BIOLOGICAL OVERVIEW

The Drosophila grain gene, coding for one of Drosophila's three GATA transcription factors, is required during development for shaping the adult legs and the larval posterior spiracles. Mutant legs are short and wide rather than long and thin, while the spiracles are flat instead of dome-shaped. Analysis of loss-of-function mutations at the cellular level indicates that grain affects organ shape by locally controlling cell rearrangement. Ectopic grain expression causes major morphogenetic movements, resulting in the invagination of the posterior segments into the embryo. This is the first gene that has been shown to affect epithelial morphogenesis by controlling cell rearrangements. Thus, grn belongs to a new category of genes that controls the morphogenesis of organs without affecting their early specification during development (Brown, 2000).

The posterior spiracles are composed of two structures: the filzkörper, an internal tube that forms a refringent filter, and the stigmatophore, an external protruding structure that develops by cell re-arrangement (Hu, 1999). The development of the posterior spiracles of Drosophila may serve as a model to link patterning genes and morphogenesis. A genetic cascade of transcription factors downstream of the Hox gene Abdominal-B subdivides the primordia of the posterior spiracles into two cell populations that develop using two different morphogenetic mechanisms. The inner cells that give rise to the spiracular chamber invaginate by elongating into 'bottle-shaped' cells. The surrounding cells give rise to a protruding stigmatophore by changing their relative positions in a process similar to convergent extension. In the larvae the spiracular chamber forms a very refractile filter, the filzkörper. The opening of the spiracular chamber, the stigma, is surrounded by four sensory organs; the spiracular hairs. Clones labeling the spiracular hairs show that each one is formed by four cells related by lineage, two neural and two support cells, the typical structure of a type I external sensory organ. When the larva is buried in the semi-liquid medium on which it feeds, the stigmatophore periscopes out of the medium allowing the larva to continue breathing. The genetic cascades regulating spiracular chamber, stigmatophore, and trachea morphogenesis are different but they coordinate to form a functional tracheal system (Hu, 1999).

The cells of the stigmatophore express the Spalt (Sal) zinc finger transcription factor before their morphogenesis starts. Using this marker in fixed embryos it has been shown that from stage 12 to stage 15, the location of the sal expressing cells changes in a manner consistent with the cells rearranging their position with their neighbours. This movement is a major morphogenetic mechanism for spiracle formation, since it both contributes to close the spiracle and to elongate the stigmatophore (Hu, 1999). The process is similar to the change of shape produced by folding a pearl necklace twice: the internal area decreases while the necklace width increases from one to four tiers (Brown, 2000).

The use of living embryos allowed for the examination of the speed of the cell rearrangements. The spiracle cells were marked using a GFP.GAP fusion protein that localizes to the cell membrane. The fusion protein was expressed in the stigmatophore using the 459.2GAL4 line, an enhancer trap line where GAL4 has probably inserted in the sal gene. The overall shape of the stigmatophore cells does not change during the period of rearrangement. Using time-lapse recording in a confocal microscope the position of individual cells were studied from stage 14. The study was complicated by the various morphogenetic movements occurring at this time (final stages of germ band retraction, dorsal closure and spiracle formation movements) that take the cells in and out of the focal plane. In four embryos, groups of cells could be followed for at least 30 minutes. The cell rearrangement movements do not occur at a uniform speed. In all cases, some cells keep their relative positions during the period analysed, while others rearranged. These results show that the spiracle cell rearrangements occur discontinuously (Brown, 2000).

To find out the identity of the genes affecting cell rearrangement, mutants affecting the formation of the stigmatophore were studied. grn has been identified in a mutagenesis designed to find zygotic lethal mutations in Drosophila (Jürgens, 1984). Two embryonic lethal alleles with similar phenotypes were isolated, grn^7L12 and grn^7J86. To determine the primary cause of the defects in grn mutants it was first determined whether this gene is required for the specification or the morphogenesis of the posterior spiracles. sal and cut (ct) are the first genes required for posterior spiracle patterning. In grn mutants, the activation of both genes was not affected, showing that the patterning of the spiracles was normal. Analysing the position of sal-expressing cells at different stages is a good way to follow the cell rearrangements in the stigmatophore (Hu, 1999). In the wild type, the number of cells directly abutting the invaginating spiracular chamber is approximately 25 at stage 12. After 5-7 hours of development (depending on the temperature) the number of stigmatophore cells abutting the spiracular chamber decreases to eight and the number of tiers increases because the cells have shifted position. In grn mutants, the spiracle does not close and similar cell counts show that neither the number of cells abutting the spiracular chamber nor the number of cell tiers change during development. This result suggests that grn is required for cell rearrangements during morphogenesis (Brown, 2000).

Small amounts of cell rearrangement may have strong effects on organ shape. In the posterior spiracle of Drosophila, when the cells that form the tubule connecting to the trachea invaginate, the epithelial integrity remains intact by rearrangement of the surrounding cells. This rearrangement process in the stigmatophore results in a decrease from 25 to eight in the number of cells abutting the invaginating tissue. This decrease can be obtained if cells rearrange twice between stage 12 and stage 15, a period that lasts 5 hours at 25C° (or 7.5 hours at 20C°, the temperature at which the in vivo analysis was carried out). In vivo observations show that cell rearrangement in the spiracle does not occur synchronously. Cells alternate static periods with others of rearrangement that can last about 15 minutes. The spatial distribution and position of cells in the stigmatophore of grn mutant embryos remained almost unchanged from stage 12 to stage 16. This observation suggests that grn may be required for the cell rearrangements (Brown, 2000).

Four lines of evidence suggest that the leg, which requires grn, also develops using cell rearrangement. (1) Mosaic analysis shows that clones generated in the leg imaginal disc always occupy a long and thin stripe that runs from proximal to distal in the adult leg. These clones can be 100 cells in length, without exceeding three cells in width. This shape is markedly different from the shape of clones in the imaginal disc. (2) In many cases leg clones are split by neighboring cells that have intercalated breaking the continuity of the clone. (3) It has been shown that bristles adjacent in the adult leg originate from cells that are not adjacent in the leg primordium. (4) Direct scanning electron microscope (SEM) observations at the stage when the leg discs evert, suggest that there is cell intercalation. The similarity of the effects controlled by grn in leg and spiracle suggest that both phenotypes are due to the control of cell rearrangement. The short and wide shape of legs formed by grn mutant cells can be accounted for by an alteration in cell rearrangement in the primordium, resulting in a leg with altered shape. This hypothesis is supported by the shorter and broader appearance of grn clones compared with wild-type clones generated at the same stage. This effect is most evident when the clones are generated simultaneously in twin analysis (Brown, 2000).

The different shape of these clones is not due to cell death, as shown by the fact that grn clones can form most of the leg. The involvement of grn in the control of cell shape can also be discarded as a cause for the different shape of grn clones by observing the distribution of the cells in mutant clones. In Drosophila, each cell in the leg forms a single trichome; therefore, an altered distribution of the trichomes in the leg would indicate a change in the apical shape of the cells forming them. Close observation of the cuticle in legs with grn clones shows that trichome density in all axes is unaffected, indicating that the cell shape has not changed. Most of the leg cell divisions have ended by third instar larva, and grn is required after this stage, discarding a function of grn in controlling spindle orientation. Similarly, grn is required in the spiracle after the divisions have finished. These results suggest the conclusion that grn has a role in the control of cell rearrangement during morphogenesis and that abnormal convergent extension in the epidermis is the cause for the defects observed in the shape of the leg and the posterior spiracle (Brown, 2000).

During normal gastrulation the posterior 10% of the blastoderm invaginates, forming the primordium of the posterior midgut and the hindgut. The area immediately anterior to that (from 10% to 20% of the blastoderm length as measured from the posterior end) does not invaginate, forming the external cuticle of A8, A9, A10 and the unsegmented telson. Ectopic grn expression results in the invagination of this area, doubling the amount of blastoderm tissue that invaginates at the posterior end of the embryo. This major morphogenetic movement occurs without a change in the specification of the segments, which form the normal cuticle structures but in internal positions. This observation and the normal patterning of legs with mutant clones indicates that the main function of grn is to control the morphogenetic movements of the cells rather than to specify the structure they will form. This contrasts with the function of serpent and the Hox genes, which affect morphogenesis indirectly by changing the fate of the cells. Thus, grn belongs to a new category of genes that control the morphogenesis of organs without affecting their early specification during development (Brown, 2000).

GENE STRUCTURE

cDNA clone length - 3112

Bases in 5' UTR - 223

Bases in 3' UTR - 1428

PROTEIN STRUCTURE

Amino Acids - 486

Structural Domains

The predicted Grain polypeptide sequence contains two copies of the Cys-Cys type zinc finger that is unique to the GATA factor gene family. Comparison of all the available Drosophila GATA finger sequences reveals that 43 of the 48 (90%) amino acids are identical for the N-terminal fingers of the Pannier and Grain proteins. In contrast, a slightly longer sequence is conserved for the C-terminal Pannier and C-terminal Grain fingers or the single finger of Serpent. Of the 54 aligned amino acid residues of this region, 38 (70%) are identical, and the remainder show only conservative change. Interestingly, the four point mutations identified as Pannier null mutants are restricted to the conserved residues of the N-terminal finger (Lin, 1995).

date revised: 12 January 2022

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.