The initial establishment of the tracheal network in the Drosophila embryo is beginning to be understood in great detail, both in its genetic control cascades and in its cell biological events. By contrast, the vast expansion of the system during larval growth, with its extensive ramification of preexisting tracheal branches, has been analyzed less well. The mutant phenotypes of many genes involved in this process are probably not easy to reveal, as these genes may be required for other functions at earlier developmental stages. We therefore conducted a screen for defects in individual clonal homozygous mutant cells in the tracheal network of heterozygous larvae using the mosaic analysis with a repressible cell marker (MARCM) system to generate marked, recombinant mitotic clones. We describe the identification of a set of mutants with distinct phenotypic effects. In particular we found a range of defects in terminal cells, including failure in lumen formation and reduced or extensive branching. Other mutations affect cell growth, cell shape, and cell migration.

---

BRANCHED epithelial tubes are the structural and functional units of many animal organs, such as the lung, the kidney, or blood vessels in vertebrates. They enable the transport of gases or liquids in the body. The Drosophila tracheal (respiratory) system with its structural simplicity and accessible genetics has become an excellent model in recent years to study the events that control tube morphogenesis (GHABRIAL et al. 2003; UV et al. 2003; KERMAN et al. 2006). Several recent studies have shed light on the complexity of the genetic control and cellular processes of tracheal development. Yet we still do not have a full understanding of the molecular basis of all of the cell biological events involved. We do not know the effectors of signaling pathways, the molecules involved in cell remodeling, or the molecular mechanisms responsible for the sprouting of cytoplasmic outgrowth and lumen formation in terminal branching.

The basic network of the tracheal epithelial tubes is constructed during embryogenesis (reviewed in GHABRIAL et al. 2003; UV et al. 2003; KERMAN et al. 2006). It arises from 10 clusters of ectodermal cells on either side of the embryo. After two cycles of cell division, 80 cells of each tracheal placode invaginate to form an elongated sac, from which six major buds grow in different directions, giving rise to the different primary branches. A set of stereotypic migration, branching, and differentiation events, largely directed by the fibroblast growth factor (FGF) homolog Branchless (Bnl) and its receptor Breathless (Btl) (SUTHERLAND et al. 1996), produces secondary and terminal branches. Cell fusions interconnect the branches.

Further differentiation occurs in the postembryonic period. To support the massive growth during the larval period, the terminal tracheal cells continue to ramify, forming extensive arrays that cover and support large areas of target tissues (MANNING and KRASNOW 1993). In contrast to the genetically predetermined branching pattern in the embryo, formation of terminal branches in the larva is more flexible, reacting directly to the oxygen requirement of the surrounding tissues (for review see GHABRIAL et al. 2003; KERMAN et al. 2006), which is signaled to the tracheal cells by Branchless (JARECKI et al. 1999). However, flexibility in branching does not mean lack of organization. Spacing between branching points is regular and branches never cross each other.

Although some players involved in these processes have been identified, it is not fully understood what happens in the terminal cells upon FGF signal activation by Bnl binding to the receptor Btl. There may be a number of reasons why previous genetic screens have identified only a subset of the players. The zygotic loss-of-function of molecules involved in tracheal morphogenesis may not lead to detectable defects in the embryo if a significant maternal component of the molecule is available, masking the loss of the zygotic gene product. Conversely, the function of molecules needed for branch morphogenesis during larval stages may be difficult or impossible to analyze in mutants if the same molecule is also required for other essential processes during embryogenesis. We therefore set up a screen to analyze defects in homozygous mutant clones within heterozygous larvae using the mosaic analysis with a repressible cell marker (MARCM) system (LEE and LUO 1999, 2001). In this technique, expression of a UAS construct with a reporter gene is suppressed by the presence of the repressor GAL80, which is placed on the homologous chromosome of the one carrying the mutation to be analyzed. When a mitotic recombination between these chromosomes occurs, the reporter gene is expressed only in clones homozygous for the mutation, i.e., those that have lost the GAL80-carrying chromosome arm. The system has been successfully applied to analyze the roles of known genes in the later stages of trachea formation (CABERNARD and AFFOLTER 2005; GHABRIAL and KRASNOW 2006; LEVI et al. 2006).

Here we report the identification of 343 strains carrying mutations on the left arm of the second chromosome, affecting formation of the tracheal system and the characterization of the phenotypes they cause. A further analysis of a subset of these mutants and their phenotype in a different part of the respiratory system, the air sac primordium, is reported in the accompanying article by CHANUT-DELALANDE et al. (2007, this issue).

Drosophila stocks:

Drosophila melanogaster lines were raised at 25° in standard conditions. To generate ethyl methanesulfonate (EMS) mutant lines the following strains were used: y^d2w¹¹¹⁸P{ey-FLP.N}2 P{GMR-lacZ.C(38.1)}TPN1;P{neoFRT}40A (BERGER et al. 2001) and P{Hs-hid}sp/CyO⁸ftz lacZ (KNOBLICH and LEHNER 1993; GRETHER et al. 1995; kindly provided by R. Lehmann). The following MARCM line was used during the tracheal screen: yw,hsFlp1.22;tub-Gal80,FRT40A;btl-Gal4, UAS-GFP (kindly provided by S. Luschnig). For complementation analysis the following mutant lines were used: w^–; conv^[K6507]/CyO wg-lacZ, w^–; nrv2 ^[23B]/CyO wg-lacZ, vari ^[K3953]/CyO (kindly provided by G. Beitel) as well as a mmy deficient line Df(2L)BSC6, dp^ov1 cn¹/SM6a (PARKS et al. 2004) (Bloomington Stock Center). Clonal analyses in wing discs were performed with w^–,y, hsFlp 1.22; FRT40A, ubi-GFP/SM6-TM6 and w^–,y,hsFlp1.22 tubGal4 UAS nlsGFP; FRT40A Gal80 strains (kindly provided by T. Klein). The following mutant line was used for MARCM analysis in the tracheal system: w^–; nrv2^[K13314]40AFRT/CyO, act-GFP (kindly provided by R. Fohen). For mutant characterization the w¹¹¹⁸; ; P{UAS-myr-mRFP}2/TM6B, Tb¹ strain was used (Bloomington Stock Center). Deficiency lines for 2L generated by Exelixis (PARKS et al. 2004) were used for complementation.

EMS mutagenesis and stock establishment:

The mutant lines were generated in batches of 1000. In each round 300 2- to 3-day-old males from an isogenized stock with an FRT site on the left arm of the second chromosome (FRT40A) were mutagenized with 30 mM EMS (standard protocol) (GRIGLIATTI 1998). EMS-fed males were crossed with P{Hs-hid}sp/CyO⁸ftz lacZ females (mass cross). To select progeny with the required genotype (FRT40A, mutation/CyO⁸ftz lacZ) larvae were heat-shocked on days 5 and 6 for 2 hr at 38° (to induce the hs-hid transgene expression). Only larvae carrying the mutated FRT chromosome and not the hs-hid transgene survived this procedure. One thousand males from the progeny were used to establish single lines by crossing them to P{Hs-hid}sp/CyO⁸ftz lacZ females. Flies with the correct genotype were selected by heat-shock treatment.

Generation of MARCM clones in the trachea:

Male progeny of FRT40A mutant lines were crossed with females carrying all components of the MARCM system (Figure 1A). A UAS-GFP construct was used as a reporter gene, driven by the trachea-specific btl-GAL4. The eggs were collected for 4 hr at 25° and heat-shocked for 30 min at 38° to induce mitotic recombination. Live third instar larvae were observed under a Leica fluorescence stereo microscope MZFLIII (Fluo-Combi). We note that whereas 50% of the larvae were expected to have the genotype in which MARCM clones can be generated, the proportion of larvae that actually showed clones ranged from 30 to 40%, even when FRT chromosomes without known mutations were used. For live observation larvae were first immobilized by adding water to the vials and after 1 hr placed on microscopic slides featuring a plastic grid that prevents disturbance of larval structure by a coverslip. Eight larvae were analyzed per line. For each line showing defects the phenotype was confirmed by analyzing a further 16 larvae. For detailed analysis of phenotypes Zeiss (Thornwood, NY) Axioplan 2 Imaging was used. Results were documented with an AxioCam HRm camera and Axiovision software (Zeiss).

View larger version (33K): In this window In a new window Download PPT slide   FIGURE 1.— MARCM analysis in wild-type larval trachea and screen-crossing scheme. (A) Crossing scheme for MARCM analysis. (B) Third instar larva with wild-type MARCM clones in the tracheal system; green, GFP. Bar, 1 mm. (C) Mutagenesis scheme: males carrying an isogenic FRT40A chromosome were mutagenized with EMS and crossed en masse to females heterozygous for a hs-hid transgene and the CyO ftz-LacZ balancer chromosome. The asterisk denotes an EMS-induced mutation. After heat-shock-based selection of progeny carrying a balanced mutagenized FRT40A chromosome, single males of such a genotype were individually crossed to hs-hid/CyO ftz-LacZ females. A heat-shock treatment of the progeny from these crosses allowed the establishment of balanced mutant stocks. Males and females from each line were crossed to establish stocks; in parallel, males were crossed to females carrying all the components of the MARCM system and the progeny were analyzed in third instar larvae.

 

Complementation test:

Preceding the complementation tests, all lines were checked for lethality. The homozygous lethal lines were identified by the absence of non-CyO flies after four generations. After an initial grouping into phenotypic classes all lines within a group were tested for complementation. The crosses were kept at 25° and progeny were scored for presence of non-CyO flies.

Immunostainings:

Embryos from mutant lines were collected at 25° for 16 hr, fixed according to standard protocols, and stained with mouse monoclonal antibody 2A12 (1:20, N. Patel) that recognizes an unknown antigen in the tracheal lumen and with rabbit anti ß-Gal antibody (1:500, Cappel) to distinguish between heterozygous and homozygous embryos. The secondary antibodies were biotin-labeled goat anti-mouse and biotin-labeled goat anti-rabbit (1:500; Jackson ImmunoResearch Laboratories, West Grove, PA).

Third instar larvae were cut open along the ventral midline, fixed in 4% paraformaldehyde in PBS, and stained with anti-armadillo (1:3000, from E. Wieschaus; RIGGLEMAN et al. 1990) and goat anti-mouse Alexa 568 (1:500; Invitrogen, San Diego) antibodies.

Clonal analysis in wing imaginal disc:

Larvae from the crosses FRT40A mut/CyO x w^–,y, hsFlp 1.22; FRT40A, ubi-GFP/SM6-TM6 or w^–,y,hsFlp1.22 tubGal4 UAS nlsGFP; FRT40A Gal80 were heat-shocked 24–48 hr after egg lay at 37° for 1 hr and left to develop. Wing imaginal discs from third instar larvae were dissected, fixed in 4% paraformaldehyde, and analyzed under a Zeiss Axioplan 2 microscope.

Conditions of the screen:

To identify genes involved in tracheal morphogenesis we performed a genetic mosaic MARCM clone screen (LEE and LUO 1999) for recessive mutations displaying defects in larval tracheal development. We restricted our analysis to genes located on the left arm of the second chromosome as the other chromosomes were screened by a similar approach by others (M. KRASNOW, personal communication; GHABRIAL and KRASNOW 2006; LEVI et al. 2006). We therefore used a chromosome carrying the FRT recombination site, FRT40A, at the base of this chromosome arm. Males carrying an FRT40A chromosome were crossed to "FRT40A MARCM females" (Figure 1A). These females carry a heat shock-flipase (hs-flp) transgene, an FRT40A chromosome with a tubulin-Gal80 (tub-Gal80) construct, and a third chromosome bearing the breathless-Gal4 (btl-Gal4) and UAS-green fluorescent protein (UAS-GFP) constructs (Figure 1A). A heat-shock treatment of the progeny of the crosses leads to flp-mediated mitotic recombination at the FRT40A sites, resulting in the loss of the GAL4 inhibitor GAL80 from cells homozygous for the mutation, which are therefore positively marked with GFP. As a time point for induction of FLP-mediated recombination we chose the late blastoderm/early gastrulation stage because thereafter proliferation in the tracheal primordium ceases (SAMAKOVLIS et al. 1996). The drawback of such early heat shock is an interference with blastoderm formation. However, we found that a heat shock of 30 min at 38° allowed survival and led to sufficient mitotic recombination to generate on average of 30–80 homozygous marked clones per L3 larva when control FRT40A chromosomes with no known mutations were used.

We first analyzed control clonal cells carrying no known mutations in the tracheae of wild-type third instar larvae to determine the wild-type features and to establish screening criteria. In such larvae the clonal cells were distributed randomly in all types of branches (Figure 1B). Usually the number of clonal cells was slightly higher in the anterior part of the larva. The clonal cells in the dorsal trunk were large and hexagonal (Figure 3A), except for fusion cells, which had a ring-like shape. Cells on secondary branches were elongated and varied in size depending on position. The number and length of branches in clonal terminal cells were highly variable with a higher branching rate in the anterior part of the animal. A properly formed lumen filled with gas was found in all clonal cells. The activity of the breathless-GAL4 transgene is not absolutely restricted to the tracheal system, and therefore GFP-positive clonal cells were also found in the epidermis, or in the heart, where they always occurred as a pair in a specific position (not shown).

View larger version (107K): In this window In a new window Download PPT slide   FIGURE 3.— Phenotypes of class D mutant clones. (A, C, and E) Wild-type clonal cells on the dorsal trunk are hexagonally shaped and flat. The shape of the dorsal trunk is not affected at the site of the clonal cell. (B) Clonal cells on the dorsal trunk in mutant line 2L3090 (subclass D0) clonal cells are smaller and rounded up. (D) In mutants from class DI, the dorsal trunk shows a bend at the position of the clonal cell. (F) In mutants from class DIII clonal cells in the dorsal trunk, but not in other branches, are smaller than in wild type. Insets in A, B, E, and F show higher magnifications of the regions marked with rectangles. Phenotypic classes are indicated next to allele names. Bars, 100 µm.

 
We also tested mutations that were known or expected to influence tracheal morphogenesis for their effects in third instar MARCM clones. The strongest effect was seen for mutations in components of the FGF-signaling pathway. For example, larvae in which clones were induced that were mutant for the FGF-signal transduction adaptor Downstream of FGF (Dof, also named Stumps or Heartbroken; MICHELSON et al. 1998; VINCENT et al. 1998; IMAM et al. 1999) showed only half as many clones as control larvae, and the homozygous mutant cells were always confined to the dorsal trunk and the primary or the secondary branches, but not to the terminal branches, as has also been described recently by others (GHABRIAL and KRASNOW 2006). To our surprise, cytoskeletal regulators known to play important roles in other cell shape changes or cell migration, such as DRhoGEF2 (DrhoGEF^4.1) (BARRETT et al. 1997), Dreadlocks (dock⁴), PAK-kinase (Pak¹⁴) (NEWSOME et al. 2000), or Capulet (capt¹⁰) (BAUM et al. 2000), showed minor or no effects apart from a slight reduction in the number of clones (not shown).

Screen for mutations on the left arm of the second chromosome:

We followed the mutagenesis scheme in Figure 1C to establish mutant fly stocks carrying random EMS-induced mutations on the second chromosome. Males carrying an isogenic FRT40A chromosome were treated with EMS and crossed en masse to females heterozygous for a heat-shock-hid (hs-hid) construct and the CyO ftz-LacZ; balancer chromosome. The hs-hid construct allows the heat-inducible elimination of larvae carrying this chromosome to select for progeny carrying a balanced mutagenized FRT40A chromosome. Six thousand males selected in this way were then individually crossed to hs-hid/CyO ftz-LacZ females in two successive screens, a pilot screen and a full-scale screen. Again, a heat-shock treatment of the progeny of this cross allowed the establishment of 4779 heterozygous, balanced mutant lines (500 in the pilot screen). The loss of lines occurred at two steps of the procedure: 16% of single crosses were sterile and 6% did not survive the heat shock.

Screen results:

To screen for defects in MARCM mutant clones, 10 males of each line were crossed en masse to 50 FRT40A MARCM females. Clones were induced in the progeny as described above, and the homozygous mutant clonal cells in third instar larvae were examined for their number, size, shape, and distribution. Also the appearance of the lumen within clonal cells was taken into account.

In 93% (4452) of the lines we observed the wild-type phenotype, whereas in 343 we detected defects in one or more of the analyzed criteria. We classified 339 of these mutant lines into five distinct phenotypic classes (Table 1). Five mutations caused phenotypes that did not fall into any of these classes. The phenotypes are discussed below.

View this table: In this window In a new window   TABLE 1 Phenotypic classes of mutant clones

 

Lethality and complementation:

To estimate the number of genes identified and the level of saturation of the screen we performed complementation tests. We could not use the tracheal phenotype in MARCM clones for complementation analysis, as it is impossible to create trans-heterozygous clones. Hence we tested complementation using lethality. First we analyzed the lethality of the mutant lines from the phenotypic classes B–E and found 140 chromosomes (78%) to be homozygous lethal. We tested the lethal lines within each of the phenotypic classes B–E for complementation (a limited complementation analysis within group A is described below). Although it was not a priori clear in each case whether the lethal mutations were necessarily the same as the mutations responsible for the phenotypes, this analysis revealed a large number of complementation groups. The chance is not high that two independent chromosomes each with a mutation resulting in the same tracheal phenotype also have additional lethal mutations that happen not to complement each other. This suggests that at least in these cases the lethality and the morphological phenotypes are due to the same mutation. The lethality can therefore be used for further analysis and mapping.

The distribution of the number of alleles per complementation group varied between the four phenotypic classes tested (Table 2, Figure 2A). In each class, most genes are represented by a single allele only. On the basis of allele frequencies and the recovery rates of new loci in the course of the screen (Figure 2B) we conclude that our screen did not reach saturation.

View this table: In this window In a new window   TABLE 2 Mutant lines are listed in phenotypic classes

 

View larger version (22K): In this window In a new window Download PPT slide   FIGURE 2.— Allele frequencies and recovery curve. (A) Distribution of allele frequencies for loci from different phenotypic classes. (B) The numbers of isolated mutants and recovered new loci were plotted against the total number of analyzed chromosomes. The curve of identified loci has not leveled, indicating that saturation has not yet been reached.

 

Phenotypic classes:

Class A:

Class A consists of 163 mutant lines in which no GFP-positive clonal cells were found either in the tracheal system or in the epidermis. Sixteen lines that were found in the pilot screen were analyzed in more detail. They were first reanalyzed for their phenotype using larger numbers of larvae, to test whether larvae with clones were missed for stochastic reasons. Indeed, 3 lines turned out in this rescreen to be able to produce a few GFP-positive clones. For the remaining 13 lines no positively marked mutant clones could be identified.

To test whether the genes mutated in these 13 lines were also required at early stages of tracheal development, we analyzed the phenotypes of embryos that were homozygous for the mutated chromosomes. In most cases we found no specific defects, indicating either that the gene products were not required during embryogenesis or that they were maternally supplied in sufficient amounts. One line, 2L0058, showed a convoluted dorsal trunk. We therefore tested it for complementation of mutations in known genes on the second chromosome that show this mutant phenotype [convoluted, varicose, nervana2 (BEITEL and KRASNOW 2000), and mummy (cystic) (ARAUJO et al. 2005; TONNING et al. 2006)] but found it not to be allelic to any of them. However, the observation that a line from phenotypic class A had a convoluted dorsal trunk suggested that other lines from this class might also have such a phenotype and be allelic to the genes mentioned above. We therefore performed complementation tests against the whole set of class A mutants (both from the pilot and the main screen) and found that it contained two new alleles of varicose and five further mutations that were allelic to 2L0058. Both of the two new varicose alleles also showed the typical embryonic phenotype of a convoluted dorsal trunk, but the other lines that were allelic to 2L0058 unexpectedly did not (data not shown).

We can envisage two reasons for the lack of mutant clones for the 13 mutants. Either no recombination had occurred (for example, because the FRT site had been damaged) or mutant clones were unable to develop and survive in the tracheal system up to the third larval instar. To distinguish between these possibilities, and to test whether clonal lethality was specific for the tracheal system, we generated mutant clones in the wing imaginal disc of 10 of the lines. In this case, we did not use MARCM clones, but used standard clonal analysis in which the mutant clone is marked by loss of GFP, and the wild-type sister clone can be identified by the high level of GFP. We observed three different outcomes. For one mutant line we found no recombinant clones at all: neither mutant, GFP-negative clones nor homozygous wild-type clones marked by high levels of GFP in the heterozygous epithelium were found. This indicates that no recombination had taken place, presumably because the FRT site was damaged. Imaginal discs from 7 mutant lines showed clones expressing high levels of GFP, but no adjacent clones lacking GFP, indicating that recombination had taken place, but the homozygous mutant cell had not developed into a clone. These genes must therefore be regarded as cell lethal. One line produced mutant clones that were much smaller than the sister clone, and one line produced clones of the same size as the sister clones, showing that cell lethality was not general, but was restricted to tracheal cells. This analysis indicates that the majority of genes identified in this class appear to be genes required for general cell survival and we therefore did not analyze the remaining 147 lines of this class.

The observation that varicose mutant clonal cells failed to contribute to the larval tracheal system was unexpected, because varicose is not essential for the formation of the tracheal system as such (BEITEL and KRASNOW 2000; WU et al. 2007). This suggests that mutant cells might be unable to remain in the tracheal epithelium during the restructuring and growth of the larval stages. However, the Varicose protein is a member of a group of basolateral proteins required for septate junctions and tracheal morphogenesis (WU et al. 2007), and clones with mutations in other septate junction proteins (cor, Nrg, Nrv2) have previously been shown to be able to survive and remain in the epithelium of wing imaginal discs (GENOVA and FEHON 2003). Thus these mutations are not cell lethal and also do not interfere with the cells remaining in epithelia per se. This raises the question whether varicose differs in principle from the other septate junction genes or whether the requirements for septate junction proteins are different in tracheal and imaginal disc cells. We tested this by a MARCM clonal analysis of the new varicose alleles (2L1623 and 2L3215) in the wing imaginal disc and by creating MARCM clones mutant for nrv2 (nrv2^k13315) (GENOVA and FEHON 2003) in the tracheal system.

We found no varicose mutant clones in wing imaginal discs, indicating that the failure of these cells to contribute to the tracheal system was not specific, but that varicose mutant cells are generally unable to remain in epithelia up to the third larval instar. In parallel, we induced nervana2 mutant clones in the trachea. In 77 larvae we analyzed we found only 5 with clones. Thus, although not essential for the ability of cells to be maintained in the epithelium of the disc, nervana2 makes a significant contribution. This is consistent with the observation that clones for mutations in SJ components can be recovered only if induced late during development and that they are usually very small (R. FEHON, personal communication).

Class B:

Class B contains 20 mutations resulting in a low number of clonal cells. We separated them into two phenotypic subclasses, one with the number of clonal cells varying from 10 to 30 per larva (10 lines, subclass B0) and one with <10 clonal cells per larva (10 lines, subclass BI). Thirteen of the mutations are lethal, defining 12 complementation groups (one containing an allele from class CII).

Class C:

Class C includes 47 mutants that showed small clonal cells. Here we also identified more than one type of defect. In 7 lines the size reduction was very strong (subclass CI) and in 20 lines the small size of the clonal cells was combined with their low number (subclass CII). In addition to reduced size, the clonal cells often had irregular shapes, especially in the dorsal trunk. Forty-one of the mutant chromosomes are homozygous lethal. In complementation tests 14 were found to be single alleles, while the remaining 27 each failed to complement one or two other mutations. However, they did not define clear complementation groups. Ten mutations show noncomplementation with more than one complementation group, but in each case with only one allele from a given complementation group and not the others (Table 2). This complementation behavior might be due to interactions with background mutations on the original chromosome (which, by themselves, cannot be lethal, as the starting chromosome was homozygous viable). Alternatively, it is possible that the mutations were indeed in the same gene, but that they affected different domains or differentially spliced exons in such a way that the combination of two mutant isoforms was nevertheless able to provide wild-type function; i.e., the isoforms were able to complement each other. In this case, each isoform should be lethal over a complete null allele, for example, a deficiency. We therefore set up crosses against deficiencies to map the lethality of a subset of the mutants. We found that the chromosome carrying the 2L1663 allele was lethal over Df(2L)BSC4. We then tested 2L2572, 2L1554, and 2L526 against this deficiency and found that all three were complementing, indicating that at least one lethal hit on the 2L1663 chromosome maps to a different locus that the lethal hits on the other chromosomes. We also found one deficiency [Df(2L)Exel7014] that failed to complement 2L2572. Again, when crossed to two other "alleles" of 2L2572 (2L1663 and 2L1554), this deficiency was able to complement them. Thus, either the mutations responsible for the complementation behavior between the alleles are not the same as those uncovered by the deficiency or there are multiple mutations on each chromosome that interact with each other.

Class D:

Mutant clones in class D showed defects in the dorsal trunk. The 28 lines from this group were classified into five subclasses. Class D0 consists of 6 lines with diverse defects not observed in any other subgroup: 2L0128, clonal cells in DT larger than wild type; 2L0196, DT narrowed within clonal cell; 2L0372, smaller lumen and irregular shape of clonal cells in DT; 2L3090, rounded and smaller clonal cells in DT and secondary branches (Figure 3D); 2L3191, irregular lumen in DT at clonal position; 2L4771, bends in fusion cells of DT if clonal, no clonal cells in terminal branches, and no clonal cells in other fusion cells.

Class DI includes 10 lines exhibiting bends of the dorsal trunk at the position of each mutant cell (Figure 3B) whereas other branches were unaffected. Eight of the mutations in this class form a single complementation group, group D1, which we analyzed in more detail. In the mutants from this group, clonal cells in the dorsal trunk were smaller than wild-type clonal cells and their shape was irregular. The total number of clonal cells as well as the number of clonal cells in the dorsal trunk was not affected in mutant larvae (supplemental Table S1 at http://www.genetics.org/supplemental/). The number of clonal cells showing the defects ranged from 55.87% (line 2L3696) to 80% (line 2L4333; supplemental Table S1).

When larvae were stained with antibody against armadillo/beta-catenin, the difference in size between the mutant cells in the dorsal trunks and their surrounding wild-type neighbors was clearly visible (Figure 4F). In addition, the clonal cells had lost the hexagonal shape typical for dorsal trunk cells. No obvious defects in subcellular architecture were seen and armadillo was localized apically, as in the neighboring wild-type cells (Figure 4).

View larger version (48K): In this window In a new window Download PPT slide   FIGURE 4.— Complementation group D1 cell shape and junctions. Staining against armadillo of the dorsal trunk with wild-type (A–C) and mutant clonal cells (D–I). Mutant clonal cells, marked with GFP, are misshapen and much smaller than neighboring wild-type cells. Phenotypic classes are indicated next to allele names. Bars, 50 µm.

 
We examined the early tracheal development in homozygous mutant embryos of all eight alleles. Embryos died during late embryogenesis. Staining for the luminal marker 2A12 revealed that all but one line showed wild-type tracheae. In line 2L0439 breaks in the dorsal trunk were observed and the lumen in the dorsal trunk stayed discontinuous throughout development (data not shown). Although this phenotype is potentially interesting, the presence of defects in the embryonic tracheae in only one line suggests that it is caused by an additional mutation on the second chromosome in line 2L0439 and that the phenotype is not related to the defects observed in clones in the larva. In conclusion, the product of the gene mutated in these lines either is not required during embryogenesis or is supplied maternally in sufficient amounts to allow normal tracheal development.

In the four lines in class DII clonal cells were found exclusively in the dorsal trunk. This might indicate that the mutations interfere with cell migration into the branches of the network or that the mutant cells cannot be maintained within the epithelium of the branches.

Class DIII comprises five lines with small clonal cells in the dorsal trunk (Figure 3F). Clonal cells in other branches were indistinguishable from wild-type clones.

Class DIV was characterized by a low number or a complete lack of clonal cells in the dorsal trunk and consists of three lines.

Complementation tests with the homozygous lethal lines from phenotypic class D revealed that in addition to group D1 there were two more complementation groups with more than one allele, one containing three alleles that had initially been assigned to the different phenotypic subclasses (see Table 2).

Class E:

The 81 mutants in class E showed defects only in terminal extensions and fell into six phenotypic subclasses. Only three of the lethal hits in this class were represented by more than one allele (see Table 2).

Class E0 included 10 lines with complex defects: abnormal formation of terminal extensions (2L1668, 2L1923, 2L2806, 2L3047, 2L3193, 2L3393; Figure 5B), larger diameter of terminal extensions (2L2084, 2L2944, 2L3807; not shown), smaller diameter of terminal extensions (2L4105, 2L4515; not shown), abnormal ratio of clonal cells in terminal branches and other types of branches (2L4155, 2L3807; not shown), fewer branches in terminal cells (2L4155, 2L2944), and no lumen in terminal extensions (2L4155).

View larger version (66K): In this window In a new window Download PPT slide   FIGURE 5.— Phenotypes of class E mutant clones. Four examples of terminal branching defects are shown. (A) Wild-type dorsal terminal clonal cells. (B) In line 2L1668 (subclass E0) the nucleus of the terminal cell is wrongly positioned. Rather than being at the center of the cell from which all branches arise, the nucleus appears to have moved into one of the branches. (C and D) In the wild type, branches of neighboring terminal cells respect a virtual boundary between their territories (C). In the mutant clone of line 2L1629 (subclass EI) such a restriction is not observed (D). (E) In subclass EII terminal clonal cells had a reduced number of branches. (F) Mutant clones from subclass EIV showed more extensive branching than wild-type terminal cells at similar positions. (G and H) In mutants from subclass EIII no gas-filled lumen can be seen in bright-field view at the position of the clonal cell (dashed line). All cells except for C and D are dorsal terminal cells (C and D are lateral terminal cells). Phenotypic classes are indicated next to allele names. Bars, 50 µm.

 
Class EI was characterized by "terminal branch crossing" (Figure 5D). Two types of this defect could be distinguished: crossing of sister branches of a single terminal cell or crossing of branches of two different cells. As shown in Figure 6C, in the wild type the branches of neighboring terminal cells, even if growing into the same region of a tissue, never access the same or overlapping target territories and thus never cross each other (GHABRIAL et al. 2003). In 24 mutant lines such a restriction was not observed. Also the phenomenon of "crossing" or fusion between branches of single terminal cells found in these mutants (not shown) did not occur in wild-type larvae. However, mutant dorsal terminal cells never crossed the dorsal midline.

View larger version (106K): In this window In a new window Download PPT slide   FIGURE 6.— Phenotype of complementation group E2: terminal cells. (A and B) In wild-type clonal cells the lumen of the cell can be seen in bright-field image (gray branches in B). Cell area is outlined with a dashed line. (C–F) In mutant clonal cells no lumen could be seen (cell area outlined with a dashed line) whereas the lumen of neighboring nonclonal cells is readily discernable (arrowheads). Phenotypic classes are indicated next to allele names Bar, 50 µm.

 
Class EII comprises 11 lines in which mutant terminal cells showed reduced branching (Figure 5E). The strength of this phenotype ranged from terminal cells branching into a tree with a reduced branch length to a reduced number or the lack of branches.

Class EIII included 18 lines showing problems with the formation of the lumen within the terminal extensions. The strongest phenotype in this group was a complete abolition of lumen formation within the branch as depicted in Figure 5, G and H (6 lines). Also milder phenotypes with partially formed lumen were found (10 lines). In two cases we observed properly formed but mispositioned lumen. In the wild type, the lumen was placed centrally in the cell whereas in these lines it was on the side. Most of the lines also had a reduced number and length of branches.

Two lines (2L3637 and 2L4501) lacked a lumen not only in terminal cells but also in clonal cells in secondary branches. These two lines were allelic, defining one of the three complementation groups in class E that were represented by two alleles (group E2). We analyzed these lines in more detail.

The mutant terminal cells lacked a lumen either throughout the cell or in major parts of the cell (Figure 6, C–F, supplemental Table S2 at http://www.genetics.org/supplemental/). Also in clonal cells on secondary branches gaps in gas filling were found to span either part of or the whole cell (Figure 7, E–L, supplemental Table S3). Clonal cells in the dorsal trunk and transverse connectives had no defects.

View larger version (74K): In this window In a new window Download PPT slide   FIGURE 7.— Phenotype of complementation group E2: secondary branches. (A–D) In wild-type clonal cells the lumen is seen throughout the length of the cell both in bright-field and in fluorescence images. (E–L) Mutant clones show gaps in the lumen over part of their length (brackets). Myristoylated RFP was expressed in wild-type (J) and group E2 mutant (F and J) clonal cells. In wild type the signal is excluded from the lumen whereas in the mutant within the gaps the signal of myristoylated RFP was diffused, with no clear membrane localization. Phenotypic class is indicated next to allele names. Bar, 10 µm.

 
An absence of lumen could arise because the lumen is never formed, or because it collapses after having formed, or the lumen could appear to be missing because the liquid originally present in it was never replaced by gas. The latter possibility is ruled out in these mutants by the distribution of the GFP expressed in the mutant cells. If the lumen was present but filled with liquid rather than gas, then the lumen might not be visible by light microscopy, but GFP should still be seen only in the cytoplasm surrounding the lumen and not throughout the whole width of the tracheal branch. However, the GFP we used in our screen, as well as a myristoylated RFP that is enriched (but not restricted to) cell membranes, is not excluded from a central area. Thus the lumen in these cells was never formed properly or has collapsed.

To distinguish between these possibilities we analyzed homozygous mutant embryos from both lines. These embryos did not show gaps in the tracheal system when stained with the luminal marker 2A12 (data not shown). Thus, the gaps in the larval clonal cells in the secondary branches are most likely due to a collapse at later stages.

Class EIV consisted of four lines showing excessive branching (Figure 5F). In all lines, terminal cells were more strongly ramified than wild-type cells at similar positions.

Class EV was characterized by the absence or a reduced number of clones in terminal cells in comparison to the number of clones in other branches and compared to clones induced using a nonmutated FRT chromosome control (not shown). This subclass contains 13 lines.

Classes F–I:

The remaining four phenotypic classes were each defined by only one or two mutations showing a phenotype that did not fall into any of the other phenotypic classes.

In class F (line 2L0419) the cells that will develop into the adult tracheal system, the tracheal histoblasts, were smaller than in wild-type clones. The line is homozygous lethal.

Class G (line 2L0445) mutant clones in the dorsal trunk and secondary branches (e.g., transverse connective) were abnormally shaped and too small. The line is homozygous viable.

Class H contains two lines (2L1281, 2L1296) in which no clonal cells were seen in the epidermis. Line 2L1281 is homozygous lethal, and 2L1296 is viable.

In class I (line 2L3574) clonal cells were present only in the epidermis but not in the trachea. The line is homozygous lethal.

The development of the Drosophila tracheal system is a process based on cell migration and extensive cell rearrangements. Much insight into its genetic control and cellular processes has been gained through genetic screens (SAMAKOVLIS et al. 1996; BEITEL and KRASNOW 2000; HEMPHALA et al. 2003; CELA and LLIMARGAS 2006), the use of live fluorescent imaging techniques (RIBEIRO et al. 2002, 2004; JAZWINSKA et al. 2003; KATO et al. 2004), and detailed analysis of known components of different signaling pathways. Since most of the studies have concentrated on tracheal development in the embryo, not much is known about later stages, especially the molecular basis of terminal branching events that occur in response to oxygen needs of the surrounding tissue. The screen presented here allowed us to identify a series of mutations with phenotypes in the larval tracheal system that are apparently largely distinct from previously known genes: with the exception of line 2L0058 and varicose we identified only genes that did not have an essential function during embryonic tracheal morphogenesis.

A number of classes of genes and phenotypes one might have expected to recover were in fact not found. For example, we did not find any mutants with phenotypes resembling those caused by loss of components of the integrin signaling pathway. Recent studies identified tendrils/rhea, the Drosophila talin to play a role in the maintenance of the lumen in terminal cells (LEVI et al. 2006). Although a significant number of genes that participate in integrin signaling map on chromosome arm 2L [such as certain laminins, the integrin-v chain, and intracellular adapters as well as other as yet unidentified genes found in screens for integrin signaling components (WALSH and BROWN 1998)] none of the mutant phenotypes from our screen is similar to that of talin or ß-integrin (myospheroid) and -integrins (mew, inflated) (LEVI et al. 2006). It remains to be tested whether other components result in different phenotypes and whether mutations in subclass EIII (problems with lumen formation) may be in some of the genes mentioned above.

We would have expected that clones with mutations in genes known to play a role in cell shape changes or cell migration, such as DRhoGEF2, Dreadlock, PAK-kinase, or Capulet, would not behave normally in the tracheal system. However, we observed no defects in these clones. Similarly, the expression of dominant-negative forms of Rac1 and Cdc42 in the embryo did not cause major tracheal defects (LEE and KOLODZIEJ 2002). Nevertheless, these results were surprising, especially as RhoA has been shown to be needed in tracheal fusion (LEE and KOLODZIEJ 2002) and Rac to be involved in the maintenance of epithelial architecture during embryonic tracheal development (CHIHARA et al. 2003). Perhaps the maternal dose in the embryo is sufficient to carry the clones through early development, and once a stable tracheal epithelial network has formed these GTPases become dispensable.

Finally, there is little overlap in the identification of mutations affecting formation of the air sac primordium (see accompanying article by CHANUT-DELALANDE et al. 2007). A total of 1084 of the 4435 lines that showed no defects in the larval tracheae were screened; 44 were found to have defects in the air sac. Thus, these genes required for air sac development were not needed for tracheal development. To test whether the converse was also true, a random subset of 39 mutants from phenotypic classes B–E was analyzed. Thirty-six lines showed no defects, and the remaining 3 had very weak phenotypes, showing that these 39 genes had no important function in air sac development. Thus, the two processes under investigation require largely different sets of molecules.

It is worth noting that the presence of cells that are too small or abnormally shaped does not necessarily lead to defective tracheal morphology. Most of the "small cell" clones are able to form part of tracheal branches or dorsal trunks in which the overall shape is not affected. Thus, in these cases the neighboring cells simply accommodate the mutant cells. By contrast, mutations in the D1 complementation group, which also cause small irregular cells, have an additional morphological effect on the shape of the dorsal trunk. This indicates either that a degree of coordination between tracheal cells is necessary to form correctly shaped tubes and the gene mutated in these lines participates in this process or that the gene acts cell autonomously, perhaps in regulating physical properties of the cell that do not allow it to contribute to the correct curvature, stiffness, or other parameters of tube architecture.

While the new mutant phenotypes we have found already indicate the existence of as yet unknown morphogenetic mechanisms, it is obvious that the full potential of the mutants can be exploited only once the molecular identity and cell biological functions of the genes have been established.

We thank R. Fohen, G. Beitel, C. Samakovlis, R. Schuh, and the Bloomigton Stock Center for fly stocks; S. Luschnig for fly stocks and discussions; K. van Mark and P. Radermacher for technical support during the screen; and K. Johnson and J. Nair for critical comments on the manuscript. This work was supported by the Deutsche Forschungsgemeinschaft (LE546/3-1; LE546/3-2). M.M.B. had a fellowship from the Cologne International Graduate School in Genetics and Functional Genomics.

¹ These authors contributed equally to this work.

ARAUJO, S. J., H. ASLAM, G. TEAR and J. CASANOVA, 2005 mummy/cystic encodes an enzyme required for chitin and glycan synthesis, involved in trachea, embryonic cuticle and CNS development—analysis of its role in Drosophila tracheal morphogenesis. Dev. Biol. 288: 179–193.[CrossRef][Medline]

BARRETT, K., M. LEPTIN and J. SETTLEMAN, 1997 The Rho GTPase and a putative RhoGEF mediate a signaling pathway for the cell shape changes in Drosophila gastrulation. Cell 91: 905–915.[CrossRef][Medline]

BAUM, B., W. LI and N. PERRIMON, 2000 A cyclase-associated protein regulates actin and cell polarity during Drosophila oogenesis and in yeast. Curr. Biol. 10: 964–973.[CrossRef][Medline]

BEITEL, G. J., and M. A. KRASNOW, 2000 Genetic control of epithelial tube size in the Drosophila tracheal system. Development 127: 3271–3282.[Abstract]

BERGER, J., T. SUZUKI, K. A. SENTI, J. STUBBS, G. SCHAFFNER et al., 2001 Genetic mapping with SNP markers in Drosophila. Nat. Genet. 29: 475–481.[CrossRef][Medline]

CABERNARD, C., and M. AFFOLTER, 2005 Distinct roles for two receptor tyrosine kinases in epithelial branching morphogenesis in Drosophila. Dev. Cell 9: 831–842.[CrossRef][Medline]

CELA, C., and M. LLIMARGAS, 2006 Egfr is essential for maintaining epithelial integrity during tracheal remodelling in Drosophila. Development 133: 3115–3125.[Abstract/Free Full Text]

CHANUT-DELALANDE, H., A. C. JUNG, L. LIN, M. M. BAER, A. BILSTEIN et al., 2007 A genetic mosaic analysis with a repressible cell marker screen to identify genes involved in tracheal cell migration during Drosophila air sac morphogenesis. Genetics 176: 2177–2187.[Abstract/Free Full Text]

CHIHARA, T., K. KATO, M. TANIGUCHI, J. NG and S. HAYASHI, 2003 Rac promotes epithelial cell rearrangement during tracheal tubulogenesis in Drosophila. Development 130: 1419–1428.[Abstract/Free Full Text]

GENOVA, J. L., and R. G. FEHON, 2003 Neuroglian, Gliotactin, and the Na+/K+ ATPase are essential for septate junction function in Drosophila. J. Cell Biol. 161: 979–989.[Abstract/Free Full Text]

GHABRIAL, A. S., and M. A. KRASNOW, 2006 Social interactions among epithelial cells during tracheal branching morphogenesis. Nature 441: 746–749.[CrossRef][Medline]

GHABRIAL, A., S. LUSCHNIG, M. M. METZSTEIN and M. A. KRASNOW, 2003 Branching morphogenesis of the Drosophila tracheal system. Annu. Rev. Cell Dev. Biol. 19: 623–647.[CrossRef][Medline]

GRETHER, M. E., J. M. ABRAMS, J. AGAPITE, K. WHITE and H. STELLER, 1995 The head involution defective gene of Drosophila melanogaster functions in programmed cell death. Genes Dev. 9: 1694–1708.[Abstract/Free Full Text]

GRIGLIATTI, T., 1998 Mutagenesis, pp. 58–60 in Drosophila: A Practical Approach, edited by D. B. ROBERTS. Oxford University Press, Oxford.

HEMPHALA, J., A. UV, R. CANTERA, S. BRAY and C. SAMAKOVLIS, 2003 Grainy head controls apical membrane growth and tube elongation in response to Branchless/FGF signalling. Development 130: 249–258.[Abstract/Free Full Text]

IMAM, F., D. SUTHERLAND, W. HUANG and M. A. KRASNOW, 1999 stumps, a Drosophila gene required for fibroblast growth factor (FGF)-directed migrations of tracheal and mesodermal cells. Genetics 152: 307–318.[Abstract/Free Full Text]

JARECKI, J., E. JOHNSON and M. A. KRASNOW, 1999 Oxygen regulation of airway branching in Drosophila is mediated by branchless FGF. Cell 99: 211–220.[CrossRef][Medline]

JAZWINSKA, A., C. RIBEIRO and M. AFFOLTER, 2003 Epithelial tube morphogenesis during Drosophila tracheal development requires Piopio, a luminal ZP protein. Nat. Cell Biol. 5: 895–901.[CrossRef][Medline]

KATO, K., T. CHIHARA and S. HAYASHI, 2004 Hedgehog and Decapentaplegic instruct polarized growth of cell extensions in the Drosophila trachea. Development 131: 5253–5261.[Abstract/Free Full Text]

KERMAN, B. E., A. M. CHESHIRE and D. J. ANDREW, 2006 From fate to function: the Drosophila trachea and salivary gland as models for tubulogenesis. Differentiation 74: 326–348.[CrossRef][Medline]

KNOBLICH, J. A., and C. F. LEHNER, 1993 Synergistic action of Drosophila cyclins A and B during the G2-M transition. EMBO J. 12: 65–74.[Medline]

LEE, S., and P. A. KOLODZIEJ, 2002 The plakin Short Stop and the RhoA GTPase are required for E-cadherin-dependent apical surface remodeling during tracheal tube fusion. Development 129: 1509–1520.[Medline]

LEE, T., and L. LUO, 1999 Mosaic analysis with a repressible cell marker for studies of gene function in neuronal morphogenesis. Neuron 22: 451–461.[CrossRef][Medline]

LEE, T., and L. LUO, 2001 Mosaic analysis with a repressible cell marker (MARCM) for Drosophila neural development. Trends Neurosci. 24: 251–254.[CrossRef][Medline]

LEVI, B. P., A. S. GHABRIAL and M. A. KRASNOW, 2006 Drosophila talin and integrin genes are required for maintenance of tracheal terminal branches and luminal organization. Development 133: 2383–2393.[Abstract/Free Full Text]

MANNING, G., and M. A. KRASNOW, 1993 Development of the Drosophila tracheal system, pp. 609–686 in The Development of Drosophila melanogaster, edited by M. A. A. BATE. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

MICHELSON, A. M., S. GISSELBRECHT, E. BUFF and J. B. SKEATH, 1998 Heartbroken is a specific downstream mediator of FGF receptor signalling in Drosophila. Development 125: 4379–4389.[Abstract]

NEWSOME, T. P., S. SCHMIDT, G. DIETZL, K. KELEMAN, B. ASLING et al., 2000 Trio combines with dock to regulate Pak activity during photoreceptor axon pathfinding in Drosophila. Cell 101: 283–294.[CrossRef][Medline]

PARKS, A. L., K. R. COOK, M. BELVIN, N. A. DOMPE, R. FAWCETT et al., 2004 Systematic generation of high-resolution deletion coverage of the Drosophila melanogaster genome. Nat. Genet. 36: 288–292.[CrossRef][Medline]

RIBEIRO, C., A. EBNER and M. AFFOLTER, 2002 In vivo imaging reveals different cellular functions for FGF and Dpp signaling in tracheal branching morphogenesis. Dev. Cell 2: 677–683.[CrossRef][Medline]

RIBEIRO, C., M. NEUMANN and M. AFFOLTER, 2004 Genetic control of cell intercalation during tracheal morphogenesis in Drosophila. Curr. Biol. 14: 2197–2207.[CrossRef][Medline]

RIGGLEMAN, B., P. SCHEDL and E. WIESCHAUS, 1990 Spatial expression of the Drosophila segment polarity gene armadillo is posttranscriptionally regulated by wingless. Cell 63: 549–560.[CrossRef][Medline]

SAMAKOVLIS, C., N. HACOHEN, G. MANNING, D. C. SUTHERLAND, K. GUILLEMIN et al., 1996 Development of the Drosophila tracheal system occurs by a series of morphologically distinct but genetically coupled branching events. Development 122: 1395–1407.[Abstract]

SUTHERLAND, D., C. SAMAKOVLIS and M. A. KRASNOW, 1996 branchless encodes a Drosophila FGF homolog that controls tracheal cell migration and the pattern of branching. Cell 87: 1091–1101.[CrossRef][Medline]

TONNING, A., S. HELMS, H. SCHWARZ, A. E. UV and B. MOUSSIAN, 2006 Hormonal regulation of mummy is needed for apical extracellular matrix formation and epithelial morphogenesis in Drosophila. Development 133: 331–341.[Abstract/Free Full Text]

UV, A., R. CANTERA and C. SAMAKOVLIS, 2003 Drosophila tracheal morphogenesis: intricate cellular solutions to basic plumbing problems. Trends Cell Biol. 13: 301–309.[CrossRef][Medline]

VINCENT, S., R. WILSON, C. COELHO, M. AFFOLTER and M. LEPTIN, 1998 The Drosophila protein Dof is specifically required for FGF signaling. Mol. Cell 2: 515–525.[CrossRef][Medline]

WALSH, E. P., and N. H. BROWN, 1998 A screen to identify Drosophila genes required for integrin-mediated adhesion. Genetics 150: 791–805.[Abstract/Free Full Text]

WU, V. M., M. H. YU, R. PAIK, S. BANERJEE, Z. LIANG et al., 2007 Drosophila Varicose, a member of a new subgroup of basolateral MAGUKs, is required for septate junctions and tracheal morphogenesis. Development 134: 999–1009.[Abstract/Free Full Text]

Communicating editor: K. V. ANDERSON

Related articles in Genetics:

ISSUE HIGHLIGHTS

Genetics 2007 176: NP. [Full Text]  

This article has been cited by other articles:

				M. Affolter and E. Caussinus Tracheal branching morphogenesis in Drosophila: new insights into cell behaviour and organ architecture Development, June 15, 2008; 135(12): 2055 - 2064. [Abstract] [Full Text] [PDF]

				H. Chanut-Delalande, A. C. Jung, L. Lin, M. M. Baer, A. Bilstein, C. Cabernard, M. Leptin, and M. Affolter A Genetic Mosaic Analysis With a Repressible Cell Marker Screen to Identify Genes Involved in Tracheal Cell Migration During Drosophila Air Sac Morphogenesis Genetics, August 1, 2007; 176(4): 2177 - 2187. [Abstract] [Full Text] [PDF]

International Access Link

Online ISS 1091-6490
Copyright 2008 by the Genetics Society of America
phone: 412-268-1812   fax: 412-268-1813   email: genetics-gsa{at}andrew.cmu.edu