DNA binding has not been examined. As these sequences are still present in DfdScrSMδ23, it may explain why this chimera retains some Dfd-specific functions, such as the formation of cirri and ability to activate EAE-lacZ. The picture that emerges from all of these data is that Hox proteins may use different motifs to interact with cofactors such as Exd, depending on the specific in vivo function and target gene being regulated (Joshi, 2010).

In general, the sequences surrounding Hox YPWM motifs and the N-terminal arms of their homeodomains are highly conserved, from invertebrates to vertebrates, in a paralog-specific manner (Joshi, 2007). Thus, based on the results presented in this study, it is hypothesized that these sequences, which are referred to as Hox specificity modules, may in general be used for the recognition of specific DNA-binding sites in a cofactor-dependent manner. In the case of Scr binding to fkh250, an X-ray crystal structure revealed that the histidine and arginine side chains recognize an unusually narrow minor groove that is an intrinsic feature of the fkh250-binding site. Without the benefit of a Dfd-Exd-site I crystal structure, it cannot be know with certainty if Dfd's His-15 and Arg3 also read the shape of a narrow minor groove. However, the fact that the same two basic residues are required for both Scr and Dfd suggests the possibility that this is the case for Dfd binding to EAE site I as well (Joshi, 2010).

DfdScrSMδ23, which has the specificity module of Scr in place of Dfd's, exhibited clear Scr-like functions in vivo, as assayed by fkh250-lacZ activation and larval cuticle transformation. Other attempts to swap Hox specificities by generating chimeric Hox proteins have had variable success. For example, when the linker and N-terminal arm of Scr is used to replace the equivalent region of Antennapedia (Antp), the chimera behaved like Scr. This finding supports the importance of specificity modules in conferring Hox specificity. When the homeodomain and C-terminal region of Ubx were replaced by the equivalent domains from Antp, the chimera behaved like Antp, suggesting that the identity of the linker region may not be critical in all cases. Other Hox chimeras have generated less clear changes of specificity. For example, chimeras between Ubx and Dfd generated a cuticle phenotype that was dissimilar to that produced by either parent protein. Similarly, a chimera between Ubx and Abd-B had novel properties that were unlike those produced by either parent protein. It is noteworthy that the cleanest changes in specificity occurred when the chimera was generated between Hox genes that are adjacent to each other in the Hox complex. This correlation may be due to the fact that adjacent Hox genes are more similar to each other, both in sequence and in function, than nonadjacent Hox genes. This higher degree of similarity is likely a consequence of how these genes are thought to have duplicated during evolution (Joshi, 2010).

Previous work on the regulation of fkh by Scr, the reporter gene used to study the activity of the Exd-Scr-binding site had a multimerized version of the minimal 37-bp fkh250 element. In contrast, in the work described in this study, an intact regulatory element from the Dfd gene was characterized, revealing significantly more complexity. In particular, the 570-bp modC element contains a single 'classical' Exd-Hox composite site, but also four additional Dfd sites and several additional Exd-Hth-binding sites. Mutagenesis studies suggest that all of these inputs are important for the full activity of this enhancer. Also noteworthy is that there are additional Dfd-Exd-binding sites in the larger 2.7-kb EAE element that, in principle, could also be used in vivo. Thus, the picture that emerges from this analysis is that native enhancer elements may use a combination of classical Exd-Hox-binding sites together with additional arrangements that may not always conform to the classical spacing of the Exd and Hox half-sites. This picture raises the question of how the linker and N-terminal arm residues are positioned correctly in these nonclassical arrangements. The answer may lie in the fact that, in vivo, the assembly of the complete multiprotein complex -- which is likely to include factors in addition to Dfd, Exd, and Hth -- promotes the recognition of Dfd-binding sites in ways that are not fully revealed by experiments that examine binding to individual or small groups of binding sites in isolation (Joshi, 2010).

Depending on the context, most transcription factors have the capacity to activate and repress transcription. In most cases, it is not understood how this choice is made. One established scenario is that other proteins that get recruited to an enhancer element determine the sign of the regulation. However, this type of model is not sufficient to explain the results presented in this study. The results suggest that the DNA-binding properties of the Exd-Hox complex influence the regulatory output of the bound protein-DNA complex. Deletion of two motifs (γ23) from the N-terminal region of DfdScrSM converted this protein from a repressor of fkh250-lacZ to an activator of fkh250-lacZ, while deletion of the same motifs from DfdWT did not change the regulatory output: The protein retained its ability to repress fkh250-lacZ. The only difference between Dfdγ23 (represses) and DfdScrSMγ23 (activates) is the specificity module, and the only difference between DfdScrSMγ23 (activates) and DfdScrSM (represses) is the presence or absence of motifs 2 and 3. These results imply that the relevance of motifs 2 and 3, which are far from the DNA-binding domain, depends on the identity of the specificity module. These findings lead to a suggestion that the DNA-binding site, together with how it is read by the specificity module, plays an important role in determining the overall conformation of the Hox-Exd complex, which eventually determines whether there will be recruitment of a coactivator or corepressor. This idea fits well with a DNA allostery model that was supported recently by cell culture experiments with the glucocorticoid receptor. In these experiments, it was discovered that small differences in the DNA-binding site lead to differences in conformation and the degree of transcriptional activation. This study extends this idea by showing that Hox proteins with different specificity modules, and therefore with slightly different DNA recognition properties, result in unique regulatory outputs in an in vivo context. Furthermore, in these experiments, a complete change was observed in the sign of the regulation from repression to activation, instead of a more subtle change of activation amplitude. Thus, the transcriptional output of a Hox-cofactor complex depends both on the ability of these complexes to bind to their binding sites with high specificity, in part by reading structural features of the DNA, and on the three-dimensional architecture of the bound complex, which is a consequence of both protein-DNA and protein-protein interactions. An important goal for the future will be to use structural biology methods to see how different Hox specificity modules result in distinct conformations of Exd-Hox complexes (Joshi, 2010).

Hox proteins display a common and ancestral ability to diversify their interaction mode with the PBC class cofactors

Hox transcription factors control a number of developmental processes with the help of the PBC class proteins. In vitro analyses have established that the formation of Hox/PBC complexes relies on a short conserved Hox protein motif called the hexapeptide (HX). This paradigm is at the basis of the vast majority of experimental approaches dedicated to the study of Hox protein function. This study questioned the unique and general use of the HX for PBC (Extradenticle in Drosophila) recruitment by using the Bimolecular Fluorescence Complementation (BiFC) assay. This method allows analyzing Hox-PBC interactions in vivo and at a genome-wide scale. It was found that the HX is dispensable for PBC recruitment in the majority of investigated Drosophila and mouse Hox proteins. HX-independent interaction modes are uncovered by the presence of Meis class cofactors, a property which was also observed with Hox proteins of the cnidarian sea anemone Nematostella vectensis. Finally, it was revealed that paralog-specific motifs convey major PBC-recruiting functions in Drosophila Hox proteins. Altogether, these results highlight that flexibility in Hox-PBC interactions is an ancestral and evolutionary conserved character, which has strong implications for the understanding of Hox protein functions during normal development and pathologic processes (Hudry, 2012).

In Drosophila, it is interesting to note that the only Hox proteins which were described to achieve their regulatory activities in absence of HD-containing isoforms of Hth were Lab and Scr. Accordingly, the formation of Lab/Exd/Hth or Scr/Exd/Hth complexes in vitro is more sensitive to the DNA-binding of Exd than to the DNA-binding of Hth. On the contrary, Ubx and AbdA are more sensitive to the loss of Hth DNA-binding for trimeric complex assembly in vitro and for regulating their respective physiological target enhancers in vivo. These data suggest that Meis DNA-binding could be more critical for Hox proteins displaying alternative PBC interaction modes than for Hox proteins displaying a unique HX-dependent interaction mode (Hudry, 2012).

The understanding of the molecular mechanisms by which Meis proteins could influence Hox-PBC interactions will require the resolution of Hox/PBC/Meis/DNA structures. It is speculated that Hox-PBC interactions that are strongly remodeled by Meis likely rely on PBC-Meis and Hox-Meis interactions. Although the formation of PBC/Meis complexes is well established, interactions between Hox and Meis proteins were rarely described. Meis proteins can form cooperative DNA-binding complexes with vertebrate Hox proteins of posterior paralog groups, but interactions with more anterior Hox proteins have only been described in a DNA-binding-independent context. In an EMSA experiments, no Hox-Meis DNA-binding complex was observed, except with the HX-mutated form of AbdB. Hox-Meis interactions could thus require the presence of the PBC cofactor to be stabilized, eventually leading to alternative Hox-PBC contacts. In that 'ménage-à-trois,' the existence of Hox-PBC and Hox-Meis interactions have the advantage to expand the range of molecular strategies that could be used by Hox proteins to assemble into a trimeric complex (Hudry, 2012).

Distinct genetic requirements for BX-C mediated specification of abdominal denticles. Dev Dyn. PubMed ID: 24155218

Hox genes encode transcription factors playing important role in segment specific morphogenesis along the anterior posterior axis. Most work in the Hox field aimed at understanding the basis for specialised Hox functions, while little attention was given to Hox common function. In Drosophila, genes of the Bithorax complex [Ultrabithorax (Ubx), abdominalA (abdA) and AbdominalB (AbdB)] all promote abdominal identity. While Ubx and AbdA share extensive sequence conservation, AbdB is highly divergent, questioning how it can perform similar functions than Ubx and AbdA. This study investigated the genetic requirement for the specification of abdominal-type denticles by Ubx, AbdA and AbdB. The impact of ectopic expression of Hox proteins in embryos deprived for Exd as well as for Wingless or Hedgehog signaling involved in intrasegmental patterning was analyzed. Results indicated that Ubx and AbdA do not require Exd, Wg and Hh activity for specifying abdominal-type denticles, while AbdB does. These results support that distinct regulatory mechanisms underlie Ubx/AbdA and AbdB mediated specification of abdominal-type denticles, highlighting distinct strategies for achieving a similar biological output. This suggests that common function performed by distinct paralogue Hox proteins may also rely on newly acquired property, instead of conserved/ancestral properties (Sambrani, 2013b).

This study relies on a gain of function strategy, scoring phenotypes in the thorax, where none of these three BX-C proteins are expressed, but where Exd, Wg, and Hh are expressed. This allows circumventing the difficulty resulting from the incapacity to unambiguously identify posterior most abdominal segments, where AbdB acts, and avoid complications in interpreting results that would arise from cross regulation between BX-C genes in the abdomen. However the approach also questions whether the conclusion of this study applies for BX-C proteins activity in their endogenous expression domains. Regarding Exd requirements, maternal and zygotic loss results in abdominal segment fusion, where segments are fused by pairs: A1/A2-3/A4-5/A6-7/A8. Although denticle belts are highly disorganized, denticles are clearly of abdominal type in anterior abdominal segments. When present in A8, the belt of denticles is very much reduced, and in most embryos is absent. This indicates that Ubx and AbdA do not require Exd for the specification of abdominal-type denticles in their endogenous expression domain, while AbdB does. The complete segment fusion resulting from loss of Wg and Hh signaling make it impossible to unambiguously identify the A8 segments, and therefore does not allow addressing if AbdB also requires Wg and Hh signaling in its endogenous expression domain. Denticles are found in continuous lawn, that encompasses most abdominal segments. Therefore denticles in the Ubx and AbdA expression domains are clearly of abdominal types, indicating that Ubx and AbdA do not require Wg and Hh signaling for the specification of abdominal-type denticles. Thus, whenever possible, resident BX-C Hox protein activity in loss of exd, wg and hh function are consistent with the conclusion raised in the gain of function approach, further supporting that Ubx/AbdA and AbdB use distinct regulatory mechanisms for achieving a common function (Sambrani, 2013b).

Such a conclusion was recently reached by studying the molecular mechanisms underlying repression of the limb-promoting gene Dll by Ubx and AbdB. It was shown that the cofactor requirement and intrinsic protein domain requirement for Ubx versus AbdB repression of Dll was distinct. Ubx represses Dll by binding DNA cooperatively with the Exd and Hth cofactors, which relies on the UbdA domain, a domain specific to Ubx and AbdA and located C-terminal to the HD (Sambrani, 2013a). Surprisingly, Ubx DNA binding is dispensable, probably due to cooperative binding to DNA with Exd and Hth, DNA binding proteins that likely compensate for Ubx loss of DNA binding. By contrast, AbdB represses Dll without the help of the Exd and Hth, and DNA binding of AbdB is strictly required for repression. It was further established that in specifying posterior spiracles and regulating empty spiracles expression, Exd/Hth antagonize AbdB activity, showing that the AbdB/Exd partnership depends on the biological context. Mechanisms at the origin of cooperativity/antagonism are still to be discovered (Sambrani, 2013b).

The present study corroborates the conclusion reached by the analysis of Dll repression by Ubx and AbdB and extends it in several ways: first by using a distinct Hox biological activity as functional readout; second by including in the analysis the AbdA Hox protein; and third by examining additional genetic requirements (Wg and Hh signaling). The work therefore provides further support for the view that distinct molecular strategies underlie an apparent unicity in BXC protein controlled biological function (Sambrani, 2013b).

Given the observation that Ubx and AbdA are very similar, sharing a highly conserved HD as well as additional protein domains such as the HX and UbdA motifs, while AbdB lacks these domains and has a highly divergent HD, it is not surprising that the genetic requirements are similar for Ubx/AbdA and distinct for AbdB. More unexpected was the finding that Ubx and AbdA do not require Exd for specifying abdominal-type denticles, while AbdB does. This indeed contrasts with the known and previously described Exd requirement for Ubx in A1 segment identity specification and Dll repression, and also contrasts with the dispensability of Exd for A8 segment identity specification, posterior spiracle specification and Dll repression (Sambrani, 2013a). This highlights that requirement of Exd for Hox activity depends on the specific function examined, rather than being a general and universal requirement (Sambrani, 2013b).

A salient difference between the central Ubx/AbdA and posterior AbdB Hox proteins is the mode of Hox DNA binding. Posterior paralogue Hox proteins have usually a stronger affinity for DNA when binding as monomer than central class Hox proteins. This difference mainly stems from the ability of posterior but not central class Hox proteins to make extensive contacts with the DNA backbone. These differences provide a frame to understand the requirement of Exd/Pbx cofactor for central class Hox proteins, which upon interaction with Hox proteins raises their DNA binding affinity. In the case of specification of abdominal-type denticles, the contribution of Exd is likely different, as required for AbdB and not Ubx/AbdA activity. This suggests that Exd may be involved in regulating the activity, rather than DNA binding, a function previously suggested in the regulation of Deformed Hox protein function (Sambrani, 2013).

In summary, this work together with the study of Dll repression by BX-C proteins highlights that distinct regulatory mechanisms and molecular strategies underlies common Hox protein functions. Thus while sequence divergence following gene duplication promotes functional divergence, it also generates novel gene regulatory mechanisms and molecular strategies that yet promotes a common biological output (Sambrani, 2013).

A flexible extension of the Drosophila Ultrabithorax homeodomain defines a novel Hox/PBC interaction mode

The patterning function of Hox proteins relies on assembling protein complexes with PBC proteins (the homeodomain proteins encoded by the Drosophila extradenticle (exd) and vertebrate pbx genes), which often involves a protein motif found in most Hox proteins, the so-called Hexapeptide (HX). Hox/PBC complexes likely gained functional diversity by acquiring additional modes of interaction. This study structurally characterized the first HX alternative interaction mode of the Hox protein Ultrabithorax based on the paralogue-specific UbdA motif and further functionally validates structure-based predictions. The UbdA motif folds as a flexible extension of the homeodomain recognition helix and defines Hox/PBC contacts that occur, compared with those mediated by the HX motif, on the opposing side of the DNA double helix. This provides a new molecular facet to Hox/PBC complex assembly and suggests possible mechanisms for the diversification of Hox protein function (Foos, 2015).

extradenticle: Biological Overview | Evolutionary Homologs | Regulation | Developmental Biology | Effects of Mutation | References

Home page: The Interactive Fly © 1997 Thomas B. Brody, Ph.D.

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