Stereotypic founder cell patterning and embryonic muscle formation in Drosophila require nautilus (MyoD) gene function

Abstract

nautilus is the only MyoD-related gene in Drosophila. Nautilus expression begins around stage 9 at full germ-band extension in a subset of mesodermal cells organized in a stereotypic pattern in each hemisegment. The muscle founder cell marker Duf-LacZ, produced by the enhancer trap line rP298LacZ, is coexpressed in numerous Nautilus-positive cells when founders first appear. Founders entrain muscle identity through the restricted expression of transcription factors such as S59, eve, and Kr, all of which are observed in subsets of the nautilus expressing founders. We inactivated the nautilus gene using homology-directed gene targeting and Gal4/UAS regulated RNAi to determine whether loss of nautilus gene activity affected founder cell function. Both methods produced a range of defects that included embryonic muscle disruption, reduced viability and female sterility, which could be rescued by hsp70-nautilus cDNA transgenes. Our results demonstrate Nautilus expression marks early founders that give rise to diverse muscle groups in the embryo, and that nautilus gene activity is required to seed the correct founder myoblast pattern that prefigures the muscle fiber arrangement during embryonic development.

Each hemisegment of the Drosophila embryo contains 30 syncitial muscle fibers with 3–25 nuclei, which are organized in a stereotypic pattern with specified shapes, ectodermal attachment points and innervation patterns. Michael Bate and Manfred Frasch and coworkers (1, 2) were the first to propose a cellular basis for the pattern of muscle organization in Drosophila based on the identification of a special class of myoblast, called the founder cell. Founders arise at specific locations in the somatic mesoderm. Each larval muscle is prefigured by a single founder cell that seeds muscle formation and entrains muscle identity after rapidly fusing with the surrounding fusion competent myoblasts. Work by several groups has now provided an outline for the early events in myogenesis (reviewed in refs. 3–7). Two types of myoblasts are determined in the early mesoderm from clusters of cells that define a myogenic field marked by high levels of Twist expression. Specific cells within the myogenic field express Lethal of scute to establish myogenic equivalence groups, which are further delineated through lateral inhibition and the activation of the Notch-Delta signaling pathway. High levels of Delta signaling from the future founder progenitors activate the Notch pathway in surrounding cells to generate the fusion-competent myoblasts that cannot fuse or express myosin in the absence of founder myoblasts. Founder progenitors divide asymmetrically in response to Numb gene activity to generate a pair of founder cells that each entrain the position and identity of one of the ≈30 muscles in each hemisegment.

The myogenic role of the Drosophila MyoD homolog, nautilus, is debated, even though ectopic expression can induce myogenesis in Drosophila, and mouse fibroblasts and a nautilus transgene can rescue the Caenorhabditis elegans MyoD (hlh-1) null (8, 9). Nautilus is expressed in a stereotypic pattern in a subset of ≈30 mesodermal cells per hemisegment that are incorporated into every somatic muscle in the embryo (10, 11). These cells are required for muscle formation because their toxin ablation abrogates myogenesis in the embryo (12). Previous studies by Abmayr and associates, using ethane methyl sulfonate (EMS) nautilus-null alleles, concluded nautilus was not essential for myogenesis or viability and was responsible only for the specification of a minor subset embryonic muscles, DA3 and DO4 (13). By contrast, the injection of nautilus dsRNA into embryos resulted in the RNAi-induced loss of nautilus function and the disruption and reduction of muscle formation in the embryo (12).

To study this issue further, we used two independent strategies to disrupt nautilus gene function. First, we targeted the insertion of an armadillo-GFP (armGFP) reporter into the first exon of nautilus by ends-out homologous recombination (14, 15). Second, we designed a Gal-4-inducible RNAi vector to degrade nautilus mRNA during development (16). Both methods result in the loss of nautilus gene function accompanied by severe muscle disruption in up to 30% of the embryos. The absence of nautilus impacts every stage of development and results in reduced viability, decreased mobility in the larval and adult stages as well as female sterility. Rather than specifying a minor group of fibers, Nautilus is shown to mark diverse subsets of founders that express the identity genes for different embryonic muscles. Nautilus gene function is needed to establish the stereotypic founder cell pattern in each hemisegment that prefigures the formation and spatial organization of somatic muscle in the embryo.

Results and Discussion

Targeted Disruption of the nautilus Gene by Homologous Recombination.

An armGFP reporter was inserted into the first exon of the nautilus gene using ends-out gene targeting by homologous recombination (14, 15). One candidate gave the expected XbaI (5.5- and 4.8-kb) and BstZ17 I (5.5- and 3.1-kb) fragments diagnostic for correct targeting when probed with the nautilus 8.2-kb HindIII fragment [supporting information (SI) Fig. 4]. The precision of the targeting event was verified by sequence analysis of PCR fragments that span the junction regions and was determined to be precise (see SI Text). The targeted insertion is subsequently noted as nau^armGFP.

nau^armGFP Is Null for nautilus mRNA and Protein.

To verify that homozygous nau^armGFP flies do not express nautilus mRNA, RT-PCR was performed on total RNA from stage 10–14 embryos, second- and third-instar larvae, and mixed adults, and was compared with nautilus mRNA levels in wild-type embryos and mixed adults. Because the third chromosomal balancer for nau^armGFP is marked with TM6B and a strong ubiquitin-GFP reporter, the weaker armGFP signal was selected to obtain homozygous nau^armGFP embryos. Tb+ larvae and Tb+ pupae were picked to obtain later stages. The homozygous nau^armGFP animals showed no mRNA expression at any stage using a PCR primer pair that spans all three introns and covers the entire coding region (SI Fig. 4). nau^armGFP-null embryos were also stained with Nautilus rabbit polyclonal antibody, and no evidence for the characteristic repetitive segmental Nautilus nuclear expression pattern was observed (SI Fig. 5). By these criteria, the targeted insertion of armGFP into the nautilus gene is a bona fide null.

The nau^armGFP Mutation Affects Embryonic Muscle Formation.

nau^armGFP homozygous embryos were collected and stained with polyclonal antibody to myosin heavy chain raised to exon-17, an exon conserved in all skeletal muscle myosins (17, 18). Approximately one-third of the null embryos (n = 89/319) showed a severely disrupted muscle phenotype reminiscent of embryos injected with nautilus dsRNAs (12). In severe disruptions, the muscles often appeared as rounded balls or thin disorientated fibers, whereas in the less severe cases, various subsets of fibers were missing (see below). The precise fiber loss reported for muscles DA3 and DO4 in the EMS nautilus null alleles was not observed (13).

To confirm that the muscle phenotype was because of a nautilus loss of function and not other mutations on the third chromosome, nau^armGFP was placed over an Exelixis deficiency that removes the nautilus gene along with ≈33 additional genes, none of which are known to be important for myogenesis [Bloomington stock center (no. 7674, Exel 6195, 95A4;95B1)]. nau^armGFPover the deficiency (nau^armGFP/Df) produced the same percentage and range of embryonic muscle phenotypes as the homozygous nau^armGFP mutant (Fig. 1A). The nau^armGFP/Df embryos with severe muscle disruptions still had developed gut structure, as revealed by staining with a monoclonal antibody to position-specific (PS) β-integrin; however, the gut constriction pattern was often abnormal with fewer or misplaced constrictions (SI Fig. 6). Epidermal muscle attachment sites, marked by PS β-integrin, were clearly evident and were located at the ends of the misplaced muscle fibers, supporting the previous proposal that signals from the muscle ends to the epidermis regulate the maturation of these sites (SI Fig. 7) (19). The disrupted fibers also appear innervated or able to attract neurons, based on the coassociation of neurons, as seen with the mAb 22C10 (SI Fig. 7). The expression patterns for Eve and Tinman, which mark cells of the nervous system and the future dorsal vessel, respectively, were also examined and were minimally affected in nautilus null embryos (20, 21) (SI Fig. 6).

Fig. 1.

Disruption in the embryonic muscle pattern in the nau^armGFP mutant, the nau¹⁸⁸ EMS allele, and the +/− Gal-4-induced nau pINT-1 transgene. (A) The embryonic muscle pattern in the homozygous nau^armGFP mutant (nau^armGFP) and in the nau^armGFP mutation over the Exelixis deficiency (nau^armGFP/Df) show the same range of muscle disruptions (nau^armGFP/nau^armGFP, n = 89/319). (B) The EMS allele nau¹⁸⁸ over the Exelixis deficiency or the nau^armGFP mutant has the same muscle disruption phenotype (n = 31/146) as nau^armGFP/Df embryos. The homozygous nau¹⁸⁸ allele is the same (not shown). The nau¹⁸⁸ allele is balanced over Sb LacZ (Upper Left), so loss of the balancer follows loss of LacZ. (C) Induction of nautilus RNAi with pAct5C-Gal4 (+gal4) in the nau-pINT-1 background phenocopies the muscle disruptions seen with nau^armGFP/Df and nau1⁸⁸/Df. Dorsal is the top, and anterior is to the left.

The nau¹⁸⁸ EMS Allele Has the Same Phenotype as the nau^armGFP Mutant.

The EMS allele, nau¹⁸⁸ (kindly supplied by S. Abmayr), was placed over nau^armGFP as well as the Exelixis deficiency. The nau¹⁸⁸ allele is over a balancer marked with Sb^LacZ, so nau¹⁸⁸ can be followed easily by the loss of β-galactosidase expression (Fig. 1B Upper Left). The nau¹⁸⁸ homozygous embryos showed minor to severe muscle disruptions, identical to the nau^armGFP homozygous embryos (data not shown). As with the nau^armGFP mutant, the nau¹⁸⁸ allele was placed over the Exelixis deficiency to rule out effects of possible nonspecific mutations accumulated in the stock. There were no discernable differences in the either the pattern or the degree of muscle disruption between nau¹⁸⁸/nau^armGFP, nau^armGFP/Df, and nau¹⁸⁸/Df; ≈22% of the nau¹⁸⁸/Df embryos showed severe muscle disruption (n = 31/146) (Fig. 1B Upper Right). Less-severe muscle disruptions affecting fewer fibers were also observed in the nau¹⁸⁸ mutant, but the pattern of fiber loss was not specific, as reported previously (Fig. 1B Lower Left) (13). Approximately 50% of the nau¹⁸⁸ embryos did not survive to the pupal stage, as indicated by the ratio of Tb to Tb+ pupae; a 2:1 ratio is expected under normal conditions, but we observed ≈4:1 ratio, similar to the ratio observed with the homozygous nau^armGFP mutant (Table 1).

Table 1.

Viability of nau^armGFP and nau¹⁸⁸ embryos and larvae

Cross	nau−/+ (Tb) pupae	nau−/− (Tb+)pupae	Ratio (Tb/Tb+) (Expect 2:1)	Death, % (embryo + larvae)
nauarmGFP/TM6B × D f(3R)Exe16195/TM6B	347	80	4.3	58
188/TM6B × nauamGFP/TM6B	271	70	3.9	48
188/TM6B × Df(3R)Exel6195/TM6B	375	105	3.6	40

The viability of the nau^armGFP mutant was analyzed at different stages of development to determine when the loss of nautilus had the greatest impact. Most lethality, ≈70%, occurred in the embryo and larval stages, as reflected by a significant reduction in the number of pupae (Table 2). Larvae that survived were weak and moved in uncoordinated fashion, often staying in one position, oscillating back and forth. Slightly more than half of the pupae hatched with extended eclosion times of >12 h compared with a few minutes for normal pupae, whereas the remaining pupae did not open, or flies only partially exited the pupa case. Most of the flies that hatched remained at the bottom of the vial with shriveled wings, unable to move or stand, and eventually died. Temperature affected survival rates, because 64/108 pupae survived to adulthood at 25°C, but only 6/34 survived at 16°C, whereas embryo and larval survival percentages were minimally affected by temperature (Table 2).

Table 2.

Viability of nau^armGFP mutant at different stages of development

Genotype	Total embryos	Larvae (%)	Pupae (%)	Dead pupae (%)	Hatch (survival) (%)
Wild type (w1118)	204	(96 (195)	94 (191)	0	94 (191)
nauamGFP/nauarmGFP	334	70 (235)	32 (108)	13 (44)	19 (64)
nauarmGFP/nauarmGFP(at 16°C)	141	66 (93)	24 (34)	20 (28)	4 (6)
nauarmGFP/Df(3R)	451	67 (302)	30 (134)	4 (20)	25 (114)

Surviving adult nau^armGFP and nau¹⁸⁸ females developed distended abdomens filled with eggs that remained held >20 days, even when they were mated in groups or in single crosses (SI Table 4 and SI Fig. 8). The ovary from nautilus mutant females did not show the normal progression of egg development, and all of the egg chambers were of a similar but smaller size compared with normal mature egg chambers. This is observed in single ovarioles stained with phalloidin, where one can see the smaller, uniform-size egg chambers in the mutant (nau^armGFP), whereas there is a clear progression in egg chamber sizes in the normal ovariole (SI Fig. 8). This likely reflects a reduced function in the musculature of the ovarian sheath but will have to be confirmed in future studies. The “held egg” phenotype was also observed in nau¹⁸⁸/Df and nau^armGFP/nau¹⁸⁸ female populations (SI Table 4). Surviving males are slow-moving and weak, but they are fertile. The loss of nautilus function affects all stages of Drosophila development, particularly somatic myogenesis in the embryo, maturation beyond the third instar larval stage, and female fertility.

A gal4-inducible nau RNAi Transgene Phenocopies the nau^armGFP Mutant.

An independent strategy was developed that did not disrupt the gene but targeted nautilus mRNA using an Gal4-inducible RNAi transgene. We used a modified pUAST vector, called pINT-1 (pUAST intron-1), containing the intron from the 5′ UTR of the actin 5C gene as a spacer for inverted repeats, in this case the nautilus cDNA lacking the initiator codon (SI Fig. 9) (16, 22, 23). In S2 cells transfected with nau-pINT-1 and pACT-nautilus, nuclear Nautilus was evident but with the addition of pACT-Gal4, Nautilus protein was no longer detected (not shown), and silencing correlated with the appearance of nautilus-specific siRNAs (SI Fig. 9).

In the absence of Gal4 induction, the homozygous nau-pINT-1 embryos displayed a normal embryonic muscle pattern (Fig. 1C, −gal). A homozygous nau-pINT-1 line on the third chromosome was crossed to a pACT Gal4/CyO stock to obtain progeny with one copy of the Gal4 inducer and one copy of the nau-pINT-1 RNAi transgene. This stock was then crossed back to the homozygous nau-pINT-1 stock to generate progeny with two copies of nau-pINT-1 and one copy of pACT-Gal4. Slightly >20% of the Gal4 induced embryos had severe muscle disruptions (n = 39/196), identical to those observed with the nau^armGFP or nau¹⁸⁸ mutants (Fig. 1C,+gal 4). In addition, many of the adult females had enlarged abdomens and were unable to lay eggs, a phenocopy of the “held egg” phenotype observed in the nau^armGFP mutant (SI Fig. 8). In the surviving adult population from the first cross, there was also a 2-fold increase in the number of CyO compared with normal flies (P < 0.0001, Fisher's exact test), indicating that approximately half of the induced nau-pINT-1 embryos died at earlier stages, similar to the early death observed in the nau^armGFP mutant (Table 3). Importantly, the induced nau-pINT-1 resulted in a phenotype identical to the nau^armGFP mutant, as well as the previous phenotype observed with the injection of nautilus dsRNAs into the embryo (12).

Table 3.

Viability of nau-pINT-1 embryos

Cross	Genotype	Genotype	Ratio (Expect: Cyo: + = 1:1)
pAct-gal4/Cyo × w; +/+	+/pAct-gal4 (191)	+/Cyo (197)	1:1
pAct-gal4/Cyo; +/+ × +/+; nau-pINT/nau-pINT	pact-gal4/+; nau-pINT/+ (245)	+/Cyo; +/nau-pINT (457)	1:1.9*

*Fisher's exact test or χ² test, P < 0.0001.

hsp-70-nau cDNA Transgene Rescues the nau^armGFP Mutant.

To confirm that the nau^armGFP mutant phenotype was because of the loss of nautilus protein expression, a Casper hsp-70-nau cDNA transgene was used as a rescue construct, because the extent of the fully active nautilus promoter is not known. The nautilus cDNA was used previously to rescue female sterility attributed to nautilus loss of function in strains with overlapping nautilus deficiencies (8, 13). With the rescue transgene and nau^armGFP balanced over a third chromosome marked with TM6B, the ratio of Tb:Tb+ pupae could be followed as a measure of lethality in the embryo and larval stages. A ratio of 2:1 for Tb:Tb+ pupae was expected in the absence of early lethality, but a ratio of 4.6:1was observed, indicating significant death at earlier stages (SI Table 5). The viability of embryos and larvae increased 2-fold with the introduction of heat-shock nautilus transgene, going from 35% to 77% (SI Table 5). Approximately 90% of the rescued mutant females had normal abdomens (41 of 46), and roughly half of these (22 of 41) were fertile and produced normal armGFP-positive progeny when mated to wild-type males (SI Table 6). Embryonic muscle disruption was greatly reduced in the rescued stock (data not shown). Similar results were obtained for hsp-70-nau cDNA on the second chromosome (data not shown). Rescue was achieved in the absence of heat shock, because overexpression of Nautilus was lethal, as reported for overexpression using the twist driver, 24B-GAL4, and UAS-nau cDNA (8). This confirmed that the loss of nautilus gene function in nau^armGFP was responsible for the range of phenotypes affecting myogenesis, viability, and fertility.

Reevaluation of nautilus Gene Function.

Injection of dsRNA representing the entire ORF or subregions of nautilus mRNA into the developing embryo resulted in severe disruption of the embryonic muscle pattern, whereas the injection of dsRNA for β-galactosidase or buffer did not have a similar affect on myogenesis. Control injections of twist, engrailed, and daughterless dsRNA phenocopied the null for each gene (12). A subsequent genetic study of EMS-induced nautilus loss-of-function alleles concluded nautilus had a minimal zygotic role in embryonic myogenesis and viability and was responsible for the specification of a subset of dispensable embryonic muscles, DA3 and DO4 (13). This was an extension of an earlier study where transheterozygous flies for overlapping deficiencies that remove the nautilus gene survived to adulthood at rates similar to their heterozygous siblings (24).

A comparison between the nau¹⁸⁸ EMS allele and the nau^armGFP mutant, either in combination or with the Exelixis nautilus deficiency, has now demonstrated that both null mutants have the same range of defects at all stages of development: there is severe disruption of the embryonic muscle pattern in roughly one-quarter to one-third of the embryos, and less severe disruptions do not correlate with the loss of a distinct set of muscle fibers (13); in the nau¹⁸⁸ and nau^armGFP mutants, 50–70% of the embryos die by the pupal stage; surviving adults show reduced mobility, and female progeny are infertile with enlarged abdomens and unable to oviposit. This result is difficult to reconcile with the previous report that females transheterozygous for the overlapping nautilus deficiencies were able to oviposit at rates equivalent to wild-type females (24). We have never observed this with either the nau^armGFP or nau¹⁸⁸ mutants, and the basis for this discrepancy is not clear.

We have also been able to phenocopy both the nau^armGFP and nau¹⁸⁸ mutations using a Gal-4-inducible nautilus RNAi transgene. Although off-target effects in RNAi have become a concern, nautilus dsRNA has not been reported to have off-target issues in genomic wide screens (see http://flyrnai.org), and the similarity between the genetic and RNAi data supports this interpretation (25, 26).

The loss of nautilus gene function impacts all stages of development and demonstrates the Drosophila MyoD homolog is essential for normal myogenesis and viability. The select fiber loss for muscles DA3 and DO4 previously reported for the nau¹⁸⁸ allele was not observed in either the nau^armGFP or nau¹⁸⁸ mutants and is not representative of the range of the phenotypes shown above. This earlier conclusion may have been the result of a biased analysis of “escapers” in the EMS nau-null embryo population that displayed segmentation and some movement rather than the analysis of the entire embryo population (13).

The appearance of a small number adult flies in nautilus nulls indicates there are redundant or compensatory mechanisms that can sustain a low level of survival, even though these flies have movement and coordination problems, and the females are sterile. Recent evidence for redundancy comes from C. elegans where SRF and HAND proteins work in concert with the worm nautilus homolog, hlh-1, as additional myogenic factors, because overexpression of either can convert naïve blastomeres to muscle, and HAND alone can trigger myogenic conversion in the absence of hlh-1 function (27). Alternatively, survival may reflect a stochastic outcome that selects for the minimal disruption in the founder cell pattern.

The nau^armGFP Mutation Affects the Founder Cell Pattern.

To determine whether the disorganization in the embryonic muscle observed in the nau^armGFP mutant was because of a change in the underlying stereotypic founder cell pattern, the nau^armGFP mutation was introduced into the enhancer trap line, rP298LacZ, that marks the expression of the founder cell-specific protein, Dumbfounded (Duf), with nuclear β-galactosidase beginning from stage 11 (28, 29). In normal rP298LacZ embryos, Duf-LacZ (red) was expressed in a highly organized, stereotypic pattern in a subset of mesodermal cells before MHC expression (Fig. 2A, Wt). In the presence of the nau^armGFP mutation, the normal early Duf-LacZ expression pattern was disrupted, even in stage 11 embryos (Fig. 2A Upper, compare Upper and Lower Left). Founder disruption was more evident in slightly later stage embryos, still in the absence of MHC expression (Fig. 2A, MHC absence not shown). Myoblast fusion occurs within a matter of a few hours and is coincident with the onset of muscle-specific protein synthesis, so the absence of myosin expression is a good correlate for minimal muscle differentiation (30–32). Twenty to thirty percent of the embryos showed severe disruption in the founder cell pattern in the absence of myosin synthesis, similar to the percentage of embryos that demonstrated prominent muscle defects. In general there was no discernable decrease in the number of Duf-LacZ founders when comparing nau^armGFP and normal rP298LacZ embryos, although the LacZ intensity was sometimes reduced. In wild-type stage 13 embryos, myosin expression (green) and syncytia formation are first observed in the posterior segments in the ventral mesoderm adjacent to the CNS, particularly in abdominal segments 5–7 (Fig. 2B, Wt; see refs. 1 and 3). In the nau^armGFP mutant, embryo myosin synthesis first appeared around the misplaced Duf-LacZ-positive founder cells (Fig. 2B, Mut), demonstrating the absence of nautilus gene function correlates with a disruption in the founder cell pattern and the loss of muscle integrity in the embryo. Disruption of the Kr founder pattern, which marks at least 13 muscle groups in both dorsal and ventral regions of the embryo (4), was also evident from stages 11–13 in the nau^armGFP mutant embryos (SI Fig. 6). This was not unexpected, because Kr and Duf-LacZ are coexpressed in numerous founders (SI Fig. 6), and subsets of Nautilus-positive founders also express Kr (see below; Fig. 3).

Fig. 2.

Muscle formation follows the disrupted founder cell pattern. (A) The nau^armGFP mutation was introduced into the enhancer trap line, rP298LacZ, that marks the expression of the founder cell-specific protein, Duf, with nuclear β-galactosidase (Duf-LacZ, red). All embryos were myosin-negative, a marker for fusion and muscle differentiation. The highly ordered reiterative founder cell pattern is disrupted in the mutant stage 11 embryos (compare Left Upper and Lower) and in later stages before muscle formation (Wt vs. Mut). (B) At the onset of myogenesis, the new muscle (green) appears around the misplaced founders: the normal (Wt, Upper, arrows) or disrupted (Mut, Lower, nau^armGFP) founder cell pattern, with merge in right images. Dorsal is to the top, and anterior is to the left.

Fig. 3.

Nautilus is expressed before Duf-LacZ and subsequently marks founders expressing Kr, Eve, and S59, as well as Duf. (Top) Nautilus protein (green) is first detected at stage 9–10 when Duf-LacZ (red) is barely evident in posterior segments. The merge indicates that a few cells coexpress Nautilus and Duf-LacZ (arrows). (Middle) A majority of the Nautilus-positive cells in the posterior segments show coexpression of Duf-LacZ as founders emerge at stage 12 (Inset). (Bottom) Enlarged images; by stage 13, Nautilus (green) is coexpressed in subsets of founders (red) positive for Kr (K-n), Eve (E-n), and S59 (S59-n), each marked with arrows. Dorsal is to the top, and anterior is to the left.

The loss of founder cell patterning suggests the nautilus gene is required for proper cell–cell interaction, possibly involving cell adhesion components or transmembrane receptors that position founder cells at particular sites in the epidermis of each hemisegment (33). Extracellular matrix and cell adhesion components have been identified in a screen enriched for genes expressed in founder cells, including nidogen and tartan, but we have not yet determined whether they are potential targets for nautilus regulation (34). Gut constrictions are often absent or abnormal in the nau^armGFP mutant (SI Fig. 6). Nautilus protein was previously shown to be expressed in the gut constrictions in wild-type embryos (11), and this fits with the subsequent findings from Bate and coworkers (35) that founders and fusion-competent myoblasts are involved in the formation of visceral muscle and gut constriction. The founder cell pattern also prefigures adult myotube locations, so the weakness and uncoordinated movements seen in nau^armGFP larvae and adults may reflect alterations in the normal muscle fiber organization and attachments because of a general disruption of the founder cells throughout development (36, 37).

Nautilus Is an Early Marker for Most Muscle Founders.

Nautilus protein expression commences in a stereotypic pattern just as myogenic competence is being established, around stage 9–10 (Fig. 3Top) (5, 10, 11). The founder cell marker, Duf-LacZ, is not highly expressed until stage 11, when muscle progenitors and fusion-competent myoblasts are segregated, but weak expression can be seen in a few Nautilus-positive cells in the posterior segments in the early embryo (Fig. 3 Top Right, arrows) (28, 29, 34). By stage 12, founders (Duf-LacZ) organized into a segmentally repeated pattern began to appear in the posterior segments and by this stage, most founders were Nautilus-positive, as indicated by the yellow in the merged image and shown in the enlarged inset (Fig. 3 Middle Right). In a slightly magnified view of stage 13 embryos, subsets of Nautilus-positive founders were also seen to coexpress diverse muscle identity genes specifying both dorsal and ventral muscle groups; Kr and Nautilus were coexpressed in numerous more dorsally positioned founders in each hemisegment (Fig. 3 Bottom Left, arrows); coexpression of Nautilus and Eve was seen only in a group of founders located near the dorsal edge of the mesoderm in a region corresponding to the eventual position of the Eve-positive dorsal muscle, DA1 (Fig. 3 Bottom Center, arrows); S59-LacZ and Nautilus were coexpressed in founders destined to form the ventral muscle cluster II composed of muscles 26/27/19/VaP (Fig. 3 Bottom Right, arrows), as reported, which also express Kr, Msh, and Apterous (38). Thus Nautilus marks subsets of founders that specify diverse muscle groups throughout the hemisegment, not just the minor dorsal muscles, DA3 and DO4 (39), and these are disrupted in the nautilus null, as evidenced by the altered Duf-LacZ and Kr patterns (Fig. 10).

Even though a majority of Duf-expressing founders are Nautilus-positive at stage 11, by stage 13, Nautilus expression is more restricted to subsets of Kr, S59, and Eve-positive cells. However, the transient expression of Nautilus likely underestimates its role in the specification of founder identity, because every muscle in the embryo contains one or more nuclei that initially expressed Nautilus, suggesting the loss of nautilus function could impact most muscles (4, 11). Regardless, Nautilus expression is not detected in all founders once the muscle identity genes are activated, and this may determine, in part, the degree of disruption in the nautilus null. It is interesting to note that the integrity of the more dorsal muscle groups, where Nautilus, Kr, and Eve are coexpressed, is most often affected first in the least-penetrant nautilus null embryos and presents a phenotype similar to the nau¹⁸⁸ mutant (Fig. 1).

In summary, we have demonstrated not only does nautilus gene function play an important role in Drosophila myogenesis and viability as a determinant in founder cell patterning, but Nautilus expression is also an early marker for muscle founders involved in the specification diverse muscle groups in the embryo.

Materials and Methods

Drosophila Strains.

The following strains were used in this study: wild-type w¹¹¹⁸ for P element transformation; the rP298LacZ enhancer trap line [provided by A. Nose (28)]; the nau¹⁸⁸(nau¹⁸⁸/Sb^LacZ) EMS-induced nautilus loss-of-function allele [provided by Susan Abmayr (13)]; RRHS59-LacZ flies [provided by Manfred Frasch (38)]; w¹¹¹⁸ carrying heat-shock inducible FLP recombinase and I-SceI endonuclease for targeted recombination (14, 15); w¹¹¹⁸ carrying constitutively active FLP recombinase (15); pActGal-4 (16) and the Exelixis deficiency (stock no. 7674, Exel 6195, break points: 95A4;95B1) from the Bloomington Stock Center (Bloomington, IN). Introduction of nau^armGFP into the rP298LacZ background is described in SI Text along with other details on stocks. Flies were reared at 25°C on standard yeast-cornmeal-agar medium unless otherwise specified.

Plasmid Constructions.

To construct the donor vector for ends-out targeted disruption of the nautilus gene, 2.9- and 5.3-kb DNA fragments spanning the unique PstI site in exon 1 and the 5′ and 3′ HindIII sites flanking the nautilus gene, respectively, were PCR- amplified from w¹¹¹⁸ genomic DNA and combined with an armGFP reporter into the P element targeting vector pw30. To prepare the Gal-4-inducible nautilus RNAi transgene, nau-pINT-1, a PCR fragment containing the 5′ noncoding intron from the actin 5C gene was ligated into pUAST in the BglII and XhoI sites to generate the Gal-4-inducible RNAi transgene vector, pINT-1. PCR-amplified inverted repeats for the nautilus cDNA (without the initiator ATG) were cloned into the EcoRI-BglII and XhoI-Asp-718 sites of pINT-1, respectively. The nautilus rescue transgene was prepared by using pCasper-hs P-element vector and the ORF of the nautilus cDNA cloned into the EcoRI-BamHI sites (see SI Text).

Germ-Line Transformation, Gene Disruption, and Genetic Crosses.

Transgenic lines were established by using standard procedures (40, 41); gene targeting was performed by injecting pw30-nau armGFP into w¹¹¹⁸ embryos, as described (14, 15); appropriate genetic crosses were used to establish transgenic lines, the targeted insertion, homozygous lines, deficiency stocks, and balanced insertion stocks, as indicated with details in SI Text.

Immunohistochemical Staining.

Embryos were fixed and stained as described (42), using antibodies as stated in the text and SI Text. TSA (Molecular Probes, Carlsbad, CA) amplification was used for Nautilus and Kr detection.

Southern Blot, RT-PCR, and siRNA Detection

Southern blots, RT-PCRs, and siRNA analysis are described in SI Text.

Abbreviations

armGFP

armadillo-GFP

EMS

ethane methyl sulfonate.