BIOLOGICAL OVERVIEW

Drosophila onecut homeobox gene encodes a protein product with one cut domain (see Drosophila Cut for more information on cut domain function) and one homeodomain. Onecut can bind to similar DNA sequences with the same high specificity and affinity as Onecut proteins from other species through the highly conserved cut domain and homeodomain. Interestingly, the cut domain alone can mediate DNA-binding, but the homeodomain cannot. However, depending upon the promoter context, cooperative interactions are observed between the two domains to confer high DNA-binding affinity and specificity. Onecut appears to be a moderate transcriptional activator and functions as a nuclear protein in neuronal tissues of both the CNS and PNS during development and in the adult. In the eye, Onecut expression is independent of glass, a transcriptional regulator of R cell differentiation. Taken together, these results suggest a role for Onecut in the regulation of some aspects of neural differentiation or maintenance. In support of this notion, overexpression of a putative dominant negative form of Onecut during eye development does not affect early cell fate specification, but severely affects photoreceptor differentiation (Nguyen, 2000).

A filter DNA-binding screen was carried out to identify transcriptional regulators of rhodopsin (rh) gene expression using a Drosophila adult retina cDNA expression library with defined cis-acting regulatory sequences from the rh promoters as probes. Among the genes identified was a novel gene, termed A1 (Fortini, 1991 and references therein). A blast search of the subsequently isolated full length cDNA encoded sequence revealed two highly conserved regions in the carboxyl half of the conceptual protein: a single cut domain followed by a homeodomain. These motifs characterize a rapidly expanding group of homeodomain proteins known as the Onecut proteins (Lannoy, 1998). The founding member is hepatocyte nuclear factor-6 (HNF-6), which was originally identified among a group of nuclear factors required for liver gene expression (Lemaigre, 1996; Samadani, 1996). No homolog of HNF-6 has yet been identified in the fly, thereby, making Onecut the only known Drosophila Onecut family member (Nguyen, 2000).

The bipartite DNA-binding domains of Onecut and of other cut-homeodomain members illustrate a successful strategy for assembling multiple DNA-binding domains that function independently or cooperatively with the homeodomain to achieve a greater level of specificity or to enhance DNA-binding affinity. The cut domain of Onecut alone can clearly mediate high affinity DNA-binding in a sequence-specific manner. The homeodomain, however, is incapable of DNA-binding by itself. This is also true for the HNF-6 homeodomain as shown by mutant proteins that have either a deleted or mutated homeodomain but are still capable of binding (Lannoy, 1998). Likewise, among the Cut proteins, the homeodomain of mammalian CDP/Cux displays very low DNA-binding affinity and generally binds in a non-specific manner. Therefore, the Onecut homeodomain has no or relatively very low DNA-binding affinity. This may reflect sequence differences in the third DNA-recognition helix of the homeodomain. For example, the Onecut members have the defining F48 and M50 residues that are atypical divergences from the classical homeodomain (Nguyen, 2000).

However, the apparent inability to bind does not rule out a role for the homeodomain in Onecut DNA-binding. It is evident that the cut domain alone is not sufficient for binding, but depends on the presence of the homeodomain for the DNA-binding activity. Changes in DNA-binding affinity are also influenced by the homeodomain. This indicates a significant involvement of cooperative interactions between the cut domain and the homeodomain in determining Onecut DNA-binding activity. A similar case is seen for the POU-class homeodomain proteins. The POU homeodomain alone is insufficient for high affinity DNA-binding, but requires the POU-specific domain for effective interaction with target sites. Analogously, the Paired domain and the Paired homeodomain can function as independent DNA-binding domains, where the latter can operate through dimerization. However, the Paired domain and the homeodomain can also cooperate together to specify DNA-binding activity. Thus, cooperative interaction between bipartite DNA-binding domains appears to be an important mechanism for achieving higher DNA-binding affinity and sequence specificity, and is not exclusive to the cut-homeo domain proteins (Nguyen, 2000).

The indirect role of the homeodomain in Drosophila Onecut binding also points to the possibility that the homeodomain may participate in protein-protein interaction with other transcription factors to effect target specificity and transcriptional activity. One line of evidence in support of this is the observation that mutations in the F48M50 dyad of homeodomain of Onecut homolog HNF-6 do not abolish binding, but affect transcriptional activity in a target-dependent manner (Lannoy, 1998). For target sites that do not utilize the homeodomain for binding, the homeodomain may be involved in promoting transcriptional activation either directly or indirectly by recruiting other factors. In the case of Onecut, its weak transactivation activity may suggest a potential interaction with additional activators, Glass or Eyeless for example, in order to promote high levels of transcriptional activation (Nguyen, 2000).

Ectopic expression of onecut in early eye development appears insufficient to specify photoreceptor cell fate. Although Onecut is expressed in the nervous system throughout development, the developing eye was used as a sensitive assay for examining onecut function. Two independent UAS-onecut transgenic lines were generated that allow overexpression of full-length onecut protein using the Gal4/UAS binary expression system. The GMR-Gal4 driver was used to induce Onecut expression in all cells behind the morphogenetic furrow, and the sev-Gal4 driver was used for expression in a subset of cells that includes the R-cell and the cone cell precursors. If the expression of Onecut were sufficient to specify the photoreceptor cell fate, the formation of extra R cells would be expected in GMR-Gal4/UAS-D-onecut flies or the neural transformation of the cone cells in the sev-Gal4/UAS-onecut flies. However, eye development was normal in both lines examined. This is consistent with the idea that onecut is not involved in establishing R cell identity, but is required for some aspects of neural differentiation that occur subsequent to the specification of the R cell fate (Nguyen, 2000).

Thus, as with Glass, ubiquitous expression or expression of Onecut in the non-neuronal cells, such as cone cell precursors, has no apparent effect on eye development. Intriguingly, genes such as Chaoptin, which are directly regulated by Glass and are normally restricted to the R cells, do not respond to ectopic Glass expression in non-photoreceptors. Thus, Glass is necessary for the differentiation of the R cells but not sufficient to drive neural differentiation in the non-photoreceptor cells. This may indeed be the case for ectopically expressed Onecut. Two possibilities may explain these observations: (1) the R cells may have developed along a different developmental history, as compared to the non-photoreceptor cells, thus endowing them with other factors necessary for the initiation and maintenance of neural differentiation in response to regulatory molecules such as Onecut and Glass; (2) although Glass is expressed in all cells behind the morphogenetic furrow in the eye disc, Glass- dependent transcription is only restricted to the R cells. This suggests the existence of some inhibitory function in the non-photoreceptors. Interestingly, this restriction has been shown to be sufficiently directed by the conserved ATTG motif in the Rhodopsin1 promoter (Rh1PE). When the ATTG is mutated, Glass activity is no longer R cell-specific but is found in all cells. An inhibitor could bind to this ATTG to block Glass activity in the non-photoreceptors. If this inhibitor is ubiquitously expressed, an R cell-specific positive factor must exist to prevent this inhibition of Glass-mediated transcriptional activation. Given its ability to bind to the Rh1PE, its affinity for the ATTG motif of this promoter element, and expression in the photoreceptors, Onecut is a likely candidate for this R-cell specific positive factor (Nguyen, 2000).

Without a loss-of-function mutation in onecut it is difficult to investigate its role in neural development. In an attempt to address this issue, a dominant negative approach was tried, by fusing the DNA-binding domain of Onecut to the repressor domain of Engrailed and expressing the fusion protein under the control of the UAS promoter (UAS-EnCH). This approach has been used successfully to reveal the possible functions of transcriptional activators during development. Such fusion protein is expected to compete with endogenous wild type protein for target genes and to interfere with gene expression by active repression. When expressed in all cells behind the morphogenetic furrow during eye development as driven by GMR-Gal4, the EnCH fusion protein causes a dramatic reduction of the adult eye, giving it a rough, glossy, and flattened appearance. A histological section of the mutant eye reveals an apparent lack of R cells and a remnant of support cells such as pigment cells and cone cells. However, cryosections stained for the neuronal marker Elav, suggest that photoreceptor nuclei are still present, but that the retina does not fully extend as in wild type eyes. Morphologically, the optic lobes are reduced in size (Nguyen, 2000).

To determine when the initial developmental defect occurs, GMR-Gal4/+;UAS-EnCH/+ third instar larval eye discs were stained with an antibody to Elav, an early neuronal cell marker, and with an antibody to Cut, a marker for non-neuronal cone cells. The result shows that the EnCH protein does not affect early cell fate specification events, as pattern formation occurs normally for both R cells and cone cells. Abnormal R cell differentiation must have occurred during later stages of eye development. This observation is remarkably reminiscent of the loss-of-function glass mutants, in which early cell fate determination proceeds normally, but a defect in R cell differentiation leads to the subsequent degeneration of the retina (Nguyen, 2000 and references therein).

To gain insight on how the EnCH fusion protein may affect R cell differentiation, candidate genes that are known to be required for proper R cell differentiation were examined. Unlike glass mutants, the EnCH fusion protein does not affect the expression of Chaoptin, an R cell-specific protein that is under glass regulation. The expression of the homeodomain gene, orthodenticle (otd), also is not affected as determined by the expression of a lacZ-reporter gene under the control of an eye-specific otd enhancer. Surprisingly, a Rh1-promoter lacZ transgene does not respond to overexpression of the EnCH fusion protein and neither does a Rh1PE-lacZ reporter line. The response of a Rh4-lacZ transgene to the EnCH fusion protein was examined. In this case, occasionally only a few cells express the lacZ reporter. It cannot be distinguished whether this is due to an absence of R7 cells or that these cells are more sensitive to the EnCH fusion protein. Nevertheless, these findings are still consistent with the idea that onecut and other regulators such as glass may control different aspects of R cell differentiation by impinging on different target genes that are required during late stages of eye development (Nguyen, 2000).

With a distinct nuclear expression in neuronal cells throughout development and in the adult, Onecut is likely to play a role in regulating neural differentiation or maintenance by controlling neural-specific gene expression. In the eye, one candidate target gene is rhodopsin, as suggested by specific DNA-binding in vitro to two cis-acting elements of Rhodopsin1: RCSI and Rh1PE. Interestingly, the RCSI element is conserved not only in the opsin genes of many different species but also in the promoters of many R cell-specific genes, and is required for their expression in photoreceptors. Therefore, onecut might regulate other R cell-specific genes containing RCSI-like binding sites in addition to rhodopsin for proper R cell differentiation (Nguyen, 2000).

In addition to Onecut, several other transcription factors have been shown to be required for R-specific gene expression. For example, orthodenticle appears to be required for late differentiation of R cells. The loss of otd function causes a reduction in rh gene expression and abnormal morphogenesis of photoreceptor rhabdomeres. In addition to its role in initiating eye development, the Pax-6/eyeless gene has been implicated in regulating late differentiation events during Drosophila eye development by directly regulating target genes such as the Rh1 gene. Therefore, Onecut, Glass, and additional transcription factors may form a cross-regulatory network that coordinates retina-specific gene expression during Drosophila eye development (Nguyen, 2000).

The striking conservation of the cut domain and homeodomain among the Onecut genes suggests that their functions in sequence-specific DNA-binding and transcriptional regulation are conserved during the course of evolution. However, the genes or classes of genes they regulate may be entirely different. For example, the mammalian Onecut proteins are expressed in numerous tissues such as endodermal and neural derivatives (Landry, 1997; Rausa, 1997). In contrast, Onecut has exclusive neuronal expression. Interestingly, similar to Onecut, HNF-6 is also expressed in the retina, which first appears at embryonic stage E17 and implicates a potential role in photoreceptor development (Landry, 1997). The observation that Drosophila Onecut and the mammalian Onecut proteins share common expression in neuronal tissues may point to an ancestral site of action of the Onecut proteins (Nguyen, 2000).

In C. elegans there are five Onecut members (Lannoy, 1998). Interestingly, Ceh-21 and Ceh-39 are capable of binding to HNF-6 binding sites (Lannoy, 1998), which suggests that the C. elegans genes may function similarly to Drosophila Onecut and the mammalian Onecut proteins. The Ceh-38 gene has been reported to be expressed in multiple tissues throughout development like the mammalian genes, with particular expression in endodermal derivatives and in many types of neurons. The C17H12 gene perhaps represents the most ancestral member within this lineage since it is more like Onecut as indicated by sequence homology and by the possession of an STP box. Finally, since all but C17H12 show much greater divergence among the Onecut proteins, it will be interesting to see whether these genes will have overlapping expression patterns, like the HNF-6 and OC- 2, or have restricted expression pattern like onecut (Nguyen, 2000 and references therein).

GENE STRUCTURE

The onecut primary transcription unit is encoded by four exons. Interestingly, the cut domain is encoded entirely within the second exon; whereas the homeodomain is split and is encoded by exons 3 and 4, which is the case for some homeodomains, for example, the orthodenticle gene. Promoter prediction analysis of genomic sequences surrounding the 5' end of the full-length onecut cDNA suggests a highly probable transcription start site 112 nucleotides upstream and a TATA box 5' to it. Approximately 3.1 kb upstream of the onecut gene is a transcription unit encoding the Drosophila Eph receptor tyrosine kinase, which is transcribed in an opposite orientation (Nguyen, 2000).

cDNA clone length - 4.8 kb

Exons - 4

PROTEIN STRUCTURE

Amino Acids - 1081

Structural Domains

Structurally, Drosophila Onecut is more closely related to the mammalian Onecut proteins than to the Cut proteins. The cut domain and homeodomain of Drosophila Onecut share a high degree of amino acid identity with those of the mammalian Onecut proteins, 95% and 83%, respectively. Outside of the onecut subfamily, they are only about 28% to 46% identical to the corresponding domains of the other cut-homeodomain proteins. On further inspection, the homeodomain of Drosophila Onecut possesses amino acid residues that are conserved only among the Onecut members, such as the characteristic residues F48 and M50 in the third DNA recognition helix. Residue 50 has been shown to contact DNA in the major groove and is particularly important in conferring DNA-binding specificity. In the majority of homeodomains, residue 48 is always a tryptophan (W). The cut domain of Onecut is also remarkably conserved. For example, when compared to mammalian Onecut proteins, only four of the 74 residues differ, and of these four residues three are conservative changes. A third region that is highly conserved between Drosophila Onecut and other Onecut proteins is a serine/threonine/proline-rich sequence of about 28 amino acids called the STP box. Some amino acid residues within this sequence motif are found in all Onecut members including C17H12, the only C. elegans member that has a STP box. The STP box has been implicated to play a role in transcriptional activation based on experiments carried out on HNF-6 (Jacquemin, 1999; Lannoy, 2000). Finally, a conserved LSDLL motif within the cut domain of the Onecut proteins is present. This motif has been found to play a role in transcriptional stimulation as well and is capable of interacting with coactivators such as the CREB-binding protein (CBP) (Lannoy, 2000). In addition, Drosophila Onecut contains five unique tandem repeats, which has been called DOR (for Drosophila Onecut Repeats); each consists of 21 amino acid residues and the sequence is not found in any of the other Onecut members nor in any proteins present in the sequence database. An alignment of these repeats reveals that some residues are highly conserved, even in DOR2, which appears to be the most divergent of the repeats. Finally, several other features that have been described in a subset of Onecut proteins are not found in Drosophila Onecut, such as the polyglycine tract at the amino-terminus or the polyhistidine stretch located immediately downstream of the STP box. Whether any of these sequences have a functional role will have to await further studies, but clearly they are not evolutionarily conserved and may represent a recent acquisition by the different Onecut members during the divergence of the various phylogenetic lineages (Nguyen, 2000).

date revised: 22 March 2001

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.