The role of epididymal sperm-binding proteins in reproductive tract immunity is now well recognized in addition to their role in sperm maturation. Spermatozoa acquire forward motility and fertilizing ability during their passage through the epididymis, where they acquire a wide variety of proteins belonging to different classes. Previously, we demonstrated that EPPIN (epididymal protease inhibitor), an androgen-regulated, sperm-binding protein containing protease-inhibitory motifs, is expressed specifically in the testis and epididymis. In the present study, we investigated the antibacterial activity of EPPIN against Escherichia coli and the mechanism of antimicrobial action. EPPIN exhibited dose- and time-dependent antibacterial activity that was relatively insensitive to salt. However, EPPIN lost its antibacterial activity completely on reduction and alkylation of its cysteines, indicating the importance of disulfide bonds for its activity. EPPIN permeabilized the outer and inner membranes of E. coli, which is consistent with its ability to induce striking morphological alterations of E. coli membranes as shown by scanning electron microscopy. EPPIN did not cause disruption of eukaryotic membranes in the rat erythrocyte hemolytic assay. The present results indicate that EPPIN has a role in the innate immune system of human epididymis.
The innate immune system forms the first barrier against invading microorganisms. Among the important components of the innate immune system are the antimicrobial proteins and peptides. A wide variety belonging to different families, including defensins, cathelicidins, and protease inhibitors, have been identified in both plants and animals . In recent years, the role of antimicrobial proteins and peptides in male reproductive tract immunity has emerged as a major area of interest in efforts to identify reproductive tract-specific proteins that control sexually transmitted diseases. These diseases may be controlled in part by antimicrobial proteins, such as cathelicidins [2, 3], β-defensins [4, 5], bovine seminal plasmin , and the protease-inhibitors serum leukocyte protease inhibitor (SLPI)  and cystatin 3 , produced in the male reproductive tract. Furthermore, the antimicrobial activity of other epididymal proteins has been demonstrated [9, 10]. Previously, we demonstrated that the epididymis specific β-defensin 118 (DEFB118) , members of the HE2 family (HE2α, HE2β1, and HE2β2) and their C-terminal peptides , and cystatin 11  exhibit potent antibacterial activity by a membrane-permeabilizing mechanism, suggesting a role for epididymal proteins in innate immunity in addition to sperm maturation. Of these, hCAP18 , members of the HE2 family , DEFB118 , and cystatin 11  bind to the surface of sperm. Spermatozoa acquire these proteins during passage through the initial segment and caput epididymis, where they develop mature functions, such as forward motility and the ability to fertilize the egg.
We recently identified and characterized human EPPIN (epididymal protease inhibitor), a cysteine-rich protein containing both Kunitz-type and whey acidic protein (WAP)-type, four-disulfide core protease-inhibitor consensus sequences (Fig. 1) . EPPIN is expressed specifically in the epididymis and testis and is bound to the ejaculated spermatozoa , suggesting that it has a role in sperm maturation. The EPPIN gene, which is located within a protease-inhibitor gene cluster on chromosome 20, contains four exons and produces three splice variants (EPPIN-1, -2, and -3). Similarly, we identified and characterized mouse EPPIN and a cluster of similar protease inhibitors on mouse chromosome 2 . This cluster contains genes expressing protease-inhibitor domains with homology to WAP . The most studied WAP motif proteins in this cluster are SLPI and protease inhibitor 3 or elafin (elastase-specific inhibitor) [19, 20]. The WAP motif is a 50-amino-acid protein motif with eight cysteine residues at defined positions. The WAP motif-containing proteins are functionally diverse. They include sodium-potassium ATPase inhibitors , calcium-transport inhibitors of spermatozoa , and proteins responsible for Kallman syndrome , wound healing , and metastasis . Furthermore, earlier reports demonstrated that the WAP motif-containing proteins, such as SLPI [7, 26, 27], single WAP motif proteins 1 and 2 (SWAM1 and SWAM2) , and elafin , exhibit antimicrobial activities. The WAP motif-containing proteins including EPPIN  are expressed in reproductive tissues. For example, SLPI is expressed in the prostate, seminal vesicle, epididymis, testis , and endometrium . Elafin is expressed in human endometrium , though to our knowledge, its presence in the male reproductive tract has not been reported. However, elafin-like protein 1 expression was shown in the epididymis using DNA microarray analysis . The human epididymis-specific gene product HE4, a WAP motif-containing protein with similarity to elafin and SLPI, was found initially to be epididymis specific  and was thought to play a role in sperm maturation.
The antimicrobial properties of other WAP motif-containing proteins raised the possibility that EPPIN has a role in the innate immunity of the male reproductive tract. EPPIN contains multiple sulfhydryl linkages and a cationic nature similar to other WAP motif proteins. The mechanism of the antimicrobial action of WAP motif proteins is not understood. More is known about other cationic antimicrobial proteins and peptides, and their activities depend on their tertiary structure and are influenced by ambient salt concentrations. They interact with the negatively charged membranes of target organisms, resulting in membrane disruption and cell death [35, 36]. They are also reported to interact with specific targets inside bacteria, thereby disrupting normal cellular metabolism and viability [37, 38]. In the present study, we demonstrate the antibacterial activity of EPPIN and investigate its mechanism of action. Our results indicate that EPPIN exhibits structure-dependent and salt-tolerant antibacterial activity, and its mechanism of action involves permeabilization of bacterial membranes.
Materials and Methods
Recombinant Human EPPIN Production
Human EPPIN-1 cDNA (nucleotides 76–432) lacking part of the N-terminal secretory sequence was generated by polymerase chain reaction using the EPPIN-1/Bluescript clone as template. This construct was designed to contain a 5′ BamHI site and a 3′ KpnI site to facilitate in-frame cloning into pQE-30 (Qiagen, Valencia, CA). The expression vector was sequenced to confirm fidelity and transformed to Escherichia coli strain M15 [pREP-4] according to the Qiagen protocol. Fusion protein expression was induced with 1 mM isopropyl-1-thio-β-d-galactoside and the recombinant protein purified from a bacterial lysate with nickel-nitrilotriacetic acid-agarose (Qiagen, Valencia, CA) and then dialyzed against PBS .
The antibacterial activity of EPPIN against E. coli XL-1 blue (Stratagene, La Jolla, CA) was tested using the colony-forming unit (CFU) assay as described previously . Briefly, 2 × 106 CFU/ml of E. coli suspended in 10 mM sodium phosphate buffer (pH 7.4) were incubated with 10–100 μg/ml of recombinant EPPIN, and aliquots of the assay mixture were serially diluted with 10 mM sodium phosphate buffer (pH 7.4) after a 15- to 180-min incubation at 37°C with shaking. One-hundred microliters of each diluted samples were plated on Luria-Bertani agar plates and incubated at 37°C overnight to allow colony development. The resulting colonies were counted by hand. Bacterial survival was expressed as percentage survival using the following formula:% survival = (number of colonies surviving after treatment with the antibacterial protein/number of colonies surviving in the absence of antibacterial protein) × 100. The effect of salt on EPPIN antibacterial activity was tested by incubating E. coli in 10 mM sodium phosphate buffer containing NaCl ranging from 25 to 300 mM.
Reduction and alkylation of EPPIN were carried out according to a protocol described earlier . Antibacterial assays were performed with the denatured EPPIN to test the importance of structural integrity for its activity. A control containing sodium phosphate buffer without EPPIN was treated in parallel to test whether residual β-mercaptoethanol and iodoacetamide affected bacterial growth in the assays.
Membrane Permeabilization Assays
The ability of EPPIN to permeabilize outer and inner membranes of E. coli was tested using the dyes N-phenyl-1-napthylamine (NPN; Molecular Probes, Eugene, OR) and 3,5-dipropylthiadicarbocyanine iodide (diSC3-5; Molecular Probes), respectively, according to standard protocols described earlier [12, 40, 41]. Midlog-phase E. coli washed twice with 5 mM Hepes buffer (pH 7.4) containing 5 mM glucose and resuspended in the same buffer were used for the permeabilization assays. For the outer membrane permeabilization assay, 10–100 μg/ml of EPPIN were added to E. coli suspension containing 10 μM NPN, and the increase in fluorescence was measured using an Aminco SLM 8100 fluorescence spectrophotometer (SLM Instruments, Inc., Urbana, IL) with the excitation monochromator set at 350 nm and the emission set at 420 nm. For the inner membrane permeabilization assay, diSC3-5 was added to the bacterial suspension to a final concentration of 0.4 μM and incubated until the uptake was maximal, as indicated by a stable reduction in fluorescence. To this suspension, varying concentrations of EPPIN were added, and the increase in fluorescence was measured in an Aminco SLM 8100 fluorescence spectrophotometer with the excitation monochromator set at 622 nm and the emission set at 670 nm.
Hemolytic activity of EPPIN was determined as described previously . Rat erythrocytes washed three times with 0.9% saline were resuspended to a concentration of 5% in saline. Aliquots of the suspended erythrocytes were treated with varying concentrations (10–100 μg/ml) of EPPIN for 1 h at 37°C in a 96-well plate. After incubation, samples were centrifuged at 1000 × g for 10 min, and supernatants were transferred to a fresh plate and absorbance read at 560 nm. Hemolysis caused by saline and 1% Triton X-100 served as 0% and 100% controls, respectively.
Scanning Electron Microscopy
The effect of EPPIN on E. coli was demonstrated by scanning electron microscopy as described previously . Briefly, E. coli resuspended (108 CFU/ml) in 10 mM sodium phosphate buffer were treated with 25 μg/ml of EPPIN and incubated for 2 h at 37°C. After incubation, bacterial cells were washed and fixed overnight at 4°C with an equal volume of 4% glutaraldehyde. Following fixation, bacterial cells were vacuum-filtered onto 0.1-μm polycarbonate membrane filters (Poretics Corporation, Livermore, CA). The filters were rinsed with 0.15 M sodium phosphate buffer and dehydrated through a series of ethanol (30%–100%) concentrations. Using a critical point dryer (Balzers CPD-020; Bal Tec AG, Vaduz, Lichtenstein), the filters were dried, mounted, and coated with a 15-nm thickness of gold:palladium metal (60:40 alloy) in a Hummer-X sputter coater (Anatech Ltd., Alexandria, VA). Samples were examined on a Cambridge Stereoscan 200 scanning electron microscope (LEO Electron Microscopy, Inc., Thornwood, NY) using an accelerating voltage of 20 kV.
Statistical analyses using Student t-test were performed using Sigma Plot software (SPSS, Inc., Chicago, IL). Values are shown as the mean ± SD.
The antibacterial activity of EPPIN was tested in the CFU assay. In a concentration- and time-dependent manner, EPPIN was bactericidal against E. coli (Fig. 2). Though EPPIN, at a concentration of 10 μg/ml, did not cause any bacterial killing even after 3 h of incubation, bacterial survival declined with a concentration of 25 μg/ml for the same incubation period. Bacterial survival was reduced further with higher concentrations and shorter incubation times. Incubation of E. coli with 50 and 100 μg/ml of EPPIN for 30–180 min resulted in several log units of bacterial killing. As negative controls, the epididymis-specific lipocalin LCN6  and BSA were included in the assays. Neither LCN6 nor BSA at a concentration of 100 μg/ml exhibited any bactericidal activity even after a 2-h incubation (data not shown).
The killing activity of some cationic antimicrobial proteins and peptides is influenced by ambient salt concentrations because of the interference of monovalent or divalent cations with the electrostatic interactions involving bacterial surfaces. However, EPPIN retained most of its antibacterial activity over the range of NaCl concentrations tested. In the absence of salt (0 mM NaCl), less than 1% of the bacteria survived in the presence of EPPIN. Bacterial survival increased only up to 4% with increasing NaCl concentrations (Fig. 3A), indicating that EPPIN exhibits salt-tolerant antibacterial activity that may rely on other mechanisms in addition to electrostatic interactions. The presence of NaCl alone increased bacterial growth substantially (Fig. 3A).
Earlier reports suggest that the activity of antimicrobial proteins and peptides is dependent on the structural integrity conferred by their disulfide bond linkages. To test whether disulfide linkages are essential for EPPIN antibacterial activity, they were reduced and the cysteines alkylated. Reduced and alkylated EPPIN failed to kill E. coli even when incubated for 2 h at a concentration of 100 μg/ml (Fig. 3B), indicating the requirement of disulfide bond integrity for its antibacterial activity.
Membrane permeabilization has been implicated in the mechanism of action of cationic antimicrobial proteins and peptides. To determine whether EPPIN-mediated bacterial killing involves disruption of bacterial membranes, the outer membrane permeabilization ability of EPPIN was investigated using a fluorescent dye, NPN. This dye fluoresces weakly in an aqueous environment but strongly in the hydrophobic interior of cell membranes. Membrane disruption by antimicrobial agents facilitates the binding of NPN to the damaged membrane, where it emits strong fluorescence. EPPIN, when added to the bacterial suspension, caused a rapid increase in fluorescence within 1 min of addition in a dose-dependent manner, indicating its ability to permeabilize the outer membrane of E. coli (Fig. 4A). The inner membrane permeabilization capacity of EPPIN was analyzed using diSC3-5 dye, which distributes between the cells and the medium depending on the membrane potential. Once the dye enters the cells, it forms self-quenching aggregates, resulting in a steady decrease in fluorescence. Disruption of the inner membrane by antimicrobial proteins and peptides releases the dye back into the medium, resulting in an increase in fluorescence. We observed a dose-dependent increase in fluorescence within 2 min of addition of EPPIN to the bacterial suspension, indicating its inner membrane-permeabilizing ability (Fig. 4B). These results suggest that the antibacterial activity of EPPIN involves interaction with and permeabilization of outer and inner membranes of E. coli.
Antimicrobial proteins and peptides that disrupt bacterial membranes are sometimes toxic to eukaryotic cells. Using an erythrocyte assay, we tested whether eukaryotic membranes are disrupted by EPPIN. Hemolysis was not observed when rat erythrocytes were incubated with 10–100 μg/ml of EPPIN for 1 h (Fig. 5). The inability of EPPIN to damage the rat erythrocyte plasma membrane suggests that eukaryotic membranes, including the epithelial cells of the epididymis, likely are resistant to the antimicrobial membrane-damaging properties of EPPIN.
Scanning Electron Microscopy
Because EPPIN caused membrane disruption as evidenced by membrane permeabilization studies, we examined the morphological effects of EPPIN on the target E. coli cells. The E. coli were incubated with EPPIN as in the CFU assay and were depicted by scanning electron microscopy. Untreated control E. coli had smooth and normal surface morphology (Fig. 6, A–C). In contrast, E. coli exposed to 25 μg/ml of EPPIN for 2 h showed pronounced membrane wrinkling and surface blebbing (Fig. 6, D–F). They appeared shrunken and collapsed, with surface blebs mostly at the dividing septa. Such structural changes induced by EPPIN on E. coli provide further evidence of its membrane-dependent bacterial killing.
The activity of antimicrobial proteins and peptides is generally influenced by a variety of factors, such as their amino acid composition, charge of the peptide, disulfide bond pairing, salinity of the environment, and composition of the bacterial membranes. Structural integrity conferred by the disulfide bonds plays a vital role in maintaining their antimicrobial activity. In the present study, disulfide bond-reduced and alkylated EPPIN had no activity against E. coli. Similar effects of disulfide bond reduction on antimicrobial activity is reported for tachylepsins , protegrins , sepacins , members of the HE2 family , DEFB118 , and other antimicrobial peptides [46, 47]. High concentrations of monovalent or divalent salts inhibit the activity of some antimicrobial proteins and peptides, because their cations interfere with the electrostatic interactions between the peptide and bacterial membrane. However, EPPIN displayed little or no loss of antibacterial activity in the presence of NaCl concentrations ranging from 25 to 300 mM. Salt-resistant antimicrobial activity was also demonstrated for cathelicidin-derived peptides , members of HE2 family , DEFB118 , and other peptides [49–52], whereas β-defensins [53–55] and horse cathelicidins  were salt sensitive. Thus, EPPIN likely maintains its antibacterial activity throughout the normal physiological NaCl concentration in human seminal plasma, which is reported to range from 75 to 222 mM [57, 58].
The mechanism of cationic antimicrobial peptide activity including β-defensins involves interaction with and disruption of target organism membranes in addition to their interaction with critical intracellular factors [37, 38]. Though antimicrobial activity of some WAP motif-containing proteins is reported, the mechanism of their antimicrobial action is largely unknown. The WAP motif-containing proteins, including EPPIN, are cationic in nature and could bind to the negatively charged bacteria. Such a membrane binding and permeabilization-mediated mechanism of action has been reported for other cationic antimicrobial proteins and peptides, including defensins , bactenectins , and indolicidins . EPPIN caused rapid permeabilization of E. coli outer and inner membranes, similar to other reported cationic epididymal antimicrobial proteins and peptides, such as DEFB118 and members of the HE2 family [11, 12]. Consistent with membrane permeabilization, EPPIN induced structural changes, such as membrane wrinkling and blebbing, in E. coli. Membrane wrinkling and blebbing similar those shown in Figure 6, D and E, were earlier demonstrated for E. coli treated with salmon antimicrobial protein and magainin II . Structural changes induced by other antimicrobial proteins [63, 64] and the epididymal DEFB118  have been reported as well.
The membrane permeabilization activity of EPPIN is not necessarily a general feature of other WAP motif proteins, because they differ in their amino acid composition and the three-dimensional arrangement of amino acids. Proper positioning of cationic and hydrophobic amino acids in an antimicrobial protein or peptide may render it amphipathic, thus allowing it to bind and penetrate bacterial membranes . Antimicrobial mechanisms of other WAP motif proteins will need to be tested individually. Whether membrane permeabilization is the only event involved in EPPIN-mediated antibacterial action or whether it interacts with critical factors inside the cell, as reported for other antimicrobial proteins and peptides [35, 36], needs further investigation.
Several WAP motif proteins are thought to have antiprotease activity because of their similarity to SLPI and elafin. To our knowledge, however, experimental evidence of this activity is lacking. Similarly, the antiprotease activity of EPPIN is not yet documented. The antimicrobial activities of other WAP motif-containing proteins are not dependent on their antiprotease activities. For example, recombinant SWAM1 and SWAM2 lacked protease-inhibitory activity  but displayed antimicrobial activity. The antimicrobial activities of SLPI and elafin are independent of their antiprotease activity. In the case of SLPI, the amino terminus, which lacks the proteinase-inhibitory domain, displayed antibacterial activity against E. coli and Staphylococcus aureus . The antimicrobial activity of elafin is reported to be independent of its antielastase activity . Further studies are required to decipher the contribution of EPPIN’s antiprotease activity, if any, to its antimicrobial action. Similarly, studies employing mutational analysis of EPPIN’s WAP motif will provide further definitive evidence regarding the role of WAP motif toward the antimicrobial activity of EPPIN.
Antimicrobial peptides that kill pathogenic organisms can be cytotoxic to eukaryotic cells . The results of the present study, however, indicate the preferential disruption of bacterial membranes and not eukaryotic membranes by EPPIN, as shown by the membrane permeabilization and hemolysis assays and as might be expected of secretory proteins. Other human epididymal antimicrobial proteins and peptides also lacked any disruptive effect on eukaryotic cells [11, 12]. It is proposed that cholesterol, which stabilizes the lipid bilayer, in eukaryotic membranes, may prevent the membrane-damaging effect of EPPIN.
In conclusion, we report that EPPIN exhibits potent antibacterial activity that is structure dependent and salt tolerant. Its mechanism of action involves interaction with bacterial membranes, resulting in their disruption and, finally, leading to cell death. Taken together with earlier reports, our results suggest that WAP motif proteins are part of the innate immune system of the male reproductive tract.
We thank Dr. Ashutosh Tripathy, Macromolecular Interaction Facility (MaCInFac), University of North Carolina, for his help with the fluorescence studies. We also thank Victoria Madden, Microscopy Services Laboratory (MSL), University of North Carolina, for her help with the scanning electron microscopy. Finally, we thank Dr. R.E.W. Hancock, Department of Microbiology and Immunology, University of British Columbia (Vancouver, BC, Canada), for his suggestions regarding the permeabilization assays.