The Neural Receptor Protein Tyrosine Phosphatase DPTP69D Is Required During Periods of Axon Outgrowth in Drosophila

Abstract

We have isolated and characterized a series of 18 chemically induced alleles of Ptp69D ranging in strength from viable to worse than null, which represent unique tools for probing the structure, function, and signaling pathway of DPTP69D. Three alleles are strongly temperature sensitive and were used to define the developmental periods requiring DPTP69D function; adult health requires DPTP69D during the mid- to late-pupal stage, eclosion requires DPTP69D during the early to mid-larval stage, and larval survival requires DPTP69D during embryogenesis. Mutations predicted to abolish the phosphatase activity of the membrane proximal D1 domain severely reduce but do not abolish DPTP69D function. Six alleles appear null; only 20% of null homozygotes pupate and <5% eclose, only to fall into the food and drown. One allele, Ptp69D⁷, confers axon and viability defects more severe than those of the null phenotype. Sequence analysis predicts that Ptp69D⁷ encodes a mutant protein that may bind but not release substrate. Like mutations in the protein tyrosine phosphatase gene Dlar, strong Ptp69D alleles cause the ISNb nerve to bypass its muscle targets. Genetic analysis reveals that the bypass defect in Dlar and Ptp69D mutants is dependent upon DPTP99A function, consistent with the hypothesis that DPTP69D and DLAR both counteract DPTP99A, allowing ISNb axons to enter their target muscle field.

AXONS navigate through a highly varied and dynamic embryonic environment to reach their synaptic targets. Along the way, differentially expressed proteins provide positional information that helps guide developing axons along appropriate pathways (Tessier-Lavigne and Goodman 1996). Growth cones at the tip of elongating axons express a large variety of receptor molecules that bind to these guidance cues. Binding of receptor to ligand in turn regulates the actin-based growth cone cytoskeleton via a variety of signal transduction mechanisms.

Tyrosine phosphorylation, one such signaling pathway, is emerging as a key regulator of actin polymerization during axon guidance. For example, Drosophila Ena, which promotes actin polymerization (Krause  et al. 2002), is a substrate of the receptor tyrosine phosphatase DLAR in vitro and the cytoplasmic tyrosine kinase Abl in vitro and in vivo (Comer  et al. 1998; Wills  et al. 1999). Embryonically, Ena, DLAR, and Abl are all expressed in axons, suggesting that signal transduction involving tyrosine phosphorylation can affect axon guidance by modulating actin polymerization through Ena (Bennett and Hoffmann 1992; Gertler  et al. 1995; Krueger  et al. 1996; Comer  et al. 1998).

Tyrosine phosphorylation also regulates axon guidance through the small GTPases Rac, Rho, and Cdc42. These G proteins regulate actin filament polymerization, branching, and bundling in part by regulating the actin depolymerizing protein Cofilin and the actin filament capping complex Arp2/3 (Meyer and Feldman 2002). Several lines of evidence indicate that Rac, Rho, and Cdc42 function in tyrosine-kinase-stimulated axon outgrowth. In Drosophila, Abl and DLAR interact genetically with Trio, a guanine nucleotide exchange factor for Rac, during embryonic motor axon guidance (Bateman  et al. 2000; Liebl  et al. 2000). In PC12 cells, nerve growth factor stimulation of the Trk tyrosine kinase induces neurite formation in a Rac- and Cdc42-dependent manner (Daniels  et al. 1998). In cerebellar granule cells, stimulation of neurite outgrowth by L1, a neural cell adhesion molecule (CAM), requires the Src cytoplasmic tyrosine kinase, activation of the mitogen-activated protein kinase pathway, and Rac (Schmid  et al. 2000). Finally, Ena, Abl, and DLAR, among several other tyrosine kinases and phosphatases, have been shown to be directly involved in axon guidance in Drosophila, and many of their vertebrate homologs also function in this process (Callahan  et al. 1995; Desai  et al. 1997b; Bashaw  et al. 2000).

Although the mechanisms by which tyrosine kinases affect axon guidance are relatively well studied, much less is known about their enzymatic antagonists, the receptor-like protein tyrosine phosphatases (RPTPs). RPTPs comprise a diverse protein family and are expressed in the developing nervous systems of many animals (Van  Vactor  et al. 1993; Yan  et al. 1993; Desai  et al. 1994; Gershon  et al. 1998; Ledig  et al. 1999). Extracellularly, RPTPs resemble CAMs, often containing multiple immunoglobulin (Ig) domains and/or fibronectin (Fn) type III repeats. Intracellularly, RPTPs consist of one or two phosphatase domains. The ability of RPTPs to reverse the catalytic action of tyrosine kinases implicates both protein families in many of the same processes, including cell proliferation, adhesion, and migration. In Drosophila five RPTPs, DLAR, DPTP10, DPTP52F, DPTP69D, and DPTP99A, facilitate motor and central nervous system (CNS) axon guidance during embryogenesis (Krueger  et al. 1996; Desai  et al. 1997a; Sun  et al. 2000; Schindelholz  et al. 2001). DLAR and DPTP69D are also required for retinal axon guidance during late larval and pupal stages (Garrity  et al. 1999; Clandinin  et al. 2001; Maurel-Zaffran  et al. 2001). To date, the molecular mechanisms that mediate these effects remain unclear, although the interactions among Dlar, Trio, Abl, and Ena indicate that regulation of actin polymerization is a key outcome.

To further elucidate RPTP signaling pathways, we have undertaken a detailed genetic structure-function analysis of DPTP69D. DPTP69D contains two Ig and 3 Fn domains extracellularly and two phosphatase domains intracellularly (Figure 1B). DPTP69D is cleaved at an extracellular RDKR site ∼90 residues from the trans-membrane segment, but the cleavage products remain associated (P. Snow, personal communication). DPT-P69D is particularly interesting because unlike Ptp10D, Ptp99A, and Dlar, Ptp69D is an essential gene. Mutations in Ptp69D confer mild motor axon defects in the embryo and highly penetrant retinal axon defects in the larva (Desai  et al. 1996; Garrity  et al. 1999; Newsome  et al. 2000), while Ptp10D and Ptp99A mutants appear wild type. Interestingly, Ptp69D Ptp99A and Ptp10D;Ptp69D double mutants display severe and highly penetrant defects in embryonic motor and central axons, respectively (Desai  et al. 1996; Sun  et al. 2000). In this article, we describe the isolation and characterization of 18 new alleles of Ptp69D. Utilizing these alleles, we demonstrate that the membrane-proximal phosphatase domain is critical for function and that DPTP69D is required during developmental stages coinciding with periods of axon outgrowth. This panel of alleles will also serve as genetic tools useful for delineating the DPTP69D signal transduction pathway.

MATERIALS AND METHODS

Mutagenesis: About 100 3- to 5-day-old Oregon R or w; 11F3 males were places into bottles containing paper saturated with a solution of 3% sugar, 4% green food color, and 25 mm EMS (Sigma, St. Louis). After 6 hr, the flies were examined and those with green abdomens were returned to the bottle. After 24 hr, the greenest males were placed on fresh food and allowed to recover. Twenty-four hours later, nongreen males, usually between 10 and 30 in number, were mated en mass with 50–100 virgin w; TM3/TM6B females. Males and females were removed after 4 and 8 days, respectively. F₁ males were individually mated to w; Df(3L)^8ex34/TM6B or w; Ptp69D¹/TM6B virgins. Vials producing no unbalanced F₂ progeny were tested further.

Genetics: Strains that failed to complement all four deletion alleles were considered to harbor new Ptp69D alleles, propagated as balanced stocks, and backcrossed by recombining the markers ru h th st cu sr e ca onto the mutant chromosomes. Single w; ru h Ptp69D* th st cu sr e ca/TM6B males were then mated with w¹¹¹⁸ females and the markers were removed by recombination. This procedure resulted in two recombination events within 12% genetic distance of Ptp69D on the left and two events within 4% to the right, removing most or all nonselected mutations induced during mutagenesis. The Ptp69D transgene used to test for extraneous lethal mutations is described in Desai  et al. (1996).

Immunohistochemistry: Twenty-four-hour embryo collections were fixed and prepared for immunohistochemistry as

described (Patel 1994). Antibody binding was visualized using horseradish-peroxidase-conjugated goat anti-mouse or anti-rabbit antibodies and diaminobenzidine immunohistochemistry. Ptp69D mutant embryos were identified by their failure to stain with anti-DPTP69D mAbs 3F11, 3F12, or 5A6 (Desai  et al. 1996), or anti-β-galactosidase sera (Jackson Laboratories, Bar Harbor, ME; expressed by TM6B T8-lacZ balancer chromosome, gift of Peter Kolodziej); sorted; restained with mAb 1D4 to visualize motor axons; cleared in 70% glycerol::30% PBS; and dissected. Embryos were examined on an Olympus AX-70 microscope using differential interference contrast optics and photographed using a Spot-cooled CCD camera.

Temperature-shift experiments: Four vials containing 15–20 w; Ptp69D¹/TM6B virgin females were crossed to 10–15 w;  Ptp69D¹⁰/TM6B males at 25°. These flies were transferred to new food every 12 hr, and dead flies were replaced to maintain progeny counts. When F₁ flies began to eclose from the first group of four vials, the entire set was transferred to 18°. Flies were counted and removed daily from each group, until no new flies eclosed in a 24-hr period. At that time, hatched and unhatched pupae were counted. Similar experiments were carried out with crosses of w; Df(3L)^8ex34/TM6B females with w; Ptp69D¹²/TM6B or w; Ptp69D¹⁸/TM6B, except that these crosses were first maintained at 18° and the parents were transferred to new food every 24 hr.

Molecular biology:  Genomic DNA from single embryos: Stage 16 embryos were bleach dechorionated, rinsed, picked individually into siliconized microtiter plate wells (Sigmacote), devitellinized in 5 μl of 1:1 heptane/methanol, and washed twice with 10 μl 50 mm KCl, 10 mm Tris pH 8.3, and 0.05 mm EDTA. After removing the fluid, embryos were pulverized with siliconized toothpicks, resuspended in 4 μl 50 mm KCl, 10 mm Tris pH 8.3, 0.5% Tween, 0.5% NP-40, and transferred to a tube containing 4 μl “last buffer” (50 mm KCl, 10 mm Tris pH 8.3). Each well was rinsed with 4 μl last buffer and this was added to the appropriate tube, which was stored on dry ice. Embryo lysates were digested with 8 μl last buffer

containing 0.25 mg/ml Proteinase K (Sigma) for 1 hr at 50°, incubated at 94° for 10 min, and stored at –20°.

Embryo genotyping: DNA from individual embryos was used in tandem PCR reactions designed to detect the presence of the TM6B-T8 lacZ balancer. One microliter of lysate was added to 19 μl standard PCR buffer (Herculase; Stratagene, La Jolla, CA) containing 250 nm each lacZ primer (GCA GGC TTC TGC TTC AAT CAG CGT GCC and GTT TGC CGT CTG AAT TTG ACC TGA GCG). PCR reactions that lacked the ∼500-bp lacZ PCR product were diluted 1:100 and reamplified. Lysates that lacked the lacZ PCR product, but generated a Ptp69D PCR product, were considered to be mutant.

Sequencing: A total of 1–2 μl lysate from mutant embryos was used in a series of PCR reactions designed to amplify overlapping ∼1-kb fragments of the Ptp69D gene. These primary PCR reaction products were purified using the Ultra-Clean DNA purification kit (CLP), diluted 1:100, and reamplified using nested Ptp69D primers. Purified secondary PCR product (0.5 μl) was used in a BigDye (ABI, Columbia, MD) sequencing reaction containing 4 μl SeqSaver (Sigma) in a volume of 20 μl. Sequencing samples were purified by ethanol precipitation and submitted to the sequencing core at Vanderbilt University Medical Center. Sequences were aligned to a GenBank sequence of Ptp69D using Vector (Burlingame, CA) NTI software. Mutations were confirmed by forward and reverse primers in the same embryo and by sequencing the same region of the gene in a second mutant embryo.

Ranking of alleles: Alleles are ranked in Tables 1,2,3 according to penetrance and expressivity of their viability defects with the strongest alleles at the top. The rankings were determined by comparing the sum of the expected percentages of healthy free-walking adults, total adults, and total pupae produced by each genotype. The ranking in Table 5 was determined by comparing the sum of the rates of total and severe ISNb defects, ISN defects, SNa defects, and SNc defects.

RESULTS

Existing  Ptp69D  alleles: Previous genetic studies of Ptp69D utilized four deletion alleles derived from the imprecise excision of two P elements (Desai  et al. 1996). Two derive from 59A, an insertion 3′ to Ptp69D in the nearby kinesis light chain (KLC) gene (Figure 1A; Gindhart  et al. 1998). These are Df(3L)^8ex25, which lacks sequences that encode most of the cytoplasmic domains, and Df(3L)^8ex34, which lacks the entire Ptp69D gene and many other essential complementation groups. Both also lack two genes located between Ptp69D and KLC, CG32112 and CG10984, at least one of which is essential (our unpublished observations). The other two deficiencies derive from 11F3, an insertion into the 5′ untranslated region of Ptp69D:Ptp69D¹, which lacks all but the first 115 bp of the 5′ untranslated region (UTR), the translational start site, and sequences that encode most of the extracellular domain; and Ptp69D⁴, which lacks the first 115 bp of the 5′ UTR, the entire promoter, and the nearby small nuclear ribonucleoprotein (snRNP-69D) gene. Ptp69D¹ and Ptp69D⁴ are also mutant for mirror (mirr), a nearby transcription factor gene. Embryos homozygous for each of the deletion alleles fail to express DPTP69D, as assayed using monoclonal antibodies to the extracellular domain. These four alleles were used to screen for and analyze our panel of 18 EMS-generated Ptp69D alleles.

Isolation of new  Ptp69D  alleles: As described in materials and methods, male progeny of mutagenized males were crossed individually to Ptp69D¹, Ptp69D⁴, or Df(3L)^8ex34 virgin females. Candidates that failed to complement these three deletions could bear mutations in Ptp69D, KLC, snRNP69D, mirr, or other essential genes uncovered by Df(3L)^8ex34. To distinguish among these possibilities, all candidates were crossed to the other three deletion alleles. Using these tests, six alleles of KLC, two alleles of mirr, and 11 other lethal mutations defining six complementation groups uncovered by Df(3L)^8ex34 (data not shown) were isolated in addition to 18 potential Ptp69D alleles. All 18 of these candidates map to the genetic interval between hairy and thread, phenotypic markers that flank the Ptp69D locus. A Ptp-69D transgene was able to rescue animals homozygous for 17 of the candidates, confirming their identity as Ptp69D alleles. The remaining allele, Ptp69D¹⁶, bears a closely linked unselected lethal mutation (see Temperature-sensitive Ptp69D alleles).

Strength of  Ptp69D  deletion alleles: To examine phenotypic differences among the original deletion alleles, flies heterozygous for each of the 18 EMS alleles were crossed with flies heterozygous for each of the four deletion alleles at 25° and 18° and the fate of the cross progeny was examined. As can be seen in Tables 1 and 2, the progeny of crosses with Ptp69D⁴ and Df(3L)^8ex25 generally have a greater probability of developing to the pupal and adult stages than do offspring of crosses with Df(3L)^8ex34 or Ptp69D¹. Summing the data from all 25° crosses reveals that Ptp69D⁴ and Df(3L)^8ex25 trans-heterozygous animals are twice as likely to develop into healthy adults and 1.5 times as likely to eclose as are Df(3L)^8ex34 and Ptp69D¹ trans-heterozygotes (data not shown). Df(3L)^8ex34 and Ptp69D¹ display little or no difference in viability in trans to any of the new alleles. We conclude that the four deletion alleles vary in strength, with Df(3L)^8ex34 = Ptp69D¹ > Df(3L)^8ex25 > Ptp69D⁴ in order of decreasing strength.

Strength of  Ptp69D  EMS alleles: To determine the strength of the new alleles and to identify temperature-sensitive alleles, we examined the fate of animals homozygous for each allele at 18° and 25° (Table 3). One striking result of these experiments is that Ptp69D⁷, Ptp-69D¹¹, and Ptp69D¹⁶ homozygotes rarely, if ever, form pupae. The observation that animals hemizygous for Ptp69D⁷ or Ptp69D¹¹ can form pupae (Tables 1 and 2),

albeit infrequently, indicates that their viability phenotype is significantly worse than the null phenotype and suggests that these may be neomorphic alleles. Alternatively, it is possible that Ptp69D⁷ and Ptp69D¹¹ harbor closely linked unselected mutations that decrease the viability of homozygotes. Both alleles have been backcrossed and can be rescued by a Ptp69D genomic transgene, however, supporting the conclusion that Ptp69D⁷ and Ptp69D¹¹ are unusually strong alleles. Ptp69D¹⁶ is discussed below (see Temperature-sensitive Ptp69D alleles). At the other end of the spectrum, Ptp69D¹⁰ and Ptp69D²⁰ are so weak that homozygotes can survive and propagate at either 18° or 25°, and Ptp69D⁵ is a weak, semiviable allele. Thus these 18 new mutations define an allelic series ranging in strength from subliminal to worse than null.

Ptp69D  null phenotype: To identify when animals mutant for Ptp69D die, we balanced the new alleles and deletion alleles with TM6B, which confers a dominant Tubby pupal phenotype, allowing the easy identification of trans-heterozygous (non-Tubby) pupae and subsequent evaluation of their fate. On one hand, in crosses of Ptp69D¹⁰ or Ptp69D²⁰ with any of the deficiencies >50% of trans-heterozygous progeny survive to adulthood at 25° (Table 1). By contrast, <20% of animals hemizygous or homozygous for the strongest of the new alleles form pupae, and ≤5% eclose. These few adults eclose 1–3 days after their heterozygous siblings, are feeble, and become trapped in the food like ancient mammals in the La Brea tar pits (tar pit phenotype). These observations suggest that alleles conferring these phenotypes, including Ptp69D¹, Ptp69D⁷, Ptp69D⁸, Ptp69D¹¹, Ptp69D¹⁶, and Ptp69D²¹, as well as Ptp69D^Y98 and Ptp69D^Y100, described in Newsome  et al. (2000), represent complete loss of Ptp69D function at 25°. Note that Ptp69D¹ and Ptp69D^Y100 are also presumed to be null alleles by molecular criteria (see Figure 1). In all cases, crosses of new EMS alleles with either strong deficiency produce fewer-than-expected trans-heterozygous pupae. Moreover, in all cases, a lower proportion of trans-heterozygous pupae hatch than do balanced siblings. Taken together, these observations indicate that most animals lacking DPTP69D die as larvae, a minority survive to the pupal stage, and rare individuals develop into feeble adults.

18°  null alleles: On the basis of the 18° phenotypes of the molecular null alleles Ptp69D¹ and Ptp69D^Y100, the Ptp69D null phenotype is temperature independent (Table 2). Six of the eight alleles identified as putative null alleles at 25°, Ptp69D¹, Ptp69D⁷, Ptp69D⁸, Ptp69D¹¹, Ptp69D¹⁶, and Ptp69D^Y100, also appear to be null at 18° (Tables 2 and 3).

Temperature-sensitive  Ptp69D  alleles: The viability phenotype of Ptp69D¹⁰, Ptp69D¹², and Ptp69D¹⁸ is especially temperature sensitive. Although Ptp69D¹⁰ adults survive at both temperatures, Ptp69D¹² and Ptp69D¹⁸ adults are viable at 18°, but do not eclose or suffer the tar pit phenotype at 25°. Ptp69D¹⁷, Ptp69D¹⁹, Ptp69D²¹, and Ptp69D²² are also temperature sensitive; twice as many animals homozygous for these alleles survive to the pupal stage at 18° than at 25° (Table 3). These results indicate that Ptp69D¹² and Ptp69D¹⁸ are ideal alleles for temperature-shift experiments as adults bearing these alleles are quite healthy at 18° but die at 25° (see below).

One allele, Ptp69D¹⁶, gives anomalous complementation results. In trans to Ptp69D¹ and Ptp69D⁴, Ptp69D¹⁶ behaves like a temperature-sensitive allele, with 25° pupation rates of 13 and 27%, respectively, and 18° pupation rates of 28 and 52%, respectively. In contrast, Ptp69D¹⁶ behaves like an extremely strong allele in trans to Df(3L)^8ex34 and Df(3L)^8ex25, having a pupation rate of <1% at all temperatures. Similarly, Ptp69D¹⁶ homozygotes die before the pupal stage. One interpretation of these results is that Ptp69D¹⁶, Df(3L)^8ex34, and Df(3L)^8ex25 share a common second-site lethal mutation, presumably in KLC, CG32112, or CG10984. This interpretation is supported by the observation that a Ptp69D transgene rescues Ptp69D¹⁶/Ptp69D¹ but not Ptp69D¹⁶/Df(3L)^8ex34 (data not shown).

Adult health requires DPTP69D until the pupal stage: To determine the developmental requirement for DPTP69D function, the fate of animals bearing temperature-sensitive alleles raised at permissive or nonpermissive temperatures was examined. Figure 2 shows the effect of temperature on the health of adults. When raised entirely at 25°, <20% of Ptp69D¹² and Ptp69D¹⁸ hemizygotes eclose, and those that do are feeble, fall into the food, and die. By contrast, >75% of Ptp69D¹² and Ptp69D¹⁸ hemizygotes survive to become healthy, free-walking adults when raised at 18°. Ptp69D¹⁰/Ptp69D¹ animals are similarly temperature sensitive with 20% developing into healthy adults at 25° as compared with 100% at 18°. Temperature-shift experiments involving any of these genotypes demonstrate a pupal requirement for DPTP69D function. Specifically, to survive after eclosion, Ptp69D¹² and Ptp69D¹⁸ hemizygotes had to be mid-stage pupae prior to the 25° shift, and Ptp69D¹⁰/ Ptp69D¹ animals had to be shifted to 18° as early pupae. Even late-stage pupae that are hemizygous for Ptp69D¹² or Ptp69D¹⁸ suffer from shifting, indicating an ongoing requirement for DPTP69D function throughout the pupal stage.

Eclosion requires DPTP69D during embryogenesis: Nonpermissive-to-permissive shifts involving Ptp69D¹⁰/ Ptp69D¹ animals also reveal an embryonic requirement for DPTP69D function, as animals reared for as little as 12 hr at 25° are less likely to survive as adults than animals reared entirely at 18°. The pupal and embryonic requirements for DPTP69D function are also reflected in eclosion rates (Figure 3). The rate of eclosion drops from 100% at 18° to 75% when Ptp69D¹⁰/Ptp69D¹ animals are reared at 25° for their first 24 hr. Eclosion drops further to 40% if Ptp69D¹⁰/Ptp69D¹ animals are raised until the early pupal stage at the high temperature. Similarly, the eclosion rate of Ptp69D¹² and Ptp69D¹⁸ hemizygotes drops from >80% when raised at 18° to <20% when shifted to the nonpermissive temperature prior to the onset of pupation. Unlike adult health, the eclosion rates of Ptp69D¹² and Ptp69D¹⁸ hemizygotes are unaffected by late pupal shifts to the nonpermissive temperature. Thus we conclude that adult health requires DPTP69D function throughout the pupal stage whereas eclosion requires DPTP69D function during embryonic and early to mid-pupal stages.

Larval survival requires DPTP69D during late embryogenesis: In addition to its requirement for eclosion, the underrepresentation of hemizygous pupae among the progeny of flies bearing Ptp69D¹² or Ptp69D¹⁸ crossed to Df(3L)^8ex34 indicates a function for DPTP69D prior to the pupal stage. To determine the developmental requirement for DPTP69D function for larval survival, Ptp69D¹⁰/Ptp69D¹, Ptp69D¹²/Df(3L)^8ex34, and Ptp69D¹⁸/Df (3L)^8ex34 pupal counts were correlated with temperature as shown in Figure 4. As can be seen, 100% of Ptp69D¹⁰/Ptp69D¹ animals develop into pupae when maintained entirely at the permissive temperature. A similar result was obtained for animals shifted to the permissive temperature after 12 hr. Strikingly, development at the nonpermissive temperature for as little as 24 hr caused a 25% decrease in the expected number of pupae. Complementary temperature-shift experiments confirm this embryonic period of sensitivity, as survival to the pupal stage increases from 40 to 80% when Ptp69D¹² and Ptp-69D¹⁸ hemizygotes are shifted to the nonpermissive temperature after embryogenesis. These results reveal that larval survival requires DPTP69D function during late embryonic development.

Protein expression in  Ptp69D  mutants: To examine the effects of the EMS alleles on DPTP69D expression, mutant embryos were examined with anti-DPTP69D monoclonal antibodies. These experiments revealed that the alleles fell into three categories of DPTP69D expression. While Ptp69D⁷, Ptp69D¹⁰, Ptp69D²⁰, and Ptp-69D²¹ embryos express wild-type levels of DPTP69D, DPT-P69D levels are reduced in Ptp69D⁵, Ptp69D¹³, Ptp69D¹⁷, and Ptp69D¹⁹ embryos and undetectable in embryos homozygous for the remaining nine alleles (Table 4). In no case did we detect changes in DPTP69D localization.

Ptp69D  alleles cause motor axon defects: The role of DPTP69D in axon guidance has been inferred from the examination of the neuromuscular system of Ptp69D¹/  Df(3L)^8ex25 trans-heterozygous embryos (Desai  et al. 1996). Although Ptp69D¹ appears to represent a null allele, our results suggest that Df(3L)^8ex25 may not completely abrogate DPTP69D function, as Ptp69D¹/Df(3L)^8ex25 animals develop to the pupal and adult stages more frequently than do Ptp69D¹/Df(3L)^8ex34 animals (Tables 1 and 2). Indeed, several of the new alleles confer stronger motor axon phenotypes. For example, embryos homozygous for Ptp69D⁷, Ptp69D¹⁰, Ptp69D²¹, and Ptp69D²² display motor axon defects that are both more severe and of higher penetrance than those of Ptp69D¹/Df(3L)^8ex25 embryos

(Table 5). In all cases ISNb nerve outgrowth is most sensitive to the loss of DPTP69D. Embryos homozygous for strong alleles (excluding Ptp69D⁷; see DPTP69D and DLAR allow ISNb muscle innervation) display defects in 10–33% of ISNb nerves, including premature termination, ectopic or reduced defasciculation, and defects in target recognition (Figure 5). Other motor nerves display less penetrating defects in mutant embryos: 5–22% of SNas are defective, most lacking a dorsal or posterior branch; 3–14% of ISNs are defective, often defasciculating abnormally; and 0–12% of SNcs are defective. See Figures 5,6,7,8 for representative defects. These observations indicate that DPTP69D functions to facilitate motor axon guidance.

Ptp69D  motor axon defects are partially penetrant: The relatively normal appearance of most nerves in embryos presumed to lack zygotic DPTP69D indicates that ISNb axons are not completely dependent upon DPTP69D function. One caveat to this conclusion is that DPTP69D, presumably of maternal origin, can be detected in early embryos (Fitzpatrick  et al. 1995). However, embryos lacking both maternal and zygotic sources of DPTP69D appear generally similar to those lacking only embryonic DPTP69D (C. Desai, unpublished observations). These results support the idea that molecules such as the other neural RPTPs partially compensate for the loss of DPTP69D.

DPTP69D and DLAR allow ISNb muscle innervation: Bypass defects arise when some or all ISNb axons fail to enter the target ventro-lateral muscle (VLM) field, but instead grow dorsally within or next to the ISN nerve (Figure 5). As can be seen in Table 6, these defects occur at a rate of ∼15% in Ptp69D null embryos and 30% in Dlar mutants (Krueger  et al. 1996; Desai  et al. 1997a). Strikingly, one of our new alleles, Ptp69D⁷, confers the bypass phenotype at rates approaching those seen in Dlar mutants (Table 6). Previous results demonstrated that the Dlar bypass phenotype requires a third RPTP, DPTP99A (Desai  et al. 1997a). The ISNb bypass phenotype observed in Ptp69D embryos is likewise dependent upon DPTP99A (Table 6). Note that we removed only one copy of Ptp99A in these experiments because Ptp69D Ptp99A double-mutant embryos display severe ISNb defects that can preempt the bypass phenotype (Desai  et al.  1996, 1997a). These results suggest that DPTP69D and DLAR function together to allow ISNb axons to innervate the VLM by counteracting DPTP99A at this choice point.

Molecular defects in  Ptp69D  alleles: To determine the molecular basis for reduced or absent function of the Ptp69D gene, selected alleles were sequenced. For the most part, alleles directing the expression of immunologically detectable DPTP69D were chosen for sequence analysis. The results are illustrated in Figure 1B and listed in Table 7. A shared molecular defect may be responsible for the similar phenotypes exhibited by two temperature-sensitive alleles, Ptp69D¹² and Ptp69D¹⁸. Both alleles bear a missense mutation, glycine to aspartate, in the extracellular region of the protein, C-terminal to the cleavage site 47 residues from the start of the trans-membrane segment. Two of our EMS alleles, Ptp69D²⁰ and Ptp69D²¹, as well as Ptp69D^Y125 (Newsome  et al. 2000), have missense mutations in the active site of the D1 phosphatase domain. The mutation in Ptp69D²¹ changes the invariant catalytic cysteine residue to tyrosine (Figure 1B) and almost certainly abrogates the activity of this domain (Table 8). In Ptp69D²⁰ and Ptp-69D^Y125, conserved active-site glycines are mutated to alanine and arginine, respectively (Figure 1B). It will be interesting to determine the effects of these changes on phosphatase activity and then to correlate activity with the ability to support axon guidance. The two alleles that cause the most severe axon guidance defects have strikingly different mutations. Ptp69D⁷ has a small three-amino-acid deletion in the D1 phosphatase domain that removes a conserved aspartate residue. In other phosphatases, this residue is required to complete catalysis; mutants in which this residue is changed to alanine bind but do not release their substrates. By

contrast, the change in Ptp69D¹⁰ converts a valine between the immunoglobulin domains to a glutamate. Finally, embryos homozygous for Ptp69D⁹ and Ptp69D¹⁷ express low or undetectable levels of DPTP69D. Both alleles have changes in the trans-membrane region, suggesting that the C-terminal cleavage product may stabilize the DPTP69D extracellular portion and/or anchor it to the cell surface.

DISCUSSION

The 18 new chemically induced mutations in Ptp69D form an allelic series. A matrix of crosses between these 18 alleles and 4 existing alleles derived from the imprecise excision of nearby P elements reveals differences not only among the new alleles, but also among the deletion alleles. Among the deletion alleles, Ptp69D⁴ appears weakest but Df(3L)^8ex25 also appears to retain residual DPTP69D activity. Specifically, animals bearing the new alleles in trans to Ptp69D⁴ or Df(3L)^8ex25 develop to later stages more frequently than do Ptp69D¹ and Df(3L)^8ex34 trans-heterozygotes. These results lead to two conclusions. First, levels of DPTP69D undetectable by immunohistohemistry are functionally relevant. Second, the Df(3L)^8ex25 allele apparently directs the expression of an N-terminal fragment that lacks phosphatase domains but retains residual biological activity. These observations also indicate that previous descriptions of the Ptp69D mutant phenotype, based upon Ptp69D¹/  Df(3L)^8ex25 embryos, did not describe the effects of complete loss of zygotic DPTP69D expression.

Ptp69D⁷  may be a neomorphic allele: The severe axon and viability defects caused by Ptp69D⁷ could be due to closely linked unselected modifier mutations or to neomorphic activity of the mutant DPTP69D encoded by this allele. Because Ptp69D transgenes rescue the viability of Ptp69D⁷ and because Ptp69D⁷ directs the expression of wild-type levels of DPTP69D, we favor the second hypothesis. The molecular defect in the Ptp69D⁷ allele supports this hypothesis: lacking a conserved aspartate residue required for catalysis, Ptp69D⁷ might bind and sequester substrate proteins. Previous studies show that the cytoplasmic domains of DLAR and DPT-P69D can substitute for one another in retinal and motor axon guidance, implying that they can utilize the same substrates (Desai  et al. 1997a; Maurel-Zaffran  et al. 2001). Thus the defects observed in Ptp69D⁷ embryos could be due to loss of DPTP69D coupled with impaired signaling through DLAR and/or other neural RPTPs.

The phenotypic and genetic similarities of the ISNb bypass defect in Dlar and Ptp69D⁷ mutants support this conclusion.

Temperature-sensitive alleles:  Ptp69D¹² and Ptp69D¹⁸ confer very similar phenotypes at all temperatures, consistent with their shared molecular defect. At 25°, <5% of Ptp69D¹² or Ptp69D¹⁸ homozygotes survive to the adult stage, all of which fall into the food and die. By contrast, at 18° most Ptp69D¹² and Ptp69D¹⁸ homozygotes develop into healthy flies. The high frequency of adult survival at this permissive temperature indicates that DPTP69D¹² and DPTP69D¹⁸ attain close to wild-type activity at 18°.

The third strongly temperature-sensitive allele, Ptp69D¹⁰,

confers a very weak viability defect. In fact, Ptp69D¹⁰ homozygotes survive even at 25°. The temperature-sensitive nature of Ptp69D¹⁰ is unmasked in hemizygotes; although two-thirds survive to the adult stage, only 11% remain free of the food at 25°, whereas adult survival is 100% at 18°. The observation that Ptp69D¹⁰ homozygotes survive while hemizygotes do not indicates that boosting the residual DPTP69D activity in Ptp69D¹⁰ hemizygotes by a factor of 2 or less is sufficient to support viability at 25°. Moreover, it is likely that reducing DPTP69D activity by a small amount could decrease the hemizygous eclosion and/or pupation rates. Therefore, mutations in genes whose products function in or regulate DPTP69D signal transduction are likely to dominantly modify the lethal phenotype of Ptp69D¹⁰ hemizygotes.

D1 inactive mutants retain residual activity:  Ptp69D²¹ is weakly temperature sensitive as the pupation rate jumps from 6% at 25° to 32% at 18° for hemizygotes and from 11 to 35% for homozygotes. (Interestingly, the pupation rate of animals hemizyogous for another active site mutant, Ptp69D^Y125, is also weakly temperature sensitive, rising from 11% at 25° to 25% at 18°.) It is surprising that Ptp69D²¹ is temperature sensitive because the cysteine-to-tyrosine change in the D1 domain should eliminate catalytic activity at all temperatures. Several explanations could account for this temperature sensitivity. For example, the function of DPTP69D could be more critical at higher temperatures, although several other strong alleles, including Ptp69D⁷, Ptp69D⁸, Ptp69D⁹, and Ptp69D¹⁴, confer temperature-independent phenotypes. Alternatively, the introduction of a bulky tyrosine residue into the active site could disrupt the structure of the D1 domain at higher temperatures. Such disruption could destabilize the intracellular domain or inhibit noncatalytic aspects of DPTP69D function, such as docking. It is therefore of interest to determine if the loss of D1 activity is sufficient to mimic the viability defects

conferred by Ptp69D²¹ at 25°. Two other mutants, DPTP69D²⁰ and DPTP69D^Y125, have active site substitutions predicted to decrease or eliminate phosphatase activity and should enable us to address this question. In any case, it appears that elimination of D1 phosphatase activity does not completely abolish DPTP69D function, at least at 18°.

Developmental requirement for DPTP69D function: DPTP69D is required during embryogenesis and throughout the pupal stage for full viability. The coincidence of the pupal and embryonic requirements for DPTP69D with axonogenesis suggests that axon defects underlie the morbidity of Ptp69D mutants. This conclusion is supported by the suppression of both viability and axon defects of Ptp69D mutants by selective expression of Ptp69D cDNA in the nervous system (Garrity  et al. 1999). Motor axon defects found in Ptp69D mutants could account for their decreased fitness, as essential behaviors such as feeding and locomotion depend upon proper neuromuscular development. However, uncharacterized defects, such as CNS axon guidance errors, could also debilitate Ptp69D mutants. In particular, the requirement for function in late-stage pupae could indicate a role for DPTP69D in synaptogenesis.

Two primary goals motivated this work. First, we wanted to probe the structure of DPTP69D by analyzing mutants with reduced or otherwise altered function. Second, we wished to generate tools with which to investigate the DPTP69D signaling pathway. By characterizing our panel of new alleles, we have been able to define the developmental stages during which Ptp69D function is required. We have identified allelic combinations with a threshold of DPTP69D function, which are being used to screen for mutations that modify DPTP69D signaling. We have also discovered a putative “substrate trap” mutant with a severe axon guidance phenotype, which is being used to purify substrates of DPTP69D. Thus the alleles described in this article open the door to genetic and biochemical study of DPTP69D signal transduction.

Footnotes

Acknowledgement

We thank Anna Ingalls, Sharon Sams, Amir Lagstein, and Hong Wan for help with fly culture and mutagenesis; David Miller and Peter Kolodziej for critical reading of this manuscript and helpful comments; Kai Zinn for MAbs 3F11, 3F12, and 5A6; and Cory Goodman for MAb 1D4. J.P. was supported by a Training Program in Developmental Biology grant from the National Institutes of Health (NIH; 5T32HD0750205). This work was supported by a grant from the NIH (NSHD38141) to C.D. as well as a Basil O'Conner Starter Scholars Award (5-FY99-0111) from the March of Dimes Birth Defects Foundation.

LITERATURE CITED