runt

Promoter Structure

The runt promoter has a TATA box (Kania, 1990). Runt has an extended cis-regulatory region. There are multiple elements that make quantitative contributions to runt regulation during segmentation. Sequences that are more than 8.5 kb upstream of the runt promoter are necessary for normal expression during the post-blastoderm stages of embryogenesis (Butler, 1992).

Analysis of the runt promoter reveals three different modes of regulation responsible for segmental expression. Upstream elements resemble the stripe-specific enhancers of even-skipped. The interval between -14 and -9 kb contains site specific elements for stripes 1, 3 and 5. Gap genes giant, hunchback, knirps, and Krüppel function as the major input for these stripe specific enhancers. A second mode of regulation is carried out by DNA extending 5 kb upstream of the transcription initiation site. Termed the "7-stripe region", this interval reveals few distinct subelements, suggesting that regulatory sites responsible for early stripe formation are spread diffusely over the whole region. The pair-rule gene even-skipped shows the strongest input for the expression regulated by the 7-stripe region, with additional contributions of bicoid and tailless. Of the pair-rule gene regulatory elements identified so far, the one most similar to the runt 7-stripe region is the zebra element of fushi tarazu. A third mode of regulation is carried out by a late element between +3 and +6 kb (i.e., downstream of the runt transcription start site). This "six-stripe element" is responsible for the narrow late stripes appearing between the 7 early stripes, giving rise to the late segment polarity pattern of runt expression. The 6-stripe element is not an independent enhancer, as it requires promoter-proximal sequences (Klingler, 1996).

Pair-rule genes serve two important functions during Drosophila development: they first initiate periodic patterns, and subsequently interact with one another to refine these patterns to the precision required for definition of segmental compartments. A pair-rule input region of the runt gene characterizes this through the use of reporter gene constructs and by comparison with corresponding sequences from Drosophila virilis. Many but not all regulatory properties of this '7-stripe region' are functionally conserved. All primary pair-rule genes (hairy, eve and runt itself) are known to affect the expression from this element. Nevertheless, the interspecies comparision reveals that the conservation between homologous sequences is surprisingly low. Fourteen conserved blocks can be identified which together comprise 419 bp, i.e. 6.7% of the total sequence. The largest of these blocks encompasses the basal promoter region with the transcription start site and TATA box. Without the TATA box area, i.e. when just regarding the upstream region, only 346 bp appear to be conserved, which corresponds to 5.6% of the total. Hairy and Bicoid are shown to bind to conserved sequence blocks in vitro, and Tailless binding sites are also conserved. While expression in the early central domain and the early stripe pattern are largely conserved between the 7-stripe regions of the two species, this is not the case for the narrowing of the stripes and the transition to a segmental expression. This function change in the upstream region of the runt genes in the two species is reflected by the phylogenetic footprinting results, which do not identify conserved sites in the sequence that mediates this regulation in Drosophila melanogaster. When compared to similar data for gap gene input elements in eve and hairy, these data suggest that pair-rule target sequences of runt are less constrained during evolution, and that functional elements mediating pair-rule interactions can be dispersed over many kilobases (Wolff, 1999).

With respect to runt's function in segmentation, promoter constructs identify two subelements that are essential for the stripe pattern. (1) A general activating sub-element is situated close to the TATA box, between -0.7 and -0.1 kb. Constructs without the DNA lack expression in stripes almost completely, and they also do not form an early central domain. (2) Sequences between -.07 and -1.3 kb negatively regulate the stripe pattern during gastrulation. Experiments suggest that the DNA between -1.2 and -.07 kb mediates the inhibitory effects of fushi tarazu on runt expression. At later stages of development, the 7-stripe region also gives rise to neural expression, first in head spots, and also at the fully extended germ band stage in the ventral nerve cord. Both of these expression aspects can be attributed to distinct elements. The regulation in the CNS is provided by DNA between 5 and 6 kb upstream of the transcription start site, and the head spot expression by DNA between -1.3 and -1.9 kb. In both cases, expression does not require the presence of the general activating element immediately upstream of the basal promoter. These regulatory regions are functionally conserved, for the most part, in D. virilis. Nevertheless, the subelements mediating repression by ftz and activation in the head spots appear not to be conserved (Wolff, 1999).

Transcriptional Regulation

There are several distinct phases of runt expression in the early embryo. Each phase depends on a different set of regulators. The first phase of expression is a broad-field of mRNA accumulation in the central regions of syncytial blastoderm stage embryos. This pattern is due to terminal repression by the anterior and terminal maternal systems. The effect of the terminal system, even at this early stage, is mediated by two zygotic gap genes, tailless and huckebein. A 7 stripe pattern of Runt mRNA accumulation emerges during the process of cellularization. The initial formation of this pattern depends on position-specific repression by zygotic gap genes.

In a second phase, three pair-rule genes, hairy, even-skipped, and runt itself, affect Runt's 7 stripe pattern. The autoregulatory effects of runt are stripe specific; the effects of hairy are more uniform; and the patterns obtained in even-skipped mutant embryos show a combination of both stripe specific and uniform regulatory effects.

In a third distinct phase of expression, at the onset of gastrulation, runt is expressed in 14 stripes . fushi tarazu plays a negative regulatory role in generating this pattern, whereas the pair-rule genes paired and odd-paired are required for either activating or maintaining runt expression during these stages (Klingler, 1993).

The asymmetric distribution of the gap gene knirps (kni) in discrete expression domains is critical for striped patterns of pair-rule gene expression in the Drosophila embryo. To test whether these domains function as sources of morphogenetic activity, the stripe 2 enhancer of the pair-rule gene even-skipped was used to express kni in an ectopic position. Manipulating the stripe 2-kni expression constructs and examining transgenic lines with different insertion sites led to the establishment of a series of independent lines that display consistently different levels and developmental profiles of expression. Individual lines show specific disruptions in pair-rule patterning that are correlated with the level and timing of ectopic expression (Kosman, 1997).

It is likely the KNI functions as a repressor to set the posterior border of eve stripe three. To test whether the early repression of eve stripe 3 is mediated through the eve stripe three enhancer, stripe 2-kni constructs were crossed with a line carrying lacZ under the control of this enhancer. Ectopic kni specifically represses the stripe 3 enhancer in a dose-dependent manner. Stripe 2-kni causes disruption of runt stripes 2 and 3, but has no effect on stripe 1. The repression of stripe 3 increases in proportion to the level of ectopic kni, a response similar to that seen for eve stripe 3. Different levels of ectopic kni cause disruptions of fushi tarazu stripes 2 and 3, but have no effect on the expression of ftz stripe 1. It is possible that these effects are indirect and may be mediated through other segmentation genes but this possibility is made unlikely by the fact that hairy expression is virtually unaffected in stripe 2-kni embryos. These results suggest that the ectopic domain of kni acts as a source for morphogenetic activity that specifies regions in the embryo where pair-rule genes can be activated or repressed. Evidence is presented that the level and timing of expression, as well as protein diffusion, are important for determining the specific responses of target genes (Kosman, 1997).

The early bell-shaped gradient of even-skipped expression is sufficient for generating stable parasegment borders. The anterior portion of each early stripe has morphogenic activity, repressing different target genes at different concentrations. These distinct repression thresholds serve to both limit and subdivide a narrow zone of paired expression. Within this zone, single cell rows express either engrailed, where runt and sloppy-paired are repressed, or wingless, where they are not (Fujioka, 1995).

Although many of the genes that pattern the segmented body plan of the Drosophila embryo are known, there remains much to learn in terms of how these genes and their products interact with one another. Like many of these gene products, the protein encoded by the pair-rule gene odd-skipped (Odd) is a DNA-binding transcription factor. Genetic experiments have suggested several candidate target genes for Odd, all of which appear to be negatively regulated. Pulses of ectopic Odd expression have been used to test the response of these and other segmentation genes. Three different phenotypes are generated in embryos in which odd is expressed from a heat shock promoter: head defects only, a pair-rule phenotype and a pair-rule phenotype restricted to the dorsal half of the embryo. The head defects only phenotype prevails when Odd is induced between 2:10 and 2:30 hours after egg laying (AEL). The second phenotype is generated when Odd is induced between 2:30 and 2:50 AEL, while the third phenotype prevails when heat shocks are administered between 2:50 and 3:10 AEL. The results are complex, indicating that Odd is capable of repressing some genes wherever and whenever Odd is expressed, while the ability to repress others is temporally or spatially restricted (Dréan, 1998).

Two of the seven pair-rule genes tested do not show significant changes in expression at the stages examined. These include the genes odd-paired (opa) and, surprisingly, ftz. In odd minus embryos, ftz stripes do not resolve properly, remaining about 3 cells wide until well into the process of germ band extension. This suggests that Odd may be a repressor of ftz. However ectopic Odd does not repress ftz expression. Also unexpected was the fact that ectopic Odd has effects on all three of the ëprimaryí pair-rule genes. These were previously thought not to be regulated by Odd. In stage 5 embryos, stripe 1 of hairy is efficiently repressed by ectopic Odd. The first stripe of eve is also repressed at this stage. Repression of h stripe 1 continues in older embryos and is accompanied by weaker repression of stripes 2-6. These effects of Odd on h correlate with what appears to be a modest broadening of h stripes in odd-minus embryos, particularly stripe 1. Early repression of the first stripes of h and eve likely accounts for the cuticular head defects that arise from early pulses of ectopic Odd expression. Interestingly, in odd-minus embryos, the entire 7-stripe pattern of h appears to expand, both anteriorly and posteriorly. This is also true of eve and runt stripes. These data provide no explanation for this, but it may explain the fairly consistent spacing of h stripes, despite their apparent broadening (Dréan, 1998).

The gene, hopscotch (hop), the Drosophila JAK kinase homolog, is required maternally for the establishment of the normal array of embryonic segments. In hop mutant embryos, although expression of the gap genes appears normal, there are defects in the expression patterns of the pair-rule genes even-skipped, runt, and fushi tarazu. The effect of hop on the expression of these genes is stripe-specific (Binari, 1994).

Ectopic expression of the 69 kDa TTK protein significantly represses even-skipped, odd-skipped, hairy and runt. The 88 kDa form does not act to repress these genes (Brown, 1993).

Pair-rule gene expression is disrupted in Dichaete mutants. Expression of the gap genes Krüppel, knirps, and giant are normal, indicating that Dicaetae acts in parallel or downstream of these gap genes. the so-called primary pair-rule gene even-skipped, Hairy, and runt each show reductions in levels of expression in Dichaete mutants, with variable stripe specific effects on eve, fushi tarazu, hairy and runt. Since the stripes of pair rule genes generally occur in the correct anterior-posterior position in Dichaete mutants, the gene is unlikely to provide key positional information; it is more likely to be required in the maintainance or establishment of appropriate levels of pair-rule gene expression in the central region of the embryo (Russell, 1996 and Nambu, 1996).

Neuroblast expression of runt is restricted to single cells through the action of the Notch pathway (Kania, 1990).

At least some aspect of dosage compensation is not carried out by Male-specific-lethals, including MSL-2. Early runt dosage compensation is directed by the product of the early promoter of Sex lethal. Thus the early transcripts of Sex lethal have a role in addition to splicing, that is, in directing the early stages of dosage compensation. runt dosage compensation is a consequence of early Sex lethal expression in females. Since MSL-1 and MSL-2 begin to associate with the X chromosome during the cellular blastoderm, it is likely that MSL-independent compensation of genes such as runt and MSL-mediated compensation of other early-acting X-linked genes could either be separated by a very short developmental period or could occur simultaneously. The mechanism of early Sex lethal directed dosage compensation is unknown (McDowell, 1996 and Bernstein, 1994).

Control of photoreceptor axon target choice by transcriptional repression of Runt

Drosophila photoreceptor neurons (R cells) project their axons to one of two layers in the optic lobe, the lamina or the medulla. The transcription factor Runt (Run) is normally expressed in the two inner R cells (R7 and R8) that project their axons to the medulla. The relationship between Run and the ubiquitously expressed nuclear protein Brakeless (Bks), which has previously been shown to be important for axon termination in the lamina, has been examined. Bks represses Run in two of the outer R cells: R2 and R5. Expression of Run in R2 and R5 causes axonal mistargeting of all six outer R cells (R1-R6) to the inappropriate layer, without altering expression of cell-specific developmental markers (Kaminker, 2002).

As an axon navigates toward a target region during development, it alters its course based on attractive or repulsive molecular signals in its environment. There are at least two phases in the establishment of neuronal connections. First, axons project to and distinguish between regions or layers and second, once within the target layer, axons fine-tune their projections. This second step involves precise interactions between growth cones and target cells and has been well studied, particularly for the participation of cell-surface molecules and their associated signal transduction machinery. This study investigates the role of two transcription factors, Run and Bks, in the first phase of target layer selection for differentiating R cells in the Drosophila optic lobe (Kaminker, 2002).

The expression pattern of several R cell-specific differentiation markers is normal in bks mutants. In striking contrast to other markers, however, Run is ectopically expressed in two extra R cells per cluster in somatic loss-of-function clones of bks mutant tissue. In bks clones, Run expression is expanded from its normal R7/R8 pattern to also include R2 and R5. This suggests that Bks represses Run in R2 and R5 cells. Cells along the edges of bks clones were analyzed for the expression of Run. In 196 ommatidia counted along clone boundaries, R2/R5 expression of Run was never seen in a cell that is wild type for bks. It is concluded that the repression of run by Bks is cell-autonomous (Kaminker, 2002).

R1-R6 photoreceptor axons misproject to the medulla in bks loss-of-function mutants. To determine whether this axonal targeting defect is due to the relief of Run repression in R2 and R5, the GAL4/UAS system was used to express Run in these cells. When Run is expressed in R2, R5 and R8 using the MT14-GAL4 driver, all innervating R-cell axons bypass the lamina and projected through to the medulla. When Run is misexpressed in R2 and R5 alone, the defect is as severe as when Run is misexpressed in all R cells using the GMR-GAL4 driver. Run over-expression in R8 alone, where it is normally expressed, does not affect axonal projections. In addition, misexpression of Run in R1, R6 and R7 using lz-GAL4 or in R3 and R4 using sal-GAL4 does not give rise to a comparable axonal misprojection phenotype. The severe axonal mistargeting phenotype in MT14-GAL4/UAS-run flies is attributable to Run expression in R2 and R5 (Kaminker, 2002).

Unlike the ordered wild-type array, thick bundles of axons are seen entering the medulla when Run is mis-expressed in R2 and R5. The axons do not project into deeper areas of the brain, but stop within the medulla. The phenotype observed in this genetic background is very similar to that for bks. Therefore, in both bks loss-of-function and MT14-GAL4/UAS-run genetic backgrounds, Run expression in R2 and R5 results in the mistargeting of all retinal axon types to the medulla. This also suggests that the targeting of R2 and R5 axons affects axonal pathfinding of other outer R cells. For technical reasons, it was not possible to generate marked, double-lethal clones of run and bks (Kaminker, 2002).

The mistargeting of R-cell axons could, in principle, result from the conversion of all R-cell fates to R7 and R8. Markers for every R-cell type and for cone cells were therefore analyzed, both in mosaic clones of null bks mutant tissue and in the context of MT14-GAL4/UAS-run. In each of these backgrounds, the Run-expressing R2 and R5 cells did not express the R8-specific antigen, Bride of Sevenless (Boss) or the R7 marker, Prospero (Pros). Therefore, the projection phenotype of R cells to the medulla in these backgrounds does not result from the conversion of these R cells to the R7/R8 type during their development. The R1/R6 marker Bar and the R2/R5/R3/R4 marker Rough are also unaffected in these backgrounds. It is concluded that Run reprograms the projection pattern of outer R cells without affecting the expression of developmental markers of cell identity (Kaminker, 2002).

Consistent with the R cell marker expression in larval tissue, plastic sections of adult eyes show that Run misexpression in R2 and R5 during development does not perturb adult R cells or ommatidial structure. The correct complement and arrangement of R cells was found. Strikingly, these seemingly normal adult R cells misproject their axons to the inappropriate optic layer. Axon termination in the lamina region is virtually absent and all R-cell axons project to the medulla. This adult phenotype is also identical to that reported in bks mutant clones in which R cells are unchanged in the expression pattern of Rhodopsins. These data provide strong evidence that Run expression in R2 and R5 causes mistargeting of all outer R-cell axons without changing their individual R-cell fates, as determined by developmental markers and by the adult morphology of rhabdomeres. In spalt (sal) mutants, cell fate is altered without changes in axonal connectivities. Similarly, in sal-GAL4/UAS-run flies, some change in R3/R4 fate to R7 cell type is evident without a significant perturbation of outer R-cell projections. Hence, transcriptional events that control cell identity are separable from those that control axonal targeting (Kaminker, 2002).

R-cell axons provide critical anterograde signals to the lamina target region to induce the proliferation and differentiation of lamina neurons, and to induce the differentiation and migration of glial cells to their correct position adjacent to the lamina plexus. In turn, the glial cells provide positional information that directs R1-R6 axons to terminate in the lamina, and in the absence of glia these axons project into the medulla. The development of the lamina target region was assessed using neuronal and glial cell differentiation markers. Wild-type brains stained with anti-Dachshund (Dac) antibody show a large area of labeled cells corresponding to maturing lamina precursor cells (LPCs) and differentiated lamina neurons. This remains unchanged in GMR-GAL4/UAS-run brains, although the R-cell axons do not target properly. Additionally, the three rows of glial cells that delineate the lamina plexus in wild type also remain unaltered when Run is misexpressed. It is concluded that the abnormal pathfinding of R cells is due to defects intrinsic to R cells and not to a global disruption of lamina target neurons or glia (Kaminker, 2002).

Thus, this study highlights two important characteristics of neuronal pathfinding in the optic lobe. (1) A mechanism involving Bks exists in wild-type flies for repressing Run expression specifically in R2 and R5 cells. The bks loss-of-function causes relief of run repression in these two cells, which entirely abolishes the distinct targeting of R cells to the two optic ganglia. Perhaps additional genes are responsible for repressing run in the other R cells. (2) The mistargeting of R2 and R5 alone is sufficient for all outer R cells to project to the medulla. It appears that R2 and R5 cells, the first two outer R cells to be specified, provide pioneering axons whose tracts the other axons follow. In rough loss-of-function, R2 and R5 are converted to R8 cells. The resulting phenotype, however, is not as severe as that described in this study. Presumably, a residual pioneering function is maintained. The mechanism underlying the interaction between R2/R5 and R3/R4/R1/R6 axons in regulating target layer selection is unclear, although interaction between R1-R6 axons regulates a later step in axonal pathfinding: the fine-tuning of R-cell connections in the lamina (Kaminker, 2002).

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.