Genetics, Vol. 155, 1253-1265, July 2000, Copyright © 2000
Genetic Analysis Demonstrates a Direct Link Between Rho Signaling and Nonmuscle Myosin Function During Drosophila Morphogenesis
Susan R. Halsella,
Benjamin I. Chua, and
Daniel P. Kieharta
a Department of Cell Biology, Duke University Medical Center, Durham, North Carolina 27710
Corresponding author:
Daniel P. Kiehart, Duke University Medical Center, Department of Cell Biology, Research Dr., 307 Nanaline Duke Bldg., Durham, NC 27710., d.kiehart{at}cellbio.duke.edu (E-mail)
Communicating editor: R. S. HAWLEY
 | ABSTRACT |
|---|
A dynamic actomyosin cytoskeleton drives many morphogenetic events. Conventional nonmuscle myosin-II (myosin) is a key chemomechanical motor that drives contraction of the actin cytoskeleton. We have explored the regulation of myosin activity by performing genetic screens to identify gene products that collaborate with myosin during Drosophila morphogenesis. Specifically, we screened for second-site noncomplementors of a mutation in the zipper gene that encodes the nonmuscle myosin-II heavy chain. We determined that a single missense mutation in the zipperEbr allele gives rise to its sensitivity to second-site noncomplementation. We then identify the Rho signal transduction pathway as necessary for proper myosin function. First we show that a lethal P-element insertion interacts genetically with zipper. Subsequently we show that this second-site noncomplementing mutation disrupts the RhoGEF2 locus. Next, we show that two EMS-induced mutations, previously shown to interact genetically with zipperEbr, disrupt the RhoA locus. Further, we have identified their molecular lesions and determined that disruption of the carboxyl-terminal CaaX box gives rise to their mutant phenotype. Finally, we show that RhoA mutations themselves can be utilized in genetic screens. Biochemical and cell culture analyses suggest that Rho signal transduction regulates the activity of myosin. Our studies provide direct genetic proof of the biological relevance of regulation of myosin by Rho signal transduction in an intact metazoan.
MORPHOGENESIS encompasses a complex array of cell shape changes, rearrangements, and movements at many stages throughout the life cycle of an organism. The dynamic cytoskeleton drives many of these cellular events, and regulation of the cytoskeleton during morphogenesis is likely a multistep process. Morphogenetic cytoskeletal changes or movements may occur in a cell-autonomous fashion in response to differentiation cues. Alternatively, these cytoskeletal changes may be induced downstream of extracellular cues. Indeed, it is likely that any given morphogenetic process requires a combination of both cell-autonomous and nonautonomous processes. In either case, intracellular signal transduction leading to direct reconfiguration of cytoskeletal structure and activity would be required. As such, understanding morphogenesis requires investigation of the interplay between upstream regulation and cytoskeletal dynamics in an intact animal.
The actin cytoskeleton plays a special role in epithelial sheet morphogenesis. During vertebrate neurulation, neural tube formation is a multistep process, and cell shape changes occur at the dorsal and medial hingepoints (reviewed in SCHOENWOLF and SMITH 1990 
). Electron microscopy reveals actin filaments in these neuroepithelial cells undergoing wedging; additional studies reveal that cytochalasin D treatment can disrupt dorsal-lateral hingepoint cell wedging, followed by disruption of neural fold convergence. Pharmacological and genetic studies in Drosophila reveal that the actin cytoskeleton is critical for many developmental processes, including mRNA localization, nuclear migrations, and cellularization in the early embryo (reviewed in MILLER 1995 ). The role of the actin cytoskeleton during morphogenesis has also been examined in Caenorhabditis elegans. During the early morphogenetic events of ventral enclosure and elongation, specific subcellular actin arrays are observed (PREISS and HIRSH 1986 ; WILLIAMS-MASSON et al. 1997 ). Cytochalasin D treatment disrupts these arrays and both morphogenetic movements fail. These studies suggest that contraction of the actin cytoskeleton drives the completion of ventral enclosure and elongation. Nevertheless, the molecular motor that drives cell shape changes and rearrangements in C. elegans has not been identified.
Conventional, nonmuscle myosin-II (henceforth, myosin), a chemomechanical motor, drives contraction of the actin cytoskeleton. Myosin functions throughout phylogeny, driving cell shape changes required for cytokinetic furrow formation, cell movement, and tissue morphogenesis. The role of myosin in these processes may be inferred from its subcellular location. In addition, mutagenesis of myosin genes in several organisms (including Drosophila, see below) reveals its critical function in these processes. The MYO1 gene encodes the Saccharomyces cerevisiae myosin-II heavy chain, and MYO1 mutants exhibit retarded cell growth (WATTS et al. 1987 ). Mutants also fail to form actomyosin contractile rings, but cytokinesis by alternative means still occurs (BI et al. 1998 ). Myosin function in Dictyostelium discoideum has been disrupted by mutating the heavy chain gene via homologous recombination and by antisense RNA inactivation (DE LOZANNE and SPUDICH 1987 ; KNECHT and LOOMIS 1987 ; MANSTEIN et al. 1989 ). These studies reveal that fruiting body morphogenesis is disrupted and that lack of myosin function blocks formation of the stalk at the mound stage. Microscopic analysis suggests that cell shape changes driven by myosin function are critical during Dictyostelium morphogenesis (KNECHT and SHELDEN 1995 ; SHELDEN and KNECHT 1996 ). Similar to the yeast study, adhesive Dictyostelium myosin-II null cells can undergo cytokinesis in the absence of myosin-II function; however, cells in suspension absolutely require myosin-II function for cytokinesis (NEUJAHR et al. 1997 ; ZANG et al. 1997 ).
Genetic and experimental studies in Drosophila melanogaster reveal multiple steps in the life cycle that require myosin function. Myosin is subcellularly localized in embryonic cells undergoing apical constriction (YOUNG et al. 1991 , YOUNG et al. 1993 ; EDWARDS and KIEHART 1996 ; KIEHART et al. 2000 ). Zygotic activity of the single copy, nonmuscle myosin-II heavy chain gene (zipper) is absolutely required for enclosure of the amnioserosa by the lateral epidermis during dorsal closure (YOUNG et al. 1993 ). It was noted that ventral enclosure in C. elegans resembles dorsal closure and may also require a myosin motor (WILLIAMS-MASSON et al. 1997 ). Embryonically, zipper also functions during head involution and axon guidance (COTE et al. 1987 ; ZHAO et al. 1988 ; JACK and MYETTE 1997 ; PEDERSON 1997 ; BLAKE et al. 1998 ). Further genetic manipulation reveals that border cell migration and nurse cell transport during oogenesis, nuclear migrations in the early embryo, larval cytokinesis, and pupal imaginal disc morphogenesis also require myosin function (GOTWALS and FRISTROM 1991 ; KARESS et al. 1991 ; WHEATLEY et al. 1995 ; EDWARDS and KIEHART 1996 ; HALSELL and KIEHART 1998 ). In fact, depletion of myosin during leg imaginal disc morphogenesis results in deformed adult legs, referred to as the malformed phenotype (mlf, described below; EDWARDS and KIEHART 1996 ).
Pupal leg imaginal disc morphogenesis results from myosin-driven cell shape changes, and it is particularly sensitive to perturbation (reviewed in VON KALM et al. 1995 ; GOTWALS and FRISTROM 1991 ; EDWARDS and KIEHART 1996 ; HALSELL and KIEHART 1998 ). No significant cell proliferation or rearrangements (relative to the cells' contacts with their nearest neighbors) are observed in the prepupal leg imaginal disc; instead the early phase of leg disc elongation appears to rely on the change in cell shapes from anisometric to isometric (CONDIC et al. 1990 ). Disruption of these cell shape changes gives rise to the mlf phenotype that is readily identifiable in the adult (Fig 3). We exploited the sensitivity of leg imaginal disc morphogenesis by testing whether Deficiencies for genomic regions uncover loci that genetically interact with zipper (HALSELL and KIEHART 1998 ). In that study, we found 2 deficiencies that genetically interact strongly and 17 that interact at intermediate levels. Within this collection of deficiencies, we have thus far identified 3 whose interaction with zipper can be explained by individual mutant loci. These three loci include cytoplasmic tropomyosin, viking (encoding a collagen IV), and a single complementation group represented by two EMS-induced alleles (E3.10 and J3.8).
View larger version (53K):
In this window
In a new window
Download PPT slide
|
Figure 2.
Immunoblot analysis reveals myosin heavy chain protein levels in zipEbr homozygous mutant animals. Animals homozygous mutant for zipEbr die as late embryos/early larvae; therefore, myosin heavy chain levels were examined in both stages. Protein levels were compared between comparably staged wild-type, zipEbr, and zip2 homozygous mutant animals. The wild-type embryo controls (lane A, WT) were derived from Canton-S stocks, while the phenotypically wild-type larvae (lane D, Ebr/CyO) were the zipEbr/CyO, y+-bearing siblings of the mutant larvae. In addition to staining with polyclonal anti-myosin heavy chain antibody (mhc), we also stained with a polyclonal antibody directed against  -spectrin (spec) as a loading control that suggests minor variation in total protein loaded. Levels of myosin heavy chain protein in zipEbr homozygous mutant animals (lane C, Ebr/Ebr embryos and lane E, Ebr/Ebr larvae) are somewhat reduced as compared to the levels in wild type. This level of 205-kD heavy chain is significantly greater than that observed in zip2 homozygotes (lane B, 2/2 embryos). The molecular lesion in zip2 is a premature stop codon that should give rise to a ~85-kD protein (MANSFIELD et al. 1996 ); any detectable full-length myosin heavy chain may result from read-through or more likely reflects perdurance of the maternally encoded protein.
|
|
View larger version (83K):
In this window
In a new window
Download PPT slide
|
Figure 3.
Genetic interactions between RhoGEF2 and zipper and between RhoA and zipper cause defects in leg imaginal disc morphogenesis that give rise to the malformed leg phenotype. (A) The third leg of a wild-type adult consists of tarsal segments, the tibia, and the femur. The tibia and femur (arrows) are long, slender structures, and morphogenetic cell shape changes drive their elongation (CONDIC et al. 1990 ). (B and C) The malformed leg phenotype exhibits variable expressivity. (B) In less severe cases, a dent in the femur or tibia (arrowhead) is diagnostic. This fly was genotypically RhoGEF24.1 +/+ zipEbr. (C) In more severe cases, a dent and twisting of the femur is apparent, along with shortening and thickening of the tibia (arrowheads). This fly was double heterozygous for RhoAE3.10 and zipEbr. Note that flies from any given cross giving rise to progeny double heterozygous for either RhoGEF2 or RhoA and zipEbr show the entire range of malformed phenotypes. Similar results are obtained with all of the tested RhoGEF2 and RhoA alleles. These images were captured using brightfield microscopy and a 5x Neofluar objective (0.15 NA). The scale is the same in A–C.
|
|
Here, we determined that the zipperEbr (zipEbr) allele used in our screens has a single missense mutation within the region encoding the globular head domain. Further, we identified two zipEbr genetic interactors that encode members of the Rho signal transduction pathway. First, we screened lethal P-element transposon insertions and found that a P-element insertion that disrupts the RhoGEF2 (Rho Guanine Exchange Factor 2) locus interacts genetically with zipper. We also show that the complementation group defined by the E3.10 and J3.8 mutations encodes RhoA. DNA sequence analysis of these two EMS-induced mutations reveals single point mutations within the RhoA gene. Finally, we detected genetic interactions between the RhoA mutation and a flanking chromosomal deficiency. Overall, these results verify a direct, in vivo link between the Rho signal transduction pathway and myosin function.
| MATERIALS AND METHODS |
|---|
Stocks:
We obtained a collection of second chromosome lethal P-element insertions, wa Nfa-g; Df(2R)Jp1/CyO, wa Nfa-g; Df(2R)Jp4/CyO and wa Nfa-g; Df(2R)Jp8/CyO stocks from the Bloomington Stock Center (Bloomington, IN). Bill Saxton (Indiana University) provided wa Nfa-g; E3.10/CyO (SAXTON et al. 1991 ). Ruth Steward (Rutgers University) provided J3.8/CyO (R. STEWARD, personal communication). We obtained the w; RhoGEF21.1/CyO and w; RhoGEF24.1/CyO stocks from J. Settleman (Massachusetts General Hospital Cancer Center, Harvard Medical School; BARRETT et al. 1997 ). M. Mlodzik (European Molecular Biology Laboratory) provided y w; RhoA72F/CyO and y w; RhoA72O/CyO (STRUTT et al. 1997 ). zipEbr/SM5 was obtained from Jim Fristrom (University of California, Berkeley; GOTWALS and FRISTROM 1991 ; GOTWALS 1992 ).
Genetic screens and complementation analysis:
All crosses were performed at 25° on standard cornmeal/molasses fly food. The second-site noncomplementation screen was performed as previously described (HALSELL and KIEHART 1998 ). Five zipEbr/SM5 virgin females were mated to three to five mutation- or deficiency-bearing males. Crosses were brooded every 4–5 days. All progeny were scored for the malformed phenotype (GOTWALS and FRISTROM 1991 ). Penetrance of the malformed phenotype in flies double heterozygous for the mutation or deficiency and zipEbr (e.g., m +/+ zipEbr) was compared directly to sibling flies singly heterozygous for either of the mutations (e.g., m +/SM5 or + zipEbr/Balancer). In no cases did flies singly heterozygous for the mutation or deficiency result in a significant number of malformed flies.
Genetic complementation was performed by crossing five balanced virgin females to three to five balanced males. All progeny were scored. Since Balancer/Balancer progeny are embryonic and/or early larval lethal, the Mendelian expectation for mutations that complement one another is that 33% of the adult progeny will carry both mutations.
P-element reversion:
Specificity of the l(2)04291 interaction with zipEbr was verified by excision of the P element and reversion of (1) the malformed phenotype in double heterozygous flies and (2) the homozygous lethality of the original l(2)04291 chromosome. Briefly, cn l(2)04291/CyO; ry506 males were mass-mated to Sp/CyO; Sb 
2-3/TM6 virgins. In the F1 generation, individual cn l(2)04291/CyO; Sb 2-3/ry506 males were mated to Bc Elp/CyO; ry virgins. In the F2 generation, excision events were identified as a loss of the ry+ marker within the P element. From each line exhibiting ry males, individual cn l(2)04291rev/Bc Elp or CyO; ry males were mated to cn zipEbr/SM5 virgins. In the next and subsequent generations, reversion of the genetic interaction with zipEbr was observed as a loss of the malformed phenotype in cn l(2)04291rev +/cn + zipEbr double heterozygous flies. Stocks were established for each of the l(2)04291rev lines, and each was tested for homozygous viability and in complementation assays with the original l(2)04291 chromosome.
Plasmid rescue:
Genomic DNA flanking the P-element insertion site was recovered by plasmid rescue, using established methods. In brief, genomic DNA was isolated from cn l(2)04291/CyO flies and quantified by its absorbance at 260 nm (LIS et al. 1983 ). A total of 1 µg was digested with XbaI and SpeI for 4 hr, and the enzymes were heat inactivated by incubation at 65° for 20 min, followed by ethanol precipitation. The DNA pellet was resuspended in 15 µl TE, and 5 µl was ligated overnight at 15° in a 200-µl reaction containing 3 units of T4 DNA ligase (GIBCO BRL, Gaithersburg, MD). The ligated product was ethanol precipitated, resuspended in 10 µl H2O, and the entire resuspension was used to transform XL-1 Blue-competent bacteria (Stratagene, La Jolla, CA). Transformants were selected on LB-kanamycin plates. The insert DNA was subcloned in Bluescript (Stratagene) as HindIII fragments, and these subclones were sequenced with Bluescript-specific primers at the Duke University Comprehensive Cancer Center DNA analysis facility (see below).
DNA sequencing of the zipEbr, E3.10, and J3.8 mutations:
Homozygous mutant animals were collected as follows. Overnight egglays were performed, and the embryos were allowed to age for at least 24 hr at 25°. zipEbr homozygotes were collected from a y; zipEbr/CyO, y+ stock and were identified as y embryos and/or larvae. Mutant E3.10 and J3.8 embryos were collected on the basis of the characteristic head involution defect assayed at 20x with a stereomicroscope (Fig 5). Genomic DNA was then isolated as follows. Twenty animals were homogenized in 50 µl of 10 mM Tris-Cl, pH 8.2, 1 mM EDTA, 25 mM NaCl, 200 µg/ml proteinase K by pressing the embryos against the tube wall with a P-200 ("yellow") tip. The homogenate was digested at 37° for 20 min, and then the proteinase K was heat inactivated by incubation for 2 min at 98° (GLOOR et al. 1993 ). The debris was pelleted in the microfuge and the DNA-containing supernatant was used in the PCR reactions.
View larger version (109K):
In this window
In a new window
Download PPT slide
|
Figure 5.
RhoA mutant embryos exhibit dorsal anterior holes. Cuticle preparations were made of (A) a newly hatched wild-type first instar larva; and (B, E, and F) RhoAE3.10/RhoAE3.10 homozygous, (C) RhoAE3.10/Df(2R)Jp8 trans-heterozygous, and (D) RhoAE3.10/RhoA72O trans-heterozygous embryos. (A) The wild-type larva exhibits an internalized head skeleton and intact cuticle at the anterior (arrow) and at the posterior, elongated posterior spiracles (arrowhead). (B–D) Embryos mutant for RhoA exhibit similar mutant phenotypes. In all cases a hole on the dorsal side at the anterior of the embryos is apparent (B–D, arrow in B). This hole extends from the anterior to approximately the second or third thoracic segment. (E) Higher magnification of the anterior of the RhoAE3.10/RhoAE3.10 homozygous embryo shown in B. The head skeleton has been extruded out of the dorsal opening. In addition to the anterior holes, the shape of the posterior spiracles is aberrant in RhoA mutants. (F) Higher magnification of the RhoAE31.0/RhoAE3.10 homozygous embryo in B is shown. In contrast to the elongated spiracle in the wild-type larva, RhoA mutant embryos exhibit compacted posterior spiracles (B–D, arrowhead in B). The compaction of the posterior spiracle is readily observed in the kinked filzkörper (dark structures within the spiracle). In A–F, anterior is to the left and dorsal is up. The cuticles were imaged by phase contrast microscopy, using a 10x Plan-Neofluar objective (0.3 NA; A–D) or a 20x Achroplan objective (0.45 NA), optivar setting 2x (E and F). Scales are the same in A–D and E and F.
|
|
Fragments of DNA spanning either the zipper or RhoA genes were generated by PCR. zipper-specific primers were based on the complete genomic sequence (MANSFIELD et al. 1996 ). RhoA-specific primers were based on the previously published cDNA sequence for RhoA and RhoL and intron sequences determined in this study (HARIHARAN et al. 1995 ; MURPHY and MONTELL 1996 ). Each 30-µl PCR reaction included 3 µl of the 50-µl mutant genomic DNA preparation, 40 mM KCl, 60 mM Tris-Cl, pH 8.5, 166 µm dNTPs, 25 ng of each primer, 1.5–6.0 mM MgCl2, 0.2 µl of Perfect Match (Stratagene), and 0.3 µl of BIO-X-ACT DNA polymerase (ISC BioExpress, Kaysville, UT). The PCR reactions were performed in a MJ Research thermocycler as follows: 5 min at 95°, followed by 35 cycles of 1 min at 95°, 1 min at 58°, 5 min at 72°, and finally a cycle of 10 min at 72°. Typically, at least three PCR reactions were combined, run on an agarose gel, and eluted from gel slices with QIAGEN spin columns (QIAGEN, Valencia, CA). DNA sequencing was performed by the Duke University Comprehensive Cancer Center DNA analysis facility using an ABI Prism 377XL DNA sequencer and dRhodamine/BigDye terminator cycle sequencing reagents (PE Applied Biosystems, Foster City, CA).
Protein sample preparation and immunoblotting:
Twenty appropriately staged embryos or larvae were collected from egg-lay plates, homogenized in 100 µl 1x sample buffer, and boiled for 3 min. SDS-PAGE and immunoblotting were performed essentially as in KIEHART and FEGHALI 1986 , with the following modifications. Aliquots (15 µl) of total embryo homogenate in sample buffer (~3 embryo equivalents) were separated in 3-mm lanes on 6%/0.6% acrylamide gels by SDS-PAGE. Blots were blocked with Blotto (5% nonfat milk, 0.02% sodium azide, 0.2% Tween-20), and all antibody incubations were in Blotto. Primary antibody dilutions were 1:500 and included rabbit polyclonal antisera directed against myosin heavy chain (656; KIEHART and FEGHALI 1986 ) or -spectrin (905; BYERS et al. 1987 ). Signal detection was performed by incubation with horseradish peroxidase conjugated secondary antibodies (1:500; Chemicon International, Temecula, CA), followed by incubation with SuperSignal substrate (Pierce, Rockford, IL) according to manufacturer's directions. Exposures were typically 1–2 min.
Mounting of adult legs and wings and embryonic cuticles:
Malformed adults were preserved in 70% ethanol. Legs and/or wings were removed and mounted in CMPC10 (PolySciences, Warrington, PA). The legs and wings were observed by brightfield microscopy, using a 5x Neofluar objective [0.15 numerical aperture (NA)] on a Zeiss Axioplan microscope. Background subtracted images were captured with a Hamamatsu 4880 cooled CCD camera and Metamorph software (KIEHART et al. 1994 ). Contrast was adjusted using Adobe Photoshop software and images were montaged and labeled.
Embryonic cuticles were mounted as follows. Overnight egg-lays were performed on standard grape juice plates. The embryos were subsequently aged for 36–48 hr at 25°. Unhatched, brown embryos were hand dechorionated on double-stick tape and mounted in a drop of 8.4% polyvinyl alcohol (Sigma, St. Louis), 22% lactic acid, and 22% phenol (PVLP; PEDERSON et al. 1996 ). Alternatively, recently hatched Canton-S larvae were mounted in a drop of Hoyers mountant, diluted 1:1 with lactic acid (WIESCHAUS and NUSSLEIN-VOLHARD 1986 ). The mounts were weighted and allowed to harden overnight at 65°. The cuticles were imaged by phase-contrast microscopy, using a 10x Plan-Neofluar objective (0.3 NA) or 20x Achroplan objective (0.45 NA). Image capturing was as above.
| RESULTS |
|---|
zipEbr is the zipper allele most sensitive to second-site noncomplementation in genetic interaction screens:
Previous studies revealed that putative null alleles of zipper do not exhibit second-site noncomplementation behavior; in contrast, zipEbr and two postembryonic lethal zipper alleles (zip2.1 and zip6.1) do (HALSELL and KIEHART 1998 ; J. FRISTROM and S. R. HALSELL, unpublished results). zipEbr is more sensitive than either zip2.1 or zip6.1. The zipEbr allele was generated by EMS mutagenesis and was identified in screens for mutations that interact genetically with the Broad-Complex (GOTWALS and FRISTROM 1991 ).
To determine the molecular lesion associated with the genetic behavior of zipEbr, we sequenced genomic DNA isolated from zipEbr homozygotes. All intron/exon boundaries in the mutant zipEbr locus are wild type. We identified five silent mutations, relative to the sequence determined by the Berkeley Drosophila Genome Project (accession no. AC006244, Berkeley Drosophila Genome Project). Significantly, within the entire open reading frame we found only a single missense mutation located within exon 9 (Fig 1). zipEbr exhibits a transition mutation at nucleotide 14,374 (relative to the corrected sequence reported in MANSFIELD et al. 1996 , accession no. UU35816U), such that G is changed to A. Such transitions are consistent with those generated by EMS (ASHBURNER 1989 ), and in this case, result in an arginine-to-histidine replacement at amino acid 276 (with respect to the corrected, short transcript published in KETCHUM et al. 1990 ). This is equivalent to the arginine found at position 274 in the chicken fast skeletal muscle myosin heavy chain (COPE and HODGE 2000 ). The arginine at this position is highly conserved in all members of the myosin superfamily. Among 130 different myosins, 128/130 of them conserve the arginine residue at this position (COPE and HODGE 2000 ). On the basis of comparison to the chicken skeletal myosin heavy chain structure determined crystallographically, this residue lies near the wall of the ATP-binding pocket (RAYMENT et al. 1993 ).
We performed immunoblot analysis to determine if the level of myosin heavy chain protein is altered in zipEbr mutants. zipEbr mutants die as late embryos or early larvae, and both stages were examined (Fig 2). The level of myosin heavy chain protein in zipEbr homozygous mutant embryos is somewhat reduced in comparison to the level seen in wild-type embryos (Fig 2A and Fig C). Further, the level of myosin heavy chain observed in the zipEbr homozygous larvae is comparable to that observed in their heterozygous zipEbr siblings (Fig 2D and Fig E). This is in marked contrast to that of the apparent null allele, zip2, which is predicted to produce a truncated protein due to a premature stop codon at residue 750 and exhibits little or no detectable myosin heavy chain protein in homozygous mutant embryos (Fig 2B; YOUNG et al. 1993 ; MANSFIELD et al. 1996 ). Any residual, full-length myosin heavy chain protein in zip2 homozygotes is due to read-through or more likely reflects perdurance of wild-type protein loaded maternally. zip2/+ heterozygous flies do not display second-site noncomplementation in our assay. We conclude that the dominant behavior of the zipEbr allele in the mlf second-site noncomplementation is probably due to the unique structure of the R276H substituted zipEbr gene product and not likely to arise from a reduction in myosin heavy chain protein levels.
RhoGEF2 interacts genetically with zipper:
To identify loci encoding gene products that collaborate with nonmuscle myosin during morphogenesis, we performed second-site noncomplementation screens for the mlf leg phenotype (Fig 3; HALSELL and KIEHART 1998 ). We extended our previous chromosomal deficiency screens by screening a collection of 268 single, lethal P-element insertional mutations on the second chromosome for genetic interactions with the zipEbr allele. Fourteen insertions failed to complement zipEbr. Previously, we arbitrarily defined the strength of the genetic interaction on the basis of the percentage of flies of the appropriate genotype that exhibit the malformed phenotype: weak interactions show penetrance of 10–25% while intermediate interactions are 25–75% penetrant (HALSELL and KIEHART 1998 ). Eleven of the lethal P-element insertions we identified are weak interactors. Three of the insertions are intermediate interactors. Two of these intermediate interactors are not second-site noncomplementing loci but are new zipper alleles, exhibiting intraallelic complementation (S. R. HALSELL, unpublished observations). The third intermediate interacting mutation, l(2)04291, causes mlf flies in trans to zipEbr with a penetrance of 38% (Table 1).
Genetic reversion analysis confirmed that the interaction observed between zipEbr and l(2)04291 is a direct consequence of the P-element insertion. After mobilizing the P element by crossing in a transposase source, we established 57 lines that had lost the rosy+ marker. Each line was tested for genetic interaction with zipEbr. Of these lines, 29 no longer interacted genetically with zipEbr. All 29 reverted lines are homozygous viable and all complement the original l(2)04291 P-element insertion, suggesting that each represents a precise excision of the P element. This result demonstrates that the observed genetic interaction with zipEbr is specific to the transposon insertion. The remaining 28 lines appear to be imprecise excision events. All interact genetically with zipEbr, are homozygous inviable, and fail to complement the original P allele.
We subsequently determined that the P-element insertion disrupts the RhoGEF2 locus. We recovered genomic DNA flanking the P-element insertion by plasmid rescue, sequenced flanking DNA, and discovered that the P element lies within an intron that interrupts the 5' UTR of the RhoGEF2 gene. We made this observation concomitant with the first publications of this RhoGEF2 locus (BARRETT et al. 1997 ; HACKER and PERRIMON 1998 ). To further confirm that the genetic interaction observed with zipEbr results from a mutation in RhoGEF2, two EMS-induced mutant RhoGEF2 alleles, 1.1 and 4.1 (BARRETT et al. 1997 ), were tested in the malformed leg assay. Both alleles interacted with zipEbr; the penetrance of the malformed phenotype in double heterozygous flies was 33% with the RhoGEF21.1 allele and 27% with the RhoGEF24.1 allele (Table 1), comparable to that seen with the original P-insertional allele.
In addition to the malformed legs observed in flies double heterozygous for mutant RhoGEF2 and zipEbr, we observed malformed wings at comparable frequencies (Fig 4). Between 80 and 97% of the flies exhibiting a malformed leg phenotype also exhibited malformed wings. In contrast, most other loci that interact with zipper do not exhibit significant wing defects (S. R. HALSELL, unpublished observation). We rarely observed malformed wings when the legs were wild type. Taken together, these data indicate a requirement for RhoGEF2 during myosin-driven leg and wing imaginal disc morphogenesis.
Myosin-driven imaginal disc morphogenesis requires RhoA Function:
In a previous screen for zipper interactors, we identified a single complementation group, represented by two independently derived EMS mutations (E3.10 and J3.8), as the locus within Df(2R)Jp8 responsible for this deficiency's strong second-site noncomplementation and partial synthetic lethality in trans to zipEbr (Table 1; HALSELL and KIEHART 1998 ). These mutations were recovered in screens for lethals uncovered by Df(2R)Jp8, which deletes the chromosome between cytogenetic positions 52F5-9; 52F10-53A1 (SAXTON et al. 1991 ; Z. LIU and R. STEWARD, personal communication).
We have characterized E3.10 and J3.8 further. These alleles are recessive, embryonic lethals; homozygous or trans-heterozygous E3.10/J3.8 mutant embryos have dorsal anterior holes in their cuticle (Fig 5B). Further, mutant allele E3.10 behaves genetically like a null. When E3.10 is in trans Df(2R)Jp8, mutant embryos show similar cuticular defects as E3.10 homozygotes (cf. Fig 5B and Fig C).
Interestingly, polytene chromosome in situ analysis placed the RhoA locus at cytogenetic position 52E3-6 (STRUTT et al. 1997 ), adjacent to but not included in the interval removed by Df(2R)Jp8. Nevertheless, we suspected that the E3.10 and J3.8 mutations might be allelic to RhoA. Specifically, like the E3.10 and J3.8 mutations, RhoA excision alleles are recessive embryonic lethals that have defects in the anterior cuticle (STRUTT et al. 1997 ).
To test the hypothesis that E3.10 and J3.8 are alleles of RhoA, we performed complementation analysis between each of the two point mutations, E3.10 and J3.8, and each of two RhoA excision alleles, 72F and 72O (STRUTT et al. 1997 ). Both E3.10 and J3.8 fail to complement the RhoA excision alleles 72F and 72O (Fig 6). Trans-heterozygotes for either of the EMS alleles and either one of the excision alleles are embryonic lethal and give rise to embryos with dorsal anterior holes in the cuticle (Fig 5D). We also mapped the EMS and RhoA excision alleles genetically with chromosomal deficiencies. If the previously reported position of RhoA (52E3-6) is correct, then RhoA72F and RhoA72O should fail to complement Df(2R)Jp1, which uncovers cytogenetic interval 51C3; 52F8-9, but should complement Df(2R)Jp8. Flies carrying either RhoA excision allele are 100% viable in trans to the Df(2R)Jp1 chromosome, while Df(2R)Jp8 fails to complement both alleles (Fig 6). This result is consistent with the observed complementation behavior of E3.10 and J3.8 with respect to these deficiencies. Finer genetic mapping revealed that Df(2R)Jp4 (which uncovers cytogenetic interval 51F13; 52F8-9) interacts genetically with zipEbr at a penetrance of 90% (Table 1). This deficiency also uncovers all four RhoA alleles (Fig 6). These results place RhoA within cytogenetic interval 52F8-9 (Fig 6). Taken together, these data strongly suggest that mutations E3.10 and J3.8 are due to lesions in the RhoA locus.
View larger version (44K):
In this window
In a new window
Download PPT slide
|
Figure 6.
Finer genetic mapping of the RhoA locus. (A) Fail-to-complement tests were performed with the mutations shown. Above the parentheses is the total number of adults recovered that are trans-heterozyogous for the indicated mutations (e.g., RhoAE3.10/RhoAJ3.8). The number in parentheses indicates the total number of progeny recovered in the cross. Because the parental flies are balanced, complementation between mutations is indicated when 33% of adult progeny carry both mutations (see MATERIALS AND METHODS). On the basis of this criterion, the E3.10 and J3.8 mutations fail to complement known RhoA alleles. (B) On the basis of the complementation data in A, the RhoA locus fails to complement Df(2R)Jp4 and Df(2R)Jp8 and fully complements Df(2R)Jp1. This places the RhoA locus at cytogenetic position 52F8-9, on the basis of the cytology defined for these deficiencies (SAXTON et al. 1991 ).
|
|
On the basis of these data, we tested the excision RhoA alleles for genetic interaction with zipper. Both RhoA72F and RhoA72O genetically interact with zipEbr. Penetrance of the malformed phenotype observed with the RhoA mutations ranges from 92 to 96%, comparable to that seen with Df(2R)Jp4 and slightly lower than that observed with Df(2R)Jp8, E3.10, and J3.8 (Table 1). As in the case of the RhoGEF2-zipEbr genetic interactions, 80–90% of the RhoA-zipEbr double heterozygous flies with a malformed leg also exhibit a malformed wing (Fig 3 and Fig 4).
To characterize the nature of the E3.10 and J3.8 RhoA alleles, we examined them for molecular lesions in the RhoA locus. We isolated genomic DNA from homozygous E3.10 and J3.8 embryos. Genomic fragments were PCR amplified using RhoA-specific primers based on the published cDNA sequences (HARIHARAN et al. 1995 ; MURPHY and MONTELL 1996 ) and were subsequently sequenced. All intron/exon boundaries spanning the coding region were examined and were found to be wild type. The gene structure differs from that previously published (Fig 7; STRUTT et al. 1997 ), but agrees with that reported by MAGIE et al. 1999 . In the Strutt study, five exons were identified, excluding one alternative 5' exon that we did not examine. We identified six exons of the sizes and organization shown in Fig 7. We definitively defined the exon boundaries by DNA sequencing. These exons include the 5' UTR, all coding sequences, and the most 5' portion of the 3' UTR. However, we did not examine the exon structure of the entire 847-bp 3' UTR. E3.10 and J3.8 each have a single point mutation within the coding sequence in exon 5 (Fig 7). Each mutation is a base substitution of the transition class, consistent with their induction by EMS (Fig 7; ASHBURNER 1989 ). J3.8 is a nonsense mutation at amino acid 180 that results in a protein truncated for the last 13 amino acids. E3.10 is a missense mutation that encodes a tyrosine residue instead of the cysteine at position 189, which is normally modified by prenylation (ZHANG and CASEY 1996 ; SEABRA 1998 ; see DISCUSSION). In summary, our genetic and molecular analyses reveal that mutations E3.10 and J3.8 are in the RhoA gene. We now refer to them as RhoAE3.10 and RhoAJ3.8
View larger version (17K):
In this window
In a new window
Download PPT slide
|
Figure 7.
Genomic organization of the RhoA locus and the molecular lesions in RhoAE3.10 and RhoAJ3.8. (A) We sequenced the coding region and exon/intron boundaries of the RhoA gene. We identified 5 exons. The sixth exon, encoding the end of the 3' UTR, was not analyzed. Sizes of the introns were first estimated on the basis of the lengths of PCR products generated from genomic DNA templates or by sequencing through the entire intron. The exact sizes for all of the introns were determined by aligning the exon sequences with genomic sequence generated by the Berkeley Drosophila Genome Project (accession no. AC004248; Berkeley Drosophila Genome Project, unpublished results). In addition, this alignment confirms the exon/intron structure we have determined. The open boxes indicate coding sequences and shaded boxes represent the untranslated sequences. In contrast, STRUTT et al. 1997 identified only four exons, two of which include protein coding sequence. Asterisks at the end of the coding sequence in exon 5 indicate the site of the molecular lesions in RhoAJ3.8 and RhoAE3.10. (B) Single base changes in the RhoA gene are found in the RhoAJ3.8 (Q180stop) and RhoAE3.10 (C189Y) alleles. The last 13 amino acids of RhoA are shown.
|
|
RhoA mutations interact genetically with Df(2R)Jp1:
In the course of mapping the RhoA mutations genetically, we observed that RhoA mutations interact genetically with Df(2R)Jp1 (Table 2). This interaction does not require the presence of zipEbr. All four tested RhoA alleles behave as second-site noncomplementors; RhoAE3.10 shows the highest penetrance of the malformed phenotype (42%). Df(2R)Jp1 may delete a portion of the RhoA regulatory sequence, leading to interallelic complementation in terms of viability but noncomplementation in terms of mlf, or the Df(2R)Jp1 chromosome may carry a tightly linked, hypomorphic RhoA allele. Alternatively Df(2R)Jp1 may remove an additional gene or genes required for leg imaginal disc morphogenesis. Interestingly, zipEbr also interacts genetically with Df(2R)Jp1 (HALSELL and KIEHART 1998 ). We have also found that RhoA mutations exhibit second-site noncomplementation with another locus encoding a gene product in the Rho signal transduction pathway (S. R. HALSELL, unpublished observations). Taken together, these results suggest that a gene product encoded by a locus within Df(2R)Jp1 might function in the Rho signal transduction pathway to regulate nonmuscle myosin function.
| DISCUSSION |
|---|
We have taken a genetic screening approach to identify loci encoding gene products critical for myosin function during morphogenesis (this study; HALSELL and KIEHART 1998 ). We examined flies double heterozygous for zipper and another mutation for second-site noncomplementation behavior giving rise to a malformed adult leg.
zipEbr is the most sensitive zipper allele in screens for second-site noncomplementation:
In pilot screens, we tested a variety of zipper alleles for their sensitivity to second-site noncomplementation (J. FRISTROM and S. R. HALSELL, unpublished results; HALSELL and KIEHART 1998 ). We found that severe zipper alleles (null or near null, hereafter referred to as null) do not exhibit second-site noncomplementation behavior. However, zipEbr and two other hypomorphic alleles did exhibit second-site noncomplementation behavior (HALSELL and KIEHART 1998 ). At least one of the null alleles, zip2, contains a premature stop codon leading to a truncation of the protein within the head region (MANSFIELD et al. 1996 ). zip2 homozygous mutant embryos have little or no discernable myosin heavy chain protein (Fig 2B; YOUNG et al. 1993 ). To determine the molecular basis of the genetic behavior of zipEbr, we examined the level of myosin heavy chain protein in homozygous mutant animals by immunoblotting. While the level of protein in mutant embryos is slightly reduced as compared to wild-type animals, homozygous mutant larvae exhibit similar levels when compared to their phenotypically wild-type, heterozygous siblings (Fig 2D and Fig E). This contrasts with the loss of myosin heavy chain protein found with the genetically noninteracting zipper null alleles. This result suggests that a simple reduction in the level of myosin heavy chain in heterozygous flies does not give rise to second-site noncomplementation behavior.
In addition to determining myosin heavy chain protein levels, we also sequenced the zipEbr allele. We identified a single missense mutation that changes an arginine to histidine at amino acid position 276. Interestingly, this arginine is highly conserved in both conventional and unconventional myosin heavy chains (COPE and HODGE 2000 ). Mutations in the human myosin VIIA gene that change this arginine residue (human: amino acid 241) to serine give rise to the autosomal recessive congenital deafness and blindness disorder, Usher syndrome type 1B (JANECKE et al. 1999 ). In the mouse myosin VII gene, shaker-1, this same arginine residue (mouse: amino acid 241) is mutated to proline and this mutation is also associated with deafness (GIBSON et al. 1995 ). Based on the solved structure of the chicken skeletal myosin head domain, this arginine (chicken: amino acid 274) lies near the wall of the ATP-binding pocket and may be required to maintain the integrity of this key structure in the native myosin molecule (RAYMENT et al. 1993 ). As such, this mutated amino acid may affect myosin activity. Myosin heavy chains dimerize in the native myosin molecule. Considering the immunoblot and DNA sequence data, we predict that zipEbr heterozygous flies carry a mixed population of myosin heavy dimers, including wild-type homodimers, mutant homodimers, and wildtype/mutant heterodimers. We hypothesize that this mixed population of myosin heavy chain dimers gives rise to the genetic sensitivity observed in zipEbr heterozygous flies.
Our studies reveal that the Rho signal transduction pathway and myosin play essential roles during Drosophila leg and wing morphogenesis:
Additional morphogenetic processes are likely to require collaboration between myosin and the Rho signaling pathway. We previously showed that viability depends on myosin and RhoA function; flies double heterozygous for zipEbr and RhoA alleles E3.10 (C189Y) or J3.8 (Q180stop) exhibit substantially reduced viability (16 and 50% of Mendelian expectation, respectively; HALSELL and KIEHART 1998 ). We have not yet determined the phenotype of arrest in these animals. Most significantly, our experiments provide direct, genetic evidence supporting the prediction that Rho-mediated regulation of myosin activity is critical for morphogenetic cell shape change.
The mutation in the RhoAE3.10 allele disrupts the CaaX box:
RhoAE3.10 genetically behaves as a severe allele, yet molecularly results from a single amino acid change that converts a cysteine at position 189 to a tyrosine residue. This missense mutation causes severe effects because it alters the first residue, cysteine, in the CaaX box. The CaaX box is a common feature of members of the Ras-superfamily of small GTPases (reviewed in VALENCIA et al. 1991 ). Functionally, the cysteine residue is the site of a post-translational prenylation modification (reviewed in ZHANG and CASEY 1996 ; SEABRA 1998 ). Subsequent to this modification further lipid modifications may occur, and in most cases, the final three amino acids are removed. These modifications are required for proper association of the small GTPase and the membrane; without this association, the GTPase is nonfunctional. These functional relationships have been demonstrated for numerous Ras superfamily members, including Rho. Site-directed mutagenesis that changes the CaaX box cysteine to serine of the S. cerevisiae RhoA homolog, Rho1, results in the failure of the mutated Rho1 protein to repartition from the cytosolic compartment to the membrane (YAMOCHI et al. 1994 ). Further, these Rho1 mutant cells fail to grow. In mammalian tissue culture, CaaX box-mutated RhoB cannot be lipid modified, and these cells lose their ability to become transformed in sensitized backgrounds (LEBOWITZ et al. 1997 ). Therefore, it is likely that the RhoAE3.10-encoded protein cannot be post-translationally modified, resulting in a complete loss of RhoA function. Similarly, the nonsense mutation at residue 180 in the J3.8 allele would remove the CaaX box and an additional nine amino acids and, therefore, would also behave as a severe RhoA allele.
However, on the basis of the differences observed in their genetic interactions with Df(2R)Jp1 and their levels of reduced viability in trans to zipEbr, RhoAE3.10 appears to be a more severe allele than RhoAJ3.8. We hypothesize that the protein encoded by RhoAE3.10 may have a partial dominant-negative effect because it does not repartition properly. On the other hand, the premature stop codon in RhoAJ3.8. may give rise to an unstable gene product. Since appropriate antibodies directed against Rho are not yet available, we cannot adequately evaluate this alternative.
Myosin, RhoGEF2, and RhoA function in multiple morphogenetic processes:
Studies reveal that multiple processes require myosin function throughout Drosophila development, including oogenic cell migrations, larval cytokinesis, and imaginal disc morphogenesis (GOTWALS and FRISTROM 1991 ; KARESS et al. 1991 ; EDWARDS and KIEHART 1996 ; HALSELL and KIEHART 1998 ; this study). Strong or null alleles of zipper are embryonic lethal, fail during dorsal closure, and give rise to embryos with dorsal cuticular holes (YOUNG et al. 1993 ). Additionally, myosin immunolocalization studies suggest that myosin is required during stages not yet tested functionally, including embryonic cellularization and gastrulation (YOUNG et al. 1991 ). RhoGEF2 and RhoA also function at least during a subset of the morphogenetic processes that require myosin.
Mutations in the Drosophila RhoGEF2 gene have been identified by three distinct means: phenotypic suppression of ectopically expressed RhoA (BARRETT et al. 1997 ), genetic screens for maternally encoded molecules required during early Drosophila embryogenesis (PERRIMON et al. 1996 ; HACKER and PERRIMON 1998 ), and genetic screening for molecules required for myosin function (this study). Maternal depletion of RhoGEF2 results in defects during gastrulation (BARRETT et al. 1997 ; HACKER and PERRIMON 1998 ). Specifically, embryos lacking maternal RhoGEF2 fail during apical constriction of ventral furrow cells. Interestingly, myosin localizes to the apical ends of these ventral furrow cells (YOUNG et al. 1991 ; LEPTIN et al. 1992 ; P. E. YOUNG and D. P. KIEHART, unpublished observations). This observation coupled with the genetic interaction between RhoGEF2 and myosin during leg morphogenesis suggests that RhoGEF2 may exert some of its effect during gastrulation via the activity of myosin in these cells.
RhoA mutations are recessive embryonic lethals. Zygotic depletion of RhoA results in an anterior dorsal hole in the cuticle (Fig 5; STRUTT et al. 1997 ). This defect has been characterized as a dorsal closure phenotype. Dorsal closure is an embryonic morphogenetic event in which the lateral epidermis moves over the dorsal side of the embryo, ultimately fusing along the midline (YOUNG et al. 1993 ; CAMPOS-ORTEGA and HARTENSTEIN 1997 ). If dorsal closure fails, then cuticular holes result. Typically, these holes are more posteriorly localized than those observed in RhoA mutants. However, certain zipper alleles give rise to cuticular holes that extend from the posterior one-third of the embryo to the anterior end (S. R. HALSELL, unpublished observations). These extensive cuticular holes are consistent with the head involution defects observed in zipper mutants and may reflect combined defects in head morphogenesis and dorsal closure. Therefore RhoA loss-of-function mutations may more accurately represent a particular sensitivity in head morphogenesis to perturbation rather than being dorsal closure mutants per se.
Nonetheless RhoA function during dorsal closure has been implicated by analysis of embryos expressing dominant negative RhoA transgenes (HARDEN et al. 1999 ; LU and SETTLEMAN 1999 ). In wild-type embryos, the leading-edge cells and the adjacent lateral cells elongate during dorsal closure (YOUNG et al. 1993 ). When dominant-negative RhoA is driven in the leading edge by utilizing the GAL-4 UAS system (BRAND and PERRIMON 1993 ), stretching of the leading cells initiates but is ultimately lost, and the lateral cells never elongate (LU and SETTLEMAN 1999 ). The Jun-kinase signal transduction cascade acts during dorsal closure and induces expression of the TGFß gene, decapentaplegic (dpp), in the leading-edge cells (reviewed in NOSELLI 1998 ). Leading-edge dpp expression is a prerequisite for elongation of the flanking lateral cells. In the dominant-negative RhoA embryos, dpp expression is wild type, therefore the authors suggest that RhoA acts upstream of a separate transcriptional pathway (LU and SETTLEMAN 1999 ). On the basis of our and other data, we suggest that RhoA may function directly upstream of myosin in the leading edge. First, we show here that RhoA signaling is necessary for myosin-driven cell shape changes during leg imaginal disc morphogenesis. Second, zipper mutants lose myosin in the leading-edge cells, and, subsequently, the leading-edge cells fail to elongate (YOUNG et al. 1993 ). Finally, myosin is delocalized in leading-edge cells expressing dominant negative RhoA (HARDEN et al. 1999 ). Taken together, these results suggest that RhoA signaling may have a direct cellular output at the level of myosin activity in the leading-edge cells and may not exert its effect via a transcriptional pathway.
Evidence that actomyosin dynamics are regulated by Rho:
Numerous pharmacological, cell culture, and biochemical studies implicate the Rho subfamily of GTPases as signal transducers upstream of actin cytoskeleton rearrangements and myosin regulation (reviewed in VAN AELST and D'SOUZA-SCHOREY 1997 ). In Drosophila, injection of mutant forms of Rho or Cdc42 proteins induces gross malformations in the actomyosin cytoskeleton, disrupting a specialized embryonic cytokinesis known as cellularization (CRAWFORD et al. 1998 ). When dominant-negative Rac1 is expressed at later stages of embryogenesis, the actomyosin cytoskeleton is disrupted in the leading-edge cells during dorsal closure (HARDEN et al. 1995 ). In Swiss 3T3 cells, the Rho GTPase induces the formation of actin stress fibers (RIDLEY and HALL 1992 ). Further, it has been demonstrated that contractility of the actin cytoskeleton, presumably mediated by myosin, is required for stress fiber formation and that this contractility is downstream of Rho signal transduction (CHRZANOWSKA-WODNICKA and BURRIDGE 1996 ).
In metazoans, nonmuscle myosin and smooth muscle-based contractility depend on the phosphorylation state of the noncovalently bound regulatory light chain (reviewed in TAN et al. 1992 ; JORDAN and KARESS 1997 ; J. CRAWFORD, K. A. EDWARDS and D. P. KIEHART, unpublished results). Molecularly, activated Rho may modulate the phosphorylation state of the regulatory light chain. Biochemical analysis reveals that activated Rho binds and activates a variety of effectors, including a group of serine/threonine kinases known as Rho kinase/ROK and p160ROCK/ROKß (LEUNG et al. 1995 , LEUNG et al. 1996 ; ISHIZAKI et al. 1996 ; MATSUI et al. 1996 ). In vitro biochemical assays reveal that Rho kinase can phosphorylate the regulatory light chain at its activating sites and induce myosin activity (AMANO et al. 1996 ; KUREISHI et al. 1997 ). Further, Rho kinases phosphorylate the myosin binding subunit of myosin phosphatase and thus repress its activity; the net result is a further increase in the phosphorylation state of the regulatory light chain (KIMURA et al. 1996 ).
Genetic screens in C. elegans indirectly implicate regulation of myosin by Rho:
Genetic screens for morphogenesis defects in C. elegans have identified mutations in loci encoding Rho signal transduction components (WISSMANN et al. 1997 ). Mutations in the C. elegans Rho kinase locus, let-502, disrupt embryonic elongation, while mutations in the regulatory subunit of the myosin phosphatase gene, mel-11, suppress the let-502 morphogenetic defect (WISSMANN et al. 1997 ). These results suggest that Rho signal transduction is upstream of myosin-driven morphogenesis in C. elegans. This hypothesis cannot be tested directly because myosin mutations that affect cell sheet morphogenesis have not been identified in C. elegans. Nonmuscle myosin is encoded by more than one locus and functional redundancy of these loci may preclude the isolation of morphogenetic myosin mutations. In contrast, we were able to screen for mutations that dominantly interact with the zipper locus that encodes the myosin heavy chain. Our approach biases our screens toward identifying loci whose gene products may directly affect myosin function. Thus, our studies provide genetic evidence consistent with Rho signal transduction acting directly upstream of myosin activity and bridge the previous in vitro and in vivo analyses.
Specificity of the RhoA and zipper interaction:
Our genetic analyses reveal that RhoA and RhoGEF2 genetically interact with zipper. Comparison of the cytogenetic locations of other loci encoding Rho subfamily components to genomic regions shown to uncover loci that interact genetically with zipper (HALSELL and KIEHART 1998 ) indicate that these loci probably do not interact in this screen on the basis of leg morphogenesis. This includes Rac1 (map position 61F5) and Rac2 (map position 66A1; HARDEN et al. 1995 ). Thus, null alleles of Rac are unlikely to interact with zipper. Drosophila Cdc42 maps to cytogenetic position 18E1-2, and two strong loss-of-function alleles and one dominant-negative allele of Cdc42 have been identified (FEHON et al. 1997 ). Neither of the two strong loss-of-function alleles of Cdc42 exhibit second-site noncomplementation with zipper (S. R. HALSELL, unpublished results). Interestingly, the dominant-negative allele of Cdc42 does interact with zipper (S. R. HALSELL, unpublished results). This result suggests that Cdc42 function may be required for leg morphogenesis or that the dominant-negative form may titrate factors shared with RhoA.
A closely related RhoA gene, Rho-like (RhoL), maps to cytogenetic interval 85D10-12 (SASAMURA et al. 1997 ). Previous screens for zipper interactions with this genomic region do not reveal an interacting locus (HALSELL and KIEHART 1998 ). Similarly, a Rho-type guanine exchange factor, rtGEF, also maps to a genomic region that is noninteracting with zipper (WERNER and MANSEAU 1997 ; HALSELL and KIEHART 1998 ). Finally, a putative Drosophila Rho/Rac effector, protein kinase N (pkn), has been identified (LU and SETTLEMAN 1999 ). With low penetrance, pkn mutants give rise to embryos with cuticular defects that resemble those of RhoA mutants (LU and SETTLEMAN 1999 ). This locus maps to 45C1, but our previous genomic screens did not test this region (HALSELL and KIEHART 1998 ). However, its binding to the GTP-bound forms of Rac1, Rac2, and RhoA and its slight effect on dorsal closure suggest that pkn may function as an effector of Rac or that it may be a redundant effector during dorsal closure. As such, it may not function during leg imaginal disc morphogenesis.
The genetic interaction between the RhoA mutant alleles and Df(2R)Jp1 suggests that additional components of the RhoA signaling pathway may be identified in future screens for the malformed leg phenotype. For example, Rho kinase represents a good candidate locus for interaction. Thus, the specificity of the interactions observed in this study suggests that future screens based on the malformed leg phenotype could identify gene products that function in concert with the RhoA signaling pathway during morphogenetic cell shape changes.
Conclusion:
Our genetic screening methods prove highly efficient in identifying mutations that disrupt morphogenesis and in linking them in cellular pathways. These genetic studies greatly extend biochemical and tissue culture analyses, providing direct biological relevance to these assays.
TOP
ABSTRACT
MATERIALS AND METHODS
