Available online at www.sciencedirect.com
Biochimie 91 (2009) 141e154 www.elsevier.com/locate/biochi
Research paper
Novel lactoferrampin antimicrobial peptides derived from human lactoferrin Evan F. Haney a, Kamran Nazmi b, Fanny Lau a, Jan G.M. Bolscher b, Hans J. Vogel a,* b
a Structural Biology Research Group, Department of Biological Sciences, University of Calgary, Calgary, Alberta, Canada T2N 1N4 Department of Oral Biochemistry, Academic Centre for Dentistry Amsterdam (ACTA), VU University, 1081 BT Amsterdam, Netherlands
Received 21 March 2008; accepted 30 April 2008 Available online 15 May 2008
Abstract Human lactoferrampin is a novel antimicrobial peptide found in the cationic N-terminal lobe of the iron-binding human lactoferrin protein. The amino acid sequence that directly corresponds to the previously characterized bovine lactoferrin-derived lactoferrampin peptide is inactive on its own (WNLLRQAQEKFGKDKSP, residues 269e285). However, by increasing the net positive charge near the C-terminal end of human lactoferrampin, a significant increase in its antibacterial and Candidacidal activity was obtained. Conversely, the addition of an N-terminal helix cap (sequence DAI) did not have any appreciable effect on the antibacterial or antifungal activity of human lactoferrampin peptides, even though it markedly influenced that of bovine lactoferrampin. The solution structure of five human lactoferrampin variants was determined in SDS micelles and all of the structures display a well-defined amphipathic N-terminal helix and a flexible cationic C-terminus. Differential scanning calorimetry studies indicate that this peptide is capable of inserting into the hydrophobic core of a membrane, while fluorescence spectroscopy results suggest that a hydrophobic patch encompassing the single Trp and Phe residues as well as Leu, Ile and Ala side chains mediates the interaction between the peptide and the hydrophobic core of a phospholipid bilayer. Ó 2008 Elsevier Masson SAS. All rights reserved. Keywords: Antimicrobial peptide; Human lactoferrin; Lactoferrampin; Nuclear magnetic resonance; Micelle-bound peptide structure
1. Introduction Antimicrobial peptides are currently being considered as a potential alternative for conventional antibiotics because of their prevalence throughout nature [1,2]. Moreover, as they often have non-specific modes of action [3,4], it is thought that most pathogenic organisms are unable to develop resistance to the effects of antimicrobial peptides. This makes them attractive candidates for pharmaceutical applications to combat the ever increasing problem of antibiotic resistance [5,6]. However, the wide diversity of amino acid sequences and three-dimensional structures that have been reported for antimicrobial peptides [7,8] has created a significant challenge in this area of research because no concrete rules have
* Corresponding author. Tel.: þ1 403 220 6006; fax: þ1 403 289 9311. E-mail address:
[email protected] (H.J. Vogel). 0300-9084/$ - see front matter Ó 2008 Elsevier Masson SAS. All rights reserved. doi:10.1016/j.biochi.2008.04.013
emerged that completely characterize the mechanism of action for all antimicrobial peptides. It is therefore necessary to examine each peptide individually to determine which residues are essential for its activity and to identify additional residues that can be mutated to increase the activity for potential commercial use. Such an approach has been used, for example, to characterize the antimicrobial activities of nisins [9,10] and defensins [11,12]. As a further extension of this strategy, antimicrobial peptides have also been generated from peptide sequences with no significant endogenous antimicrobial activity [13]. Lactoferrin is a multifunctional iron-binding protein found in milk and mucosal secretions of mammals [14]. It has proven to be a valuable source of antimicrobial peptide sequences. The first antimicrobial peptide isolated from this large immuno-modulatory protein was generated following pepsin digestion of bovine lactoferrin and was named bovine lactoferricin [15]. This 25-residue cationic disulphide cross-linked
142
E.F. Haney et al. / Biochimie 91 (2009) 141e154
List of abbreviations LFampB bovine lactoferrampin. LFampH human lactoferrampin. DPPC 1,2-dipalmitoyl phosphatidylcholine. DPPG 1,2-dipalmitoyl phosphatidylglycerol. DSC differential scanning calorimetry. PI propidium iodide. SDS sodium dodecyl sulfate. DSS 2,2-dimethyl-2-silapentane-5-sulphonic acid. CFUs colony forming units. SDA sabouraud dextrose agar. RP-HPLC reverse phase e high performance liquid chromatography. NMR nuclear magnetic resonance. NOESY nuclear Overhauser effect spectroscopy. TOCSY total correlation spectroscopy. PPB potassium phosphate buffer. PBS phosphate buffered saline.
peptide forms a beta hairpin structure in aqueous solution [16] and demonstrates bactericidal activity against a wide range of bacterial species [17]. Human lactoferricin can also be generated from the pepsin digestion of the human lactoferrin protein but the antibacterial activity of this 49-residue peptide is significantly lower than that of the bovine peptide [18]. On the other hand, portions of human lactoferricin appear to have important immuno-stimulating [19] or endotoxin binding properties [20,21]. Human lactoferricin is partially folded in a membrane mimetic solution, maintaining a similar conformation to that seen in the intact lactoferrin protein [22]. For a complete review of the properties of the lactoferricin peptides, see Gifford et al. [23]. Recently, a novel cationic antimicrobial peptide sequence was identified in the N-terminal lobe of bovine lactoferrin and it was termed lactoferrampin [24]; in the remainder of this article it will be referred to as LFampB. This 17-residue peptide corresponds to residues 268e284 of bovine lactoferrin and although it is not a naturally occurring peptide, it displays significant antibacterial and candidacidal activity in vitro [24]. Interestingly the cationic charges at the C-terminal end of LFampB are essential for the antimicrobial activity of this peptide while the N-terminal portion of this peptide was thought to form an amphipathic helix [25,26]. NMR structural characterization of LFampB bound to detergent micelles confirmed the presence of a well structured N-terminal amphipathic helix and a flexible cationic C-terminal region [27]. Following the initial discovery of LFampB, a slightly longer sequence was published that included an additional N-terminal helix cap region, comprised of Asp-Leu-Ile. It was found that the longer peptide (cap-LFampB) had a higher activity than the regular LFampB peptide [26]. The importance of the Nterminal DLI helix cap motif was examined in antibacterial assays on various bacterial species. This work confirmed that
the cap, in spite of the added negative charge, gave rise to increased antimicrobial activity in LFampB, which was attributed to an increased stability of the alpha-helix at the Nterminus of the peptide [28]. This is consistent with the known stabilizing effects of helix cap motifs. (For a recent review of helix cap motifs see Fonseca et al. [29].) In this study, synthetic human lactoferrampin (LFampH) peptides were characterized to determine which residues were important for the antimicrobial activity of these related human lactoferrin-derived peptide sequences. LFampH was modeled after the same region of the LFampB peptide and it corresponds to residues 269e285 in human lactoferrin. Two key components are examined: the addition of a C-terminal lysine (residue 286 in human lactoferrin), which increases the overall positive charge of the peptide. We also studied the addition of the N-terminal helix cap, whose corresponding sequence in human lactoferrin is DAI. In addition, the Asp residue that is found near the C-terminal end of LFampH was mutated to an Asn residue to further increase the net positive charge, making this part of the sequence more similar to the C-terminal portion of LFampB. The amino acid sequences of all the lactoferrampin peptides used in this study are shown in Table 1. 2. Materials and methods 2.1. Peptide synthesis LFampB (WKLLSKAQEKFGKNKSR), cap-LFampB (DLIWKLLSKAQEKFGKNKSR), LFampH (WNLLRQAQE KFGKDKSP), LFampH-K (WNLLRQAQEKFGKDKSPK), cap-LFampH (DAIWNLLRQAQEKFGKDKSP), cap-LFamp H-K (DAIWNLLRQAQEKFGKDKSPK) and cap-LFampHK D17N (DAIWNLLRQAQEKFGKNKSPK) were synthesized with a MilliGen 9050 peptide synthesizer (MilliGen/ Biosearch, Bedford, MA) on pre-loaded PEG-PS solid phase supports (Applied Biosystems, Foster City, CA) with N-Fmoc-protected amino acids (ORPEGEN Pharma GmbH, Heidelberg, Germany). Peptide synthesis was achieved using Fmoc-chemistry according to the procedures outlined by the manufacturers. The peptides were purified by RP-HPLC (Jasco Corporation, Tokyo, Japan) to a purity of at least 95% and the authenticity of the peptides was confirmed by Table 1 Sequences of the human lactoferrampin peptides used in this study Peptide
Sequence
LFampB cap-LFampB LFampH LFampH-K cap-LFampH cap-LFampH-K cap-LFampH-K D17 N
WKLLSKAQEKFGKNKSR DLIWKLLSKAQEKFGKNKSR WNLLRQAQEKFGKDKSP WNLLRQAQEKFGKDKSPK DAIWNLLRQAQEKFGKDKSP DAIWNLLRQAQEKFGKDKSPK DAIWNLLRQAQEKFGKNKSPK
Number of AA’s
Net charge
17 20 17 18 20 21 21
þ5 þ4 þ2 þ3 þ1 þ2 þ3
The sequence of LFampB and cap-LFampB has been included for comparison.
E.F. Haney et al. / Biochimie 91 (2009) 141e154
ion trap mass spectrometry with a LCQ Deca XP (Thermo Finnigan, San Jose, CA). 2.2. Antimicrobial activity Candida albicans 315 (ATCC 10231) was cultured in Sabouraud dextrose broth and on Sabouraud dextrose agar (SDA) plates at 30 C under aerobic conditions. Streptococcus sanguis SK4 were cultured anaerobically in Brain Heart Infusion (BHI) medium at 37 C. Escherichia coli K12 was grown aerobically in BHI media at 37 C. To examine the peptide-mediated killing abilities of the lactoferrampin peptides, a propidium iodide (PI) (Molecular Probes Inc. Eugene, OR) permeability assay was used as described previously [24]. The PI assay is a quick assay to compare relative antimicrobial activities under a variety of conditions followed by validation by a microdilution viability assay. Briefly, 100 mL aliquots from a mid-log phase culture of C. albicans (approximately 107 colony forming units (CFUs)/mL) or 100 mL aliquots of mid-log phase bacterial suspensions (approximately108 CFUs/mL) were placed in 96-well U-bottom low affinity plates (Greiner), supplemented with PI (final concentration 6 mM), and incubated with equal volumes of peptide solutions (0.2e50 mM) in 1 mM potassium phosphate buffer (PPB) at pH 7. The bacterial suspensions were incubated at 37 C while the Candida assays were incubated at 30 C. PI fluorescence was measured at 5 min intervals for 1 h, at excitation and emission wavelengths of 544 and 620 nm, respectively, in a Fluostar Galaxy microplate fluorimeter (BMFG Labtechnologies, Offenburg, Germany). Fluorescence was recorded every 5 min for 1 h. The correlation between PI and CFUs has been previously described [24,30] and in the present experiment, only the maximum PI fluorescence is assayed to verify 100% killing. To verify killing, the number of CFUs in representative wells was determined after 1 h by diluting the cells in buffer and plating them on appropriate growth media. Following an incubation period of 48 h, the number of CFUs was counted. 2.3. Hemolytic activity Hemolytic activity of cap-LFampB, LFampH-K, capLFampH-K and cap-LFampH-K D17 N was obtained according to protocols described previously [24]. Erythrocytes from four healthy individuals, who had given their informed consent, were collected by centrifugation at 123 g for 15 min. Cells were washed three times in phosphate buffered saline (PBS) or isotonic glucose phosphate buffer containing 1 mM PPB at pH 7.0, supplemented with 287 mM glucose as an osmoprotectant. All of the cell suspensions were normalized to a hemoglobin concentration of 2 mM. Hundred microliters of these suspensions were placed in a 96-well V-bottom microtiter plate and 100 mL of a series of peptides in the same buffer was added, starting with a concentration of 50 mM. Each peptide concentration was tested in triplicate. After 45 min of incubation at 37 C, the microtiter plates were centrifuged at 550 g for 5 min and the absorbance in the supernatant
143
was measured at 450 nm. Complete hemolysis was achieved by the addition of 1% Tween-20. The percentage of hemolysis was calculated according to the following formula: [(A450 of the peptide treated sample A450 of buffer treated sample)/ (A450 of Tween-20 treated sample A450 of buffer treated sample)] 100%. 2.4. Fluorescence spectroscopy Fluorescence emission spectra were collected on a Varian Cary Eclipse (Varian Inc., Palo Alto, CA) fluorimeter equipped with a multicell sample holder and a temperature control unit that was maintained at 20 C. Each 2 mL sample contained 1 mM peptide in buffer (10 mM Tris, 150 mM NaCl, 1 mM EDTA, pH 7.4) and samples containing detergent were prepared to a final SDS concentration of 25 mM. Using an excitation wavelength of 280 nm, emission spectra were recorded between 300 and 500 nm and slit widths of 10 nm were used for both the excitation and emission settings. In total, 10 spectra were recorded and then averaged in the final analysis. 2.5. Differential scanning calorimetry Samples were prepared according to a modified protocol from that originally described by Prenner et al. [31]. The correct amount of peptide was taken from stock solutions prepared in water and placed in a small glass vial which was frozen and lyophilized overnight to remove the aqueous phase. The following day, the peptide was resuspended in pure ethanol. 0.5 mg of 1,2-dipalmitoyl phosphatidylglycerol (DPPG) or 0.5 mg of 1,2-dipalmitoyl phosphatidylcholine (DPPC) (Avanti Polar Lipids, Alabaster, AL), dissolved in chloroform, was added to the peptide solution in ethanol and then the organic phase of this mixture was evaporated under a stream of nitrogen gas. Pure lipid films of DPPG or DPPC were prepared in a similar fashion as described above without any peptide or ethanol in the mixture. All of the lipid films were placed under vacuum overnight to remove any remaining organic solvent. Lipid films and buffer (20 mM phosphate buffer, 130 mM NaCl, pH 7.4) were heated to approximately 10 C above the main phase transition temperature and then the lipid film was hydrated with 1 mL of the heated buffer to give a final lipid concentration of 0.5 mg/mL. The samples were vortexed to ensure complete resuspension of the lipid film. DSC data were collected using a Microcal high-sensitivity VP-DSC (Microcal, Northampton, MA). Five thermograms were recorded for each sample between 20 C and 60 C employing a heating scan rate of 10 C/h and the second, third and fourth scans were averaged in the final analysis. All of the data acquisition and analysis was done using the Microcal Origin Software (version 7.0). 2.6. NMR sample preparation Lyophilized peptide samples were dissolved in 90% H2O:10%D2O to a final volume of 500 mL. At this point,
E.F. Haney et al. / Biochimie 91 (2009) 141e154
144
10 mL of 10 mM DSS (2,2-dimethyl-2-silapentane-5-sulphonic acid) was added to each sample as an internal chemical shift standard. The pH of the samples was adjusted using dilute HCl and NaOH to values between 3.5 and 4.5. Once the final pH was established, the peptide concentration in a diluted aliquot was determined using a theoretical extinction coefficient of 3(280) 5500 M1 cm1. Following NMR data collection of the aqueous peptide samples, deuterated SDS (Cambridge Isotope Laboratories, Inc., Andover, MA) was added to a final concentration of 200 mM and the pH was readjusted to between 3.5 and 4.5. Final peptide concentrations and pH values for each of the NMR samples used in this study are summarized in Table 2. All NMR experiments were carried out at 298 K.
2.7. NMR experiments All spectra were recorded at 298 K on a Bruker Avance Ultrashield-Plus 600 MHz spectrometer (Bruker, Fa¨llanden, Switzerland). Two-dimensional (2-D) nuclear Overhauser enhancement (NOESY) spectra were collected for the aqueous samples using a mixing time of 250 ms. 2-D NOESY and 2D total correlation (TOCSY) spectra were recorded for the micelle-bound peptide samples using mixing times of 100 ms for both experiments. Spectra were obtained with 4096 512 data points in the F2 and F1 dimensions respectively at a sweep width of 8503.401 Hz. Water suppression in the NOESY and TOCSY spectra was achieved using the excitation sculpting technique [32]. The 2-D NOESY and TOCSY spectra were processed with the NMRPipe software package [33] on a workstation running the Redhat 7.1 version of the Linux operating system. All spectra were zero-filled in both dimensions and Fourier transformed with a shifted sineebell curve. Spectra were visualized and analyzed using NMRView 5.0.1.4 [34].
2.8. Structure calculations Proton chemical shifts for all of the human lactoferrampin derivatives were assigned according to Wuthrich [35]. These chemical shifts along with NOE peaks picked from the 2-D NOESY spectra were used as input values for automated structure calculations by CYANA 2.0 [36]. Analysis of the structures was performed through built in CYANA protocols and using PROCHECK [37]. All of the resulting structures were visualized and analyzed using MOLMOL [38]. Table 2 NMR sample parameters including peptide concentration and pH of each sample Peptide
Concentration (mM)
pH
pH in SDS
LFampH LFampH-K Cap-LFampH Cap-LFampH-K Cap-LFampH-K D17N
1.53 1.49 1.16 1.36 1.35
3.90 3.94 4.36 3.62 3.87
3.61 4.44 4.47 4.41 4.25
Fig. 1. Peptide concentration dependent activity against Candida albicans (A), Escherichia coli (B) and Streptococcus sanguis (C) of cap-LFampB (filled diamonds), LFampB (open diamonds), LFampH (open triangles), LFampH-K (open circles), cap-LFampH (filled triangles), cap-LFampH-K (filled circles) and cap-LFampH-K D17 N (filled squares) as determined by monitoring the PI fluorescence enhancement. This figure shows the PI fluorescence observed after 60 min incubation with the peptide and represents the average of duplicate trials for a typical experiment carried out at least two times. See methods and materials for a complete description of the PI assay protocols.
E.F. Haney et al. / Biochimie 91 (2009) 141e154 1000
1000
A
B
800
Intensity (a.u.)
Intensity (a.u.)
800
600
400
600
400
200
200
0 300
350
400
450
0 300
500
350
Wavelength (nm) 1000
1000
C
450
500
450
500
D
800
Intensity (a.u.)
600
400
200
0 300
400
Wavelength (nm)
800
Intensity (a.u.)
145
600
400
200
350
400
450
0 300
500
350
Wavelength (nm) 1000
400
Wavelength (nm)
E
Intensity (a.u.)
800
600
400
200
0 300
350
400
450
500
Wavelength (nm) Fig. 2. Fluorescence emission spectra of LFampH (A), LFampH-K (B), cap-LFampH (C), cap-LFampH-K (D) and cap-LFamp-K D17 N (E) in buffer (black) or in the presence of 25 mM SDS (grey). All spectra were recorded at a peptide concentration of 1 mM using an excitation wavelength of 280 nm.
3. Results 3.1. Antimicrobial activity The two bovine peptides, LFampB and cap-LFampB were the most active against C. albicans, E. coli and S. sanguis as demonstrated by the largest fluorescence signals from the PI assays (Fig. 1). Interestingly, the two corresponding human peptides, LFampH and cap-LFampH were essentially inactive at all of the peptide concentrations tested as there is no increase in the fluorescence indicating that PI is not crossing the yeast
membrane. However, the addition of a Lys residue to the Cterminus of these two peptides, giving LFampH-K and capLFampH-K, led to a significant increase in PI permeability in the presence of C. albicans (Fig. 1A). This increase in membrane permeability was not observed for E. coli and S. sanguis (Fig. 1 B and C). Finally, the cap-LFampH-K D17N peptide, with the negatively charged Asp in the C-terminus mutated to a neutral Asn residue, was more active against all three organisms tested compared to all of the other human variants. The effect on E. coli is modest (Fig. 1B) however, against S. sanguis and C. albicans, the activity of the cap-LFampH-K D17N
E.F. Haney et al. / Biochimie 91 (2009) 141e154
146
A
4.0
Cp (10-3 cal/°C)
Cp (10-3 cal/°C)
4.0
3.0
2.0
1.0
0.0
B
3.0
2.0
1.0
0.0 20
30
40
50
60
20
Temperature (°C)
C
4.0
3.0
2.0
1.0
0.0
60
D
3.0
2.0
1.0
30
40
50
60
20
Temperature (°C)
30
40
50
60
Temperature (°C)
E
4.0
3.0
Cp (10-3 cal/°C)
Cp (10-3 cal/°C)
50
0.0 20
4.0
40
Temperature (°C)
Cp (10-3 cal/°C)
Cp (10-3 cal/°C)
4.0
30
2.0
1.0
0.0
F
3.0
2.0
1.0
0.0 20
30
40
50
60
Temperature (°C)
20
30
40
50
60
Temperature (°C)
Fig. 3. Differential scanning calorimetry thermograms of pure DPPG (A), and 10:1 lipid to peptide ratios containing DPPG mixed with LFampH (B), LFampH-K (C), cap-LFampH (D), cap-LFampH-K (E) and cap-LFampH-K D17 N (F). Each thermogram is an average of three scans and each sample contains 0.5 mg/mL of DPPG.
peptide begins to approach that of the LFampB peptides (Fig. 1A and C). In general, the killing kinetics of all the peptides are quite fast under the conditions tested as the maximum values are reached within 10 min of adding the peptide. 3.2. Hemolytic activity Only those peptides with appreciable candidacidal activity were tested for their hemolytic activity. The hemolytic activity of LFampB has been published previously [24] and it was shown to be relatively non-hemolytic. None of the peptides tested here showed hemolytic activity in PBS. In the PPB containing isotonic glucose, the most hemolytic peptide was cap-LFampB
where less than 10% hemolysis was observed at a peptide concentration of 6.2 mM and an HC50 at approximately 25 mM. Two peptides, cap-LFampH-K and cap-LFampH-K D17 N, demonstrated about 10% hemolysis in PPB with isotonic glucose at a peptide concentration of 50 mM while LFampH-K again showed no appreciable hemolytic activity at any of the peptide concentrations tested in this buffer. 3.3. Fluorescence The intrinsic tryptophan fluorescence of each of the LFampH derivatives was measured to determine if the presence of SDS micelles altered the environment surrounding
E.F. Haney et al. / Biochimie 91 (2009) 141e154 18.0
A
18.0 16.0
14.0
14.0
12.0
12.0
Cp (cal/°C)
Cp (10-3 cal/°C)
16.0
10.0 8.0 6.0
8.0 6.0 4.0
2.0
2.0
0.0
0.0
-2.0
-2.0
30
40
50
60
B
10.0
4.0
20
20
Temperature (°C) 18.0
C
18.0 16.0
14.0 12.0 10.0 8.0 6.0 4.0
60
D
12.0 10.0 8.0 6.0 4.0 0.0
-2.0
-2.0
20
30
40
50
60
20
Temperature (°C)
30
40
50
60
Temperature (°C)
E
18.0 16.0
14.0
Cp (10-3 cal/°C)
Cp (10-3 cal/°C)
50
14.0
0.0
16.0
40
2.0
2.0
18.0
30
Temperature (°C)
Cp (10-3 cal/°C)
Cp (10-3 cal/°C)
16.0
147
12.0 10.0 8.0 6.0 4.0
F
14.0 12.0 10.0 8.0 6.0 4.0 2.0
2.0 0.0
0.0
-2.0
-2.0
20
30
40
50
60
Temperature (°C)
20
30
40
50
60
Temperature (°C)
Fig. 4. Differential scanning calorimetry thermograms of pure DPPC (A), and 10:1 lipid to peptide ratios containing DPPC mixed with LFampH (B), LFampH-K (C), cap-LFampH (D), cap-LFampH-K (E) and cap-LFampH-K D17 N (F). Each thermogram is an average of three scans and each sample contains 0.5 mg/mL of DPPC.
the tryptophan residue found at or near the N-terminus of the human lactoferrampin peptides. The emission spectra of the human lactoferrampin peptides in aqueous buffer was similar among all of the peptides with an emission maximum centered around 358 nm for the uncapped peptides and 356 nm for the capped versions. All of the observed intensities were similar ranging from 400 to 450 a.u. (Fig. 2). Upon the addition of SDS micelles, the effect on the Trp emission was dependent on the presence of the N-terminal helix cap, DAI, or if the peptide started with Trp as the first amino acid. The emission spectra for LFampH and LFampH-K, both lacking the helix cap, showed a blue shift of the maximum wavelength to
338 nm and a decrease in the intensity of the emitted light (Fig. 2A and B). The emission spectra for the peptides that possessed the N-terminal helix cap Asp-Ala-Ile also demonstrated a blue shift, although smaller than that observed for the previously mentioned peptides, to approximately 348 nm, however the intensity of the emitted light increased significantly in comparison to these peptides in buffer (Fig. 2CeE). 3.4. Differential scanning calorimetry Differential scanning calorimetry was used to examine the effects of the human lactoferrampin derivatives on DPPG and
E.F. Haney et al. / Biochimie 91 (2009) 141e154
148 0.6
A
1.1 1.6 2.1 2.6 3.1 3.6 4.1 4.6 8.8 8.7 8.6 8.5 8.4 8.3 8.2 8.1 8.0 7.9 7.8 7.7 7.6 0.6
B
1.1 1.6 2.1 2.6 3.1
transition is broadened and the temperature of this peak is slightly shifted to a lower temperature. Interestingly, LFampH appears to have a more significant effect on the main phase transition of DPPG than LFampH-K as evidenced by the irregular shape of the thermogram around 40 C. For the three peptides with the helix cap: cap-LFampH, cap-LFampH-K and capLFampH-K-D17 N, the heating profiles of 10:1 lipid to peptide ratios are rather different from the thermogram of pure DPPG (Fig. 3DeF). In all three cases, the pre-transition has disappeared and the main phase transition is extremely broad. The effect of the human lactoferrampin peptides on the lipid phase behavior of DPPC is not as dramatic compared to the results observed in DPPG (Fig. 4). For all the peptides, the Tm of the main phase transition does not change from 41 C and most of the differences are seen in the shape of the pre-transition peak and the intensity of the main transition. LFampH has very little effect on the shape of the thermogram in comparison to pure DPPC (Fig. 4B). LFampH-K has a slightly larger effect with the pre-transition at 34 C flattening out and the main phase transition decreasing in intensity (Fig. 4C). Both of the capped LFampH peptides caused the pre-transition to disappear and a large decrease in the intensity of the main phase transition (Fig. 4D and E). Interestingly, the thermogram of the cap-LFampH-K D17 N peptide mixed with DPPC resembles the results seen for LFampH-K with the majority of the pre-transition disappearing and a main phase transition of similar intensity (Fig. 4F). 3.5. NMR spectroscopy
3.6 4.1 4.6 8.8 8.7 8.6 8.5 8.4 8.3 8.2 8.1 8.0 7.9 7.8 7.7 7.6
Fig. 5. Fingerprint regions of 2D 1H-NOESY spectra of LFampH in 9:1 H2O:D2O (A) and in 200 mM SDS (B). Note the increased number of peaks and the larger dispersion in the sample containing SDS micelles which is indicative of a more defined peptide structure.
DPPC lipids. These phospholipids are useful model lipids for this experiment because the negatively charged head groups of the DPPG molecules are considered a good mimic for the bacterial membrane while DPPC is a better model for a eukaryotic membrane as they are both zwitterionic [39]. In addition, the palmitoyl acyl chains are common in biological membranes, which allows for extrapolation of these results to biological systems. In agreement with the fluorescence results, the DSC results can be separated into the peptides that posses the helix cap and those that do not. The DSC thermogram for pure DPPG shows a pre-transition at approximately 32 C and a main phase transition at 40 C (Fig. 3A). At lipid to peptide ratios of 10:1, the effect of LFampH and LFampH-K on DPPG is very significant (Fig. 3B and C). In both cases the pre-transition at 32 C disappears in the presence of peptide and the main phase
All of the peptides were relatively unstructured in aqueous solution but in the presence of deuterated SDS micelles, the peptides adopted a more definite conformation as evidenced by the increased number and distribution of observed NOE cross peaks in the 2D-NOESY spectrum (Fig. 5). Statistics from all of the CYANA structure calculations for the peptides in SDS micelles are shown in Table 3. A significant number of distance restraints were used to generate the three-dimensional solution structures of the human lactoferrampin peptides and most of the phi and psi angles of the resulting structures fall into the most favored and additionally allowed regions of the Ramachandran plot [40]. The few residues that lie outside of these regions of the Ramachandran plot are found in the flexible C-terminus of the LFampH (see below) where a large degree of freedom is associated with the conformation of the peptide backbone. Upon examination of the backbone conformations of all the human lactoferrampin variants, there are similar structural features that emerge in all of the peptide structures. LFampH and LFampH-K adopt a well-defined helix stretching from Trp1 to Gly12 (Fig. 6A and B) as demonstrated by the low backbone ˚ and 0.54 A ˚ , respectively (Table 3). RMSD for this region of 0.48 A Adding the DAI helix cap to the N-terminus does not significantly alter the conformation of the longer lactoferrampin peptides and in reality, the length of the well-defined helical region is simply extended by three residues (Fig. 6CeE). Backbone RMSDs for cap-LFampH, cap-LFampH-K and
E.F. Haney et al. / Biochimie 91 (2009) 141e154
149
Table 3 Structural Statistics for the SDS bound micelle structures of LFampH derivatives determined by CYANA LFampH
LFampH-K
Cap-LFampH
Cap-LFampH-K
Cap-LFampH-K D17N
Peaks With assignment Without assignment
381 6
435 7
382 5
396 9
424 5
Upper distance limits Total Short-range ji jj 1 Medium range 1 < ji jj < 5 Long range ji jj 5
258 184 73 1
323 220 102 1
266 186 80 0
281 201 80 0
295 206 89 0
RMSD (residue range) ˚) Backbone (A ˚) Backbone (A
(1e12) 0.48 1.06
(1e12) 0.54 1.03
(1e15) 0.53 1.09
(1e15) 0.65 1.25
(1e15) 0.44 1.00
Ramachandran spacea Most favored Additionally allowed Generously allowed Disallowed
82.5 17.1 0.0 0.4
72.5 26.8 0.7 0.0
85.9 13.8 0.0 0.3
74.7 25.0 0.0 0.3
67.6 31.8 0.0 0.6
a
As determined by PROCHECK.
cap-LFampH-K D17 N in the presence of SDS micelles and fit across the first 15 residues of these peptides were calculated to ˚ , 0.65 A ˚ and 0.44 A ˚ , respectively (Table 3). This corbe 0.53 A responds to a well-defined helix stretching from Asp1 to Gly15. This glycine residue is the same as the Gly in position 12 in the uncapped peptides. It would appear that the glycine residue in the human lactoferrampin variants serves as a helix breaker and allows for conformational flexibility in the Cterminal region of the micelle-bound peptide (Fig. 6). Closer examination of the side chain orientations of the human lactoferrampin peptides reveals a conserved hydrophobic patch in all of SDS bound structures (Fig. 7). The hydrophobic patch in LFampH and LFampH-K is bordered by Trp1 and Phe11 and encompasses Leu3, Leu4 and Ala7. In capLFampH, cap-LFampH-K and cap-LFampH-K D17 N, this hydrophobic surface is made up of the same five residues (Trp4, Leu6, Leu7, Ala10, and Phe11), but in addition Ala2 and Ile3 are also found close to this hydrophobic face. The hydrophobic patch corresponds to one face of the welldefined helical region that has been described for all of the human lactoferrampin peptides. This helix is amphipathic in nature with two positively charged residues (Arg and Lys) and one negatively charged (Glu) residue on the opposite face. It is also interesting to note the orientation of Asp1 in all of the structures of the capped peptides as its side chain appears to reside closer to the hydrophobic face of the amphipathic helix as opposed to the charged surface. The remaining charged amino acid side chains are found in the unstructured C-terminal regions of the peptides. Mapping the charged residues onto the surface of the human lactoferrampin peptides further demonstrates the presence of the hydrophobic patch and the highly charged nature of the flexible C-terminus (Fig. 8). 4. Discussion Many of the physiological activities associated with bovine and human lactoferrin have been localized to the cationic N-
terminal lobe of this iron-binding protein [41]. Already present within this region of lactoferrin is lactoferricin with enhanced antimicrobial activity [23]. LFampB and cap-LFampB are antimicrobial peptides that are also derived from the N-terminal lobe of bovine lactoferrin with activity against a wide range of gram positive, gram negative and yeast [24,28]. The key determinants underlying the mechanism of activity of these two peptides has been narrowed to the presence of an N-terminal amphipathic helix and a highly flexible and cationic C-terminus [26,27]. In this study, we have examined related lactoferrampin sequences derived from human lactoferrin with the intent of establishing a new set of antimicrobial peptides derived from the basic N-terminal lobe of human lactoferrin. While synthetic peptides directly corresponding to the LFampB and capLFampB sequences (LFampH and cap-LFampH) did not display any antimicrobial activity against C. albicans, E. coli or S. sanguis (Fig. 1), peptides with a larger net cationic charge in the C-terminus did show enhanced Candida membrane perturbation in the PI assay. The addition of a Lys residue to the C-terminus led to a significant increase in the observed fluorescence at peptide concentrations higher than 25 mM (Fig. 1). It appears that this extra positive residue is analogous to the C-terminal Arg residue that is present in the bovine forms of lactoferrampin [24]. This increased cationic character of the C-terminus may account for the increased candidacidal activity of LFampH-K and cap-LFampH-K. Continuing with the idea that the cationic nature of the Cterminus is essential for the antimicrobial activity of LFampB, mutating Asp17 in cap-LFampH-K to an Asn residue removes a carboxylic acid group resulting in an even stronger net increase in the antimicrobial activity of this human variant. The C-terminus of this mutated human peptide now resembles the C-terminus of LFampB and interestingly, the antimicrobial activity of this peptide is now beginning to approach that observed in LFampB (Fig. 1). While the importance of the positively charged C-terminus of LFampH is evident from these results, there is no significant increase in antimicrobial
E.F. Haney et al. / Biochimie 91 (2009) 141e154
150
A C N
B C N
C N
C
D N
C
E N
C
Fig. 6. Backbone overlays of the NMR structures determined for the human lactoferrampin variants in the presence of SDS micelles. For LFampH (A) and LFampH-K (B), the structures are superimposed across the first 12 residues of the peptides. For cap-LFampH (C), cap-LFampH-K (D) and capLFampH-K D17 N (E) the backbones are fitted across the first 15 residues of the peptides. The backbone and heavy atom RMSD values for each of the peptides can be seen in Table 3. This figure was generated using MOLMOL [38].
activity observed in any of the human peptides when the helix cap is added to the peptide. This is different from the results seen for LFampB [25,28]; however further analysis of the other biophysical data for the human lactoferrampin peptides still indicates that the N-terminal region of this peptide forms an amphipathic helix that is likely to interact with the microbial phospholipid bilayer. Examination of the SDS bound structures of all the LFampH variants reveals an N-terminal helix stretching from the first residue of a given peptide through to the conserved Gly residue in each of the peptides (Fig. 6). This glycine residue appears to induce a conformational flexibility in the C-terminal region of
human lactoferrampin which is consistent with the micellebound structures of LFampB determined previously [27]. Glycine residues in other antimicrobial peptides, such as chicken cathelicidin fowlicidin-1, have been shown to induce kinks in helical peptides [42]. In this case, the glycine induced kink appears to separate the various functions into different regions of the fowlicidin peptide. This may be similar to what is occurring for the lactoferrampin peptides as the glycine residue separates the region responsible for interacting with the membrane, the amphipathic helix, from the cationic region responsible for attraction to the negatively charged membrane. The importance of conformationally flexible regions of antimicrobial peptides has also been observed in other antimicrobial peptides. For instance, Caerin1.1, an amphibian antimicrobial peptide, is a helical peptide with two proline residues that cause a flexible hinge between two well structured helices [43]. Mutating these two proline residues to Ala residues stabilizes the helical structure of caerin1.1 along the entire length of the peptide but this mutation essentially abolishes the antimicrobial activity for this peptide [44]. A caerin1.1 mutant with both Pro residues replaced with Gly residues reintroduces some of the C-terminal flexibility although at an intermediate level compared to the native peptide and consequently only recovers a fraction of the original activity of the peptide [44]. All of the LFampH structures have a hydrophobic surface on the amphipathic helix that is anchored on either end by Trp and Phe (Fig. 7). This hydrophobic patch is identical to that observed in the micelle-bound structures of LFampB [27]. Intrinsic tryptophan fluorescence was used to examine the molecular environment of the Trp residue in the absence and presence of SDS micelles. For all of the human lactoferrampin variants, a blue shift was observed in the Trp emission spectra when the peptide was mixed with SDS micelles. This can be attributed to the indole moiety inserting into the hydrophobic core of the detergent micelles. For LFampH and LFampH-K, there is an attenuation of the fluorescence signal in the presence of SDS micelles which is akin to the results reported for LFampB [27]. We previously attributed this observed quenching to the proximity of the Trp side chain to an interaction of the indole ring with the sulfur atoms in the head groups of the SDS molecules [27]. This was based on earlier Trp fluorescence measurements performed on the protein calmodulin where the sulfur atoms of methionine residues were shown to cause quenching of the Trp fluorescence [45]. With this in mind, it is interesting to examine the capped LFampH peptides where the emission intensity is dramatically increased in the presence of SDS micelles. This would suggest that the DAI helix cap causes the N-terminal helix of LFampH to insert more deeply into the hydrophobic core of the membrane, which increases the distance between Trp and the sulfur atom in the detergent molecule and consequently abrogates the quenching effect observed in the uncapped peptides. Also the positive charge on the N-terminal end of the peptide would influence the depth of insertion, particularly when it is part of the Trp1 residue in the uncapped peptides. The idea that the capped LFampH peptides insert more deeply into the hydrophobic core of a bilayer is supported
E.F. Haney et al. / Biochimie 91 (2009) 141e154
151
by the results from the DSC experiments. DSC is a valuable tool to examine the changes that occur in the phospholipid bilayer structure upon binding by an antimicrobial peptide [46,47]. In the case of LFampH and LFampH-K, there are significant changes that are observed in the thermogram, most notably the reduction in the intensity of the main phase transition peak and a slight broadening of this region of the thermogram. This suggests that LFampH and LFampH-K are interacting with the DPPG lipids and are interacting to some extent with the hydrophobic core of the bilayer. The addition of the DAI helix cap to the LFampH peptides induces even more dramatic changes in the DPPG thermograms with significant broadening of the main phase transition peak. In the case of cap-LFampH-K D17 N, the most active human variant of lactoferrampin, the main phase transition is extremely disrupted and the Tm has shifted to a much lower temperature. The main phase transition is the point at which the phospholipid bilayer converts from a gel phase to the liquid crystalline phase and it is strongly influenced by the intermolecular packing between the acyl chains of the lipid species [48]. These results suggest that the capped versions of the LFampH peptides penetrate more deeply into the hydrophobic core of the phospholipid bilayer and significantly disrupt intermolecular interaction between the acyl chains of the DPPG molecules leading to broad Tm’s for the main phase transition and even lowering the amount of energy required to move from the rippled gel phase to the liquid crystalline phase. The effect of the LFampH peptides on the thermotropic phase behavior of DPPC is not as significant with the largest decrease in the height of the main phase transition being observed for LFampH peptides containing the helix cap of DAI. This is in agreement with the idea that the helix cap promotes the interaction between the N-terminal amphipathic helix and the phospholipid membrane, which leads to a disruption of the acyl chain packing in the hydrophobic core of the bilayer. Because the differences in the thermograms are not as significant for DPPC compared to DPPG, this suggests that these peptides are potentially selective for the negatively charged bacterial membrane and indicates the possibility of a good therapeutic index for potential pharmaceutical applications. This difference in membrane perturbation properties of the lactoferrampin variants is further reflected in their antimicrobial and hemolytic activity. Bacterial membranes contain lipids with negatively charged head groups [49] as do Candida membranes [50]. This confers an overall negative charge to the surface of these membranes. Mammalian membranes on the other hand are composed predominantly of zwitterionic lipids [49] which gives these membranes no net charge and most of
Fig. 7. Ribbon diagrams of NMR structures of LFampH (A), LFampH-K (B), cap-LFampH (C), cap-LFampH-K (D) and cap-LFampH-K D17 N (E) in the presence of SDS micelles. Hydrophobic residues (Trp, Leu, Ile, Ala and Phe) have been highlighted in green, cationic residues (Lys and Arg) are shown in blue and anionic residues (Asp and Glu) are shown in red. In all the peptide structures, the hydrophobic residues cluster along one face of the amphipathic helix while the charged residues are found on the opposite face of the helix or in the flexible C-terminus. This figure was generated using MOLMOL [38].
152
E.F. Haney et al. / Biochimie 91 (2009) 141e154
Fig. 8. Surface charge distribution for the SDS bound structures of LFampH (A), LFampH-K (B), cap-LFampH (C), cap-LFampH-K (D), and cap-LFampH-K D17 N (E). Positively charged regions are shown in blue, negatively charged regions are red and neutral regions appear white. The panel on the left corresponds to the same orientation of the peptide as seen in Fig. 7 while the panel on the right is looking at the opposing face. This figure was generated using MOLMOL [38].
E.F. Haney et al. / Biochimie 91 (2009) 141e154
the peptides tested in this study exhibit limited hemolytic activity. It is an attractive hypothesis that the changes in the membrane structure seen in the DSC experiments help to explain the results of the PI and hemolytic activity assays and may account for the selectivity against bacterial and yeast species over mammalian cells. Generating antimicrobial peptides from inactive sequences is not a new phenomenon in antimicrobial peptide research. In a recent study, a non-toxic membrane associated portion from the N-terminal region of the E. coli enzyme IIAGlc was mutated in a series of steps to generate the antimicrobial peptide, aurein 1.2 [51]. The solution structures of these mutant peptides revealed features important for antibacterial activity such as a larger hydrophobic surface and the distribution of charges on the hydrophilic face. Concerning LFampH, the antimicrobial activity of the peptide is enhanced by increasing the net positive charge in the flexible C-terminal region of the peptide while the hydrophobic patch mediates the interaction with the phospholipid bilayer. The most active peptide based on the human lactoferrampin sequence, cap-LFampH-K D17 N, is closer in sequence to cap-LFampB but the antimicrobial activity of this peptide is still not as high as the bovine form. This underscores the idea that despite the striking structural similarities between the human lactoferrampin peptides compared to LFampB, the antimicrobial activity is not dictated by the micelle-bound structure alone and that there are other factors that must contribute to the observed antimicrobial activity of this class of peptides. 5. Conclusion Novel antimicrobial peptides from the N-terminal lobe of human lactoferrin were identified and their solution structures bound to SDS micelles were determined using NMR spectroscopy. Fluorescence and DSC experiments provided further information about the mechanism through which these human lactoferrampin peptides interact with lipid species and gave insight into their possible mechanism of action. Interestingly, LFampH and cap-LFampH, two peptides that correspond directly to previously reported bovine lactoferrampin sequences, were relatively inactive. By adding a Lys residue to the C-terminus or mutating an Asp to an Asn, the net positive charge of the flexible C-terminal end increased and this led to a significant improvement in the antimicrobial and Candidacidal activity of these peptides. All of the LFampH peptides formed a well structured N-terminal helix but this did not directly correlate to antimicrobial activity for all of these peptides. The amphipathic Nterminal helix appears to be responsible for mediating the interaction between the peptide and the phospholipid bilayer but the antimicrobial activity of the LFampH peptides is dictated by the net cationic charge of the flexible C-terminal region. Acknowledgements This project was supported by the Canadian Institutes for Health Research through the ‘Novel alternatives to antibiotics’
153
program. HJV is the recipient of a Scientist award from the Alberta Heritage Foundation for Medical Research. The BioNMR center received financial support from the Canadian Institutes for Health Research and the University of Calgary. We are indebted to Dr. Deane McIntyre for NMR spectrometer upkeep. References [1] M. Zasloff, Antimicrobial peptides of multicellular organisms, Nature 415 (2002) 389e395. [2] H. Jenssen, P. Hamill, R.E.W. Hancock, Peptide antimicrobial agents, Clin. Microbiol. Rev. 19 (2006) 491e511. [3] D.I. Chan, E.J. Prenner, H.J. Vogel, Tryptophan- and arginine-rich antimicrobial peptides: structures and mechanisms of action, Biochim. Biophys. Acta 1758 (2006) 1184e1202. [4] K.A. Brogden, Antimicrobial peptides: pore formers or metabolic inhibitors in bacteria? Nat. Rev. Microbiol. 3 (2005) 238e250. [5] H.P. Stallmann, C. Faber, A.V.N. Amerongen, P.I.J.M. Wuisman, Antimicrobial peptides: review of their application in musculoskeletal infections, Injury 37 (2006) 34e40. [6] R.E.W. Hancock, H.G. Sahl, Antimicrobial and host-defense peptides as new anti-infective therapeutic strategies, Nat. Biotechnol. 24 (2006) 1551e1557. [7] R.M. Epand, H.J. Vogel, Diversity of antimicrobial peptides and their mechanisms of action, Biochim. Biophys. Acta 1462 (1999) 11e28. [8] P.M. Hwang, H.J. Vogel, Structureefunction relationships of antimicrobial peptides, Biochem. Cell Biol. 76 (1998) 235e246. [9] R. Rink, J. Wierenga, A. Kuipers, L.D. Kluskens, A.J.M. Driessen, O.P. Kuipers, G.N. Moll, Dissection and modulation of the four distinct activities of nisin by mutagenesis of rings A and B and by C-terminal truncation, Appl. Environ. Microbiol. 73 (2007) 5809e5816. [10] J. Yuan, Z.Z. Zhang, X.Z. Chen, W. Yang, L.D. Huan, Site-directed mutagenesis of the hinge region of nisinZ and properties of nisinZ mutants, Appl. Microbiol. Biotechnol. 64 (2004) 806e815. [11] M. Pazgier, A. Prahl, D.M. Hoover, J. Lubkowski, Studies of the biological properties of human beta-defensin 1, J. Biol. Chem. 282 (2007) 1819e1829. [12] D.M. Hoover, Z.B. Wu, K. Tucker, W.Y. Lu, J. Lubkowski, Antimicrobial characterization of human beta-defensin 3 derivatives, Antimicrob. Agents Chemother. 47 (2003) 2804e2809. [13] H.S. Ahn, W. Cho, S.H. Kang, S.S. Ko, M.S. Park, H. Cho, K.H. Lee, Design and synthesis of novel antimicrobial peptides on the basis of alpha helical domain of Tenecin 1, an insect defensin protein, and structureeactivity relationship study, Peptides 27 (2006) 640e648. [14] P. Valenti, F. Berlutti, M.P. Conte, C. Longhi, L. Seganti, Lactoferrin functions e current status and perspectives, J. Clin. Gastroenterol. 38 (2004) S127eS129. [15] M. Tomita, W. Bellamy, M. Takase, K. Yamauchi, H. Wakabayashi, K. Kawase, Potent antibacterial peptides generated by pepsin digestion of bovine lactoferrin, J. Dairy Sci. 74 (1991) 4137e4142. [16] P.M. Hwang, N. Zhou, X. Shan, C.H. Arrowsmith, H.J. Vogel, Threedimensional solution structure of lactoferricin B, an antimicrobial peptide derived from bovine lactoferrin, Biochemistry 37 (1998) 4288e4298. [17] W. Bellamy, M. Takase, H. Wakabayashi, K. Kawase, M. Tomita, Antibacterial spectrum of lactoferricin-B, a potent bactericidal peptide derived from the N-terminal region of bovine lactoferrin, J. Appl. Bacteriol. 73 (1992) 472e479. [18] L.H. Vorland, H. Ulvatne, J. Andersen, H.H. Haukland, O. Rekdal, J.S. Svendsen, T.J. Gutteberg, Lactoferricin of bovine origin is more active than lactoferricins of human, murine and caprine origin, Scand. J. Infect. Dis. 30 (1998) 513e517. [19] D. Legrand, E. Elass, M. Carpentier, J. Mazurier, Interactions of lactoferrin with cells involved in immune function, Biochem. Cell Biol. 84 (2006) 282e290.
154
E.F. Haney et al. / Biochimie 91 (2009) 141e154
[20] B. Japelj, P. Pristovsek, A. Majerle, R. Jerala, Structural origin of endotoxin neutralization and antimicrobial activity of a lactoferrin-based peptide, J. Biol. Chem. 280 (2005) 16955e16961. [21] P.H.C. vanBerkel, M.E.J. Geerts, H.A. vanVeen, M. Mericskay, H.A. deBoer, J.H. Nuijens, N-terminal stretch Arg(2), Arg(3), Arg(4) and Arg(5) of human lactoferrin is essential for binding to heparin, bacterial lipopolysaccharide, human lysozyme and DNA, Biochem. J. 328 (1997) 145e151. [22] H.N. Hunter, A.R. Demcoe, H. Jenssen, T.J. Gutteberg, H.J. Vogel, Human lactoferricin is partially folded in aqueous solution and is better stabilized in a membrane mimetic solvent, Antimicrob. Agents Chemother. 49 (2005) 3387e3395. [23] J.L. Gifford, H.N. Hunter, H.J. Vogel, Lactoferricin: a lactoferrin-derived peptide with antimicrobial, antiviral, antitumor and immunological properties, Cell. Mol. Life Sci. 62 (2005) 2588e2598. [24] M.I.A. van der Kraan, J. Groenink, K. Nazmi, E.C.I. Veerman, J.G.M. Bolscher, A.V.N. Amerongen, Lactoferrampin: a novel antimicrobial peptide in the N1-domain of bovine lactoferrin, Peptides 25 (2004) 177e183. [25] M.I.A. van der Kraan, C. van der Made, K. Nami, W. Van’t Hof, J. Groenink, E.C.I. Veerman, J.G.M. Bolscher, A.V.N. Amerongen, Effect of amino acid substitutions on the candidacidal activity of LFampin 265e284, Peptides 26 (2005) 2093e2097. [26] M.I.A. van der Kraan, K. Nazmi, A. Teeken, J. Groenink, W. ’t Hof, E.C.I. Veerman, J.G.M. Bolscher, A.V.N. Amerongen, Lactoferrampin, an antimicrobial peptide of bovine lactoferrin, exerts its candidacidal activity by a cluster of positively charged residues at the C-terminus in combination with a helix-facilitating N-terminal part, Biol. Chem. 386 (2005) 137e142. [27] E.F. Haney, F. Lau, H.J. Vogel, Solution structures and model membrane interactions of lactoferrampin, an antimicrobial peptide derived from bovine lactoferrin, Biochim. Biophys. Acta 1768 (2007) 2355e2364. [28] M.I.A. van der Kraan, K. Nazmi, W. ’t Hof, A.V.N. Amerongen, E.C.I. Veerman, J.G.M. Bolscher, Distinct bactericidal activities of bovine lactoferrin peptides LFampin 268e284 and LFampin 265e284: Asp-Leu-Ile makes a difference, Biochem. Cell Biol. 84 (2006) 358e362. [29] N.A. Fonseca, R. Camacho, A.L. Magalhaes, Amino acid pairing at the N- and C-termini of helical segments in proteins, Proteins 70 (2008) 188e196. [30] E.C.I. Veerman, K. Nazmi, W. van Hof, J.G.M. Bolscher, A.L. den Hertog, A.V.N. Amerongen, Reactive oxygen species play no role in the candidacidal activity of the salivary antimicrobial peptide histatin 5, Biochem. J. 381 (2004) 447e452. [31] E.J. Prenner, R.N.A.H. Lewis, L.H. Kondejewski, R.S. Hodges, R.N. McElhaney, Differential scanning calorimetric study of the effect of the antimicrobial peptide gramicidin S on the thermotropic phase behavior of phosphatidylcholine, phosphatidylethanolamine and phosphatidylglycerol lipid bilayer membranes, Biochim. Biophys. Acta 1417 (1999) 211e223. [32] T.L. Hwang, A.J. Shaka, Water suppression that works e excitation sculpting using arbitrary wave-forms and pulsed-field gradients, J. Magn. Reson. A 112 (1995) 275e279. [33] F. Delaglio, S. Grzesiek, G.W. Vuister, G. Zhu, J. Pfeifer, A. Bax, Nmrpipe e a multidimensional spectral processing system based on unix pipes, J. Biomol. NMR 6 (1995) 277e293. [34] B.A. Johnson, R.A. Blevins, Nmr view e a computer-program for the visualization and analysis of NMR data, J. Biomol. NMR 4 (1994) 603e614.
[35] K. Wuthrich, NMR of Proteins and Nucleic Acids, John Wiley & Sons Inc., New York, 1986. [36] P. Guntert, C. Mumenthaler, K. Wuthrich, Torsion angle dynamics for NMR structure calculation with the new program DYANA, J. Mol. Biol. 273 (1997) 283e298. [37] R.A. Laskowski, J.A.C. Rullmann, M.W. MacArthur, R. Kaptein, J.M. Thornton, AQUA and PROCHECK-NMR: programs for checking the quality of protein structures solved by NMR, J. Biomol. NMR 8 (1996) 477e486. [38] R. Koradi, M. Billeter, K. Wuthrich, MOLMOL: a program for display and analysis of macromolecular structures, J. Mol. Graph. 14 (1996) 51. [39] K. Lohner, The role of membrane lipid composition in cell targeting of antimicrobial peptides, in: Lohner (Ed.), Development of Novel Antimicrobial Agents: Emerging Strategies, Horizon Scientific Press, Norfolk, England, 2001, pp. 149e165. [40] G.N. Ramachandran, C. Ramakrishnan, V. Sasisekharan, Stereochemistry of polypeptide chain configurations, J. Mol. Biol. 7 (1963) 95. [41] H.J. Vogel, D.J. Schibli, W.G. Jing, E.M. Lohmeier-Vogel, R.F. Epand, R.M. Epand, Towards a structureefunction analysis of bovine lactoferricin and related tryptophan- and arginine-containing peptides, Biochem. Cell Biol. 80 (2002) 49e63. [42] Y.J. Xiao, H. Dai, Y.R. Bommineni, J.L. Soulages, Y.X. Gong, O. Prakash, G.L. Zhang, Structureeactivity relationships of fowlicidin1, a cathelicidin antimicrobial peptide in chicken, FEBS J. 273 (2006) 2581e2593. [43] H. Wong, J.H. Bowie, J.A. Carver, The solution structure and activity of caerin 1.1, an antimicrobial peptide from the Australian green tree frog, Litoria splendida, Eur. J. Biochem. 247 (1997) 545e557. [44] T.L. Pukala, C.S. Brinkworth, J.A. Carver, J.H. Bowie, Investigating the importance of the flexible hinge in caerin 1.1: solution structures and activity of two synthetically modified caerin peptides, Biochemistry 43 (2004) 937e944. [45] T. Yuan, A.M. Weljie, H.J. Vogel, Tryptophan fluorescence quenching by methionine and selenomethionine residues of calmodulin: orientation of peptide and protein binding, Biochemistry 37 (1998) 3187e3195. [46] V.V. Andrushchenko, H.J. Vogel, E.J. Prenner, Interactions of tryptophanrich cathelicidin antimicrobial peptides with model membranes studied by differential scanning calorimetry, Biochim. Biophys. Acta 1768 (2007) 2447e2458. [47] E.J. Prenner, R.N.A.H. Lewis, K.C. Neuman, S.M. Gruner, L.H. Kondejewski, R.S. Hodges, R.N. McElhaney, Nonlamellar phases induced by the interaction of gramicidin S with lipid bilayers. A possible relationship to membrane-disrupting activity, Biochemistry 36 (1997) 7906e7916. [48] B. Klajnert, J. Janiszewska, Z. Urbanczyk-Lipkowska, A. Bryszewska, R.M. Epand, DSC studies on interactions between low molecular mass peptide dendrimers and model lipid membranes, Int. J. Pharm. 327 (2006) 145e152. [49] S.E. Blondelle, K. Lohner, M.I. Aguilar, Lipid-induced conformation and lipid-binding properties of cytolytic and antimicrobial peptides: determination and biological specificity, Biochim. Biophys. Acta 1462 (1999) 89e108. [50] M. Abdi, D.B. Drucker, Phospholipid profiles in the oral yeast Candida, Arch. Oral Biol. 41 (1996) 517e522. [51] G.S. Wang, Y.F. Li, X. Li, Correlation of three-dimensional structures with the antibacterial activity of a group of peptides designed based on a nontoxic bacterial membrane anchor, J. Biol. Chem. 280 (2005) 5803e5811.
