D-form KLKLLLLLKLK-NH2 peptide exerts higher antimicrobial properties than its L-form counterpart via an association with bacterial cell wall components

The antimicrobial peptide KLKLLLLLKLK-NH2 was developed based on sapesin B, and synthesized using D-amino acids. Biochemical properties of the D-form and L-form KLKLLLLLKLK-NH2 peptides were compared. In order to limit the effects due to bacterial resistance to proteolysis, antimicrobial activities of the peptides were evaluated after short-term exposure to bacteria. D-form KLKLLLLLKLK-NH2 exhibited higher antimicrobial activities than L-form KLKLLLLLKLK-NH2 against bacteria, including Staphylococcus aureus and Escherichia coli. In contrast, both the D-form and L-form of other antimicrobial peptides, including Mastoparan M and Temporin A, exhibited similar antimicrobial activities. Both the D-form KLKLLLLLKLK-NH2 and L-form KLKLLLLLKLK-NH2 peptides preferentially disrupted S. aureus-mimetic liposomes over mammalian-mimetic liposomes. Furthermore, the D-form KLKLLLLLKLK-NH2 increased the membrane permeability of S. aureus more than the L-form KLKLLLLLKLK-NH2. Thus suggesting that the enhanced antimicrobial activity of the D-form was likely due to its interaction with bacterial cell wall components. S. aureus peptidoglycan preferentially inhibited the antimicrobial activity of the D-form KLKLLLLLKLK-NH2 relative to the L-form. Furthermore, the D-form KLKLLLLLKLK-NH2 showed higher affinity for S. aureus peptidoglycan than the L-form. Taken together, these results indicate that the D-form KLKLLLLLKLK-NH2 peptide has higher antimicrobial activity than the L-form via a specific association with bacterial cell wall components, including peptidoglycan.

Scientific REPORTs | 7:43384 | DOI: 10.1038/srep43384 Sapecin B is an antimicrobial peptide that was originally isolated from the culture medium of an embryonic cell line, NIH-Sape-4, derived from Sarcophaga peregrine (flesh fly). It displays potent activity against Gram-positive bacteria 13 . Two other related proteins, sapecin and sapecin C, were also isolated from culture medium of NIH-Sape-4 [13][14][15] . Sapecin B has significant sequence similarity to a scorpion venom toxin, charybdotoxin 13,16 . Structural comparison of sapecin B and charybdotoxin identified the undecapeptide RSLCLLHCRLK-NH 2 , which corresponds to amino acid residues 7 to 17 of sapecin B with C-terminal amidation 16,17 . The peptide fragment RSLCLLHCRLK-NH 2 showed significant antimicrobial activity, suggesting that this region is responsible for the antimicrobial activity of the peptide 17 . The undecapeptide KLKLLLLLKLK-NH 2 was developed by modifying the primary structure of RSLCLLHCRLK-NH 2 . In addition to its activity against Gram-positive bacteria, Gram-negative bacteria, and fungi 18 , KLKLLLLLKLK-NH 2 has been shown to enhance mammalian immune responses via undefined molecular mechanisms [19][20][21] . The antimicrobial activity of the D-form KLKLLLLLKLK-NH 2 , which was synthesized using D-amino acids, persisted longer than the L-form because of its resistance to proteolytic degradation 18 .
In this study, we examined the antimicrobial properties of D-form KLKLLLLLKLK-NH 2 . D-form KLKLLLLLKLK-NH 2 displays higher antimicrobial activity against bacteria than its L-form; however, this elevated activity could not be explained by resistance to proteolytic degradation. It is important to note that other D-form antimicrobial peptides did not show higher antimicrobial activity than their L-form counterparts. Furthermore, D-form KLKLLLLLKLK-NH 2 showed higher affinity for bacterial cell wall components, such as peptidoglycan, than its L-form. Thus, the enhanced antimicrobial activity of the D-form KLKLLLLLKLK-NH 2 relative to its L-form is due to direct interactions with bacterial cell surface components.

Results
MICs of D-form KLKLLLLLKLK-NH 2 were lower than those of L-form KLKLLLLLKLK-NH 2 .
Previously, D-form KLKLLLLLKLK-NH 2 was shown to persist longer in bacterial culture medium and it showed higher antimicrobial activity to Staphylococcus aureus than the L-form 18 . In order to further examine the antimicrobial properties of D-form KLKLLLLLKLK-NH 2 , we determined the MICs of the peptides against S. aureus, Escherichia coli, and Candida albicans. MICs of D-form KLKLLLLLKLK-NH 2 were lower than those of its L-form, especially against S. aureus where the MIC of the D-form was 16-fold lower than the L-form (Table 1). We determined minimum inhibitory concentrations (MICs) of other antimicrobial peptides, including KLKLLLKLK-NH 2 , a derivative of KLKLLLLLKLK-NH 2 18 , FIKRIARLLRKIF-NH 2 (Kn2-7) derived from Buthus martensii scorpion venom 22 , INLKAIAALAKKLL-NH 2 (Mastoparan M) derived from hornet venom 23 , and FLPLIGRVLSGIL-NH 2 (Temporin A) derived from Rana temporariareference 24 against S. aureus. All of these peptides are expected to form a helical structure similar to KLKLLLLLKLK-NH 2 16,17,22-24 . MIC of D-form KLKLLLKLK-NH 2 against S. aureus is more than 32-fold lower than that of the L-form (Table 2). In contrast, the MIC of D-forms and L-forms of Mastoparan M, Kn2-7, and Temporin A against S. aureus (Table 2) were similar. These observations indicate that KLKLLLLLKLK-NH 2 and its related peptide KLKLLLKLK-NH 2 are unique because these D-form peptides display lower MICs against S. aureus than their L-forms.   peptide was not observed (Fig. 1g). This observation suggests that the higher antimicrobial activity of D-form KLKLLLLLKLK-NH 2 was not due to its resistance to proteolytic degradation. In addition, in order to exclude the possibility that bovine serum albumin or some components from culture medium specifically affect the antimicrobial activity of KLKLLLLLKLK-NH 2 , we performed experiments without culture medium and/or bovine serum albumin in the assay mixture. D-form KLKLLLLLKLK-NH 2 also showed higher antimicrobial activity to S. aureus than L-form KLKLLLLLKLK-NH 2 in the absence of culture medium and/or bovine serum albumin (Fig. 1f). It is noteworthy that antimicrobial activity of both L-form and D-form peptide in the absence of culture medium and bovine serum albumin were lower than those in our standard assay condition ( Fig. 1a and f). The antimicrobial activity of D-form KLKLLLKLK-NH 2 was also higher than that of its L-form counterpart (Fig. 2a).
In contrast, the D-forms and L-forms of Kn2-7, Mastoparan M, and Temporin A peptides displayed similar antimicrobial activities against S. aureus ( Fig. 2b-d). These results indicate that KLKLLLLLKLK-NH 2 and its derivative KLKLLLKLK-NH 2 are unique in that their D-forms have antimicrobial activities than their L-forms.
D-form KLKLLLLLKLK-NH 2 increased bacterial membrane permeability. Cationic antimicrobial peptides bind to the negatively charged bacterial surface and penetrate into the bacterial membrane. Therefore, their effects on bacterial membrane permeability closely correlate with antimicrobial activity. Effects of KLKLLLLLKLK-NH 2 and Mastoparan M on membrane permeability of S. aureus were monitored by ethidium bromide influx rates. As shown in Fig. 3a, both D-form KLKLLLLLKLK-NH 2 (20 μ g/ml) and L-form KLKLLLLLKLK-NH 2 (20 μ g/ml) increased ethidium bromide influx rates; however, the rates were higher in response to D-form KLKLLLLLKLK-NH 2 than the L-form KLKLLLLLKLK-NH 2 . In contrast, D-form and L-form Mastoparan M (20 μ g/ml) increased ethidium bromide influx rates to a similar extent (20 μ g/ml) (Fig. 3b). These observations are consistent with the findings that the antimicrobial activity of D-form KLKLLLLLKLK-NH 2 against S. aureus was higher than that of its L-form KLKLLLLLKLK-NH 2 (Fig. 1a). However, that antimicrobial activity of D-form Mastoparan M against S. aureus was similar with that of its L-form (Fig. 2c).

S. aureus peptidoglycan and E. coli lipopolysaccharide preferentially inhibited the antimicrobial activity of D-form KLKLLLLLKLK-NH 2 .
Most cationic antimicrobial peptides interact with bacterial membranes. Previously, sapecin was shown to have a high affinity for cardiolipin 25 . This observation encouraged us to examine whether D-form KLKLLLLLKLK-NH 2 specifically disrupts liposomes that mimic the cellular membrane of S. aureus. Both D-form and L-form KLKLLLLLKLK-NH 2 released calcein from S. aureus-mimetic liposomes 17,26 , which consisted of phosphatidylglycerol and cardiolipin (Fig. 4a). On the other hand, neither D-form nor L-form KLKLLLLLKLK-NH 2 was able to release calcein from mammalian-mimetic liposomes 27 that consisted of phosphatidylcholine, phosphatidylethanolamine, and cholesterol ( Fig. 4a). Mammalian-mimetic liposomes demonstrated similar sensitivity to Triton X-100 as S. aureus-mimetic liposomes, excluding the possibility that mammalian-mimetic liposome are resistant to chemical treatments (Fig. 4b). These observations indicate that both D-form and L-form KLKLLLLLKLK-NH 2 preferentially disrupt S. aureus-mimetic liposomes, which likely contributes to the antimicrobial activity of KLKLLLLLKLK-NH 2 . Thus, the ability to disrupt S. aureus-mimetic liposomes is not the cause of higher antimicrobial activity of D-form KLKLLLLLKLK-NH 2 relative to its L-form.
To identify a specific target of D-form KLKLLLLLKLK-NH 2 , we analyzed whether bacterial cell wall components were able to inhibit the antimicrobial activities. A comparison of the antimicrobial activities of D-form and L-form KLKLLLLLKLK-NH 2 revealed that 1.9 μ g/ml of D-form and 7.5 μ g/ml of L-form displayed similar antimicrobial activity to S. aureus. The antimicrobial effect of D-form KLKLLLLLKLK-NH 2 was almost inhibited by 40 μ g/ml of S. aureus peptidoglycan, but the same concentration failed to abrogate the antimicrobial  S. aureus-type (S. aureus) and mammalian-type (Mammalian) liposomes containing calcein were exposed to Triton X-100. The amount of calcein that leaked from the liposomes was measured using a spectrofluorophotometer and normalized to determine the % release relative to 0.1% Triton X-100. The error bars represent the mean ± standard deviations from triplicate assays.
activity of L-form (Fig. 5a). These observations highlight the potential for a specific interaction between D-form KLKLLLLLKLK-NH 2 and peptidoglycan. In order to exclude the possibility that some contaminants, such as proteases, in peptidoglycan samples might affect the inhibitory effects, heat-treated peptidoglycan was used for the analysis. As shown in Fig. 5f, heat-treated peptidoglycan showed similar inhibitory effects on the antimicrobial activities with those of untreated peptidoglycan. To further confirm that peptidoglycan is a specific target of D-form KLKLLLLLKLK-NH 2 , the antimicrobial effects were investigated in the presence of lysozyme-digested Antimicrobial activities of D-form KLKLLLLLKLK-NH 2 (1.9 μ g/ml) and L-form KLKLLLLLKLK-NH 2 (7.5 μ g/ml) against S. aureus were examined in the presence of the indicated concentrations of peptidoglycan from S. aureus (a), lipopolysaccharide from E. coli (b), lipid A (c), lipoteicoic acid from S. aureus (d), and peptidoglycan from E. coli (e). (f) Antimicrobial activities of D-form KLKLLLLLKLK-NH 2 (1.9 μ g/ml) and L-form KLKLLLLLKLK-NH 2 (7.5 μ g/ml) against S. aureus were examined in the absence or presence of peptidoglycan (40 μ g/ml) or heat-treated peptidoglycan (40 μ g/ml) from S. aureus. (g) Antimicrobial activities of D-form KLKLLLLLKLK-NH 2 (2.0 μ g/ml) against S. aureus were examined in the absence or presence of 40 μ g/ml of peptidoglycan treated with lysozyme (digested peptidoglycan), 40 μ g/ml of peptidoglycan treated without lysozyme (peptidoglycan), or control buffer treated with lysozyme (lysozyme). Antimicrobial activities of D-form and L-form peptides of Kn2-7 (6.25 μ g/ml) (h) or Mastoparan M (8 μ g/ml) (i) against S. aureus were examined in the presence of the indicated concentrations of peptidoglycan from S. aureus. Gray bars and white bars represent CFUs in assay mixtures treated with D-form and L-form peptides, respectively. Black bars represent CFU in assay mixtures treated without peptide. The error bars represent the mean ± standard deviations from triplicate plates. Concentrations of dimethyl sulfoxide in the assay mixtures were 0.15%. peptidoglycans (Fig. 5g). The D-form did not show an inhibitory effect on antimicrobial activity. The antimicrobial activity of D-form KLKLLLLLKLK-NH 2 was preferentially inhibited by lipopolysaccharide prepared from E. coli (Fig. 5b). Furthermore, antimicrobial activity of D-form KLKLLLLLKLK-NH 2 was also preferentially inhibited by synthetic E. coli lipid A, a membrane anchor region of lipopolysaccharide (Fig. 5c). In contrast, lipoteichoic acid prepared from S. aureus inhibited the antimicrobial effect of both D-form and L-form peptides similarly, indicating that the inhibitory effect was not specific for the D-form peptide (Fig. 5d). Peptidoglycan prepared from E. coli had a weak inhibitory effect on the antimicrobial activity of D-form and L-form peptides (Fig. 5e). Taken together, these observations indicate that some cell surface components, such as S. aureus peptidoglycan, preferentially associate with D-form KLKLLLLLKLK-NH 2 rather than its L-form. Moreover, this preferential association accounts for higher antimicrobial activity of D-form KLKLLLLLKLK-NH 2 than that of the L-form.
In addition, inhibitory effects of peptidoglycan against Kn2-7 and Mastoparan M were examined. As shown in Fig. 5h, peptidoglycan shows significant inhibitory effects against the antimicrobial activities of both D-forms and L-forms of Kn2-7. Furthermore, peptidoglycan shows weak inhibitory effects to antimicrobial activities of Mastoparan M, and the inhibitory effect was not specific for the D-form peptide (Fig. 5i). D-form KLKLLLLLKLK-NH 2 showed higher affinity for S. aureus peptidoglycan than L-form KLKLLLLLKLK-NH 2 . The inhibitory effect of S. aureus peptidoglycan on the antimicrobial activity of D-form KLKLLLLLKLK-NH 2 suggested a specific interaction between these two molecules. To determine whether there was a direct association, direct binding between KLKLLLLLKLK-NH 2 and S. aureus peptidoglycan was examined. Biotin-labeled D-form or L-form KLKLLLLLKLK-NH 2 was added to multi-well plates that were coated with immobilized S. aureus peptidoglycan. Binding of biotin-labeled peptides to the D-form or L-form KLKLLLLLKLK-NH 2 was quantified using avidin-labeled peroxidase. As shown in Fig. 6, D-form KLKLLLLLKLK-NH 2 has a higher affinity for S. aureus peptidoglycan than the L-form counterpart.

Discussion
Incorporation of D-amino acids into antimicrobial peptides has been shown to improve their therapeutic efficacy; however, little is known about how the underlying mechanisms make them distinct from their L-form counterparts (reviewed in ref. 28). In this study we found that D-form KLKLLLLLKLK-NH 2 showed higher antimicrobial activity against both Gram-positive and Gram-negative bacteria, including S. aureus and E. coli, relative to its L-form counterpart. Moreover, the enhanced antimicrobial activity of the D-form was not due to its resistance to proteolytic degradation. D-form KLKLLLLLKLK-NH 2 showed higher affinity for S. aureus peptidoglycan than the L-form counterpart. Peptidoglycan and lipopolysaccharide prepared from S. aureus and E. coli, respectively, selectively inhibited the antimicrobial activities of D-form KLKLLLLLKLK-NH 2 . Thus, specific interactions between D-form peptides and components of the bacterial cell wall may contribute to its elevated antimicrobial activity.
Cationic antimicrobial peptides target the negatively charged cell surface of microorganisms. In some cases, D-forms of naturally-occurring antimicrobial peptides have antimicrobial activities similar to those of L-form counterparts, and it is believed that the interaction between antimicrobial peptide and microbial cell surface is not due to specific, close interactions 10,12 . This general notion is consistent with our observations of similar antimicrobial activities of the D-forms and L-forms of Mastoparan M, Kn2-7, and Temporin A. In addition, D-form KLKLLLLLKLK-NH 2 showed similar activity to disrupt S. aureus-mimetic liposomes when compared to the L-form. These observations indicate that the interaction between antimicrobial peptides and anionic bacterial-type liposomes does not require close contact based on the structure, but charge-based interactions are important for antimicrobial activities. In contrast to the previous studies, our results showed that D-form KLKLLLLLKLK-NH 2 had a higher affinity for some cell surface compounds than its L-form counterpart, and the affinity of the D-form for bacterial surface components contributed to its antimicrobial activity. Our observations indicated that some specific, close contact between antimicrobial peptides and bacterial cell surface components increase antimicrobial activities in addition to charge-based contact. Peptidoglycan is consisted of sugars and peptides, and they are chiral components. The chiral portions of peptidoglycan might be involved in the association of D-form KLKLLLLLKLK-NH 2 . It is noteworthy that high affinity of D-form KLKLLLLLKLK-NH 2 to cell surface components including peptidoglycan does not necessary indicate direct targeting. There might be mechanisms to facilitate peptide transfer to the plasma membrane, which determine the effective concentration.
Comparison of the D-4Leu and L-4Leu antimicrobial peptides revealed that the D-form had a greater tendency to bind to the biofilm exopolysaccharide alginate 29 . This current study of KLKLLLLLKLK-NH 2 largely recapitulated these findings. To date, the molecular basis for the close interaction of D-form peptides with bacterial cell surface components remains unknown; however, the importance of precise structures of the bacterial molecules involved in these interactions has been shown. Antimicrobial activities of D-form KLKLLLLLKLK-NH 2 were preferentially inhibited by S. aureus peptidoglycan but not by E. coli peptidoglycan. This difference is likely based on the structural differences between S. aureus peptidoglycan and E. coli peptidoglycan.
Based on our observations, replacement of all L-amino acids with D-amino acids in an antimicrobial peptide may introduce structural changes that are beneficial for antimicrobial activity. It is important to note that not all antimicrobial peptides have distinct activities based on whether they are expressed as a D-form or L-form, and the number of these peptides may be fairly low. Future studies should focus on elucidating the specific interactions of the D-form modification with bacteria as well as the molecular basis underlying this this phenomenon. This will aid in the development of peptide therapeutics.

Methods
Reagents and antimicrobial peptides. Dimethyl sulfoxide, bovine serum albumin (fraction V), cardiolipin, L-α -phosphatidyl-DL-glycerol, peptidoglycan purified from S. aureus, lysozyme, and lipoteichoic acid purified from S. aureus were purchased from Sigma-Aldrich. Cholesterol, 2-dioleoyl-sn-glycero-3-phosphocholine, and 2-dioleoyl-sn-glycero-3-phosphoethanolamine were purchased from Avanti Polar Lipids Inc. Peptidoglycan purified from E. coli was purchased from InvivoGen. Calcein was purchased from Dojindo. Triton X-100 was purchased from Thermo Fisher Scientific. Lipopolysaccharide purified from E. coli 0111:B4 was purchased from List Biological Laboratories, Inc. Synthetic lipid A was purchased from Peptide Institute Inc. Ruby protein gel stain and Any kD TM precast polyacrylamide gels were purchased from Bio-Rad.
Antimicrobial peptides and biotin-labeled antimicrobial peptides were commercially synthesized by Hayashi Kasei, Thermo Fisher Scientific, and the Toray Research Center. C-terminals of the synthetic peptides were modified by amidation. All peptides were initially suspended in dimethyl sulfoxide.

Determination of MIC.
Bacterial suspensions in Muller-Hinton II medium were adjusted to an optical density of 550 nm (OD 550 ) = 0.0011. C. albicans suspensions in YM medium were adjusted to OD 650 = 0.033. Peptides were serially diluted in 10 mM phosphate buffer (pH 6.0) containing 130 mM sodium chloride, 0.2% bovine serum albumin, and 2.56% dimethyl sulfoxide. The peptide solution (100 μ l) was mixed with 100 μ l of bacteria or C. albicans suspensions. Bacterial cultures were incubated for one day at 37 °C. C. albicans cultures were incubated for two days at room temperature. Cell growth was monitored optically and the MIC was determined.
Assay for antimicrobial activity. Bacteria and C. albicans were suspended in growth medium. Peptides suspended in dimethyl sulfoxide were serially diluted in 10 mM phosphate buffer (pH 6.0) containing 130 mM sodium chloride, 0.2% bovine serum albumin as described previously 13 . Concentrations of dimethyl sulfoxide in the assay mixtures are indicated in the figure legends. In order to examine the effects of bovine serum albumin on the assay, 10 mM phosphate buffer (pH 6.0) containing 130 mM sodium chloride, was used for the dilution of peptides. Peptide solution (500 μ l) was added to 500 μ l of bacteria suspensions and then the mixture was incubated at 37 °C for 10 min. In order to examine the effects of culture medium components on the assay, bacteria suspension was prepared with 10 mM phosphate buffer (pH 6.0) containing 130 mM sodium chloride. Alternatively, 500 μ l of peptide solution was added to 500 μ l of C. albicans suspensions and the mixture was incubated at room temperature for 10 min. The inhibitory effects of the bacterial components were analyzed by incubating 450 μ l of S. aureus suspension with 500 μ l of peptide solution plus 50 μ l of inhibitor samples at 37 °C for 10 min. Then, the peptide/ bacteria suspensions were diluted and plated onto LB agar, LB agar containing 0.5% glucose, or YM agar. After cultivation of the plates, colony forming units (CFU) in the peptide/bacteria suspension were calculated based on the average of triplicate plates. Assay for membrane permeability. To examine membrane permeability, ethidium influx rates were examined as previously described 30,31 . S. aureus suspension cultures were adjusted to an OD 600 of 0.4 in 10 mM phosphate buffer (pH 6.0) containing 130 mM sodium chloride and 0.2% bovine serum albumin. Then, peptide in dimethyl sulfoxide (8 μ l) or dimethyl sulfoxide alone (8 μ l) was added to 2 ml of S. aureus suspensions. At 30 sec after the addition of peptide, ethidium bromide was added to a final concentration of 5 μ g/ml, and fluorescence of the ethidium-nucleic acid complex was monitored using a RF-5300PC spectrofluorometer (Shimadzu). Excitation and emission wavelengths were 545 nm with 5 nm slits and 600 nm with 10 nm slits, respectively.
Liposome suspensions were prepared by diluting of 1 μ l of liposomes into 40 ml of 10 mM phosphate buffer (pH6.0) containing 130 mM sodium chloride. Peptides were serially diluted in 10 mM phosphate buffer (pH 6.0) containing 130 mM sodium chloride and 1% dimethyl sulfoxide. Peptide samples (20 μ l) were added to 2 ml of liposome suspension, and the mixtures were incubated at room temperature for 10 min. Calcein leakage from the liposomes was examined using a RF-5300PC spectrofluorometer. Excitation and emission wavelengths were 490 nm and 520 nm (with a 5 nm slit width), respectively 32 .
Digestion and heat-inactivation of peptidoglycan. Peptidoglycan (120 μ g) prepared from S. aureus was added to 1 mg/ml of lysozyme in phosphate buffered saline (150 μ l) 33 . Peptidoglycan without lysozyme and lysozyme without peptidoglycan were also prepared as controls. Samples were incubated overnight at 37 °C, and then incubated at 100 °C for 15 min to inactivate lysozyme. For heat-inactivation of peptidoglycan, 600 μ g of peptidoglycan suspended in water (300 μ l) was incubated at 100 °C for 15 min. The samples were sonicated for 10 sec at setting 1 using a Branson sonifier model S-150D. These samples were used as inhibitor samples for antimicrobial activity assays.
Peptidoglycan-binding assay. Peptidoglycan-binding assays were performed as previously described with some modifications [34][35][36] . Peptidoglycan from S. aureus (100 μ g/ml) was suspended in 0.2% trifluoroacetic acid and sonicated twice for 10 sec at setting 1 using a Branson sonifier model S-150D. The peptidoglycan suspension (50 μ l) was used to coat the wells of a flat bottom 96-well microplate (Thermo Fisher Scientific). The plate was incubated at room temperature until the water evaporated. The plate was placed at 60 °C for 1 h to dry out completely, and then blocked with 200 μ l of 5 mg/ml bovine serum albumin in binding buffer (10 mM phosphate buffer (pH 6.0) containing 130 mM sodium chloride, 0.05% Tween 20, and 0.01% trifluoroacetic acid) at 37 °C for 2 h. The plate was washed four times with 200 μ l of binding buffer. Biotin-labeled peptides in 100 μ l of binding buffer containing 0.5% dimethyl sulfoxide were added to the wells and incubated at 37 °C for 2 h. Detection of biotin-labeled peptides was performed using Vectastain ABC reagent (Vector Laboratories) according to manufacturer's instructions. The wells were washed four times with binding buffer, then 100 μ l of avidin-labeled peroxidase was added to each well, and the plate was incubated at 37 °C for 1 h. The wells were washed again as described above. After washing, 100 μ l of 3, 3′ , 5, 5′ -tetramethylbenzide substrate was added and the plate was incubated at room temperature. After 10 min, the reaction was stopped by the addition of 100 μ l of 0.5 M sulfuric acid. Absorbance was measured at 450 nm.