The Major Bactericidal Activity of Human Seminal Plasma Is Zinc-Dependent and Derived from Fragmentation of the Semenogelins¹

Seminal plasma contains a wide range of inorganic and organic constituents produced by the seminal vesicles, the prostate, the epididymis, and the bulbourethral glands (1). The major proteins in seminal plasma are semenogelin I and II (SgI and SgII).³ They are synthesized in the seminal vesicles and are present in seminal plasma with concentrations of ∼50 g/L of SgI and ∼10 g/L of SgII. Semenogelins are thought to affect the sperm capacitation and motility (2). They are expressed in the seminal vesicles, and also a small amount of SgII is expressed in the epididymis (3). The sequence similarity between SgI and SgII is 78% and they contain similar repeats of 60 aa. SgI (50 kDa) contains six such repeats and SgII (63 kDa) contains eight repeats (3). After ejaculation, SgI, SgII, and fibronectin aggregate and form a coagulum that traps and immobilizes the spermatozoa (4, 5). The coagulum is then liquefied within 15–20 min as the semenogelins are degraded into smaller fragments, mainly by prostate-specific Ag (PSA) (6, 7), thus freeing the spermatozoa.

One of the major roles of seminal plasma is to provide nutrients and protection for the spermatozoa in the female reproductive tract. The protection of the spermatozoa is of vital importance for reproduction. To ensure fertilization, the integrity of the spermatozoa needs to be maintained both while still in the male genital tract and, after ejaculation, in the female genital tract. While the male reproductive tract is sterile, the healthy female vaginal tract is colonized mostly by lactobacilli that produce lactic acid and thereby cause an acidic environment in the vagina. The lactic acid produced by the lactobacilli is the major antibacterial component of vaginal fluid (8), and the low pH constitutes a major barrier to colonization with pathogenic bacteria. However, since the vagina is not sterile, it reasons that seminal plasma contains antibacterial components to ensure that the spermatozoa ascend into the sterile uterine environment devoid of microbes.

Many antibacterial components have been identified in seminal plasma, such as lysozyme (9), lactoferrin (10), phospholipase A₂ (11), secretory leukocyte protease inhibitor (SLPI) (12), and semenogelin I-derived peptides (13). Nevertheless, there is scant information available on the role of these polypeptides in host defense during and after coitus. A recent study from our group demonstrated that, upon interaction of seminal plasma and vaginal fluid, an antimicrobial peptide ALL-38 is generated following coitus (14). In seminal plasma, ALL-38 exists in abundant amounts as the inactive precursor protein hCAP-18 (15). Proteolytic cleavage of hCAP-18 by semen-derived gastricsin occurs when the seminal plasma pH drops in the vagina after coitus, which subsequently liberates the mature and active ALL-38 peptide (14). However, since gastricsin is activated by low pH 2–7 h after sexual intercourse (16), and most viable spermatozoa enter the uterus ∼30 min after sexual intercourse (17), these sperm might not benefit from the full antimicrobial potential of coitally activated ALL-38.

Herein, we sought to identify which components of seminal plasma are biologically relevant for its intrinsic bactericidal activity. We found that most of the bactericidal activity of seminal plasma was a result of peptides derived proteolytically from semenogelin (SgI and SgII). This bactericidal activity was dependent on zinc ions, which are present in seminal plasma in high concentrations and have previously been reported to bind the histidine-rich semenogelins (18).

Monoclonal mouse anti-semenogelin Abs were generated as previously described (19). Polyclonal goat anti-SLPI Abs were from R&D Systems. Polyclonal rabbit anti-hCAP-18 Abs were generated as previously described (20). HRP-conjugated Abs were from DakoCytomation, and PSA was from Sigma-Aldrich.

Overnight cultures of the bacterial strains Streptococcus pyogenes M5 Manfredo, Escherichia coli (strain 37.4), Staphylococcus aureus (strain 5120), Streptococcus agalactiae SB1, Pseudomonas aeruginosa (strain 1553), and Enterococcus faecalis (strain 2374) were grown in THY broth (Todd Hewitt with addition of yeast). Neisseria gonorrhoeae (strain 34327) was grown over night on GC agar base plates.

Freshly ejaculated semen was collected from healthy volunteers and patients at Malmö University Hospital Fertility Clinic. The semen was allowed to liquefy for 1 h at room temperature followed by centrifugation at 1,000 × g at 4°C for 10 min. The supernatant was collected, aliquoted, and stored at −20°C. Four pools of seminal plasma from four to five healthy donors were used in the experiments. Furthermore, seminal plasmas from six individual donors were used in verification experiments. Before the seminal plasma was used in a CFU assay it was dialyzed (molecular weight cutoff of 3500 Da) against 25 mM ammonium acetate (pH 7) at 4°C. Postcoital seminal plasma in vaginal fluid was collected from healthy donors 10 h following sexual intercourse and stored at −20°C as previously described (14).

Seminal plasma was dialyzed against 25 mM ammonium acetate (pH 7), mixed with pre-equilibrated CM beads (MacroPrep CM cation exchange support; Bio-Rad Laboratories), and incubated overnight at 4°C. The beads were washed with 25 mM ammonium acetate (pH 7) and cationic peptides were eluted with 30% acetic acid. The eluate was dried down and dissolved to its starting volume in 10 mM Tris, 5 mM glucose (pH 7.5), and this was repeated until the acetic acid was removed and pH of the solution was 7.5.

SgI and SgII were purified from human seminal plasma as previously described (18) and the purified SgI was cleaved by incubation with PSA in 50 mM Tris-HCl, 0.5 M urea (pH 8.0) as previously described (21).

Divalent ions were depleted from seminal plasma by incubation with 125 mM EDTA. Subsequently, the EDTA was removed by dialysis. Ni-NTA agarose (Qiagen) was equilibrated with coupling buffer (50 mM NaH₂PO₄, 2 M urea (pH 7.4)) and the seminal plasma sample was diluted in coupling buffer and incubated under rotation with the Ni-agarose overnight at 4°C. The flowthrough was collected and the agarose washed with coupling buffer, and the semenogelin was eluted with coupling buffer with addition of 500 mM imidazole. The eluate was dried down to its starting volume and then dialyzed against coupling buffer to remove the imidazole.

Gel overlay assay was performed essentially as described by Lehrer et al. (22). Briefly, duplicate samples were run on nondenaturing acid urea (AU-PAGE) gels in 5% acetic acid at 100 V for 1 h 15 min. Bacteria were grown overnight in THY broth, inoculated, and grown for 3 h. The bacteria were washed and resuspended in 10 mM NaPO₄ (pH 7.4). Bacteria (4 × 10⁶) were added to 12 ml melted underlay agarose (10 mM NaPO₄ (pH 7.4), 0.03% TH broth, 1% agarose type 1 (Sigma-Aldrich)) and poured into a square petri dish. One AU gel was stained with Coomassie brilliant blue and one AU gel was washed three times for 4 min in 10 mM NaPO₄ (pH 7.4) and then placed on top of the agarose and incubated for 3 h at 37°C. The AU gel was then removed and an overlay agarose (6% THY broth, 1% agarose type 1 (Sigma-Aldrich)) was poured on top of the underlay and incubated overnight at 37°C. To make the clearing zones more visible, the agarose was stained with Coomassie brilliant blue and then destained.

Except for N. gonorrhoeae, bacteria were grown overnight, inoculated, and grown for 3 h, washed, and resuspended in 10 mM Tris, 5 mM glucose (pH 7.5). Approximately 8 × 10⁴ CFU of bacteria were incubated together with different seminal plasma samples in Tris-glucose buffer for 1 h at 37 °C. After incubation the suspension was diluted and plated. The plates were incubated overnight and the colonies counted. N. gonorrhoeae was grown on GAB plates and bacteria were taken directly from the plates and resuspended in 0.7% casamino acids. The antimicrobial assays were performed in the same way as for the other bacteria with the exceptions of using casamino acids buffer instead of Tris-glucose and incubating for 3 h instead of 1 h. The longer incubation time was chosen to allow proper killing of the slow-growing N. gonorrhoeae.

Samples were run on an AU gel and transferred onto a polyvinylidene difluoride membrane. The membrane was fixed in 0.05% glutaraldehyde in PBS, washed in PBST (PBS with 0.05% Tween 20), and blocked in 3% skimmed milk in PBST. The membrane was incubated with primary Abs in 1% skimmed milk for 1 h at room temperature, washed in PBST and incubated with secondary HRP-conjugated Abs in 1% skimmed milk for 1 h at room temperature, washed in PBST, and developed with SuperSignal West Pico chemiluminescent substrate (Pierce).

Seminal plasma samples were run on two identical AU gels. One was stained with Coomassie brilliant blue and the other was used in a gel overlay assay. Bands corresponding to the clearing zones in the agarose gel were cut out from the stained gel and sent to Alphalyse, where the gel band was treated with trypsin that cleaves the protein after lysine and arginine residues, and the resulting peptides were analyzed by MALDI-TOF/TOF mass spectrometry. The masses of the peptides were used to query sequence databases for proteins with matching peptide masses.

Two peptides from SgII with the sequence H-KQEGRDHDKSKGHFHMIVIHHKGGQAHHG-OH (SgII peptide A) and H-QDQEHGRKAHKISYPSSRTEERQLHHGE-OH (SgII peptide B) were purchased from Schafer-N. The peptides were synthesized using the Fmoc (N-(9-fluorenyl)methoxycarbonyl) strategy, purified by reverse-phase chromatography, and analyzed by HPLC coupled to a QP800 spectrometer. The purities were >95%.

An overnight culture of S. pyogenes was inoculated and grown for 3 h, washed twice in 10 mM Tris, 5 mM glucose (pH 7.5), and resuspended in the same buffer. The bacteria (2 × 10⁷ CFU) were incubated for 1 h at 37°C in buffer only or with addition of cationic peptides extracted from seminal plasma. Samples of S. pyogenes suspensions were adsorbed onto carbon-coated copper grids for 2 min, washed briefly on two drops of water, and negatively stained on two drops of 0.75% uranyl formate. The grids were rendered hydrophilic by glow discharge at low pressure in air. Specimens were observed in a Jeol JEM 1230 electron microscope operated at 60 kV accelerating voltage. Images were recorded with a Gatan MultiScan 791 CCD camera.

Two-dimensional gel electrophoresis was performed essentially as described (23). The samples were mixed 1/1 with (0.2% cetyltrimethylammonium bromide (CETAB) and 0.1% acetic acid) and then 1/1 with 3× AU loading dye (9 M urea in 5% acetic acid with methyl green). Seven microliters of each sample was electrophoresed on a first-dimension AU-PAGE gel. First-dimension strips were stained in 10% amido black, washed in double-distilled H₂O, and 50 mM Tris (pH 8.8) and then stored in equilibrium buffer containing 10 mg/ml DTT at −80°C. First-dimension strips were subsequently electrophoresed on a second-dimension Tris-tricine SDS-PAGE gel, and the gel was stained with SYPRO Ruby stain (Bio-Rad) and imaged under UV light.

We examined the bactericidal activity of seminal plasma by gel overlay and CFU assays. In the gel overlay assay we found several bands with activity against E. coli (Fig. 1,A) while little or no antibacterial activity was observed in seminal plasma by the CFU assay (Fig. 1,B). The activity of antimicrobial (poly)peptides is often inactivated by physiological salt concentration in vitro. To investigate whether the lack of antibacterial activity in the CFU assay was due to the salt concentration, seminal plasma was dialyzed against ammonium acetate before performing the CFU assay. After dialysis we found prominent antibacterial activity of seminal plasma even after diluting the sample to 2% of the original concentration (Fig. 1,B). In the gel overlay assay seminal plasma displayed similar antibacterial activity before and after dialysis (Fig. 1,A). To test whether the antibacterial activity was due to cationic (poly)peptides, these cationic (poly)peptides were extracted from seminal plasma and CFU assays were performed. The cationic fraction was found to be equally active to dialyzed seminal plasma (Fig. 1, B and C). No antibacterial activity was found in the seminal plasma after extraction of the cationic fraction or in protein free seminal fluid (Fig. 1 C), demonstrating that the cationic fraction was solely responsible for the antibacterial activity of seminal plasma.

The antibacterial activity of dialyzed seminal plasma as well as the cationic fraction was tested against a variety of microbes. We found prominent activity against group A and group B streptococci, E. coli, S. aureus, and E. faecalis as well as modest activity against P. aeruginosa (Fig. 1 C) by CFU assay. No activity was found against N. gonorrhoeae. Increasing the concentration of cationic peptides did not increase the antibacterial activity against P. aeruginosa or N. gonorrhoeae (data not shown).

To examine the bacterial killing mediated by the seminal plasma group A streptococci, S. pyogenes M5 was incubated with cationic fraction from seminal plasma and analyzed by electron microscopy. The antibacterial components of seminal plasma caused lysis of the bacteria (Fig. 1 D).

To identify the bactericidal (poly)peptides, seminal plasma and extracted cationic peptides were tested in gel overlay assays against E. coli, and bands corresponding to the zones of clearance were excised and identified by trypsin digestion followed by mass spectrometry. Twenty bands in total were identified. Of these, two bands were identified as SLPI, two bands as lactoferrin, and one band as phospholipase A_2. However, 15 bands were identified as semenogelins. Nine bands were SgI, and six bands were SgII. Except for SgII, all these (poly)peptides have known antibacterial activity (13, 12, 10, 11). To demonstrate that a peptide derived from SgII was functionally antimicrobial, we tested the antibacterial activity of a synthetic histidine-rich peptide derived from the sequence of SgII (H-KQEGRDHDKSKGHFHMIVIHHKGGQAHHG-OH), named SgII peptide A. SgII peptide A exhibited antibacterial activity in CFU assay in a concentration of 0.32 μM against S. pyogenes (Fig. 2). In CFU assays the antibacterial activity of the peptide against E. coli was modest but potentiated by low pH or the addition of zinc as seen with other histidine-rich antimicrobial peptides (Fig. 2). Since SgII is present in a concentration of ∼16 μM in seminal plasma, single fragments of SgII will be present in antimicrobial relevant concentrations in seminal plasma. The SgII peptide A was selected because of its high histidine content and positive charge at neutral pH. This made it likely that the peptide would be antibacterial.

Not all histidine-rich fragments of semenogelins display these characteristics. We chose to test the antibacterial activity of another synthetic SgII fragment with histidines (H-QDQEHGRKAHKISYPSSRTEERQLHHGE-OH), SgII peptide B. This peptide was very similar to SgII peptide A but had both a lower content of histidines and a lower positive charge at neutral pH. SgII peptide B did not display any significant antimicrobial activity even after addition of zinc or at low pH, demonstrating that not all semenogelin fragments are antibacterial (data not shown). The demonstration of antibacterial activity of SgII indicated a major role of the semenogelins in the antibacterial activity of seminal plasma.

After ejaculation the semenogelins are continuously degraded to smaller and smaller fragments mainly by PSA-mediated proteolysis (21). Our method for identification of antibacterial (poly)peptides did not allow us to discriminate between the semenenogelin holoproteins or the degradation products. To determine whether the semenogelin holoproteins or only the degradation products were antibacterial, we tested the antibacterial activity of purified SgI and SgII, PSA, and PSA-cleaved SgI in a gel overlay assay. We chose a gel overlay assay since this allowed us both to analyze samples of semenogelin holoprotein stored in urea that prevents spontaneous degradation of the semenogelins and, importantly, to determine whether we had any degradation of the semenogelins. We did not find any antibacterial activity of either the semenogelin holoproteins (5 μg) or PSA (5 μg) (Fig. 3,A), but semenogelin cleaved by PSA was antibacterial (Fig. 3 B). Thus, only degradation products of the semenogelins were antibacterial and contributed to the antibacterial activity of seminal plasma.

To determine the origin of the major bactericidal components in seminal plasma, we compared the antibacterial activity of normal seminal plasma with seminal plasma from two vasectomized patients (contains no material from testicles and epididymis) and seminal plasma from three patients with seminal vesicle dysfunction (contains no material from seminal vesicles). The seminal plasma samples were tested both in a CFU assay (after dialysis) and in a gel overlay assay. The seminal plasma from patients with dysfunctional seminal vesicles had no antibacterial activity (Fig. 4, A and B). Apart from a diminished activity against E. coli, the seminal plasma from vasectomized patients had similar antibacterial activity as compared with normal seminal plasma (Fig. 4 B). Thus, the major bactericidal activity in seminal plasma seemed to originate from the seminal vesicles.

To assess which of the identified (poly)peptides was responsible for the major antibacterial activity, we analyzed normal seminal plasma, seminal plasma from vasectomized patients, and seminal plasma from patients with dysfunctional seminal vesicles by two-dimensional gel analysis and compared for protein content as previously described (23). Normal seminal plasma and plasma from vasectomized patients did not differ appreciably in protein content (Fig. 5 and Table I). However, seminal plasma from patients with dysfunctional seminal vesicles had much lower protein content and was practically devoid of semenogelins. By spot matching, only one protein spot of SgI was identified in seminal plasma from patients with seminal vesicle dysfunction compared with 15 spots in both normal seminal plasma and seminal plasma from vasectomized patients. For SgII only, one spot out of six was identified (Table I). Since SgII is also expressed in the epididymis, this was not surprising. These results were corroborated by Western blots that demonstrated that normal seminal plasma and seminal plasma from vasectomized patients contained semenogelin immunoreactivity, whereas seminal plasma from patients with dysfunctional seminal vesicles did not (Fig. 4,C). Furthermore, immunoblotting was performed for two antimicrobial (poly)peptides not detected by the two-dimensional gel analysis, hCAP-18 and SLPI. hCAP-18 is synthesized in the epididymis (15) and was found in normal seminal plasma and seminal plasma from patients with dysfunctional seminal vesicles but not in seminal plasma from vasectomized patients (Fig. 4,C). SLPI is synthesized throughout the male genital tract (24) and was found in all samples (Fig. 4 C). These data supported a major role for the semenogelins in the antibacterial activity of seminal plasma.

Ejaculate was collected and samples were stored in urea at different time points to inhibit further degradation of the semenogelins. The samples were electrophoresed on AU-PAGE followed by either immunoblotting with anti-semenogelin Abs or antibacterial gel overlay assay. The immunoreactive bands of semenogelins corresponded to the clearing zones on the overlay assay (Fig. 6,A), and the immunoreactivity of the semenogelins was lost over time as the peptides were degraded. The loss of semenogelin immunoreactivity was paralleled by the disappearance of clearing zones on the overlay assay (Fig. 6,A). These data were confirmed by a CFU assay whereby seminal plasma had lost antibacterial activity after incubation for 24 h at 37°C (Fig. 6,B). The most prominent clearing zones were found after 30 min to 2 h after ejaculation, indicating that the antibacterial activity is at a maximum after resolution of the seminal clot. In CFU assays the antibacterial activity of seminal plasma persisted for a longer period of time than in the gel overlay assay. This is most likely due to the facts that the CFU assay is much more sensitive and that in the CFU assay the antibacterial activity is a readout of all the active antibacterial (poly)peptides present in the sample, while in overlay assay a single (poly)peptide must be present in antibacterial concentration for antibacterial activity to be found. In contrast to the semenogelins, another antibacterial polypeptide susceptible to proteolytic processing, hCAP-18, was not degraded by incubation for 24 h at 37 °C (Fig. 6,C). Incubation at 4°C for 24 h did not cause loss of antibacterial activity (Fig. 6,B) or loss of the immunoactivity against semenogelins (Fig. 6 C). To validate these data antibacterial activity was examined by overlay assays with seminal plasma from six individual donors not contributing to the seminal plasma pools. For all six donors prominent bands with antibacterial activity were found corresponding to semenogelin immunoreactivity. After 48 h incubation at 37°C both the antibacterial activity and the semenogelin immunoreactivity were lost (data not shown).

Between different pools of seminal plasma and different experiments we found that the bands of semenogelins and the bacterial clearing zones had migrated differently. This is most likely due to the various degradations of the semenogelins in these experiments. To confirm that prolonged degradation of the semenogelins led to loss of antibacterial activity, we incubated the holoprotein of SgI with PSA for 24 h at 37°C. This resulted in complete loss of antibacterial activity as well as a complete degradation of SgI (Fig. 3 C).

To investigate whether the semenogelins are degraded postcoitally in vivo, we examined a sample of seminal plasma collected from vagina 10 h after sexual intercourse as previously described (14). Compared with seminal plasma, only small amounts of semenogelins were found in the postcoital sample even after adjusting for a 2-fold dilution (Fig. 6,D). In contrast, SLPI was readily detected both in seminal plasma and in the postcoital sample (Fig. 6 D). Quantification of the intensity of the SLPI bands revealed only a loss of 20% in immunoreactivity in the postcoital sample compared with the diluted seminal plasma. We have previously also detected large amounts of processed hCAP-18 in the postcoital sample (14). Thus, the loss of semenogelin immunoreactivity seemed to be selective for the semenogelins.

Furthermore, no antibacterial activity was found in the postcoital sample by gel overlay assay (Fig. 6 D), indicating the antibacterial activity of seminal plasma also after sexual intercourse is dependent on the semenogelins.

Semenogelins are histidine-rich proteins and bind zinc. Zinc is present in seminal plasma in a concentration of 2 mM and is bound to the semenogelins. The binding of zinc has implications for semenogelin-mediated inhibition of PSA activity (18) Since it is known that binding of zinc is important for the antimicrobial activity of other histidine-rich peptides (25), we wanted to determine whether zinc was important for the antibacterial activity of seminal plasma. Seminal plasma was treated with EDTA to remove all divalent cations and then dialyzed in ammonium acetate to remove the EDTA. Seminal plasma depleted of divalent cations was not antibacterial in CFU assays (Fig. 7). The addition of zinc (unbound zinc removed by dialysis) fully restored the antibacterial activity. Addition of two other divalent cations, calcium or magnesium, did not fully restore the antibacterial activity of divalent cation-depleted seminal plasma (Fig. 7).

Some histidine-rich antimicrobial peptides, whose antibacterial activity is dependent on zinc at neutral pH, are capable of exerting antimicrobial activity at low pH without the presence of zinc (26). However, using a similar experimental setup we found no antibacterial activity of divalent cation-depleted seminal plasma in low pH, demonstrating that binding of zinc is essential for the antibacterial activity of seminal plasma even at low pH (data not shown).

Due to the importance of zinc, we tested whether the decrease in antibacterial activity of seminal plasma from vasectomized patients could be restored by addition of zinc. Indeed, addition of zinc increased the antibacterial activity of seminal plasma from vasectomized patients against E. coli (Fig. 4 B). No increase in antibacterial activity was seen when zinc was added to the seminal plasma from patients with dysfunctional seminal vesicles (data not shown).

To selectively remove semenogelin from seminal plasma, the seminal plasma was treated with EDTA, dialyzed, and incubated with nickel-agarose. Flowthrough representing the semenogelin-depleted seminal plasma was collected and the bound proteins were eluted. Western blots with semenogelin Abs showed that semenogelin was depleted from the semenogelin-depleted seminal plasma and present in the semenogelin eluate (Fig. 8,A). SLPI was used as a control and found to be present in semenogelin-depleted seminal plasma but not in the eluate containing the semenogelins (Fig. 8,A). Zinc was added to both the semenogelin-depleted seminal plasma and to the semenogelin-containing eluate, and the antibacterial activity was analyzed by gel overlay assay. The semenogelin-depleted seminal plasma was not antibacterial whereas the semenogelin-containing eluate demonstrated significant activity (Fig. 8,B). The clearing zones on the antibacterial overlay assay corresponded to semenogelin immunoreactivity apart from the fastest migrating zone. As expected, the semenogelin-containing eluate was not antibacterial without addition of zinc (data not shown). To verify that the antimicrobial activity was due to semenogelins, we analyzed the semenogelin eluate by two-dimensional electrophoresis and bands in the cationic antibacterial fraction were identified by spot matching (Fig. 8,C and Table II). Of 24 spots, 11 were SgI and 6 were SgII. Most of the other spots were also found in seminal plasma from the seminal vesicle-deficient patient devoid of antimicrobial activity, and none of the other proteins are known to have roles in host defense.

Purified semenogelins are rapidly degraded if not stored in urea (21). Consequently, we found that the eluate from the nickel-agarose column had to be stored in urea to avoid degradation of the semenogelins and preserve the antibacterial activity. Indeed, in absence of urea both the antibacterial activity and immunoreactivity toward the semenogelins were very rapidly and totally lost. The necessity of urea to preserve the antimicrobial activity of the semenogelin-containing eluate indicates that the fastest migrating clearing zone found by the antibacterial overlay assay not detected by the semenogelin Ab may be a semenogelin fragment not recognized by the mAb used. Due to the presence of urea, which is antibacterial, antibacterial testing was not performed by CFU assay.

Collectively, our data demonstrate that the major bactericidal activity of seminal plasma was transient and generated as the semenogelins were degraded. As the semenogelins were further degraded they lost the bactericidal activity, and the bactericidal activity of seminal plasma was consequently greatly reduced. Although other antibacterial (poly)peptides like SLPI and lactoferrin are present in seminal plasma, we found that the contribution of these (poly)peptides to the bactericidal activity was minor compared with the activity of the semenogelin-derived peptides.

Our data demonstrate that the holoproteins of the SgI and SgII that trap the spermatozoa after ejaculation are not bactericidal. However, the semenogelin-derived peptides formed by cleavage are the major bactericidal components in seminal plasma. When the seminal coagulum is liquefied the spermatozoa are exposed to a high concentration of semenogelin-derived antibacterial peptides. Accordingly, any microbes associated with the spermatozoa after ejaculation are trapped together with spermatozoa and are exposed to the semenogelin-derived peptides as the spermatozoa are released. It has previously been found that SgI associates with eppin, a cysteine-rich protein from testis and epidydimis, on the surface of spermatozoa and thereby provides the spermatozoa with antibacterial protection (13, 27) even after the spermatozoa has left the vagina. Interestingly, even though seminal plasma displayed both rather broad and potent bactericidal activity, using the same assays, seminal plasma was not bactericidal against N. gonorrhoeae, an important pathogen causing sexually transmitted disease.

We did not find any antibacterial activity of either SgI or SgII holoprotein in a gel overlay assay using 5 μg of both proteins. This is in contrast to a previous report by Bourgeon et al. where the holoprotein of SgI was reported to be antibacterial in antibacterial radial diffusion assay (13). However, details of the purification were not provided nor was there any verification that the holoprotein used was intact. To keep the purified holoproteins of the semenogelins intact, the proteins must be stored in urea to prevent spontaneous aggregation and degradation (21). However, urea is antibacterial and is not compatible with the radial diffusion assay, and consequently we used the antibacterial gel overlay assay. In this assay, samples are eletrophoresed in the presence of urea. Afterwards, the urea is washed away and the gels are immediately put on a bacterial lawn, thus minimizing degradation. Furthermore, Coomassie staining of duplicate gels allowed verification that the intact holoproteins were used.

The antibacterial activity of seminal plasma greatly decreased over time, and we found this to parallel further degradation of the semenogelin-derived peptides, but not other antibacterial (poly)peptides. As the semenogelins were degraded to smaller and smaller fragments, they lost their antibacterial activity. It is possible that prolonged antibacterial effect of the seminal plasma may disturb the normal vaginal flora. Accordingly, it is tempting to hypothesize that the transient effect of semenogelin-derived peptides in seminal plasma offers suitable antibacterial protection of the spermatozoa without imposing long-term effects on the vaginal microflora. Hence, after the spermatozoa have left the vagina the antibacterial activity of seminal plasma is gradually reduced as the semenogelins are further degraded.

The zinc concentration in seminal plasma is ∼2 mM compared with 10–18 μM in plasma (1). Semenogelins are histidine-rich proteins, much like the histatins in saliva. Indeed, the sequence of the PSA-generated semenogelin fragments (21) resemble closely in charge and histidine content other histidine-rich peptides with known antimicrobial activity. Histidine-rich (poly)peptides are typically antimicrobial in a zinc-dependent manner (25). Free zinc is directly antibacterial at concentrations as low as 8 μM (our unpublished data), but the zinc in seminal plasma is protein-bound mainly to the semenogelins (18), and we found that protein-free seminal fluid was devoid of antibacterial activity, confirming that the most of zinc present in seminal plasma is protein-bound. Interestingly, we found that binding of zinc was necessary for the antibacterial activity of the semenogelin-derived peptides in seminal plasma. Other histidine-rich peptides lose the dependence on zinc for antimicrobial activity at low pH (26). However, using a similar experimental setup, we found that the antibacterial activity of seminal plasma was strictly dependent on zinc even at low pH. Thus, zinc is mandatory for the antibacterial activity of human seminal plasma even in the low pH found in the vagina.

In conclusion, we found that the major antibacterial activity of seminal plasma was transient and dependent on semenogelin-derived peptides in a zinc-dependent manner. We propose that the immediate protection of spermatozoa against microbes before entering the uterus as well as preventing ascension of bacterial pathogens along with the spermatozoa into the female upper genital tract are due to the semenogelin-derived peptides and that this antibacterial activity may be one of the most important physiological effects of these peptides. These findings ascribe a novel role for semen coagulation and liquefaction in the protection of spermatozoa and the female genital tract against bacteria.

The outstanding technical assistance of Maria Baumgarten and Malgorzata Berlikowski is greatly appreciated.

The authors have no financial conflicts of interest.