BIOLOGICAL OVERVIEW

Su(H) is an integral part of the Notch signaling pathway and is neurogenic, like Notch. Su(H) protein is both cytoplasmic and nuclear. In the cytoplasm it interacts with the transmembrane Notch receptor that receives neurogenic signals from outside the cell. Su(H) then carries the Notch signal into the nucleus. Once there, Su(H) acts as a transcription factor regulating neurogenesis. The target of Su (H) is the Enhancer of split complex of genes which acts to suppress neural development. This suppressive function of the E(spl) complex is at the heart of the lateral inhibition function of the Notch pathway.

The names for Drosophila genes often run counter-intuitive to their actual function. One example of this perversity is the gene Hairless. One would assume a gene named "hairless" acts to restrict the number of bristles. Just the opposite is true. It seems that Hairless functions to enhance the formation of adult sense organs. Only when mutated does Hairless 's name match the results of its (in)activity, that is, the loss of adult kinesthetic sense organs. A fully functioning Hairless, novel in sequence, acts to oppose the antineural effect of Su(H), that is, the tendency of Su(H) to inhibit neurogenesis. Hairless physically interacts with Su(H), interferring with Su(H)'s ability to activate transcription of the Enhancer of split complex. And here too the naming is perverse. Enhancer of split genes oppose neurogenesis, yet they are grouped correctly among the neurogenic genes. Thus Hairless can be considered a neural patterning gene, carrying out its normal function by direct inhibition of Su(H) (Brou, 1994). One can speculate whether or not Hairless also interacts positively with proneural genes to enhance the neural fate.

Suppressor of Hairless has an integrase domain, homologous to a mammalian gene responsible for DNA site specific integration in the generation of immunoglobulin diversity. Site-directed mutagensis of the integrase-related region of Su(H) reveals that the protein is functional when the integrase domain is mutated (Schweisguth, 1994). What is the function in the fly of the integrase domain? An answer to this question is reserved for the future.

Aspects of Drosophila development seem to resemble development of C. elegans more and more at every turn. The assumption of alternative sister cell fates in asymmetric division and the role of cleavage plane in determining positional fate of cells are two common aspects that keep cropping up. Su(H) expression in sensory organ precursors (SOP) has an integral role in establishment of alternative cell fates, and also indicates the importance cleave plane in determining cell fate. SOPs divide once to give rise to pIIb and pIIa progeny. PIIb give rise to neuron and sheath (glial theocogen) cells and pIIa give rise to shaft (stimulus receptive) and socket cells. These four progeny of SOPs make up the mechanosensory bristle of the adult fly.

Specific accumulation of Su(H) in the socket cell, which is the receiver of lateral inhibition, is delayed, suggesting that the differential expression of Su(H) is a consequence of a pre-existing asymmetry. In any case, a high level of Su(H) is both necessary and sufficient for socket cell fate (Schweisguth, 1994).

Division of pIIa occurs before division of pIIb, resulting in a three-cell stage. The orientation of these three-cell clusters along the anteroposterior axis and the position of the socket cell within this cluster is reproducibly defined in the developing notum. The presumptive socket cell is always the most posterior of the three cells. It is probable that this alignment is primarily determined by the orientation of the cleavage plane. It is not yet known what factors are responsible for the positional clues determining cleavage plane (Gho, 1996).

A new hypomorphic numb mutant not only displays a double-socket phenotype, due to a hair cell to socket cell transformation, but also a double-sheath phenotype, due to a neuron to sheath cell transformation. This provides direct evidence that numb functions in the neuron/sheath cell lineage as well. These results, together with the observation from immunofluorescence analysis that Numb forms a crescent in the dividing IIa and IIb cells suggest that asymmetric localization of Numb is important for the cell fate determination in all three asymmetric cell divisions in the sensory organ lineage. In the hair/socket cell lineage, but not the neuron/sheath cell lineage, a Suppressor of Hairless mutation acts as a dominant suppressor of numb mutations whereas Hairless mutations act as enhancers of numb. Therefore Su(H) and Hairless are required for determining the hair/socket cell lineage but not the neuron/sheath cell lineage. The neuron/sheath cell lineage is not affected by the loss or increase of Su(H) activity. Epistasis analysis indicates that Suppressor of Hairless acts downstream of numb. Results from in vitro binding analysis show a physical interaction between Numb and Hairless and suggest that the genetic interaction between numb and Hairless may occur through direct protein-protein interaction. These studies reveal that Suppressor of Hairless is required for only a subset of the asymmetric divisions that depend on the function of numb and Notch. Since numb and Notch are involved in an antagonistic manner in all three asymmetric divisions of the sensory organ precursor lineage, and Su(H) is involved in only a subset of these, there must exist a Su(H)-idependent Notch signaling that has not yet been well characterized (Wang, 1997).

The receptor protein Notch plays a conserved role in restricting neural-fate specification during lateral inhibition. Lateral inhibition requires the Notch intracellular domain to coactivate Su(H)-mediated transcription of the Enhancer-of-split Complex. During Drosophila eye development, Notch plays an additional role in promoting neural fate independent of Su(H) and E(spl)-C, and this finding suggests an alternative mechanism of Notch signal transduction. Genetic mosaics were used to analyze the proneural enhancement pathway. Proneural enhancement involves upregulation of proneural gene expression in single cells that will become neurons. In Drosophila eye development, Notch (N) is required for proneural enhancement in addition to lateral inhibition. The molecular mechanism of proneural enhancement has not been determined. As in lateral inhibition, the metalloprotease Kuzbanian, the EGF repeat 12 region of the Notch extracellular domain, Presenilin, and the Notch intracellular domain are required. By contrast, proneural enhancement becomes constitutive in the absence of Su(H), and this leads to premature differentiation and upregulation of the Atonal and Senseless proteins. Ectopic Notch signaling by Delta expression ahead of the morphogenetic furrow also causes premature differentiation. It is concluded that proneural enhancement and lateral inhibition use similar ligand binding and receptor processing but differ in the nuclear role of Su(H). Prior to Notch signaling, Su(H) represses neural development directly, not indirectly through E(spl)-C. During proneural enhancement, the Notch intracellular domain overcomes the repression of neural differentiation. Later, lateral inhibition restores the repression of neural development by a different mechanism, requiring E(spl)-C transcription. Thus, Notch restricts neurogenesis temporally to a narrow time interval between two modes of repression (Li, 2001).

In the developing eye, lateral inhibition restricts the proneural gene atonal (ato) to individual R8 photoreceptor cells, which found each ommatidium. Earlier, ato must first have reached levels of activity sufficient to sustain expression by autoregulation, in conjunction with its bHLH heterodimer partner encoded by daughterless (da) and with a zinc-finger protein encoded by senseless (sens). Such 'proneural enhancement' depends on N and Dl but not on Su(H) or E(spl)-C. Clones of cells mutant for the E(spl)-C or for Su(H) lead to neural hyperplasia because they lack lateral inhibition, but clones of cells mutant for N or Dl show reduced neural differentiation because they lack proneural enhancement. These divergent phenotypes show that proneural enhancement occurs by a mechanism distinct from that of lateral inhibition (Li, 2001).

Mosaic analysis with Notch pathway mutations have been used to elucidate the mechanism of proneural enhancement. Requirements similar to those of canonical N signaling for processed forms of Dl, Notch EGF repeats 10-12, and proteolytic processing of the N intracellular domain have been found. Proneural enhancement is independent of any Su(H)-mediated gene activation but is mimicked by the complete absence of Su(H) protein, and this indicates that proneural enhancement depends on the disruption of Su(H)-mediated gene repression (Li, 2001).

The phenotypes of other mutations can be compared to the E(spl) or N phenotypes. A neurogenic mutant phenotype indicates a role in lateral inhibition, not in proneural enhancement. A hyponeural phenotype indicates a requirement in proneural enhancement (Li, 2001).

The neurogenic phenotype of the metalloprotease kuz suggests that processed Dl might be important for lateral inhibition and that unprocessed, transmembrane Dl may not be sufficient. It is unknown what form of Dl is required for proneural signaling. Clones mutant for kuz show neural hyperplasia. The distribution of R8 cells labeled by Boss antibody is intermediate between the distributions of clones null for E(spl) and for N. This indicates either partial loss of lateral inhibition or a weak proneural phenotype that still permits some neurogenesis to occur. Ato expression was examined to distinguish these possibilities. In kuz clones, Ato protein appears at the same time as it does in neighboring wild-type regions, but it remains at a low level. Posterior to the furrow, small clusters of R8 cells express Ato at a higher level, but many fewer cells do so than in E(spl) clones. This shows that proneural enhancement is affected in kuz mutant clones, but to a lesser degree than in N null clones, so that more cells go on to take the R8 cell fate. An intermediate phenotype associated with small clusters of R8 cells results in combination with the kuz lateral-inhibition defect. This is consistent with a role for processed Dl in proneural enhancement as well as in lateral inhibition, although it is important to note that kuz might have roles besides Dl processing. Such roles might include other aspects of N function (Li, 2001).

EGF repeats 10-12 bind Dl and are important during lateral inhibition because a glutamic acid-to-valine substitution in EGF repeat 12 in the N^M1 mutant is embryonic lethal and neurogenic. Clones of N^M1 mutant cells in the eye affect proneural enhancement and lateral inhibition, as does kuz, and this finding indicates that Dl interacts with the EGF repeat 12 region of N for proneural enhancement as well as for lateral inhibition (Li, 2001).

Clones mutant for the psn mutation were examined to test whether the novel proneural pathway requires proteolytic processing of N. Clones of psn exhibit an intermediate phenotype. Small patches of R8 cells differentiate, as in N^M1 or kuz clones but unlike in E(spl) clones. Ato expression initiates normally but never elevates to the same levels seen in the wild type. Lateral inhibition is deficient in psn clones as judged by the loss of E(spl) expression [E(spl) mDelta], so the intermediate psn phenotype indicates an effect on proneural enhancement in addition (Li, 2001).

In lateral inhibitory signaling, the processed intracellular domain enters the nucleus. Clones mutant for the N^CO mutation were examined to test whether proneural enhancement is also mediated by the released intracellular domain or, alternatively, by other parts of the processed protein. In place of Gln-1865, N^CO encodes a termination codon that truncates the N intracellular domain close to the transmembrane domain. Eye clones of N^CO almost completely lack R8 cells or other neurons. Ato expression is greatly reduced, and only rare R8 cells form posterior to the furrow. Expression of the Senseless protein, a marker for Ato activity, is also greatly reduced, and this finding confirms the failure to establish high levels of Ato expression and function. These results show that the N intracellular domain is required for proneural enhancement. Similar results were obtained with N^60g11, which truncates the intracellular domain carboxy-terminal to the ankyrin/CDC repeats (Li, 2001).

It is noteworthy that the N^CO phenotype is 'stronger' than clones of the N protein null, in which occasional patches of neurogenesis are seen. If this is attributed to the dominant-negative effect of the protein encoded by N^CO, then residual neurogenesis in N null clones must reflect residual N protein, perhaps persisting from before the mitotic recombination event (Li, 2001).

The N intracellular domain converts nuclear Su(H) protein from a transcriptional repressor into a transcriptional activator during lateral inhibition. What is its role in proneural enhancement? It has been concluded that proneural enhancement does not require Su(H) based on the neurogenic phenotype of Su(H) mutant clones. However, the original Su(H) mutants seem not to have eliminated the Su(H) repressor function. Recently, deletion alleles of the Su(H) gene have been recovered that eliminate all Su(H) function (Li, 2001).

Clones homozygous for the Su(H)^Delta47 allele are neurogenic, as described previously for other alleles. In addition, however, Su(H)^Delta47 mutant cells differentiate prematurely. Ato expression begin earlier in Su(H)^Delta47 clones than in neighboring tissue, and it soon reaches high levels. The senseless gene is expressed in response to ato activity. Senseless is also expressed prematurely in Su(H)^Delta47 clones. Daughterless protein is ubiquitous but upregulated in ato-expressing cells of the furrow. It was hard to see premature elevation of Daughterless in Su(H)^Delta47 clones, and this must be subtle if it occurs (Li, 2001).

Premature differentiation in Su(H)^Delta47 clones might be explained if Su(H) normally antagonizes proneural enhancement. Then, in the total absence of Su(H) protein, Ato would enhance prematurely and initiate eye differentiation. Accelerated differentiation would in turn accelerate the progress of the morphogenetic furrow, induce Atonal expression more anteriorly, and begin the cycle again. To investigate the effect of N signaling on this Su(H) function, Dl was misexpressed ahead of the morphogenetic furrow. A transposon insertion in the hairy gene provided GAL4 protein expression. Ato expression is expanded anteriorly throughout the domain of h expression in h^GAL4; UAS-Dl eye discs. The sca gene, which is expressed in response to ato activity, is also expressed more anteriorly in response to ectopic Dl. Neural differentiation begins normally in the most posterior part of h^GAL4;UAS-Dl eye discs but becomes progressively disorganized more anteriorly as differentiation accelerates (Li, 2001).

The similiarity between activating N signaling ahead of the morphogenetic furrow and deleting Su(H) indicates that N signaling overcomes repression mediated by Su(H). If Su(H) antagonizes proneural enhancement by activating gene transcription, activating N ahead of the furrow should have released the N intracellular domain, elevated gene transcription, and antagonized morphogenetic furrow progression and differentiation, opposite that of what was observed (Li, 2001).

Different forms or complexes of N intracellular domain might be required to antagonize Su(H)-mediated repression during proneural enhancement from those that coactivate Su(H)-mediated gene transcription. The possible role of bib, mam, and neur in proneural enhancement has not been assessed. The bib gene encodes a transmembrane protein required for lateral inhibition in embryonic neurogenesis. Ommatidia that are mutant for bib contain occasional extra photoreceptor cells, and some ommatidia have multiple R8 cells. Ato expression begins and progresses normally, but posterior to the morphogenetic furrow small clusters of two or three cells, instead of single cells as in the wild type, often retain Ato expression. Sections through the adult retinas often reveal ommatidia with extra photoreceptor cell rhabdomeres, both of the R8/R7 small rhabdomere class and of the larger R1-R6 outer photoreceptor class. Since bib affects lateral inhibition only slightly, it is possible that an equally subtle requirement for bib in proneural enhancement might be undetected in these experiments (Li, 2001).

These findings suggest a model for proneural enhancement. The release of N intracellular domain in response to Dl derepresses genes that are repressed by Su(H). The relevant targets do not require Su(H)-mediated transcriptional activation, so deletion of Su(H) mimics N signaling. The mechanism contrasts with lateral inhibition. N signaling provides N intracellular domain as a coactivator for Su(H), which is essential for the transcription of E(spl)-C. Lateral inhibition cannot proceed in the absence of Su(H) because blocking repression by Su(H) is not sufficient for E(spl)-C transcription (Li, 2001).

The ato gene could be a direct target of proneural enhancement. ato regulatory sequences have been examined for activity control regions, but possible repression sites have not been assessed. Another candidate is daughterless, which encodes a bHLH heterodimer partner of Ato that is required for Ato function in eye development. A third candidate is senseless, a zinc finger protein that enhances and maintains proneural gene expression. Expression of ato and sens is prematurely elevated in the absence of Su(H), which is consistent with regulation by Su(H)-R. However, each might depend on Su(H)-R only indirectly because elevated expression of ato or sens requires the function of all three genes (Li, 2001).

Why does proneural enhancement precede lateral inhibition if both depend on Su(H) and nuclear N intracellular domain? (1) Multiple lines of evidence indicate that proneural enhancement requires less N activation than does lateral inhibition. These include the greater sensitivity of lateral inhibition to the N^ts mutation, nonautonomous rescue of proneural enhancement by Dl over distances for which lateral inhibition go unrescued, and neurogenesis in N mutant clones due to undetectably low levels of N protein (which is eliminated by dominant-negative protein from the N^CO allele). Therefore proneural enhancement is expected to occur sooner in response to N signaling. (2) The evolving transcriptional regulation of ato changes sensitivity to lateral inhibition over time. Even recombinant N intracellular domain expression does not prevent initial ato expression ahead of the furrow, but ato is exquisitely sensitive later when its expression depends on autoregulation (Li, 2001).

The main result of this study is that neural development in the Drosophila eye depends on two functions of the N intracellular domain in response to ligand binding: (1) N relieves Su(H)-mediated repression to enhance ato expression and function and to permit neurogenesis (proneural enhancement); (2) later, another pathway requires N to coactivate Su(H)-dependent E(spl) transcription (lateral inhibition). No genes or regions of N have yet been found to be required to affect one function but not the other. By means of these two functions stimulated by the same ligand, N signaling coordinates the upregulation of ato in proneural cells and represses ato in cells not specified as neural precursor cells, and N restricts neural patterning to a narrow time interval between two distinct modes of repression (Li, 2001).

GENE STRUCTURE

Bases in 3' UTR - four polyadenylation signals 711bp

PROTEIN STRUCTURE

Structural Domains

Suppressor of Hairless contains an integrase domain that is 82% homologous to mouse J kappa-recombination-signal-binding-protein (JkRBP). A 40 amino acid integrase domain, in the center of a region of high interspecies homology is conserved by a family of site-specific recombinases. The integrase function is responsible for VDJ recombination in immunoglobulin maturation (Furukawa, 1991).

date revised: 20 March 2001 

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

Please e-mail comments/corrections to brodyt@codon.nih.gov