BIOLOGICAL OVERVIEW

Buttonhead is a transcriptional activator related to vertebrate Sp1. It is a gap gene localized to three head segments: antennal, intercalary and mandibular [Image], whose expression domain physically overlaps the more restricted expression of empty spiracles and orthodenticle.

The transcription factors of the head create developmental compartments from which the special features of each head segment normally develop. BTD regulates transcription of segment polarity genes wingless and engrailed, which are under the control of pair-rule genes in the trunk. The fact that wingless and engrailed are dependent on gap genes in the head, points to the antiquity of the gap gene as a means of regulating segmentation. It also highlights the complexity of region specific transcriptional regulation of critical genes for segmentation (Mohler, 1995).

It has been proposed that the gap-like segmentation genes orthodenticle, empty spiracles and btd contribute to a combinatorial code for head development. The expression domain of each gene corresponds to two or three segment anlagen that fail to develop in each mutant. It is though that these overlapping expression domains mediate head metamerization and generate a combinatorial code to specify segment identity. To test this model, a system was developed for targeted gene expression of buttonhead in the early embryo. Misexpression of btd in the anterior half of the blastoderm embryo directed by the hunchback proximal promoter rescues the btd mutant head phenotype to wild-type. This indicates that while btd activity is required for the formation of specific head segments, its ectopic expression does not disturb head development. It is concluded that the spatial limits of btd expression are not instructive for metamerization of the head region and that btd activity does not contribute to a combinatorial code for specification of segment identity (Wimmer, 1997).

buttonhead is required for the formation of the mandibular, intercalary and antennal head segments of the embryo. The Btd protein is functionally and structurally related to the human C2H2 zinc finger transcription factor Sp1. A second Sp1-like Drosophila gene, termed Drosophila Sp1 (D-Sp1), had been identified on the basis of a partial sequence showing that the gene encodes a characteristic zinc finger domain, composed of three finger motifs similar to both Sp1 and btd. D-Sp1 is located in the same cytological location as btd in chromosome band 9A on the X-chromosome. It had been proposed that D-Sp1 and btd are likely to act as a gene pair and function in an at least partially redundant manner. D-Sp1 acts as a transcriptional regulator. Lack-of-function analysis combined with rescue and gain-of-function studies indicates that btd and D-Sp1 play essential and redundant roles for mechanosensory organ development. However, D-Sp1 lacks the specific features of Btb required for embryonic intercalary and antennal segment formation (Schock, 1999b).

In contrast to btd, D-Sp1 transcripts are observed in low amounts throughout the early embryo up to the preblastoderm stage. Thus, D-Sp1 is expected to be expressed maternally. D-Sp1 transcripts are shown to accumulate during oogenesis in the nurse cells and are transported into the growing oocyte. This observation explains the presence of maternal transcripts in eggs. The transcripts are likely to have a relatively short half life since no D-Sp1 transcripts are detected at blastoderm stage. A further undescribed aspect of the D-Sp1 expression is its distinct pattern in the developing peripheral nervous system, covering the lateral region of the embryo that corresponds to the cluster of cells expressing proneural genes. In addition, D-Sp1 expression is found in the larval brain and in ventral imaginal discs of the larva (i.e. in the antenna but not the eye imaginal disc), and in the leg but not the wing imaginal discs. In the antenna disc, the pattern of D-Sp1 expression is similar to atonal expression, which is necessary for sensory organ development. In the leg imaginal disc, the D-Sp1 expression pattern is similar to Distal-less. Thus it is possible that D-Sp1 is required for the expression of these genes or conversely, ato and Dll are necessary to express D-Sp1. D- Sp1 and ato as well as D-Sp1 and Dll could be part of a common genetic pathway, a proposal to be addressed by future experiments (Schock, 1999b).

btd participates in chordotonal organ development. Chordotonal organs serve a multitude of proprioreceptive, tactile and auditory functions during the life cycle of the fly. They are involved in various behaviors such as the larval withdrawal from touch and the coordinated movements of larvae and adults. Chordotonal organs are required for the reception of stretching and vibration, and each organ consists of a defined number of scolopidia. This study has focused on the lateral pentascolopidial chordotonal organ (lch5), which is segmentally repeated in the abdominal segments of the embryo. It is composed of five scolopidia, which are easily identified by morphological means due to their characteristic size and shape. In embryos carrying a btd point mutation, the number of scolopida in the lch5 varies between two and five. The variability in the phenotype suggeststhat btd participates in chordotonal development, but that the development of the remaining scolopidia is likely due to the activity of other genes that potentially compensate for the loss of btd activity in the developing scolopidia (Wimmer, 1996). In view of the common expression domains of btd and D-Sp1 in the developing peripheral nervous system, it was reasoned that D-Sp1 represents a candidate for carrying a btd-like function for lch5 development. This implies that the absence of both genes in deficiency mutant embryos should enhance the lch5 phenotype (Schock, 1999b).

In order to obtain such double mutant embryos, chromosomal deficiencies and duplications uncovering the 9A region of the X-chromosome were examined by in situ hybridization with antisense RNA probes for the two genes, asking whether the chromosomes carry or have lost one or both genes. Deficiency Df(1)C52 lacks both genes. Embryos that are hemizygous for the deficiency Df(1)C52 allow for the examination of the contribution of D-Sp1 to chordotonal organ development. Such embryos indeed develop a stronger lch5 phenotype than observed with btd mutants. Whereas btd mutants develop at least two, and usually an average of three scolopidia, the deficiency mutants lacking both btd and D-Sp1 develop no scolopidia in most cases and up to a maximum of two scolopidia in a few chordotonal organs per embryo. Thus, D-Sp1 or another gene within the deficiency is necessary for scolopidia development in addition to btd. Furthermore, since some scolopidia were found in btd/D-Sp1-deficient embryos, their development must be promoted by the activity of at least a third gene outside the deficient region (Schock, 1999b).

A 5.2-kb cis-acting element drives expression in the precursors of the peripheral nervous system (Wimmer, 1996). This feature of the btd enhancer allowed for the assessment of whether D-Sp1 in addition to btd participates in lch5 development. Transgene-dependent expression of D-Sp1 in btd mutant embryos shows at least four and frequently five scolopidia per lch5 and is therefore similar to the rescue obtained with btd (Wimmer, 1996). The results show that btd and D-Sp1 carry overlapping or even redundant functions for scolopidia development, but also indicate that D-Sp1 is more efficient than Btd in rescuing scolopidia development (Schock, 1999b).

Expression of D-Sp1 and btd in the developing nervous system and the rescue of the lch5 suggested that D-Sp1 and Btd act in the same neurogenic or neural pathway. In order to see whether the two genes can generate neuronal cell fates in tissues that normally do not express the genes, the Gal4/UAS system was employed. To drive ectopic expression of D-Sp1 or btd in proneural clusters of the wing imaginal disc, a scabrous-Gal4 driver line was used. Ectopic expression of D-Sp1 causes development of mechanosensory stout bristles in ectopic places within the wing, including the posterior margin, veins and intervein regions. In addition, and probably due to scabrous-Gal4 driving D-Sp1 expression in a double row of cells in the anterior wing margin, ectopic stout bristles develop in rows 1 and 3. Ectopic expression of Btd caused a similar but somewhat weaker phenotype as D-Sp1. Since scolopidia and stout bristles are both mechanosensory cells, the results suggest that D-Sp1 and btd are redundant or functionally overlapping components of a genetic circuit that specifies the mechanosensory fate of cells (Schock, 1999b).

The largely identical expression domains and the close relationship with respect to sequences encompassing the zinc finger region (72% identity at the amino acid level) suggests that D-Sp1 and btd constitute a gene pair. Due to their partially redundant function in the peripheral nervous system, D-Sp1 and btd would be expected to carry similar or even redundant functions for head development. This proposal was tested by asking whether D-Sp1 can rescue head segmentation in btd mutant embryos. D-Sp1 cDNA was expressed under the control of a 5.2-kb btd enhancer fragment. This cis-acting region conducts gene expression in the btd blastodermal expression domain and rescues all head defects when driving expression of a btd cDNA in btd mutants. In btd mutant embryos, transgene-derived D-Sp1 expression rescues only mandibular structures as shown by mandibular engrailed (en) expression and the presence of mandibular head structures. Antennal or intercalary structures are not rescued. Mandibular rescue and the failure to rescue the pregnathal segments lacking in btd mutant embryos is independent of the number of D-Sp1 transgene copies. Thus, as has been observed with human Sp1, D-Sp1 lacks the specific features of Btd necessary for pregnathal segment development. The results allow the conclusion that D-Sp1 acts as a transcriptional regulator, which is able to control the Btd target genes necessary for mandibular development. However, Btd must contain additional protein domains that are missing in Sp1 as well as in D-Sp1 and are necessary for the promotion of pregnathal segments (Schock, 1999b).

The DNA-binding properties of the Btd protein are indistinguishable from the human transcription factor Sp1. Furthermore, Btd and Sp1 are capable of activating transcription in transfected cultured cells through interaction with the same DNA target sites. Btd and Sp1 functionally interact with the same TATA box-binding protein-associated factors and support in vitro transcription activation through these contacts. Transcriptional activation by human Sp1 involves the general RNA polymerase II transcription factor TFIID (Smale, 1990). TFIID is composed of the TATA box-binding protein (TBP) and a set of TBP-associated factors (TAFs). TAFs provide interfaces for enhancer-bound transcription factors that contact the basal transcription apparatus and direct the activation of transcription. Sp1 selectively interacts with TAFII110 (Hoey, 1993), one of the eight TAFs of Drosophila that include TAFII250, TAFII150, TAFII110, TAFII80, TAFII60, TAFII40, TAFII30alpha, and TAFII30beta. Partial TFIID complexes composed of TBP, TAFII250, TAFII150, and TAFII110 are sufficient to support transcriptional activation in vitro, whereas partial complexes lacking TAFII110 or mutant Sp1 lacking the interaction surface fail to activate transcription (Gill, 1994 and Chen, 1994). Furthermore, different transcription factors, such as NTF-1, functionally interact with other TAFs. This suggested that the different subunits of TFIID, each with distinct structural characteristics, provide specific substrates for enhancer-bound transcriptional activators to contact the basal transcription apparatus and thereby direct the activation of gene expression (Schock, 1999a and references).

Biochemical interaction assays and functional in vitro studies provided critical support for the notion that transcriptional activation by Sp1 is mediated by binding to TAFII110 (Hoey, 1993). Protein-protein interactions (i.e., Sp1 and TAFII110) are sufficient to support transcriptional activation in vitro, whereas partial complexes lacking TAFII110 fail to do so (Chen, 1994). Because the C2H2 zinc finger DNA-binding motif and glutamine- and serine/threonine-rich N-terminal domains are conserved between Sp1 and Btd (Kadonaga, 1998), a test was performed to see whether Btd exerts Sp1-like biochemical properties. For this, binding studies were performed involving components of the transcriptional apparatus, including TFIIA and the eight Drosophila TAFs. Affinity beads containing either the purified N-terminal or C-terminal region of Btd, excluding the zinc finger domain, or full-length purified Sp1 are able to specifically retain TAFII110 and TAFII150. In addition, weak binding is observed with TAFII60 and TFIIA, whereas the other four TAFs are not retained on the Btd- or Sp1-coated affinity resins. This indicates that both Btd and Sp1 can target the same components of the TFIID complex in vitro (Schock, 1999a).

Do the conserved contacts of Btd also support the activation of transcription in a cell-free reaction system in an Sp1-like fashion? In vitro reactions containing endogenous Drosophila TFIID support Btd-dependent activation of transcription, whereas reactions containing only TBP fail to do so. This result suggests that Btd, like Sp1, requires TAF subunits of the TFIID complex to mediate activation of transcription in vitro. To assess which TAF subunits mediate Btd-dependent activation of transcription, in vitro-assembled TBP-TAF complexes were used instead of the endogenous TFIID, previously shown to mediate Sp1-dependent transcriptional activation (Chen, 1994). Reactions containing in vitro-assembled TBP-TAF complexes composed of TBP, TAFII250, and various combinations of TAFII60, TAFII150, and TAFII110 were assayed. Strongest activation of transcription in response to Sp1 and Btd is mediated by partial complexes containing TAFII150 and TAFII110. This observation is consistent with the earlier finding that TAFII110 and TAFII150 mediate Sp1-dependent activation in a synergistic manner (Chen, 1994). The results suggest further that the presence of TAFII60 does not provide additional transcriptional activation when acting in a combination with TAFII110 or TAFII150, meaning that there is no synergistic interaction between TAFII110 or TAFII150 and TAFII60. These findings demonstrate indistinguishable properties of Sp1 and Btd with respect to their abilities to activate transcription in a TAF-mediated manner (Schock, 1999a).

btd mutant embryos lack three adjacent head segments: the mandibular, the intercalary, and the antennal segments. The development of these segments can be identified by a distinct set of head sensory organs in the embryo or the larval cuticle and by the expression domains of the segment polarity gene engrailed (en) in the germ-band-extended embryo. btd transgene expression under the control of the btd 5.2-kb cis-acting element provides the proper spatial and temporal expression in the three btd-dependent head segment anlagen and rescues the btd mutant head phenotype. In contrast, a transgene containing the Sp1 cDNA in place of the btd cDNA provides a partial mandibular segment only, even when multiple copies of the Sp1 cDNA were supplied. Thus, Sp1 can activate Btd target genes required for mandibular development but is unable to support intercalary and antennal development (Schock, 1999a and Wimmer, 1995).

One possibility to account for the different in vivo properties of Btd and Sp1 is that they exhibit different DNA-binding characteristics that allow them to distinguish different sets of target sites. To assess this possibility directly, an Sp1-derived protein was derived that contains the Btd zinc finger DNA-binding domain (Sp1Btdzf). This hybrid gene was placed under the control of the btd enhancer sequence and expressed in transgenic btd mutant embryos. Expression of Sp1Btdzf transgene results in an Sp1-like rescue of the mandibular segment. Thus, differences in target site specificity and/or binding affinity of the Sp1 and Btd DNA-binding domains are not critical with respect to the biological responses caused by the two proteins. Furthermore, taking into account their identical TAF requirement in vitro, the relevant differences between Sp1 and Btd should also not be dependent on their ability to stimulate transcriptional activation of the biologically relevant target genes (Schock, 1999a).

This proposal was tested by a genetic interaction assay, involving a dominant negatively acting Drosophila mutant of TAFII110 and a homozygous btd mutant that was rescued by a single copy of the btd-expressing transgene. The rationale behind this mutant combination was that the btd activity in such embryos was limiting and that an interference with the TAFII110-BTD interaction should further decrease btd action and thus impair btd-dependent head development. However, trans-heterozygous TAFII110/btd mutant embryos develop a normal head segmentation pattern. BTD must therefore be able to properly activate its target genes by means other than synergistic interactions involving TAFII110 and TAFII150 exclusively. However, lowering the level of BTD or TAFII110 further or reducing the TAFII150 level in addition to BTD and TAFII110 might result in head defects and thereby reveal the necessity of interactions between the two factors (Schock, 1999a).

These data provide evidence that BTD and human Sp1 interact in vitro with the same subset of TAFs and use the same TAFs for transcriptional activation in a reconstituted cell-free assay. In addition, Sp1 containing the Btd DNA-binding domain supports partial mandibular development but fails to provide intercalary and antennal structures. The same is true for a Btd miniprotein composed of the Btd DNA-binding domain fused to the strong transcriptional VP16 activation domain (F. Schock, unpublished results cited in Schock, 1999). Because Sp1 and the VP16 activation domain contact different TAFs (i.e., TAFII40 and the basal factor TFIIB in the case of VP16), binding of the transcription factor and its contacts with basal transcription factors appears to be sufficient to establish mandibular pattern elements. This minimal requirement explains why BTD-dependent mandibular development can be achieved in response to a biologically unrelated transcription factor such as human Sp1. However, the transgene expression study also establishes that this property of BTD is not sufficient to establish segments that require empty spiracles (intercalary segment) or empty spiracles and orthodenticle (antennal segment) in addition to Btd. This suggests that Btd contains properties to interact with Ems and Otd or with their downstream factors. Furthermore, because Sp1 is not sufficient to establish all aspects of mandibular development, it also appears likely that the proper formation of this segment requires an additional Btd-specific component (Schock, 1999a).

PROTEIN STRUCTURE

Amino Acids - 644

Structural Domains

The transcriptional activation function of BTD is encoded by two serine threonine rich domains, alternating with a single glutamine rich domain. In addition, BTD contains an 11 amino acid motif (Btd box) N-terminal to the triple zinc finger domain (Wimmer, 1993)

date revised: 4 August 97 

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

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