org Interactive Fly, Drosophila gurken: Biological Overview | Regulation | Factors affecting Gurken RNA localization and translation | Developmental Biology | Effects of Mutation | References

Gene name - gurken

Synonyms -

Cytological map position - 29C

Function - ligand for Torpedo

Keywords - Dorsal group

Symbol - grk

FlyBase ID:FBgn0001137

Genetic map position - 2-30

Classification - TGF-alpha-like

Cellular location - cytoplasmic

NCBI links: Precomputed BLAST | Entrez Gene | UniGene

Recent literature
Norvell, A., Wong, J., Randolph, K. and Thompson, L. (2015). Wispy and Orb cooperate in the cytoplasmic polyadenylation of localized gurken mRNA. Dev Dyn [Epub ahead of print]. PubMed ID: 26214278
In Drosophila, the dorsal-ventral (D-V) axis of the oocyte is dependent on Gurken (Grk) protein distribution. This is achieved through the cytoplasmic localization of grk mRNA and regulation of its translation. During mid-late stages of oogenesis, grk mRNA and protein are localized to the dorsal-anterior of the oocyte, while unlocalized grk transcripts are translationally silenced. As females carrying mutations in the gene encoding the CPEB protein Orb lay ventralized eggs due to insufficient Grk levels, it seems likely that cytoplasmic polyadenylation of grk transcripts may play a role in their translational regulation. This study found that grk is polyadenylated throughout oogenesis, with poly(A) tails of approximately 30-50 A residues. Hyperadenylated grk transcripts, with poly(A) tails of 50-90 As, are detected in late stage egg chambers, but they fail to accumulate in oocytes deficient in Orb or the poly(A) polymerase Wispy (Wisp). wisp females also lay weakly ventralized eggs, demonstrating that they produce inadequate amounts of Grk. Finally, unlocalized grk transcripts are also not appropriately hyperadenylated. Thus, localized cytoplasmic polyadenylation of grk mRNA by Wisp and Orb is necessary to achieve appropriate Grk protein accumulation in the D/A corner of the oocyte during mid to late oogenesis.

Two signals are responsible for dorsoventral patterning in the egg and embryo (Schümpbach, 1994). The first signal requires both the ligand Gurken and the receptor Torpedo (EGF-R). Carried from the oocyte to follicle cells (maternal cells that surround the egg), it is a signal to enter a dorsal differentiation program. This signal delimits the ventral zone within the follicle cell epithelium. The formation of the ligand of the second signal (Spätzle) is thus spatially restricted to a zone in the ventral part of the egg. When fertilization takes place, Spätzle is activated and triggers the receptor Toll to activate the transcription factor Dorsal in the ventral cells of the developing embryo.

Gurken mRNA becomes localized to the anterior dorsal part of the oocyte, between the nucleus and the overlying oocyte membrane. A number of genes are required to localize GRK mRNA including cornichon, fs(1)K10, orb, squid, cappuccino and spire. Gurken is the ligand for the Torpedo/EGF receptor, located on the surface of follicle cells that envelope the oocyte. GRK triggers EGF-R/Torpedo signaling and the consequent modification of cell fate in the follicle cells.

One follicle cell gene regulated by Torpedo signaling is rhomboid, a gene involved in the specification of dorsal cell fate in dorsal follicle cells (Ruohola-Baker, 1993). Gurken signaling involves a feedback from follicle cells to the oocyte, resulting in a correct polarization of the oocyte's anterior-posterior microtubule cytoskeleton. The mechanism underlying this feedback is currently not known (Roth, 1995).

The establishment of dorsal cell fate in follicle cells is important in the later differential ability of ventral follicle cells to activate the second dorsal-ventral signal triggered upon fertilization, signaling through the Toll receptor to activate the Dorsal transcription factor in ventral cells.

Thus, through the action of Gurken, the dorsal-ventral polarity of the egg is established well in advance of fertilization and embryonic development. The interaction between the oocyte and its enveloping follicle cells is critical for the future development of the embryo.

In addition to its role in establishing dorsal-ventral polarity, the Gurken-EGF-R interaction has a prior role in the induction of posterior follicle cell fate and the specification of the egg's anterior-posterior polarity. Early Gurken is localized to the posterior pole of the oocyte. Gurken mutant egg chambers lack posterior follicle cells and show a mirror-image duplication of the AP axis of the oocyte. In addition, a large proportion of Egf-r mutants develop an AP phenotype very similar to that produced by grk mutation. It is believed that AP polarity arises from the movement of the oocyte to the posterior of the nurse cells within the egg chamber. The germinal vesicle (egg nucleus) and GRK mRNA both become localized to the posterior of the oocyte, leading to the polarized production of Gurken. EGF-R becomes activated in adjacent polar follicle cells to determine posterior follicle cell fate. The posterior follicle cells subsequently signal back to the oocyte to repolarize the oocyte microtubule cytoskeleton, probably by inducing the dissembly of the microtubule organizing center at the posterior of stage 6 oocytes. A new microtubule nucleating activity is generated at the anterior margin of the oocyte, and as a consequence the germinal vesicle moves to the anterior margin of the oocyte, ready for the subsequent involvement of Gurken in establishing dorsal cell fate (González-Reyes, 1995 and Roth, 1995).

decapentaplegic is required for patterning of anterior eggshell structures, reflecting expression of dpp in anterior somatic follicle cells. In grk mutant females, ectopic expression of dpp occurs at the posterior pole of 77% of the mutant egg chambers. sax mutant egg chambers exhibit similar posterior dpp expression. This suggests that SAX receptor function is required for acquisition of at least some posterior fate, and also suggests that dpp is the target of gurken-Egfr signaling (Twombly, 1996).

Gurken is here shown to induce two different follicle cell fates because the follicle cells at the termini of the egg chamber differ in their competence to respond to Gurken from the main-body follicle cells in between. Anterior follicle cells are known to become subdivided into three distinct follicle cell types along the anterior-posterior axis: border cells, stretched follicle cells and centripetal follicle cells. The border cells are a group of 6-10 cells that delaminate from the follicular epithelium at the anterior tip of the egg chamber and migrate between the nurse cells to the anterior of the oocyte. At the same time, the adjacent stretched follicle cells spread to cover the nurse cells as the rest of the follicular epithelium moves posteriorly to envelop the oocyte. The centripetal follicle cells just posterior to the stretched follicle cells come to lie over the anterior of the oocyte after these movements are complete, and these cells then migrate between the oocyte and the nurse cells toward the center of the egg chamber during stage 10b (Gonzalez-Reyes, 1998).

It is argued that the terminal follicle cell populations (consisting of both anterior and posterior follicle cell populations) are equivalent prior to gurken signaling. To explain how Gurken can induce two different follicle cell fates, it has been proposed that the follicle cell layer is divided into two cell types during early oogenesis: the terminal follicle cells at each end of the egg chamber, which become posterior if they receive the Gurken signal and anterior if they do not, and the main-body follicle cells, which are induced to become dorsal rather than ventral. To determine whether the main-body follicle cells can adopt a posterior fate in response to Gurken, the original dicephalic mutation was analyzed. Unlike the spindle mutants, dicephalic alters the position of the oocyte without affecting Gurken signaling. This allows for a test of the fate of anterior, main-body and posterior follicle cells by allowing for the exposure of each of these populations to the Gurken signal. The induction of posterior follicle cell fate is followed directly, using a recently identified enhancer trap line that specifically labels these cells. When the oocyte is correctly positioned at the posterior of dicephalic mutant egg chambers, the enhancer trap line is strongly expressed in the posterior follicle cells just as it is in wild type. In contrast, no expression is observed when the oocyte lies in the middle of the mutant egg chamber. Thus, misplaced oocytes cannot induce main-body follicle cells to adopt a posterior fate, arguing that main-body and posterior follicle cells possess two different fates (Gonzalez-Reyes, 1998).

Although it has been clearly demonstrated that the posterior follicle cells adopt an anterior fate if they do not receive the Gurken signal, it has not been shown that the anterior terminal cells are competent to adopt a posterior fate. The dicephalic mutant egg chambers where the oocyte lies at the anterior of the cyst have been examined. In this case, the posterior marker is expressed in the anterior follicle cells of these egg chambers; this confirms that misplaced oocytes can still signal to induce posterior fate and confirms the equivalence of anterior and posterior follicle cells. Furthermore, these egg chambers develop completely normally although they have a reversed polarity with respect to the anterior-posterior axis of the whole ovariole. The terminal follicle cells at the opposite end of the oocyte become subdivided into the three anterior follicle cell types which undergo their normal migrations, whereas the 'anterior' cells in contact with the oocyte not only express the posterior marker, but also signal to polarize the oocyte cytoskeleton, since all of the maternal mRNAs tested localize to the appropriate position. These results demonstrate that the main-body follicle cells and the terminal follicle cells do indeed constitute two distinct populations that differ in their competence to respond to Gurken and prove that the terminal follicle cells at each end of the egg chamber are equivalent prior to Gurken signaling (Gonzalez-Reyes, 1998).

The Egfr, as receptor of the posterior Gurken signal, is required cell autonomously to repress anterior fate in all posterior follicle cells. Although the expression of several markers at the termini of developing egg chambers suggests the existence of populations of terminal follicle cells, it is not clear how many cells respond to Gurken directly by adopting a posterior rather than an anterior fate. To define this population, a mapping was performed to determine which cells revert to the default anterior fate when they cannot respond to Gurken because they lack its putative receptor. Small marked clones of cells were generated that are homozygous for top CO, a null allele of the Egfr, and their fate was followed by staining for the beta-gal activity of the L53b enhancer trap line, which labels all three subpopulations of anterior follicle cells from stage 9 onwards. When the clones are generated (at approximately stage 2 of oogenesis) and scored at stage 10, mutant cells that lie near the posterior of the oocyte are seen to always express L53b, whereas clones over the middle of the oocyte do not. Thus, removal of the Egfr causes a cell-autonomous transformation from posterior to anterior fate, indicating that Gurken signals directly to induce posterior fate in the whole terminal follicle cell population. With one exception, all Egfr- cells that fall within 10-11 cell diameters of the posterior end of the egg chamber express L53b, whereas mutant cells that fall anterior to this boundary do not. This analysis indicates that about 200 terminal follicle cells receive the Gurken signal directly, ruling out a model in which only the polar follicle cells (the most posterior cell population) are competent to respond to Gurken by becoming posterior. The cells that become anterior if they cannot respond to Gurken constitute the entire population of follicle cells that contact the oocyte during previtellogenic stages. Thus, all of the cells that receive the 'posteriorizing' Gurken signal are competent to respond to it (Gonzalez-Reyes, 1998).

In mutants such as gurken in which the induction of posterior follicle cell fate is blocked, the terminal follicle cells at the posterior develop like their anterior counterparts by forming border, stretched and centripetal follicle cells. This raises the question of whether the anterior follicle cells are subdivided into three cell types after the decision between anterior and posterior is taken, or whether there is a symmetric prepattern in the terminal follicle cells at both ends of the egg chamber. The ability to generate small clones of anterior cells at the posterior by removing the Egfr makes it possible to distinguish between these possibilities. If the latter model is correct, isolated patches of anterior cells should still respond to the symmetric prepattern correctly and form the appropriate type of anterior cell, even though they are surrounded by posterior cells, whereas the former model predicts that these cells should be unable to interpret their position. To follow the fate of small patches of anterior cells at the posterior of the egg chamber, small Egfr- clones were generated, but in this case, clone generation took place in the presence of enhancer trap lines that are expressed specifically in each of the three anterior follicle cell types. Egfr- cells that fall within a region 8-11 cell diameters from the posterior pole show staining for a centripetal cell marker, whereas clones that fall either proximal or distal to this 3-cell-wide belt do not activate this marker. Thus, anterior cells at the posterior express the anterior BB127 centripetal cell marker autonomously in a region that is the exact posterior counterpart of the anterior centripetal follicle cell domain. Furthermore, clones of as few as 4 cells express BB127 if they fall within this region, indicating that anterior cells can correctly interpret their position with respect to the posterior pole, although all of the surrounding cells are posterior. The same conclusion applies to a border cell and a stretched cell marker. The results demonstrate that small posterior clones of anterior cells can interpret their position with respect to the posterior pole by adopting the appropriate anterior follicle cell fate: the most terminal Egfr- cells behave like border cells, the subterminal Egfr- cells behave like stretched follicle cells, and the least terminal like centripetal cells. Thus, the positional information that specifies the positions of these distinct cell types at the anterior pole is also present at the posterior, strongly suggesting that there is a symmetric prepattern within the terminal follicle cell population that is independent of the decision between anterior and posterior fate (Gonzalez-Reyes, 1998).

Notch is shown to be required for the correct subdivision of the terminal follicle cells. Although the phenotype of Notch and Delta mutants provided the first evidence that the posterior follicle cells play a role in the polarization of the oocyte, it is still not known at which step in anterior-posterior axis formation Delta/Notch signaling is required. To address this question, an examination was performed to determine whether the N ts mutant disrupts this pathway before or after the induction of the posterior follicle cells by the oocyte. Since the most sensitive assay for a failure in posterior follicle cell determination is transformation to anterior fate, the expression of the border cell enhancer trap line slbo 1 was examined in stage 10 N ts egg chambers that had been maintained at the restrictive temperature of 32°C. These conditions produce a penetrant oocyte polarization phenotype in which the germinal vesicle often remains at the posterior of the oocyte. Surprisingly, slbo is expressed in neither the anterior nor posterior follicle cells of these egg chambers. In addition, the anterior most follicle cells in N ts mutant egg chambers never round up or migrate between the nurse cells towards the oocyte, indicating that Notch activity is required for border cell development (Gonzalez-Reyes, 1998).

To determine whether Notch activity is required for the specification of other anterior follicle cell types, this experiment using enhancer trap line s to label all anterior follicle cells. N ts egg chambers lack both border and centripetal follicle cells, but these missing cells do not appear to be transformed into stretched follicle cells. Notch therefore seems to play a role in determining the size of the anterior terminal follicle cell population, as well as its subdivision into distinct cell types. The N ts mutant fails to disrupt the patterning of the posterior terminal follicle cells, confirming that Notch is not required for the determination of posterior identity. However, many fewer cells express a posterior marker than in the control egg chambers. This reduction in posterior follicle cell number indicates that Notch plays a role in specifying the size of the terminal follicle cell population that is competent to respond to Gurken and is consistent with the decrease in terminal follicle cell number observed at the anterior of these egg chambers (Gonzalez-Reyes, 1998).

These results suggest a three-step model for the anterior-posterior patterning of the follicular epithelium that subdivides this axis into at least five distinct cell types. Altogether, these observations support a stepwise model for the patterning of the follicle cell layer along the AP axis. In the first step, the follicle cell epithelium is divided into terminal and main-body follicle cell populations. There is no lineage restriction boundary between the posterior terminal follicle cells and the main-body follicle cells at a stage in development that is four cell divisions before stage 6, indicating that the distinction between these two cell types arises after stage 1. Because the terminal cells have to be specified before Gurken signaling occurs, this restricts the time at which this population is determined to between stages 2 and 5. Although the data do not suggest a mechanism for how these cells are specified, their position suggests a simple model in which they are induced by a 'terminalizing' signal that spreads from the two poles of the egg chamber. The most likely sources for such a signal are the two polar follicle cells at each end of the egg chamber, since these cells lie in the center of the terminal domain and adopt a terminal fate themselves. The next step in the patterning of the follicular epithelium is the formation of a symmetric prepattern within each terminal follicle cell population. How this prepattern is established is unknown, but the geometry of the egg chamber again suggests that it might involve signals that emanate from the poles. Indeed, it is possible that the terminal follicle cells are specified and patterned by the same process, since both events require Notch activity. For example, the 'terminalizing' signal could induce distinct terminal fates at different distances from the pole. The third step in the patterning of the follicle cell layer occurs when the oocyte induces one population of terminal follicle cells to adopt a posterior fate, thereby breaking the symmetry of the follicle cell layer. As a consequence, the symmetric prepattern in the terminal follicle cells is interpreted differently in the anterior and posterior populations. The anterior cells become subdivided into border, stretched and centripetal follicle cells, while the posterior cells may undergo a similar subdivision into posterior cell types. In this way, the sequential patterning of the terminal follicle cells gives rise to at least five different cell types along the anterior-posterior axis (Gonzalez-Reyes, 1998).

Determination of EGFR signaling output by opposing gradients of BMP and JAK/STAT activity

A relatively small number of signaling pathways drive a wide range of developmental decisions, but how this versatility in signaling outcome is generated is not clear. In the Drosophila follicular epithelium, localized epidermal growth factor receptor (EGFR) activation induces distinct cell fates depending on its location. Posterior follicle cells respond to EGFR activity by expressing the T-box transcription factors Midline and H15, while anterior cells respond by expressing the homeodomain transcription factor Mirror. This study shows that the choice between these alternative outputs of EGFR signaling is regulated by antiparallel gradients of JAK/STAT and BMP pathway activity and that mutual repression between Midline/H15 and Mirror generates a bistable switch that toggles between alternative EGFR signaling outcomes. JAK/STAT and BMP pathway input is integrated through their joint and opposing regulation of both sides of this switch. By converting this positional information into a binary decision between EGFR signaling outcomes, this regulatory network ultimately allows the same ligand-receptor pair to establish both the anterior-posterior (AP) and dorsal-ventral (DV) axes of the issue (Fregoso Lomas, 2016).

This study shows that the choice between two alternative Grk/EGFR signaling outcomes in the follicular epithelium depends on positional input provided by Upd and Dpp. At the posterior, the presence of Upd allows Grk to induce Mid/H15 while, at the anterior, Grk together with Dpp positively regulates Mirr. In this context, Upd and Dpp serve to define the response to Grk/EGFR signaling, since they are not sufficient to induce Mid/H15 and Mirr, respectively, in the absence of Grk (Fregoso Lomas, 2016).

Mutual repression is demonstrated between Mid/H15 and Mirr that is proposed to generate a double-negative feedback circuit that toggles the system between anterior and posterior outcomes. Moreover, in addition to their mutual regulation, analysis of double-mutant clones reveals that Upd and Dpp each regulate both Mid/H15 and Mirr and, thus, each provides input to both sides of this circuit. Upd is required for the expression of Mid and H15 even in the absence of a functional mirr gene, demonstrating that Upd is required for Mid/H15 expression independent of its ability to repress Mirr. Similarly, Dpp signaling can repress Mid independently of its positive effect on Mirr. The choice of Grk/EGFR signaling outcome in this context thus depends not only on mutual repression between these alternate targets but also on their opposing regulation by Upd and Dpp (Fregoso Lomas, 2016).

It is proposed that these elements define a bistable network that controls the choice between two alternative outcomes of Grk/EGFR signaling. These outcomes are irreversible -- e.g., posterior EGFR signaling in later stages cannot induce Mirr unless Mid and H15 are absent - and mutually exclusive, and the factors described in this study include two key elements found in bistable networks: feedback and non-linearity. The feedback in this case is provided by the reciprocal repression between Mirr and Mid/H15, generating a double-negative feedback loop that reinforces the choice of signaling outcome. In addition, bistability requires non-linearity in the response of the circuit to its upstream regulators, which makes the switch more sensitive to graded inputs. It is proposed that, in the follicular epithelium, this is achieved by the joint opposing regulation of the feedback circuit by both Dpp and Upd; each activates one side of the switch while repressing the other, biasing the outcome in the same direction (Fregoso Lomas, 2016).

These alternative responses to Grk are separated in time, as the source of Grk moves from posterior to anterior during the course of development. An important element that determines the choice between them is the early pattern of Mirr expression. In early stages of oogenesis, Mirr is Grk independent and is restricted to the main body follicle cells, due to its repression in the terminal regions by Upd. These main-body follicle cells correspond to the future anterior region of the columnar epithelium, and it is proposed that this early expression of Mirr predisposes them to express Mirr instead of Mid/H15 when Grk adopts its final dorsal anterior localization. Such a role for the early phase of Mirr expression is also consistent with the DV asymmetry of the Mid expression domain; as Grk moves anteriorly, leaving the range of posterior Upd and entering this domain of early Mirr expression and Dpp pathway activity, only the peak dorsal levels of Grk are capable of inducing Mid (Fregoso Lomas, 2016).

These observations also provide an example of how antiparallel signaling gradients can be integrated during epithelial patterning. Tissue patterning by opposing morphogen gradients is observed in developmental contexts as diverse as the Drosophila blastoderm and vertebrate neural tube, where they engage downstream transcriptional networks whose dynamic properties generate reproducible gene expression boundaries. Mutual repression between downstream transcription factors helps define the position and boundaries of cell fate domains, but how the opposing gradients are integrated is not well understood. This study shows that, in the follicular epithelium, the opposing Upd and Dpp gradients are integrated at the level of the Mirr-Mid/H15 feedback circuit. This integration occurs not only at the level of the mutual repression between Mid/H15 and Mirr but also by the ability of each gradient to regulate both sides of this circuit (Fregoso Lomas, 2016).

Together, these elements define the framework of a regulatory network that integrates localized positional information to regulate a binary choice of EGFR signaling outcome. The results allow construction of a model that both accounts for how an individual cell responds to Grk/EGFR signaling and explains how these spatial inputs are integrated across the epithelium to generate a defined pattern of Mid/H15 and Mirr expression, ultimately defining the pattern of the eggshell. Mirr is required for the generation of the high- and low-Broad domains, and Mid/H15 expression is required to define the posterior limit of these domains. The ability of Dpp and Upd to influence the outcome of EGFR signaling allows a single signaling input, namely localized secretion of Grk by the oocyte, to generate multiple distinct outputs that are localized in space and time, thus establishing both the AP and DV polarity of the epithelium and generating a complex and reproducible pattern of cell fates (Fregoso Lomas, 2016).


cDNA clone length - 1707

Bases in 5' UTR -370

Bases in 3' UTR - 454


Amino Acids - 294

Structural Domains

The Gurken protein has an N-terminal signal peptide involved in the secretion of Gurken and a single EGF repeat. The protein has a potential transmembrane domain and a cytoplasmic region. In addition there are two potential N-glycosylation sites. The region between the signal sequence and the EGF repeat contains PEST sequences tagging the protein for rapid turnover. There is a set of six conserved cysteine residues (Neuman-Silberberg, 1993).

GRK is 24% identical to Spitz protein; 22% identical to Neu differentiation factor, and 28% identical to TGF-alpha. The EGF-like region is 35% identical to repeats found in EGF, Notch and Xotch, the Xenopus homolog of Notch (Neuman-Silberberg, 1993).

The EGF-like domain in Vein is 43 amino acids long and has the six invariant cysteines and highly conserved glycine and arginine residues characteristic of the motif. The cysteines are thought to form disulfide bonds, thereby producing a looped structure. The EGF motif in VN is between 30% and 44% identical to other EGF-like ligands. It shares 37% identity with Spitz and 33% identity with Gurken. Amino-terminal to the EGF domain of VN is an Ig-like domain of the C2 type that includes nonimmunological proteins (Schnepp, 1996).

gurken: Regulation | Factors affecting Gurken RNA localization and translation | Developmental Biology | Effects of Mutation | References

date revised: 30 August 98

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.