BIOLOGICAL OVERVIEW

The gene karst, which codes for the protein ß_Heavy-spectrin (ß_H) may be confused with, and must be distinguished from the ß-spectrin gene, which codes for the protein ß-spectrin. This overview will first consider the biochemistry and cell biology of karst and then discuss its effects on phenotype. ß spectrins are polymerization partners of alpha Spectrin. an elongated molecule that is a constituent of the submembrane cytoskeleton of epithelial cells. Epithelial cells constitute many tissues of the fly.

The combinatorial association of Drosophila spectrin monomers produces either alpha2/ß2 (conventional) or alpha2/ß_H2 (Karst-containing) tetramers (Dubreuil, 1990). In this notation the 2's refer to the fact that each tetramer contains two alpha and two beta subunits. Both tetramers crosslink actin and contain ß-chain pleckstein homology (PH) and alpha-chain SH3 domains, supporting the notion that spectrins interact with several cellular components other than actin. Karst, or ß_Heavy-spectrin, with 30 spectrin repeats and an Mr = 470 x10³, is longer and more massive than conventional ß-spectrins (Mr~250 x10³ and 17 repeats). ß_Heavy is also unique among ß-isoforms because it (1) contains an SH3 domain and (2) it probably does not bind to ankyrin (Thomas, 1997). Although ß_Heavy-spectrin was first recognized in the fly, it is of ancient origin (Thomas, 1997), has recently been cloned from Caenorhabditis elegans (McKeown, 1998) and may have a vertebrate homolog. In polarized epithelia, alpha-spectrin is generally distributed on plasma membranes; its polymerization partner ß-spectrin is restricted to the basolateral surface, and its second partner, ß_H is localized to the apical domain at the adherens junction, on the free apical surface and under brush borders. Recruitment of ß-spectrin to the membrane can occur after adhesion events in fly and vertebrate tissue culture cells. However, during early development in Drosophila, the ß_H membrane skeleton becomes polarized prior to formation of mature adherens junctions. This suggests the hypothesis that different ß isoforms may be recruited in different ways (Thomas, 1998 and references).

karst received its name because of the resemblance of the rough eye phenotype to certain karst landforms: domed hills, separated by flat valleys. karst mutant eyes appear roughest towards the posterior margin where ommatidia are sometimes absent. Although the phenotype of the most severe karst mutant eyes resembles that of the EGF receptor mutation DER Ellipse, no dominant interaction has been observed between these loci. karst mutant eyes also exhibit a variable frequency in the percentage of ommatidia that lack photoreceptor R7 (up to 61% exhibit this phenotype). The fact that R7 is the missing photoreceptor has been confirmed by examining serial sections through karst mutant eyes. Ommatidia with only six photoreceptors in distal sections (relative to the brain) have seven photoreceptors in proximal sections. Since R1-6 extend the entire depth of the retina, while R7 extends only through distal regions and R8 replaces R7 in more proximal regions, this result is consistent with the specific absence of R7. karst mutant eyes sometimes (at low frequency) contain ommatidia with very abnormal numbers of photoreceptors. The morphology of such clusters suggests that they may have arisen through fusion of ommatidia to another cluster containing only a small number of photoreceptors. Some photoreceptors have rhabdomeres that are greatly expanded in size. Serial sectioning reveals that these photoreceptors are broader throughout the depth of the retina and do not result from a failure of such cells to extend properly during pupation. The roughening of the eye is 100% penetrant but exhibits variable expressivity. Within ommatidia that do have eight photoreceptors, the arrangement of these cells is not as regular as in wild-type eyes. However, the presence of a recognizable trapezoid in many ommatidia indicates that this patterning event is intact in karst mutants and reveals an equatorial line, so cluster rotation still occurs (Thomas 1998).

To determine if karst belongs in the Sevenless signaling pathway, an epistasis experiment was performed in which the constitutively activated Sevenless receptor, Sev S11 was crossed into a karst mutant background. In a wild-type background, Sev S11 produces an excess of R7 cells because it is expressed in more than one cell per precluster. Siblings with three phenotypes (karst, S11 and S11/karst) were analyzed. If karst is epistatic to Sev S11 (i.e. blocking signaling when penetrant), the mean number of R7s per cluster in S11/karst double mutant flies would be expected to be lower than the mean in S11 flies. Unfortunately, the wider effects of the karst mutation on cell position and rhabdomere morphology prevent a reliable counting of the number of R7-type photoreceptors in S11/karst flies. Nevertheless, the mean number of photoreceptors per ommatidium for the S11/karst flies is not significantly different from the S11 flies. This is consistent with the interpretation that Sev S11 is epistatic to karst, suggesting that karst is upstream of Sevenless or in a parallel pathway contributing to the ligand interaction (Thomas 1998).

It is proposed that the loss of R7 photoreceptors seen in karst mutant eyes is indicative of a reduction in cell adhesion. The commitment of the R7 precursor (pre-R7) to a photoreceptor cell fate requires that the membrane-bound ligand encoded by the bride of sevenless gene (Boss) in R8 must directly contact the Sevenless receptor (SEV) on the surface of pre-R7, and this interaction must sustain a signal for some time to cause this cell fate decision. A subpopulation of Sevenless is seen to concentrate at the adherens junctions, specifically in regions in contact with R8 (Tomlinson, 1987); however, the bulk of both Boss and Sev are more apically restricted (Krämer, 1991; Tomlinson, 1987). If cell adhesion is decreased at the adherens junction, two hypotheses could account for the loss of R7 in karst mutant eyes. (1) A reduction in Boss-Sev encounters might cause signaling to drop below the threshold required to trigger R7 commitment. (2) A pre-R7 or R8 cell might move out of position, preventing contact and thus signaling. In either case, it seems reasonable to expect that the phenotype would be variable as is the case in karst mutant eyes (Thomas 1998).

Two further aspects of the phenotype are consistent with a reduction in cell-cell adhesion: the abberant arrangement of photoreceptors in karst mutant ommatidia and the leakage of hemolymph from abdominal spiracles. With regard to the first condition, enough properly rotated, trapezoidal ommatidia can be identified to indicate that this is probably not a patterning defect. Loosened cell adhesion could readily explain why the cells do not hold their correct relative positions. The second phenotype, a lack of tracheal integrity and hence, hemolymph leakage, might arise because the cells adhere less well to one another. Indeed, it has been observed that dissection to remove the gut of 3rd instar larvae is considerably easier in karst mutant individuals than their heterozygous siblings, because the tracheae are more easily disrupted. Alternatively, the appropriate adhesive contacts may develop incorrectly during pupation leaving gaps in the network (Thomas 1998).

The apical subcellular distribution of ß_H/Karst in both the wing and eye imaginal discs colocalizes with Shotgun (DE-cadherin) at the adherens junction, as it does in embryonic epithelia. This association appears intimate: in regions of the eye disc where Shotgun is more abundant, ß_H is also more prominent. The mutual exclusivity between ß_H and the conventional ß-spectrin isoform is the norm and may be important for cell polarization. While ß_H is localized to the adherens junction in imaginal disc epithelia, alpha-spectrin (presumably partnered by ß-spectrin) extends into the basolateral domain. This is consistent with the apical restriction of ß_H in other epithelia and the consistent restriction of ß-spectrin to the basolateral membrane (de Cuevas, 1996). To establish such a situation, different proteins must recruit the different spectrins to each domain. In the fly, conventional ß-spectrin is recruited to the membrane by ankyrin (Dubreuil, 1996), while ß_H does not colocalize with ankyrin (Lee, 1997) and has no conserved ankyrin-binding site (Thomas, 1997), suggesting that its interaction with the membrane is not ankyrin dependent. Binding to ankyrin could thus be used to specifically recruit ß-spectrin to the basolateral membrane. Presumably the reciprocal situation with some as yet unidentified receptor for ß_H results in its apical restriction. This mutually exclusive targeting makes it unlikely that the variable nature of the karst mutation is due to partial redundancy of the two ß-isoforms (Thomas 1998).

PROTEIN STRUCTURE

Amino Acids - 4097 (Thomas, 1997)

Structural Domains

The published sequence of ß_Heavy-spectrin (Dubreuil, 1990), now termed Karst, is partial and represents about 5 kb of a mature 13 kb mRNA. The conceptual translation of the open reading frame in this sequence predicts a peptide with an Mr = 189 x10³ which has been shown to be part of a larger protein with an Mr = 430 x10³. The sequence of this clone contains a consensus actin binding domain followed by twelve and a half copies of a typical 106 amino acid repetitive domain, and as such, is representative of an N-terminal fragment of a member of the spectrin/a-actinin/dystrophin gene superfamily of actin cross-linking proteins. The sequences of the C-terminal region of dystrophins are extraordinarily conserved and could, in principle, be used to determine the precise relationship of ß_H and dystrophin. Another partial cDNA has been recovered that represents the final, approx. 600 amino acids of the ß_H protein. There is no detectable similarity to dystrophin in this C-terminal clone, beyond that expected for a member of the a-actinin/spectrin/dystrophin superfamily. Thus Karst is classified as a spectrin and not a dystrophin (Thomas, 1994).

ß_Heavy-spectrin is distinguished from the other ß spectrin forms because it contains an SH3 domain; in addition, the PH domains in the C-terminus are unique to ß_Heavy-spectrin. Whereas ß-spectrin contains 16 spectrin repeats, ß_Heavy-spectrin contains 29 repeats. A major feature found in conventional ß-spectrins is the ankyrin binding site, absent from ß_Heavy-spectrin. The ß_Heavy-spectrin and ß-spectrin appear to be colinear over the initial 17 spectrin repeats, with an additional conserved 12 spectrin repeats present in the C-terminal half of ß_Heavy-spectrin. The unique (nonrepeat) features present in the ß_Heavy-spectrin sequence are as follows:

• segment 1 - conserved actin-binding domain
• segment 2 - probable dimer nucleation site resembling that of conventional ß-spectrin
• segment 7 - Src homology 3 (SH3) domain inserted between the b and c helices of segment 6
• segment 20 - an unusual extension of four amino acids between helices a and b of unknown significance
• segment 32 - spectrin repeat truncated after the b helix and probably forming the head-to-head interaction site with alpha-spectrin during tetramer formation
• segment 33 - nonrepetitive C-terminal segment containing a pleckstrin homology (PH) domain and a lysine-rich terminus characteristic of all nonerythroid ß-spectrins (Thomas, 1997).

Analysis of the amino acid sequences of alpha-actinin, spectrins, and dystrophin proteins has suggested that all three protein families arose from a single common ancestral protein that was alpha-actinin-like (Dubreuil, 1991). Specifically, alpha-actinin has an N-terminus resembling that of ß-spectrins and dystrophins, a short repeat alpha-helical motif common to the whole family, and Ca2+-binding EF-hands at the C-terminus related to those of alpha-spectrins and dystrophins. The spectrin repeats are reiterated a distinct number of times in each protein, resulting in a characteristic actin-crosslinking distance: alpha-Actinin has 4 repeats; ß-spectrin 17; alpha-spectrin 20; dystrophin 24, and ß_Heavy-spectrin 30 (Thomas, 1997)

date revised: 21 July 98

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

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