Genetic Control of Cuticle Formation During Embryonic Development of Drosophila melanogaster

Corresponding author: Amy Bejsovec, B336 LSRC, Box 91000, Research Dr., Durham, NC 27708-1000., bejsovec{at}duke.edu (E-mail)

Communicating editor: T. C. KAUFMAN

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The embryonic cuticle of Drosophila melanogaster is deposited by the epidermal epithelium during stage 16 of development. This tough, waterproof layer is essential for maintaining the structural integrity of the larval body. We have characterized mutations in a set of genes required for proper deposition and/or morphogenesis of the cuticle. Zygotic disruption of any one of these genes results in embryonic lethality. Mutant embryos are hyperactive within the eggshell, resulting in a high proportion reversed within the eggshell (the "retroactive" phenotype), and all show poor cuticle integrity when embryos are mechanically devitellinized. This last property results in embryonic cuticle preparations that appear grossly inflated compared to wild-type cuticles (the "blimp" phenotype). We find that one of these genes, krotzkopf verkehrt (kkv), encodes the Drosophila chitin synthase enzyme and that a closely linked gene, knickkopf (knk), encodes a novel protein that shows genetic interaction with the Drosophila E-cadherin, shotgun. We also demonstrate that two other known mutants, grainy head (grh) and retroactive (rtv), show the blimp phenotype when devitellinized, and we describe a new mutation, called zeppelin (zep), that shows the blimp phenotype but does not produce defects in the head cuticle as the other mutations do.

PATTERNING of the embryonic epidermis in Drosophila has been studied extensively by examining the cuticular pattern elements deposited by epidermal cells late in development. These structures serve as indelible markers of cell fates within the epidermis and have proven invaluable in genetic screens designed to identify mutations that disrupt cell fate decisions (NUSSLEIN-VOLHARD and WIESCHAUS 1980 ). However, little is known about the deposition of the cuticular layers during late stages of embryogenesis. Here we report the characterization of a set of Drosophila genes required for proper formation of the embryonic cuticle. We have isolated mutations in the Drosophila chitin synthase enzyme and demonstrate that these mutations are allelic with krotzkopf verkehrt (kkv), a gene identified in the Heidelberg screens for mutations affecting cuticle pattern (JURGENS et al. 1984 ).

kkv mutants show severe disruption of the head cuticle, which was the basis of their isolation in the earlier screen. The genetic screen in which our alleles were isolated involved devitellinizing embryos as well as dechorionating them, and this procedure reveals an additional phenotype associated with mutations in kkv. Embryos develop fully and move within the eggshell, but fail to hatch; when the vitelline membrane of these mutant eggs is broken by mechanical pressure, the embryos stretch to several times the size of a wild-type embryo, suggesting that the cuticle is much more elastic. We refer to this very distinctive embryonic phenotype as the "blimp" phenotype. Subsequently we found mutations in four other loci that yield similar blimp phenotypes. These are knickkopf (knk; JURGENS et al. 1984 ), grainy head (grh; NUSSLEIN-VOLHARD et al. 1984 ), and retroactive (rtv; WIESCHAUS et al. 1984 ) from the Heidelberg screens, and zeppelin (zep), a new mutation identified in our genetic screen.

We believe that further analysis of this collection of genes will provide insight into the process of cuticle formation and the orchestration of interactions between discrete epidermal cells during the formation of the continuous cuticular layers. Morphogenesis of the cuticle defines the structure of the larval body, forming its exoskeleton. This process is interesting not only from a developmental standpoint, but also from an economic one. Disruption of chitin synthesis provides a means of targeting insects without harming humans and other animals. For example, the active ingredient in a popular flea control medication for pets is lufenuron, a potent inhibitor of the chitin synthase enzyme (DRYDEN et al. 1999 ).

Insect cuticle is comprised of fibrils of chitin, a ß1-4 linked polymer of N-acetyl-D-glucosamine, embedded in a matrix of proteins and lipids (COHEN 1987 ). Synthesis of the chitin chains, catalyzed by chitin synthase, appears to occur intracellularly. Analysis of the sheep blowfly (Lucila caprina) chitin synthase suggests an integral membrane protein with the catalytic domain located on the cytosolic face of the plasma membrane (TELLAM et al. 2000 ). Thus the polymers must be extruded across the membrane, perhaps through a pore formed by the transmembrane domains of chitin synthase. The transport and deposition of chitin polymers may be the process disrupted by lufenuron and other pesticides in this class of benzoylphenyl ureas. These compounds are ineffective in blocking chitin synthesis in cell-free systems, suggesting that they do not directly inhibit the polymerization step of chitin synthase (COHEN 1985 ).

Lufenuron has been shown to disrupt molting of Drosophila larvae and to block development of eggs laid by mothers that were fed the pesticide (WILSON and CRYAN 1997 ). We find that when such eggs are dechorionated and mechanically devitellinized, they show a blimp phenotype similar to our mutant embryos. This observation, along with the discovery that kkv encodes a chitin synthase, supports the idea that the stretchy cuticle aspect of our blimp phenotypic class results from disruption of the chitin biosynthetic pathway. Further analysis of the set of genes defined by these mutations may give insight into the biochemical and genetic cascade required for proper chitin deposition and cuticle formation during Drosophila embryogenesis.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Drosophila stocks and culture:
Previously described alleles of kkv, knk, rtv, grh, and shg were obtained from the Bloomington (kkv¹, knk¹, rtv¹¹, grh^IM, and shg²) and Umea (kkv², kkv³, knk², and knk⁴) Stock Centers. All deficiency stocks used in this work were obtained from the Bloomington Stock Center. l(3)82Fh alleles were obtained from Adelaide Carpenter (University of Cambridge). P-element line EP(3)3092 was obtained from Exelexis, EP(3)0974 from Todd Laverty, and the other P-element lines described in this work from the Bloomington Stock Center. Data regarding the different alleles of the genes described in this article are summarized in Table 1.

Flies were reared on standard cornmeal-agar-molasses medium unless otherwise noted. For analysis of embryonic stages, eggs were collected on apple juice-agar plates. To examine embryonic cuticles, eggs were allowed to develop fully (24 hr at 25°), dechorionated in bleach, and then transferred to a microscope slide bearing a drop of Hoyer's medium mixed 1:1 with lactic acid (WIESCHAUS and NUSSLEIN-VOLHARD 1986 ). Mechanical devitellinization was performed by exerting gentle pressure on the coverslip to burst the vitelline membrane. Cuticle preparations were heated at 65° overnight before viewing.

Isolation of mutant alleles:
New alleles of kkv (kkv^DP14, kkv^DZ8, kkv^JH9, and kkv^LX5) and the defining allele of zeppelin (zep^LP13) were generated in the course of an EMS mutagenesis designed to recover suppressors of wingless (wg) mutant phenotypes. This was a standard F₃ lethal screen performed in a wg mutant background and involved examining cuticle preparations from individual isogenized lines. Thus recessive mutations either linked to wg on the second chromosome or segregating independently could be assessed for their ability to alter the wg mutant phenotype. The unhatched eggs from 5304 independently derived mutagenized lines were examined and 42 lines showing altered cuticle pattern were retained for further analysis. Several mutations suppressed the wg phenotype and their characterization revealed new Wingless pathway components (VAN DE WETERING et al. 1997 ; CAVALLO et al. 1998 ; MCCARTNEY et al. 1999 ; A. BEJSOVEC, unpublished data). Other unlinked mutations were found that do not modify the wg mutant phenotype but that affect the cuticle pattern. Among this class were the five mutations, representing two complementation groups, that generate a blimp phenotype when the embryos are mechanically devitellinized.

Mapping of mutations:
The blimp mutations, DZ8 and LP13, were mapped by meiotic recombination against rucuca, a multiply marked third chromosome. Both mutations were found to map between st and cu at an approximate genetic position of 47.0. DZ8 and the other mutations from the screen with which it had been found to be allelic, DP14, JH9, and LX5, disrupt the embryonic head cuticle, as do kkv and knk, two previously identified embryonic lethal mutations that map close to this position. Therefore, we obtained alleles of these genes from the Bloomington Stock Center to perform complementation tests and found that our blimp mutations fail to complement kkv¹, kkv², and kkv³. We also discovered that knk shows a blimp phenotype when it is mechanically devitellinized. This aspect of the kkv and knk phenotypes had been overlooked previously because prior work had not included mechanical devitellinization in the preparation of cuticles.

kkv and knk had been erroneously mapped to Df(3R)Tpl10 and Df(3R)by10, respectively (JURGENS et al. 1984 ). We determined that kkv maps instead to Df(3R)3-4, which removes 82F3-4;82F10-11, and also that it is allelic to l(3)82Fh (CARPENTER 1999 ). knk maps to Df(3R)by62, which removes 85D11-13;85F6. We mapped both loci and the zep^LP13 mutation more precisely by using P-element-mediated male site-specific recombination (CHEN et al. 1998 ). P elements used for mapping kkv were the following: P{lacW}l(3)L3051, P{PZ}l(3)02255, P{PZ}l(3)09904, EP(3)0974, P{PZ}l(3)03644, and P{PZ}Itp-r83A. P elements used for mapping knk were the following: P{PZ}l(3)10477, P{lacW}l(3)s2681, EP(3)3092, P{PZ}tws01436, and P{PZ} tws02414. P elements used for mapping zep were P{PZ}shk1, P{lacW}l(3)L7251, and P{lacW}l(3)j4D1.

kkv and knk mutant chromosomes were marked with st and ca, crossed into a 2-3 background to provide transposase, and placed in trans to the various P-bearing chromosomes. Males of the proper genotype were crossed back to a strain carrying an independently isolated allele of kkv or knk, respectively, on a chromosome also marked with st and ca. Thus only one class of recombinant, either st or ca, will be recovered in a given cross since kkv and knk homozygous progeny die as embryos. If st recombinants are viable, then the mutation must lie distal to the P element, and if ca recombinants are viable, the mutation is proximal to the P element. We scored 2000 flies from each cross. For each P element tested, we observed at least five recombinants of one class with one or no recombination events of the other class. Rearrangements can occur during the recombination event (CHEN et al. 1998 ), with duplications of the wild-type counterpart of the mutant locus producing a recombinant that would otherwise be inviable. Thus, some crosses produced both st and ca recombinant offspring. However, such rearrangements are rare events; by scoring enough recombinants we were able to determine and exclude the recombinant class that represented the exception.

Plasmid rescue:
Genomic sequence flanking the P elements used in the male recombination experiments was obtained from the Berkeley Drosophila Genome Project, with the exception of P{PZ}l(3)09904. We performed plasmid rescue from this strain as described (WILSON et al. 1989 ). Genomic DNA from 10 flies was digested with XbaI and added to a 200-µl ligation reaction. The reaction was phenol/chloroform extracted, ethanol precipitated, and resuspended in 10 µl Tris-EDTA. DH5- cells (GIBCO-BRL, Gaithersburg, MD) were transformed and plated on Luria broth-kanamycin. DNA was extracted from six resulting colonies and subjected to restriction analysis to verify that all were identical. Two were subsequently chosen for sequencing.

Sequence analysis of mutations:
Candidate genes within the intervals defined by the male site-specific recombination experiments were examined for mutational changes between wild-type (Oregon-R) and the mutant stocks. PCR primers were designed using GCG software and tested using the Amplify program (W. R. Engels, University of Wisconsin-Madison). PCRs were performed on genomic DNA, products were purified using Wizard DNA preps (Promega, Madison, WI), and sequencing was performed on an automated sequencer (IBI, New Haven, CT). To enhance the quality of certain sequences, dITP was incorporated into the PCR reaction (DIERICK et al. 1993 ). For kkv, the following PCR primers were used to amplify a 2-kb fragment that encodes the most highly conserved region of the protein: 5'-tatccatttagagcgcccac-3' and 5'-tgaacgaggttggagaactg-3'. For knk, two sets of primers were used to amplify two overlapping fragments spanning the entire knk open reading frame: 5'-agcgaacccgcgaaaaaacc-3' and 5'-gacccaaaacctgaacgcac-3' were used to amplify a 1.8-kb fragment containing the proximal part of the knk coding region, while 5'-cttggactcaaagaccattc-3' and 5'-agcatacattctacacacgc-3' were used to amplify a 1.8-kb fragment containing the distal portion of the knk sequence.

Sequence similarity searches were performed using the BLAST program (http://www.ncbi.nlm.nih.gov/BLAST/; ALTSCHUL et al. 1990 ). Sequence alignments were performed using the ClustalW program (http://www.ebi.ac.uk/clustalw/; THOMPSON et al. 1994 ). Transmembrane predictions were generated using the TMHMM program (http://www.cbs.dtu.dk/services/TMHMM-2.0; SONNHAMMER et al. 1998 ) and the TMPRED program (http://www.ch.embnet.org/software/TMPRED_form.html).

Lufenuron treatment:
Lufenuron (Sigma, St. Louis/Fluka) was dissolved in 95% ethanol at a stock concentration of 10 mg/ml. Working concentrations ranged from 0.1 to 10 µg/ml and were achieved by diluting into 95% ethanol and adding 100 µl to a standard fly vial containing 5 ml of solidified medium, which had been scored with a spatula to accelerate absorption of the supplement. Control vials were handled the same way but with no lufenuron added to the 95% ethanol supplement.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Characterization of mutations that disrupt embryonic cuticle formation:
Mutations in two genes, originally named blimp and zeppelin, were isolated coincidentally in an F₃ lethal genetic screen for mutations that modify the wg mutant phenotype. This screen identified important new components in the Wg/Wnt signaling pathway, such as dTCF (VAN DE WETERING et al. 1997 ; CAVALLO et al. 1998 ) and dAPC2 (MCCARTNEY et al. 1999 ). However, the screen also yielded several mutations that did not suppress wg but that did alter the appearance of the embryonic cuticles. Homozygous mutant embryos fail to hatch but are initially hyperactive, often reversing themselves within the eggshell (Fig 1A and Fig B). When the mutant embryos are mechanically devitellinized, the resulting cuticle preparations stretch to a much greater extent than do wild-type cuticles (Fig 1C and Fig 2). We refer to the inflated appearance of these embryonic cuticles as the blimp phenotype, and this is the source of the gene names. Both genes were mapped meiotically against the rucuca multiply marked chromosome and found to lie between st and cu, roughly at genetic map position 47.0.

View larger version (55K): [in this window] [in a new window] [Download PPT slide]   Figure 1. The "retroactive" and blimp phenotypes. (A) kkv homozygous mutant embryo oriented correctly within the eggshell, with anterior to the left. Arrow points to the micropyle, a specialization of the eggshell that marks the anterior of the egg. (B) zep homozygous mutant embryo showing reversed orientation. The micropyle of the eggshell is to the left (arrow), but anterior head structures in the embryo are to the right. Note that the head structures of the zep embryo appear normal, whereas the kkv mutant head skeleton is degenerated. (C) A devitellinized cuticle preparation of embryos from a kkvDZ8 stock shows the blimp phenotype of a kkv homozygote (bottom) compared with a wild-type kkv/+ heterozygous sibling (top).

View larger version (93K): [in this window] [in a new window] [Download PPT slide]   Figure 2. The blimp class of mutations. (A) A wild-type mechanically devitellinized cuticle preparation shows the normal extent to which the embryonic cuticle is flattened by a coverslip. (B) An embryo derived from a wild-type female that has been fed on lufenuron-treated food shows much greater stretching of the cuticle after devitellinization, comparable to the kkvDZ8 mutant phenotype (C). Defects in the head skeleton of the lufenuron-treated specimen also resemble those observed in the kkv mutant embryo. (D) knk1 mutant embryos also show cuticle stretching although the head skeleton defects are milder. (E) zepLP13 mutant embryos show stretching to the same extent as kkv mutant embryos, but the head skeleton is much more wild type in appearance. rtv11 mutant embryos (F) and grhIM embryos (not shown) also show excessive stretching when flattened under a coverslip. The degree of cuticle stretching is somewhat variable between preparations and so is not as accurate a measure of allelic severity as is degree of head cuticle disruption. A–F are at the same scale.

In addition to the blimp mutant phenotype, all four independently isolated mutant alleles of blimp show a defect in the embryonic head skeleton and slight denticle belt abnormalities (Fig 1C and Fig 2A and Fig C). Two previously identified genes, kkv and knk, had been mapped to genetic map positions 47.5 and 49.1, respectively, and both produce defects in the embryonic head skeleton when mutated. Consequently we tested both mutations for failure to complement our blimp and zeppelin mutations. All four of our blimp alleles fail to complement kkv, indicating that they are new mutations in this known gene. However, our zeppelin mutation complements both knk and kkv, and therefore zep represents a new gene.

In the course of our complementation tests, we discovered that the original kkv and knk mutations produce a blimp phenotype when the embryos are mechanically devitellinized (Fig 2D). The genetic screens in which these alleles were isolated (JURGENS et al. 1984 ) did not involve mechanical devitellinization of the cuticle preparations, and so the inflated-cuticle aspect of their phenotypes had been missed. For this reason, we reexamined other mutations from these screens, which had been described as having defects in the embryonic head skeleton. We found that in addition to kkv and knk, rtv and grh produce blimp phenotypes (Fig 2F and Table 1). Thus we have defined five different genes that are required for normal cuticle integrity.

The retroactive gene, as the name suggests, was also noted to show reversal within the eggshell. We find that kkv embryos show roughly the same frequency of reversal as do retroactive embryos: 8–9% are reversed relative to wild type (Table 2). knk mutants show a lower frequency of reversal, 2%, whereas zeppelin mutants show a far greater frequency of reversal, approaching 27% (Table 3). This phenotype probably reflects the increased elasticity of the cuticle, allowing the mutant embryos to move more freely within the confines of the eggshell than do wild-type embryos.

The cuticle defects, particularly the disruption of the head skeleton, are most severe in kkv and grh mutants. All alleles of kkv, both those isolated previously and those identified in this screen, produce similar phenotypes. When removed from the vitelline membrane, kkv and grh mutant embryos are very flaccid and are not motile although they are able to contract their body wall muscles. All three alleles of knk and the one available allele of rtv produce milder defects in the head skeleton and denticle belts. When removed from the vitelline membrane they are more robust than the kkv and grh mutants, and they are motile but die within hours after removal from the eggshell. The head skeleton and denticle belts of zep mutants are almost wild type (Fig 2A and Fig E) and these embryos are sometimes able to hatch on their own, although they die at roughly the same stage as the knk and rtv mutants. The degree of cuticle expansion can vary among cuticle preparations due to uncontrollable differences in the mechanical devitellinization process. However, the severity of head defects, flaccidity, and motility are consistent within the alleles of each complementation group. Thus the phenotypic effects of the blimp class mutations can be ranked from most to least severe: kkv, grh > rtv, knk > zep.

We suspect that at least one more gene may be associated with a blimp phenotype. Embryos homozygous for the small deficiency Df(3R)ry615, which deletes cytological positions 87B12–87E8, show a phenotype similar to that of zep. Thus a sixth gene involved in cuticle formation may reside within this interval.

Mapping mutations that produce the blimp phenotype:
We chose to focus our further analysis on kkv, knk, and zep, since these three genes lie close together on the third chromosome. First, we refined our mapping by analyzing deficiencies from the Bloomington Stock Center. We find that kkv fails to complement Df(3R)3-4 (82F3-4;82F10-11) but does complement the overlapping deficiency Df(3R)ME15 (81F3-6;82F5-7). This narrowed down the interval where kkv maps to between 82F5 and 82F11 (Fig 3), contradicting earlier work that had placed it between 83C1 and 84B2 (JURGENS et al. 1984 ). Likewise, we have determined a new cytological position for knk: knk fails to complement Df(3R)by62 (85D11-13;85F6) but it complements Df(3R)by10 (85D8; 85E10-13), the deficiency to which it had been previously assigned (JURGENS et al. 1984 ). Thus its new interval is 85E10–85F6 (Fig 3). zep fails to complement Df(3R)4-75, a small deficiency within the centromeric heterochromatin (81Fab). In all cases, the phenotypes generated by the mutation in trans to the noncomplementing deficiency are indistinguishable from the homozygous mutant phenotype. Therefore our new alleles of kkv, kkv^DP14, kkv^DZ8, kkv^JH9, and kkv^LX5, and of zep, zep^LP13, as well as the previously isolated allele of knk, knk¹, are likely to represent null mutations at these loci.

View larger version (28K): [in this window] [in a new window] [Download PPT slide]   Figure 3. Genetic mapping of kkv, knk, and zep. All three genes map to the third chromosome near the centromere. zep (left) appears to be located within the centromeric heterochromatin, as it maps within the interval deleted by Df(3R)4-75 (81Fab) and is not included within the interval deleted by Df(3R)ME15 (81F3-6;82F5-7). kkv (middle) lies within the interval deleted by Df(3R)3-4 (82F3-4;82F10-11) and has been positioned between P{PZ}09904 (82F8-9) and EP(3)0974 (83A1-2). knk (right) maps to the interval deleted in Df(3R)by62 (85D11-13;85F6) but is not included in the interval deleted in Df(3R)by10 (85D8;85E10-13). It has been more precisely positioned between EP(3)3092 (which is erroneously assigned to 87C8-9 by the Berkeley Drosophila Genome Project) and two P insertions at 85F12-13, P{PZ}tws01436 and P{PZ}tws02414.

To correlate these cytological positions with the molecular map, we used male site-specific recombination to place our genes with respect to P-element insertions within each interval. Although meiotic recombination does not occur in Drosophila males, crossing over can be induced at the position of P-element insertions when transposase is present in the male germline (PRESTON and ENGELS 1996 ). This observation has been adapted for mapping mutations rapidly at molecular resolution, since P-element insertions can be positioned on the genomic scaffold by performing plasmid rescue techniques and analyzing the recovered flanking DNA (CHEN et al. 1998 ).

Unfortunately, the heterochromatic region where zep is located has not yet been sequenced and so we have not yet identified the gene responsible for this mutant phenotype. However, both kkv and knk lie within characterized regions of the genome and we were able to identify the transcription units corresponding to these genetic loci.

Molecular analysis of kkv, a chitin synthase gene:
We determined that kkv lies distal to P{PZ}l(3)09904 (82F8-9) and proximal to EP(3)0974 (83A1-2; Fig 3). We identified the genomic sequence flanking P{PZ}l(3) 09904 by sequencing a plasmid rescue fragment, and we obtained genomic sequences flanking EP(3)0974 from the Berkeley Drosophila Genome Project. By comparing these sequences with the Drosophila genome sequence (ADAMS et al. 2000 ), we refined the location of kkv to a 40-kb interval containing nine predicted genes. Within this region we identified a candidate gene that could explain the elastic cuticle phenotype behind the blimp phenotype: a predicted transcript, CG2666, encoding a putative chitin synthase.

Chitin synthase genes have been identified not only in sheep blowfly (TELLAM et al. 2000 ) and mosquito (IBRAHIM et al. 2000 ), but also in Saccharomyces cerevisiae as chitin is a major component of the fungal cell wall (CABIB et al. 2001 ). All of these enzymes share two conserved domains (Fig 4 and Fig 5A and Fig B), referred to as con1 and con2, each of which contains several invariant residues essential for the catalytic activity of the yeast chitin synthase (NAGAHASHI et al. 1995 ; YABE et al. 1998 ). We performed sequence analysis on our kkv mutants, concentrating on a 2-kb exon of CG2666 that contains the con1 and con2 domains of this fly chitin synthase. In our kkv^DZ8 allele, we detected a mutation that changes tryptophan 947 of the predicted amino acid sequence to a stop codon (tgg tag). This would result in a truncated protein that lacks the con2 domain. The kkv^JH9 allele contains a mutation that converts glycine 1063 to aspartate (ggc gcc). This substitution disrupts an invariant SWG motif in the con2 domain (Fig 4). Thus both of these mutations would be expected to abolish catalytic activity of the chitin synthase enzyme. We also examined the sequence spanning the con1 and con2 domains from the existing kkv alleles kkv¹, kkv², and kkv³, but were unable to find any nucleotide changes within this region. Presumably the genetic lesions in these alleles lie elsewhere in the gene, either 5' or 3' to the highly conserved region examined (Fig 5A).

View larger version (44K): [in this window] [in a new window] [Download PPT slide]   Figure 4. Comparison of conserved domains from kkv and yeast chitin synthases. (A) Alignment of the con1 domain from kkv and from the three S. cerevisiae chitin synthases, CHS1, CHS2, and CHS3. This domain is highly conserved among all chitin synthases and contains several clusters of invariant residues. (B) Alignment of the con2 domain from kkv and the yeast CHS1, CHS2, and CHS3. Although this domain is less highly conserved than con1 overall, there are several invariant residues, including the C-terminal SWG motif that is required for catalytic activity. kkvDZ8 introduces a stop codon at Trp 947, so that the protein is truncated before the con2 domain, and kkvJH9 changes an invariant glycine within the SWG motif (arrow) to an aspartic acid.

View larger version (183K): [in this window] [in a new window] [Download PPT slide]   Figure 5. Predicted gene structure and amino acid sequence of kkv. (A) Schematic diagram of predicted intron-exon structure for the kkv genomic region and the relative positions of con1 and con2, the two protein domains that are highly conserved among all chitin synthases. The 2-kb region that was examined for mutations is indicated by a bracket above the diagram. (B) Predicted amino acid sequence of the complete kkv protein. This differs from the sequence available from the Berkeley Drosophila Genome Project. However, similar conclusions about the chitin synthase-1 gene structure were reached by R. Tellam and are cited in TELLAM et al. 2000 as well as in a Flybase error report for CG2666, dated July 2000, at http://flybase.bio.indiana.edu/.bin/fbidq.html?FBrf0129248. The con1 domain is underscored and the con2 domain is doubly underscored. (C) ClustalW alignment of kkv sequence with other insect chitin synthases: the sheep blowfly, L. caprina, and the mosquito, A. aegypti, as well as a second chitin synthase found in the Drosophila genome, CG7464. kkv shows greater similarity to the enzyme from L. caprina than to the second Drosophila chitin synthase.

The gene annotation for CG2666 released by the Berkeley Drosophila Genome Project predicts a chitin synthase enzyme, but was not accurate with regard to the predicted transcript. To determine the correct kkv gene structure, we performed a sequence similarity search against the protein database (ALTSCHUL et al. 1990 ). Using regions of high homology with other chitin synthases, we deduced the proper coding regions for kkv from the genomic sequence available from the Berkeley Drosophila Genome Project. As expected, kkv shows sequence similarity to a large number of chitin synthases from various species, with highest similarity to Caenorhabditis elegans, mosquito, sheep blowfly, and a second Drosophila gene, CG7464 (Fig 5C). We used two different computer programs, TMHMM and TMPRED, to determine the topology of the predicted multiple transmembrane domains. As with the sheep blowfly chitin synthase, the kkv sequence appears to have 17–18 transmembrane domains oriented such that the catalytic domains of the enzyme face the cytosol.

Molecular characterization of knk, a novel protein gene:
We determined that knk lies distal to the P-element insertion EP(3)3092 and proximal to P{PZ}tws01436 and P{PZ}tws02414 (Fig 3). Genomic sequences from these insertions were obtained from the Berkeley Drosophila Genome Project, and they were found to flank a 20-kb interval containing three predicted transcription units. We performed sequence analysis on the three existing knk mutations for all three candidate genes: CG6217, CG6208, and CG3940. All three knk alleles carry nonsense mutations within the predicted reading frame of CG6217 (Fig 6A). knk¹ converts tryptophan 299 to a stop codon (tgg tag), knk² changes lysine 656 to a stop codon (aag tag), and knk⁴ mutates leucine 631 to a stop codon (ttg tag). None of the mutant stocks showed significant nucleotide changes in the other two genes. Therefore we propose that CG6217 represents the knk gene.

View larger version (60K): [in this window] [in a new window] [Download PPT slide]   Figure 6. Predicted gene structure and amino acid sequence of knk. (A) Schematic diagram of predicted intron-exon structure for the knk genomic region. Arrows mark the positions of nonsense mutations found in knk mutant strains. knk1 changes the Trp 299 codon from UGG to UAG. Interestingly, both knk2 and knk4 are base transversions rather than the more typical transitions generated by EMS: knk2 changes the Lys 655 codon from AAG to UAG, and knk4 changes the Leu 632 codon from UUG to UAG. (B) ClustalW alignment of the knk (CG6217) gene product (protein product accession no. AAF54497) and the two most closely related Drosophila proteins, encoded by Skeletor (CG14681) and CG12492.

The transcription start site has not been defined for CG6217/knk, and so we do not know if the entire 5' untranslated region is included within the first exon predicted for the gene. The gene spans 2695 bp and consists of six exons, with conceptual translation yielding a protein of 689 amino acids. The protein sequence shows similarity to two other Drosophila genes: Skeletor and CG12492 (Fig 6B). Skeletor encodes a protein thought to be a component of the nuclear spindle matrix, but its true function is not known as no mutant alleles have yet been isolated (WALKER et al. 2000 ). CG12492 is a predicted gene of unknown function. Thus it is not obvious from molecular analyses what function knk may play in the process of embryonic cuticle formation.

We believe that other loci in the fly genome are also involved in cuticle integrity. Our genetic screen was not designed specifically to recover blimp class phenotypes, and so our isolation of new kkv alleles was purely fortuitous. Since we did not recover knk alleles, it is clear that our small-scale screen did not approach saturation for this type of mutant phenotype. Furthermore, in the course of our deficiency analysis of the blimp mutations, we found a deficiency, Df(3R)ry516, which complements kkv, knk, and zep, but that produces a blimp phenotype when homozygous. This suggests that a blimp class gene resides within the interval defined by the breakpoints of this deficiency, cytological positions 87B12–87E8. The Drosophila genome annotations place the knk homologs, Skeletor at 86C2 and CG12492 at 95C5-7, as well as the second fly chitin synthase CG7464 at 79B2-3, outside of this cytological region. However, it is worth noting that determining the exact breakpoints of deficiencies by cytology is difficult and that Skeletor in particular might be the gene responsible for the phenotype associated with this deficiency. In the absence of mutations in Skeletor, we cannot eliminate this possibility.

Finally, knk function does not appear to be additive with kkv function; knk kkv double-mutant embryos show a phenotype indistinguishable from the kkv single- mutant embryos (data not shown). Since kkv is epistatic, knk appears to function upstream of chitin deposition in the cuticle formation process. Alternatively, knk may act parallel to the kkv chitin synthase in a pathway that independently influences chitin deposition. Although kkv mRNA can be detected in the epidermis of late-stage embryos (not shown), knk mRNA was not detected at significant levels at any stage of embryogenesis, using standard digoxygenin whole mount in situ hybridization protocols. This suggests that very low levels of gene product are sufficient to mediate wild-type knk gene activity.

Lufenuron treatment produces a kkv phenocopy:
Disruption of chitin synthesis by treatment with lufenuron, a potent inhibitor, kills Drosophila larvae during ecdysis to the next larval instar (WILSON and CRYAN 1997 ). Feeding lufenuron to mature adult Drosophila does not affect the health of the individuals but does disrupt the development of eggs laid by treated females. The embryos were reported to show hyperactivity within the eggshell but fail to hatch from the egg perhaps because of defects in the head skeleton. We repeated these experiments to determine if the embryos would also show a blimp phenotype when mechanically devitellinized. We find that cuticle preparations of embryos from lufenuron-treated mothers are very similar in appearance to those of kkv mutant embryos (Fig 2, A–C). Thus inhibition of chitin synthesis by artificial means produces a phenotype like that of loss-of-function mutations in a chitin synthase gene. We presume that lufenuron fed to the mothers is incorporated into the eggs for inhibition of chitin synthase to occur in the embryo.

Genetic interactions with shotgun, the Drosophila E-cadherin:
The identification of kkv as a chitin synthase and the ability of a chitin synthesis inhibitor to phenocopy kkv shows that disrupting synthesis or deposition of chitin alone can account for the blimp phenotype. However, we believe that two of the blimp class genes, knk and zep, may function in the epidermis prior to cuticle deposition because both interact genetically with mutations in the Drosophila E-cadherin, encoded by shotgun (shg).

shg is provided both maternally and zygotically and is required for oogenesis as well as for subsequent embryonic development (TEPASS et al. 1996 ; GODT and TEPASS 1998 ). Zygotic loss of shg function produces defects only in those tissues that undergo dramatic morphogenesis, indicating that maternal shg product provides sufficient cadherin-mediated adhesion for most embryonic epithelia but not for those subject to mechanical stress. The embryonic ventral epidermis undergoes substantial cell rearrangements and degenerates in shg mutant embryos, resulting in a fragmented cuticle (NUSSLEIN-VOLHARD et al. 1984 ). shg loss-of-function mutations are completely recessive, but in embryos homozygous for either knk or zep, heterozygosity for shg leads to a fragmented cuticle phenotype similar to that of a shg homozygote (Table 2). This effect is not observed for kkv, rtv, or grh, the other blimp class genes. Thus knk and zep seem likely to encode proteins that are more broadly involved in epidermal development. In the absence of these gene functions, a single gene dose of shg becomes inadequate for epithelial cell adhesion in the ventral epidermis.

The epidermal tissue of the blimp class mutants appears structurally indistinguishable from that of wild-type embryos. Epidermal cell membrane-associated proteins such as Coracle (FEHON et al. 1994 ), Armadillo (PEIFER et al. 1991 ), and dAPC2 (MCCARTNEY et al. 1999 ) show normal localization and reveal no discernible difference between mutant and wild-type embryos in cell size, shape, or number (data not shown). Thus it remains to be determined what aspect of epidermal cell function is disrupted in knk and zep mutant embryos.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We have defined a set of genes that are required for proper development of the Drosophila embryonic cuticle. Mutations in these genes are zygotic lethal and result in flaccid embryos with very elastic cuticles that stretch to a remarkable degree when flattened beneath a coverslip. Three of the genes associated with this blimp phenotype, kkv, knk, and zep, map to a small region near the centromere of the third chromosome, but since at least two appear to be structurally unrelated, they do not appear to represent a gene cluster. We have determined the molecular identities of kkv and knk, with kkv encoding a chitin synthase enzyme and knk encoding a novel protein of unknown function.

The blimp phenotype can be accounted for by a failure of the epidermal cells to deposit cuticle properly. Loss-of-function mutations in kkv, the chitin synthase, or inhibition of chitin synthesis with the benzoylphenyl urea, lufenuron, both produce this stretchy cuticle effect. A second Drosophila chitin synthase, encoded by CG7464, has not been characterized genetically and so we do not know if it participates in formation of the embryonic cuticle. The chitin synthases in yeast, CHS1, -2, and -3, each perform a distinct function in cell wall formation: CHS1 repairs damaged chitin during cell separation, CHS2 is required for primary septum formation, and CHS3 is involved in all other chitin synthesis (CABIB et al. 2001 ). Specialization among the insect chitin synthases is likely as well; the enzyme isolated from the mosquito Aedes aegypti appears to be specific for the midgut epithelium, where it contributes to the peritrophic matrix that surrounds the blood bolus after a meal (IBRAHIM et al. 2000 ). The authors postulate that a second as-yet-unidentified chitin synthase in A. aegypti directs cuticular synthesis of chitin. In contrast, in the sheep blowfly, L. caprina, chitin synthase is expressed in larval epidermal cells and so is likely to be involved in formation of the exoskeleton (TELLAM et al. 2000 ).

Another blimp class gene, grainy head, was previously known to affect head skeleton and embryonic cuticle (BRAY and KAFATOS 1991 ). grh encodes a GATA family transcription factor and activates the transcription of a number of genes during development, one of which is Dopa-decarboxylase (Ddc). This enzyme is synthesized in the cuticle-secreting layer of cells at the end of embryogenesis; the dopamine produced undergoes further metabolism and oxidation to produce quinones that crosslink cuticular proteins (SCHOLNICK et al. 1983 ). Thus, loss of grh function would result in weakening of the cuticle indirectly through its failure to activate Ddc expression.

The contribution of the remaining blimp class genes to cuticle formation is not clear. rtv and zep remain uncharacterized at the molecular level; we have characterized knk, but its molecular identity does not provide insight into its possible role. We suspect that further analysis of knk and zep may reveal a function within the epidermal cell layer that involves cell-cell adhesion, because their loss of function sensitizes the embryo to reduction in zygotic DE-cadherin dose. This could indicate that knk and zep have a general role in maintaining epidermal integrity, which indirectly affects cuticle deposition. Since kkv is epistatic to knk, knk could function upstream of the chitin synthase, for example, by localizing the enzyme to the correct membrane domain. Alternatively, knk and zep may be directly involved in the extrusion or assembly of cuticle matrix components, and disruption of this process in mutant embryos places abnormal stress on the adherens junctions that link epidermal cells together. In any case, further study of these two genes may shed light on the interplay between the epidermal epithelium and the overlying cuticular layers that it must construct during late stages of embryogenesis.

These late-stage events appear to be coordinated by the steroid hormone, 20-hydroxyecdysone. Genes that regulate ecdysone levels in the embryo influence both epidermal integrity and cuticle deposition (CHAVEZ et al. 2000 ). However, mutations in ecdysone regulatory genes produce a broader spectrum of defects than do the blimp class mutations, including severe disruption of morphogenetic movements throughout embryogenesis. In contrast, knk, zep, and the other blimp class mutations do not disrupt early stages of development and these mutant embryos have normal morphology. Thus, the blimp class genes may be a subset of the targets of ecdysone regulation during late stages of embryogenesis, where they orchestrate the final stages of epidermal development and cuticle secretion.

	FOOTNOTES

¹ Present address: Alzheimer Research Laboratory, Department of Neurosciences, Case Western Reserve University School of Medicine, Cleveland, OH 44106.
² Present address: The Neurosciences Institute, 10640 John Jay Hopkins Dr., San Diego, CA 92121.

	ACKNOWLEDGMENTS

We are extremely grateful to the Bloomington Stock Center for their unending patience and endless supply of essential fly stocks. We also thank Adelaide Carpenter and the Berkeley Drosophila Genome Project for sharing information and reagents. Thanks also to members of the Bejsovec lab at Duke for comments on the manuscript, and to Tracie Quarles for technical assistance. This work was supported by National Institutes of Health grant GM-59068 and National Science Foundation grant IBN 97-34072 to A.B. and a Macy Scholar Undergraduate Fellowship to S.O.

Manuscript received September 11, 2001; Accepted for publication February 8, 2002.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ADAMS, M. D., S. E. CELNIKER, R. E. HOLT, C. A. EVANS, and J. D. GOCAYNE et al., 2000  The genome sequence of Drosophila melanogaster.. Science 287:2185-2195.

ALTSCHUL, S. F., W. GISH, W. MILLER, and E. W. MYERS, 1990  Basic local alignment search tool. J. Mol. Biol. 215:403-410.

BRAY, S. J. and F. C. KAFATOS, 1991  Developmental function of Elf-1: an essential transcription factor during embryogenesis in Drosophila.. Genes Dev. 5:1672-1683.

CABIB, E., D.-H. ROH, M. SCHMIDT, L. B. CROTTI, and A. VARMA, 2001  The yeast cell wall and septum as paradigms of cell growth and morphogenesis. J. Biol. Chem. 276:19679-19682.

CARPENTER, A. T. C., 1999  Saturation mutagenesis of region 82F. Dros. Inf. Serv. 82:113-114.

CAVALLO, R. A., R. T. COX, M. M. MOLINE, J. ROOSE, and G. A. POLEVOY et al., 1998  Drosophila TCF and Groucho interact to repress Wingless signaling activity. Nature 395:604-608.

CHAVEZ, V. M., G. MARQUES, J. P. DELBECQUE, K. KOBAYASHI, and M. HOLLINGSWORTH et al., 2000  The Drosophila disembodied gene controls late embryonic morphogenesis and codes for a cytochrome P450 enzyme that regulates embryonic ecdysone levels. Development 127:4115-4126.

CHEN, B., T. CHU, E. HARMS, J. P. GERGEN, and S. STRICKLAND, 1998  Mapping of Drosophila mutations using site-specific male recombination. Genetics 149:157-163.

COHEN, E., 1985  Chitin synthetase activity and inhibition in different insect microsomal preparations. Experientia 41:470-472.

COHEN, E., 1987  Chitin biochemistry: synthesis and inhibition. Ann. Rev. Entomol. 32:71-93.

DIERICK, H., M. STUL, W. DE KELVER, P. MARYNEN, and J. J. CASSIMAN, 1993  Incorporation of dITP or 7-deaza dGTP during PCR improves sequencing of the product. Nucleic Acids Res. 18:4427-4428.

DRYDEN, M. W., H. R. PEREZ, and D. M. ULITCHNY, 1999  Control of fleas on pets and in homes by use of imidacloprid or lufenuron and a pyrethrin spray. J. Am. Vet. Med. Assoc. 1:36-39.

FEHON, R. G., I. A. DAWSON, and S. ARTAVANIS-TSAKONAS, 1994  A Drosophila homolog of membrane-skeleton protein 4.1 is associated with septate junctions and is encoded by the coracle gene. Development 120:545-557.

GODT, D. and U. TEPASS, 1998  Drosophila oocyte localization is mediated by differential cadherin-based adhesion. Nature 395:387-391.

IBRAHIM, G. H., C. T. SMARTT, L. M. KILEY, and B. M. CHRISTENSEN, 2000  Cloning and characterization of a chitin synthase cDNA from the mosquito Aedes aegypti.. Insect Biochem. Mol. Biol. 30:1213-1222.

JÜRGENS, G., E. WIESCHAUS, C. NÜSSLEIN-VOLHARD, and H. KLUDING, 1984  Mutations affecting the pattern of the larval cuticle in Drosophila melanogaster: II. Zygotic loci on the third chromosome. Wilhelm Roux's Arch. Dev. Biol. 193:283-295.

MCCARTNEY, B. M., H. A. DIERICK, C. KIRKPATRICK, M. M. MOLINE, and A. BAAS et al., 1999  Drosophila APC2 is a cytoskeletally-associated protein that regulates Wingless signaling in the embryonic epidermis. J. Cell Biol. 146:1303-1318.

NAGAHASHI, S., M. SUDOH, N. ONO, R. SAWADA, and E. YAMAGUCHI et al., 1995  Characterization of chitin synthase 2 of Saccharomyces cerevisiae. Implication of two highly conserved domains as possible catalytic sites. J. Biol. Chem. 270:13961-13967.

NÜSSLEIN-VOLHARD, C. and E. WIESCHAUS, 1980  Mutations affecting segment number and polarity in Drosophila.. Nature 287:795-801.

NÜSSLEIN-VOLHARD, C., E. WIESCHAUS, and H. KLUDING, 1984  Mutations affecting the pattern of the larval cuticle in Drosophila melanogaster: I. Zygotic loci on the second chromosome. Wilhelm Roux's Arch. Dev. Biol. 193:267-282.

PEIFER, M., C. RAUSKOLB, M. WILLIAMS, B. RIGGLEMAN, and E. WIESCHAUS, 1991  The segment polarity gene armadillo affects the wingless signalling pathway in both embryonic and adult pattern formation. Development 111:1028-1043.

PRESTON, C. R. and W. R. ENGELS, 1996  P-element-induced male recombination and gene conversion in Drosophila.. Genetics 144:1611-1622.

SCHOLNICK, S. B., B. A. MORGAN, and J. HIRSH, 1983  The cloned Dopa decarboxylase gene is developmentally regulated when reintegrated into the Drosophila genome. Cell 34:37-45.

SONNHAMMER, E., G. VON HEIJNE, and A. KROGH, 1998  A hidden Markov model for predicting transmembrane helices in protein sequences. Proc. Int. Conf. Intell. Syst. Mol. Biol. 6:175-182.

TELLAM, R. L., T. VUOCOLO, S. E. JOHNSON, J. JARMEY, and R. D. PEARSON, 2000  Insect chitin synthase: cDNA sequence, gene organization and expression. Eur. J. Biochem. 267:6025-6043.

TEPASS, U., E. GRUSZYNSKI-DEFEO, T. A. HAAG, L. OMATYAR, and T. TOROK et al., 1996  shotgun encodes Drosophila E-cadherin and is preferentially required during cell rearrangement in the neurectoderm and other morphogenetically active epithelia. Genes Dev. 10:672-685.

THOMPSON, J. D., D. G. HIGGINS, and T. J. GIBSON, 1994  CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22:4673-4680.

VAN DE WETERING, M., R. CAVALLO, D. DOOIJES, M. VAN BEEST, and J. VAN ES et al., 1997  Armadillo co-activates transcription driven by the product of the Drosophila segment polarity gene dTCF.. Cell 88:789-799.

WALKER, D., D. WANG, Y. JIN, U. RATH, and Y. WANG et al., 2000  Skeletor, a novel chromosomal protein that redistributes during mitosis provides evidence for the formation of a spindle matrix. J. Cell Biol. 151:1401-1411.

WIESCHAUS, E., and C. NÜSSLEIN-VOLHARD, 1986 Looking at embryos, pp. 199–227 in Drosophila: A Practical Approach, edited by D. B. ROBERTS. IRL Press, Oxford.

WIESCHAUS, E., C. NÜSSLEIN-VOLHARD, and G. JURGENS, 1984  Mutations affecting the pattern of the larval cuticle in Drosophila melanogaster: zygotic loci on the X-chromosome and the fourth chromosome. Wilhelm Roux's Arch. Dev. Biol. 193:296-307.

WILSON, T. G. and J. R. CRYAN, 1997  Lufenuron, a chitin-synthesis inhibitor, interrupts development of Drosophila melanogaster.. J. Exp. Zool. 278:37-44.

WILSON, C., R. KURTH PEARSON, H. J. BELLEN, C. J. O'KANE, and U. GROSSNIKLAUS et al., 1989  P-element-mediated enhancer detection: an efficient method for isolating and characterizing developmentally regulated genes in Drosophila. Genes Dev. 3:1301-1313.

YABE, T., T. YAMADA-OKABE, T. NAKAJIMA, M. SUDOH, and M. ARISAWA et al., 1998  Mutational analysis of chitin synthase 2 of Saccharomyces cerevisiae: identification of additional amino acid residues involved in its catalytic activity. Eur. J. Biochem. 258:941-947.

This article has been cited by other articles:

				M. Narasimha, A. Uv, A. Krejci, N. H. Brown, and S. J. Bray Grainy head promotes expression of septate junction proteins and influences epithelial morphogenesis J. Cell Sci., March 15, 2008; 121(6): 747 - 752. [Abstract] [Full Text] [PDF]

				X. Guan, B. W. Middlebrooks, S. Alexander, and S. A. Wasserman Mutation of TweedleD, a member of an unconventional cuticle protein family, alters body shape in Drosophila PNAS, November 7, 2006; 103(45): 16794 - 16799. [Abstract] [Full Text] [PDF]

				A. Tonning, S. Helms, H. Schwarz, A. E. Uv, and B. Moussian Hormonal regulation of mummy is needed for apical extracellular matrix formation and epithelial morphogenesis in Drosophila Development, January 15, 2006; 133(2): 331 - 341. [Abstract] [Full Text] [PDF]

				B. Moussian, E. Tang, A. Tonning, S. Helms, H. Schwarz, C. Nusslein-Volhard, and A. E. Uv Drosophila Knickkopf and Retroactive are needed for epithelial tube growth and cuticle differentiation through their specific requirement for chitin filament organization Development, January 1, 2006; 133(1): 163 - 171. [Abstract] [Full Text] [PDF]

				W. P. Devine, B. Lubarsky, K. Shaw, S. Luschnig, L. Messina, and M. A. Krasnow Requirement for chitin biosynthesis in epithelial tube morphogenesis PNAS, November 22, 2005; 102(47): 17014 - 17019. [Abstract] [Full Text] [PDF]

				N. Ren, C. Zhu, H. Lee, and P. N. Adler Gene Expression During Drosophila Wing Morphogenesis and Differentiation Genetics, October 1, 2005; 171(2): 625 - 638. [Abstract] [Full Text] [PDF]

				C. Cenci and A. P. Gould Drosophila Grainyhead specifies late programmes of neural proliferation by regulating the mitotic activity and Hox-dependent apoptosis of neuroblasts Development, September 1, 2005; 132(17): 3835 - 3845. [Abstract] [Full Text] [PDF]

				K. A. Mace, J. C. Pearson, and W. McGinnis An Epidermal Barrier Wound Repair Pathway in Drosophila Is Mediated by grainy head Science, April 15, 2005; 308(5720): 381 - 385. [Abstract] [Full Text] [PDF]

				S. B. Ting, J. Caddy, N. Hislop, T. Wilanowski, A. Auden, L.-l. Zhao, S. Ellis, P. Kaur, Y. Uchida, W. M. Holleran, et al. A Homolog of Drosophila grainy head Is Essential for Epidermal Integrity in Mice Science, April 15, 2005; 308(5720): 411 - 413. [Abstract] [Full Text] [PDF]

				J. Tao, E. Kuliyev, X. Wang, X. Li, T. Wilanowski, S. M. Jane, P. E. Mead, and J. M. Cunningham BMP4-dependent expression of Xenopus Grainyhead-like 1 is essential for epidermal differentiation Development, March 1, 2005; 132(5): 1021 - 1034. [Abstract] [Full Text] [PDF]

				H. Merzendorfer and L. Zimoch Chitin metabolism in insects: structure, function and regulation of chitin synthases and chitinases J. Exp. Biol., December 15, 2003; 206(24): 4393 - 4412. [Abstract] [Full Text] [PDF]

				A. T. Chao, H. A. Dierick, T. M. Addy, and A. Bejsovec Mutations in Eukaryotic Release Factors 1 and 3 Act as General Nonsense Suppressors in Drosophila Genetics, October 1, 2003; 165(2): 601 - 612. [Abstract] [Full Text] [PDF]

J. Hemphala, A. Uv, R. Cantera, S. Bray, and C. Samakovlis Grainy head controls apical membrane growth and tube elongation in response to Branchless/FGF signalling