BIOLOGICAL OVERVIEW

Trying to make evolutionary sense of the gastrulation process [Images], looking for a unity between insects and vertebrates with regard to the genes involved and the movements of tissues, simply does not work. Vertebrate Brachyury has a major role in differentiation of the notochord and in the formation of the mesendoderm. A similar role for Brachyenteron, or T-related gene, a Drosophila Brachyury homolog, is not found in insects. Brachyenteron is required for specification of hindgut and anal pads. A second Brachyury homolog, Optomotor blind shows anterior activity in the brain and central nervous system, again not functionally identical to Brachyury. Thus the fly has specialized Brachyury homologs for the anterior and posterior domains serving different functions from Brachyury in chordates (Kispert, 1994a and Murakami, 1995).

The formation of mesoderm and endoderm requires one gastrulation process in frogs and fish, while insects use separate processes for each type of tissue formed. In zebrafish, cells of the future mesoderm and endoderm involute beneath the margin of neighboring blastomeres, with separate fates for different involuting cells (Schulle-Merkes, 1994). In Xenopus there is mass movement through a single blastpore (Keller, 1992). In Drosophila however, mesoderm is formed by ventral invagination while endoderm is formed by invagination from separate anterior and posterior primordia.

The apparent unity of the gastrulation process in fish and frogs belies its complexity, and the complex origins of mesoderm and endoderm. Regardless of the divergent morphogeneic movements when comparing insects and vertebrates, the level of complexity remains high in either case. For example, in spite of the seemingly unitary origin of mesoderm, an examination of zebrafish goosecoid, coding for a homeodomain protein and no tail, the fish Brachyury homolog, reveals a complex origin: initially expressions overlap, but with time, these genes take on separate roles. goosecoid is expressed in a precoidal area while no tail is expressed in the presumptive mesoderm.

Whereas Xenopus and fish Brachyury homologs are involved in induction of mesoderm, it has been thought neither of the Drosophila Brachyury homologs has anything to do with this process (Schulte-Meeker, 1994). optomotor blind is expressed in neuronal and glial cells in the developing nervous systerm (Poeck, 1993), while Trg is expressed in the hindgut. Both mark the future site of involution, and continue to be expressed as the involution process proceeds (Kispert, 1994a and Murakami, 1995). Thus Trg appears to be involved in the development of the proctodeum, having no immediately apparent phylogenetic relationship to the notochord and optomotor blind shows neural expression, again far removed from a mesodermal function. Homologs of Trg are also expressed in the developing hindgut of Tribolium and Locusta embryos (see Tribolium early embryonic development) suggesting a highly conserved function of Trg in insects. This conservation and the high similarity of T and Trg raise the question of a common evolutionary origin of the hindgut of insects and the notochord of chordates (Kispert, 1994b).

Functions for Drosophila brachyenteron and forkhead in mesoderm specification and cell signalling

This conundrum has largely disappeared in the face of an excellent study by Kusch (1999) on the origin of the caudal visceral mesoderm of Drosophila (CVM). The Drosophila Brachyury homolog brachyenteron (byn) is essential for the development of hindgut, anal pads and Malpighian tubules. byn is activated by the terminal gap gene tailless (tll) in a region of 0%-20% egg length of the syncytium (0% = posterior tip). With completion of cellularization, the byn expression becomes downregulated in the posteriormost cap of the embryo, which will later form the posterior midgut, by the terminal gap gene huckebein (hkb). Thus, the expression of byn is confined to a ring of cells from about 10%-20% egg length. The dorsal and the lateral aspects of that ring correspond to the proctodeum, from which the hindgut, the anal pads and the Malpighian tubules later develop. Intriguingly, hkb also determines the posterior extent of the ventral mesoderm primordium by repressing the mesodermal determinant snail (sna). This suggests that the ventralmost aspect of byn expression might comprise the posterior tip of the mesoderm primordium (Kusch, 1999).

The visceral musculature of the larval midgut of Drosophila has a lattice-type structure and consists of an inner stratum of circular fibers and an outer stratum of longitudinal fibers. The longitudinal fibers originate from the posterior tip of the mesoderm anlage, which has been termed the caudal visceral mesoderm. The CVM migrates in an orderly movement anteriorly and eventually forms an outer layer of longitudinal muscle fibers surrounding the midgut. The progenitors of a second tissue, the inner sheet of circular muscles of the midgut, are recruited from 11 parasegmentally arranged clusters of dorsal mesoderm in the trunk region and are therefore referred to as trunk visceral mesoderm (TVM) (Kusch, 1999).

In this study, the specification of the CVM has been investigated and particularly the role of the Drosophila Brachyury-homolog brachyenteron. Supported by fork head, brachyenteron mediates the early specification of the CVM along with zinc-finger homeodomain protein-1. This is the first function described for brachyenteron or fork head in the mesoderm of Drosophila. The mode of cooperation resembles the interaction of the Xenopus homologs Xbra and Pintallavis. Another function of brachyenteron is to establish the surface properties of the CVM cells; these properties are essential for orderly cell migration along the trunk-derived visceral mesoderm. During this movement, the CVM cells, under the control of brachyenteron, induce the formation of one muscle/pericardial precursor cell in each parasegment. It is here proposed that the functions of brachyenteron in mesodermal development of Drosophila are comparable to the roles of the vertebrate Brachyury genes during gastrulation (Kusch, 1999).

During germband retraction and midgut closure, the progenitors of the the outer, longitudinally oriented fibers of the visceral mesoderm, the CVM, perform an ordered movement that can be subdivided into three phases. The first migratory phase starts at early germband retraction when the cells begin to move anteriorly from their position at the posterior tip of the mesodermal germ layer and split into two tightly packed, bilaterally symmetrical clusters on each side of the posterior midgut primordium. When these clusters have reached the anterior tip of the posterior midgut primordium, the cells detach from each other and disperse anteriorly as two rows along the germband, the second phase of the migration. During this movement, the cells are arranged along the dorsal and ventral edge of the midgut primordia and are in close contact with the band of progenitors of the circular muscle fibers. The band seems to serve as a migration substratum. During the last phase of the migration, which takes place as the midgut encloses the yolk, the progenitors of the longitudinal muscle fibers spread regularly over the underlying circular muscle fibers. The cells acquire a spindle shape, then stretch in an anteroposterior direction and form about 16-20 regularly spaced longitudinal muscle fibers. These fibers reach from the proventriculus to the midgut-hindgut transition where the ureters of the Malpighian tubules insert. The foregut and the hindgut lack any longitudinal muscles and are solely covered by the inner layer of circular muscles (Kusch, 1999).

The specification of the CVM and its fate were monitored by the detection of Byn protein or the expression of CVM-specific markers like croc-lacZ and cpo-lacZ. The initial byn expression at the posterior pole is regulated by tll and hkb. Thus it is likely that the CVM cells are specified under the control of the same genes. In fact, in hkb embryos, the size of the CVM primordium is enlarged and comprises more cells than normal. This corroborates the notion that the CVM primordium constitutes the most posteriorly located mesoderm primordium. tll expression reaches more anteriorly than the hkb domain and encompasses the primordia of the proctodeum and of the CVM. One would therefore expect that the formation of the CVM is entirely dependent on tll. Indeed, this is the case: the CVM is missing in tll mutant embryos. Part of the function of tll seems to be mediated by byn. In byn mutants, a significantly reduced number of CVM cells is seen, and these few cells form clusters that are less compact and migrate significantly slower than in wild type. Later, they fail to contact the TVM and do not distribute along the germband. During stage 11, most of the cells acquire a condensed appearance resembling apoptotic bodies. A high level of apoptosis is detected in the proctodeum of byn embryos as well as in the posteriormost mesoderm. By stage 13, cells with the properties of the CVM are not detectable any longer in the mutants and, as expected from this, the dissected midguts of byn embryos lack the outer, longitudinal muscle fibers (Kusch, 1999).

byn embryos show morphological aberrations at a time before the CVM begins to migrate anteriorly. The severely shortened hindgut causes a significant shift in the spatial relationship of the various primordia at the posterior region of the embryo and thereby might indirectly affect the migration of the CVM. In order to exclude such an indirect influence, byn embryos were generated that expressed byn in the CVM precursors, but not in the hindgut. In such embryos, the CVM survives and disperses virtually the same as in wild type along the TVM, whereas the proctodeum remains rudimentary as in ordinary byn mutants. These results demonstrate that the defective migration and the death of the CVM cannot be attributed to the disordered morphology of the posterior gut structures. It has therefore been concluded that byn in the mesoderm is essential for the adhesive and migratory properties of the CVM precursors. byn cannot be the only gene that mediates the function of tll in the specification and further development of the CVM since the lack of tll causes a far stronger phenotype than the lack of byn. In addition to byn, the gene fkh is known to act downstream of tll in the posterior gut. fkh is expressed in a large domain at the posterior pole that encompasses the byn expression domain including the ventral, mesodermal aspect. In fkh mutants, the CVM specification seems less impaired than in byn mutants: the number of CVM cells is initially quite normal. However, as in byn mutants, the cells fail to migrate along the germband although differentiation of the migration substratum, the TVM, is not affected. By stage 14, most of the CVM cells have been eliminated by apoptosis. On this level of analysis, fkh mutants resemble embryos homozygous for weak byn alleles. However, the phenotype of byn fkh double mutants shows that byn and fkh either have distinct functions in the specification of the CVM or act synergistically. In double mutants, no CVM cells are distinguishable, just as in tll mutants. Therefore, the function of tll in the specification of the CVM appears to be mediated by byn and fkh (Kusch, 1999).

Only the anterior and the posterior mesoderm are competent to be specified by byn as CVM, in conjunction with fkh. Therefore, at least one other gene must exist that confines the competence to form CVM to these two regions. A good candidate for this gene is zinc finger homeodomain protein-1 (zfh-1). At the blastoderm stage, zfh-1 is expressed in high levels in the terminal regions of the mesoderm including the primordium of the CVM. zfh-1 is essential for the migration of the CVM: in zfh-1 mutant embryos, CVM-specific gene expression such as croc-lacZ is deleted. From the restricted effects of ectopic byn /fkh, it has been proposed that the two genes are capable of specifying CVM development only in the region of high zfh-1 expression. zfh-1, byn and fkh act in parallel downstream of tll. High levels of caudal zfh-1, as with byn and fkh, are dependent on tll, and there is no crossregulation between zfh-1, byn and fkh (Kusch, 1999).

The caudal visceral mesoderm is specified independent of twist function. The overlapping expressions of the two zygotic genes twi and snail (sna) are essential for gastrulation and specification of the mesoderm. twi is an activator of mesodermal gene expression, whereas sna mainly functions as a repressor of neuroectodermal gene expression in the mesoderm. Deviating from this rule, mesodermal zfh-1 expression is missing in sna mutants, whereas high zfh-1 expression is unaffected in the termini of the mesoderm in twi mutants. The expression of byn and posterior fkh does not depend on sna or twi function, raising the question whether the CVM might be at least partially specified in twi or sna mutants. In sna embryos, no byn-expressing cells can be detected in a mesoderm-specific position, i.e. between the epidermis and the midgut epithelium. Such embryos lack CVM-specific gene expression, which is in accordance with the findings that zfh-1 depends on sna and CVM development on zfh-1. Strikingly, in twi mutants, byn-expressing cells can be found at an internal, mesoderm-typical position in the tail region. It is not clear how the cells find their way into the twi embryos, which are characterized by the failure to form a ventral furrow. The cells probably immigrate after they have been internalized together with the adjoining posterior gut anlagen. Later the cells begin to express CVM-specific marker genes and undergo the first phase of the normal migration movement: they arrange as two clusters ventrolaterally to the posterior midgut. The cells initiate the second phase of migration as well; they become migratory, disperse and acquire the typical spindle shape of normal CVM cells. However, most likely because of the absence of their putative migration substratum, the TVM, they merely spread over the posterior midgut in twi embryos. Thus, the CVM is not only internalized during gastrulation independent of twi function, but also acquires at least some of its adhesive and migratory properties. This view is consistent with the finding that the expressions of byn, fkh and zfh-1, which are required for the specification of the CVM, are not affected in twi mutants (Kusch, 1999).

The defects in the CVM of a byn mutant suggest that byn is not only involved in the early specification of the CVM, but also plays an essential role in establishing the adhesive and migratory properties of the CVM cells. These do not properly form the two bilaterally symmetrical clusters and later fail to disperse along the TVM in byn mutants. Moreover, the ubiquituous mesodermal byn expression strongly affects the normal migratory behaviour of the CVM. The cells do not exclusively migrate along the TVM towards the anterior, but attach to any other cell in the mesoderm. As a consequence, migration is not restricted dorsoventrally and the cells do not reach the anterior half of the midgut. In these experimental embryos, it is impossible to physically separate splanchopleura and somatopleura, since they are firmly attached to each other. Normally this is not the case. It is concluded from these observations that ectopic byn expression changes the adhesive properties of the mesoderm. The specific effects of ectopic byn on the surface properties of other mesoderm cells also include the rescue of the germ cell migration defect of byn mutants. Normally, the germ cells pass through the posterior midgut epithelium as it becomes mesenchymal, prior to germband retraction. From the basal side of the endoderm, the germ cells migrate to the adjoining mesoderm of the body wall. Later, they migrate anteriorly and intermingle with the somatic gonadal mesoderm. It has been suggested that the CVM plays an important role in directing or facilitating this transition of the germ cells to the mesoderm. During the first phase of migration, the CVM cell clusters are located at a position where the germ cells pass through the epithelium of the posterior midgut before they migrate to the gonadal mesoderm. In byn mutants, the germ cells pass through the midgut epithelium as in wild type, but virtually all cells distribute over the ventral surface of the posterior midgut rather than contacting the mesoderm. byn is neither expressed in the germ cells nor in the gonadal mesoderm and the latter develops normally in byn mutants. Therefore, the observed phenotype is most likely due to defects in a byn-dependent signaling or the adhesive properties of the CVM. This idea is supported by the finding that ubiquitous expression of byn in the mesoderm rescues the defective germ cell migration in byn mutants almost completely: only a few germ cells do not coalesce with the somatic gonadal precursors. Thus, either the few rescued CVM cells of byn embryos at the normal position (close to the migrating germ cells) or the adhesive or signaling properties that ectopic byn confers to other mesodermal cells, are sufficient for the transition of the germ cells from the midgut to the gonadal mesoderm (Kusch, 1999).

A dynamic functional evolutionary diversification has recently been proposed for the Brachyury genes. The Brachyury proteins are a subfamily of the T-domain transcription factors, which are classified by a highly conserved DNA-binding domain. Brachyury relatives have been found in the diploblastic Cnidarian as well as triploblasts like chordates, protochordates, echinoderms, hemichordates and insects, suggesting that evolutionarily conserved mechanisms are regulated by these factors. However, initial comparisons of the expression and function of Brachyury genes between deuterostomes and protostomes indicate diverse rather than common functions. In deuterostomes, Brachyury seemed to play a critical role in mesoderm development, whereas the protostome Brachyury proteins have been shown to be involved in hindgut development. For instance, in vertebrates, Brachyury is transiently expressed in all nascent mesendodermal cells. As development proceeds, the expression becomes restricted to the notochord and tailbud. In cephalochordates and urochordates, the distinct aspects of mesodermal expression have been divided between two paralogs, one expressed in the posterior mesoderm and the other expressed in the notochord, whereas no expression is detectable in the gut. In echinoderms, Brachyury has only been reported to be expressed in the migrating secondary mesenchyme cells. However, it has recently been shown that the hemichordate Brachyury homolog is not only expressed in the mesoderm but also in the oral and anal region of the gut. Since Brachyury is transiently expressed in vertebrates in the developing hindgut epithelium of the tailbud region, it has been proposed that the expression of Brachyury in the posterior gut reflects its original function in development. Intriguingly, the Hydra Brachyury gene is expressed in the cells of the gut opening. In the light of the findings reported in this paper, it is concluded that the role of byn in the mesoderm of Drosophila is reminiscent of and likely to be homologous to the proposed mesodermal functions of vertebrate Brachyury genes (Kusch, 1999).

The homeodomain transcription factor Orthopedia is involved in development of the Drosophila hindgut

The Drosophila hindgut is commonly used model for studying various aspects of organogenesis like primordium establishment, further specification, patterning, and morphogenesis. During embryonic development of Drosophila, many transcriptional activators are involved in the formation of the hindgut. The transcription factor Orthopedia (Otp), a member of the 57B homeobox gene cluster, is expressed in the hindgut and nervous system of developing Drosophila embryos, but due to the lack of mutants no functional analysis has been conducted yet. This study shows that two different otp transcripts, a hindgut-specific and a nervous system-specific form, are present in the Drosophila embryo. Using an otp antibody, a detailed expression analysis during hindgut development was carried out. otp was not only expressed in the embryonic hindgut, but also in the larval and adult hindgut. To analyse the function of otp, the mutant otp allele otp^GT was generated by ends-out gene targeting. In addition, two EMS-induced otp alleles were isolated in a genetic screen for mutants of the 57B region. All three otp alleles showed embryonic lethality with a severe hindgut phenotype. Anal pads were reduced and the large intestine was completely missing. This phenotype is due to apoptosis in the hindgut primordium and the developing hindgut. These data suggest that otp is another important factor for hindgut development of Drosophila. As a downstream factor of byn otp is most likely present only in differentiated hindgut cells during all stages of development rather than in stem cells (Hildebrandt, 2020).

The Drosophila embryonic hindgut is a single-layered ectodermally derived epithelium surrounded by visceral musculature. It arises from a group of cells at the posterior part of the blastoderm stage embryo referred to as the hindgut primordium. The hindgut primordium is a ring of about 200 blastoderm cells that is internalised during gastrulation to form a short, wide sac. In a relatively short time this epithelium sac is transformed into a long tube containing approximately 700 cells. The growth of the hindgut starting at stage 12 is not due to cell divisions, but a twofold endoreplication that leads to an increase in cell size, and as a consequence total length of the hindgut. During this process, the developing hindgut becomes subdivided along the anterior posterior (AP) axis and the dorsoventral (DV) axis. Along the AP axis, the hindgut forms three morphologically distinct regions: the small intestine, large intestine, and rectum. The small intestine is the most anterior part of the hindgut and is connected to the posterior midgut, whereas the large intestine is the central part of the hindgut and forms three distinct regions along the DV axis. The dorsal and ventral regions constitute the outer and inner portions of the hindgut loop, respectively. Two rows of boundary cells are organised between these two regions and as two rings at the anterior and posterior borders of the large intestine. The most posterior-most portion of the hindgut is the rectum, which connects to the anal pads (Hildebrandt, 2020).

Several genes are required to establish the hindgut primordium and to pattern the hindgut along the AP axis. At the blastoderm stage a group of posterior cells, called the proctodeal primordium, will later on give rise to the hindgut. In these cells the transcription factor Tailless (Tll) is expressed and subsequently activates other transcription factors like Brachyenteron (Byn), Fork head (Fkh) and Bowel (Bowl) as well as the signalling protein Wingless (Wg), which are all necessary for hindgut development. The transcription factor Caudal (Cad) is also expressed in the proctodeal ring, but independently of Tailless. Tll and Wg are necessary to establish the primordium, whereas Cad is necessary for the internalisation of the hindgut primordium later on. Proper gene expression in and maintenance of the large intestine requires byn, Dichaete (D), raw, lines (lin) and mummy (mmy), while bowl and drumstick (drm) are required for gene expression in and maintenance of the small intestine (Hildebrandt, 2020).

The Drosophila T-box gene brachyenteron (byn) is expressed in the ring of cells that internalise to form the embryonic hindgut and expression is maintained in the hindgut throughout embryogenesis. In byn mutants the hindgut is shortened due to apoptosis and the large intestine is missing. The Drosophila homeobox gene orthopedia (otp) is also expressed in the hindgut, anal pads and along the central nervous system (Simeone, 1994). It is located in 57B region of the second chromosome in close vicinity to the other homeobox genes Drosophila retinal homeobox (Drx) and homeobrain (hbn). In the hindgut, otp is directly activated by byn in a dose-dependent manner via multiple binding sites present in a regulatory element of otp (Hildebrandt, 2020).

Otp is highly conserved through evolution and has been identified in most multicellular organisms. Among these are several invertebrates such as sea urchins, the mollusc Patella vulgata, the annelid Platynereis dumerilii and several vertebrates such as zebrafish, that have two genes namely otp1 and otp2, chicken, mouse and human. otp genes of vertebrates have a major function in the development of the hypothalamic neuroendocrine system (see Del Giaccom 2008 for review) (Hildebrandt, 2020).

The function of otp during Drosophila development has been unknown so far as no mutants have been described. The present study shows that otp is required for proper hindgut development in Drosophila. One otp allele was generated by ends-out gene targeting and two additional otp alleles were isolated in an EMS-mutagenesis screen. All three otp alleles are characterised by a dramatically reduced hindgut lacking the complete large intestine. This reduction in hindgut length is due to apoptosis in the hindgut primordium and the developing hindgut (Hildebrandt, 2020).

This paper analysed the function of the transcription factor Orthopedia during hindgut development. In the embryo, otp is expressed in the hindgut primordium, the developing hindgut, and the anal pads. This expression is dependent on several upstream regulators, such as lines and byn. Byn directly activates otp through modular binding sites upstream of hindgut specific promoter of otp in a cooperative fashion. otp is then expressed in the large intestine, rectum and anal pads, unlike Byn, which is also expressed in the small intestine. Byn alone is therefore not sufficient to activate otp; lines expression might also be needed. In the small intestine where byn is expressed, lines is repressed by drumstick preventing otp activation. If lines is overexpressed in the small intestine, the repressive effect of drumstick can be overruled and otp expression can take place, supporting the proposed model that lines and byn are necessary for otp expression in the hindgut. Once otp is activated in the embryonic hindgut, its expression continues until the adult stage. The only region where otp in contrast to byn is not expressed is the larval hindgut proliferation zone in the anteriorly located pylorus which supposedly generates hindgut stem cells capable of replacing the larval hindgut cells undergoing apoptosis and building the adult hindgut. The presence of adult hindgut stem cells has been questioned when it was shown that proliferation only takes place in response to tissue damage. The current view is that all parts of the Drosophila intestinal tract maintain stem cells that could migrate across organ boundaries. Since otp expression was never shown in areas where intestinal stem cells are present, otp is rather expressed in and a marker for differentiated hindgut cells (Hildebrandt, 2020).

To analyse the function of otp, a mutant allele was generated by gene targeting via homologous recombination and using this targeting strain, two additional mutant alleles were identified by complementation among an EMS-induced collection of mutants from the 57B region. In the gene targeting mutant, part of exon 4 and exon 5 were missing resulting in an N-terminal deletion of the otp-PC protein form including most of the homeodomain. In the otp¹⁰²⁴ mutant, the N-terminus and most of the homeodomain were present, but the C-terminal part of the protein was missing. In both cases no otp protein expression was detected since the anti-Otp antibody was directed against the C-terminal part of Otp. In the case of otp¹³⁰⁶⁴, no protein was detected with anti-Otp antibody nor was a sequence alteration in the coding region detected. The splice sites were intact, but it cannot be ruled out that a cryptic splice site might be generated. The generation of cryptic splice sites by mutations is often the case in human genetic disorders like Neurofibromatosis type I (NFI) or Cystic fibrosis where such mutations generate pseudo-exons. Another possibility might be a mutation in a regulatory region of the gene. All otp alleles showed embryonic lethality with a strong hindgut phenotype. The loss of the large intestine led to a dramatic reduction in hindgut length to about one third of the wildtype length. This phenotype is comparable to the byn phenotype, since byn is directly regulating otp. The three transcriptional regulators drm, bowl and lin are required for patterning and cell rearrangements during elongation in the hindgut, but when compared to otp, showed only a reduction to 40% (drm and bowl) or 50% (lin) in the mutant phenotype, suggesting that the loss of otp function is much more severe compared to these genes. A gut specific function of otp like seen in Drosophila is not known for otp genes in higher organisms, where an expression in the nervous system and a function in various aspects of nervous system development is known. In Drosophila, otp is also expressed in the ventral nerve cord and the brain. Expression in the nerve cord seems to be post-transcriptionally regulated, since, in contrast to the mRNA expression posterior to segment A2, the otp protein is not expressed there. This might be due to the regulation via a miRNA as it was seen for other developmental processes (Hildebrandt, 2020).

The nervous system function of otp in higher organisms has been mainly analysed in various model organisms like zebrafish and mice. It was shown that otp is expressed in the hypothalamus that exerts influence on physiological function in various processes like blood pressure, circadian rhythmicity, energy balance and homeostasis. In zebrafish otp in the hypothalamus is required in the preoptic area for the production of the neurohypophysial peptide arginine vasotocin, for dopaminergic and neuroendocrine cell specification, regulation of stress response and through neuropeptide switching that impacts social behavior. In mice, a loss of otp leads to a progressive impairment of neuroendocrine cells in the hypothalamus, and homozygous mutant animals die soon after birth with a failure of terminal differentiation of neurosecretory cells. Recently, it was shown that a mutation in otp is associated with obesity and anxiety in mice. The otp gene from humans was cloned some time ago, but only during the last few years, it could be shown that otp has a high diagnostic value concerning pulmonary neuroendocrine tumours. Even if these tumours accounted for only 1%-2% of all lung tumours, their occurrence increased over the last decades. People with poor survival rates showed a strong downregulation of otp. otp that is normally located in the nucleus (nOTP) could also be detected in the cytoplasm (cOTP) or not be present at all. Patients with nOTP have a favourable disease outcome, those with cOTP have an intermediate outcome and those with no otp expression have the worst disease outcome demonstrating the diagnostic value of OTP. Due to these very interesting aspects of otp function in the nervous system of higher organisms, it would be interesting to analyse the function of otp during embryonic brain development of Drosophila, as well as later functions of otp during larval development and in the adult, using the newly generated otp alleles in the future (Hildebrandt, 2020).

Using gene expression analysis and newly generated mutant otp alleles, this study has shown that the Drosophila homeodomain transcription factor Orthopedia is an important factor for hindgut development. These findings demonstrate a requirement of otp to build the large intestine of the hindgut and also in the correct formation of the anal pads. This phenotype is caused by apoptosis at the beginning of hindgut development. otp as a downstream factor of byn is most likely present only in differentiated hindgut cells during all stages of development rather than in stem cells (Hildebrandt, 2020).

Dynamics of an incoherent feedforward loop drive ERK-dependent pattern formation in the early Drosophila embryo

Positional information in development often manifests as stripes of gene expression, but how stripes form remains incompletely understood. This study used optogenetics and live-cell biosensors to investigate the posterior brachyenteron (byn) stripe in early Drosophila embryos. This stripe depends on interpretation of an upstream ERK activity gradient and the expression of two target genes, tailless (tll) and huckebein (hkb), that exert antagonistic control over byn. High or low doses of ERK signaling were found to produce transient or sustained byn expression, respectively. Although tll transcription is always rapidly induced, hkb converts graded ERK inputs into a variable time delay. Nuclei thus interpret ERK amplitude through the relative timing of tll and hkb transcription. Antagonistic regulatory paths acting on different timescales are hallmarks of an incoherent feedforward loop, which is sufficient to explain byn dynamics and adds temporal complexity to the steady-state model of byn stripe formation. It was further shown that 'blurring' of an all-or-none stimulus through intracellular diffusion non-locally produces a byn stripe. Overall, this study provides a blueprint for using optogenetics to dissect developmental signal interpretation in space and time (Ho, 2023).

This study has dissected the regulation of the byn stripe by combining precise optogenetic inputs in space and time with live biosensors of target gene expression. Using ectopic activation of Ras on the ventral side of wild-type embryos, high- and low-amplitude OptoSOS inputs were defined that induce distinct byn transcriptional dynamics – a pulse of expression in early NC14 versus more sustained expression – that match its endogenous responses in the posterior terminus and stripe-forming region. These conditions were then used to characterize the tll and hkb inputs that explain these byn dynamics in space and time (Ho, 2023).

This approach yielded novel insights about both the temporal and spatial interpretation of ERK inputs to pattern the byn stripe. First, differences in signal amplitude are interpreted through the timing of tll and hkb expression. The onset of tll expression is always rapid, occurring as quickly as 4 min after signaling onset, whereas there is a dose-dependent delay in the onset of hkb expression. This delay in hkb expression is a function of Ras/ERK input amplitude, not of developmental time. These data are consistent with previous observations in OptoSOS embryos that hkb RNA only accumulates to high levels in response to blue light inputs over 30 min. They also broaden the conception of the thresholds for tll and hkb expression: tll and hkb can be induced by inputs of the same amplitude, but hkb requires that the signal persist for a longer time. If the amplitude is low enough, the signal must persist longer than the developmental window allows, and hkb is never expressed. Thus, cumulative dose of ERK input (amplitude integrated over time) appears to be the relevant feature sensed by the circuit, as has been proposed for the terminal pattern as well as other systems. Integration of signal over time has similarly been shown to be important for interpretation of several morphogen pathways including Hedgehog, Wnt, Nodal and BMP. The byn circuit then processes this input through the relative timing of tll and hkb, rather than simply their presence, to determine local byn expression (Ho, 2023).

This more nuanced understanding of byn regulation resolves a conundrum of the endogenous pattern: how can the transient pulse of expression of byn in the high-ERK, Hkb-positive domain be reconciled with the presence of its inhibitor? It is shown in this study that at the high light levels which produce a comparable pulse of byn transcription, hkb transcription is delayed relative to tll and this delay is also evident in the accumulation of their protein products. Thus, there is a temporal window in which only the positive regulator is present, allowing for a pulse of byn expression, before accumulation of the repressor. The sequential appearance of Tll and Hkb was hypothesized during the initial characterizations of posterior patterning but has only now been directly shown. It is interesting to note that Tll has been characterized as a transcriptional repressor, implying that there is an intermediate node between tll and byn. However, the identity of this node and how it affects the timing of byn activation and repression remain unknown (Ho, 2023).

Improved understanding of byn regulation also explains how a byn stripe can form in conditions where tll and hkb transcription have the same spatial domain. The current study revisits these results with improved tools, in particular endogenously tagged transcriptional reporters of tll and hkb that are able to clearly resolve differences in transcriptional dynamics that were obscured by enhancer-based reporter constructs.It was found that stimulus conditions that support sustained byn can also support hkb expression in NC14, but under these conditions hkb expression is largely absent from earlier nuclear cycles. The co-expression of sustained byn with hkb under low light differs from the wild-type pattern, where the byn stripe forms in a region only expressing tll. Presumably the endogenous ERK gradient induces tll expression at even lower activity levels than optogenetic inputs. It is noted that the shortened bursting duration of sustained byn at the ectopic position (~25 min) compared with the endogenous stripe (~45 min) suggests that the late-appearing hkb under low light does ultimately repress byn in late NC14. It is also possible that the network dynamics reported in this study provide robustness to the byn circuit, allowing it to produce different outputs for even a narrow range of input strengths (Ho, 2023).

This study reveals that the tll-hkb-byn circuit can be classified as an incoherent feedforward loop with rapid activation and delayed repression, a circuit with well-characterized pulse-generation and stripe-forming properties. A unique feature of this circuit however is that the delay in hkb expression is dose-dependent, meaning that differences in signal amplitude are converted to differences in hkb dynamics and thus different byn responses (i.e. transient if hkb onset is fast, sustained if hkb onset is slow). Interestingly, similar dose-dependent delays in transcriptional onset were recently shown for Dorsal and BMP signaling targets. What is the mechanism underlying this delay in hkb onset? The dose-dependence of tll and hkb has been a longstanding open question even without the complexity of temporal dynamics. ERK signaling activates transcription of both tll and hkb through relief of the same repressor, Cic, and it is unclear why these genes would require different doses of ERK signaling. The experiments rule out a few possible explanations. Developmental time does not appear to be crucial, given that the delay in hkb transcription is observed regardless of when light is applied and both the tll and hkb loci are known to be accessible early. It is also possible to rule out interactions with other components of the anterior-posterior patterning machinery given that this study was able to produce an ectopic byn stripe rotated 90° from its endogenous counterpart. One intriguing possibility, supported by previous ChIP-seq results, is that Cic leaves the enhancers of hkb more slowly than those of tll. It is also possible that signaling-dependent chromatin changes are involved. These models will be tested in future studies (Ho, 2023).

The second major finding is that the boundary of a uniform OptoSOS input is blurred in space downstream of Ras to produce two domains from a single input – a transient byn domain within the high-ERK illuminated region and a sustained domain in the low-ERK unilluminated region. These non-local effects of a local Ras input are most likely mediated by diffusion of active intracellular components, a well-established contributor to developmental patterning in the syncytial Drosophila embryo. It remains unknown whether the endogenous terminal dpERK gradient is produced from a similar gradient of active Torso receptors, or is due to the combination of a discrete domain of Torso activity at the poles and cytoplasmic diffusion of downstream components. If the latter model is correct, the developmental rescue by an all-or-none OptoSOS input may not be an example of a simple input replacing the function of a complex one, but rather a good approximation of endogenous activation in the terminal system. A number of systems once thought to depend strictly on input concentration have similarly been shown to depend on an unexpectedly simple form of the input (Ho, 2023).

Several limitations of the optogenetic system reveal opportunities for future investigation. These experiments were performed at an ectopic position in the embryo where position-specific gene expression may influence ERK interpretation differently than at the poles. For example, the gap gene knirps has been shown to repress tll in the center of the embryo, and it was observed that the total domain of tll and byn expression was smaller under low light. Because of these positional differences in ERK sensitivity, it is not possible to make absolute comparisons about input and output strengths with the endogenous terminal pattern. In the future it will be interesting to investigate this circuit in embryos lacking other sources of positional information, preventing localized gap gene expression . Also, it is possible that the methods left some transcriptional bursts undetected, and it is not possible to distinguish whether an upper bound of ~75% transcriptionally active nuclei represents true transcriptional heterogeneity or an experimental limit of detection. These limitations could be overcome by future studies using techniques that simultaneously label the target DNA locus and measure transcription in live embryos, or advances in high-quality volumetric imaging and machine-learning approaches. Finally, many questions remain about the precise temporal relationships between ERK activation, gene transcription and protein accumulation. What is the relative influence of tll and hkb transcripts produced by early versus late nuclear cycles, and what is the delay between RNA production and protein accumulation? Combining transcriptional and protein reporters in the same embryo with mathematical models will allow these questions to be addressed (Ho, 2023).

Altogether, this work provides a blueprint for dissecting a developmental circuit with optogenetic tools to reveal new insights about network architecture. This study has manipulated amplitude, duration, timing and spatial pattern of the signal to understand the contributions of each factor to signal interpretation. This framework will be an effective strategy for dissecting other developmental circuits in the future (Ho, 2023).

GENE STRUCTURE

Base pairs in 5' UTR - 336

Base pairs in 3' UTR - 311

PROTEIN STRUCTURE

Amino Acids - 721

Structural Domains

TRG has a stretch of 200 amino acids with extensive homology to chordate T (Brachyury) gene products from mouse, Xenopus, zebrafish and ascidians (phylum Urochordataq). Drosophila Optomotor-blind also contains a T-domain, but OMB T-domain is divergent from that of the other species in that it contains insertions. Functional homology of Drosophila Trg with Brachyury is not apparent. Whereas Brachyury has a major role in differentiation of the notochord and in the formation of the mesendoderm, a similar role is not found in insects (Kispert, 1994a).

The mouse Brachyury (T) gene is the prototype of a growing family of so-called T-box genes that encode transcriptional regulators and have been identified in a variety of invertebrates and vertebrates, including humans. Mutations in Brachyury and other T-box genes result in drastic embryonic phenotypes, indicating that T-box gene products are essential in tissue specification, morphogenesis and organogenesis. The T-box encodes a DNA-binding domain of about 180 amino-acid residues: the T domain. The X-ray structure of the T domain from Xenopus laevis was examined in complex with a 24-nucleotide palindromic DNA duplex. The protein is bound as a dimer, interacting with the major and the minor grooves of the DNA. A new type of specific DNA contact is seen, in which a carboxy-terminal helix is deeply embedded into an enlarged minor groove without bending the DNA. Hydrophobic interactions and an unusual main-chain carbonyl contact to a guanine account for sequence-specific recognition in the minor groove by this helix. Thus the structure of this T domain complex with DNA reveals a new way in which a protein can recognize DNA (Muller, 1997).

date revised: 5 August 2023

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.