Fes/Fer non-receptor tyrosine kinases regulate cell adhesion and cytoskeletal reorganisation through the modification of adherens junctions. Unregulated mammalian Fes/Fer kinase activity has been shown to lead to tumours in vivo. Drosophila Fer localises to adherens junctions in the dorsal epidermis and regulates a major morphological event, dorsal closure. Mutations in Src42A cause defects in dorsal closure similar to those seen in dfer mutant embryos. Furthermore, Src42A mutations enhance the dfer mutant phenotype, suggesting that Src42A and DFer act in the same cellular process. DFer is required for the formation of the actin cable in leading edge cells and for normal rates of dorsal closure. A gain-of-function mutation in dfer (dfer^gof) expresses an N-terminally fused form of the protein, similar to oncogenic forms of vertebrate Fer. dfer^gof blocks dorsal closure and causes axon misrouting. In dfer loss-of-function mutants ß-catenin (Armadillo) is hypophosphorylated, whereas in dfer^gof ß-catenin is hyperphosphorylated. Phosphorylated ß-catenin is removed from adherens junctions and degraded, thus implicating DFer in the regulation of adherens junctions (Murray, 2006),

Fes/Fer non-receptor tyrosine kinases regulate numerous cellular processes, including cell adhesion, cytoskeletal reorganisation and intracellular signalling (Greer, 2002). Mammalian Fer has been implicated in the regulation of adherens junctions (AJs) in tissue culture (Kim, 1995; Piedra, 2003; Rosato, 1998; Xu, 2004). Adherens junctions (AJ) are responsible not only for adhesion between neighbouring cells, but also for providing a link between the plasma membrane and the actin cytoskeleton. AJs are rich in phosphotyrosine residues, and several tyrosine kinases and phosphatases are known to regulate their assembly, stability and function (Murray, 2006),

Cell adhesion and cytoskeletal reorganisation are the driving force behind many morphogenetic movements of embryonic development. For example, during dorsal closure in the Drosophila embryo, the epidermal sheets on each side of the embryo extend dorsally to meet and fuse at the dorsal midline, thereby enclosing the amnioserosa and yolk sac. Cell adhesion in the dorsal-most row of cells, the leading edge cells, is regulated in part through the modification of adherens junctions (Murray, 2006),

Intercellular adhesion is mediated by the transmembrane protein E-Cadherin, a homophilic adhesion molecule that binds to E-Cadherin on neighbouring cells. ß-catenin binds to the cytoplasmic tail of E-Cadherin. ß-catenin can also interact with the F-actin binding protein α-catenin. α-catenin acts as a molecular switch regulating F-actin assembly, binding alternately to ß-catenin and to F-actin. p120-catenin (p120ctn) binds to the juxtamembrane domain (JMD) of E-Cadherin, and is thought to act as a regulator of adherens junction assembly and disassembly (Murray, 2006),

Both ß-catenin and p120-catenin are regulated by phosphorylation. Fer binds constitutively to p120-catenin and can phosphorylate it in vitro (Kim, 1995). Phosphorylation of p120-catenin increases its affinity for Cadherin. When overexpressed or activated, Fer can also phosphorylate ß-catenin and thereby disrupt adhesion (Piedra, 2003; Rosato, 1998): phosphorylation on Tyr654 disrupts its association with E-Cadherin; phosphorylation on Tyr142 blocks interaction with α-catenin. In C. elegans, the Fer orthologue associates with ß-catenin in vivo and depends upon ß-catenin for localisation to cell-cell junctions (Putzke, 2005). Fer can also stabilise the cadherin complex by phosphorylating and activating the phosphatase PTP1B, which in turn keeps ß-catenin (Tyr654) dephosphorylated (Xu, 2004). Thus Fer has the capacity to both positively and negatively regulate cadherin complex stability (Murray, 2006),

Several of the putative substrates of Fer are also substrates of the Src family of cytoplasmic tyrosine kinases (SFKs). Both p120ctn and Cortactin have been identified as being substrates of Src. Like Fer, the SFK Fyn binds constitutively to p120-catenin, and can phosphorylate ß-catenin at Tyr142. Src is able to phosphorylate ß-catenin at Tyr654. Given the overlap in substrate specificity, Fer and SFKs may play overlapping or redundant roles in the regulation of cell adhesion and motility. Functional redundancy between different SFKs has been demonstrated in mammals and in insects. In mice, double mutations in SFKs lead to overt phenotypes, where single mutations do not. In Drosophila, members of the Src kinase family cooperate to regulate JUN kinase (JNK) activity: double mutations in Src42A;tec29, and Src42A;Src64 give a dorsal open phenotype, whereas single mutations do not (Murray, 2006),

Only a single member of the Fes/Fer family is found in Drosophila, DFer (Fps85D). The canonical form of DFer, p92dfer, was identified by similarity to other family members (Katzen, 1991), and shares equal homology with Fes and Fer. Subsequently, a second cDNA encoding a short isoform, p45dfer, was discovered (Paulson, 1997). Paulson showed that DFer can transform vertebrate cells, suggesting that the molecular pathways through which Fer signals are likely to be conserved (Murray, 2006),

This study shows that DFer acts in conjunction with Src42A in the process of dorsal closure. dfer mRNA is specifically upregulated in the leading edge cells of the dorsal epidermis. DFer localises to adherens junctions and is required for the formation of the F-actin cable in leading edge cells, and for normal rates of dorsal closure. When mutations in dfer are combined with a mutation in Src42A, dorsal closure fails completely. A gain-of-function dfer mutant (dfer^gof) blocks dorsal closure and causes axon misrouting. The dfer^gof mutant expresses an N-terminally fused form of DFer, similar to oncogenic forms of Fer. ß-catenin phosphorylation levels are reduced in dfer loss-of-function mutants and increased in dfer^gof mutants. This supports a role for DFer in the regulation of AJs and cell-cell adhesion, and may begin to explain its role in oncogenesis (Murray, 2006),

This study shows that DFer localises to adherens junctions during Drosophila embryogenesis, where it regulates leading edge morphology and dorsal closure. In dfer null mutants, P-Tyr staining is reduced at the leading edge, in particular at the actin-nucleating centers. Two potential substrates of DFer, p120ctn and ß-catenin, are localised to adherens junctions. It was found that in dfer^Δ1 mutants ß-catenin phosphorylation is reduced. Conversely, ß-catenin is more highly phosphorylated in dfer^gof mutants, demonstrating that the role of Fer in the phosphorylation of ß-catenin is conserved in Drosophila. Interestingly, the overall level of ß-catenin at cell-cell junctions is lower in dfer^gof mutants, suggesting that phosphorylated ß-catenin is lost from AJs and subsequently degraded (Murray, 2006),

dfer mutants also exhibit a disorganised and reduced F-actin cable at the leading edge. Formation of the F-actin cable appears to depend on adherens junctions, as F-actin nucleation begins at the level of the AJs and the F-actin cable is disrupted in DE-Cadherin mutants. It has been suggested that elevated levels of cytoplasmic α-catenin near stable AJs could favour the formation of F-actin bundles. DFer may contribute to the formation of the F-actin cable by phosphorylating ß-catenin, reducing its affinity for α-catenin, and thereby increasing the local levels of cytoplasmic α-catenin. If DFer promotes stable F-actin bundles then the regulated loss of DFer from the leading edge at stage 14 may enable the more motile Arp2/3 regulated F-actin filopodia to form and complete dorsal closure by 'zipping up' (Murray, 2006),

DFer and Src42A cooperate during dorsal closure. DFer localises to AJs and regulates ß-catenin phosphorylation. In Drosophila, Src42A binds and phosphorylates ß-catenin, although this may not be direct. Consequently, the more severe phenotypes seen in the dfer;Src42A loss-of-function mutants are most likely due to a combined loss of phosphorylation on at least two different tyrosine residues of ß-catenin (Murray, 2006),

dfer mRNA is upregulated in leading edge cells. This, together with reports that vertebrate v-Fps and Fes mediate JNK pathway activation (Li, 1998), suggested that dfer might activate the JNK pathway during dorsal closure. Although DFer itself cannot induce dpp expression, it does play a supporting role in the maintenance of dpp levels, as revealed in the Src42A mutant background. A similar failure in the maintenance of dpp, as opposed to its induction, is seen in mutants of the Wnt pathway. Given the comparable phenotypes, and the fact that phosphorylated ß-catenin is reduced in dfer mutants, it is possible that DFer contributes to the maintenance of dpp via the Wnt pathway (Murray, 2006),

This study isolated a novel, gain-of-function mutation, dfer^gof, in which a fragment of the White protein is fused to the N terminus of Dfer. This protein, Wex1-DFerRB, is analogous to oncogenic forms of Fps in which part of the viral GAG protein is fused to the N terminus of the endogenous proto-oncogene, generating an activated kinase. Although dfer^gof mutants express DFer at higher levels, this alone seems unlikely to account for the observed defects, as overexpression of DFerRB gives no obvious embryonic phenotype (Murray, 2006),

In dfer^gof mutants, the leading edge cells fail to elongate and the F-actin-rich filopodia are greatly reduced. The overall levels of the AJ junction components DE-Cadherin and ß-catenin are reduced, and ß-catenin is hyperphosphorylated. This suggests that AJs are disrupted in dfer^gof mutants. By contrast, it is interesting that the morphology of amnioserosal cells is shifted to a more motile appearance: F-actin is reduced at the cortex and there is an increase in the number of filopodia, perhaps because of a disruption of cell-cell junctions. In vertebrates, Fer has the capacity to both positively and negatively regulate cadherin-complex stability. This dual function may reflect a difference in binding partners present at AJs in different tissues (Murray, 2006),

Although loss of dfer does not appear to affect axon guidance, dfer^gof mutants have a clear CNS phenotype in which axons aberrantly cross the midline. A similar phenotype is seen with overexpression of the Abelson tyrosine kinase, which antagonises the receptor Robo. dfer^gof mutants also disrupt axon guidance in the PNS, with some general misrouting of motor nerves and some overly large synapses. In vertebrates, Fer associates with N-Cadherin in elongating neurites, where it can coordinately regulate N-Cadherin and integrin adhesion (Arregui, 2000). Fer has been shown to be concentrated in growth cones of stage 2 hippocampal neurons and is required for neuronal polarisation and neurite development (Lee, 2005). Similar to the leading edge, DFer may be required at growth cones to regulate filopodial extensions. In chick retinal cells, the phosphatase PTP1B when phosphorylated by Fer, localises to the catenin-binding domain of N-Cadherin (Xu, 2004). Interestingly, the Drosophila homologue of PTP1B, DPTP61F, is expressed in the CNS and binds to the axon guidance molecule Dock (Murray, 2006),

Strikingly, all of the phenotypes associated with dfer^gof mutants are rescued by expression of the Puckered tyrosine phosphatase. Given that JNK pathway activity appears normal in dfer^gof mutants, Puckered may target DFer itself, or its substrates, at least one of which is hyperphosphorylated in dfer^gof mutants. A role for DFer during Drosophila embryonic development in the regulation of AJ stability, in the formation of the contractile leading edge during dorsal closure, and in axon guidance. It cooperates with Src42A to regulate ß-catenin phosphorylation at AJs. A gain-of-function mutant with structural similarity to oncogenic forms of vertebrate Fer was isolated. Unregulated Fer activity leads to oncogenesis, possibly through unregulated epidermal to mesenchymal transition. This study has shown that activated DFer, or loss of DFer together with Src42A, disrupts AJs. This may provide a model for studying oncogenesis in the whole organism (Murray, 2006),

GENE STRUCTURE

cDNA clone length - 4959 (transcript variant A)

Bases in 5' UTR - 339

Exons - 14

Bases in 3' UTR - 642

PROTEIN STRUCTURE

Amino Acids - 1325 (isoform A)

Structural Domains

The complete nucleotide sequence for the 3.2-kbp cDNA included an open reading frame that could encode a protein of 803 amino acids with a calculated molecular weight of 92,505. The protein encoded by dfps85D bears hallmarks of PTKs, including a 30-kDa catalytic domain that composes the carboxy-terminal third of the protein, an amino acid sequence characteristic of ATP-binding sites (residues 548 to 553 and lysine at 570), motifs of sequence that serve as signatures of tyrosine-specific kinases, a tyrosine (residue 691) whose phosphorylation appears to be involved in enzymatic activation of the fps protein, and a domain known as SH2 that is conserved in the cytoplasmic PTKs and that is thought to serve regulatory functions for the enzymes. The amino acid sequence of the protein encoded by dfps85D was compared with the sequences of all known PTKs. Greatest resemblance was found with the products of vertebrate fps and a related gene known as fer. The extents of the resemblances to fps and fer were virtually identical (Katzen, 1991).

dfer is a complex locus encoding at least three protein isoforms: DFerRA, DFerRB/p92dfer and DFerRC. DFerRB/p92dfer was originally identified by Katzen (1991), and has the canonical structure of Fes/Fer family kinases, consisting of an FCH motif at the N terminus and three coiled-coil regions, an SH2 domain and a kinase domain. The longest isoform, DFerRA, shares the promoter and exon structure of DFerRB, but includes an additional exon, which encodes a novel protein domain that lies between the canonical N-terminal domain and the SH2 domain. A short isoform, DFerRC, transcribed off a second promoter, encodes a truncated form of DFerRB, lacking most of the N-terminal domain but including the third coiled-coil domain. A fourth isoform, DFerRD/p45dfer, has been described (Paulson, 1997), which shares the SH2 and kinase domain, but has a distinct 10 residue N-terminal domain. Recent genome and EST sequence data, however, raise some doubt over the validity of this isoform (Murray, 2006),

date revised: 10 December 2006

Home page: The Interactive Fly © 2006 Thomas Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.