BIOLOGICAL OVERVIEW

An important feature of development is the precise positioning of cell types in relation to one another, specifically with reference to the alternation of naked cuticle with cells bearing denticles in the dorsal epidermis of the abdomen of the fly, a phenomenon known as segment polarity. How is pattern achieved during development? Patterning can be viewed from two perspectives: in terms of the cell groups that instruct the fate of distant cells, and in terms of the specific proteins (segment polarity genes) responsible for signaling. Many genes involved in pattern formation were identified by genetic analysis of cuticle patterning in Drosophila (Nusslein-Volhard, 1980). These genes include a small number of ligands, their receptors, and signal transducers. This essay deals with the segment polarity gene lines (lin), originally identified by Nusslein-Volhard (1984) because it affects the dorsal epidermal pattern. Evidence is provided that Lin is involved in stage specific Wingless signaling activity; it acts in the cell receiving the Wg signal. In addition, Lin can localize to the nuclei of cells signaled by Wg-secreting cells. It is hypothesized that Lin interacts with nuclear Wg signal transducers and confers stage specificity to the pathway. Lin also localizes to the cytoplasm of cells receiving the Hedgehog signal, suggesting that Hh competes with Wg signaling by exporting Lin from the nucleus (Hatini, 2000).

The Wg pathway acts in several patterning processes, eliciting varied responses. In fact, even in the same tissue, Wg input elicits distinct responses at different times. For example, early in embryogenesis, Wg input consolidates parasegmental boundaries by maintaining Engrailed (En) expression in adjacent epidermal cells. Late in embryogenesis, Wg input is no longer needed for En maintenance, but rather specifies cell fate. In the ventral epidermis, it specifies the smooth cuticle cell type. It accomplishes this specification by repressing expression of genes required for denticle fate specification, including veinlet (ve; also known as rhomboid), serrate, and shaven-baby. In the dorsal epidermis, Wg specifies a fine hair cell type but the genes mediating this response are not known. The molecular basis for the stage-specific response to Wg signaling is unclear. Because two redundant receptors Frizzled and Frizzled2 mediate all Wg signaling, specificity is not likely to be conferred by multiple receptors with distinct specificities. Thus, transducers of Wg signaling may trigger specific responses by interacting with tissue-specific regulatory proteins, and/or signal transducers activated by another pathway. One such example is the gene teashirt (tsh), that modulates Wg signaling by binding the transcription regulatory domain of Armadillo. Lin is essential for late Wingless signaling activity, acting downstream of Armadillo but upstream of Wg target genes. Because Lin is not required for the early role of Wg signaling, its requirement for late Wg-dependent cell-type specification is stage specific. Moreover, because Lin is required only minimally for late signaling ventrally, but is essential dorsally, its role in Wg-dependent cell-type specification is tissue specific, or specific to a subregion within a tissue. Thus Lin is used at developmentally different times and places to modulate the effects of different factors (Hatini, 2000).

The fly embryonic body plan is subdivided into parasegmental units. Dorsally, rows of epidermal cells adopt four distinguishable cell fates, 1°-4° (primary to quaternary), depending on their position along the parasegment, generating a precise and reproducible cuticle pattern. A single row of cells differentiates large, pigmented denticles: the 1° fate. This row is followed by two to three cell rows producing smooth cuticle: the 2° fate. The next two to three rows secrete pigmented thick hairs, the 3° fate, which are shorter than 1° denticles. The following seven to eight rows secrete fine hairs, the 4° fate, which are longer and less pigmented than 3° cells. Three to four cell rows of smooth cuticle complete the pattern. These cell types reflect alternative fate decisions made by the underlying epithelial cells. Wg and Hh inputs specify these fates and organize the pattern. Hh and the homeodomain protein, Engrailed, are co-expressed in stripes, the posterior row of which produces the 1° fate, whereas the anterior rows adopt the smooth fate. Hh is required across half the parasegment, where cells adopt the 1°-3° fates. When late hh function is blocked, the 1°-3° fates are missing and are replaced by excess 4° fates. Wg is also required across half the parasegment, where cells adopt the 4° fate. Wg is expressed in the anterior, adjacent to the En/Hh domain, in a subset of the cells producing the 4° cell type. To distinguish a possible role for Wg in specifying fate from its earlier role in maintaining En/Hh expression, two different conditions were used, each of which provides for the Wg-dependent maintenance of En/Hh but inactivates or reduces Wg signaling during fate specification. (1) A wg temperature-sensitive allele was inactivated after sufficient Wg signaling had been delivered to maintain En/Hh [6 hr after egg laying (AEL)]. (2) Dominant-negative Pangolin (Pan, the intracellular transducer of the wingless signal) was expressed using Ptc-GAL4, which blocks Wg signaling only in cells flanking the En/Hh domain. When late Wg function is blocked in either of these ways, the 4° cell fate is missing, and the anterior En cells adopt ectopic 1° fate. These cells are flanked anteriorly by smooth cuticle. The parasegment is also narrower, suggesting cell loss. In addition, upon careful examination of these mutant embryos, occasionally it has been found that the domain of 3° cell fates is extended. Compared with two to three rows of 3° fate in wild type, five to six rows are produced in embryos blocked for Wg signaling. To identify additional components mediating Hh and Wg function, a screen was carried out of an existing mutant collection. The lin mutation was selected because the 4° fate is missing, and the domain producing the 3° cell fate is extended from two or three rows to five or six. In addition, the anterior En-expressing cells adopt 1° fate instead of smooth cuticle, and smooth cuticle is produced anterior to these cells. Thus, the pattern in lin mutants is reminiscent of the pattern in embryos blocked for late Wg signaling, suggesting a role for lin in the Wg pathway (Hatini, 2000).

Because Wg expression decays in lin mutants, it was asked whether Wg decay could explain the pattern defects. However, Wg expression decays at 10 hr AEL, while Wg activity is required earlier, between 6 and 9 hr AEL, for dorsal patterning. Nevertheless, it was tested directly whether restoring Wg, by driving UAS-Wg using Arm-GAL4 in lin mutant embryos can restore dorsal patterning, and it does not. As expected, activating Wg signaling in wild-type embryos affects the pattern. Because targeted expression of Wg cannot restore the lin phenotype, it is concluded that lin is required downstream of Wg expression (Hatini, 2000).

The pattern defect in dorsal epidermis suggests that lin mutations affect only late-stage Wg signaling. Because a maternal contribution of lin could mask an earlier role in Wg signaling, embryos were examined that lack both maternal and zygotic lin activity. However, the cuticle phenotype of these embryos is indistinguishable from zygotic lin mutants. Thus, lin is necessary only for late, Wg-dependent cell-type specification in dorsal embryonic epidermis (Hatini, 2000).

Lin and late Wg activities are required for 4° but not 1°-3° cell fates. To determine whether Lin and late Wg function are sufficient for specifying the 4° cell fate, Arm-GAL4 was used to drive ubiquitous expression of either Lin, Wg, or activated Wg signal transducers. Expression of Wg or activation of the Wg pathway globally by driving activated Arm expression leads to the replacement of 1°-3° cell fates with 4° fates, consisting of secreted fine hairs that are longer and less pigmented than 3° cells. Similarly, driving of global Lin expression elicits an identical response, replacing the 1°-3° fates with 4° fates. Thus, Wg and Lin are sufficient for specifying the 4° cell fate when provided at higher levels. One portion of the parasegment is less sensitive to ectopic Lin or Wg activity. Driving Lin or Wg to high levels of expression in cells posterior to the En domain through the use of Ptc-GAL4 is sometimes not sufficient to transform the 2° fate to 4°. This domain is where Hh activity is normally highest, suggesting that Hh signaling competes with Lin and Wg in fate selection (Hatini, 2000).

Lin may contribute to Wg signaling either by controlling the production of a signal or by transducing a signal. These possibilities can be distinguished, because if Lin acts by producing a signal, restoration of restricted expression of Lin in lin mutants will restore the 4° cell fate non-autonomously. If Lin is required for transducing the signal, the 4° fate will be restored autonomously only in cells expressing Lin. In lin mutants, the anterior- and posterior-most En rows differentiate the 1° denticle type, while two to three cell rows internal to the En domain differentiate smooth cuticle. Use of En-GAL4 and UAS-Lin to target Lin to the En domain transforms all the En-expressing cells cell-autonomously to the 4° cell fate. Cells flanking the En domain are not affected. Similarly, targeting of Lin to the Wg domain in lin mutants by use of Wg-GAL4 restores the 4° fate, but only in that portion of the 4° field that expresses Wg. Two to three rows of cells anterior to the Wg domain still adopt 3° fate. Restoration of Lin activity in Wg-producing cells does not restore the full pattern. Thus, these two experiments indicate that Lin generally does not regulate expression of a signal, but rather acts cell-autonomously to specify the 4° cell fate (Hatini, 2000).

There is one instance where Lin may regulate a signal. Expression of Lin in the Wg domain rescues the ectopic 1° denticle row found in the En domain of lin mutants. Thus, Lin has a second role within Wg-expressing cells, and, in this case, it acts non-autonomously to specify the (smooth) fate of the posteriorly adjacent row of En cells (Hatini, 2000).

Evidence for the suggestion that Lin acts in some way to specify dorsal cell fate has to this point relied on cuticle analysis, which represents the final differentiated state of the cells, first visible at about 13 hr AEL. However, Wg signaling specifies these cell fates between 6 and 9 hr AEL. Thus, to test whether or not Lin acts in concert with Wg, earlier molecular markers for Wg patterning need to be identified and the effects of Lin activity on these markers need to be tested. The first Wg-dependent target gene is wg itself. If Wg function is blocked late by expression of dominant-negative Pan with the Ptc-GAL4 driver, late wg expression is lost from both the dorsal and the ventral epidermis. Reciprocally, if Wg signaling is activated by driving of activated Arm expression with the Ptc-GAL4 driver, an ectopic Wg stripe is induced posterior to the En domain. Thus, wg gene expression depends on Wg input and provides a molecular readout for the pathway. In lin mutants, late wg expression fades from the dorsal epidermis of fully retracted embryos (10 hr AEL). Reciprocally, overexpression of Lin using the Ptc-GAL4 driver activates wg expression posterior to the En domain in the dorsal epidermis. The ectopic expression of wg is identical to that obtained by expression of activated Arm posterior to the En domain. The only distinction is that the effect of Lin is restricted to the dorsal epidermis, while global activation of Wg signaling affects the ventral epidermis as well. Thus, Wg input and Lin are both necessary and sufficient for activation of wg gene expression in dorsal epidermis (Hatini, 2000).

The second Wg-dependent target gene is ve (veinless or rhomboid), which is expressed in a row of cells posteriorly adjacent to the En/Hh-expressing cells. This spatially restricted pattern is regulated in ventral epidermis by Wg signaling. Wg regulates ve expression dorsally as well. For example, if Wg function is inactivated at late stages, ve is ectopically expressed anterior to the En domain. Reciprocally, if the Wg pathway is broadly activated, ve expression is repressed. Thus, ve expression also provides a molecular readout for the Wg pathway. In lin mutant embryos, a second stripe of ve is induced anterior to the En/Hh cells in the dorsal epidermis. Thus, Lin and Wg function are similarly required to repress ve gene expression. It is concluded that Lin acts in concert with Wg in regulating target genes and consequently patterning the dorsal cell types (Hatini, 2000).

Because Lin acts cell-autonomously upstream of Wg target genes, it was determined where along the Wg signaling pathway Lin function is required. This determination was accomplished by forcing activation of cytoplasmic Wg signal transducers in lin mutants and testing cuticle pattern and ve gene regulation. Wg signaling was activated by either overexpressing Dishevelled or using constitutively activated Arm, with the UAS-GAL4 system. Activation of Wg signaling in wild-type embryos leads to global specification of the 4° cell fate across the dorsal epidermis. In contrast, this specification does not occur if lin function is removed. In fact, the pattern resembles the lin mutant pattern. For example, rather than adopting 4° fate, 1° and 2° fates are still established, as is the ectopic 1° fate and the smooth fates anterior to these cells. Cells in the remainder of the parasegment resemble immature 3° rather than 4° cells. For example, the cuticle protrusions are shorter and more pigmented than 4° cells, but their base is narrower than fully mature 3° cells. The expression of a molecular marker is consistent with the cuticle pattern, because now ve is not repressed by increased Wg signaling if lin activity is removed. Ventrally, Wg activation still represses ve even when lin activity is removed, further demonstrating the restriction of the role of Lin to dorsal Wg signaling. Thus, in dorsal epidermis, Lin is crucial for completing Wg signal transduction, acting downstream or in parallel to Arm (Hatini, 2000).

The epistasis also is consistent with Lin being a downstream target gene of the Wg pathway. However, this model is unlikely since lin is still expressed in embryos blocked for Wg signaling, and lin expression is not up-regulated in embryos globally activated for Wg signaling. Therefore, Lin could act either in a parallel pathway to the Wg pathway, or may cooperate with Wg signal transducers (Hatini, 2000).

If Lin is acting strictly in parallel to Arm, then both Wg input and Lin must act for normal patterning. If, however, Lin is required downstream of Wg signaling then overexpression of Lin in embryos blocked for Wg signaling will bypass the need for Wg input and restore Wg-dependent readouts. The dorsal epidermis in embryos null for Wg signaling is poorly differentiated and occasionally missing. Expression of Lin in these embryos does not rescue the pattern. Because lin is not needed for early Wg signaling perhaps it is not surprising that Lin is not capable of restoring En expression in the absence of Wg. Expression of Lin in embryos blocked for late Wg signaling may not allow distinguishing between the two models posed above because these embryos might contain residual Wg signaling activity. To examine this issue, Lin was overexpressed using Ptc-GAL4 in wg^ts embryos shifted to the nonpermissive temperature at 6 hr AEL, or in embryos co-expressing Lines and dominant-negative Pan, and their cuticle pattern was examined. In either experiment, the usual effect on patterning caused by the defect in Wg signaling is blocked in embryos co-expressing Lin because the 4° cell fate is restored. For example, anterior to the En domain, smooth cuticle is replaced with the 4° type. In addition, the 4° cell fate is restored posterior to the 2° domain. Expression of Lin in embryos expressing dominant-negative Pan with the ubiquitous Arm-GAL4 driver replaces all cell fates with 4° cell fate. This result may have two interpretations. High levels of Lin may trigger a response independent of Wg input. Alternatively, any residual Wg signaling that may be present in these embryos allows Lin to function (Hatini, 2000).

In the current model for Wg signaling, Arm regulates Wg target gene expression in association with the DNA-binding protein Pangolin. Because the epistasis shows that Lin regulates Wg target gene expression downstream of Arm, Lin protein would be expected to be found in nuclei. Crude and purified polyclonal antisera weakly detects Lin protein by Western blot from embryonic extracts, but these sera fail to detect the protein in situ. The lin genomic rescue transgene was modified to include a Myc tag at the carboxyl terminus of the Lin protein. The modified transgene rescues lin mutants to adulthood; however, Myc-tagged Lin protein could not be detected in situ. It is concluded that Lin is expressed below the detection levels of both the anti-Lin and the commercial anti-Myc antibodies. However, by use of the UAS-GAL4 system, Lin protein can be detected by antibodies in situ in embryos overexpressing Lin. For instance, embryos carrying UAS-Lin and the Ptc-GAL4 driver show the expected high levels of Lin protein both anterior and posterior to the En expression domain at 10 hr AEL. However, Lin exhibits a different subcellular localization in each expression domain. Lin is nuclear in cells anterior to the En domain, in exactly those cells dependent on Lin and Wg function. Lin is preferentially cytoplasmic posterior to the En domain, which may help explain why excess Lin sometimes does not alter patterning in this domain. However, it must be assumed that low levels of Lin do enter the nucleus, as sometimes these cells are transformed to the 4° fate. Expression of Lin with Arm-GAL4 shows that Lin is nuclear in half of the parasegment anterior to the En domain, and cytoplasmic in the other half posterior to the En domain (Hatini, 2000).

To determine whether the differential localization is signal-dependent Lin was co-expressed with Wg to ectopically activate Wg signaling. In this case, Lin accumulates exclusively in nuclei both anterior and posterior to the En domain. To decrease Wg signaling, dominant-negative Pan was expressed with Ptc-GAL4. Due to feedback regulation, this expression leads to reduced Wg expression, and therefore reduced Wg signaling. In such embryos, the differential localization of Lin is altered, but mostly at lateral positions in the embryo, where Lin occasionally localizes to the cytoplasm, anterior to the En domain. In the dorsal epidermis, Lin is still nuclear, which is consistent with the fact that although Pan function is compromised in these embryos, excess Lin still promotes the 4° fate. It is concluded that Wg signaling can influence the subcellular localization of Lin. In addition, the nuclear localization of Lin supports the genetic model that Lin is needed to complete transduction of the Wg signal, probably by cooperating with Arm/Pan in the nucleus. Interestingly, in the ventral epidermis, where Lin is required on either side of the En domain, Lin is nuclear both anterior and posterior to that domain. Thus, Lin subcellular localization is correlated with its spatial requirement (Hatini, 2000).

Three models are proposed for Lin function in Wg-dependent cell-fate specification. (1) Because Lin protein is nuclear in cells in which it acts, Lin may directly regulate the transcription of Wg target genes by interacting with DNA or DNA-bound proteins. For instance, Arm and Pan may act at one site on target genes, while Lin acts directly at an adjacent site. Because Lin does not have an obvious DNA-binding domain, it may act in combination with another DNA-binding protein. Pan bound to one site could easily be imagined to modulate the function of Lin at a second site, since LEF1/TCF proteins related to Pan can play an architectural role, regulating the assembly of multiprotein enhancer complexes. This scenario is reminiscent of the midgut Ubx enhancer, which contains a Wg-responsive Lef-1 DNA-binding site, and a Dpp-dependent cyclic AMP-responsive element. The identification of response elements in the target genes ve and wg will be needed to investigate whether the regulation by Arm/Pan and Lin is direct and to reveal the precise mechanism at work. It is also possible that Lin interacts directly with Pan and/or Arm, rather than participating in a distinct complex. However, initial experiments with the two-hybrid system or immunoprecipitation have not provided evidence for direct interactions with Arm (Hatini, 2000).

(2) In a second model, Lin may regulate transcription by creating a state permissive for Wg signal transducers to act. For example, Lin itself may act by remodeling chromatin structure, allowing efficient access of transcription factors to Wg-regulated promoters, leading to an increased activity of DNA-bound Pan/Arm complexes (Hatini, 2000).

(3) Finally, Lin may regulate transcription factor activity by affecting the modification state of factors such as Arm and or Pan. Precedent exists for this: in Caenorhabditis elegans, the Lit1 MAP kinase regulates the function of Wrm1 and Pop1, proteins related to Arm and Pan, respectively (Hatini, 2000 and references therein).

Although Lin can promote Wg signaling in the dorsal epidermis, it is not sufficient to do so ventrally. Thus, Lin can promote Wg signaling only in a competent region of cells. The activity of Lin dorsally must depend on additional factors that provide this region with the competence to modulate Wg signaling in response to Lin. Such factors could provide a biochemical link between Lin and the Arm/Pan complex (Hatini, 2000).

Whichever model holds, note that the role of Arm and Pan is more clearly understood for transcriptional activation events, where Arm is proposed to confer an activation function to the otherwise negative regulator Pan. Only a few genes have been characterized as being repressed by Wg signal transduction, for example, dpp and wg in developing leg and wing, respectively, and veinlet. Unfortunately, response elements mediating Wg-dependent repression have not yet been identified. It is likely that the Arm/Pan complex can recruit repressors or activators to Wg target genes, perhaps in a manner analogous to Smad proteins, which can activate by recruiting the co-activator CBP/p300 or repress by recruiting co-repressors such as TGIF or Ski, which, in turn, bind histone deacetylases. The identification of ve as a gene repressed by Wg input will eventually provide tools to investigate negative regulation by the Wg pathway (Hatini, 2000).

Because Lin is essential for Wg-dependent cell-type specification acting downstream or in parallel to Arm, and because Lin nuclear localization is influenced by Wg, a particularly attractive model involves Arm in the localization of Lin. Arm/beta-catenin share homology with the importin/karyopherin family of transport receptors and can bind the nuclear pore machinery (Fagotto, 1998), raising the possibility that one function of Arm/beta-catenin is to import regulatory proteins to the nucleus. Perhaps at the same time as Wg signaling stabilizes Arm, Wg associates with Lin in the cytoplasm and transports Lin across the nuclear membrane. The association of Arm with other regulatory proteins may also promote their nuclear import. Alternatively, because the phosphorylation state of a protein can lead to transport into or out of the nucleus, perhaps Wg signaling affects Lin localization through the inhibition of the serine/threonine kinase Shaggy. In this case, Lin nuclear import would be regulated by an upstream component of the Wg pathway and then interact with other downstream components (Arm/Pan) in the nucleus (Hatini, 2000).

Wg signaling has two distinct roles in patterning the embryonic epidermis. The first is to prevent an En-dependent signal from crossing over anteriorly. When Wg signaling is deficient, a signal from the En domain, such as Hh, crosses over and specifies cell fate anterior to it, creating a mirror image ve expression and cuticle pattern (smooth-1°/1°-smooth). In lin mutants, ve expression and pattern are similarly reorganized, indicating that lin is necessary for this role of Wg. The second role of Wg is to specify cell fate in half of the parasegment, an event that is also blocked in lin mutants (Bokor, 1996). Thus, these two patterning roles of Wg are compromised in lin mutants (Hatini, 2000).

Normally, Hh patterns posteriorly from its source, specifying the 1°-3° cell fates, whereas Wg patterns to the anterior, specifying the 4° cell fate. The domains of Wg and Hh influence meet in the middle of the parasegment, at approximately the interface of 3° and 4° cell types. A balance between Wg and Hh signaling is essential for proper patterning in this region since changing the strength of either can alter the pattern in the middle of the parasegment. For example, reduction of Wg function results in excess 3° cells at the expense of 4° cells. Similarly, a gradual increase in Hh signaling leads to a gradual increase in the number of 3° cells at the expense of 4° cells. It is proposed that Lin helps govern the balance between Wg and Hh influence by promoting Wg signaling activity, which can thereby better compete with Hh function. An increase of Lin activity can alter this balance, leading to excess 4° cells at the expense of 3° cells. At higher levels of Lin, the entire domain under Hh influence differentiates as 4° cells. Aside from the contribution of Lin levels to the pattern, the level of Hh is also crucial. Hh may contribute to this balance by antagonizing Lin function. The antagonism will be highest posterior to the En domain where Hh signaling is the highest. In fact, further evidence is found for this antagonism, because the expression of high levels of Lin in this region is not always sufficient to replace the 2° with 4° fate (Hatini, 2000).

It is proposed that, in wild type, Hh antagonizes Lin activity in the middle of the parasegment, thereby toning down Wg signaling activity in this region, and allowing these cells to adopt 3° rather than 4° fate. One way by which Hh may antagonize Lin function, and thereby Wg signaling, is by exporting Lin to the cytoplasm (Hatini, 2000).

GENE STRUCTURE

mRNA length - 3416

Bases in 5' UTR - 280

Bases in 3' UTR - 559

PROTEIN STRUCTURE

Amino Acids - 858

Structural Domains

Lines is a novel protein with no similarity to any other characterized proteins in the NCBI database (Hatini, 2000).

date revised: 30 June 2000

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.