BIOLOGICAL OVERVIEW

The mechanism by which the Wingless signal is received and transduced across the membrane is not completely understood. The arrow gene function is essential in cells receiving Wingless input. arrow acts upstream of Dishevelled and encodes a single-pass transmembrane protein; this indicates that it may be part of a receptor complex with Frizzled class proteins. Arrow is a low-density lipoprotein (LDL)-receptor-related protein (LRP), strikingly homologous to murine and human LRP5 and LRP6. Thus, a new and conserved function is suggested for the LRP subfamily in Wingless/Wnt signal reception (Wehrli, 2000).

In each embryonic segment, Wingless is expressed anterior to cells expressing Engrailed which co-express the secreted signaling protein Hedgehog. Most of the pattern across each segmental field is organized by these two signals. For example, Wg has two roles in patterning the segment. (1) Between 3 and 6 hours of embryonic development, early Wg signaling is necessary for the continued expression of En and Hh in the adjacent cell row. This ensures the maintenance of the parasegmental subdivisions of the body axis, generating the segmental body plan observed in larval cuticle preparations. (2) Between 6 and 9 hours of development, Wg signaling assigns specific cell fates. Within each segment, ventral epidermal cells secrete either smooth cuticle or protrusions called denticles. The double row of En/Hh-expressing cells straddles the boundary between smooth cuticle and the first denticle row. Wg signaling instructs anterior En cells to adopt smooth cell fate. When Wg function is artifically eliminated after 8 hours of development, the anterior En cell incorrectly adopts a denticle fate. Similarly, anterior En cells adopt denticle fates in arrow mutants. These embryos have wild-type segmentation because early Wg signaling proceeds normally in zygotic arrow mutants. However, when both maternal and zygotic arrow function is removed (referred to as arr^null embryos), the resulting embryos are indistinguishable from wg null mutants. As the maintenance of En expression depends on the reception of the early Wg signal, arr^null embryos were assayed for En maintenance. En expression is initiated properly in wg and arr^null embryos but then fades. In addition, Wg protein is distributed normally in arr^nullmutant embryos, as compared with wild-type embryos. In particular, in arr^null embryos Wg is still detectable within cells distant from the Wg-expressing cells. Thus, arr mutations do not affect the production of ligand, nor its distribution across the epithelium, but nevertheless block the two Wg-dependent patterning events in embryonic epidermis. Furthermore, arrow function is required for all other Wg signaling events that have been tested, such as in embryonic midgut, as well as in all facets of imaginal disc patterning. Thus, arrow conforms to the expectation for a gene encoding an essential component of the Wg pathway (Wehrli, 2000).

Since Arrow appears not to be involved in the production of Wg, Arrow might be required in cells receiving Wg input. To test this, the expression of Wg target genes was examined in tissues mosaic for arrow function. If Arrow is required to produce active signal, then arrow mutant cells will be rescued by signal produced from adjacent Arrow-positive cells. However, if Arrow is necessary in the receiving cell, then all the arrow mutant cells will exhibit a defect in Wg target gene expression. Wg is expressed along the dorso-ventral boundary of the wing disc. Secreted from these cells, Wg activates the transcription of Distalless (Dll), and acts in a concentration-dependent manner to pattern the wing. In the wing pouch, none of the arrow mutant cells within a clone expresses Dll. Since mutant cells at the edge of the clone are not rescued by Wg secreted from their wild-type neighbors, arrow function is necessary in the cell receiving Wg. This is supported by analysis in the leg disc, where Wg is expressed in a ventral quadrant, and specifies ventral and ventro-lateral fates. Proper patterning also requires that dpp, a secreted factor necessary for dorsal and dorso-lateral fates, is repressed in cells receiving Wg. If Wg signal transduction is blocked, as in clones of cells mutant for dsh, dpp expression is de-repressed and patterning is disrupted. Similarly, dpp expression is de-repressed in clones of cells mutant for arrow in the ventral anterior quadrant. These analyses show that arrow is necessary in cells responding to Wg input, for both positive and negative gene regulation (Wehrli, 2000).

Having established that arrow acts in the responding cell, the requirement for arrow was ordered relative to the intracellular transduction cascade. Smooth cuticle is restored to arr^null embryos in alternate segments when Dsh is expressed using Prd-GAL4, indicating rescue of Wg signal transduction. This contrasts with the overexpression of Wg, which has no effect in arr^null embryos. These data suggest that Arrow acts downstream of Wg but upstream of Dsh, because signaling, once activated by Dsh, no longer requires Arrow. It remains possible that Arrow might normally act as a scaffold and concentrate Dsh to an appropriate subcellular location; in this model, flooding the cell with Dsh simply bypasses the requirement for Arrow. In either case, Arrow is unlikely to act in a pathway parallel to Wg signal transduction since (1) the arrow mutant phenotype is rescued by Dsh, a canonical Wg signal transducer, and (2) loss of arrow function blocks signaling even when excess Wg is presented. In addition, the fact that the arrow phenotype is not suppressed by excess Wg distinguishes arrow from the genes involved in proteoglycan-assisted presentation of the Wg ligand. Each of the mutants affecting this step, sugarless, sulfateless and dally, is substantially suppressed by providing excess Wg ligand, showing that glycosaminoglycans are not essential for reception of the signal but only increase the efficiency of ligand presentation to the receptor. Together, these data argue that Arrow is absolutely essential for Wg to signal, and are consistent with a role for Arrow in reception rather than presentation of signal (Wehrli, 2000).

The mosaic and epistasis analyses place the requirement for Arrow activity upstream of Dishevelled in responding cells. Because arrow encodes a putative transmembrane protein, epistasis tests between Arrow and Fz proteins would be of interest, but no activating mutations exist in either Fz or Arrow that would lead to signal transduction in the absence of ligand. There are, however, several suggestive similarities between Fz proteins and Arrow. Arrow and DFz2 transcription is modulated similarly, whereas that of the Dsh and Arm signal transducers is not. In addition, ectopic DFz2 expression can mildly sensitize cells to Wg signaling, for example in the wing, where overexpression of DFz2 produces ectopic margin bristles -- a Wg-dependent cell type. This potentiation of Wg signaling is ligand-dependent, and ectopic bristles are only found near the wing margin, which is a source of high levels of Wg. As is expected for a gene essential for Wg signaling, loss of arrow function in clones leads to loss of margin, similar to that observed for fz Dfz2 clones. Overexpression of Arrow also produces ectopic bristles near the wing margin. Thus, arrow is required for Wg-dependent signaling at the margin, and can potentiate Wg signaling in a manner similar to that of DFz2. The potentiation of signaling caused by excess DFz2 depends on arrow function, as shown by inducing arrow mutant clones while overexpressing DFz2. If excess DFz2 were able to bypass Arrow and restore signaling on arrow mutant cells then these cells should retain the ability to form wing margin and produce marked arrow mutant margin bristles, neither of which was observed. This result is consistent either with Arrow functioning after DFz2 engages ligand, or with Arrow and the Fz class proteins functioning as co-receptors, although this epistasis must be confirmed when activating mutations become available (Wehrli, 2000).

The extensive homology between Arrow and both mouse and human LRP5 and LRP6, indicates that the role ascribed to Arrow in Wg signaling may extend to these LRPs for Wnt signaling. Indeed, an insertion in the mouse LRP6 gene has been identified that leads to several Wnt-like phenotypes (Pinson, 2000). Arrow, and by extension LRP5 and LRP6, have specific roles for the Wg/Wnt pathway, as both Dpp and Hh can signal to arrow null cells (Wehrli, 1998). arrow null mutant cells do not survive as well as their wild-type sister cells when twin spot clones are made, suggesting some role in viability. Whether this can be attributed to a deficit in Wnt signaling is not known, although cells doubly mutant for fz and dfz2 also do not survive (Wehrli, 2000).

The simplest model is that Arrow and a Fz-class protein act together as a receptor complex. Alternatively, Arrow might assist in recycling the Fz receptors to the plasma membrane after Wg/Wnt ligand binding and internalization, thereby providing unoccupied receptors to allow efficient, extended signaling. This idea is suggested by the fact that Arrow is related to LDL receptors, which are sometimes involved in recycling proteins from the plasma membrane. If this is the case, then overexpression of Fz proteins should suppress the need for arrow function by supplying excess nascent receptors. However, overexpressing DFz2 at the wing margin does, in fact, not suppress the Wg signaling defect caused by loss of arrow function (Wehrli, 2000).

Since several components required for Wg signaling, such as Dsh, Sgg and Fz, are also necessary for the determination of tissue polarity, a test was performed to see whether Arrow also acts in this process. The only tissue where arrow related polarity defects are found is the eye. However, the role for Arrow in polarity in this tissue is probably indirect, reflecting an early, global role for a Wnt pathway upstream of planar polarity determination. In other tissues, Arrow is not involved in tissue polarity. For example, arr mutant clones in the dorsal leg exhibit no such defects. This result also shows that the Fz receptor is available to respond to tissue polarity signals in the absence of arrow function and one would therefore expect it to be at the cell surface in arrow mutant cells. This finding suggests that Arrow does not merely act as a chaperone to guide Fz-class proteins to the cell surface. This again is consistent with the model that Fz and Arrow have a similar role in the reception of the Wg signal and that they may both be part of a single receptor complex. Recent work raises the possibility that other LRPs are part of a signal reception complex, just as is postulated for Arrow. The VLDL and ApoE receptors (both LRPs) are essential during mouse cerebellar development, where they bind the ligand, Reelin, while intracellularly binding to, and inducing the phosphorylation of, the adapter protein, Disabled-1. In addition to the LRPs, a second receptor—the Cadherin-related neuronal receptor—also binds Reelin, while its intracellular domain associates with the Fyn tyrosine kinase. This suggests that wheen the two receptor subunits bind Reelin, two proteins are brought into proximity inside the cell, Disabled-1 and Fyn. Perhaps Arrow and the Fz proteins similarly bind ligand and consequently bring together proteins in the cytoplasm that initiate Wg/Wnt signaling (Wehrli, 2000).

PROTEIN STRUCTURE

Amino Acids - 1678

Structural Domains

arrow encodes a protein that demonstrates striking sequence conservation to the mammalian LDL-receptor related proteins LRP5 and LRP6. Arrow is a putative type I transmembrane protein containing four epidermal growth factor (EGF)-like repeats, each preceded by six YWTD spacer domains. Between the EGF/YWTD region and the membrane are LDL-receptor (LDLR) type A repeats. Both classes of repeats have been implicated in ligand binding in other contexts. For example, the LDLR repeats in the LDL receptor are central to binding the LDL particle, whereas EGF-like repeats contribute to ligand binding by some LRPs. The cytoplasmic tail contains proline-rich regions that are potential targets for SH3-domain-containing proteins, and the sequence YKII is a potential signal for internalization (Wehrli, 2000).

Arrow, LRP5 and LRP6 form a distinct group of LRPs based on overall sequence conservation. The conservation among Arrow, LRP5 and LRP6 is not restricted to the repeats, but is found throughout the proteins. In addition, the blocks of EGF-like and LDL-receptor repeats are in reversed order as compared with other LRP members. Finally, the EGF-like repeats of Arrow are slightly more closely related to those of LRP5 and LRP6 than to those in other LRP proteins. The same is true for the LDL receptor repeats. These comparisons indicate a significant functional conservation between Drosophila Arrow and human LRP5 and LRP6 (Wehrli, 2000).

date revised: 22 February 2022

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.