Antimicrobial Peptide

BIOLOGY OF REPRODUCTION 67, 804–813 (2002) DOI 10.1095/biolreprod.102.004564

Novel Antimicrobial Peptide of Human Epididymal Duct Origin1
Hans Henning von Horsten, Petra Derr, and Christiane Kirchhoff 2
IHF Institute for Hormone and Fertility Research at the University of Hamburg, D-22529 Hamburg, Germany
ABSTRACT
HE2, a gene expressed specifically in human epididymis, gives rise to multiple mRNAs that encode a group of small cationic secretory peptides. Localization of HE2 within the defensin gene cluster and prediction that -defensin-like modules exist suggest that these peptides have antimicrobial activity and represent components of the innate epithelial defense system of the epididymal duct. Reverse transcription-polymerase chain reaction analysis confirmed the occurrence of eight human HE2derived transcripts, including minor mRNA variants, that had previously been shown only in animal species. Employing isoform-specific antibodies against the predicted HE2 products, multiple 4- to 8-kDa peptides were detected in human epididymal epithelium, epididymal fluid, and ejaculate. N-terminal microsequencing has suggested a proteolytic processing of these peptides by a furin-like proprotein convertase, which cleaves a propiece from the longer precursor peptides. HE2 and HE2 1, representing major peptide isoforms in the human epididymis, were recombinantly expressed, and their susceptibility to furin cleavage was demonstrated in vitro and in vivo. Processed recombinant peptides and chemosynthetic fragments were included in antimicrobial tests. In addition to the -defensin-like HE2 1 with its expected antibacterial function, HE2 C-terminal fragments showed antibacterial activity against Escherichia coli, although it showed no significant similarity to -defensins nor to any other known protein family.

epididymis, gene regulation, sperm

INTRODUCTION

Infections of the genitourinary tract are a major cause of morbidity and may contribute to infertility, but possible defense mechanisms are largely unknown. Because the luminal contents of the testis and epididymis are isolated from adaptive immune responses for obvious reasons (for a review see [1]), it seems compelling that their epithelia have developed alternative mechanisms to fight ascending infections, including local production of antimicrobial peptides and proteins. Among the most potent innate immune effectors of mammalian epithelia are the -defensins, small cationic peptides with six cysteines [2]. HE2, a gene expressed specifically in human epididymis [3], is located within the human defensin gene cluster and predicts -defensin-like modules [4], suggesting that it may be part of the epididymal epithelial defense system. A gene product
Supported by the German Research Association (DFG contracts Ki 317/5 and Ki 317/8) and by a graduate stipend from the Ernst Schering Research Foundation to H.H.vH. Results presented in this publication are part of the doctoral thesis of H.H.vH. 2 Correspondence: Christiane Kirchhoff, IHF Institute for Hormone and Fertility Research, Grandweg 64, D-22529 Hamburg, Germany. FAX: 49 40 56190864; e-mail: [email protected]
1

Received: 14 February 2002. First decision: 11 March 2002. Accepted: 4 April 2002. 2002 by the Society for the Study of Reproduction, Inc. ISSN: 0006-3363. http://www.biolreprod.org

of the rat epididymal duct, bin1b, has recently been cloned [5] and has been suggested to represent the rat counterpart of the human HE2 gene. Bin1b has a predicted peptide with the structural characteristics of -defensins. The highly restricted expression pattern of the mRNA, which was found only in the proximal parts of the rat epididymis, led to the assumption that the encoded peptide was effective against invasive pathogenic microorganisms while simultaneously preserving spermatozoa from cytotoxic damage during epididymal transport, maturation, and storage [5]. HE2 transcripts had originally been cloned from a human epididymal cDNA library by differential screening [3]. The cDNA clones obtained were derived from abundant epithelial transcripts and showed a highly restricted hybridization pattern at levels detectable by Northern blot, suggesting that HE2 was produced in large amounts only by the proximal parts of the human epididymal duct epithelium [3, 6]. It is now obvious that the initially identified HE2 mRNA is a member of an entire family of human mRNA variants [7] that result from alternative mRNA splicing. Differential expression patterns of HE2 variants have been described and suggest that the encoded peptides may exert their effects consecutively during epididymal transit [7]. Part of the HE2 mRNA variants have predicted cationic peptides with a pattern of six cysteine residues near their C-termini [7, 8], which is a hallmark of the -defensin family of antimicrobial peptides [2]. Moreover, the human genome project localized the HE2 gene on the short arm of chromosome 8, directly adjacent to the -defensin genes for hBD-2 and hBD-3 (segment 8p22-p23; [4, 9]). The prediction of the existence of -defensin-like modules and their location within the human defensin gene cluster suggest that HE2 might encode novel antimicrobial peptides of the human male genital tract that are suitable for fighting specific genital tract infections. However, very little is known about the production and secretion of the predicted human peptides. A detailed characterization at the peptide level has been hampered because HE2 expression is largely restricted to the proximal human epididymis. The antisera employed previously [3, 7] were not well suited to detect all major human peptide isoforms. Thus, proof of the existence of the HE2 peptide still does not exist. Also, any direct functional evidence of the encoded peptides is lacking. Indeed, despite the prediction of -defensin-like modules in rat bin1b and a number of HE2derived peptides, an antimicrobial activity of these peptides has not been shown. Moreover, the HE2 sequence is entirely different from the defensins and from any other known peptide, and a function cannot be inferred. To characterize the HE2-derived peptides whose origin is the human epididymal duct, isoform-specific antibodies were raised and human epididymal tissue, luminal fluid, and ejaculates were analyzed by immunohistochemistry, Western blotting, immunoprecipitation, and peptide microsequencing. Chemosynthetic and recombinantly expressed HE2-derived peptides were included in tests of antimicrobial activity. 804

ANTIMICROBIAL PEPTIDE

805

MATERIALS AND METHODS

Epididymal Tissues, Fluids, and Ejaculates
Human testes and epididymides were obtained from local hospitals where patients were undergoing orchidectomy for prostatic carcinoma (courtesy of Drs. Ching-Hei Yeung and Trevor G. Cooper, University of Munster). Two patients had received long-term treatment with GnRH and ¨ antiandrogen before surgery. Informed consent was obtained from all patients and, in all cases, investigations were conducted in accordance with the guidelines of the Helsinki Declaration, as revised in 1983. Epididymal fluids were collected by cannulation and retrograde perfusion of two additional organs as described elsewhere [10]. Fluid samples were dispensed to aliquots before storage. Permission was sought from the Hansestadt Ethics Committee, in Hamburg, Germany, in order to perform studies involving human ejaculates from young, healthy volunteers. Rat tissues were obtained from freshly killed laboratory animals. All tissue and fluid samples were shock-frozen in liquid nitrogen and stored at 80 C.

AF168620) by standard f-moc solid-phase procedures (courtesy of Dr. Markus Koppitz, Schering AG, Berlin, Germany). Sequences were as follows: P1, H2N-GELRERAPGQGTNGC-COOH; P2, H2N-KRDLLPPR TPPYQVC-COOH; P3, H2N-ISHREARGPSFRICVDFLGPRWARGCST GN-COOH (representing the amino acid sequence of HE2 2 ([3]; GenBank accession number X67697); synHE2 cycl with a disulfide bond, H2N-VISHREARGPSFRICVDFLGPRWARGCSTGN-COOH; P4, H2NCVSNTDEEGKEKPEM-COOH; and P5, H2N-GDVPLGIRNTIC-COOH. All peptides eluted as single peaks upon reverse-phase high-performance liquid chromatography (HPLC). Mass spectroscopy of peptides was kindly performed by Dr. Jurgen Harder, University of Kiel. Electrospray ioniza¨ tion-mass spectrometry (ESI-MS) analyses (Q-Tof2; Micromass, Almere, Netherlands) revealed a mass of 3346.88 Da for the P3 peptide (3346.79 Da calculated for linear synHE2 ) and 3444.00 Da for the synHE2 cycl peptide (the calculated mass of the reduced peptide is 3445.93 Da), which confirmed the presence of a disulfide bond.

Antipeptide Antisera
Peptides P1 and P2 were synthesized to contain an artificial C-terminal cysteine for subsequent coupling to keyhole-limpet-hemocyanine (KLH; Sigma, Deisenhofen, Germany) as a carrier; P4 and P5 peptides contained an endogenous C-terminal cysteine for coupling. The N-terminal isoleucine of P3 was used for coupling to KLH. Rabbits were immunized with an s.c. injection, and immune sera were obtained after 60, 90, and 125 days (Pineda-Antikorper-Service, Berlin, Germany). Monospecific purifi¨ cation of polyclonal antibodies was performed by affinity chromatography after peptide coupling to epoxy-activated sepharose 4B (Amersham-Pharmacia-Biotech, Freiburg, Germany).

Preparation of Human Sperm Membrane Proteins
A salt wash and 0.3 M lithium 3,5-diiodosalicylate (LIS) extracts of sperm proteins were prepared from liquefied human ejaculates. Seminal plasma was removed by centrifugation at 500 g for 30 min, and sperm pellets were resuspended. LIS extracts were prepared as described elsewhere [11]. To obtain salt wash proteins, sperm pellets were resuspended in a cold 0.6 M NaCl solution containing 10 mM PMSF. After 10 min of incubation on ice, cold ethanol was added to give a final concentration of 60% (v/v) ethanol. After 10 min on ice, suspensions were centrifuged for 10 min at 4000 g. Supernatants were precipitated with four volumes of acetone at 20 C overnight. Precipitated proteins were harvested by centrifugation for 20 min at 4000 g at 4 C. Supernatants were decanted and pellets were dried under vacuum. Protein pellets were resuspended in deionized water, and insoluble proteins were separated by centrifugation. The soluble fraction was employed in SDS-PAGE.

Immunohistochemistry
Cryosections of human epididymides were postfixed in 4% paraformaldehyde and used in immunohistochemistry as described elsewhere [12]. Immunolocalization of HE2 peptides was achieved by indirect Cy2 immunofluorescence employing P3 and P4 antisera as first antibodies and Cy2-conjugated anti-rabbit antibodies (1:100; Jackson ImmunoResearch Laboratories, West Grove, PA) as second antibodies. Tissue sections were investigated using an epifluorescence microscope (Nikon, Japan) equipped with a 450–490 nm excitation wavelength filter (Omega Optics, Brattleboro, VT).

Preparation of Seminal Plasma Proteins
Thirty-five milliliters of liquefied human seminal plasma were poured into 200 ml of acetone and stirred at room temperature. To accelerate suspension formation, up to 5% of n-tributylphosphate was added to the mixture. Suspensions were centrifuged at 4000 g for 10 min, and the proteinaceous pellets were dried under vacuum. The powder was resuspended in sterile deionized water containing a protease inhibitor cocktail (Complete, Boehringer-Roche, Mannheim, Germany). Samples were recentrifuged at 27 000 g for 30 min, the supernatants were dispensed in 500- l aliquots and lyophilized, and the dried samples were stored at 80 C.

Immunoprecipitation
All steps were carried out at 4 C. Protein samples were titrated to pH 7. Twenty microliters of normal rabbit serum or preimmune serum were added and the samples were incubated for 1 h on a rotating wheel. Thirty microliters of water-equilibrated protein A-agarose slurry (BoehringerMannheim) were added, taking care not to compress the beads, and the tubes were again incubated on the rotating wheel for 1 h, followed by a brief centrifugation at 7000 g in an Eppendorf centrifuge. Supernatants were removed into fresh tubes and recentrifuged for 5 min at 14 000 g. Seventy microliters of affinity-purified rabbit antibodies were added and the samples were incubated for 1 h on the rotating wheel. Fifty microliters of equilibrated protein A-agarose were then added, followed by overnight incubation on the rotating wheel. Samples were centrifuged for 30 sec at 7000 g, and the supernatants were discarded. Pellets were washed with 1 ml of radioimmunoprecipitation assay buffer (150 mM NaCl, 1.0% v/v NP-40, 0.5% sodium deoxycholate, 0.1% SDS, and 50 mM Tris pH 8.0) supplemented with a proteinase inhibitor cocktail (Complete; BoehringerMannheim). Tubes were centrifuged intermittently for 30 sec at 7000 g. Washing was repeated three times. After the final centrifugation step, supernatants were removed and 30 l of 1 loading buffer (see below) was added to the pellets. Before electrophoresis (see below) samples were denatured for 2 min at 97 C.

RNA Extraction and Reverse Transcription-Polymerase Chain Reaction
RNA from human and rat tissues was extracted into 15–20 volumes of chaotropic solution as described elsewhere [12]. Oligo(dT)-primed cDNA was synthesized from 2–5 g of total RNA according to the GeneScript protocol (Genecraft, Munster, Germany) in the presence of 0.5 ¨ M betaine (Sigma, Deisenhofen, Germany). Sequences of oligonucleotide primers were 5 -AGACATGAGGCAACGATTGCTCC-3 (the forward primer common to HE2 , HE2 1, and HE2C), 5 -GGCAGGGAGGTTCAACGGAC-3 (the forward primer specific to HE2B and HE2E), 5 GGGATCAGAGCAAATGTCACGC-3 (the reverse primer common to HE2 , HE2 1, HE2B, and HE2E), and 5 -CATCAGTTTTAATGTAAACAGCAGGCGTC-3 (the reverse primer specific to HE2C as described by Frohlich et al. [8]). Reactions were carried out using 1 unit of Bio¨ therme Taq polymerase (Genecraft) per reaction in a touchdown polymerase chain reaction (PCR) program that dropped from 61 C to an annealing temperature of 51 C. Amplicons were separated and visualized on 1% Tris-acetate/EDTA agarose/ethidium bromide gels. Identity of inserts was confirmed by subcloning them into the pGEM-T-Easy plasmid vector (Promega, Mannheim, Germany) and plasmid sequencing by standard procedures.

Western Blot Analysis and N-Terminal Peptide Sequencing
Samples were separated on 15% Laemmli gels or on 16.5% Tris-Tricine gels containing 6 M urea [13] and transferred to polyvinylidene difluoride membranes (Millipore, Eschborn, Germany) in a continuous buffer system using a semidry blotter (Phase, Lubeck, Germany). Molecular ¨ weight standards employed in Laemmli gels were Rainbow low-range molecular weight markers (Amersham-Pharmacia); in Tris-Tricine gels, molecular weight standards (BioRad, Hercules, CA) were used. The protein transfer procedure was modified as described in [14] to efficiently blot

Chemosynthetic Peptides
Peptides were synthesized according to the amino acid sequences deduced from the HE2 cDNAs (GenBank accession numbers X67697, AF168616, AF170797, AF168617, AF168618, AF168619, and

806

VON HORSTEN ET AL.
Parallel control gels were blotted and analyzed as described [19]. The antibiotic dye, crystal violet, served as a positive control. After electrophoresis, the gel was washed for 15–25 min in 10 mM sodium phosphate buffer pH 7.2, and the buffer was changed every 5 min. The pH was controlled using pH strips in the 0–14 range (Macherey & Nagel, Duren, ¨ Germany). As the gel reached pH 7 it was removed and placed on top of the prepared E. coli agar plates. Plates were incubated for 3 h at 37 C. CAU-PAGE gels were then removed and the bacteria-containing agar was overlayed with Luria-Bertani (LB) agar (Difco, Detroit, MI). The plate was incubated overnight at 37 C. Zones of clearing were readily apparent after this time. To perform the CFU assay as described by Harder et al. [20], 3 103 E. coli DH5 (Life Technologies) was incubated with the linear synHE2 and the synHE2 cycl peptide at increasing concentrations in 100 l of 10 mM sodium phosphate buffer (pH 7.4) containing 10% (v/v) LB broth. While the linear HE2 1 was dissolved in sterile water, a mixture of 99.5% dimethyl sulfoxide/0.5% trifluoracetic acid was employed primarily to dissolve the cyclic peptide. The primary stock solutions were then diluted to concentrations that were suitable for antimicrobial assays.

cationic peptide isoforms. Immunodetection of proteins was carried out by standard procedures as described in [15], employing the CL-HRP substrate system (Pierce Chemical Company, Rockford, IL) at a dilution of 1:10 and exposure to x-ray-film (Kodak, Rochester, NY). Specificity of antibody binding was confirmed by comparing it with the corresponding preimmune serum and by competition with the peptide as described in [10]. Corresponding peptide bands were out from Coomassie-stained parallel blots and subjected to N-terminal peptide sequencing (TopLabGmbH, Munchen, Germany). ¨

Recombinant Protein Expression in Escherichia coli
An HE2 1 cDNA fragment [3] lacking the intrinsic signal peptide coding region was subcloned into the pMAL-c2x vector (New England Biolabs, Schwalbach/Ts, Germany) after addition by PCR amplification of a 3 -HindIII restriction site. Oligonucleotide primers were as follows: sense oligomer, 5 -TCTCAAGCCAGACATGTG-3 ; antisense oligomer, 5 -TACTAAAGCTTCTAATTCCCAGTGGAAC-3 . 3 -HindIII-digested 5 -blunt-end fragments were ligated into the XmnI/HindIII-digested vector. The recombinant plasmid was amplified in E. coli DH5 (Life Technologies, Rockville, MD) and verified by sequencing. The cytosolic proteasedeficient E. coli strain ER2508 (New England Biolabs) was transformed and induced for protein expression according to suggestions of the supplier (New England Biolabs). Bacteria were harvested by centrifugation and the cells were lysed. Supernatants were diluted and loaded onto an amylose resin column by gravity flow. The column was washed and the recombinant protein was competitively eluted. Elutes containing recombinant HE2 1-maltose-binding protein (MBP) fusion were dialyzed against deionized water and lyophilized.

RESULTS

RT-PCR Analysis of Human Epididymal HE2 mRNA Variants

Recombinant Protein Expression in Insect Cells
Complementary DNA fragments without intrinsic signal peptide-coding regions were subcloned into the pMelBacB transfer vector (Invitrogen, Leek, Netherlands). BamHI and HindIII restriction sites were added by PCR amplification for insertion. The sense oligomer for the HE2 1 variant was 5 -CAACGGATCCAGACATGAGGCAACGATTGCTC-3 ; the antisense oligomer was 5 -CAACAAGCTTAGATCCCAGATCTGCCATCC3 . Recombinant virions were selected by plaque assay and propagated in Sf9 cells. Virions were used to infect High Five insect cells at a multiplicity of three plaque-forming units per cell, and recombinant peptides were recovered from the culture medium after 96 h of infection. Supernatants containing the HE2 1 peptides were adsorbed to an Oasis MAX mixed-mode anion exchange column (Waters-GmbH, Eschborn, Germany). Fifty microliters of culture supernatant were loaded onto a methanolconditioned 500 mg Oasis MAX column. The column was washed with 5 ml of 5% methanol in 20 mM Tris-HCl pH 8.0, followed by 5 ml of 100% methanol. Seven milliliters of 50% (v/v) 100 mM acetic acid/acetone were used to elute the recombinant peptide. Elutes were heated to 56 C for 30 min to remove excess acetone and lyophilized.

Protease Digestion of Recombinant HE2 1
Recombinant MBP-HE2 1 from E. coli was digested using 1.5 units of recombinant truncated human furin (New England Biolabs) per microgram of fusion protein in 100 mM Hepes pH 7.5 at 25 C, 0.5% Triton-X100, 2 mM CaCl2, and 1 mM -mercaptoethanol reaction buffer for 6 h at 30 C. For the double digest with furin and Factor Xa (New England Biolabs), the digest mix was chilled and incubated overnight at room temperature using the same activity of recombinant furin as above in combination with 1 g of factor Xa per 50 g of MBP-HE2 1. The digests were extracted with StrataClean resin (Stratagene) [16] in order to obtain salt-free samples for subsequent Western blotting.

Previous analyses have described six human HE2 mRNA variants ([3, 7]; GenBank accession numbers X67697, AF168616, AF170797, AF168617, AF168618, AF168619, and AF168620). Three additional variants have been described for Pan troglodytes by Frohlich et al. [8] ¨ (GenBank accession numbers AF263552, AF263553, and AF263555). Based on these sequences, HE2-specific primers were chosen to span at least two exons to be able to distinguish the amplification of contaminating DNA and unspliced precursors (Fig. 1). Amplicons were subcloned and their identity was verified by sequencing (data not shown). In addition to HE2 and HE2 1, three minor PCR products, HE2B, HE2C, and HE2E were identified (Fig. 1), which were homologous to the EP2B, EP2C, and EP2E variants of chimpanzee [8]. Besides HE2 1, the HE2C and HE2E mRNA variants also predicted -defensin-like peptides. A quantitative comparison of the HE2 mRNA variants was difficult to perform from the standard PCR runs, however, analysis of cDNAs from two different patient tissues at nonsaturating conditions suggested that HE2 and HE2 1 transcripts represented major splice variants (Fig. 1). HE2 was also found at low levels in testicular cDNA, however, the minor transcripts were not detected. Comparing the staining intensities of amplicons derived from contaminating genomic DNA, amplification of HE2C fragments appeared to be less efficient (Fig. 1), possibly not reflecting true relative levels of the epididymal HE2C mRNA variant.
Characterization of HE2-Related Peptides in Human Epididymal Tissue and Fluid

In Vitro Antibacterial Assays
Gel overlay assays were performed essentially as described in [17, 18]. E. coli DH5 were grown to log-phase and harvested by centrifugation at 4000 g for 3 min. The pellet was washed once in 10 mM sodium phosphate buffer pH 7.2 and resuspended as described. A volume containing 4 106 bacterial colony forming units (CFUs) was added to 10 ml of 42 C sodium phosphate buffer that contained 1% low electroendosmosis, low sulfate agarose (Metaphor; Biozym, Hamburg, Germany). Bacteria-containing agar was poured to a depth of 2 mm into a Petri dish. Continuous acid urea (CAU)-PAGE of protein samples to be tested for antimicrobial activity was performed as described by Lehrer et al. [18].

A panel of monospecific antibodies against five chemosynthetic peptides (see Materials and Methods) was raised to identify HE2-related antigens in human epididymal tissue and in flushed epididymal secretions. P3 and P4 antibodies specifically directed against HE2 /HE2B and HE2 1/HE2E, respectively (Fig. 2), were chosen to detect HE2-related antigens in cryosections of human epididymis. From our previous in situ transcript hybridization results [6], maximum levels of HE2 expression were expected in the distal caput and proximal corpus regions. Indirect Cy2 immunofluorescence showed that both antibodies reacted specifically with epithelial antigens in this region (Fig. 3). In tissues from two patients with no known history of an-

ANTIMICROBIAL PEPTIDE

807

tiandrogen treatment, P3- and P4-reactive antigens were localized in the apical part of the duct epithelium; a weak labeling of luminal contents was observed in some of the ductal cross-sections. Epididymides from two other patients who had received long-term antiandrogen treatment showed a deviating morphology in that the duct lumen was largely absent. They also showed a deviating pattern of HE2-related antigen expression (Fig. 3). Whereas high levels of P3-reactive antigen appeared to be associated with the apical stereocilia of the duct epithelium, P4 immunoreactivity was weak in these tissues, persisting only in the basal parts of the epithelium (Fig. 3). Luminal fluid flushed from two human epididymides was analyzed by Western blotting by employing antibodies P1–P4 (see Materials and Methods). The P1 antiserum directed against a sequence common to six predicted HE2 peptide isoforms (Fig. 2) detected no antigens in the expected molecular weight range (not shown). The P2 antibody, directed against another peptide sequence common to all HE2 isoforms except HE2B and HE2E (Fig. 2), reacted with multiple peptides in the range of approximately 5 to 8 kDa (Fig. 4). This was in agreement with the masses predicted for HE2 1 and HE2 2, as well as for HE2 1 and HE2 2 (Fig. 2). The most prominent products separated either as a broad band or as a doublet of approximately 8 kDa. The P3 antibody, specific for HE2 1/2 and HE2B (calculated masses of 8.5 and 3.7 kDa, respectively; Fig. 2) detected an antigen of approximately 8 kDa as well as a smaller band, the higher apparent mass being in agreement with that of the HE2 form. Unexpectedly, the P4 antibody, specific for HE2 1 and HE2E (Fig. 2), revealed two close bands of approximately 8 kDa as well (Fig. 4). This apparent mass corresponded to the calculated mass of HE2E (7.2 kDa), but not HE2 1 (12.2 kDa; Fig. 2). Specificity of antibody binding was confirmed by competition experiments employing the corresponding chemosynthetic oligopeptides and custom peptide sequencing of the immunoreactive protein bands in the range of 8 kDa (see below). From our RT-PCR analysis (Fig. 1) the HE2 1 isoform, rather than HE2E, was assumed to represent the predominant P4-reactive -defensin-like isoform. None of the antisera, however, reacted with peptides in the molecular mass range of 12 kDa as predicted from the HE2 1 cDNA sequence after cleavage of the signal peptide ([7]; Fig. 2). To explain this apparent discrepancy, a peptide band was cut in the range of 8 kDa from a parallel blot and prepared for N-terminal amino acid sequencing. The sequences obtained revealed a mixture of HE2-related peptides with different N-termini. These were either RHV[N]HSA, corresponding to the N-terminus of the majority of HE2 isoforms after signal peptide cleavage, or DLLPPRT, corresponding to an internal peptide sequence common to most HE2 isoforms and representing a proprotein convertase cleavage consensus. From this result, limited endoproteolysis of proximal promoter-derived HE2 peptide isoforms, including HE2 1, was suggested, resulting in processed peptides of approximately 8 kDa. Edman sequencing further suggested that the asparagine at position 4 of the RHV[N]HSA sequence was at least partially occupied by N-linked glycans, possibly related to the failure of P1 antibody reactivity on human epididymal fluid (see above). The internal DLLPPRT motif is part of the P2 epitope and is common to all peptide isoforms predicted by transcripts from the proximal HE2 promoter, including HE2 and HE2 1 (see Fig. 1 and Fig. 2), however, excluding HE2B and HE2E. A truncated Nterminus starting at this motif would result from cleavage

FIG. 1. Detection by RT-PCR analysis of HE2 , HE2 1, and three minor HE2-derived mRNAs, HE2B, HE2C, and HE2E (named by Frohlich et al. ¨ [8]) in human epididymal cDNA. a) Separation of amplicons obtained by RT-PCR on ethidium-bromide stained agarose gel. The specificity of the primer pairs employed is indicated above each lane; identity of amplicons was confirmed by sequencing. Two different patient tissues are represented by two sets of human epididymal cDNA data. M indicates a 100base pair (bp) ladder. Lanes 1–3, RT-PCR of human testicular cDNA; 1/ 1 primers detect HE2 (404 bp amplicon). Lanes 4–6 show human epididymal cDNA from patient 1 and lanes 7–9 show human epididymal cDNA from patient 2; in both tissues, HE2 (404 bp) and HE2 1 (328 bp) are predominant (lanes 4 and 7). HE2B (269 bp) and HE2E (193 bp) are separated in lanes 5 and 8; HE2C (354 bp) appears in lanes 6 and 9. Lanes 10–12 are negative controls for each pair of primers. b) Structure of the human HE2 gene locus (modified after Jia et al. [9]) as proven here by RT-PCR, containing eight exons and two promoters, plus a diagrammatic representation of the nine human epididymal HE2 mRNA variants identified to date. Open reading frames are highlighted by gray shading, -defensin domains by dark shading. Primers (positions marked by arrows) were chosen to span at least two exons to discriminate amplicons originating from contaminating genomic DNA by their increased size (larger than the approximately 400-bp HE2 1 fragment).

of the peptide isoforms between arginine35 and aspartate36 by a furin-like endoprotease. To specifically reveal the presence of HE2 in human epididymal fluid and to determine the N-terminal sequence of the secreted product, immunoprecipitation of protein samples was performed by employing the P3 antibody. Separation of the precipitated proteins was performed on SDS gels, the proteins were blotted, and a band in the range of 8 kDa was cut out. The major N-terminal amino acid sequence obtained was again DLLP, corresponding to the Nterminus of a proteolytically processed HE2 peptide. Other HE2-related N-termini, for example, predicted from the HE2E cDNA sequence, were not found. Still, the occurrence in human epididymal fluid of blocked N-termini or still-undetected N-termini of the minor HE2 peptide iso-

808
FIG. 2. Diagrammatic representation of nine human HE2 peptide isoforms as predicted from the cloned cDNAs, and the location of peptide sequences chosen for oligopeptide synthesis and antibody production (P1–P5). HE2 1 and HE2 2 peptides differ only by amino acid exchanges as marked by three dotted lines (compare with Osterhoff et al. [3]). Signal peptidase and proprotein convertase cleavage sites are represented by arrows (↓). Signal peptides are shown as blocks with a white background; mature peptides are shown as light gray shaded blocks; the location of peptide sequences P1–P5 is shown as dark gray shaded blocks; numbers indicate corresponding amino acids in HE2-peptides; and -defensin-like modules are shown as blocks with a dark background. The table shows molecular mass and isoelectric point values as calculated from the peptide sequences of all isoforms.

VON HORSTEN ET AL.

forms (i.e., HE2B and the bin1b-homologous HE2E [Fig. 2]) is not excluded.
Detection of HE2-Related Peptides in Human Ejaculate

LIS extracts and salt wash membrane protein preparations of human ejaculated sperm were analyzed by employing P3 and P4 antibodies (Fig. 2). With the P3 antiserum, only inconsistent labeling in the range of approximately 8 kDa was observed (data not shown), corroborating previous observations that a Western blot detection of HE2 on human ejaculated sperm was difficult to perform [7]. However, using the P4 antiserum directed against isoforms HE2 1, HE2E, or both, a faint band of approximately 8 kDa was detected in LIS extracts, and a prominent band of a similar apparent mass was detected in salt wash preparations of cationic sperm proteins (Fig. 5a). The presence of HE2-related antigens in human seminal plasma was confirmed by immunoprecipitation employing isoform-specific antibodies. Proteins were precipitated by the P2 antiserum and subsequently analyzed by Western blotting using a combination of P2, P3, P4, and P5 antisera (Fig. 5b). The P3 antiserum was negative (data not shown). A prominent 8-kDa band was detected by the other three antisera, probably representing the processed form of HE2 1 (Fig. 2). The P2 antiserum detected additional minor peptides of 6 kDa (Fig. 5b). Because these smaller peptides reacted with neither P4 nor P5 antiserum (Fig. 2), they were assumed to represent alternative, smaller HE2 peptide isoforms rather than degradation products of HE2 1.
Recombinant Expression and Processing of HE2 and HE2 1

bonds. The recHE2 1/MBP fusion protein from E. coli was affinity-purified and processed by protease treatment in vitro. Factor Xa was used to cleave the fusion protein so that no vector-derived residues were attached. To obtain a peptide containing the DLLPP motif at its N-terminus, recombinant human furin was employed, either alone or in combination with factor Xa. Protease treatment of recHE2 1 in vitro and subsequent Western blot analysis confirmed that the peptide was indeed a substrate for furin-like proprotein convertases (Fig. 6a). A structural model for the mature, processed HE2 peptide was proposed (Fig. 7A), starting with the DLLPP motif and containing an intramolecular disulfide bond. The -defensin-like HE2 1 peptide (Fig. 7B) was produced in High Five cells and enriched from the cell culture supernatants by anion exchange chromatography. Recombinantly expressed proteins were analyzed by employing P1, P2, P4, and P5 antisera (Fig. 2). Although multiple bands were detected with all antisera, including P1, the fastest migrating (approximately 8 kDa) band reacted with all antisera except P1 (Fig. 6b). Subsequent N-terminal sequencing of the corresponding peptide band cut from parallel blots revealed the DLLPP motif. From this result we concluded that the 8-kDa band contained the processed HE2 1 propeptide, and that processing of the prepropeptide

An HE2 1 cDNA fragment [3] and an HE2 1 cDNA fragment, both encoding the complete propeptides ([7]; Figs. 1 and 2) were chosen for recombinant protein production. HE2 1 was expressed in E. coli as a maltose-binding protein (MBP) fusion. The cysteine-rich, -defensinlike HE2 1, on the other hand, was expressed in insect cells to allow correct peptide folding and formation of disulfide

FIG. 3. Localization of HE2-related peptides in the human epididymal duct epithelium. Indirect Cy2 immunofluorescence was performed in 8m duct cryosections of human distal caput/proximal corpus epididymis. A and C) Immunofluorescence employing P3 antibody (1:200; HE2 / HE2B-specific) as a primary antibody; E) corresponding preimmune serum (1:200); B and D) indirect Cy2 immunofluorescence employing P4 antibody (1:200; HE2 1/HE2E-specific) as a primary antibody; F) corresponding preimmune serum (1:200). Panels A and B are tissue sections from a patient with no known history of antiandrogen treatment; the apical cytoplasm and the apical stereocilia were strongly immunopositive. Panels C and D are tissue sections from a patient after long-term antiandrogen treatment; the duct lumen has collapsed. Lu, Duct lumen; Ep, duct epithelium. Employing the HE2 /HE2B-specific P3 antibody, apical stereocilia were still strongly immunopositive, whereas the HE2 1/HE2E-specific P4 antibody reacted only with the basal part of the duct epithelium. Bar 50 m.

ANTIMICROBIAL PEPTIDE

809

810

VON HORSTEN ET AL.

FIG. 4. Detection of secreted HE2-related peptides in human epididymal fluid. a) Western blot analysis after standard Laemmli PAGE of flushed epididymal fluid (ef). Lane 1, 4 l of ef reacted with P2 antibody after competition with P2 peptide; lane 2, 4 l of ef reacted with P2 antibody (1:250); lane 3, 8 l of ef reacted with P2 antibody; lane 4, 8 l of ef reacted with pre-P4; lane 5, 8 l of ef with P4 antiserum (1:250). b) Western blot analysis after Tris-Tricine PAGE of flushed ef. Lane 1, 8 l of ef reacted with P3 antiserum (1:250); lane 2, 8 l of ef reacted with P4 antiserum (1:250).

also occurred in vivo by a furin-like endoprotease present in insect cells [21]. By optimizing the multiplicity of infection, most of the recombinant protein products corresponded to the fully processed form (data not shown).
In Vitro Antibacterial Activity of HE2-Derived Peptides

Recombinant HE2 1 protein was enriched from insect cell culture supernatants as processed form and run on acid-

urea gels. Gels were subsequently included in the bacterial gel overlay assay [17, 18] against E. coli DH5 . In the agar overlying the recHE2 1 lane, a bacteria-free, clear zone was observed. Parallel acid-urea gels were processed for Western blot analysis. Employing the P4 antiserum specifically directed against the HE2 1 peptide, an immunoreactive band with a migration pattern corresponding to that exhibiting antibacterial activity was identified (Fig. 8). Proteolytic processing, disulfide bond formation, and subsequent HPLC-purification of the E. coli-derived recHE2 peptide proved difficult at a preparative scale. To provide the larger amounts of pure peptide necessary for in vitro antibacterial assays, HE2 2 was obtained as linear and cyclic chemosynthetic C-terminal peptide fragments, the latter containing an intramolecular disulfide bond (Fig. 7A). Because there is yet no information on the minimal length of an in vivo, bioactive HE2 peptide, the C-terminal part of the sequence common to the HE2 and HE2B peptide isoforms (compare Figs. 2 and 7) was chosen for oligopeptide synthesis. The linear peptide was readily dissolved in water; however, the cyclic form was difficult to dissolve. On acid-urea PAGE gels, the linear and cyclic synHE2 2 peptides showed different migration patterns (Fig. 9A). The linear peptide, despite its elution as a single HPLC peak and its homogenous mass as determined by ESI-MS, showed several bands, which is indicative of oligomer formation. The cyclic peptide, equally pure as determined by HPLC and ESI-MS, showed similar bands in the lower part of the gel, however, a large proportion of molecules migrated as aggregates in the upper part of the gels. In the acid-urea PAGE bacterial gel overlay assay, zones of clearing corresponded to the peptide oligomers present in the lower part of the gels, whereas the aggregates in the upper part showed no killing of E. coli (Fig. 9A). Quantitative CFU assays were performed by employing the HPLC peak-purified, synthetic HE2 2 peptides at increasing con-

FIG. 5. Detection of HE2-related peptides in a human ejaculate. a) Western blot analysis of human sperm membrane protein preparations employing the P4 antibody. Lane 1, 8 g of sperm membrane proteins from LIS extracts; lane 2, 80 g of sperm membrane proteins from LIS extracts reacted with P4 antiserum (1:250); lane 3, 10 g of sperm salt wash proteins with P4 antiserum; lane 4, 40 g of sperm salt wash proteins with P4 antiserum. b) Immunoprecipitation of HE2-related peptides from human seminal plasma. Lane 1, preclearing and detection with P2antiserum, 1:250; lane 2, immunoprecipitation with P2 antiserum and detection with P2 antiserum; lane 3, immunoprecipitation with P2 antiserum and detection with P4 antiserum; lane 4, immunoprecipitation with P2 antiserum and detection with P5 antiserum.

FIG. 6. Proteolytic processing of recombinantly expressed HE2 in vitro and in vivo. a) Furin peptidase cleavage of E. coli-expressed MBP-HE2 1 fusion protein and subsequent Western blot analysis with P3 antiserum (1:500). Lane 1, Factor Xa/furin-double digest of 50 g of MBP-HE2 1; lane 2, furin digest of 50 g of MBP-HE2 1. Furin cleavage of HE2 1 was more effective in the double digest. b) Recombinantly expressed HE2 1 as recovered from approximately 200- l supernatants of transfected High Five insect cell cultures. Proteins were characterized by Western blot analysis with isoform-specific antibodies. Lane 1, detection with P1 antiserum (1:250); lane 2, detection with P4 antiserum (1:250); lane 3, marker proteins.

ANTIMICROBIAL PEPTIDE

811
FIG. 7. Diagrammatic representation of the major human epididymal HE2 peptide isoforms. A) Proposed processing and structure of the processed HE2 peptide compared with the synthetic HE2 peptide fragments employed in the antibacterial assays. Signal peptide is indicated as a block with a white background; the cleaved prodomain is shown as light gray shaded blocks; the processed peptide is shown as blocks with a dark background. Ciphers refer to the numbers of amino acids starting with methionine 1 (1–25, signal peptide; 25–60, prolonged propeptide; 61–103, mature HE2 ; 73–103, synthetic peptide). Amino acid differences between the two allelic forms are indicated; plus and minus symbols mark charged amino acids; the proposed loop structure is stabilized by a cysteine (black bar) and an ionic (black dots) bond. B) Deduced amino acid sequences of processed HE2 1 and HE2C have been aligned with the known human -defensins (hBD1–hBD4) and the -defensin consensus. Dashes have been inserted to improve alignment.

centrations. The results suggested that the minimal concentrations of both peptides necessary to kill 90% of E. coli were 60 g/ml at a sodium concentration of 20 mM (Fig. 9B). This minimal inhibitory peptide concentration may, however, be an underestimate because of the tendency of the peptides to form high molecular weight aggregates that were biologically inactive (Fig. 9A).
DISCUSSION

sins is predicted (Fig. 7B). To corroborate this assumption, we have shown that recombinant HE2 1 containing a defensin-like motif has antibacterial activity. A weak similarity to the defensin propeptide consensus also resides

Immunohistochemistry, Western blot analyses employing isoform specific antibodies, and peptide microsequencing confirmed for the first time the occurrence of multiple human HE2-encoded peptides in the human male genital tract, including the HE2 isoform. We have shown that HE2 and HE2 1 represent the major isoforms produced by the epididymal epithelium, processed at the peptide level, and secreted apically into the duct lumen. These results are different from those in other mammalian species [5, 7, 8], but they are congruent with previous results on mRNA levels in humans [3, 7]. Minor HE2-derived peptides were also detected, but they cannot yet unequivocally be assigned to minor mRNA variants. The presence of HE2B, HE2C, and HE2E transcripts in human epididymal RNA had not been shown before; rather, the existence of corresponding exons in the human HE2 gene had been inferred by extrapolation from the homologous chimpanzee transcripts [8]. Epididymis-restricted expression and sperm binding properties of HE2-derived peptides [3, 7] had suggested that the HE2 gene may be involved in the process of posttesticular sperm maturation. More recent results suggest that it is related to the -defensin gene family [4, 5, 9]. Indeed, in three human variants (i.e., HE2 1, HE2C, and HE2E), the six-cysteine motif reminiscent of the -defen-

FIG. 8. Antibacterial activity of recombinant HE2 1 protein expressed in High Five insect cells. A) CAU-PAGE separation and Coomassie staining of a supernatant from untransfected cultures (lanes are referred to as ) and of HE2 1 peptide enriched from the supernatant of a transfected cell culture (lanes are referred to as 1). B) Subsequent bacterial gel overlay assay of the same gel against E. coli DH5 . Locations of substances with antimicrobial activity in the gel are visualized by bacteria-free, clear zones in the agar. They are accentuated by arrows. C) Western blot analysis of recombinantly expressed HE2 1 run on a parallel acid-urea gel employing the P4 antibody (1:250).

812

VON HORSTEN ET AL.

FIG. 9. Antibacterial activity of synthetic HE2 peptide fragments. A) CAU-PAGE separation and Coomassie staining of C-terminal peptides with and without an intramolecular cysteine bond. Because this separation system did not contain SDS, the polypeptide standard (BioRad) served only as a landmark. M indicates the polypeptide standard. Lane 1, approximately 50 g of synHE2 2cycl; lane 2, 10 g of linear synHE2 ;2; lane 3, crystal violet was included as a positive control. B) Bacterial gel overlay of parallel acid-urea gel containing the same amounts of synthetic HE2 peptide fragments. Antimicrobial activity is visualized by the bacteria-free, clear zones. C) Quantification of antibacterial activity by CFU assay of linear synHE2 2 ( ) and synHE2 2cycl ( ) against E. coli DH5 . Bacteria were incubated for 3 h at 37 C in 100 l of 10 mM sodium phosphate buffer pH 7.2, containing 10% LB broth with increasing concentrations of peptide (20 mM total sodium concentration). Antibacterial activity was measured by plating serial dilutions of each incubation mixture and counting of the colony forming units the following day. The limit of detection (1 colony per plate) corresponded to 1 102 CFU/ml.

within the propeptide sequence predicted from HE2 mRNA. However, its processed isoform, besides being strongly cationic, showed no other reminiscence to any of the known classes of peptide antibiotics. Still, our results show that HE2 , like the -defensin-like HE2 1, has antibacterial activity, indicating that the human gene produces antimicrobial peptides with a sequence that is entirely different from the known defensins. Thus, we may have detected a novel class of antibacterial peptides that had not been observed before in other species or tissues, and which may be involved in defending the epithelia of the male genital tract in humans. Complementary DNA cloning of bin1b, which was proposed as representing the rat HE2 homologue [5], and identification of a homologous mouse EST clone (GenBank accession number AK020333), allows the generation of deficient animals by gene targeting. Targeted gene mutation will certainly help to promote a better understanding of the functional role of the -defensin-like HE2 isoforms in vivo. However, HE2E, which shows closest similarity to bin1b [5], appeared to represent only a minor product in the human epididymis. Other variants, including a counterpart HE2 , have not yet been found in rodents, and it is possible that they do not exist. Still, alternative mRNA splicing is a potent mechanism of expression regulation, and the human HE2 isoforms, which are constitutively expressed only as minor forms, may be induced during infection or inflammation. Our results suggest that in addition to extensive processing at the mRNA level, proteolytic processing of proximal promoter-driven HE2 peptides occurs by a ubiquitous furin-like serine protease, by cleaving a glycosylated 35amino acid propiece from the precursor peptides. Activation by endoproteolytic cleavage has been shown only for the -defensins, which are synthesized as longer preprodefensins [22]. A similar posttranslational processing has not previously been shown for -defensins, but may apply to all HE2 peptides that contain the corresponding proprotein convertase cleavage consensus. Limited endoproteo-

lysis at sites marked by this consensus sequence is a widespread process by which biologically active peptides are produced within the secretory pathway of eukaryotic cells (for a review see [23]). HE2 processing was observed in the same way in human epididymis and in insect cell cultures, and could play a role in regulating cellular transport and peptide function. Additional putative cleavage sites for other endoproteases are present in some of the HE2 isoforms as well. Regionspecific expression in the epididymis of various proteases and their presence in epididymal fluid has been recently demonstrated [24]. Matrilysin, a matrix metalloprotease implicated in -defensin activation [25] is secreted luminally by the epididymis, seminal vesicles, and prostate [26]. It may also cleave prosegments of HE2 peptides, possibly further increasing the number of potentially bioactive HE2derived peptides. Regulation of defensins occurs via signal transduction pathways that are common to other immune responses (for a review see [27]. However, bin1b expression in rat epididymis [28] and HE2 gene expression in nonhuman primates [7, 8] seems to depend on androgens. Therefore, it was an unexpected finding that HE2-related peptides were still present in the epididymides of two patients after long-term antiandrogen treatment. No differences were observed at the mRNA level by Northern blot analysis (data not shown), and it is possible that these patients were not completely androgen-deprived. Still, the duct morphology in both tissues exposed to antiandrogens was dramatically altered compared with untreated tissues. Thus, our observation is interesting in that the -defensin-like peptides (i.e., HE2 1/HE2E), were located only basally in androgen-deprived ducts. Location of the defensin-unrelated peptides (i.e., HE2 /HE2B), on the other hand, was comparable to that in tissues of untreated patients. The human male genital tract is a recognized site of antimicrobial peptide secretion [29], and it also expresses the -defensins hBD-1 [30] and hBD-4 [31]. It may still use a more tissue-specific repertoire of antimicrobial pep-

ANTIMICROBIAL PEPTIDE

813

tides; the most compelling evidence for this is the in vitro antibacterial activity of HE2 and HE2 1 peptides shown here. The antimicrobial activity of HE2 , which shows no similarity to the known classes of peptide antibiotics, was nevertheless surprising. However, the functions and requirements of the epididymis during transport, maturation, and preservation of spermatozoa are different from those of other epithelia. Thus, the in vivo functions of HE2-encoded peptides may be more complex. Specifically, their presence in the ejaculate and on the sperm surface point to additional functions other than epithelial defense. Other antimicrobial peptides also revealed additional functions. Besides their antimicrobial activity, a chemoattractant role for cells of the adaptive immune system has been described [32]. Enteric defensins form anion-conductive pores in phospholipid bilayers and stimulate chloride secretion [33]. Defensins are cytotoxic to both bacterial and normal eukaryotic cells (for a review see [2]). Moreover, CAP37 has been reported to immobilize sperm [34]. Some cases of male idiopathic infertility thus may be caused by endogenous peptide antibiotics, which may bear a contraceptive potential.
ACKNOWLEDGMENTS
We are indebted to Drs. Ching-Hei Yeung and Trevor G. Cooper, Institute of Reproductive Medicine, Munster, for their generous help in pro¨ viding human epididymal tissue and fluid. We thank Drs. Richard Ivell and Caroline Osterhoff, IHF, Hamburg, and Drs. Jens-Michael Schroder ¨ and Jurgen Harder, University of Kiel, for helpful discussions. Dr. F. Lei¨ denberger, IHF, Hamburg, provided excellent facilities and continued interest and support. Dr. Marcus Koppitz, Schering AG, Berlin, synthesized most peptides employed in this study.

REFERENCES
1. Pollanen P, Cooper TG. Immunology of the testicular excurrent ducts. ¨ J Reprod Immunol 1994; 26:167–216. 2. Lehrer RI, Ganz T. Antimicrobial peptides in mammalian and insect host defence. Curr Opin Immunol 1999; 11:23–27. 3. Osterhoff C, Kirchhoff C, Krull N, Ivell R. Molecular cloning and characterization of a novel human sperm antigen (HE2) specifically expressed in the proximal epididymis. Biol Reprod 1994; 50:516–525. 4. Frohlich O, Po C, Young LG. Organization of the human gene en¨ coding the epididymis-specific EP2 protein variants and its relationship to defensin genes. Biol Reprod 2001; 64:1072–1079. 5. Li P, Chan HC, He B, So SC, Chung YW, Shang Q, Zhang YD, Zhang YL. An antimicrobial peptide gene found in the male reproductive system of rats. Science 2001; 291:1783–1785. 6. Krull N, Ivell R, Osterhoff C, Kirchhoff C. Region-specific variation of gene expression in the human epididymis as revealed by in situ hybridization with tissue-specific cDNAs. Mol Reprod Dev 1993; 34: 16–24. 7. Hamil KG, Sivashanmugam P, Richardson RT, Grossman G, Ruben SM, Mohler JL, Petrusz P, O’Rand MG, French FS, Hall SH. HE2beta and HE2gamma, new members of an epididymis-specific family of androgen-regulated proteins in the human. Endocrinology 2000; 141: 1245–1253. 8. Frohlich O, Po C, Murphy T, Young LG. Multiple promoter and splic¨ ing mRNA variants of the epididymis-specific gene EP2. J Androl 2000; 21:421–430. 9. Jia HP, Schutte BC, Schudy A, Linzmeier R, Guthmiller JM, Johnson GK, Tack BF, Mitros JP, Rosenthal A, Ganz T, McCray PB. Discovery of new human beta-defensins using a genomics-based approach. Gene 2001; 263:211–218. 10. Derr P, Yeung CH, Cooper TG, Kirchhoff C. Synthesis and glycosylation of CD52, the major ‘maturation-associated’ antigen of rat spermatozoa, in the cauda epididymidis. Reproduction 2001; 121:435– 446. 11. Alexander NJ, Bearwood D. An immunosorption assay for antibodies to spermatozoa: comparison with agglutination and immobilization tests. Fertil Steril 1984; 41:270–276.

12. Kirchhoff C, Osterhoff C, Young LG. Molecular cloning and characterization of HE1, a major secretory protein of the human epididymis. Biol Reprod 1996; 54:847–856. 13. Schagger H, von Jagow G. Tricine-sodium dodecyl sulfate-polyacryl¨ amide gel electrophoresis for the separation of proteins in the range from 1 to 100 kDa. Anal Biochem 1987; 166:368–379. 14. Szewczyk B, Kozloff LM. A method for the efficient blotting of strongly basic proteins from sodium dodecyl sulfate-polyacrylamide gels to nitrocellulose. Anal Biochem 1985; 150:403–407. 15. Schroter S, Derr P, Conradt HS, Nimtz M, Hale G, Kirchhoff C. Male¨ specific modification of human CD52. J Biol Chem 1999; 274:29862– 29873. 16. Ziegler J, Vogt T, Miersch O, Strack D. Concentration of dilute protein solutions prior to sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Anal Biochem 1997; 250:257–260. 17. Steinberg DA, Lehrer RI. Designer assays for antimicrobial peptides. Disputing the one-size-fits-all theory. Methods Mol Biol 1997; 78: 169–186. 18. Lehrer RI, Rosenman M, Harwig SS, Jackson R, Eisenhauer P. Ultrasensitive assays for endogenous antimicrobial polypeptides. J Immunol Methods 1991; 137:167–173. 19. Wang MS, Pang JS, Selsted ME. Semidry electroblotting of peptides and proteins from acid-urea polyacrylamide gels. Anal Biochem 1997; 253:225–230. 20. Harder J, Bartels J, Christophers E, Schroder JM. Isolation and char¨ acterization of human beta-defensin-3, a novel human inducible peptide antibiotic. J Biol Chem 2001; 276:5707–5713. 21. Veenstra JA. Mono- and dibasic proteolytic cleavage sites in insect neuroendocrine peptide precursors. Arch Insect Biochem Physiol 2000; 43:49–63. 22. Valore EV, Martin E, Harwig SS, Ganz T. Intramolecular inhibition of human defensin HNP-1 by its propiece. J Clin Invest 1996; 97: 1624–1629. 23. Steiner DF. The proprotein convertases. Curr Opin Chem Biol 1998; 2:31–39. 24. Metayer S, Dacheux F, Dacheux JL, Gatti JL. Comparison, characterization, and identification of proteases and protease inhibitors in epididymal fluids from domestic animals. Biol Reprod; 2002; 66:1219– 1229. 25. Wilson CL, Heppner KJ, Rudolph LA, Matrisian LM. The metalloproteinase matrilysin is preferentially expressed by epithelial cells in a tissue-restricted pattern in the mouse. Mol Biol Cell 1995; 6:851– 869. 26. Wilson CL, Ouellette AJ, Satchell DP, Ayabe T, Lopez-Boado YS, Stratman JL, Hultgren SJ, Matrisian LM, Parks WC. Regulation of intestinal alpha-defensin activation by the metalloproteinase matrilysin in innate host defense. Science 1999; 286:113–117. 27. Kaiser V, Diamond G. Expression of mammalian defensin genes. J Leukoc Biol 2000; 68:779–784. 28. Ibrahim NM, Young LG, Frohlich O. Epididymal specificity and an¨ drogen regulation of rat EP2. Biol Reprod 2001; 65:575–580. 29. Malm J, Sorensen O, Persson T, Frohm-Nilsson M, Johansson B, Bjartell A, Lilja H, Stahle-Backdahl M, Borregaard N, Egesten A. The human cationic antimicrobial protein (hCAP-18) is expressed in the epithelium of human epididymis, is present in seminal plasma at high concentrations, and is attached to spermatozoa. Infect Immunol 2000; 68:4297–4302. 30. Valore EV, Park CH, Quayle AJ, Wiles KR, McCray PB, Ganz T. Human beta-defensin-1: an antimicrobial peptide of urogenital tissues. J Clin Invest 1998; 101:1633–1642. 31. Garcia JR, Krause A, Schulz S, Rodriguez-Jimenez FJ, Kluver E, Adermann K, Forssmann U, Frimpong-Boateng A, Bals R, Forssmann WG. Human beta-defensin 4: a novel inducible peptide with a specific salt-sensitive spectrum of antimicrobial activity. FASEB J 2001; 15: 1819–1821. 32. Yang D, Chertov O, Bykovskaia SN, Chen Q, Buffo MJ, Shogan J, Anderson M, Schroder JM, Wang JM, Howard OM, Oppenheim JJ. ¨ Beta-defensins: linking innate and adaptive immunity through dendritic and T cell CCR6. Science 1999; 286:525–528. 33. Lencer WI, Cheung G, Strohmeier GR, Currie MG, Ouellette AJ, Selsted ME, Madara JL. Induction of epithelial chloride secretion by channel-forming cryptdins 2 and 3. Proc Natl Acad Sci U S A 1997; 94:8585–8589. 34. D’Cruz OJ, Pereira HA, Haas GG Jr. Sperm immobilizing activity of a synthetic bioactive peptide 20–44 of 37-kDa cationic antimicrobial protein (CAP37) of human neutrophils. J Androl 1995; 16:432–440.