fray expression was examined using both the enhancer trap reporter and tissue in situ hybridization. The overall expression patterns observed using these assays were similar. Expression of fray is first apparent in preblastoderm embryos, consistent with a maternal contribution. Zygotic expression is first observed in stage 12 embryos, in anterior and posterior regions that will develop into the gut. By stage 15, the proventriculus and hindgut show very high expression, while cells of the midgut show moderate expression. In the abdominal bodywall, only muscle fiber 17 is stained. Just outside the CNS, the dorsal median cells in the thoracic segments show expression. These cells, mesodermal in origin, located on the dorsal midline of the CNS, have been shown to pioneer the transverse nerve, and are thought to function like glia. By stage 17, all the thoracic and abdominal dorsal median cells are stained, as are muscle fiber 17 and the gut (Leiserson, 2000).
The expression most relevant to the nerve phenotype is found in a set of cells associated with the CNS nerve roots. Based on their location and by their coexpression of Repo, a glial-specific transcription factor, it is concluded that these fray-positive cells are the exit glia, the cells that give rise to the peripheral ensheathing glia. Later in development, expression is observed in a small subset of nerve-associated cells (Leiserson, 2000).
The restriction of the nerve phenotype and fray expression to ensheathing glia suggests that Fray plays an essential role in these cells for normal axonal ensheathment. To investigate this, rescue experiments were undertaken using GAL4 enhancer trap lines to restore Fray expression in different subsets of cells in a mutant background. Ubiquitous expression of the fray cDNA is able to rescue the nerve bulge phenotype and the larval lethal phenotype. Three drivers with expression in peripheral glial cells are all able to rescue the mutant nerve phenotype, providing strong evidence that Fray is required in peripheral glial cells for normal nerve ensheathment (Leiserson, 2000).
The rescue experiments also provided evidence for other developmental roles of Fray, besides glial ensheathment of axons. Driving expression with a weak, constitutive driver produces adults that have visible defects, such as reduced wings, eyes, and legs and abnormally patterned bristles. In contrast, a strong constitutive driver fully rescues the mutant phenotype. Driving fray expression with either driver in a wild-type background produces normal-looking adults, suggesting that the visible defects are due to loss of function of fray and not the result of the overexpression of the rescue construct (Leiserson, 2000).
The fray gene was discovered in an enhancer trap screen in which line PZ4624 expressed the reporter lacZ gene in the embryonic neuromuscular system as well as in the midline dorsal median cells. To clone the gene, a genomic clone was isolated from the region by plasmid rescue, which was used to initiate a chromosomal walk through the region. Screening by tissue in situ hybridization, a genomic fragment was identified that reveals a pattern of mRNA expression that mirrors the enhancer trap reporter gene expression (Leiserson, 2000).
Three mutant alleles of the gene (denoted r1, r2, and r3) were isolated by excising the P element from line PZ4624. All three mutations result in lethality by the late third instar larval stage, fail to complement one another, and exhibit similar nerve phenotypes. Subsequently, another allele was obtained from the same enhancer trap screen, PZ7551, which is an insertion into the 5' region of the transcription unit and results in larval lethality during the first and second instars (Leiserson, 2000).
To learn whether fray plays an important role in the nervous system, fillet preparations of third instar mutants, labeled with the neuronal marker anti-HRP, were examined. The only abnormalities observed by light microscopy were localized bulges or swellings in the nerves, inside of which extensive axon defasciculation was observed. This is referred to as the 'frayed nerve' phenotype. Detailed ultrastructural analysis has revealed that the fray phenotype also involves other, nonbulging regions of the nerve (Leiserson, 2000).
The fray mutation is 100% penetrant; bulges and associated axon defasciculation were found in every larva examined. By light microscopy, the defects appear localized, ranging from mild bulging to immense bulges that can exceed 300 µm in length and 50 µm in diameter. The number of these bulges ranges from 8 to 22 per animal. The neuron-specific anti-HRP labeling within the large bulges reveals defasciculated axons. In contrast, in wild-type larvae, nerves are of uniform width (7 µm) with straight and tightly bundled axons, as revealed by anti-HRP labeling and electron microscopy (Leiserson, 2000).
Directly imaging live larvae through the cuticle with enhanced video microscopy reveals frayed nerves in recently hatched first instar larvae, indicating that the bulges arise early in development. Using this same technique, the same phenotype was observed in mutants bearing the severe allele PZ7551; such mutants die before the end of the second instar. The similarity in phenotype suggested that the frayed nerve phenotype in hypomorphs is not a secondary effect but is closely linked to the loss of fray function (Leiserson, 2000).
By labeling the nuclei in third instar fray mutant peripheral nerves, it was possible to examine whether the nerve bulges result from a loss of the ensheathing perineurial cells, leading to the weakening and hence bulging of the nerve. Since the perineurial cells far outnumber the six to eight ensheathing peripheral glia in each nerve, the vast majority of the nuclei labeled in these preparations are perineurial. The perineurial layer was found to be continuous; perineurial nuclei are associated with the bulges, suggesting that the bulges are not due to openings in the ensheathment or to localized cell loss. This result is in agreement with transmission electron microscopy (TEM) analysis that shows that the perineurial layer is continuous, even in the largest bulges (Leiserson, 2000).
No strict relation between the presence of a bulge and its location along the nerve was found. Bulges were found in every nerve and in nearly every segment. There was, however, a higher tendency for bulges to occur near the ventral ganglion. For the A8 nerve, a second hot spot was found at the levels of segments A6 and A7 (Leiserson, 2000).
The extensive homology among members of the PF kinase family suggests that they are functional homologs. To test this, a rat PASK cDNA was used to see whether it could substitute for fray function in fray mutants. The rat homolog was chosen over other mammalian homologs because it had already been characterized and had been demonstrated to possess kinase activity in vitro. For these experiments, the two strongest GAL4 drivers were selected; they exhibit essentially complete rescue of the fray lethal and nerve phenotypes when driving expression of the Drosophila fray cDNA. Expressing PASK using either of these drivers reduces the severity of the fray lethal phenotype, as evidenced by a shift of the lethal phase from larva to pupa. The degree of rescue obtained with PASK is less than that obtained using the Drosophila cDNA. Remarkably, these results show that, despite being separated by hundreds of millions of years of evolution, rat PASK is capable of substituting for fray function (Leiserson, 2000).
Gli is a receptor with an inactive serine esterase domain, whose molecular function is unknown but has been hypothesized to participate in signal transduction. Because a kinase such as Fray almost certainly has a signal transduction function, it would be of interest to see whether Fray and Gli operate in the same signaling cascade. Toward this end, evidence of genetic interactions between the two genes was sought by comparing the phenotype of a hypomorphic mutant with or without a severe allele of the other gene. No evidence of genetic interaction has yet been found. For example, the mutant phenotypes of frayr1/frayr1;gliAEDelta45/+ and frayr1/frayr1;+/+ animals appear indistinguishable -- both survive to the third instar and have the frayed nerve phenotype. If fray and gli operate in separate pathways, one would not expect them to interact genetically. The lack of a genetic interaction, however, could be due to other factors and by itself does not rule out the possibility that they function in the same pathway (Leiserson, 2000).
Leiserson, W. M., Harkins, E. W. and Keshishian, H. (2000). Fray, a Drosophila serine/threonine kinase homologous to mammalian PASK, is required for axonal ensheathment. Neuron 28: 793-806. 11163267
Petitt, M., Melnick, M. and Perrimon, N. (1998). The Drosophila gene fray encodes a STE20-related kinase that controls multiple microtubule-based movements. Molec. Biol. Cell 9: 43a
Tsutsumi, T., Ushiro, H., Kosaka, T., Kayahara, T. and Nakano, K. (2000). Proline- and alanine-rich Ste20-related kinase associates with F-actin and translocates from the cytosol to cytoskeleton upon cellular stresses. J. Biol. Chem. 275: 9157-9162. 10734050
date revised: 30 June 2001
Home page: The Interactive Fly © 1997 Thomas B. Brody, Ph.D.
The Interactive Fly resides on the
Society for Developmental Biology's Web server.