Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Production, characterization, and elucidation of structure-function relationship of lactoferricin and… Chan, Judy Chuk Kwan 2006

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2006-199503.pdf [ 8.12MB ]
Metadata
JSON: 831-1.0092929.json
JSON-LD: 831-1.0092929-ld.json
RDF/XML (Pretty): 831-1.0092929-rdf.xml
RDF/JSON: 831-1.0092929-rdf.json
Turtle: 831-1.0092929-turtle.txt
N-Triples: 831-1.0092929-rdf-ntriples.txt
Original Record: 831-1.0092929-source.json
Full Text
831-1.0092929-fulltext.txt
Citation
831-1.0092929.ris

Full Text

P R O D U C T I O N , CHARACTERIZAT ION, AND ELUCIDATION OF S T R U C T U R E - F U N C T I O N RELATIONSHIP OF LACTOFERRICIN AND O T H E R PEPTIDES DERIVED F R O M F O O D - G R A D E BOVINE L A C T O F E R R I N by J U D Y C H U K K W A N C H A N B. Sc. , The University of British Columbia, 1997 M. Sc. , The University of British Columbia, 1999 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF T H E R E Q U I R E M E N T S F O R T H E D E G R E E OF D O C T O R OF PHILOSOPHY in T H E F A C U L T Y OF G R A D U A T E STUDIES (Food Science) T H E UNIVERSITY OF BRITISH COLUMBIA September 2006 © Judy Chuk Kwan Chan, 2006 Abstract Lactoferricin (Lfcin), a cationic antimicrobial peptide, was purified by peptic digestion of food grade bovine lactoferrin (LF) followed by fractionation on an industrial grade cation exchange resin with stepwise salt gradient elution. A 2-step process using competitive displacement cation exchange chromatography led to 35% recovery of peptides with masses of 3124 and 3196 Da corresponding to Lfcin. Other cationic peptides produced concurrently with Lfcin were tentatively identified. Varying iron saturation levels of LF had no effect on the production of Lfcin. Two approaches were used to understand how Lfcin acts as an antimicrobial agent. Homology Similarity Analysis (HSA) was used to evaluate the impact of peptide pattern similarities of various Lfcin derivatives on their effects as antimicrobial agents. Helical property of residues 4-9 in the Lfcin sequence was the most important in determining the antimicrobial activity of Lfcin against Escherichia coli, followed by cationic charge pattern of residues 4-9 and residues 1-3. Raman spectroscopy in the C-H stretching (2800-3000 cm"1) region was used to study interactions between Lfcin and bacterial membrane models. Presence of Lfcin in 1,2-dipalmitoyl-sn-glycerol-3-phosphocholine (DPPC) multilamellar liposomes restricted the lateral chain-chain interaction along the acyl chain throughout the temperature range examined, and increased the main transition temperature (Tm) from 38-40 °C. Lfcin had little effect on 1,2-dipalmitoyl-sn-glycerol-3-[phosphor-rac-(1-glycerol)] (DPPG) liposomes at temperatures below the T m . In contrast, mobility of the acyl chains in 1,2-dipalmitoyl-sn-glycerol-3-phosphoethanolamine (DPPE) was increased in the presence of Lfcin at temperatures below 32 °C. Although Lfcin had little effect on the zwitterionic unsaturated phospholipids, 1-palmitoyl-2-oleoyl-sn-glycerol-3-phosphocholine (POPC) and 1-palmitoyl-2-oleoyl-sn-glycerol-3-phosphoethanolamine (POPE) , Lfcin lowered the lateral chain-chain interaction along the acyl chains of negatively charged 1-palmitoyl-2-oleoyl-sn-glycerol-3-[phosphor-rac-(1-glycerol)] (POPG) . Raman spectral analyses also showed that the removal of arginine promoted interactions of Lfcin derivatives with hydrophobic acyl chains and methyl ends of negatively charged D P P G liposomes. Furthermore, addition of two extra tryptophan residues in Lfcin derivatives stabilized acyl chains and methyl ends in D P P E liposomes. HSA and Raman spectroscopy are two useful tools for understanding the antimicrobial mechanisms of Lfcin and will facilitate the application and utilization of Lfcin as a naturally occurring antimicrobial agent. Table of Contents Abstract .'. ii Table of Contents iii List of Tables vii List of Figures ; ix List of Abbreviations xi Acknowledgements xii Dedication ....xiii Co-Authorship Statements xiv CHAPTER 1. GENERAL INTRODUCTION AND LITERATURE REVIEW 1 1.1. Lactoferrin 1 1.1.1. Sources of Lactoferrin 1 1.1.2. Structure of Lactoferrin 1 1.1.3. Biological Functions of Lactoferrin 2 1.1.3.1. Antimicrobial Properties of Lactoferrin 2 1.1.3.2. Antiviral Functions of Lactoferrin 3 1.1.3.3. Other Biological Functions of Lactoferrin 3 1.1.3.4. Other Functions of Lactoferrin 4 1.2. Lactoferricin & Its Derivatives 5 1.2.1. Production and Purification of Lactoferricin and Related Peptides 5 1.2.1.1 Discovering Lactoferricin 5 1.2.1.2. Identifying Lactoferricin 5 1.2.1.3. Production and Purification Strategies 6 1.2.2. Antimicrobial Spectrum of Lactoferricin and Related Peptides 8 1.2.3. Structure - Function Relationship Analysis of Lactoferricin and Its Derivatives 8 1.2.3.1. Basic Structural Properties of Bovine Lactoferricin 8 1.2.3.2. Roles of Secondary Structure 9 1.2.3.3. Roles of Basic Amino Acids 10 1.2.3.4. Roles of Hydrophobic Amino Acids 11 1.3. Cationic Antimicrobial Peptides 12 1.3.1. Structures and Modes of Action of Other Cationic Antimicrobial Peptides 12 1.3.2. Proposed Antimicrobial Mechanisms of Lactoferricin 13 1.4. Phospholipid Bilayer 14 1.5. Raman Spectroscopy 15 1.5.1. Raman Spectroscopy: Structure - Function Relationship of Proteins 15 1.5.2. Raman Spectroscopy: Structure - Function Relationship of Peptides and Phospholipid Bilayers 16 1.6. Multivariate Data Analysis Techniques 17 1.6.1. Principal Component Analysis 18 1.6.2. Partial Least Square Regression 18 1.6.3. Principal Component Similarity Analysis 18 1.6.4. Multivariate Data Analysis: Structure - Function Relationship of Lactoferricin and Its Derivatives 20 1.7. Hypotheses 22 1.8. Objectives 22 1.9. References 30 CHAPTER 2. PRODUCTION OF LACTOFERRICIN AND OTHER CATIONIC PEPTIDES FROM FOOD GRADE BOVINE LACTOFERRIN WITH VARYING IRON SATURATION LEVELS - 3 9 2.1. Introduction 39 2.2. Materials and Methods 40 2.2.1. Preparation of Lactoferrin with Varying Iron Saturation Levels 40 2.2.2. Digestion of Lactoferrin 40 2.2.3. Purification of Lactoferricin and Other Peptides 40 2.2.4. Peptide Mass Analysis 41 2.2.5. Identification of Peptide Sequences 41 2.2.6. Iron Content in Lactoferrins 41 2.2.7. Protein Structure by Raman Spectroscopy 41 2.3. Results 42 2.3.1. Production of Lactoferricin and Other Cationic Peptides from Lactoferrin 42 2.3.2. A Two-Step Process for Lactoferricin Isolation 42 2.3.3. The Identification of Lactoferricin and Other Cationic Peptides....! 43 2.3.4. Effects of Iron Saturation Levels of Lactoferrin 44 2.4. Discussion 44 2.4.1. Production of Lactoferricin and Other Cationic Peptides from Lactoferrin 44 2.4.2. Identification of Other Cationic Bioactive Peptides 46 2.4.3. Effects of Iron Saturation Levels of Lactoferrin 47 2.5. References 60 CHAPTER 3. HOMOLOGY SIMILARITY ANALYSIS OF SEQUENCES OF LACTOFERRICIN AND ITS DERIVATIVES 62 3.1. Introduction 62 3.2. Materials and Methods 63 3.2.1. Lactoferricin Sequences 63 3.2.2. Peptide Property Scales . . .63 3.2.3. Homology Similarity Analysis (HSA) 63 3.2.4. Principal Component Similarity (PCS) Analysis 64 3.2.5. Artificial Neural Networks (ANN) 64 3.3. Results 64 3.3.1. Homology Similarity Analysis of Lactoferricin Sequences 64 3.3.2. Principal Component Similarity Analysis 65 3.3.3. Regression Analysis Using Artificial Neural Networks 66 3.4. Discussion 66 3.5. References 81 CHAPTER 4. RAMAN SPECTROSCOPIC STUDY OF THE EFFECTS OF LACTOFERRICIN ON THERMAL TRANSITIONS OF PHOSPHOLIPIDS IN LIPOSOMES 83 4.1. Introduction 83 4.2. Materials and Methods 84 4.2.1. Materials 84 4.2.2. Preparation of Liposomes 85 4.2.3. Raman Spectroscopy 85 4.2.4. Raman Spectral Analysis 85 4.3. Results 86 4.4. Discussion 88 4.4.1. Temperature-induced Changes in Raman Spectra of Model Membranes 88 4.4.2. Effects of Lactoferricin on Model Membranes 91 4.5. References 98 CHAPTER 5. RAMAN SPECTROSCOPIC STUDY OF THE EFFECTS OF LACTOFERRICIN DERIVATIVES ON PHOSPHOLIPID LIPOSOMES 101 5.1. Introduction 101 5.2. Materials and Methods 102 5.2.1. Materials , 102 5.2.2. Preparation of Liposomes ...103 5.2.3. Raman Spectroscopy 103 5.2.4. Raman Spectral Analysis 103 5.2.5. Properties of Lactoferricin Derivatives 103 5.3. Results 103 5.3.1. Effects of Lactoferricin Derivatives on D P P C Liposomes 103 5.3.2. Effects of Lactoferricin Derivatives on D P P G Liposomes 104 5.3.3. Effects of Lactoferricin Derivatives on D P P E Liposomes 105 5.4. Discussion 106 5.4.1. Effects of Arginine on Lactoferricin on the Thermal Behaviour of Phospholipid Liposomes 106 5.4.2. Effects of Tryptophan on the Thermal Behaviour of Lactoferricin of Phospholipid Liposomes 107 5.4.3. Effects of Peptide Length on the Thermal Behaviour of Lactoferricin of Phospholipid Liposomes 109 5.4.4. Proposed Antimicrobial Actions 110 5.5. References 117 CHAPTER 6. CONCLUSION 120 6.1. Review of Results as Related to Working Hypotheses 120 6.2. Strengths and Significance of Results to the Field of Study 121 6.3. Recommendations for Future Studies 121 6.4. References 123 v i List of Tables Table 1.1. Percentage of phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylglycerol (PG), and 1,3-bis(sn-3-phosphatidyl)-sn-glycerol (cardiolipin, CL) in bacterial membranes 29 Table 2.1. Possible cationic peptides produced from the peptic digestion of bovine lactoferrin at 37 °C for 4 hours at pH 2.0. Cationic amino acids are indicated in bold 58 Table 2. 2. Secondary structure of bovine lactoferrin preparations (LF) at pH 2.0 as determined using Raman spectroscopy in this study and human lactoferrin (LFH) as reported in the literature 59 Table 3.1. Homology similarity coefficient and average property values for helix (Hx), charge (Ch), hydrophobicity (Hp) and bulkiness (Bk) calculated for three segments in the sequences of Group I lactoferricin derivatives, with minimum inhibitory concentrations (MIC) against E. coli as reported in the literature. Segment I = position 1-3, Segment II = position 4-9, and Segment III = position 10-15 in the reference (Peptide 13) with 15-amino acid sequence identical to bovine lactoferricin. Shaded zones in the sequences represent regions with high similarity to the reference peptide, while the shading of Hx II indicates the high similarity coefficient for helix propensity in Segment II for Group I derivatives. Peptide #13 was used as the reference peptide and its values were shaded for easy comparison 69 Table 3.2. Homology similarity coefficient and average property values for helix (Hx), charge (Ch), hydrophobicity (Hp) and bulkiness (Bk) calculated for three segments in the sequences of Group II lactoferricin derivatives, with minimum inhibitory concentrations (MIC) against E. coli as reported in the literature. Segment I = position 1-3, Segment II = position 4-9, and Segment III = position 10-15 in the reference (Peptide 13) with 15-amino acid sequence identical to bovine lactoferricin. Shaded zones in the sequences represent regions with high similarity to the reference peptide, while the shading of Ch I indicates the high average values for charge in Segment I characteristic of Group II derivatives 71 Table 3.3. Homology similarity coefficient and average property values for helix (Hx), charge (Ch), hydrophobicity (Hp) and bulkiness (Bk) calculated for three segments in the sequences of Group III lactoferricin derivatives, with minimum inhibitory concentrations (MIC) against E. coli as reported in the literature. Segment I = position 1-3, Segment II = position 4-9, and Segment III = position 10-15 in the reference (Peptide 13) with 15-amino acid sequence identical to bovine lactoferricin. Shaded zones in the sequences represent regions with high similarity to the reference peptide. Group III derivatives included exceptional cases with unexplained high or low MIC values 72 Table 3.4. Homology similarity coefficient and average propensity values for helix (Hx), charge (Ch), hydrophobicity (Hp) and bulkiness (Bk) calculated for three segments in the sequences of Group IV lactoferricin derivatives, with minimum inhibitory concentrations (MIC) against E. coli as reported in the literature. Segment I = position 1-3, Segment II = position 4-9, and Segment III = position 10-15 in the reference (Peptide 13) with 15-amino acid sequence identical to bovine lactoferricin. Shaded zones in the sequences represent regions with high similarity to the reference peptide, while the shading of Ch II indicates low similarity coefficient and average values of charge in Segment II of Group IV derivatives 73 Table 3.5. Homology similarity coefficient and average property values for helix (Hx), charge (Ch), hydrophobicity (Hp) and bulkiness (Bk) calculated for three segments in the sequences of Group V lactoferricin derivatives, with minimum inhibitory concentrations (MIC) against E. coli as reported in the literature. Segment I = position 1-3, Segment II = position 4-9, and Segment III = position 10-15 in the reference (Peptide 13) with 15-amino acid sequence identical to bovine lactoferricin. Shaded zones in the sequences represent regions with high similarity to the reference peptide, while the shading of Hxll indicates low similarity coefficient values for helix propensity in Segment II of Group V derivatives 75 Table 3.6. Amino acid structural propensity values used for Homology Similarity Analysis 77 78 Table 5.1. Structural properties of lactoferricin derivatives used in the study. 116 List of Figures Figure 1.1. Primary amino acid sequence (single letter code) of bovine lactoferrin. * denotes residues 17 to 42, representing the antimicrobial peptide, lactoferricin. + denotes the major contributors to one of the two iron-binding sites, and A denotes the major contributors to the second iron-binding site (adapted from Pierce et al., 1991) 23 Figure 1.2. Three-dimensional structure of diferric bovine lactoferrin at 2.8 A resolution where the arrows indicate the location of bound iron (Reprinted from Moore et al., 1997, Copyright 1997, with permission from Elsevier) '.. 24 Figure 1.3. Antimicrobial peptide fragments isolated from bovine lactoferrin as reported in the literature; Peptides I, II, III are from Dionysius and Milne (1997) and Peptides, 1a, 1b, 2, and 3 are from Hoek et al. (1997) 25 Figure 1.4. Hydrophobic (a), hydrophilic (b), and positive (c) side chains along the B-strands of bovine lactoferricin. The backbone of all residues is shown in gray, while the highlighted side chains are in black (Reprinted from Hwang et al., 1998, Copyright 1998, with permission from American Chemical Society) 26 Figure 1.5. Structural representation of the arrangement of phospholipids. Arrangement of the head group and fatty acid domains of phospholipids in the fluid liquid crystalline state (La), the ordered gel state (LB), and the reversed hexagonal state (HM) (Reprinted from Dowhan, 1997, Copyright 1997, with permission from the Annual Reviews www.annualreviews.org) 27 Figure 1.6. Raman spectra of dipalmitoylphosphatidylglycerol (DPPG) in the C-H stretching mode region (2800-3000 cm"1) between 25 and 55 °C (a). Peak height intensity ratios (b): bsso/bsso (A) and I2935/I2880 (B) (adapted from Vincent et al., 1993) 28 Figure 2.1. MALDI-ToF mass spectrum of lactoferrin digest produced by peptic digestion for 4 hours at 37 °C at pH 2.0 50 Figure 2.2. Purolite C-106 cation exchange chromatography of peptic LF digest (Solid line, Absorbance at 280nm; dotted line, conductivity). The supernatant of a 0.5 g of peptic LF digest was loaded onto a 10 mL column and eluted with 50 mM sodium phosphate buffer (pH 6.5) containing 0, 0.2, 0.5, 0.8, 1.0, and 2.0 M of NaCI at a flow rate of 1.0 mL per minute. Fraction peaks were collected and further analyzed 51 Figure 2.3. MALDI-ToF mass spectra of LF digest fractions eluted from a 10 mL cation exchange column with Purolite C-106 resin using 50 mL phosphate buffer at pH 6.5 containing no NaCI (a), 0.2 M NaCI (b), 0.5 M NaCI (c), 0.8 M NaCI (d), 1.0 M NaCI (e), and 2.0 M NaCI (f) 52 Figure 2.4. Purolite C-106 cation exchange chromatography in competitive displacement mode of peptic LF digest (a: Solid line, Absorbance at 280nm; dotted line, conductivity). The supernatant of a 5.0 g of peptic LF digest was loaded onto a 5.0 mL column and eluted with 50 mM sodium phosphate buffer (pH 6.5) containing 0, 1.0, and 2.0 M NaCI (a). Fraction peaks were collected and further analyzed by MALDI-ToF (b, i: fraction collected after sample loading was completed; ii: 1.0 M NaCI fraction; iii: 2.0 M NaCI fraction.) 53 Figure 2.5. MALDI-ToF post source decay mass spectrum of peak with mass of 3196 Da (a) and possible peptides that were resulted from the post source decay (b) 54 Figure 2.6. MALDI-ToF mass spectra of lactoferrin digest resulting from peptic digestion of apo-LF (a) and holo-LF (b) 55 Figure 2.7. MALDI-ToF mass spectra of 1.0 M NaCI fractions resulting from peptic digestion of apo-LF (a) and holo-LF (b) 56 Figure 2.8. MALDI-ToF mass spectra of 2.0 M NaCI fractions resulting from peptic digestion of apo-LF (a) and holo-LF (b). 57 Figure 3.1. Principal component similiarity (PCS) scattergram of lactoferricin derivatives using data obtained from homology similiarity analysis (HSA). Group l-V are listed in Tables 3.1 - 3 . 5 79 Figure 3.2. Plots of predicted values vs. observed values by artifical neural network: (a) Log MIC and (b) Reciprocal MIC 80 Figure 4.1. Raman spectra of dipalmitoylphosphatidylcholine (DPPC) liposomes in the presence of lactoferricin showing changes in the major peaks near 2850, 2880, and 2935 cm' 1 as a function of temperature 94 Figure 4.2. bsso/bsso ratio of D P P C (a), I2935/I2880 ratio of D P P C (b), bsso/bsso ratio of P O P C (c), and I2935/I2880 ratio of P O P C (d) in the absence (O) or presence ( • ) of Lfcin measured between -10 °C and 70 °C by Raman spectroscopy 95 Figure 4.3. bsso/bsso ratio of D P P G (a), I2935/I2880 ratio of D P P G (b), bsso/bsso ratio of P O P G (c), and I2935/I2880 ratio of P O P G (d) in the absence (O) or presence ( • ) of Lfcin measured between -10 °C and 70 °C by Raman spectroscopy 96 Figure 4.4. bsso/bsso ratio of D P P E (a), I2935/I2880 ratio of D P P E (b), bsso/bsso ratio of P O P E (c), and I2935/I2880 ratio of P O P E (d) in the absence (O) or presence ( • ) of Lfcin measured between -10 °C and 70 °C by Raman spectroscopy 97 Figure 5.1. Thermal behaviours of D P P C as indicated by bsso/bsso (a) and I2935/I2880 (b) in the absence of lactoferricin (Lfcin, • ) , the presence of L fc in 1 5 (i, • ) , Lfc in 1 5 R- (ii, • ) , Lfc in 1 5 W- (iii, • ) , Lfcin 1 5W+ (iv, X ) , and LfcinCore (v, * ) measured between -10 °C and 70 °C by Raman spectroscopy 113 Figure 5.2. Thermal behaviours of D P P G as indicated by bsso/bsso (a) and I2935/I2880 (b) in the absence of lactoferricin (Lfcin, O), the presence of Lfc in 1 5 (i, • ) , Lfc in 1 5 R- (ii, • ) , Lfc in 1 5 W- (iii, A ) , Lfcin 1 5W+ (iv, X ) , and LfcinCore (v, * ) measured between -10 °C and 70 °C by Raman spectroscopy 114 Figure 5.3. Thermal behaviours of D P P E as indicated by bsso/bsso (a) and I2935/I2880 (b) in the absence of lactoferricin (Lfcin, O), the presence of L fc in 1 5 (i, • ) , Lfc in 1 5 R- (ii, • ) , Lfc in 1 5 W- (iii, A ) , Lfcin 1 5W+ (iv, X ) , and LfcinCore (v, * ) measured between -10 °C and 70 °C by Raman spectroscopy 115 List of Abbreviations ANN Artifical Neural Network C A P Cationic Antimicrobial Peptide D P P C 1,2-dipalmitoyl-sn-glycerol-3-phosphocholine D P P E 1,2-dipalmitoyl-sn-glycerol-3-phosphoethanolamine D P P G 1,2-dipalmitoyl-sn-glycerol-3-[phospho-rac-(1-glycerol)] HSA Homology Similarity Analysis LF Lactoferrin from Bovine, unless otherwise stated LFB Lactoferrin from Bovine Lfcin Lactoferricin from Bovine, unless otherwise stated Lfc in 1 5 Lactoferricin derivative with the following sequence: F K C R R W Q W R M K K L G A Lfc in 1 5 R- Lactoferricin derivative with the following sequence: F K C A A W Q W A M K K L G A Lfc in 1 5 W- Lactoferricin derivative with the following sequence: F K C R R A Q A R M K K L G A Lfcin 1 5W+ Lactoferricin derivative with the following sequence: F K C W R W Q W R W K K L G A LfcinB Lactoferricin from Bovine LfcinCore Lactoferricin derivative with the following sequence: R R W Q W R-NH 2 LfcinH Lactoferricin from Human LFH Lactoferrin from Human MIC Minimum Inhibitory Concentration P C A Principal Component Analysis P C S Principal Component Similarity P L S Partial Least Square P O P C 1 -palmitoyl-2-oleoyl-sn-glycerol-3-phosphocholine P O P E 1-palmitoyl-2-oleoyl-sn-glycerol-3-phosphoethanolamine P O P G 1-palmitoyl-2-oleoyl-sn-glycerol-3-[phospho-rac-(1-glycerol)] Acknowledgements Many, many thanks go to many people who worked with me in the Department. They listened when I faced hurdles during my research progress; they cheered when I made important breakthroughs. They know who they are and they understand why they are not named here. Sincere thanks go to my supervisor, Dr. Eunice Li-Chan, who is knowledgeable, dedicated, and caring. It is a true joy to be her student. Funding for this research was provided by the Dairy Farmers of Canada and the Natural Sciences and Engineering Research Council of Canada. J C K C was the recipient of the University Graduate Fellowship provided by the University of the British Columbia. x i i This thesis is dedicated to: Dad who encourages me to follow my dreams Mom who makes sure that I study hard so that I can follow my dreams Dennis who is always there with me when I follow my dreams Co-Authorship Statements Part of the work presented in Chapter 2 was submitted for publication in the Journal of Agricultural and Food Chemistry, titled "Production of lactoferricin and other cationic peptides from food grade bovine lactoferrin with varying iron saturation levels". J C K Chan, the thesis author, was the principal author who identified and designed the research program, performed the research, analyzed data, and prepared the manuscript. E C Y Li-Chan, J C K Chan's supervisor, was the co-author and provided close guidance on the above research. Part of the work presented in Chapter 3 was published in the Journal of Agricultural and Food Chemistry, 2003: 51, 1215-1223, titled "Homology similarity analysis of sequences of lactoferricin and its derivatives". S Nakai, member of J C K ' s supervisory committee, was the principal author who identified and designed the research program. J C K Chan, the thesis author, assisted in data collection, verified data analyses, and aided in manuscript preparation. J Dou, co-author of the study, developed the computer programs used in the study. E C Y Li-Chan and M Ogawa, co-authors of the study, provided guidance on the design of the research program and manuscript preparation. Part of the work presented in Chapter 4 will be submitted for publication in Biochimica et Biophysica Acta (BBA) - Biomembranes, titled "Raman spectroscopic study of the effects of lactoferricin on thermal transitions of phospholipids in liposomes". J C K Chan, the thesis author, was the principal author who identified and designed the research program, performed the research, analyzed data, and prepared the manuscript. E C Y Li-Chan, J C K Chan's supervisor, was the co-author and provided close guidance on the above research. Part of the work presented in Chapter 5 will be submitted for publication in Biochimica et Biophysica Acta (BBA) - Biomembranes, titled "Raman spectroscopic study of the effects of lactoferricin derivatives on phospholipid liposomes". J C K Chan, the thesis author, was the principal author who identified and designed the research program, performed the research, analyzed data, and prepared , the manuscript. E C Y Li-Chan, J C K Chan's supervisor, was the co-author and provided close guidance on the above research. J C K Chan E C Y Li-Chan x i v CHAPTER 1. GENERAL INTRODUCTION AND LITERATURE REVIEW Lactoferrin (LF) is an iron-binding protein in bovine whey, with a broad spectrum of antimicrobial activity (Bellamy et al., 1992b). Studies have shown that pepsin digestion of bovine LF yields a peptide fragment that has more potent activity than undigested LF (Tomita et al., 1991). This fragment is named bovine lactoferricin (Lfcin) and is a 25-amino acid peptide corresponding to residues 17-41 from the N-terminal loop region of LF with the sequence F K C R R W Q W R M K K L G A PSITC V R R A F (Bellamy et al., 1992b). This 25-amino acid peptide includes 8 basic amino acids that contribute to a net charge of +6.85 at pH 7 and forms an anti-parallel, amphipathic, B-sheet structure (Rekdal et al., 1999; Hwang et al., 1998). Lfcin has been the subject of intense research in past decades in regard to its antimicrobial and immunomodulatory activities, as well as its inhibitory effects on carcinomas and tumor metastasis (Gifford et al., 2005; Mader et al., 2005). 1.1. Lactoferrin 1.1.1. Sources of Lactoferrin Lactoferrin (LF), a member of the transferrin family of iron-binding proteins, was first reported in bovine milk in 1939 (Sorensen and Sorensen, 1939). Since then, LF has been found in and isolated from various mammalian milks including human (LFH) (Johansson, 1960), bovine (LFB) (Groves, 1960), and camel (Kappeler et al., 1999; Masson and Heremans, 1971). LF levels in milks depend on the species, as well as the lactation period. The amount of LF is 5 to 7 mg • mL"1 in human colostrum and 1 to 3 mg • mL"1 in mature milk (Hennart et al., 1991), while the content of LF in bovine milk is markedly lower at 0.8 mg • mL"1 in the colostrum and 0.1 to 0.4 mg • mL"1 in mature milk (Sanchez et al., 1988). In addition to mammalian milks, LF has also been found in various exocrine secretions such as tears, seminal fluid, cervical mucus, and saliva (Levay and Viljoen, 1995). Furthermore, LF was detected in the lymphatic system (Beljaars et al., 2002). 1.1.2. Structure of Lactoferrin Bovine lactoferrin consists of a single polypeptide chain of 689 amino acids with a molecular weight of 76,000 Da (Figure 1.1) (Pierce et al., 1991). Crystallographic analyses have revealed that both bovine (Moore et al., 1997) and human (Anderson et al., 1987 and 1989) lactoferrins are composed of two globular lobes of similar size which are connected by a single short bridging peptide (Figure 1.2). These studies demonstrated that LFB is an a/B protein containing approximately 4 1 % a-helix and 24% B-sheet. Results from Fourier transform infrared (FTIR) spectroscopic analysis of LFH also showed similar results; the secondary 1 structural content of LFH was estimated to be approximately 44% a-helix, 28% B-sheet, and 21 % turns (Hadden et al., 1994). Crystallographic studies indicated that each lobe of the LF consists of two domains that form a cleft enclosing the binding site for a metal ion (Moore et al., 1997). Hence, each LF molecule has the ability to bind reversibly two ferric ions (Fe 3 + ) per molecule (Moore et al., 1997). In addition to these specific iron binding sites, it has been reported that LFB was capable of binding 700 times the amount of iron that could be bound to the specific binding sites (Nagasako et al., 1993). It was postulated that LFB binds iron at sites other than the specific binding sites, possibly on the surface of the globular protein (Nagasako et al., 1993). The binding of iron to lactoferrin is an important factor in the thermal stability of its structure. Paulsson et al. (1993) reported that iron-saturated (holo-) bovine lactoferrin was more resistant to heat-induced changes than was the iron-depleted (apo-) LF. Results from differential scanning calorimetry demonstrated that lactoferrin isolated from human milk had a maximum peak temperature at 67.0 °C, and transition enthalpy, and activation energy of 2276 kJ • mol"1, and 275.5 kJ • mol"1, respectively, while holo-lactoferrin saturated with iron had a maximum peak temperature at 90.6 °C, a transition enthalpy of 3209 kJ • mol"1, and an activation energy of 387.6 kJ • mol"1 (Mata et al., 1998). Although results from a FTIR spectroscopy study showed that iron status in LF did not affect the secondary structural content, i.e. a-helix, B-sheet, and turns, of LFH, removal of iron from LFH led to increases in the extent of 1 H - 2 H exchange (Hadden et al., 1994). Hadden et al. (1994) postulated a "loosening" of LF upon iron removal. This could explain the lower heat stability demonstrated by apo-lactoferrin. Bovine LF possesses five N-linked glycosylation sites at Asn-233, -281, -368, -476, and -545 (Pierce et al., 1991). The carbohydrate moieties of bovine LF consist of complex and high-mannose type glycans and the relative proportions of these glycans vary with the lactation period (Hutchens et al., 1994). It was reported that the structures of these glycans are highly heterogeneous (Wei et al., 2000). 1.1.3. Biological Functions of Lactoferrin 1.1.3.1. Antimicrobial Properties of Lactoferrin A physiologically diverse range of pathogenic Gram-positive and Gram-negative bacteria was reported to be susceptible to inhibition and inactivation by bovine lactoferrin (Antoniniet al., 1997; Bellamy et al., 1992a and 1992b; Ellison and Giehl, 1991; Tomita et al., 1991 and 1994; Yamauchi et al., 1993; Yamazaki et al., 1997). In addition, LF was used to enhance the effectiveness of antifungal agents against various Candida species (Wakabayashi 2 et al., 1996 and 1998). LF also inhibited Trichophyton mentagrophytes in a guinea pig model with dermatophytosis (Wakabayashi et al., 2000). Several early studies demonstrated that the antimicrobial property of LF is iron dependent. Nonnecke and Smith (1984) reported that the addition of ferric iron to the assay system diminished the inhibitory activity of apo-LF. The antibacterial effect of LF on enterotoxigenic Escherichia coli was also affected by the iron status of the LF. The presence of 1.0 mg • mL"1 of holo-LF did not inhibit the growth of any of the 19 E. coli strains tested. On the contrary, the presence of apo-LF acted against all strains tested (Dionysius and Milne, 1997). Hence, the strong iron-binding capacity of LF could be responsible for the antimicrobial actions of LF simply by removing iron from the environment and thus limiting microbial growth (Bullen et al., 1972). Nevertheless, increasing evidence has suggested that additional mechanisms are involved in the antimicrobial actions of LF. For instance, bactericidal effects exhibited by apo-LF could not be reversed by the addition of exogenous iron in excess of its chelating capacity and such antimicrobial effects could not be reproduced by exposing the microorganisms to iron-deficient medium alone (Arnold et al., 1980 and 1982). Furthermore, immunofluorescence studies have shown that LF bound directly to the surface of microorganisms, causing damage to the outer membrane and the release of lipopolysaccharides (Ellison et al., 1988; Ellison and Giehl, 1991). 1.1.3.2. Antiviral Functions of Lactoferrin Lactoferrin is capable of inhibiting replication of a wide range of viruses, namely, hepatitis C (Ikeda et al., 1998, 2000), rotavirus (Superti et al., 1997), poliovirus (Marchetti et al., 1999), human immunodeficiency virus (HIV) (Harmsen et al., 1995; Puddu et al., 1998), and (3-herpes virus (Andersen et al., 2001). Evidence has shown that LF prevents viral infection of the target cell by direct binding to either the virus particles or the viral receptors of the host cell that the virus uses for cell entry. Furthermore, LF could indirectly exert its antiviral mode of action through the upregulation of the antiviral response of the immune system (van der Strate etal . , 2001). 1.1.3.3. Other Biological Functions of Lactoferrin Besides milk, lactoferrin can be found in various body fluids and it is therefore postulated to have significant roles in immunoregulation. The release of LF in the body is activated by the presence of neutrophils. LF in the body is believed to play several significant roles in the host defense system including the regulations of the growth and differentiation of various immune cells, as an anti-inflammatory agent, prevention of oxidative tissue damage, and regulation of iron availability (Brock, 1995). 3 Bovine LF exerted at least a 87% inhibitory effect on the adhesion of enterotoxigenic Escherichia coli to human epithelial cells in vitro (Kawasaki et al., 2000). In an in vivo mouse study, the counts of adherent bacteria in various locations of the intestinal tract were lower in the LFB treated group than in the control group. While LF could be used to lower the presence of pathogenic bacteria in the gut lining, LF could also act as a growth promoter of Bifidobacterium spp. that could be commonly found in the intestinal tracts of breast-fed infants (Miller-Catchpole et al., 1997; Petschow et al., 1999). In vivo studies also showed that when administered orally, LFB helped to balance the gut flora of experimental mice (Hentges et al., 1992) and human full-term infants (Roberts et al., 1992). The levels of Enterobacteriaceae, Streptococcus, Staphylococcus, and Clostridium in the fecal flora of the infants were decreased by almost 20% after feeding LFB-enriched formula for two weeks. Meanwhile, the percentages of Bifidobacterium in the fecal flora increased by almost 50%. Results from these studies indicated that lactoferrin supplemented formula assisted in the establishment of a Bifidobacterium-predom'mani intestinal flora in infants. Other studies showed that LF exhibited chemopreventative effects in the development and progression of bladder carcinogenesis (Masuda et al., 2000) and esophagus and lung carcinogenesis (Ushida et al., 1999; ligo et al., 1999). 1.1.3.4. Other Functions of Lactoferrin Due to its iron-binding capacity, lactoferrin also exhibits strong antioxidative ability by chelating excess metal ions in food systems. Huang et al. (1999) examined the effects of lactoferrin on the oxidation of a corn oil emulsion and a liposome model system. Results showed that the antioxidant activity against iron-catalyzed autoxidation increased with LF concentration. A mixture of 1 (aM LF and 0.5 (iM ferrous ions exhibited better antioxidative ability than 1 (iM LF alone. However, LF became a prooxidant at concentrations higher than 15 nM and was unable to inhibit copper-catalyzed autoxidation. Results showed that the antioxidant or prooxidant activity of LF depended on the lipid system, buffer, LF concentration, the concentration of metal ions, and oxidation time. The antioxidant activity of LF was studied in a commercially modified infant formula supplemented with iron at a level of 12 mg per liter (Satue-Gracia et al., 2000). Results demonstrated that in the presence of increasing concentrations of lactoferrin, the induction periods for oxidation of the infant formulas increased and the rates of oxidation decreased. Lactoferrin inhibited oxidation in a concentration-dependent manner even at concentrations beyond its capacity to bind iron at its two high affinity binding sites. 4 It is postulated that lactoferrin can be used as a multi-purpose additive in infant formulae for its ability to improve intestinal flora of infants, antioxidant property, and antimicrobial abilities (Farnaud and Evans, 2003). 1.2. Lactoferricin & Its Derivatives Lactoferricin (Lfcin) was first described as a "potent antimicrobial peptide" that could be generated from bovine lactoferrin by heat treatment at acidic pH (Saito et al., 1991) and by pepsin digestion (Tomita et al., 1991). LF hydrolysates produced using these two methods exhibited higher antimicrobial ability than untreated lactoferrin. Since then, extensive studies have been conducted for the production and purification of Lfcin as well as the understanding of its structure-function relationships. 1.2.1. Production and Purification of Lactoferricin and Related Peptides 1.2.1.1 Discovering Lactoferricin The presence of the antimicrobial peptide sequence was first noticed when heat-treated lactoferrin displayed increased antibacterial activity (Saito et al., 1991). A peptide fraction with strong activity was identified after fractionation of heat-treated LF by reverse-phase H P L C . It was suggested that at least one bactericidal domain would be released by heating LF at pH 2.0 at 120 °C for 15 minutes. At almost the same period of time, hydrolysates of LF were generated by digestion with porcine pepsin, cod pepsin, or aspartic protease at a concentration of 3% (w/w of substrate) at 37 °C and pH 2.5 for 30 minutes (Tomita et al., 1991). The antimicrobial activity of the resulting LF hydrolysates against Escherichia coli 0111 were at least eightfold stronger than undigested LF. The minimal inhibitory concentration (MIC) of LF hydrolysates remained unaffected even when the peptic hydrolysis continued for more than 4 hours, indicating that the active peptides were resistant to further cleavage by pepsin. The antibacterial property of pepsin digested LF hydrolysates was not affected in the presence of 0.1 mM F e S 0 4 ; however, the antimicrobial property of untreated LF was completely abolished, suggesting that the antibacterial domain functions by a mechanism distinct from chelation of iron or other metal ions. LF hydrolysates prepared by cleavage with porcine trypsin, papain, or other neutral proteases were much less active than those produced by peptic digestion. 1.2.1.2. Identifying Lactoferricin Pepsin hydrolyzed human lactoferrin and bovine lactoferrin were fractionated by reverse-phase H P L C and one antimicrobial peptide (named human lactoferricin, LfcinH; and bovine lactoferricin, LfcinB) was isolated in each hydrolysate (Bellamy et al., 1992b). Results 5 from amino acid sequence analysis indicated that LfcinH has a molecular weight of 5,558 Da and consists of two sub-fragments, residues 1 to 11 and residues 12 to 47, linked by a disulfide linkage between Cys10 on sub-fragment 1 and Cys46 on sub-fragment 2. Another disulfide linkage was also observed between Cys20 and Cys37 on sub-fragment 2 (Bellamy et al., 1992b). Bovine lactoferricin is a relatively shorter peptide corresponding to residues 17 to 41 near the N-terminus of LFB. A disulfide linkage between Cys19 and Cys 36 was reported and the peptide has a molecular weight of 3,124 Da as calculated from the amino acid sequence (Bellamy etal . , 1992b). 1.2.1.3. Production and Purification Strategies Pepsin digestion of LF has been the most commonly utilized method to produce bovine lactoferricin in most of the studies reported so far. However, the composition and antimicrobial potency of the resulting peptide mixtures vary depending upon the source and composition of the starting material, as well as the digestion conditions (e.g. digestion pH, temperature, and duration). Furthermore, purity of the resulting Lfcin would rely on the chromatographic techniques used. A number of production and purification strategies for lactoferricin have been conducted in attempt to isolate Lfcin in larger quantities. Some of these strategies will be discussed below. Dionysius and Milne (1997) reported that three LfcinB related peptides could be isolated, from pepsin digests of LFB, by cation exchange chromatography. The LF hydrolysate was prepared as previously described (Bellamy et al., 1992a and b) and fractionated through a cation exchange column equilibrated with 0.05 M phosphate buffer at pH 7.5. Peptides of interest were eluted with the equilibrating buffer containing 1 M NaCI. Using reverse-phase H P L C , mass spectrometry, N-terminal amino acid sequencing, and amino acid analysis, three peptides originating from the N-terminal of the LF were identified; the amino acid sequences and molecular weights of these peptides are shown in Figure 1.3. Peptide I had a mass of 3195 and corresponded to residues 17 to 42 of LFB with a disulfide linkage between Cys19 and Cys36. Peptide II consisted of two peptides, residues 1 to 16 and residues 43 to 48, with a disulfide bond at Cys9 and Cys45. Peptide III had a mass of 5851 Da and corresponded to residues 1 to 48 of LFB with a cleavage between residues 42 and 43; disulfide bonds could be found between Cys9 and 45 and between Cys19 and 36 (Dionysius and Milne; 1997). All three peptides exhibited improved antimicrobial activity over undigested lactoferrin. However, only the activity of peptide I was comparable to the Lfcin previously reported (Bellamy et al., 1992a, b). The activity of peptides II and III were much weaker. In fact, peptide I was almost identical to LfcinB that was previously described (Bellamy et al., 1992b) except for the presence of an additional alanine at the C-terminus. 6 The isolation of Lfcin and other related peptides has also been reported using a bead-based cation-exchange chromatography on SP-Sepharose Fast Flow resin (Recio and Visser, 1999). Two major peptides of mass 3123 and 3194 were eluted with 7 M ammonia and 2 M NaCI. These peptides corresponded to residues 17-41 and 17-42 of LF. However, other components were also found in the fraction, including the 27-amino acid peptides and the oxidized form of the 26-amino acid peptide. Shimazaki et al. (1998) also reported that a peptide with molecular mass of 3195.5 Da was isolated using affinity chromatography with an immobilized heparin column. The peptide of interest was eluted from the column with 1 M NaCI in various buffer systems. Hoek et al. (1997) reported the production of LfcinB and other peptides sharing high sequence homology with LfcinB by recombinant chymosin digestion of bovine LF. Specifically, LFB (5 g) was dissolved in 200 mL of water, and the solution was adjusted to pH 3.0 by HCI. To the solution, 200 mg of recombinant chymosin was added. It was then incubated at 60 °C for 1 hour. The resulting solution was pH adjusted, centrifuged and passed through a cation-exchange membrane filtration system. Peptides eluted with 2M NaCI were further fractionated by R P - H P L C and analyzed by mass spectrometry. The molecular weights and amino acid sequences of four highly homologous peptides were determined as shown in Figure 1.3. Peptides 1a, 1b, and 2 were single peptides with amino acid sequences starting at residue 17 of the LF. All three peptides formed a cyclic structure with a disulfide linkage between Cys 19 and Cys 36. Peptide 3 consists of two peptides joined by a disulfide bond; its main peptide shares sequence homology with the other three and is also disulfide linked. All four peptides exhibited antimicrobial activity against £ coli L361. Minimum inhibitory concentrations (MIC) reported for peptides 1a, 1b, 2, and 3 were 12.5, 12.5, 13.2, and 23.4 ng • mL" 1, respectively. However, when expressed in molar concentration, MIC for all four peptides were 4 nM (Hoek et al., 1997). A patent for large-scale production of Lfcin from bovine LF claims that a yield of 10.5 g of Lfcin with 99% purity could be obtained from 600 g of LFB (Tomita et al., 1999). LF was first digested with an enzyme and resulting hydrolysate was passed through a hydrophobic or cation exchange chromatography medium. After the column was washed with water and a citric acid-sodium phosphate buffer at pH 7.0, the adsorbed Lfcin was desorbed with the same buffer at pH 5.0. To date, research on Lfcin production has been limited to the uses of laboratory-grade LFB and chromatographic media. The development of a fractionation strategy for the large 7 scale production of Lfcin using food-grade materials is much needed for the incorporation of Lfcin in human food system. 1.2.2. Antimicrobial Spectrum of Lactoferricin and Related Peptides As mentioned before, both purified LfcinH and LfcinB exhibited stronger antimicrobial activity against Escherichia coli 0111 than their undigested counterparts (Bellamy et al., 1992b). The MIC against E. coli 0111 was 3000, 500, and 100 ug • mL"1 for LFH, LF hydrolysate, and LfcinH, respectively, and 2000, 100, and 6 ug • mL"1 for LFB, LFB-hydrolysate, and LfcinB, respectively. Furthermore, a diverse range of Gram-positive and Gram-negative bacteria was found to be susceptible to LfcinB at MIC ranging from 6 to 12 ug • mL"1 (in 1% peptone) for most of the Gram-negative bacteria tested and from 0.3 to 12 ug • mL' 1 for most of the Gram-positive bacteria tested. Bifidobacterium bifidum strains were resistant to LfcinB (Bellamy et al., 1992a). With MIC ranging from 3 to 18 ug • mL"1, various types of yeasts were inactivated by LfcinB, while 3 to 45 ug • mL"1 of LfcinB were needed to inhibit the growth of filamentous fungi. Lfcin could enhance the effectiveness of antifungal agents against Candida albicans (Wakabayashi et al., 1998). Skin disease-causing fungi could also be inactivated by the addition of 6 to 30 ug • mL"1 of Lfcin (Bellamy et al., 1993; Bellamy et al., 1994; Tomita eta l . , 1994). However, the antimicrobial effectiveness of LfcinB was decreased in the presence of N a + , K + , C a 2 + , M g 2 + , F e 3 + , and various buffer salts including K 2 H P 0 4 " , K H 2 P 0 4 " , Tris-HCI, H E P E S - N a O H , and P IPES-NaOH (Bellamy et al., 1992a; Yamauchi et al., 1993). The antimicrobial activity of LfcinB against Salmonella enteritidis was inhibited in nutrient rich media such as tryptic soy broth (Facon and Skura, 1996). Studies have shown that the antimicrobial activity of LfcinB in ground beef (Venkitanarayanan et al., 1999) and carrot juice (Chantaysakorn and Richter, 2000) diminished due to the presence of high levels of cations in the systems. 1.2.3. Structure - Function Relationship Analysis of Lactoferricin and Its Derivatives 1.2.3.1. Basic Structural Properties of Bovine Lactoferricin As described in previous sections, bovine lactoferricin (LfcinB) is a 25-amino acid peptide corresponding to residues 17-41 from the N-terminal loop region of LF with the sequence 1 7 F K C R R W Q W R M K K L G A PSITC V R R A F 4 1 ; a disulfide linkage between Cys19 and Cys36 was also reported (Bellamy et al., 1992b). This 25-amino acid peptide includes 8 basic amino acids that contribute to a net charge of +6.85 at pH 7 (Rekdal et al., 1999). Nuclear magnetic resonance (NMR) analysis of Lfcin revealed that Lfcin formed a slight distorted anti-parallel 6-sheet structure with a loop formed between residues 27 to 30 (Hwang et al., 1998). 8 Due to the presence of the disulfide linkage between Cys19 and Cys36 and the orientation of Trp22, two noticeable kinks were observed between Arg20 and Arg21 and between Cys19 and Trp22 (Hwang et al., 1998). Results from circular dichroism analysis (CD) also demonstrated that LfcinB displayed no a-helical conformation, but consisted of 41% (3-sheet in an aqueous environment. The presence of trifluoroethanol (TFE, 50% v/v) increased the a-helical content of LfcinB to 6.5%. When the disulfide bond was reduced, less p-sheet was detected in both the aqueous environment and in the presence of T F E (Shimazaki et al., 1998). Once the LfcinB was folded into a p-sheet structure, side chains of Phe17, Cys19, Trp22, Trp24, Leu29, Ala31, Pro32, Ile24, and Cys36 comprised a prominent hydrophobic surface on one side of the (3-sheet structure (Figure 1.4a). Outside of this hydrophobic strip, most side chains are hydrophilic, including Ser17, Gln23, Thr35, and Val37, or positively charged, including Lys18, Arg20, Arg21, Arg25, Lys27, Lys28, Arg38, and Arg39 (Figure 1.4b and c). The high degree of amphipathic structure was thought to be responsible for the antimicrobial activity of the peptide by interacting with and thus disrupting the cellular membrane. Recent studies using chemically synthesized lactoferricin and analogs with different structural variations have provided new information on the structure - function relationship of this cationic antimicrobial peptide. 1.2.3.2. Roles of Secondary Structure In an aqueous environment LfcinB forms a p-sheet structure stabilized by a disulfide bond. When the cyclic structure was cleaved with CNBr and when the disulfide bond was reduced in the presence of p-mercaptoethanol, little change in the antimicrobial activity of Lfcin against a number of bacteria was observed (Hoek et al., 1997; Strom et al., 2000). This suggested that neither a cyclic nor a linear structure of the peptide is particularly critical for its effectiveness as an antimicrobial agent. Results from several studies showed that shorter peptide fragments composed of LfcinB residues 17-28 (Tomita et al., 1994), residues 20-30 (Kang et al., 1996) and residues 17-31 (Rekdal et al., 1999), or LfcinH residues 20-35 (Odell et al., 1996) and residues 20-30 (Odell et al., 1996; Chappie et al., 1998) demonstrated similar or greater antimicrobial activity against Escherichia coli and other microbes than their native counterparts. C D analysis of LfcinB 20-30 (Kang et al., 1996) showed that this short LfcinB derivative adopted an a-he|ical structure in the presence of T F E and sodium dodecyl sulfate (SDS). When the helical region of LfcinH 20-30 was disrupted by replacing the methionine at position 26 with a proline, the antibacterial activity of the peptide against E. coli, Staphylococcus aureus, and an Acinetobacter sp. was diminished (Chappie et al., 1998). 9 Furthermore, using NMR spectroscopy, Schibli et al. (1999) observed that the antimicrobial core of Lfcin (RRWQWR, Tomita et al., 1994) adopted an orderly conformation with three Arg residues on one side of the structure and two Trp residues on the other side of the structure. They further illustrated that the aromatic Trp residues were more deeply buried in the S D S micelles with the Arg and Gin exposed to the solvent. Although Lfcin and its derivatives may form either a-helix or (3-sheet in different environments, evidence suggests that the orientation of both basic and hydrophobic amino acids in the peptide chain is critical to the antimicrobial properties of shorter LfcinB derivatives. 1.2.3.3. Roles of Basic Amino Acids In a study investigating the heparin-binding ability of LfcinB analogs, it was reported that the absence of basic amino acids such as arginine and lysine would decrease the heparin binding ability of LfcinB 17-31 (Shimazaki et al., 1998). Since this heparin binding ability was thought to be responsible for the inhibitory effects of lactoferrin on inflammation, the presence of basic amino acids in LfcinB might also play a critical role in its antimicrobial activity. In order to examine the importance of basic amino acids in the antimicrobial activity of LfcinB, all the basic amino acids in LfcinB 20-30 were replaced with glutamic acids. The modified peptides showed lower antimicrobial activity against Escherichia coli 0111 and Bacillus subtilis than the unmodified peptide (Kang et al., 1996). Furthermore, CD spectra of the modified LfcinB in 30 mM S D S were similar to the ones in aqueous environment, suggesting that the modified LfcinB did not adopt an a-helical conformation. Similar results were reported when alanine replaced arginine or lysine residues; higher MIC against E. coli and S. aureus were observed (Strom et al., 2000). In contrast, the antimicrobial activity of the peptide increased when the glutamic acid in position 30 was replaced with an alanine. Although the net charge seems to be important in the antimicrobial activity of LfcinB, effects of the size and the position of the basic amino acids should also be considered. When Arg were replaced with Lys, it was found that the Lys substituted LfcinB 20-30 exhibited less antimicrobial activity (Kang et al., 1996). The replacement of Lys with Ala had less effect on MIC than the replacement of Arg (Strom et al., 2000). These results suggested that the larger size and the positive charge of the guanidinium group of the arginine residues were important in the antimicrobial property, with a more severe disruption of the bacterial cell membrane. The position of the basic amino acids along the p-sheet region also affects the biological activity of LfcinB. For instance, when the Arg in positions 20, 21, and 25 of LfcinB was replaced with an Ala, the MIC against E. coli were 35, 61, and 28 uM, respectively (Strom et al., 2000). Each Arg may have a specific role in the interaction with the cellular membrane. 10 As demonstrated by Schiffer-Edmundson helical wheel models, the asymmetry in charge distribution along the p-sheet is important in improving the antimicrobial activity (Rekdal et al., 1999); LfcinB analogs having all of the basic amino acids located along one face of the sheet demonstrated stronger antimicrobial activities than analogs with more basic amino acids but distributed at different locations. 1.2.3.4. Roles of Hydrophobic Amino Acids Hydrophobic residues on the cationic antimicrobial peptides are also essential in the formation of an amphipathic structure. Alanine substitution of all the hydrophobic residues in LfcinB 20-30 resulted in no antimicrobial activity in the concentration range tested (Kang et al., 1996). Similar results were reported by Strom et al. (2000). When one of the two tryptophan residues in position 22 and 24 was replaced by alanine, no inhibitory effect against E. coli or S. aureus was observed at concentration less than 100 uM, while the MIC of the unmodified peptides for these two microorganisms were 24 and 48 uM, respectively. Replacing Met at position 26 with alanine also increased the MIC to 70 uM. The importance of tryptophan in the antibacterial activity was investigated by substituting Trp in pentadecapeptides derived from human (LfcinH), caprine (LfcinC), murine (LfcinM), and porcine (LfcinP) origin (Strom et al., 2000). These Lfcins each have only one Trp residue in the Lfcin segments tested. Upon adding a second Trp residue to LfcinH, LfcinC, or LfcinP, but not to LfcinM, the antimicrobial activity was improved by 1.5 to 6-fold compared to the unmodified peptides. These results indicated that the hydrophobic tryptophan plays a key role in the antimicrobial activity of Lfcin. The specific role of tryptophan was examined by replacing the Trp residues in LfcinB 17-31 with natural and unnatural aromatic amino acids (Haug and Svendsen, 2001). The replacement of nitrogen in Trp with a sulfur atom in the side chain of (B-(benzothien-3-yl)alanine (Bal) enhanced the antibacterial activity of LfcinB 17-31. These results suggested that hydrogen bonding between the peptide and the cellular membrane was not essential in the antimicrobial properties of the peptide. Similarly, the antibacterial activity of the resulting peptide was also increased when the indole side chain of Trp was replaced with naphthalene. Again, these results suggested that hydrophobic aromatic side chains were more favorable for the interaction between the peptides and the cellular membrane. Antimicrobial activity was also improved when the Trp was replaced with unnatural hydrophobic aromatic amino acids that were larger in size (Haug et al., 2001). Conversely, the antimicrobial effect diminished when the smaller phenylalanine was substituted for Trp. Also, substituting hydrophobic aromatic side chains with a more elongated shape led to increased antimicrobial activities. These improvements were further enhanced when both Trp residues in positions 22 and 24 were 11 replaced (Haug et al., 2001). These size and shape effects suggested that the hydrophobic side chains of Lfcin were essential for the antimicrobial activity. Larger and elongated side chains may be able to penetrate deeper into the cellular membrane and serve as a stronger anchor, making the overall peptide a more efficient antimicrobial agent. The combined effect of hydrophobic and basic amino acid side chains was examined using LfcinM (Strom et al., 2001). As was mentioned above, the addition of one extra Trp in LfcinM was inefficient in providing an effective antimicrobial peptide. However, MIC was considerably improved when the Glu residues were replaced with Arg or Ala, and the Val residue with Tyr. These results showed that the presence of both aromatic and basic amino acids were essential in establishing an amphipathic structure with antimicrobial property. 1.3. Cationic Antimicrobial Peptides In order to gain a better understanding of the possible antimicrobial mechanism of lactoferricin, it is useful to examine modes of action of other similar peptides. Some general information of other cationic antimicrobial peptides (CAPs) will be briefly discussed. More extensive reviews may be found in Bechinger (1997), Hwang and Vogel (1998), Epand and Vogel (1999), Lohner and Prenner (1999), Shai (1999), and Chan and Li-Chan (2006). 1.3.1. Structures and Modes of Action of Other Cationic Antimicrobial Peptides More than 140 naturally occurring C A P s have been identified. They are between 11 and 50 amino acids long and have a net positive charge of +2 or more. Most of these C A P s have certain properties in common. They all have affinity for membrane lipids and their specificity for microbial membranes in many cases has been shown to be related to the positive charge on the peptide favoring interaction with the exposed anionic lipids of microorganisms (Epand and Vogel, 1999). In general, these C A P s can fit into four major classes, namely, B-sheet structure stabilized by one to three disulfide bridges, a-helices, extended helices with a predominance of one or more amino acids (including unusual amino acids such as lanthionine derivative, 3-methyllanthionine derivative, and dehydroalanine in nisin), and loop structures (Hancock, 1997; Hwang and Vogel, 1998). The a-helical antimicrobial peptides were the first to be identified and characterized. The antimicrobial mechanisms of these helical peptides are therefore the most well understood. Most C A P s in this class exert their activity by causing membrane permeation and cell lysis via one of two general mechanisms: the 'barrel-stave' mechanism and the 'carpet' mechanism (Shai, 1999). The 'barrel-stave' model describes the formation of transmembrane channels and pores. Bundles of a-helices interact with the outer-membrane followed by the insertion of the hydrophobic face of the helices into the hydrophobic core of the membrane. In the 'carpet' 12 model, helices are not inserted into the hydrophobic core. Instead, they remain in contact with the phospholipid head group on the membrane and disintegrate or solubilize the membrane by disrupting the bilayer curvature. As has been discussed in previous sections, lactoferricin forms a hairpin (3-sheet structure, and thus belongs to the (3-sheet class of C A P s . However, little is known about the mechanism by which this class of antimicrobial peptides produces membrane damage. It appears that different (3-sheet C A P s have different antimicrobial mechanisms. No unique mode of action is identified for this group of C A P s (Wu et al., 1999). In addition to the peptide structure, the chemical composition of the phospholipids of the microbial membrane is also a major determinant of the effectiveness of antimicrobial peptide. Charge-charge interactions, membrane curvature strain, and hydrophobic mismatch between peptides and lipids are all important parameters. For instance, a threshold peptide to lipid ratio is needed for the pore formation on the microbial membrane and the threshold level is specific for different microorganisms depending on their phospholipid composition (Lohner and Prenner, 1999; Huang, 2000). 1.3.2. Proposed Antimicrobial Mechanisms of Lactoferricin A mechanism for the antimicrobial activity of lactoferricin derivatives has been postulated (Strom et al., 2001). Like other C A P s , Lfcin must first penetrate through the cell wall of the bacteria. Haukland et al. (2001) reported similar MIC values of Lfcin for E. coli, Gram negative bacteria, and S. aureus, Gram positive bacteria, suggesting that the presence of the cell wall had no effect of the antimicrobial activity of Lfcin. Once Lfcin penetrated through the cell wall, Lfcin was able to establish interaction with its target microorganism via the positively charged amino acid residues on the peptide and the negatively charged outer membrane of the bacteria. LfcinB would interact with the lipopolysaccharides on the outer membrane of E. coli, and with teichoic acid on S. aureus (Voland et al., 1999). Once the Lfcin was attached to the membrane, electrostatic attraction might bring the hydrophobic residues such as tryptophan and tyrosine close to the lipophilic portions of the bacterial membrane. The hydrophobic amino acids on the Lfcin might migrate into the phospholipid layer, causing a depolarization of the bacterial inner membrane, and therefore interrupting the regular membrane function of the target bacteria (Ulvatne et al., 2001). Electron micrograph images showed that Lfcin induced the formation of blebs on the surface of E. coli (Ulvatne et al., 2001). However, unlike most other C A P s , Lfcin did not cause increases in membrane permeability of intact bacteria. A complete model for the bactericidal mechanisms of LfcinB has not been elucidated yet. 13 It was reported that Lfcin induced fusion of liposomes, and the fusion process may represent a means of transport of the antimicrobial peptide to cross the lipid barrier and eventually interact with cytoplasmic targets (Ulvatne et al., 2001). Tryptophan-containing antibacterial peptides have been shown to have the ability to inhibit DNA synthesis and protein synthesis (Subbalakshmi et al., 1996; Subbalakshmi and Sitaram, 1998). Lactoferrin from human milk has the ability to bind DNA and it was demonstrated that the DNA-binding site is in fact the antimicrobial domain of the protein (Kanyshkova et al., 1999). It was suggested that Lfcin could exert its antimicrobial activity by interacting with nucleic acids and preventing either replication or transcription of protein. 1.4. Phospholipid Bilayer Since it is widely accepted that the antimicrobial mechanism of lactoferricin is dependent on its initial contact with the external side of the bacterial membrane, it is important to gain more insight into the structure of the phospholipid bilayer of the bacterial membrane. The cell membrane is composed of a phospholipid bilayer that is interspersed by other lipid molecules such as cholesterol. This phospholipid bilayer functions to separate the aqueous environment inside and outside of the cell, whereas there is no water in the space between the phospholipid layers. The outside parts of the bilayer are described as being hydrophilic and the inside parts are described as being hydrophobic. Water soluble molecules entering the cell must find ways to pass through this phospholipid hydrophobic barrier (Dowhan, 1997). The head-group composition of the phospholipids of E. coli is relatively invariant under a broad spectrum of growth conditions. On average, £. coli is composed of 70-80% of phosphatidylethanolamine (PE), 20-25% of phosphatidylglycerol (PG), and less than 5% of cardiolipin (CL) (Table 1.1). On contrary, P G is the most abundant phospholipids in Micrococcus sp. (Dowhan, 1997), responsible for 60 to 70% of the membrane. P E , P G , and CL are three major phophoslipids found in the membrane of Bacillus sp. (Dowhand, 1997). Although less than 1 % of the phospholipids in Micrococcus and Bacillus sp. are composed of P C , P C makes up 35% of Bacillus megaterium (Huijbregts et al'., 2000). The lipid composition of membranes is not constant (Opekarova and Tanner, 2003). The composition is tightly regulated to perform its physiological functions. For instance, the regulation of the acyl chain composition in wild-type cells of E. coli is necessary for the organism to be able to grow in a 'window' between a lamellar gel phase and reversed non-lamellar phase (Morein et al., 1996). The cells respond to a higher growth temperature by synthesizing lipids with shorter and more saturated acyl chains to maintain a semi-lamellar 14 phase. Under normal physiological conditions, position 1 of the phospholipid backbone is occupied predominantly by palmitic acid, and position 2 by the dominant monounsaturated fatty acids palmitoleic and c/s-vaccenic acids. In cells grown at 37°C, the ratio of saturated to unsaturated fatty acids is about 1:1 (Dowhan, 1997). 1.5. Raman Spectroscopy The phenomenon of Raman scattering was first described by Sir C. V. Raman in 1928. He observed that the wavelength of a small fraction of the visible radiation scattered by certain molecules differed from that of the incident beam and the shifts in wavelength depended upon the chemical structure of the molecules responsible for the scattering. In Raman spectroscopy, the sample is first excited by a radiation source having a wavelength that is away from any absorption peaks of the analyte. The molecules will then interact with a photon from the energy source and be excited to a "virtual state" between the ground state and the first electronic excited state. Most of the excited molecules will return to their original energy level before excitation, producing a phenomenon known as Rayleigh scattering. However, approximately one in each million of photons will return to the first excited vibrational state and emit an energy lower than that of the original excitation laser. The lower emitted energy known as a Stokes-Raman shift would provide information on molecular vibrations that, in turn, yield data on structural conformation and the surrounding environment (Carey, 1999). The vibrational states probed by Raman spectroscopy and infrared spectroscopy are very similar. In fact, these two vibrational spectroscopy techniques are complementary. Vibrations that involve strong dipole moments such as hydroxyl- or amine-stretching, and carbonyl groups are strong in an infrared spectrum and weak in a Raman spectrum. On the contrary, stretching vibrations of carbon double or triple bonds and symmetric vibrations of aromatic groups (non-polar) give very strong Raman bands but usually result in weak infrared signals. 1.5.1. Raman Spectroscopy: Structure - Function Relationship of Proteins One of the major advantages of Raman spectroscopy is that both normal water (H 2 0) and heavy water (D 2 0) generate very weak Raman spectra. Thus, this analytical technique could be used to examine virtually any sample morphology. For proteins, this includes solutions, suspensions, precipitates, gels, films, fibers, single crystals, and polycrystalline and amorphous solids. The presence of water, which is almost unavoidable in food protein samples, would therefore create little interference in the Raman spectrum. Consequently, Raman spectroscopy is of particular interest as a tool to monitor in situ changes of proteins in 15 food systems during processing or storage, or accompanying phase transitions from solutions to emulsions, gels or precipitates (Pelton and McLean, 2000). Raman spectroscopy is a valuable technique for the understanding of protein structures and conformations. The vibrational spectra of various amino acid side chains and the - C O - N H -peptide bond are sensitive to chemical changes and the surrounding microenvironments. For instance, the amide I and III bands (near 1660 cm" 1 and 1240 cm" 1 , respectively) are most commonly used for secondary structure characterization of the protein molecule including the levels of a-helix, antiparallel 6-sheet, and disordered structure (Li-Chan, 1996). The intensity and location of Raman bands assigned to various stretching or bending vibrational modes of functional groups in the side chains of amino acid residues can be used to provide information on the chemistry and microenvironment of those residues. In particular, valuable information obtained for Raman bands near 530, 650, 725, & 2550 cm" 1 provide conformational information about the S S , S H , and S C groups of cystine, cysteine, and methionine residues, respectively. Microenvironment around C H structures such as those on aliphatic residues could be determined by Raman bands near 1450 and 2800-3000 cm" 1 . Side chain structure of tyrosine, tryptophan, phenylalanine, and histidine could also be analyzed (Li-Chan et al., 1994). 1.5.2. Raman Spectroscopy: Structure - Function Relationship of Peptides and Phospholipid Bilayers Since the presence of water creates little interference in the Raman spectrum, Raman spectroscopy is being recognized as a valuable analytical tool for the understanding of protein and peptide structures in the biological system. For instance, model systems of poly(L-lysine) on phospholipid membranes were examined by Laroche et al. (1988 and 1990). Several studies have also been conducted to examine the conformational changes of tryptophan residues in gramicidin, a cationic antimicrobial pentadecapeptide, in phospholipid micelles and other membrane mimicking environments (Bouchard and Auger, 1993; Maruyama and Takeuchi, 1997 & 1998). The mechanisms of membrane pore formation induced by a 23-residue peptide, magainin, was also examined by Jackson et al. (1992) and Ludtke et al. (1996). In addition to antimicrobial peptides, Raman spectroscopy has been used to examine conformational behaviours of other bioactive peptides, such as the secondary structure of a neuropeptide bound to liposomes (Williams and Weaver, 1990), the effects of synthetic human pulmonary surfactants on phospholipid bilayer (Vincent et al., 1993), and interactions between DNA and the telomere binding protein subunit (Laporte et al., 1999). Raman spectroscopy can also be used to study the structural arrangement of phospholipids. Phospholipids commonly exist in three states (Figure 1.5). Classical 16 representations of phospholipid organization in membranes focus on lamellar or bilayer structures such as the ordered gel state (Lp) or fluid liquid crystalline state (L a). In vitro, some natural phospholipids can.assume nonbilayer structures such as the reversed hexagonal (HM) phase. The temperature at the mid-point of transition (transition temperature, 7~m) for the L B - to-L a transition, is a measure of a distinct collective property of the phospholipids making up a particular bilayer (Dowhan, 1997). Two distinct Raman features observed in the C-H stretching mode region (2750 to 3050 cm"1) of the hydrocarbon chains are used to elucidate the structural behaviour of the phospholipid bilayer. The 2850 and 2880 cm" 1 Raman bands are assigned, respectively, to the symmetric and the asymmetric C-H stretching modes for the coupled methylene moieties of the hydrocarbon chains in the bilayer interior (Gaber and Peticolas, 1977). As temperature increases and upon melting, the intensity of the 2850 cm" 1 band (bsso) increases in intensity relative to the 2880 cm" 1 band (l288o) (Figure 1.6a). A greater relative peak-height intensity ratio bsso/bsso reflects weaker lateral interchain interactions in the bilayer (Figure 1.6bA). This intensity ratio is thus a sensitive probe for the intermolecular coupling and for the lateral packing of the acyl chains. As temperature increases and as the bilayer structure is transitioning from the ordered gel state (L p) to fluid liquid crystalline state (L a), the bsso/bsso ratio will experience a sudden change. Changes in the I2845/I2880 ratio indicate the transition temperature, T m (Snyder et al., 1978, 1980, and 1982). The band around 2935 cm" 1 mainly arises from the symmetric C-H stretching mode of terminal methyl groups in the hydrophobic center of the bilayer. As the temperature increases, the intensity of the 2935 cm" 1 (I2935) Raman band increases (Figure 1.6bB). The relative peak-height intensity ratio is therefore indicative of both the lattice packing order and the intrachain conformational order. This is an index of the overall disorder of the lipid acyl chain matrix (Gaber and Peticolas, 1977; Snyder et al., 1980; Vincent et al., 1993). 1.6. Multivariate Data Analysis Techniques Multivariate data analysis techniques are needed for simultaneous consideration of several variables in complex systems and identification of patterns of interest. Multivariate data analysis techniques have gained much importance in medicinal chemistry and have been used in the understanding of protein signaling in apoptosis (Janes et al., 2003), tumor classification (Nguyen and Rocke, 2002), and the early detection of cancer (Chen et al., 2004; Skates and lliopoulos, 2004) and heart diseases (Mateos-Caceres et al., 2004). Among the various statistical methods, principal component analysis (PCA) and partial least squares (PLS) regression are the most commonly used techniques for the analysis of 17 structure-function relationships of peptides. Edman et al. (1999) used P C A and P L S to characterize sequence patterns and properties in signal peptides from mycoplasmas, Gram-positive bacteria, and E. coli. Structure-function relationships of anti-hypertensive peptides isolated from milk proteins were also studied using P C A (Pripp et al., 2004). 1.6.1. Principal Component Analysis Principal component analysis (PCA) is one of the most popular unsupervised multivariate methods that decrease the complexity of a data set with no prior knowledge of the systems. It uses all the measured variables to examine the inter-relationship of the data by combining a large number of original variables to produce uncorrelated indices known as "principal components" (PC); minimal loss of information could be achieved. In addition, since the calculated P C s are uncorrelated, the overall data set could be adequately represented by a small number of P C s (Aishima and Nakai, 1991). It is a common practice to visualize and interpret P C A results by plotting the first few major P C s . Sample classification could then be achieved by observing the grouping arrangement of the data points on the P C plots. Although it is possible to include up to three P C s in a 3-dimensional plot, visualization of 3-D plots could be difficult and might lead to poor or even incorrect interpretation of the P C A data. Hence, P C plots usually only depict two P C s at a time (e.g. PC1 versus P C 2 , PC1 versus P C 3 , P C 2 versus P C 3 , etc.) Since only two P C s could be interpreted at a time, only limited total numbers of P C s would be analyzed. Furthermore, depending on which two P C s were plotted, different observations and interpretations could be made on the same set of data. 1.6.2. Partial Least Square Regression Partial least squares (PLS) regression analysis is a type of supervised multivariate technique that can be used for pattern recognition and prediction (Aishima and Nakai, 1991). The emphasis of P L S is on predicting the responses based on many dependent variables and not on trying to understand the underlying relationship between the variables (Geladi and Kowalski, 1986). P L S is based on double principal component analysis. The analysis of one data set is guided by the structure of a second data set (the dependent variables), which is in turn used for modeling both the first and second sets of data. In a comparative study reported by Garthwaite (1994), when there were a large number of variables and random error variance, P L S was shown to be more powerful in predicting equations than principal component analysis. 1.6.3. Principal Component Similarity Analysis A relatively new unsupervised multivariate analysis technique, principal component similarity (PCS) analysis, combines principal component analysis (PCA) and pattern similarity 18 computation (Aishima et al., 1987). The procedure for P C S analysis described by Horimoto et al. (1997) is adapted below: Step 1. Apply P C A to the original set of samples for k parameters and n samples to derive k P C s and k eigenvalues E and then steps 2 to 5 are followed for each sample. Step 2. Compute independent variable variance V as follows: P, = Ej / I E k Equation 1 V = 100 (1 - 1 Pi) Equation 2 where i is the principal component number. In general, only P C s with E| higher than 1.0 would be included in the computation. Step 3. Compute dependent variables Y as follows: Y, = Vi + M (PCi - P C q ) Equation 3 where q is the reference and M is percentage of variance of each principal component factor. Step 4. Carry out linear regression analysis for Y versus V for each sample at each time to compute correlation coefficient (r2) and slope (S). Step 5. Plot S (slope) versus r2. Step 6. The resultant plot could be used for classification of the original set of samples. P C S can effectively reduce the number of original variables needed for analysis to just a few selected principal components as P C A is applied (Step 1). The most important advantage of P C S over the traditional P C A technique is that results of several P C s could be considered and interpreted at the same time (Steps 2 & 3). Thus, P C S could be used to explain maximum variability observed in the available data. Since P C S is a relatively new multivariate statistical method, some of its applications in food samples will be described. P C S was reported for the first time to analyze gas chromatographic data to distinguish mangoes of different cultivars (Vodovotz et al., 1993). It was later applied to analyze H P L C data of Cheddar cheese to identify quality defects (Furtula et al., 1994a and 1994b). P C S analysis was also capable of distinguishing milk samples inoculated with Pseudomonas fragi, P. fluorescens, Lactococcus lactis, Enterobacter aerogenes, and their mixed culture from uninoculated samples (Horimoto et al., 1997) and identifying an unexpected new type of beany flavor in soymilk that was not detected by a sensory panel (Wang et al., 1997). Detection of Escherichia coli and Staphylococcus aureus contaminated salmon patties and hamburger patties was also possible using P C S to analyze gas chromatography data of volatile compounds in the samples (Nakai et al., 1999). 19 1.6.4. Multivariate Data Analysis: Structure - Function Relationship of Lactoferricin and Its Derivatives Principal component analysis (PCA) was used to analyze the structure - function relationship of eight Lfcin analogs using 10 structural property descriptors, namely, number of amino acids, mean hydrophobicity, Kyte-Doolittle hydrophobicity, mean hydrophobic moment, hydrophobic moment, mean charge moment, charge moment, Gamier a, Chou-Fasman a, and Emini surface probability (Rekdal et al., 1999). Results showed that helicity, charge moments, and peptide length were important factors in the antibacterial activity of Lfcin derivatives. Nevertheless, P C A alone was unable to link the structural properties with the biological activities of the peptides. Hence, partial least squares regression (PLS) was applied in addition to P C A to correlate peptide structure to antimicrobial ability of 19 synthetic 15-residue derivatives of murine lactoferricin (Strom et al., 2001). A total of 12 structural property descriptors including 6 helical properties, 4 hydrophobicity data, and two charge parameters were utilized to define the entire peptide sequence. The MIC against E. coli and S. aureus were used as the dependent variables. P C A results showed that the descriptors used explained 82% of the variation in antimicrobial activity; net charge and micelle affinity were the most important structural parameters. When the structural parameters were used to predict the antimicrobial activity of these derivatives using P L S , a good correlation was obtained between the observed and predicted activity against E. coli. In order to determine if theoretical descriptors of each amino acid in a peptide could be used to represent the entire peptide sequence, a matrix of 29 physicochemical variables for 20 coded amino acids were analyzed using P C A (Hellberg et al., 1987). Three principal components, z-,, z 2 , and z 3 , were obtained and used in P L S analysis to relate chemical properties to the biological activities. This approach successfully predicted the biological activity of oxytocin, pepstatin, and bradykinin (Hellberg et al., 1987). The z-i, z 2 , and z 3 descriptors and the P L S approach were adapted to predict the antimicrobial activity of Lfcin analogs (Lejon et al., 2001a). Results showed that in addition to charge and micelle affinity as reported in Strom et al. (2000), lipophilicity of the peptide was equally important. It was concluded that the activity of peptides predicted by using P C A derived descriptors, z-i, z 2 , and z 3 , of each amino acid in.a peptide provided better antimicrobial activity predictions than using the structural properties of the entire peptides (Lejon et al., 2001a). However, the above approach would only consider the amino acid composition of the entire peptide; sequential locations and arrangements were not considered. Based on the results of previous studies, it was noted that the effects of substituting an amino acid with another amino acid on the antimicrobial properties could be significantly affected by the specific 20 location of substitution along the peptides. Hence, it is crucial to consider the sequence of the amino acids while assessing the structure -function relationship of peptides. It was suggested that the position of a specific amino acid in a peptide is of importance to the physico-chemical behaviour of the peptide and information that distinguishes between positions within the peptide is needed to improve results from P C A and P L S analysis (Lejon et al., 2001b). To consider the position and sequential information of specific amino aicds in a peptide, a new statistical program, Homology Similarity Analysis (HSA) was developed. Using HSA, specific structural properties (hydrophobicity, helicity, strand propensity, turn propensity, bulkiness, charge, and hydrogen bonding propensity) of each amino acid in a designated sequence of a sample peptide are compared the corresponding amino acid in the corresponding location in the sequence of a reference peptide (Nakai et al., 2003). Property indices of the amino acids in the sample peptide are plotted against the amino acids in the reference peptide. Linear regression is carried out and the resulting regression coefficient (r2) is computed as the Homology Similarity Coefficient. H S A is therefore focused on the similarity within segments around conserved sites in the sequences, which could be playing important roles in functions, such as active site and binding site. Hence, information that distinguishes between positions of a specific amino acid with a peptide could be more accurately evaluated and the bioactivity of a peptide can be more clearly explained using H S A (Nakai et al., 2003). 21. 1.7. Hypotheses 1. Lactoferricin and other cationic peptides can be produced by enzymatic digestion of food-grade lactoferrin. 2. The production of lactoferricin and its related peptides are affected by the level of iron saturation of lactoferrin. 3. Structure-function relationships of lactoferricin and its derivatives can be elucidated using multivariate data analysis techniques. 4. Lactoferricin can induce changes in model phospholipid bilayers. 5. Presence and absence of particular amino acids, namely arginine and tryptophan, in lactoferricin can affect the behaviour of lactoferricin on model phospholipid bilayers. 1.8. Objectives The specific objectives are to: 1. Produce lactoferricin by peptic digestion of food-grade bovine lactoferrin, 2. Evaluate the effects of iron saturation on the profile of cationic peptides produced by peptic digestion of lactoferrin, 3. Elucidate the structure-function relationship of lactoferricin and its derivatives using homology similarity analysis and multivariate data analysis techniques including principal component similarity and artificial neural network, 4. Examine the changes in phospholipid bilayers at different temperatures in the presence or absence of lactoferricin by vibrational spectroscopic techniques, and 5. Observe the effects of arginine and tryptophan moieties in Lfcin on the structures of phospholipid liposomes. 22 1 11 21 31 41 51 APRKNVRWCT ISQPEWFKCR RWQWRMKKLG APSITCVRRA FALECIRAIA EKKADAVTLD **** ********** ********** * + 61 • • GGMVFEAGRD PYKLRPVAAE IYGTKESPQT HYYAVAVVKK GSNFQLDQLQ GRKSCHTGLG + 121 RSAGWIIPMG ILRPYLSWTE SLEPLQGAVA KFFSASCVPC IDRQAYPNLC QLCKGEGENQ 181 CACSSREPYF GYSGAFKCLQ DGAGDVAFVK + 241 AFKECHLAQV PSHAVVARSV DGKEDLIWKL + 301 KDSALGFLRI PSKVDSALYL GSRYLTTLKN 361 WSQQSGQNVT CATASTTDDC IVLVLKGEAD 421 SSLDCVLRPT EGYLAVAVVK KANEGLTWNS 481 CAFDEFFSQS CAPGADPKSR LCALCAGDDQ 541 • ' AFVKNDTVWE NTNGESTADW AKNLNREDFR 601 SDRAAHVKQV LLHQQALFGK NGKNCPDKFC ETTVFENLPE KADRDQYELL CLNNSRAPVD LSKAQEKFGK NKSRSFQLFG SPPGQRDLLF LRETAEEVKA RYTRVVWCAV GPEEQKKCQQ ALNLDGGYIY TAGKCGLVPV LAENRKSSKH LKDKKSCHTA VDRTAGWNIP MGLIVNQTGS GLDKCVPNSK EKYYGYTGAF RCLAEDVGDV LLCLDGTRKP VTEAQSCHLA VAPNHAVVSR LFKSETKNLL FNDNTECLAK LGGRPTYEEY 661 LGTEYVTAIA NLKKCSTSPL LEACAFLTR Figure 1.1. Primary amino acid sequence (single letter code) of bovine lactoferrin. * denotes residues 17 to 42, representing the antimicrobial peptide, lactoferricin. + denotes the major contributors to one of the two iron-binding sites, and A denotes the major contributors to the second iron-binding site (adapted from Pierce et al., 1991). 23 Figure 1.2. Three-dimensional structure of diferric bovine lactoferrin at 2.8 A resolution where the arrows indicate the location of bound iron (Reprinted from Moore et al., 1997, Copyright 1997, with permission from Elsevier). 24 Peptide I. 1 7 F K C R R W Q W R M K K L G A PSITC V R R A F A 4 2 3195 Da, Peptide II. 1 A P R K N V R W C T ISQPE W 1 6 4 3 L E C I R A 4 8 2673 Da, I I and Peptide III. 1 A P R K N V R W C T ISQPE W F K C R R W Q W R M K K L G APSIT C V R R A F A 4 2 4 3 L E d l R A 4 8 5851 Da Peptide 1 a. 1 7 F K C R R W Q W R M K K L G A PSITC V R R A F 4 1 3123 Da, Peptide 1 b. 1 7 F K C R R W Q W R M K K L G A PSITC V R R A F A 4 2 3194 Da, Peptide 2. 1 7 F K C R R W Q W R M K K L G A PSITC V R R A F A L 4 3 3308 Da, and Peptide 3. 1 A P R K N V R W C T I S Q P E W 16 17, I I F K C R R W Q W R M K K L G A PSITC V R R A F ALECI A 48 5850 Da Figure 1.3. Antimicrobial peptide fragments isolated from bovine lactoferrin as reported in the literature; Peptides I, II, III are from Dionysius and Milne (1997) and Peptides, 1a, 1b, 2, and 3 are from Hoek et al. (1997). 25 (c) Figure 1.4. Hydrophobic (a), hydrophilic (b), and positive (c) side chains along the B-strands of bovine lactoferricin. The backbone of all residues is shown in gray, while the highlighted side chains are in black (Reprinted from Hwang et al., 1998, Copyright 1998, with permission from American Chemical Society). 26 Figure 1.5. Structural representation of the arrangement of phospholipids. Arrangement of the head group and fatty acid domains of phospholipids in the fluid liquid crystalline state (La), the ordered gel state (LB), and the reversed hexagonal state (HM) (Reprinted from Dowhan, 1997, Copyright 1997, with permission from the Annual Reviews www.annualreviews.org). 27 CO c CD 2800. 3020 b. 55 co g a: CO CD 1.1 0 .9 H 0.8-Q.7 0 . 6 ' 0 . 5 H 0. 4 H 0.3 ' I ' ' ' ' I ' 1_J_L A r r ^ m &0&^ 30 40 50 Temperature °C Figure 1.6. Raman spectra of dipalmitoylphosphatidylglycerol (DPPG) in the C-H stretching mode region (2800-3000 cm" 1) between 25 and 55 °C (a). Peak height intensity ratios (b): I2850/I2880 (A) and I2935/I2880 (B) (Reprinted from Vincent et al., 1993, Copyright 1993, with permission from American Chemical Society). 28 Table 1.1. Percentage of phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylglycerol (PG), and 1,3-bis(sn-3-phosphatidyl)-sn-glycerol (cardiolipin, CL) in bacterial membranes. Organism Percentages of Total Phospholipids P C P E P G C L References Gram-positive Micrococcus sp. Bacillus sp. Bacillus megaterium Gram-negative Azotobacter sp. Rhodopseudomonas sp. Escherichia coli E. coli wild-type strain K12 < 1 < 1 < 1 20-45 35 48 60-70 4-20 25-45 10-50 6 11 1-2 60-70 25-30 2-5 10-15 45-50 40-50 <1 70-80 20-25 < 5 70-80 15-20 5-20 Dowhan, 1997 Dowhan, 1997 Huijbregts et al. 2000 Dowhan, 1997 Dowhan, 1997 Dowhan, 1997 Morein et al., 1996 29 1.9. References Aishima, T.; Nakai, S. Chemometrics in flavour research. Food Rev. Int. 1991, 7, 33-101. Aishima, T.; Wilson, D. L ; Nakai, S . Application of simple algorithm to flavour optimization based on pattern similarity of G C profile. In Flavour Science and Technology. (Eds M. Martens, G. A. Dallen, and H. Russwrum) 1987. Chichester UK: John Wiley & Sons Ltd., pp. 501-508. Andersen, J . ; Osbakk, S.; Vorland, L.; Traavik, T.; Gutteberg, T. Lactoferrin and cyclic lactoferricin inhibit the entry of human cytomegalovirus into human fibroblasts. Antiviral Res. 2001, 57, 141-149. Anderson, B. F.; Baker, H. M.; Dodson, E. F.; Norris, G . E.; Rumball, S. V.; Waters, J . M.; Baker, E. N. Structure of human lactoferrin at 3.2A resolution. Proc. Natl. Acad. Sci. USA. 1987, 84, 1769-1773. Anderson, B. F.; Baker, H. M.; Norris, G . E.; Rice, D. W.; Baker, E. N. Structure of human lactoferrin: crystallographic structure analysis and refinement at 2.8A resolution. J. Mol. Biol. 1989, 209,. 711-734. Antonini, G . ; Catania, M. R.; Greco, R.; Longhi, C ; Pisciotta, M. G. ; Seganti, L.; Valenti, P. Anti-invasive activity of bovine lactoferrin against Listeria monocytogenes. J. Food Prot. 1997, 60, 267-271. Arnold, R. R.; Brewer, M.; Gauthier, J . J . Bactericidal activity of human lactoferrin: sensitivity of a variety of microorganisms. Infect. Immun. 1980,28,893-898. Arnold, R. R.; Russell, J . E.; Champion, W. J . ; Brewer, M.; Gauthier, J . J . Antibacterial activity of human lactoferrin: differentiation from the stasis of iron deprivation. Infect. Immun. 1982, 35, 792-799. Bechinger, B. Structure and functions of channel-forming peptides: magainins, cecropins, melittin and alamethicin. J. Membrane Biol. 1997, 756, 197-211. Beljaars, L.; Bakker, H.; van der Strate, B.; Smit, C ; Duijvestijn, A.; Meijer, D.; Molema, G. The antiviral protein human lactoferrin is distributed in the body to cytomegalovirus (CMV) infection-prone cells and tissues. Pharm. Res. 2002, 19, 54-62. Bellamy, W.; Takase, M.; Wakabayashi, H.; Kawase, K.; Tomita, M. Antibacterial spectrum of lactoferricin B, a potent bactericidal peptide derived from the N-terminal region of bovine lactoferrin. J. Appl. Bacteriol. 1992a, 73, 472-479. Bellamy, W.; Takase, M.; Yamauchi, K.; Wakabayashi, H.; Kawase, K.; Tomita, M. Identification of the bactericidal domain of lactoferrin. Biochim. Biophys. Acta 1992b, 7727,130-136. Bellamy, W.; Wakabayashi, H.; Takase, M.; Kawase, K.; Shimamura, S.; Tomita, M. Killing of Candida albicans by lactoferricin B, a potent antimicrobial peptide derived from the N-terminal region of bovine lactoferrin. Med. Microbiol. Immunol. 1993, 782, 97-105. Bellamy, W.; Yamauchi, K.; Wakabayashi, H.; Takase, M.; Takakura, N.; Shimamura, S.; Tomita, M. Antifungal properties of lactoferricin B, a peptide derived from the N-terminal region of bovine lactoferrin. Lett. Appl. Microbiol. 1994, 78, 230-233. 30 Bouchard, M.; Auger, M. Solvent history dependence of gramicidin-lipid interactions: a Raman and infrared spectroscopic study. Biophys. J. 1993, 65, 2484-2492. Brock, J . Lactoferrin: a multifunctional immunoregulatory protein? Immunol. Today 1995, 16, 417-419. Bullen, J . J . ; Rogers, H. J . ; Leigh, L. Iron-binding proteins.in milk and resistance to Escherichia coli infections in infants. Br. Med. J. 1972, 1, 69-75. Carey, P. R. Raman spectroscopy, the sleeping giant in structural biology, awakes. J. Biol. Chem. 1999, 274, 26625-26628. Chan, J . C. K.; Li-Chan, E. C. Y. Antimicrobial Peptides. In Nutraceutical Proteins and Peptides in Health and Disease. (Eds. Y. Mine and F. Shahidi.) 2006. Boca Raton: CRC/Taylor and Francis, pp. 99-136. Chantaysakorn, P.; Richter, R. L. Antimicrobial properties of pepsin-digested lactoferrin added to carrot juice and filtrates of carrot juice. J Food Prot. 2000, 63, 376-380. Chappie, D. S. ; Mason, D. J . ; Joannou, C. L.; Odell, E. W.; Gant, V.; Evans, R. W. Structure-function relationship of antibacterial synthetic peptides homologous to a helical surface region on human lactoferrin against Escherichia coli serotype 0111 . Infect. Immun. 1998, 66, 2434-2440. Chen, Y.; Shi, G. ; Xia, W.; Kong, C ; Zhao, S.; Gaw, A. F.; Chen, E. Y.; Yang, G. P.; Giaccia, A. J . ; Le, Q.-T.; Koong, A. C. Identification of hypoxia-regulated proteins in head and neck cancer by proteomic and tissue array profiling. Cancer Res. 2004, 64, 7302-7310. Dionysius, D. A.; Milne, J . M. Antibacterial peptides of bovine lactoferrin: purification and characterization. J. Dairy Sci. 1997, 80, 667-674. Dowhan, W. Molecular basis for membrane phospholipid diversity: why are there so many lipids? Ann. Rev. Biochem. 1997, 66, 199-232. Edman, M.; Jarhede, T.; Sjostrom, M.; Wieslander, A. Different sequence patterns in signal peptides from mycoplasmas, other Gram-positive bacteria, and Escherichia coli: a multivariate data analysis. Proteins. 1999,35,195-205. Ellison, R. T. I.; Giehl, T. J . Killing of gram-negative bacteria by lactoferrin and lysozyme. J. Clin. Invest. 1991, 88, 1080-1091. Ellison, R. T. I.; Giehl, T. J . ; LaForce, F. M. Damage of the outer membrane of enteric gram-negative bacteria by lactoferrin and transferrin. Infect. Immun. 1988, 56, 2774-2781. Epand, R. M.; Vogel, H. J . Diversity of antimicrobial peptides and their mechanisms of action. Biochim. Biophys. Acta 1999, 7462, 11-28. Facon, M. J . ; Skura, B. J . Antibacterial activity of lactoferricin, lysozyme and EDTA against Salmonella enteritidis. Int. Dairy J. 1996, 6, 303-313. Farnaud, S.; Evans, R. W. Lactoferrin - a multifunctional protein with antimicrobial properties. Mol. Immunol. 2003, 40, 395-405. Furtula, V.; Nakai, S. ; Amantea, G. F.; Laleye, L. Reverse-phase H P L C analysis of cheese samples aged by a fast-ripening process. J. Food Sci. 1994a, 59, 528-532, 567. 31 Furtula, V.; Nakai, S. ; Amantea, G. F.; Laleye, L. Reverse phase H P L C analysis of reference cheddar cheese samples for assessing accelerated cheese ripening. J. Food Sci. 1994b, 59, 533-538. Gaber, B. P.; Peticolas, W. L. On the quantitative interpretation of biomembrane structure by Raman spectroscopy. Biochim. Biophys. Acta 1977, 465, 260-274. Garthwaite, P. H. An interpretation of partial least squares. J. Am. Stat. Assoc. 1994, 89, 122-127. Geladi, P.; Kowalski, B. Partial least-squares regression: A tutorial. Anal. Chimica Acta 1986, 185, 1-17. Gifford, J . L ; Hunter, H. N.; Vogel, H. J . Lactoferricin: a lactoferrin-derived peptide with antimicrobial, antiviral, antitumor and immunological properties. Cell Mol. Life Sci. 2005, 62, 2588-2598. Groves, M. L. The isolation of a red protein from milk. J. Am. Chem. Soc. 1960, 82, 3345-3350. Hadden, J . M.; Bloemendal, M.; Haris, P. I.; Srai, S . K. S.; Chapman, D. Fourier transform infrared spectroscopy and differential scanning claorimetry of transferrins: human serum transferrin , rabbit serum transferrin and human lactoferrin. Biochim. Biophys. Acta 1994, 1205, 59-67. Hancock, R. E. W. Peptide antibiotics. Lancet 1997, 349, 418-422. Harmsen, M. C ; Swart, P .J . ; de Bethune, M.P., Pauwels, R.; De Clercq, E.; The, T.H.; Meijer, D. K. Antiviral effects of plasma and milk proteins: lactoferrin shows potent activity against both human immunodeficiency virus and human cytomegalovirus replication in vitro. J. Infect. Dis. 1995, 172, 380-388. Haug, B. E.; Svendsen, J . S. The role of tryptophan in the antibacterial activity of a 15-residue bovine lactoferricin peptide. J. Pept. Sci. 2001, 7, 190-196. Haug, B. E.; Skar, M. L.; Svendsen, J . S . Bulky aromatic amino acids increase the antibacterial activity of 15-residue bovine lactoferricin derivatives. J. Pept. Sci. 2001, 7, 425-432. Haukland, H. H.; Ulvatne, H.; Sandvik, K.; Vorland, L. H. The antimicrobial peptides lactoferricin B and magainin 2 cross over the bacterial cytoplasmic membrane and reside in the cytoplasm. FEBS Lett. 2001, 508, 389-393. Hellberg, S.; Sjostrom, M.; Skagerberg, B.; Wold, S. Peptide quantitative structure-activity relationships, a multivariate approach. J. Med. Chem. 1987, 30, 1126-1135. Hennart, P. F.; Brasseur, D. J . ; Delogne-Desnoeck, J . B.; Dramaix, M. M.; Robyn, C. E. Lysozyme, lactoferrin, and secretory immunoglobulin A content in breast milk: influence of duration of lactation, nutrition status, prolactin status and parity of mother. Am. J. Clin. Nutr. 1991, 53, 32-39. Hentges, D. J . ; Marsh, W. W.; Petschow, B. W.; Thai, W. R.; Carter, M. K. Influence of infant diets on the ecology of the intestinal tract of human flora-associated mice. J. Pediatr. Gastroent. Nutr. 1992, 14, 146-152. 32 Hoek, K. S.; Milne, J . M.; Grieve, P. A.; Dionysius, D. A.; Smith, R. Antibacterial activity of bovine lactoferrin-derived peptides. Antimicrob. Agents Chemother. 1997, 41, 54-59. Horimoto, Y.; Lee, K.; Nakai, S . Classification of microbial defects in milk using a dynamic headspace gas chromatograph and computer-aided data processing. 1. Principal component similarity analysis. J. Agric. Food Chem. 1997, 45, 733-742. Huang, H. W. Action of antimicrobial peptides: two-state model. Biochemistry 2000, 39, 8347-8352. Huang, S.-W.; Satue-Gracia, T.; Frankel, E. N.; German, J . B. Effect of lactoferrin on oxidative stability of corn oil emulsions and liposomes. J. Agric. Food Chem. 1999, 47, 1356-1361. Huijbregts, R.; Anton, I.; de Kroon, A.; de Kruijff, B. Topology and transport of membrane lipids in bacteria. Biochim. Biophys. Acta 2000, 1469, 43-61. Hutchens, T. W.; Rumball, S. V.; Lonnerdal, B. Lactoferrin: Structure and Function. In Advances in Experimental Medicine and Biology. (Eds T. W. Hutchens, S. V. Rumball, B. Lonnerdal) 1994,357. New York: Plenum Press, pp. 21-32. Hwang, P. M.; Vogel, H. J . Structure-function relationships of antimicrobial peptides. Biochem. Cell Biol. 1998, 76, 235-246. Hwang, P. M.; Zhou, N.; Shan, X. ; Arrowsmith, C. H.; Vogel, H. J . Three dimensional solution structure of lactoferricin B, an antimicrobial peptide derived from bovine lactoferrin. Biochemistry 1998, 37, 4288-4298. ligo, M.; Kuhara, T.; Ushida, Y.; Sekine, K.; Moore, M. A.; Tsuda, H. Inhibitory effects of bovine lactoferrin on coon carcinoma 26 lung metastasis in mice. Clin. Exp. Metastasis 1999, 77, 35-40. Ikeda, M.; Sugiyama, K.; Tanaka, T.; Tanaka, K.; Sekihara, Hi; Shimotohno, K.; Kato, N. Lactoferrin markedly inhibits hepatitis C virus infection in cultured human hepatocytes. Biochem. Biophys. Res. Commun. 1998, 245, 549-553. Ikeda, M.; Nozaki, A.; Sugiyama, K.; Tanaka, T.; Naganuma, A.; Tanaka, K.; Sekihara, H.; Shimotohno, K.; Saito, M.; Kato, N. Characterization of antiviral activity of lactoferrin against hepatitis C virus infection in human cultured cells. Virus Res. 2000, 66, 51-63. Jackson, M.; Mantsch, H. H.; Spencer, J . H. Conformation of magainin-2 and related peptides in aqueous solution and membrane environments probed by fourier transform infrared spectroscopy. Biochemistry 1992, 37, 7289-7293. Janes, K. A.; Albeck, J . G. ; Peng, L. X. ; Sorger, P. K.; Lauffenburger, D. A.; Yaffe, M. B. A high-throughput quantitative multiplex kinase assay for monitoring information flow in signaling networks: Application to sepsis-apoptosis. Mol. Cell. Proteomics 2003, 2, 463-473. Johansson, B. Isolation of an iron containing red protein from human milk. Acta Chem. Scand. 1960, 14, 510-512. Kang, J . H.; Lee, M. K.; Kim, K. L.; Hahm, K.-S. Structure-biological activity relationships of 11-residue highly basic peptide segment of bovein lactoferrin. Int. J. Pept. Protein Res. 1996, 48, 357-363. 33 Kanyshkova, T. G. ; Semenov, D. V.; Buneva, V. N.; Nevinsky, G. A. Human milk lactoferrin binds two DNA molecules with different affinities. FEBS Lett. 1999, 451, 235-237. Kappeler, S . R.; Ackermann, M.; Farah, Z.; Puhan, Z. Sequence analysis of camel (Camelus dromedarius) lactoferrin. Int. Dairy J. 1999,9,481-486. Kawasaki, Y.; Tazume, S.; Shimizu, K.; Matsuzawa, H.; Dosako, S.; Isoda, H.; Tsukiji, M.; Fujimura, R.; Muranaka, Y.; Isihida, H. Inhibitory effects of bovine lactoferrin on the adherence of enterotoxigenic Escherichia colito host cells. Biosci. Biotechnol. Biochem. 2000, 64, 348-354. Laporte, L ; Benevides, J . M.; Thomas, G. J . J . Molecular mechanism of DNA recognition by the a subunit of the Oxytricha telomere binding protein. Biochemistry 1999, 38, 582-588. Laroche, G . ; Carrier, D.; Pezolet, M. Study of the effect of poly(L-lysine) on phosphatidic acid and phosphatidylcholine/ phosphatidic acid bilayers by Raman spectroscopy. Biochemistry "WW, 27, 6220-6228. Laroche, G . ; Dufourc, E. J . ; Pezolet, M.; Dufourcq, J . Coupled changes between lipid order and polypeptide conformation at the membrane surface. A 2H NMR and Raman study of polylysine-phosphatidic acid systems. Biochemistry 1990, 29, 6460-6465. Lejon, T.; Strom, M. B.; Svendsen, J . S . Antibiotic activity of pentadecapeptides modelled from amino acid descriptors. J. Pept. Sci. 2001a, 7, 74-81. Lejon, T.; Strom, M. B.; Svendsen, J . S . Is information about peptide sequence necessary in multivariate analysis? Chemometrics Intelligent Lab. Sys. 2001b, 57, 93-95. Levay, P. F.; Viljoen, M. Lactoferrin: a general review. Haematologica 1995, 80, 252-267. Lewis, R. N. A. H.; McElhaney, R. N. Vibrational spectroscopy of lipids. 2002. John Wiley & Sons Ltd. Li-Chan, E. C. Y. The applications of Raman spectroscopy in foods science. Trends Food Sci. Technol. 1996, 7, 361-370. Li-Chan, E. C. Y.; Nakai, S. ; Hirotsuka, M. Raman spectroscopy as a probe of protein structure in food systems. In Protein Structure-Function Relationships in Foods. (Eds R. Y. Yada, R. L. Jackman, and J . L. Smith) 1994. New York: Blackie Academic & Professional, pp. 163-197 Lohner, K.; Prenner, E. J . Differential scanning calorimetry and X-ray diffraction studies of the specificity of the interaction of antimicrobial peptides with membrane-mimetic systems. Biochim. Biophys. Acta 1999, 1462, 141-156. Ludtke, S . J . ; He, K.; Heller, W. T.; Harroun, T. A.; Yang, L.; Huang, H. W. Membrane pores induced by magainin. Biochemistry 1996,35,13723-13728. Mader, J . S. ; Salsman, J . ; Conrad, D. M.; Hoskin, D. W. Bovine lactoferricin selectively induces apoptosis in human leukemia and carcinoma cell line. Mol. Cancer Ther. 2005, 4, 612-624. Maruyama, T.; Takeuchi, H. Water accessibility to the tryptophan indole N-H sites of gramicidin A transmembrane channel: Detection of positional shifts of tryptophans 11 and 13 along the channel axis upon cation binding. Biochemistry 1997, 36, 10993-11001. 34 Maruyama, T.; Takeuchi, H. Raman linear intensity difference of membrane-bound peptides: indole ring orientations of tryptophans 11 and 13 in the gramicidin A transmembrane channel. Biospectroscopy 1998,4,171-184. Marchetti, M.; Superti, F.; Ammendolia, M.; Rossi, P.; Valenti, P.; Seganti, L. Inhibition of poliovirus type 1 infection by iron-, manganese- and zinc-saturated lactoferrin. Med. Microbiol. Immunol. .1999, 187, 199-204. Masson, P. L ; Heremans, J . F. Lactoferrin in milk from different species. Comp. Biochem. Physiol. 1971, 39B, 119-129. Masuda, C ; Wanibuchi, H.; Sekine, K.; Yano, Y.; Otani, S.; Kishimoto, T.; Tsuda, H.; Fukushima, S. Chemopreventive effects of bovine lactoferrin on N-butyl-N-(4-hydroxybutyl)-nitrosamine-induced rat bladder carcinogenesis. Jpn. J. Cancer Res. 2000, 90, 262-267. Mata, L.; Sanchez, L.; Headon, D. R.; Calvo, M. Thermal denaturation of human lactoferrin and its effect on the ability to bind iron. J. Agric. Food Chem. 1998, 46, 3964-3970. Mateos-Caceres, P. J . ; Garcia-Mendez, A.; Lopez Farre, A.; Macaya, C ; Nunez, A.; Gomez, J . ; Alonso-Orgaz, S.; Carrasco, C ; Burgos, M. E.; de Andres, R.; Granizo, J . J . ; Farre, J . ; Rico, L. A. Proteomic analysis of plasma from patients during an acute coronary syndrome. J. Am. Coll. Cardiol. 2004, 44, 1578-1583. Miller-Catchpole, R.; Kot, E.; Haloftis, G.,;Furmanov, S.; Bezkorovainy, A. Lactoferrin can supply iron for the growth of Bifidobacterium breve. Nutr. Res. 1997, 17, 205-213. Moore, S. A.; Anderson, B. F.; Groom, C. R.; Maridas, M.; Baker, E. N. Three-dimensional structure of diferric bovine lactoferrin at 2.8 A resolution. J. Mol. Biol. 1997, 274, 222-236. Morein, S.; Andersson, A . -S . ; Rilfors, L.; Lindblom, G. Wild-type Escherichia coli cells regulate the membrane lipid composition in a 'window' between gel and non-lamellar structures. J. Biol. Chem. 1996, 271, 6801-6809. Nagasako, Y.; Saito, H.; Tamura, Y.; Chimamura, S.; Tomita, M. Iron-binding properties of bovine lactoferrin in iron-rich solution. J. Dairy Sci. 1993, 76, 1876-1881. Nakai, S. ; Wang, Z. H.; Dou, J . ; Nakamura, S.; Ogawa, M.; Nakai, E.I.; Vanderstoep, J . Gas chromatography / principal component similarity system for detection of E. coli and S. aureus contaminating salmon and hamburger. J. Agric. Food Chem. 1999, 47, 576-583. Nakai, S. ; Chan, J . ; Li-Chan, E.; Dou, J . ; and Ogawa, M. Homology similarity analysis of sequences of latoferricin and its derivatives. J. Agric. Food Chem. 2003, 51, 1215-1223. Nguyen, D. V.; Rocke, D. M. Tumor classification by partical least squares using microarray gene expression data. Bioinformatics 2002, 18, 39-50. Nonnecke, B. J . ; Smith, K. L. Inhibition of mastitic bacteria by bovine milk apo-lactoferrin evaluated by in vitro microassay of bacterial growth. J. Dairy Sci. 1984, 67, 606-613. 35 Odell, E. W.; Sarra, R.; Foxworthy, M.; Chappie, D. S.; Evans, R. W. Antibacterial activity of peptides homologous to a loop region in human lactoferrin. FEBS Lett. 1996, 382, 175-178. Opekarova, M.; Tanner, W. Specific lipid requirements of membrane proteins—a putative bottleneck in heterologous expression. Biochim. Biophys. Acta 2003, 1610, 11-22. Paulsson, M. A.; Svensson, U.; Kishore, A. R.; Naidu, S. Thermal behaviour of bovine lactoferrin in water and its relation to bacterial interaction and antibacterial activity. J. Dairy Sci. 1993, 76, 3711 -3720. Pelton, J . T.; McLean, L. R. Spectroscopic methods for analysis of protein secondary structure. Anal. Biochem. 2000, 277, 167-176. Petschow, B. W.; Talbott, R. D.; Batema, R. P. Ability of lactoferrin to promote the growth of Bifidobacterium spp. in vitro is independent of receptor binding capacity and iron saturation level. J. Med. Microbiol. 1999, 48, 541-549. Pierce, A.; Colavizza, D.; Benaissa, M.; Maes, P.; Tartar, A.; Montreuil, J . ; Spik, G . Molecular cloning and sequence analysis of bovine lactotransferrin. Eur. J. Biochem. 1991, 796, 177-184. Pripp, A. H.; Isaksson, T.; Stepaniak, L.; Sorhaug, T. Quantitative structure-activity relationship modeling of ACE-inhibitory peptides derived from milk proteins. Eur. Food Res. Technol. 2004,279,579-583. Puddu, P.; Borghi, P.; Gessani , S. ; Valenti, P.; Belardelli, F.; Seganti, L. Antiviral effect of bovine lactoferrin saturated with metal ions on early steps of human immunodeficiency virus type 1 infection. Int. J. Biochem. Cell Biol. 1998, 30, 1055-1062 Recio, I.; Visser, S . Two ion-exchange chromatographic methods for the isolation of antibacterial peptides from lactoferrin in situ enzymatic hydrolysis on an ion-exhange membrane. J. Chromatogr. A. 1999, 837, 191-201. Rekdal, 0 . ; Andersen, J . ; Vorland, L. H.; Svendsen, J . S. Construction and synthesis of lactoferricin derivatives with enhanced antibacterial activity. J. Pept. Sci. 1999, 5, 32-45. Roberts, A. K.; Chierici, R.; Sawatzki, G. ; Hill, M. J . ; Volpato, S.; Vigi, V. Supplementation of adapted formula with lactoferrin: 1. Effects on the infant faecal flora. Acfa Pediatr. 1992.87,119-124. Saito, H.; Miyakawa, H.; Tamura, Y.; Shimamura, S.; Tomita, M. Potent bactericidal activity of bovine lactoferrin hydrolysate produced by heat treatment at acidic pH. J. Dairy Sci. 1991, 74, 3724-3730. Sanchez, L., Aranda, P., Perez, M. D. and Calvo, M. Concentration of lactoferrin and transferrin throughout lactation in cow's colostrum and milk. Biol. Chem. Hoppe Seyler. 1988,369,1005-1008. Satue-Gracia, M. T.; Frankel, E. N.; Rangavajhyala, N.; German, J . B. Lactoferrin in infant formulas: effect on oxidation. J. Agric. Food Chem. 2000. 48, 4984-4990. Schibli, D. J . ; Hwang, P. M.; Vogel, H. J . The structure of the antimicrobial active center of lactoferricin B bound to sodium dodecyl sulfate micelles. FEBS Lett. 1999, 446, 213-217. 36 Shai, Y . Mechanism of the binding, insertion and destabilization of phospholipid bilayer membranes by a-helical antimicrobial and cell non-selective membrane-lytic peptides: Biochim. Biophys. Acta 1999, 1462, 55-70. Shimazaki, K.; Tazume, T.; Uji, K.; Tanake, M.; Kumura, H.; Mikawa, K.; Shimo-Oka, T. Properties of a heparin-binding peptide derived from bovine lactoferrin. J. Dairy Sci. 1998, 81, 2841-2849. Skates, S.; Iliopoulos, O. Molecular markers for early detection of renal carcinoma: investigative approach. Clin. Cancer Res. 2004, 10, 6296S-6301S. Snyder, R.; Hsu, S.; Krimm, S. Vibrational spectra in the C - H stretching region and the structure of the polymethylene chain. Spectrochim. Acta 1978, 34A, 395-406 Snyder, R.; Scherer, J . ; Gaber, B. Effects of chain packing and chain mobility on the Raman spectra of biomembranes. Biochim. Biophys. Acta 1980, 607, 47-53. Snyder, R. Strauss, H. and Elliger, C. C - H stretching modes and the structure of n-alkyl chains. I. Long, disordered chains. J. Chem. Phys. 1982, 86, 5145-5150. Sorensen, M.; Sorensen, S. P. L. The proteins in whey. Compt. Rend. Lab. Carlsberg 1939, 23, 55-99. Strom, M. B.; Rekal, 0 . ; Svendsen, J . S. Antibacterial activity of 15-residue lactoferricin derivatives. J. Pept. Res. 2000, 56, 265-274. Strom, M. B.; Rekdal, 0 . ; Stensen, W.; Svendsen, J . S . Increased antibacterial activity of 15-residue murine lactoferricin derivatives. J. Pept. Res. 2001, 57, 127-139. Subbalakshmi, C ; Krishnakumari, N.; Nagaraj, R.; Sitaram, N. Requirements for antibacterial and hemolytic activities in the bovine neutrophil derived 13-residue peptide indolicidin. FEBS Lett. 1996, 395, 48-52. Subbalakshmi, C ; Sitaram, N. Mechanism of antimicrobial action of indolicidin. FEMS Microbiol. Lett. 1998, 760, 91-96. Superti, F.; Ammendolia, M.; Valenti, V.; Seganti, L. Antirotaviral activity of milk proteins: lactoferrin prevents rotavirus infection in the enterocyte-like cell line HT-29. Med. Microbiol. Immunol. 1997, 786, 83-91. Tomita, M.; Bellamy, W.; Takase, M.; Yamauchi, K.; Wakabayashi, H.; Kawase, K. Potent antibacterial peptides generated by pepsin digestion of bovine lactoferrin. J. Dairy Sci. 1991, 74, 4137-4142. Tomita, M.; Takase, M.; Bellamy, W.; Shimamura, S. A review: The active peptide of lactoferrin. Acta Paediatr. Jpn. 1994, 36, 585-591. Tomita, M.; Shimamura, S.; Kawase, K.; Fukuwatari, Y. ; Takase, M.; Bellamy, W.; Hagiwara, T.; Matukuma. A process for large-scale production of antimicrobial peptide in high purity. 1999. Canada: Canadian Intellectual Property Office, C A 2070882. Ulvatne, K ; Haukland, H. H.; Olsvik, 0 . ; Vorland, L. H. Lactoferricin B causes depolarization of the cytoplasmic membrane of Escherichia coli ATCC 26922 and fusion of negatively charged liposomes. FEBS Lett. 2001, 492, 62-65. 37 Ushida, Y.; Sekine, K.; Kuhara, T.; Takasuka, N.; ligo, M.; Maeda, M.; Tsuda, H. Possible chemopreventative effects of bovine lactoferrin on esophagus and lung carcinogenesis in the rat. Jpn. J. Cancer Res. 1999, 91, 582-588. van der Strate, B.; Beljaars, L ; Molema, G. ; Harmsen, M.; Meijer, D. Antiviral activities of lactoferrin. Antiviral Res. 2001, 52, 225-239. Venkitanarayanan, K. S.; Zhao, T.; Doyle, M. P. Antibacterial effect of lactoferricin B on Escherichia coli 0157:H7 in ground beef. J. Food Protect. 1999, 62, 747-750. Vincent, J . S.; Revak, S. D.; Cochrane, C. D.; Levin, I. W. Interactions of model human pulmonary surfactants with a mixed phopholipid bilayer assembly: Raman spectroscopic studies. Biochemistry 1993, 32, 8228-8238. Vodovotz, Y.; Arteaga, G. E.; Nakai, S. Principal component similarity analysis for classification and its application to G C data of mango. Food Res. Int. 1993, 26, 355-363. Voland, L. H.; Ulvatne, H.; Rekdal, 0 . ; Svendsen, J . S . Initial binding sites of antimicrobial peptides in Staphylococcus aureus and Escherichia coli. Scand. J. Infect. Dis. 1999, 31, 467-473. Wakabayashi, H.; Abe, S.; Okutomi, T.; Tansho, S.; Kawase, K.; Yamaguchi, H. Cooperative anti-Candida effects of lactoferrin or its peptides in combination with azole antifungal agents. Microbiol. Immunol. 1996, 40, 821-825. Wakabayashi, H.; Abe, S. ; Teraguchi, S. ; Hayasawa, H.; Yamaguchi, H. Inhibition of hyphal growth of azole-resistant strains of Candida albicans by triazole antifungal agents in the presence of lactoferrin-related compounds. Antimicrob. Agents Chemother. 1998, 42, 1587-1591. Wakabayashi, H.; Uchida, T.; Yamauchi, K.; Teraguchi, S. ; Hayasawa, H.; Yamaguchi, H. Lactoferrin given in food facilitates dermatophytosis cure in guinea pig models. J. Antimicrob. Chemother. 2000, 46, 595-601. Wang, Z. H.; Dou, J . ; Macura, D.;Durance, T.D.; Nakai, S . Solid phase extraction for G C analysis of beany flavours in soymilk. Food Res. Int. 1997, 30, 503-511. Wei, Z.; Nisimura, T.; Yoshida, S. Presence of a glycan at a potential N-glycan site, Asn-281, of bovine lactoferrin. J. Dairy Sci. 2000, 83, 683-689. Williams, R. W.; Weaver, J . L. Secondary structure of substance P bound to liposomes in organic solvents and in solution from Raman and C D spectroscopy. J. Biol. Chem. 1990,265,2505-2513. Wu, M.; Maier, E.; Benz, R.; Hancock, R. E. W. Mechanism of interaction of different classes of cationic antimicrobial peptides with planar bilayers and with the cytoplasmic membrane of Escherichia coli. Biochemistry 1999, 38, 7235-7242. Yamauchi, K.; Tomita, M.; Giehl, T. J . ; Ellison, R. T. I. Antibacterial activity of lactoferrin and a pepsin-derived lactoferrin peptide fragment. Infect. Immun. 1993, 61, 719-728. Yamazaki, N.; Yamauchi, K.; Kawase, K.; Hayasawa, H.; Nakao, K.; Moto, I. Antibacterial effects of lactoferrin and a pepsin-generated lactoferrin peptide against Helicobacter pylori in vitro. J. Infect. Chemother. 1997,3,85-89. 38 CHAPTER 2. PRODUCTION OF LACTOFERRICIN AND OTHER CATIONIC PEPTIDES FROM FOOD GRADE BOVINE LACTOFERRIN WITH VARYING IRON SATURATION LEVELS 1 2.1. Introduction The presence of an antimicrobial peptide sequence in bovine lactoferrin (LF) was first noted when heat treatment of LF increased its antibacterial activity (Saito et al., 1991). It was suggested that at least one bactericidal domain was released by heating LF at pH 2.0 at 120 °C for 15 minutes. Heat-treated LF was later fractionated by reverse-phase H P L C , and a peptide fraction with strong activity, namely lactoferricin (Lfcin), was identified. Peptic digestion of LF at 37 °C also led to similar results (Bellamy et al., 1992). Since then, a number of production and purification strategies have been conducted for the isolation of Lfcin, including the use of a cation exchange column followed by reverse-phase H P L C (Dionysius and Milne, 1997), bead-based cation-exchange chromatography on SP-Sepharose Fast Flow resin (Recio and Visser, 1999), affinity chromatography with an immobilized heparin column (Shimazaki et al., 1998), and chymosin digestion of LF followed by ion-exchange membrane filtration and reverse-phase H P L C (Hoek et al., 1997). All of these studies were conducted on a laboratory scale using laboratory-grade reagents and conditions, producing a small quantity of Lfcin for scientific research purposes. In 1999, Tomita et al. (1999) patented a method using hydrophobic or cation exchange medium to produce Lfcin from peptic digested LF in large scale. Pepsin is a non-specific protease and could cleave at multiple locations on LF. Since LF is a basic protein with a pi of 8.69, it is very likely that cationic peptides other than Lfcin are concurrently produced during the peptic digestion. Furthermore, LF is an iron-binding protein which is 20-25% saturated with iron under normal and healthy physiological conditions. The binding of iron to lactoferrin is an important factor in the thermal stability of its structure. Hadden et al. (1994) observed that the removal of iron from human LF led to increases in the extent of 1 H - 2 H exchange and postulated a "loosening" of human LF upon iron removal. Likewise, Paulsson et al. (1993) reported that iron-saturated (holo-) bovine lactoferrin was more resistant to heat-induced changes than was the iron-depleted (apo-) LF. However, the potential effect of varying iron levels on the profile of peptides produced upon peptic digestion has not been previously reported. A version of this chapter has been presented. Chan, J .C.K. , and Li-Chan, E. (March 2004). Production of lactoferricin B and other cationic peptides from lactoferrin. Selected finalist at the Withycombe-Charalambous Graduate Student Symposium, 227 t h American Chemical Society National Meeting, Anaheim, CA. A version of this chapter has been submitted for publication. Chan, J . C. K., and Li-Chan, E. C. Y. Production of lactoferricin and other cationic peptides from food grade bovine lactoferrin with varying iron saturation levels. Journal of Agricultural and Food Chemistry. 39 The objectives of this study were to develop an isolation strategy with potential for industrial scale production of Lfcin, to identify bther cationic peptides generated during the peptic digestion of LF, and to investigate the effects of initial iron-saturation levels of LF on the composition of the resultant peptide mixture produced by peptic digestion. 2.2. Materials and Methods 2.2.1. Preparation of Lactoferrin with Varying Iron Saturation Levels Food-grade bovine lactoferrin (LF, a gift from DMV International Nutritional, Fraser, NY) was used to prepare samples with different levels of iron saturation based on previously described methods (Nagasako et al., 1993; Marchetti et al., 1999; and Moore et al., 1997). Apo-lactoferrin (apo-LF, iron-depleted) was prepared by dialyzing LF against 0.1 M citric acid at pH 2.0. Holo-lactoferrin (holo-LF, iron-saturated) was prepared by adding a 10 fold molar excess of FeCI 3 to LF. Excess citric acid or iron was then removed by exhaustive dialysis against d d H 2 0 . All dialyses were conducted overnight at 4 °C until there was no change in the conductivity of the external d d H 2 0 . 2.2.2. Digestion of Lactoferrin Peptic digestion of LF was based on methods previously described (Tomita et al., 1991; Dionysius and Milne, 1997). The LF (5%, w/v) was digested with porcine pepsin (3% of LF, 2,500-3,500 units mg"1 protein, Sigma Product Number P7012, Sigma-Aldrich, Saint Louis, MO) at 37 °C for 4 hours. The digest was maintained at pH 2.0 using 2 M HCI and was terminated by adjusting the pH to 7.0 using 2 M NaOH and heating at 80 °C for 15 minutes. The resultant digest was centrifuged at 15,000 x g for 20 minutes at 4 °C and the supernatant was stored at -25 °C until further analysis 2.2.3. Purification of Lactoferricin and Other Peptides The supernatant from the peptic digest was thawed at 4 °C and applied at a flow rate of 0.25 mL per minute to a column packed with an industrial grade, carboxylic acid cation exchanger (Purolite C-106 E P , Purolite, Bala Cynwyd, PA). The resin was pre-equilibrated with 50 mM sodium phosphate buffer (PB) at pH 6.5. After washing off unbound peptides with P B at 1.0 mL per minute, bound (cationic) peptides were eluted with PB containing stepwise gradient of 0.2, 0.5, 0.8, 1.0, and 2.0 M NaCI, at 1.0 mL per minute. The fractionation process was monitored by measuring the conductivity and UV absorbance at 280 nm of the eluant. Fractions were lyophilized and stored at -25 °C. 40 2.2.4. Peptide Mass Analysis The whole peptic digest as well as fractions separated by cation exchange chromatography were analyzed after desalting (ZipTipC is, Millipore, Bellerica, MA) by matrix assisted laser desorption ionization time of flight mass spectrometry (MALDI-ToF MS) using an Applied Biosystems Voyager System (Voyager-DE ™ S T R Biospectrometry™, A M E Bioscience Ltd., London, UK) under the following conditions: voltage, 20,000 V; laser intensity, 2500; laser rep rate, 20.0 Hz; calibration matrix, a-cyano-4-hydroxycinnamic acid. Post source decay following MALDI-ToF was performed to obtain peptide sequence information. All MALDI-ToF analyses were performed by Ms. Suzanne Perry at the LMB-BL Proteomic Core Facility, University of British Columbia. 2.2.5. Identification of Peptide Sequences Peptide sequences were identified using the FindPept software tool (http://ca.expasy. org/tools/findpept.html) (Gattiker et al., 2002). 2.2.6. Iron Content in Lactoferrins Iron content of lactoferrin solutions containing different iron saturation levels was determined using inductively coupled plasma-mass spectroscopy (ICP-MS) with an ultrasonic nebulization pulse membrane desolvation inlet Elemental Research Inc., North Vancouver, Canada. 2.2.7. Protein Structure by Raman Spectroscopy Raman spectra of each LF sample (10% w/v, adjusted to pH 2.0) were measured using a laser Raman spectrometer (Model NR-1100, J A S C O Inc., Tokyo, Japan) with excitation from the 488 nm line of an argon ion laser (Coherent Innova 70C series, Coherent Laser Group, Santa Clara, CA, cooled with the Coherent Laser Pure heat exchanger system). The Raman scattering of samples placed in a hematocrit capillary tubes in a transverse arrangement (capillary held horizontally and incident laser beam perpendicular to the capillary axis) was measured under the following conditions: laser power, 200 mW; slit height, 4 mm; spectral resolution, 5 cm" 1 at 19,000 cm" 1 ; sample speed, 120 cm" 1 min"1 with data collected at every cm" 1 . Spectra from at least 6 scans were averaged for each duplicate sample. Secondary structure composition of LFs was estimated by applying the algorithm of Williams (1983) to analyze the Raman spectra in the amide I region, using the R S A P program (version 2.1) of Przybycien and Bailey (1989). 41 2.3. Results 2.3.1. Production of Lactoferricin and Other Cationic Peptides from Lactoferrin Pepsin digestion of food grade LF resulted in a mixture of peptides with a wide range of masses (Figure 2.1). A peptide with mass of 968 Da was the most dominant product, followed by peptides with masses of 4430 and 4268 Da. Peptides with masses around 3100 and 1900 Da were also detected as broad peaks by MALDI-ToF analysis. The cation exchange chromatogram of pepsin-digested LF (Figure 2.2) shows that most of the LF digestion products that had pi below neutral pH did not bind to the industrial grade cation exchange resin at pH 6.5, and were recovered in the flow through fraction. The majority of the bound LF digestion products were eluted using buffers containing 0.2 M, 0.5 M, and 0.8 M NaCI. Elutions with 1.0 M and 2.0 M NaCI'released more tightly bound peptides with • higher pi from the column. A 965 Da peptide was the most dominant peptide found in the flow through fraction as revealed on a MALDI-ToF mass spectrum (Figure 2.3a). Peptides with masses of 4275 and 4433 Da were also detected in the flow through fraction. The MALDI-ToF mass spectrum of the 0.2 M NaCI fraction showed similar components to those of the flow through fraction (Figure 2.3b); however, the peptides with masses 4268, and 4430 Da were more dominant constituents in the 0.2 M NaCI fraction, compared to the peptide at 965 Da. Peptides with masses near 2700 Da also became noticeable in this fraction. The addition of 0.5 M NaCI to phosphate buffer (PB) eluted a peptide mixture with very different mass profile (Figure 2.3c). In addition to the 4268 and 4430 Da peptides that were previously detected in other fractions, broad peaks near 1971 and 3196 Da were observed, indicating that many peptides with masses near 1971 and 3196 Da were present in the fraction. The 0.8 M NaCI eluate contained peptides with masses close to 3196 Da. In addition, peptides with lower masses (1971 and 2675 Da) were also noted (Figure 2.3d). Peptides with masses of 1052, 1971, 3196, 4275, and 4430 Da were found in the 1.0 M NaCI fraction (Figure 2.3e). Peptides with masses of 3124 and 3196 Da dominated the 2.0 M NaCI fraction (Figure 2.3f). These masses correspond to the 25- and 26-mer of Lfcin, residues 17-41 or 42 of LF, respectively (Recio and Visser, 1999). The peptides in the 2.0 M NaCI fraction with masses of 3124 and 3196 Da represent 7% of the theoretical amount of Lfcin in the pepsin hydrolysates of food-grade bovine LF. 2.3.2. A Two-Step Process for Lactoferricin Isolation The aforementioned results indicated that the peptides of interest, Lfcin with a mass of 3124 or 3196 Da, were eluted by PB containing high NaCI levels, suggesting that Lfcin was 42 highly cationic and bound strongly with the cation exchange resin. Hence, a large amount of LF digest was loaded to exceed the column capacity to allow a competitive displacement of the other peptides by highly cationic Lfcin in the LF digest. LF peptic hydrolysate from 5.0 g of LF was applied to 5.0 mL cation exchange resin and fractionated using conditions described in the previous section, except that only PB containing 1.0 M and 2.0 M NaCI were used to elute Lfcin. Figure 2.4a shows the elution profile of the fractionation. A large amount of peptides passed through the column as LF digest was being loaded on to the column. These peptides include those which passed through the column without binding to the cation exchange resin, as well as some which might have formed weak bonds with the resin and were later displaced by other stronger cationic peptides in the digest. Peptides that formed stronger bonding with the cation exchange resin were later eluted off the column by 1.0 M NaCI and 2.0 M NaCI in PB . Figures 2.4b, c, and d illustrate the MALDI-ToF mass spectra of fractions eluted at various points in time. Towards the end of the sample loading, the flow through fraction contained primarily peptides with masses around 3200 and 4200 Da (Figure 2.4b). Subsequent elution using 1.0 M NaCI buffer produced a mixture of peptides with masses of 3196 and 3124 Da, along with lower levels of peptides with masses around 1970 and 1053 Da (Figure 2.4c). The MALDI-ToF mass spectrum of the 2.0 M NaCI fraction (Figure 2.4d) shows that the fraction was predominantly composed of peptides of masses of 3196 and 3124 Da. The signal intensity observed in Figure 2.4d indicates that the purity of Lfcin recovered in the 2.0 M fraction is comparable to the purity (> 99%) of the standard used for the calibration of the MALDI-ToF mass spectrometer (data not shown). Assuming that the peptides with 3196 and 3124 Da corresponded to the 26- and 25-mer of Lfcin, a total of 70 mg (35% of the theoretical yield from 5 g of LF) of highly purified Lfcin was recovered in the 2.0 M NaCI fraction using this purification strategy. 2.3.3. The Identification of Lactoferricin and Other Cationic Peptides Identification of the fractionated peptides using FindPept software was based on the mass data obtained, cleavage properties of pepsin, digestion conditions, and known sequence information of LF. The results indicated that peptides with masses of 3196 and 3124 Da could represent the 26- and 25-mer peptide segments of residues 17-42 and 17-41, respectively in LF, with the following sequences, 1 7 F K C R R W Q W R M K K L G A PSITC V R R A F 4 1 A 4 2 . These findings were verified by post source decay performed on the 3196 and 3124 Da peptides detected in the 2.0M NaCI fraction, yielding many smaller peptides with masses of 122, 445, 485, 488, 2319, 2372, 3124, and 3196 Da (Figure 2.5a). These masses correspond to peptide fragments isolated from F K C R R W Q W R M K K L G A PSITC V R R A F A (Figure 2.5b). 43 Peptlden tool was used to identify potential peptide sequences in LF containing cationic amino acids with high pi values and whose calculated mass values were within ± 2 Da of the mass measured by MALDI-ToF MS. Sequence information obtained by Peptlden of other cationic peptides produced during peptic hydrolysis of LF was summarized in Table 2.1. Although many of these potentially bioactive peptides are located near the current peptide of interest (lactoferricin), a number of them are at different locations on the lactoferrin sequence. 2.3.4. Effects of Iron Saturation Levels of Lactoferrin ICP-MS analyses showed the iron saturation level of the LF solution dialyzed against 0.1 M citric acid (apo-LF) was decreased to 16%, compared to 20% in the original untreated LF. When excess F e 3 + was added to the LF (holo-LF), iron saturation was increased to 350%. Structural analyses of the LFs determined by Raman spectroscopy show that the Raman amide I band near 1500 - 1800 cm" 1 of apo-LF was less intense than the spectra of LF and holo-LF (data not shown). In addition, as determined using the R S A P program, holo-LF contained little a-helical structure accompanied with more B-sheet and overall random content at pH 2.0 (Table 2.2). The Raman spectra and estimated secondary structural content of LF and apo-LF appeared to be similar. Removing some of the Fe from LF decreased the a-helix content from 42 to 34%, increased the p-sheet content from 32 to 47%, and lowered the overall random content from 27 to 19% (Table 2.2). Results from the MALDI-ToF analyses show that digests of LF (Figure 2.1) and especially of apo-LF (Figure 2.6a) contained more peptide fragments with lower masses (less than 1000 Da) than the digest of holo-LF (Figure 2.6b). Similarly, peptides eluted by 1.0 M NaCI in PB from the LF (Figure 2.3e) and apo-LF digest (Figure 2.7a) also contained more peptide fragments with lower masses (less than 1000 Da) than those of holo-LF (Figure 2.7b). However, the compositions of the 2.0 M fractions produced from LF's with different levels of iron saturation were similar (Figure 2.8 and Figure 2.3f). The 3196 Da peptide remained the most dominant peptide in all three preparations, whereas the 3124 Da peptide was the second most abundant peptide. 2.4. Discussion 2.4.1. Production of Lactoferricin and Other Cationic Peptides from Lactoferrin Peptic digestion of food grade bovine lactoferrin (LF) produced a mixture of peptides with various masses (Figure 2.1). Pepsin is an enzyme with low substrate specificity (Keil, 1992), and may cleave at the P1 or P1 ' position of Phe, Tyr, Trp, and Leu, corresponding to over 200 possible cleavage sites on LF. Furthermore, its cleavage specificity is lost at pH greater than or equal to 2.0 (Keil, 1992). Slight changes of the pH during the digestion process 44 would result in varying peptic cleavage patterns. Hence, the profile of the resulting peptides would vary greatly depending on the digestion conditions used. Under the digestion conditions used in the present study, 5% (w/v) LF solution with 3% (w/w) pepsin at pH 2.0 and 37 °C for 4 hours, an array of peptides with wide ranges of size and charge properties were produced. Due to differences in amino acid composition of these peptides, they carried different charges and were bound to the cation-exchange column with different strengths. The use of buffer containing various levels of salt allowed the fractionation of peptides produced from the pepsin digested LF as shown in Figure 2.2. MALDI-ToF analyses showed that fractions eluted with buffers containing different levels of salt were composed of peptides with differing masses (Figure 2.3). Peptides with mass of 3196 Da could be fractionated with a buffer containing 0.8 M NaCI PB along with many other peptides with lower molecular masses (Figure 2.3d). The use of a 1.0 M NaCI PB allowed the purification of the 3196 Da component together with a few other peptides (Figure 2.3e). Subsequently, eluting the cation exchange column with 2.0 M NaCI PB at pH 6.5 led to the production of a peptide mixture containing two predominant peptides with masses of 3196 and 3124 Da from the digest of food-grade LF (Figure 2.3f); these peptides were verified to be Lfcin using post source decay analysis (Figure 2.5). Lfcin was thus successfully isolated from a peptide mixture using a stepwise salt gradient on an industrial cation exchange resin. Although the digestion and isolation of Lfcin has been reported in the past, many of these purifications were performed using laboratory-grade reagents resulting in only small quantities of recovered Lfcin. For example, both Dionysius and Milne (1997) and Hoek et al. (1997) reported the isolation of Lfcin from peptic digest of bovine LF followed by cation exchange chromatography on a column of S-Sepharose Fast Flow H P L C using a preparative C18 reverse-phase column, and H P L C with an analytical C18 column. Lfcin was produced at the microgram level following the method described by Recio and Visser (1999). Affinity chromatography (Shimazaki et al., 1998) was also attempted, but again, only microgram levels of Lfcin were produced. Tomita et al. (1999) described a patented method using 3000 mL of butyl moiety-containing hydrophobic or carboxymethyl moiety-containing cation exchange chromatography medium to purify Lfcin from 600 g of peptic digested LF. Lfcin was desorbed from the hydrophobic medium using 10 mM of HCI along with a mixture of 0.1 M citric acid and 0.2 M N a 2 H P 0 4 at pH 5.0. When cation exchange chromatography was used, 1 to 4 M salt solutions were used to elute Lfcin off the medium. Following the methods described, 10.5 g of highly pure (>90%) Lfcin was produced from 600 g of LF, corresponding to approximately 40% of the theoretical yield. In the present study, the concept of competitive displacement chromatography (Li-Chan et al., 1990) was applied to isolate the highly positively charged Lfcin 45 from other cationic peptides in a two-step process. First, 5.0 g of pepsin digested LF was loaded on to 5.0 mL of weak acid cation exchanger, Purolite C-106 E P , at beyond the capacity of the cation exchange column. Because of the high pi of Lfcin, other less positively charged peptides were continuously displaced by Lfcin during the loading of the pepsin digested LF, eventually resulting in a cation exchange column loaded primarily with highly purified Lfcin. It was therefore possible to produce highly purified Lfcin in the subsequent elution step. Results from the present study showed that peptic digestion of food-grade LF followed by cation exchange chromatography on an industrial-grade resin allowed the production of highly pure Lfcin in milligrams quantity with a 35% recovery of its theoretical yield. 2.4.2. Identification of Other Cationic Bioactive Peptides LF itself is a cationic protein with antimicrobial activity and previous study has shown that the digestion of LF could enhance its antimicrobial functions (Tomita et al., 1991). Although it is commonly believed that the enhanced antimicrobial property is mainly due to the release of Lfcin during the digestion, it is possible that other functional peptides may also be released during peptic treatment. In the present study, MALDI-ToF analyses of the peptic digest of food-grade LF showed that a wide range of cationic peptides were indeed released during the digestion. Based on the molecular masses detected by MALDI-ToF and knowledge of the sequence of LF, the potential candidate sequences and positions of these cationic LF peptides were identified. For instance, the mass of 4268 Da fragment found in the 0.2M and 0.5M NaCI fractions could be one or more of the following peptides: 1 1 0 Q G R K S C H T G L G R S A G WIIPM GILRP Y L S W T E S L E P L Q G A 1 4 8 with a pi of 9.31, 6 8 G R D P Y L K R P V AAEIY G T K E S P Q T H Y Y A V A V V K K G S N F Q 1 0 5 with a pi of 9.52, or 1 0 5 Q L D Q L Q G R K S C H T G L G R S A G WIIPM GILRP Y L S W T E S L 1 4 2 with a pi of 9.31. In fact, all or some of these peptides with similar masses and pi's could have been eluted at the same time, as suggested by the broad bases on the MALDI-ToF mass spectra (Figures 2.3b and 2.3c). The broad 3196 bands appearing in the 0.5M, 0.8M, and 1.0M NaCI fractions (Figure 2.3c, 2.3d, and 2.3e, respectively) could consist of more than one peptide. As demonstrated using the proteomic tools on www.expasy.com, peptic digestion of LF could result in multiple peptides with similar mass of 3196 Da, but with different pi. Results presented in this chapter indicated that some of the peptides, with a mass of 3196 Da but lower pi than Lfcin, could be eluted by NaCI PB containing lower ionic strengths. Among the possible cationic peptides that could be produced from the peptic digestion of bovine LF, many of the putative peptides contain either arginine (R) or lysine (K) and have an overall pi above 7.0. The presence of these positively charged amino acids, R and K, is a common feature among many bioactive peptides and is responsible for their bioactive functions 46 (Lehrer and Ganz, 2002). Future research should be conducted to investigate the biological activities of these cationic peptides produced during peptic hydrolysis of food-grade LF. 2.4.3. Effects of Iron Saturation Levels of Lactoferrin Under normal physiological conditions, bovine LF contains approximately 20-25% iron (Shimazaki, 2000). The food-grade LF used in this study was consistent with the reported level and had an iron saturation level of 20% as determined by ICP-MS analysis. At this iron saturation level and at pH 2.0, Raman spectroscopic analyses showed that approximately 32% of LF folded into 6-sheet structures, 41% of LF had a-helical structures, and the rest of the molecule was present as random coil structures (Table 2.2). After some iron was removed from LF by dialyzing LF against citric acid, apo-LF (16% as determined by ICP-MS) contained less a-helical structure and random coil. Consequently, more B-sheet conformation was found in apo-LF. When LF was saturated with iron, a dramatic drop in the a-helical content with a concurrent increase in the B-sheet structure was detected. The Raman spectroscopic results showed that the presence of iron was important for the overall three dimensional structure of LF. Previous data from the literature supported results observed in the present study. The differential scanning calorimetry (DSC) thermogram of native LF revealed two denaturation temperatures at 65 °C and 92 °C (Paulsson et al., 1993). Apo-LF and holo-LF showed prominent thermal denaturation peaks at 71 °C and 93 °C, respectively, indicating that the presence or absence of iron affected the thermal stability of LF. The transition at high temperature, near 92 °C, indicated the unfolding of the iron-binding domains of LF (Paulsson et al., 1993). Furthermore, apo-LF that was 30% saturated also denatured faster than holo-LF that was 100% saturated with Fe (Sanchez et al., 1992). The binding of iron to human LF has been reported to be an important factor in the stabilization of its structure. Iron-depleted human LF demonstrated an increase in the extent of 1 H - 2 H exchange, suggesting a "loosening" of human LF upon iron removal (Hadden et al., 1994). Using differential scanning calorimetry, Mata et al. (1998) showed significant increases in maximum peak temperature, transition enthalpy, and activation energy of human milk LF which was fully saturated with iron (90.6 °C, 3209 kJ • mol"1, and 387.6 kJ • mol" 1, respectively), compared to the corresponding values for LF that was less than 15% saturated with iron (67.0 °C, 2276 kJ • mol"1, and 275.5 kJ • mol"1, respectively). Crystallographic analysis of human LF also showed that the binding of iron to LF caused conformational changes and the molecule became more compact. Iron entered into the open interdomain cleft in each lobe and then the domains closed over the iron atoms (Anderson et al., 1989). This change in lactoferrin 47 structure explained why iron-saturated LF was more resistant to denaturation and proteolysis than the apo form (Anderson, 1990). Based on results reported in this chapter, it can be speculated that removal of iron from LF decreased its overall stability and increased accessibility of pepsin to potential cleavage sites on LF. Hence, more cleavages led to the production of a larger amount of low-mass peptide fragments in the digest of apo-LF. In contrast, the addition of iron increased the (3-sheet content in holo-LF, making it more resistant to structural changes under the thermal and acidic conditions for peptic digestion. Decreased accessibility of peptic cleavage sites on holo-LF thus may have led to more larger-sized peptide fragments in the digest of holo-LF. These effects of iron saturation levels are deemed important when specific peptides are needed. For instance, when peptides of higher masses (e.g. 4065, 4275 Da) are desired, the addition of F e 3 + to the LF solution would allow a higher yield of larger peptides than untreated LF or iron-depleted solutions (Figure 2.7). Nevertheless, the level of iron saturation of LF did not affect the production and purity of Lfcin. This is probably due to the fact that the Lfcin sequence, residues 17-41 or residues 17-42, is located far away from the iron-binding domains of LF (D60-Y92-Y192-H254 and D395-Y433-Y526-H595). Hence, changes in the micro-environment of these two iron-binding sites exert little effect on the structure and stability around residues 17-42. The disulfide bond between C19 and C36 on Lfcin further enhances the stability of this peptide fragment, making it resistant to further peptic digestion regardless of the iron status in the environment. Furthermore, the 2.0M NaCI fractions recovered by cation exchange chromatography from the peptic digests of all three LF preparations (untreated-LF, apo-LF, and holo-LF) demonstrated similar peptide mass profiles. All other peptides produced during the digestion processes, regardless of the initial iron saturation levels, were eluted by buffers with lower ionic strengths, leaving only the highly cationic Lfcin on the cation exchange resin. When eluted with the 2.0 M NaCI buffer, the 26-mer and 25-mer peptides corresponding to Lfcin were the most dominant peptides, resulting in a highly pure Lfcin fraction. In summary, the production of Lfcin was not affected by the iron-status of LF. However, the removal of iron from LF increased the production of small-size peptides from LF and the addition of iron to LF increased the production of large-size peptides. An economically feasible isolation and purification strategy for the production of Lfcin from food-grade LF using an industrial-grade cation exchange resin was presented. At the time of the present study, it costs approximately US$450 for each kilogram of LF (quote obtained from DMV International Nutritional, Fraser, NY, 1999) and US$2000 for each gram of Lfcin (quote obtained from the Centre for Food Technology, Hamilton, Australia, 2001). Using the purification strategy 48 described in the present study, 14 grams of Lfcin can be produced from each kilogram of LF. 49 Figure 2.1. MALDI-ToF mass spectrum of lactoferrin digest produced by peptic digestion for 4 hours at 37 °C at pH 2.0. 50 o CO 2.500 2.000 A 1.500 1.000 0.500 0.000 160.00 140.00 120.00 100.00 0.0 100.0 200.0 300.0 400.0 500.0 600.0 700.0 Volume (mL) E o CO E o 3 •o c o O Figure 2.2. Purolite C-106 cation exchange chromatography of peptic LF digest (Solid line, Absorbance at 280nm; dotted line, conductivity). The supernatant of a 0.5 g of peptic LF digest was loaded onto a 10 mL column and eluted with 50 mM sodium phosphate buffer (pH 6.5) containing 0, 0.2, 0.5, 0.8, 1.0, and 2.0 M of NaCI at a flow rate of 1.0 mL per minute. Fraction peaks were collected and further analyzed. 51 100 ^ 80 '55 I 60. 0> jg <u ¥ 20 0 y 965 4275': 4433 500 2400 4300 6200 '8100 Mass (m/z) (a) 100 g 80 | 60 .> 4 0 a: 20 0 4430 10000 500 2400 4300 6200 8100 Mass (m/z) (c) 10000 500 2400 4300 6200 8100 10000 Mass (m/z) (b) 100 8- 80 to I 60 £ .2 40 a> * 20 0 500 2400 4300 6200 8100 10000 Mass (m/z) (d) 500 2400 4300 .6200 8100 Mass (m/z) (e) 10000 100 80 60 40 .20 ' 0 3124 3196 500 2400 4300 6200 8100. 10000 Mass (m/z) (0 Figure 2.3. MALDI-ToF mass spectra of LF digest fractions eluted from a 10 mL cation exchange column with Purolite C-106 resin using 50 mL phosphate buffer at pH 6.5 containing no NaCI (a), 0.2 M NaCI (b), 0.5 M NaCI (c), 0.8 M NaCI (d), 1.0 M NaCI (e), and 2.0 M NaCI (f). 52 (a) Mass (m/z) Figure 2.4. Purolite C-106 cation exchange chromatography in competitive displacement mode of peptic LF digest (a: Solid line, Absorbance at 280nm; dotted line, conductivity). The supernatant of a 5.0 g of peptic LF digest was loaded onto a 5.0 mL column and eluted with 50 mM sodium phosphate buffer (pH 6.5) containing 0, 1.0, and 2.0 M NaCI (a). Fraction peaks were collected and further analyzed by MALDI-ToF (b, i: fraction collected after sample loading was completed; ii: 1.0 M NaCI fraction; iii: 2.0 M NaCI fraction.) 53 100 80 S 60 c a> J : | 40 CD 20 0 3124 ! 3196 106 122 445 i J 485 488 2319 2372 Mass (Da) 3196 0 1000 2000 3000 Mass (m/z) (a) Sequences and L o c a t i o n s 17 - .FKCRR WQWRM KKLGA PSITC VRRAF -A - 42 4000 106 .05 S 122 12.2 . 03-.0-3 C C •4 44 444 25 25 GA LGA PS I PS 4-4 5 41 KKLG 485 3/1 KLGA P 488 27 A PS IT 231.9. 2319.i 22 22 CRR WQWRM QWRM KKLGA KKLGA PSITC PSITC V VRRAF A 2 37.2. 2372. 31 31 RR R WQWRM WQWRM KKLGA KKLGA PSITC PSITC VR VRR 3124. 69 FKCRR WQWRM KKLGA PSITC VRRAF (b) Figure 2.5. MALDI-ToF post source decay mass spectrum of peak with mass of 3196 Da (a) and possible peptides that were resulted from the post source decay (b). 54 CO cr CD CD > 100 80 60 =3 40 20 0 4275 I 3196 S" • f t 4430 500 2400 4300 6200 Mass (m/z) 8100 10000 (a) w c 0) CP > jo CD 100 80 60 40 20 0 965 4275 3196 4430 500 2400 4300 6200 Mass (m/z) 8100 10000 (b) Figure 2.6. MALDI-ToF mass spectra of lactoferrin digest resulting from peptic digestion of apo-LF (a) and holo-LF (b). 55 100 80 'w 60 c CD > CD a: 40 20 0 3196 3124 1972 4275 5 0 0 2400 4300 6200 Mass (m/z) 8100 10000 (a) CO c CD CD .> J O CD a: 100 80 6,0 4.0 20 0 3124 1972 3196 4275 500 24Q0 4300 6200 Mass (m/z) 81,00 10000 (b) Figure 2.7. MALDI-ToF mass spectra of 1.0 M NaCI fractions resulting from peptic digestion of apo-LF (a) and holo-LF (b). 56 03 T l V c i - i Tl <D 03 00 03 M n SB • Q) co v> co T3 CD O -1 03 o —h b 0) O 03 O o' CO CD CO c_ 5' CQ o 3 T3 CD X) Q . CQ' CD CO 5" Relative Intensity (%) Relative Intensity (%) o o G3 O oo o o o N3 O 4V o CD O CO o o o 03 CO c/> N3 o o CD K3 O O 00 O O O O O O C*3 N3 CO CD CD 03 o o o o 4^ co _ o •S o 03 CO CO 0> K3 O O CO o . o o o o o CO M 4*. co CO CD cn Table 2.1. Possible cationic peptides produced from the peptic digestion of bovine lactoferrin at 37 °C for 4 hours at pH 2.0. Cationic amino acids are indicated in bold. B u f f e r NaCI O b s e r v e d C a l c u l a t e d P u t a t i v e P e p t i d e Sequence 2 P e p t i d e p i 3 C o n c e n t r a t i o n 1 Mass (Da) Mass (Da) P o s i t i o n 0 2 M NaCI & 4268 4267 94 QGRKS CHTGL GRSAG WIIPM GILRP YLSWT ESLEP LQGA 110- 148 9 31 0 5 M NaCI 4268 4269 83 GRDPY KLRPV AAEIY GTKES PQTHY YAVAV VKKGS NFQ 68- 105 9 52 4268 4269 95 QLDQL QGRKS CHTGL GRSAG WIIPM GILRP YLSWT ESL 105- 142 9 31 4430 4427 95 EEVKA RYTRV VWCAV GPEEQ KKCQQ WSQQS GQNVT CATA 336- 374 8 02 4430 ' 4427 95 TAEEV KARYT RVVWC AVGPE EQKKC QQWSQ QSGQN VTCA 334-372 7 62 0 8 M NaCI 1971 1971 27 APRKN VRWCT ISQPE W 1-.16 9 51 1971 1973 43 WRMKK LGAPS ITCVR RA 24- 40 11 72 2675 2675 21 GAPS I TCVRR AFALE CIRAI AEKKA 30-54 9 50 3162 3162 69 APRKN VRWCT ISQPE WFKCR RWQW 1-24 10 86 3162 3162 88 KCRRW QWRMK KLGAP SITCV RRAFA L 18-33 11 84 3162 3163 64 AVAPN HAWS RSDRA AHVKQ VLLHQ QALF 590- 618 10 84 1 0 M NaCI 1053 1051 17 GKNKS RSFQ 279-287 11 17 1973 1973 43 WRMKK LGAPS ITCVR RA 24- 40 11 72 2675 2675 21 GAPS I TCVRR AFALE CIRAI AEKKA 30-54 9 50 3196 3196 90 FKCRR WQWRM KKLGA PSITC VRRAF A 17-42 11 84 3196 3197 68 NSLKD KKSCH TAVDR TAGWN IPMGL IVNQ 447-477 ' 9 20 3196 3198 63 GKNKS RSFQL FGSPP GQRDL LFKDS ALGF 279-307 10 28 4275 4276 80 SPQTH YYAVA VVKKG SNFQL DQLQG RKSCH TGLGR SAGW 97-125 9 86 2 0 M NaCI 3124 3125 82 FKCRR WQWRM KKLGA PSITC VRRAF 17-41 11 84 3196 3196 90 FKCRR WQWRM KKLGA PSITC VRRAF A 17-42 11 84 3196 3197 68 NSLKD KKSCH TAVDR TAGWN IPMGL IVNQ 447-477 9 20 3196 3198 63 GKNKS RSFQL FGSPP GQRDL LFKDS ALGF 27 9-307 10 28 1 In 50 mM phosphate buffer at pH 6.5 " 2 As determined using PeptideCutter tool available at http://ca.expasy.org/tools/peptidecutter/ (Gasteiger et al., 2005) 3 As determined using Compute pl/Mw tool available at http://ca.expasy.org/tools/pi_tool.html (Bjellqvist et al., 1993) en 00 Table 2. 2. Secondary structure of bovine lactoferrin preparations (LF) at pH 2.0 as determined using Raman spectroscopy in this study and human lactoferrin (LFH) as reported in the literature. Raman X-ray 1 FTIR 4 a-helix B-sheet coil a-helix B-sheet a-helix B-sheet coil Untreated LF 42 32 27 Apo-LF 34 47 19 Holo-LF 8 57 35 Apo-LFH 43 27 22 Holo-LFH 41 24 44 28 21 1 X-ray data from Anderson et al. (1989). 2 FTIR data from Hadden et al. (1994). 2.5. References Anderson, B. F. Apolactoferrin structure demonstrates ligand-induced conformational change in transferrins. Nature 1990, 344, 784-787. Anderson, B. F.; Baker, H. M.; Norris, G.E. ; Rice, D. W.; Baker, E. N. Structure of human lactoferrin: crystallographic structure analysis and refinement at 2.8 A resolution. J. Mol. Biol. 1989, 209, 711-734. Bellamy, W.; Takase, M.; Wakabayashi, H.; Kawase, K.; Tomita, M. Antibacterial spectrum of lactoferricin B, a potent bactericidal peptide derived from the N-terminal region of bovine lactoferrin. J. Appl. Bacteriol. 1992, 73, 472-479. Bjellqvist, B.; Hughes, G. J . ; Pasquali, Ch . ; Paquet, N.; Ravier, F.; Sanchez, J . -Ch. ; Frutiger, S. ; Hochstrasser, D.F. The focusing positions of polypeptides in immobilized pH gradients can be predicted from their amino acid sequences. Electrophoresis 1993, 14, 1023-1031. Dionysius, D. A.; Milne, J . M. Antibacterial peptides of bovine lactoferrin: purification and characterization. J. Dairy Sci. 1997, 80, 667-674. Gasteiger, E.; Hoogland, C ; Gattiker, A.; Duvaud, S.; Wilkins, M.R.; Appel, R.D.; Bairoch, A. Protein identification and analysis tools on the ExPASy server. In. The Proteomics Protocols Handbook (Ed. J . M. Walker). 2005. New Jersey, Humana Press, pp. 571-607 Gattiker, A.; Bienvenut, W.V.; Bairoch, A.; Gasteiger E. FindPept, a tool to identify unmatched masses in peptide mass fingerprinting protein identification. Proteomics 2002, 2, 1435-1444. Hadden, J . M.; Bloemendal, M.; Haris, P. I.; Srai, S. K. S.; Chapman, D. Fourier transform infrared spectroscopy and differential scanning calorimetry of transferrrin: human serum transferrin, rabbit serum transferrin and human lactoferrin. Biochim. Biophys. Acta 1994, 7205, 59-67. Hoek, K. S.; Milne, J . M.; Grieve, P. A.; Dionysius, D. A.; Smith, R. Antibacterial activity of bovine lactoferrin-derived peptides. Antimicrob. Agents Chemother. 1997, 41, 54-59. Keil, B. Specificity of Proteolysis. 1992. Berlin, Springer-Verlag. Lehrer, R. I.; Ganz, T. Cathelicidins: a family of endogenous antimicrobial peptides. Curr. Opin. Hematol. 2002, 9,18-22. Li-Chan, E.; Kwan, L.; Nakai, S. Isolation of immunoglobulins by competitive displacement of cheese whey proteins during metal chelate interaction chromatography. J. Dairy Sci. 1990, 73, 2075-2086. Marchetti, M.; Superti, F.; Ammendolia, M.; Rossi, P.; Valenti, P.; Seganti, L. Inhibition of poliovirus type 1 infection by iron-, manganese- and zinc-saturated lactoferrin. Med. Microbiol. Immunol. 1999, 787, 199-204. Mata, L.; Sanchez, D.; Headon, D. R.; Clavo, M. Thermal denaturation of human lactoferrin and its effect on the ability to bind iron. J. Agric. Food Chem. 1998, 46, 3964-3970. Moore, S. A.; Anderson, B. F.; Groom, C. R.; Maridas, M.; Baker, E. N. Three-dimensional structure of diferric bovine lactoferrin at 2.8 A resolution. J. Mol. Biol. 1997, 274, 222-236. 60 Nagasako, Y.; Saito, H.; Tamura, Y.; Chimamura, S.; Tomita, M. (1993). Iron-binding properties of bovine lactoferrin in iron-rich solution. J. Dairy Sci. 1993, 76, 1876-1881. Paulsson, M A ; Svensson, U.; Kishore, A.R. ; Naidu, S. Thermal behaviour of bovine lactoferrin in water and its relation to bacterial interaction and antibacterial activity. J. Dairy Sci. 1993, 76, 3711-3720. Przybycien, T.M.; Bailey, J .C . Structure-function relations in the inorganic salt-induced precipitation of a-chymotrypsin. Biochim. Biophys. Acta 1989, 995, 231-245. Recio, I.; Visser, S . Two ion-exchange chromatographic methods for the isolation of antibacterial peptides from lactoferrin in situ enzymatic hydrolysis on an ion-exchange membrane. J. Chromatogr. A. 1999, 831, 191-201. Saito, H.; Miyakawa, H.; Tamura, Y.; Shimamura, S.; Tomita, M. Potent bactericidal activity of bovine lactoferrin hydrolysate produced by heat treatment at acidic pH. J. Dairy Sci. 1991, 74, 3724-3730. Sanchez, L ; Peiro, J . M.; Castillo, H.; Perez, M. D.; Ena, J . M.; Calvo, M. Kinetic parameters for denaturation of bovine milk lactoferrin. J. Food Sci. 1992,57, 873-879. Shimazaki, K.-l. Lactoferrin: a marvelous protein in milk? Anim. Sci. J. 2000, 71, 329-347. Shimazaki, K.; Tazume, T.; Uji, K.; Tanake, M.; Kumura, H.; Mikawa, K.; Shimo-Oka, T. Properties of a heparin-binding peptide derived from bovine lactoferrin. J. Dairy Sci. 1998, 81, 2841-2849. Tomita, M.; Bellamy, W.; Takase, M.; Yamauchi, K.; Wakabayashi, H.; Kawase, K. Potent antibacterial peptides generated by pepsin digestion of bovine lactoferrin. J. Dairy Sci. 1991, 74, 4137-4142. Tomita, M.; Shimamura, S.; Kawase, K.; Fukuwatari, Y.; Takase, M.; Bellamy, W.; Hagiwara, T.; Matukuma, H. A process for large-scale production of antimicrobial peptide in high purity. 1999. Canada: Canadian Intellectual Property Office, C A 2070882. Williams, R.W. Estimation of protein secondary structure from the laser amide I spectrum. J. Mol. Biol. 1983, 752, 783-813. 61 CHAPTER 3. HOMOLOGY SIMILARITY ANALYSIS OF SEQUENCES OF LACTOFERRICIN AND ITS DERIVATIVES2 3.1. Introduction To understand the roles of individual amino acids on the antimicrobial activity of lactoferricin, Lfcin from different sources and synthetic Lfcin derivatives have been examined in various studies (Kang et al., 1996; Odell et al., 1996; Chappie et al., 1998; Rekdal et al., 1999; Strom et al., 2000 and 2001; Haug et al., 2001 a and b). Amino acid composition of different Lfcin derivatives was compared to their antimicrobial activity. Specific amino acids of Lfcin were replaced, deleted, or modified to understand their roles in Lfcin. For instance, Strom et al. (2000) replaced positively charged amino acids such as arginine and lysine in Lfcin with alanine, and Haug et al. (2001a) replaced tryptophan residues with alanine and unusual amino acids. Although these studies were effective in examining the roles of individual amino acids in the antimicrobial activity of Lfcin, they did not account for the importance of the overall amino acid composition and their sequential order along Lfcin. Quantitative structure-activity relationships (QSAR) are important in the study of functionality of bioactive peptides. The structure-function relationship of 15-residue murine Lfcin derivatives were examined using principal component analysis (PCA) and partial least squares (PLS) regression analysis (Strom et al., 2001). However, the importance of the position of specific amino acid residues in a peptide sequence was not fully taken into consideration in their Q S A R computation (Lejon et al., 2001). In an exhaustive Q S A R study of protein sequences, Klein et al. (1986) stressed the importance of using attributes relating to hydrophobicity, charge and their distributions as represented by frequency of occurrence, periodicity of appearance, in addition to propensity of secondary structure, for the accurate classification of protein functionality. To date, however, there has been no established method to determine the distribution of amino acid attributes along protein sequences (Hellberg et al., 1987; Siebert, 2001). In order to recognize the importance of the position of amino acid residues within a peptide sequence, a computer program, Homology Pattern Similarity (HSA), was developed to compare structural properties of amino acid sequences; such as helix propensity, bulkiness, hydrophobicity, and overall charge (Nakai et al., 2002). H S A data could then be further analyzed using Principal Component Similarity (PCS) analysis, which allows the simultaneous analysis of multiple P C s (Vodovotz et al., 1993). P C S analysis has been successfully used to classify several cationic antimicrobial peptides into different groups according to charge and helical propensity parameters of Avers ion of this chapter has been published. Nakai, S. , Chan, J . C. K., Li-Chan, E. C. Y., Dou, J . , and Ogawa, M. 2003. Homology similarity analysis of sequences of lactoferricin and its derivatives. Journal of Agricultural and Food Chemistry 51: 1215-1223. Reproduced in part with permission from the Journal of Agricultural and Food Chemistry, Copyright 2003, American Chemical Society. 62 amino acids in their sequences. LfcinB and indolicidin were classified to have low charge but high helical content, while protamine was classified as being high charge (Nakai et al., 2003). The objectives of this study were to apply Homology Similarity Analysis (HSA) and Principal Component Similiarity (PCS) to elucidate the structural mechanisms for the antimicrobial activity of lactoferricin and its derivatives. 3.2. Materials and Methods 3.2.1. Lactoferricin Sequences The sequences of lactoferricin and 70 Lfcin derivatives, as well as their antimicrobial activity expressed as minimal inhibitory concentration (MIC) against E. coli, were compiled from the literature and are listed in Tables 3.1 to 3.5. Peptide #2 is the 25-residue bovine lactoferricin peptide, while peptide #13 (Table 3.1) is the corresponding 15-residue lactoferricin derivative that was used as a reference in the H S A and P C S computation in this study. 3.2.2. Peptide Property Scales The properties of amino acid side chains used in H S A are listed in Table 3.6 and included helix propensities, charge, hydrophobicity, and bulkiness, as previously reported for optimization of site-directed mutagenesis (Nakai et al., 1998). Helix propensity was described as the free energy required to fix the different amino acids in.the a-helical region (Muhoz et al., 1994); helix propensity values ranged from 0.617 for alanine to 1.780 for proline. The isoelectric points of the amino acids were used to represent charge of side chains (Gromiha, 1993). The amino acid hydrophobicity scale was statistically derived from a large set of experimental chromatographic retention data, with values ranging from -2.24 for histidine to 3.50 for leucine (Data Set 1 in Wilce et al., 1995). Lastly, bulkiness values of amino acids ranged from 3.40 of glycine (the least bulky) to 21.61 of tryptophan (the most bulky) (Gromiha, 1993). 3.2.3. Homology Similarity Analysis (HSA) The software packages for HSA used in this study can be downloaded from ftp://ftp.landfood.ubc.ca/ foodsci/. Since it was reported that a 15-mer Lfcin derivative exhibited similar antimicrobial activity to the 25-residue native peptide (Strom et al., 2000), and the hexamer corresponding to residues 4 to 9 was reported to be the antimicrobial core of Lfcin (Tomita et al., 1994), HSA was conducted on three selected segments of the peptide sequences, corresponding to residues 1 to 15 of bovine lactoferricin, namely residues 1 to 3 (Segment I), residues 4 to 9 (Segment II), and residues 10 to 15 (Segment III) of bovine lactoferricin. In HSA, specific structural properties of designated amino acids along the sample peptide and the reference peptide (peptide #13) were plotted against each other and the procedure was 63 repeated for each amino acid pair along the peptide fragment. Linear regression on the reference was carried out and the resulting coefficient of determination (r2) was referred to as the homology similarity coefficient. Average property indices within the segments were also calculated. 3.2.4. Principal Component Similarity (PCS) Analysis The software packages for P C S used in this study can be downloaded from . ftp://ftp.landfood.ubc.ca/ foodsci/. The principle of the P C S program has been previously described (Vodovotz et al., 1993). HSA data obtained for various structural properties of different segments from all 70 peptides were analyzed. 3.2.5. Artificial Neural Networks (ANN) After P C A computation, major P C scores with eigen values above 1.0 were selected for the subsequent A N N computation using STATISTICA Neural Networks (Statsoft, 1999). Two output variables, log MIC and 1000 . M I C - 1 , were tested to reflect how MIC was measured (2-fold decrease in each sequential dilution series) and to increase the sensitivity in detecting low MIC. 3.3. Results 3.3.1. Homology Similarity Analysis of Lactoferricin Sequences Results of HSA analysis of lactoferricin and its derivatives were classified into five groups (Tables 3.1 to 3.5). Lactoferricin derivatives with high antimicrobial activity (low MIC) and having similarity coefficients for helical propensity in segment II (residues 4 to 9) that were close to 1.0 (representing high similarity to the reference sample) were classified into group I (Table 3.1). Lfcin derivatives with high antimicrobial activity, but which were highly cationic (average values of charge over 8) in segment I (residues 1 to 3), were classified into group II (Table 3.2). Group III (Table 3.3) was the group exceptional to the general rules herein adopted, consisting of peptides with unexpectedly high and low antimicrobial activities that could not be easily correlated to structural attributes. Group IV (Table 3.4) consisted of Lfcin derivatives with low antimicrobial activity and low similarity coefficients for charge in segment II. Lastly, Lfcin derivatives in group V (Table 3.5) were also low in antimicrobial activity, and showed low or negative similarity coefficients for helical propensity jn segment II. Segment II in the sequences appears to play an important role in the antimicrobial activity of Lfcin derivatives against E. coli. Lfcin derivatives with higher similarity coefficient values for helical propensity in segment II demonstrated greater antimicrobial activity. When the helical pattern similarity was low in segment II (Group V), Lfcin derivatives exhibited lower inhibitory effect against E. coli. Conversely, lactoferricin derivatives with high similarity in helical propensity in segment II (Group 64 I) exhibited higher antimicrobial effect. However, it is important to note that high similarity in helical propensity of derivatives compared to the reference peptide is not necessarily correlated with high average helical propensity values per se. In addition to the helical pattern, the pattern of charge distribution in segment II is of significance in the antimicrobial ability of Lfcin derivatives. Although peptides in group IV (Table 3.4) exhibited similar helical pattern in segment II to the reference, they exhibited lower antimicrobial activity. The low pattern similarity in charge distribution and lower average cationic values in segment II of these peptides might have been responsible for their higher MIC against E. coli. Nevertheless, when the cationic values in segment I were high (Group I), the low cationic values in segment II could be compensated and led to improved antimicrobial activity. With some exceptions, homology similarities of the charge, hydrophobicity and bulkiness patterns in segments III did not appear to be important structural parameters that would affect the antimicrobial ability of Lfcin and its derivatives. Loadings of P C scores 1-5 are shown in Table 3.7. The importance of helix structure is explicitly shown in the higher loading of similarity constants of helix values. Charge distribution at segment I (positions 1-3) is another major determinant in PC1 and P C 2 (Table 3.7). Similar results are found when the average values were used in P C A . Average helix propensity of the amino acids in segment II were the major determinants of all P C s . Average charge values in segment I were also important in the calculation of P C 2 . Loadings of P C scores demonstrate the importance of helix propensity properties of the peptide sequence in segment II and the charge of the peptide sequence in segment I. 3.3.2. Principal Component Similarity Analysis Based on the information obtained by HSA, the major structural factors that influenced the MIC values of Lfcin derivatives were similarity coefficients and average values for helical propensity in segment II, charge in segment II, and charge in segment I. Thus, these six variables were used for P C A followed by P C S analysis. The sensitivity ranks of the five input variables (PC scores 1 to 5) were 1:4:2:5:3 and 1:4:5:3:2 for log MIC and reciprocal MIC, respectively. This implies that the importance of variables in correlation is in this order. As shown in Figures 3.1 to 3.5, using P C S , peptides in groups I, II, IV and V were successfully classified. Peptides in group III were distributed into group I (peptides 33, 34 and 35) and group IV (peptides 31 and 32). These results confirm the importance of helicity and charge in the antimicrobial activity of lactoferricin. 65 3.3.3. Regression Analysis Using Artificial Neural Networks When the reciprocal of MIC (1000 • MIC" 1) was used to train the A N N , correlation coefficients between the predicted MICs and observed MICs ranged from 0.88 to 0.90. Log MICs were also used to train the A N N , yielding correlation coefficients between predicted and observed MICs ranging from 0.95 to 0.97. The predicted values vs. observed values illustrated in Figure 3.2 show correlation coefficients of 0.900 and 0.950, for log MIC and 1000 • MIC" 1, respectively. 3.4. Discussion In the pioneering work on peptide Q S A R (Hellberg et al., 1987), 29 variables were employed to characterize amino acid side chains, the dimension of which was reduced using P C A into three P C s : z1, z2, and z3. However, potentially important conformational effects of peptides were ignored. To overcome these shortcomings, Strom et al. (2001) employed two parameters of a-helices representing micelle affinity, three parameters describing a-helical propensities, two parameters related to charge, four hydrophobicity parameters, and one parameter relating to the surface property. However, the use of several similar parameters created a problem of collinearity in subsequent regression analysis. Although P C A can extract important information to be used for regression analysis for Q S A R , it is difficult to identify and isolate the true influencing factors in the antimicrobial activity of lactoferricin and its derivatives. Unlike the three-z method of Hellberg et al. (1987), the present study used only one scale for each side chain property. Selection of only one index for each structural property was made by choosing the most reliable values with respect to analytical principles. Furthermore, by comparing segments rather than entire sequences, more detail in the underlying mechanisms of functions may be disclosed since the distribution of property patterns within sequences are taken into consideration. The new approach employed in this study, namely HSA, can incorporate potential distribution effects of side chain properties within the peptide sequences. It may be useful in addressing the question posed by Lejon et al. (2001) that the position of a specific amino acid in a peptide is of importance to the structure-function behaviours of peptides. HSA is a way to improve multivariate data analysis by distinguishing positioning information between sample and reference peptides. Uniformity in group I as seen in Table 3.1 is demonstrated by the pattern similarities values all being near 1.0. The lesser importance of segment III (positions 10 to 15) appears as a departure of the similarity values from 1.0 in some derivatives in this segment, as well as weaker contributions to P C scores 1 to 5 as represented in the loadings (Table 3.7). This does not necessarily imply a less important role played by segment III in MIC. Examples of the importance of segment III in particular derivatives are the smaller charge value of peptide #33 (classified into group III), the lower similarity in hydrophobicity of peptides #36 and #39 (classified into group IV), and the lower bulkiness 66 similarity in peptides #34 and #44 (in groups III and IV, respectively). Peptide #31 (in group III) with multiple lower similarities is also another example of similar deviations. When the reciprocal MIC was assigned as the output variable in A N N computation in contrast to log MIC, better comparison of higher antimicrobial activities (low MIC) was allowed. This could be useful for searching for cationic antimicrobial peptides with greater antimicrobial activity. For instance, the low antimicrobial activity of peptide #35 (Group III) may be attributed to the simultaneous replacement of two W at positions 6 and 8 in segment II. In contrast, the corresponding single replacements made separately in peptide #23 and #24 in group I did not appreciably affect the activity. The specific function of tryptophan residues can be seen in indolicidin (Hwang and Vogel, 1998). Indolicidin activity was elucidated based on overall charge and amphipathic character of the extended-helix (Falla and Hancock, 1997); unlike lactoferricin, charge plays a more important role in indolicidin (Nakai et al., 2003). To avoid the most crucial problem of A N N , namely over-learning, the number of input variables should be restrained to the minimum. The SD ratio (prediction error SD / data SD) may be considered as an index to demonstrate the effectiveness of A N N (Statsoft). Results from the present study show no appreciable difference between the S D ratio between the training and verification sets obtained (0.26-0.33 vs. 0.38-0.46 and 0.46-0.55 vs. 0.26-0.4, for logMIC and rcpMIC, respectively). This indicates the absence of over-learning in our A N N regression computation in the present study. The most popular technique used for dimension reduction has been P C A , as shown by Strom et al. (2000). However, as discussed by Lejon et al. (2001), there is a further problem in that the selected major P C scores may not be always related to the objective function (MIC in this case). The same is true in the case of the result of P C S analysis shown in Figure 3.1, which yields classification without considering the actual relation with MIC. The use of HSA and P C S provides a rapid method to predict the structure-activity relationship of Lfcin derivatives; nevertheless, it does not illustrate how the structural properties and antimicrobial activity are related. More research using complementary analytical tools are needed to fully understand how Lfcin works as an antimicrobial agent. It is likely that unknown factors are still influential in the MIC data. Furthermore, potential errors may have arisen from using the output data of MIC values reported from different sources. At any rate, the accuracy in prediction of output variables could be evaluated. To enhance the validity of the conclusions, additional data should be included for the reciprocal MIC computation using ANN in the future. Since the literature values of MIC reported by different researchers in different laboratories were used in this study, the reliability of the observed trends is dependent on consistency of these reported MIC values. However, this problem may be common as long as large scale databases are 67 the source data for data mining. Hence, the regression results shown in this study should be regarded, as an approximation of a trend, and more detailed interpretation may have to await collection of an adequate number of reliable data through collaborative study among researchers in the future. By using both HSA and P C A in the above data mining system, the present approach could be more effective in elucidating the function mechanism as it provides more detailed contributions of major P C scores to functions as found in this study than when using P C A alone. In conclusion, the new data mining system using pattern similarity data based on properties of amino acid residues of segments in homology profiles can enhance reliability in predictability of Q S A R compared to using the properties of the whole sequences. 68 Table 3.1. Homology similarity coefficient and average property values for helix (Hx), charge (Ch), hydrophobicity (Hp) and bulkiness (Bk) calculated for three segments in the sequences of Group I lactoferricin derivatives, with minimum inhibitory concentrations (MIC) against E. coli as reported in the literature. Segment I = position 1-3, Segment II = position 4-9, and Segment III = position 10-15 in the reference (Peptide 13) with 15-amino acid sequence identical to bovine lactoferricin. Shaded zones in the sequences represent regions with high similarity to the reference peptide, while the shading of Hx II indicates the high similarity coefficient for helix propensity in Segment II for Group I derivatives. Peptide #13 was used as the reference peptide and its values were shaded for easy comparison. Peptide # Segment I Segment I Segment I I -R :. - R • I - - • W • R F '"'>•"[=" K K C - R i f f ! " R'< " W • C w R M "p s s " • R " R W Q W ' R " M " K K L G A " p -S R R wT " Q W; ' R "U K ,-K ' ' w Q W. K w K r K L~ R K K w - " . W Q C W- ' w ' R " . •R _JVl ' R ' "M K L G F • K" C " R • •R W Q W 'R " M " K K L G A - p -S 9 P E W F K C - " R " " " • R • • W c w ^ . r 1$$?^ ~M~~ K A 10 11 12 13 K C - R ' w Q W .'. RJ "W K K c A c' " R " R W.. Q W R " M " " " K K' L ~ G A R ' . W -Q W R M K-t [re G A ' K d *R 4 R w Q A R I'm k d G A1 T C V R R A F T S V R R A F T C V R R A F MIC (ug • mL 1 ) Literature 1 50 A 23 B 30 B 23 B 30 B 11 B 30 B 30 E 70 E 40 E 80 E 100 E 50 F Hxll Ch I Coefficient Average Ch I Ch Hp 1 0; 000 1; 808^ 0; 000 1.000K* 1.000 808 ** 0.808?%)'808', 0 987 1 000 0 982 1.000 1 .OOO^^§OO^.'iO00|5'-«.1. 0.824 .0.808 .0.819 >, 0.808 ^ - O ^ S ^ B f s M l t J J ^ O S l f O ; 000 808" 1.000 0.808 Coefficient Average i1.000 8.400 1.000 8.400 1.000 8.400 1.000 8.400 1.000 7.850 1.000 8.400 0.984 8.033 1.000 8.400 1.000 8.400 1.000 8.400 1.000 8.400 i Coefficient Average hi 0.000 6.000 1.000 6.767 0.982 7.067 0.000 6.000 0.000 6.000 0.000 6.000 0.000 6.000 1.000 6.767 1.000 6.767 1.000 6.767 0.994 6.933 111 Coefficient Average in 0.000 6.000 1.000 7.233 1.000 7.233 1.000 7.233 1.000 7.233 1.000 7.600 1.000 7.233 1.000 7.233 1.000 7.233 1.000 7.233 1.000 7.233 Coefficient Average in 0.000 -10.000 1.000 0.113 1.000 0.113 0.425 -1.563 0.425 -1.563 0.361 -1.307 0.425 -1.563 1.000 0.113 1.000 0.113 1.000 0.113 1.000 0.113 i i Coefficient Average 0.000 0.000 1.000 13.925 1.000 13.925 0.832 12.008 0.832 12.008 0.834 11.535 0.832 12.008 1.000 13.925 1.000 13.925 1.000 13.925 1.000 13.925 .000| l.400r ,1.000 8.400 0.000^ 1.000 6.000,# 6jM oocf* iTooq 233 7.233 l.113i____3 .ooq 1.000 ..925 13.925 1 Literature Sources: A, Tomita et al., 1994; B, Kang et al., 1996; E, Rekdal et al., 1999; and F, Strom et al., 2000. 69 Table 3.1. (cont'd). Peptide # 14 15 16 17 18 19 20 21 22 23 24 25 •• F F F F F A f j § I P I I P F Segment I k _ 3 i K K H g 9 | K K c C A C c C F R' R R R R R R R R p R R R R R R R R R Segment II w . w Q W Q W A W H i l l -.Wm 0 W * F § P P S S l i - W G W Q MBit! W W W B l l w w — w - W F W • h i R R i l l l R R R H R R R R R R "M~ M lipllil IV! M JHf M M- \" fi§§§Pi " " M i l l i K K K K-' K K*"~~ R | | § | § | i K Segment III k l i l l i i i i K L K H i * k ^SfSlft^JliliPliP A -L i l l K-A L i i l S P i l l K S p iffllltlit G G G G / ! • G G G G l i l l l l • A A A •mm* A A G A A A MIC (ug • mL"1) Literature 1 70 F 80 .F 25 F 30 F 70 F 50 F 50 F 25 F 50 G 50 H 25 H 15 Hxl l Ch I Coefficienl 1.000 1.000 Average 10^808.^0.808 Ch I Ch I Coefficient Average Coefficient Average III Coefficient Average Hp III Bk III Coefficient Average Coefficient Average 1.000 8.400 0.994 6.933 1.000 7.233 1.000 0.113 1.000 8.400 0.928 5.533 1.000 7.233 1.000 0.113 1.000 1.000 13.925 13.925 1.00C5f ;0 806 1.000 8.400 0.982 7.067 1.000 7.233 •1.000 0.113 1.000 13.925 0.840 0.783 1.000 8.400 1.000 6.767 1.000 7.233 1.000 0.113 1.000 13.925 ••Y.000 0 3GP 1.000 8.400 1.000 6.767 0.632 6.617 0.931 0.393 0.959 13.225 1.000 1.000 6.767 1.000 0.808 1.000 1.000 8.400 8.400 1.000 6.767 0.931 0.694 0.393 -0.460 1.000. 0.80b 1.000 8.400 1.000 6.767 0.632 1.000 1.000 6.617 7.233 7.233 1.000 0.098 0.959 0.763 0.882 13.225 12.342 15.275 0 998 0 ' 0.806 0 818 0 818 0.707 7.600 0.994 6.933 0.988 7.417 0.966 -0.012 998 0.998 317 8.317 000 1.000 767 6.767 000 233 1.000 7.233 000 1.000 113 0.113 0.879 1.000 1.000 11.843 13.925 13.925 1.000 0 808 1.000 8.400 0.997 6.900 1.000 7.233 1.000 0.113 1.000 13.925 1 Literature Sources: F, Strom etal., 2000; G, Strom et al., 2001; H, Hang et al., 2001a; and I, Haug etal., 2001b. 70 Table 3.2. Homology similarity coefficient and average property values for helix (Hx), charge (Ch), hydrophobicity (Hp) and bulkiness (Bk) calculated for three segments in the sequences of Group II lactoferricin derivatives, with minimum inhibitory concentrations (MIC) against E. coli as reported in the literature. Segment I = position 1-3, Segment II = position 4-9, and Segment III = position 10-15 in the reference (Peptide 13) with 15-amino acid sequence identical to bovine lactoferricin. Shaded zones in the sequences represent regions with high similarity to the reference peptide, while the shading of Ch I indicates the high average values for charge in Segment I characteristic of Group II derivatives. Peptide # 26 27 28 29 30 K R R R R Segment ". "• K ' . C. K C K C K C ' ' ,R" L L L L Segment I :*'" ,' R-> . , , >, - i W , - . . ' Q *", ; • w R • w- • ; R*' r W , . Q , - w . w Q W R W W, '"' • ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ V. , ' R f , ' 'R' E A!. R„ \ A M *' - M. **?* ~ »M* M M R R R R Segment II k • , ' * „ « ' ' ' " . " K • Ll V Y Y Y - C- 1 G A G G G G MIC (ug • mL"1) 60 20 40 25 20 Literature 1 E G G G G Hx II Coefficient 1.000 0.998 0.955 0.892 0.998 Average 0.808 0.806 0.793 0.783 0.806 Ch II Coefficient 1.000 0.707 0.149 0.447 0.707 Average 8.400 7.600 6.333 6.800 7.600 Ch I Coefficient i>*" 0.566 ' 0 407'" ~0~407~ 0 407 ""0.407" Average - 8.167 8.533 - 8.533": 8 533 ," .'8.533 Ch III Coefficient 1.000 0.988 0.988 0.988 0.988 Average 7.233 7.417 7.417 7.417 7.417 Hp III Coefficient 1.000 0.950 0.966 0.966 0.966 Average 0.113 -0.062 -0.012 -0.012 -0.012 Bk III Coefficient 1.000 0.900 0.879 0.879 0.879 Average 13.925 12.422 11.843 11.843 11.843 1 Literature Sources: E, Rekdal et al., 1999; and G, Strom et al., 2001. 71 Table 3.3. Homology similarity coefficient and average property values for helix (Hx), charge (Ch), hydrophobicity (Hp) and bulkiness (Bk) calculated for three segments in the sequences of Group III lactoferricin derivatives, with minimum inhibitory concentrations (MIC) against E. coli as reported in the literature. Segment I = position 1-3, Segment II = position 4-9, and Segment III = position 10-15 in the reference (Peptide 13) with 15-amino acid sequence identical to bovine lactoferricin. Shaded zones in the sequences represent regions with high similarity to the reference peptide. Group III derivatives included exceptional cases with unexplained high or low MIC values. Peptide # 31 32 33 34 35 Segment I Segment II Segment III' L' • • F- n k {• K- K-CT" c C C R F, R R R R R R ; R W W F f Q Q 4 ; Q R w W *'W, F ' R' " R [ M M M A* u K w f l f l l l l i l s i K ~K ' k  K . ; J ,yk d G f l l i g 1 1 ? 111111 G A A A A MIC(ug-mL" 1) 23 20 200 140 300 Literature1 B E E F H Hxl l Ch Ch Hp Coefficient Average II 0.213 0.808 0.512 0.844 1.000 0.808 1.000 0.808 1.000 0.828 II Coefficient Average i 0.333 8.400 0.651 7.517' 1.000 8.400 1.000 8.400 0.998 8.233 i Coefficient Average in 0.000 6.000 1.000 6.767 0.566 5.700 1.000 6.767 1.000 6.767 Coefficient Average HI 1.000 7.233 1.000 7.233 1.000 7.233 1.000 7.233 1.000 7.233 in Coefficient Average III 0.425 -1.563 1.000 0.113 1.000' 0.113 0.999 0.088 1.000 0.113 in . Coefficient Average 0.832 12.008 1.000 13.925 1.000 13.925 0.947 13.133 1.000 13.925 1 Literature Sources: E, Rekdal etal., 1999; F, Strom et al., 2000; and H, Haug etal., 2001a. 72 Table 3.4. Homology similarity coefficient and average propensity values for helix (Hx), charge (Ch), hydrophobicity (Hp) and bulkiness (Bk) calculated for three segments in the sequences of Group IV lactoferricin derivatives, with minimum inhibitory concentrations (MIC) against E. coli as reported in the literature. Segment I = position 1-3, Segment II = position 4-9, and Segment III = position 10-15 in the reference (Peptide 13) with 15-amino acid sequence identical to bovine lactoferricin. Shaded zones in the sequences represent regions with high similarity to the reference peptide, while the shading of Ch II indicates low similarity coefficient and average values of charge in Segment II of Group IV derivatives. Peptide # 36 37 38 39 40 41 42 43 44 45 46 Segment I Segment II Segment III S K C " R " Q W a s" . A V I p) F I K C v. K C K A«t» R >WT Q , ,W'> R K • K L G A Q K K L" G A ~-R 1 R W •. i s p l l l j S a W''" *> A E , * M K< R k .', k u* V G " G A1 G s K C R " Q" Ws Q W R R^ t " n : p E c " L R Wf Q, WJ E " M v R_ K. _V G A K C "1 " e __R ~-"k v W G MIC (ug • mL"1) Literature 1 120 B 200 E 1000 F 1000 F 70 F 120 F 55 F 1000 F 440 F 1000 G 800 G Hxl Ch Ch Hp Bk Coefficient Average II 0.971 0.769 0.910 0.823 0.910 0.823 0.963 0.828 0.883 0.786 0.883 0.786 0.883 0.786 0.955 0.793 0.988 0.816 0.955 0.793 0.955 0.793 Coefficient Average -1.000 4.600 0.191 6.233 0.191 6 233 0.698 i 7 417 "0 707 7 600 0.707 4 7 600 0 707 , 7 600 6 333 0 698 7.417 0 149 6.333 "6TT4§ 6:333 I Coefficient Average HI 0.000 6.000 0.933 6.000 0.933 6.000 0.994 6.933 1.000 6.767 1.000 6.767 1.000 6.767 0.933 6.000 0.994 6.933 0.933 6.000 0.994 6.933 in Coefficient Average HI -1.000 5.067 0.988 7.417 0.988 7.417 0.996 7.500 1.000 7.233 1.000 7.233 1.000 7.233 0.988 7.417 0.996 7.500 0.988 7.417 0.988 7.417 in Coefficient Average in 0.297 -1.057 0.950 -0.062 0.950 -0.062 0.409 0.485 1.000 0.113 1.000 0.113 1.000 0.113 0.950 -0.062 0.409 0.485 0.950 -0.062 0.950 -0.062 Coefficient Average 0.833 11.298 0.900 12.422 0.900 12.422 0.381 16.002 1.000 13.925 1.000 13.925 1.000 13.925 0.900 12.422 0.381 16.002 0.900 12.422 0.900 12.422 Literature Sources: B, Kang etal., 1996; E, Rekdal etal., 1999; F, Strom etal., 2000; and G, Strom etal., 2001. 73 Table 3.4. (cont'd). Peptide # 47 48 49 50 51 52 53 54 55 56 E A R E A R A E A E Segment c " k C - K C 1 l l l l l K p i l i i l i t t i i K f C i i l l l i I f S p l I K L L L L L L L L L L Segment 1 R W Q W I K l i i •W w • R W . G W . R W Q . W' •-I W w R W G i i i i § R W' Q * < ,W R , ' w Q R W Q W R i < w: A A B8§|| R *R A E A A l R M T M U 7 M R _ r R R R R R R R R Segment III K V S l l l l V M U M K "v" V K V K ~ Y K Y " • I l l i Y K Y W3BEL G G K I S G T ^ * G G G 181113*' G G G G G G G G G G G MIC (ug • mL"1) Literature 1 1000 H 270 H 75 H 200 H 62 H 62 H 300 H 1000 H 175 H 125 H Hx II Coefficient Average Ch II 0.892 0.783 0.892 0.783 0.955 0.793 0.998 0.806 0.998 0.806 0.892 0.783 0.955 0.793 0.892 0.783 0.892 0.783 0.998 0.806 Coefficient Average 0 447 6 800 o- • (5 800 d: i49 3 333 0.707. 7.600, .0.707 7:600 " 0.447 ~ 6 800 , . 0^49% 6.333' 0.447 6.800" • 0 447 6.800 0.707^ 7.600 Ch I Coefficient Average Ch III Coefficient Average Hp III Coefficient Average Bk III Coefficient Average 0.933 6.000 0.988 7.417 0.994 6.933 0.988 7.417 0:407 8.533 0.988 7.417 0.933 6.000 0.988 7.417 0.994 6.933 0.988 7.417 0.407 8.533 0.988 7.417 0.994 6.933 0.988 7.417 0.933 6.000 0.988 7.417 0.994 6.933 0.988 7.417 0.933 6.000 0.988 7.417 0.950 -0.062 0.950 -0.062 0.950 -0.062 0.950 -0.062 0.950 -0.062 0.950 -0.062 0.966 -0.012 0.966 -0.012 0.966 -0.012 0.966 -0.012 0.900 12.422 0.900 12.422 0.900 12.422 0.900 12.422 0.900 12.422 0.900 12.422 0.879 11.843 0.879 11.843 0.879 11.843 0.879 11.843 1 Literature Source: H, Haug et al., 2001a. 74 Table 3.5. Homology similarity coefficient and average property values for helix (Hx), charge (Ch), hydrophobicity (Hp) and bulkiness (Bk) calculated for three segments in the sequences of Group V lactoferricin derivatives, with minimum inhibitory concentrations (MIC) against E. coli as reported in the literature. Segment I = position 1-3, Segment II = position 4-9, and Segment III = position 10-15 in the reference (Peptide 13) with 15-amino acid sequence identical to bovine lactoferricin. Shaded zones in the sequences represent regions with high similarity to the reference peptide, while the shading of Hxll indicates low similarity coefficient values for helix propensity in Segment II of Group V derivatives. Peptide # 57 58 59 60 61 62 63 S K S . 5jF T S Segment I ^9liii8i&8tii K K K C i i i i s * C C c c" Y F Y **R R F Y Q "~**"R Q >R R Q Q Segment II W ' A ! 1 ! J S l l l " *:w A l l -w W . . W Q i l l l l i l Q Q Q , Q R w Rip A l l i i i V l j VV R R S * J i l J l R ml. R M "*"M -M-», M, . i i s t l l l M R K R K ~ R . R Segment III K I B l i l l K** | p ^ : i f i l | L : - C iJ is • l p t § 1 G G S P l l I i ! G R G 1 M M ^ i A A G A p s 1 T C V R R T S MIC (ug • mL"1) 750 . 200 500 200 200 150 350 Literature 1 E E F F F F F Hxl l Coefficient 0 007 0.512 0 007 0 01C 0.393, Average 0 82S Wimm 0 903 , 0.854 Ch II . Coefficient 0.000 0.651 0.000 1.000 1.000 -0.703 0.447 Average 7.600 7.517 7.600 8.400 8.400 5.817 6.800 Ch I Coefficient 0.994 0.566 0.994 1.000 1.000 0.994 0.994 Average 6.933 8.167 6.933 6.767 6.767 6.933 6.933 Chi l l Coefficient 0.988 1.000 0.988 1.000 1.000 0.640 0.988 Average 7.417 7.233 7.417 7.233 7.233 8.217 7.417 Hp III Coefficient 0.987 ' 1.000 0.987 1.000 1.000 0.915 0.987 Average , 0.242 0.113 0.242 0.113 0.113 -0.228 0.242 Bk III Coefficient 0.995 1.000 0.995 1.000 1.000 0.477 0.995 Average 13.688 13.925 13.688 13.925 13.925 14.235 13.688 1 Literature Sources: E, Rekdal et al., 1999; and F, Strom et al., 2000. 75 Table 3.5. (cont'd). Peptide* 64 65 66 67 68 69 70 71 Segment I L . i . . . c .._ F F F F F F F Q Q Q Q Q Q Q Segment II W W W W A'. C Q . Q Q - ,Q Q d A R R R R R R R i — - n f i N N N N N N N A' *M * , M,,. '* Mi P> < V ; , . M '.<;<' M S t e w R R R R R R R Segment III l l i l § t K l K K K * T K K K . k A V V V V V V V R R R R R R R G G G G P P P P P P V V V S S S C 1 I K R D S MIC (ug • mL 1) 120 1000 1000 1000 580 2000 200 1000 Literature 1 B C C D D D E F Hx II Coefficient -0.764 -0 297 -0 297-:? -0.297 -0.297. ;-0:297 "•-0.297 ' -0.297, Average 0 685 0 877 0 877?; 0.877 0.877 0.877 0.877, Ch II Coefficient 1.000 -0.521 -0.521 -0.521 -0.521 -0.521 -0.521 -0.521 Average 8.400 6.617 6.617 6.617 6.617 6.617 6.617 6.617 Ch I Coefficient 0.000 0.000 0.000 0.000 0.000 0.000 0.994 0.994 Average 6.000 6.000 6.000 6.000 6.000 6.000 6.933 6.933 Ch III Coefficient 1.000 0.640 0.640 0.640 0.640 0.640 0.640 0.640 Average 7.233 8.217 8.217 8.217 8.217 8.217 8.217 8.217 Hp III Coefficient 0.173 0.915 0.265 0.915 0.265 0.263 0.915 0.915 Average -2.162 -0.228 -1.920 -0.228 -1.920 -1.837 -0.228 -0.228 Bk III Coefficient 0.658 0.477 0.431 0.477 0.431 0.438 0.477 0.477 Average 9.633 14.235 13.668 14.235 13.668 13.865 14.235 14.235 1 Literature Sources: B, Kang et al., 1996; C, et al., 2000. Odell et al., 1996; D, Chappie et al., 1998; E, Rekdal et al., 1999; and F, Strom 76 Table 3.6. Amino acid structural propensity values used for Homology Similarity Analysis (HSA). Helix Charge Hydrophobicity Bulkiness Ala 0.617 6.0 0.06 11.50 Arg 0.753 10.8 -0.85 14.28 Asn 1.089 5.4 0.25 12.82 Asp 0.932 2.8 -0.20 11.68 Cys 1.107 5.1 0.49 13.46 Gin 0.770 6.0 0.31 14.45 Glu 0.675 3.2 -0.10 13.57 Gly 1.361 6.0 0.15 3.40 His 1.034 7.6 -2.24 13.67 He 0.876 6.0 3.00 21.40 Leu 0.740 6.0 3.50 21.00 Lys 0.784 9.7 -1.62 15.70 Met 0.730 6.0 0.21 16.25 Phe 0.968 5.5 4.80 19.80 Pro 1.780 6.0 0.71 17.43 Ser 0.980 6.0 -0.62 9.47 Thr 1.053 6.0 0.65 15.80 Trp 0.910 6.0 2.29 21.61 Tyr 1.009 6.0 1.89 18.03 Val 0.940 6.0 1.59 21.50 Literature 1 A B C D 1 Literature Sources: A, Munoz et al., 1994; B, Isoelectric Point; C, Wilce et al., 1995; and D, Gromiha and Ponnuswamy, 1993. 77 Table 3.7. Loadings of input variables, homology similarity constant (SimCnst, determined using homology similarity analysis, HSA) and Average (Ave) of Helical Propensity in position 4 to 9, charge in position 4 to 9, charge in position 1 to 3, and charge in position 10 to 15, determined by Principal Component Analysis (PCA). Helix 4-9 Charge 4-9 Charge 1-3 Charge 10-15 SimCnst Ave SimCnst Ave SimCnst Ave SimCnst Ave PC1 -0.55 0.51 -0.24 -0.01 0.35 0.15 0.47 0.09 P C 2 0.42 -0.62 0.19 0.06 0.47 0.34 0.06 -0.21 P C 3 -0.66 0.60 0.42 -0.02 0.04 0.10 -0.01 0.05 PC4 -0.49 0.44 0.64 0.06 0.03 0.25 -0.26 -0.08 P C 5 -0.67 -0.20 -0.14 0.36 0.39 -0.27 -0.27 0.24 78 Coefficient at Determination Figure 3.1. Principal component similiarity (PCS) scattergram of lactoferricin derivatives using data obtained from homology similiarity analysis (HSA). Group l-V are listed in Tables 3.1 - 3.5. 79 3.50 3.00 g 2.50 _> CO > o 2.00 o "O CD 1.50 =5 1.00 _ Q. 0.50 0.00 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 (a) Experimental Log MIC Values 70.00 CO CD __ CD > o CO o o o CD ^ 20.00 CD | 10.00 20 40 60 80 (b) Experimental Reciprocal MIC Values 100 Figure 3.2. Plots of predicted values vs. observed values by artifical neural network: (a) Log MIC and (b) Reciprocal MIC. 80 3.5. References Chappie, D. S. ; Mason, D. J . ; Joannou, C. L ; Odell, E. W.; Gant, V.; Evans, R. W. Structure-function relationship of antibacterial synthetic peptides homologous to a helical surface region on human lactoferrin against Escherichia coli serotype 0111 . Infect. Immun. 1998, 66, 2434-2440. Falla, R. J . ; Hancock, R. E. W. Improved activity of a synthetic indolicidin analog. Antimicrob. Agents Chemother. 1997, 41, 771-775. Gromiha, M. M.; Ponnuswamy, P. K. Relationship between amino acid properties and protein compressibility. J. Theor. Biol. 1993, 165, 87-100. Haug, B. E.; Svedsen, J . S . The role of tryptophan in the antibacterial activity of a 15-residue bovine lactoferricin peptide. J. Pept. Sci. 2001a, 7, 190-196. Haug, B.E.; Skar, M. L ; Svendsen, J .S . Bulky aromatic amino acids increase the antibacterial activity of 15-residue bovine lactoferricin derivatives. J. Pept. Sci. 2001b, 7, 425-432. Hellberg, S. ; Sjostrom, M.; Skagerberg, B.; Wold, S. Peptide quantitative structure-activity relationships, a multivariate approach. J. Med. Chem. 1987, 30, 1126-1135. Hwang, P. M.; Vogel, H. J . Structure-function relationships of antimicrobial peptides. Biochem. Cell Biol. 1998, 76, 235-246. Kang, J . H.; Lee, M. K.; Kim, K. L.; Hahm, K.-S. Structure-biological activity relationships of 11-residue highly basic peptide segment of bovine lactoferrin. Int. J. Pept. Protein Res. 1996, 48, 357-363. Klein P.; Jacquez, I. A.; Delis, C. Prediction of protein function by discriminant analysis. Math. Biosci. 1986, 81, 177-189. Lejon, T.; Strom, M. B.; Svendsen, J . S. Is information about peptide sequence necessary in multivariate analysis? Chemometr. Intel. Lab. Syst. 2001,57, 93-95. Munoz, V.; Serrano, L. Intrinsic secondary structure propensities of the amino acids, using statistical </>-<p matrices: Comparison with experimental scales. Proteins 1994 ,20,301-311. Nakai, S. ; Nakamura, S.; Seaman, C. H. Optimization of site-directed mutagenesis. 2. Application of random-centroid optimization to one-site mutation of B. stearothermophilus neutral protease for improving thermostability. J. Agric. Food Chem. 1998, 46, 1655-1661. Nakai, S. ; Amantea, G. ; Nakai, H.; Ogawa, M.; Knanagawa, S. Definition of outliers using unsupervised principal component similarity analysis for sensory evaluation of foods. Int. J. Food Prop. 2002, 5, 289-306. Nakai, S. ; Ogawa, M.; Nakamura, S. ; Dou, J . ; Funane, K. A computer-aided strategy for structure-function relation study of food proteins using unsupervised data mining. Int. J. Food Prop. 2003, 6, 25-47. Odell, E. W.; Sarra, R.; Foxworthy, M.; Chappie, D. S. ; Evans, R. W. Antibacterial activity of peptides homologous to a loop region in human lactoferrin. FEBS Lett. 1996, 382, 175-178. 81 Rekdal, 0 . ; Andersen, J . ; Vorland, L. H.; Svendsen, J . S. Construction and synthesis of lactoferricin derivatives with enhanced antibacterial activity. J. Pept. Sci. 1999, 5, 32-45. Siebert, K. Quantitative structure-activity relationship modeling of peptide and protein behaviour as a function of amino acid composition. J. Agric. Food Chem. 2001, 49, 851-858. Statsoft, Addendum for Version 4, In Statistica Neural Networks, 1999. Tulsa, Statsoft. www.statsoft.com. Strom, M. B.; Rekdal, 0 . ; Svendsen, J . S . Antibacterial activity of 15-residue lactoferricin derivatives. J. Pept. Res. 2000, 56, 265-274. Strom, M. B.; Rekdal, 0 . ; Stensen, W.; Svendsen, J . S . Increased antibacterial activity of 15-residue murine lactoferricin derivatives. J. Peptide Res. 2001, 57,127-139. Tomita, M.; Takase, M.; Bellamy, W.; Shimamura, S. A review: The active peptide of lactoferrin. Acta Paedeatr. Jpn. 1994 ,36,585-591. Vodovotz, Y. ; Arteaga, G. E.; Nakai, S . Principal component similarity analysis for classification and its application to G C data of mango. Food Res.lnt. 1993, 26, 355-363. '• Wilce, M. C. J . ; Aguilar, M.-l.; Heam, M. T. Physicochemical basis of amino acid hydrophobicity scales: Evaluation of four new scales of amino acid hydrophobicity coefficients derived from R P - H P L C of peptides. Anal. Chem. 1995 ,67,1210-1219. 82 CHAPTER 4. RAMAN SPECTROSCOPIC STUDY OF THE EFFECTS OF LACTOFERRICIN ON THERMAL TRANSITIONS OF PHOSPHOLIPIDS IN LIPOSOMES 3 4.1. Introduction Cationic antimicrobial peptides (CAPs) play a critical role in the host defense system of many higher organisms such as plants, insects, and mammals, and are being considered as a potential source of new antibiotics on the basis of their fast killing action, selective toxicity and low potential for resistance. Studies on C A P s have shown that the presence of cationic and hydrophobic amino acids was responsible for their interactions with the bacterial membrane (Epand and Vogel, 1999), leading to the disruption of the bacterial membrane and possible action on intracellular targets, such as DNA and RNA (Kanyshkova et al., 1999). There is evidence that C A P s show specificity for particular membrane lipid components. Herbig et al. (2005) detected strong interactions of the peptide with negatively charged phospholipids. The peptide cinnamycin caused an increase in permeability and aggregation of liposomes containing phosphatidylethanolamine (PE), but liposomes containing phosphatidylserine (PS), phosphatidylinositol (PI) or cardiolipin (CL) were not appreciably reactive with cinnamycin (Choung et al., 1988). Magainins also demonstrated preferential interaction with negatively charged phospholipids, effectively permeabilizing phosphatidylglycerol (PG)-rich membranes and killing bacteria whose inner membranes contain higher amounts of P G (Matsuzaki et al., 1997). Similar results were also demonstrated when LL-37, an antimicrobial peptide found in humans, was readily incorporated into 1,2-dipalmitoyl-sn-glycerol-3-[phospho-rac-(1-glycerol)] (DPPG) monolayers (Neville eta l . , 2006). Phospholipid liposomes are widely used as a model system for biomembranes. Studies on 1,2-dipalmitoyl-sn-glycerol-3-[phospho-rac-(1-choline)] (DPPC) liposomes are commonly found in the literature because phosphatidyl choline (PC) or lecithins are the major component of most mammalian biomembranes (Opekarova and Tanner, 2003.) Phospholipid compositions of individual genera and species of Gram-negative and Gram-positive bacteria are very different from those of mammals (Yeagle, 1993). For instance, P C is often absent from Gram-positive bacteria. In contrast, P G and its derivatives are the only phospholipids found in the membrane of Staphylococcus aureus and P E is the dominant component of the membrane of Escherichia coli, a Gram-negative species (Opekarova and Tanner, 2003). Studies of the interaction of antimicrobial peptides with model membranes can provide some important insights into the antimicrobial mechanisms and how the peptides may affect the structure of 3 A version of this chapter will be submitted for publication. Chan, J . C. K., and Li-Chan, E. C. Y. Raman spectroscopic study of the effects of lactoferricin on thermal transitions of phospholipids in liposomes. Biochimica et Biophysica Acta (BBA) - Biomembranes. 83 the membranes. Even if the site of action of an antimicrobial peptide is intracellular, the peptide first has to pass through the bacterial membrane, and this process is not understood at the molecular level at present. To date, there is a paucity of literature on the interaction of Lfcin with different phospholipids. An understanding of the molecular interactions between Lfcin and different phospholipids would shed light on the antimicrobial mechanisms of Lfcin. It could also facilitate the design of Lfcin derivatives or other C A P s that exhibit effective antimicrobial action without the undesirable interactions with mammalian membranes that can lead to hemolysis (Kang et al., 1996). Raman spectroscopy can be useful to study the structural arrangement of phospholipids. The Raman spectra of phospholipids are strongly modified by environmental factors such as temperature and interactions with other substances in the bilayer structure. Sharp changes are observed in correspondence to endothermic gel and liquid crystal transition. Remarkable agreement between observations made using the Raman technique and those obtained by calorimetric measurements or theoretical calculations has been reported (Gaber and Peticolas, 1977). Raman spectroscopy has been used to study the interactions of model phospholipid membrane and bioactive peptides (Susi et al., 1979; Lafleuret al., 1989; Bouchard and Auger, 1993). In a series of studies on pulmonary surfactants, Vincent et al. (1991 and 1993) used Raman spectroscopy to measure the carbon-hydrogen stretching mode region (2800 to 3100 cm"1) of D P P C and D P P G liposomes. Peak heights of Raman peaks near 2850, 2880, and 2935 cm" 1 at various temperatures were evaluated to assess the effects of human surfactant peptides on the thermotropic behaviour on these liposomes. In this study, we use Raman spectroscopy in the C-H stretching region over the temperature range from - 1 0 °C to 70 °C to investigate the effects of Lfcin on liposomes composed of P G , P C , and P E as a model bacterial cell membrane system. 4.2. Materials and Methods 4.2.1. Materials Sodium salts of 1,2-dipalmitoyl-sn-glycerol-3-phosphocholine (DPPC), 1-palmitoyl-2-oleoyl-sn-glycerol-3-phosphocholine (POPC) 1,2-dipalmitoyl-s/i-glycerol-3-[phospho-rac-(1-glycerol)] (DPPG), 1-palmitoyl-2-oleoyl-sn-glycerol-3-[phospho-rac-(1-glycerol)] (POPG) , 1,2-dipalmitoyl-sn-glycerol-3-phosphoethanolamine (DPPE), and 1-palmitoyl-2-oleoyl-sn-glycerol-3-phosphoethanolamine (POPE) were purchased from Avanti Polar Lipids (Alabaster, AL) in lyophilized powder forms. Purified bovine lactoferricin (Lfcin) was purchased from the Centre for Food Technology (Hamilton, Australia). MALDI-ToF mass spectrometry confirmed that the Lfcin preparation had a mass of 3124 Daltons with purity over 90%. 84 4.2.2. Preparation of Liposomes The. method for preparation of multilamellar liposomes was adapted from methods previously reported in the literature examining interactions between peptides and different phospholipid systems (Laroche et al., 1988; Devlin and Levin, 1989; Laroche et al., 1990). Phospholipids were mixed with buffer (0.15M Na Tris-acetate buffer, pH 7.2) or buffer containing Lfcin at a molar ratio of lipid to peptide of 60:1. The mixture was rehydrated in pre-warmed buffer for 1 hour at 72 °C. The solution was then agitated with a vortex mixer, heated for 10 minutes, agitated again, and cooled down below the gel to liquid-crystalline phase transition temperature by incubating at -18 °C for 10 minutes. This cycle was repeated at least 3 times to ensure uniform packing of different racemates of the phospholipids in the bilayer. The samples were vortexed again, transferred to glass capillary tubes and concentrated in a hematocrit centrifuge for 5 minutes, then aged overnight at 4 °C to improve the homogeneity of the size distribution (Avanti Polar Lipid). Incorporation of Lfcin into the liposomes was inferred by the absence of Lfcin in the supernatant, as confirmed by protein assay using bicinchoninic acid. 4.2.3. Raman Spectroscopy Raman spectra of each sample were measured on a J A S C O NR-1100 Raman spectrometer ( J A S C O Inc., Tokyo, Japan) with excitation at 488 nm from an argon ion laser (Coherent Innova 70C series, Coherent Laser Group, Santa Clara, CA). Wavenumber was calibrated everyday using the 1050 cm" 1 band of a standard potassium nitrate solution. Capillary tubes containing the samples were held inside the constant temperature measuring accessory (model RT-1C, J A S C O Inc.) and equilibrated for at least 5 minutes at the desired temperature prior to spectral data collection. Spectra were collected in order of ascending temperature, with spectral data from at least 6 scans averaged for each sample at each temperature. 4.2.4. Raman Spectral Analysis OMNIC Version 6.0a (Thermo Nicolet Corporation) and G R A M S / 3 2 Version 4.14 (Spectral Notebase, Galactic Industries Corporation) were used to process the spectral data obtained in the present study. The spectra were baseline corrected (2-point system with anchors near the ends of each spectrum), smoothed (Savitsky-Golay function with a second order polynomial function with 5 points), and offset (minimum point at zero). Peak positions, heights, areas, widths at half height of peaks near 2850, 2880, and 2935 cm" 1 were recorded and compared. Curve fitting of the three peaks near 2850, 2880, and 2935 cm" 1 was performed. Properties of the fitted peaks were also measured. Typical coefficients of variation for replicate spectra of selected samples were less than 10%. The statistical method described by Badii and Howell (2003) was used to detect differences between treatments at the 5% significance level. 85 4.3. Results The C-H stretching region of D P P C liposomes in the presence of Lfcin at various representative temperatures from - 10 °C to 60 °C are shown in Figure 4.1. A gradual broadening of the peak near 2880 cm" 1 was observed with increasing temperature of the system. At temperatures below 40 °C, the peak near 2880 cm" 1 remained very distinct and had a higher intensity than the peak near 2850 cm" 1 . At temperatures above 40 °C, the peak near 2880 cm" 1 widened, decreased in intensity, and gradually shifted to higher wavenumber. The peak near 2935 cm" 1 also increased in relative intensity as temperature rose. Similar temperature dependent changes of the spectra were observed in the absence of Lfcin. A more detailed analysis of the effects of the temperature dependence of the acyl chain packing characteristics and phase transition cooperativities on the various Raman spectroscopic features can be achieved through analyzing physical properties of peaks near 2850, 2880, and 2935 cm" 1 . Preliminary studies showed that ratios of peak-height, peak-area, and peak-width at mid-height illustrated similar trends. Hence, only peak-height intensity ratios of peaks near 2850, 2880, and 2935 cm" 1 , commonly used parameters in the literature (Snyder et al., 1980 and 1982; Devlin and Levin, 1989; Vincent et al., 1991 and 1993), are shown in Figures 4.2, 4.3, and 4.4 for D P P C & P O P C , D P P G & P O P G , and D P P E & P O P E , respectively. Statistical differences between peak-height intensity ratios were confirmed by a statistical method described by Badii and Howell (2003) at 5% significance level. The intensity ratio bsso/bsso of D P P C liposomes gradually increased over the temperature range tested, with a broad transition between 20 °C and 60 °C (Figure 4.2a). The presence of Lfcin in the liposomes resulted in lower bsso/bsso ratios throughout the temperature range while exhibiting generally similar temperature dependent behaviour as the spectra without Lfcin. The presence of Lfcin increased the temperature at which the transition was observed; in other words, an increase in the intensity ratio was primarily between 20 and 40, and at 50 °C for D P P C with Lfcin, whereas D P P C without Lfcin showed an increase in the ratio starting already at 5 °C and more pronounced at 10, 15 °C, and had almost reached a plateau by 20-30 °C. The I2935/I2880 ratio depicted a sharper change as function of temperature increases in the terminal - C H 3 group than the bsso/bsso ratio (Figure 4.2b). In the absence of Lfcin, a sharp change in I2935/I2880 ratio of D P P C liposomes occurred at approximately 38 °C. A minor increase in the I2935/I2880 ratio was also detected at above 0 °C. The presence of Lfcin in D P P C liposomes resulted in a general lowering in I2935/I2880 ratio over the temperature range tested and broadening of the transition stage. Addition of the Lfcin to D P P C liposomes caused the I2935/I2880 ratio to increase from around 20 °C to 50 °C and delay the transition temperature to 40 °C. Furthermore, the presence of Lfcin in D P P C liposomes also led to the disappearance of the transition near 0 °C. 86 P O P C liposomes experienced a very mild change over the temperature range tested as indicated by the I2850/I2880 ratio (Figure 4.2c). A gentle upward shift in the bsso/bsso ratio was observed as temperature increased. Nevertheless, a transition near 0 °C was exhibited by the I2935/I2880 ratio (Figure 4.2d). The presence of Lfcin had little effect on the thermal behaviour of P O P C . The I2850/I2880 ratio of D P P G exhibited a gradual increase as temperature increased (Figure 4.3a). The ratio increased steadily at temperature below 25 °C and increased more rapidly between 25 and 40 °C, and then reached a plateau above 40 °C. The addition of Lfcin to the D P P G system had little effect as indicated by the I2850/I2880 ratio. Sharper changes of the temperature dependent behaviour of D P P G were observed using the I2935/I2880 ratio (Figure 4.3b). In the absence of Lfcin, D P P G liposomes demonstrated an increase in the I2935/I2880 ratio from approximately 0.6 to 1.0 between 30 and 40 °C. The addition of Lfcin to the D P P G system also led to an increase of the I2935/I2880 ratio starting at around 30 °C. However, the I2935/I2880 ratio only increased to approximately 0.8 at near 40 °C in the presence of Lfcin. Changes in temperature had no marked effect on the C-H stretching intensity ratios of P O P G in either the absence or the presence of Lfcin (Figures 4.3c and 4.3d). The addition of Lfcin to P O P G significantly lowered the I2850/I2880 ratio from approximately 1.0 to 0.7 over the temperature range investigated (Figure 4.3c), but had no effect on the I2935/I2880 ratio (Figure 4.3d). The I2850/I2880 ratio of D P P E remained steady at approximately 0.4 at temperature below 32 °C in the absence of Lfcin with the appearance of a small hump near 20 °C (Figure 4.4a). The I2850/I2880 ratio increased to 0.8 between 30 °C and 60 °C, and fluctuated between 0.8 and 1.0 at temperatures above 60 °C. The presence of Lfcin in D P P E liposomes led to an overall increase in the I2850/I2880 ratio, particularly at temperature below 40 °C. A hump was also observed between 0 °C and 20 °C. The I2935/I2880 ratio of the D P P E liposomes in either the absence or the presence of Lfcin remained stable at around 0.4 at temperature below 40 °C (Figure 4.4b). At about 40 °C, the ratio increased. In the absence of Lfcin, the ratio increased to above 0.8; however, with the addition of Lfcin to the system, the ratio started to rise at 42 °C and only reached to approximately 0.7. P O P E exhibited very little response to temperature changes. Only gradual increases were observed in both the I2850/I2880 arid the I2935/I2880 ratios (Figures 4.4c and 4.4d). The presence of Lfcin led to little changes in the thermal behaviour of P O P E liposomes. 87 4.4. Discussion 4.4.1. Temperature-induced Changes in Raman Spectra of Model Membranes The two dominant Raman features at 2850 and 2880 cm" 1 are assigned, respectively, to the symmetric and the asymmetric C-H stretching modes for the coupled methylene moieties of the hydrocarbon chains in the bilayer interior (Lewis and McElhaney, 2002). The band around 2935 cm" 1 mainly arises from the symmetric C-H stretching mode of terminal methyl groups in the hydrophobic center of the bilayer (Devlin and Levin, 1989). The bsso/bsso ratio directly monitors acyl chain disorder/order arising from lateral chain-chain interactions, while the I2935/I2880 ratio also furnishes an index of the degree of interchain disorder/order (Bunow and Levin, 1977). I ncreases in b 9 3 s / b 8 8 o ratio also indicate increases in the number of gauche conformations along the lipid chains (Huang et al., 1982). In other words, increases in bsso/bsso arid b 9 3 s / b s 8 o ratios reflect the loosening or increased fluidity and mobility of the phospholipid chains. As observed in the Raman spectra for all the phospholipids in the present study, these bsso/bsso and I2935/I2880 ratios increased as temperature increased, indicating a loosening effect. The thermal behaviours of the phospholipids observed in the Raman spectra are in agreement with those previously reported. Bonora et al. (2003) reported that D P P C exhibited two endothermic transitions in the 25-50 °C temperature range: a broad pre-transition, with a low enthalpy change at about 35 °C and a major, sharp transition at about 41.7 °C. Raman spectra of planar supported lipid bilayers of D P P C showed large thermal changes over the range of 25 to 50 °C as indicated by the bsso/bsso ratio. A broad and gradual increase in the bsso/bsso ratio of the D P P C lipid bilayers between 20 and 60 °C and a sharp increase in the I2935/I2880 ratio at around 40 °C were detected. Lee and Bain (2005) reported that the spectra of P O P C showed only small changes over the range of 14 to 41 °C. Similar results were observed in the present study. In comparison, it was previously reported that dimyristoyl P E exhibited only one main sharp transition near 49.9 °C in the thermal range of 30-60 °C (Bonora et al., 2003). The bsso/bsso ratios in the present study showed that D P P E has a transition temperature at approximately 42 °C, whereas the I2935/I2880 ratios indicated a transition temperature of D P P E at 50 °C. Furthermore, the bsso/bsso ratios indicated a pre-transition stage between 0 and 20 °C in D P P E liposome. The origin of the pre-transition thermal behaviour observed in liposomes is not very well established; it has been proposed to arise from the conversion of a lamellar gel phase (Lp) to a rippled gel phase (P p). When the phospholipid bilayer is heated, increased rotational motion of the glycerophospholipids occurs, especially about the C-C bonds of the acyl chains, causing the acyl chains to melt and to initiate the first stage of the transition to the liquid state through the conversion of L p to Pp. As the temperature is further increased, the main transition arises from the conversion of the Pp gel phase to a lamellar liquid-crystal L a phase. Further temperature increases lead to more 88 extensive disordering of the acyl chains and rearrangements into the nonlamellar, cubic and hexagonal phases. The I2850/I2880 and I2935/I2880 ratios observed in the present study supported these phenomena; the transition temperatures indicated by the bssrAsso ratios of D P P C , D P P G , and D P P E liposomes were lower than those indicated by the I2935/I2880 ratios. For instance, the I2850/I2880 ratios of D P P C began to increase at around 10 °C and indicated a transition temperature at around 20 °C (Figure 4.2a), whereas the I2935/I2880 ratios remained stable at temperatures between 10 and 20 °C, started to increase at about 25 °C, and showed a transition temperature near 40 °C (Figure 4.2b). Similar results could be detected in D P P G (Figures 4.3a and 4.3b) and D P P E (Figure 4.4a and 4.4b) l iposomes. This indicated that as temperature increased, the lateral chain-chain interactions began to increase, increasing distance among chains in the hydrophobic interior. Intra-chain trans/gauche isomerizations became possible only after further increases in temperatures when the P p gel phase was ready to be converted to a lamellar liquid-crystal L a phase. The lamellar (or liquid crystalline) state of phospholipids is considered to be the biologically active state of the membrane. In the lamellar state, the lipid molecules are melted and trans-gauche isomerizations are able to propagate freely up and down the acyl chains; however, rotation of the first 8 to 10 carbons in the acyl chains is restricted. Distortions caused by acyl chains prevent a perfect packing order from being established and contribute to the level of membrane fluidity observed. With the lamellar state, a variety of acyl chain organizations can be accommodated, allowing for a gradient of phase states to occur within the membrane. Upon transition to the gel state, the lipid chains become stiff and the frequency of these rotations is reduced.. Upon increasing membrane disturbance, transition to the cubic and inverted hexagonal (HM) phases may also occur (Denich et al. 2003). However, a greater severity of disturbance is required to change the bilayer towards H M phase than is needed to cause gel-to-liquid-crystalline transition. The size and charge of the head groups may affect the packing of glycerophospholipids in the bilayer. Comparison between D P P C and D P P E liposomes shows that D P P E has a higher transition temperature than D P P C . The I2935/I2880 ratios of D P P C began to rise at around 35 °C and showed a transition temperature at around 40 °C (Figure 4.2b), while the I2935/I2880 ratios of D P P E began to rise at around 40 °C and continued to increase above 60 °C (Figure 4.4b). This indicates that D P P E molecules pack more tightly than D P P C molecules. The structures of the two molecules are similar, differing only by their terminal functional groups. D P P C has a N ( C H 3 ) 3 + head group and D P P E has N H 3 + as its headgroup. Since.both head groups carry a positive charge, the differences in head group size would account for the differences in their thermal behaviour. Molecules of D P P E have a smaller polar head group, meaning that charge derealization occurs over a smaller area. This creates a stronger charge on the D P P E head group and allows for a more stable membrane structure 89 to exist as compared with membranes composed of D P P C molecules. The smaller head group allows D P P E to interact with adjacent head groups through intermolecular interactions. More energy would therefore be needed to disturb the acyl chains in D P P E liposomes than those in D P P C liposomes. D P P G has a neutral, C H 2 C H ( O H ) C H 2 O H , head group, and therefore a overall negative charge contributed from the phosphate group. The glycerol moiety is similar in size to the positively-charged choline head group of D P P C , which is a zwitterion carrying no net charge. Results from the present study showed that the I2935/I2880 ratios of both D P P C and D P P G liposomes began to increase at approximately 35 °C and exhibited a gel to liquid-crystalline transition near 40 °C (Figures 4.2b and 4.3b, respectively). However, at lower temperatures, the b s s o / b s s o ratios of D P P G were about 0.9 (Figure 4.3a), whereas the b s s o / b s s o ratios of D P P C were about 0.7 (Figure 4.2a). An opposite trend was observed in the I2935/I2880 ratios of D P P G and D P P C . At low temperatures, D P P G had I2935/I2880 ratios near 0.4 (Figure 4.3b) and D P P C had b g s s / b s s o ratios near 0.5 (Figure 4.2b). This indicated that the charge of the head group had an impact on the structural properties of the gel phase of phospholipids. The overall negativity of the D P P G headgroup might have caused a disturbance in the acyl chains, leading to the higher b s s o / b s s o ratios as observed (Figure 4.3a). The disturbance near the surface of the phospholipids bilayer could have led to a more tightly packed hydrophobic centre, leading to more ordered intra-chain packaging and lower I2935/I2880 ratios in the D P P G liposomes at low temperatures. Results from the present study showed that the negatively charged headgroup had an effect on the packing properties of phospholipid liposomes at low temperatures. However, the overall thermal dependence of the gel to liquid-crystalline phase transition was little affected by the charge of the headgroup. In contrast, the size of the headgroup could significantly affect the phase transition of phospholipid liposomes. The small head groups of P E allow strong hydrophobic interaction among its acyl chains, forming a more stable structure. All three phospholipids containing unsaturated fatty acids tested in the present study exhibited gradual increases in b s s o / b s s o and I2935/I2880 ratios over the temperature range examined. Phase transition was observed in P O P C near 0 °C as indicated by the I2935/I2880 ratios (Figure 4.2d). I2935/I2880 ratios of P O P E also indicated a gradual phase transition between 0 and 10 °C (Figure 4.4d). These results are in agreement with those reported (Caffrey and Hogan, 1992). P O P C , P O P G , and P O P E had lower transition temperatures than their saturated counterparts due to the presence of the cis-double bond in the oleoyl hydrocarbon chains. The cis-double bond introduces a kink in the acyl chains which hinders close packing, de-stabilizes the liposomes, and causes the liposomes to remain in a liquid-crystalline phase even at low temperatures. Since the majority of the temperature range examined in the present study was higher than the transition temperatures of the phospholipids 90 studied, only the thermal behaviour of the liquid crystalline phase of these phospholipids was observed. The introduction of heat at above the transition temperatures continued to induce lateral chain-chain interactions and disorder interchain structure. 4.4.2. Effects of Lactoferricin on Model Membranes The results of the present study showed that Lfcin exhibited different effects on the thermal behaviour of phospholipids depending on the headgroup and acyl chain composition. Figures 4.2a and 4.2b indicated that the presence of Lfcin in D P P C liposomes led to an overall decrease in both the I2850/I2880 and I2935/I2880 ratios, indicating that both the intermolecular and intramolecular interactions of D P P C became more ordered in the presence of Lfcin over the temperature range examined. In other words, D P P C , a common phospholipid found in mammalian cells, became more ordered in the presence of Lfcin. Similar effects have not been reported for other cationic peptides. Melittin, for instance, led to the conformational disorder of D M P C as examined by Raman spectroscopy (Dasseux et al., 1984). The Raman spectra showed decreases in C H 2 groups in the trans conformation and intermolecular order of the chains, without significant changes in the C H 2 groups trans/gauche ratio. Acyl chain symmetry was also lowered in the presence of melittin. Similar results were reported for natural sphingomyelin. In contrast, in the present study, Lfcin increased the order of both the intermolecular and intramolecular movement of the acyl chains in D P P C liposomes. This could be due to the presence of hydrophobic amino acids in Lfcin, which might have been incorporated into the hydrophobic interior of the phospholipid bilayer, increasing the order of the acyl chains. However, the exact mechanism is not known. In contrast to its effects on D P P C liposomes, Lfcin affected the D P P G liposomes only at temperatures above the transition temperatures. Figure 4.3a indicates that the presence of Lfcin lowered the I2850/I2880 ratio from 1.1 to 1.0 at temperatures above 38 °C, while Figure 4.3b indicates that the I2935/I2880 ratios at above 38 °C were also lowered to 0.8 in the presence of Lfcin. These changes in the l2 8 5o/l288o and I2935/I2880 ratios indicated that Lfcin restricted both inter- and intra-chain . interactions of D P P G ' s acyl chains above transition temperature. At low temperature, the interaction of the negatively charged head groups of P G with the positively charged amino acid residues on Lfcin appear to have restricted the position of Lfcin near the polar head group layer surface, and thereby inhibited the interaction of Lfcin with the hydrophobic core of the lipid bilayer. In contrast, at temperatures above the transition temperature, electrostatic repulsion among the charged head groups seems to have been minimized and hydrophobic attraction between the acyl chains and hydrophobic amino acids was strengthened, leading to the lowering of both l 2 8 5 o / l 2 8 8 o and l 2 9 3 5 / l 2 8 8 o ratios. Dumas et al. (1999) stated that neutral and polar head groups may regulate membrane proteins through electrostatic interactions. For example, the head groups of phosphatidylglycerol, phosphatidic acid, phosphatidylserine, and phosphatidylinositol are acidic and may interact with positively charged amino acid residues (Epand, 1998). Nevertheless, effects of temperatures on the 91 electrostatic interactions between acidic head groups and positively charged peptides/proteins have not been reported in the literature. The influence of Lfcin on D P P E liposomes also differed from those observed for D P P C and D P P G . The addition of Lfcin increased the intermolecular fluidity of the D P P E acyl chains and. lowered the transition temperature of D P P E as indicated by changes in the bsso/bsso ratios (Figure 4.4a). D P P E is characterized by a cone-shaped molecular geometry and possesses a positive spontaneous curvature. D P P E are prone to adopt an inverse hexagonal (HM) phase and exhibit the largest hydrophobic cross-section (Lohner and Prenner, 1999). Hence, this could have allowed greater interactions between the hydrophobic amino acids of Lfcin and the acyl chains and led to the formation of peptide/lipid pores or nonlamellar lipid structures (Lohner and Prenner, 1999). Both these mechanisms will destabilize the acyl chains and result in massive membrane perturbation (Lohner and Prenner, 1999). Vogel et al. (2002) reported that at high Lfcin to lipids ratios (10:1), the characteristic melting transitions of D P P C and D P P E in multilamellar liposomes were not affected by further addition of the Lfcin. In contrast, for D P P G , the researchers noted that a broad pre-transition around 32 °C disappeared. When more peptide was added to D P P G liposomes, the peak for the main transition also shifted to a slightly higher temperature. These results show that the peptide interacted more strongly.with the overall negatively charged P G headgroup than the zwitterionic P C or P E . Results from fluorescence spectrophotometry also showed that lactoferricin peptides bound more tightly and penetrated somewhat more deeply into vesicles containing 40% negatively charged P G headgroup (Vogel et al., 2002). It was estimated that the Trp aromatic rings resided close to the glycerol portion of the phospholipids (Schibli et al., 2002). Vogel et al. (2002) suggested that D P P E can undergo a transition from a bilayer to a hexagonal phase at around 43.5 °C. Perturbation of the bilayer structure by the antimicrobial peptide magainin through an effect called positive membrane curvature leads to an increase in the transition temperature (Matsuzaki et al., 1997 and 1998). It was shown that Lfcin and other antimicrobial peptides including tritrpticin (Schibli et al., 2002) and indolicidin (Selsted et ai., 1992) are capable of perturbing the regular bilayer structure, perhaps promoting the formation of a pore structure with the lipids having positive curvature around the rim of the pore in the direction of the bilayer.normal. It is possible that the membrane-destabilizing effects of Lfcin may allow it to spontaneously cross bacterial membranes. Unsaturated phospholipids, P O P C and P O P E , were relatively unaffected by the presence of Lfcin (Figures 4.2c, 4.2d, 4.4c, and 4.4d). However, the addition of Lfcin to P O P G liposomes decreased the bsso/bsso ratio over the range of temperatures tested. In the absence of Lfcin, P O P G liposomes had a bsso/bsso ratio around 1.0, while in the presence of Lfcin, the ratio dropped to around 92 0.7 (Figure 4.3c). Similar observations were reported in the presence of V4, an amphipathic cationic peptide, in P O P C , P O P G , and P O P E (Yu et al., 2005). Fluorescence correlation spectrophotometry showed that V4 had a higher affinity for P O P G than for P O P C and P O P E . These results indicated that electrostatic interaction is the major driving force for the binding process in phospholipid liposomes containing unsaturated acyl chains. The I2935/I2880 ratios of P O P G , however, were not affected by the presence of Lfcin (Figure 4.3d). Results from a 2 H NMR experiment (Thennarasu et al., 2005) also showed that the effect of a lipopeptide, MSI-843, on P O P C was maximum for C D 2 groups closer to the head group region and decreased along the acyl chain towards a minimum for the other end of the chain. In the present study, positively charged amino acids oh Lfcin interacted with the anionic head groups on the surface of the lipid bilayer of P O P G . Once Lfcin molecules were anchored on the surface of the P O P G liposomes, hydrophobic amino acids on Lfcin were allowed to interact with the hydrophobic acyl chains near the head group of the acyl chains. Results from the present study showed that Lfcin can interact with phospholipids and affect the thermal stability of the system. However, depending on the charge and size of the head groups of the phospholipids and the saturation level of the acyl chains, the effects could vary. Lohner and Prenner (1999) suggested in a review that charge-charge interactions, membrane curvature strain and hydrophobic mismatch between peptides and lipids are important parameters in determining the mechanism of membrane perturbation. Depending on the molecular properties of both lipid and peptide, antimicrobial peptides could cause membrane thinning, pore formation, promotion of nonlamellar lipid structures, or bilayer disruption. Results from this study showed that Lfcin, an amphipathic cationic peptide, was particularly effective for disrupting P G containing micelles, a negatively charged phospholipid. A better understanding of the mutual dependence of these parameters will help to elucidate the molecular mechanism of membrane damage by antimicrobial peptides and their target membrane specificity, keys for the rational design of novel types of peptide antibiotics. 93 2850 2880 2935 40 °C 20 °C 0 °C -10 °C 2800 2900 3000 Raman Shift, cm" 1 Figure 4.1. Raman spectra of dipalmitoylphosphatidylcholine (DPPC) liposomes in the presence of lactoferricin showing changes in the major peaks near 2850, 2880, and 2935 cm" 1 as a function of temperature. 94 1 . 6 1 . 4 1 . 2 1 . 0 0 . 8 0 . 6 0 . 4 0 . 2 0 . 0 4-(a) Intensity 2850/2880 D P P C O O oo o Sm *4 IS • - 2 0 1 . 4 1 . 2 1 . 0 0 . 8 0 . 6 0 . 4 0 . 2 H 0 . 0 0 2 0 4 0 6 0 Temperature (°C) (b) Intensity 2935/2880 D P P C OO O O O O s « ^ p O _ m_ 8 0 1 1 - 2 0 0 2 0 4 0 6 0 Temperature (°C) 8 0 1 . 6 1 . 4 1 . 2 1 . 0 0 . 8 0 . 6 0 . 4 0 . 2 0 . 0 (c) Intensity 2850/2880 P O P C - 2 0 1 . 4 - i 1 . 2 -o 1 . 0 -<- o 00 CN 0 . 8 -CO CM 0 . 6 -0 . 4 -0 . 2 -o . o - -0 2 0 4 0 Temperature (°C) (d) Intensity 2935/2880 P O P C 6 0 8 0 o - 2 0 0 2 0 4 0 6 0 Temperature (°C) 8 0 Figure 4.2. bsso/bsso ratio of D P P C (a), I2935/I2880 ratio of D P P C (b), bsso/bsso ratio of P O P C (c), and I2935/I2880 ratio of P O P C (d) in the absence (O) or presence ( • ) of Lfcin measured between -10 °C and 70 °C by Raman spectroscopy. 1.6 1.4 1.2 1.0 0.8 o CO CO CM _ _ S 0.6 in CO CM 0.4 0.2 0.0 (a) Intensity 2850/2880 D P P G ml — i 1 1 1 — -20 0 20 40 60 Temperature (°C) 80 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 (c) Intensity 2850/2880 P O P G < > o o o o ^ o o o 0 o -20 0 20 40 60 80 Temperature (°C) 1.4 1.2 1.0 J 0.8 J 0.6 0.4 0.2 0.0 (b) Intensity 2935/2880 D P P G o o o . o o I 1 1 1 — -20 0 20 40 60 Temperature (°C) 80 1.4 1.2 1.0 0.8 H 0.6 0.4 0.2 0.0 (d) Intensity 2935/2880 P O P G • • o o ° -20 20 40 60 Temperature (°C) 80 Figure 4.3. bssr/booo ratio of D P P G (a), I2935/I2880 ratio of D P P G (b), bssrAsso ratio of P O P G (c), and I2935/I2880 ratio of P O P G (d) in the absence (O) or presence ( • ) of Lfcin measured between -10 °C and 70 °C by Raman spectroscopy. 96 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 (a) Intensity 2850/2880 DPPE O 1.6 1.4 1.2 1 1.0 CM J 0.8 0.6 0.4 0.2 0.0 -20 1.4 1.2 1.0 0.8 0.6 0^4 0.2 0.0 0 20 4 0 6 0 80 Temperature (°C) (b) Intensity 2935/2880 DPPE O o o . 0 ST ' -20 0 20 40 60 80 Temperature (°C) (c) Intensity 2850/2880 POPE • n 1 -20 0 20 4 0 60 . 80 Temperature (°C) 1.4 1.2 1.0 0.8 •0.6 0.4 0.2 0.0 (d) Intensity 2935/2880 POPE O -20 20 40 60 80 Temperature (°C) Figure 4.4. bsso/bsso ratio of D P P E (a), bcWbsso ratio of D P P E (b), bsso/bsso ratio of P O P E (c), and I2935/I2880 ratio of P O P E (d) in the absence (O) or presence (• ) of Lfcin measured between -10 °C and 70 °C by Raman spectroscopy. 97 4.5. References Badii, F.; Howell, N. K. Elucidation of the effect of formaldehyde and lipids on frozen stored cod collagen by FT-Raman spectroscopy and differential scanning calorimetry. J. Agric. Food Chem. 2003, 26, 1440-1446. Bonora, S.; Torreggiani, A.; Fini, G . D S C and Raman study on the interaction between polychlorinated biphenyls (PCB) and phospholipid liposomes. Thermochimica Acta 2003, 408, 55-65. Bouchard, M.; Auger, M. Solvent history dependence of gramicidin-lipid interactions: a Raman and infrared spectroscopic study. Biophys. J. 1993, 65, 2484-2492. Bunow, M. R.; Levin, I. W. Comment on the carbon-hydrogen stretching region of vibrational Raman spectra of phospholipids. Biochim. Biophys. Acta 1977, 487, 388-394. Caffrey, M.; Hogan, J . LIPIDAT: a database of lipid phase transition temperatures and enthalpy changes. D M P C data subset analysis. Chem. Phys. Lipids 1992, 61, 1-109. Choung, S.-Y.; Kobayashi, T.; Takemoto, K.; Ishitsuka, H.; Inoue, K. Interaction of a cyclic peptide, Ro09-0198, with phosphatidylethanolamine in liposomal membranes. Biochim. Biophys. Acta 1988,940,180-187. Dasseux, J . L.; Faucon, J . -F. ; Lafleur, M.; Pezolet, M.; Dufourcq, J . A restatement of melittin-induced effects on the thermotropism of zwitterionic phospholipids. Biochim. Biophys. Acta 1984, 775, 37-50. Denich, T. J . ; Beaudette, L. A.; Lee, H.; Trevors, J . T. Effect of selected environmental and physico-chemical factors on bacterial cytoplasmic membranes. J. Microbiol. Methods 2003, 52, 149-182. Devlin, M. T.; Levin, I. W. Raman spectroscopic studies of the packing properties of mixed dihexadecyl-and dipalmitoylphophatidylcholine bilayer dispersion. Biochemistry 1989, 28, 8912-8920. Dumas, F.; Lebrun, M. C ; Tocanne, J . F. Is the protein/lipid hydrophobic matching principle relevant to membrane organization and functions? FEBS Lett. 1999, 458, 271 -277. Epand, R. M. Lipid polymorphism and protein-lipid interactions. Biochim. Biophys. Acta 1998, 7376, 353-368. Epand, R. M.; Vogel, H. J . Diversity of antimicrobial peptides and their mechanisms of action. Biochim. Biophys. Acta 1999, 7462, 11-28. Gaber, B. P.; Peticolas, W. L. On the quantitative interpretation of biomembrane structure by Raman spectroscopy. Biochim. Biophys. Acta 1977, 465, 260-274. Herbig, M. E.; Fromm, U.; Leuenberger, J . ; Krauss, U.; Beck-Simkinger, A. G. ; Merkle, H. P. Bilayer interaction and localization of cell penetrating peptides with model membranes: a comparative study of human calcitonin (hCT)-derived peptide with p V E C and pAntp(43-58). Biochim. Biophys. Acta 2005, 7772, 197-211. 98 Huang, C. -H. ; Lapides, J . R.; Levin, I. W. Phase-transition behaviour of saturated, symmetric chain phospholipid bilayer dispersions determined by Raman spectroscopy: Correlation between spectral and thermodynamic parameters. J. Am. Chem. Soc. 1982, 104, 5926-5930. Kang, J . H.; Lee, M. K.; Kim, K. L.; Hahm, K. S. Structure-biological activity relationships of 11-residue highly basic peptide segment of bovine lactoferrin. Int. J. Pept. Protein Res. 1996, 48,357-363. Kanyshkova, T. G. ; Semenov, D. V.; Buneva, V. N.; Nevinsky, G. A.. Human milk lactoferrin binds two DNA molecules with different affinities. FEBS Lett. 1999, 457, 235-237. Lafleur, M.; Faucon, J . F.; Dufourcq, J . ; Pezolet, M. Perturbation of binary phospholipid mixtures by melittin: a fluorescence and Raman spectroscopy study. Biochim. Biophys. Acta. 1989, 980,85-92. Laroche, G. ; Carrier, D.; Pezolet, M. Study of the effect of poly(L-lysine) on phosphatidic acid and phosphatidylcholine/phosphatidic acid bilayers by Raman spectroscopy. Biochemistry 1988, 27, 6220-6228. Laroche, G. ; Dufourc, E. J . ; Pezolet, M.; Dufourcq, J . Coupled changes between lipid order and polypeptide conformation at the membrane surface. A 2H N M R and Raman study of polylysine-phosphatidic acid systems. Biochemistry 1990, 29, 6460-6465. Lee, C ; Bain, C. D. Raman spectra of planar supported lipid bilayers. Biochim. Biophys. Acta 2005, 1711, 59-71. Lewis, R. N. A. H.; McElhaney, R. N. Vibrational spectroscopy of lipids. In. Handbook of Vibrational Spectroscopy, Vol. 5, Chalmers, J . M. and Griffiths, P. R. (Eds), Wiley, Chichester, UK, 2002, pp. 3447-3464. Lohner, K.; Prenner, E. J . Differential scanning calorimetry and X-ray diffraction studies of the specificity of the interaction of antimicrobial peptides with membrane-mimetic systems." Biochim. Biophys. Acta 1999, 7462, 141-156. Matsuzaki, K.; Sugishita, K.-l.; Harada, M.; Fujii, N.; Miyajima, K. Interactions of an antimicrobial peptide, magainin 2, with outer and inner membranes of Gram-negative bacteria. Biochim. Biophys. Acta 1997, 7327, 119-130. Matsuzaki, K.; Sugishita, K.-l.; Ishibe, N.; Ueha, M.; Nakata, S.; Miyajima, K.; Epand, R. M. Relationship of membrane curvature to the formation of pores by magainin 2. Biochemistry 1998,37, 11856-11863. Neville, F.; Cahuzac, M.; Konovalov, O.; Ishitsuka, Y.; Lee, K. Y . F.; Kuzmenko, I.; Kale, G . M.; Gidalevitz, D. Lipid headgroup discrimination by antimicrobial peptide LL-37: Insight into mechanism of action. Biophys. J. 2006,90,1275-1287. Opekarova, M.; Tanner, W. Specific lipid requirements of membrane proteins - a putative.bottleneck in heterologous expression. Biochim. Biophys. Acta 2003 ,767^11-22. Schibli, D. J . ; Epand, R. F.; Vogel, H. J . ; Epand, R. M. Tryptophan-rich antimicrobial peptides: comparative properties and membrane interactions. Biochem. Cell Biol. 2002, 80, 667-677. Selsted, M. E.; Novotny, M. J . ; Morris, W. L.; Tang, Y. Q.; Smith, W.; Cullor, J . S . Indolicidin, a novel bactericidal tridecapeptide amide from neutrophils. J. Biol. Chem. 1992, 267, 4292-4295. 99 Snyder, R.; Scherer, J . ; Gaber, B. Effects of chain packing and chain mobility on the Raman spectra of biomembranes. Biochim. Biophys. Acta 1980, 607, 47-53. Snyder, R.; Strauss, H.; Elliger, C. C - H stretching modes and the structure of n-alkyl chains. I. Long, disordered chains. J. Chem. Phys. 1982, 86, 5145-5150. Susi , H.; Sampugna, J . ; Hampson, J . W.; Ard, J . S. Laser-Raman investigation of phospholipid-polypeptide interactions in model membranes. Biochemistry 1979, 18, 297-301. Thennarasu, S.; Lee, D.-K.; Tan, A.; Prasad Kari, U.; Ramamoorthy, A. Antimicrobial activity and membrane selective interactions of a synthetic lipopeptide MSI-843. Biochim. Biophys. Acta 2005,7777,49-58. Vincent, J . S. ; Revak, S. D.; Cochrane, C. G. ; Levin, I. W\ Raman spectroscopic studies of model human pulmonary surfactant systems: phospholipid interactions with peptide paradigms for the surfactant protein S P - B . Biochemistry 1991,30,8395-8401. Vincent, J . S. ; Revak, S. D.; Cochrane, C. D.; Levin, I. W. Interactions of model human pulmonary surfactants with a mixed phospholipid bilayer assembly: Raman spectroscopic studies. Biochemistry 1993, 32, 8228-8238. Vogel, H. J . ; Schibli, D. J . ; Jing, W.; Lohmeier-Vogel, E. M.; Epand, R.F.; Epand, R. M. Towards a structure-function analysis of bovine lactoferricin and related tryptophan- and arginine-containing peptides. Biochem. Cell Biol. 2002, 80, 49-63. Yeagle, P. L. The Lipids of Cell Membranes. In. The Membranes of Cells. 1993. New York, Academic Press Inc.: pp. 22-28. Yu, L.; Ding, J . L.; Ho, B.; Wohland, T. Investigation of a novel artificial antimicrobial peptide by fluorescence correlation spectroscopy: An amphipathic cationic pattern is sufficient for selective binding to bacterial type membranes and antimicrobial activity. Biochim. Biophys. Acta 2005, 7776, 29-39. 100 CHAPTER 5. RAMAN SPECTROSCOPIC STUDY OF THE EFFECTS OF LACTOFERRICIN DERIVATIVES ON PHOSPHOLIPID LIPOSOMES 4 5.1. Introduction Lactoferrin (LF) is an iron-binding protein that is present in milk (Sorensen and Sorensen, 1939) and exhibits a wide range of antimicrobial activities. It has been shown that pepsin hydrolysate of LF has higher antimicrobial potency than undigested LF. A 25-residue peptide named lactoferricin (Lfcin), released from the amino-terminal region of LF during peptic digestion, has been identifited as being responsible for the antimicrobial properties (Bellamy et al., 1992). The peptide is highly basic, containing many Arg and Lys. Furthermore, a number of Trp and Phe are also present in Lfcin. Studies have also shown that fragments, such as 1 7 F K C R R W Q W R M K K L G A PSITC V R R A F 4 1 (Bellamy et al., 1992), 1 7 F K C R R W Q W R M K K L G A 3 1 (Rekdal et al., 1999; Strom et al., 2000), 2 0 R R W Q W M K K L G 3 0 (Kang et al., 1996), and 2 0 R R W Q W R 2 5 - N H 2 (Tomita et al., 1994) of Lfcin possess similar antimicrobial activity. The presence of hydrophobic and positively charged amino acids has been reported to be particularly important for the bioactivity of Lfcin. Q S A R analyses (Lejon et al., 2001; Strom et al., 2000) of Lfcin revealed that net charge, charge asymmetry, and micelle affinity of Lfcin were the most important structural parameters affecting their antimicrobial activity. Using principal component similarity analysis followed by artificial neural network, helix propensity and net charge of the peptides were demonstrated to have a significant effect on the antimicrobial activity of Lfcin and its derivatives against Escherichia coli (Chapter 3). Franaud et al. (2004) showed that the replacement of Trp by Ala (W22A or W24A) led to a decrease in antimicrobial activity against wild-type E. coli and L P S mutant strains. An alanine-scan experiment on Lfcin (Strom et al., 2000) revealed that the presence of Trp22 orTrp24 was mandatory for antimicrobial activity. A loss of antibacterial activity against S. aureus was observed when Phe replaced the Trp residues (Haug and Svendsen, 2001). Aside from the hydrophobic Trp residues, cationic Arg residues also play an important role in the bioactivity of Lfcin. Even "conservative" replacements of Arg with Lys produce peptides of much lower antimicrobial potency (Strom et al., 2000). Strom et al. (2001) identified the key amino acids in Lfcin by synthesizing a series of Lfcin derivatives in which amino acids were successively replaced with alaine (Ala). Similar effects have been observed for other Trp- or Arg-rich antimicrobial peptides. For example, Blondelle et al. (1995), Blondelle and Houghton (1996), and Blondelle and Lohner (2000) found that only Trp/Arg-rich hexapeptides were potent antimicrobials, while hexapeptides 4 A version of this chapter will be submitted for publication. Chan, J . C. K., and Li-Chan, E. C. Y. Raman spectroscopic study of the effects of lactoferricin derivatives on phospholipid liposomes. Biochimica et Biophysica Acta (BBA) - Biomembranes. 101 composed of Phe/Arg, Tyr/Arg, Trp/Lys, or Tyr/Lys and Phe/Lys exhibited much lower antimicrobial activities. Early studies suggested that binding to the microbial membrane surfaces was important for the action of Lfcin (Bellamy et al., 1993). However, it is not clear exactly where and how Lfcin interacts with bacterial and host membranes of different compositions. Phosphatidylglycerol (PG) and phosphatidylethanolamine (PE) are the major components of bacterial membranes (Opekarova and Tanner, 2003). In contrast, erythrocyte membranes and most other eukaryotic membranes are generally made up of phosphatidylcholine (PC). The positively charged choline ((CH3) 3 N + CH 2 CH2-) on P C and ethanolamine ( H 3 N + C H 2 C H 2 - ) on P E , together with the negatively charged phosphate moiety, form zwitterions leading to overall neutral phospholipids. In contrast, P G carries a glycerol ( C H 2 O H C H O H C H 2 O H - ) on its polar head, forming overall negatively charged phospholipids. Raman spectroscopy is a very useful tool to examine the microenvironment of C-H groups in the fatty acyl hydrocarbon chains of phospholipids. Results from Raman spectroscopic study reported in Chapter 4 indicated that Lfcin affected the thermal behaviour of P C , and destabilized the acyl chains and decreased the degree of order of P E . Furthermore, negatively charged P G may have interacted with the positive residues on Lfcin through electrostatic force and facilitated the interaction of hydrophobic fatty acyl chains with the hydrophobic amino acid residues on Lfcin. The objective of the present study was to examine the influence of insertion or replacement of tryptophan and arginine residues in a 15-residue fragment of Lfcin, on the effects of these derivatives on the thermal transition profiles of phospholipids using Raman spectroscopy in the C-H stretching region. The effect of a 6-residue peptide sequence containing the active core of Lfcin was also investigated. 5.2. Materials and Methods 5.2.1. Materials Sodium salts of D P P C , D P P G , and D P P E were obtained from Avanti Polar Lipids (Alabaster, AL) in lyophilized powder forms. Lactoferricin derivatives, F K C R R W Q W R M K K L G A (Lfcin 1 5), F K C A A W Q W A M K K L G A (Lfcin 1 5R-), F K C R R A Q A R M K K L G A (Lfcin 1 5W-), F K C W R W Q W R W K K L G A (Lfcin 1 5W+), and R R W Q W R - N H 2 (LfcinCore), were purchased from Sigma-Genosys (The Woodlands, Texas) (Table 5.1). H P L C , mass spectrometry, and amino acid analyses were performed by Sigma-Genosys to confirm that the Lfcin derivatives had the expected masses and amino acid compositions. All samples were used without further purification. Other chemicals used were of reagent grade quality. 102 5.2.2. Preparation of Liposomes As described in Section 4.2.2. 5.2.3. Raman Spectroscopy As described in Section 4.2.3. 5.2.4. Raman Spectral Analysis As described in Section 4.2.4. 5.2.5. Properties of Lactoferricin Derivatives Structural properties of the lactoferricin derivatives were calculated. Molecular weight, pi, and hydropathicity index of the peptides were determined using the ProtParam tool available on http://ca.expasy.org/tools/protparam.html (Kyte and Doolittle, 1982). Helix and charge pattern similarities of residues 4 to 10 of the derivatives were calculated using the Homology Similarity Analysis (HSA) described in Chapter 3. L fc in 1 5 was used as the reference peptide in HSA. 5.3. Results 5.3.1. Effects of Lactoferricin Derivatives on D P P C Liposomes Raman spectroscopy in the C-H stretching region was used to investigate the effects of Lfcin derivatives on temperature transitions of D P P C , D P P G , and D P P E . Analyses of the effects of temperature dependence of the acyl chain packing characteristics and phase transition cooperativities on the various Raman spectroscopic features can be achieved through peak-height intensity ratios of peaks near 2850 cm" 1 ( l 2 85o). 2880. cm" 1 (l 2s8o). and 2935 cm" 1 (I2935). The bssr/bsso ratio provides insight into the intermolecular chain-chain interactions among the acyl chains in the hydrophobic core of the phospholipid bilayer, while the I2935/I2880 ratio indicates the degree of disorder and order of the acyl chains as well as the presence of intrachain trans/gauche isomerizations (Huang et al., 1982). The bssrAsso and I2935/I2880 ratios of the Raman spectra of D P P C , D P P G , and D P P E as a function of temperature are shown in Figures 5.1, 5.2, and 5.3, respectively. The intensity ratio, I2850/I2880. of D P P C liposomes gradually increased over the temperature range tested (Figure 5.1a). In the absence of Lfcin (open square symbols), a broad gel-to-liquid crystalline phase transition was detected between 30 °C and 50 °C, with a pre-transition stage at about 15 °C to 20 °C (Figure 5.1a). With the presence of L fc in^ in the liposomes (Figure 5.1 ai), I2850/I2880 ratios were lowered throughout the temperature range while exhibiting similar behaviour as the spectra without Lfcin. The presence of Lfc in 1 5 in the liposomes led to a decrease in the transition temperature at around 38 °C. The transition was also sharper in the presence of Lfcin than that without Lfcin, and a smaller pre-transition stage was seen between 15 °C and 20 °C. The intensity 103 ratio, I2935/I2880. showed a slightly different trend (Figure 5.1b). The I2935/I2880 ratios for both D P P C liposomes with and without Lfcin-|5 increased gradually at temperature above 20 °C, although the presence of Lfc in 1 5 attenuated the extent of this increase (Figure 5.1 bi). Effects of Lfc in 1 5 derivatives were less obvious when Arg was replaced with Ala (Figures 5.1 aii and 5.1 bii). The presence of Lfcini 5 R- led to a slight decrease in the bsso/bsso ratio at temperatures below 15 °C, and thermal transition of the D P P C liposome was observed to be initiated at a lower temperature near 32 °C (Figure 5.1 aii). The I2935/I2880 ratio was not affected by the presence of Lfc in 1 5 R- (Figure 5.1 bii). The addition of Lfc in 1 5 W- to the D P P C liposomes led to a decrease in bsso/bsso ratio at temperatures below 15 °C (Figure 5.1 aiii). The thermal behaviour of D P P C was unaffected by Lfc ini 5 W- at higher temperature as shown by the bsso/bsso ratio. However, differences in the b9 3 5 / l 2 88o ratios were observed (Figure 5.1 biii). In the absence of Lfcin derivatives, D P P C exhibited I2935/I2880 ratios near 0.4 at temperature below 20 °C, which gradually increased to 0.6 between 20 and 40 °C and continued to increase to 0.8 between 40 and 70 °C. The presence of Lfc in 1 5 W- in D P P C liposomes resulted in a very steady I2935/I2880 ratio at 0.2 at temperatures below 40 °C. A sharp increase in the I2935/I2880 ratio occurred between 40 and 55 °C and the ratio reached 0.8 at temperatures above 55 °C (Figure 5.1 biii). Lfcin 1 5W+ exhibited little effect on the thermal behaviour of D P P C as indicated by both I2850/I2880 and I2935/I2880 ratios (Figures 5.1 aiv and 1biv). The addition of LfcinCore to the D P P C liposomes lowered the bsso/bsso ratio at temperatures below 30 °C from 0.8 to 0.7 (Figure 5.1 av). The I2935/I2880 ratio of D P P C liposomes was unaffected by LfcinCore at temperatures below 40 °C, but at temperatures above 40 °C, LfcinCore lowered the ratio from 0.8 to 0.6. 5.3.2. Effects of Lactoferricin Derivatives on D P P G Liposomes D P P G exhibited higher ratios than those of D P P C (Figure 5.2). In the absence of any Lfcin derivatives (open diamond symbols), the bsso/bsso ratio of D P P G liposomes increased from 0.7 at temperatures below 0 °C to about 1.0 between 10 and 40 °C. A sharp increase in the ratio from 1.0 to 1.4 was observed around 40 °C. A similar pattern was observed in the I2935/I2880 ratio (Figure 5.2bi); a sharper and narrower increase in the I2935/I2880 ratio occurred at around 38 °C. The presence of Lfc in 1 5 had no observable effect on the thermal behaviour of D P P G liposomes at temperature below 40 °C as shown by the bsso/bsso ratio (Figure 5.2ai) and at temperatures below 30 °C as indicated by the I2935/I2880 ratio (Figure 5.2bi). The bsso/bsso ratio of D P P G in the presence of Lfc in 1 5 fluctuated between 0.8 and 1.2 between 20 and 70 °C (Figure 5.2ai). The I2935/I2880 ratio remained steady at around 0.4 to 0.6 between 0 and 40 °C in the presence of Lfc in 1 5 (Figure 5.2bi) and increased to 0.7 at around 42 °C. The presence of Lfc in 1 5 in D P P G 104 delayed the shift in I2935/I2880 ratio from 38 °C to 40 °C and led to a smaller change in the I2935/I2880 ratio. When the R in the Lfc in 1 5 was replaced with A, the effects of Lfc in 1 5 on the thermal behaviour of D P P G liposomes diminished (Figures 5.2aii and 5.2bii). Lfc in 1 5 R- caused a slight decrease of the I2850/I2880 ratio at temperatures above 40 °C. The most noticeable effect of Lfc in 1 5 R- occurred between 40 and 50 °C as indicated by the I2935/I2880 ratio. Lfcini 5 R- caused a major decrease of the I2935/I2880 ratio between 40 and 50 °C. In addition, Lfc in 1 5 R- also lowered the I2935/I2880 ratio at temperatures below 10 °C (Figure 5.2bii). The thermal stability of D P P G was increased by the presence of Lfc in 1 5 W- (Figure 5.2aiii). The l 2 85o/l288o ratio exhibited a very gradual increase from 0.6 to 0.8 between 0 and 65 °C. The I2935/I2880 ratio also exhibited a gradual increase from 0.2 to 0.6 between 15 and 40 °C. Effects of Lfcin 1 5W+ and LfcinCore on the I2850/I2880 ratio of D P P G liposomes were very similar. The I2850/I2880 ratio of D P P G fluctuated in a very narrow range in the presence of these two Lfcin derivatives (Figures 5.2aiv and 5.2av). However, Lfcin 1 5W+ and LfcinCore caused different effects on the D P P G liposomes as indicated by the I2935/I2880 ratio (Figures 5.2biv and 5.2bv). Lfcin-|5W+ had no effect on D P P G liposomes below 25 °C and the ratio remained steady at 0.4. The ratio experienced a small shift from 0.4 to 0.8 between 26 and 30 °C and remained stable at 0.8 at higher temperatures. Comparing to the thermal behaviour of D P P G in the absence of Lfcin derivatives, Lfcin 1 5W+ caused a smaller shift at lower temperature (Figure 5.2biv). In contrast, the presence of LfcinCore led to a very steady and significant increase in the I2935/I2880 ratio of D P P G liposomes from 5 to 60 °C (Figure 5.2bv). 5.3.3. Effects of Lactoferricin Derivatives on D P P E Liposomes As indicated by the bssrAsso and the I2935/I2880 ratios, D P P E liposomes exhibited very little response to temperature changes between -10 and 60 °C (open circle symbols in Figures 5.3ai and 5.3bi). Sharp increases in both ratios only occurred at temperatures around 60 °C. The intensity ratios of D P P E were also lower than those of D P P C and D P P G . The addition of Lfc in 1 5 derivatives caused a general upward shift of the bsso/bsso ratio of D P P E . The presence of Lfc in^ or Lfc in 1 5 W- in D P P E led to increases in the bsso/bsso ratio (Figures 5.3ai and 5.3aiii, respectively). The ratios also demonstrated very gradual increases between -10 and 50 °C, and moderate increases at temperatures between 50 and 70 °C. In comparison, Lfc in 1 5 R-had less effect on the bsso/bsso ratio of D P P E below 40 °C, but led to significant increase in the ratios starting at around 42 °C (Figure 5.3aii). LfcinCore had a similar effect on D P P E liposomes and caused a sharp increase in the bsso/bsso ratio at around 38 °C (Figure 5.3av). Among all Lfcin derivatives, Lfcin 1 5W+ was the only exception. Unlike other derivatives that increased the bsso/bsso 105 ratio of D P P E , Lfcin 1 5W+ caused a significant decrease in the ratio over the temperature range tested, with little change as a function of increasing temperature until a sharp increase was observed near 52 °C (Figure 5.3aiv). The I2935/I2880 ratio of D P P E was unaffected by the presence of any Lfcin derivatives in the conditions studied in the present experiment (Figure 5.3b). 5.4. Discussion 5.4.1. Effects of Arginine on Lactoferricin on the Thermal Behaviour of Phospholipid Liposomes The presence of Lfc in 1 5 affected the thermal transition behaviour of D P P C , D P P G , and D P P E . Replacing the positively charged Arg amino acid in Lfc in 1 5 with Ala (Lfcin 1 5R-) reduced the impact of Lfc in 1 5 on D P P C and D P P E . The thermal behaviour of the acyl chains of D P P G was mildly affected by Lfcin 1 5R-, while Lfcin-|5R- caused a higher gel-to-liquid crystalline phase transition temperature at the methyl ends of D P P G . The presence of positively charged amino acids, especially R, has been demonstrated to be a key factor in many cationic antimicrobial peptides including Lfcin (Strom et al., 2001). However, the discussion on the exact role of R in facilitating the antimicrobial function is inconclusive. Results from Chapter 4 showed that Lfcin had the most impact on D P P G possibly due to the interactions between the negatively charged head groups on P G and the positively charged amino acid side chains on Lfcin. Results from the present study showed that when Arg was replaced with Ala, Lfc in 1 5 R- no longer exhibited the same effects exerted by Lfc in 1 5 on D P P G (Figures 5.2H). This indicates that Arg may be important for the anchoring of Lfcin on the polar head group of P G and facilitating further hydrophobic interactions between the acyl chains and hydrophobic amino acids on Lfcin. Lfc in 1 5 R- also decreased the intermolecular chain-chain interactions of D P P E (Figure 5.3ii). Such a result is quite unexpected based on results reported in Chapter 4. However, a closer look at the structural properties (Table 5.1) of Lfc in^R- shows that although the helix propensity of Lfc in 1 5 R-was similar to that of Lfc in 1 5 , the replacement of Arg with Ala minimized the charge similarity of Lfcin 1 5R-, lowered its pi, but also increased its overall hydrophobicity. As discussed in Chapter 4, due to the positive curvature stress, D P P E often exists as hexagonal type II (cone shaped, Hn) and exposes its hydrophobic acyl chains. In the presence of the highly hydrophobic Lfcin 1 5 R-, the acyl chains of D P P E were more readily disturbed, resulting in an increase in the bsso/bsso ratio when at higher temperature. Although the presence of arginine on Lfc in 1 5 seemed to play an important role in its interaction with pathogenic cell membranes, studies on other cationic antimicrobial peptides showed that the actual number of positive charges within the polypeptide chain was less significant. For 106 instance, temporins appeared to exhibit antibiotic action with only a single positively charged residue (Mangoni et al., 2000) and there was an optimal positive charge for the lytic activity of magainin 2 analogues (Blondelle and Houghten, 1992; Kiyota et al., 1996). The decrease in activity above a certain peptide charge was suggested to be the result of instability and reduced lifetime of the pore formed by the peptide chains in the membrane (Matsuzaki et al., 1997). Investigation of synthetic pardaxin in model membranes by solid-state NMR showed a decrease in the ability of the peptide to disrupt lipid bilayers in the presence of acidic phospholipids (Hallock et al., 2002). The results suggested that preferential interactions between the positively charged residues of pardaxin and the anionic lipid headgroups had prevented the peptide from inserting into the acyl chain region of the bilayer and causing morphological changes in the membrane, a phenomenon observed for zwitterionic vesicles. Temporin L (Rinaldi et al., 2002) and melittin (Ladokhin and White, 2001) lost their lytic activities in the presence of negatively charged phospholipids (Rinaldi et al., 2002). Indeed, an oriented C D study revealed that melittin was able to adopt a transmembrane orientation only in zwitterionic bilayers (Ladokhin and White, 2001). In negatively charged membranes, the strong electrostatic interaction between the cationic peptide chain and the lipid headgroups seemed to suppress peptide reorientation that in turn hindered the formation of well-defined pores. In a recent surface plasmon resonance study of melittin, Papo and Shai reached similar conclusions (Papo and Shai, 2003). Results from the present study suggest that the replacement of Arg with Ala on Lfc in 1 5 R- affected how Lfcin binds to negatively charged D P P G liposomes. The shift in the bsso/bsso and I2935/I2880 ratios of D P P G in the presence of Lfc in 1 5 R- likely indicated that the absence of Arg prevented the Lfcin derivative from interacting with the acyl chains as well as the methyl ends of the hydrophobic core of the D P P G bilayers. 5.4.2. Effects of Tryptophan on the Thermal Behaviour of Lactoferricin of Phospholipid Liposomes Tryptophan has been shown to be a key amino acid in the antimicrobial activity of Lfcin (Strom et al., 2001). It was suggested that the hydrophobic amino acid was needed to stabilize the peptide in the hydrophobic bilayer on bacterial membrane. Strom et al. (2000) reported that the addition of Trp to Lfcin from human, caprine, murine, and porcine sources significantly improved the antimicrobial activity of these Lfcins. Haug and Svendsen (2001) demonstrated that the hydrophobic aromatic side chains of Trp were responsible for the antimicrobial activity. Furthermore, they reported that increasing the bulkiness of the side chains would further enhance the antimicrobial activity (Haug and Svendsen, 2001). Results from the present study show that the presence or absence of Trp on Lfc in 1 5 had variable effects on different phospholipid liposomes. The absence of tryptophan (Lfcin 1 5W-) made little difference to the effect of L fc in^ on D P P C (Figure 5.1 iii). Effects of Lfc in 1 5 W- on D P P E were similar to those of Lfc in 1 5 (Figures 5.3iii and 5.3i). In contrast, Lfc in 1 5 W- had a major impact on 107 D P P G , lowering both the bsso/bsso and I2935/I2880 ratios throughout the temperature range tested (Figure 5.2iii). The insertion of two extra tryptophan residues on Lfc in 1 5 (Lfcin 1 5W+) had little effect on Lfcin 1 5 's effect on D P P C . Effects of Lfcini 5W+ on D P P C and D P P G were similar to those of Lfc in 1 5 W- (Figures 5.1 iv and 5.2iv), but Lfcin 1 5W+ decreased the bsso/bsso ratio of D P P E in the temperature range tested (Figure 5.3iv). The importance of Trp on the interaction of Lfcin with D P P G was therefore confirmed by the present study. Results depicted in Figure 5.2iii showed that Lfcin-i5W- indeed increased the degree of order and decreased the formation of gauche conformations of the acyl chains of D P P G at high temperature. The replacement of Trp with Ala would be expected to decrease the bulkiness, and calculated hydropathicity G R A V Y values indicate that Lfcini 5 W- was less hydrophobic than Lfc in 1 5 (Table 5.1). Furthermore, homology similarity analysis of Lfc in 1 5 W- indicated that the helix propensity of Lfcin-15 was altered after the Trp was replaced. The results imply that the presence of Ala and absence of Trp created a peptide that would likely promote greater order of rather than destabilize the D P P G bilayer. Interestingly, results presented here unexpectedly showed that the extra tryptophan residues on Lfcin 1 5W+ had the least effect on the thermal behaviour of D P P E . In fact, the insertion of two additional Trp minimized the disordering effects on D P P E exerted by Lfcin 1 5 , Lfcin 1 5 R-, Lfcini 5W-, and LfcinCore, and Lfcin 1 5W+ appeared to increase the intermolecular interactions and order of the acyl chains. These results seem to disagree with those reported in Chapter 4 and Section 5.4.1 where it was suggested that hydrophobicity and the presence of hydrophobic amino acids of Lfcin were responsible for the decrease in the degree of order of D P P E . However, Lfcin 1 5W+ behaved very differently and promoted the degree of order of D P P E in the present model as shown on Figure 5.3av. The increase in bulkiness of Lfcin 1 5W+ could have contributed to the promotion of the order of the acyl chains of D P P E . Furthermore, the present observations could be a result of the removal of the Arg and Met residues while adding two Trp to make Lfcin 1 5W+. Results presented in Section 5.4.1. showed that the removal of Arg, Lfcin 1 5 R-, had reduced the effect of Lfcin 1 5 . Hence, the presence of the appropriate number of Arg and Trp residues is critical in designing a peptide that would affect the thermal behaviour of D P P E . Increases in hydrophobicity alone without the presence of Arg would indeed minimize the effects of Lfcin on D P P E . Farnaud et al. (2004) also suggested that a balance of the number of tryptophan residues and positively charged amino acids are essential in the activities of C A P s . Tachi et al. (2002) reported that increasing peptide hydrophobicity may alter the leakage mechanism of magainin. Although increasing hydrophobicity generally increased the affinity of magainin peptides for zwitterionic phospholipids and erythrocytes, more hydrophobic analogues appeared to form less stable pores. Furthermore, while magainins are known to promote a positive membrane curvature, large increases in hydrophobicity 108 promoted the penetration of peptide chains into the lipid bilayer and expanded the hydrophobic core region. The increase in hydrophobicity imposed a negative curvature strain on the membrane, affecting pore formation rate, pore size, and pore stability (Tachi et al., 2002). The location and distribution of the Arg and Trp residues along the peptide also play important roles in the final antimicrobial activity of the peptide. Tachi et al. (2002) suggested that the mode of action of magainin is dependent on the position of hydrophobic amino acids in the peptide sequence. Multiple regression analysis of thirteen model peptides suggested that the hydrophobic moment of the peptide chain and the positioning of the hydrophobic amino acids may be the parameters that have the most significant effect on the antimicrobial activity of lytic peptides (Pathak et al., 1995). Rearrangement of the sequence in the hydrophilic C-terminal region of melittin led to increased amphipathicity, which parallels the antimicrobial activity of the peptide (Subbalakshmi et al., 1999). Wieprecht et al. (1997) suggested that Arg residues in C A P s are responsible for the establishment of electrostatic interactions and binding affinity between the C A P s and bacterial membranes. The hydrophobic amino acids are then responsible for the interactions between the peptide chain and the lipid bilayer. This interdependence of both kinds of interactions, electrostatic and hydrophobic, for antimicrobial activity was confirmed by observations made in the present study. 5.4.3. Effects of Peptide Length on the Thermal Behaviour of Lactoferricin of Phospholipid Liposomes As indicated by the bsso/bsso ratio, LfcinCore had similar effects on D P P E as those exerted by Lfc in 1 5 (Figure 5.3v). On the other hand, LfcinCore promoted the degree of order of the acyl chains of D P P G (Figure 5.2v). Jing et al. (2003) reported that a hexapeptide A c - R R W W R F - N H 2 had a large influence on the thermotropic phase behaviour of D P P G and had little effect on model membranes containing the zwitterionic D P P C . This behaviour is consistent with its biological activity and with its affinity to these membranes as determined by titration calorimetry, implying that peptide-lipid interactions play an important role in this process. Jing et al. (2003) suggested that the amphipathic structure may allow this peptide to penetrate deeper into the interfacial region of negatively charged membranes, leading to local membrane destabilization. However, Hunter et al. (2004) showed that fragments from human (RAWVARW-NH 2 ) and chicken ( IVSDGNGMNAWVAWR-NH 2 ) lysozyme bound strongly to the interface of a bacterial membrane and did not interact with the hydrophobic acyl chains in the phospholipid bilayer. R A W V A R W - N H 2 is too short to span the membrane and form channels. Hunter et al. (2004) proposed that its antimicrobial activity arises from the ability to either immediately alter the membrane characteristics and/or traverse the membrane spontaneously and act on intracellular • targets. 109 LfcinCore is too short to span the phospholipid bilayer; hence, its action on D P P G phospholipid bilayer is limited to the region near the negatively charged head groups on the bilayer interface. This indicated that changes on the bilayer surface could affect the packing of the acyl chains in the hydrophobic core. In the present study, LfcinCore stabilized the D P P G acyl chains especially at temperatures above the transition temperature. In fact, the stabilization was so strong that the conversion of a lamellar gel phase (LB) to a liquid-crystal (L a) phase was not detected in the temperature range investigated (Figure 5.2v). Knowledge about the importance of electrostatic interactions of Arg and the role of Trp residues as a membrane interface anchor will provide insight into the future design of potent antimicrobial peptidomimetics. 5.4.4. Proposed Antimicrobial Actions Results from the present study suggest a few key characteristics of the actions of Lfcin and its derivatives on phospholipid bilayers. In general, Lfcin and its derivatives exerted less effect on D P P C , a major phospholipid in mammalian cells, than D P P G and D P P E , major phospholipids in bacterial membranes. Results presented here are consistent with other observations reported in the literature. For example, as monitored by differential scanning calorimetry (DSC), the thermal transition behaviours of D P P C and D P P E were not affected by the addition of Lfcin at high concentration. In contrast, a broad pre-transition around 32 °C disappears when Lfcin was added to D P P G . When more Lfcin is added, the peak for the main transition also shifts to a slightly higher temperature (Vogel et al. 2002). These results show that Lfcin interacts more strongly with the overall negatively charged P G headgroup than the zwitterionic P C or P E headgroups (Liu and Deber, 1997). Fluorescence spectroscopy study of lactoferricin B and some of its shorter fragments have shown that Lfcin and its derivatives bind more tightly and penetrate somewhat deeper into vesicles containing 40% negatively charged P G headgroups (Schibli et al., 1999). This also explains why cationic antimicrobial peptides such as Lfcin typically prefer to bind to bacterial membranes rather than to those of the host. In the present study, changes in the amino acid composition of L fc in 1 5 had little effect on how Lfcin and its derivative affected D P P E , with the exception of Lfcin 1 5W+ (Figure 5.3). The replacement of Arg (Lfcin 1 5R-) and Trp (Lfcin 1 5W-) with Ala had little effect on how Lfc in 1 5 interacted with D P P E (Figure 5.3ii and iii). However, the addition of two Trp residuess with the removal of an Arg and a Met in Lfcin 1 5W+ minimized the effects observed in other Lfcin derivatives (Figure 5.3iv). This indicated that the presence of both Arg and Trp is important in the action of Lfcin. Furthermore, the presence of two bulky Trp might in fact hinder the electrostatic interactions between Arg and the negatively charged phospholipid head groups on the membrane interface, weakening the anchoring of Lfcin 1 5W+ on D P P E membrane and minimizing the effect of Lfcin derivatives on D P P E . 110 Interactions between D P P G and Lfcin derivatives observed in the present study provided a wealth of information towards the understanding of how Lfcin interacts with D P P G . The replacement of Arg with Ala in Lfcin-|5R- did not affect how Lfc in 1 5 interacted with D P P G (Figure 5.2ii). The absence of Trp, Lfcin 1 5W-, altered the thermal behaviour of D P P G and increased its degree of order throughout the temperature range investigated (Figure 5.2iii). On the contrary, the addition of Trp with simultaneous removal of one Arg, Lfcin 1 5W+, showed the same effects on D P P G (Figure 5.2iv). Acyl chains of D P P G became more ordered in the presence of LfcinCore (Figure 5.2v). Removing Trp from Lfcin 1 5 , Lfcin 1 5 W-, stabilized the D P P G , indicating that the presence of Trp was important to destabilize the D P P G (Figure 5.2iii). Nevertheless, the presence of too many Trp without the appropriate support from Arg would not destabilize D P P G (Figure 5.2iv). This indicates that the presence of too many bulky Trp could hinder the electrostatic interactions between the cationic amino acids and the negatively charged membrane, minimizing the abaility of Lfcin to disrupt D P P G . The importance of the electrostatic interaction on the surface of the phospholipid bilayer is further supported by the actions of LfcinCore (Figure 5.2v). Since LfcinCore is too short to span across the phospholipid bilayer, the increases in the degree of order of the D P P G acyl chains in the hydrophobic core were likely the result of the action of the LfcinCore near the negatively charged headgroups. All of these pieces of evidence show the interdependence of the electrostatic and hydrophobic groups in the interactions of Lfcin with phospholipid. The presence of both electrostatic and hydrophobic interactions is needed in the design of antimicrobial peptides. Although the presence of both electrostatic and hydrophobic actions are essential in the disruption of the degree of order of D P P G acyl chains, it is important to understand that increasing the number of Arg and Trp alone will not lead to the creation of a more "powerful" antimicrobial peptide. For instance, the addition of Trp in Lfcin 1 5W+ indeed led to the stabilization of D P P G (Figure 5.2iv). On the contrary, the removal of Arg in Lfc in 1 5 R- had no effect on the behaviour of Lfc in 1 5 on D P P G . In the present study, Lfc in 1 5 R- was designed by replacing Arg with Ala, creating a weak cationic peptide (pi = 9.79) with strong hydrophobicity (GRAVY = 0.053) (Table 5.1). Results from these two Lfcin derivatives indicate that the design of an antimicrobial peptide requires a delicate balance of the number of cationic and hydrophobic amino acids in the peptide. Other structural properties of the peptides are important to their antimicrobial functions. Results presented in Chapter 4 indicated that the helix propensity and charge property in selected segments of Lfcin derivatives were particularly important for the prediction of their antimicrobial property against E. coli. H S A analysis of the Lfcin derivatives used in the present study showed that the helix propensity of Lfc in 1 5 was altered when Trp was replaced by Ala (Lfcin^W-). Charge property of Lfc in 1 5 was also changed when Arg was replaced by Ala (Lfcin 1 5R-) (Table 5.1). These 111 changes might be responsible for the effects of Lfcin derivatives on model bacterial membranes observed in the present study. Other structural properties have also been found to be important for the bioactivity of peptides. For instance, the presence of a (3-fold is more essential than the number of cationic charges for protegrin-1 (Mani et al., 2005) and polyphemusin I (Powers et al., 2004). The importance of the (3-hairpin fold may result from the fact that it places the hydrophobic and charged residues in these peptides in an amphipathic fashion, in order to promote the insertion of the peptides into the lipid bilayers without permeabilizing the membrane. The number of charges is a secondary factor that regulates the strength of the interaction between the anionic membrane and the cationic peptides (Mani et al., 2005). Raman spectra of phospholipid liposomes with and without the presence of Lfcin derivatives observed in the present study indicated that Lfcin and its derivatives exerted less effect on D P P C , a major phospholipid in mammalian cells, than on D P P G and D P P E , major phospholipids in bacterial cells. Although the presence of both electrostatic and hydrophobic interactions is needed to establish interactions between cationic antimicrobial peptides and phospholipid membranes, the presence of positively charged and hydrophobic amino acids is not the only necessary characteristic in determining the actions of Lfcin and its derivatives on D P P G and D P P E . Other structural properties, such as the total number of cationic and hydrophobic amino acids, their overall location and distribution along the peptides, and secondary structures are also important in determining the overall antimicrobial properties of cationic antimicrobial peptides. 112 1.6 1.4 o 1-2 CO J ? 1.0 § 0.8 0.6 0.4 ( a ) I2850A288O • IT • _cn -20 0 20 40 60 80 Temperature (°C) 1.6 1.4 1.2 1.0 0.8 0.6 0.4 • 0 T , 1.6 1.4 1.2 I 1.0 CM CO - 2 " 0.6 0.4 1.6 1.4 3 1 2 CO - a 1.0 in CO _£< 0.8 0.6 .0.4 -20 0 20 40 60 80 .Temperature (°C) -20 0 20 40 60 80 Temperature (°C) [X DDD>< £ X X X >x - a -20 0 20 40 60 80 Temperature (°C) 1.6 0 1-4 CO J ? 1.2 1 1 0 CM 0.8 0.6 0.4 • X X » X > x * x x x x ' a x • - a -20 0 20 . 40 60 80 Temperature (°C) 1.0 0.8 0.6 0.4 0.2 0.0 1.0 0.8 0.6 0.4 0.2 0.0 1.0 0.8 0.6 0.4 0.2 0.0 1.0 0.8 0.6 0.4 0.2 0.0 1.0 0.8 0.6 0.4 0.2 0.0 (b) I2935/I288O •D 1 1 1 1 1 20 0 20 40 60 80 Temperature (°C) -20 0 20 40 60 80 Temperature (°C) -20 0 20 40 60 80 Temperature (°C) X x x x n -20 0 20 40 60 80 Temperature (°C) n • ^ i r * x * ° x x % * g ^ -20 0 20 40 60 80 Temperature (°C) Figure 5.1. Thermal behaviours of D P P C as indicated by bsso/bsso (a) and I2935/I2880 (b) in the absence of lactoferricin (Lfcin, • ) , the presence of Lfc in 1 5 (i, • ) , Lfc in 1 5 R- (ii, • ) , Lfc in 1 5 W- (iii, A ) , Lfcin 1 5W+ (iv, X), and LfcinCore (v, * ) measured between -10 °C and 70 °C by Raman spectroscopy. 113 (i) (ii) (iii) (iv) (v) 1.8 ! 1.6 1.4 1.2 1.0 0.8 0.6 0.4 (a) '2850'I2880 -20 20 40 60 80 Temperature (°C) 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 -20 0 20 40 60 Temperature (°C) 80 18 1.6 1.4 1.2 1.0 0.8 0.6 0.4 o ^ > o ° o o A * v O O O A * A A A A ± A \ A A -20 0 20 40 60 Temperature (°C) 80 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 O o ° o o ^ X ^ X ^ X X X 3oo x x 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 -20 0 20 40 60 Temperature (°C) o 80 x X o oox g> o<>Oo -20 0 20 40 60 Temperature (°C) 80 1.2 1.0 0.8 0.6 0.4 0.2 0.0 (b) 12935^ 12880 O -20 0 20 40 60 80 Temperature (°C) 1.2 1.0 0.8 0.6 0.4 0.2 0.0 1.2 1.0 0.8 0.6 0.4 0.2 0.0 1.2 1.0 0.8 0.6 0.4 0.2 0.0 1.2 1.0 0.8 0.6 0.4 0.2 0.0 O O -20 0 20 40 60 80 Temperature (°C) A A A A V * ^ -20 0 20 40 60 80 Temperature (°C) o x x x ^ x / \ > x ° > ^ x x X x 20 0 20 40 60 80 Temperature (°C) x 0 < » x X x X x x ^ * * x ^ S * x * . ^ x ,<r x $ x ^ ° -20 0 . 20 40 60 80 Temperature (°C) Figure 5.2. Thermal behaviours of D P P G as indicated by I2850/I2880 (a) and I2935/I2880 (b) in the absence of lactoferricin (Lfcin, O), the presence of Lfc in 1 5 (i, • ) , Lfc in 1 5 R- (ii, • ) , Lfc in 1 5 W- (iii, A ) , Lfcin 1 5W+ (iv, X), and LfcinCore (v, * ) measured between -10 °C and 70 °C by Raman spectroscopy. 114 (i) (ii) (iii) (iv) (v) (a) I2850/1 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0'I2880 , g l 0 0 0 d&Z^o -20 20 40 60 80 Temperature (°C) 1.6 1.4 1.2 1.0 H 0.8 0.6 0.4 0.2 o -20 20 40 60 80 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 A A ooo TemDerature rCI o oC®3cjp o o o -20 0 20 40 60 80 Temperature (°C) 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 -20 0 20 40 60 80 Temperature (°C) 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 o X X X s^tX o x x x x ^ M ^ ^ s ^ P 0 0 0 0 cfi&Zt^o o (b) 12935/12880 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 • I S ' -20 0 20 40 60 80 Temperature (°C) 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 -20 0 20 40 60 80 TemDerature (°C1 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 -20 0 20 40 , 60 80 Temperature (°C) 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 o, o X 1 1 1 1 1 20 0 20 40 60 80 Temperature (°C) xo * x ^x?xo -20 20 40 60 80 -20 0 20 40 60 80 Temperature (°C) Temperature (°C) Figure 5.3. Thermal behaviours of D P P E as indicated by I2850 / I2880 (a) and I2935/I2880 (b) in the absence of lactoferricin (Lfcin, O), the presence of Lfc in 1 5 (i, • ) , Lfc in 1 5 R- (ii, • ) , Lfc in 1 5 W- (iii, • ) , Lfcin 1 5W+ (iv, X), and LfcinCore (v, * ) measured between -10 °C and 70 °C by Raman spectroscopy. 115 Table 5.1. Structural properties of lactoferricin derivatives used in the study. Code M W 1 Pi 1 G R A V Y1 ' 2 Helix Similarity 3 Charge Similarity 3 FKCRR WQWRM KKLGA Lfc in 1 5 1994.46 11.74 -1.207 1.000 1.00 FKCAA WQWAM KKLGA' Lfcin 1 5 R- 1739.13 9.79 0.053 0.883 0.00 FKCRR AQARM KKLGA Lfcin 1 5 W- 1764.19 11.74 -0.847 -0.948 1.00 FKCWR WQWRW KKLGA Lfcin 1 5W+ 2079.50 11.11 -1.153 0.463 0.73 RR WQWR LfcinCore 987.13 12.30 -3.133 ProtParam Tool: http://ca.expasy.org/tools/protparam.html G R A V Y : Grand average of hydropathicity index: positive G R A V Y (hydrophobic), negative G R A V Y (hydrophilic). Homology Similarity Analysis: Helix and charge similarity of residues 4 to 10. L fc in^ was used as the reference peptide. 116 5.5. References Bellamy, W.; Takase, M.; Yamauchi, K.; Wakabayashi, H.; Kawase, K.; Tomita, M. Identification of the bactericidal domain of lactoferrin. Biochim. Biophys. Acta 1992, 1121, 130-136. Bellamy, W.; Wakabayashi, H.; Takase, M.; Kawase, K.; Shimamura, S.; Tomita, M. Killing of Candida albicans by lactoferricin B, a potent antimicrobial peptide derived from the N-terminal region of bovine lactoferrin. Med. Microbiol. Immunol. 1993, 182, 97-105. Blondelle, S. E.; Houghten, R. A. Design of model amphipathic peptides having potent antimicrobial activities. Biochemistry 1992, 31, 12688-12694. Blondelle, S. E.; Takahashi, E.; Dinh, K. T.; Houghten, R.A. The antimicrobial activity of hexapeptides derived from synthetic combinational libraries. J. Appl. Bacteriol. 1995, 78, 39-46. Blondelle, S. E.; Houghten, R. A. Novel antimicrobial compounds identified using synthetic combinatorial library technology. Trends Biotechnol. 1996, 74,60-65. Blondelle, S . E.; Lohner, K. Combinatorial librarieis: A tool to design antimicrobial and antifungal peptide analogues having lytic specificities for structure-activity relationship studies. Pept. Sci. 2000, 55, 74-87. Devlin, M. T.; Levin, I. W. Raman spectroscopic studies of the packing properties of mixed dihexadecyl- and dipalmitoylphophatidylcholine bilayer dispersion. Biochemistry 1989, 28, 8912-8920. Farnaud, S.; Spiller, C ; Moriarty, L. C ; Patel, A.; Gant, V.; Odell, E. W.; Evans, R. W. Interactions of lactoferricin-derived peptides with L P S and antimicrobial activity. FEMS Microbiol. Lett. 2004,233,193-199. Hallock, K. J.;Lee, D. K.; Omnaas, J . ; Mosberg, H. I.; Ramamoorthy, A. Membrane composition determines pardaxin's mechanism of lipid bilayer disruption. Biophys. J. 2002, 83, 1004-1013. Haug, B. E.; Svendsen, J . S . The role of tryptophan in the antibacterial activity of a 15-residue bovine lactoferricin peptide. J. Pept. Sci. 2001, 7, 190-196. Huang, C.-H. ; Lapides, J . R.; Levin, I. W. Phase-transition behaviour of saturated, symmetric chain phospholipid bilayer dispersions determined by Raman spectroscopy: Correlation between spectral and thermodynamic parameters. J. Am. Chem. Soc. 1982, 704, 5926-5930. Hunter, H. N.; Jing, W.; Schibli, D. J . ; Trinh, T.; Park, I. Y.; Kim, S. C ; Vogel, H. J . The interactions of antimicrobial peptides derived from lysozyme with model membrane systems. Biochim. Biophys. Acta 2004, 7688, 175-189. Jing, W.; Hunter, H. N.; Hagel, J . ; Vogel, H. J . The structure of the antimicrobial peptide Ac-R R W W R F - N H 2 bound to micelles and its interactions with phospholipid bilayers. J. Pept. Res. 2003, 61, 219. Kang, J . H.; Lee, M. K.; Kim, K. L.; Hahm, K.-S. Structure-biological activity relationships of 11-residue highly basic peptide segment of bovine lactoferrin. Int. J. Pept. Protein Res. 1996, 48, 357-363. 117 Kiyota, T.; Lee, S. ; Sugihara, G . Design and synthesis of amphiphilic a-helical model peptides with systematically varied hydrophobic-hydrophilic balance and their interaction with lipid- and bio-membranes. Biochemistry 1996, 35, 13196-13204. Kyte, J . ; Doolittle, R. F. A simple method for displaying the hydropathic character of a protein. J. Mol. Biol. 1982, 757, 105-32. Ladokhin, A. S.; White, S. H. 'Detergent-like' permeabilization of anionic lipid vesicles by melittin. Biochim. Biophys. Acta 2001, 7574, 253-260. Laroche, G. ; Carrier, D.; Pezolet, M. Study of the effect of poly(L-lysine) on phosphatidic acid and phosphatidylcholine/phosphatidic acid bilayers by Raman spectroscopy. Biochemistry 1988, 27, 6220-6228. Laroche, G. ; Dufourc, E. J . ; Pezolet, M.; Dufourcq, J . Coupled changes between lipid order and polypeptide conformation at the membrane surface. A 2H NMR and Raman study of polylysine-phosphatidic acid systems. Biochemistry 1990, 29, 6460-6465. Lejon, T.; Strom, M. B.; Svendsen, J . S. Is information about peptide sequence necessary in multivariate analysis? Chemometrics Intelligent Lab. Sys. 2001, 57, 93-95. Liu, L. P.; Deber, C. M. Anionic phospholipids modulate peptide insertion into membranes. Biochemistry 1997, 36, 5476-5482. Mangoni, M. L.; Rinaldi, A. C ; Di Giulio, A.; Mignogna, G. ; Bozzi, A.; Barra, D.; Simmaco, M. Structure-function relationships of temporins, small antimicrobial peptides from amphibian skin. Eur. J. Biochem. 2000, 267, 1447-1454. Mani, R.; Waring, A. J . ; Lehrer, R. I.; Hong, M. Membrane-disruptive abilities of 6-hairpin antimicrobial peptides correlate with conformation and activity: A 31P and 1H NMR study. Biochim. Biophys. Acta 2005, 7777, 11-18. Matsuzaki, K.; Nakamura, A.; Murase, O.; Sugishita, K.-l.; Fujii, N.; Miyajima, K. Modulation of magainin 2-lipid bilayer interactions by peptide charge. Biochemistry 1997, 36, 2104-2111. Opekarova, M.; Tanner, W. Specific lipid requirements of membrane proteins—a putative bottleneck in heterologous expression. Biochim. Biophys. Acta 2003, 7670, 11-22. Papo, N.; Shai, Y. Exploring peptide membrane interaction using surface plasmon resonance: differentiation between pore formation versus membrane disruption by lytic peptides. Biochemistry 2003, 42, 458-466. Pathak, N.; Salas-Auvert, R.; Ruche, G. ; Janna, M. H.; McCarthy, D.; Harrison, R. G. Comparison of the effects of hydrophobicity, amphiphilicity, and alpha-helicity on the activities of antimicrobial peptides. Proteins 1995, 22, 182-186. Powers, J . -P . S. ; Rozek, A.; Hancock, R. E. W. Structure-activity relationships for the (3-hairpin cationic antimicrobial peptide polyphemusin I. Biochim. Biophys. Acta 2004, 7698, 239-250. Rekdal, 0 . ; Andersen, J . ; Vorland, L. H.; Svendsen, J . S. Construction and synthesis of lactoferricin derivatives with enhanced antibacterial activity. J. Pept. Sci. 1999, 5, 32-45. Rinaldi, A. C ; Mangoni, M. L.; Rufo, A.; Luzi, C ; Barra, D.; Zhao, H.; Kinnunen, P. K.; Bozzi , A.; Di Giulio, A.; Simmaco, M. Temporin L: antimicrobial, haemolytic and cytotoxic activities, and effects on membrane permeabilization in lipid vesicles. Biochem. J. 2002, 368, 91-100. 118 Schibli, D. J . ; Hwang, P. M.; Vogel, H. J . The structure of the antimicrobial active center of lactoferricin B bound to sodium dodecyl sulfate micelles. FEBS Lett. 1999, 446, 213-217. Sorensen, M.; Sorensen, S. P. L. The proteins in whey. Compt. Rend. Lab. Carlsberg 1939, 23, 55-99. Strom, M. B.; Rekal, 0 . ; Svendsen, J . S . Antibacterial activity of 15-residue lactoferricin derivatives. J. Pept. Res. 2000, 56, 265-274. Strom, M. B.; Rekdal, 0 . ; Stensen, W.; Svendsen, J . S . Increased antibacterial activity of 15-residue murine lactoferricin derivatives. J. Pept. Res. 2001, 57, 127-139. Subbalakshmi, C ; Nagaraj, R.; Sitaram, N. Biological activities of C-terminal 15-residue synthetic fragment of melittin: design of an analog with improved antibacterial activity. FEBS Lett. 1999, 448, 62-66. Tachi, T.; Epand, R. F.; Epand, R. M.; Matsuzaki, K. Position-dependent hydrophobicity of the antimicrobial magainin peptide affects the mode of peptide-lipid interactions and selective toxicity. Biochemistry 2002, 41, 10723-10731. Tomita, M.; Takase, M.; Bellamy, W.; Shimamura, S. A review: The active peptide of lactoferrin. Acta Paediatr. Jpn. 1994, 36, 585-591. Vogel, H. J . ; Schibli, D. J . ; Jing, W.; Lohmeier-Vogel, E. M.; Epand, R. F.; Epand, R. M. Towards a structure-function analysis of bovine lactoferricin and related tryptophan- and arginine-containing peptides. Biochem. Cell Biol. 2002, 80, 49-63. Wieprecht, T.; Dathe, M.; Beyermann, M.; Krause, E.; Maloy, W. L ; MacDonald, D. L ; Bienert, M. Peptide hydrophobicity controls the activity and selectivity of magainin 2 amide in interaction with membranes. Biochemistry 1997,36,6124-6132. 119 CHAPTER 6. CONCLUSION 6.1. Review of Results as Related to Working Hypotheses The fractionation procedure described in Chapter 2 resulted in the production of two peptides with masses of 3196 and 3124 Da from peptic digest of food-grade lactoferrin, which corresponded to the 26- and 25-amino acid peptides, lactoferricin (Lfcin), F K C R R W Q W R M K K L G A PSITC V R R A F (A). An industrial grade cation exchange resin with stepwise salt gradient elution was used in the procedure. A fraction eluted using 2.0 M NaCI contained predominantly two peptides containing the Lfcin sequence. Based on the competitive displacement concept, a 2-step process using industrial grade cation exchange resin led to 35% recovery of Lfcin and also produced other cationic peptides. Putative sequences of cationic peptides in other eluted fractions included F K N K S R S F Q , W R M K K L G A P S ITCVR RA, and GAPSI T C V R R A F A L E CIRAI A E K K A . Iron saturation level of LF had no effect on the production of Lfcin. Nevertheless, the digestion of LF containing lower iron content led to the production of a higher quantity of low molecular weight cationic peptides. Homology sequence analysis (HSA) was performed on Lfcin and 71 Lfcin derivatives followed by multivariate data analysis using techniques including principal component similarity and artificial neural network to examine structure-function relationship of Lfcin and its derivatives. Results presented in Chapter 3 showed that helical property of position 4 to 9 in the Lfcin sequence was the most important in determining the antimicrobial activity of Lfcin against E. coli, followed by cationic charge property at positions 4 to 9 and 1 to 3. Raman spectra presented in Chapter 4 showed that the effects of Lfcin on phospholipid liposomes depend on the saturation of the acyl chain and the properties of the hand groups. Lfcin had little effect on P O P C and P O P E , i.e., phospholipids with unsaturated acyl chains and zwitterion head groups. On the contrary, Lfcin lowered the lateral chain-chain interaction along the acyl chains of P O P G with negatively charged head groups. The presence of Lfcin in D P P C liposomes restricted the lateral chain-chain interaction along the acyl chain throughout the temperature range examined, and shifted the main transition temperature from 38 °C (in the absence of Lfcin) to 40 °C. Lfcin had little effect on D P P G liposomes at temperatures below the transition temperature, but increased the order of the acyl chain terminal methyl group in the hydrophobic core of the phospholipid bilayer at temperatures above the transition temperature (~ 40 °C). In contrast, mobility of the acyl chains in D P P E was increased in the presence of Lfcin at temperatures below 32 °C. Based on results obtained in Chapters 3 and 4 and data reported in the literature (Farnaud et al., 2004; Strom et al., 2000 and 2001), synthetic Lfcin 15-mer derivatives with special features were designed for investigation of their interactions with phospholipid liposomes. Results in Chapter 5 120 indicated that the removal of cationic R residues from Lfc in 1 5 led to a peptide with low charge homology similarity and decreased its effects on negatively charged D P P G , a phospholipid that makes up 20 to 25 % of the phospholipids in E. coli. This is in agreement with results reported in Chapter 3 that the charge properties in positions 1 to 3 and 4 to 9 were important in determining Lfcin derivatives' antimicrobial activity against E. coli. The removal of hydrophobic Trp residues from Lfcin-15 increased the degree of order of D P P G . This indicated that the presence of Trp in L fc in 1 5 was needed for the disruption of D P P G . Moreover, as determined by HSA, replacing Trp by Ala in L fc in 1 5 caused a major change in the helix propensity of the peptide. Changes in the helical structure of Lfcin derivatives may have affected how Lfcin affected D P P G liposomes. D P P E , a highly ordered phospholipid, was disturbed in the presence of Lfcin and its derivatives. Nevertheless, the addition of Trp and removal of Arg from Lfc in^ minimized the effects of Lfc in 1 5 on the thermal behaviours of D P P E . 6.2. Strengths and Significance of Results to the Field of Study Lactoferricin is a peptide with antimicrobial activity against a wide range of microorganisms. It is also one of the very few antimicrobial peptides originating from a human food source (Chan and Li-Chan, 2006). The fractionation strategy presented in Chapter 2 can be easily adapted for the large-scale production of Lfcin, hence facilitating the usage of Lfcin for human health. Homology similarity analysis and principal component similarity analysis in conjunction with artificial neural network provided an effective strategy to predict the antimicrobial activity of peptides of interest. This approach allows for a rapid screening of bioactive peptides without the expensive costs associated with conducting bioassays. Results from Chapters 4 and 5 of this thesis demonstrated the effects of Lfcin and its derivatives on model phospholipid liposomes and allowed further understanding of the interaction between antimicrobial peptides and phospholipid membranes. Although Lfcin has been the subject of intense research in the last two decades, very few studies have been conducted to elucidate the molecular mechanism of Lfcin on its targeted microorganisms. The use of Raman spectroscopy provides an insight on how Lfcin may interact with phospholipid bilayers during its entry to bacterial cell. Such understanding could facilitate the design of Lfcin derivatives or other C A P s that exhibit effective antimicrobial action, without the undesirable interactions with mammalian membranes that can lead to hemolysis (Kang et al., 1996). 6.3. Recommendations for Future Studies To further expand the potential applications of food-grade bovine lactoferrin, it is recommended that a more detailed analysis of the peptides produced during peptic digestion is 121 needed. Since the identification of Lfcin as an antimicrobial peptide by Tomita and colleagues in 1991 (Tomita et al., 1991), Lfcin has been the centre of research attention in the past decade. Numerous studies have been conducted to examine its bioactivity (e.g. Gifford et al., 2005 and Mader et al., 2005). Results presented in Chapter 2 indicated that many cationic peptides were produced during the purification of Lfcin. It is highly possible that some of these peptides would carry health promoting properties. A systematic procedure should be developed to screen for their bioactivities and allow for the production of additional value-added products from food-grade bovine lactoferrin. Results from the present study showed that Lfcin, an amphipathic cationic peptide, was particularly effective against P G , an anionic phospholipid. A better understanding of the mutual dependence of these parameters will help to elucidate the molecular mechanism of membrane damage by antimicrobial peptides and their target membrane specificity. Future studies should also include the use of liposomes composed of phospholipids extracted from bacteria of interest. Such studies would allow direct comparison of results from bioassays and those from spectroscopic studies and facilitate the understanding of peptide-phospholipid interactions. Interactions between phospholipids of epithelial cells along the digestive tract and Lfcin should be investigated to understand how Lfcin affect the protective properties of phospholipids in human body. Raman spectroscopic analysis of phospholipids also facilitated the development of using liposomes as a delivering medium for antimicrobial peptides, an essential key for the rational design of novel peptide antibiotics. 122 6.4. References Chan, J . C. K.; Li-Chan, E. C. Y. Antimicrobial Peptides. In Nutraceutical Proteins and Peptides in Health and Disease. (Eds. Y. Mine and F. Shahidi.) 2006. Boca Raton: CRC/Taylor and Francis, pp. 99-136. Farnaud, S. ; Spiller, C ; Moriarty, L. C ; Patel, A.; Gant, V.; Odell, E. W.; Evans, R. W. Interactions of lactoferricin-derived peptides with L P S and antimicrobial activity. FEMS Microbiol. Lett. 2004, 233, 193-199. Gifford, J . L.; Hunter, H. N.; Vogel, H. J . Lactoferricin: a lactoferrin-derived peptide with antimicrobial, antiviral, antitumor and immunological properties. Cell Mol. Life Sic. 2005, 62, 2588-2598. Kang, J . H.; Lee, M. K.; Kim, K. L.; Hahm, K.-S. Structure-biological activity relationships of 11-residue highly basic peptide segment of bovine lactoferrin. Int. J. Pept. Protein Res. 1996, 48, 357-363. Mader, J . S.; Salsman, J . ; Conrad, D. M.; Hoskin, D. W. Bovine lactoferricin selectively induces apoptosis in human leukemia and carcinoma cell lines. Mol. Cancer Ther. 2005, 4, 612-624. Strom, M. B.; Rekal, 0 . ; Svendsen, J . S. Antibacterial activity of 15-residue lactoferricin derivatives. J. Pept. Res. 2000, 56, 265-274. Strom, M. B.; Rekdal, 0 . ; Stensen, W.; Svendsen, J . S . Increased antibacterial activity of 15-residue murine lactoferricin derivatives. J. Pept. Res. 2001, 57, 127-139. Tomita, M.; Bellamy, W.; Takase, M.; Yamauchi, K.; Wakabayashi, H.; Kawase, K. Potent antibacterial peptides generated by pepsin digestion of bovine lactoferrin. J. Dairy Sci. 1991, 74,4137-4142. 123 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0092929/manifest

Comment

Related Items