BIOLOGICAL OVERVIEW

Kekkon-1 was identified in a search for genes that may play a role in axonal outgrowth in the central nervous system (CNS). A collection of enhancer trap insertion lines was examined for beta-galactosidase expression in subsets of CNS cells. Among these the line 15A6 was selected because it expresses lacZ initially in a segmentally repeated set of cells at the CNS midline at early stages. At later stages, lacZ eventually becomes expressed in a large set of CNS cells and some peripheral nervous system cells. Kekkon2, an unlinked, independent gene, was isolated in a search for genes whose sequence resembles that of Kek1. Like Kek1, Kek2 is expressed in the CNS (Musacchio, 1996).

The same enhancer trap line, 15A6, was identified in a search for enhancer trap lines that exhibit expression patterns in follicle cells in a pattern consistent with the marked gene being a regulatory target of the epidermal growth factor receptor (Egfr) pathway, since the pattern of expression undergoes a transition during stages 8-10 when it relocalizes from a posterior to a dorsal-anterior gradient. Kekkon-1 was identified as an inhibitor of the Egfr; Kek1 acts in a negative feedback loop to modulate the activity of the Egfr tyrosine kinase. During oogenesis, kek1 is expressed in response to the Gurken/Egfr signaling pathway, and loss of kek1 activity is associated with an increase in Egfr signaling. Consistent with loss-of-function studies, ectopic overexpression of kek1 mimics a loss of Egfr activity. The extracellular and transmembrane domains of Kek1 can inhibit and physically associate with the Egfr, suggesting potential models for this inhibitory mechanism (Ghiglione, 1999).

In the complete absence of kek1 gene activity, flies are viable, fertile, and do not exhibit any overt morphological defects (Musacchio, 1996). Kek1 has the features of cell adhesion molecules (CAMs), and loss-of-function mutations of many CAMs have subtle mutant phenotypes. However, when ectopically overexpressed, some CAM molecules can generate striking mutant phenotypes that are revealing of their functions. Thus, to gain insights into the possible function of kek1 during oogenesis, the effect of overexpressing kek1 in the follicle cells was examined using the GAL4/UAS system When kek1 is expressed under the control of the GAL4 line T155, which drives expression all over the follicle cell epithelium beginning at stage 9, 100% of the resulting eggs are longer than the WT eggs and show a reduction or a complete loss of dorsal appendages. This loss of dorsal appendage material is due to a ventralization of the eggshell, a phenotype associated with egfr, or grk loss-of-function mutations. A similar ventralized eggshell phenotype is obtained when a dominant-negative form of the Egfr, EgfrDN, is expressed under the control of T155 (Ghiglione, 1999).

In addition to its role in establishing the dorsal characteristics of the eggshell, the Grk/Egfr pathway controls embryonic dorsoventral patterning by restricting a ventralizing signal. As a result, loss of Egfr function leads to ventralized embryos. Embryonic cuticles derived from T155; UAS-kek1 females shows a ventralized cuticle phenotype. To analyze the extent of this ventralization, embryos were stained for Twist (Twi) mRNA and protein that label a ventral domain ten cells wide. Overexpression of kek1 in follicle cells results in an expansion of the Twi expression domain. Similarly, embryos derived from T155; UAS-egfrDN females are also strongly ventralized. Altogether, these results indicate that overexpression of kek1 in follicle cells ventralizes both the eggshell and the embryo by most likely interfering with the activity of the Grk/Egfr signaling pathway. Interestingly, kek1 is expressed in the eye and wing imaginal discs in a pattern that is highly suggestive of induction by the Egfr (Musacchio, 1996). Although no obvious phenotypes have been described in these tissues in kek1 mutants, overexpression of kek1 in these tissues also generate phenotypes reminiscent of loss of Egfr activity. This suggests that the negative regulation of the Egfr by Kek1 is not only restricted to oogenesis (Ghiglione, 1999).

To determine whether Kek1 blocks the signaling activity of the Egfr, a test was performed to see if ectopic expression of kek1 blocks the transcriptional activation of target genes regulated by the Egfr/Ras/Raf/MEK/MAPK pathway in follicle cells. Since kek1 is a target of this pathway, kek-lacZ was used as a reporter for this experiment. Overexpression of kek1 using the Gal4 driver T155 is associated with a strong but not complete reduction of kek-lacZ expression. Using a stronger Gal4 line, CY2, a complete disappearance of kek-lacZ expression is observed. These data are consistent with the model that Kek1 downregulates the activity of the Ras/Raf/MEK/MAPK pathway. Consistent with this, epistasis experiments indicate that the inhibitory effect of Kek1 is overridden by the constitutive activation of the Egfr or Draf. In addition, the interaction between the transmembrane protein Rhomboid (Rho) and Kek1 was tested. Rho has been proposed to play a role in the activation of the Egfr, and overexpression of Rho leads to dorsalized eggshells. Overexpression of rho does not override the ventralization phenotype of UAS-kek1, indicating that Kek1 can block the effect of Rho on Egfr activation. Altogether, these results place Kek1 upstream of the Egfr and downstream of Rho (Ghiglione, 1999).

Because of the severe inhibition of Egfr activity by Kek1, revealed by the overexpression experiment, the effect of loss of kek1 function during oogenesis was reexamined. A subtle egg morphology phenotype may have been missed by simply using fertility as an assay (Musacchio, 1996). Indeed, the spacing between the dorsal appendages of eggs derived from kek1 mutant females is increased when compared to WT. Further, these eggs are also mildly shorter and rounder, a phenotype consistent with a hyperactivation of the Grk/Egfr pathway. These features do not interfere with hatching rates and patterning that is consistent with the normal Twi expression found in kek1 mutant embryos. Interestingly, when kek1 mutant flies are raised at 29°C, 5% (n = 257) of the embryos derived from kek1 mutant females showed a mild reduction in Twi expression. It is concluded that loss of kek1 activity during oogenesis leads to mildly dorsalized eggs (Ghiglione, 1999).

During oogenesis, Grk derived from the oocyte activates in a paracrine fashion the Egfr in dorsal follicle cells. Recent work (Wasserman, 1998) has shown that this paracrine signaling leads to the activation of a second phase of signaling, whereby the Egfr activity is amplified among follicle cells themselves. During this second phase, the Egfr activates a number of target genes that include both positive (rhomboid and vein) and negative (argos) regulators of the pathway. Activation of Rho in follicle cells presumably leads to the activation of the Spitz ligand for Egfr, while activation of Aos within the peak of Egfr activity at the dorsal anterior leads to repression of Egfr, effectively splitting in two the initial peak of Egfr activity. This splitting of Egfr activity eventually defines the domains where the dorsal appendages will form.

The identification of Kek1, together with the studies of Aos, indicates that there are at least two different negative regulators of Egfr activity in follicle cells. However, the regulation and function of Kek1 is distinct from those of Aos. aos is expressed only in response to high levels of Egfr activity, while kek1 is expressed in a graded fashion. Further, loss-of-function phenotypes of kek1 and aos in follicle cells are different. In the absence of kek1 activity, the spacing between the dorsal appendages is increased, while in the absence of aos, the appendages are fused dorsally. Aos has been proposed to split the initial peak of Egfr activity into two (Wasserman, 1998). It is proposed that the function of Kek1 is to restrict the lateral spreading of Egfr activation by Spi. Thus, in the absence of kek1 activity, Rho/Spi activation could spread more laterally, explaining the enhancement of the spacing between the two dorsal appendages (Ghiglione, 1999).

Thus, Kek1 acts as a potent negative regulator of Egfr activity when overexpressed. Further, it has been shown that the extracellular and transmembrane domains of Kek1 are sufficient for this inhibition. The extracellular domain of Kek1 contains one Ig-like domain and five LRRs, both of which can mediate protein-protein interactions. A number of mechanisms can underlie the mechanism by which this extracellular domain acts as an inhibitor. For example, the Kek1 extracellular domain could mask the accessibility of the extracellular domain of the Egfr to all ligands. Conversely, it could form a heterodimer with Egfr monomers and block their dimerization. Dimerization is a prerequisite to the activation of downstream signaling events by the RTK. Alternatively, Kek1 could be involved in bringing a transmembrane tyrosine phosphatase to the vicinity of the Egfr and thus lead to its deactivation. Consistent with these models, it has been observed that Kek1 can inhibit mammalian Egfr molecules from becoming tyrosine phosphorylated in response to growth factor treatment in infected insect cells (L. T. Amundadottir, et al., unpublished observations cited in Ghiglione, 1999). Finally, it is envisioned that Kek1 could target the Egfr to a degradation pathway through endocytosis or bind to additional proteins involved in their subcellular localization (Ghiglione, 1999).

In Drosophila, two other putative Kek molecules that share extensive homologies with kek1 have been identified: kek2 (Musacchio, 1996) and kek3 (unpublished data of Ghiglione, 1999). The function of these additional Kek-like proteins is not known. However, it is interesting to note that despite extensive saturation of the region containing kek3, no mutant alleles have been recovered. Whether or not this and the subtle effect of loss of Kek1 activity is due to redundancy within the Kekkon family remains to be determined. Thus, it will be important to characterize the expression patterns as well as the loss-of-function and overexpresssion phenotypes of kek2 and kek3 in order to evaluate any potential abilities to modulate the Egfr activity. Further, chimeric proteins between these molecules may help to further define the Kek1 domain(s) required for the inhibitory mechanism. Finally, there are putative transmembrane proteins in vertebrates (Suzuki, 1996) and invertebrates that show similar arrangements of LRRs and Ig motifs. This raises the possibility that Kek1 is a member of a family of structurally related Egfr inhibitors. Because any alteration in the activity of the various members of the human Egfr/ErbB family has strong links to oncogenesis, it will be important to determine if vertebrate LRR/Ig molecules share functional, in addition to sequence, similarities to Kek1. It is anticipated that the continued characterization of the Kek and related molecules will provide novel approaches to the design of inhibitors of the Egfr/ErbB family for therapeutic use in oncogenesis (Ghiglione, 1999).

GENE STRUCTURE

PROTEIN STRUCTURE

Amino Acids - 880 amino acids for Kek1 and 892 amino acids for Kek2

Structural Domains

Two members of a novel class of genes have been identifed in Drosophila that encode putative transmembrane proteins with six leucine-rich repeats and a single immunoglobulin loop. The Kek1 extracellular (428 amino acids mature) and the intracellular region (413 amino acids) are separated by a single transmembrane region. A signal peptide cleavage site is predicted between residues 20 and 21. The extracellular portion contains motifs for N-linked glycosylation at three sites and a single glycosaminoglycan attachment site. The mature extracellular domain of Kek2 is 492 amino acids, and the intracellular domain is 492 amino acids. These two molecules, Kek1 and Kek2, show striking conservation in their extracellular domains and have large and more divergent intracellular regions. The longest stretch of sequence identity between the two molecules occurs in the largely basic amino-flanking cysteine-rich region and is followed by significant homologies spanning the amino-end boundaries of the first and third repeats and also in the last repeat. These regions are known to adopt a beta-strand structure. Cysteine residues are found at position 10 of the fifth leucine-rich repeat, which is unusual in Drosophila. The spacing of leucines in the first two repeats diverges from the consensus yet is conserved between Kek1 and Kek2. The putative intracellular domains of both proteins are large and more divergent than their extracellular domains, containing several short regions of similarity with a 19% overall amino acids sequence identity. Potential phosphorylation sites occur throughout both sequences (Musacchio, 1996).

Studies on the C. elegans LET-23 Egfr have well illustrated the critical role of subcellular localization and PDZ proteins in signaling by this RTK. Interestingly, both Kek1 and the Egfr contain a TXV motif at the C terminus. This S/TXV motif has been implicated in protein-protein interactions and suggests that Kek1 and/or the Egfr may interact with PDZ-containing proteins (Ghiglione, 1999).

date revised: 20 April 99

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.