SignalP, a computational tool available on the world-wide web [22] was used to predict the amino terminal signal peptide cleavage sites of HBD-19, HBD-23, HBD-27 and HBD-29. SignalP predictions were compared with known mature β-defensin peptide sequences to predict which amino terminal variants were most likely to yield biologically active β-defensins. Similarly, the mature carboxy termini of HBD-19, HBD-23, HBD-27 and HBD-29 were predicted by comparison with the known mature β-defensin peptide sequences.

HBD-23, HBD-27, and HBD-29 were chemically synthesized by Biosynthesis Inc. (Louisville, TX, USA) and HBD-19 was synthesized and kindly provided by Drs. Saskia Milton and Charles Glabe (University of California, Irvine). After reducing 4 mg of synthetic defensins with 1 µg/mL dithiothreitol (DTT) at 55 °C for 4 h, the peptides were passed through Sep-Pak cartridges with a C-8 sorbent, washed, and eluted with acetonitrile plus 0.1% trifluoroacetic acid. Peptides were dried overnight and subsequently dissolved in 0.1 M ammonium carbonate buffer supplemented with 1 mM cysteine and 1 mM cystine under nitrogen gas [23]. The mixtures were stirred slowly at 25 °C, allowing the peptides to refold. Peptides were examined every 24 h by C-18 reverse-phase HPLC to assay the presence of foldamers. Peptides were further examined by reducing and non-reducing SDS-PAGE to assay the presence of multimers. Each lane was loaded with 3 µg of peptide. Gels were stained with Coomassie blue dye and destained in (5% (v/v) methanol, 7% (v/v) acetic acid, 88% H2O. After seven days peptides were found to be >95% monomeric, single foldamers.

The colony forming unit (CFU) reduction assay was performed using Escherichia coli ML35 (provided by Dr. Andre Ouellette, University of Southern California), Staphylococcus aureus ATCC 25923, and Pseudomonas aeruginosa ATCC 27853 as described by Dr. Ouellette and colleagues [23]. Briefly, various concentrations of defensins were mixed with 1 × 10⁶ to 5 × 10⁶ bacteria in the logarithmic growth phase in 200 µL of 10 mM PIPES [piperazine-N,N’-bis (2-ethanesulfonic acid), pH 7.4] buffer and 1% (vol/vol) trypticase soy broth. These samples were incubated at 37 °C for 4 h. After this incubation, samples were diluted 100-fold with cold 10 mM PIPES and plated in triplicate on Luria broth agar. The plates were then incubated for 12 to 16 h at 37 °C and the number of colonies was recorded. In order to assess the effect of salt and a reducing agent on defensin activity, the PIPES buffer was supplemented with either 150 mM NaCl or 2 mM dithiothreitol (DTT). As a positive bactericidal control, 20 mM lysozyme was utilized. Statistical tests were performed with Student’s two-tailed t-test.

Human peripheral blood mononuclear cells (PBMCs) were purified from normal donor blood by centrifugation over Histopaque 1077 (Sigma) and 10⁵ cells were cultured in 100 μL of media per well in 96-well plates. RPMI-1640 medium supplemented with 10% fetal bovine serum, 50 μg/mL gentamicin and 20 units/mL IL-2 was used. PBMC were incubated for three days with various concentrations of refolded or crude defensins. Thereafter, an MTS based metabolic assay (Promega) was utilized to examine the cytotoxicity of the defensins. The assay was conducted according to the manufacturer’s specifications.

The SignalP software tool was used to predict the amino terminal signal peptide cleavage sites of HBD-19, HBD-23, HBD-27 and HBD-29. Subsequently, these predictions were compared with the known mature peptide sequences of HBD-1, HBD-2, HBD-3 and HBD-4 to predict which amino terminal variants were most likely to yield biologically active β-defensins (Figure 1). Similarly, the mature carboxy termini of HBD-19, HBD-23, HBD-27 and HBD-29 were predicted by comparison with the known mature β-defensin peptide sequences. The resulting predicted mature HBD-19, HBD-23, HBD-27 and HBD-29 peptides contain the core 6-cysteine motif preceded by four to ten amino-terminal amino acid residues and followed by four to six carboxy terminal residues. The predicted mature amino terminal residue of HBD-19 and HBD-23 is glycine, which is also found at the amino terminus of HBD-1, HBD-2 and HBD-3. The predicted mature amino terminal residue of HBD-27 is glutamic acid like HBD-4 and the predicted mature amino terminal residue of HBD-29 is valine. All four of the predicted mature human β-defensins terminate with a positively charged amino acid, as do HBD-1 and HBD-3.

After reducing the peptides and allowing them to slowly refold, we analyzed their structural homogeneity by comparing the folded and unfolded forms on C-18 reverse-phase HPLC. We also analyzed the folded forms by SDS-PAGE under non-reducing (NR) and reducing conditions (R) to detect the presence of multimers. More than 95% of each peptide refolded to a single monomeric foldamer as indicated by the near uniform mobility of each peptide on C-18 reverse-phase HPLC (data not shown) and in SDS-PAGE under reducing and non-reducing conditions (Figure 2).

The antibacterial activity of each peptide, for E. coli, S. aureus and P. aeruginosa was evaluated using the reduction of colony forming unit (CFU) assay. Each peptide was tested at concentrations ranging from 3 to 100 μg/mL in a low ionic strength buffer. Lysozyme was used as a positive control for antibacterial activity. Moreover, each peptide was tested at one concentration in a high ionic strength solution and separately, in the presence of a reducing agent in order to probe the structural requirements of their antibacterial activities.

HBD-19 had antibacterial activity against E. coli ML35, and S. aureus ATCC 25923, but little against P. aeruginosa ATCC 27853 (Figure 3). At 100 μg/mL HBD-19 reduced the number of E. coli colonies 134-fold, the number of S. aureus colonies 20-fold and the number of P. aeruginosa colonies almost 10-fold. Its activity towards E. coli was significantly reduced when 150 mM NaCl (p = 0.001) or 2 mM DTT (p = 1.4 × 10⁻⁸) was added. Only DTT significantly attenuated the anti-S. aureus activity of HBD-19 (p = 4.5 × 10⁻⁵). In contrast, neither 150 mM NaCl nor 2 mM DTT had a significant effect on E. coli colony numbers alone. Both agents inhibited S. aureus colony formation significantly alone, but the inhibition was less than 2.5 fold in both cases so it did not obscure the attenuation of HBD-19 antibacterial activity against S. aureus. Neither 150 mM NaCl nor 2 mM DTT had a significant effect on P. aeruginosa colony numbers alone, p = 0.11 and p = 0.31 respectively, but paradoxically both agents potentiated HBD-19’s activity against P. aeruginosa colony formation, albeit modestly, 4.7 and 8.1 fold respectively.

Among the four defensins tested, HBD-23 exerted the strongest antibacterial activity (Figure 4). At 100 μg/mL, HBD-23 reduced the number of E. coli colonies by nearly five logs (5.6 × 10⁴-fold), the number of S. aureus colonies 185-fold and the number of P. aeruginosa colonies by more than five logs (1.4 × 10⁵-fold). The anti-E. coli activity of 30 μg/mL HBD-23 was decreased under physiological salt conditions (p = 0.004) and enhanced by 2 mM DTT (p = 7.7 × 10⁻⁵). As stated above, neither agent affected E. coli colony formation alone. The addition of 2 mM DTT did not significantly affect the anti-S. aureus activity of HBD-23 but addition of 150 mM NaCl attenuated this activity (p = 7.6 × 10⁻⁶). As above for HBD-19, the modest inhibition of S. aureus colony formation mediated by NaCl alone did not mask its ten-fold attenuation of the anti-S. aureus activity of HBD-23. Both NaCl and DTT significantly inhibited the anti-P. aeruginosa activity of HBD-23 (p = 0.007 and 0.001 respectively), but as stated above neither agent significantly affected P. aeruginosa colony formation alone.

Unlike HBD-23, HBD-27 showed little antibacterial activity against E. coli or P. aeruginosa (Figure 5). HBD-27 was weakly active, however, against S. aureus, but only at high concentrations. At 100 μg/mL HBD-27 reduced the number of P. aeruginosa colonies by nearly 22-fold. This modest antibacterial activity of HBD-27 against P. aeruginosa was attenuated by the addition of 150 mM NaCl (p = 0.001) or by the addition of 2 mM DTT (p = 0.02).

Lastly, HBD-29 exhibited antibacterial activity against E. coli and P. aeruginosa, but had little to no activity against S. aureus (Figure 6). At 100 μg/mL HBD-29 reduced the number of E. coli colonies 112-fold and the number of P. aeruginosa colonies 26-fold. The antibacterial activity of HBD-29 against E. coli was attenuated by NaCl (p = 0.0006), but not by DTT. In contrast, its activity against P. aeruginosa was blocked by DTT (p = 0.002), but not by NaCl.

Previous studies reported that defensins interact electrostatically with bacterial membranes and thereby destroy their structure [3,12,14]. Thus, it is important to know whether these peptides can similarly interfere with human cellular membranes. The cellular toxicity of HBD-19, HBD-23, HBD-27, and HBD-29, both in their crude and folded states, was investigated using a colorimetric metabolic-activity assay on human peripheral blood mononuclear cells (PBMC; Figure 7).

Folded HBD-19 showed a low level of cellular toxicity that was concentration dependent, reaching 20% at 30 μg/mL. Crude HBD-19, however, had little effect on PBMC viability. Likewise, crude and folded HBD-23 and HBD-27 demonstrated little or no cytotoxicity. In fact, folded HBD-27 was somewhat stimulatory, particularly at 30 μg/mL. In contrast, folded HBD-29 exhibited concentration dependent toxicity for PBMC; up to 23% at 30 μg/mL, while crude HBD-29 stimulated PBMC metabolism slightly.

This study examined the antibacterial properties of four human β-defensins—HBD-19, 23, 27, and 29—that are normally expressed, at least at the RNA level, in the male reproductive tract [19,20,21]. Because the epididymis is anatomically continuous with the urethra, host defense against ascending bacterial pathogens is pertinent in the protection of spermatozoa from bacterial contamination. Previous studies showed HBD-1 and HBD-4 are also expressed in the urinary tract and are known to be microbicidal [7,24].

The results of this study showed HBD-23 to be the most potent antibacterial agent among the four defensins tested. However, we found the antibacterial activity of HBD-23 was reduced in the presence of salt, an effect that has been documented by Shimoda et al, who observed that ultrastructural changes on the S. aureus surface caused by defensins were prevented at high-salt concentration [14]. This supports the hypothesis that HBD-23, like other previously studied antimicrobial defensins, permeabilizes the bacterial cell membrane by an electrostatic charge-based mechanism [12,25,26]. Antimicrobial peptides can also damage eukaryotic cells in the same manner [3]. Our data revealed that refolded HBD-19 and HBD-29 decreased cellular viability in a dose-dependent manner. However, HBD-23, and HBD-27 demonstrated low levels of cytotoxicity on human peripheral blood mononuclear cells, which is in agreement with some other defensins demonstrating low levels of hemolytic activities [16]. The presence of cholesterol and lipid-anchored proteins may interfere with defensin binding, preventing host damage [27,28].

We found that HBD-23 killed E. coli more efficiently in vitro when its disulfide bridges were reduced. This may be due to the positive charges being more exposed to anionic bacterial membrane and thereby killing at a faster rate. Nevertheless, it remains to be determined whether these disulfide bridges have a functional role in vivo, as one study reported that they stabilize defensins against enzymatic digestion [29].

Interestingly, Hwang et al showed that certain defensins are more efficient in killing bacteria when more positive charges are present within the cysteine core region [30]. This may explain why HBD-23 was the most potent as it contained the most net positive charges within the six cysteine motif, followed by HBD-29, HBD-19, and finally HBD-27, which showed the least amount of killing. These observations were consistent with the data presented by Rodríguez-Jiménez, who found that synthetic HBD-23 and HBD-29 had broad spectrums of antibacterial activity, while HBD-27 did not [20].

The human β-defensins studied here may also have other roles in immune defense including antiviral activity like HBD-2 and HBD-3 [4,31,32]. Moreover, multiple studies have reported that β-defensins recruit monocytes, T-cells, immature dendritic cells and activated neutrophils by acting as chemokines that interact with the chemokine receptors CCR2 and CCR6 [15,16,17,18,33]. Human β-defensins 1–4 activate keratinocytes to produce IL-18 and HBD-3 also activates macrophages and myeloid dendritic cells via the toll-like receptors (TLR) 1 and 2 [34,35]. Further investigation of HBD-19, 23, 27 and 29 may therefore shed more light on the intricate functions of human immunity.