BIOLOGICAL OVERVIEW

Through years of research, considerable insight has been garnered into the regulatory mechanisms of photoreceptor cell specification and the mechanisms of phototransduction. However, the events that occur between the specification of a photoreceptor neuron and the final organization and function of the phototransduction signaling cascade are poorly defined. PvuII-PstI homology 13 (Pph13), also known as Munster, is a homeodomain transcription factor expressed only in photoreceptor cells. Pph13 expression correlates with the differentiation and not specification of photoreceptor cells. In agreement with its expression profile, Pph13 is found to be required for both rhabdomere morphogenesis and for the proper detection of light. In addition, Pph13 exerts its effect by the regulation of photoreceptor specific gene expression (Zelhof, 2003).

Upon specification, photoreceptor neurons immediately send axonal projections into the optic lobe of the Drosophila brain. The outer photoreceptor cells project into the lamina whereas the inner photoreceptor cells (R7, R8) send axonal projections deeper into the optic lobe and terminate in the medulla. Furthermore, the differentiation of the photoreceptor neurons is not complete until 4 days later, at the end of metamorphosis. One unique feature of vertebrate and Drosophila photoreceptor neurons is the creation of a specialized light-sensing organelle on the apical cell surface. In Drosophila, the rhabdomere is the photoreceptor light-sensing organelle and is the functional equivalent of the outer segment of vertebrate rod and cone cells. Each rhabdomere consists of 60,000 tightly-packed microvilli, each 50 nm in diameter and 1-2 µm in length. This results in a tremendous increase in surface area to house the tens of millions of rhodopsin molecules and associated signaling molecules that are responsible for the detection of light (Zelhof, 2003).

A key aspect of eye structure is the organization of the phototransduction machinery into the rhabdomere. As the rhabdomere develops, the signaling molecules required for the detection and translation of light into a receptor potential are expressed and localized to the rhabdomere. In Drosophila, the activation of rhodopsin leads to the activation of Phospholipase C (PLC) via a coupled heterotrimeric G protein. PLC catalyzes the breakdown of phosphatidyl 4,5-bisphosphate [PtdIns(4,5)P₂] into the two intracellular messengers inositol triphosphate [Ins(1,4,5)P₃] and diacylglycerol (DAG). This reaction leads to the opening of light sensitive cation-selective channels (TRP, TRPL and TRPgamma) and the generation of a depolarizing receptor potential (Zelhof, 2003).

To identify genes required for photoreceptor cell differentiation, a screen was performed for the presence or absence of the deep pseudopupil (DPP) in Drosophila adult eyes. The presence of the DPP is an indication of the overall integrity of the photoreceptor cells and their associated rhabdomeres. Such screens have been effective in isolating mutations that affect eye structure and development. To limit the search for those mutants that affect aspects of differentiation and not specification, mutants were excluded that had incorrect external morphology, particularly rough or irregular shaped eyes. Consequently, 6,000 viable second chromosome EMS mutated lines, generated from 38,000 F3 lines were screened for the absence of a DPP and 33 mutant stocks were isolated that represent 18 complementation groups (Zelhof, 2003).

Since import of the signaling components into the developing rhabdomere occurs late in photoreceptor differentiation, it was reasoned that flies that lacked a DPP and could not correctly respond to light would be the best candidates for mutants defective in photoreceptor terminal differentiation. Using electroretinogram assays (ERGs) that measure the capacity of photoreceptor cells to convert light into a receptor potential, the collection of 33 mutants was screened for those that had defects in light perception. The results indicated that among the group that had irregular ERGs, one had a severe deficiency in the detection of light. This mutant responds reproducibly only to long durations of high intensity light. The characteristic on/off transients of wild-type responses are undetectable and by 10 days post eclosion, the mutants appear to lose all responses to light (Zelhof, 2003).

While weak or no response to light can be the result of numerous factors (e.g. missing phototransduction molecules or subsequent retinal degeneration), the absence of a DPP in newly eclosed flies suggests that the rhabdomeres did not form. To examine whether the rhabdomeres are present, an EM-ultrastructural analysis was performed on the adult mutant photoreceptor cells. This examination revealed that all photoreceptor cells and associated rhabdomeres were present. However, the rhabdomeres were severely malformed. They were significantly smaller in size and often the microvilli within each rhabdomere were misaligned. In addition, the rhabdomeres did not consistently extend the entire thickness of the retina. Given the malformed rhabdomeres and the inability of this mutant to detect light, this mutant was tentatively name hazy; subsequently hazy was found to represent a null allele of the photoreceptor specific homeodomain gene Pph13 (Zelhof, 2003).

Pph13 is expressed only in photoreceptor cells, the larval cells of Bolwig's organ and the adult R1-R8 photoreceptor cells, and no defect has been detected in specification of these cells. Pph13 does not have the ability to activate transcription outside the context of a photoreceptor cell nor does ectopic expression of Pph13 affect the fate of any of the photoreceptor or associated accessory cells. The only phenotypes observed are in photoreceptor cell morphogenesis and function, suggesting Pph13 function is restricted to photoreceptor cell terminal differentiation (Zelhof, 2003).

Pph13^hazy mutants have two striking defects: the ability of the photoreceptor cell to detect light and the biogenesis of the light sensing organelle, the rhabdomere. Are the two phenotypes connected? The possibility that the malformed rhabdomeres are contributing to the inability of these mutants to detect light or vice versa cannot be eliminated. However, the severity of the rhabdomere defect cannot be solely responsible. For example, mutants have been isolated that result in malformed rhabdomeres equal to those seen in Pph13^hazy but have a normal ERG. In addition, the loss of Chaoptin and NinaC both result in a considerable loss of rhabdomeric size and rhodopsin levels but they have a better response to light then Pph13^hazy mutants (Zelhof, 2003).

In addition, Pph13 is required for the transcription of phototransduction proteins. Clearly, trpl, trpgamma and Gß are not expressed in mutant photoreceptor cells, and the absence of Pph13 affects the full expression of several other signaling components. This is clearly observed with Arr1 expression. (1) The data demonstrate that in the inner (R7/R8) wild-type photoreceptor cells have a considerable higher expression of Arr1 when compared with the outer photoreceptor cells (R1-R6). (2) The loss of Pph13 does not eliminate expression of Arr1 in photoreceptor cells. Arr1 expression can be seen in mutant R7/R8 photoreceptor cells and the lack of signal in the outer photoreceptors is not due to the absence of Arr1 expression but rather the fact that these cells start out with lower levels of Arr1. Taken together, while all of the detected protein aberrations can explain the severe loss of light sensitivity, the results do not eliminate the possibility of a yet unidentified molecule required for proper phototransduction. Selective rescue and identification of any other missing components will be needed to explain the complete molecular mechanisms responsible for the decrease in light sensitivity (Zelhof, 2003).

The molecular mechanisms for rhabdomere biogenesis are for the most part unknown. Nonetheless, the data do provide a few insights into rhabdomere biogenesis. Pph13 is required for the generation or execution of a late acting signal necessary for the elaboration and growth of the microvilli into a rhabdomere. Immunofluorescent and EM analyses demonstrate that the defects observed in Pph13^hazy mutants are the result of a developmental flaw and not of retinal degeneration. The disorganized rhabdomeres do not show any of the characteristic signs of degeneration and more significantly a clear halt is detected in rhabdomere development by 72 hours APF. In addition, by all measurements, the early events (36 to 60 hours APF) of rhabdomere biogenesis occur normally (Zelhof, 2003).

The data also indicate the failure of growth is not due to the improper localization or delivery of proteins to the rhabdomere. For example, Chaoptin, which is required for the cross-linking of microvilli still localizes to the developing rhabdomere before and after the rhabdomere has stalled in development. In addition, the proteins composing the phototransduction machinery, especially rhodopsin, which has a role in phototransduction and in maintaining the structural integrity of the rhabdomere, are imported and stabilized within the malformed rhabdomeres. The characteristic expansion of the endoplasmic reticulum associated with defects in rhabdomeric protein cell trafficking was not detected (Zelhof, 2003).

What is responsible for the flaw in rhabdomere biogenesis? Most notably for a transcription factor believed to be necessary for the activation and not repression of gene transcription, a grossly abnormal accumulation of Rac1 is seen in Pph13^hazy mutant photoreceptor cells. However, the presence of Rac1 is in agreement with the fact that the terminal web does form in Pph13^hazy mutants. Given that the exact function of Rac1 has not been resolved in rhabdomere biogenesis and that small Rho GTPases have been implicated in mediating signals required for actin reorganization, future experiments will address the function of Rac1 in photoreceptor terminal differentiation and determine how the misregulation of Rac1 accumulation and activity may be contributing to the Pph13^hazy rhabdomere phenotypes (Zelhof, 2003).

The molecular cloning of Pph13^hazy has identified another homeodomain gene required for photoreceptor morphogenesis. Previous reports have established or implicated eyeless (Pax6), orthodenticle (otd) and Onecut homeodomain genes in eye development. What is the relationship between these various homeodomain transcription factors and how do they coordinate photoreceptor terminal differentiation? Numerous possibilities exist in which each of these transcription factors could control a unique subset of molecular mechanisms required for a functional photoreceptor cell; alternatively, they could act in concert on the same genes to promote differentiation. To eliminate or confirm any one of these possibilities would be premature and further extensive characterization of each of these genes in photoreceptor development is necessary (Zelhof, 2003).

Nevertheless, the preliminary data does allow for some speculation. It is clear that eyeless is required for photoreceptor cell specification and without it a photoreceptor cell or a gene like Pph13 could not function. Besides its early role in photoreceptor cell specification, eyeless is also necessary for rhodopsin expression and superficially, characterization of the late transcriptional targets of Eyeless and Pph13 appears to be different. Pph13 is absolutely required for trpl, trpgamma and Gß expression but is not necessary for rhodopsin expression. This result would suggest that once a cell has committed to a photoreceptor cell fate, both Eyeless and Pph13 have separate and distinct molecular pathways that contribute to photoreceptor differentiation (Zelhof, 2003).

However, comparison of otd and Pph13 mutants suggest a more complex mode of coordination for photoreceptor differentiation. The rhabdomere defects observed otd^uvi and Pph13^hazy mutants are similar. In each case, the defects appear not to be the result of degeneration but a failure in their biogenesis. The rhabdomere terminal web does form in both cases but the overall size and morphology are abnormal. Both Otd and Pph13 are required in the same developmental time window for rhabdomere morphogenesis, but neither is necessary for the expression of the other. Whether Otd and Pph13 represent two parallel pathways directing the expression of the same genes or two distinct pathways with different genetic targets to promote rhabdomere biogenesis will require further investigation. In addition, they do not share a defect in phototransduction. otd^uvi photoreceptor cells exhibit normal ERGs and no loss of phototransduction proteins downstream of rhodopsin, as seen in Pph13 mutants. Clearly, Pph13 is responsible for two aspects of photoreceptor cell differentiation: phototransduction and rhabdomere morphogenesis (Zelhof, 2003).

Given that the molecular mechanisms orchestrating the differentiation of photoreceptor cells remain largely undefined, the goal of the genetic approach was to isolate genes required for photoreceptor terminal differentiation. Work with Pph13^hazy has shed some light on the regulation of this process. However, additional studies that combine the accessibility and genetic amenability of Drosophila eye development, with whole genome expression profiling techniques in both wild-type and Pph13^hazy mutant photoreceptor cells, should identify additional transcriptional targets necessary for photoreceptor cells to achieve and maintain a functional state (Zelhof, 2003).

GENE STRUCTURE

The Munster (Mu) genomic clone was originally isolated during a low stringency screen for Drosophila homeobox containing genes (Dessain, 1993). The original EMBL3 phage (called phage 1) contains another partial homeobox sequence that belongs to goosecoid (gsc). Mapping of this region reveals that Mu and gsc homeobox are only separated by 6 kb of genomic sequence. Moreover, both gsc and Mu map on the left arm of the second chromosome at position 21C5-6, very close to another homeobox-containing gene, aristaless (al). Mapping performed with P1 clones suggests that gsc and Mu are about 150 kb away from al (Goriely, 1999).

cDNA clone length - 1878

Bases in 5' UTR - 505

Exons - 3

Bases in 3' UTR - 299

PROTEIN STRUCTURE

Amino Acids - 357

Structural Domains

Sequence comparison restricted to the homeobox reveals a close homology between the Mu, Gsc and Al homeodomains. All three belong to the Paired-class. However, although both Al and Mu bear a glutamine at position 50 of their homeobox (Q50), Gsc possesses a lysine (K50). Residue 50 has been shown to confer DNA binding specificity to the homeodomain. This suggests that al, Mu and gsc might be part of a cluster of related homeobox genes that could be involved in the formation of sense organs or appendages in the Drosophila embryo. However, sequences outside the homeobox bears little similarity suggesting an ancient separation of the genes, or alternatively, rapid divergence of the region outside the homeodomain. A short region that shows homology to the Octapeptide/ GEH domain (Goosecoid Engrailed homology domain) is found in the C-terminal part of the Mu protein. This region has been shown to be required for active transcriptional repression by homeodomain-containing proteins. Repo is another homeodomain-containing molecule whose homeobox is closely related to Mu. Interestingly, the third helix of the homeodomain of Repo, Mu and Gsc has been conserved during evolution. There is a perfect conservation at the DNA level between Mu and repo over 44 bp, and only one base pair difference between Mu/repo and gsc. This difference accounts for one residue change that is responsible for the Q50 (encoded by the CAG codon) to K50 (encoded by the AAG codon) transition between these molecules. Al homeodomain exhibits a more divergent DNA sequence over this stretch with 11 differences, despite the excellent conservation at the amino acid level. This suggests that Gsc, Mu and Repo could derive from a common ancestor. It should be noted, however, that repo maps to a different chromosomal location than that of gsc, Mu and al. The high degree of homology between Mu, gsc and repo DNA sequences is likely to explain the cross-hybridization observed when a probe containing the Mu homeobox sequence is used for in situ hybridization (Goriely, 1999).

date revised: 25 November 2003

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