Gene name - vrille
Cytological map position - 25D6
Function - transcription factor
Symbol - vri
FlyBase ID: FBgn0016076
Genetic map position - 2-17
Classification - basic leucine zipper protein
Cellular location - nuclear
Vrille is a bZIP transcription factor related to other leucine zipper proteins found in all vertebrate species. vrille plays a role in two complex Drosophila pathways. Initially described as a maternal effect lethal mutation, various alleles produce impaired zygotic development resulting in ventralization. vrille acts as enhancer of decapentaplegic (ddp) mutations during the development of dorsal/ventral polarity in the embryo (George, 1997). vrille was later found in a screen for genes whose expression is controlled by the biological clock of adult flies (Blau, 1999). The closest mammalian homolog to VRI is E4BP4, a transcriptional repressor (Cowel, 1992).
Mutants in dpp result in the ventralization of the embryo. In such mutants the dorsal region, where dpp is normally expressed, is ventralized and develops characteristics of the ventral neuroectoderm. Genetic screens have been performed to identify genes interacting with dpp. Genetic interactors might act (1) upstream of dpp, regulating dpp expression either directly or indirectly, or (2) downstream, regulating genes coding for protein involved in the transmission of the dpp signal, or regulating various targets of the dpp pathway. Genes discovered in such screens include Mothers against dpp (Mad) and Medea, a MAD homolog: both genes appear to be essential components of the dpp pathway. vrille is one such dpp interactor, acting as an enhancer of dpp and easter mutations. Easter is involved in triggering a protease cascade that leads to the activation of Dorsal, which functions to repress dpp in the ventral portion of the embryo during early development.
Putative truncated Vri proteins with no bZIP domain, as well as P element induced mutations and P excision secondary mutations, are able to maternally aggravate the embryonic ventralized ea and dpp phenotypes; this suggests a maternal participation of vrille in early establishment of dorsoventral polarity of the embryo. vri mutations alter wing vein differentiation, which leads to a phenotype very similar to that of shortvein alleles of dpp. Furthermore, vri enhances dpp phenotypes in wing and Mad mutation acts as a dominant enhancer of the vri wing phenotype. vri is expressed both in embryonic-specific regions of the gut, and in the larval gut and imaginal discs. Although the function of the gene remains to be elucidated, vri appears to be active during various stages of development and in different cell types (George, 1997).
bZIP transcription factors can act as heterodimers, acting together with partner bZIP proteins. Genetical analysis suggests that Vri acts in concert with another factor. It is noteworthy that all the vri alleles tested behave like antimorphs, since they show a dominant enhancement of dpp phenotypes stronger than a deficiency. In other words, an altered Vri protein shows a stronger effect than the complete absence of Vri. Therefore, a non-functional Vri protein, which is still able to bind to another factor, might be expected to have a stronger effect than the simple absence of one copy of the gene. The same observation has been made with tkv. Df(2L)tkvSz2, which totally deletes the nearby tkv genes and vri as well, does not enhance dpp, whereas tkv point mutations do. The occurrence of a large number of antimorphic alleles is consistent with the nature of the two proteins, a bZIP transcription factor and a serine-threonine kinase receptor, which both act as dimers (George, 1997).
What other protein could partner with Vri in regulating the Dpp pathway? Three of the known D. melanogaster bZIP transcription factors are implicated in the dpp pathway. The homolog of the mammalian fos oncogene, Drosophila Fos related antigen (Fra or Kayak), is expressed in embryos (in the head, dorsal ectoderm, amnioserosa, a subset of cells of the peripheral nervous system, a portion of the midgut, the hindgut and in the anal pads). The Fra partner, Jun related antigen (Jra or Djun), is expressed throughout development. It is noteworthy that, in mammals, these genes are induced in response to TGFbeta. A Drosophila homolog of CREB, Cyclic-AMP response element binding protein A (dCREB-A), is expressed in ovarian columnar follicular cells and in male reproductive organs, embryonically in salivary glands and the brain, in the optic lobe, and in the midgut of the adult. It is interesting to note that it has been shown that dCREB-A is required for dorsoventral patterning of the larval cuticle; it was proposed that it functions near the end of both the DPP-and SPI-signaling cascades. Some localizations of these bZIP factors coincide with the Vri expression pattern. Furthermore, putative Fos/Jun and CREB binding sites have been identified close to the Zerknullt (a Hox transcription factor of the dpp pathway) binding sites in the RACE (an amnioserosa specific gene) promoter. It has been suggested that Hox-bZIP synergy is a common feature of dpp signaling. It will therefore be interesting to test for a genetic and molecular interaction between these factors and vri (George, 1997 and references therein).
Several other genes in the dpp pathway exhibit maternal ventralizing effects. These include schnurri, which encodes a zinc finger transcription factor, and punt , a type II TGFbeta receptor. However, the level of ventralization is stronger with vri than with punt, schnurri or thick veins. Therefore, partial reduction of vri function in mothers enhances the slight ventralization induced by a weakly ventralizing easter allele. No additional increased ventralization is observed when a vri allele is also provided zygotically. Furthermore, the maternal effect of the vri gene enhances dpp ventralizing embryonic phenotypes. This is observed with all the available lethal vri alleles and two different dpp alleles. In the strongest interaction, no vri/dpp embryos survive and moderately ventralized embryos are observed, whereas in the same context with tkv, only the more moderately ventralized V4 embryos are recovered. With some vri alleles a zygotic effect is observed (George, 1997).
Thus, vri plays a role in dorsal/ventral patterning. Its exact position in this pathway is not known, particularly because there are likely to be many targets of Dpp signaling and because Vri is likely to act with a partner in regulating multiple genes both maternally and zygotically.
vrille, in its role as a clock gene, was cloned in a differential display protocol designed to discover genes expressed in the adult head. The circadian clock runs using a transcriptional negative feedback loop involving Period/Timeless and Clock/Cycle dimeric complexes. It was hypothesized that other genes with important clock roles may be regulated by this loop. Differential display was used to search for genes whose RNA levels in adult Drosophila heads respond to nuclear entry of the Per/Tim complex. A comparison was made of the profiles, over time, of wild-type flies and per01 mutants to select for clock-controlled rather than light-regulated transcripts. per01 mutants make no Per protein, and therefore no Per/Tim dimeric complex enters the nucleus (Blau, 1999).
vri is expressed in circadian pacemaker cells in the brain; VIR mRNA oscillates in phase with per and tim expression in wild-type flies, and vri is regulated by the same transcriptional loop that controls PER and TIM mRNA levels. Flies with altered levels of vri show a range of behavioral rhythm phenotypes. Reducing vri gene dosage causes period shortening. Suppression of the normal cycle of vri expression generates long-period rhythms or arrhythmicity. These latter phenotypes are associated with a block in per and tim expression, indicating that vri regulates the central clock. Accumulation of pigment dispersing factor (PDF), a neuroactive peptide hormone, is also suppressed, suggesting that vri additionally connects the clock to behavior (Blau, 1999).
Since VRI mRNA oscillates, is expressed in pacemaker cells, and encodes a transcription factor, it seemed possible that vri would regulate patterns of gene activity involved in behavioral rhythmicity. Effects of Vrille on behavior were monitored either by decreasing the gene dosage of vri, or increasing gene dosage. vri homozygotes are lethal, dying late in embryogenesis, thereby preventing testing of the phenotypes of adult flies with no functional Vri. To test whether vri mutations affect rhythmic behavior, the locomotor activity rhythms of adult Drosophila heterozygous for null mutations of vri were examined under three conditions: (1) flies heterozygous for vri1, a null allele with a single point mutation that introduces a stop codon upstream of the bZIP domain; (2) alleles of vri5, a lethal P element inserted in the vri gene; and (3) Df(2L)-cl(h3), in which the vri chromosomal region is deleted. In all cases, reducing the dosage of vri by half shortens the period length of the locomotor activity rhythm by 0.4 to 0.8 hr. These effects on period length are comparable in magnitude to those associated with a heterozygous deletion or a duplication of the per locus. These results show that a particular level of vri activity is required to set the period of the Drosophila clock to 24 hr (Blau, 1999).
If a VRI mRNA or protein oscillation of specified amplitude is required for wild-type function of the Drosophila clock, behavioral rhythms might also be disturbed by eliminating cyclical expression of vri. Since vri is essential for proper development, the GAL4-UAS system was used to continuously express vri only in clock cells. The driver P element contained the timeless promoter for clock cell specificity, with five GAL4-binding sites (UAS) inserted 333 bases upstream of the putative start site of transcription. This was fused to the yeast transcriptional activator GAL4 to make the P element tim(UAS)-gal4. It was reasoned that with such a construct, endogenous factors would activate expression from the tim promoter, and subsequently GAL4 proteins should positively influence their own expression. In this way, oscillations from the tim promoter might be negated. Flies were transformed with either this P element or with a UAS-vri cDNA P element, and one tim(UAS)-gal4 line was crossed to three independently isolated UAS-vri lines (V1, V2, and V3) (Blau, 1999).
Initially tested were the effects of these transgenes on vri oscillations. In the presence of driver but no UAS-vri transgene, VRI mRNA levels show a 9-fold oscillation in flies held in constant darkness for 1 day. In contrast, in flies that also contain the UAS-vri P element, vri levels are close to the peak of the wild-type oscillation at all times of day. There is still a 2.5-fold oscillation in VRI mRNA levels, but the phase of this weak oscillation is reversed with the peak now at CT2 (CT indicates circadian time: time in constant darkness). These results demonstrate that the normal vri oscillations have been removed. The reason for the weak antiphase oscillation is not known, but it may contribute to the behavioral phenotypes observed (Blau, 1999).
Next the behavioral rhythms of the three UAS-vri lines crossed to the tim(UAS)-gal4 driver were tested and a series of phenotypes were observed. Line V1 flies are all rhythmic, but the period of the behavioral rhythm is lengthened to 25.5 hr. Most flies in lines V2 and V3 are arrhythmic (78% and 92%, respectively). A few flies have either long-period rhythms of ~28 hr or weak rhythms of 26.5 to 29 hr. However, visual inspection of the records of V3 flies considered rhythmic by computer analysis indicate that these are atypical long-period rhythms. All of these mutant phenotypes require the tim(UAS)-gal4 driver, since control flies homozygous for the UAS-vri transgenes alone uniformly show wild-type rhythms. These rhythm phenotypes are not simply due to overexpression of any bZIP protein, since flies expressing either D-Jun or D-Fos from the tim(UAS)-gal4 transgene have wild-type rhythms (Blau, 1999).
Thus perturbing vri levels causes different kinds of rhythm phenotypes. Flies missing one functional copy of vri have short-period rhythms, while continuous vri expression causes long-period rhythms or arrhythmia. The latter phenotypes are associated with changes in the molecular cycles of per and tim, and the strengths of the molecular and behavioral phenotypes are perfectly correlated. It is concluded that a normal vri oscillation is essential for a 24 hr clock at both the molecular and behavioral levels. These results indicate that vri is a clock gene as well as a clock-controlled gene (Blau, 1999).
One of the most interesting questions in circadian biology is how a molecular cycle is translated into time of day information for the behaving organism. The neuropeptide vasopressin is directly regulated in the mouse by cycling activity of the CLK/BMAL-1 complex (Jin, 1999). Since vasopressin is a well known modulator of neuroendocrine function, its regulation by the CLK/BMAL-1 complex indicates how some physiological responses can be directly programmed by a circadian clock. This report shows that pdf expression is regulated by the Drosophila clock and requires cycling vri expression. pdf encodes a neuropeptide expressed in the axons of the pacemaker cells, and these projections connect the LNs with target cells in the dorsal brain. PDF protein has been shown to accumulate in the LN axons with a circadian rhythm. The period of this rhythm is shortened by the perS mutation, and continuous accumulation of PDF in the dorsal brain is associated with arrhythmia and a variety of period changes in adult locomotor activity. PDF mRNA levels do not cycle in wild-type flies. Since continuous expression of vri suppresses PDF protein accumulation without affecting accumulation of pdf mRNA, cycling Vri expression in wild-type Drosophila is likely to contribute to the observed cycling of Pdf protein. Vri may affect Pdf levels by specifying rhythmic expression of a factor involved in translation, maturation, stabilization, transport, or release of the neuropeptide (Blau, 1999 and references therein).
All of these observations point to a likely role for Pdf in coupling a molecular clock to timed behavior; this study has demonstrated that vri conveys essential regulatory signals from the clock to Pdf. There is also evidence that Pdf can in turn influence function of the clock. In the cockroach, microinjection of PDF produces time-dependent shifts in the phase of the locomotor activity rhythm (Petri, 1997). The magnitude of these phase shifts (up to 4 hr) is similar to that produced by light (Petri, 1997). This indicates that a transient change in Pdf level will cause a stable change in molecular components of a clock that regulates behavior in at least some insects. Possibly, the novel pathway of per and tim suppression observed in V2 and V3 Drosophila lines is a direct consequence of eliminating Pdf (Blau, 1999).
A striking conservation of some features of vri expression is seen in flies when compared to members of the related PAR domain family of mice transcription factors. Drosophila vri and murine dbp, tef, and hlf RNAs all show circadian oscillations (Wuarin, 1990; Falvey, 1995; Fonjallaz, 1996); loss of dbp affects mouse locomotor activity rhythms (Lopez-Molina, 1997). Thus, a role in circadian rhythmicity may be conserved for this family of transcription factors throughout much of the animal kingdom (Blau, 1999).
Three cDNAs (cDNA2, 7 and 8) were isolated from a 3 to 12 hour cDNA library. Two introns are predicted in the vrille gene. Intron 1 is located after position 459 and is 1396 bp long; intron 2 maps after position 807 and is 108 bp long. In both cases the predicted splice donor and acceptor sites (GT at 5' and AG at 3' splice sites) fit the invertebrate splice junction consensus. cDNA7 contains a polyA tail (24A) preceded by a consensus polyadenylation site (AATAAA) at position 3764. Two other potential polyadenylation sites were found at positions 3373 and 3385 but no corresponding polyadenylated cDNAs were isolated. The 5' end of cDNA1 begins with a G, not present in the genomic sequence. This residue may have been introduced by the reverse transcriptase copying the cap site, indicating that cDNA1 is probably full-length. No TATA or CAT box is found, however, within the 300 bp mapping upstream from the 5' cDNA1 end. At the 3' intron 1 junction, -11 nucleotides upstream from the 5' end of cDNA3, the CAGTTC sequence that fits the cap site consensus is found. Within intron 1 at positions -86 and -104 upstream from the 5' cDNA3 end are found two TATA box consensus sites, and at position -121 is found a ACAAT sequence that fits the CAT box consensus. These potential transcriptional regulatory domains and the potential cap site sequence are consistent with the prediction that cDNA3 is an almost full-length cDNA. This hypothesis is fully supported by the fact that an identical 5' cDNA end was found when using a 4 to 8 h cDNA library. The largest cDNA predicted from cDNA1 and cDNA7 is 3807 bp and encodes a putative 729 amino acid protein. The first ATG in frame at position 318, however, is preceded by the ACATT sequence, which does not perfectly fit the C/AAAA/C consensus of Drosophila translation initiation sites. A second type of cDNA, 3340 bp long is predicted from cDNA3 and cDNA7 and encodes a putative 610 amino acid protein starting at position 675 and lacking the first 119 amino acids (George, 1997).
The predicted Vrille protein is rich in serine (14%), glutamine (8%) and proline (8%) residues. Two runs of histidines and glutamines (OPA repeats), frequently found in eukaryotic transcription factors, are detectable between positions 61-68 (7 over 8) and between positions 413-434 (17 over 22). A GS box consisting of a serine and glycine repeat is found between positions 325 and 342 (17 over 18). GS boxes have been described in serine-threonine kinase type I TGFbeta receptors and are juxtamembrane intracellular domains with potential sites for phosphorylation on the serine residues. Searches in the GenPept protein database reveal (between positions 234 and 297) a motif characteristic of the leucine zipper class of DNA binding proteins common to the two putative proteins. These proteins contain two contiguous domains, a basic DNA binding domain and the leucine zipper involved in homo- or hetero-dimerization of the proteins. The most closely related proteins are the bZIP encoded by the gene 9 of X. laevis and the human NF-IL3A/E4BP4 protein. The Gene 9 protein (similar to Gene 8) is induced by thyroid hormone during metamorphosis in the tadpole resorption program and is 63% identical and 87% similar to Vri within 68 amino acids of the bZIP domain. E4BP4, a placental human transcription factor, is identical to the human NF-IL3A protein that binds and transactivates the interleukin-3 (IL-3) promoter in T cells. NF-IL3A/E4BP4 shares 60% identity and 93% similarity with Vri, across the 68 amino acids within the bZIP domain. NF-IL3A/E4BP4 binds specifically to regulatory sequences in the interleukin-3 promoter, the adenovirus E4 promoter and human gamma interferon promoter. NF-IL3A/E4BP4 is a multiple site phosphorylated bZIP factor and has been also isolated as a binding protein of the CREB/ATF-like sequence of the upstream regulatory region of human interleukin-1b gene. Vri is also related to the PAR subfamily of bZIP proteins, which includes VBP/TEF, DBP and HLF (hepatic leukaemia factor). The PAR proteins display a high identity, not only within the bZIP domain but also over a conserved proline and acidic-amino-acid-rich (PAR) adjacent domain, at the amino terminal to the basic region. NF-IL3A/E4BP4, Gene 9, Drosophila Giant, and the product of the cell-death specification gene of C. elegans, CES-2 and Vrille share a high identity within the bZIP domain. Unlike the PAR proteins, however, Vri does not contain a PAR domain. Nevertheless, NF-IL3A/E4BP4, Giant, CES-2 and the PAR proteins share similar consensus binding sites. It is noteworthy that E4BP4, Giant and CES-2 have been shown to act as transcriptional repressors (George, 1997 and references therein).
Strikingly, vri shows strong conservation over its bZIP region to the PAR domain family of mammalian transcription factors, three of which, DBP, TEF, and HLF, show circadian oscillations (Wuarin, 1990; Falvey, 1995; Fonjallaz, 1996). Homozygous deletion of mouse dbp shortens the period of locomotor activity rhythms (Lopez-Molina, 1997). However, the mammalian gene with the most homologous bZIP domain to vri is the transcriptional repressor, E4BP4 (Cowell, 1992), which, like vri, has no PAR domain. E4BP4 has not been studied in relation to mammalian circadian rhythms (Blau, 1999).
date revised: 15 March 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.