Inhibition of HIV-1 envelope glycoprotein-mediated cell fusion by a DL-amino acid-containing fusion peptide: possible recognition of the fusion complex.

The N-terminal fusion peptide (FP) of human immunodeficiency virus-1 (HIV-1) is a potent inhibitor of cell-cell fusion, possibly because of its ability to recognize the corresponding segments inside the fusion complex within the membrane. Here we show that a fusion peptide in which the highly conserved Ile(4), Phe(8), Phe(11), and Ala(14) were replaced by their d-enantiomers (IFFA) is a potent inhibitor of cell-cell fusion. Fourier transform infrared spectroscopy confirmed that despite these drastic modifications, the peptide preserved most of its structure within the membrane. Fluorescence energy transfer studies demonstrated that the diastereomeric peptide interacted with the wild type FP, suggesting this segment as the target site for inhibition of membrane fusion. This is further supported by the similar localization of the wild type and IFFA FPs to microdomains in T cells and the preferred partitioning into ordered regions within sphingomyelin/phosphatidyl-choline/cholesterol giant vesicles. These studies provide insight into the mechanism of molecular recognition within the membrane milieu and may serve in designing novel HIV entry inhibitors.

Specific fusion proteins, located on the surface of viral membranes, mediate membrane fusion (1). The envelope glycoprotein gp160 from HIV, 1 containing two non-covalently associated subunits, gp120 and gp41 (2), mediates the membrane fusion activity of the virus. The binding of the gp120 subunit to target cell receptors (3-7) induces a conformational change in the glycoprotein, which results in the exposure of a previously hidden hydrophobic N-terminal stretch of gp41, designated the "fusion peptide" (8,9). Evidence supporting the role of the FP domain in mediating membrane fusion came from studies with intact envelope proteins (10 -12), as well as synthetic peptides used in model and biological systems (13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25). However, the molecular mechanism of membrane fusion is still poorly understood despite extensive studies done in both biological and model systems. Nevertheless, the results of many studies suggest common motifs for the diverse biological and model fusion reactions (13,26).
Virus-induced membrane fusion is highly sensitive to single amino acid mutations in the FP domain (27,28). This can lead to the conclusion that the structure of FPs is a crucial parameter in the fusion process. Indeed, several studies suggest that the structure of FPs plays a major role in their activities (12,29). To distinguish between the effects of structure and hydrophobicity, we compared an all-L-amino acid FP with its enantiomer composed entirely of D-amino acids (25). Both had the same activity in liposome fusion assays, although having mirror image structures (25). Furthermore, the FP of HIV-1 was shown to inhibit cell-cell fusion (30 -32). Cumulative evidence suggests that the mechanism of inhibition is through interaction with the corresponding region in the intact gp41 (23,25). This inhibitory activity of the peptide is chirality-independent, eliminating the possibility of interaction at the receptor level since chirality is crucial for recognition of soluble proteins (25). The chirality independence of membrane-inserted peptide-peptide interaction in vivo has been recently verified by using a glycophorin A transmembrane domain (33).
The non-chiral nature of the FP-mediated inhibition of HIV-1 cell fusion reinforces the need to understand the role of the precise structure of the FP in the fusion process. To alter the structure of the FP while preserving the hydrophobicity (i.e. same amino acid composition), we introduced D-amino acids into the FP of HIV-1. For this purpose, we synthesized two 33-residue peptides, WT, which corresponds to the all-L sequence of the HIV-1 FP, and a diastereomer analogue. In the latter, termed IFFA, 4 highly conserved amino acids, Ile 4 , Phe 8 , Phe 11, and Ala 14 , were replaced by their D-enantiomers. The ability of IFFA to inhibit HIV-mediated cell-cell fusion was investigated. The results were correlated with structural studies performed by using ATR-FTIR and fluorescence spectroscopy. The observed functional and structural properties of the peptides shed light on the molecular recognition process within the membrane milieu. Moreover, the ability of the diastereomer to maintain the inhibition properties of the WT FP makes it a promising candidate for designing new anti-HIV inhibitors.
Peptide Synthesis and Fluorescent Labeling-The peptides were synthesized by a solid phase method on phenylacetamidomethyl-amino acid resin (0.15 moles/gram), as described previously (23,30). Labeling of the N terminus of the peptides was achieved as described previously (35,36). The synthetic peptides were purified (Ͼ95% homogeneity) by reverse-phase high pressure liquid chromatography on a C 18 column using a linear gradient of 25-80% acetonitrile in 0.05% trifluoroacetyl for 40 min. The peptides were subjected to amino acid analysis and mass spectrometry to confirm their composition. Unless stated otherwise, stock solutions of concentrated peptides in Me 2 SO were used to avoid aggregation of the peptides prior to their use. The final concentration of Me 2 SO in each experiment had no effect on the system under investigation.
HIV-mediated Cell-Cell Fusion Assay-A ␤-galactosidase assay (37) was used to measure the inhibitory effect of the peptides. TF228 and SupT1 cells were infected overnight at 10 multiplicity of infection with different recombinant vaccinia viruses. TF228 cells were transfected with vCB4R containing the T7 promoter and lacZ gene, which expresses ␤-galactosidase enzyme from Escherichia coli. SupT1 cells were transfected with vTF7-3, having T7 RNA polymerase under the control of the natural P7.5 early-late vaccinia virus promoter. Fusion was analyzed by detecting ␤-galactosidase activity. Since the cytoplasm of SupT1 cells contains T7 RNA polymerase and the cytoplasm of TF228 cells contains the ␤-galactosidase enzyme under the control of the T7 promoter, the enzyme was produced only in cytoplasmically fused cells. Cells were mixed and incubated for 3 h at 37°C in the presence of peptides. Then cells were lysed with detergent, and ␤-galactosidase activity was detected colorimetrically, using the substrate chlorophenol red-␤-galactopyranoside. The rate of hydrolysis was measured by reading absorbance at 595 nm. The assay was repeated in the presence of different concentrations of the FP, its IFFA diastereomer, and several other control peptides. Inhibition of the fusion process resulted in a reduction of ␤-galactosidase activity. There was about 1% fusion (␤-galactosidase expression) in the controls: target cells without HIV receptors or cells expressing uncleaved gp160 (37).
Fusion Peptide Localization on T Cells-Activated T cells (39) (10 4 ) were incubated for 30 min at room temperature with NBD-labeled WT or IFFA FPs (0.5 M final concentration). The cells were washed twice with phosphate-buffered saline (100 l) to remove excess unbound peptide. The cells were then observed under a fluorescent confocal microscope. NBD excitation was set at 488 nm with the laser set at 2% power to prevent bleaching of the fluorophore. Fluorescence data were collected from 525 nm and higher.
Vesicle Preparation-Giant unilamellar vesicles (GUV) were prepared as described previously by Moscho et al. (40). Briefly, a dried film of lipids containing a total of 2 mg of PC:SM:CH:Rho-PE (1:1:1:0.001) was dissolved in 30 l of chloroform. The lipids were added to a round flask containing 970 l of chloroform and 150 l of methanol. Next, we carefully added 7 ml of double-distilled H 2 O along the flask walls. The mixture was evaporated under nitrogen flow (2 bar) while rotating the flask at 30 -50 rpm for 5-10 min. Small unilamellar vesicles (SUV) were prepared by sonication from PC:CH (10:1 w/w). The cholesterol was included to reduce the curvature of the small unilamellar vesicles (23).
Fusion Peptide Localization in Model Membranes-GUV labeled with Rho-PE were observed under a confocal microscope with excitation set at 488 and 543 nm; emission was collected from 505 to 525 nm and from 610 nm and higher. Unordered regions were observed with excitation set at 543 nm but not with excitation at 488 nm at 610 nm and up. No signal was detected at 505-525 nm. The NBD-labeled FP and IFFA were added to the GUV, carefully, at a final concentration of 0.5 M. Binding of the peptides to the GUV was followed until maximal binding was observed after ϳ10 -15 min for PC:SM:CH and 1 h for PC:CH. To eliminate the possibility of quenching the NBD fluorophore due to FRET with the rhodamine, we bleached the rhodamine using the 543-nm laser at 100% intensity for 30 s. No change in NBD emission at 505-525 nm was observed, whereas the rhodamine emission at 610 nm was abolished.
Peptide Binding to SUV-The degree of peptide association with lipid vesicles was measured by adding lipid vesicles to 0.1 M NBD-labeled peptides at 28°C, as has been described previously (41). The fluorescence intensity was measured as a function of the lipid:peptide molar ratio, with excitation set at 467 nm (10-nm slit) and emission set at 530 nm (5-nm slit).
Resonance Energy Transfer Measurements-Fluorescence resonance energy transfer was measured using NBD-labeled peptides serving as donors and with Rho-labeled peptides serving as energy acceptors (42,43). Fluorescence spectra were obtained at room temperature with excitation set at 467 nm using a 10-nm slit width. In a typical experiment, donor peptide (final concentration 0.04 M) was added to a dispersion of PC:CH. (10:1 w/w) SUV (400 M) in phosphate-buffered saline followed by the addition of acceptor peptide in several sequential doses. Fluorescence spectra were obtained before and after the addition of the acceptor. The efficiency of energy transfer (E) was determined as described previously by Kliger et al. (23).
Attenuated Total Reflectance-Fourier Transform Infrared (ATR-FTIR)-Experiments were performed as described previously (44). Briefly, 1 mg of lipids mixed with chloroform, WT FP, or IFFA (1:50) was spread on the prism and dried under vacuum for 30 min. The sample was also tested under hydrated conditions by deuteration.

RESULTS
A peptide, representing the N-terminal 33-amino-acid segment of gp41 of HIV-1 (LAV1a strain), WT, and its diastereomer analogue IFFA, were synthesized and labeled at their N-terminal amino acid with either NBD or rhodamine fluorescent probes. The sequences of the peptides and their designations are shown in Table I. The secondary structure of the FP at the site of the substitutions is a ␤-sheet (44). The rationale behind the diastereomer is that the D-amino acids will destabilize the structure of the FP (the side chains will switch sides) while maintaining the original sequence. We investigated the peptides for their ability to inhibit HIV-1-mediated cellcell fusion. The results were correlated with structural information.
Inhibition of Cell-Cell Fusion Induced by WT and IFFA-The peptides were tested for their inhibitory effect on HIV-1-mediated cell fusion as described under "Experimental Procedures." The WT peptide and its diastereomer exhibited marked inhibitory effects at concentrations Ͻ10 ng/ml. The data also suggest that IFFA is slightly more active than the WT (Fig. 1A). We used a 13-mer peptide corresponding to the N-terminal region of the FP, as well as the FP of the Sendai virus, as negative controls. We further investigated the inhibitory effect of the peptide when added to the ENV-expressing cells prior to adding target cells. An inhibitory effect was observed, similar to that seen in Fig. 1A, when the peptide was added to the mixture of cells. Thus, we assume that the peptide may associate with the surface of the cells and remain there until the fusion process is initiated. At

HIV-1 Inhibition by DL-Amino Acids Fusion Peptide
that point, it inhibits the fusion process. Alternatively, the FP may bind directly to the ENV complex. However, this is less likely since the wild type FP is presumably hidden within the gp120/gp41 complex until receptor binding. The inhibition of HIV ENV-mediated cell fusion by FP was demonstrated to be HIV-specific. We studied the virus specificity of the IFFA diastereomer in inhibiting viral fusion. Although the IFFA is a potent inhibitor of HIV-1, it could not inhibit influenza HA-mediated cell fusion at concentrations up to 3 M (Fig. 1B). This result confirms that the inhibitory effect of IFFA is specific to HIV, similar to the WT FP. Furthermore, neither the WT FP nor the IFFA could reach complete inhibition of cell-mediated fusion. The inhibition increases fast at the lower nM range, but even up to 1 M, it cannot pass the 75% inhibition. This observation may be accounted for by the nature of the FP itself. At high concentrations, the FP increases negative curvature and promotes membrane destabilization (23). Thus, it is possible that as the concentrations are raised, the FP and IFFA start to promote fusion as well as inhibit it.
Binding of WT and IFFA to the Surface of the T Cells-We investigated the binding of the WT FP and IFFA to the surface of T cells. Such binding supports a model in which the inhibition observed can be through direct interaction of the FPs with protein complexes on or within the T cell membranes. NBDlabeled WT and IFFA were incubated for 15 min with activated T cells and then washed three times with phosphate-buffered saline. Binding of the peptides to T cells and their localization was investigated by confocal fluorescence microscopy. Both the WT and the IFFA demonstrated high affinity toward T cells (Fig. 2). The peptides were localized at the plasma membrane and formed patches of high intensity. The results suggest that both the WT and the IFFA preferentially bind to certain domains on the surface of T cells (Fig. 3). When sliding across the z axis, the peptides obviously do not penetrate the cytoplasmic membrane, and in fact, no peptide was observed within any other membrane compartment in the cells.
Binding of IFFA to Unordered Regions on GUVs-We prepared sphingomyelin:phosphatidylcholine:cholestrol:phosphatidyl-ethanolamine-rhodamine (SM:PC:CH:Rho-PE) (1:1:1: 0.001) giant unilamellar vesicles (GUVs) labeled with rhodamine, as described under "Experimental Procedures," and observed them under a confocal microscope (45). Rho-PE is known to partition into unordered regions. Indeed, the fluorescence of Rho-PE was located in patches, indicating that it segregated out of the more ordered SM/CH regions and into the less ordered PC regions (Fig. 4A). We then added the NBDlabeled IFFA. The peptide partitioned into ordered regions of the GUV, as demonstrated in Fig. 4B by the segregation of the green (NBD-IFFA) and red (Rho-PE) colors, suggesting an intrinsic affinity of the peptides for SM/CH domains. We verified that there was no NBD quenching due to FRET between the fluorophores by bleaching the rhodamine. The fluorescence of the NBD-labeled peptide remained constant, even when the rhodamine signal was completely abolished. Similar results were obtained with the WT FP (data not shown). Thus, we can conclude that the IFFA diastereomer partitions preferentially into ordered regions, similarly to the WT FP. A similar experiment with PC:CH:Rho-PE (10:1:0.001) GUVs, with no microdomains present, demonstrated a uniform binding pattern for the IFFA (Fig. 4C). The absence of ordered regions is demonstrated by the uniform distribution of the Rho-PE over the surface of the GUVs. However, the time to reach maximal intensity was about four times slower than that observed for SM:PC:CH GUVs.
Co-assembly of WT Fusion Peptide and Its Diastereomer in the Membrane-bound State-The self-association of the peptides in their membrane-bound state was monitored by measuring resonance energy transfer. A dose-dependent quenching of the emission of the NBD-WT donor, consistent with energy transfer, was observed when Rho-WT or Rho-IFFA acceptor (final concentration of 0.025-0.35 M) was added to a mixture of NBD-WT (0.05 M) and SUV (400 M). The energy transfer was calculated and plotted versus the acceptor:lipid molar ratio (Fig. 5). The acceptor peptide was added only after the donor peptide was already bound to the membrane, thus decreasing association in solution. The free energy of peptide association with SUV was measured as described under "Experimental Procedures." It was found to be Ϫ9.3 kcal/mole for the WT FP and Ϫ8.3 kcal/mole for the IFFA. Accordingly, the lipid:peptide molar ratio in these experiments was kept high enough to ensure low surface density of donors and acceptors while reducing the energy transfer between unassociated peptide monomers to a minimum. To confirm that the observed energy transfer is due to peptide oligomerization, the transfer efficiencies observed in the experiments were compared with the energy transfer expected for randomly distributed membranebound donors and acceptors (Fig. 5, dashed line). The random distribution was calculated as described earlier (42), using 51 Å as the R 0 value for the NBD/Rho donor/acceptor pair (43). The levels of energy transfer between the different pairs are significantly higher than those expected for a random distribution of donors and acceptors.
Secondary Structure of the Peptides as Determined by ATR-FTIR Spectroscopy-We studied the FTIR spectra of the amide I region of the WT and IFFA peptides in the presence of PC:CH

HIV-1 Inhibition by DL-Amino Acids Fusion Peptide
multibilayers. The samples were exposed to D 2 O vapors followed by hydration until equilibrium was achieved. The amide I frequencies are known to decrease by up to 10 cm Ϫ1 upon H/D exchange. The peaks corresponding to random coil and ␣-helix partly overlap, but random coils undergo H/D exchange at higher rates than ␣-helices. Thus, hydration with D 2 O vapors can enable better distinction between the two components. The contributions of the various secondary structure elements to the amide I peak were obtained by using PEAKFIT (Jandel Scientific, San Rafael, CA) and comparison with accepted values from the literature (Fig. 6) (46, 47). The WT and IFFA both have mainly a ␤-sheet secondary structure, represented by peaks within the 1620 -1640 cm Ϫ1 range (Table II). Deconvolution of the IR spectra resulted with 78 and 70% ␤-sheet content for WT and IFFA, respectively. A second peak, located at ϳ1648 cm Ϫ1 , most likely represents a random coil segment (47). A peak at around 1670 cm Ϫ1 may represent either a high frequency helix or a ␤-turn. The area of that peak is 12 and 11% for WT and IFFA, respectively. These results fit well with recent FTIR (44), as well as NMR studies of the FP (48). The studies demonstrated that the FP can associate strongly within a membrane environment and adopt both parallel or antiparallel ␤-sheet structures. This is in contrast to the FTIR by Gordon et al. (49), which demonstrated an ␣-helical structure for the FP at low concentrations and a ␤-sheet structure only at high loading. However, they used a shorter 23-residue peptide.
The amide II band is even more susceptible to D 2 O hydration. Upon H/D exchange, its frequency shifts by about 100 cm Ϫ1 . After exposure of the samples to D 2 O vapors, a strong effect was detected for the diastereomer peptide, whereas a smaller shift was observed for the WT (data not shown). These results indicate that D-amino acid substitutions destabilize the secondary structure of the peptide in the membrane-bound state to a certain extent. Although the overall structure of IFFA is maintained, there is higher flexibility, and thus, lower aggregation and an increase in the exposure to D 2 O. It is interesting to note that the peaks corresponding to the ␤-sheet components of the diastereomer are wider than those of the WT. This phenomenon was observed previously for diastereomers with primarily ␣-helix structures (50).
Orientation of the Lipid Multibilayers and the Effect of the Peptides on Lipid Order-The symmetric (v sym ϳ2850 cm Ϫ1 ) and antisymmetric (v antisym ϳ2920 cm Ϫ1 ) vibrations of the lipid methylene C-H bond are perpendicular to the molecular axis of a fully extended hydrocarbon chain. Thus, measurements of the infrared absorbance dichroism indicate the order and the orientation of the lipid sample relative to the prism surface. The R values for the lipid and lipid/peptide samples were calculated from the stronger v antisym (CH 2 ) peaks. The results show that the lipid samples were well ordered and the peptides induced similar changes in the ordering of the acyl chains; lipids alone had an R ratio of 1.14, lipids mixed with WT FP had an R ratio of 1.29, and lipids mixed with IFFA had an R ratio of 1.27 when deuterated. This suggests a similar penetration of the peptides into the membrane bilayer. DISCUSSION Recently, a peptide named T20/Fuzeon/enfuvirtide corresponding to the C-terminal helical region of gp41 was approved by the FDA for use as a fusion inhibitor (51). This peptide highlighted a new family of inhibitors that target the viral entry into the host. The FP of HIV gp41 has similar properties to the Fuzeon in that it targets a transient triggered state of gp41. It was previously shown to inhibit HIV infection in vitro with EC 50 concentrations in the nM range (23). This inhibition proved to be chirality-independent when an all-D-analogue of the FP demonstrated similar antiviral activity (25). Nevertheless, the FP inhibitory activity is sequence-specific; the HIV-1 FP cannot inhibit the activity of Sendai virus (data not shown), and the FP of Sendai virus could not inhibit HIV-1 ENVmediated cell fusion (23). Furthermore, no cross-reactivity was observed between HIV-1 FP and HIV-2 ENV-mediated cell fusion (25).
We synthesized a diastereomer analogue of the WT FP with 4 residues in a highly conserved region replaced by D-amino acids. The motivation behind the IFFA diastereomer synthesis was to verify the importance of the precise structure of the FP when compared with its inhibitory potential. The major secondary structure of the WT FP is a ␤-sheet at the sites of the substitutions. Introduction of D-amino acids will result in a stretch of amino acids with their side chains facing the same direction instead of alternating. This is bound to cause some steric hindrance and thus affect the structure. Nevertheless, the IFFA diastereomer was able to maintain a high inhibitory activity against HIV ENV-mediated cell-cell fusion. This inhibition is specific to HIV since the IFFA could not inhibit influenza HA-mediated fusion (Fig. 1B).
Interestingly, we found that the overall ATR-FTIR spectra of both the WT and the IFFA FPs were similar. The main secondary  structure of both peptides was a ␤-sheet. This suggests that the hydrophobic environment of the surroundings can facilitate a "wild type"-like ␤-sheet structure. This is comparable with a recent report showing a similar structure of an ␣-helical peptide and its diastereomer (52,53). Deconvolution of the spectra revealed some differences in ␤-structure contributions between the FP and its IFFA diastereomer. Specifically, the IFFA displayed a lower aggregated ␤-sheet peak at 1620 cm Ϫ1 and a higher ␤-sheet peak at 1636 cm Ϫ1 when compared with the WT FP. This difference suggests a lower aggregation state for the IFFA diastereomer, which does not interfere with its inhibitory activity. Furthermore, a broadening of the ␤-structure peaks of IFFA is reminiscent of the ␣-helical broadening that is commonly observed with short antimicrobial diastereomer peptides. This phenomenon is usually accompanied by the higher solubility of the peptides (50,52,54), which was also observed for the IFFA (data not shown). The localization of both peptides seems to be mainly on the surface of the bilayers since both exhibited only a minor disturbance in the lipid order. This is in agreement with the current accepted model, which places the FP in an oblique orientation on the surface of the membrane (55)(56)(57)(58)(59)(60). Imaging of the WT and IFFA, in the presence of T cells, revealed that both peptides were localized to the plasma membrane. Moreover, the pattern of patches formed around the membrane for both peptides suggests that they localize within specific regions of the membrane. We further demonstrated the localization of both peptides to ordered regions on PC:SM:CH (1:1:1) GUVs in vitro. This is consistent with previous studies of a shorter segment of the WT FP (16 amino acids) that showed a similar pattern of binding to the cytoplasmic membrane of T cells, suggesting perhaps localization on lipid rafts (61). Note also that a recent study demonstrated that raft localization of influenza HA is crucial for fusion (62).
The IFFA diastereomer was able to inhibit HIV-1-mediated cell-cell fusion (Fig. 1A). This inhibition is similar to that reported for the WT FP (23). Previously, it was shown that the inhibitory action of the FP and its derivatives does not occur at the receptor level and most likely results from the ability of these peptides to associate with the FP domain of gp41 (23,25). Peptides from other regions of gp41 were also shown to exert their inhibitory activity by association with the gp41 domains (63). In our study, the diastereomer derivative retained the ability to associate with the WT FP in the membrane-bound state in vitro (Fig. 5). Furthermore, the binding pattern on the surface of T cells is in agreement with the mechanism proposed above for the inhibition of cell-cell fusion. The fusion peptide may assemble with itself after membrane insertion, resulting in the high intensity regions observed. We believe that within the membrane, the structure of the diastereomer is similar enough to the wild type to allow the necessary wild type interaction to be conserved. Thus, the functionality of the peptide is preserved, although the inhibition is sequence-specific, as reported previously (23,25).
We believe that the IFFA diastereomer inhibits fusion through the following pathway. The cell surface contains negatively charged components, such as sialic acid, heparan sulfate or other sugars. Binding to heparan sulfate was demonstrated previously for a short fragment of the WT HIV FP. This initial binding promoted a localized increase in concentration, which drives the insertion into the membrane in a cooperative manner (61). The hydrophobic nature of the membrane induces secondary structure similar to the WT peptide despite the D-amino acid modifications. In turn, this is followed by the assembly of IFFA with the WT protein, which may create an inactive fusion complex.
The need for widening the antiviral, especially anti-HIV, arsenal is great. Fuzeon addresses just such a need. It inhibits a new step in the virus life cycle, and thus, it has a synergistic effect with the customary anti-HIV cocktails. However, aggregation and susceptibility to proteolytic cleavage are the main drawbacks of peptide inhibitors derived from membrane proteins. The short biological half-life of the Fuzeon peptide demonstrates this problem (several shots of large doses are needed per day). We demonstrated that the solubility of the FP can be increased by incorporating D-amino acids while retaining full inhibitory activity within the membrane. Furthermore, altering the chirality of membrane-inserted peptides is not as important as in aqueous solution, merely resulting in minor modifications to the secondary structure. This makes the diastereomer peptide inhibitors attractive alternatives to the regular all-L-peptide inhibitors. Moreover, placing the D-amino acids in strategic positions can increase proteolytic resistance. This, in turn, might improve the pharmacological properties of the peptides.