The Drosophila slamdance Gene: A Mutation in an Aminopeptidase Can Cause Seizure, Paralysis and Neuronal Failure

Corresponding author: Mark Tanouye, Policy, and Management, 201 Wellman Hall, University of California, Berkeley, CA 94720., tanouye{at}uclink4.berkeley.edu (E-mail)

Communicating editor: T. F. C. MACKAY

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We report here the characterization of slamdance (sda), a Drosophila melanogaster "bang-sensitive" (BS) paralytic mutant. This mutant exhibits hyperactive behavior and paralysis following a mechanical "bang" or electrical shock. Electrophysiological analyses have shown that this mutant is much more prone to seizure episodes than normal flies because it has a drastically lowered seizure threshold. Through genetic mapping, molecular cloning, and RNA interference, we have demonstrated that the sda phenotype can be attributed to a mutation in the Drosophila homolog of the human aminopeptidase N (APN) gene. Furthermore, using mRNA in situ hybridization and LacZ staining, we have found that the sda gene is expressed specifically in the central nervous system at particular developmental stages. Together, these results suggest that the bang sensitivity in sda mutants is caused by a defective APN gene that somehow increases seizure susceptibility. Finally, by using the sda mutation as a sensitized background, we have been able to identify a rich variety of sda enhancers and other independent BS mutations.

MEMBRANE peptidases are a group of ectoenzymes that are widely distributed in animal tissues and have been implicated in a variety of biological functions. They have been shown to be essential for maturation of proteins, activation and inactivation of hormonal peptides, degradation of nonhormonal peptides, and determination of protein stability. They can also function as receptors and as molecules involved in cell adhesion and signal transduction (SANDERINK et al. 1988 ; TAYLOR 1993 ). Aminopeptidase N (APN; EC 3.4.11.2) is a transmembrane ectoenzyme that catalyzes the removal of neutral and basic amino acids from the N termini of a number of small peptide substrates (SHIPP and LOOK 1993 ; RIEMANN et al. 1999 ). The catalytic domain faces the exterior of the plasma membrane and is anchored by a transmembrane-spanning domain. Human APN is identical to CD13, a cluster antigen expressed on the surface of myeloid progenitors, monocytes, granulocytes, and myeloid leukemia cells (LOOK et al. 1989 ). Depending on the species, the APN protein is composed of 963–967 amino acids with a short N-terminal tail in the cytoplasm (9–10 amino acids), a transmembrane segment (23- to 24-amino-acid residues), and a large extracellular ectodomain containing the active site (LUCIANI et al. 1998 ). APN is a member of the M₁ family of zinc-dependent metallopeptidases, which includes related enzymes such as aminopeptidase A, aminopeptidase B, leukotriene A4 hydrolase, puromycin-sensitive aminopeptidase, thyrotropin-releasing hormone-degrading enzyme, the rat vesicle protein Vp 165, and Escherichia coli pepN (MCCAMAN and GABE 1986 ; FUNK et al. 1987 ; SCHAUDER et al. 1991 ; NANUS et al. 1993 ; CONSTAM et al. 1995 ; KELLER et al. 1995 ; RAWLINGS and BARRETT 1995 ; FUKASAWA et al. 1996 ; CADEL et al. 1997 ).

APN is highly expressed in liver, brush borders of kidney, small intestine, and placenta (SHIPP and LOOK 1993 ; RIEMANN et al. 1999 ). It has also been found in the brain, lung, blood vessels, and primary cultures of fibroblasts. APN has been implicated in a variety of tissue-specific functions. In the intestinal brush border, APN functions in the final hydrolysis of ingested nutrients and in amino acid scavenging. On vascular cells, APN serves to metabolize particular vasoactive peptides (WARD et al. 1990 ). In malignant neoplasms, APN is widely considered to influence the invasion mechanism by catalyzing the degradation of collagen type IV and enabling tumor-cell invasion through the basement membrane during metastasis of a primary tumor to vital organs (SAIKI et al. 1993 ; KIDO et al. 1999 ; ISHII et al. 2001 ). APN is thought to play a role in antigen processing and presentation (FALK et al. 1994 ). Cell cycle control and cell differentiation of macrophages/monocytes and mitogenic activation of lymphocytes are also associated with surface APN activity (AMOSCATO et al. 1989 ; KOCH et al. 1991 ). APN appears to be directly linked to signal transduction pathways in monocytes by mediating the release of Ca²⁺ from intracellular stores and the extracellular milieu (SANTOS et al. 2000 ). In humans, APN is thought to act as a receptor for coronavirus 229E and it is a receptor for transmissible gastroenteritis virus in pig (DELMAS et al. 1992 , DELMAS et al. 1994 ; YEAGER et al. 1992 ). In the brain, APN is involved in the enzymatic cascade of the renin-angiotensin system through the cleavage of angiotensin III (ZINI et al. 1996 ). In synaptic membranes, it is widely believed to play the principal role in the inactivation of enkephalin signaling by catalyzing the release of N-terminal tyrosine from the peptide (SOLHONNE et al. 1987 ). In mouse brain cortical slices, APN appears to be associated with the metabolism of nociceptin/orphanin FQ, the natural ligand for the nucleolus organizing region (NOR) receptor (MONTIEL et al. 1997 ; TERENIUS et al. 2000 ). Thus, these studies show that in a variety of different tissues, APN displays an amazing diversity of physiological functions.

In this study, we present evidence for a novel APN function: a role in behavior and nervous system excitability as revealed by Drosophila mutants. One class of behavioral mutants, the "bang-sensitive" (BS) mutant class, has especially intriguing behavioral and electrophysiological phenotypes. The BS class includes several mutants such as bangsenseless (bss), easily shocked (eas), slamdance (sda), and technical knockout (tko). All BS mutants suffer from cycles of intense behavioral hyperactivity and temporary paralysis caused by a mechanical shock, such as a tap of the culture vial on the bench top or brief vortex mixing (a "bang"; BENZER 1971 ; GANETZKY and WU 1982 ). The hyperactivity phenotype is characterized by intense, uncoordinated motor activity featuring wing flapping, leg shaking, and abdominal muscle contractions; the paralytic phenotype, on the other hand, is observed as a cessation of all physical activity (BENZER 1971 ; PAVLIDIS et al. 1994 ). The hyperactivity and paralysis can be mimicked on the electrophysiological level by stimulating and recording from the central nervous system (CNS; PAVLIDIS and TANOUYE 1995 ; KUEBLER and TANOUYE 2000 ). These analyses show that BS mutants have enhanced seizure sensitivity, being 5–10 times more sensitive to seizures than wild-type flies.

So far, only two BS genes have been fully characterized: tko, which encodes a mitochondrial protein, and eas, which encodes an ethanolamine kinase (ROYDEN et al. 1987 ; PAVLIDIS et al. 1994 ). In this article we present findings from two lines of study on sda, a previously uncharacterized BS mutant. First, we report the results of behavioral and electrophysiological testing on adult sda flies, which shows that sda mutants have a dramatically reduced seizure threshold compared to wild-type flies. We then present genetic and molecular data that suggest a mutation in the structural gene for a Drosophila homolog of human APN is responsible for the bang sensitivity in sda mutants. In situ hybridization and LacZ staining reveal sda gene expression, most notably in the CNS. Finally, we describe the results of a genetic screen in which the sda mutation was used to generate a sensitized genetic background: We were able to obtain various enhancers of sda as well as other independent BS mutations. Since it is surprising that a modification in an APN gene can alter neuronal excitability, our analysis of sda may encourage new ways of thinking about factors influencing seizure sensitivity.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Fly stocks and genetic mapping of sda:
Stocks were maintained on standard cornmeal-molasses medium at 22°. Wild-type flies were the Canton-Special (CS) strain. Three BS mutants were used: eas, bss, and sda. The eas gene is located at map position 1-53.5 and encodes an ethanolamine kinase (PAVLIDIS et al. 1994 ). The bss gene is located at 1-54.6; its gene product has not been described (GANETZKY and WU 1982 ). The mutant sda allele, sda^iso7.8, was identified by Dr. Tim Tully (Cold Spring Harbor Laboratory) as a spontaneous mutation that caused behavioral paralysis following a mechanical bang stimulus; it was provided to us as a generous gift. The sda gene was mapped to an apparent map position of 3-95.9 on chromosome 3R on the basis of recombination relative to ebony (e), rough (ro), and claret (ca) markers. Cytogenetic analysis placed the sda locus within a five-band region (97D1–5) on the basis of its inclusion in the deficiencies Df(3R)ro-XB3 = Df(3R) 97D1-2; 97D9 and Df(3R)ro-z1 = Df(3R) 97D1-2; 97D15 and its exclusion from Df(3R)Bd = Df(3R) < 97D5; 97F1-98A1.

Isolation of the sda^HZ.P1 mutation:
The sda^HZ.P1 allele was isolated in a screen utilizing P-element hybrid dysgenesis. It is a recessive lethal of sda that fails to complement the behavioral paralysis phenotype of sda^iso7.8. The sda^HZ.P1 mutation was isolated in a cross utilizing ry P(ry⁺ LacZ)(97D6-9)/ry Sb P(ry⁺ delta2.3) females crossed with ry sda^iso7.8 males. These females contain a P element located at 97D6–9, close to the map position of sda^iso7.8 that might facilitate mutation of sda by local hopping. The starting transposon insert itself does not cause BS phenotypes and complements sda^iso7.8. The female is dysgenic due to the overproduction of transposase by the P(ry⁺ delta2.3). Exceptional ry P(ry⁺ LacZ)(97D6-9)/ry sda^iso7.8 male and female progeny from this cross that show bang-sensitive paralysis are individually crossed to set up appropriate stocks. The screen examined 20,000 flies, and two mutations were identified, one of which, sda^HZ.P1, failed to complement sda^iso7.8. Both the lethality and the failure to complement paralysis phenotypes reverted when the P element of sda^HZ.P1 was lost upon remobilization. The transposon of sda^HZ.P1 was mapped to 97D2–8 by in situ hybridization to polytene salivary gland chromosomes. The sda^HZ.P1 and sda^iso7.8 mutations are tightly linked and recombination experiments have not been able to separate them: Among 800 progeny of heterozygous females, no wild-type recombinants were identified.

Isolation of dominant enhancer mutations for sda^iso7.8/+:
Dominant enhancer mutations of an sda^iso7.8/+ paralytic phenotype were identified in a screen utilizing P-element hybrid dysgenesis. Enhancers were isolated in a mating of X^X,8:P(w⁺ LacZ)/Y; ry Sb P(ry⁺ delta2.3)/+ females crossed to w/Y; sda^iso7.8 males. These females contain an attached-X chromosome with eight mobile P-element transposons, each marked with w⁺ and containing a LacZ reporter, an origin of replication, and an ampicillin-resistance gene to allow cloning via plasmid rescue. The females are dysgenic due to the overproduction of transposase by the P(ry⁺ delta2.3). Exceptional w/Y; sda^iso7.8/+ male progeny that showed bang-sensitive paralysis were individually crossed to set up appropriate stocks. Of 12,000 flies examined, 15 were bang sensitive. Linkage to the second or third chromosome was determined by segregation using T(2;3)ap^Xa and Cyo and TM3 balancers. Six mutations segregated with the second chromosome and nine segregated with the third chromosome.

Behavioral testing:
Testing for BS paralysis was performed on flies 2–3 days posteclosion. Flies were rested for >2 hr after exposure to CO₂ anesthesia before testing. Ten flies were then placed into a clean vial (Applied Scientific) and allowed to rest for an additional 30 min. These flies were vortexed on a VWR vortex at maximum setting for 10 sec and for those flies that showed paralysis, the recovery process was monitored. To test for refractory period, the flies were vortexed again 4–20 min later to see if the flies were still bang sensitive. To minimize data variation due to experimental setting or handling, a large number of flies (n > 100) were analyzed for each strain in this study.

Electrophysiology:
Electrophysiology was performed on flies 2–3 days posteclosion using methods previously described to stimulate and record giant fiber (GF)-driven muscle potentials and seizures (KUEBLER and TANOUYE 2000 ). In brief, the fly was taken from a vial by sucking onto its head with a 23-gauge needle attached to a vacuum line. Another needle attached to a vacuum line was used to suck onto the abdomen, thereby completely immobilizing the fly. The fly was then affixed to a mounting needle by cyanoacrylate adhesive. Two types of stimulation were delivered to the brain, using bipolar tungsten stimulating electrodes: single-pulse stimuli and high-frequency (HF) wavetrains. Single-pulse stimuli (0.5 msec duration, 0.8 Hz) were used to drive the GF. GF-driven muscle potentials were recorded from the dorsal longitudinal muscles (DLMs), using tungsten recording electrodes. GF thresholds were determined as the lowest voltage at which the short-latency GF pathway responded. During the course of each experiment, the GF was stimulated continuously to assess GF system circuit function. To elicit seizures, short wavetrains of HF electrical stimuli (0.5-msec pulses delivered at 200 Hz for 300 msec) were delivered to the brain and the intensity (voltage) of the HF stimulus was varied as noted. Previously, we have shown that seizures in many genotypes, including sda^iso7.8, are elicited in an all-or-nothing manner (KUEBLER and TANOUYE 2000 ). Seizures consist of high-frequency activity in at least seven different muscle groups and >30 muscle fibers in the thorax. The activity in each muscle fiber corresponds to seizure activity in the motoneuron that innervates it. In this article, recordings of DLM muscle potentials were used to denote the occurrence of seizures as described previously (KUEBLER and TANOUYE 2000 ).

Molecular mapping of sda:
Standard molecular techniques were employed for the manipulation of DNA and RNA (SAMBROOK et al. 1989 ). Molecular access to the sda region was achieved via plasmid rescue of genomic DNA from the P-element insertion in sda^HZ.P1 (WILSON et al. 1989 ). Genomic DNA (2–5 µg) was digested with XbaI alone or in combination with either SpeI or NheI, and the fragments were self ligated with T₄ DNA ligase. Ligated products were transformed into JS5 electrocompetent cells (Bio-Rad, Hercules, CA) and the transformants were selected on kanamycin (10 mg/ml) plates plus X-Gal and isopropyl thiogalactoside (IPTG). Plasmid DNA from positive clones (kan^r, LacZ+) was isolated, cut with restriction enzymes, and the fragments were used to probe Southern blots and Northern blots. The intron-exon boundaries of the sda gene were determined by sequencing the cDNA clone LP 11029 (Berkeley Drosophila Genome Project) and aligning the results to sequences in the genome database.

Analysis of mutant sequences:
To examine the molecular basis of the sda^iso7.8 mutation, PCR primers were designed to amplify the coding region of the sda gene in wild-type and mutant flies in six overlapping sections. The predicted products corresponded to the following nucleotide positions: primer set one, 477–1304; primer set two, 1199–2039; primer set three, 1648–2488; primer set four, 2270–3116; primer set five, 2803–3767; and primer set six, 3223–4031. These primers were used to amplify the sda coding sequence by reverse-transcribed PCR using adult whole RNA. The PCR products were cloned into TOPO vectors and sequenced (Invitrogen, San Diego). The insertion mutation in sda^iso7.8 (contained within the PCR product of primer set one) was confirmed by sequencing at least two subclones of reverse-transcribed (RT)-PCR reactions. Other regions were also sequenced at least twice. The insertion site of the P allele, sda^HZ.P1, was determined by first using plasmid rescue to isolate the genomic fragment flanking the insertion site and then sequencing this fragment using a primer complementary to a site near the end of the P-element sequence (ATACTTCGGTAAGCTTCGGC).

Northern blots:
For the developmental Northern blot, whole RNA was extracted from wild-type (w¹¹¹⁸) embryos (0–24 hr), third instar larvae, and adults, using Trizol reagent (GIBCO BRL, Gaithersburg, MD). The RNA was separated on a denaturing gel, blotted onto nitrocellulose membrane, and probed with radiolabeled DNA fragments using standard molecular biology techniques. The probe used here was an 848-bp fragment covering base pairs 62–908 of the sda gene sequence; labeling was done with [-³²P]dCTP [New England Nucleotides, Stratagene (La Jolla, CA) Prime-It kit]. To compare sda^iso7.8 mutants to wild-type flies, whole RNA was isolated from adults and analyzed in the same manner described above.

RNA interference:
RNA interference (RNAi) was performed as described previously (KENNERDELL and CARTHEW 1998 ). To synthesize the DNA template, forward (base pairs 62–79) and reverse (base pairs 908–893) primers flanking an 846-bp region of the sda sequence were made. Each primer had on the 5' end a 30-bp sequence 5'-CGGATCCTAATACGACTCACTATAGGGAGA-3', which contains a T7 promoter sequence (underlined) for in vitro transcription and extra nucleotides to optimize efficiency. The transcription was performed using the RNA transcription kit from Stratagene. The double-stranded RNA (dsRNA) product was purified, precipitated, and assayed using A₂₆₀ as described (KENNERDELL and CARTHEW 1998 ). The dsRNA was dissolved in TE buffer to the desired concentration and was injected into w¹¹¹⁸ embryos (less than stage 3), using a needle with a tip opening of 0.5–2.5 µm. Injected embryos were covered with Halocarbon 600 oil and placed at 18° in a moist box for recovery. After hatching the surviving larvae were transferred to a culture vial, grown to adults at room temperature, and tested for BS behavior. For control experiments, separate collections of wild-type embryos were prepared and mock injected, injected with TE buffer, or injected with dsRNA (2.0 or 3.0 µM) prepared against the sticky ch1 (stich1) coding sequence (PROKOPENKO et al. 2000 ).

Embryo and larval mRNA in situ hybridization:
The experiments were conducted according to methods previously reported (KOPCZYNSKI et al. 1998 ). The DNA template used spans base pairs 62–908 of the sda LP11029 transcript. The antisense (reverse) primer has a T7 promoter sequence to synthesize specifically antisense sda RNA. RNA synthesis was performed using the GIBCO Riboprobe synthesis kit that uses digoxygenin-labeled uridine. The labeled antisense RNA molecules were hybridized to whole embryos, and color visualization was produced using anti-digoxygenin alkaline phosphatase-conjugated antibodies. The final stained embryos were mounted on slides and examined under a digital compound microscope. Photos were taken using the program Spot and prepared using Adobe Photoshop. For larval mRNA in situ hybridization, the third instar larval CNS was prepared using previously described methods (MASSUCI et al. 1990 ). Following CNS preparation, the rest of the experiment (probe synthesis, hybridization, and staining) was done as described for embryos.

LacZ reporter staining in sda^HZ.P1 adult CNS:
Adults of sda^HZ.P1 were decapitated using forceps and the tissues immobilized by freezing in O.C.T. (Tissue-Tek). The frozen block was then sliced into 10-µm sections by cryostat, and the sections were blotted onto slides pretreated with poly L-lysine (Sigma, St. Louis). The tissues were fixed in 2% glutaraldehyde for 15 min and then washed three times in 1x PBS buffer for 5 min at room temperature. The slides were placed in staining solution with 1/30 volume X-Gal [composition of staining solution: 1.8 ml of 0.2 M Na₂HPO₄, 0.7 ml of 0.2 M NaH₂PO₄, 1.5 ml of 5.0 M NaCl, 50 µl of 1.0 M MgCl₂, 3.0 ml of 50 mM K₃(Fe(CN)₆), and 3.0 ml of 50 mM K₄ (Fe(CN)₆); total volume is brought to 50 ml with H₂O]. The sections were then stained until the desired intensity was obtained. Final results were photographed with a digital compound microscope and the pictures processed via Adobe Photoshop. A control experiment was performed with w¹¹¹⁸ flies.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The sda behavior:
The behavioral phenotypes of sda mutants are generally similar to other mutants of the BS paralytic class such as bss, eas, and tko. Undisturbed sda flies do not show notable defects in specific behaviors: They eat, walk, jump, fly, groom, court, and mate normally; they show usual positive-phototaxis and negative-geotaxis behaviors. There are no apparent alterations in the overall levels of activity such as hyperactivity or sluggishness. Behavioral abnormalities are induced in all homozygous sda mutants by a mechanical shock (a bang). The resulting behavioral phenotype is complex with five distinguishable phases: initial seizure, paralysis, recovery seizure, recovery with refractory period, and complete recovery. The initial behavioral seizure is characterized by leg shaking, abdominal muscle contractions, wing flapping, and proboscis extension; this phase usually lasts several seconds. This is followed by complete paralysis with no physical activity observed and distinguished by a relaxed state of the wings, legs, body, and proboscis; paralysis lasts 20 sec. Each sda mutant then shows a postparalysis hyperactive phase or recovery seizure, characterized by massive uncoordinated motor activity somewhat similar to the initial seizure. Finally, the flies right themselves and resume normal behavior.

The recovery time (from the start of the bang to when the flies stand back up again) varies among different BS mutants (Fig 1). For example, in sda flies the time for 50% recovery is 37 sec, faster than the recovery times for eas and bss, which are 140 and 150 sec, respectively (Fig 1). All of the different BS strains are similar in initial seizure, paralysis, and recovery seizure. However, following the recovery seizure, only sda mutants recover immediately. Other strains, most notably bss, undergo additional bouts of paralysis and seizure that resemble tonic-clonic activity in human epilepsy and that can last for many minutes, thereby increasing the time of recovery. The relatively rapid recovery of sda mutants appears to be entirely due to the lack of any tonic-clonic activity. Following recovery, sda mutants resume normal behaviors. Interestingly, immediately following recovery, sda mutants cannot be reparalyzed by mechanical stimulation; that is, the mutants are no longer BS. This is termed "the refractory period" and is a transient period present in all BS genotypes, although it varies in duration among the different strains. For sda flies, the refractory period is 7 min and is shorter than those for eas and bss, which are 10 and 12 min, respectively.

View larger version (15K): In this window In a new window Download PPT slide   Figure 1. Behavioral analysis of sda and other BS mutants. Flies of the genotype indicated were shocked mechanically in batches of 10 flies in an empty food vial on a vortex mixer (10 sec at maximum setting) and allowed to recover from paralysis. The number of flies standing at intervals following the shock was recorded until the entire population had recovered. A given population of flies recovers in a reproducible way, although there is tremendous variability in recovery times among the different genotypes. Shown are the percentage of flies recovered with time and a cumulative measure that includes the initial behavioral seizure, the paralysis period, and the recovery seizure, which are not indicated separately. Genotypes include CS (n = 100 flies), sda (n = 160 flies), bss (n = 100 flies), and eas (n = 149 flies).

The sda mutation is a weak semidominant in behavioral tests. Heterozygous sda/+ flies show mostly normal behavior, although a few (1–2%) are BS. The semidominant BS phenotype is more readily apparent if tests are performed exclusively on very young flies. For example, tests on young flies 1–2 days posteclosion show that as many as 45% can show some BS paralytic behavior. However, this phenomenon is not consistent across the other BS genotypes; for example, old eas and bss flies (>4 days) actually show a stronger paralytic phenotype than do young flies (1–2 days posteclosion; P. PAVLIDIS and M. TANOUYE, unpublished observations).

Seizure and failure in the GF pathway of sda adults:
The electrophysiological phenotypes of sda mutants are generally similar to other mutants of the BS paralytic class such as bss, eas, and tko. In tests of the general properties of the GF system, sda flies responded normally. Stimulation of sda GFs with single stimulus pulses (0.5 msec, 0.8 Hz) produces DLM responses that are normal in appearance, threshold (2.2 ± 0.38 V), and latency (1.3 msec). When tested with twin pulses, sda DLMs followed GF stimuli separated by a minimum of 10 msec, similar to wild-type flies (TANOUYE and WYMAN 1980 ). When tested for following frequency by 20 stimulus pulses, sda DLMs followed at least 19 of the stimuli at a frequency of 131 ± 17.6 Hz, similar to wild-type flies (KUEBLER and TANOUYE 2000 ). Thus, under conditions of mild to moderate stimulation, the GF-DLM responses of sda are normal.

Electrophysiological analysis of sda mutants with HF stimuli shows that seizures may be induced in individual flies and that these mutants are particularly seizure sensitive compared to wild-type, similar to other mutants in the BS mutant class. We investigated the electrophysiological basis of sda seizure and paralysis using a standard protocol for stimulating and recording from the adult fly GF pathway that has been described previously (KUEBLER and TANOUYE 2000 ). The features of sda mutant electrophysiology described here are qualitatively similar to those reported previously for other fly genotypes (Fig 2A; PAVLIDIS et al. 1994 ; PAVLIDIS and TANOUYE 1995 ; KUEBLER and TANOUYE 2000 ; KUEBLER et al. 2001 ). "Seizures" in sda mutants following HF stimuli consist of aberrant high-frequency firing (>100 Hz) lasting for 2–3 sec and present in all the muscle fibers and motoneurons examined (KUEBLER and TANOUYE 2000 ). For sda males and females, seizure thresholds are 6.2 ± 0.8 and 6.8 ± 1.0 V, respectively. These seizure thresholds are considerably lower than those for male and female CS flies whose values are 30.1 ± 3.8 and 44.5 ± 4.4 V, respectively. Thus, using the criterium of HF stimulus threshold, sda mutants are five to seven times more seizure sensitive than wild-type flies. The next aspect of the phenotype is the sudden failure of GF stimulation to evoke DLM potentials (Fig 2D). Such failure in the nervous system is due to synaptic failure in many central synapses and is likely the underlying cause of behavioral paralysis in sda mutants (PAVLIDIS and TANOUYE 1995 ). The period of synaptic failure has been measured as 38.0 ± 7.0 sec in sda (PAVLIDIS and TANOUYE 1995 ). Following the period of synaptic failure, a recovery seizure is observed that consists of aberrant high-frequency firing of the DLM motoneurons, similar to the initial seizure in scale and duration (Fig 2D).

View larger version (29K): In this window In a new window Download PPT slide   Figure 2. Seizures in sda and CS flies. The mutant sda fly is more susceptible to seizures than the wild-type (CS) fly and therefore has a much lower seizure threshold. (A) A seizure is elicited in a sda fly by a high-frequency stimulus of low strength (8 V) and displayed at a high sweep speed. The HF stimulus (HFS) is a short wavetrain (0.5 msec pulses at 200 Hz for 300 msec) of electrical stimuli delivered to the brain. Recording is from a DLM muscle fiber and reflects the activity of the single DLM motoneuron that innervates it. The seizure is widespread as similar activity can be found in recordings from seven different muscle groups in the fly following HF stimulation (KUEBLER and TANOUYE 2000 ). (B) A low-voltage HF stimulus of 8 V fails to elicit a seizure in a wild-type CS fly because the stimulus is below the seizure threshold. Following the HF stimulus artifact, no seizure activity is observed in this recording displayed at a high sweep speed. Note also that there is no period of synaptic failure and single-pulse stimulation of the GF (0.5 Hz) continues to evoke DLM potentials. Two such effective single-pulse stimuli are depicted in this trace; each was effective in evoking a DLM potential. (C) A seizure is elicited in a wild-type CS fly by a high-voltage HF stimulus (30 V), which is above the threshold for seizure. The seizure in this recording begins within the large stimulus artifact and is displayed at a high sweep speed. (D) Same recording as for A from a sda fly, but displayed at a slower sweep speed. In this recording, the HF stimulus and seizure are followed by a quiescent period (SYNAPTIC FAILURE) that is characterized by synaptic failure within the GF circuit (PAVLIDIS and TANOUYE 1995 ). During this period, there are stimulus artifacts (downward-going) from continuous single-pulse stimulation of the GF (0.5 Hz), but no evoked DLM potentials. Spontaneous activity or "recovery seizure" appears as additional seizure-like activity occurring just after the synaptic failure period and just prior to recovery (recovery not evident in this trace). Vertical calibration bars are 20, 40, 40, and 10 mV for A, B, C, and D, respectively. Horizontal calibration bars are 300 msec, 1.2 sec, 1.2 sec, and 1.5 sec for A, B, C, and D, respectively (modified from KUEBLER and TANOUYE 2000 , Figure 1).

The sda mutation is semidominant in HF electrophysiology. Heterozygous sda^iso7.8/+ mutants show a lower seizure threshold than do wild type (30.6 ± 4.5 V in heterozygous females vs. 44.5 ± 4.4 V in CS females). This is interesting to us as it suggests that, although appearing to be largely wild type in behavior and electrophysiology, sda/+ heterozygous mutants may be fairly close to expressing seizure-sensitive phenotypes. This suggests that heterozygotes could be used to provide a sensitized genetic background for detecting other mutations affecting seizure susceptibility such as weak BS alleles and BS enhancers.

Cloning and characterization of sda:
We mapped sda to a small region on the third chromosome (97D1–5) defined by three closely spaced deletion breakpoints, Df(3R)ro-XB3, Df(3R)ro-z1, and Df(3R)Bd (Fig 3). Initial identification of the sda gene was made possible by the isolation of a lethal P-element allele, sda^HZ.P1 (Fig 3). Molecular access to the sda gene was via the sda^HZ.P1 mutation by the method of plasmid rescue of genomic DNA flanking the transposon insertion. Sequencing of the rescued genomic DNA fragment and comparison of the results with information available in the Drosophila database further confirmed a 97D location for the sda gene. Using the genomic fragment as a probe, we identified a genomic DNA clone from a -phage library; we sequenced this 8-kb fragment and used it to identify cDNA clone LP11029 from the Berkeley Drosophila Genome Project (Fig 4). Two transcripts are encoded by sda. One is a 4.8-kb transcript that corresponds to the full-length LP11029 cDNA. A second is a 2.2-kb transcript. Screening of multiple cDNA libraries has failed to identify a cDNA corresponding to the 2.2-kb transcript so its characterization is unavailable presently. Southern and Northern blot analysis of the LP11029 cDNA and 4.8-kb transcript shows that sda is a large gene that spans 30 kb of genomic DNA and contains eight exons (Fig 3). Introns II and III are especially large intervening sequences of 8 and 13 kb, respectively.

View larger version (15K): In this window In a new window Download PPT slide   Figure 3. Genetic and molecular representation of the 97D region that contains sda. (A) Cytogenetic analysis places the sda locus in a five-band region on the basis of its inclusion in Df(3R)ro-xB3 = Df(3R)97D1-2; 97D9 and Df(3R)ro-z1 = Df(3R)97D1-2; 97D15 and its exclusion from Df(3R)Bd = Df(3R)<97D5; 97F1-98A1. The sda gene is located on the same scaffold as Tl (cytogenetic location 97D1–2) and ro (97D2–3) and maps to the right of both. Depicted is the 97D1–97D5 interval, the approximate locations of the defining deletion breakpoints marked with solid triangles, and the apparent order of Tl ro sda within the interval. (B) Depicted is a representation of scaffold AE003758 (Berkeley Drosophila Genome Project), showing 200 kb and the relative locations of the Tl, ro, and sda genes. Arrow for sda indicates the direction of transcription. (C) A blowup showing the genomic organization of the full-length cDNA LP11029. Roman numerals designate exons. The stippled bars within the cDNA represent coding sequences. Small open triangles indicate the location of the sdaHZ.P1 P-element transposon insertion in exon I and the sdaiso7.8 insertion in exon III; both insertions are located in the 5' UTR.

View larger version (135K): In this window In a new window Download PPT slide   Figure 4. Sequence of a sda cDNA. Shown is the complete sequence of cDNA clone LP 11029 with its predicted amino acid translation product. The insertion site of the sdaHZ.P1 P-element transposon between nucleotides 61 and 62 is indicated by a box, as is the insertion site of sdaiso7.8 between nucleotides 671 and 672. The predicted membrane-spanning segment is indicated by the underline from amino acid position 34 to 58. Shading indicates the glucinzin Zn2+-binding motif [HEXXH-(18X)-E] beginning at amino acid position 486 and the GAMEN motif (AAMEN in the sda sequence) beginning at amino acid position 447.

Sequence analysis of LP11029 revealed that it consists of 4811 nucleotides, 3213 of which code for a putative protein of 1071 amino acids. Comparison of the deduced protein with sequence databases revealed significant similarity to previously identified human APN (Fig 3 and Fig 4; LOOK et al. 1989 ; WATT and WILLARD 1990 ; YEAGER et al. 1992 ). The aminopeptidase encoded by sda is distinct from a recently described Drosophila sequence located at 84F6–85A3, which encodes for a dipeptidyl aminopeptidase that degrades the insect neuropeptide proctolin (MAZZOCCO et al. 2001 ). The deduced sda protein sequence displays canonical features present in human APN and other members of the M₁ family of zinc-dependent ectopeptidases. The sequence predicts an N-terminal cytoplasmic tail of 33 amino acids for sda. This cytoplasmic tail is longer than human APN due to the presence of 24 unconserved amino acids at the sda N terminus. The predicted sda protein contains a transmembrane segment of 25 amino acids and an ectodomain of 1013 amino acids. The gluzincin Zn²⁺-binding motif [HEXXH-(18X)-E] that constitutes the core of the active site of both gluzincin aminopeptidases and endopeptidases is present in the sda protein, starting at amino acid position 486. In this segment of the protein, there is 84% identity between sda and human APN (26 identical residues out of 31 total amino acids in the segment). The GAMEN motif, identified as another conserved sequence motif in the gluzincin aminopeptidases, is present in the sda protein as AAMEN and is positioned 39 amino acids N-terminal to the Zn²⁺-binding motif. In human APN, the conserved glutamate residue of this motif has been shown to be a crucial residue in an anionic binding site recognizing the free amino group at the N terminus of the substrate (LUCIANI et al. 1998 ). The overall amino acid similarity shows that the Drosophila protein is 33% identical and 51% similar to human APN and 31% identical and 50% similar to mouse APN. Substantially greater identity is observed in the catalytic portions of the ectodomain (Fig 4).

Molecular basis of sda mutations:
We determined the molecular basis of the known sda alleles: the original isolate sda^iso7.8 and sda^HZ.P1 acquired in the course of this work. The exact insertion site of the P element of sda^HZ.P1 was found by sequencing the genomic DNA fragment from plasmid rescue using a primer targeting the end of the P-element sequence. This analysis showed that sda^HZ.P1 had inserted in exon I between nucleotides 61 and 62 in the 5' untranslated region (UTR) of the gene. We determined the molecular basis for the spontaneous allele sda^iso7.8 by using RT-PCR to amplify the coding sequence of the sda gene in wild-type and mutant flies. The resulting products were sequenced and compared. This analysis revealed a 2-bp insertion in exon III between nucleotides 671 and 672 in the 5' UTR of the gene (Fig 5). The molecular basis of both of the sda alleles is consistent with the LP11029 cDNA we identified as representing the sda gene. These results suggest that sda mutant phenotypes most likely arise from underexpression or perhaps misexpression of Drosophila APN.

View larger version (118K): In this window In a new window Download PPT slide   Figure 5. Sequence homology of the predicted sda gene product. Sequence homology of the predicted sda protein product with human APN (APN-h) and mouse APN (APN-m) members of the M1 family of zinc-dependent metallopeptidases (OLSEN et al. 1988 ; STRAUSBERG 2001 ). Amino acid identities are indicated by solid boxes and similarities are indicated by shaded boxes. Overall, sda is 33% identical and 51% similar to human APN and 31% identical and 50% similar to mouse APN. The thick solid bar indicates the predicted membrane-spanning segment of sda. Open boxes indicate the glucinzin Zn2+-binding motif [HEXXH-(18X)-E] and the GAMEN motif (AAMEN in the sda sequence).

RNAi of Drosophila APN causes BS phenotypes:
We attempted to generate sda phenotypes in non-BS flies by altering normal levels of APN expression using the method of RNAi (CARTHEW 2001 ). RNAi was performed using dsRNA corresponding to an 824-bp segment of the sda coding transcript. We injected w¹¹¹⁸ embryos with a 1.8-µM solution of dsRNA and assayed for BS behavior among the surviving adults. Results from the experiment showed that 31 of the 34 surviving adults exhibited a BS behavioral phenotype: They became paralyzed following mechanical stimulation. The BS behavioral phenotype persisted in these flies for at least 10 days. There were some quantitative differences between the RNAi-treated flies compared to sda^iso7.8 mutants. The recovery time was shorter for the RNAi-treated flies (15 sec) compared to the mutant (37 sec); also, the refractory period was shorter for the RNAi-treated flies (3 min) compared to the mutant (7 min). Control injections using non-APN dsRNA (2.0 or 3.0 µM) yielded flies with no discernible BS phenotype. The indication from this analysis is that specific alteration of APN expression by RNAi is sufficient to generate BS behavioral phenotypes in wild-type animals. We also tried a lower concentration of APN dsRNA (0.9 µM). Surviving adults (30 flies tested) did not become paralyzed by the mechanical stimulus, but many of them showed an unusual neurological defect of "wobbliness" manifest as unsteady legs and shaky movement. This type of phenotype has not previously been associated with BS mutants, but has been observed in other BS genetic combinations in our collection (D. HEKMAT-SCAFE and M. TANOUYE, unpublished observations). These results suggest that the BS phenotype observed in sda^iso7.8 mutants is the consequence of the lesion found in the sda gene, for two reasons: (1) When we inhibit the expression of the gene in non-BS flies by introducing dsRNA, we can mimic the BS phenotype observed in sda mutants, suggesting that the gene product is directly related to bang sensitivity, and (2) the fact that interference appears to be proportional to the concentration of dsRNA used implies that the observation is a true representation of gene interference and not just a coincidence.

sda mRNA: Northern blot analysis:
From our results of the RNAi experiment and analysis of the nature of molecular lesion in the sda mutants, it appears that the molecular defect associated with the sda mutation is very likely an abolition or at least a downregulation of sda gene expression. To test our hypothesis, we performed a Northern blot comparing sda gene expression between wild-type and sda^iso7.8 adults (Fig 6). The results demonstrate that while the 2.2-kb variant is expressed at low levels in both wild-type and mutant flies, the 4.8-kb variant is present only in the former but not the latter. This observation supports our RNAi results and provides a further link between the sda gene and the BS phenotype.

View larger version (41K): In this window In a new window Download PPT slide   Figure 6. Northern blotting of wild-type vs. sda adults. Whole RNA was extracted from both wild-type (wt) and sda mutant (sda-) adults, run on a denaturing gel, blotted onto nitrocellulose membrane, and probed with a fragment specific for the 5' UTR of the sda CDS. The results indicate that the 4.8-kb transcript of the sda gene (LP11029) is expressed only in wild-type adults but not in mutant adults. A second variant (2.2 kb) also appears to be expressed in both mutant and wild-type flies, but its constituency and identity are currently unknown (see text). We also probed for ribosomal protein 49 (rp49) to ensure that equal amounts of RNA were loaded for both wild-type and sda strains.

We also performed a developmental Northern blot of wild-type flies (embryos, third instar larvae, and adults; data not shown) to characterize the temporal expression of the gene. Our results show that the 4.8-kb sda transcript is expressed throughout development, with expression highest in the adult stage and lowest in the embryonic stage; the 2.2-kb transcript is also expressed in each of the three stages examined, although the expression pattern differs notably from that of the longer splice form. For the 2.2-kb transcipt, expression is lowest in the adult and highest in third instar larva. In all three stages examined, expression of the 4.8-kb transcript is greater than that of the 2.2-kb variant.

mRNA in situ and LacZ reporter analysis:
To better understand how sda might function in the organism, we examined its gene expression at various developmental stages using mRNA in situ hybridization and LacZ reporter staining. In our embryo in situ experiment, we found that sda is initially expressed at very low levels (stages 1–12), but starting at stage 13 its expression drastically increases (Fig 7). During stages 13 and 14, sda is expressed profusely in the spiracle region, proventriculus, and, most interestingly, in distinct patterns in the CNS. This pattern changes from stage 15 onward as expression in the CNS and spiracle regions decreases (but is still notable) while expression in the proventriculus and gut regions increases (Fig 8). We compared the expression pattern of sda between wild-type flies and sda mutants and found no gross differences in terms of the locality in which sda is expressed or the intensity of its expression (Fig 6 and Fig 7).

View larger version (82K): In this window In a new window Download PPT slide   Figure 7. Expression of sda in early and middle embryonic stages. The expression of sda in the embryonic stages may be categorized into three phases: early stage (stages 1–11), middle stage (stages 12–14), and late stage (stages 15–17, shown in Fig 8). Expression was probed using mRNA in situ with an antisense RNA probe specific to the sda gene. (A–C) The early-stage expression for wild-type embryos, sda mutant embryos, and a blowup of the stained area in wild-type embryos, respectively. Here we observed very little expression of sda, with only a barely notable staining in the putative CNS region. (D–F) The middle-stage expression pattern for wild-type embryos, sda mutant embryos, and a blowup of regions of prominent expression from the wild-type picture, respectively. During this period sda is expressed profusely in the CNS (arrow 1), putative gut (arrow 2), and the spiracles (arrow 3). Note in particular the expression in the CNS. Taking into account possible discrepancies due to experimental preparations and photographic techniques, we cannot discern any gross differences between wild-type and sda mutant embryos in the locality or intensity of sda expression.

View larger version (53K): In this window In a new window Download PPT slide   Figure 8. Expression of sda in late embryonic stages and in the larval CNS. Expression was probed using mRNA in situ with an antisense RNA probe specific to the sda gene. The larval CNS was dissected from third instar larvae and stained as for embryos; only the ventral ganglion is shown since staining is restricted to this region of the CNS. (A–C) The late-stage expression for wild-type embryos, sda mutant embryos, and a blowup comparison, respectively. During this interval sda expression seems to be reduced compared to the middle stage, especially in the CNS and the spiracle region; in the putative gut, however, expression remains very strong. When compared to each other, again we do not see any gross discrepancies in sda expression between wild-type and mutant embryos. (D–F) The expression of sda in the wild-type larval CNS. Interestingly, here we see that sda is expressed in three tiny clusters bilaterally symmetric on the surface of the ventral ganglion. Although we do not know currently what they are, they are very likely neuronal precursors to certain subclasses of neurons. This three-cluster staining is observed in all the larval CNS dissected and stained (n = 12) and must be distinguished from nonspecific staining that occurs randomly and occasionally among the specimens.

Since sda expression is observed in embryonic CNS, we were curious to see if it is expressed in the CNS of later stages. Therefore, we performed a mRNA in situ hybridization specifically on excised CNS from third instar larvae (Fig 8). We found, to our curiosity, that sda is expressed prominently in the ventral ganglion in three small but distinct clusters. Although at present we do not know for certain what these clusters represent, it is likely that they are groups of neural precursors.

And finally, since the sda BS phenotype is an adult phenomenon, we sought to analyze the gene expression in the adult CNS. To achieve this we sectioned heads of sda^HZ.P1 flies and utilized the LacZ reporter gene located in the P element as a reporter to see how sda is expressed in the CNS. The results we obtained are shown in Fig 9. There is prominent staining in the mushroom bodies, protocerebrum, and antennae (although the exact identity of the cells stained is not known), and the results are very consistent when compared across similar sections. Rough estimates suggest that 40–50 cells in the protocerebrum, 20–30 cells in the mushroom bodies, and 30–40 cells in the antennae express sda. Control experiments performed with w¹¹¹⁸ adults showed no discernible staining and eliminate the possibility that the staining from sda^HZ.P1 heads arises from endogenous ß-galactosidase activity.

View larger version (613K): In this window In a new window Download PPT slide   Figure 9. LacZ reporter staining in adult CNS. sdaHZ.P1 adults were decapitated, their heads placed in an embedding medium, and the tissues immobilized by freezing. The heads were then sliced into 10-µm sections and mounted on pretreated slides. (A–C) Various sections of the head; note the peculiar semicircular nature of sda expression in the CNS, which is located in the mushroom bodies and protocerebrum. Finally, D shows an antenna with substantial staining in the areas directly beneath the cuticles. Blue spots denote the region of observed LacZ staining.

sda enhancer screen:
We examined the possibility that sda^iso7.8/+ heterozygotes could provide a sensitized genetic background, facilitating the identification of new BS paralytic mutants. An interesting feature of the BS mutant class is that each of the mutants thus far examined has been extremely seizure sensitive. The characterization of new BS mutations in Drosophila could improve our understanding of factors influencing seizures and give new insight into the difficult problem of human seizure disorders. As shown, the sda^iso7.8 mutation acts as a semidominant allele in electrophysiology and behavioral tests. Thus, we hypothesized that sda^iso7.8/+ could provide a background for detecting new mutations affecting seizure sensitivity. P elements were used as a mutagen and mobilized in dysgenic attached-X females; female progeny from this mobilization were then crossed with sda^iso7.8 males. The offspring were tested by mechanical stimulation, selecting for those displaying behavioral paralysis.

The resulting progeny are doubly heterozygous for sda^iso7.8 and the transposon insertion. We examined 12,000 chromosomes and isolated 15 mutant lines (Table 1). These doubly heterozygous lines varied in the percentage of flies susceptible to seizures. For example, line M showed the weakest phenotype with 40% paralysis for the double heterozygotes (genotype M +/+ sda^iso7.8). Apparently, the transposon of M acts as a weak dominant enhancer of the antimorphic nature of sda^iso7.8. Line A showed a stronger phenotype with 100% paralysis for the double heterozygotes (genotype A +/+ sda^iso7.8). Apparently, the transposon of A acts as a strong dominant enhancer of the antimorphic character of sda^iso7.8.

For each of the isolated lines, we performed the appropriate crosses to remove sda^iso7.8 from the background and homozygose the transposon. Five lines (lines A, D, F, J, and O; Table 1) each had flies that paralyzed following mechanical stimulation (10–100%), suggesting that their respective P elements may have identified new members of the BS paralytic class. Four lines (lines G, H, I, and N; Table 1) yielded no homozygous mutant progeny, suggesting that the transposon caused a recessive lethal mutation. For four other lines (lines B, C, L, and M; Table 1) viable, homozygous mutant flies displayed no discernible phenotypes; the mechanisms by which these act to enhance the sda/+ phenotype remain unclear. P elements were mapped to various second and third chromosome locations by chromosomal in situ hybridization as indicated in Table 1. For each line, genomic DNA was isolated by plasmid rescue. DNA sequence analysis has identified several gene candidates, including a filamin actin-binding protein gene at 59A (line D), a helix-loop-helix protein gene at 86B (line A), and an RNA-binding protein gene at 90D (line N).

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The sda gene encodes a Drosophila APN:
Molecular mapping has localized the Drosophila sda mutations sda^iso7.8 and sda^HZ.P1 to a transcription unit that displays similarity to the human APN gene. RNAi analysis has shown that interference with Drosophila APN expression in wild-type animals causes behavioral abnormalities that resemble those observed in sda^iso7.8 mutants. These observations taken together strongly suggest that sda is Drosophila APN and that defects in its normal gene function are responsible for all of the observed mutant phenotypes, including a lowered threshold to seizures induced by HF stimuli. This role of altered excitabilities in the nervous system has not been suggested previously and adds to the extensive list of functions that are known for human and mouse APN.

The nature of sda mutations:
Molecular examination of sda^iso7.8 and sda^HZ.P1 indicates their molecular lesions as a small deletion and insertion into the 5' UTR of sda, respectively. Our expectation is that both mutations should cause alterations in sda expression in the form of downregulation or misexpression. The former possibility would fit in well with the RNAi results leading to BS phenotypes. This idea is indeed strongly supported by our Northern blot analysis comparing sda expression of wild-type vs. sda^iso7.8 mutant adults (Fig 6). It is interesting to note here, however, that our embryo mRNA in situ results do not demonstrate any gross differences in expression of sda between the wild-type and mutant strains. There are several possible explanations for this. The simplest reason could just be that sda expression (the 4.8-kb variant) in mutants is not downregulated until a later stage. A second possibility might be that the staining performed in the in situ hybridization was indiscriminate between the 4.8- and the 2.2-kb splice variants (since the probe spans the 5' UTR, a region that appears to be shared by both variants) so that the results cannot distinguish the absence of one or the other; as a result, staining might appear similar although one variant may be entirely absent. Nonetheless, the fact that sda is not expressed in mutant adults is very strong evidence for the gene being responsible for the BS phenotype and also is consistent with our RNAi data. We are currently trying to raise antibodies to the SDA protein to conduct antibody-staining experiments to further document the localization of the sda gene product as well as fine-tune our mRNA in situ results.

A role for APN in nervous system excitability:
There are several possibilities for how APN might alter nervous system function or structure and thereby contribute to seizure sensitivity. One possibility is suggested by its involvement in mammalian neuropeptide processing and degradation (SOLHONNE et al. 1987 ; ZINI et al. 1996 ; MONTIEL et al. 1997 ; TERENIUS et al. 2000 ). Neuropeptides have not, thus far, been widely implicated in epilepsy, although a knockout mutation of neuropeptide Y has led to epileptic phenotypes in the mouse (BARBAN et al. 1997 ). Another possibility is suggested by APN involvement in malignant neoplasms (SAIKI et al. 1993 ; KIDO et al. 1999 ; ISHII et al. 2001 ). Tumor-cell invasion is a complex process involving cell adhesion, motility, and degradation of tissue and extracellular matrix barriers. This process resembles neuronal cell migration and growth cone outgrowth that have been implicated in epileptic syndromes in human and mouse (for example, see AIGNER et al. 1995 ; EKSIOGLU et al. 1996 ; FOX et al. 1998 ).

An especially interesting possibility for how APN might act to alter nervous system excitability comes from a recent report implicating APN in Ca²⁺-mediated signal transduction in monocytes (SANTOS et al. 2000 ). Anti-APN mAbs that inhibit enzyme activity induced a transient rise in intracellular Ca²⁺ when incubated with monocytes. The Ca²⁺ increase begins at 30 sec and peaks at 60 sec. A Ca²⁺ increase was not observed with control anti-APN mAbs that did not inhibit enzyme activity or with mAbs that are directed against another myeloid marker (CD33). Subsequent experiments showed that the increase arose from two separate Ca²⁺ sources. An early response was due to release from intracellular Ca²⁺ stores, possibly the sarcoplasmic reticulum; a more sustained Ca²⁺ response was due to an influx of external Ca²⁺. Tyrosine kinase inhibitors were able to inhibit the rise in Ca²⁺ induced by ligation of APN, as were inhibitors of the phosphatidylinositol 3-kinase. It was suggested that normally in vivo peptides, as yet unidentified, act as ectopeptidase ligands to cause signal transduction directly via APN.

Although a similar function in brain APN has not been described yet, a parallel in epilepsy investigations shows an important role for Ca²⁺ signaling. Spontaneous mutations in mouse at the tottering, lethargic, and stargazer loci have each been shown to cause generalized absence epilepsy and cortical spike-wave discharges. The tottering locus has been shown to encode a Ca²⁺ channel -subunit (FLETCHER et al. 1996 ); the lethargic locus encodes a Ca²⁺ channel ß-subunit (BURGESS et al. 1997 ); and stargazer encodes a Ca²⁺ channel -subunit (LETTS et al. 1998 ). In addition to Ca²⁺ channels, Ca²⁺ signaling appears to contribute to epilepsy syndromes in other ways. For example, seizure phenotypes are observed in a mouse knockout mutation affecting calcium calmodulin kinase -subunit (BUTLER et al. 1995 ). Mutations in synapsin I and synapsin II, molecules that mediate Ca²⁺-dependent synaptic vesicle release, also cause seizure phenotypes (ROSAHL et al. 1995 ). Also interesting are knockout mutations in mouse inositol 1,4,5-trisphosphate receptors that cause seizures and in the tyrosine kinase receptors that suppress seizures (CAIN et al. 1995 ; MATSUMOTO et al. 1995 ).

Drosophila BS mutants:
Genetic and molecular analysis of Drosophila behavioral mutants has been an effective way to identify molecules regulating nervous system excitability, such as ion channel genes (WU and GANETZKY 1992 ). It is thought that knowledge gained from the Drosophila nervous system can be applied to the study of mammalian nervous system function and pathology. The underlying assumption is that even though there are differences between insect and mammalian nervous systems at the gross anatomical level, many of the fundamental cellular and molecular mechanisms regulating excitability are conserved (BENZER 1971 ). A further assumption is that these fundamental mechanisms, when altered by mutation, become manifest as behavioral phenotypes. Studies of Shaker behavioral mutants and their resulting identification and analysis of potassium channels, for instance, have given support to this notion (KAMB et al. 1987 ; TSENG-CRANK et al. 1990 ).

One class of Drosophila behavioral mutants, the BS mutant class, has not been studied extensively although its behavioral and electrophysiological defects are particularly intriguing. The BS mutants have enhanced seizure sensitivity, and by studying them we may increase our understanding of what influences seizure susceptibility, a central issue in such prominent maladies as human epilepsy (MCNAMARA 1994 ). A mutant screen has been described here that uses sda/+ heterozygous mutants as a sensitized genetic background for the isolation of new BS mutations, a procedure that is quite difficult in wild-type backgrounds. Preliminary results from this screen show that it is very efficient in identifying BS mutations using P elements as mutagen. Our expectation is that with other mutagens such as ethyl methanesulfonate results might even be more spectacular since such agents can sample the genome more impartially for mutations that can cause seizure sensitivity.

In conclusion, we feel that there are several unique advantages in using the Drosophila BS mutants to study seizure susceptibility. First, the BS mutants can be used in conjunction with a diverse selection of other Drosophila excitability and behavioral mutants to examine the types of molecular defects that can suppress or enhance seizure susceptibility. In addition, many excellent methodologies are available for Drosophila, such as P-element-mediated cloning as well as the use of the completed fly genome database, to aid in the molecular characterization of seizure sensitivity. Finally, we have developed useful electrophysiology protocols for quantifying seizures and paralysis, using the adult GF pathway (TANOUYE and WYMAN 1980 ; PAVLIDIS and TANOUYE 1995 ; KUEBLER and TANOUYE 2000 ). This allows us to evaluate and compare different mutations and their combinatorial effects on seizure susceptibility. Together, these conditions make the BS mutants an attractive model with which to study seizure susceptibility.

Conclusion:
The results in this article provide the first evidence that an aminopeptidase can influence the seizure susceptibility of Drosophila. We have shown here that the bang-sensitive mutant sda, when stimulated either mechanically or electrically can experience hyperactivity alternated with paralysis. Using various genetic and molecular analyses, we have revealed that sda mutants have a lesion in an aminopeptidase gene that leads directly to a drastic increase in seizure sensitivity. In future studies it would be interesting to analyze biochemically the function of this particular aminopeptidase, other molecules it may interact with, and the mechanisms by which it influences seizure sensitivity.

	FOOTNOTES

Sequence data from this article have been deposited with the EMBL/GenBank Data Libraries under accession no. AF480087.
¹ These authors contributed equally to this work.

	ACKNOWLEDGMENTS

We thank Diana Ho for assistance in maintenance of Drosophila stocks, Tim Tully for generously providing the sda^iso7.8 allele, and John Ngai's lab for use and assistance of their cryostat apparatus. We also thank members of the Tanouye lab and others for comments on the manuscript and contributions to this work, especially Charlie Oh, Jeremy Lee, Daria Hekmat-Scafe, Xiaoyun Ren, Ed Glasscock, Aaron Yeow, Robert Newman, and David Bentley. This work was supported by a National Institutes of Health (NIH) research grant and an Epilepsy Foundation grant to M.T., an NIH postdoctoral fellowship to H.Z., and a National Science Foundation predoctoral fellowship to J.T.

Manuscript received February 20, 2002; Accepted for publication August 15, 2002.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AIGNER, L., S. ARBER, J. P. KAPFHAMMER, T. LAUX, and C. SCHNEIDER et al., 1995  Overexpression of the neural growth-associated protein GAP-43 induces nerve sprouting in the adult nervous system of transgenic mice. Cell 83:269-278.[Medline]

AMOSCATO, A., J. ALEXANDER, and G. BABCOCK, 1989  Surface aminopeptidase activity of human lymphocytes. I. Biochemical and biologic properties of intact cells. J. Immunol. 142:1245-1252.[Abstract]

BARBAN, S., G. HOLLOPETER, J. ERICKSON, P. SCHWARTZKROIN, and R. PALMITER, 1997  Knock-out mice reveal a critical antiepileptic role for neuropeptide Y. J. Neurosci. 17:8927-8936.[Abstract/Free Full Text]

BENZER, S., 1971  From the gene to behavior. J. Am. Med. Assoc. 218:1015-1022.[Medline]

BURGESS, D., J. JONES, M. MEISLER, and J. NOEBELS, 1997  Mutation of the Ca²⁺ channel ß-subunit gene Cchb4 is associated with ataxia and seizures in the lethargic (lh) mouse. Cell 88:385-392.[Medline]

BUTLER, L., A. SILVA, A. ABELIOVICH, Y. WATANABE, and S. TONEGAWA et al., 1995  Limbic epilepsy in transgenic mice carrying a Ca²⁺/calmodulin-dependent kinase II alpha-subunit mutation. Proc. Natl. Acad. Sci. USA 92:6852-6855.[Abstract/Free Full Text]

CADEL, S., T. FOULON, A. VIRON, A. BALOGH, and S. MIDOL-MONNET et al., 1997  Aminopeptidase B from the rat testis is a bifunctional enzyme structurally related to leukotriene-A4 hydrolase. Proc. Natl. Acad. Sci. USA 94:2963-2968.[Abstract/Free Full Text]

CAIN, D., S. GRANT, D. SAUCIER, E. HARGREAVES, and E. KANDEL, 1995  Fyn tyrosine kinase is required for normal amygdala kindling. Epilepsy Res. 22:215-