BIOLOGICAL OVERVIEW

Tissue growth during animal development is tightly controlled so that the organism can develop harmoniously. The salvador (sav) gene, which encodes a scaffold protein, restricts cell number by coordinating cell-cycle exit and apoptosis during Drosophila development. Hippo (Hpo), the Drosophila ortholog of the mammalian MST1 and MST2 serine/threonine kinases, is a partner of Sav. Hippo was described in five publications that appeared simutaneously: Pantalacci (2003) identified Hippo in a yeast two-hybrid screen in a search for Salvador interacting proteins, Udan (2003) identifed and positionally cloned hippo in a mutagenesis screen for genes that regulate tissue growth, and Harvey (2003), Jia (2003) and Wu (2003) identified hippo in screens for genes that restrict growth and cell number. Loss of hpo function leads to sav-like phenotypes, whereas gain of hpo function results in the opposite phenotype. Whereas Sav and Hpo normally restrict cellular quantities of the Drosophila inhibitor of apoptosis protein DIAP1 (Thread), overexpression of Hpo destabilizes DIAP1 in cell culture. DIAP1 is phosphorylated in a Hpo-dependent manner in S2 cells and Hpo can phosphorylate DIAP1 in vitro. Thus, Hpo may promote apoptosis by reducing cellular amounts of DIAP1. In addition, Sav is an unstable protein that is stabilized by Hpo. It is proposed that Hpo and Sav function together to restrict tissue growth in vivo (Pantalacci, 2003; Harvey, 2003; Jia, 2003; Udan, 2003 and Wu, 2003). This biological overview will deal with the Pantalacci study, while subsequent sections will deal with the other studies.

The final size of each organ in an animal is determined by cell growth and by the balance between proliferation and cell death, which determines cell number. Each of these parameters is controlled and coordinated by developmental cues so that the body parts reach their characteristic sizes and shapes. It remains unclear how cells in a developing organ sense their limits and respond by modulating growth, division and apoptosis. The gene sav (also known as shar-pei) limits organ size in vivo. Initial loss-of function (LOF) analyses have shown that sav is required for restricting cell division rates, for timely exit from the cell cycle and for programmed cell death during Drosophila eye development. In the absence of sav, an increase in cell number generated through excessive proliferation and reduced apoptosis leads to an increase in organ size. Timely cell-cycle exit during Drosophila development is contingent on the downregulation of Cyclin E and cyclin-dependent kinase activity, which is achieved through transcriptional and posttranscriptional events. In sav mutant clones, increases in Cyclin E protein and messenger RNA persist after their disappearance from wild-type cells. Thus, the effect of sav on proliferation can, at least in part, be explained by its ability to repress cellular quantities of Cyclin E. Apoptosis in Drosophila is triggered by developmentally regulated expression of the pro-apoptotic proteins Head involution defective (Hid), Grim and Reaper. These act by sequestering DIAP1 through their inhibitor of apotosis protein (IAP)-binding motif, and also by repressing quantities of DIAP1 protein through various mechanisms. When inactivated, DIAP1 no longer represses upstream caspases such as Dronc, and cell death ensues. In sav mutant clones, Hid- or Reaper-induced cell death is impaired. In addition, DIAP1 protein is increased in sav clones. Thus, sav promotes apoptosis (at least in part) by repressing DIAP1. sav encodes a protein containing two WW domains and a coiled coil, suggesting that it carries out its functions by interacting with partner proteins (Pantalacci, 2003).

Mutations in the warts (wts; also known as Large Tumour Suppressor) Ser/Thr kinase were initially shown to cause tissue overgrowth. Further studies established that the wts mutations phenocopy the sav mutants, suggesting that these genes function in the same pathway. Indeed, Wts and Sav associate biochemically in vitro; however, the role of wts is unclear and provides little insight into sav function (Pantalacci, 2003).

To further understanding of the biochemical mechanisms underlying sav function, the yeast two-hybrid system to was used to search for Sav binding partners. One of the recovered clones encoded the carboxy-terminal 67 amino acids of the Drosophila ortholog of the mammalian Ste20-like kinases MST1 and MST2 (MST1/2); the Drosophila protein is referred to as Hippo. When expressed in Drosophila S2 tissue-culture cells, epitope-tagged Sav and Hpo co-imunoprecipitate, confirming that these two proteins associate in Drosophila cells. In addition, epitope-tagged Hpo co-precipitate with Wts, the other Sav partner kinase (Pantalacci, 2003).

RNA-mediated interference (RNAi) was used to analyse the hpo LOF phenotype. Transgenic animals were generated bearing a fragment of the hpo complementary DNA in the pSympUAS12 vector to drive the expression of a double-stranded RNA under the control of the Gal4 transcription factor. Large hpo RNAi clones marked with green fluorescent protein (GFP) were induced by using the flp-out GAL4 technique, and the consequences were examined of hpo LOF on tissue growth. hpo inactivation induces an increase in tissue and organ size, leading, for example, to oversized halteres, or to an overgrowth of imaginal discs, the epithelial structures that give rise to adult body parts. Thus, like sav, hpo restricts tissue and organ growth in vivo (Pantalacci, 2003).

The expression pattern of hpo mRNA was examined by in situ hybridization; hpo is expressed uniformly in larval eye imaginal discs. To test whether hpo is required for cellular differentiation, eye discs stained for the neuronal marker Elav were examined. Like sav mutations, hpo LOF does not impair neuronal differentiation. Cells in the eye imaginal disc normally proliferate actively during early larval development in the first mitotic wave, and then temporarily arrest in the morphogenetic furrow where differentiation begins. After the passage of the morphogenetic furrow, a subpopulation of cells undergoes one additional round of divisions (the second mitotic wave) before permanently arresting. BrdU incorporation and Cyclin E staining were used to monitor S phases in larval eye imaginal disc. In hpo RNAi clones, cells fail to arrest in the morphogenetic furrow and, after the second mitotic wave, undergo additional rounds of division, similar to sav mutant cells. As observed for sav mutants, hpo mutant cells fail to downregulate Cyclin E. Conversely, transgenic animals were generated to express hpo ectopically under the control of the GAL4 promoter. When overexpressed in clones, hpo prevents cells from entering the second mitotic wave (Pantalacci, 2003).

Apoptosis was examined in hpo RNAi clones. Overexpression of the proapoptotic protein Hid in the eye by using the GMR promoter potently induces apoptosis, which is blocked by the caspase inhibitor p35 or by sav mutations. GMR-hid-induced apoptosis, as assayed by TdT-mediated dUTp nick end labelling (TUNEL) staining, is also blocked in hpo RNAi clones. In addition, DIAP1 expression is increased in hpo RNAi clones in a wild-type background, as has been observed in sav clones. Conversely, ectopic expression of hpo potently induces apoptosis. Apoptosis triggered by hpo overexpression is not impaired in sav mutant clones, suggesting that hpo overexpression can bypass the absence of sav. Thus, hpo inactivation phenocopies sav LOF, suggesting that in vivo the two genes act in concert to control apoptosis and proliferation (Pantalacci, 2003).

MST1/2, like most Ste20 family members, are potent activators of the JNK pathway (Dan, 2001). The ability of Hpo to trigger Drosophila JNK (DJNK; Basket) activation was measured by using an enhancer trap inserted in the DJNK target gene puckered (puc). Overexpression of Hpo in the eye robustly activates the puc-lacZ reporter gene. The MAPK pathway has been reported to protect cells from apoptosis by downregulating hid function in the Drosophila eye. Using an antibody against phosphorylated MAPK, it was found that Hpo overexpression neither represses nor activates MAPK. Thus, like its mammalian counterparts, Hpo can activate JNK but not the MAPK pathway in vivo (Pantalacci, 2003).

Numerous studies have shown that the JNK pathway can participate in the apoptotic process in Drosophila and mammals. The possibility exists that Sav and Hpo might act through DJNK to promote apoptosis. Therefore, apoptosis markers were examined in clones mutant for basket (bsk), which encodes the only Drosophila JNK. Quantities of DIAP1 were found to be normal in bsk mutant cells. Accordingly, developmental apoptosis in the pupal retina, which is blocked in sav clones, is unimpaired in bsk mutant clones. Finally, apoptosis triggered by combined overexpression of Sav and its partner Wts in the eye is not affected by bsk mutations. These data suggest that, at least in the eye, sav and hpo induce apoptosis through a DJNK-independent pathway (Pantalacci, 2003).

A common aspect of sav and hpo function seems to be the downregulation of DIAP1. This protein, like other IAP family members, is a RING-finger protein that directly binds to and inhibits upstream caspases, such as Dronc, and promotes their ubiquitination. DIAP1 quantities seem to be crucial in the apoptosis-triggering pathway and are therefore tightly regulated through distinct mechanisms controlled by the pro-apoptotic proteins Hid, Grim and Reaper, as well as by the caspases. (1) DIAP1 is highly unstable and is targeted for degradation by auto-ubiquitination as well as by at least two other ubiquitin ligases, Morgue and UbcD1. (2) Grim and Reaper repress the translation machinery, which primarily affects unstable proteins such as DIAP1. DIAP1 is also cleaved by the downstream caspases Drice and Dcp-1, which expose an N-terminal asparagine and make DIAP1 susceptible to degradation by the N-end rule pathway. It was therefore considered that the Hpo and Sav regulation of DIAP1 might be crucial to their pro-apoptotic function. Indeed, Hippo was shown to phosphorylate and decreases the stability of DIAP1 (Pantalacci, 2003).

Therefore, Hippo, a partner of the Sav, controls cell proliferation and apoptosis during Drosophila development. sav and hpo share similar LOF phenotypes, suggesting that they function in the same pathway. Downregulation of Hpo or Sav in vivo leads to an upregulation of DIAP1. Evidence in cultured cells is provided that Hpo-dependent phosphorylation of DIAP1 can occur, and that this phosphorylation is correlated with a reduction in DIAP1 stability. Further work will be required to determine whether DIAP1 phosphorylation is responsible for its destabilization; however, the results do not exclude additional modes of DIAP1 regulation by Hpo. Hpo interaction with Sav results in Sav phosphorylation and stabilization. However, because kinase-dead Hpo still elicits these effects, Hpo may not phosphorylate Sav directly. Thus, Hpo may recruit a kinase that phosphorylates Sav and stabilizes it. Alternatively, Hpo binding may prevent Sav degradation, leading to an increase in Sav levels that is independent of its phosphorylation. Because Sav may function in a complex with Hpo, this might constitute a positive feedback mechanism. Since Sav and Hpo have well-conserved mammalian orthologs, it will be interesting to investigate whether these orthologs trigger apoptosis by regulating IAPs. Sav, Hpo and Wts seem to form the basis of a previously unknown tissue-growth-suppressing pathway in flies, and the study of their mammalian orthologs will probably provide valuable insight into tumor formation (Pantalacci, 2003).

---

GENE STRUCTURE

cDNA clone length - 2828 bp

Bases in 5' UTR - 119

Exons - 2

Bases in 3' UTR - 699

PROTEIN STRUCTURE

Amino Acids - 669

Structural Domains

Like its mammalian counterparts, dMST contains an N-terminally situated kinase domain and C-terminally located regulatory domains. The dMST kinase domain exhibits 84% and 83% identity to that of hMST1 and dMST2, respectively. The C-terminally situated dimerization and kinase inhibitory domains present in MST1 and MST2 are also highly conserved in dMST (Jia, 2003).

Hpo belongs to the Ste20 family of kinases, which has over 20 members, including the prototypical kinase Ste20. These kinases have been shown to trigger the activation of mitogen-actived protein kinase (MAPK) pathways in eukaryotes. The two mammalian members, MST1/2, are highly similar to Hpo in the amino-terminal Ser/Thr kinase domain as well as in the C-terminal Sav-binding region. MST1 is activated in response to various proapoptotic stimuli, such as Fas ligand or staurosporine, and is cleaved and further activated by caspases. On cleavage, the MST kinase domain translocates to the nucleus, where it phosphorylates histone 2B, possibly triggering chromosome condensation. Overexpression of MST1 in mammalian tissue-culture cells induces cell death and activation of the c-Jun N-terminal kinase (JNK) and p38 MAPK pathways. MST1 is also upregulated in arresting cells in tissue cultures. Thus, the mammalian ortholog of Hpo is involved in cell death and proliferation (Pantalacci, 2003).

To identify the hpo gene, the lethality associated with the mutation was mapped through meiotic mapping by using P-elements and single nucleotide polymorphisms (SNPs) as markers; different mutations were found in CG11228 in all four hpo alleles. Transgenes containing the whole genomic region of CG11228 between the proximal and distal neighboring genes rescue the lethality of hpo trans-heterozygotes, showing that CG11228 indeed encodes Hpo. hpo is ubiquitously expressed during imaginal disc development, consistent with its requirement for proper proliferation arrest in these tissues (Udan, 2003).

Hpo is an ortholog of the mammalian Ste20-family kinases MST1 and MST2 (MST1/2) (Dan, 2001). MST1/2 are implicated in the activation of mitogen-activated protein kinase (MAPK) cascades and in regulating apoptosis. Hpo contains three highly conserved domains: an amino-terminal kinase domain; a central autoregulatory domain that inhibits kinase activity, and a domain implicated in dimerization and binding to Sav at its very carboxy terminus. Notably, the protein produced from all four hpo alleles is truncated after the kinase domain, and the highly conserved dimerization domain that mediates complex formation with Sav is deleted. All four alleles contain recessive loss-of-function mutations and are homozygous lethal at the first-second larval instar stage (Udan, 2003).

The N-terminal kinase domain is highly conserved between Drosophila and human, as is the C terminus, which contains a putative nuclear localization signal. Hpo contains a polyglutamine stretch and in a less-conserved region, both Drosophila and human proteins contain potential caspase cleavage sites. BLAST searches of the fly genome do not reveal a second Mst1-like gene, suggesting that hpo may represent both Mst genes in Drosophila. A conserved C-terminal region includes a dimerization domain. Human Mst proteins have been reported to be cleaved by caspases during apoptosis (Lee, 1998). A potential caspase cleavage site exists in Hpo, although the regions surrounding this site do not share much sequence similarity with human Mst1 and Mst2 (Harvey, 2003).

Human Mst kinases were originally characterized as proteins that were activated in response to cellular stresses (Taylor, 1996). Hpo and Mst kinases belong to a large family of proteins that have a kinase domain similar to that of the yeast Sterile 20 protein, which is thought to be a mitogen-activated protein kinase kinase kinase kinase (Map4K) (Dan, 2001). MAP4Ks possess a conserved kinase domain and are postulated to function at the apical point in Map kinase signaling cascades, which have been reported to regulate diverse biological functions including cellular growth, proliferation, differentiation, and apoptosis (Chang, 2001). Hpo may trigger activation of one or more MAPK pathways and indeed Mst1 and/or Mst2 have been shown to activate Jun N-terminal kinase (JNK) in mammalian cells (Graves, 1998). Recently human Mst1 has been shown to directly phosphorylate histone H2B and may mark chromatin for condensation during apoptosis (Cheung, 2003). Mst can also bind via its C terminus to a family of proteins lacking known enzymatic domains that include NORE and can function in a proapoptotic pathway downstream of Ki-Ras (Khokhlatchev, 2002) in mammalian cells (Harvey, 2003 and references therein).

date revised: 20 February 2004

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.