The Drosophila gene twins encodes the regulatory B subunit of type 2A protein phosphatase. Here we report that its partial loss-of-function mutations caused abnormal morphogenesis in the adult peripheral nervous system. In wild-type flies, the mechanoreceptor, one major class of sensory organs, is composed of four specialized cells (one neuron and three accessory cells) that are derived from a single precursor cell. The hypomorphic twins mutations did not block division of this precursor, but most likely altered cell fate in this lineage to produce only accessory cells that form sensory structures. Stepwise reductions of twins protein enhanced this transformation. In these mutants, another regulatory subunit, A, and the catalytic subunit, C, of the phosphatase were expressed at normal levels. Therefore, the modulation of the phosphatase activity by the B subunit appears to be crucial for specification of neural cell identity.

Protein phosphorylation and dephosphorylation are crucial post-translational modifications, and they have drastic effects on various cellular processes. In contrast to numerous protein kinases, serine/threonine protein phosphatases constitute only four major groups that are classified on the basis of their dependence on divalent cations and responsiveness to specific inhibitors (reviewed by Cohen, 1989; Cohen et al., 1990; Peruski et al., 1993). Biochemical characterization of one of these groups, type 2A protein phosphatase (PP2A), revealed several different subunits. The common core comprises a 36×103Mr catalytic subunit (C subunit) and a 65×103Mr regulatory subunit (A subunit). One additional modulator, B (55×103Mr), B′ (54×103Mr) or B″ (72×103Mr), can bind to the AC complex. The B subunit plays a steering role in the catalytic activity; toward various substrates studied, it enhances or suppresses the dephosphorylation action of the enzyme (Imaoka et al., 1983; Usui et al., 1988; Kamibayashi et al., 1991; Agostinis et al., 1992).

PP2A displays a broad substrate specificity in vitro, and is assumed to carry out pleiotropic actions in vivo (Lee et al., 1991; Clarke et al., 1993; Ferrigno et al., 1993; Sontag et al., 1993). Biochemical knowledge alone is not sufficient for an understanding of the specific cellular functions of PP2A, because PP2A and other groups of protein phosphatases share many substrates in vitro. Through genetic approaches carried out primarily in yeast, distinct physiological roles of PP2A have been demonstrated; these include control of cell cycle, cellular morphogenesis, and transcription (Kinoshita et al., 1990; Healy et al., 1991; Ronne et al., 1991; Van Zyl et al., 1992; Kinoshita et al., 1993; Blacketer et al., 1993). These experimental systems, however, do not allow us to test the function of PP2A in the development of multicellular organisms. One can imagine that protein dephosphorylation may play critical roles in inter- and intracellular signaling necessary for cell differentiation and other developmental processes. In fact, three Drosophila loci encoding protein phosphatases were found to be essential for establishing or maintaining ordered tissue structures. One is retinal degeneration C (rdgC), which is necessary to prevent light-induced retinal degeneration (Steele et al., 1992). Its product has a central domain that shares 30% identity with types 1, 2A and 2B serine/threonine protein phosphatases. Another is corkscrew (csw), a maternally derived gene required for normal determination of cell fates at the termini of the embryo (Perkins et al., 1992). The csw protein is a putative nonreceptor protein tyrosine phosphatase. The third gene, twins (tws), was identified through our previous mutant search for malformation of imaginal discs, and it encodes the regulatory B subunit of PP2A (Uemura et al., 1993).

The original tws allele was induced by insertion of a P element construct, PlacW (Bier et al., 1989), and designated as twsP. This is the strongest loss-of-function allele isolated to date, although it seems to retain a very low level of tws+ activity. The homozygotes die at an early pupal stage, and an abnormal pattern formation is observed in the imaginal discs. The imaginal disc is a closed monolayered epithelium that grows in larvae and produces adult structures at metamorphosis (reviewed by Bryant, 1978; Whittle, 1990). In the wing disc, the twsP mutation leads to a mirror-symmetrical duplication of the wing anlage, and this extra structure appears to be formed at the expense of a small domain in the future notum. A mutation in the same gene was independently isolated and found to cause aberrant mitosis of neuroblasts in the larval central nervous system (Mayer-Jaekel et al., 1993). This mutation is called abnormal anaphase resolution1 (aar1, Gomes et al., 1993). These results indicate differential effects of tws/aar mutations on various cell types.

By remobilization of PlacW of twsP, we isolated weaker alleles and examined their phenotypes with regard to morpho-genesis. Here we report a highly local pattern duplication in the peripheral nervous system in homozygous hypomorphic adults, that is, twinning of the external sensory structures (bristles). The development of sensory organs (sensilla) can be divided into three steps (reviewed by Ghysen and Dambly-Chaudière, 1989; Jan and Jan, 1990). First, precursor cells appear on the epidermal cell layer in a characteristic pattern, and then they undergo a number of mitotic divisions. Finally each of the progeny cells differentiates into an individual specialized type. One major class of sensilla is the mechanoreceptor, which consists of one neuron and three accessory cells (Fig. 1A; Lawrence, 1966; Zacharuk, 1985). Two of the three are the support cells that produce either a shaft or a socket of the bristle, and the third accessory cell is the sheath cell. Each mechanoreceptor in most categories is composed of sister cells derived from a single precursor (Bodmer et al., 1989; Harten-stein and Posakony, 1989; Huang et al., 1991). Our analyses at the level of single cells suggest that the tws mutations affect this lineage of the sensillum development, making progeny of the precursor cells enter only the differentiation pathway into support cells. This study provides the first evidence for an essential function of ser/thr dephosphorylation in establishing neural cell identities.

Fig. 1.

Duplication of mechanosensory bristles in the twins55 mutant. (A) Cellular architecture of an adult mechanoreceptor that has a bristle (modified from Uemura et al., 1989). A neuron extends a single dendrite, which innervates a socketed shaft in the cuticle. The dendrite is wrapped by a sheath cell (thecogen), indicated here by broken lines. DNA endoreplication occurs in two support cells that produce a shaft and a socket of the bristle (trichogen and tormogen, respectively), making them larger than a neuron or a sheath cell. (B) Nomarski microscope photographs showing mechanosensory bristles on the wild-type abdominal cuticle. Every bristle consists of a shaft and a socket. (C) Duplicated bristles in a homozygous tws55 adult. Many of the doubled sockets have fused to each other (two examples are indicated by arrows). Exceptionally, twin sockets do not touch each other (asterisk). Scale bar, 50 μm.

Fig. 1.

Duplication of mechanosensory bristles in the twins55 mutant. (A) Cellular architecture of an adult mechanoreceptor that has a bristle (modified from Uemura et al., 1989). A neuron extends a single dendrite, which innervates a socketed shaft in the cuticle. The dendrite is wrapped by a sheath cell (thecogen), indicated here by broken lines. DNA endoreplication occurs in two support cells that produce a shaft and a socket of the bristle (trichogen and tormogen, respectively), making them larger than a neuron or a sheath cell. (B) Nomarski microscope photographs showing mechanosensory bristles on the wild-type abdominal cuticle. Every bristle consists of a shaft and a socket. (C) Duplicated bristles in a homozygous tws55 adult. Many of the doubled sockets have fused to each other (two examples are indicated by arrows). Exceptionally, twin sockets do not touch each other (asterisk). Scale bar, 50 μm.

Fly stocks

The origin of tws alleles is described in Uemura et al. (1993). Df(3R)by62 uncovers a chromosomal region from 85D11-14 to 85F16 (Kemphues et al., 1983). To follow sensory organ precursors and their progeny in tws55, we made recombinant third chromosomes containing both the tws55 mutation and the B52 lacZ insert (Ueda et al., 1992). Recombinants between an R7 marker, XA12 (Van Vactor et al., 1991), and a pupal lethal allele, tws60, were made; and expression patterns of lacZ were studied in XA12 tws60 homozygotes or in XA12 tws60/twsP.

Histochemistry

Mouse monoclonal antibody 22C10 and rat polyclonal anti-Cut antibody F2 were provided from Shin Togashi and Karen Blochlinger (Blochlinger et al., 1990), respectively. Anti-HRP antibodies (Jan and Jan, 1982) and DCAT1 (Oda et al., 1993) were used to stain pho-toreceptor cells in eye imaginal discs. X-gal staining was performed essentially as described in Ueda et al. (1992) except that the reaction was done in the absence of Triton X-100. For bromodeoxyuridine (BrdU) labeling, we basically followed the protocols of Hartenstein and Posakony (1989) and Usui and Kimura (1992). We found that labeled cells of nonepithelial origin often adhered to the disc epithelium and obscured signals of sensory cells. Dead cells were identified by staining with acridine orange or trypan blue under the conditions described by Spreji (1971).

Polymerase chain reaction

Three primers were synthesized to amplify genomic DNA of the tws locus from single flies: primer no. 1 (5′-GAATTCTCGTTGCTTC-TAGTGCT-3′), primer no. 2 (5′-GAATTCCACAATACTCCACT-GTA-3′), and primer no. 3 (5′-GAATTCATCAATCTAGT-TGATAG-3′). A protocol for preparation of fly homogenates was provided by Gloor and Engels (University of Wisconsin, Madison). PCR was done under conditions of 35 cycles of denaturation at 95°C (1.5 minutes), annealing at 45°C (2 minutes), and extension at 72°C (3 minutes). We amplified cDNA clones of the A and C subunits of PP2A from an embryonic cDNA library (Zinn et al., 1988). Primers for amplification of the entire coding region of the C subunit were 5′-GAATTCATGGAGGATAAAGCAACA-3′ and 5′-GAATTCAAG-GAAATAATCGGGTGT-3′. A partial cDNA fragment of the A subunit (amino acids 335-591) was amplified with the following primers: 5¢-CGGGATCCTTGTCTCGGACCCCAAT-3′ and 5′-CGGGATCCCGCTGCAGCTATGCCGG-3′.

Antibodies against PP2A subunits

Rabbit antiserum against a gene10-tws fusion protein (TW1; amino acids 80-231) were previously described (Uemura et al., 1993). A GST-tws fusion protein (amino acids 80-231) was used for affinity purification of the rabbit antiserum. Rat antibodies were generated against inclusion bodies of TW-1. Antibodies to the catalytic subunit of PP2A were raised to a peptide (CPAPRRGEPHVTRRTPDYFL) that corresponds to the C-terminal amino acids except for its artificial cysteine. A 0.46 kb EcoRI fragment (amino acids 439-591) of the regulatory A subunit cDNA was subcloned into pGEMEX-1 (Promega), and the expressed fusion protein was injected into rats.

Hypomorphic alleles of twins are adult lethal

Previously, we carried out a reversion test of the original allele twsP and isolated adult lethal lines as well as revertants (Uemura et al., 1993). Since all of the adult lethal strains showed similar morphological defects in the peripheral nervous system (PNS) and were in the same complementation group, we reasoned that they represented mutations in tws. They were shown to be hypomorphic alleles of tws by complementation tests with twsP and by a chromosomal deficiency in the tws region (see below). tws55 was chosen as a representative of 22 adult lethal mutations, and we also used tws59 for detailed analysis of the phenotype, because tws59 exhibited a less severe PNS phenotype than the other strains. The tws55 homozygotes developed into pharate adults, and many of them eclosed but immediately fell into the food. They showed entirely uncoordinated movement in walking and overturned; they did not jump or fly. In the mutant adults, the only obvious pattern defect that we have found so far is formation of extra sensory structures.

Duplication of mechanosensory bristles

The adult body is covered with external sensory structures, most of which are bristles of the mechanoreceptors (Fig. 1A). Each bristle can be easily recognized in a high magnification view of the cuticle (Fig. 1B). In tws55, the sensory structures were duplicated; two sets of the shaft and the socket were often formed instead of a single pair of each structural component (Fig. 1C). Almost all the twin sockets were physically connected or fused to each other. The above result suggests that the number of support cells per abnormal sensory organ increased to four in contrast to two in the wild type. In the wild-type notum, large macrochaete bristles occur at invariant locations and therefore are individually named (Fig. 2A). The bristles were duplicated or triplicated in the mutants, but they always appeared at defined positions where they were supposed to be (Fig. 2B). It is unlikely that the extra structures in the mutants resulted from the formation of supernumerary sensory organ precursors (SOPs). This was shown by staining macrochaete SOPs with an enhancer trap lacZ marker B52 (Ueda et al., 1992) in the tws55 background. In wing discs of homozygous mutant larvae, no additional SOP signal was detected (data not shown).

Fig. 2.

Bristle phenotype in the notum and penetrance of defects in twins alleles. (A) A magnified view of the posterior part of the wild-type notum. Large bristles of macrochaetes labeled in this figure are anterior dorsocentral (aDC), posterior dorsocentral (pDC), anterior scutellar (aSC), and posterior scutellar (pSC) bristles. Short hairs are structural components of smaller mechanoreceptors (microchaetes). (B) In the notum of the tws55 homozygous adult, many bristles and hairs are duplicated or triplicated (arrowheads). Scale bar, 100 μm. (C) Expressivity of the bristle phenotype at each location in tws59 homozygous flies (open bars), tws59 hemizygotes (tws59/Df, stippled bars), tws55 homozygotes (hatched bars), and tws55 hemizygotes (tws55/Df, crosshatched bars). In every column, the lower portion enclosed by thicker lines represents the frequency of duplication plus triplication. The ‘triple bristle’ phenotype occurred in less than 5% of all bristles at each position. Upper parts in the individual columns indicate the proportion of the ‘single shaft with a large socket’ phenotype (see text). ‘Bristle loss’ phenotype was seen in a small fraction of macrochaetes (2% or less) only in tws55 hemizygotes, and is not included in this figure. Sample number of macrochaetes in each genotype was over two hundred.

Fig. 2.

Bristle phenotype in the notum and penetrance of defects in twins alleles. (A) A magnified view of the posterior part of the wild-type notum. Large bristles of macrochaetes labeled in this figure are anterior dorsocentral (aDC), posterior dorsocentral (pDC), anterior scutellar (aSC), and posterior scutellar (pSC) bristles. Short hairs are structural components of smaller mechanoreceptors (microchaetes). (B) In the notum of the tws55 homozygous adult, many bristles and hairs are duplicated or triplicated (arrowheads). Scale bar, 100 μm. (C) Expressivity of the bristle phenotype at each location in tws59 homozygous flies (open bars), tws59 hemizygotes (tws59/Df, stippled bars), tws55 homozygotes (hatched bars), and tws55 hemizygotes (tws55/Df, crosshatched bars). In every column, the lower portion enclosed by thicker lines represents the frequency of duplication plus triplication. The ‘triple bristle’ phenotype occurred in less than 5% of all bristles at each position. Upper parts in the individual columns indicate the proportion of the ‘single shaft with a large socket’ phenotype (see text). ‘Bristle loss’ phenotype was seen in a small fraction of macrochaetes (2% or less) only in tws55 hemizygotes, and is not included in this figure. Sample number of macrochaetes in each genotype was over two hundred.

We carried out a quantitative analysis of these phenotypes in combination of tws alleles and a chromosomal deficiency (Df) for the tws region (Fig. 2C). The mutant phenotypes in tws59 homozygotes were weaker than those in the others. The adult lethal alleles represent partial loss-of-function, as each of these alleles in a trans relation to the deficiency chromosome tended to show a more severe phenotype. Enhancement of penetrance was also observed when placed trans to twsP (data not shown). Thus, the alleles were ordered from strong to weak in the following series: twsP, tws55 and tws59. Besides the ‘double or triple bristle’ phenotype, we found cases where a single shaft protruded from a large socket with a bottle gourd shape. This structure was reminiscent of fusion of at least two sockets. As for aSC macrochaetes, the proportion of this ‘one shaft with a large socket’ form significantly increased in tws55/Df. The most minor phenotype was the ‘bristle loss’; i.e., neither socket nor shaft was found at locations where they should have formed in the wild type. This was seen only in tws55 hemizygous individuals at a very low frequency (at best 2% of aSC). Except for this rare phenotype, sensory precursor cells were born, underwent mitosis, and gave rise to an excess number of support cells in the mutants.

Neurons and sheath cells are probably transformed into support cells

To distinguish cell types of the mechanoreceptor, we employed two molecular markers. One was the monoclonal antibody 22C10, which labels the cytoplasm of neurons and shaft cells at an appropriate pupal stage (Hartenstein and Posakony, 1989). The other was a polyclonal antibody against the product of the Cut locus (α-Cut Ab), which differentially labels nuclei of the four cells constituting the mechanosensory organ; and the level of product expression in the support cells is considerably higher than in the neuron and sheath cells (Blochlinger et al., 1990, 1993). The results obtained with either marker showed that the abnormal bristles in tws mutants were not innervated by neurons.

22C10 stained the shaft cell and the neuron with its extended axon in the wild type (Fig. 3A). Beneath the twin shafts in homozygous tws55 pupae, we could find two large 22C10-binding cells and no other signal on any focal planes (Fig. 3B). Triplicated bristles were not associated with neurons, either. We surveyed 65 macrochaetes with double or triple bristles, and found that neurons were missing in 64 organs. Loss of neurons was also the case with smaller mechanosensory organs (microchaetes). In the mutant, hair duplication was found in 15% of microchaetes on the average. Eighty-four ‘double hair’ microchaetes were examined, and all of them were free of neurons. Labeling with α-Cut Ab showed that not only neurons but also sheath cells were missing in abnormal sensory organs in the mutant (Fig. 3C-F). Fig. 3C and D are photographs of a normal macrochaete taken at two different focal planes. Staining patterns, nuclear size, and the positions of nuclei relative to the epidermal layer could distinguish the four cell types. In the wild type, the socket cell was located on the shaft cell in which a cluster of granular bodies was seen. Smaller nuclei of a neuron and a sheath cell were also visualized. What we could detect in the mutant organ were four large nuclei; two were of subepidermally located socket cells and the other two nuclei exhibited the features of shaft cells (Fig. 3E,F). We investigated cellular compositions in a large number of macrochaetes by the B52 marker, which stains shaft cells more intensely than the other three cell types. Of 63 ‘double bristle’ macrochaetes examined, 60 organs contained neither neurons nor sheath cells, which was consistent with the results described above.

Fig. 3.

Cellular composition of the ‘double bristle’ mechanoreceptor. (A,B) Staining patterns of macrochaetes with mAb 22C10. (A) The normal macrochaete at 37 hours after puparium formation (APF). Underneath a shaft (out of focus), a large support cell is present and is attached to a neuron that extends an axon. (B) Duplicated shafts in a tws55 homozygote. No neuronal cell is present by the two shaft cells. An arrow points to a neuron and a shaft cell of a microchaete with normal appearance. (C,D) Subepidermal (C) and deeper (D) focal planes of a wild-type macrochaete stained with anti-Cut Ab at 40 hours APF. Closed arrowheads point to a sheath cell nucleus, and the open arrowhead in D indicates a neuronal nucleus. The large nucleus in C is of a socket cell; and that in D, of a shaft cell. The shaft cell nucleus contains condensed bodies with high expression of Cut. (E,F) An abnormal mechanoreceptor that has two shafts in the tws55 homozygote. Focal planes in E and F correspond to those in C and D, respectively. Only two socket cells (E) and two shaft cells (F) are labeled by the antibody. Scale bar, 20 μm.

Fig. 3.

Cellular composition of the ‘double bristle’ mechanoreceptor. (A,B) Staining patterns of macrochaetes with mAb 22C10. (A) The normal macrochaete at 37 hours after puparium formation (APF). Underneath a shaft (out of focus), a large support cell is present and is attached to a neuron that extends an axon. (B) Duplicated shafts in a tws55 homozygote. No neuronal cell is present by the two shaft cells. An arrow points to a neuron and a shaft cell of a microchaete with normal appearance. (C,D) Subepidermal (C) and deeper (D) focal planes of a wild-type macrochaete stained with anti-Cut Ab at 40 hours APF. Closed arrowheads point to a sheath cell nucleus, and the open arrowhead in D indicates a neuronal nucleus. The large nucleus in C is of a socket cell; and that in D, of a shaft cell. The shaft cell nucleus contains condensed bodies with high expression of Cut. (E,F) An abnormal mechanoreceptor that has two shafts in the tws55 homozygote. Focal planes in E and F correspond to those in C and D, respectively. Only two socket cells (E) and two shaft cells (F) are labeled by the antibody. Scale bar, 20 μm.

We proposed two models that could explain the twofold increase in support cells and loss of neurons and sheath cells (Fig. 4). In the wild-type fly, a single precursor cell divides twice to generate four non-equivalent daughter cells (Fig. 4A, Hartenstein and Posakony, 1989; Huang et al., 1991). The simplest explanation for the ‘double bristle’ mechanoreceptor is transformation of the neuron and the sheath cell into support cells (Fig. 4B). In tws55 homozygotes, all progeny of the SOP assume the fate of the support cells, leading to formation of two pairs of a shaft cell and a socket cell. The tws55 hemizygous condition led to a prominent increase of aSC macrochaetes with a single shaft and a large socket (Fig. 2C). B52 expression patterns indicated that a combination of three socket cells and one shaft cell may be responsible for many of these abnormal aSC bristles (data not shown). Therefore, the transformation might have a preference for the socket cell over the shaft cell.

Fig. 4.

Cell lineage in the wild type and those proposed for the mutant. (A) Cell lineage for a mechanoreceptor in the wild-type fly. The first division of the sensory organ precursor (SOP) produces two secondary precursor cells, and one of them gives rise to a neuron and a sheath cell, while the other generates two support cells that are indicated by socket and shaft in the diagram. (B,C) Models of the lineage in the tws mutant. (B) Transformation of the neuron and the sheath cells into support cells leads to two pairs of the socket-forming cell and the shaft-making cell. (C) The neuron and the sheath cell are not formed either because of cell death or mitotic arrest of one of the secondary precursor cells. In addition, four support cells are made by extra cell divisions.

Fig. 4.

Cell lineage in the wild type and those proposed for the mutant. (A) Cell lineage for a mechanoreceptor in the wild-type fly. The first division of the sensory organ precursor (SOP) produces two secondary precursor cells, and one of them gives rise to a neuron and a sheath cell, while the other generates two support cells that are indicated by socket and shaft in the diagram. (B,C) Models of the lineage in the tws mutant. (B) Transformation of the neuron and the sheath cells into support cells leads to two pairs of the socket-forming cell and the shaft-making cell. (C) The neuron and the sheath cell are not formed either because of cell death or mitotic arrest of one of the secondary precursor cells. In addition, four support cells are made by extra cell divisions.

Another possible but more complicated hypothesis is suggested in Fig. 4C. The neuron and the sheath cell are eliminated either by cell death or by mitotic arrest, and then additional divisions of the would-be support cells occur. This model seems less likely under the sensitivity of our analyses. We traced temporal division patterns of SOPs in the mutant by using the B52 marker. Microchaete SOPs were chosen to follow mitosis, because cell divisions tend to be synchronized in each longitudinal row of microchaetes (Usui and Kimura, 1993). For example, SOPs within the most medial row in a heminotum divide earlier than those within a neighboring lateral line (Fig. 5A). We found that every SOP in the most medial row gave rise to two daughters (secondary precursor cells) in tws55 as in the wild type (Fig. 5B). Thus, it was unlikely that the first mitosis of the SOP was delayed or one of the two daughter cells died prior to the second division. Mitotic patterns of the secondary precursor cells were too com-plicated to compare between the wild type and mutant, mainly because the two sister cells in each cluster do not divide syn-chronously (Hartenstein and Posakony, 1989; Usui and Kimura, 1993). This made it difficult to interpret a cluster of three cells in the mutant. Labeling of dividing cells with bro-modeoxyuridine was also employed to follow the sensory lineage, but we found that this method was less reliable for our purpose (see Materials and Methods). We tried to detect dead cells with vital dye staining, and could not find any significant signals on the mutant notum.

Fig. 5.

Cell division patterns of sensory organ precursors. Cell divisions of microchaete SOPs were followed using the lacZ marker B52 in the wild type (A) or in the tws55 mutant (B). The broken line in each panel corresponds to the dorsal midline. Photographs are oriented with the anterior up. At 16 hours APF, first divisions of SOPs are complete in two medial longitudinal rows both in the wild type and in the mutant. Arrows represent SOPs in more lateral lines, and their mitotic phase is delayed. Scale bar, 20 μm.

Fig. 5.

Cell division patterns of sensory organ precursors. Cell divisions of microchaete SOPs were followed using the lacZ marker B52 in the wild type (A) or in the tws55 mutant (B). The broken line in each panel corresponds to the dorsal midline. Photographs are oriented with the anterior up. At 16 hours APF, first divisions of SOPs are complete in two medial longitudinal rows both in the wild type and in the mutant. Arrows represent SOPs in more lateral lines, and their mitotic phase is delayed. Scale bar, 20 μm.

Other sensory organs that were not grossly affected

Surplus mechanosensory structures are found all over the body including wing margins, heads and eyes. In contrast to singly innervated mechanoreceptors, chemoreceptors are innervated by multiple neurons. The lineage relationship among cells in the chemosensory organ has been established for the taste receptor in the labellum (Ray et al., 1993). Dupli-cation of the taste bristle was very rare in the tws55 adult. Almost all chemosensory bristles on the anterior wing margin showed no altered morphology (Fig. 6A,B). Long hairs along the posterior wing margin are excluded from the sensory system, because they are not innervated in the wild type. Curiously, some of these structures were duplicated in the mutant (Fig. 6C,D).

Fig. 6.

Wing margins of the wild type and the mutant. (A) The dorsal row of an anterior wing margin of the wild type comprises recurved chemosensory bristles (arrows) and a line of stout mechanosensory bristles. Chemoreceptors are spaced apart from each other by an average of four stout bristles. (B) The anterior wing margin of a tws55 homozygote contains double or triple stout bristles, but chemosensory bristles look normal. (C) Uninnervated hairs on the posterior margin of the wild type. Some of those structures are duplicated in tws55 (D). Scale bar, 50 μm.

Fig. 6.

Wing margins of the wild type and the mutant. (A) The dorsal row of an anterior wing margin of the wild type comprises recurved chemosensory bristles (arrows) and a line of stout mechanosensory bristles. Chemoreceptors are spaced apart from each other by an average of four stout bristles. (B) The anterior wing margin of a tws55 homozygote contains double or triple stout bristles, but chemosensory bristles look normal. (C) Uninnervated hairs on the posterior margin of the wild type. Some of those structures are duplicated in tws55 (D). Scale bar, 50 μm.

The Drosophila compound eye is a large adult sense organ and is generated from a larval eye imaginal disc. A single unit of the eye, the ommatidium, is an assembly of 20 cells and eight of those are photoreceptor cells (Cagan and Ready, 1989; reviewed by Tomlinson, 1988). Although formation of photoreceptors is not dependent on lineage restriction but on cellular interactions, we studied the effects of tws mutations. This was because the Drosophila Raf-1 serine/threonine kinase functions in specification of the photoreceptor R7 fate (Dickson et al., 1992). Eye discs were isolated from third instar larvae homozygous for pupal lethal tws alleles, and stained with antibodies for all photoreceptors and/or with a marker for R7. The arrangement of photoreceptor clusters seemed to be normal, and the R7 signal was seen in each cluster. Homozygous tws55 adults exhibited a slightly rough eye phenotype, and alteration of the number of the outer pho-toreceptors (R1-6) was observed in 8% of the mutant ommatidia. Most of the abnormal ommatidia contained only five outer cells (data not shown).

Imprecise excision of the P element reduced protein level of the B subunit of PP2A

The adult lethal hypomorphs of tws were recovered in a reversion test of twsP. Characterization of their genomic structures by the polymerase chain reaction (PCR) demonstrated that in every line PlacW had been imprecisely excised, leaving a short fragment at the insertion point. In twsP, the P element is inserted into a tip of the second intron, 4 bp down-stream from an exon-intron boundary (Fig. 7A, Uemura et al., 1993). Initially we carried out Southern blot analysis, but the resolution was not high enough to reveal lesion in any of 22 hypomorphic strains except tws59. Then tws59 and nine randomly chosen lines were examined by PCR with primers that were derived from genomic sequences on either side of the P insertion site. In all of the ten genomes studied, we detected insertions that ranged from approximately 20 to 40 base pairs, and the determined sequences of four inserts were identical to those of a PlacW terminal. In tws59, a deletion of 250 bp in the intron II was found as well as a 19 bp insert that came from PlacW. We also studied the genomes of ten revertants that were perfectly viable and showed no abnormal PNS phenotype. In every case, the PCR reaction yielded a fragment of the size to be expected if the transposon had been precisely excised.

Fig. 7.

Genomic structures and northern analysis. (A) Genomic structure and major twins transcripts in the wild type. Boxes represent the open reading frame. twsP has one copy of PlacW (broken line) in its genome, which is inserted into an edge of the second intron. PlacW is 11 kb long. Primer no. 1 is positioned 5′ upstream from the insertion point by 80 bp; no. 2 and no. 3 are located 145 bp and 420 bp downstream, respectively. Genomic fragments I and II were used for northern analysis. (B) Northern blot analysis of poly(A)+ RNA prepared from whole bodies of the wild-type or homozygous mutant third-instar larvae. In each lane, 2 μg of poly(A)+ RNA was loaded. The wild-type (WT) fly produced three major messages (3.0, 4.0 and 4.6 kb). In the tws55 and tws59 lanes, three high-molecular-mass RNAs were present, and they hybridized with probe II. A predominant RNA was 9.0 kb long. Multiple bands were detected with probe I in twsP, which were previously shown to be tws-lacZ fusion transcripts (Uemura et al., 1993). mRNAs of the two other PP2A subunit genes were present in similar abundance in the wild type and in the mutants. In the wild type, transcripts for the A or C subunit were several times as abundant as those for B.

Fig. 7.

Genomic structures and northern analysis. (A) Genomic structure and major twins transcripts in the wild type. Boxes represent the open reading frame. twsP has one copy of PlacW (broken line) in its genome, which is inserted into an edge of the second intron. PlacW is 11 kb long. Primer no. 1 is positioned 5′ upstream from the insertion point by 80 bp; no. 2 and no. 3 are located 145 bp and 420 bp downstream, respectively. Genomic fragments I and II were used for northern analysis. (B) Northern blot analysis of poly(A)+ RNA prepared from whole bodies of the wild-type or homozygous mutant third-instar larvae. In each lane, 2 μg of poly(A)+ RNA was loaded. The wild-type (WT) fly produced three major messages (3.0, 4.0 and 4.6 kb). In the tws55 and tws59 lanes, three high-molecular-mass RNAs were present, and they hybridized with probe II. A predominant RNA was 9.0 kb long. Multiple bands were detected with probe I in twsP, which were previously shown to be tws-lacZ fusion transcripts (Uemura et al., 1993). mRNAs of the two other PP2A subunit genes were present in similar abundance in the wild type and in the mutants. In the wild type, transcripts for the A or C subunit were several times as abundant as those for B.

These small inserts partially blocked RNA splicing of tws premessenger. This was demonstrated by northern analysis of poly(A)+ RNA from homozygous mutant larvae (Fig. 7B). In tws55 or tws59, three high-molecular-mass bands were detected and the major one was 9.0 kb long. These large products were precursors of tws mRNAs, as they hybridized with an intron II-specific probe. Each of the unprocessed mRNAs would correspond to one of the three major mature messages (3.0, 4.0, and 4.6 kb). The large RNA bands from tws59 larvae migrated slightly faster than those from tws55 animals, which was consistent with the 250 bp deletion within the tws59 intron. Detection of incorrectly processed transcripts was suggested in aar1, which is allelic to tws (Mayer-Jaekel et al., 1993).

The tws protein is a Drosophila homologue of the regulatory subunit B of type 2A protein phosphatase (PP2A) (Uemura et al., 1993). The genes of the two other subunits of Drosophila PP2A have been cloned, those of the regulatory subunit A and the catalytic subunit C (Orga et al., 1990; Mayer-Jaekel et al., 1992). We compared the level of each subunit in the wild type with that in the mutant homozygotes by immunoblot analysis. Residual amounts of B (tws) were detected in two of the three alleles examined, and the amounts correlated with the severity of the phenotype (Fig. 8A). The mild adult lethal mutant tws59 produced 10% or less of the wild-type level of protein; and the more severe tws55 homozygote, only roughly 2%. In tissues isolated from twsP larvae, the protein was, if expressed, at a level below the limit of our detection. However, levels of the A and C subunits were not changed by any of these three mutations (Fig. 8B). All of the three subunits appeared to be abundant in the cytoplasm and expressed ubiquitously in the disc; there seemed to be no significant concentration in SOPs or in any particular progeny, compared with surrounding epidermal cells (data not shown).

Fig. 8.

Immunoblot tests for PP2A subunits. From homozygous mutant larvae of each tws allele, imaginal discs and other tissues that are rich in PP2A subunits were isolated, and their extracts were analyzed with antibodies to individual subunits. Samples of the wild type and the mutants were equalized for protein. (A) A series of diluted wild-type extracts (WT, 1/10, 1/50, 1/100) and the mutant homogenates (tws59, tws55 and twsP) were probed with rabbit antibodies against tws protein. In the tws59 and tws55 lanes, one or two bands were detected, and they migrated slightly faster than the 51×103Mr wild-type product. To visualize the low-level expression in the mutants, this blot was overexposed. (B) A blot of the wild-type and mutant samples was incubated with antibodies to the regulatory A subunit of PP2A (60×103Mr). Another blot of the same set was used for detection of the catalytic subunit (40×103Mr). Every mutant allele expressed similar levels of these two components.

Fig. 8.

Immunoblot tests for PP2A subunits. From homozygous mutant larvae of each tws allele, imaginal discs and other tissues that are rich in PP2A subunits were isolated, and their extracts were analyzed with antibodies to individual subunits. Samples of the wild type and the mutants were equalized for protein. (A) A series of diluted wild-type extracts (WT, 1/10, 1/50, 1/100) and the mutant homogenates (tws59, tws55 and twsP) were probed with rabbit antibodies against tws protein. In the tws59 and tws55 lanes, one or two bands were detected, and they migrated slightly faster than the 51×103Mr wild-type product. To visualize the low-level expression in the mutants, this blot was overexposed. (B) A blot of the wild-type and mutant samples was incubated with antibodies to the regulatory A subunit of PP2A (60×103Mr). Another blot of the same set was used for detection of the catalytic subunit (40×103Mr). Every mutant allele expressed similar levels of these two components.

Lack of the regulatory B subunit of PP2A probably alters neural cell fates

We studied peripheral neurogenesis in tws mutant adults at single-cell resolution, and have shown an essential role for type 2A protein phosphatase (PP2A) in the determination of cell fates. A major phenotype in the mutant was duplication of mechanosensory bristles. The morphological as well as molecular features of the mechanoreceptors can be explained most easily by transformation of neurons and sheath cells into support cells. Occasional triplications of the bristles would require such transformation, and extra mitosis of support cells as well. The molecular mechanism that drives this unusual division remains to be studied. Among mechanosensory organs in the wild type, those of the anterior wing margin may be exceptional; cells that constitute each wing margin mechanoreceptor are not always derived from a single SOP (Hartenstein and Posakony, 1989). Nonetheless, one of the two secondary precursor cells gives rise to support cells, and the other generates neurons and/or sheath cells. Thus, the transformation model can be at least partially applicable to twinning of the stout bristle.

Because extensive pattern duplication takes place in the wing disc in early pupal lethal alleles like twsP, we had expected that mutant homozygotes, if they could survive to adults, would produce wings and nota of aberrant shape. However, we found no reproducible malformation of wings and nota in hypomorphic adults. Therefore, the adult lethal alleles appear to keep the minimal zygotic tws+ activity necessary to establish the large-scale body patterns. In other words, isolation of the weak loss-of-function mutations enabled us to focus on indispensable functions of PP2A in neural development.

The twins PNS phenotype is separable from aar mitotic defects

Our findings of no identifiable mitotic arrest of SOPs present a striking contrast to the phenotype of abnormal anaphase resolution (aar), which is allelic to tws (Mayer-Jaekel et al., 1993; Gomes et al., 1993). In the aar1, progression through mitosis of larval neuroblasts is blocked; mitotic cells with high chromosome condensation increase, and stretched or lagging chromatids are observed in anaphase. The cell division pattern of the neuroblast is different from that of the SOP. The neuro-blast is a stem cell that produces a smaller daughter cell (ganglion mother cell); this division results in cells of unequal size. However, the SOP, its two primary daughters, and the undifferentiated sensillum cells are indistinguishable from one another with regard to size, shape and level within the cell layer (Hartenstein and Posakony, 1989). Although we did not directly examine chromosomal segregation in dividing SOPs in tws55 homozygotes, the SOP division was not prevented (Fig. 5), and the progeny were viable and underwent differentiation. Moreover, some SOPs were capable of additional cell divisions to generate triple bristles. Thus, our observations strongly suggest that the general mitotic machinery is not impaired in the tws55 SOP. The rare ‘bristle loss’ phenotype in tws55 hemizygotes might be a consequence of cell division arrest in the precursor cells, which was shown to be the case for a Cyclin-A mutant (Ueda et al., 1992). However, because of the very low penetrance, it would not be easy to specify which step of the sensillum development is responsible for this phenotype.

Possible roles for PP2A in development of the sensory nervous system

The molecular basis of asymmetric cell division has been studied in a variety of species and systems (reviewed by Horvitz and Herskowitz, 1992). Our study has added a new clue, modulation of PP2A activity, to gain access to the mechanisms that produce lineage-related cells with distinct identities. The association of the B subunit with the AC complex is known to control the catalytic activity of PP2A towards a number of substrates. Since the expression of A and C was not affected by any of the analyzed tws alleles, the protein level of B by itself appears to determine the severity of the PNS phenotype. Loss of tws may alter the dephosphorylation state of a certain molecule(s) crucial for sensory development, leading to disruption of their molecular functions. All of the PP2A subunits are ubiquitously present over an imaginal epithelium where only a limited number of cells become SOPs. It was unexpected and surprising that highly localized pattern defects were caused by mutations of a gene such as tws that is widely expressed and considered to be required for general cell metabolism. The hypothetical target molecule(s) may be expressed and active in a stage-or cell type-specific manner, and could function as an indispensable cytoplasmic factor(s) for the cell fate decision in mechanoreceptors.

Our analyses did not clarify at which stage of the sensory lineage the cell fates were changed in the tws mutants. One possibility is that each SOP underwent a symmetrical division instead of an asymmetrical one, producing only precursors of support cells. This scheme would be consistent with our observations that four support cells in the ‘double bristle’ always consist of two pairs of a shaft cell and a socket cell and that socket-free shafts were never formed. Another hypothesis is that some of the final progeny of SOPs failed to differentiate into neurons or sheath cells and entered into the alternative pathway to become support cells. These two models would be testable if markers that clearly discriminate between the two sister secondary precursor cells were available. The first explanation, if demonstrated, would suggest that tws+ is necessary to generate asymmetry of daughter cells rather than to simply promote neuronal differentiation.

Mutants of several loci have been reported to produce duplication of sensory structures, and underlying mechanisms have been studied for some of these mutants. These include numb, scabrous, Hairless, and asense (Uemura et al., 1989; Mlodzik et al., 1990; Bang and Posakony, 1992; Domínguez and Campuzano, 1993). Comparison of tws and numb mutants is particularly interesting; complete loss-of-function numb mutations are embryonic lethal, and transformation of neurons and sheath cells into support cells is observed as in homozygous tws adults (Uemura et al., 1989). Recent research on numb suggested that its protein product displayed a polarized sub-cellular localization in SOPs only during the mitotic phase and, more curiously, that the protein was unequally partitioned into daughter cells (Rhyu et al., 1994). Furthermore, results of ectopic expression of numb showed that this differential segregation confers distinct fates to daughter cells. It would be intriguing to determine whether the asymmetric distribution of the numb protein is altered by tws mutations, or whether numb is phosphorylated and a good substrate of PP2A.

We are grateful to Karen Blochlinger and Shin Togashi for anti-bodies and to Ryu Ueda, Daisuke Yamamoto, and Larry Zipursky for Drosophila strains. We thank Yasuhisa Adachi and Hiroshi Shima for technical advice. Finally we wish to thank Masao Takeda and Kazuya Usui for helpful discussions; and Yuh Nung Jan, Ken-ichi Kimura, and Michelle Rhyu for critical reading of the manuscript. This work was supported by a Fellowship of the Japan Society for the Promotion of Science for Junior Scientists (K. S.) and by research grants from the Ministry of Education, Science, and Culture of Japan.

Agostinis
,
P.
,
Derua
,
R.
,
Sarno
,
S.
,
Goris
,
J.
and
Merlevede
,
W.
(
1992
).
Specificity of the polycation-stimulated (type-2A) and ATP, Mg-dependent (type-1) protein phosphatases toward substrates phosphorylated by P34cdc2 kinase
.
Eur. J. Biochem
.
205
,
241
248
.
Bang
,
A. G.
and
Posakony
,
J. W.
(
1992
).
The Drosophila gene Hairless encodes a novel basic protein that controls alternative cell fates in adult sensory organ development
.
Genes Dev
.
6
,
1752
1769
.
Bier
,
E.
,
Vaessin
,
H.
,
Shepherd
,
S.
,
Lee
,
K.
,
McCall
,
K.
,
Barbel
,
S.
,
Ackerman
,
L.
,
Carretto
,
R.
,
Uemura
,
T.
,
Grell
,
E.
,
Jan
,
L.
and
Jan
,
Y. N.
(
1989
).
Searching for pattern and mutation in the Drosophila genome with a P-lacZ vector
.
Genes Dev
.
3
,
1273
1287
.
Blacketer
,
M. J.
,
Koehler
,
C. M.
,
Coats
,
S. G.
,
Myers
,
A. M.
and
Madaule
,
P.
(
1993
).
Regulation of dimorphism in Saccharomyces cerevisiae: involvement of the novel protein kinase homolog Elm1p and protein phosphatase 2A
.
Mol. Cell. Biol
.
13
,
5567
5581
.
Blochlinger
,
K.
,
Bodmer
,
R.
,
Jan
,
L. Y.
and
Jan
,
Y. N.
(
1990
).
Patterns of expression of cut, a protein required for external sensory organ development in wild-type and cut mutant Drosophila embryos
.
Genes Dev
.
4
,
1322
1331
.
Blochlinger
,
K.
,
Jan
,
L. Y.
and
Jan
,
Y. N.
(
1993
).
Postembryonic patterns of expression of cut, a locus regulating sensory organ identity in Drosophila
.
Development
117
,
441
450
.
Bodmer
,
R.
,
Carretto
,
R.
and
Jan
,
Y. N.
(
1989
).
Neurogenesis of the peripheral nervous system in Drosophila embryos: DNA replication patterns and cell lineages
.
Neuron
3
,
21
32
.
Bryant
,
P. J.
(
1978
).
Pattern formation in imaginal discs
. In
Genetics and Biology of Drosophila
, Vol.
2C
. (eds.
M.
Ashburner
and
T. R. F.
Wright
), pp.
229
335
.
New York/London
:
Academic Press
.
Cagan
,
R. L.
and
Ready
,
D. F.
(
1989
).
The emergence of order in the Drosophila pupal retina
.
Dev. Biol
.
136
,
346
362
.
Clarke
,
P. R.
,
Hoffmann
,
I.
,
Draetta
,
G.
and
Karsenti
,
E.
(
1993
).
Dephosphorylation of cdc25-C by a type-2A protein phosphatase: specific regulation during the cell cycle in Xenopus egg extracts
.
Mol. Biol. Cell
4
,
397
411
.
Cohen
,
P.
(
1989
).
The structure and regulation of protein phosphatases
.
Annu. Rev. Biochem
.
58
,
453
508
.
Cohen
,
P. T. W.
,
Brewis
,
N. D.
,
Hughes
,
V.
and
Mann
,
D. J.
(
1990
).
Protein serine/threonine phosphatases; an expanding family
.
FEBS Lett
.
2
,
355
359
.
Dickson
,
B.
,
Sprenger
,
F.
,
Morrison
,
D.
and
Hafen
,
E.
(
1992
).
Raf functions downstream of Ras1 in the Sevenless signal transduction pathway
.
Nature
360
,
600
603
.
Domínguez
,
M.
and
Campuzano
,
S.
(
1993
).
asense, a member of the Drosophila achaete-scute complex, is a proneural and neural differentiation gene
.
EMBO J
.
12
,
2049
2060
.
Ferrigno
,
P.
,
Langan
,
T.
and
Cohen
,
P.
(
1993
).
Protein phosphatase 2A1 is the major enzyme in vertebrate cell extracts that dephosphorylates several physiological substrates for cyclin-dependent protein kinases
.
Mol. Biol. Cell
4
,
669
677
.
Ghysen
,
A.
and
Dambly-Chaudière
,
C.
(
1989
).
Genesis of the Drosophila peripheral nervous system
.
Trends Genet
.
5
,
251
255
.
Gomes
,
R.
,
Karess
,
R. E.
,
Ohkura
,
H.
,
Glover
,
D. M.
and
Sunkel
,
C. E.
(
1993
).
Abnormal anaphase resolution (aar): a locus required for progression through mitosis in Drosophila
.
J. Cell Sci
.
104
,
583
593
.
Hartenstein
,
V.
and
Posakony
,
J. W.
(
1989
).
Development of adult sensilla on the wing and notum of Drosophila melanogaster
.
Development
107
,
389
405
.
Healy
,
A. M.
,
Zolnierowicz
,
S.
,
Stapleton
,
A. E.
,
Goebl
,
M.
,
Depaoli-Roach
,
A. A.
and
Pringle
,
J. H.
(
1991
).
CDC55, a Saccharomyces cerevisiae gene involved in cellular morphogenesis: identification, characterization, and homology to the B subunit of mammalian type 2A protein phosphatase
.
Mol. Cell. Biol
.
11
,
5764
5780
.
Horvitz
,
H. R.
and
Herskowitz
,
I.
(
1992
).
Mechanisms of asymmetric cell division: two Bs or not two Bs, that is the question
.
Cell
68
,
237
255
.
Huang
,
F.
,
Dambly-Chaudière
,
C.
and
Ghysen
,
A.
(
1991
).
The emergence of sense organs in the wing disc of Drosophila
.
Development
111
,
1087
1095
.
Imaoka
,
T.
,
Imazu
,
M.
,
Usui
,
H.
,
Kinohara
,
N.
and
Takeda
,
M.
(
1983
).
Resolution and reassociation of three distinct components from pig heart phosphoprotein phosphatase
.
J. Biol. Chem
.
258
,
1526
1535
.
Jan
,
L. Y.
and
Jan
,
Y. N.
(
1982
).
Antibodies to horseradish peroxidase as specific neuronal markers in Drosophila and in grasshopper embryos
.
Proc. Natl. Acad. Sci. USA
79
,
2700
2704
.
Jan
,
Y. N.
and
Jan
,
L. Y.
(
1990
).
Genes required for specifying cell fates in Drosophila embryonic sensory nervous system
.
Trends Neurosci
.
13
,
493
498
.
Kamibayashi
,
C.
,
Estes
,
R.
,
Slaughter
,
C.
and
Mumby
,
M. C.
(
1991
).
Subunit interactions control protein phosphatase 2A
.
J. Biol. Chem
.
266
,
13251
13260
.
Kemphues
,
K. J.
,
Raff
,
E. C.
and
Kaufman
,
T. C.
(
1983
).
Genetic analysis of B2t, the structural gene for a testis-specific β-tubulin subunit in Drosophila melanogaster
.
Genetics
105
,
345
356
.
Kinoshita
,
N.
,
Ohkura
,
H.
and
Yanagida
,
M.
(
1990
).
Distinct, essential roles of type I and 2A protein phosphatases in the control of the fission yeast cell division cycle
.
Cell
63
,
405
415
.
Kinoshita
,
N.
,
Yamano
,
H.
,
Niwa
,
H.
,
Yoshida
,
T.
and
Yanagida
,
M.
(
1993
).
Negative regulation of mitosis by the fission yeast protein phosphatase ppa2
.
Genes Dev
.
7
,
1059
1071
.
Lawrence
,
P. A.
(
1966
).
Development and determination of hairs and bristles in the milkweed bug, Oncopeltus fasciatus (Lygaeidae, Hemiptera)
.
J. Cell Sci
.
1
,
475
498
.
Lee
,
T. H.
,
Solomon
,
M. J.
,
Mumby
,
M. C.
and
Kirschner
,
M. W.
(
1991
).
INH, a negative regulator of MPF, is a form of protein phosphatase 2A
.
Cell
64
,
415
423
.
Mayer-Jaekel
,
R. E.
,
Baumgartner
,
S.
,
Bilbe
,
G.
,
Ohkura
,
H.
,
Glover
,
D. M.
and
Hemmings
,
B. A.
(
1992
).
Molecular cloning and developmental expression of the catalytic and 65-kDa regulatory subunits of protein phosphatase 2A in Drosophila
.
Mol. Biol. Cell
3
,
287
298
.
Mayer-Jaekel
,
R. E.
,
Ohkura
,
H.
,
Gomes
,
R.
,
Sunkel
,
C. E.
,
Baumgartner
,
S.
,
Hemmings
,
B. A.
and
Glover
,
D. M.
(
1993
).
The 55 kd regulatory subunit of Drosophila protein phosphatase 2A is required for anaphase
.
Cell
72
,
621
633
.
Mlodzik
,
M.
,
Baker
,
N. E.
and
Rubin
,
G. M.
(
1990
).
Isolation and expression of scabrous, a gene regulating neurogenesis in Drosophila
.
Genes Dev
.
4
,
1848
1861
.
Oda
,
H.
,
Uemura
,
T.
,
Shiomi
,
K.
,
Nagafuchi
,
A.
,
Tsukita
,
S.
and
Takeichi
,
M.
(
1993
).
Identification of a Drosophila homologue of β-catenin and its association with the armadillo protein
.
J. Cell Biol
.
121
,
1133
1140
.
Orga
,
S.
,
Brevis
,
N. D.
,
Alphey
,
L.
,
Axton
,
J. M.
,
Dudai
,
Y.
and
Cohen
,
P.
(
1990
).
The structure of protein phosphatase 2A is as highly conserved as that of protein phosphatase 1
.
FEBS Lett
.
275
,
44
48
.
Perkins
,
L. A.
,
Larsen
,
I.
and
Perrimon
,
N.
(
1992
).
corkscrew encodes a putative protein tyrosine phosphatase that functions to transduce the terminal signal from the receptor tyrosine kinase torso
.
Cell
70
,
225
236
.
Peruski
Jr.,
L. F.
,
Wadzinski
,
B. E.
and
Johnson
,
G. L.
(
1993
).
Analysis of the multiplicity, structure, and function of protein serine/threonine phosphatases
.
Adv. Prot. Phosphatases
7
,
9
30
.
Ray
,
K.
,
Hartenstein
,
V.
and
Rodrigues
,
V.
(
1993
).
Development of the taste bristles on the labellum of Drosophila melanogaster
.
Dev. Biol
.
155
,
26
37
.
Rhyu
,
M. S.
,
Jan
,
L. Y.
and
Jan
,
Y. N.
(
1994
).
Asymmetric distribution of numb protein during division of the sensory organ precursor cell confers distinct fates to daughter cells
.
Cell
76
,
477
491
.
Ronne
,
H.
,
Carlberg
,
M.
,
Hu
,
G. Z.
and
Nehlin
,
J. O.
(
1991
).
Protein phosphatase 2A in Saccharomyces cerevisiae: effect on cell growth and bud morphogenesis
.
Mol. Cell. Biol
.
11
,
4876
4884
.
Sontag
,
E.
,
Fedorov
,
S.
,
Kamibayashi
,
C.
,
Robbins
,
D.
,
Cobb
,
M.
and
Mumby
,
M.
(
1993
).
The interaction of SV40 small tumor antigen with protein phosphatase 2A stimulates the map kinase pathway and induces cell proliferation
.
Cell
75
,
887
897
.
Spreji
,
T. H.
(
1971
).
Cell death during the development of the imaginal disks of Calliphora erythrocephala
.
Neth. J. Zool
.
21
(
3
),
221
261
.
Steele
,
F. R.
,
Washburn
,
T.
,
Rieger
,
R.
and
O’Tousa
,
J. E.
(
1992
).
Drosophila retinal degeneration C (rdgC) encodes a novel serine/threonine protein phosphatase
.
Cell
69
,
669
676
.
Tomlinson
,
A.
(
1988
).
Cellular interactions in the developing Drosophila eye
.
Development
104
,
183
193
.
Ueda
,
R.
,
Togashi
,
S.
,
Takahisa
,
M.
,
Tsurumura
,
S.
,
Mikuni
,
M.
,
Kondo
,
K.
and Miyake
, (
1992
).
Sensory mother cell division is specifically affected in a cyclin-A mutant of Drosophila melanogaster
.
EMBO J
.
11
,
2935
2939
.
Uemura
,
T.
,
Shepherd
,
S.
,
Ackerman
,
L.
,
Jan
,
L. Y.
and
Jan
,
Y. N.
(
1989
).
numb, a gene required in determination of cell fate during sensory organ formation in Drosophila embryos
.
Cell
58
,
349
360
.
Uemura
,
T.
,
Shiomi
,
K.
,
Togashi
,
S.
and
Takeichi
,
M.
(
1993
).
Mutation of twins encoding a regulator of protein phosphatase 2A leads to pattern duplication in Drosophila imaginal discs
.
Genes Dev
.
7
,
429
440
.
Usui
,
H.
,
Imazu
,
M.
,
Maeta
,
K.
,
Tsukamoto
,
H.
,
Azuma
,
K.
and
Takeda
,
M.
(
1988
).
Three distinct forms of type 2A protein phosphatase in human erythrocyte cytosol
.
J. Biol. Chem
.
263
,
3752
3761
.
Usui
,
K.
and
Kimura
,
K.
(
1992
).
Sensory mother cells are selected from among mitotically quiescent cluster of cells in the wing disc of Drosophila
.
Development
116
,
601
610
.
Usui
,
K.
and
Kimura
,
K.
(
1993
).
Sequential emergence of the evenly spaced microchaetes on the notum of Drosophila
.
Roux’s Arch. Dev. Biol
.
203
,
151
158
.
Van Vactor
,
J. D. L.
,
Cagan
,
R. L.
,
Kramer
,
H.
and
Zipursky
,
S. L.
(
1991
).
Induction in the developing compound eye of Drosophila: multiple mechanisms restrict R7 induction to a single retinal precursor cell
.
Cell
67
,
1145
1155
.
Van Zyl
,
W.
,
Huang
,
W.
,
Sneddon
,
A. A.
,
Stark
,
M.
,
Camier
,
S.
,
Werner
,
M.
,
Marck
,
C.
,
Sentenac
,
A.
and
Broach
,
J. R.
(
1992
).
Inactivation of the protein phosphatase 2A regulatory subunit A results in morphological and transcriptional defects in Saccharomyces cerevisiae
.
Mol. Cell. Biol
.
12
,
4946
4959
.
Whittle
,
J. R. S.
(
1990
).
Pattern formation in imaginal discs
.
Seminars Cell Biol
.
1
,
241
252
.
Zacharuk
,
R. Y.
(
1985
).
Antennae and sensilla
. In
Comprehensive Insect Physiology, Biochemistry and Pharmacology
. (eds.
G. A.
Kerkut
and
L. I.
Gilbert
), pp.
1
69
.
Oxford
:
Pergamon Press
.
Zinn
,
K.
,
McAllister
,
L.
and
Goodman
,
C. S.
(
1988
).
Sequence analysis and neuronal expression of fasciclin I in grasshopper and Drosophila
.
Cell
53
,
577
587
.