Coordination between cell proliferation and cell death is essential to maintain homeostasis in multicellular organisms. In Drosophila, these two processes are regulated by a pathway involving the Ste20-like kinase Hippo (Hpo) and the NDR family kinase Warts (Wts; also called Lats). Hpo phosphorylates and activates Wts, which in turn, through unknown mechanisms, negatively regulates the transcription of cell-cycle and cell-death regulators such as cycE and diap1. Yorkie (Yki), the Drosophila ortholog of the mammalian transcriptional coactivator yes-associated protein (YAP), has been identified as a missing link between Wts and transcriptional regulation. Yki (named for its loss-of-function phenotype after a very small breed of dog, the Yorkshire Terrier) is required for normal tissue growth and diap1 transcription and is phosphorylated and inactivated by Wts. Overexpression of yki phenocopies loss-of-function mutations of hpo or wts, including elevated transcription of cycE and diap1, increased proliferation, defective apoptosis, and tissue overgrowth. Thus, Yki is a critical target of the Wts/Lats protein kinase and a potential oncogene (Huang, 2005).

The increase in cell number that accompanies the growth of an organ or organism results from the balanced coordination of three simultaneous processes, including cell growth, cell proliferation, and cell death. Cell growth is a prerequisite for cell proliferation during normal organ growth, and sustained cell proliferation must be coupled to appropriate cell growth. With appropriate cell growth, a net increase in cell number in a growing organ depends on the rate at which they are generated via cell proliferation, as well as the rate at which they are eliminated by cell death (apoptosis). How cell proliferation and cell death are coordinated during tissue growth and homeostasis is yet to be completely understood, and this mechanism must be intact throughout life to prevent diseases such as cancer (Huang, 2005).

Recent studies in mice and fruit flies have revealed two distinct modes in which cell proliferation and cell death could be coupled. In the first mode, increased proliferation, such as that resulting from activation of the Myc oncogene, is coupled in an obligatory fashion to increased cell death. Such coupling between proliferation and apoptosis provides an important failsafe mechanism to prevent inappropriate proliferation of somatic cells. In the second mode, increased proliferation, such as that resulting from activation of the microRNA bantam, or inactivation of the tumor suppressors hippo (hpo), salvador (sav), and warts (wts), is accompanied by an inhibition of cell death. Here, suppression of cell death might allow the overproliferating cells to overcome proliferation-induced apoptosis, thus resulting in a robust increase in organ size. In many aspects, these circumstances resemble certain cancer cells, which display both increased cell proliferation and suppressed cell death (Huang, 2005 and references therein).

hpo, sav, and wts (also called lats) were identified from genetic screens in Drosophila for negative regulators of tissue growth. Inactivation of any of these genes results in increased cell proliferation and reduced apoptosis. hpo encodes a Ste20 family protein kinase, sav encodes a protein containing WW and coiled-coil domains, and wts encodes an NDR (nuclear Dbf-2-related) family protein kinase. Studies have suggested that these genes function in a common pathway that coordinately regulates cell proliferation and apoptosis by targeting the cell-cycle regulator CycE and the cell-death inhibitor DIAP1. Using a combination of genetic and biochemical assays, it has been shown that Hpo, Sav, and Wts define a novel protein kinase cascade wherein Hpo, facilitated by Sav, phosphorylates Wts (Wu, 2003). It was further demonstrated that this pathway, hereafter referred to as the Hpo pathway, negatively regulates the transcription of diap1 (Wu, 2003). It is worth noting that this model differs significantly from an alternative model by others that suggests that this pathway regulates DIAP1 posttranscriptionally through phosphorylation of DIAP1 by Hpo. Another unresolved issue in Hpo signaling concerns the molecular mechanism of the Wts/Lats kinase. While previous studies have identified a number of putative targets for this tumor suppressor, including the G2/M regulator cdc2 and the actin regulators zyxin and LIMK1, none of them could account for the excessive overgrowth associated with wts mutant clones. Thus, the most critical target of the Wts/Lats kinase has remained elusive (Huang, 2005 and references therein).

yorkie (yki) has now been identified as the elusive target of the Wts/Lats tumor suppressor. yki encodes the Drosophila ortholog of yes-associated protein (YAP), a transcriptional coactivator in mammalian cells (Yagi, 1999; Strano, 2001; Vassilev, 2001). Yki is required for normal tissue growth and diap1 transcription and is phosphorylated and inactivated by Wts. Overexpression of yki phenocopies loss-of-function mutations of hpo, sav, or wts. Taken together, these studies identify a missing link between Hpo signaling and transcriptional control and provide further support for the model implicating the Hpo signaling pathway in transcriptional regulation of diap1. These studies further reveal a functional conservation between YAP and Yki and implicate YAP as a potential oncogene in mammals (Huang, 2005).

Activation of yki leads to massive tissue overgrowth that resembles the loss-of-function phenotype of hpo, sav, or wts. To probe the physiological function of yki, the “flip-out” technique was used to generate clones of cells in which yki is overexpressed during development. yki-overexpressing clones lead to marked overgrowth in adult epithelial structures. Wing imaginal discs containing multiple yki-overexpressing clones reach up to eight times the area of control wing discs raised under identical conditions. Besides the overgrowth phenotype, adult cuticles secreted by yki-overexpressing cells display an unusual texture. In yki-overexpressing clones on the notum, the apical surface of the epidermal cells is domed such that cell-cell boundaries are visible between adjacent cells, whereas cell boundaries are not visible in the neighboring wild-type tissues. Both the overgrowth and the abnormal cell morphology caused by yki overexpression closely resemble those shown previously for hpo and wts mutant cells, suggesting that these genes might function in a common pathway (Huang, 2005).

Cell-doubling time for control and yki-overexpressing cells in the wing imaginal disc was determined by analyzing well-separated flip-out clones 48 hr post clone induction. The cell-doubling time for wild-type and yki-overexpressing clones (30 pairs of clones analyzed) was 16.1 hr and 12.0 hr, respectively. Thus, like mutant clones of hpo or wts, yki-overexpressing cells multiply faster. Notably, while cells in the control clones intermingle with their neighbors and form wiggly borders, yki-overexpressing cells minimize their contacts with their neighbors and form round smooth borders. This phenotype indicates distinct adhesive properties of the yki-overexpressing cells and resembles that seen with loss-of-function wts clones. FACS analysis shows that yki-overexpressing cells have a similar cell-cycle profile and cell size distribution as compared to wild-type cells. Thus, like loss-of-function of hpo (Wu, 2003), activation of yki does not accelerate a particular phase of the cell cycle. Rather, each phase of the cell cycle is proportionally accelerated (Huang, 2005).

Activation of yki in the eye imaginal disc leads to increased number of interommatidial cells without affecting photoreceptor differentiation. Focus was placed on the eye imaginal disc, a pseudostratified epithelium in which cell differentiation, proliferation, and apoptosis occur in a highly stereotyped manner. In the third instar, the morphogenetic furrow (MF) traverses the eye disc from posterior to anterior. Cells anterior to the MF are undifferentiated and divide asynchronously, whereas cells in the MF are synchronized in the G1 phase of the cell cycle. Posterior to the MF, cells either exit the cell cycle and differentiate or undergo one round of synchronous division (second mitotic wave, SMW) before differentiation. These cells assemble into approximately 750 ommatidia, leaving behind approximately 2000 superfluous cells that are eliminated by a wave of apoptosis ~36 hr after puparium formation (APF) (Huang, 2005).

To investigate whether activation of yki perturbs photoreceptor differentiation, the neuronal marker Elav was examined. yki-overexpressing ommatidial clusters have the normal complement of photoreceptor cells. The spacing between adjacent ommatidial clusters is increased due to the presence of extra interommatidial cells. The formation of extra interommatidial cells is most evident in pupal eye discs, when yki-overexpressing clones contain many additional cells between photoreceptor clusters. Thus, like loss-of-function of hpo, sav, or wts, yki overexpression results in an increased number of uncommitted, interommatidial cells without affecting early retina patterning (Huang, 2005).

Activation of yki leads to increased cell proliferation and decreased apoptosis. To pinpoint the developmental cause of yki-induced overgrowth, cell proliferation and apoptosis were monitored in eye imaginal discs. In wild-type eye discs, cells posterior to the MF undergo a synchronous second mitotic wave (SMW) that can be revealed as a band of BrdU-positive cells. Few BrdU-positive cells are found posterior to the SMW. In yki-overexpressing clones, cells fail to undergo cell-cycle arrest posterior to the SMW and continue cell cycles as shown by BrdU incorporation as well as M phase marker phospho-histone H3 (PH3). Thus, yki overexpression results in increased cell proliferation (Huang, 2005).

Using the TUNEL assay, cell death was monitored in the pupal retina at a point when a wave of apoptosis normally removes excess interommatidial cells around 36 hr APF. Strikingly, cell death was significantly suppressed in yki-overexpressing clones, even though abundant apoptosis was detected in the neighboring wild-type cells. Thus, normal developmental cell death is largely inhibited by yki overexpression (Huang, 2005).

The mechanisms of how body and organ size are regulated are just beginning to be understood. Recent studies in Drosophila have implicated a number of pathways in the coordinate control of cell growth, proliferation, and apoptosis, which ultimately regulate body and organ size. The insulin/Tsc/TOR signaling network, for example, plays a major role in coordinating organ growth with environmental cues such as nutrients. The Hpo signaling pathway, in contrast, might contribute to an intrinsic size 'checkpoint' that normally stops growth when a given organ reaches its characteristic size. Thus, molecular elucidation of the Hpo signaling pathway should provide important insights into size-control mechanisms in development (Huang, 2005).

Previous studies of the Wts/Lats tumor suppressor have failed to identify any target of this kinase that could account for its potent growth-regulatory activity. This study has provided genetic and biochemical evidence implicating Yki, the Drosophila ortholog of the mammalian coactivator protein YAP, as a direct, critical target of Wts/Lats in the Hpo pathway. Yki associates with and is phosphorylated by Wts. Moreover, Wts-mediated phosphorylation of Yki is stimulated by upstream components of the Hpo pathway, and the extent of Yki phosphorylation induced by Hpo pathway components in vitro correlates with the severity of the overexpression phenotype caused by the respective transgenes in vivo. Most importantly, overexpression of yki phenocopies loss of hpo, sav, or wts, while loss of yki results in the opposite phenotype, and epistasis analyses unambiguously places yki downstream of hpo, sav, and wts. Taken together, these results provide compelling evidence that Yki is a critical target of Wts/Lats in the Hpo pathway. It is further speculated that the relationship between Yki and Hpo signaling is likely conserved during evolution since overexpression of mammalian YAP is able to rescue the lethality associated with hyperactivation of the Hpo pathway in Drosophila. The functional conservation between Yki and YAP further suggests that YAP might function as an oncogene in mammals (Huang, 2005).

Yki is the first substrate identified for NDR family kinases, which include, besides Wts/Lats, Cbk1, Dbf2, and Dbf20 in budding yeast, Sid2 and Orb6 in fission yeast, Cot-1 in Neurospora, Sax-1 in C. elegans, Trc in Drosophila, and NDR1 and NDR2 in mammals (reviewed by Tamaskovic, 2003). The NDR family kinases are involved in diverse events in cell-cycle and cell morphogenesis, such as maintaining cell polarity (Cbk1 and Orb6), coordinating CDK inactivation and cytokinesis (Dbf2, Dbf20, and Sid2), and neuronal morphogenesis (Sax-1). Despite their diverse cellular functions, all NDR family kinases share similar structural features, such as the insertion of 30-60 amino acids between kinase subdomains VII and VIII, the presence of conserved activation loop and hydrophobic motif, and the presence of N-terminal noncatalytic domain (Tamaskovic, 2003). These common features suggest that NDR family kinases may employ similar mechanisms to interact with their substrates and regulators. Along this line, it is suggested that the approach described in this study, which uses the N-terminal noncatalytic domain of Wts as yeast two-hybrid bait, might provide a general method to discover substrates for other NDR family kinases (Huang, 2005).

A model has been proposed whereby Hpo, somehow facilitated by Sav, phosphorylates Wts (Wu, 2003). While this model implied that phosphorylation of Wts leads to activation of its kinase activity, it was not possible to directly test this due to the lack of an appropriate assay that measures pathway activity downstream of Wts. The identification of Yki as a Wts substrate provides a new tool to evaluate the earlier model. Consistent with the previous model implicating Hpo as an activating kinase of Wts, it has been shown that in S2 cells, the phosphorylation of Yki induced by transfected Wts is dependent on the endogenous Hpo protein. Furthermore, the in vitro kinase activity of Wts toward Yki is strongly stimulated when Wts is coexpressed with Hpo-Sav. It is suggested that such a relationship between Hpo and Wts is likely conserved during evolution. Indeed, a recent study (Chan, 2005) has demonstrated the activation of the mammalian Lats1 kinase by the mammalian Hpo homologs Mst1/Mst2 (Huang, 2005).

It is worth noting that several Ste20-like kinases have been implicated in the activation of NDR kinases. Such examples include the activation of Wts by Hpo (Wu, 2003; Chan, 2005), the activation of Dbf2 by Cdc15, the regulation of Orb6 by Pak1, and the regulation of Sid2 by Sid1. Thus, activation by Ste20-like kinases might represent a general mechanism for regulating NDR kinases. In retrospect, the difficulties in identifying substrates for NDR kinases might be due to their substrate specificity in conjunction with a requirement for activation by upstream kinases. Another emerging feature of the NDR kinases concerns their regulation by the Mob family of small regulatory proteins, which have been found to associate with multiple NDR family kinases, such as Dbf2, Orb6, Sid2, Cbk1, NDR1, and NDR2 (Tamaskovic, 2003). In Drosophila, Mats, a Mob family protein, has recently been identified as a tumor suppressor gene that likely regulates Wts in the Hpo signaling pathway (Lai, 2005). Thus, regulation by Mob family proteins likely represents an important and shared feature of modulating NDR family kinases (Huang, 2005 and references therein).

Previous studies of Hpo signaling have suggested two contrasting models on how this pathway regulates the cell-death regulator DIAP1. Using a diap1-lacZ reporter to follow diap1 transcription, elevated diap1 transcription was observed in mutant clones of hpo, sav, or wts that closely matches the increase in DIAP1 protein levels. Based on these results, it was proposed that the Hpo pathway negatively regulates diap1 at the level of transcription (Wu, 2003). However, an alternative model suggested that Hpo regulates DIAP1 posttranscriptionally by directly phosphorylating DIAP1, thus promoting its degradation. This model was largely based on two lines of evidence, including in situ hybridization showing unchanged diap1 mRNA level in mutant clones and the ability of Hpo to phosphorylate DIAP1 in vitro. It is noted, however, that in situ hybridization used in the latter studies did not involve the marking of mutant clones and thus may be less definitive than the diap1-lacZ reporter. A major drawback of the posttranscriptional model is that it cannot easily account for the involvement of Wts in the Hpo pathway. A direct link between Hpo and DIAP1 inevitably implies Wts as acting upstream or in parallel with Hpo, which is contradictory to other studies of the NDR kinases that generally place them downstream of the Ste20-like kinases (Huang, 2005 and references therein).

If the Hpo signaling pathway regulates diap1 via a transcriptional mechanism, then there should exist transcriptional regulator(s) that control diap1 transcription and whose activity may be regulated by the Hpo signaling pathway. Furthermore, such transcriptional regulator(s) must account for the mutant phenotypes resulting from deregulation of the Hpo pathway, such as changes in diap1 transcription and overgrowth. This current study demonstrates that Yki represents such a regulator, thus further supporting the previous model implicating the Hpo pathway in regulating diap1 transcription (Huang, 2005).

Understanding the molecular mechanisms by which the Hpo pathway regulates diap1 transcription will provide important insights into the developmental coordination of tissue growth and apoptosis. Like other coactivators, Yki presumably functions by interacting with DNA binding transcription factors. YAP, the mammalian homolog of Yki, is known to function as coactivator for a number of transcription factors, such as the p53 family member p73 (Strano, 2001), the Runt family member PEBP2α (Yagi, 1999), and the four TEAD/TEF transcription factors (Vassilev, 2001). This interaction is generally mediated by WW domains of YAP and PPxY motifs of the cognate transcription factors. At present, it is unclear whether any of the reported mammalian proteins represents the physiological partner for YAP. Along this line, it is worth noting that while the reported ability of YAP to transactivate p73 in cultured mammalian cells is more suggestive of a tumor suppressor function for YAP (Basu, 2003), these studies clearly implicate Yki and YAP as potential oncogenes. One interesting possibility (Lowe, 2004) is that the reported coupling of mammalian YAP to p73 might represent a fail-safe mechanism to limit the oncogenic potential of YAP in much the same way as cell death is obligatorily linked to oncogene activation (Huang, 2005).

An important direction in the future is to identify the DNA binding transcription factor that partners with Yki to regulate gene transcription; identifying the factor should provide critical insights into how Yki (and likely YAP as well) could function as a potent oncogene. This effort should be facilitated by the dissection of the diap1 promoter and the identification of a minimal Hpo-responsive element that confers transcriptional regulation of diap1 by the Hpo pathway. With such a DNA element, one should be able to identify the DNA binding transcription factor that partners with Yki to regulate the transcription of diap1 and other Hpo-pathway-responsive genes (Huang, 2005).

Many components of the Hpo pathway are conserved throughout evolution, suggesting that this emerging pathway might play a similar role in mammals. Indeed, previous studies have shown that human homologs of wts, hpo, and mats could rescue the respective Drosophila mutants. Moreover, mice lacking a wts homolog are prone to tumor formation, and the human orthologs of sav and mats are mutated in several cancer cell lines. Such conservation is further extended in the current study, showing that Yki and YAP have similar biological activity when assayed in Drosophila. These results suggest that the Hpo signaling pathway might play a conserved role in mammalian growth control. Furthermore, inactivation of growth suppressors of the Hpo pathway, such as Hpo, Sav, Wts, and Mats, and hyperactivation of growth promoters of the pathway, such as YAP, are likely to contribute to mammalian tumorigenesis (Huang, 2005).

GENE STRUCTURE

cDNA clone length - 1427

Bases in 5' UTR - 170

Exons - 4

Bases in 3' UTR - 112

PROTEIN STRUCTURE

Amino Acids - 418

(AAZ42161 )

Structural Domains

Yki is most closely related to the human yes-associated protein (YAP, also called YAP65) (Sudol, 1994), with 31% identity between the two proteins. Both proteins contain two WW domains, protein-protein interaction modules of 35-40 amino acids that are known to interact with PPXY-containing polypeptides. The similarity between Yki and YAP extends beyond the WW domains and includes a stretch of sequence similarity at the N-terminal part of the proteins. The WW domains of Yki are required for its interaction with Wts. While initially isolated as a protein that interacts with the SH3 domain of the Yes proto-oncogene, the involvement of YAP in Yes signaling has not been validated (Sudol, 1994). Notably, the corresponding SH3 binding region (Sudol, 1994) is absent in the Drosophila Yki protein. In contrast, YAP has been implicated as a transcriptional coactivator, a class of transcriptional regulators that do not bind to DNA themselves but associate with DNA binding transcription factors and supply or stimulate transcriptional activation of the cognate transcription factors. Specifically, YAP has been shown to function as a coactivator for a number of transcription factors, such as the p53 family transcription factor p73 (Strano, 2001), the Runt family protein PEBP2α (Yagi, 1999), and the TEAD/TEF family transcription factors (Vassilev, 2001). However, these studies have been performed exclusively in cultured mammalian cells and little is known about the physiological function of YAP (Huang, 2005).

date revised: 20 November 2005

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.