Interactive Fly, Drosophila

The 5-HT2 mRNA stripes appear in phase with those of cells expressing the pair-rule gene fushi tarazu. The 5-HT2 pattern in ftz mutants displays important modification; a partial loss of stripe restriction, shift of the first stripe to the anterior of the cephalic furrow, and enlargement of the low-expressing middle region to the third stripe. Therefore, 5-HT2 transcription is not strictly dependent of ftz expression. The 5-HT2 pattern is uneffected for mutated loci known not to affect ftz, such as engrailed and odd-skipped, and the mesodermic exclusion persists in snail mutants. Therefore, the 5-HT2 gene is located in the vicinity of ftz within the hiearchy of segmentation genes (Colas, 1995).

Serotonin synthesis and germband extension

Given the gastrulation impairments observed in the 5-HT2 receptor mutant, one might predict that similar phenotypes would be observed in 5-HT synthesis mutants. Various Drosophila mutants in which a lack of 5-HT has been detected at the gastrulation stage have embryonic lethality and display abnormal gastrulation movements associated with specific cuticular defects. After the initial description of the 5-HT2 mRNA expression at the early gastrula stage (Colas, 1995), the genomic locus of the 5-HT2 gene was mapped to 82C-E, and two small overlapping deletions were generated. Df(3R)HTR6 and Df(3R)HTRI are homozygous lethal and only the latter deletes the receptor gene. In homozygous Df(3R)HTRI embryos [but not in Df(3R)HTR6], the pushing force generated by ectodermal cell intercalation is impaired or lacking. This is associated with a desynchronization of germband extension from mesoderm and endoderm invaginations. Extreme desynchronization leads to a complete extension arrest. Although the mesoderm invagination is apparently normal in homozygous deficient embryos, defects in the ventral midline closure are frequently observed (Colas, 1999a). Although the phenotype displayed by the dead embryos is variable, a common characteristic is the original differentiation of their cuticle. In addition to 24% of homozygous balancer embryos, which die as weak first instar larvae, 20% of the individually polymerase chain reaction (PCR)-genotyped homozygous Df(3R)HTRI embryos secrete a cuticle with very few structures suggesting that it results from embryos arrested at an early stage (ghost). In the other homozygous Df(3R)HTRI dead embryos, all segments are present and appear identical in contrast to pair-rule mutants. Such a segment, termed 'double-line', consists of only two rows of thick denticles interspersed by clear spaces. Homozygous Df(3R)HTRI embryos and larvae show, within a range of increasing severity of embryonic defects, the double line phenotype either in normal sized crawling trachealess larvae (26%); in small late-hatching and non-crawling first instar larvae (14%) (petite); in non-hatched embryos similar to petite (32%) or in embryos lacking complete cuticular head structures (8%) (punchy). Furthermore, Df(3R)HTR6, j7E8 or j3A4 transposon insertion homozygous embryos do not display any of these characteristic embryonic abnormalities and die later during larval stages.

Examination of morphologically staged, wild-type gastrulae has revealed that the peaks of 5-HT and of 5-HT2 receptor precisely coincide with stage 7 when the rapid phase of germband extension begins (Colas, 1999a). A search was carried out for documented mutations that could specifically affect this peak of 5-HT synthesis in the Drosophila gastrula. The first focus was placed on alleles of the Punch locus, which encodes the GTP-CH enzyme (Reynolds, 1987). This enzyme synthesises the pteridin cofactor required for the enzymatic activity of aromatic amino acid hydroxylases, including tryptophan hydroxylase (TPH). The embryo-specific class of alleles of the Pu locus affects maternal and/or early zygotic GTP-CH activity. It has been suggested that this lethality is due to a deficit in early pteridin function. Using a capillary electrophoresis technique to evaluate 5-HT content in single class V Punch rWE67 embryos, three populations (in stage-7 embryos) can be distinguished: 25.5% with no detectable 5-HT (less than 5 attomoles); 25.5% with an average of 43,80 attomoles and 49.0% with an average of 22 attomoles, while the control embryos contain 44.5 attomoles. Consistently, three different embryo populations are found by PCR-genotyping after 48 h in the rWE67 progeny: 47% give rise to normal larvae (heterozygous rWE67/CyO) and 24% die at the first-instar larval stage (homozygous balancer CyO/CyO). The cuticle of the third population (29%) (homozygous Punch rWE67/rWE67) is abnormal with a distribution of embryos with all segments having an identical pattern, the double line or ghost phenotype. The complete range of these cuticular phenotypes is present in homozygous 5-HT2 null [Df(3R)HTRI] embryos with the exception of cellularization defects that appear to be specific to maternal functions of biopterins (Chen, 1994). In n homozygous rWE67 embryos, the intercalation of ectoderm cells that drives germband extension fails to occur. There is an associated desynchronization of germband extension from mesoderm and endoderm invaginations (Colas, 1999a). The common cuticular defects observed in the 5HT2 null and in this 5-HT deficient allele of Punch thus constitute an additional phenotypic correlation between the absence of 5-HT2 and the lack of its endogenous ligand 5HT in early gastrulae (Colas, 1999b).

The Df(2R)F36 deficiency uncovers most of the Pu transcribed region (McLean, 1993). Nevertheless, early embryonic lethality is not observed in homozygous F36 embryos; instead they show a larval lethality with unpigmented cuticle but no double lines phenotype (Reynolds, 1987). When assessed for the presence of 5-HT, it appears that 100% of each single tested embryo from heterozygous F36 progeny contain approximately 34 0:37 attomoles of 5-HT. This again supports a link between rWE67 early embryonic lethality and the lack of 5-HT (Colas, 1999b).

Previous molecular characterization of the Punch locus has shown that Pu transcripts and protein are expressed during oogenesis and BH4 levels have been reported to be abnormal in oocytes of a viable class V allele (O'Donnell, 1993). It has been hypothesized that defects in maternal deposition of biopterins may be responsible for early embryonic lethality. To verify this hypothesis a test was carried out for the presence of BH4 in early embryos before zygotic transcription starts. Newly laid embryos from an F36 heterozygous cross contain an average of 25 fmol of BH4/mg of total protein, whereas newly laid embryos from a similar rWE67 cross contain no detectable BH4 and control embryos have a BH4 level averaging 61 fmol of BH4/mg of proteins. The prezygotic BH4 level diminishes with time, suggesting it corresponds to a pool of maternally deposited biopterin which is not stable and that zygotically synthesised GTP-CH is required to maintain this level during subsequent embryonic development. However, a maternally deposited pool of BH4 in F36 embryos appears sufficient to satisfy the early embryonic requirement for biopterins and allows them to reach the larval stage. This result strongly supports the notion that the impairment of 5HT synthesis caused by the absence of maternal BH4 is at the origin of gastrulation defects in rWE67 (Colas, 1999b).

In order to confirm that the absence of a maternal pool of BH4 is directly linked with the lack of 5-HT and consequent gastrulation defects, an investigatation was carried out to see if similar embryonic defects could be observed in other 5-HT synthesis mutants. Mutants in the 5-HT biosynthetic pathway, Df(2R)PblX1 lacking tryptophan hydroxylase (TPH), and Df(2R) TW130 lacking dopa decarboxylase (DDC) were selected. Deficiencies that remove the entire locus were selected in order to eliminate the possibility of the expression of compensatory transcripts. Initially, 5HT levels were tested in single embryos from crosses of heterozygous parents. In crosses from both of these deficiencies, one quarter of all stage-7 embryos lack 5-HT and the rest show 5-HT levels not significantly different from wild type. This confirms that zygotically active enzymes are responsible for the peak of 5-HT synthesis observed at gastrulation. The cuticle of the homozygous dead embryos again showed a distribution of double-line and ghost embryos in Df(2R) PblX1 and in Df(2R) TW130. These data support the notion that double-line cuticular segments are the manifestation of a desynchronized extension and suggest that 5-HT signaling is a mechanism regulating cell intercalation. It ensures that the movements of ectoderm are strictly in phase with those of the other germ layers, a necessity for the reproducibility of development (Colas, 1999b).

The Interactive Fly resides on the
Society for Developmental Biology's Web server.