Gene name - tribbles
Cytological map position - 77B6--9
Function - signal transduction
Keywords - cell cycle, protein degradation pathway
Symbol - trbl
FlyBase ID: FBgn0028978
Genetic map position - 3-
Classification - serine/threonine kinase-like domain
Cellular location - cytoplasmic and nuclear
Tribbles activity regulates cell cycle by directly and posttranscriptionally affecting String expression. The Cdc25 protein String achieves mitotic activation by hydrolyzing phospho-tyrosine 15 of the cyclin dependent kinase Cdc2, thus activating Cdc2. string is transcribed in a spatial pattern controlled by the anterior-posterior and dorsoventral patterning systems. Expression of String mRNA in a given mitotic domain precedes mitosis by a few minutes. By analyzing the exception to this rule found in domain 10 on the ventral side at the embryo, the tribbles mode of regulation was uncovered. Although string is expressed in these cells, they do not divide until they are internalized. This delay depends on the activity of the tribbles gene (Seher, 2000) named after the small, round, fictional organisms (from the television series "Star Trek") that proliferate uncontrollably when they contact water. The tribbles effect is restricted to the ventral furrow, even though TRBL mRNA is also present outside of this domain and the trbl mutation can be rescued by uniform exogeneous expression. This suggests that trbl activity is triggered by an input which is present only in the ventral furrow region (Großhans, 2000). Tribbles acts by specifically inducing degradation of the CDC25 mitotic activators String and Twine via the proteosome pathway. By regulating CDC25, Tribbles serves to coordinate entry into mitosis with morphogenesis and cell fate determination (Mata, 2000).
tribbles was identified by three screens: (1) a screen designed to identify loci required for a delayed mitosis in domain 10 (Großhans, 2000); (2) a deficiency screen for mutations that disrupt gastrulation (Seher, 2000), and (3) a screen for genes that when overexpressed in the germline would affect oogenesis (Mata, 2000), In the first screen, embryos deficient for trbl, and for a second uncharacterized mutant frühstart, mitotis takes place prematurely in the ventral domain, such that their pattern of String mRNA expression and the mitotic pattern match each other (Großhans, 2000). In the second screen, the deficiency of tribbles results in the improper formation of the ventral furrow (Seher, 2000). In the third screen (Mata, 2000), an insertion in a vector designed to be overexpressed in the germline produced 32 cell oocytes (cysts) instead of the usual 16 cell oocyte (1 egg and 15 nurse cells).
Tribbles function was examined by assessing the effects of overexpression in imaginal disc cells. The effect of expressing tribbles in wing imaginal discs was assessed using engrailed-GAL4 to drive expression in the posterior compartment. The anterior compartment served as control tissue. The resulting wings had a posterior compartment of close to normal size and pattern, but apparently made up of fewer, larger cells. This interpretation was confirmed by looking directly at the disc, where the cells expressing tribbles, the posterior cells, were found to be fewer and larger than control cells. This type of effect is typical of severe cell cycle delays. Since the effect is confined to one compartment, the wild-type (anterior) cells cannot out-compete the tribbles overexpressing cells. To confirm that tribbles affects the cell cycle and to determine what phase is being affected, cells size analysis, using green fluorescent protein marked cells, was performed on tribbles-expressing cells. Expression of tribbles results in a 4- to 5-fold decrease in the number of cells. DNA content indicates that almost all of these cells are in G2/M. It is concluded that tribbles overexpression slows the G2 phase of wing imaginal disc cells substantially (Mata, 2000).
String/CDC25 phosphatase is required for G2/M progression in imaginal disc cells. Cells homozygous for a hypomorphic string mutation can grow but they divide very slowly and thus become very large, similar to the tribbles overexpression phenotype. Given that string is a limiting regulator in G2 of imaginal disc cells, it was asked whether increased expression of string would affect the tribbles GOF phenotype. Expression of UAS-string under control of engrailed-GAL4 did not give an overt wing phenotype. When UAS-string is coexpressed with UAS-tribbles, the posterior compartment reverts to the normal cell number and density. Thus, overexpression of string suppresses the effect of tribbles overexpression in imaginal disc cells. In the germline of the ovary, overexpression of string also suppresses the extra cystocyte division phenotype caused by tribbles expression. Thus, despite the apparent differences in cell cycle control mechanisms, tribbles opposes or downregulates string/CDC25 in both the wing disc and the germline (Mata, 2000).
To see whether overexpression of tribbles affects String protein levels, discs were double-stained with anti-String and anti-Tribbles antibodies. Cells that overexpress tribbles do not show normal levels of endogenous String protein. String expression in third instar larval discs is nonuniform due to asynchronous cycling of the cells. By counting expressing and nonexpressing cells in several discs, it was verified that String-expressing cells do not express Tribbles. The lack of cells coexpressing String and Tribbles could not be explained by cells being in the wrong phase of the cell cycle to accumulate String, since overexpression of tribbles delays cells in G2, the phase of the cell cycle in which String protein normally accumulates. Expression of tribbles also affects the level of String protein when both were overexpressed by the GAL4/UAS system. The overall level of overexpressed String protein is clearly reduced by Tribbles, although not to background levels. The simplest interpretation of these results is that Tribbles directly and posttranscriptionally affects String expression (Mata, 2000).
To investigate the interaction between tribbles and string under more controlled conditions, tissue culture experiments were carried out. The CDC25 homologs String and Twine, and the mitotic Cyclins A and B, were epitope tagged at the N terminus and cloned into the metallothionine expression vector. When cotransfected into Schneider cells, expression of Tribbles severely reduced the level of HA-String. This reduction in protein level was observed without any change in HA-String mRNA levels. Tribbles has a similar effect on HA-Twine protein levels, although less severe, but has no effect on the mitotic Cyclins A and B. Given that the promoter, the 5'UTR and 3'UTRs as well as the initiator methionine and N-terminal epitope tag were identical in the HA constructs, it was reasoned that Tribbles might be acting by increasing String (and Twine) protein turnover. To test this hypothesis, the proteosome inhibitor lactacystin was used. Addition of lactacystin reverses the effect of Tribbles on HA-String protein. This indicates that Tribbles induces degradation of String protein via a proteosome-dependent pathway. To confirm this, a pulse-chase experiment was performed. As expected, metabolically labeled HA-String disappears more rapidly in the presence of Tribbles. Mitotic cyclins are also degraded via the proteosome, stimulated by anaphase-promoting complex (APC). Because Cyclin levels are unaffected by Tribbles, it has been concluded that Tribbles regulates CDC25 protein turnover in a specific manner, not just by generally increasing proteosome activity (Mata, 2000).
The involvment of tribbles in regulating mitosis during gastrulation is documented by Großhans, 2000 and Seher, 2000 (for a description of the Seher study, see Effects of Mutation). During gastrulation, new cells, formed during the first 13 cell cycles, change their shape, and morphogenetic movements rearrange their positions relative to one another. One of the most prominent of these morphogenetic movements is the formation of the ventral furrow, which brings the mesoderm anlage into the interior of the embryo. After the first 13 divisions, cell division ceases, only to resume during gastrulation. During this time mitosis occurs in an asynchronous manner, independently, in at least 25 domains. This asynchrony allows morphogenesis and cell division to occur simultaneously in different regions of the embryo and sequentially in specific primordia. For example, the ventral-most cells first form the ventral furrow and only after this invagination is completed, do they enter mitosis (Großhans, 2000).
Entry into mitosis is positively controlled by String, the homolog of Cdc25, which is both necessary and sufficient for mitosis during gastrulation. The expression pattern of String mRNA closely matches the mitotic pattern. During the cleavage stage String mRNA is uniformly distributed, but then degraded at the pause in mitosis and the transition to cellularization. String mRNA subsequently reappears at the beginning of gastrulation in a pattern preceding the mitotic domains. In all domains except one, mitosis starts a few minutes after String mRNA is expressed. Mitotic domain 10, which comprises most of the mesoderm anlage, behaves differently: the gap between String mRNA expression and entry into mitosis is much longer (Großhans, 2000).
The delay in their mitosis suggests that ventral cells contain a factor lengthening the gap between appearance of String mRNA and entry into mitosis. This delay involves a subtle titration of string activity, since it can be shortened by addition of two more copies of the string chromosomal region, raising the copy number of string to four. Under these conditions the ventral cells divide at about the same time as the cells of domains 1 to 3, which matches the String mRNA pattern more closely than it does in wild-type embryos. Only the mitosis in domain 10 is shifted in these experiments, the order mitosis in the other mitotic domains is not changed (Großhans, 2000).
To examine more stringently whether the factor counteracting string is specific for ventral cells, exogeneous String mRNA was expressed at the same level in all cells of the embryo, using a UAS-String transgene driven by a maternally provided Gal4 in embryos otherwise homozygous for a string deletion. Four copies of maternally provided Gal4 produce high levels of string activity, indicated by the uniform entry of all cells into mitosis immediately at the beginning of gastrulation. In these embryos ventral furrow formation is inhibited. Using females with three or two Gal4 insertions, expression of string was gradually lowered. This shifts the onset of mitosis to a time when the first mitoses normally occur in wild-type embryos. Under these conditions differences in the behavior of the cells become apparent. In spite of the uniform string expression, the ventral cells undergoing cell shape changes to form the ventral furrow enter mitosis later than the other cells. This special behavior of the ventral cells is not observed in string heterozygous embryos that have endogeneous as well as exogeneous String mRNA. It is concluded from these experiments that ventral cells contain a dosage-sensitive factor, the ventral inhibitor, that counteracts string activity and that the delay of mitosis in domain 10 of wild-type embryos is due to this factor (Großhans, 2000).
In order to identify components that constitute the ventral inhibitor, a genome-wide screen was carried out for loci that are required for a delayed mitosis in domain 10. By screening 99% of the genome, two novel loci, frühstart and tribbles, were identified. In embryos deficient for either of these genes, cells in the ventral domain are the first to enter mitosis, such that their pattern of String mRNA expression and the mitotic pattern match one another. The order of the other mitotic domains is not altered, suggesting that frs and trbl act specifically in the ventral cells. The double mutant frs trbl shows the same phenotype as the single mutants, suggesting that frs and trbl are nonredundant genes in a common process (Großhans, 2000).
As a consequence of the early mitosis, the mesodermal precursors remain on the surface and do not form a proper ventral furrow. This defect is similar to that observed in embryos in which all cells have been forced into mitosis by increased string dosage or string overexpression. Although other zygotically active genes are known to affect formation of the ventral furrow, frs and trbl are unique in that their defects solely depend on the premature mitosis. In double mutant frs string and trbl string embryos, no mitosis takes place during gastrulation, and the ventral furrow forms as in wild-type. The premature mitosis in frs or trbl embryos is not caused by overexpression of String mRNA in the ventral region, since String mRNA is present in comparable amounts in mutant embryos and with a similar pattern as in their heterozygous siblings or wild-type embryos. Since String is the rate-limiting factor for entry into mitosis during gastrulation, this observation suggests that frs and trbl counteract string via a posttranscriptional mechanism (Großhans, 2000).
In embryos mutant for either snail or twist, no ventral furrow forms and cells are shifted to more lateral fates. String mRNA is not present in domain 10 and mitotic patterns in the ventral region of these mutant embryos are difficult to evaluate. String mRNA is restored to wild-type levels in the prospective domain 10 of snail mutants carrying three copies of wild-type twist. In such mutants, the ventral cells are the first ones to divide, indicating that snail is required for the function of the ventral inhibitor. One possibility would be that the persistence of trbl expression in the ventral region requires mesodermal determination and thus wild-type snail activity. However, snail mutants show a normal pattern of trbl expression and maintain trbl expression in the ventral domain. Similarly, in twist homozygous mutants and in embryos homozygous for deficiencies for frs the expression of trbl is not changed. Because snail embryos do not show a ventral mitotic inhibition, even though their trbl expression is normal, it is concluded that some aspect of mesodermal determination mediated by snail is required for Trbl activation (Großhans, 2000).
To directly investigate the activity of trbl on mitosis, the rapid cell cycle of cleavage stage embryos was used as an assay. trbl is not yet expressed during this stage: it appears only at the beginning of cellularization when the rapid nuclear divisions stop. Large amounts of synthetic Trbl mRNA were injected into the posterior half of the embryos: the subsequent nuclear divisions were followed, and the embryos were fixed before reaching gastrulation. In addition, the progression of the cell cycle after injection was recorded in embryos that express a GFP-Histone fusion protein that labels interphase and mitotic chromosomes. The nuclei at the injection site do not participate in the last mitosis (13th division) resulting in larger nuclei and lower density at the injection site. As a negative control for this assay, mRNA of the unrelated serine-threonine protein kinase Pelle was injected. This did not inhibit mitosis. Since presence of trbl alone is not sufficient to inhibit mitosis and since Trbl requires an additional trigger in the ventral cells, it is possible that this requirement is overridden by providing Trbl mRNA in excess in the microinjection assay (Großhans, 2000).
tribbles encodes a predicted 483 amino acid protein that shows homology throughout most of its length to a canine protein called c5fw and to closely related proteins (Wilkin, 1996, Wilkin, 1997). The function of these mammalian proteins is unknown. The only notable structural feature of c5fw is its similarity to two subdomains of protein kinases, a feature conserved in the Drosophila Tribbles protein. However, the c5fw protein, as well as Tribbles, lacks key residues usually required for kinase activity and c5fw does not appear to be a kinase (Wilkin, 1997). The sequence of the Tribbles protein does not suggest how it is acting biochemically. The only notable structural feature of Tribbles is the similarity to kinase subdomains VIII and IX, which may prove to be important (Mata, 2000).
Sequencing of the cDNA shows that tribbles encodes a protein with extensive similarity to the SNF1 class of serine/threonine kinases. Domains IV to XI are highly conserved, whereas domains I to III diverge significantly from the consensus. Four vertebrate proteins (including two human ESTs), which are the closest known relatives of Tribbles, share these characteristics. Perhaps the most significant property shared by these proteins that distinguishes them from other members of the protein kinase family is the fact that an asparagine in domain VIb (N171 of protein kinase A-Calpha) that occurs in all known protein kinases is replaced by an arginine (R269 in Tribbles). Thus, together these proteins define a new subgroup within the protein kinase superfamily. Whether they act as kinases remains to be tested (Seher, 2000).
Comparison of the primary structure of Tribbles with the consensus sequence of protein kinases shows severe deviations from the kinase consensus sequence. Three of the invariant residues are changed in Trbl: a KR change in the ATP binding site in subdomain II; an NR change in the catalytic loop within subdomain VIB, and a DS exchange in subdomain VII. In addition, the highly conserved histidine of the catalytic loop is substituted by a leucine. These severe deviations from the consensus make it unlikely that Trbl is a functional protein kinase. A second striking feature of the primary sequence is the abundance of serine residues in the N-terminal region of Trbl. Twenty-two of the first 49 residues are serines, which includes a piece with nine consecutive serines (Großhans, 2000).
The deviations from the consensus of protein kinases make it unlikely that Trbl is a functional protein kinase. To test this possibility further, an allele of Trbl (K266R) was generated, mutated at a crucial lysine of the catalytic center thought to govern the specificity of kinases for their targets. This lysine was changed to an arginine, a change which should maintain the configuration of the active site but alter its specificity. A lysine is present in all serine-threonine protein kinases, whereas an arginine is characteristic of conventional tyrosine kinases. When injected into embryos, this allele also induces a premature pause in the cell cycle (20 out of 41 embryos scored), confirming that Trbl does not function as a serine-threonine protein kinase (Großhans, 2000).
date revised: 1 September 2000
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.