Roles for beta(pat-3) integrins in development and function of Caenorhabditis elegans muscles and gonads.

Heterodimeric integrin receptors for extracellular matrix (ECM) play vital roles in bidirectional signaling during tissue development, organization, remodeling, and repair. The beta integrin subunit cytoplasmic domain is essential for transmission of many of these signals and overexpression of an unpaired beta tail in cultured cells inhibits endogenous integrins. Unlike vertebrates, which have at least nine beta subunit genes, the nematode Caenorhabditis elegans expresses only one beta subunit (betapat-3), and a null mutation in this gene causes embryonic lethality. To determine the functions of integrins during larval development and in adult tissues, we have taken a dominant negative approach by expression of an HA-betatail transgene composed of a hemagglutinin (HA) epitope tag extracellular domain connected to the betapat-3 transmembrane and cytoplasmic domains. Expression of this transgene in muscle and gonad, major sites of integrin expression, caused a variety of phenotypes dependent on the level of transgene expression. Abnormalities in body wall and sex muscles led to uncoordinated movement and egg-laying defects. Significant anomalies in migration and pathfinding were caused by tissue-specific expression of HA-betatail in the distal tip cells (DTC), the cells that direct gonad morphogenesis. A pat-3 gene with Tyr to Phe mutations in the cytoplasmic domain was able to rescue pat-3 null animals but also showed DTC migration defects. These results show that betapat-3 plays important roles in post-embryonic organogenesis and tissue function.

The extracellular matrix (ECM) 1 is a complex network of glycoproteins and proteoglycans that determines tissue organization and function through interactions with cell surface receptors. Mechanical and chemical signals provided by the ECM are transduced primarily through integrin adhesion receptors (1)(2)(3). Integrins are transmembrane heterodimers composed of ␣ and ␤ subunits that play a central role in integrating bidirectional information between the ECM and the inside of the cell. Binding of integrins to ECM ligands leads to receptor clustering at focal adhesions where connections to the actin cytoskeleton are formed (4). In vitro binding studies and expression of mutant integrins have implicated specific ␤ cytoplasmic tail sequences in the linkage to actin filaments (5).
In the nematode Caenorhabditis elegans, two ␣ (PAT-2 and INA-1) integrins and one ␤ (PAT-3) integrin are expressed in many tissues, including muscles, neurons, and gonad (6 -8). In body wall muscles, ␤pat-3 integrins are localized to muscle attachment sites called dense bodies (9). Attachment of muscle cells to the nearby hypodermis is essential for muscle filament assembly and function and attachment defects caused by ␣pat-2 and ␤pat-3 loss-of-function alleles give a paralyzed and arrested at two-fold (Pat) embryonic phenotype (10). In addition to integrins, many cytoskeletal and signaling molecules such as DEB-1/vinculin (11), talin (12), and UNC-97/PINCH (13) are colocalized to muscle dense bodies showing their similarity to vertebrate focal adhesions.
␣ and ␤ cytoplasmic tails appear to regulate each other's activities and, in the absence of ligand, the ␤ integrin tail is unable to interact with cytoskeletal components (14). However, the ␤ integrin cytoplasmic domain alone contains sufficient information for focal adhesion localization of single subunit chimeric proteins consisting of the ␤ cytoplasmic domain attached to a heterologous extracellular and transmembrane domain (15). Furthermore, these chimeric ␤ integrins induce FAK phosphorylation and inhibit endogenous integrin functions such as cell adhesion, migration, and matrix assembly (15)(16)(17)(18)(19). In vivo, a similar chimeric ␤ integrin was shown to perturb mammary gland development (20). These observations indicate that the autonomous expression of the ␤ integrin cytoplasmic domain acts as a dominant negative inhibitor of endogenous integrins by effects on intracellular pathways.
Loss of ␤ 1 integrin function in mammals (21,22) and related integrins ␤ PS in fruit flies (23) and ␤pat-3 in nematodes (10) causes early embryonic lethality. In C. elegans, lethality results from defects in muscle cell attachment to ECM and in muscle filament assembly. The dearth of non-lethal mutations in pat-3, the sole ␤ integrin gene in C. elegans (8), has prevented the identification of roles for this integrin in other tissues and at larval and adult stages. Furthermore, known non-lethal mutations in C. elegans integrin genes reside in the extracellular domains (7,24) and probably affect ligand binding or protein stability precluding analysis of changes in intracellular signaling. In this study, we have disrupted integrin functions by expression of the PAT-3 cytoplasmic tail and by rescue of pat-3 embryonic lethality with various mutant pat-3 alleles. Phenotypes were characterized in C. elegans larvae and adults. We find that integrin cytoplasmic domain function is important for development and function of multiple nematode tissues throughout the life cycle.

EXPERIMENTAL PROCEDURES
Production of ost-1::HA-␤tail Plasmids and Transgenic Nematodes-Primers homologous to positions 2228 -2245 and 2445-2464 and spanning the transmembrane and cytoplasmic domains of the ␤pat-3 cDNA (6) were used in PCR amplification with the cosmid ZK1058 (25) as a template. The primers introduced an NheI site at the 5Ј-end and a PstI site at the 3Ј-end. The resulting 289-bp NheI-PstI PCR product, which includes a 53-bp intron between 2388 and 2389, was ligated to the 3Ј-end of a 96-bp XbaI fragment containing three tandem hemagglutinin (HA) epitope tags (26). The XbaI-PstI fragment was completely sequenced and inserted along with a PstI-XbaI fragment containing the unc-54 3Ј-UTR and transcription termination sequences (27) at an XbaI site between exon 1 and intron 1 of the plasmid pOST-X1. ost-1 regulatory sequences promote expression in muscles and gonads (28,29). Site-directed mutagenesis of ost-1::HA-␤tail to generate plasmids encoding HA-␤⌬ and HA-␤⌬15 used degenerate oligonucleotide primers to insert TAA termination codons at the indicated positions (see Fig. 1). HA-␤YYFF, HA-␤Y792F, and HA-␤Y804F were produced using degenerate primers with TAC to TTC codons changing tyrosine to phenylalanine.
HA-␤tail cDNA was also inserted downstream of the lag-2 promoter in the plasmid pJK590 (gift of Dr. Judith Kimble). Forward and reverse primers, 5Ј-ATGGTACCTCATTCGCCGTCCAAGATGCG3-Ј and 5Ј-AGGGTACCTTTTTTTCTAGAGCACGTACG3-Ј, respectively, were used in PCR amplification of the N-terminal HA tag, the transmembrane and cytoplasmic domains of ␤pat-3, and unc-54 3Ј-UTR and transcription termination sequences using pOST-HA␤ as a template. The primers introduced KpnI sites at the 5Ј-and 3Ј-ends. This PCR product was ligated into a KpnI site between the lag-2 promoter and the GFP coding sequence. The resulting construct, pLAG-HA␤, does not express GFP.
Nematode Culture and Analysis of Phenotypes-Nematodes were cultivated on nematode growth medium agar plates with OP50 bacteria according to standard techniques (32). To characterize phenotypes caused by transgene expression, F2 progeny of F1 Rol transgenic hermaphrodites were scored. For each injected plasmid, progeny from two or more F1 Rol adults were scored for transgenic phenotypes. Phenotypes included Pat (paralyzed arrested at two-fold) embryos, Unc (uncoordinated) animals, which became paralyzed during larval or adult stages, Egl (egg-laying defective) animals with the "bag of worms" phenotype, and Emb eggs, which did not hatch and were arrested before the two-fold stage.
To analyze gonad morphology, young adult transgenic hermaphrodites were mounted in a drop of M9 buffer containing 0.25 mM levamisole (Sigma Chemical Co.) on a coverslip coated with 2% agarose and examined using a Nikon Diaphot microscope with DIC optics. Images were captured with an NEC video camera using IMAGE (version 1.61, National Institutes of Health) software. DNA organization in the gonads was monitored by staining transgenics with DAPI (0.1 g/ml) and visualizing them with a Nikon Optiphot microscope equipped for epifluorescence. Images were captured using an Optronics cooled chargecoupled device camera and Adobe PhotoShop (version 5.0). DTC migration defects and oocyte accumulation in the proximal gonads were identified using DIC microscopy of young adult hermaphrodites. Defects such as inappropriate or extra turns, migration in the wrong direction, or aberrant stops were counted as DTC migration anomalies. Oocyte accumulation was scored when oocytes were present in a non-linear arrangement in the proximal gonad.
Rescue of pat-3-A 12-kb genomic DNA fragment containing the entire pat-3 gene plus 4.8 and 1.7 kb of 5Ј-and 3Ј-flanking sequence, respectively, was excised from cosmid ZK1058 with PstI and BsrB1 restriction enzymes and cloned between PstI and SmaI sites of pSP73 to generate pPAT3-PB12K. Site-directed mutagenesis of pPAT3-PB12K to generate pPAT3-Y804F and pPAT3-YYFF was performed using the overlap extension method. Briefly, pOST-HA␤Y804F and pOST-HA␤YYFF were used as templates for PCR amplification of the segment encoding the mutant cytoplasmic tails. A second PCR product encoding the 3Ј-UTR of pat-3 was amplified. After overlap extension, the product was digested with MscI and EcoRI, and the resulting 754-bp MscI-EcoRI fragment was inserted into pPAT3-PB12K.
Protein Expression and Localization-To determine the level of transgene expression, five transgenic adults from each phenotypic category (Unc, Egl, and Rol) were lysed in SDS sample buffer and analyzed by SDS-polyacrylamide gel electrophoresis and immunoblotting. Ascites fluid containing 12CA5 anti-HA antibodies was used at a dilution of 1:1000 and detected with horseradish peroxidase-conjugated goat antimouse IgG and ECL reagents (Pierce).
To localize HA-␤tail, N2 and transgenics were frozen on poly-Llysine-coated slides and fixed with methanol and acetone (Ϫ20°C) for 5 min. Fixed animals were stained with 12CA5 anti-HA antibodies (1:400 dilution) or MH25 (1:250 dilution) anti-PAT-3 antibodies (34) for 4 h at room temperature followed by fluorescein isothiocyanate-conjugated goat anti-mouse IgG (Molecular Probes) at a dilution of 1:750 overnight at 4°C. Cytoskeletal organization of muscle cells was examined by staining fixed animals with 0.2 unit/ml rhodamine-conjugated phalloidin (Molecular Probes) for 2 h at room temperature. HA-␤tail, PAT-3, and actin filaments were visualized with a Nikon Optiphot microscope equipped for epifluorescence.

Expression of an HA-␤tail Transgene Induces
Integrin-associated Phenotypes-To determine the tissue-specific functions of ␤pat-3 cytoplasmic domain interactions in C. elegans, a chimeric transgene was prepared containing the PAT-3 transmembrane and cytoplasmic domains connected to a heterologous hemagglutinin extracellular domain. The transgene was inserted into ost-1 in-frame with and immediately following the first exon, which encodes the signal sequence and first five amino acids of the extracellular matrix protein SPARC (29). SPARC is expressed at high levels in nematode gonads and body wall and sex muscles in embryonic, larval, and adult animals (28,29). These tissues are also major sites of pat-3 expression (6). Thus, this system will allow us to examine the transdominant effects of HA-␤tail transgene expression on endogenous integrin functions. In addition to HA-␤tail, we also generated HA-␤⌬ lacking the entire ␤pat-3 cytoplasmic domain and HA-␤⌬15 missing the C-terminal 15 amino acids of the tail (Fig. 1). Transgenic nematodes were generated by co-injection of individual HA-␤tail constructs with pRF4 containing rol-6(su1006dm) as dominant marker into N2 wild type nematodes. Transgenic F1 animals were allowed to self-fertilize, and F2 progeny were examined for phenotypes associated with expression of the HA-␤tail transgene. Phenotypes of varying severity were observed in more than half of the transgenics (Fig. 2, Table I). In the most severe defect, embryos showed a Pat phenotype with arrested elongation at about the 2-fold stage ( Fig. 2A). The Pat phenotype paralleled the defects observed in pat-3 animals indicating that overexpression of the transgene has a dominant negative effect on the function of endogenous integrins. More than a quarter of the transgenics were uncoordinated (Unc) and developed a progressive dystrophy and adult paralysis with significantly reduced mobility throughout larval and adult stages. The least severe overt phenotype was an egg-laying (Egl) defect. Sex muscles appeared properly formed and localized as viewed by fluorescent phalloidin staining but were dysfunctional in that egg laying was not increased after stimulation with serotonin. Pat, Unc, and Egl phenotypes result from defects in tissues coexpressing endogenous integrins and the HA-␤tail transgene. Mutations in other integrin-associated proteins such as vinculin (11), perlecan (35), and laminin (36) cause similar phenotypes.
To verify the specificity of the defects, HA-␤⌬ and HA-␤⌬15 F2 animals were also analyzed. More than half of the HA-␤tail transgenic nematodes had either Unc or Egl phenotypes with a lesser number of Pat. In the complete absence of a ␤ cytoplasmic domain, HA-␤⌬ transgenic animals showed no significant phenotypes (Table I). In cultured cells, deletion of the C-terminal 15 amino acids of the ␤ 1 subunit results in decreased localization to focal adhesions (37). A lower percentage of HA-␤⌬15 animals displayed Pat, Unc, and Egl phenotypes compared with HA-␤tail F2 progeny (Table I). These results demonstrate a requirement for the cytoplasmic tail in development of these phenotypes and show that a deletion that can affect integrin subcellular localization yields less dramatic effects.
Phenotype Severity Correlates with Transgene Expression Levels-The relationship between transgene expression and phenotype was determined by examining the level of HA-␤tail protein in animals with different phenotypes. Equal numbers of Unc, Egl, Rol, and wild type N2 animals were picked, lysed in SDS gel sample buffer, and used for immunoblots with an anti-HA antibody. Fig.  2B shows that Unc animals had the highest level of HA-␤tail protein whereas lower levels were found in Egl and still lower in Rol animals. Thus the severity of the phenotype correlates with the level of transgene expression. Two HA-␤tail-integrated lines were established by gamma irradiation, mwIs25[ost-1::HA-␤tail] and mwIs34[ost-1::HA-␤tail]. Although mwIs25 animals had a low penetrance of overt phenotypes, mwIs34 transgenics showed reduced fecundity and severe defects in morphogenesis and locomotion, typical of Unc animals. The expression levels of HA-␤tail in these established lines also corresponded with phenotype severity, because mwIs34 had a significantly higher level of HA-␤tail protein than mwIs25 (not shown).
HA-␤tail Affects Muscle Cell Organization-HA-␤tail protein was distributed throughout the body wall muscle in alignment with parallel muscle filaments in mwIs34 Unc animals (Fig.  3A). Although some diffuse staining was apparent, there was also a dotted pattern indicative of localization to dense bodies. A regularly repeated pattern of dense body staining was also observed in wild type nematodes stained for endogenous integrins with the MH25 monoclonal antibody against the PAT-3 extracellular domain (Fig. 3B). In HA-␤tail transgenic muscle, endogenous ␤pat-3 localization appeared more irregularly spaced and less well ordered than in wild type (Fig. 3, B and C) suggesting an effect of the transgene product on dense body organization. mwIs34 muscles also had severely disorganized and unevenly distributed actin filaments (Figs. 4, A and B). Together, these results show that the HA-␤tail protein disrupts endogenous integrin and body muscle filament organization, at least in part through co-localization to dense bodies.
Integrin Activity Is Required for Gonad Development and Function-The adult hermaphrodite gonad consists of two Ushaped tubular arms surrounded by sheath cells and connected proximally to the spermathecae (38). At the distal end of each arm, the distal tip cell (DTC) navigates gonad migration during development and determines germ cell polarity from proliferative in the distal arm through meiosis to oocyte formation in the loop region (39) (Fig. 5A). Comparison of gonads from mwIs34 nematodes with wild type N2 and HA-␤⌬ animals demonstrated that HA-␤tail expression induced significant defects in gonad morphology. In one-third of HA-␤tail hermaphrodite gonad arms, misdirected DTC migration along the dorsal trajectory was observed (Fig. 5, B and C, Table II). Typical pathfinding defects included extra turns, dorsal to ventral migration, looping back, and posterior instead of anterior migration. Gonad elongation along the ventral body wall in response to UNC-6 cues appeared normal as did the change in trajectory from ventral to dorsal.
DTC pathfinding defects could arise from disruption of integrin functions within the DTC or from effects on the cells that form the DTC migratory path. To determine whether the HA-␤tail effects on DTC migration are cell autonomous, transgenic animals were prepared with the HA-␤tail transgene under control of the DTC-specific promoter from lag-2. lag-2::HA-␤tail animals also showed a significant proportion of DTC migration defects (Table II, Fig. 5D). Together, these results indicate that endogenous integrin function is required in the DTC for pathfinding along the dorsal body wall during gonad morphogenesis.
Gonad migration defects have also been attributed to certain mutant alleles of ina-1, which encodes an ␣ integrin subunit (7), and ced-10, which encodes an Rac GTPase (40). Comparison of gonad morphologies of ced-10 (not shown) and ina-1 mutant  hermaphrodites (Fig. 5E) to mwIs34 animals showed similar pathfinding defects along the dorsal body wall. Furthermore, similar percentages of defective gonad arms were counted for transgenic and mutant hermaphrodites (Table II). The similarities in gonad morphology between ina-1 mutants and the HA-␤tail overexpression phenotype indicate that integrins are important for normal DTC pathfinding. Furthermore, perturbation of either ligand binding, as in the ina-1 mutant, or integrin-mediated intracellular signaling with HA-␤tail expression caused similar defects.
Gonad function was also compromised by blocking endogenous integrins. More than 80% of mwIs34 animals showed oocyte accumulation in the proximal gonad (Fig. 6A). A corresponding defect in DNA organization was observed in these oocytes by DAPI staining. Wild type germ nuclei in the pachytene stage showed the characteristic DNA staining pattern (38) and mwIs34 germ cell DNA appeared normal in the distal gonad arm (Fig. 6, B and C). In sharp contrast, proximal arms were filled with clumped nuclear staining (Fig. 6C) typical of polyploid nuclei that result from failure of oocytes to progress through the spermatheca during ovulation (41). As   both pat-3 and ost-1::HA-␤tail have been shown to be expressed in the spermatheca (6,28), this phenotype could result from defective dilation of the spermatheca at ovulation.
Cytoplasmic Tyrosine Residues Contribute to PAT-3 Function-Integrins have been shown to be directly involved in cell migration in cell culture systems and may function similarly during DTC migration and gonadogenesis. Two conserved tyrosine residues in the cytoplasmic tail have been implicated in regulating chemotaxis (42). Both tyrosines reside in the phosphorylation motif NPXY, and vertebrate ␤ 1 and ␤ 3 integrins can be phosphorylated on tyrosine under appropriate conditions (43)(44)(45)(46)(47). To determine whether these tyrosines are important for migration of the DTC, pat-3-null animals were rescued with a wild type pat-3 transgene or a pat-3 transgene mutated at both tyrosines or at Y804 alone (Fig. 1). pat-3(YYFF) and pat-3(Y804F) rescued animals showed significant levels of DTC pathfinding defects, 49 and 52% respectively, whereas mwEx443 animals rescued with the wild type pat-3 had low levels of DTC pathfinding defects (Table II). In addition, 19% of pat-3(YYFF) adult hermaphrodites had a "notched head" Unc phenotype. These results suggest that cytoplasmic domain tyrosines play a role in integrin-mediated signaling during gonad and body wall morphogenesis.
In contrast to the loss-of-function phenotypes with mutations in pat-3, similar mutations in the HA-␤tail transgene appeared to have a gain-of-function effect. HA-␤YYFF and HA-␤Y792F progeny developed a severe embryonic lethal (Emb) phenotype that stopped development earlier than the Pat phenotype observed with HA-␤tail. As many as 50% of offspring died as embryos after the start of morphogenesis (Table I). This observation suggests that, in the absence of these tyrosines, a gainof-function phenotype increases the activity of the integrin cytoplasmic tail thus improving its ability to disrupt endogenous integrin functions during early morphogenesis. Surviving larvae and adults developed the same phenotypes as HA-␤tail transgenics, including Unc, Egl, and DTC migration defects (Tables I), indicating that this gain-of-function phenotype does not have a detectable impact on later stages of the nematode life cycle. DISCUSSION Overexpression of integrin ␤ tails has been shown to affect intracellular activities of endogenous integrins in cultured cells and mouse mammary tissue, making this a feasible approach to examine the effects of reduction in integrin signaling in other organisms. Loss of function of the C. elegans ␤ integrin (pat-3) causes an embryonic lethal phenotype. Therefore, to determine the potential effects of loss of integrin activity in larval and adult stages, we applied the transdominant inhibition approach by expression of an HA-␤tail transgene in muscles and gonad, major sites of integrin expression. Fitting with a role for this transgene product in inhibiting endogenous integrins, we found co-localization of HA-␤tail with PAT-3 in muscle dense bodies, a requirement for the intact ␤ cytoplasmic domain, defects in muscle filament organization and cell migration during gonad morphogenesis, and correlation between severity of the phenotypes and level of transgene expression. In addition, our results provide new information regarding integrin function in C. elegans: 1) integrins are important for directing DTC migration, and conserved tyrosine residues in the cytoplasmic tail of PAT-3 are necessary for correct morphogenesis of the distal arm of the gonad, 2) integrin function is also required for normal oocyte development and progression through the spermatheca, and 3) tyrosine to phenylalanine mutations in HA-␤tail lethal phenotype suggesting that these residues have a role in integrin function during early embryogenesis.
HA-␤tail expression in body wall and sex muscles appears to affect mainly cytoskeletal organization and muscle filament contraction. These defects could result from reduced muscle cell attachment to the basement membrane in correspondence with the deficiencies reported for expression of similar chimeric ␤ tails in cultured cells (5). Localization of the transgene product to dense bodies, the C. elegans muscle equivalent of the focal adhesion, positioned it appropriately to affect intracellular interactions with cytoskeletal-associated proteins. However, muscle cell adhesion was not completely ablated as actin filaments were present, albeit disorganized, and body wall muscle cells showed the typical elongated spindle shape. Mosaic analysis showed that pat-3(-) muscles also had disorganized filaments, which were accompanied by characteristic body deformities (24). An additional factor that may contribute to the muscle defects pertains to their role in basement membrane assembly. These cells express many basement membrane components, including perlecan (35), type IV collagen (48), SPARC (29), and laminin (36). Inhibition of ␣ 5 ␤ 1 integrin function in fibroblasts blocks fibronectin matrix assembly (15), and laminin was inappropriately deposited between cells in mammary glands expressing a ␤ 1 integrin chimera (20). This raises the possibility that HA-␤tail may perturb basement membrane assembly sufficiently to reduce muscle cell adhesion, especially during movement or egg laying when muscle cells are under significant tension.
Gonad morphogenesis in C. elegans is dependent on regulated DTC migration and pathfinding taking cues from the body wall ECM (39, 41). HA-␤tail transgenic animals showed misguided migration of DTCs. Unlike the complete block of migration on fibronectin by expression of a ␤ 1 chimera (15), in transgenic nematodes, the DTC showed significant levels of migration but with a major defect in pathfinding. This phenotype resulted from a deficiency within the DTC itself as gonad morphogenesis in lag-2::HA-␤tail transgenic nematodes was also impaired. Mutations in INA-1, an ␣ integrin subunit that is expressed in nerve cells and the DTC (7), show a similar deficiency further strengthening the role of integrins in gonad morphogenesis.
Our results also show a role for the cytoplasmic domain tyrosine residues in DTC migration and gonad morphogenesis. . This arrangement contrasts with wild type oocytes, which are found in a linear arrangement (see Fig. 5A). B, DAPI staining was used to visualize normal DNA organization in the germ cells and oocytes (arrows) of wild type N2. C, in mwIs34 transgenic animals, distal germ cell DNA appears normal but oocyte DNA was found in large clumps in the proximal gonad (arrows). Mislocalization of the DTC is shown (arrowhead).
Compared with nematodes expressing wild type pat-3, mutant pat-3(YYFF) or pat-3(Y804F) expression induced significant levels of aberrant DTC migration. Because no other overt defects were observed in adult hermaphrodites, these tyrosine residues appear to function in a tissue-specific manner in C. elegans. Cell-and tissue-specific activities have also been attributed to these tyrosines in mammalian ␤ 1 and ␤ 3 integrins where these residues have been shown to participate in chemotactic migration in fibroblasts (42), in activation-dependent lymphoid cell adhesion (49), and in platelet function during hemostasis (47).
Proper formation of the U-shaped gonad arms also depends on signals from the secreted proteins UNC-6/netrin (50), UNC-129/TGF-␤ (51), and the matrix metalloproteases GON-1 (52) and MIG-17 (53). Both UNC-129 and MIG-17 play significant roles in migration along the dorsal body wall. HA-␤tail-induced defects are also localized to the dorsal region suggesting that signals from the secreted proteins may be integrated with DTC responses to integrin ligation. Thus, HA-␤tail expression could affect the ability of the DTC to integrate signals received from multiple sources, including endogenous integrins. It is also possible that this transgene changes DTC interactions with the basement membrane resulting in an inability to read environmental cues exposed by matrix metalloproteases.
The defects in DTC-directed migration induced by HA-␤tail and by mutants pat-3(YYFF) and pat-3(Y804F) resemble the effects of mutations in ced-2, ced-5, and ced-10, signaling molecules involved in cell death and related to mammalian CrkII, DOCK180, and Rac, respectively (40,54). These three genes constitute a pathway that regulates cell surface and/or cytoskeletal changes required for engulfment of cell corpses and that is recruited by the integrin ␣ v ␤ 5 during phagocytosis by mammalian cells (55). This pathway is also activated during DTC migration, and our results indicate that this ced pathway also regulates ovulation. In the proximal gonad arms, oocyte accumulation is due to aberrant ovulation, which usually results from defects in spermatheca dilation (41). Dilation is induced by signals from the maturing oocyte (56) implicating ␤pat-3 integrins in transmitting or receiving these activating signals. Hermaphrodites expressing both a mutant CED-5 and HA-␤tail showed a significant enhancement in oocyte accumulation in the proximal gonad arms compared with parental animals with either gene alone. 2 Apparently, integrin-mediated signals are integrated with a CED-5/DOCK180 pathway during ovulation. Although tyrosine to phenylalanine mutations in pat-3 caused a loss-of-function phenotype, the same mutations in HA-␤tail had no detectable untoward effects on DTC migration. In contrast, there appeared to be a gain-offunction embryonic phenotype characterized by early embryonic death. This observation suggests that the absence of these tyrosines in the HA-␤tail protein improves the ability of the transgene product to disrupt endogenous integrin functions, possibly through titration of intracellular components, an effect that only has obvious consequences during early morphogenesis.
In addition to the ␤ integrin requirement during embryogenesis, our results show that ␤pat-3 plays important roles in post-embryonic organogenesis and tissue functions. Furthermore, tyrosine mutations indicate that interactions involving these residues are important mediators of tissue-and stagespecific functions. This information will be very useful in developing animals with cytoplasmic tail mutations in endogenous integrins and in identifying other mutations that affect integrin signaling.