Mated Drosophila melanogaster Females Require a Seminal Fluid Protein, Acp36DE, to Store Sperm Efficiently

Corresponding author: Mariana F. Wolfner, Department of Molecular Biology and Genetics, 423 Biotechnology Bldg., Cornell University, Ithaca, NY 14853-2703., mfw5{at}cornell.edu (E-mail)

Communicating editor: L. PARTRIDGE

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Mated females of many animal species store sperm. Sperm storage profoundly influences the number, timing, and paternity of the female's progeny. To investigate mechanisms for sperm storage in Drosophila melanogaster, we generated and analyzed mutations in Acp36DE. Acp36DE is a male seminal fluid protein whose localization in mated females suggested a role in sperm storage. We report that male-derived Acp36DE is essential for efficient sperm storage by females. Acp36DE¹ (null) mutant males produced and transferred normal amounts of sperm and seminal fluid proteins. However, mates of Acp36DE¹ males stored only 15% as many sperm and produced 10% as many adult progeny as control-mated females. Moreover, without Acp36DE, mated females failed to maintain an elevated egg-laying rate and decreased receptivity, behaviors whose persistence (but not initiation) normally depends on the presence of stored sperm. Previous studies suggested that a barrier in the oviduct confines sperm and Acp36DE to a limited area near the storage organs. We show that Acp36DE is not required for barrier formation, but both Acp36DE and the barrier are required for maximal sperm storage. Acp36DE associates tightly with sperm. Our results indicate that Acp36DE is essential for the initial storage of sperm, and that it may also influence the arrangement and retention of stored sperm.

DROSOPHILA melanogaster females, like females of many other animal species, store the sperm they receive during mating (reviewed in GILBERT et al. 1981 ; SANDER 1985 ; BIRKHEAD and MOLLER 1993 ; NEUBAUM and WOLFNER 1999 ). Drosophila females have three organs specialized for the retention of sperm, including a long, coiled tubule called the seminal receptacle and two sac-like organs termed spermathecae (MILLER 1950 ). The female uses her stored sperm to fertilize several hundred eggs over the 2 wk after a mating (LEFEVRE and JONSSON 1962 ). The ability to store sperm is an integral part of the Drosophila pattern of reproduction, and it may increase fecundity (COOK 1970 ; HIHARA 1981 ), minimize the biochemical and environmental hazards associated with copulation (THORNHILL and ALCOCK 1983 ; CHAPMAN et al. 1995 ), and provide a milieu for sperm competition (HARSHMAN and PROUT 1994 ; CLARK et al. 1995 , CLARK et al. 1999 ; CLARK and BEGUN 1998 ; HUGHES 1997 ; PRICE 1997 ). Moreover, Drosophila females link the presence of stored sperm to postmating behaviors such as egg laying and receptivity to remating (MANNING 1962 , MANNING 1967 ; HIHARA 1981 ; SCOTT 1987 ; KALB et al. 1993 ; TRAM and WOLFNER 1998 ), which serves to further maximize progeny production.

The mechanisms that govern sperm storage in Drosophila and other animals are likely to require a combination of anatomical features and the action of molecules contributed by males and females (DEVRIES 1964 ; FOWLER 1973 ; BERTRAM et al. 1996 ; ARTHUR et al. 1998 ; NEUBAUM and WOLFNER 1999 ). While previous studies have focused on the morphology of the male and female genitalia (MILLER 1950 ; BAIRATI 1968 ), as well as the sperm storage organs and the arrangement of sperm within them (FILOSI and PEROTTI 1975 ; HIHARA and HIHARA 1993 ), relatively little is known about the molecules furnished by males or females that are important for sperm storage.

Female D. melanogaster receive ~4000–6000 sperm in a single mating and store up to ~1100 of them (GILBERT 1981 ). They use up to 600–800 sperm to fertilize eggs over the subsequent 2 weeks (KAPLAN et al. 1962 ; GILBERT et al. 1981 ). That male seminal fluid components play a role in the transfer or storage of sperm in females was first suggested by the low fertility of repeatedly mated males (LEFEVRE and JONSSON 1962 ; HIHARA 1981 ). KALB et al. 1993 obtained additional evidence favoring this idea when they observed that the mates of male flies deficient in Acps, seminal fluid proteins derived from male accessory glands, contained fewer sperm than normal at 6 hr postmating. In the accompanying article, TRAM and WOLFNER 1999 show that this phenotype results from a requirement for Acps for proper storage of sperm, rather than from insufficient sperm production or transfer by Acp-deficient males.

We wished to identify specific molecules that mediate sperm storage. In particular, one Acp, Acp36DE, was a good candidate for involvement in sperm storage on the basis of its localization in the mated female genital tract (BERTRAM et al. 1996 ) and on data from CLARK et al. 1995 , which show a correlation of alleles at this locus and levels of sperm competition in chromosomes from natural populations. We show here that Acp36DE is a major player in the process of sperm storage.

Acp36DE is a large glycoprotein of 122 kD that is processed to a 68-kD form shortly after its transfer to females during mating (BERTRAM et al. 1996 ). In mated females, Acp36DE localizes to two specific areas: Acp36DE binds to the lower oviduct on the ventral side, just above the openings to the three sperm storage organs. Acp36DE also associates with the mass of sperm and other seminal fluids in the uterus (the "sperm mass"). After mating, the sperm mass does not progress into the upper oviduct, but remains confined to the uterus, posterior to the site of Acp36DE localization. These observations led us to propose a model in which Acp36DE marks the presence of a barrier in the mated female oviduct that blocks the movement of sperm into the upper oviduct (BERTRAM et al. 1996 ). The mating plug secretions, which occupy the lower portion of the uterus, are proposed to delimit the sperm at the posterior. In this manner, sperm are confined to a small region of the anterior uterus, which may facilitate their entry into the narrow openings of the storage organs. Whether Acp36DE serves only as a marker of the barrier in the oviduct or comprises part or all of the barrier was not possible to determine in the BERTRAM et al. 1996 study.

In mated eggless females, a proper barrier does not form in the oviduct, and sperm (as well as Acp36DE) become mislocalized into the anterior oviduct near the ovaries (BERTRAM et al. 1996 ). If the formation of an oviduct barrier (and the concomitant full localization of Acp36DE at the oviduct) is important for sperm storage, we predict that sperm storage will be compromised in eggless females mated to normal males.

Here, we report the isolation and characterization of a null mutation in Acp36DE. We show that this mutation severely compromises the ability of sperm of mutant males to be stored by their mates. To probe the mode of action of Acp36DE in sperm storage, we tested whether the presence of Acp36DE is necessary for sperm to be properly restricted from entry into the upper oviduct. In addition, we used biochemical and genetic means to perturb the association of Acp36DE with the sperm mass or with the posterior oviduct wall. We find that although Acp36DE is not necessary for the oviduct sperm barrier to form, the barrier is important in sperm storage, and Acp36DE's localization correlates with the efficiency of sperm storage. Furthermore, Acp36DE was found to tightly associate with sperm and to enter the storage organs.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Fly strains and isolation of Acp36DE mutants:
A strain of Oregon-R flies made isogenic for chromosome 2 was termed iso2. Three-day-old iso2 males were treated with 25 mM ethyl methanesulfonate (EMS; LEWIS and BACHER 1968 ). A series of genetic crosses as in HERNDON and WOLFNER 1995 generated */Df males, where * represents an EMS-mutagenized chromosome and Df represents Df(2L)H2O, a deficiency that deletes 36A8-9; 36E1-2 (SIMPSON 1983 ). Individual */Df males were tested by dot blot or Western blotting (VAN VACTOR et al. 1988 ; HERNDON and WOLFNER 1995 ) for loss or truncation of Acp36DE.

Acp36DE^P, kindly provided by A. Clark and C. Langley, contains a P-element insertion within the Acp36DE coding sequence (CLARK et al. 1994 ; BERTRAM et al. 1996 ; D. M. NEUBAUM and M. F. WOLFNER, unpublished results), which results in truncation of the protein. Acp36DE^P males appear to have normal fertility.

Homozygous bw sp tud¹ females (BOSWELL and MAHOWALD 1985 ) were mated to iso2-derived males to produce eggless daughters and spermless sons. Thus, the genetic background of eggless females is quite related to their control, iso2 females. Other balancers and markers are listed in LINDSLEY and ZIMM 1992 . All phenotypic analyses described in this article were carried out using Acp36DE*/Df flies. Controls were Oregon-R (for Western blots) or Acp36DE⁺/Df males, where the Acp36DE⁺ chromosome was derived from a strain closely matched in genetic background to the mutants. This strain had been maintained in parallel to the Acp36DE¹ and Acp36DE² mutant strains (for sperm counts and phenotypic assays).

Genetic analysis:
Acp36DE alleles were maintained by backcrossing Acp36DE*/CyO males to CyO/Df females every generation. All phenotypes described here cosegregated. EMS-induced mutations at genomic locations other than chromosome 2 were segregated away from the genetic background of Acp36DE* males as the number of backcrosses increased.

Nucleic acid and protein analysis:
Recent EST sequences deposited in the databases by the Berkeley Drosophila Genome Project, in conjunction with our Northern blotting, have extended the Acp36DE sequence at the 5' end of the gene (W. SWANSON, personal communication; GenBank entry no. U85759 has been revised accordingly). The revised Acp36DE ORF spans 2736 bp, the equivalent of 912 amino acids. Amino acids 1–23 comprise a signal sequence conforming to the specifications of VON HEIJNE 1983 , leading us to believe that this is the likely start of Acp36DE's ORF. The Acp36DE mutant lesions were identified by PCR-amplifying DNA from */Df flies using two primers that include XbaI linkers and derive from 454 to 472 and 2774 to 2792 of the Acp36DE sequence (WOLFNER et al. 1997 ). Amplified products were sequenced directly or cloned into pBluescript II SK +/- (Stratagene, La Jolla, CA), and the sequences of both strands were determined using automated cycle sequencing (Perkin-Elmer, Norwalk, CT/Applied Biosystems, Foster City, CA). Acp36DE⁺ cDNA sequence (WOLFNER et al. 1997 ) or PCR-generated Acp36DE genomic sequence was used for comparison to mutant alleles.

Total poly(A)⁺ RNA was isolated from adult male Drosophila using the Total RNA isolation kit (Promega, Madison, WI). Northern blots were prepared as described in MONSMA and WOLFNER 1988 , with ~10 µg RNA per lane, and were probed with random-primed Acp36DE or, as a control, ß-tubulin (BIALOJAN et al. 1984 ) DNA probes, followed by autoradiography.

Protein extracts from the genital tracts of 3-day-old */Df (2L)H2O males were prepared and immunoblotted as described in MONSMA and WOLFNER 1988 . Blots were probed with affinity-purified Acp36DE (BERTRAM et al. 1996 ) or Acp26Aa (MONSMA and WOLFNER 1988 ) primary antibodies diluted 1:1000 or 1:500, respectively, and with horseradish-peroxidase-conjugated secondary antibody (Sigma, St. Louis) diluted 1:2000. Proteins were visualized by enhanced chemiluminescence (ECL; Amersham, Arlington Heights, IL).

Sperm motility, transfer, and quantitation:
Mature sperm were dissected from the seminal vesicle and tested for motility by observing their sinuous tail undulations in physiological buffer (CENCI et al. 1994 ).

To evaluate sperm transfer by mutant males, the genital tracts of females mated to Acp36DE¹/Df or control (Acp36DE⁺/Df) males were dissected at 30 min after the start of mating (~20 matings per genotype). After removal of the ovaries and oviduct, the mated female's uterus was squashed under a coverslip, causing the sperm mass to spread flat. In double blind experiments, we assigned sperm spreads a score of 1–5 on the basis of the size and density of the spread viewed at x40 in a Zeiss (Thornwood, NY) Axioskop. Results were confirmed in duplicate experiments.

Sperm stored in female genital tracts were fixed and stained with 2% orcein in 60% acetic acid (GILBERT 1981 ; GILBERT et al. 1981 ) and counted at x100 magnification using phase-contrast microscopy. Data were analyzed using the Statview 4.1 statistics program (Fisher's t-tests and ANOVA). Total sperm stored was calculated as the sum of those stored in the seminal receptacle and two spermathecae. Data for total sperm counts (see Table 2) were taken only from the subset of tracts that retained all three organs after squashing, although additional data for individual organs were taken from any sample that had intact and suitably squashed organs.

Egg-laying, fertility, receptivity, and life-span assays of female mates of Acp36DE mutant males:
Control lines in all assays were overtly normal Acp36DE-positive lines retained from the mutagenesis experiment. Control and mutant lines were backcrossed every generation to Df/CyO stocks to isogenize genetic background and to avoid selection of modifiers. For all assays, Oregon-R females and hemizygous Df males (mutant or control) were collected a few hours after eclosion and aged 3–5 days. Flies were maintained at constant temperature (24° ± 0.5°), humidity, and day/night cycle (12:12 hr light:dark).

Egg-laying and receptivity responses were examined as described in KALB et al. 1993 and HERNDON and WOLFNER 1995 .

The effect of Acp36DE on life span was assayed in experiments like those of CHAPMAN et al. 1995 , with the exception that microcauterized flies were not included. Each Oregon-R female was placed with a Acp36DE¹/Df or control (Acp36DE⁺/Df) male for 2 days, but every third day was housed with a control male to control for nonseminal fluid costs of mating, such as egg production (PARTRIDGE et al. 1987 ). A total of 30 females were used per experimental group.

Using the Statview 4.1 statistics program, we performed ANOVA and Fisher's t-tests on data from all assays.

Sperm-binding assays:
In vivo assay: For all assays described in this paper, males and females were held in isolation for 3–5 days before mating. Female genital tracts were dissected 30 min after the start of mating, and the ovaries were removed. The mass of sperm present in the anterior uterus was then squeezed out of the torn oviduct by applying pressure to the uterus. Sperm masses were transferred to a microfuge tube, where 500 µl (>1000x volume) of buffer was added. Buffers contained 1x PBS ± 0.1% Tween 20 + 0.0, 0.1, 0.5, or 1 M NaCl, and were adjusted to pH 5, 7, or 10 with 1 N HCl or 1 N NaOH. In some experiments, samples were vortexed vigorously for 10 sec. All samples were incubated for 1 hr and then spun in a microfuge for 5 min to pellet the sperm. Supernatants were spun in a SpeedVac to reduce volume. Pellets and supernatants were resuspended in 2x SDS-PAGE loading buffer, run on 10% SDS-PAGE gels, and Western blotted (MONSMA and WOLFNER 1988 ). After probing with anti-Acp36DE, blots were stripped and reprobed with anti-Acp26Aa (PARK and WOLFNER 1995 ). To determine whether soluble Acp36DE was precipitated by any of the buffers, a parallel experiment was performed using accessory gland homogenate (spun free of particulate matter) in each of the buffers under similar conditions.

In vitro assay: Using 40 males per treatment, mature sperm were dissected from the seminal vesicle, while the accessory glands (AG) were homogenized in protease-inhibitory buffer (PARK and WOLFNER 1995 ) and centrifuged briefly to remove particulate matter. AG extracts (100 µl) were mixed with sperm (or not, as a control) for 1 hr at 4° with gentle shaking. Samples were then layered onto 30–70% sucrose gradients and spun for 20 min at 5000 x g in a swinging bucket rotor (SW50.1) at 4°. Fractions (250 µl) were collected beginning at the top of the gradient and examined under a microscope for the presence of sperm. Samples were reduced in volume, loading buffer was added, and samples were run on gels and immunoblotted.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Isolation of Acp36DE mutants:
Since the nature and severity of the Acp36DE mutant phenotype was not predictable, we searched for mutations in Acp36DE with a screen based on protein expression rather than function. To identify mutant lines with decreased or absent Acp36DE expression, affinity-purified Acp36DE polyclonal antibodies (BERTRAM et al. 1996 ) were used to screen dot or Western blots (VAN VACTOR et al. 1988 ; HERNDON and WOLFNER 1995 ) containing extracts of males heterozygous for a mutagenized chromosome 2 and a deficiency that uncovers Acp36DE (*/Df males).

From 8805 mutagenized lines tested, we recovered two in which full-length Acp36DE was undetectable on Western blots of extracts of */Df males (Fig 1A). In contrast to wild-type males, which had a prominent 122-kD Acp36DE band (lane 1), Acp36DE¹/Df males had no detectable Acp36DE (lane 2), even when lanes were loaded with extracts of 10-fold more flies than control lanes. Acp36DE¹/Df males did not make a detectable Acp36DE RNA transcript (Fig 1B). A second Acp36DE mutant, Acp36DE², gave rise to a 56-kD Acp36DE (Fig 1A, lane 3), whose signal on Western blots was less intense than that of the wild-type Acp36DE. Acp36DE² is likely truncated and less abundant or less well recognized by the polyclonal Acp36DE antibody. Acp36DE² tested normal in all functional assays and will not be discussed further in this article.

View larger version (40K): In this window In a new window Download PPT slide   Figure 1. Molecular characterization of Acp36DE mutants. (A) A Western blot probed with anti-Acp36DE. Lanes contain extracts of accessory glands from 3-day-old virgin Oregon-R (wild-type) males (lane 1, one pair of accessory glands), Acp36DE1/Df males (lane 2, 10 pairs), or Acp36DE2/Df males (lane 3, 10 pairs). (B) Northern blot of total RNA isolated from control or Acp36DE1/Df males probed for Acp36DE mRNA (top) or, as a control, ß-tubulin mRNA (bottom). (C) Diagram of proteins encoded by the Acp36DE+ and Acp36DE1 alleles. Acp36DE1 is predicted to be truncated by a stop codon at amino acid 274 (CAA TAA).

Sequence analysis of Acp36DE¹ showed that a nonsense mutation, created by a C-to-T transition at position 823 in the ORF sequence, caused Gln²⁷⁴ to be replaced with a stop codon (Fig 1C). The shortened open reading frame could encode a secreted protein of ~28 kD, but no novel Acp36DE cross-reactive band of this or any other size was detectable on immunoblots, presumably because Acp36DE¹ mRNA was destabilized by the mutation (SACHS 1993 ; HERNDON and WOLFNER 1995 ). On the basis of the lack of detectable Acp36DE mRNA or protein, we conclude that Acp36DE¹ is a null allele.

We believe the phenotypes described in the remainder of this article are the result of mutations in the Acp36DE gene for the following reasons. The phenotypes we describe are observed in Acp36DE¹/Df males, but not in Acp36DE¹/Acp36DE⁺ males, which have normal fertility. The fully recessive nature of the Acp36DE¹ mutation indicates that the gene responsible for the phenotypes we report below must lie in the region uncovered by Df(2L)H2O. The only known gene within this region with a mutation affecting male fertility is blanks (36A8; 36B6); the bln¹ allele causes a postmeiotic differentiation defect so that bln mutants do not make sperm (CASTRILLON et al. 1993 ). Since the mutant males described in this article do make motile sperm (see below), they are unlikely to have a mutation in bln. Male sterile mutants that make motile sperm are rare (10 in 400, FITCH et al. 1998 ). Thus, the likelihood is very low that we induced simultaneously, in the small genetic region uncovered by Df(2L)H2O, two independent mutations, one that abolished production of Acp36DE and another that resulted in semisterility despite production of normal amounts of motile sperm.

Acp36DE mutant males had normal viability and produced normal amounts of sperm and seminal fluid proteins other than Acp36DE:
Acp36DE mutant flies were fully viable and displayed no abnormal visible morphological phenotypes. The accessory glands of mutants were normal in morphology and size, and produced normal amounts of other Acps (including Acps 26Aa, 26Ab, 32CD, 62F, and 63F/64A, MONSMA and WOLFNER 1988 ; WOLFNER et al. 1997 ; data not shown), as well as the ejaculatory duct protein Est-6 (GILBERT 1981 ; data not shown).

Acp36DE mutant males produced abundant, morphologically normal, motile sperm. Sperm of mutant males are functional, since eggs fertilized by sperm of Acp36DE mutant males survived to adulthood, and the number of progeny produced by Acp36DE mutant males was proportional to the number of sperm stored by their mates (see below). The time (± SE) Acp36DE¹/Df or control males spent in copulation was not significantly different (t = 20.86 ± 0.533 and 23.29 ± 1.155 min, respectively; for Acp36DE¹/Df and Acp36DE⁺/Df, P = 0.1964). An opaque mass was observed in the uteri of females recently mated to Acp36DE¹/Df males (Fig 2A), but not in virgin females. When dissected, the opaque mass was found to contain hundreds of sperm tangled together. To evaluate the efficiency of sperm transfer, we dissected and squashed uteri of mated females and scored the sperm spreads using phase-contrast microscopy. In a double-blind experiment, the mass of sperm transferred by mutant males (n = 20) was indistinguishable in size, area, and sperm density from the mass of sperm observed in females mated to control males (n = 20). This suggests that Acp36DE¹ mutant males consistently transferred a number of sperm similar to control males. TRAM and WOLFNER 1999 observed variability in sperm transfer by males lacking Acps. Our observation that Acp36DE¹ males appear to transfer sperm normally suggests that other Acps underlie effects on the uniformity of sperm transfer.

View larger version (164K): In this window In a new window Download PPT slide   Figure 2. Sperm in the genital tract of mated females. (A) A large mass of sperm (black arrow) is observed in the uterus of a female immediately after mating to a Acp36DE1/Df male. Sperm are not mislocalized to the upper oviduct (arrowheads), but, as in wild-type matings, remain largely posterior to a boundary in the lower oviduct. Bar, 0.128 mm. (B) Inside the seminal receptacles of females mated to control males, the staining of sperm heads (e.g., white arrow) reveals numerous sperm arrayed in a roughly parallel fashion. (C) In females mated to Acp36DE1/Df males, the seminal receptacle (shown) and spermathecae are mostly empty. (D) The few sperm that are present in mates of Acp36DE1/Df males tend to clump in a nonparallel arrangement. Diffusely staining round nuclei are female somatic nuclei. (B–D) Bar, 0.009 mm.

Sperm stored in mates of Acp36DE¹/Df males were few in number and were arranged in a disorganized fashion:
On the basis of its localization in mated females (BERTRAM 1996), we hypothesized that lack of Acp36DE would affect the number of sperm stored in females. Accordingly, we counted sperm in the storage organs of mated females at 6 hr after mating, by which time sperm storage is complete (GILBERT 1981 ). In seminal receptacles, on average, 17% as many sperm were stored by mates of Acp36DE¹/Df males as by mates of controls (P < 0.0001, Table 1). The few sperm that were present in the seminal receptacles of mates of mutant males were in disorganized clumps, with the sperm heads in twisted disarray (Fig 2D), in contrast to the more parallel arrangement of sperm from control matings (Fig 2B). Similarly, at this time, spermathecae of Acp36DE¹/Df males had only 11% as many sperm on average as mates of controls. A summation of total sperm stored in the seminal receptacle and two spermathecae indicates that mates of Acp36DE¹/Df males stored overall ~15% as many sperm as mates of control males.

To determine whether mates of Acp36DE¹/Df males lost (or used) sperm from the storage organs at the same rate as mates of controls, we compared the numbers of sperm stored by females at 6, 24, and 48 hr after mating (Fig 3, Table 1). Although females mated to Acp36DE¹/Df males stored far fewer sperm than females mated to controls, both types of female lost a similar proportion of stored sperm from their spermathecae over the first 48 hr after mating and from their seminal receptacle over the first 24 hr after mating. On the second day after mating, however, females mated to mutant males rapidly lost sperm from their seminal receptacles, reducing the number stored to almost zero. In contrast, females mated to controls still retained ~50% of the initially stored sperm at 48 hr after mating. Thus, though the primary effect of Acp36DE appears to be on the number of sperm that initially get stored by females, absence of Acp36DE also has an effect on the retention of sperm in the seminal receptacle by day 2 after mating.

View larger version (21K): In this window In a new window Download PPT slide   Figure 3. Sperm depletion from the seminal receptacle (SR) or spermathecae (Sp) of mates of Acp36DE1/Df or control males. Numbers are from Table 1. Note that in the first 48 hr after mating, few sperm are lost from the spermathecae of mates of Acp36DE1/Df or control males, whereas in the seminal receptacle, mates of Acp36DE1/Df males show a proportionally more rapid sperm depletion than mates of controls.

Mates of Acp36DE mutant males produced 10% as many progeny as mates of controls:
To address the role of Acp36DE in fertility, we determined the number of progeny from single matings of Acp36DE¹/Df or control males mated to wild-type females (Fig 4A). While Acp36DE¹/Df males were not sterile, their mates produced over the mating period only 10% as many progeny as females mated to control males (P < 0.001).

View larger version (30K): In this window In a new window Download PPT slide   Figure 4. Behavioral responses in females mated to Acp36DE mutant males. Data from a typical experiment are shown in each panel. In all cases, experiments were replicated three times, with similar results each time. Oregon-R females were tested as virgins (open diamonds) or after mating to a Acp36DE1/Df male (solid squares), control male (open squares), or a spermless son of a bw sp tud1 mother (solid circles). (A) Fertility: The mean number (± SE) of adult progeny obtained from virgin or mated females is shown for each day after mating. (B) Egg laying: The mean number (± SE) of eggs laid by virgin or mated females is shown for each day after mating. The numbers of females tested per group were 48 mates of Acp36DE1/Df males, 23 mates of control males, and 17 virgins. (C) Receptivity: Oregon-R females were initially mated to a Acp36DE1/Df male, a control male, or a spermless male. The cumulative percentage of females that were willing to mate again with a wild-type control male during a 1-hr challenge period is shown for each day after the initial mating. The numbers of females tested per group were 17 mates of Acp36DE1/Df, 19 mates of control, and 11 mates of spermless. (D) Life span: Survival curves for females mated to Acp36DE mutant (solid squares) or control males (open squares) are shown. A total of 30 females were tested per group.

Moreover, progeny production by mates of Acp36DE¹/Df males showed a different pattern and kinetics from that of the controls (Fig 4A). Although female mates of Acp36DE¹/Df males laid a nearly normal number of eggs on the first day after mating (see below), only 33% on average of those eggs ultimately gave rise to adults, suggesting that the remaining eggs were unfertilized. The majority of progeny from Acp36DE¹/Df matings came from eggs laid on days 1 and 2 after mating. Very few progeny resulted from eggs laid three or more days after mating, suggesting that those eggs were nearly all unfertilized. In contrast, on average, >80% of eggs laid by control mates on each of the first through ninth day after mating gave rise to adults. These data are consistent with the compromised storage and retention of sperm in mates of Acp36DE¹/Df males.

Mates of Acp36DE mutant males fail to maintain at least two sperm-storage-dependent behaviors:
Persistence of increased egg-laying rate and decreased receptivity to courtship for more than 1 day after mating depends on stored sperm in the female (MANNING 1962 , MANNING 1967 ; GARCIA-BELLIDO 1964 ; KALB et al. 1993 ). Since lack of Acp36DE decreases sperm storage, we tested whether females mated to Acp36DE mutant males can maintain these behavioral responses.

Egg laying: On the first day after mating, when the egg-laying rate is largely independent of sperm (KALB et al. 1993 ), mates of Acp36DE¹/Df males showed a robust response, laying a number of eggs equivalent to controls (P = 0.1082, Fig 4B). The high rate of egg laying was not maintained in mates of Acp36DE¹/Df males. By the third day after mating, the number of eggs laid by Acp36DE¹ mates was intermediate between and significantly different from the number laid by mates of controls (P = 0.0014) and that of virgins (P = 0.0022). On the fourth and subsequent days after mating, the number of eggs laid by Acp36DE¹ mates was indistinguishable from that of virgins (P = 0.1622) and significantly less than that of the controls (P = 0.0003, day 4; P < 0.0001, days 5 and 6). By day 7 after mating, the number of eggs laid by controls had also declined, and the number of eggs laid by Acp36DE¹ mates was indistinguishable from that of controls (P = 0.3236).

Receptivity: On the first day after mating, females mated to all classes of males remained unreceptive to remating (Fig 4C). Mates of Acp36DE¹/Df males became receptive starting on the second day after mating and remated at frequencies similar to those of virgin females by the third day after mating. This pattern of remating is similar to the behavior of mates of spermless males (KALB et al. 1993 ; Fig 4C).

Thus, in both egg-laying and receptivity assays, mates of Acp36DE¹ mutants responded normally on the first day after mating, a time when these behaviors can be driven by Acps, without stored sperm (KALB et al. 1993 ). A response on day 1 is consistent with the receipt of Acps other than Acp36DE by mates of Acp36DE¹ mutants. Acp26Aa (HERNDON and WOLFNER 1995 ) and sex peptide (Acp70A, CHEN et al. 1988 ) stimulate egg laying, and sex peptide depresses receptivity on the first day after mating. Mates of Acp36DE¹/Df males did not maintain an elevated egg-laying rate or a depressed receptivity on the second and subsequent days after mating, when these behaviors depend on the presence of stored sperm. The fall-off in egg laying and in mating recalcitrance on day 2 and later are not likely to result directly from Acp36DE, since this protein was not detectable in extracts of whole female genital tracts after 6 hr postmating (BERTRAM et al. 1996 ).

Life span of mated females was not affected by lack of Acp36DE from their mates:
CHAPMAN et al. 1995 demonstrated that the life span of females was decreased in part by receipt of Acps during mating. In tests of the effects of the Acp36DE¹ mutation on the mates of mutant males, we found no significant differences in the life span of females mated to Acp36DE¹/Df relative to control males (P = 0.1566, Fig 4D). Thus, Acp36DE does not account to any significant degree for the toxic effect of seminal fluid.

The barrier in the oviduct is important for sperm storage, but Acp36DE is not required for formation of this barrier:
The preceding experiments established a function for Acp36DE in sperm storage in females. We next addressed how Acp36DE might execute this function. Upon noting the presence of Acp36DE in the oviduct, we had hypothesized that Acp36DE marked or was part of a barrier that prevented sperm entry into the upper oviduct and corralled sperm to a region near the sperm storage organs, thus facilitating their storage (BERTRAM et al. 1996 ). To investigate whether Acp36DE was required in females for barrier formation, we mated normal females to Acp36DE¹/Df or Acp36DE⁺/Df males. The localization of sperm in females mated to control males (n = 20) or to Acp36DE¹/Df males (n = 20) did not differ. Although a large mass of sperm was observed in the anterior uterus of females mated to null males, sperm did not travel beyond this boundary to the upper genital tract (Fig 2A, arrowheads). Genital tracts of females mated to control males appeared identical to that shown in Fig 2A. Therefore, the presence of Acp36DE in the oviduct was not necessary to keep sperm from mislocalizing to the upper genital tract.

We then reexamined our initial hypothesis and asked whether the establishment of a barrier and the presence of Acp36DE near the barrier was important for Acp36DE's function in sperm storage. To test this, we evaluated sperm storage in eggless females mated to normal (Oregon-R) males. In eggless females, a greatly reduced amount of Acp36DE localizes to the oviduct, and a sperm barrier does not form (BERTRAM et al. 1996 ). We counted the number of sperm present in the storage organs of mated eggless females at 6 hr after mating, the time of maximal sperm storage (GILBERT 1981 ). Significantly fewer (P = 0.0086) sperm were stored in the seminal receptacles of eggless females relative to Oregon-R females (Table 2). Likewise, significantly fewer (P < 0.0001) sperm were found in the spermathecae at this time (Table 2). On average, eggless females stored 19% fewer sperm than control females mated to normal males. However, note that eggless females with their reduced oviductal localization of Acp36DE still stored far more sperm than females mated to Acp36DE¹/Df males, who received no Acp36DE at all. The two experiments described above demonstrate that although Acp36DE is not necessary to form the barrier, the presence of the barrier and/or the full and normal localization of Acp36DE in the female is necessary for maximal sperm storage.

Acp36DE binds to sperm:
The association of Acp36DE with the sperm mass in the mated female genital tract (BERTRAM et al. 1996 ) could be important for its function in sperm storage. Alternatively, Acp36DE may be present in the sperm mass without interacting specifically with sperm. To assess the strength of the association of Acp36DE and sperm, we tested whether Acp36DE could be dissociated from the sperm mass in the presence of high salt, detergent, or extreme pH. Sperm masses were removed from mated female genital tracts and placed in one of seven buffers (see MATERIALS AND METHODS; Fig 5 legend). After incubation, sperm were pelleted by centrifugation. The sperm fraction (pellet) and soluble fraction (supernatant) were assayed by Western blotting. Both the full-length (122 kD) and processed (68 kD) Acp36DE products remained associated with sperm. In all treatments, >80% of both forms of Acp36DE remained associated with the sperm pellet (Fig 5, lanes 1–14). Vortexing the sperm samples at the start of incubation did not release more Acp36DE to the supernatant (data not shown). Except for buffer 1, the presence of Acp36DE in the pellet was not caused by precipitation (buffers 2–7 in Fig 5), since Acp36DE present in male accessory gland homogenates did not pellet when incubated alone (Fig 5, lanes 15 and 16), or when mixed with the torn-open testes and seminal vesicles of spermless males (not shown). Moreover, a different seminal fluid protein, Acp26Aa (MONSMA and WOLFNER 1988 ; PARK and WOLFNER 1995 ), readily dissociated from sperm upon the addition of NaCl (Fig 5, lanes 5–14). This result suggests that the association of Acp36DE with sperm is caused by biochemical features of Acp36DE rather than by any nonspecific "stickiness" of sperm. We also tested the association of a functional though truncated form of Acp36DE (Acp36DE^P, CLARK et al. 1994 ) with the sperm mass. The 56-kD Acp36DE^P interacted tightly with the sperm mass in a manner similar to full-length Acp36DE (data not shown). Thus, both Acp36DE and Acp36DE^P interact with the sperm mass in vivo, and this interaction is refractory to biochemical dissociation.

View larger version (44K): In this window In a new window Download PPT slide   Figure 5. The association of Acp36DE or Acp26Aa with the sperm mass in vivo. Western blot of sperm mass preparations probed with anti-Acp36DE or anti-Acp26Aa, showing proteins present in the pellet (P), which contains sperm, or the supernatant (S), which contains proteins that have dissociated from the sperm mass (lanes 1–14). Since the sperm masses were removed from mated females at 30 min after the start of mating, they show the processed products of Acp36DE (68 kD; and the female-specific, cross-reactive 64-kD band; BERTRAM et al. 1996 ) or Acp26Aa (30, 28, and 23 kD; PARK and WOLFNER 1995 ) normally present at that time. The components of seven PBS-based buffers are shown. As a control, accessory gland extract containing Acp36DE was incubated in each buffer without sperm (lanes 15 and 16 for buffer 2, data not shown for the other buffers). No precipitation of Acp36DE was observed for buffers 2–7, though slight precipitation occurred in buffer 1.

As a tool for further examination of Acp36DE-sperm interactions, we developed a sperm-binding assay to test whether Acp36DE could associate with sperm in vitro. Mature sperm dissected from the seminal vesicle were incubated for 1 hr with accessory gland homogenates containing Acp36DE. Samples were then spun through sucrose gradients until equilibrium was reached for the sample components. Fractions taken from the gradients were assayed for colocalization of Acp36DE and sperm. As shown in Fig 6, Acp36DE and sperm were both present in fractions ~70% into the gradient (around fraction 14, Fig 6A). In control assays without sperm, Acp36DE remained in the top 10–15% of the gradient (Fig 6D). In parallel experiments, Acp36DE^P was found in the sperm-containing fractions in a manner similar to the full-length protein (Fig 6B; data not shown). Acp36DE made as a GST fusion protein in Escherichia coli also migrated with sperm in the gradient (Fig 6C), suggesting that posttranslational modifications such as glycosylation are not required for binding of sperm by Acp36DE, and that the presence of other Acps is dispensable for Acp36DE binding to sperm. A control protein, Acp26Aa (MONSMA and WOLFNER 1988 ; PARK and WOLFNER 1995 ), did not bind to sperm in parallel assays (data not shown). Thus, Acp36DE can interact with sperm in the absence of other Acps.

View larger version (55K): In this window In a new window Download PPT slide   Figure 6. The cosedimentation of endogenous Acp36DE and E. coli-made Acp36DE with sperm in vitro. Western blots of sucrose gradient fractions probed with anti-Acp36DE and the corresponding fractions that contain sperm. After spinning to equilibrium, the gradient was divided into 20 equally sized fractions ranging from 30% sucrose (fraction 1) to 70% sucrose (fraction 20). Sperm was mixed with accessory gland extracts of (A) wild-type males, (B) Acp36DEP males, or (C) Acp36DE-GST fusion protein made in E. coli. As a control, (D) accessory gland extracts were tested without sperm. The increasing amount of sucrose in the deeper fractions of the gradient slightly retards mobility of the proteins in the gels. +, sperm seen in the fraction upon microscopic examination; -, no sperm detected in the fraction.

Acp36DE enters the sperm storage organs in a normal mating:
Although large amounts of Acp36DE were not detected inside the storage organs of females in previous immunohistochemical experiments (BERTRAM et al. 1996 ), the impermeability of these chitin-lined organs to the staining reagents, and possibly the brown color of the spermathecae, prevented detection of small amounts of Acp36DE in these tissues. The tight association of Acp36DE with sperm in the uterus suggests that the protein must either accompany the sperm into the organs or rapidly dissociate from the sperm as they are stored in the first hours after mating. To distinguish between these explanations, we assayed Acp36DE entry into the storage organs by a method independent of immunohistochemistry. Using fine tweezers, we cleanly separated spermathecae and seminal receptacles from the rest of the genital tract of mated females. Spermathecal and seminal receptacle homogenates from wild-type matings (Ore-R x Ore-R) were tested for Acp36DE on Western blots. Both organs contained full-length (122 kD) and processed (68 kD) Acp36DE (Fig 7A, lanes 1 and 2) at 2 hr after the start of mating, though slightly more Acp36DE was observed in the seminal receptacle than in the spermathecae. The signal obtained from ~120 spermathecae and ~60 seminal receptacles was ~10–50 times fainter than the signal of a single female genital tract extract (uterus plus sperm storage organs). Thus, a small fraction of the total Acp36DE transferred to a female does enter each of the sperm storage organs.

View larger version (38K): In this window In a new window Download PPT slide   Figure 7. Entry of Acp36DE into the storage organs and processing of Acp36DE in the absence of sperm or eggs. (A) Western blot of female sperm storage organ extracts probed with anti-Acp36DE. Each lane contains the spermathecae (Sp) or seminal receptacles (SR) of 62 mated females of Oregon-R (OreR, lanes 2 and 3) or eggless females (lanes 4 and 5), all dissected at 2 hr after the start of mating. Lane 1 contains an extract of one female genital tract (GT) with the ovaries removed and dissected at 30 min after the start of mating. 122 kD, Acp36DE (full length); 68 kD, Acp36DE (processed); 64- and 75-kD, female-specific, cross-reactive proteins (NEUBAUM 1999). (Acp36DE is expressed only in males.) (B) Western blot of 10 mated female genital tracts from OreR x OreR (lane 1) or spermless x eggless (lane 2) matings probed with anti-Acp36DE.

Since eggless females fail to localize Acp36DE properly in the lower oviduct, we tested whether Acp36DE could enter the storage organs of eggless females. Eggless females mated to Ore-R males stored an equivalent amount of full-length Acp36DE, but its 68-kD processing product could not be detected in either the spermathecae or the seminal receptacles (Fig 7A, lanes 4 and 5). The lack of detectable 68-kD Acp36DE in the sperm storage organs of mated eggless females does not reflect a lack of processing in these females. Processing of Acp36DE in the female genital tracts, which normally requires 20 min after mating (BERTRAM et al. 1996 ), is independent of the presence of eggs or sperm (Fig 7B). Thus, we believe the lack of processed Acp36DE in the sperm storage organs of mated eggless females may indicate that storage can occur properly for only a brief time immediately after mating, when full-length but not processed Acp36DE is present in the genital tract.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Storage of sperm by mated females is an important aspect of the reproductive strategy of many animals; yet its molecular basis is poorly understood. In Drosophila, molecules required for sperm storage are made by the male's accessory gland (TRAM and WOLFNER 1999 ). Here, we have identified the first molecule essential for sperm storage, and we have shown that this protein, the Drosophila accessory gland protein Acp36DE, is a major contributor to the process of sperm storage in females.

Acp36DE¹ null mutant males made abundant, motile, morphologically normal sperm and transferred them in large numbers to females during mating (Fig 2A). However, sperm failed to accumulate properly in the storage organs of females mated to Acp36DE¹ mutant males (Fig 2C and Fig D). A small number of sperm, roughly 15% relative to wild-type controls, did enter the storage organs and were stored. Some of these sperm successfully fertilized eggs (confirming that the sperm are functional), while others were lost from the storage organs at a greater rate than wild-type controls (Fig 3). As a result of storing few sperm, females mated to males lacking Acp36DE produced few progeny, only 10% of the number produced by females mated to normal males (Fig 4A). Moreover, the egg-laying and receptivity behaviors of females mated to Acp36DE-deficient mates reverted to a premated state within 2 or 3 days after mating instead of after the normal 10–14 days (Fig 4B and Fig C). Thus, the lack of Acp36DE indirectly affected egg-laying and receptivity behaviors, and served to minimize the number of progeny produced. While mutations in Acp36DE did not result in complete sterility, the fecundity of mutant males was severely compromised. The direct and indirect phenotypes of Acp36DE mutants emphasize the importance that female sperm storage has for the reproductive strategy of D. melanogaster and the role that male-derived proteins play in this process.

Acp36DE appears to be one of the primary male proteins responsible for eliciting female sperm storage. Acp36DE¹/Df mutant males elicited approximately the same level of sperm storage as males with genetically ablated accessory glands, who produce almost no Acps at all (TRAM and WOLFNER 1999 ). In contrast, the stimulation of egg laying, another postmating effect of males upon females, results from the cumulative but apparently independent effects of at least two gene products, Acp26Aa (HERNDON and WOLFNER 1995 ) and sex peptide (CHEN et al. 1988 ; MOSHITZKY et al. 1996 ).

A protein that is needed for sperm storage might act at any of several levels: helping sperm enter the storage organs, helping sperm arrange themselves appropriately inside the organs, maintaining sperm viability over time, or controlling the release of sperm over time. Our data favor the primary role of Acp36DE as being in the entry of sperm into storage, though we describe observations that suggest a secondary effect of the protein in sperm arrangement in and release from storage. A possible role for Acp36DE in promoting the entry of sperm into storage was suggested initially by the localization of Acp36DE protein in the female (BERTRAM et al. 1996 ). Since Acp36DE localization in the oviduct of wild-type mated females was coincident with the upper limit at which sperm were observed in the genital tract, we suggested that Acp36DE marks or forms a barrier in the oviduct that confines sperm within an area from which they can be easily channeled into the storage organs (BERTRAM et al. 1996 ). Here, we present four observations important for understanding the mode of action of Acp36DE in sperm storage. (1) The presence of Acp36DE was not necessary for sperm to be restrained from entering the upper oviduct of mated females (i.e., for "barrier formation"). (2) In eggless females, inefficient barrier formation and incomplete binding of Acp36DE at the lower oviduct correlated with fewer sperm stored. (3) Acp36DE tightly associated with sperm in vivo, and sperm and Acp36DE cosedimented in vitro. (4) Acp36DE was found in the sperm storage organs of mated females in both its full-length and processed forms. These points will be discussed in turn.

While the nature of the mechanism (or "barrier") that holds sperm in the posterior genital tract is unknown, it appears to be intact in females mated to Acp36DE¹ null mutant males. Females mated to males lacking Acp36DE retained a large mass of sperm in the uterus immediately after mating, and few, if any, sperm were observed in the upper oviduct near the ovaries. Despite proper barrier formation, 85% fewer sperm were stored than when Acp36DE was present. Thus, the presence of the barrier was not sufficient alone for normal efficient sperm storage.

The reverse experiment, in which normal amounts of Acp36DE are supplied but formation of the barrier is faulty, was performed by mating normal males to eggless females. In these matings, Acp36DE localization in the lower oviduct was weakened but not eliminated (BERTRAM et al. 1996 ). Eggless females stored significantly fewer sperm (averaging 19% less overall) relative to wild-type matings. These data show that without an oviduct barrier, a full aliquot of Acp36DE is not sufficient for full sperm storage. Therefore, the barrier must play some role in sperm storage, even if it is indirect. The barrier may be important simply because it traps Acp36DE at a critical location within the genital tract, from which Acp36DE exerts its action upon sperm storage. Alternatively, the activity of the barrier in sperm storage may be independent of Acp36DE. However, both the barrier and Acp36DE are required for optimal sperm storage in D. melanogaster.

What sort of action might Acp36DE exert from its location at the lower oviduct near the sperm storage organs? From the lower oviduct, Acp36DE could act on muscles or on the female nervous system to stimulate the storage organs to open or to begin contractions that pump sperm into storage. This type of mechanism is used in Rhodnius prolixus, where accessory gland secretions are reported to stimulate the genital ducts of females to contract (DAVEY 1960 , DAVEY 1965 ). Drosophila females might require ductile contractions to move sperm into storage, since the openings to the storage organs are extremely narrow and may impose limitations on the undulatory movements of sperm (see LINLEY and SIMMONS 1981 ). Interestingly, ARTHUR et al. 1998 report that D. melanogaster flies with female bodies but masculinized central nervous systems store dramatically fewer sperm than control females, indicating that the female nervous system plays an important role in Drosophila sperm storage. Alternatively, Acp36DE might complement the action of a physical barrier by providing a chemical signal to sperm to deflect them away from the upper regions of the genital tract and into the storage organs. Acp36DE and the sperm barrier may work together to ensure the efficient storage of sperm before the advent of egg laying results in the unstored sperm being pushed out of the uterus.

In addition to associating with the oviduct wall, Acp36DE also associates in vivo and in vitro with the sperm mass. The Acp36DE-sperm interaction was strong enough to withstand vortexing and incubation in high salt, detergent, and extreme pH (Fig 5). Such strong interactions are reminiscent of carbohydrate interactions of lectins and the proteins that bind to them (KOEHLER 1978 ; NICHOLSON and YANAGIMACHI 1979 ). As a glycoprotein, Acp36DE may interact with moieties found on the surface of sperm. The in vitro association of recombinant Acp36DE produced by E. coli suggests that the proteinaceous component of Acp36DE is sufficient to bind to sperm (Fig 6), but this result does not preclude the possible interaction of Drosophila Acp36DE's carbohydrate with sperm as well.

Since Acp36DE associated tightly with sperm in the uterus, we hypothesized that Acp36DE might enter the storage organs bound to sperm. Sperm enter the seminal receptacle earlier than the spermathecae and are stored there in greater numbers (GILBERT 1981 ; TRAM and WOLFNER 1999 ). Consistent with our hypothesis, Acp36DE was detected inside the storage organs by 2 hr after mating (Fig 7), with slightly more Acp36DE observed in the seminal receptacle than in the spermathecae at this time. Thus, Acp36DE was present at a time and place from which it might influence the sperm once they are inside the organs, possibly helping sperm to assume or retain an orderly parallel arrangement, or preventing their premature loss from the organs. In this context, it is interesting that sperm in the storage organs of females mated to males lacking Acp36DE failed to assume their normal parallel configuration (Fig 2D). In addition to Acp36DE, another D. melanogaster seminal fluid molecule influences sperm storage: the ejaculatory product esterase-6, a carboxylesterase (GILBERT 1981 ; GILBERT et al. 1981 ). In contrast to Acp36DE, null mutations in esterase-6 decrease rather than increase the rate of sperm loss (GILBERT 1981 ), suggesting that this protein operates differently from Acp36DE.

The processed (68 kD) form of Acp36DE was detected in the storage organs of wild-type but not eggless females, although full-length Acp36DE entered the storage organs of both types of females in approximately equivalent amounts (Fig 7). These observations may be explained if sperm stop entering storage earlier in eggless females, perhaps because of the impaired localization of Acp36DE at the oviduct. For 15–20 min after the start of mating, sperm and full-length Acp36DE (the only form available during that interval since processed products are not detected immediately after mating, BERTRAM et al. 1996 ) begin entering the storage organs, with Acp36DE possibly bound to the sperm. As the 68-kD form of Acp36DE begins to appear, it also enters the storage organs. In eggless females, the mislocalization of sperm up the oviduct may divert sperm (and Acp36DE) from entering the storage organs.

Acp36DE is one of several Acp genes for which polymorphisms in natural populations were reported to correlate with levels of sperm competition, specifically with the ability of stored sperm to resist displacement by sperm from a subsequent mate (CLARK et al. 1994 ). Given the role we demonstrate here for Acp36DE in the storage and retention of sperm, we propose that the effect of the wild-caught Acp36DE polymorphic alleles on sperm competition may reflect the action of the protein in storing and retaining sperm rather than an active participation in resistance to competition.

In summary, Acp36DE plays a critical role in the entry of sperm into storage and influences the maintenance or release of stored sperm. To date, most mechanisms proposed to facilitate sperm storage derive from morphological and behavioral criteria (for review see BIRKHEAD and MOLLER 1993 ; EBERHARD 1996 ; NEUBAUM and WOLFNER 1999 ). Our identification of Acp36DE as essential for sperm storage now permits a molecular genetic dissection of this essential reproductive process.

	FOOTNOTES

¹ Present address: Cardiovascular Research Center, Massachusetts General Hospital/Harvard Medical School, 149 13th St., Charlestown, MA 02129.

	ACKNOWLEDGMENTS

The authors are grateful to Drs. Ken Kemphues, Volker Vogt, and Uyen Tram for comments on the manuscript. We thank Dr. Willie Swanson for Acp36DE sequence analysis, Dr. Rollin Richmond for esterase-6 antibodies, Dr. Laura Herndon and Oliver Lung for advice and protocols, and Eduardo Gonzalez for dedicated technical assistance. This work was supported by sequential National Science Foundation grants to M.F.W. (IBN94-06171 and IBN97-23356); during part of this work, D.M.N. was supported by National Institutes of Health Training Grant T32-GM07617.

Manuscript received January 4, 1999; Accepted for publication June 28, 1999.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ARTHUR, B. I., E. HAUSCHTECK-JUNGEN, R. NÖTHIGER, and P. I. WARD, 1998  A female nervous system is necessary for normal sperm storage in Drosophila melanogaster: a masculinized nervous system is as good as none. Proc. R. Soc. Lond. Ser B 265:1749-1753.

BAIRATI, A., 1968  Structure and ultrastructure of the male reproductive system of Drosophila melanogaster Meig. 2. The genital duct and accessory glands. Monit. Zool. Ital. 2:105-182.

BERTRAM, M. J., D. M. NEUBAUM, and M. F. WOLFNER, 1996  Localization of the Drosophila accessory gland protein Acp36DE in the mated female suggests a role in sperm storage. Insect Biochem. Mol. Biol. 26:971-980[Medline].

BIALOJAN, S., D. FALKENBERG, and R. RENKAWITZ-POHL, 1984  Characterization and developmental expression of ß-tubulin genes in Drosophila melanogaster.. EMBO J. 3:2543-2548[Medline].

BIRKHEAD, T. R. and A. P. MØLLER, 1993  Sexual selection and the temporal separation of reproductive events: sperm storage data from reptiles, birds and mammals. Biol. J. Linnean Soc. 50:295-311.

BOSWELL, R. E. and A. P. MAHOWALD, 1985  tudor, a gene required for assembly of the germ plasm in Drosophila melanogaster.. Cell 43:97-104[Medline].

CASTRILLON, D., P. GONCZY, S. ALEXANDER, R. RAWSON, and C. EBERHART et al., 1993  Toward a molecular genetic analysis of spermatogenesis in Drosophila melanogaster: characterization of male-sterile mutants generated by single P element mutagenesis. Genetics 135:489-505[Abstract].

CENCI, G., S. BONACCORSI, C. PISANO, F. VERNI, and M. GATTI, 1994  Chromatin and microtubule organization during premeiotic, meiotic and postmeiotic stages of Drosophila melanogaster spermatogenesis. J. Cell Sci. 107:3521-3524[Abstract].

CHAPMAN, T., L. F. LIDDLE, J. M. KALB, M. F. WOLFNER, and L. PARTRIDGE, 1995  Cost of mating in Drosophila melanogaster females is mediated by male accessory gland products. Nature 373:241-244[Medline].

CHEN, P. S., E. STUMM-ZOLLINGER, T. AIGAKI, J. BALMER, and M. BIENZ et al., 1988  A male accessory gland peptide that regulates reproductive behavior of female D. melanogaster.. Cell 54:291-298[Medline].

CLARK, A. G. and D. G. BEGUN, 1998  Female genotypes affect sperm displacement in Drosophila. Genetics 149:1487-1493[Abstract/Free Full Text].

CLARK, A. G., S. SILVERIA, W. MEYERS, and C. H. LANGLEY, 1994  Nature screen: an efficient method for screening natural populations of Drosophila for targeted P-element insertions. Proc. Natl. Acad. Sci. USA 91:719-722[Abstract/Free Full Text].

CLARK, A. G., M. AGUADÉ, T. PROUT, L. G. HARSHMAN, and C. H. LANGLEY, 1995  Variation in sperm displacement and its association with accessory gland protein loci in Drosophila melanogaster.. Genetics 139:189-201[Abstract].

CLARK, A. G., D. G. BEGUN, and T. PROUT, 1999  Female x male interactions in Drosophila sperm competition. Science 283:217-220[Abstract/Free Full Text].

COOK, R., 1970  Control of fecundity in D. melanogaster.. Dros. Info. Serv. 45:128.

DAVEY, K. G., 1960  A pharmacologically active agent in the reproductive system of insects. Can. J. Zool. 38:39-45.

DAVEY, K. G., 1965 Reproduction in the Insects. Oliver & Boyd, Edinburgh.

DEVRIES, J. K., 1964  Insemination and sperm storage in Drosophila melanogaster.. Evolution 18:271-282.

EBERHARD, W. G., 1996 Female Control: Sexual Selection by Cryptic Female Choice. Princeton University Press, Princeton, NJ.

FILOSI, M. and M. PEROTTI, 1975  Fine structure of the spermatheca of Drosophila melanogaster Meig. J. Submicrosc. Cytol. 7:259-270.

FITCH, K., G. YASUDA, K. OWENS, and B. WAKIMOTO, 1998  Paternal effects in Drosophila: implications for mechanisms of early development. Curr. Top. Dev. Biol. 38:1-34[Medline].

FOWLER, G. L., 1973  Some aspects of the reproductive biology of Drosophila: sperm transfer, sperm storage and sperm utilization. Adv. Genet. 17:293-360.

GARCIA-BELLIDO, A., 1964  Das Sekret der Paragonien als Stimulus der Fekundität bei Weibchen von Drosophila melanogaster.. Z. Naturforsch. 19:491-495.

GILBERT, D. G., 1981  Ejaculate esterase and initial sperm use by female Drosophila melanogaster.. J. Insect. Physiol. 27:641-650.

GILBERT, D. G., R. C. RICHMOND, and K. B. SHEEHAN, 1981  Studies of esterase 6 in Drosophila melanogaster. V. Progeny production and sperm use in females inseminated by males having active or null alleles. Evolution 35:21-37