Successful fertilization requires a precise series of cell-cell interactions between gametes [reviewed by [1–6]]. The molecules that mediate these cell-cell interactions need to be present in sperm and oocytes at the right time and place to carry out their functions. Changes in the cellular distribution of gamete proteins could also regulate activity and access to other interacting molecules. Sperm in particular are highly polarized cells with functionally and morphologically distinct surface domains [7, 8]. These domains represent the compartmentalization of functions such as motility, energy production and sperm-egg cell surface interacting regions [9]. Many sperm molecules display a restricted distribution that corresponds to various sperm regions and this distribution is likely to be critical for their function [10, 11].

The nematode Caenorhabditis elegans is an excellent model system for the study of the molecular mechanisms of fertilization [12, 13]. Nematode sperm, like sperm from most species, are highly polarized cells. Unlike flagellated sperm, nematode sperm are amoeboid and have a single pseudopod protruding from the cell body. The pseudopod provides cellular motility, lacks organelles and has a highly dynamic membrane surface [14]. Despite this "engine in front" mode of motility, C. elegans sperm must carry out the same basic functions that are common to all sperm.

A genetic approach and the hermaphrodite/male reproductive biology of C. elegans facilitated the identification of a number of genes required for sperm-oocyte interactions at fertilization. One of these genes encodes SPE-9, a sperm transmembrane protein that contains an array of ten EGF motifs (Fig. 1) [15]. EGF motifs function as extracellular protein-protein interaction domains and are involved in diverse functions such as adhesive and ligand-receptor interactions [16]. In order to determine the distribution of SPE-9 in sperm, we obtained specific antisera directed against two different regions of the protein.

Specific antisera were obtained for two different regions of SPE-9 denoted EX (extracellular) and C (cytoplasmic) (Fig. 1). The EX sera was directed against a 13 amino acid peptide corresponding to a region between EGF motifs 4 and 5 in the predicted extracellular portion of SPE-9. The C sera was directed against a 22 amino acid peptide corresponding to the predicted cytoplasmic tail of SPE-9. We observed identical staining patterns with both of these antisera.

SPE-9 is segregated to spermatids where, depending on the plane of focus, we see a ring of staining (Fig. 2A,2B,2C,2D). In addition to the identical staining patterns observed with the EX and C antisera, we stained sperm isolated from worms carrying a null allele of spe-9 to further validate specificity. The eb19 allele of spe-9 encodes a premature stop codon before the amino acid sequences used to generate our antibodies [15]. spe-9(eb19) animals are completely sterile and produce spermatids that show no immunoreactivity to both the EX and C sera (Fig. 2E,2F,2G,2H). To make sure that SPE-9 localization patterns are not temperature dependent, we immunostained sperm isolated from worms raised at 16°C, 20°C and 25°C. The staining pattern of SPE-9 was not altered by culture temperature (data not shown). The SPE-9 staining pattern is very different from staining patterns seen with sera that stain internal sperm structures [17, 18]. We also labeled spermatids with the monoclonal antibody 1CB4 [19] (Fig. 2I,2J). The epitope recognized by 1CB4 has been shown to reside on membranous organelles (MOs). MOs are located in the cytoplasm close to the plasma membrane. This MO vesicle-staining pattern is distinct from SPE-9 staining (compare 2B and 2D with 2J).

The staining pattern seen for SPE-9 is consistent with the hypothesis that it is a sperm cell surface molecule. All known EGF motif containing proteins function extracellularly [16, 20, 21]. SPE-9 has a well defined signal sequence and it is most closely related to cell surface Notch/LIN-12/GLP-1 ligands [15]. Finally, if we do not permeabilize cells with detergent we can observe staining with the EX sera but not the C sera or 1CB4 (Fig. 3).

In C. elegans, spermiogenesis refers to the conversion of round sessile spermatids to polar motile spermatozoa [reviewed in [22]]. Spermiogenesis occurs in vivo as sperm enter particular regions of the hermaphrodite reproductive tract. Preparations of spermatids can be activated in vitro with various reagents (see below and materials and methods) [23]. Although the molecular nature of this activation pathway is poorly understood, the cellular events of spermiogenesis are fairly well described. These events include the initial polymerization of major sperm protein (MSP) that drives the formation of spike-like protrusions. MSP is the sperm cytoskeletal protein required for pseudopod motility [24]. Shortly after their formation, the spikes fuse to form a mature pseudopod capable of propelling the cell. Additionally, MOs fuse with the plasma membrane. The function of the MOs is not known. However, mutations that block MO fusion produce sperm with short pseudopods, motility defects and result in sterility [25].

Spermiogenesis induced in vivo or in vitro with reagents such as pronase, TEA, or monensin can occur very rapidly (~5 minutes) [23, 26–28]. With in vitro activators, it was very difficult to fix cells for immunolocalization and catch transient spikes. Trifluoperazine (TFP) initiates spermiogenesis but induces spikes that do not coalesce into a pseudopod [23]. We found that SPE-9 is localized to the spikes on these arrested cells (Fig. 4A,4B). One concern is that the large and often webbed spikes induced by TFP represent a late spike intermediate. Another way to induce persistent spikes is to treat spermiogenesis mutants such as spe-12 with pronase [23, 29]. We found that SPE-9 is localized to spikes induced on spe-12 spermatids treated with pronase (Fig 4C,4D). We were initially concerned that pronase could remove immunoreactive epitopes from the surface of sperm. Under these conditions, EX staining did appear to be slightly reduced while C staining was not reduced consistent with the predicted topology of SPE-9. We conclude that either SPE-9 is a poor substrate for pronase degradation or that the concentrations used in our experiments are not sufficient to remove the bulk of immunoreactive epitopes.

For mature sperm, we found SPE-9 concentrated on the pseudopod (Fig. 4E,4F,4G,4H). We observed the same localization patterns for SPE-9 regardless of the sperm activator used or culture temperature. These results make intuitive sense since SPE-9 is previously localized to spikes that coalesce to form the pseudopod.

Figure 5 shows a schematic summary of SPE-9 distribution. Localization and dynamic redistribution of cell surface components is a fundamental feature of polarized cells that is absolutely required for their proper function in an organism. There are several major mechanisms that could drive the localization of membrane proteins. One mechanism involves the specific distribution of newly synthesized proteins to discrete cell surface domains. The other mechanism involves the redistribution of molecules after insertion into the plasma membrane. The latter mechanism is relevant to SPE-9 since there is no protein synthesis during spermiogenesis or in mature sperm [30, 31]. Our data show that SPE-9 is dynamically redistributed through spermiogenesis. This redistribution could occur by trapping SPE-9 in spikes and then pseudopods or through active translocation to these structures. Determining the underlying mechanism of SPE-9 localization will be interesting. The movement of SPE-9 is coincident with major rearrangements of the MSP cytoskeleton. Furthermore, SPE-9 has no known localization sequences.

For C. elegans, the precise mode of sperm entry into the oocyte is still poorly characterized. An important unanswered question concerns the region of C. elegans sperm (e.g. cell body or pseudopod) that interacts with the egg surface. The requirement of SPE-9 for fertilization and its localization pattern in mature sperm is consistent with the idea that C. elegans sperm interact with the oocyte cell surface "pseudopod first". This idea also makes intuitive sense since the cell body is dragged along and the motility apparatus is in the front. In contrast, flagellated sperm propel the cell body in front and interact with the egg plasma membrane in a spatially restricted manner [5]. In mammals, the sperm cell body is highly mosaic and only an equatorial region of the sperm head is important for sperm binding and fusion with the egg plasma membrane [11, 32]. Furthermore, molecules thought to mediate these events are specifically restricted to this region [33].

This is the first report describing the wild type localization pattern of SPE-9 in sperm. The distribution of SPE-9 through spermiogenesis is dynamic. The accumulation of SPE-9 in specific sperm structures is very fast and is coincident with dramatic rearrangements in the major sperm protein cytoskeleton. The mechanism of this redistribution may be novel since SPE-9 has no known localization signals (e.g. PDZ recognition sequence). The concentration of SPE-9 in the pseudopod suggests that this sperm structure is not only important for locomotion but is also important for gamete interactions. Finally, we have generated new reagents that will be useful for the future study of sperm development and function. For instance, SPE-9 antibodies can be used for biochemical studies or as cellular markers for the study of sperm mutants.