1 STRUCTURAL ANALYSIS OF THE C TERMINUS OF LUNG SURFACTANT PROTEIN B (SP B) By ANNA KUZNETSOVA A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2013
2 201 3 Anna Kuznetsova
3 To Alexanders my dear husband and my lovely son
4 ACKNOWLEDGMENTS I am very thankful to my advisors Prof. Joanna Long and Prof. Gail Fanucci for their supp ort during my doctorate studying at the University of Florida. With their research guidance and invaluable help I was able to grow on both professional and personal levels. I would like to express my gratitude to Prof. Arthur Edison and Dr. James Rocca for teaching me solution NMR techniques and data processing I would also like to thank other members of my doctorate committee Prof. Alexander Angerhofer and Prof. Nicole Horenstein for their contribution to my research. I would also like to express my great est appreciation to my colleagues from the research groups of Dr. Long and Dr. Fanucci for their i ntensive discussions, exchange of ideas. J oint research with my fellow graduate students from these two labs helped me significantly to achieve my goals. Also I would like to thank my family and friends, who supported and encouraged me in all of my beginnings. I also thank Dr. Charles S chwieters, Dr. Andrew Nieuwkoop and Dr. Donghua Zhou for helpful discussions and providing assistance with structure calculati ons. The NMR instrumentation provided by the National Science Foundation's National High Magnetic Field Laboratory is gratefully acknowledged as well as s upport from the University of Florida and Gates Foundation funds The assistance of Dr. Alfred Chung i n peptide synthesis is also gratefully acknowledged.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 8 LIST OF ABBREVIATIONS ................................ ................................ ........................... 12 ABSTRACT ................................ ................................ ................................ ................... 15 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 17 Pulmonary Surfactant as a Surface Active Agent in Lungs ................................ ..... 17 Pulmonary Surfactant in Alveoli ................................ ................................ .............. 17 Pulmonary Surfactant Cycle ................................ ................................ ............. 17 Pulmonary Surfactant Chemical Composition ................................ .................. 18 Lipid Trafficking by Lipid Associated Lung Surfactant Proteins ........................ 20 Clinical Aspects of Pulmonary Surfactant Deficiency ................................ .............. 21 Pulmonary Surfactant Replacement ................................ ................................ ....... 22 2 METHODS FOR STUDYING LIPID DYNAMICS AND MEMBRANE ACTIVE PEPTIDE STRUCTURE IN MEMBRANE MIMETIC ENVIRONMENTS ................. 29 Circular Dichroism (CD) ................................ ................................ .......................... 29 Solid State NMR Spectroscopy ................................ ................................ ............... 32 NMR of lipid bilayers ................................ ................................ ........................ 33 31 P Chemical Shift Anisotropy ................................ ................................ .......... 34 2 H Quadrupolar Coupling ................................ ................................ ................. 39 DePaking ................................ ................................ ................................ .......... 44 Solution State NMR Method of Protein Structure Determination ............................ 47 Protein Resonance Assignment ................................ ................................ ....... 48 Restraint Generation for Protein Structure Calculation ................................ ..... 53 Protein Structure Determination by Simulated Annealing Approach ................ 54 Protein Structure Validation ................................ ................................ .............. 58 Methods of Peptide Production ................................ ................................ ............... 59 Solid Phase Peptide Synthesis ................................ ................................ ......... 59 Heterologous Peptide Expression in E.Coli ................................ ...................... 61 Peptide Purification by High Performance Liquid Chromatography .................. 64 3 INTERACTION OF THE C TERMINUS OF SP B WITH LIPID BILAYERS MIMICKING LUNG IS MODULATED BY ACYL CHAIN SATURATION .................. 84
6 Materials and Methods ................................ ................................ ............................ 85 Synthesis of SP B 59 80 ................................ ................................ ...................... 85 Heterologous Expression of a SPB 59 80 Double Mutant (SP B 59 80 ) .................. 85 Preparation of Peptide:Lipid Samples ................................ .............................. 86 CD Experiments ................................ ................................ ............................... 86 Solid state NMR Analysis ................................ ................................ ................. 87 Results ................................ ................................ ................................ .................... 88 Secondary Structure ................................ ................................ ......................... 88 The Interaction of SP B 59 80 with Lipid Headgroups ................................ .......... 88 2 H NMR Studies of DPPC:POPG Lipid Miscibility on Addition of SP B 59 80 ...... 90 Effects of SP B 59 80 on Lipid Acyl Chains ................................ .......................... 92 Discussion ................................ ................................ ................................ .............. 98 4 SOLUTION NMR STRUCTURE OF THE C TERMINUS OF SP B IN METHANOL AND IN DPC MICELLES ................................ ................................ .. 117 Materials and Methods ................................ ................................ .......................... 118 Synthesis of SP B 59 80 ................................ ................................ .................... 118 Heterologous Expression of 15 N labeled SP B 59 80 Double Mutant (SP B 59 80 ) ................................ ................................ ................................ ................ 118 Circular Dichroism (CD) experiments ................................ ............................. 119 NMR samples preparation ................................ ................................ .............. 120 NMR experiments ................................ ................................ ........................... 120 NMR data analysis and structure calculation ................................ .................. 122 Results ................................ ................................ ................................ .................. 123 Secondary structure in methanol and in DPC micelles ................................ ... 123 NMR Resonance Assignments for SP B 59 80 in methanol ............................... 123 Three Dimensional Structure of SP B 59 80 in Methanol ................................ ... 125 NMR Resonance Assignment for SP B 59 80 in DPC Micelles ......................... 127 Three Dimensional Structure of SP B 59 80 in DPC Micelle ............................. 129 Discussion ................................ ................................ ................................ ............ 132 5 CONCLUSIONS AND FUTURE DI RECTIONS ................................ .................... 163 APPENDIX : SCRIPTS FOR STRUCTURE CALCULATIONS AND REFINEMENT .... 167 LIST OF REFERENCES ................................ ................................ ............................. 185 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 205
7 LIST OF TABLES Table page 4 1 1 H chemical shifts for SP B 59 80 in methanol measured at 3 00 K. ..................... 160 4 2 3 J couplings for SP B 59 80 in methanol measured at 300 K. ...................... 160 4 3 A summary of structural statistics for the 10 final structures of SP B 59 80 in methanol and SP B' 59 80 bound to DPC micelles ................................ .............. 161 4 4 1 H and 15 N chemical shifts for 15 N labeled SP B 59 80 in DPC micelles measured at 315 K. ................................ ................................ .......................... 162
8 LIST OF FIGURES Figure page 1 1 Particles in the alveolar subphase. ................................ ................................ ..... 26 1 2 Pul monary surfactant renewal cycle. ................................ ................................ .. 27 1 3 Hypothetical in vitro model for specific DPPC adsorption to the air water interface. ................................ ................................ ................................ ............ 28 2 1 Far UV CD spectra associated with various types of secondary structure. ........ 65 2 2 Chemical structures of glycerophospholipids. ................................ .................... 66 2 3 Gel to liquid crystalline phase transition of the bilayer lamellar phase. ............... 66 2 4 Types of lipid phases. ................................ ................................ ......................... 67 2 5 Origins of the powder pattern obtained in the presence of chemical shift anisotropy (CSA) interaction. ................................ ................................ .............. 68 2 6 Chemical shift anisotropy of t he rigid lipid phosphate group is a measure of the degree of order and characterizes the types of the motions that are experienced by the group in the lipid phase. ................................ ...................... 69 2 7 Polymorphisms and phosphorus NMR lineshapes. ................................ ............ 70 2 8 Effect of the quadrupole interaction on Zeeman energy levels of nucleus with spin I=1 with axial symmetry. ................................ ................................ .............. 71 2 9 The angles relating the principal axis of the EFG tensor (usually the direction of C 2 H bond) to the applied magnetic field H o direction. ................................ ................. 72 2 10 Origin of the powder pattern of 2 H NMR spectra observed for deuterium nuclei incorporated into phospholipid tails in membranes in uni and multilamellar vesicles ................................ ................................ .......................... 73 2 11 Solid state NMR provides residual quadrupolar couplings (RQCs) that directly correspond to ensemble averaged molecular structure. ........................ 74 2 12 Chemical shift pattern of different amino acids in 2D 1 H 1 H TOCSY spectrum. ................................ ................................ ................................ ........... 75 2 13 ................................ ...... 76 2 14 A summary of solid phase peptide synthesis with Fmoc N terminus protection group. ................................ ................................ ................................ ................. 78
9 2 15 pET 31b expression system designed by Novagen and SPB C peptide purification scheme. ................................ ................................ ............................ 79 2 16 Overexpression of 18 kDa KSI SPB 59 80 His 6 fusion construct in BL21(DE3)pLysS cells monitored as function of time after overexpression inducti on. ................................ ................................ ................................ ............ 80 2 17 Nickel affinity chromatogram of purification of KSI SPB 59 80 His 6 fusion construct from inclusion bodies in denaturing conditions ................................ .... 80 2 18 Cleavage efficiency of KSI SPB 59 80 His 6 fusion construct by cyanogen bromide in 80% formic acid was monitored as a function of time ....................... 81 2 19 Pure SPB 59 80 peptide in RP HPLC fractions after lyophilization.. ....................... 81 2 20 Mass spectra of expressed SP B 59 80, wild type SP B 59 80 and 15 N SP B 59 80 peptides ................................ ................................ ................................ .............. 82 3 1 CD spectra at 45 B' 59 80 at a P:L molar ratio of 1:100 in 4:1 DPPC:POPG and 3:1 POPC:POPG. ................................ ............................... 106 3 2 Phosphorous NMR spectra of 4:1 DPPC d 62 :POPG MLVs with SP B 59 80 ...... 107 3 3 Phosphorous NMR spectra of 3:1 POPC d 31 :POPG MLVs with SP B 59 80 ...... 108 3 4 DSC scans for 4:1 DPPC d 62 :POPG LUVs with SP B 59 80 .............................. 109 3 5 Deuterium NMR spectra as a function of temperature and SP B' 59 80 molar ratio for 4:1 DPPC d 62 :POPG MLVs and 4:1 DPPC:POPG d 31 MLVs ............. 109 3 6 First moment, M 1 as a function of temperature, for the 2 H NMR spectr a plotted in Fig.3 5 ................................ ................................ .............................. 110 3 7 Deuterium NMR spectra of 4:1 DPPC d 62 :POPG MLVs with SP B 59 80 at the indicated P:L molar ratios. ................................ ................................ ................ 111 3 8 Deuterium NMR spectra of 3:1 POPC d 31 :POPG MLVs with SP B 59 80 at the indicated P:L molar ratios. ................................ ................................ ............... 112 3 9 Order parameter profiles for the sn 1 chains of DPPC d 62 and POPG d 31 in 4:1 DPPC:POPG MLVs at 44 B 59 80 ................................ .............. 113 3 10 Order parameter profiles for the sn 1 chains of POPC d 31 and POPG d 31 in 3:1 POPC:POPG MLVs at 44 B 59 80 ................................ .............. 114 3 11 Changes in d 31 palmitoyl acid chain order parameters in 4:1 DPPC:POPG MLVs and 3:1 POPC:POPG MLVs on addition of SP B' 59 80 .......................... 115 3 12 Models of SP B 59 80 interactions with the two lipid environments ................... 116
10 4 1 CD spectra of SP B 59 80 in methanol at 303K and bound to DPC micelle s at 3 18K ................................ ................................ ................................ ................. 137 4 2 300ms NOESY spectrum for SP B 59 80 in methan ol at 300 K. .......................... 138 4 3 Summary of the NOE interactions and CSI measurements for SP B 59 80 in methanol. ................................ ................................ ................................ .......... 140 4 4 Summary of the NOE interactions for SP B 59 80 in methanol and for SP B' 59 80 in DPC micelles ................................ ................................ .............................. 141 4 5 SP B 59 80 structure in methan ol ................................ ................................ ........ 142 4 6 1 H 1 5 N HSQC spectrum and assignments for 15 N labeled SP B' 59 80 in DPC micelles at 315K. ................................ ................................ .............................. 143 4 7 Alternating 1 H 1 H TOCSY and 1 H 1 H NOESY from 15 N edited 3D TOCSY HSQC and NOESY HSQC spectra of 15 N labeled SP B 59 80 peptide in DPC micelles. ................................ ................................ ................................ .......... 143 4 8 Summary of the NOE resonances for SP B' 59 80 in DPC micelles. .................... 146 4 9 Three dimensional structure of SP B' 59 80 peptide in DPC micelles. ................. 147 4 10 SP B' 59 80 in DPC micelles with residues colored according to their charge ...... 148 4 11 Thre e dimensional surface models of SP B 59 80 structure in methanol and in DPC micelle ................................ ................................ ................................ ...... 149 4 12 Possible mechanism of DPPC trafficking by SP B 59 80 ................................ .... 149 4 13 The Ramachandran plot of ensemble of 10 SP B 59 80 structures in methanol .. 150 4 14 The Chi1 Chi2 plots show the chi1 chi2 sidechain torsion angle combination s for 10 structure ensemble of SP B 59 80 in methanol ................................ ......... 151 4 15 The Ramachandran plot of ensemble of 10 SP B 59 80 structures i n methanol with data points colored according to G factor ................................ .................. 153 4 16 Plots of main chain parameters for 10 SP B 59 80 structures in methanol compare d t o the well refined structures at a similar resolution. ........................ 154 4 17 The Ramachandran plot of ensemble of 6 SP B' 59 80 structures in DPC micelles ................................ ................................ ................................ ............ 155 4 18 The Chi1 Chi2 plots show the chi1 chi2 sidechain torsion angle combinations for ensemble of 6 structures of SP B' 59 80 in DPC ................................ ............ 157
11 4 19 The Ramachandran plot of ensemble of 6 SP B' 59 80 structures in DPC with data points colored according to G factor ................................ ......................... 158 4 20 Plots of main chain parameters for 6 SP B' 59 80 structures in DPC compare d t o the well refined structures at a similar resolution. ................................ ......... 159 5 1 Molecular models of SP B 59 80 directing specific DPPC adsorption to the air water interface and facilitating tight packing interactions between the bilayers of tubular myelin. ................................ ................................ .............................. 166
12 LIST OF ABBREVIATION S ARDS Acute respiratory distress syndrome C D Circular dichroism C OSY Correlation spectroscopy CS A C hemical shift anisotropy CSI C hemical shift index C TERMINUS Carboxy terminus D I P SI Decoupling In the Presen ce of Scalar Interactions D P C D odecyl phosphatidylcholine DPPC Dipalmitoylphosphatidylcholine DPP G Dipalmitoylphosphatidyl glycerol D S C Di fferential scanning calorimetry EDTA Ethylenediaminetetraacetic acid EFG Electric field gradient EPR Electron paramagnet ic resonance FTIR Fourier transform infrared GST Glutathione S transferase H C N Proton, Carbon, Nitrogen HFIP H exafluoro 2 propanol HPLC H igh performance liquid chromatography HSQC H eteronuclear single quantum coherence IVM Internal variable module IPTG I D 1 thiogalactopyranoside KSI Ketosteroid isomerase LB Lamellar body
13 LPS L ipopolysaccaride LS Lung surfactant LUV Large unilamellar vesicle MALDI TOF Matrix assisted laser desorption time of flight MAS Magic angle spinning MBP Maltose binding pr otein MD Molecular dynamics MLV Multi lamellar vesicle NMR N uclear magnetic resonance NOE N uclear Overhauser effect NOESY N uclear Overhauser enhancement spectroscopy N TERMINUS Amino terminus PAS P rincipal axis system PC P hosphatidylcholine PE P hosphatidyl ethanolamine PG P hosphatidyl glycerol PO PC Palmitoyloleoylphosphatidylcholine POPG Palmitoyloleoylphosphatidylglycerol PS Pulmonary surfactant RDS Respiratory distress syndrome R F Radiofrequency R MSD root mean square deviation R NA R ibonucleic acid RQC Resid ual quadrupolar coupling R P HPLC Reverse phase high performance liquid chromatography
14 S APLIP s S aposin like proteins SDS S odium dodecyl sulfate SP A Surfactant protein A SP B Surfactant protein B SP B 59 80 Surfactant protein B residues 59 to 80 SP B 63 78 Su rfactant protein B residues 63 to 78 SP B C Surfactant protein B C terminus (residues 59 to 80) SP C Surfactant protein C SP D Surfactant protein D SP PS S olid phase peptide synthesis TCEP T ris(2 carboxyethyl)phosphine TFA T rifluoroacetic acid TFE T rifluoroe thanol TM Tubular myelin TOCSY T otal correlation spectroscopy TSP 3 (Trimethylsilyl)propionic 2,2,3,3 d4 acid sodium salt UV Ultraviolet
15 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Ful fillment of the Requirements for the Degree of Doctor of Philosophy STRUCTURAL ANALYSIS OF THE C TERMINUS OF LUNG SURFACTANT PROTEIN B (SP B) By Anna Kuznetsova May 2013 Chair: Gail E. Fanucci Cochair: Joanna R. Long Major: Chemistry Lung surfactant i s a complex lipid protein substance that coats the inner layer of alveoli and facilitates breathing by minimizing surface tension. Surfactant protein B (SP B) is of particular importance since it may specifically enrich d ipalmitoylphosphatidylcholine ( DPP C ) at the air water interface by trafficking and sorting of lipids In vivo and in vitro studies have demonstrated that the C terminus of SP B (SP B 59 80 ) promote s lipid exchange and rapid transfer from bilayer stores into interfacial monolayers. Little is known about the atomic level structure and dynamics of this peptide in the lipid environment. S olution and solid state NMR were used to probe the structure and dynamics of SP B 59 80 in detergent micelles and lipid vesicles to compare the effects of the di fferent lipid environments on peptide helical structure and the effects of the peptide on lipid dynamics Using standard TOCSY and NOESY experiments, we were able to assign all resonances for SP B 59 80 in methanol and in DPC micelles; CD studies indicate the peptide is helical in both lipid environments. Using XPLOR NIH an ensemble of 10 B 59 80 in methanol and an
16 ensemble of 6 structures with backbone RMSD 0. 67 B 59 80 in DPC micelles. Both structures are unusual amphipathic structures with a helix turn helix motif formed due to the presence of proline at position 67. We also detect ed hydrophobic contacts in methanol and hydrophilic side chain contacts in micelles due to t he tw o helices positioned with their hydrophilic sides facing outwards in polar methanol solvent and facing inwards in the non polar environment of the DPC micelle The interchange between two conformations suggest s a possible mechanism for interaction of the C terminus of SP B with DPPC and subsequent lipid transfer T he N and C terminal helices with distinct charged and non polar faces connected by a flexible loop, D PPC molecule is bound to SP B 59 80 it could be transferred from the lipid bilayers to the monolayer at the air/water interface or between lipid leaflets. T he net result would be destabilization of local bilayer structure or ev en translocation of the DPPC lipid.
17 CHAPTER 1 INTRODUCTION Pulmonary Surfactant as a Surface Active Agent in Lungs A fundamental problem associated with breathing stems from the high surface area of the lung. Lungs require such a large surface area t o carry out gas exch ange at a physiologically appro priate level (1) This is attained through the alveoli tiny aqueous lined air sacs with a high radius of curvature and a surface area of ~200 380 m 2 (2) p = 2 /r), the smaller the alveolar radius (r) the higher the pressure (p) required to keep it infl ated There is a general need to minimize surface tension ( ) in small, gas containing, aqueous lined structures by introducing surfactants to stabilize them With a standard air/water interface it would be impossible to inflate the lungs at atmospheric pr essure. The task of reducing surface tension and increasing lung compliance is accomplished by p ul monary surfactant (PS) which covers the respiratory surface of the alveoli Pulmonary surfactant (PS) is a surface active lipid protein complex lining the al veoli that reduces alveolar surface tension to near zero values at the lung air liquid interface and allows inflation and oxygen exchange at ambient pressure. Pulmonary Surfactant in Alveoli Pulmonary Surfactant Cycle PS is synthesized by specialized cel ls of the alveolar epithelium alveolar epithelial type II cells, also called type II pneumocytes, which make up 5% of the epithelial surface. PS is stored in these cells in the form of tightly packed bilayers in vesicles called lamellar bodies (LBs). Upo n secretion into the extracellular space the LBs form a n extensive lipid rich network of bilayers termed tubular myelin (TM), that
18 covers rapidly and efficiently the whole respiratory surface (Fig.1 1, Fig. 1 2). From tubular myelin phospholipids are tran sferred to the monolayer at the air water interface to efficiently cover and stabilize this interface for adequate respiration (Fig. 1 2 and Fig, 1 3) Pulmonary Surfactant Chemical Composition PS is conserved across mammalian species and contains 80% phos pholipids 5 10% neutral lipids (mainly cholesterol), and 10% proteins by weight (3 5) The primary phospholipids of surfactant are dipalmitoylphosphatidylcholine DPPC ( ~40 50%), palmitoyloleoylphosphatidylcholine POPC (~30%) and anionic palmi toyloleoylphosphatidylglycerol POPG (~10%) (6) In an aqueous environment phospholipids self assemble into bilayers (the structural form of LBs), but at polar/non polar interfaces (e.g. the air liquid interface) phospholipids can form monolayers, with the headgroups oriented toward the aqueous pha se and the hydrophobic acyl chains pointing toward the air (7) DPPC, which has melting temperature (T m likely favored as the major phospholipid of PS at the air water interface because its fully saturated acyl chains can be very t igh t ly packed and withstand high lateral pressure (8) The higher the concentration of phospholipid molecules at the interface, the fewer the number of water molecules exposed to air providing the large reductions of surface tensions at the alveolar air liquid interface (7) However, t o quickly cover the alveolar surface, is attained by m ixing DPPC with low er melting unsaturated phospholipids (9) However, lipids alone cannot achieve the near zero surface tension present in lung surfactant. T he unique physical properties of PS a re determined by low levels of lipid associated proteins : Surfactant protein B (SP B) and Surfactant protein C (SP C) (10)
19 SP B and SP C are extremely hydrophobic, soluble in organic solvents such as chloroform/methanol mixtures, and thus co purify with lipids in chloro formic extractions of PS (11) D ue to their low concentration and high hydrophobicity, little is known about the atomic level structure and dynamics of these proteins. The other two surfactant proteins SP A and SP D are hydro philic protein complexes that protect the respiratory surface from pathogens (12, 13) Mature SP B is an extremely hydrophobic lipid associated homodimer, made up of two 79 81 amino acid monomers that are disulfide linked (14, 15) It contain s high levels of valine, leucine, isoleucine, alanine, phenylalanine, and tryptophan. (16 18) The native form of the protein contains s ix highly conserved cysteines that form three intra molecular disulfide bonds and a seventh cysteine residue that forms an intermolecular disulfide bond. The intramolecular disulfide bonds define a fold that has been conserved for an estimated 300 million years. (19) The unique pattern of intramolecular disulfide bonds, the resulting fold and its hydrophobicity place SP B in the saposin like family of proteins (SAPLIP). (20, 21) Other proteins in the saposin family include amoebapores from Entamoeba histolytica, acid sphingomyelinase, acyloxyacyl hydrolase, and sphingolipid activator proteins A D (saposins A D). (22) Like other members of the saposin superfamily, SP B is primarily helical in lipid bilayers and structure promoting solvents. (23, 24) However, SP B is unique in this family with its high hydro phobicity and integration into lipid assemblies. (22) FTIR and CD spectroscopies indicate that in lipid bilayers the amphipathic helices of SP B lie in the plane of the lipid lamellae. (23, 25) One of the interesting features of SP B and a major difference between SP B and SP C is that SP B bears intermittent highly charged residues, 9 of 12 which
20 are basic and localized in amphipathic helix regions of the N and C terminal segments of the protein, imparting a net cati onic charge to SP B of > + 5 at physiologic pH. (26, 27) A net cationic charge of SP B at physiologic pH may promote a selective interaction of SP B wit h the anionic PG, ample in surfactant (28) Lipid Trafficking by Lipid A ssociated Lung Surfactant Proteins Hydrophobic surfactant proteins B and C ( SP B (29) and SP C (30, 31) ) promote rapid transfer of phospholipids from bilayers of TM to interfacial monolayers at the lung air water interface (Fig. 1 3). SP B and SP C are required as mediators at relatively low concentrat ion (~0.1 mol%) to make the transfer of phospholipids through the aqueous hypophase rapid and efficient (10) H owever, their mechanisms of action are still lacking as are atomic resolution structures of SP B and SP C in lipid membranes or lipid like environments FTIR indicates that in lipid bilayers SP B is a helical protein with the helices in the plane of the lipid lamellae (32) In contrast, SP C adopts a helical structure with a transmembrane orientation (33) Both SP B and SP C distribute preferentially in disordered re gions of membranes and interfacial monolayers (9, 34) In contrast with SP C, SP B, which directly interacts with the phospholipids in surfactant, is particularly critical to its function. (35) SP B is absolutely required for survival and normal respiration. Hereditary SP B deficiency is lethal in humans (36) and in SP B knock out mice; (37) antibodies against SP B cause respiratory distress syndrome (RDS) in vivo (38) and insufficient SP B is a leading cause of respiratory distress in premature infants. SP B is strict ly required not only for the biogenesis and packing of PS into lamellar bodies, (39, 40) but it also underpins the unique PL packing and dynamics observed in PS. It is also thought that SP B may specifically enrich D PPC at the air water interface by trafficking and sorting of lipids in bulk PS (41) but the mechanism of
21 lipid tra fficking by SP B is unknown. SP B could promote formation and stabilization of certain non bilayer intermediate structures required for efficient phospholipid transfer. (42) Monolayer experiments with films containing SP B have shown that SP B facilitates ordered and reversible compression driven structural transitions by formation of associated bilayer patches, which can quickly reinsert into the interfacial monolayer upon expansion. (43 ) Compression of pure lipid films usually ends with an irreversible collapse. (44, 45) Thus respiratory failure associated with absence of SP B is probably caused not by impaired PS production but rather by the loss of a proper dynamic flow of lipids (46) Clinical Aspects of Pulmonary Surfactant Deficiency Deficient or dysfunctional PS leads to severe respiratory disease. Maturation of type II pneumocites and PS secretion occur late in gestation and PS is often insufficient in premature infants leading to Respiratory Distress Syndrome (RDS), a major cause of perinatal mortality (46) RDS is associated with high alveolar surface tension, increased work of breathing and alveolar collapse (47) Impaired gas exchange at the alveolar capillary barrier causes a decrease in blood oxygen levels (hypoxia) and an increase in carbon dioxide ( respiratory acidosis) (48) Adult p atients suffering from extensive lung inflammation, infec tion or trauma often develop a cute RDS (ARDS) due to inactivation of PS by inflammatory by products and blood components leaked into the airways through a deteriorated alveolar capillary barrier (46) The p ivotal role of PS in RDS and ARDS has led to the development of therapies which rely on either synthetic or modified natural PS (extracted from bovine or porcine sources). C urrent FDA approved therapy for RDS in premature infants relies on animal derived P S. Such treatment creates the risk of severe immune response if administered repeatedly but this is not as crucial for
22 RDS treatment in preterm babies since they start producing their own surfactant when they start breathing. In contrast the need for reg ular administration of surfactant preparations for treat ment of ARDS in adults and mature immune function makes n atural PS therapy impossible in older populations due to surfactant inhibition and severe immune response (49) Consequently, s ynthetic lung surfactant replacements are an attractive goal for treatment of RDS and A RDS, as they are considerably cheaper to produce, remove the immunologic risks, and may be crucial to effectively treating ARDS. However, although current s ynthetic PS replacements are less susceptible to degradation by plasma proteins and remove the risk of immune response, they reduce surface tension much less than animal derived surfactant. To better mimic the natural surfactant we need to elucidate the molecular mechanisms by which surfactant proteins and membranes modulate respiratory physiology (50) Pulmonary Surfactant Replacement The hydrophobicity and di sulfide bridges in SP B have made heterologous expression and purification of the full length protein i n large quantities impractical Synthetic, peptide based replacements of SP B for treatment of RDS and ARDS have received notable attention not only beca use they are much easier to make with high yield and purity than the full length protein, (51, 52) but also because they remove the immunologic risks associated with animal derived surfactant. (53 55) While the entire 80 amino acid SP B protein is essential for lung surfactant organization, dynamics and respiration, fragments of the native sequence have shown significant biophysical function Peptides corresponding to N termina l and C terminal 20 25 residue fragments of the native sequence have shown surface activity similar to native SP B, (17, 56) with maximal activity achieved using a chimeric construct of the two
23 termini. (57) This could be explained by the fact that b oth N terminal and C terminal peptides form cationic amphipathic helices in SDS micelles and lipid bilayers, (58 61) which makes them the most critical fragments of SP B for phospholipid binding and their rapid transfer from bilayer stores into interfacial monolayers (51, 52) but their individual roles in lipid trafficking are not well understood. B oth N terminal and C terminal peptides promote lipid exchange in vitro and restore lung function in animal models of respiratory distress. (23, 62 64) Surfactant preparations containing these synthetic peptides improve oxy genation and lung compliance in surfactant deficient animal models. (17, 27, 65 67) The preponderance of leucines and relative lack of prolines and aromatic sidechains in the C terminus distinguish it from the N ter minus, which contains four prolines as well as a phenylalanine and a tryptophan. The peptides also have varying spacing between hydrophilic and hydrophobic amino acids. These differences lead to varying secondary structures and penetration into lipid lamel lae, and we have documented they have different effects on phospholipid dynamics. (68, 69) Although to date the N terminal fragment has been more extensively studied in terms of its molecular properties and structu ral adaptation in lipid environments, many in vivo activities of SP B are fulfilled by its C terminal end, specifically residues 59 80 ( SP B 59 80 ) (51, 70) SP B 59 80 has shown efficacy in altering phospholipid prope rties based on in vitro assays measuring surface tension, such as pulsating bubble surfactometry, as well as in vivo assays demonstrating gain of function in surfactant deficient fetal rabbits (17, 57) The most su ccessful potential synthetic replacement for SP B (71 75) named KL 4 peptide, KLLLLKLLLLKLLLLKLLLLK (55) was design ed to mimic the charge distribution
24 and hydrophilic/hydrophobic ratio of the primary sequence of SP B 59 80 This simple peptide mimetic of SP B 59 80 relieves respiratory distress syndrome (76) h owever KL 4 /lipid prepara tions are not as efficient in lowering surface tension as native PS Moreover, an FTIR study of KL 4 in DPPC/DPPG concluded the peptide can form a helix spanning the bilayers and it has been posited that KL 4 might more closely mimic SP C rather than C termi nus of SP B (77) Thus, understanding how SP B 59 80 functions in the lipid environment would allow the development of mimetics which are more stable and efficient in lowering surface tension than KL 4 but less susceptible to degradation by plasma proteins than native SP B T he recent solution NMR study of residues 63 78 of the C terminus of SP B (SP B 63 78 ) reconstituted in either SDS micelles or the organic solven t HFIP found the first five residues to be unstructured and established that the rest of the sequence formed a helix in both SDS micelles and organic solvent. (60) This study and a previous CD study (59) of SP B 59 80 in TFE and SDS micelles are the only documented structural assessments of the C terminal region of SP B. SP B 59 80 is believed to form an amphipathic helix involved in lipid organization, but direct structural measurements in lipid contexts have not been documented. Detailed study of the interaction of SP B with phospholipids is necessary to create more effective clinical surfactant preparations. Based on understanding the properties of SP B 59 80 new non natural peptide analo gs of SP B that exploit the protein's qualities in altering lipid dynamics could be made to extend or enhance artificial lung surfactant therapies. Thus, understanding how SP B 59 80 affects the molecular and biophysical properties of lipids is of particula r relevance to the treatment of various forms of RDS.
25 The roles of m embrane proteins in many physiological events can be understood only by studying their structure and dynamics in their native lipid environments. Surfactant protein B minimizes surface te nsion at the alveolar air water interface by improving dynamics of the lipid redistribution at the interface through trafficking and sorting of LS lipids especially DPPC lipid enrich ed at the air water interface T he C terminus of SP B (SP B 59 80 ) may sel ect for DPPC lipid and promote its exchange and rapid transfer from bilayer store s into interfacial monolayers. The synthetic LS mimics in this study utilized binary mixture DPPC:POPG to mirror the interfacial LS phases, whereas POPC:POPG mixture was used as the mimic of bilayer storage phases of LS. Chapter 3 provides more detail on the partitioning of SP B 59 80 peptide into these LS mimics and effects of SP B 59 80 on lipid dynamics Chapter 4 discusses the structures of SP B 59 80 peptide in methanol solut ion and in DPC micelles, which mimic DPPC rich LS environments Chapter 2 introduces the theory behind the techniques used in this work. Chapter 5 summarizes all results and provides conclusions for the dissertation as well as the incentives for future exp eriments
26 Figure 1 1. Particles in the alveolar subphase. In this electron micrograph section of a rat lung, lamellar bodies (LB) are seen form ing tubular myelin (TM) (bar at used and rejected surfactant materials. Inset: detail of tubular myelin at lower left, showing small projections in the corners, thought to represent SP A 2 for 48 h, but no differences were seen from controls. Figure borrowed with permission from : Goerke J .( 199 8), Pulmonary surfactant: functions and molecular composition BBA 1408 79 89
27 Figure 1 2 Pulmonary surfactant renewal cycle: PS is synthesi zed in alveolar epithelial type II cells stored there as tightly packed bilayers in lamellar bodies (LBs). and secret ed into the extracellular space to form a n extensive lipid ri ch network of bilayers tubular myelin (TM) From TM bilayers phospholipids are transferred to the monolayer at the air water interface to efficiently cover and stabilize this interface for adequate respiration (see also next figure) Figure borrowed and adapted with permission from : Serrano A G. and Perez Gil, J. (200 6 ) Protein lipid interactions and surface activity in the pulmonary surfactant system Chemistry and Physics of Lipids 141 1 05 1 18
28 Figure 1 3 Hypothetical i n vitro model for specific DPPC adsorption to the air water interface. P hospholipids (mainly DPPC) are transferred f rom tubular myelin bilayers to the monolayer at the air water interface to efficiently cover and stabilize this interface for adequate respi ration. Adapted with permission from : Blanco, O., and Perez Gil, J. (2007) Biochemical and pharmacological differences between preparations of exogenous natural surfactant used to treat Respiratory Distress Syndrome: Role of the different components in an efficient pulmonary surfactant, European Journal of Pharmacology 568 1 15.
29 CHAPTER 2 METHODS FOR STUDYING LIPID DYNAMICS AND M EMBRANE ACTIVE PEPTI DE STRUCTURE IN MEMBRAN E MIMETIC ENVIRONMEN TS Circular Dichroism (CD) T he circular dichroism (CD) techniqu e allows structural assessment of proteins and provides quick measure ment of structural changes, which are often essential to their biological function T his method allows monitoring of structure under a variety of conditions in which proteins actually ope rate, including interacting with lipid bilayers In chapters 3 and 4 CD spectra provided prompt assessment of SP B 59 80 peptide structure in different environments (methanol, DPC micelles, lipid bilayers mimicking LS) and showed that it is helical in all o f them with some minor possible secondary structure differences due to environment. Consequently, the next chapters have been focused on high resolution structures of SP B 59 80 in methanol and DPC micelles, which could provide insights into its structure a nd interaction with lipid environment of LS. L ight is an unpolarized electromagnetic wave with randomly oriented electric vectors, which propagates equally in all directions Plane polarized light is made up of 2 circularly polarised components of equal m agnitude, one rotating counter clockwise (left handed, L) and the other clockwise (right handed, R). (78) Circular dichroism (CD) refers to the differential absorption of the se 2 components (79) If, after passage through the sample being examined, the L and R components are not absorbed or are absorbed to equal extents (for an achiral chromohore) the recombination of L and R would look like circle and would regenerate radiation polarized in the original plane However, if L and R are absorbed to different extents (for a chiral chromohore) the resulting radiation after recombination of unequal L and R vectors would look like ellipse and would be said to possess elliptical polarization which gives rise to CD signal (79, 80) Thus CD
30 signal is measure d as the difference in absorbance between the L and R circularly polarized components ( ) and lik e any light absorption value it c an be rewritten in (2 1) w here and are the molar extinction coefficients for the L and R circularly polarized components of light respectively; C is the molar concentration and is the path length is cm. or historical reasons CD signal is always report ed in degrees of ellipticity ( ) W hen there is no difference in the absorbance of right and left circular polarized light ( E R = E L ) resulting combined radiation is linearly polarized When there is complete absorbance of the circular polarized light in one direction ( E R =0 or E L =0 resulting radiat ion is circularly polarized The case in between is elliptically polarized light with ellipticity defined through its tangent which is approximately equal to the angle itself for small angles : (2 2) wh ere E R E L and E R + E L are the minor and major axes of the ellipse resulting from addition of two light components with different amplitudes and circularly polarized in different directions By analogy to the definition of molar circular dichroism molar ellipticity is circular dichroism in degrees of ellipticity corrected for concentration (in mols/L) and cell length (in cm) : (2 3) The CD signal is obtained as ellipticity in units of millidegr ees and is normalized according to the protein concentration to yield molar ellipticity per ami n o acid : (78)
31 (2 4 ) The number of amino a cids in the protein is denoted as # AA, [M] is the molar concentration, deg is the CD signal in millidegrees, and l is the path length in centimeters of the cuvette. In proteins, the chromophores of interest include the peptide bond (absorption below 240 nm), aromatic amino acid side chains (absorption in the range of 260 t o 320 nm) and disulphide bonds (weak broad absorption bands cent e red around 260 nm). (79) P rotein secondary structure is defined by the conformation of the peptide bond s and the CD a bsorption of peptide bo nd s in the far UV region (240 nm and below) is characteristic of the different types of regular secondary structure found in proteins (Fig. 2 1 ) (79) T he spectrum of a polypeptide in random conformation displays a small positive n transition at ~ t ransition at approximately 195 nm. T he spectrum of the helix vice versa displays a large negative n transition at 222 nm This usually weak forbidden n is enh anced in helix because even though an isolated amide bond are orthogonal, the orbitals of neighboring chromophores are not necessarily orth o gonal and thus the transition is more likely to occur in a polymer between neighboring amide bonds The strong transition i n helix spectrum is split into two transitions b ecause of exciton coupling with negative band at approximately 208 nm and a positive band at approximately 192 nm. (81) The C D spectrum structure ha s a negative band at 218 nm and a positive band at 195 nm also show ing evidence of exciton coupling in the owever the splitting of the transitions are different than in the case of the helix
32 Consequently, CD spectrum of the helical conformation (Fig. 2 1) has characteristic shape with two negative peaks at 208 nm and 222 nm of the equal intensity. However, CD spectra of SP B 59 80 in different environments showed that the peak at 208 nm wa s much lower than the peak at 222 nm and sug gested helical conformation of SP B 59 80 may have some deviat ions from purely helical state Solid State NMR Spectroscopy Solid state NMR was used to study the effect of SP B 59 80 on lipid dynamics in bilayers which provided some indirect evidence of the peptide partitioning depth into the lipid bilayers and allowed some assumptions about the structure of SP B 59 80 in these lipid bilayers. In solid state NMR b esides the Zeeman interaction of the nuclear spin w ith the external magnetic field, there are additional perturbations of the total Hamiltonian due to magnetic interactions (dipolar coupling, chemical shift) and electric interactions (quadrupolar coupling). The corresponding 3x3 tensors that describe t hese couplings in every direction provide information regarding both the structure and dynamics of biomolecular system. T he principal values and principal axis systems (PAS) of the various coupling tensors in their motionally averaged form yield structural inf ormation about the system and define overall lineshape observed in NMR spectr a whereas their fluctuations are related to the dynamics of the system or the mean squared amplitudes and rates of the motions and lead to characteristic relaxation behavior S tr uctural fluctuations of fluid membrane lipids give rise to averaging of the coupling tensors affect ing the NMR lineshapes and relaxation times that are determined by the corresponding coupling tensor S olid state NMR is thus a unique technique because it
33 p rovides information about both structur e and dynamic s at the level of individual atom s (82) S t udies of biological membranes and other lipid systems by solid state NMR have been mostly carried out using phosphorus ( 31 P) and deuterium ( 2 H ) nuclei of lipid molecules. Th e unique feature of the method is that it is used to study not only m acroscopic pro perties of lipid systems such as lipid phases in the system and phase transition s between them but also microscopic properties at the atomic level, such as local environment in a very small submolecular region of the membrane. The physical state of the h ydrocarbon chain region of phospholipid bilayer is studied by 2 H NMR by determining motional anisotropies of C 2 H bonds (82) whereas 31 P NMR is a valuable tool for lipid phase behavior analysis by looking at the motional properties of lipid headgroup s via 31 P chemical shift anisotropies (CSA) NMR of lipid bilayers Representative glycerophospholipi ds shown in Figure 2 2 possess polar head groups X that differ in their size, capacity for hydrogen bonding, charge and the nonpolar acyl chains which vary in their length and degree of unsaturation. (82) Biological membranes are phospholipid bilayers that form due to t he hydrophobic effect which causes the hydrophobic lipid acyl chains partition w ithin the bilayer interior to avoid contact with water ; the polar and/or charged head groups interact with water at the membrane surface. I n excess water phospholipid bilayers undergo phase transition from highly ordered lamellar gel phase with tiled chain s ( L ) to the lamellar liquid crystalline phase ( L ) a s temperature increases (Fig. 2 3 ) (82) Oth er types of phospholipid nanostructures are also possible, including the normal and inverted hexagonal phase s, cubic phases and micellar (Fig. 2 4 ) (83, 84) depending on the structures of the individual lipid molecul es. L iquid crystalline lamellar phase is classified
34 as smectic A lyotropic phase under liquid crystal classification. (82) In lyotrophic liquid crystals solvent molecules fill the space around the compounds to provide additional degree of freedom of concentration and fluidity that enable formation of a variety of different phases. In the s mectic A p hase, the molecules are oriented along the layer normal, also called director axis, which also represents the axis of motional averaging (Fig.2 3 ). (85) Phospholipids in the bilayer can be represented as r igid rods with their most probable orientation parallel to the director axis. However, if followed for a long enough time the molecules will constantly execute angular excursions around this preferred orientation. The description of liquid crystal s involve s an analysis of order: positional order (whether molecules are arranged an ordered lattice) and orientational order (whether molecules are mostly pointing in the same direction), and moreover order can be either short range or long range. O rientational or der parameter which measures angular excursions and fluctuations around the director axis, is usually quantitative ly defined based on the average of the second Legendre polynomial : (85) (2 5 ) w here S is order parameter ranging from S=1 for perfectly aligned sample and S=0 for completely completely random and isotropic sample; is instantaneous angle between the bilayer director axis and local director axis which usually is C 2 H bond vector in a 2 H NMR of lipid acyl chains 31 P Chemical Shift Anisotropy 31 P NMR is mainly applied to membrane studies to detect lipid phase change s and the degree of organization of lipid headgroups. Various phases shown in Fig. 2 4 can be adopted by hydrated phospholipids depending on the nature of the lipids, level of
35 hydration, pH, temperature, presence of ions, etc. The technique is based on t he large chemical shift anisotropy exhibited by phosphorus in phosphates, the relatively high sensitivity of this nucleus the fact that it does not require any synthetic labeling as 31 P is a 100% abundant nucleus and that there is generally only one phos phate group present per lipid molecule (86, 87) 31 P is a spin 1 /2 nucleus, which means it is dipolar with only one transition between two energy levels. Dipolar nuclei are spherical with a uniform charge distributio n over their surface, disturbing the external magnetic field independent of direction. This is in contrast to what is seen for quadrupolar, spin 1 nuclei (see below) and results in a much sharper NMR signal. (88) The two anisotropic interactions primarily affecting 1D 31 P solid state NMR spectra and contributing to the line broadening are chemical shift anisotropy (CSA) and heteronuclear dipol ar couplings to protons. The latter is removed with high power proton decoupling and thus the CSA is the dominant interaction typically measured. When a sample is placed in a strong magnetic field, the nuclei of the same NMR active isotope in different che mical environments experience different magnetic fields due to the shielding effect of the surrounding electrons and therefore resonate at different characteristic frequencies, which are determined as a result of a reduction of the externally applied magne tic field on the nucleus (B o ) by the shielding electrons. This reduction is known as the chemical shift and it is measured in units of parts per million (ppm) relative to the external magnetic field. In general form total Hamiltonian H of interaction bet ween 31 P nucleus and the surrounding magnetic field can be represented by Zeeman interaction of nucleus with B o applied magnetic field and its chemical shift shielding by the bond electrons from the applied external field: (89, 90)
36 (2 6 ) where (chemical shielding tensor) shows the extent of the nucleus screening from the effect of applied field B o by the bonding electrons in each direction in three dimensional sp ace. Consequently, the pattern of bonding electrons around the 31 P nucleus in the tetrahedral phosphate group greatly influences its chemical shift. This makes 31 P chemical shift to be highly dependent on relative orientation of the phosphate group with re spect to the applied magnetic field B o demonstrating large chemical shift anisotropy (Fig. 2 5 ) Due to rapid molecular tumbling in liquids, the electronic environments of the 31 P nuclei are spatially averaged during signal collection and chemical shift te nsor describing all 31 P nuclei becomes isotropic (i.e. the same in every direction) and collapses to a number thus giving single chemical shift for each type of chemical bonding environment irrespective to molecular orientation. H owever, in solids in the absence of such spatial averaging lead to the chemical shift remain ing anisotropy with its magnitude dependent on both molecular orientation and molecular environment. The a nisotropy of the 31 P chemical shift in lipids is defined b y a 3x3 matrix (tensor) and leads to the dependence of the resulting 31 P resonant frequency on the molecular orientation as well giving rise to the characteristic static 31 P solid state NMR lineshape (Fig. 2 5 ) In the laboratory re ference frame this tensor represents a regular 3x3 matrix with 9 components, but when represented in the nuclear reference frame, also called principle axis system, this tensor becomes diagonal. The diagonal components, labeled 11 22 33 which represe nt eigenvalues of this axis system, are also called the tensor principal values or principal components. The principal axis system for chemical shielding tensor of 31 P in phospholipid phosphate moiety is shown on f ig ures 2 5 and
37 2 6 with its components experimentally measured to be 11 =80ppm 22 =25 ppm, 33 = 110ppm. If the principal components and principal axes of the chemical shift tensor are known, then the chemical shift can be calculated for any orient ation: (2 7 ) i is the angle between the i th principal axis of chemical shielding tensor and the direction of applied magnetic field B o (89, 90) The static powder pattern, which is the sum of spectra for all possible orie ntations, can be calculated using eqn. 2 7 For microcrystalline sample, where all angles between the applied magnetic field and principal axes are populated, the powder pattern demonstrates three different 11 22 33 components (Fig.2 6 A). Also vice versa the principal components of chemical shielding tensor of 31 P can be estimated directly from the 31 P spectrum as shown on the same figure with 11 and 33 defining its edges and 22 defining the central peak. Rapid anisotropic motion can average some of the components of the chemical shift tensor as shown on Fig. 2 6 ( B,C ) for lamellar and hexagonal phases correspondingly. In membrane lamellar phase (Fig. 2 6 B) lipids rapidly rotate around the normal to the lipid bilayer averaging 22 33 components that are perpendicular to the axis of rotation, whereas 11 component does not change as it is parallel to the axis of motional averaging. The resulting tensor becomes axially symmetric: (2 8 ) The values of the perpendicula r ( ) and parallel ( ) tensor components correspond to the extremes in frequencies of line shape due to the bilayer normal of the lipid membranes being oriented either parallel or perpendicular relative to the ex ternal
38 magnetic field. This results in a 31 P powder lineshape with a high field peak and a low field shoulder, which is a characteristic lineshape observed for fluid phospholipid uni and multilamellar vesicles in a static solid state NMR experiment. Its b readth can be ), the difference between the and tensor elements. (91) The lineshape is affected by both motions and partial orientation of the bil ayers in the magnetic field. The more motions exist the narrower the lineshape and the smaller is the Each part of this powder pattern is also related to a certain orientation of the bilayer in a spherical liposome to the external magnetic f ield. In spherical lipid vesicles there are more perpendicularly oriented lipids relative to the particular angular orientation of the lipid over the sphere scale (92) At high magnetic fields, the plane of the bilayer tends to favor a perpendi cular orientation to the magnetic field leading to elongated ellipsoidal liposomes, due to the anisotropy of the magnetic susceptibility of phospholipid molecules being negative. (91) If 11 axis of the phosphate mo iety is also involved into rapid rotation (Fig. 2 6 C), then there is further averaging of and components into and effective components and the effective tensor still has axial symmetry, but the total chemical shift anisotropy is twice smaller compared to the bilayer case. The more motion of 11 axis is allowed the more it is reduced with the anisotropy collapsing to
39 sin gle isotropic peak with if 11 axis samples all possible angles fast enough. (93) 31 P ssNMR spectroscopy of phospholipids is often used to gather information about lipid polymorphisms as the spectral lineshapes are reflective of the different lipid ph ases that phospholipids can adopt: gel and liquid crystalline lamellar (bilayer) phases, the inverted hexagonal phase, and isotropic phases such as micellar and cubic phases (Figure 2 7 ). (87) If multiple phases are present, the spectrum will be a superposition of lineshapes. For lamellar phases, the liquid crystalline phase with more lipid motion gives a much narrower spectrum compared to the gel phase 31 P spectra of hexagonal phases exhibit twice narrower lineshape with reversed asymmetry compared to lamellar phases due to further averaging of 11 axis of the phosphate moiety in comparison to the bilayer phase as described in the paragraph above, and shown on Fig. 2 6 C. The cylinders in a hexagonal phase (H II ) have a very small radius (Fig. 2 3 ) and the lateral diffusion about the cylinder axis causes rotation of 11 axis of the phosphate moiety around the cylinder axis. In micelles and fluid isotrop ic phases with rapid reorientation in three dimensions anisotropic interactions are completely averaged out giving isotropic lineshapes with single resonances like those seen in solution NMR spectra. 2 H Quadrupolar Coupling An important feature of 2 H NMR s pectroscopy is that it enables both membrane lipids and membrane proteins to be studied by substitution of 2 H for 1 H. Substitution allows labeling specific individual C 2 H bonds in lipids with 2 H and thus obtaining site specific atomically resolved informa tion used for resonance assignment in fully perdeuterated acyl chains in phospholipids. (82) S p ecific lipids can also be isotopically
40 enriched either on one or both acyl chains to study the behavior of a particular lipid and a particular acyl chain in a complex lipid mixture Static 2 H solid state NMR probes motion in general and allows to detect ch anges occuring at the molecular level in a biologically relevant environment such as lipid membranes when specific variables are introduced. For example, l ipid melting and other phase changes can be studied as lipid 2 H lineshapes change with changing tempe rature, i ntroduction of proteins etc. (87) Interaction with p dynamics and assembly also triggering visible spectral changes. R elative protein insertion depth into lipid bilayer is also estimated using profile of lipid dynamic distribution along the acyl chain Deuterium has a I=1 spin nucleus and as such has a quadrupolar moment and three Zeeman energy levels and two transitions (Fig. 2 8 ). However, since the three energ y levels are equally spaced, the two transitions between them are degenerate and give rise to a single peak with o in isotropic solutions. In anisotropic systems th is degeneracy is removed due to the coupling of the quadrupole moment of the 2 H nucleus with the e lectric field gradient (EFG) created by the surrounding bonding electrons of the C 2 H bond. Quad rupolar coupling described by the Hamiltonian H Q causes the singlet splitting into doublet with the frequency separation between the two resonances defined as quadrupolar splitting Q (Fig. 2 8 ). Any nucleus with spin I=1 has a not spherical shape and thu s an uneven charge distribution represent ing a n electric quadrupole Electric quadrupole interacts with an EFG analogously to the interaction of an electric dipole with an electric field. Even though magnetic couplings, such as dipolar couplings between th e 2 H and 1 H nuclei, 2 H chemical shifts, are also
41 present in 2 H NMR their contribution is generally neglected because they are much smaller than the dominant electric quadrupolar interaction. (82) For symmetric EFG tensor, as in the case of C 2 H bond, the quadrupole splitting is defined in general form as: (2 9 ) where principal axis of EFG tensor, usually the direction of C 2 H bond, and applied external magnetic field H o For saturated C D bonds, the quadrupolar coupling, is 167 kHz in the static limit. (94, 95) The static quadrupolar splitting constant (167 kHz) was determined by measuring the splitting values for C D bonds in several deuterated alkane containing compounds such as ethane and acetonitrile in frozen solids (95) In equation 2 9 angle can be replaced by the combination of three angles which when combined also relat e the principal axis of the EFG tensor to the applied magnetic field H o direction (Fig. 2 9 ) : (96) (2 1 0 ) where a bar denotes a time average, represents the angle between the C 2 H bo nd vector and the instantaneous segmental chain orientation, represents the angle between the instantaneous segmental chain orientation and the director of the motion, taken to be the normal to the bilayer surface and is the angle between the directo r of the motion and the magnetic field direction H o The angles and are shown in Fig. 2 9 The re are two different parts that can be separated from each other in the eqn. 2
42 1 0 : the first two time averaged braces represent dynamic averaging and are denoted by the time averaged order parameter ( S C D ) ; the last term defines orientational dependence through angle The two parts of the equation are then separated and treated differently : (82) (2 1 1 ) where (2 1 2 ) is the combination of and angles represents the time varying angle between the major axis of EFG ( C 2 H bond vector ) and the director of the motion ( normal to the bilayer surface ) and is also shown on Fig. 2 9 B y definition the order parameter S C D measures two types of dynamics : fluctuation of the C 2 H bond about the in stantaneous segmental chain orientation (conformational order) and fluctuation of the long molecular axis around the director of the motion ( molecular order) O rientational term (3cos 2 1) in the eqn 2 1 0 is responsible for formation of very broad powder lineshape for liposomes (Pake powder pattern) because it defines specific resonance frequency of a deuterium doublet as function of orientation of the bilayer normal with respect t o magnetic field. (96) Since i n phospholipid vesicle many different orientations are possible ranging from perpendicular ( '=90 ) to parallel ( '=0 with some orientations more probable than the other then the signal from many lipid vesicles represents a continuum of deuterium doublets covering all values of angle and corresponding d iffer ent frequencies that depend on th e orientation (97) The shape of the Pake pattern is due to the fact that there are only two possible orientations along the external magnetic field (parallel and antiparallel), but
43 many orientations perpendicular to the B 0 field (Fig. 2 10 ) Consequently, the frequency for the 90 orientation of the bilayer normal is more intense than the 0 shoulder This follows simply from the spherical shape of lipid vesicle and thus also applies to the shape of 31 P CSA spectra of lipid vesicles discussed in the previous section (98) S eparation between two most intense peaks with '= 9 0 can be easily measured experimentally and used to calculate o rder parameter S CD from eqn. 2 1 1 assuming is and t he static quadrupolar splitting constant is 167 kHz : (2 1 3 ) Consequently, q uadrupole splitting between the most intense peaks is effectively reduced by the factor of S CD which is equal to 1 for perfectly ordered systems and approaches 0 for completely disordered systems, so acyl chain order and order parameter S CD decrease as th e quadrupolar splitting decreases, and vice versa. (98) The more motion at a specific acyl carbon the narrower the lineshape becomes and the two peaks for a particular bilayer orientation separated by the quadrupolar split ting will be closer together (Figure 2 11 ). The powder spectrum of the whole lipid molecule which is perdeuterated at every acyl chain position, is complex as it represents the superposition of individual powder spectra for every pair of deuterons at a pa rticular carbon along the deuterated acyl chain(s) with a specific quadrupolar splitting assigned to each pair Once is determined for each particular deuterated methylene group, the S CD values can be calculated and plotted against the carbon number of the DPPC acyl chain from 2 to 16 to show an increase or decrease in order in comparing various lipid samples (Fig. 2 11 ) The deuterons at the most distal carbon position 16
44 always have more motion and thus demonstrate the lowest order parameter in the deuterated lipid acyl chain due to the fa st rotation of the methyl group The signal of the terminal CD 3 group deuterons is also 1.5 times bigger than for the other (methylene) positions due to the one additional deuteron. The multiple spl ittings of perdeuterated lipid acyl chains and associated line broadening complicate lipid 2 H spectra because peaks from the different labeled positions hav e different order parameters and overlap with each o t her A ssign ment of specific frequencies to each CD 2 or CD 3 group in the 2 H lineshape is very difficult to do with any degree of accuracy without deconvoluting the whole spectr um which is known as dePaking and is described in the section below. The dePakin g transforms complicated, broad lines hapes to i ndividual frequencies and allows order parameters to be calculated using more accurate splitting values. Qualitative information like p hase or structure information can be obtained from non dePaked spectra, but quantitative analyses, such as order paramete r profile and insertion depth estimate, require more precise frequency assignments that could be obtained only from deconvoluting (dePaking) the broad powder lineshapes De P aking A broad powdertype NMR spectrum of the random multilamellar dispersion of per deuterated phospholipid consists of overlapping Pake powder patterns of deuterons at individual positions in lipid (Figure 2 1 1 A,B ) In the process of d ePaking these individual Pake patterns are extracted from the original spectrum and deconvoluted to lea ve only the peaks that correspond to the most abundant orientation A d ePaked spectrum now consists of the m ore highly resolved individual deuteron subspectra making individual frequencies more easily assigned (Figure 2 1 1 C ) In other words, t his procedur e calculates sort of (99) After
45 d ePaking resonances are still a bit broad, but overall the peaks are sharper, better resolved, and more easily assigned. Samples with membrane bilayers (i.e. MLVs or LUVs) demonstrate little change in the average orientation of the molecules relative to the bilayer normal (solid state regime) Con sequently, molecular motions do not completely averag e tensor s of the anisotropic interactions ( CSA, dipole dipole, and quadrupolar interactions mentioned above ), which retain a spatial dependent component in the form of a second order Legendre polynomial 3cos 2 1, where is the angle between the external static magnetic field and the bilayer normal. (100) The spatial component varies between 1 and 1/2 (i.e. = 0 to 90 ). When is set to the magic angle, 54.7 cos 2 is equal to 1/3 and the spatial dependence component of these anisotropic NMR interactions becomes equal to zero (Fig. 2 1 0 ) Thus anisotropy can be almost completely removed from the spectr um by spinning at the magic angle with respect to the external magnetic field, which is beneficial for the resonance assignment as the anisotropy of the molecular interactions broadens the resonances in unoriented samples and decreases resolution leading t o spectra which are broad superpositions from all the contributions of the possible orientations of the lipid molecules in the sample with respect to the external magnetic field Another way to remove anisotropy is to orient lipid bilayers, but for many sa mples this is not feasible. However, in some cases retaining spatial dependence component of the anisotropic NMR interactions is crucial for analysis because it provides information about molecular orientation in the system; this is important for distingu ishing between lipid polymorphisms in our samples However, i n order to keep resolution high the orientational distribution for the NMR interactions and the anisotropies that define their
46 strengths need to be separated. The original broad powder spectrum i s described by the following two equations : (101) or (2 1 4 ) where g(x) is the anisotropy distribution function and ( ) is the orientation distribution function. In equation s 2 14 g(x) is a lineshape function for each anisotropy, such as for a single 31 P CS A in a pure phospholipid sample; ( ), is a superposition of spectra from the individual orien ted spectra of a powde r pattern and there is one orientation distribution function for each orientation, D ePaking can be performed in d ifferent ways : either extract g(x) when ( ) is known or opposite ly cal culate ( ) from the measured data whe n g(x) is known. For example, i n a perfectly spherical lipid vesicle bilayers adopt completely random orientations with respect to the magnetic field, but the function their orientation distribution around the sphere is known and it is proportional to sin the angle between the bilayer normal and the magnetic field Since is known for the sphere, t he anisotropy distribution function g(x) can be extacted from the spectrum of lipid vesicle using a standard inversion (dePaking) procedure (102) For samples with ellipsoidal deformation of the MLVs in which the lipid bilayers align to s ome degree in the magnetic field and are called partially aligned the orientational or probability distribution function becomes: (101)
47 (2 1 5 ) where k is the square of the ratio of the long to short axes of the ellipsoids. If k is equal to 1 the lipid vesicle is a sphere. If this value is >1 the shape is more ellipsoidal with its long axis along the external magnetic field. Deconvolution of spectra of MLVs with ellipsoidal deformation is obtained with an iterative procedure which simultaneo usly determines k and dePakes the spectrum. Since our lipid samples showed some degree of align ment in the magnetic field, this iterative procedure was utilized and dePaking of NMR data was accomplished with previously published algorithms which simultane ously dePake and determine macroscopic ordering in partially aligned lipid spectra using Tikhonov regularization (101, 102) Prior to dePaking 31 P NMR spectra were referenc ed to phosphate buffer, whereas dePaked spec tra were quantitated by fitting the two peaks with Lorentzian line shapes Assignments of 2 H resonances were made based on assignments of individual deuterons in lipid vesicles reported by Petrache, et al. (103) Solution State NMR Method of Protein Structure Determination A naly sis of the nuclear magnetic resonance spectra of the particular protein requires that most of the NMR resonances observed for this protei n in the given conditions are assigned to the specific atoms in the protein sequence. Resonance assignment is said to be completed for the protein if it is determined which chemical shift corresponds to which atom. It can be achieved for proteins using tw o main types of NMR experiments : one where magnetization is transferred through the chemical bonds, and the other where the transfer is through space, irrespective of the bonding structure. The first type is used to assign the different chemical shifts to specific nucle i and the second type is mainly used to generate the distance restraints used in the
48 following structure calculation In these experiments magnetization is detected for the specific isotope, usually proton, by irradiation of the sample place d into the static magnetic field with pulses of electromagnetic energy that have specific frequency in the radio diapason equal to the frequency of Larmor precession of this isotope. T hen the magnetization is transferred between different nuclei using dela ys determined by intrinsic internuclear couplings T he whole process is controlled by the pulse sequence of the experiment and the change in the pulse sequence allow s one to investigate and select specific types of connections between nuclei. Th e process o f assignment of each amino acid in the sequence is also called backbone The exact procedure depends on whether the protein is isotopically labelled or not (see below) (104) Protein Resonance Assignment The conventional assignment procedure for unlabelled protein involves a set of two dimensional homonuclear experiments : co rrelation spectroscopy (COSY) or total correlation spectroscopy (TOCSY) and nuclear Overhauser effect spectroscopy (NOESY) (105) Both COSY and TOCSY experiments transfer magnet ization through the chemical bonds between adjacent protons and these transfers are represented by off diagonal cross peaks between adjacent protons whereas protons themselves appear on the diagonal (Fig. 2 1 2 A) This interaction or spin spin coupling o f two nuclear spins I 1 and I 2 through the chemical bonds connecting them is called J coupling and can be expressed in the form I 1 I 2 The mechanism proposed in 1951 by N.F. Ramsey and E. M. Purcell states that J coupling can be represented as the sum of th e magnetic interaction between each nucleus and the electron spin of its own atom together with the exchange coupling of the electron spins with each other. I n COSY magnetization can only be transfer red between protons on adjacent atoms whereas in TOCSY i t is
49 transferred among all protons within a given spin system, even between distant protons as long as there are J couplings between every intervening proton. All possible COSY magnetization transfers f or valine are shown as red peaks on Fig. 2 1 2, A: HN H H HN, H H H H H H H H TOCSY connects all protons within valine spin system HN H H H (red and green peaks) and thus shows links from the backbone NH group to all side chain hydrogens of that residue giv ing a so called amide trace charac teristic of the valine sidechain structure Thus a so called spin system shown on Fig 2 1 2, A is characterisic for the valine structure Different types of protons have characteristic chemical shifts, which allows amino acids to have distinct patterns in t 1 2 B). Consequently, u sing a TOCSY spectrum the amino acid type can be identified or narrowed down significantly. (106) To connect the different spin systems in a sequential order, the NOESY experiment has to be used. In NOESY magnetization is transfer red through space, consequently, this experiment will show for each NH group all 1 H resonances which are within about 5 6 regardless of whether they are in the same spin system or not (Fig. 2 1 3, A) This transfer of magnetization or spin polarization known as Nuclear Overhauser Effect (NOE) happens directly from one population of nuclear spins to another through space rather than through electro ns spins of chemical bonds connecting them. The nuclear spin polarization is t ransfer red from one nuclear spin population to another via cross relaxation mechanism Thus, atoms that are in close proximity to each other can give a NOE, whereas spin coupling is observed only when the atoms are connected by 2 3 chemical bonds. The neighboring residues are
50 inherently close in space, so the sequential assignment is done on the assumption that the two neighboring NH groups are always visible. Thus two NH groups c an be linked because they each have an NOE to the other NH group. The additional presence of NH(i+1) NOE resonances and NOEs of side chain protons from i th residue to NH groups of i+1 residue can further confirm that the residues i and i+1 are sequen tially connected (Fig. 2 1 3, B). Sequential a ssignment is completed when the spin system s of i th and i+1 amino acid residues order through the strong NOESY peaks showing that HN, H and H of i th amino acid are very close in space to HN i+1 atom of the next amino acid. Helical sections are Having a rough idea of the secondary structure an d topology of the protein can thus significantly aid backbone assignment usin g 2D proton spectra only (106) These no n sequential NOE crosspeaks show how the protein amino acid sequence is folded in space and are used to derive distance restraints for protein structure calculation after sequential assignment is finished F or example, on Fig. 2 1 3, NOEs between H H of L27 and the amide proton of G32 are not sequential The use of h omonuclear 2D 1 H 1 H TOCSY and NOESY for sequence assignment is restricted to very small proteins or peptides because different protons have very similar chemical shifts and are located in the same area of the spectrum thus causing peak overlap The larg er the protein, the more significant is the problem of peak overlap. For small proteins less than 10 kDa the problem of peak overlap can be solved by 15 N labelling the protein If magnetization is sent from the amide protons t o the other
51 protons through the 15 N nucleus directly attached to the amide protons, then the one bond scalar J couplings between the amide protons and the attached nitrogens 1 J( 1 H 15 N) can be employed to separate 2D 1 H 1 H TOCSY and NOESY traces of differen t amino acids in the 3 rd dimension according to the 15 N chemical shift of amide nitrogen of each amino acid in the process of acquiring 3D 15 N NOESY HSQC and the 15 N TOCSY HSQC spectra 1 J( 1 H 15 N) couplings are ~92Hz and are much larger then homonuclear 1 H 1 H J couplings of a few Hz consequently separation between the TOCSY and NOESY planes along the 15 N dimension is much greater than any distance between the 1 H 1 H peaks in the planes Due to the reduction of overlap in 2D 1 H 1 H TOCSY and NOESY planes more peaks can be detected in each spectrum, which allows one to do better backbone assignment and generate more distance restraints for a more precise structure calculation P rio r to sequential assignment of the 15 N labelled protein it is convenient to measure a relatively quick heteronuclear 2D single quantum correlation (HSQC) spectrum 15 N HSQC each 1 H bound to a heteronucleus in this case 15 N gives one peak consequently, there is one signal for each amino acid residu e T he exception s to this rule are the proline, which has no amide hydrogen and certain other residues (arginine, lysine, glutamine, asparagine and tryptophan) with nitrogen containing sidechains that give rise to additional signals. 15 N HSQC spectrum all ows researchers to determine whether the expected number of peaks is present and thus to identify possible problems due to multiple conformations or sample heterogeneity This simple evaluation predicts the feasibility of doing subsequent longer, more labo rious and expensive experiments It is not possible to assign peaks to specific atoms from the heteronuclear s ingle quantum correlation alone,
52 consequently, 3D 15 N NOESY HSQC and the 15 N TOCSY HSQC spectra have to be acquired for the sequential assignment and structure determination. Although an assignment process using 3D 15 N experiments (TOCSY HSQC and NOESY HSQC) is an improvement over 2D methods for larger peptides/small proteins, the method also has drawbacks which are obvious for the proteins larger than 10 kDa : (107) 1. Still im possible to unambiguously assign all the side chain proton resonances. 2. With increasing protein size the 15 N edited TOCSY HSQC correlations will become weaker due to the longer correlation time and you m ay not get a complete transfer along the side chain, thus making amino acid typ e determination impossible 3. Assignment relies on HN X NOEs, which are conformation dependent and degeneracy in the 1 H dimension becomes a problem, especially for helical prote ins and larger proteins. The excellent resolution of 3D experiments coupled with efficiency of direct scalar coupling transfers via 13 C/ 15 N made t 1 H 13 C 15 N) assignment method the standard way to study proteins larger than 10 15 kD by NMR The large single bond heteronuclear couplings (e.g. 13 C O 15 N, 1 J=15Hz) give efficient transfer through the backbone when compared to small homonuclear 1 H couplings of several Hz Also C and C chemical shifts are quite different (C 65 ppm; C 180ppm) This allows the treatment of C and C carbon atoms as separate nuclei consequently, the particular magnetization transfer along the backbone (from N to CO or to C ) can be selected and manipulated independently This approach removed the
53 use of NOEs for making sequential assignments consequently, it has the advantage of being less ambiguous and more reliable. (107) Restraint Generation for Protein Structure Calculation After the protein sequential assignment is achieved, i.e. each chemical shift is assigned to each proton in the protein sequence, each crosspeak in a 1 H 1 H NOESY spectrum of the protein starts to signif y spatial proximity between the two specific protons with t he intensity of the cross peak decayin g as the distance between the protons increases. T he distance between these specific protons can be determined from the intensity of the cross peak between them because the i ntensity is proportional to the distance to the minus 6th power Thus each peak in NOESY can be converted into a maximum distance between the nuclei, usually between 1.8 and 6 to generate a list of experimentally determined distance restraints Distance restraints are the most widely used type of restraints in protein structure calcu lation The intensity distance relationship is not exact, so usually a distance range is used in the distance restraint for structure calculation For the protein structure calculation correct sequential assignment is of great importance because the NOESY cross peaks have to be assign ed to the correct nuclei based on the chemical shifts in order to generate meaningful distance restraints (104, 105) In addition to distance restraints, one can generate angular restraints for the prot ein psi and phi backbone torsion angles using two different experimental techniques Both approaches utilize the fact that different values of the psi and phi backbone torsion angles dictate certain geometry around the alpha carbon which in turn affects t he coupling constants and chemical shifts of the nearby atoms. Consequently, if the coupling constants 3 J(HA HN) or the chemical shifts of HN, 15 N,
54 HA, CA, CB are experimentally measured then the psi and phi backbone torsion angles can be predicted The f irst approach uses the Karplus equation (108, 109) to generate angle restraints from co upling constants between the amide and alpha protons : 3 J(HA HN) = Acos 2 ( ) + Bcos ( ) + C (2 1 6 ) where is the dihedral angle between the two vicinal protons for which the coupling is measured ; A B and C are empirically derived parameters The superscript "3" indicates that a 1 H atom is coupled to another 1 H atom three bonds away. Such hydrogens bonded to neighbouring atoms are termed vicinal. T he magnitude of these couplings is generally the the largest at angles of In the second approach psi and phi torsio n angles are search ed for in the database of proteins with known structures against measured chemical shifts using the TALOS+ algorithm (110) Protein Structure Determination by Simulated Annealing Approach The expe riment ally determined restraints are organized into lists and converted to the proper text format to be used as input to the computer programs such as XPLOR NIH (111, 112) for the structure calculation process. In general any structure determination program converts the experimental restraints and the known general protein properties (bond lengths, angles etc.) into separate energy term s E j exp (X) and E k chem i (X) correspondingly, which are functions of atomic coordin ates: (2 1 7 ) This conversion is performed in such a way that energy increases if the atomic coordinates deviate from the ideal positions where the specific type of restraint is completely satisfied, i.e. an s generated for violation of the particular
55 restraint. C ertain weight ing factors j and k are assigned to the different energy terms to describe the importance of each term for the structure determination : (2 1 8 ) Then the wh ole hybrid energy E hybrid is minimized in the process of searching for the ideal set of atomic coordinates, which would satisfy all of the restraints the best possible way, while a delicate balance is maintained in satisfying different restraints due to fi ne tuning their weight ing factors Once this ideal set of coordinates is found the energy be comes the lowest possible and the global minimum energy is achieved. T he optimization process start s from the unfolded structures (extended protein strands) and res ults in an ensemble of structures that will converge only i f the sufficient experimental data were provided to the program to dictate a certain fold The calculations have to start from extended strand s rather than from entirely random structures which ma y already be folded because once the polypeptide chain has folded incorrectly in a local minimum it is unlikely to converge to the correct global minimum region. (113) In the XPLOR program t he total target function F tot ( potential energy of the system) for which the global minimum region is searched is made up of the following terms: (112, 113) (2 19 ) F covalent is the target function for maintaining correct bond lengths, angles and planes, and is given by : (113) (2 2 0 )
56 where the force constants for the bond ( k b ), angle ( k ) and improper torsions ( k ) are set to uniform high values to ensure near ly perfect stereochemistry throughout the calculations The improper torsion terms serve to maintain planarit y and chirality. F repel is the target function used to prevent unduly close nonbonded contacts and is given by : (113) if r s r min (2 2 1 ) if r < s r min The values of r min are the standard values of the van der Waals radii as r epresented by the Lennard Jones potential used in the CHARMM empirical energy function; s is a van der Waals radius scale factor, and k r the van der Waals repulsion force constant. F N MR is a complex term whose functional form depends on the difference betw een the calculated and target value of a particular experimentally measured NMR restraint It measures how well the NMR observables calculated from a molecular structure match their experimental counterparts. (112) An a dditional term RAMA used in the structure calculations is a multi dimensional torsion angle database p otential term that is derived from the Protein Data Bank and bias es structures to wards existing features seen in the database. However, it is readily overridden by experimental restraints in cases of conflict. Ideally, only F N MR class of potential terms would be necessary in structure determination, but NMR restraints alone do not cont ain sufficient structural information. (112) In the Monte Carlo algorithm the temperature of the system is temporarily increase d in order to sear ch for the global minimum region of the target function. T he simulated annealing method in XPLOR achieve s the same effect by us ing Newton's
57 equations of motion to increase temporarily the temperature of the system in order to search for the global minimum region of a target function T he simulated annealing procedure involves the simultaneous integration of Newton's equations of motion: (113) (2 2 2 ) for all n atoms in the system whose temperature is given by : (113) (2 2 3 ) Integration of the classical equations of m otion is performed using a Verlet integration algorithm with initial velocities assigned to a Maxwellian distribution at an appropriate temperature. Unbiased sampling of the conformational space consistent with the restraints is achieved by varying the ran dom number seed used to assign the initial velocities. This ensures that the different trajectories diverge during the early stages of the simulations and only converge later as more and more experimental restraints are satisfied (113) In biomolecular NMR structure determination, many internal coordinates are known or presumed to take agreed upon values and i t is generally not desirable that these known coordinates be altered For instance, bond lengths and angles a re taken from high resolution crystal structures It is also true that the configuration space of internal coordinates to be searched in a structure determination calculation can be much smaller. For instance, the size of torsion angle space is approximate ly one third that of Cartesian coordinates for proteins. For these reasons, it is desirable to perform molecular dynamics and minimization in arbitrary internal coordinates such as bond stretching, bending, and torsion angles which are specified in the i nternal variable
58 module (IVM) in XPLOR (113) IVM can also be used to perform dynamics in the full Cartesian space T he simulated annealing calculations routinely performed i n XPLOR IVM usually use torsion angles as internal coordinate system. The common procedure consist s of performing molecular dynamics in torsion angle space starting at a high temperature, and then slowly decreasing the temperature in order to find the global minimum region. In Xplor NIH annealing protocols, potential parameters are generally ramped while the temperature is decreased, such that the potential energy is initially softer, with lower barriers. As simulated annealing progresses and the temperature is decreased, force constants are incre ased such that the potential takes its desired final form at the end of the annealing protocol. (113) The example of XPLOR NIH script for simulated annealing calculation is provided in Appendix A. Protein Structure Validation It i s important to a c knowledge the fact that the ensemble of structures obtained in a structure calculation is an experimental model, i.e. a representation of experimental data measured by NMR, where all measured restraints have been satisfied i n the generated protein structure However, such experimental models of protein structures can contain errors, and it is very important to detect these errors in a process known as validation. In general the quality of a model will depend on both the quant ity and quality of experimental data used to generate it and the correct interpretation of such data. One of the common structure validation programs is PROCHECK. ( 114) PROCHECK checks the stereochemical quality of a protein structure, producing a number of plots analyzing its overall and residue by residue geometry. E xamples of such validation plots for the structure of SP B 59 80 in methanol and in DPC micelles ar e provided in chapter 4 (Figs. 4 13 to 4 20 ).
59 The precision of a model indicates the variance of the measured data set under the same conditions. Precision of the structure ensemble derived using solution NMR is determined with root mean square deviation ( RMSD ) factor of different structures in the ensemble from the mean structure The more precise is the structural model the smaller is RMSD of the structure ensemble. The accuracy of the model indicates the degree to which a measurem ent approaches its true value. In practice for an NMR structure ensemble it is evaluated by the degree of agreement between the model and a set of experimental data as there is no "standard molecule" against which to compare models of proteins. Methods of Peptide Production Th e experiments reported in chapter 3 utilized wild type SPB 59 80 peptide that was prepared synthetically and purified using acetonitrile gradient and high performance liquid chromatography (HPLC). Then pure SPB 59 80 peptide was dissolved in methanol and rec onstitut ed with model lung surfactant lipid mixtures 4:1 DPPC:POPG and 3:1 POPC:POPG Production of 15 N SP B' 59 80 uniformly labeled version of the peptide, required for the experiments in chapter 4, can be achieved only via peptide expression in E.Coli us ing minimal media with 15 NH 4 Cl as the sole nitrogen source. Solid Phase Peptide Synthesis S olid phase peptide synthesis (SPPS) pioneered by Merrifield, (115) is now t he most popular technique for production of short peptides SPPS a llows not only the synthesis of natural peptides that are hard to express, but also the incorporation of unnatural amino acids and isotopically enriched residues modification of the backbone, and the synthesis of D proteins, which consist of D amino acids In SPPS technique the small, insoluble yet porous resin beads (supports) are treated with functional units
60 ('linkers') T hen the carboxyl group of the first amino acid is covalently coupled to the linker whereas its N terminus is deprotected and then co upled to the carboxyl group of the next N protected amino acid unit Thus t he C terminus of the growing peptide is attached to the solid phase and the peptide continuously elongates one residue at a time during the cycle of coupling new amino acid, washing away reactant, deprotecting the end of the growing peptide chain, and washing again Then the cycle coupling wash deprotection wash is repeated again until the desired peptide sequence is produced C he mical peptide synthesis starts at the C terminal end o f the pep tide and ends at the N terminus, which is the opposite of protein biosynthesis in ribosomes which starts at the N terminal end. The peptide remain s covalently attached to the bead until cleaved from it by trifluoroacetic acid. The superiority of this technique lies in the ability to keep the peptide immobilized and intact on the solid phase during filtration whereas liquid phase by products and excess reagent are wash ed away after each reaction and replaced by the reagents to perform the next org anic reaction (116, 117) The N termini of amino acid monomers are protected by Fluorenylmethyloxycarbonyl ( Fmoc ) groups before t hey are added onto a deprotected amino acid chain to avoid the possib le adverse reactions during synthesis (116) In addition m any amino acids also have reactive side chain functional groups, which can interact with free termini or other side chain groups during peptide synthesis adversely affecting final peptide yield and purity. The specific s ide chain protecting groups used during the peptide synthesis are usually based on the benzyl (Bzl) or tert butyl (tBu) group (118) The s ide chain protecting groups are known as permanent or semipermanent prot ecting groups, because they can
61 withstand the multiple cycles of chemical treatment during synthesis and are only removed with strong acids after peptide synthesis is completed. Fmoc protection was used during SPPS of SPB 59 80 peptide for experiments in c hapter 3 with Fmoc group protect ing the amino group and resin linkage agents. The steps of Fmoc SPPS can be summarized as follows : (116, 117) 1. The Fmoc protected amino acid is attached to the resin via a linker. 2. The Fmoc protecting group is removed with piperidine in dimethylformamide (DMF) to deprotect the residue. 3. T he carboxyl group of t he next Fmoc protected amino ac id is activated with carbodiimide and coupled to the amino acid linked to the solid phase support at basic pH The coupling reaction is usually performed twice to insure optimal yield. 4. The deprotection/coupling cycle is repeated to yield the desired a mino acid sequence. 5. The linker/resin support and side chain protecting groups are cleaved with TFA, yielding a free peptide. Then the peptide is purified with HPLC. The prominent features of the above SPPS reaction steps are outlined in Figure 2 1 4 H eterologous Peptide Expression in E.Coli B asic science studies as well as clinical trials of synthetic LS formulations require l arge quantities of the peptides Peptide i solation from natural sources or their chemical synthesis are generally not cost effec tive, whereas a relatively cheap recombinant approach provides not only easy scale up for large scale peptide production but also the possibility of production of larger peptides and uniform isotopic labeling which becomes too expensive in traditional sol id state chemical synthesis even of relatively
62 short peptides (119) Escherichia coli is the most widely used host for the p eptide expression due to its fast growth rate low cost and the fact that peptides genera lly do not need post translational modification to be active T he fact that the most stable and hence abundantly produced p roteins in E. coli are usually bacterial in origin and protease resistant is important for the peptide e x pression as well Thus to ac hieve the best yields peptide coding sequences have to be codon optimized for expression in E. Coli and peptides are often expressed as fusion proteins, which protects the m from proteolytic degradation. This was a n important breakthrough in the biosyntheti c peptide production since a ttempts to produce small polypeptides of less than 100 amino acids generally failed unless these peptides were part of a fusion protein. (120) Soluble h ydrophilic carriers such as t hioredoxin, glutathione S transferase ( GST ) and maltose binding proteins (MBP), allow quick purifi cation from crude lysate by affinity chromatography using carrier protein as an affinity tag to purify the fusion protein H ighly hydrophobic fusion partners, such as ketosteroid isomerase (KSI) enhance the stability of peptides by driving them into the insoluble inclusion bodies which not only prevent s proteolytic degradation of the fusion complex but also a llows efficient separation from the soluble host proteins (119) Since many recombinant proteins are nonfunctional when tethered to a carrier protein, several cleavage methods have been developed in order to cleave the desired protein from the carrier protein. Site specific proteases such as thrombin or factor X, preserve the protein native structure, but often do not cleave quantitatively at the anticipated site. Chemical cleavage by CNBr, which cleaves quantitativ ely and specifically at methionine residues, require s strongly acidic conditions, which are
63 incompatible with the survival of native protein structure. Consequently, c hemical cleavage by CNBr can only be used for t he production of small polypeptides of les s than 100 amino acid residues since their native structure preservation is not require d during purification (120) Also chemical cleavage is not compatible wi th some fusion carriers, for example GST because the non specific cleavage at the carrier will substantially complic ate the subsequent purification (119) All these concepts have been applied in the method devel oped by Kuliopulos and Walsh (120) which utilizes extremely hydrophobic ketosteroid isomerase (KSI) as the carrier protein of fusion constructs that comprise multiple copies of the target peptides in addition to a His tag sequence His tag is used for the affinity purification of the fusion construct on Ni column under denaturing conditions which is the only possible way of purification of a highly hydrophobic protein T he highly hydrophobic KSI is used to drive t he fusion protein into protease resistant inclusion bodies. This strategy requires that methionine (Met) residues be included onto the sequence as separators of each individual component of the fusion protein (i.e., KSI, target peptides, and His tag) in order to allow for excision of target peptides by CNBr polypeptide digestion. Final recovery of the target peptides is accomplished through chromatographic separation of the fr agments in the cleavage mix ture by HPLC (121) In the c ommercially available plasmid pET31 b (122) designed by Novagen ( Fig. 2 1 5 ) the transcription of the fusion complex KSI Met [(peptide) Met] n His 6 is set under regulation by a s trong bacteriophage T7 RNA polymerase promotor T7 RNA polymerase gene is placed und er lacUV5 control, and T7 RNA polymerase synthesis is induced by IPTG which in turn induces fusion complex e xpression in the host cell T7 RNA polymerase is so selective and
64 erted to target gene expre ssion (123) The scheme of peptide expression and purification using pET31 expression system presented on Fig. 2 15 is followed by the necessary chromatograms and analyses of peptide purity on F igs. 2 16 to 2 20. Peptide Purification by High Performance Liquid Chromatography (124) Solid phase peptide synthesis and recombinant DNA techniques have allowed the production of large quantities of peptides and proteins which need to be highly purified. A method of choice for the isolation of pep tides and proteins from a wide variety of synthetic or biological sources is High Performance Liquid Chromatography (HPLC) because it provides the excellent resolution under a wide range of conditions for very closely related molecules, as well as structu rally quite distinct molecules. Reverse phase HPLC ( RP HPLC ), where peptides are separated by their hydrophobicity, is the most commonly used mode of peptide separation, although other techniques also find application, for example, ion exchange and size ex clusion chromatography which are based on molecule surface charge or size respectively In the case of synthetic peptides, RP HPLC is generally employed both for the initial analysis during synthesis and the final large scale purification. The isolation o f peptides and proteins from a biological cocktail, however, often requires a combination of techniques to produce a homogenous sample and HPLC techniques are introduced at the later stages following initial precipitation, clarification and preliminary sep arations. The separation in RP HPLC technique depends on the hydrophobic binding of the solute molecule from the mobile phase to the immobilized hydrophobic ligands attached to the stationary phase, i.e., the sorbent. The solute mixture is initially appli ed to the sorbent in the presence of aqueous buffers, and the solutes are eluted by the addition of
65 organic solvent to the mobile phase. Elution can proceed either by isocratic conditions where the concentration of organic solvent is constant, or by gradie nt elution whereby the amount of organic solvent is increased over a period of time. The solutes are, therefore, eluted in order of increasing molecular hydrophobicity RP HPLC is more commonly employed for the isolation of peptides rather than proteins, b ecause its harsh conditions cause the irreversible denaturation of protein and do not allow recover y of the protein in a biologically active form. In RP HPLC peptides are usually retained on n alkylsilica based sorbent and eluted with gradients of increasi ng concentrations of organic solvent such as acetonitrile containing an ionic modifier such as trifluoroacetic acid (TFA). Separations can be easily manipulated by changing the gradient slope, the operating temperature, the ionic modifier, or the organic s olvent composition. (124) Figure 2 1 Far UV CD spectra associated with various types of secondary structure. Solid line, a helix; long dashed line, anti parallel h sheet; dotted line, type I h turn; cross dashed line, extended 3 1 helix or poly (Pro) II helix; short dashed line, irregular structu re. Figure borrowed with permission from: Kelly, S. M., Jess, T. J., and Price, N. C. (2005) How to study proteins by circular dichroism, Biochimica Et Biophysica Acta Proteins and Proteomics 1751 119 139.
66 Figure 2 2 Chemica l structures of glycerophospholipids. The polar head groups vary in their size and charge: zwitterionic head groups phosphocholine (PC) and phosphethanolamine (PE), and the anionic head group phospho glycerol (P G ). The nonpolar acyl chains vary in their len gth and degree and position of unsaturation. Figure 2 3 Gel to liquid crystalline phase transition of the bilayer lamellar phase. When melting above the temperature T m phospholipids obtain more degrees of freedom and more motions in their acyl chains as well as more axial rotation compared to the gel phase.
67 Figure 2 4 Types of lipid phases. Hexagonal phase is shown in its normal and inverted types; bilayer possesses ordered gel phase which melts into disordered liquid crystalline phase upon heating Borrowed and a dapted with permission from : Koynova, R., and Caffrey, M. (2002) An index of lipid phase diagrams, Chemistry and Physics of Lipids 115 107 219. (84)
68 Figure 2 5 Origins of the powder pattern obtained in the presence of c hemical shift anisotropy (CSA) interaction The CSA interaction results in a powder pattern due to the distribution of populations of orientations the lipid bilayers adopt. The CSA is defined by a tensor with elements 11 22 and 33 in lipid bilayers that are time averaged to and c omponents This is because the average tensor is oriented with respect to the bilayer normal. With solid state NMR you can obtain orientation information that is lost in solution NMR. Each part of the powder pattern is related to a certain orientation of the bilayer and each orientation leads to a different frequency. With spherical vesicles you have more perpendicular th an parallel orientations. This is illustrated by the colored spheres showing chances are higher for an angle of rotation of bilayer normal that is 90 relative to the external magnetic field (B 0 ). As the angle decreases, chances of the lipid orienting at t hat angle in the magnetic field leads to a lower intensity. Figure adapted with permission from : Levitt, M. H. (2008) Spin dynamics: Basics of nuclear magnetic resonance 2 ed., Wiley
69 Figure 2 6 C hemical shift anisotropy of the rigid lipid phosphate group is a measure of the degree of order and characterizes the types of the motions that are experienced by the group in the lipid phase The shape and the width of the pattern are strictly dependent on the orientation of the axis of motional averaging re lative to principal components of the chemical shift tensor 11 22 33 (89, 90) A: F or microcrystalline sample, where all angles between the applied magnetic field and principal axes are populated, the static p ow der pattern is the sum of spectr a for all possible orientations and has three different 11 22 33 components. B: Rapid anisotropic motion around normal to the lipid bilayer does not change 11 as it is parallel to the axis of motional averaging but av erage s the other two components of the chemical shift tensor 22 33 that correspond to the axes perpendicular to the axis of rotation. C: If 11 axis of the phosphate moiety is also involved into rapid rotation, then there is further averaging of and components into and effective components and the effective tensor still has axial symmetry, but the total chemical shift anisotropy is reduced. Borrowed and a dapted with permission from : Smith, I. C. P., and Ekiel, I. H. (1984) Phosphorus 31 NMR: Principles and applications Elsevier Inc.
70 Figure 2 7 Polymorphisms and phosphorus NMR lineshapes. T he lineshapes for 31 P NMR spectra correspond to different phases or polymorphisms. The phase is shown on the left and its corresponding spectrum on the right. Several phases can result in an isotropic peak. Additional lipid motion further averages 31 P CSA and makes the lineshape narrower. Borrowed and a dapted with permission from: Pfeiffer H ., Weichert H ., Klose G ., and Heremans K (20 1 2) Hydration behavior of POPC/C 12 Bet mixtures investigated by sorption gravimetry, 31 P NMR spectroscopy and X ray diff raction Chemistry and Physics of Lipids 1 6 5 244 2 51 (84)
71 Figure 2 8 Effect of the quadrupole interaction on Zeeman energy levels of nucleus with spin I=1 with axial symmetry. The quadrupolar splitti ng corresponds to the difference in the transition frequencies of the single quantum transitions, and is due to the perturbing interaction of the 2 H nuclear quadruple moment with the electric field gradient (EFG) of the C 2 H bond. Adapted with permission f rom : Brown, M. F., and Nevzorov, A. A. (1999) 2 H NMR in liquid crystals and membranes, Colloids and Surfaces a Physicochemical and Engineering Aspects 158 281 298.
72 Figure 2 9 The angles used to define the motional averaging in lipid bilayers by relating the principal axis of the EFG tensor ( usually the direction of C 2 H bond ) to the applied magnetic field H o direction between the C 2 H bond vector and the instantaneous segmental chain t he angle between the instantaneous segmental chain orientation and the director of the motion, taken to be the normal to the bilayer field direction. Adapted with permission from: Dufourc, E. J., Smith, I. C. P., and Jarrell, H. C. (1983) A 2 H NMR analysis of dihydrosterculoyl containing lipids in model membranes structural effects of a cyclopropane ring, Chemistry and Physics of Lipids 33 153 177 C C 2 H C n B 0
73 Figure 2 10 Origin of the powder pattern of 2 H NMR spectra observed for deuterium nuclei incorporated into phospholipid tails in membranes in uni and multilamellar vesicles (LUVs and MLVs) Different angles between magnetic field H o and the axis of m otional averaging (dashed arrow) lead to different quadrupole splittings. Since membranes rotate slowly on the timescale of 2 H NMR splittings, all these subspectra will contribute. Figure adapted with permission from: Levitt, M. H. (2008) Spin dynamics: Ba sics of nuclear magnetic resonance 2 ed., Wiley
74 Figure 2 1 1 Solid state NMR provides residual quadrupolar couplings (RQCs) that directly correspond to ensemble averaged molecular structure. There are two spin transitions in 2 H NMR that lead to two powder patterns that overlap resulting in a characteristic pattern for deuterium nucleus in solid state NMR (A) PO PC lipid with multiple deuterated sites at sn 1 position and different motions of 2 H atoms at different carbon positi ons demonstrates broad powder pattern spectrum with several overlapping 2 H lineshapes ( B ) where e ach 2 H lineshape has different quadrupolar splitting Q which is affected by lipid mobility at the site Numerically deconvoluted (de Paked) 2 H NMR spectrum ( C ) reveals quadrupolar splittings Q ( i ) for individual methylene segments ( i ) of acyl chains because dePaking transforms powder lineshapes into indi vidual frequencies As the quadrupolar splitting increases the acy l chain order increases as well with the terminal methyl being most mobile or disordered. Order parameter profile (D) shows order parameter calculated from Q ( i ) for 2 H at each carbon posit ion in sn 1 acyl chain
75 Figure 2 1 2 Chemical shift pattern of different amino acid s in 2D 1 H 1 H TOCSY spectrum. A: The schematic of the whole TOCSY spectrum for valine amino acid that contain s backbone amide alpha, beta and g amma protons. B: Schematic representation of the amide TOCSY traces for all amino acids. Plotted using chemical shift values from table 2.3 in Wuthrich, K. (1986) NMR of proteins and nucleic acids John Wiley & Sons, Inc., USA.
76 Figure 2 A: Alternating TOCS Y (plain circles) and NOESY ( circles with blue edge ) amide traces of amino acids show that HN, H and H of amino acid i are sequentially connected through space to HN at om of the next amino acid i+1 via strong NOE correlations B (next page) : for SP B 59 80 in methanol with initial assignments shown with green arrows. Amino acids give different amide traces in TOCSY spectrum. NOESY spectrum contains the same peaks as in TOCSY and additional NOE peaks between the protons, which are close in space, but not connected by bonds inside one amino acid spin system A
77 Figure 2 13. C ontinued B
78 Figure 2 1 4 A summary of solid phase peptide synthesis with Fmoc N terminus protection group. 1) Fmoc protected amino acid attached to a resin via linker is deprotected using piperidine. 2) The next Fmoc protected amino acid is activated with carbodiimide and 3) coupled to the growing peptide chain. 4 ) The deprotection/coupling cycle is repeated. The desired amino acid sequence is deprotected and cleaved from the resin with TFA to yield a free peptide.
79 Figure 2 1 5 pET 31b expression system designed by Novagen and SPB C peptide purification scheme (122) The pET 31b vector is designed for clonin g and high level expression of peptide sequences fused with the 125 amino acid KSI protein upstream and His 6 tag downstream of the peptide sequence ; peptide is separated by methionine codons from tags (123)
80 Figure 2 16 Overexpression of 18 kDa KSI SPB 59 80 His 6 fusion construct in BL21(DE3)pLysS cells monitored as function of time after overexpression induction Lane #1 sizes of the protein standards in kDa, #2 0 hours after induction #3 2 hours, #4 4 hours, #5 5 hours, #6 19 hours A B Figure 2 1 7 Nickel affinity chromatogra m of purification of KSI SPB 59 80 His 6 fusion construct from inclusion bodies in denatur ing conditions (A); extremely hydrophobic KSI SPB 59 80 His 6 fusion construct after dialysis against water precipitates in the dialysis bag (B). 200 116 97 66 45 31 21.5 14 6.5 1 2 3 4 5 6
81 Figure 2 18 Cleavage efficiency of KSI SPB 59 80 His 6 fusion construct by cyanogen bromide in 80% formic ac id was monitored as a function of time : t 0 is 0 hours after overexpression induction, t 1 is 2 hours, t 2 is 3 hours, t 3 is 19 hours, t 4 is 2 2 hours t 5 is 23 hours Sizes of the protein standards (PS) are shown on the right side in kDa. Figure 2 1 9 Pure SPB 59 80 peptide in RP HPLC fractions after lyophilization Peptide was loaded at different concentrations to estimate its quantity. t 0 t 1 t 2 t 3 t 4 t 5 PS PS PS 26.6 16.9 14.4 6 .5 3.5 PS SPB 59 80 SPB 59 80 SPB 59 80 PS SPB 59 80 26.6 16.9 14.4 6 .5 3.5 26.6 16.9 14.4 6 .5 3.5
82 Figure 2 20 Mass spectr a of expressed SP B 59 80 (Top) wild type SP B 59 80 (Bottom) and 15 N SP B 59 80 (next page) peptides with t heoretical m/z 2578 2531 and 2610 respectively Mass (m/z) Mass (m/z) % Intensity % Intensity Peak 1 2 3 4 m/z 2538 2561 2577 2596 Peak 1 2 3 4 5 6 7 8 9 m/z 860 865 876 1139 1252 1439 1672 2513 2530
83 Figure 2 20 Continued. Mass (m/z) % Intensity Peak 1 2 3 4 5 6 7 8 9 10 11 12 m/z 795 806 861 866 877 909 1044 1124 1319 1141 2595 2612
* previously published in full version as: Antharam, V.C. Farver, R.S. Kuznetsova, A. Sippel, K.H. Mills, F.D. Elliott, D.W. Sternin, E. Long, J.R. (2008), Interactions of the C terminus of lung surfactant protein B with lipid bilayers are mod u lated by acyl chain saturation BBA 1778 2544 2554 The data was also partially published in dissertation by Antharam, V.C. (2008), Biophysical characterization of peptide mimics of lung surfactant protein B. CHAPTER 3 INTERACTION OF THE C TERMINUS OF SP B WITH LIPID BILAYER S MIMICKING LUNG IS M ODULATED BY ACYL CHA IN SATURATION To date, KL 4 that was developed based on the hydrophilic/hydrophobic pattern in SP B 59 80 is the only synthetic replacement for SP B that has enjoyed the most clinical success. (71 75) However, their primary sequences have only modest similarity There are only posi tively charged lysines in KL 4 each separated by four leucines, whereas SP B 59 80 additionally has negatively charged terminal amino acids and variable spacing of three o r four hydrophobic residues between polar or charged residues. Consequently, a significant question remaining is whether KL 4 and SP B 59 80 act similarly, despite clear deviations at the primary amino acid level. These differences could lead to varying seco ndary structures and penetration into lipid lamellae as well as different effects on phospholipid dynamics. One key to understanding how similar KL 4 is to SP B 59 80 in its behavior is to examine the binding of SP B 59 80 to model lipid membrane systems, its effects on their dynamics, and any effect in DPPC lipids; these results can then be directly compared to similar previous studies of KL 4 The study of the effects of SP B 59 80 on 4:1 DPPC:POPG and 3:1 POPC:POPG MLVs (68) w as included in this dissertation since it provided important preliminary results, which determined the course and the main goals of this dissertation. My role was to express and purify SP B' 59 80 with isoleucines incorporated in lieu of methionines to demo nstrate that the mutations cause no effects on lipid dynamics and the results are
85 identical to those of the native sequence which were obtained by other authors in the paper Lipid composition of 4:1 DPPC:POPG is similar to formulations commonly used in s tudying lung surfactant, while 3:1 POPC:POPG is a paradigm lipid system commonly employed to probe peptide/lipid interactions, particularly in studies of cationic, amphipathic helices (126 128) Lipid phases of these compositions could also be found in localized areas of the alveoli during the breathing cycle (129) Materials and Methods Synthesis of SP B 59 80 SP B 59 80 (DTLLGRMLPQLVCRLVLRCSMD) was synthesized via solid phase p eptide synthesis on a Wang resin (ABI 430, ICBR, UF) and cleaved from the resin with King's reagent and ether precipitated. The cleaved product was purified via RP HPLC using a C18 Vydac column with a water/acetonitrile gradient (containing 0.3% TFA). The fractions corresponding to SP B 59 80 were collected and the p eptide was verified by mass spectrometry with a mass to charge ratio of (m/z) of 2533. Dried peptide was weighed and dissolved in methanol to a stock concentration of approximately 1 mM, and aliq uots were analyzed by amino acid analysis for a more accurate determination of concentration (Molecular Structure Facility, UC Davis). Heterologous Expression of a SPB 59 80 Double Mutant (SP B 59 80 ) For later experiments, SP B' 59 80 was expressed using a pET31 construct (EMD Biosciences, Inc., Gibbstown, NJ) incorporating a codon optimized synthetic gene for SP B 59 80 (DNA2.0, Menlo Park, CA) in BL21(DE3) cells with subsequent purification and cleavage using established protocols (120) The expressed sequence of SP B' 59 80 was modified to incorporate isoleucines in lieu of methionines (DTLLGRILPQLVCRLVLRCSID) for compatibility with a cyanogen bromide cleavage
86 re action. Following HPLC purification of the final product, fractions corresponding to SP B' 59 80 were collected and the peptide was verified by mass spectrometry with a mass to charge ratio (m/z) of 2497. Dried peptide was weighed and dissolved in methanol to a stock concentration of approximately 1 mM, and aliquots were analyzed by amino acid analysis for a more accurate determination of concentration (Molecular Structure Facility, UC Davis). This peptide was used to obtain results shown in Figs. 3 1, 3 5, 3 6, and 3 11 as well as the POPG d31 results in Figs. 3 9 and 3 10. Its effects on lipid dynamics were ascertained to be identical to those of the native sequence. Preparation of Peptide:Lipid Samples POPC, DPPC, POPG, POPC d31, DPPC d62 and POPG d31 were purchased as chloroform solutions (Avanti Polar Lipids, Alabaster, AL) and concentrations were verified by phosphate analysis (130) (Bioassay Systems, Hayward, CA). The lipids were mixed at a molar ratio of 4:1 DPPC:POPG and 3:1 POPC:POPG in chloroform and aliquoted. For samples containing peptide, a methanol solution of SP B 59 80 was added to lipid solutions with final peptide: lipid (P:L) molar ratios ranging from < 1:1000 to > 1:50. The samples were dried under a stream of n itrogen with the sample temperature maintained at 42 50 C in a water bath; the resulting films were suspended in cyclohexane, flash frozen, and lyophilized overnight to remove residual solvent. CD Experiments 3 mg of peptide lipid powder was solubilized in 1 mL of 10 mM HEPES buffer at pH 7.4, with 140 mM NaCl, to achieve a concentration SP B 59 80 with 4 mM water bath to facilitate solubilization accompanied by 3 5 freeze thaw cycles with vortexing to achieve equilibration. Peptide lipid MLVs were extruded through 100 nm filters (Avanti Polar Lipids, Alabaster, AL) 15
87 25 times above the Tm of the lipids to form LUVs just prior to CD analysis. CD experiments were performed on an Aviv Model 215 (Lakewood, NJ) at 45 200 260 nm wavelength range, a 1 nm step size and averaging of 40 50 sc ans. Background contributions from the buffer and LUVs were removed by subtracting appropriate controls. Solid state NMR Analysis For each solid state NMR sample, 30 mg of peptide lipid powder was placed in a 5 containing 5mM HEPES at pH 7.4, 140mM NaCl, and 1mM EDTA in 2 H depleted water (Cambridge Isotopes, Andover MA) was added. Samples were made using 4:1 DPPC d62:POPG, 4:1 DPPC:POPG d31, 3:1 POPC d31 :POPG, and 3:1 POPC:POPG d31 lipid preparations. NMR samples were then subjected to 3 5 freeze thaw cycles with gentle vortexing to form MLVs. 31 P and 2 H NMR data were collected on 500 and 600 MHz Bruker Avance systems (Billerica, MA) using standard 5m m BB O probes. For 31 P NMR experiments, 25 kHz proton decoupling was employed during acquisition to remove dipolar couplings. Spectra were acquired at 34, 39, and 44 1024 2048 scans and a 5 s recycle delay between scans to minimize RF sample heating. For the 2 H NMR experiments, data were collected 90 field of 40 kHz. Spectra were acquired at 34, 39, and 44 with 1024 or 2048 scans and a 0.5 s recycle delay bet ween scans. To monitor the phase transitions of the lipids, for some samples spectra were also collected over a range of 30 44 DePaking of NMR data was accomplished with previously published algorithms which simultaneously dePake and de termine macroscopic ordering in partially aligned lipid spectra using Tikhonov regularization (101)
88 31 P NMR spectra were referenced to phosphate buffer prior to dePaking and dePaked spectra were quantitated by fitting the two peaks with Lorentzian line shapes. Assignments o f 2 H resonances were made based on published values for DPPC (103, 131) Results Secondary Structure The CD spectra at 45 B' 59 80 incorporated into lipid vesicles are shown in Fig. 3 1. The CD spectra are cha racterized by double minima at 206 208 and 222 nm, as is typically seen for peptide helices. Interestingly, the spectra have a minimum ellipticity at 206 208 significantly lower than the minimum at 222 nm. Similar spectra have been observed for KL 4 in DPPC :POPG LUVs (132) and peptides which are helices in buffer (133) Interpretation of the CD spectra in terms of helix content is complicated by the fact that SP B 59 80 does not f orm a typical helix when projected on a helical wheel. Fitting of the CD data with standard deconvolution software (134) led to secondary structure estimates of 75 helix, 17 shee t for SP B 59 80 interacting with POPC:POPG LUVs. Secondary structure estimates of 98 helix with negligible sheet populations were seen for SP B 59 80 interacting with DPPC: POPG LUVs. However, the quality of the fits was poor due to the nonstandard shape of the CD spectra and poor quality of the data below 200 nm due to light scattering from the lipids. The Interaction of SP B 59 80 with Lipid Headgroups 31 P solid state NMR for 4:1 DPPC:POPG and 3:1 POPC:POPG MLVs with varying levels of SP B 59 80 were collected to assess the effect of this peptide on lipid phases and the orientation of the phospholipid headgroups. Shown in Fig. 3 2 are 31 P spectra
89 for the DPPC d62:POPG samples at 44 allowing a clea r determination of the individual, time averaged chemical shift anisotropies (CSAs) of the PC and PG headgroups. Only lineshapes consistent with lamellar phases are observed and the resonance for the POPG lipids moves with addition of peptide; this is more evident after dePaking the spectra. As seen in other studies (132, 135, 136) spontaneous macroscopic lipid alignment occurs in the magnetic field. Due to this phenomenon, normally spherical MLVs undergo a deformati on to more ellipsoidal geometries, distorting the powder spectra. The extent of magnetic field alignment was accounted for in the dePaking algorithm (101) The difference in PG and PC CSAs is due to the difference in their preferred headgroup orientations relative to the mem brane normal (89) With increasing levels of SP B 59 80 the PC CSA is invariant, but the PG CSA lessens with increasing peptide levels. The CSA for POPG alone is significantly smaller than observed for POPG in 4:1 DPPC:POPG or 3:1 POPC:POP G mixtures prior to the addition of SP B 59 80 (132) However, for the binary 4:1 DPPC:POPG mixture containing higher concentrations of SP B 59 80 the 31 P CSAs are more comparable to those of the neat lipids. Thus, SP B 59 80 is clearly affecting the interactions of the PC and PG lipids. Shown in Fig. 3 3 are 31 P spectra for the POPC d31:POPG samples at 44 MLVs, addition of SP B 59 80 leads to a considerably smaller dec rease in the PG CSA; in contrast, KL 4 which has similar effects on the PG CSA in both DPPC:POPG and POPC:POPG mixtures (132) Based on these results, when POPG is interacting with PC headgroups, electrostatic interac tions cause the PG headgroups to reorient leading to a subsequent increase in their averaged CSA values. Addition of surfactant peptide
90 disrupts this interaction, causing the PG CSAs to move toward values seen for POPG alone; if the surfactant peptide is h ighly cationic, electrostatic interactions lead to similar decreases in the POPG CSA regardless of whether the PC lipid is fully saturated or monounsaturated (132) In the case of SP B 59 80 the negatively charged am ino acids at the termini result in less electrostatic interaction between the peptide and the PG headgroup. However the partitioning of the peptide into the DPPC:POPG lipids is sufficient to significantly alter the orientation of the PG headgroup due to it s overall effects on lipid packing and dynamics; in POPC:POPG lipids it is not. The 31 P NMR spectra for 3:1 POPC d31:POPG MLVs containing varying concentrations of SP B 59 80 are typical for lamellar phases up to 1.7 mol% peptide (Fig. 3 3 ). However, the sp ectrum at 2.5 mol% peptide suggests the onset of other bulk dynamics in the lipids, consistent with exchange between a lamellar phase and either a micellar phase or a hexagonal phase. With the onset of a phase transition at 2.5 mol% peptide, no clear disti nction between the PC and PG lipids can be made. 2 H NMR Studies of DPPC:POPG Lipid Miscibility on Addition of SP B 59 80 DSC thermograms for 4:1 DPPC d62:POPG LUVs containing varying levels of SP B 59 80 i ndicate SP B 59 80 has l ittle e ffect on DPPC:POPG Lipi d Miscibility ( Fig. 3 4 ; data by Antharam, V.C. ) with the phase transition mov ing to slightly higher temperatures with increasing peptide concentration. In contrast, KL 4 has been shown by DSC and fluorescence microscopy to promote lipid phase separation an d domain formation (132, 137, 138) The relative differences between KL 4 and SP B 59 80 are not completely unexpected as KL 4 contains solely hydrophilic residues that are cationic while SP B 59 80 contains a mixture of cationic and anionic residues. Thus, partitioning of KL 4 with POPG
91 may be more energetically favorable, enhancing phase separation in DPPC/POPG mixtures below the phase transition temperature of DPPC. To determine more precisely whether addition of SP B 5 9 80 causes phase separation of the DPPC and POPG lipids, 2 H NMR spectra were collected over the temperature range of the phase transitions for samples which contained either deuterated DPPC or deuterated POPG (Fig. 3 5). A first moment analysis of the dat a was used to determine the phase transition temperatures of the individual lipids (Fig. 3 6). From these spectra, it is clear that in the 4:1 DPPC:POPG samples the lipids are fully miscible with the DPPC and POPG melting at similar temperatures. The phase transition seen for deuterated DPPC is at a slightly lowe r temperature (midpoint of 30.6 percentage of the fatty acyl chains are deuterated (80% vs. 10%) (139) but may also reflect slight differences in POPG content between the two samples since they were made from different lipid stock solutions. For the samples that do not contain SP B 59 80 the spectra at intermediate melting temperatures are a superposition of gel phase and liquid phase spectra. Addition of SP B 59 80 affects the phase transition temperature of both lipids with the phase transition temperature for DPPC d62 increasing to 31.5 the POPG d31 phase transition temperature decreasing to 30.3 SP B 59 80 at lower temperatures the POPG d31 spectra have a larger percentage of the dynamic lipid phase evident and in the DPPC d62 spect ra the dynamic phase is absent, as can be seen by co peptide also leads to an increase in temperature at which the lipids are deuterated DPPC; 36 38 for deuterated POPG), con sistent with the DSC data.
92 Attempts to determine the fractions of liquid phase and gel phase by spectral subtractions with samples containing differing percentages of SP B 59 80 were unsuccessful due to variations in the lipid alignment and acyl chain order parameters. Thus, addition of SP B 59 80 leads to some phase separation of the lipids at lower temperatures with a significant fraction of POPG remaining in the gel phase; at higher temperatures, the lipids exhibit similar spectra consistent with the lipid s becoming fully miscible. These differences are too subtle to be observed via differential scanning calorimetry. Above the phase transition temperature of the domain with the higher Tm, the peaks in the 2 H spectra coalesce to a single resonance for each p osition in the acyl chains suggesting the absence of separate domains in the fluid phase or exchange between domains is fast on the NMR time scale. Effects of SP B 59 80 on Lipid Acyl Chains 2 H NMR spectra of DPPC d62:POPG and POPC d31:POPG MLVs at 44 h varying levels of SP B 59 80 (Figs. 3 7 and 3 8) allow determination of the effect of this membrane active peptide on lipid dynamics in the fluid phase and insight into the depth of peptide penetration. For 3:1 POPC d31:POPG MLVs, decreases are seen in th e ordering of the POPC sn 1 acyl chains on addition of SP B 59 80 This suggests the peptide binds to the interface, increasing the lateral spacing between the lipids, allowing more motion in the acyl chains. In contrast, addition of SP B 59 80 to 4:1 DPPC d 62:POPG MLVs increases the ordering of the DPPC acyl chains, particularly toward the middle of the bilayer. This suggests either 1) insertion of SP B 59 80 deeply into the DPPC:POPG bilayers, restricting the motional freedom of the acyl chains, or 2) an ele ctrostatic interaction of the peptide with the lipid headgroups resulting in a change in the lipid packing and headgroup conformation. Also shown in Figs. 3 7 and 3 8 are the
93 dePaked spectra which allow assignment of each C D bond in the acyl chain and det ermination of the time averaged order parameters, < S CD > For samples in which the lipid bilayers show a tendency to align in the magnetic field, the dePaking must take into account the effects of lipid alignment on the probability distribution function. As suming the magnetic field leads to an ellipsoidal deformation of the MLVs, the probability distribution becomes : (101) ( 3 1) where refers to square of the ratio of the long to short axes of the ellipsoids. Using Tikhonov regularization, and the dePaked spectra were determined simultaneously. Lipid acyl chain order parameter profiles (140) were generated (Figs. 3 9 and 3 10) by assigning the dePaked spectra, measuring the quadrupolar splitting for deuterium atoms at various positions along the acyl chain, and determining their orde r parameter using Eq. ( 3 2), where is the quadrupolar coupling of the deuterium nucleus to the electric field gradient. A static quadrupolar coupling of 167 kHz (94) was assumed in calculating the order parameters. ( 3 2) Comparison of these order parameter profiles to those for the in dividual lipids show that the DPPC order parameters for 4:1 DPPC:POPG MLVs with P:L ratios > 1:200 are higher than would be expected for DPPC alone (132) Within the resolution of the 2 H NMR experiment, no phase separ ation is seen. The profiles also yield a more detailed picture of how SP B 59 80 affects the lipid dynamics with carbon positions 9 16
94 the most affected in the PC lipid acyl chains for both lipid systems. The effects of SP B 59 80 on the PG lipids were simil arly monitored by collecting 2 H NMR spectra for 4:1 DPPC:POPG d31 and 3:1 POPC:POPG d31 MLVs, dePaking the spectra, and generating order parameter profiles. The order parameter profiles for the sn 1 chain on POPG in these lipid mixtures are also given in F igs. 3 9 and 3 10. In these experiments an overall decrease in ordering of the POPG acyl chains is seen on addition of SP B 59 80 Decreases in the order parameters are seen at all the acyl positions in POPG in 3:1 POPC:POPG MLVs and are slightly larger tha n the changes for POPC. The POPG acyl chain order parameters in 4:1 DPPC:POPG MLVs do not appear to be as affected by SP B 59 80 however a comparison of the results further down the acyl chain for POPG d31 when mixed with DPPC vs. POPC again demonstrates t hat the peptide interacts with these lipid systems in a manner dependent on the degree of saturation of the fatty acid chains. Since SP B 59 80 is most likely interacting with both lipid populations in the fluid phase rather than segregating the lipids and interacting with a single phase above changes in POPG d31 and DPPC d62 order parameter profiles on addition of SP B 59 80 to DPPC:POPG MLVs may seem to be contradictory. The observed changes in order parameters can be reco nciled when the data are viewed as the change in order parameter at each position along the acyl chain on addition of peptide relative to the order parameters of the lipids without the peptide. Shown in Fig. 3 11 are the changes in order parameters at part icular methylene positions in the sn 1 palmitoyl acyl chains for the individual lipids in 3:1 POPC:POPG and 4:1 DPPC:POPG MLVs, respectively, on addition of peptide at a P:L molar ratio on the order of 1:100. From these graphs, the behavior of the PG lipid s relative to the PC
95 lipids is clearly offset, but the trends with respect to acyl chain position are similar. The offset of the profiles for the PG lipids is consistent with 31 P NMR findings described above showing that association of SP B 59 80 with the l ipids leads to a change in the orientation PG headgroup and an overall decrease in the 31 P CSA for the POPG lipids. Thus, the offset in the PG profiles in Fig. 3 11 relative to the PC profiles does not reflect significantly less internal order in the PG li pid methylene chains per se, but instead reflects a change in the average orientation of the PG lipid director relative to the membrane normal. The degree of offset for the POPG profiles in the DPPC:POPG and POPC:POPG mixtures correlate with the degree to which the POPG 31 P CSA is affected by addition of peptide. Thus the addition of SP B 59 80 is most likely equally affecting the internal dynamics of the methylene chains in both lipids in a similar manner From their profiles, the meth y l enes in the plateau region are less affected than those further down the acyl chain in the POPC:POPG lipid system. This behavior is similar to changes observed on addition of antimicrobial peptides, which typically cause larger changes in order further down the methylene chai n (127, 141, 142) but is in contrast to the interaction of KL 4 (132) or saposin C (143, 144) with monounsaturated lipids, which have a larger effect on the plateau region. Of particular interest to the roles of SP B 59 80 and KL 4 in lung surfactant formulations are their similar effects on 4:1 DPPC:POPG MLVs. Previously we have observed that KL 4 increases order along the DPPC acyl chains w ith its greatest effect at the center of the bilayers, and this same trend is seen with SP B 59 80 The profile for POPG shows little change in order toward the center of the bilayers and a decrease in the plateau region. However, as discussed above, these differences reflect a change in the average orientation of the POPG headgroup and if this change in
96 orientation is taken into account, the internal order at individual methylene positions in the POPG acyl chain is likely increasing over the entire length o f the acyl chain. Increases in lipid acyl chain ordering have been observed in response to polyelectrolyte binding to lipid headgroups (145, 146) as well as in response to the addition of small molecules, such as cho lesterol (147, 148) or transmembrane peptide helices which partition into the acyl chain region of the lipids (149, 150) In the case of polyelectrolyte binding, larger change s are seen for the plateau region of the lipids reflecting a change in overall orientation of the lipids due to changes in packing of the headgroups. Smaller changes are also seen near the center of the bilayers and are interpreted as resulting from closer packing of the lipids on binding of the electrolyte. For transmembrane helices similar in length to SP B 59 80 increases in order parameters which are similar in magnitude are seen for both the methylene positions in the plateau region and toward the cent er of the bilayers; these trends are observed in both POPC and DPPC bilayers. This is expected since insertion of a transmembrane peptide of this length primarily affects the thickness of the bilayer and would interact with both the lipid acyl chains and h eadgroups across the entire span of the bilayers. In contrast, on addition of cholesterol, larger changes in order parameters are observed toward the center of the bilayer (147) since cholesterol partitions to the bilayer interior and does not strongly interact with the lipid head groups. Interestingly, the effects of SP B 59 80 and KL 4 on order in DPPC:POPG bilayers are most similar to cholesterol, suggesting the bulk of each peptide is partitioning deep within the lipid bilayers. The higher degree of ordering toward the center of the bilayer suggests the surfactant peptides are penetrating deeply into the bi layer and decreasing the mobility of the acyl chains. However, the smaller
97 changes seen in the plateau region (carbons 3 8) suggest the peptides do not adopt a transmembrane orientation. The thermodynamic penalty imposed by placing SP B 59 80 in a transmemb rane orientation would be prohibitive since it would place the hydrophilic amino acids at positions 6, 10, 14, and 16 into the hydrophobic core (151) Thus, the changes in order parameters indicate SP B 59 80 lodges into the hydrophobic region of the bilayer, while maintaining a perpendicular orientat ion to the bilayer normal. This type of interaction would lead to a negative curvature strain within the lipid bilayers, a phenomenon which has been hypothesized as being important to lung surfactant function (152) A recent solution NMR study of SP B 63 78 bound to SDS micelles found the peptide forms an amphi pathic helix at the water interface with the hydrophilic residues solvent accessible (60) consistent with our interpretation. Looking at the same sn 1 C D positions for POPC d31 in 3:1 POPC:POPG MLVs (Figs. 10 and 11), it can be seen that SP B 59 80 decreases order by as much as 20%, with t he largest decrease in ordering seen for position 15. This strongly correlates to SP B 59 80 partitioning at the interface of 3:1 POPC:POPG MLVs and increasing the area per lipid molecule leading to more motion in the acyl chains. A striking decrease in th e plateau region is seen at 2.5 mol% peptide along with the appearance of an isotropic peak in the 2 H spectrum (Fig. 3 8) consistent with the 31 P NMR data suggesting that this concentration of peptide leads to destabilization of the lamellar phase and exch ange of the lipids between a lamellar and nonlamellar phase. From these results we conclude that the penetration and interaction of SP B 59 80 with the lipids in lung surfactant is dependent on the degree of saturation in the lipids. Intriguingly, the pepti de penetrates deeply into mixtures containing a high level of
98 saturated lipids even though the order of the fatty acyl chains in these lipids is higher than those in monounsaturated lipids. When interacting with monounsaturated lipids, SP B 59 80 binds at t he interface and has effects similar to other amphipathic peptides, including antimicrobial peptides, which increase curvature strain in the lipids and can cause micelle formation at higher concentrations. The interaction of SP B 59 80 with these binary lip id systems is on the whole very similar to behavior previously seen for KL 4 suggesting the two peptides have very similar mechanisms of binding and similar effects on the dynamics of the lipids. Discussion The effects of the C terminal region of SP B, spe cifically residues 59 80, on lipid dynamics and acyl chain ordering were explored in two binary lipid systems that differ in their degree of monounsaturation. SP B 59 80 has been used as a template for the design of simple molecules for the treatment of res piratory distress syndrome (RDS), the most notable of these being the 21 amino acid peptide KL 4 which has shown exceptional efficacy in lung surfactant formulations, particularly when compared to more conventional therapies that rely upon exogenous sources of SP B (153) The SP B 59 80 peptide is presumed to be helical when interacting with lipids based on the distrib ution of charged and uncharged residues in its primary amino acid sequence and this is borne out by our CD data. To date, few studies examining the structure of SP B 59 80 and its interactions with lipids have been carried out despite its demonstrated surfa ce activity and role in the rational design of KL 4 Additionally, subsequent studies of KL 4 have suggested its properties might be more similar to SP C, which is a small, highly hydrophobic transmembrane protein in contrast to SP B, which is found at membr ane interfaces. Previously we have examined the structure of KL 4 (154 ) and found it to be
99 consistent with the peptide partitioning in bilayers with the helix axis perpendicular to the membrane normal, but its depth of penetration is dependent on the degree of saturation in the lipids (132) Whether SP B 59 80 binds similarly to lipid interfaces and affects the properties of the lipids in the same manner is the subject of this study. While the sequence of KL 4 is based on SP B 59 80 there are some notable differences in their primary ami no acid sequences. First, KL 4 contains only two types of amino acids, the hydrophobic leucine and cationic lysine; in SP B 59 80 the distribution of hydrophobic and hydrophilic residues is similar, but the hydrophilic amino acids are a mixture of anionic (D 59 and D80), polar (Q68) and cationic (R64, R72, and R76) residues at physiologic pH. The spacing of lysines in KL 4 is based on the five charged amino acids in SP B 59 80 but the preponderance of long, cationic sidechains would allow more favorable electro static interactions with the anionic POPG lipids and phospholipid interfaces in general. If both peptides form amphipathic helices at lipid interfaces, their secondary structure might also differ since the spacing of the polar residues in SP B 59 80 is prim arily every four residues rather than every five residues, as in KL 4 .We have found that KL 4 has the ability to deeply embed in DPPC rich bilayers, which we have attributed to the lysine sidechains being able to snorkel up to the phosphates at the lipid int erface (132) While arginine residues would also allow this behavior, the aspartic acid residue sidechains at the ends of SP B 59 80 would be too short and the electrostatic interactions with the phosphate moieties wo uld be unfavorable. Finally, the occurrence of a proline residue at position 67 in SP B 59 80 could affect the helical nature of the peptide. Nonetheless, both peptides have been demonstrated to be surface active and effective in lowering surface tension at air/water interfaces. Based on the DSC data, SP
100 B 59 80 does not have the same effect on the macroscopic phase properties of DPPC:POPG mixtures as KL 4 DSC experiments performed by Saenz, et al. (137) and our group (132) as well as epifluorescen ce studies (138) indicate that KL 4 mediates phase separation in DPPC:POPG environments. In contrast, DSC indicates SP B 59 80 has little effect on the thermodynamic properties of DPPC:POPG LUVs other than small effects pha se transition. However, the 2 H NMR data show there is a small degree of phase separation in the gel phase on addition of SP B 59 80 In contrast to the calorimetry data, which reports on bulk thermodynamic properties of the lipids, 31 P and 2 H NMR data monit oring lipid dynamics above the phase transition temperature show the two peptides interact similarly with lipid vesicles and have similar effects on lipid dynamics in 4:1 DPPC d62:POPG lipid environments. The 31 P NMR data indicate that neither peptide affe cts the DPPC CSA, while both peptides affect the POPG CSA. While an interaction of SP B 59 80 with the anionic PG headgroups is one plausible explanation of this data, it should be noted that SP B 59 80 does not possess the distinct cationic charge periodici ty found in KL 4 ; the sole use of lysines for charged residues in KL 4 could allow for a more enhanced electrostatic interaction with PG compared to SP B 59 80 This fact is borne out by quantitative comparison of the data for the two peptides. The changes in the POPG CSA, which report on changes in the average orientation of the POPG headgroups, with increasing concentrations of SP B 59 80 are less than half those seen for KL 4 Changes caused by SP B 59 80 can be attributed to the peptide generally affecting li pid lipid interactions rather than a direct electrostatic interaction with the POPG headgroups. However, addition of either peptide leads to increases in the time averaged deuterium order parameters for the DPPC lipids,
101 particularly at carbons 9 15. These increases suggest equally deep penetration of the peptides into the DPPC:POPG bilayers and are similar in scale for both peptides. For both peptides a transmembrane orientation is unlikely as this would place 3 4 polar amino acids into the bilayer interior In this regard, our observations support KL 4 and SP B 59 80 partitioning similarly into DPPC:POPG lipid bilayers. In 3:1 POPC d31:POPG lipids, the changes in the 31 P CSAs for the lamellar vesicles on addition of SP B 59 80 are smaller than those seen in 4: 1 DPPC:POPG samples or on addition of KL 4 to POPC:POPG mixtures. However, at 2.5 mol% SP B 59 80 a striking difference is seen in the 31 P powder patterns suggesting the onset of exchange between the lipid lamellar phase and a more dynamic lipid phase. The 2 H NMR order parameters decrease with increasing peptide concentrations. These increases in fatty acid mobility are similar to the trends found for KL 4 but the changes with SP B 59 80 are considerably greater in magnitude. These findings indicate that the SP B 59 80 remains in the interfacial region of the lipid bilayers at lower concentrations. At 2.5 mol% SP B 59 80 a large coalesced peak in the center of the 2 H spectrum appears indicating the formation of a second, non bilayer lipid phase is occurring. Th e 31 P and 2 H NMR data are consistent with the formation of small vesicles with motions that are fast on the NMR timescale, and may also implicate SP B 59 80 in lipid shuttling, lysis or degradation. Lipid lysis and degradation has been suggested as one of t he many functions of SP B that could be necessary for effective surfactant recycling and remediation (155, 156) An antimicrobial function has also been attributed to SP B, and the lipid dynamics seen at 2.5 mol% SP B 59 80 are similar to behavior seen with higher concentrations of amphipathic antimicrobial peptides such as LL37, pardaxin, magainin, and synthetic derivatives (127,
102 141, 142, 157) This function is of particular in terest as well to the saposin protein family, of which SP B is the most hydrophobic member; this family includes several proteins which bind and traffic lipids for enzymatic modification or degradation (22) However, further studies are needed to differentiate between either a role for SP B 59 80 in lipid degradation or in forming nonlamellar phases. Interestingly, we only saw this effect in 3:1 POPC:POPG, and not in DPPC:POPG at 2.5 mol% SP B 59 80 This suggests the activity of SP B 59 80 is dependent on the saturation level of the acyl chains in a concentration dependent manner and provides a means of discrimination between lipids in the dynamic lung environment. Based on the CD, 31 P NMR and 2 H N MR data, a model of how SP B 59 80 interacts with these lipid systems can be put forth (Fig. 3 12). It should be noted that in this model we assume that SP B 59 80 is helical in a lipid environment, based on our CD data, and we propose that the peptide assum es an orientation in which the helix axis is perpendicular to the membrane normal due to the occurrence of charged residues throughout the primary sequence. Such an assumption is valid given the findings in the literature pertaining to the helical nature o f SP B and in particular the C terminal region (55, 60, 158, 159) However, from examination of the primary amino acid sequence and our 2 H NMR derived order parameters it is conceivable that the type of helical struc ture SP B 59 80 helix and the pitch of the helix may change as the peptide more deeply penetrates into the saturated lipid environments. For KL 4 this type of structural transition has been observed by solid state MAS NMR exp eriments (Mills, F.D. and Long ,J.R., unpublished data). To answer the question of the type of helix SP B 59 80 forms in a lipid environment, similar MAS NMR experiments on 13 C labeled peptide in complex with lipids need to be performed.
103 Since an increase i n DPPC acyl chain ordering is seen on addition of SP B 59 80 to 4:1 DPPC:POPG lipid vesicles, we conclude the peptide penetrates deeply into the hydrophobic region of the lipids and restricts acyl chain motion. In 3:1 POPC:POPG lipid vesicles, the disorderi ng of the acyl chains indicate that the peptide is at the lipid/buffer interface in this lipid system. This is consistent with the CD data which shows solely a helical conformation for SP B 59 80 when it is interacting with 4:1 DPPC:POPG vesicles, suggestin g complete binding of the peptide, and a mixture of helix and random coil when it is interacting with 3:1 POPC:POPG vesicles, suggesting the peptide might partition between the lipid interface and aqueous phases. The results for addition of SP B 59 80 to DP PC:POPG suggest the peptide may deeply penetrate into the bilayer with the charged sidechains KL 4 The snorkeling hypothesis is particularly relevant to KL 4 since all of the charged residues are lysines. However, while SP B 59 80 also has periodically spaced charged residues, the aspartic acid sidechains are not as long and their interaction with the phosphate moieties would not be as favorabl e We are currently pursuing EPR and NMR studies whi ch will allow refinement of our model. This will yield more accurate information in terms of orientation, depth penetration and structure of the peptide at particular amino acids. However, from the current study we can conclude that SP B 59 80 interactions with lung surfactant lipids and subsequent changes in lipid dynamics are dependent on the level of fatty acid saturation. While SP B 59 80 and KL 4 behave similarly in this regard, key differences are also evident. In particular, SP B 59 80 is able to alter b ilayer structure in POPC:POPG vesicles at a concentration as low as 2.5 mol% while KL 4 is able to facilitate phase separation in DPPC:POPG vesicles. The phase
104 separation characteristics of KL 4 may be of key relevance to the peptide's success in the clinic and can most likely be attributed to the use of solely positively charged residues rather than a mixture of cationic and anionic residues. The properties of SP B 59 80 in POPC:POPG vesicles may be due to the fact that polar and charged residues are spaced e very four residues instead of every five residues leading to subtle changes in the structure and interaction of the peptide with the lipid lamellae relative to KL 4 ; alternatively electrostatic interactions may lead to the peptide being even more peripheral ly located in the lipid/water interface in comparison to KL 4 O ur CD spectra of SP B 59 80 in 4:1 DPPC:POPG and 3:1 POPC:POPG lipid environments (Fig. 3 1 ) indicate the peptide is almost entirely heli cal in both lipid preparations, h owever the peptide may n helix with s imilar CD spectra helices (160) 31 P NMR was used to examine lipid headgroup polymorphism and dynamics in 4:1 DPPC:POPG and 3:1 POPC:POPG lipid systems in the presence of SP B 59 80 peptide at different peptide:lipid (P/L) molar r atios (Fig. 3 2, 3 3 ) The difference in PG and PC resonances in 31 P spectra is due to the difference in their preferred headgroup orientations relative to the membrane normal. In both lipid mixtures, with increasing levels of peptide, the PC Chemical Shift Anisotropy (CSA) is invariant, but the PG CSA decreases towards a value characteristic for the neat POPG lipid. Electrostatic interactions of POPC with POPG reorient anionic PG headgroups which leads to an in crease in PG CSA, but addition of positively c harged peptide disrupts this interaction and causes the PG CSAs to move back toward neat POPG values. At the highest P/L ratio the POPC/POPG mixtures transition to a nonlamellar phase.
105 Deuterium NMR spectra of 4:1 DPPC:POPG and 3:1 POPC:POPG lipid samples with varying levels of SP B 59 80 allowed us to infer the depth of peptide penetration into the bilayer by looking at lipid acyl chain ordering and using DePaking of the spectra to assign each C D bond in the acyl chain and determine the time averaged order pa rameters
. Order parameters at particular methylene positions in the palmitoyl acyl chains of PC lipids depend on concentration of SP B 59 80 (Fig. 3 9 3 10 ). Peptide addition to POPC:POPG sample decreases ordering of the deuterated POPC acyl chain s. This suggests the peptide binds to the interface, increasing the lateral spacing between the lipids and allowing more motion in the acyl chains. In contrast, addition of peptide to DPPC:POPG increases the ordering of the deuterated DPPC acyl chains, par ticularly toward the middle of the bilayer. This indicates deep peptide penetration which restricts motion of acyl chains. For the second lipid, POPG, in each lipid mixture the changes in the order parameters are affected by the changes in the average orie n tation of lipid headgroup, but the overall changes in acyl chain order parameters follow those of the PC lipids. Our results indicate SP B 59 80 penetrates deeply into DPPC:POPG bilayers and binds more peripherally to POPC:POPG bilayers (Fig. 3 12 ). Simil ar behavior has been observed for KL 4 a peptide modeled after SPB 59 80 (161) Our laboratory has determined that this is because KL 4 forms unusual helical structures which are lipid depen dent (162) KL 4 changes its helical pitch and penetration as a function of lipid composition. The deep penetration of SP B 59 80 within DPPC:POPG lipid bilayers would require a similar change in helical structure. The structural pl asticity and variable penetration depth of KL 4 and potentially SP B 59 80 can affect the stability and composition of lung surfactant
106 lipid structures and may provide a mechanism for lipid trafficking from lamellar bodies and tubular myelin to the air water interface in a manner that selects for DPPC. Figure 3 1. CD spectra at 45 B' 59 80 at a P:L molar ratio of 1:100 in 4:1 DPPC:POPG ( ) and 3:1 POPC:POPG ( ). The final peptide concentration was 40 Spectra taken with 1 nm step siz e and averaging of 40 50 scans
107 Figure 3 2 Phosphorous NMR spectra of 4:1 DPPC d 62 :POPG MLVs with SP B 59 80 at the indicated P:L molar ratios. (Top) Static NMR spectra and (Bottom) DePaked spectra. Spectra were collected with 200 L of sample at 44 2048 scans.
108 Figure 3 3 Phosphorous NMR spectra of 3:1 POPC d 31 :POPG MLVs with SP B 59 80 at the indicated P:L molar ratios. (Top) Static NMR spectra and (Bottom) DePaked spectra. Spectra were collected with 200 L of sample at 44 2048 scans.
109 Figure 3 4 DSC scans for 4:1 DPPC d 62 :POPG LUVs with SP B 59 80 at the indicated P:L molar ratios Figure 3 5 Deuterium NMR spectra as a function of temperature for (a) 4:1 DPPC d 62 :POPG MLVs, (b) 4:1 DPPC d 62 :POPG MLVs with SP B' 59 80 at a P:L molar ratio of 1:100, (c) 4:1 DPPC:POPG d 31 MLVs, and (d) 4:1 DPPC:POPG d 31 MLVs with SP B' 59 80 at a P:L molar ratio of 1:100. Spectra were collected with 200 L of sample and 2048 scans.
110 Figure 3 6 First moment, M 1 as a function of temperature, for the 2 H NMR spectra plotted in Fig. 3 5 ; data for 4:1 DPPC d 62 :POPG MLVs (open squares), 4:1 DPPC d 62 :POPG MLVs with SP B' 59 80 at a P:L molar ratio of 1:100 (closed squares), 4:1 DPPC:POPG d 31 MLVs (open triangles), and 4:1 DPPC:POPG d 31 MLVs with SP B' 59 80 a t a P:L molar ratio of 1:100 (closed triangles) show the variations in the melting temperatures of the lipids on addition of peptide.
111 Figure 3 7 Deuterium NMR spectra of 4:1 DPPC d 62 :POPG MLVs with SP B 59 80 at the indicated P:L molar ratios. Spectra w ere taken at 44 spectra; (Bottom) dePaked spectra.
112 Figure 3 8 Deuterium NMR spectra of 3:1 POPC d 31 :POPG MLVs with SP B 59 80 at the indicated P:L molar ratios. Spectra were taken at 44 spectra and (Bottom) deP aked spectra.
113 Figure 3 9 Order parameter profiles for the sn 1 chains of DPPC d 62 (top) and POPG d 31 (bottom) in 4:1 DPPC:POPG MLVs at 44 B 59 80 at the indicated P:L molar ratios.
114 Figure 3 10 Order parameter profiles for the sn 1 chains of POPC d 31 (top) and POPG d 31 (bottom) in 3:1 POPC:POPG MLVs at 44 B 59 80 at the indicated P:L molar ratios.
115 Figure 3 11 Changes in d 31 palmitoyl acid chain order parameters in 4:1 DPPC:POPG MLVs (top) and 3:1 POPC:POPG MLVs (bottom) on add ition of SP B' 59 80 at a P:L molar ratio of 1:100.
116 Figure 3 12 Models of SP B 59 80 interactions with the two lipid environments studied assuming a helical peptide conformation. (Left) Based on 2 H NMR data, SP B 59 80 deeply penetrates into 4:1 DPPC:POP G lipid bilayers. (Right) In contrast, in 3:1 POPC:POPG MLVs lipids, a more peripheral interaction of SP B 59 80 with the lipid headgroup region is proposed. A transmembrane orientation of the peptide is unlikely as it would place four polar residues in the hydrophobic interior.
117 CHAPTER 4 SOLUTION NMR STRUCTU RE OF THE C TERMINUS OF SP B IN METHANOL AND IN DPC MICELLES The C terminal fragment of SP B, specifically residues 59 80 (SP B 59 80 ), can efficiently transfer phospholipids mainly DPPC, to air water inter faces to lower interfacial surface tension as it was demonstrated in in vivo studies where surfactant deficient fetal rabbits partially restored lung function after application of synthetic SP B 59 80 (17) As s hown i n the previous chapter SP B 59 80 partition s different ly within lipid bilayers with different levels of acyl chain saturation This differential partitioning within lipid bilayers provides a potential mechanism for transferring of DPPC lipid by SP B to the air water interface However, since direct structural measurements in varying lipid contexts have not been documented the exact mechanism of such lipid transfer is not known. The C terminal peptide is believed to form an amphipathic helix involved in lipid organization Further enhancement of artificial lung surfactant therapies requires knowledge of the high resolution structure of SP B 59 80 in the lipid environment of the lung and an un derstanding of the mechanism of interaction between SP B 59 80 and specific lipids. Since high resolution structural studies in bulky lipid vesicles by solid state NMR are quite expensive, technically challenging and require a lot of time, preliminary measurements were made in variou s smaller lipid bilayer mimetics ( detergent micelles bicelles amphipols and nanodiscs ) using solution NMR (163) A recent solution NMR study of residues 63 78 of the C terminus of SP B (SP B 63 78 ) reconstituted in either SDS micelles or the organic solvent HFIP found the first five residues to be unst ructured and established that the rest of the sequence formed a helix in both SDS micelles and organic solvent. (60) However, measurements of surface tension reduction upon mixing of
118 DPPC/POPG lipids with peptides of different lengths from the C terminus of SP B clearly demonstrate that the longer SP B 59 80 peptide can further reduce surface tension to 6.7 dynes/cm (vs 10.6 dynes/cm for a shorter SP B 64 80 peptide) at a peptide concentration of 1 mol% relative to the lipids. (17, 55) Moreover, micelles of denaturing detergent SDS are not a good mimic for DPPC lipid bilayer. In this chapter solution NMR is used to determine the structure of the C terminus of human SP B, residues 59 80 (SP B 59 80 ), in methanol and in micelles of the zwitterionic mild d etergent dodecylphosphocholine (DPC). Structural studies of SP B 59 80 in the more physiologically relevant DPC micelles, which mimic the major surfactant lipid DPPC, could provide a better understanding of the mechanism by which SP B may specifically trans fer DPPC in tubular myelin bilayers and to the phospholipid monolayer at the air water interface Materials and Methods Synthesis of SP B 59 80 SP B 59 80 (DTLLGRMLPQLVCRLVLRCSMD) was synthesized via solid phase peptide synthesis on a Wang resin (ABI 430, IC BR, UF) cleaved with King's reagent (164) and purified via RP HPLC with a C18 Vydac column and a wa ter/acetonitrile gradient (0.3% TFA). P urity of the product was verified by mass spectrometry (m/z = 2533 ) P eptide was dissolved in methanol to a stock concentration of ~ 1 mM, and the final concentration was determined by amino acid analysis (Molecular Str ucture Facility, UC Davis) This peptide was used for structural measurements in methanol. Heterologous Expression of 15 N labeled SP B 59 80 Double Mutant (SP B 59 80 ) 15 N labeled SP B' 59 80 was expressed using a pET31 construct (EMD Biosciences, Inc., Gibb stown, NJ) incorporating a codon optimized synthetic gene for SP B 59 80
119 (DNA2.0, Menlo Park, CA) in BL21(DE3) pLysS cells (Novagen, EMD Biosciences, Inc., Gibbstown, NJ) in minimal media (165) containing 15 N ammonium chloride as a nitrogen source with subsequent purification and cleavage using established protocols (120) The expressed sequence of SP B' 59 80 was modified to incorporate isoleucines i n lieu of methionines (DTLLGRILPQLVCRLVLRCSID) for compatibility with a cyanogen bromide cleavage reaction. SP B' 59 80 peptide contained an additional homoserine lactone residue at its C terminus, resulting from cyanogen bromide cleavage of the peptide fro m the fusion protein. Following HPLC purification of the final product, the peptide was verified by mass spectrometry (m/z = 2 610) P eptide was dissolved in methanol to a stock concentration of ~ 1mM, and the final concentration was determined by amino acid analysis (Molecular Structure Facility, UC Davis). This peptide was used for structural measurements in DPC micelles, which required 15 N editing to resolve the proton resonances. Circular Dichroism ( CD ) experiments D P C d 38 detergent was dissolved in chlo roform at 50mg/ml. The detergent solution in chloroform was mixed with a methanol solution of SP B 59 80 to obtain a final peptide: detergent molar ratio of 1: 20 0 The sample was dried under nitrogen at 315 323 K in a water bath, suspended in cyclohexane, f lash frozen, and lyophilized overnight to remove residual solvent. Dried sample was solubilized in 10 mM sodium phosphate buffer a t pH 7.4 to achieve a concentration of 40 M SP B 59 80 with 8 mM detergent. A sample of SP B 59 80 in methanol was also prepared a t a concentration of 40 M. SP B 59 80 secondary structure was monitored by far UV CD spectroscopy. CD experiments were performed on an Aviv Model 215 (Lakewood, NJ) at 303 K or 318K over the range of
120 200 260 nm with a 1 nm step size and averaging of 40 50 sc ans. Baseline scans were obtained using the appropriate peptide free controls. NMR samples preparation Methanol stock solutions of 0.7 mM SPB 59 80 and 0.5 mM 15 N labeled SP B' 59 80 were incubated for 1 2 hours with 15 fold molar excess TCEP to fully reduce disulfide bonds in the peptide. To prepare a sample of 1.2 mM SPB 59 80 peptide stock solution was dried under nitrogen gas and redissolved in methanol d 3 (99.8% Sigma Aldrich, Inc., St. Louis, MO) before transfer to a 5mm NMR tube (Norell, Inc., Landis ville, NJ) for NMR measurements. This sample was used to collect 2D TOCSY and NOESY spectra. To prepare a sample of 1.2 mM 15 N labeled SP B' 59 80 in 80 mM DPC d 38 peptide stock solution was mixed with a chloroform solution of 50mg/ml D P C d 38 (98% Cambri dge Isotope Laboratories, Andover, MA) and solvents were evaporated using nitrogen. The resulting film was suspended in cyclohexane, flash frozen and lyophilized overnight to remove residual solvent. The dry sample was solubilized in 110 L of 50 mM aqueou s sodium phosphate buffer (pH 4.5) with 10% D 2 O (99.96% Cambridge Isotope Laboratories, Andover, MA) containing 1 mM 3 (Trimethylsilyl)propionic 2,2,3,3 d4 acid sodium salt (TSP) (98% Sigma Aldrich, Inc., St. Louis, MO) as an internal reference; this s olution was transferred to a 2.5 mm NMR tube (Norell, Inc., Landisville, NJ). The 15 N labeled sample was used to collect 2D 15 N HSQC, TOCSY and NOESY spectra and 3D 15 N edited NOESY HSQC and TOCSY HSQC spectra. NMR experiments All NMR spectra were recorded on a Bruker Avance II 600 MHz spectrometer equipped with a 5mm Z gradient TXI cryo probe and Topspin 2.1 software. In all experiments WATERGATE (166, 167) and States TPPI (168) were used for solvent
121 suppression and quadrature detection in the t 1 dimension, respectively. Sequence ass ignment for SPB 59 80 in methanol d 3 solution was achieved using TOCSY (MLEV 17; (169) 70 ms mixing time) and NOESY (300 ms mixing time) spectra acquired at 300 K. 1 H spectral width in both dimensions was 11.0 ppm with chemi cal shifts referenced to TMS using residual CHD 2 OH signal at 3.31 ppm as a secondary reference. For multidimensional heteronuclear NMR studies of 15 N labeled SP B' 59 80 in DPC micelles, a set of well resolved spectra was recorded at 312 K, including 15 N HS QC 15 N edited NOESY HSQC and TOCSY HSQC spectra. 3D NOESY HSQC spectra were collected with the following parameters for each dimension listed in the order of acquisition: 1 H acquisition 2048 points, 16 scans, 14.0 ppm; 15 N dimension 40 points, 30.0 ppm; 1 H indirect 256 points, 11.7 ppm. Three different mixing times, 75, 150, and 300 ms, were used for the NOESY HSQC and spin diffusion was evaluated via NOE buildup curves for several peaks. As the resu lt of finding balance between maximum number of NOEs important for sequential assignment and structure determination and minimum spin diffusion effects we selected 300ms spectrum for sequential assignment and determination of distance constraints. The 3D T OCSY HSQC was collected using DIPSI 2 (170) with a 60ms mixing time and the following parameters: 1 H acquisition 2048 points, 8 scans, 12.0 ppm; 15 N dimensio n 40 points, 37.0 ppm; 1 H indirect 128 points, 12.0 ppm. 2D TOCSY (DIPSI 2; (170) mixing time 60 ms) and NOESY (75, 150, 200 and 300 ms mixing times) spectra were acquired for this sample under the same conditions to help with sidechain assignments in 3D spectra and to obtain additional distance restraints
122 between side chains. All chemical shifts in the DPC containing sample were referenced using TSP (3 (Trim ethylsilyl)propionic 2,2,3,3 d4 acid sodium salt) as an internal standard. NMR data a nalysis and structure calculation NMR data were processed using NMRPipe (171) and analyzed with SPARKY (172) For the SPB 59 80 structure in methanol, the major restraints were derived from the 300 msec mixing time 2D NOESY spectrum; cross peak intensities w ere converted to medium, and weak NOE cross peaks, respectively. For the DPC containing SP B' 59 80 sample cross peak intensities were categorized as strong, medium, weak and v ery extended covalent structure was used to generate starting coordinates, and an ensemble of 100 minimized structures was calculated using a standard simulated annealing protocol in Xplor NIH version 2.32. (111, 112, 173) The structure was further refined via several rounds of minimization using standard si mulated annealing protocol for refinement in Xplor NIH version 2.32 and minor adjustments to the constraints have been made. Ten structures with minimum total energy were selected out of 3000 structures to meet the following criteria: no NOE derived distan XPLOR (173) and quality of structures was evaluated with PROCHECK V 3.4.4 (114, 174) and MOLPROBITY (175) The structure, 1 H chemical shifts and 3 J coupling constants of SPB 59 80 in methanol were deposited in the Protein Data Bank (PDB entry 2m0h) and
123 BioMagResBank (BMRB entry 18809). The structure, 1 H and 15 N chemical shifts of SP B' 59 80 bound to DPC micelles were deposited in the Protein Data Bank (PDB entry 2m1t) and BioMagResBank (BMRB entry 18879). Results Secondary structure in methanol and in DPC micelles Far UV CD spectra of SP B 59 80 both in methanol and in DPC micel les are characterized by two negative maxima at 208 nm and at 222 nm as is typically observed for helical peptides (Fig. 4 1). helices the negative maxima at 208 and 222 nm typically have equal intensity, (176) b ut for both spectra in Figure 1 the negative maximum at 208 nm is more intense than the one observed at 222 nm. Previous CD studies of SP B 59 80 (68) and KL 4 (132) in lipid b ilayers are similar to the spectra in Figure 4 1. These observations clearly demonstrate that the SP B 59 80 is helical in different environments, but the exact helical structure it adopts is determined by the context of the surrounding environment. NMR Res onance Assignments for SP B 59 80 in methanol The structure of SP B 59 80 in methanol was analyzed by 1 H NMR at 3 00 K In the presence of TCEP, signals were sharper and their number decreased, indicating a loss of oligomer formation. Complete sequence specif ic proton resonance assignments of SP B 59 80 in methanol were achieved using 1 H 1 H TOCSY and 1 H 1 H NOESY spectra. (104) The NH region of the NOESY spectrum with assignments is shown in Figure 4 2 A complete list of assignments is deposited in the BMRB and given in Table 4 1 For SP B 59 80 in methanol, continuous series of H highfield shift s exceeding by 0.2 0. 6 ppm the chemical shifts predicted for random coil sequences were measured for
124 residues 61 64 and 68 78 ; the chemical shift index (CSI) is plotted in Figure 4 3 (Bottom) and is suggestive of helix formation for these regions. (177, 178) T he 3 J HN, coupling constants of 4.8 Hz observed for residues 61 62 and 69 76 are also consistent with an helical structure (179) (Table 4 2). 3 J couplings could not be measured f or residues 63, 64, 66, 68, 71, 72, 77 and 78 due to overlap The observed NN ( i i + 1 ) and N ( i i + 3 ) NOE patterns (Fig. 4 3) also suggest th at residues 61 64 and 68 78 are helical A nearly complete pattern of N(i,i + 3) connectivities was seen for residues 61 64 and 66 79 In a well str uctured helix, the N(i,i + 3) should be much stronger than either the N(i,i + 2) and/or the N(i,i + 4) NOEs (180) and Figure 4 3 s hows that this is clearly the case for the SP B 59 80 in methanol However, the i nte nsities of N(i,i + 3) NOE s are approximately the same either for 3 10 helix ( N(i,i or an helix ( N(i,i In contrast, + 3) NOE s should be prominent for an helix and weak for a 3 10 helix (104) Also N(i,i + 2) NOEs are unique for 3 10 helix and N(i,i + 4) connectivities are a unique feature of an helix (104) Figure 4 3 shows that at the N and C terminal ends both N(i,i + 2) and N(i,i + 4) N OEs are present consequently the helix type of the ends was determined by the relative intensity of the two types of NOEs for each residue. N(i,i + 4) NOEs become quite weak toward the C terminus suggesting the C terminal residues 78 80 form a 3 10 helix due to the existence of two N(i,i + 2) NOEs in this region C omparison of the N( i i + 2) and N( i i + 4) NOEs for N terminal residues 61 64 suggests a mixture of 3 10 and helix, however, this interpretation is somewhat ambiguous since t he N( i i + 2) c onnectivity can also arise from random coil and/or
125 nascent helix conformations (each of which contains isolated turns giving transient i i + 3 hydrogen bonds ) without persistent 3 10 helix (181) Many crystal structures of protein and peptide helices consistently show helix termini are subject to deviations from r helix and C terminal helices often end with a turn of 3 10 helix (182 184) Numerous NMR studies on helical peptides also report NOEs consistent with mixed populations of 3 10 helix (180, 185) Long range inter helical NOEs that position the two helices close in space are consistent with a helical hairpin structure and a few of the inter helical NOEs observed are reported in Figure 4 3. A close inspection of the per residue NOE distribution for L62 R64 and Q68 R76 residues of SP B 59 80 in methanol shows a higher number of NOE contacts than for the residues at the ends of the sequence. Also long range NOEs have been detected for pairs of residues G63 C71 and M6 5 C71, which form hydrophobic contact between the two helices ( Fig. 4 4 A). These observations are consistent with a helical hairpin structure. Three Dimensional Structure of SP B 59 80 in Methanol The structure of SP B 59 80 in methanol was calculated by restrained simulated annealing with subsequent restrained molecular dynamics and energy minimization using XPLOR NIH. (173) 459 Interproton distance restraints were used, 217 of which were intra residual, 96 were sequential, 88 were medium range 4 were long range and 54 were ambigous D ihedral restraints ( 12 ) from 3 J couplings were also used in the calculations. The 10 structures with the lowest total energies and no NOE restraint violation s greater than 0. 29 (Table 4 3 ) Figure 4 5 A shows the superposition of the ten structur es with rmsd of 0.68 for all residues and 0.28 for the more structured residues 61 78
126 (Table 4 3 ) A minimized average structure from the ensemble with well formed secondary structure elements (Fig. 4 5 B) can be cha racterized as a helical L shaped hairpin motif, where two helices are separated by a turn, which is induced by a trans conformation of P67 (Figure 4 5 C side view obtained by 90 rotation of the structure in 4 4B). The N terminal helix (residues 61 65) is a right handed 3 10 helix containing the achiral G63 residue in the middle, the C terminal helix is a well defined right handed helix extending from L69 to C 7 7 ending with a mix of random coil, 3 10 and character for the last and the most disordered turn of the helix (residues 78 to 80). In the ensemble (Fig. 4 5 A) the ends of the peptide are more mobile and slightly less defin ed than the rest of the peptide. This may be a result of not including hydrogen bonds constraints in the calculation. Residues L66 and P67 form a flexible loop between the two helices. Conformational averaging is more likely for NMR structures of small pep tides, especially in the end regions and regions containing aromatics, resulting in a few violations of nonbonded contacts and phi, psi torsion angles. As can be seen in Figure 4 5 (B and C), SP B 59 80 is amphipathic in methanol. The hydrophobic amino acid side chains shown in green reside along the bottom side of the structure thus providing an extended, uninterrupted hydrophobic surface. Positively charged residues (blue), negatively charged residues (red) and a polar Q68 residue (grey in the middle of p eptide) are all located on the other side of the structure creating a hydrophilic surface. Moreover, th e structure shown in Figure 4 5 suggests that hydrophobic interaction between N and C terminal helices G63 C71 plays an essential role in stabilizing th e bent helical hairpin structure of SP B 59 80 peptide in methanol This distribution of lipophilic / hydrophilic residues may also give rise to the oligomerization
127 tendency of SP B 59 80 peptide in methanol since the lipophilic residues would be buried in th e oligomeric form but hydrophilic residues remain exposed to the solvent, stabilizing the oligomeric structures in methanol leading to formation of intermolecular disulfide bonds in the absence of a reducing agent f or 1mM methanolic solutions of SP B 59 80 NMR Resonance Assignment for SP B 59 80 in DPC Micelles Preliminary 2D 1 H 1 H NOESY and TOCSY spectra (not shown) of unlabeled SP B 59 80 peptide in DPC micelles had cross peak line widths that were significantly broader than in the methanol sample. This pr evented unambiguous assignment of a greater portion of NOE cross peaks. However, a sample of universally 15 N labeled SP B' 59 80 peptide in DPC micelles showed significant dispersion of signals in the 15 N dimension of an 1 H 15 N HSQC spectrum (Fig. 4 6 ) and adequate resolution of almost all 1 H, 15 N cross peaks. The good resolution in the 15 N dimension allowed recording of 15 N edited NOESY HSQC and TOCSY HSQC spectra with 1 H 1 H NOESY and TOCSY traces for each residue resolved in separate 15 N planes. Almost comp lete backbone assignment of SP B' 59 80 in DPC micelles was achieved from combined analyses of 1 H 1 H NOESY and TOCSY planes of the corresponding three dimensional 15 N edited NOESY HSQC and TOCSY HSQC spectra and identif ication of the sequential cross signal s in a 2D 1 H 1 H NOESY connecting H and HN of residue i with HN of residue i +1 (Fig. 4 7 ). Sequential NOEs in the 1 H, 15 N NOESY HSQC spectra involving side chain atoms were used to confirm the backbone assignments. Due to the helical nature of the pe ptide, strong sequential NH NH cross peaks were observed in the 1 H, 15 N NOESY HSQC spectra, facilitating assignment by extended amid amide walks. Assignment of aliphatic side chain protons
128 for short side chain residues was achieved using only the 1 H, 15 N TOC SY HSQC spectrum. However, assignment of aliphatic side chain protons for longer side chain residues (L, I, V, R) required comparison of 1 H, 15 N TOCSY HSQC and 2D 1 H 1 H TOCSY spectra since the self relaxation of the nuclear spins during the mixing (spin loc k) period of the TOCSY pulse sequence did not allow efficient magnetization transfer from the HN to the terminal side chain protons of these residues. Additional distance restraints between side chains were obtained from the 2D NOESY spectra with 200 and 3 00 ms mixing times. The chemical shifts of 15 N labeled SP B' 59 80 peptide in DPC micelles at 315 K, referenced to TSP, are listed in Table 4 4 As expected, the analyses of NOE patterns from the 1 H, 15 N NOESY HSQC spectrum and the sidechain region of the 2D 1 H 1 H NOESY spectrum of SP B' 59 80 peptide in DPC micelles revealed the presence of diagnostic medium range NOEs connecting the H of residue i with HN protons of i+2, i+3, i+4 residues and i+3 H residue (Fig. 4 8 ). These NOE patterns are similar to NOE patterns obtained for the methanol sample. The L73 and L7 5 resonances were not resolved (Fig. 4 6 ) since these residues are located in the center of 7 residue palindromic CRLVLRC sequence and have very similar 15 N and NH chemical shifts. Spectra of a short er analog, SP B 63 78 taken by Booth and coworkers (60) also show that 15 N and NH chemical shifts for L73 and L75 are quite similar to each other in SDS micelles. H H (i, i + 3 ) connectivities were observed in residues L61 I65 and L66 R76 on a 2D 1 H 1 H NOESY spectrum. This pattern of NOEs suggests that in DPC micelles SP B' 59 80 acquires a helical conformation similar to the SP B 59 80 conformation in methanol, whe re P67 induces a kink separating two helices. This is in agreement with the CD data (Fig. 4 1). It is
129 noteworthy that several long range NOE interactions between the two helices, primarily involving side chains of charged residues D59, R64, Q68, R72, R76, D80 were observed defining the helical hairpin structure. SP B' 59 80 in DPC micelles additionally had H NH (i, i+3) and H NH (i, i+4) NOEs from T60 to D80. It is noteworthy that H NH (i, i+2) NOEs become stronger for the N and C terminal residues of the sequence relative to the methanol sample. No 3 J proton proton coupling constants were used in the analysis of SP B' 59 80 in DPC because their observation was not possible due to relatively broad line widths in the 1 H NMR spectra. Three Dimensional Structure of SP B 59 80 in DPC Micelle An ensemble of conformations for SP B' 59 80 in complex with D PC micelles was determined based on 433 distance constraints including 108 intra residual, 88 sequential, 142 medium range, 15 long range and 80 ambigous NOEs (Table 4 3 ). The DPC bound structure of SP B' 59 80 peptide is well defined. Figure 4 9 A shows the superposition of all backbone atoms (N, C for the 6 lowest energy structures of SP B' 59 80 in DPC micelles with rmsd of 0.67 for all residues and 0.48 for the more structured residues 61 78 (Table 4 3 ) Res idues L61 S78 of SP B' 59 80 demonstrate the lowest rmsd values for the backbone and side chain atoms. In contrast, residues D59 T60 and I79 D80 have relatively higher rmsd both for the backbone and side chain atoms. Fig. 4 9 B shows a selected representativ e structure of SP B' 59 80 bound to DPC. This structure has a short right handed helix at the N terminus (residues D59 R64), showing partial helical character common for nascent helices, and a well formed right handed helix at the C terminus (residues Q68 D80); the two helices are connected by a short loop consisting of residues I65, L66, and P67.
130 The N and C terminal helices of SP B' 59 80 assume a defined orientation relative to each other giving rise to a helical L shaped hairpin structure (Fig. 4 9 B,C ). The L shaped fold of SP B' 59 80 in DPC micelles is predominantly amphipathic and possesses a continuous hydrophobic face maintained by residues L61, L62, G63, I65, L66, P67, L69, V70, C71, L73, V74, L75, C77 and I79; residues D59, T60, R64, Q68, R72, R76 and D80 are positioned on the hydrophilic face of the peptide (Fig. 4 9 B,C ). The SP B' 59 80 structure in DPC micelles has increased long range side chain/side chain packing interactions between the charged sidechains of the N and C termini helices relative to SP B 59 80 in methanol (Fig. 4 4 B) likely due to the presence of the amphipathic micelles (Figure 4 9 and 4 10 ). These side chain interactions between D59, T60, R64, Q68, R72, R76 and D80 rotate the two helices so that arginine side chains face each other in DPC bound structure (Fig. 4 9 C) rather than tilt outward from each other like in methanol (Fig. 4 5 C). P67 forms the turn of the loop. In other words, side chains of charged residues in SP B' 59 80 appear to be involved in a strong interaction that is locking the two helices into a compact L shape conformation in the DPC micelle environment (Fig. 4 10 ). The views of the DPC bound structure of SP B' 59 80 (Fig. 4 9 B and C) and structure of SP B 59 80 in methano l (Fig. 4 5 B and C) have been selected to emphasize their amphipathic nature and overall similarity of structures. However, there are several striking differences between the two structures that result from difference in polarity of polar methanol and non polar hydrophobic core of DPC micelle. In both structures the C terminal helix (residues 68 80) is a right handed amphipathic helix, whereas the short N terminal helix (residues 61 64) forms a nascent right handed 3 10 helix in
131 methanol and adopts a nasc ent right helix in DPC micelles. Of particular interest is P67 positioned in the flexible hinge, which break s the helical conformation and allows rotation of N and C terminal helices with respect to each other. Due to flexibility of its proline induced loop the two helices are either positioned with their hydrophilic sides outwards facing the polar methanol solvent as in methanol structure (Fig. 4 5 C) or inwards hiding from the non polar environment of the DPC micelle interior (Fig. 4 9 C). It is important to point out that in refinement calculations the structure in methanol was obtained if its NMR constraints were applied to a starting structure based on the DPC data and vice versa. This indicates that each set of constraints is unique, complete and sufficient to drive folding into the specific conformer. Similar changes in orientation of amphipathic helices with environment have been shown for phospholamban, (186) turn helix conformation in organic solvents (187, 188) and in DPC micell es. (186) The two charged arginine residues (R9 and R13) as well as polar phosphorylation sites S16 and T17 are oriented toward the inside facing domain II in the structures obtained in organic solvents thus hiding from the hydrophobic environment, whereas in DPC micelles they are oriented outward facing the bulk solve nt placing the charged side chains toward the bulk solvent and the hydrophobic residues toward the micelle interior (186) The major difference between phospholamban and SP B 59 80 is the balance between hydrophobic and hydrophilic amino acids in their structures with SP B 59 80 peptide having larger hydrophobic face an d smaller hydrophilic face than phospholamban. Consequently, due to this major shift in hydrophobicity SP B 59 80 peptide partitions deeper into DPC micelle hydrophobic core and places charged amino acid side chains closer to each other to
132 shield them from hydrophobic environment of the micellar core and still allow for their efficient interaction with the DPC headgroups. SP B 59 80 also partitions deeply into 4:1 DPPC:POPG lipid bilayer. (68, 69) Discussion A previous solution NMR study of SP B 63 78 reconstituted in either SDS micelles or the organic solvent HFIP found that SP B 63 78 adopts the same structure in both environments with the first five residues being unstructured and the rest of the sequence forming an amp helix. (60) Our structur al studies of SP B 59 80 with a longer and functionally more relevant sequence have revealed essential structural differences between SP B 63 78 and SP B 59 80 that may explain why SP B 59 80 reduces surface tension in lipid monolayers more efficiently than pe ptides with the shorter sequence at comparable concentrations. (17, 55) Using methanol instead of HFIP as well as the milder DPC detergent instead of the denaturant SDS could also account for some of the observed str uctural differences between our work and previously published data for SP B 63 78 DPC micelles most closely mimic the major surfactant lipid DPPC and thus likely provide a better understanding of the mechanism of interaction between SP B and DPPC in pulmon ary surfactant The four extra residues at the N terminus and two extra residues at the C terminus of SP B 59 80 compared to SP B 63 78 include two negatively charged residues, D59 and D80, which provide not only a counter balance for the three positive argi nine residues but also stabilize the N terminal helical structure which allowed us to observe the interhelical NOEs between the N and C terminal helices and to obtain a helical L shaped hairpin structure. The helical hairpin structure also highlights the role of P67 residue in the loop between the two helices as a mobile hinge. This loop allows the peptide to change its topology with environment.
133 The h elical hairpin is a known structure for many amphipathic peptides. For example, this fold is adopted by P a4 pardaxin (189) and MSI 594 (190) in LPS micelles and by Pa4 pardaxin in DPC micelles. (142) In these helical hairpin structures the t wo helices interact through hydrophobic contacts. For example, in the structure of Pa4 in LPS micelles, F15 found in the proline containing loop between the two helices plays an important role in bringing the nonpolar faces of the two helices together. (189) In the structure of MSI 594 in LPS micelles, F5 interacts with other hydrophobic sidechains (190) and its mutation to alanine causes loss of the strong interaction between t he two helices; this opens up the helical hairpin structure of MSI 594 and decreases its antibacterial properties dramatically. (191) It is interesting that Pa4 pardaxin also changes structure with environment: the D PC bound structure of Pa4 (142) lacks the interhelical packing interactions observed in LPS micelles and thus adopts a loose helical hairpin with a sharp kink rather than a tight turn at the proline containing loop s eparating the two helices. The existence of the P67 based hinge between the N and C terminal helices in both the SP B 59 80 and the SP B' 59 80 which can bind specifically to zwitterionic detergent or phospho lipids (Fig. 4 1 1 4 1 2 ), suggesting an interesting mechanism for how it might traffic DPPC and lower surface SP B 59 80 contains the positively charged arginine side chains which electrostat ically attract the negatively negatively charged aspartates, which may be important for initial binding to the positively charged choline headgroup of DPPC. The DPPC molecul e could then remain
134 bound to the peptide while it changes conformation leading to moveme nt of the DPPC molecule as well. Both structure ensembles of SP B 59 80 in methanol and of SP B' 59 80 in DPC have been extensively refined with XPLOR NIH refine_nordc pr otocol while being continuously evaluated using good quality structure criteria via PROCHECK. The results of final evaluation shown in Figs. 4 13 to 4 20 demonstrate quite good quality of the final refined structures. While the structure of SP B has not be en solved, several more soluble members of the saposin and saposin like family of proteins (SAPLIPs) have been characterized. The saposin fold also assembles amphipathic helical hairpin structures which are important to lipid bilayer perturbation and lipid trafficking. Physiological functions of saposins and SAPLIPs can be classified by the level of lipid packing perturbation as (i) mere membrane binding accompanied by only local disordering of the lipid structures for membrane targeting; (ii) membrane pert urbation, but without permeabilization, e.g. lipid extraction and presentation for enzymatic activity; and (iii) permeabilization as a killing principle the essential activity of defense proteins. (192) Only two amino acid residues with functional group s are strictly conserved between the saposin isoforms: a tyrosine residue and proline residue, both located in the loops connecting two amphipathic helices into a helical hairpin structure inside the saposin fold. Since not only saposins but also many SAPL IPs possess a proline residue at the similar position, it might be of structural benefit for lipid perturbation and extraction. (192) The examples of SAPLIPs with antimicrobial, cytolytic activity that are able to penetrate membranes are NK lysin (193) and its human homolog gra nulysin. (194) These proteins have several positively charged amino acids, especially arginine residues, on their surfaces. Their
135 positive charges presumably form the initial contact with the membrane followed by rot generating hydrophobic patches on the protein surface and enabling interaction with the hydrophobic core of the bilayer. (194) In contra st, the saposins, which extract and present lipids from various membranes for enzymatic processing or antigen presentation, are all primarily negatively charged proteins without pronounced electropositive patches on their surface. (192) For example, sapos in C contains three clusters of negatively charged glutamate residues implicated in membrane binding. However, the negatively charged electrostatic surface of saposin C needs to be partially neutralized to trigger membrane binding. (195) Mature SP B has been shown to destabilize phospholipid membranes directly, causing eventual fusion between liposome membranes and rapid leakage of their content. (196 198) Destabilization of lipid bilayer packing could also be part of the mechanism by which SP B promotes formation of a surface associated multilayer reservoir of surfactant to stabilize the interfacial monolayer when it is laterally compressed du ring exhalation. It has been shown that upon compression of monolayer films containing SP B, buckled multilayer patches form and remain associated with the monolayer. Upon expansion of the film at a lower surface pressure, the buckled structures are quickl y reinserted into the in terfacial film to re form a flat monolayer (199 206) In contrast, compres sion of pure lipid films usually ends with an irreversible collapse. (207, 20 8) These observations suggest that SP B is a critical component for lipid trafficking from the aqueous subphase to the lipid monolayer. (209, 210)
136 Based on our two observed structures for SPB 59 80 we propose possibl e mechanism for interaction with DPPC and its trafficking between lipid lamellae by the C terminus of SP B (Fig. 4 1 2 ): the N and C terminal helices with the distinct charged and non polar faces are connected by a flexible loop, which allows the two helic around their axes to change orientation of their charged faces relative to each. If a DPPC molecule is bound to SP B 59 80 it could be SP B 59 80 helices resembling t he process of squeezing water out of clothing by two rollers in an antique laundry mangle machine. T he net result would be destabilization of local bilayer structure or even translocation of the DPPC lipid. In summary, we have determined the three dimensi onal structure of SP B 59 80 in DPC detergent micelles by three dimensional heteronuclear NMR spectroscopy and the three dimensional structure of SP B 59 80 in methanol using two dimensional NMR techniques. The overall similarity of these structures and key differences in tertiary contacts provide novel insight into the potential mechanism of DPPC trafficking by the C terminus of SP B. Our structures and the proposed mechanism for DPPC trafficking may facilitate the engineering of novel peptide analogs of SP B that will be more efficient in surface tension reduction in lungs.
137 Figure 4 1. CD spectra of SP B 59 80 in methanol at 303K (red) and bound to DPC micelle s in 10 mM sodium phosphate, pH 7.4, at 318K (green). Peptide concent ration in C containing sample the peptide :detergent molar ratio is 1:200
138 A Figure 4 2. 300ms NOESY spectr um for SP B 59 80 in methanol at 300 K A: NH H region T he continuous line represents the assignme nt pathway with sequential assignments in black and other assignments in blue. By default the assignments have the form: #residue HA #residue HN (if present, other protons are indicated by letter and number). B: sidechain region. NOEs of the proline system relevant to sequential assignment are shown.
139 B Figure 4 2. Continued.
140 Figure 4 3 Summary of the NOE interactions and CSI measurements for SP B 59 80 in methanol A : Sequential i, i+3; H NH ; H signals and i, i+4 NOE interactions for residues 61 64 and 67 77 are characteristic of an helix Black bars indicate unambiguous NOEs and gray bars indicate ambiguous NOEs. Strong, medium and weak classifications are indicated by the height of the b ar for the sequential NOEs. Some NOEs are not marked because they could not be clearly identified due to resonance overlap. B: chemical shift index (CSI) plotted as the difference between random coil and observed H chemical shifts (211) Significantly negative CSI values are indicative of a helical conformation. The number of distance constraints per residue is presented in Fig. 4 12 A. B A
141 A B Figure 4 4 Summary of the NOE interactions for SP B 59 80 in methanol (A) and for SP B' 59 80 in DPC micelles (B) The number of distance constraints per residue is presented, with white bars representing intra residue constraints, pale gray bars representing sequential constraints, dark gray bars representing medium range (i j < 5) constraints, and black bars representing long range (i An increased numbe r of long range interactions is observed for residues essential to tertiary structure formation: h ydrophobic interaction between G63 and V70 in methanol (A) and interaction between charged side chains of D59, T60, R64, Q68, R72, R76 in DPC micelles (B) In teractions of pairs of charged residues T60 R64, R64 Q68, Q68 R72, R76 D80, which are not long range but are essential for tertiary structure, contribute to an increased number of medium range interactions observed for these residues.
142 Figure 4 5 A: Superpositions of the final set of 10 refined SP B 59 80 structures in methanol with respect to the backbone atoms (N, C B and C: Representative views of SP B 59 80 in methanol illustrating the amphipathic nature of SP B 59 80 Green: hydrophobic residues L, M, C V P ; Red : negatively charged aspartate residues ; Blue: positively charged arginine residues; Grey: polar Q68 residue in the middle of peptide, hydroxyl containing residues T60, S78 at the ends and nonpolar G63 adjacent to blue R64. C A B
143 Figure 4 6 1 H 15 N HSQC spectrum and assignments for 15 N labeled SP B' 59 80 in DPC micelles at 315K. Residues L73 and L75, located in the center of the 7 residue palindromic sequence CRLVLRC are too over lapped to be assigned as separate peaks. Figure 4 7. Alternating 1 H 1 H TOCSY (red) and 1 H 1 H NOESY (blue) from 15 N edited 3D TOCSY HSQC and NOESY HSQC spectra of 15 N labeled SP B 59 80 peptide in DPC micelles. Assignments are indicated by the residue on top of each strip. Sequential NOEs of the backbone walk connecting protons of residue i with HN of residue i+1 are shown by red arrows and other NOEs by black arrows of decreasing boldness indicating increasing distance. A: Assignments for residues at the ends T60 R64 and C77 D80. B: Assignments for residues in the middle I65 R76.
144 Figure 4 7 C ontinued. A
145 Figure 4 7 C ontinued. B
146 Figure 4 8 Summary of the NOE resonances for SP B' 59 80 in DPC micelles A: T he ser ies of i, i+3 NH and signals and i, i+4 NH from residues 60 64 and 67 77 is characteristic of an helix formation for these residues Black bars indicate strong unambiguous NOEs and dotted bars indicate weak NOEs. Strong, medium and weak classifications are indicated by the height of the bar for the sequential NOEs. Some NOEs are not marked because they could not be clearly identified due to resonance overlap. B: CSI plotted as the difference between random coil and observed H chemical shifts (212) The number of distance constraints per residue is presented in Supplementary information Fig. S1, B. A B
147 Figure 4 9 Three dimensional structure of SP B' 59 80 peptide in DPC micelles. A: Sup of the 6 lowest energy structures of SP B' 59 80 in DPC micelles B and C: A representative DPC bound conformation of SP B' 59 80 showing the distribution of charged and nonpolar residues characteristic of an amphip athic peptide from the top view ( B ) and a side view (C) including all the side chains. Extensive side chain to side chain interactions involving D59, R64, Q68, R72, R76 and D80 are present in the NMR data. Green: hydrophobic residues L, I, V C, P ; Red : ne gatively charged D residues ; Blue: positively charged R residues; Grey: polar Q68 residue in the center of peptide, hydroxyl containing residues T60, S78 at the ends and non polar G63 adjacent to blue R64. C B A C
148 Figure 4 10 SP B' 5 9 80 in DPC micelles with residues colored according to their charge to show pairs of electrostatic interaction of positively charged R residues (blue) with negatively charged phosphate moiety (red) and of negatively charged D residues (red) with positivel y charged choline moiety (blue) of DPC detergent. Green: hydrophobic residues L, I, V C, P ; Grey: polar Q68 residue in the center of peptide and hydroxyl containing residues T60, S78 at the ends and non polar G63 adjacent to blue R64. DPC micelle coordina tes in PDB format were obtained from MD simulations of 65 DPC molecules with 6305 water molecules for 1100 ps (213) Coordinates obtained with permission from : Tieleman, D. P., van der Spoel, D., and Berendsen, H. J. C. (2000) Molecular dynamics simulations of dodec ylphosphocholine micelles at three different aggregate sizes: Micellar structure and chain relaxation, Journal of Physical Chemistry B 104 6380 6388
149 Figure 4 1 1 Three dimensional surface models of SP B 59 80 structure in methanol interacting with a DP C molecule (Left) and in DPC micelle (Right). Residues are colored according to their charge to show electrostatic interaction between positively charged (blue) arginine or choline moiety of DPC detergent and negatively charged (red) aspartate or phosphate moiety of DPC detergent. Green: hydrophobic residues L, I, V C, P ; Grey: polar Q68 residue in the middle of peptide and hydroxyl containing residues T60, S78 at the ends and non polar G63 adjacent to blue R64. Figure 4 1 2 P ossible mechanism of DPPC trafficking by SP B 59 80 Structures are colored according to charge of residues: positively charged arginines (blue) and negatively charged aspartates (red) and hydrophobic residues (g reen ).
150 F igure 4 13. The Ramachandran plo t of ensemble of 10 SP B 59 80 structures in methanol generated by PROCHECK (114) shows the phi psi torsion angles for all residues in the ensemble of 10 structure s (except those at the chain termini). Glycine residues are separately identified by triangles as these are not restricted to the regions of the plot appropriate to the other sidechain types. The colouring/shading represents the different regions described in (174) : A dditional Ramachandran plots were generated for each of the 20 different amino acid types (see next page ). The darkest areas (red or green) correspond to the "core" regions representing the most favorable combinations of phi psi backbone torsion angles.
151 Figure 4 14.The Chi1 Chi2 plots show the chi1 chi2 sidech ain torsion angle combinations for 10 structure ensemble of SP B 59 80 in methanol for all residue types whose sidechains are long enough to have both these angles. The darkest areas (green) correspond to the "core" regions representing the most favorable c ombinations of chi1 chi2 sidechain torsion angles.
152 Figure 4 14. Continued
153 Figure 4 15. The Ramachandran plot of ensemble of 10 SP B 59 80 structures in methanol generated by PROCHECK (114) with data points colored according to G factor with yellow as favorable and red as unfavorable.
154 Figure 4 16. Plots of main chain parameters for 10 SP B 59 80 structures in methanol show how the structures (represented by the solid square) compare with well refined structures at a similar resolution. The dark band in each graph represents the results from the well refined structures; the central line is a least squares fit to the mean trend as a function of resolution, while the width of the band on either side of it corresponds to a variation of one standard deviation about the mean. In some cases, the trend is dependent on the resolution, and in other cases it is not
155 F igure 4 17. The Ramachandran plot of ensemble of 6 SP B' 59 80 structures in DPC micelles generated by PROCHECK (114) shows the phi psi torsion angles for all resi dues in the ensemble of 6 structures (except those at the chain termini). Glycine residues are separately identified by triangles as these are not restricted to the regions of the plot appropriate to the other sidechain types. The colouring/shading represe nts the different regions described in (174) : Additional Ramachandran p lots were generated for each of the 20 different amino acid types (see next page). The darkest areas ( red or green ) correspond to the "core" regions representing the most favorable combinations of phi psi backbone torsion angles.
156 Figure 4 17 Continued
157 Figure 4 18.The Chi1 Chi2 plots show the chi1 chi2 sidechain torsion angle combinations for ensemble of 6 structures of SP B' 59 80 in DPC for all residue types whose sidechains are long enough to have both these angles. The dar kest areas (green) correspond to the "core" regions representing the most favorable combinations of chi1 chi2 sidechain torsion angles
158 Figure 4 19. The Ramachandran plot of ensemble of 6 SP B' 59 80 structures in DPC generated by PROCHECK (114) with data points colored according to G factor with yellow as favorable and red as unfavorable
159 Figure 4 20 Plots of main chain parameters for 6 SP B' 59 80 structures in DPC show how the structure s (represented by the solid square) compare with well refined structures at a similar resolution The dark band in each graph represents the results from the well refined structures; the central line is a least squares fit to t he mean trend as a function of resolution, while the width of the band on either side of it corresponds to a variation of one standard deviation about the mean. In some cases, the trend is dependent on the resolution, and in other cases it is not.
160 Table 4 1 1 H chemical shifts for SP B 59 80 in methanol measured at 300 K. Residue HN HA HB HG HD HE D59 4.435 3.066 2.962 T60 8.552 4.521 1.322 L61 8.498 4.178 1.736 1.708 1.708 0.991 0.943 L62 8.050 4.135 1.668 1.668 1.642 0.930 0.980 G63 8.166 3.765 3.971 R64 7.874 4.270 1.967 1.967 1.824 1.738 3.212 7.392 M65 7.997 4.521 2.169 2.169 2.586 2.706 L66 8.110 4.380 1.914 1.769 1.796 0.995 0.960 P67 4.269 2.355 1.897 1.996 2.200 3.810 3.764 Q68 7.869 4 .023 2.288 2.151 2.480 2.446 6.964 7.590 L69 8.025 4.074 1.896 1.794 1.640 0.924 0.951 V70 8.246 3.551 2.221 0.964 1.060 C71 8.148 3.998 3.120 2.856 2.502 R72 8.078 3.960 2.042 1.990 1.684 1.684 3.200 7.575 L73 8.174 4.090 1.935 1.750 1.821 0.890 0.913 V74 8.403 3.590 2.248 0.950 1.093 L75 8.273 4.083 1.910 1.634 1.869 0.907 0.883 R76 8.310 4.106 2.080 2.080 1.856 1.773 3.292 3.204 7.500 C77 8.138 4.313 3.080 2.932 2.670 S78 7.898 4.370 4.005 3.920 M79 7.665 4.517 2.184 2.070 2.705 2.564 D80 7.944 4.763 2.936 2.850 Table 4 2. 3 J couplings for SP B 59 80 in methanol measured at 300 K. 3 J couplings could not be measured f or residues 63, 64, 66, 68, 71, 72, 77 and 78 due to overlap 3 J HN coupling constant Mean value Min value Max value Range T60HN HA 7.4 6.4 8.4 1 L61HN HA 4.8 3.8 5.8 1 L62HN HA 4.8 3.8 5.8 1 M65HN HA 7.35 6.35 8.35 1 L69HN HA 5.0 4.0 6.0 1 V70HN HA 4.8 3.8 5.8 1 L73HN HA 4.88 3.88 5.88 1 V74HN HA 4.8 3.8 5.8 1 L75HN HA 4.45 3.45 5.45 1 R76HN HA 4.45 3.45 5.45 1 M79HN HA 7.73 6.73 8.73 1 D 80HN HA 8.1 7.1 9.1 1
161 Table 4 3 A summary of structural statistics for the 10 final structures of SP B 59 80 in methanol and SP B' 59 80 bound to DPC micelles restraints for structure calculation SP B 59 80 in methanol SP B' 59 80 bound to DPC micelles NOE based distance restraints (total) Intra residue (|i j| = 0) Sequential (|i j| = 1) Medium range (2 |i j| 4) Long range (|i j| 5) Ambigous 3 J HN, H coupling based backbone angle restraints (total) 459 217 96 88 4 54 12 433 108 88 142 15 80 a,b rmsd of j coupling violation (Hz) 0 0 0.050 0.001 0.286 0.250 0.025 3 0 0.060 0.012 0.5 rmsd from covale nt geometry b angles (deg) impropers (deg) 0.003 0.000 0.482 0.006 0.332 0.008 0.005 0.000 0.673 0.016 0.412 0.031 rmsd from the mean structure a backbone atoms of residues (61 total rms d of residues (61 0.68 1.32 0.28 1.15 0.67 1.44 0.48 1.36 Ramachandran plot c residues in the most favored region residues in the additional allowed region residues in the generously allowed region outliers 130 (72.2%) 40 (22.2%) 10 (5.6%) 0 (0%) 59 (54.6%) 43 (39.8%) 6 (5.6%) 0 (0%) a Calculated with VMD (173) when ensemble of structures are superimposed. b Calculated by XPLOR NIH. (111, 112, 173) c Calculated with Procheck V3.4.4 (114, 1 74) for the accepted ensemble of structures.
16 2 Table 4 4 1 H and 15 N chemical shifts for 15 N labeled SP B 59 80 in DPC micelles measured at 315 K. Residue N HN HA HB HG HD Ne HE D59 4.650 3.045 3.045 T60 115.1 8.767 4.510 4.480 1.275 L61 124.1 8.908 4.156 1.800 1.875 1.666 0.918 0.965 L62 117.9 8.505 4.020 1.740 1.770 1.527 0.865 0.943 G63 104.7 8.060 3.775 3.935 R64 117.2 7.670 4.400 1.760 1.952 1.650 3.204 84.300 7.363 I65 116.5 8.005 4.092 1.940 1 .255, 1.300 0.910 0.840 L66 119.7 8.367 4.245 1.800 1.910 1.615 0.873 0.920 P67 1.930 1.930 3.600 3.655 Q68 114.4 7.245 4.055 2.130 2.220 2.445 2.477 110.900 6.840 7.460 L69 119.1 8.060 4.070 1.770 1.860 1.650 0.830 0.870 V70 117.2 8.328 3.535 2.140 1.020 0.930 C71 116.8 8.058 4.660 3.465 3.140 R72 117.2 7.771 4.370 1.940 1.760 3.193 84.300 7.582 L73 115.3 7.853 4.445 1.870 1.760 1.566 0.877 V74 112.2 7.706 4.350 2.335 1.055 0.936 L75 115. 3 7.853 4.445 1.870 1.760 1.566 0.877 R76 118.2 7.913 4.475 1.850 1.710 1.630 1.560 3.178 85.000 7.313 C77 120.6 8.440 4.560 3.415 3.170 S78 116.3 8.460 4.455 3.840 I79 121.0 8.056 4.170 1.884 1.435, 1.165 0.893 0.845 D 80 121.0 8.353 4.720 2.930 2.780
163 CHAPTER 5 CONCLUSIONS AND FUTURE DIRECTION S Many lung surfactant deficiency and dysfunction pathologies could potentially be treated by therapies utilizing combinations of synthetic LS protein mimics and lipids a s LS replacement s This could ameliorate or prevent an immune response thus expanding the range of treatable respiratory distress syndromes. Presented here were insights into the molecular level behavior of LS constituents and an attempt to elucidate a po tential mechanism of interaction between LS protein mimics and LS lipids This work is a continuation of research aimed at understanding LS on a molecular level to guide development of its synthetic replacements In Chapter s 3 and 4 we showed that SP B 59 80 demonstrates helical structure in methanol and in DPC micelles solutions as well as when associated with DPPC/POPG and POPC/POPG lipid systems as determined by CD measurements. H owever, we found some minor differences due to the environment, which coul d be interesting with respect to determining peptide function and required elucidation of the high resolution structures of the peptide in these environments The lipid systems retain a lamellar organization, even with 3 mol% addition of SP B 59 80 This is in contrast to CD data for the N terminal SP B peptide, SP B 1 25 studied by my colleagues which retain s a constant secondary structure when associated with DPPC/POPG and POPC/POPG lipid systems and caus es the formation of fluid isotropic lipid phases (69) In Chapter 3 we showed that SP B 59 80 penetrates deeply into DPPC:POPG bilayers and binds more peripherally to POPC:POPG bilayers allowing it to distinguish between fully saturated and monounsaturated lipids (161) Similar behavior has been observed for KL 4 a peptide mimic of SPB 59 80 which forms unusual lipid dependent
164 helical structures. (162) KL 4 changes its helic al pitch and penetration as a function of lipid composition. The deep penetration of SP B 59 80 within DPPC:POPG lipid bilayers c ould require a similar change in helical structure. However, as described in Chapter 4, SP B 59 80 alters its structure by rotati ng two shorter helices around a proline hinge rather than having a single longer helix with variable pitch. The structural plasticity and variable penetration depth of KL 4 and SP B 59 80 can potentially affect the stability and composition of lung surfactan t lipid structures and may provide a mechanism for lipid trafficking from lamellar bodies and tubular myelin to the air water interface in a manner that selects for DPPC. With the findings presented in Chapter 4 a more thorough molecular model is establis hed that provides insights into how the C terminal end of SP B may interact with DPPC lipid and selectively transfer it to the alveolar air water interface. The proposed mechanism can drive the development of future SP B mimetics. Taken together, these st udies highlight that development of successful synthetic LS replacement formulation s probabl y requires one to specifically target peptidomimetics that affect DPPC dynamics The structural study also highlighted the unique role of the proline in the middle of SP B 59 80 sequence, which acts as a hinge allowing two helices to roll around their axes and adapt to the surrounding environment This is in contrast to KL 4 a functional mimic of SP B 59 80 which changes helical pitch to adapt to different lipid envir onments through exchange between i,i+4 H bonds and i,i+5 H bonds The rotation about the proline hinge in SP B 59 80 requires change in a smaller number of bonds compared to helical pitch changes of KL 4 and thus is probably more energetically favorable. Co nsequently, introduction of the proline residue in the middle
165 of KL 4 peptide could make its structural adaptation to the environment much more energetically favorable, which could in turn make it more efficient in RDS treatment. While we established the ef fect of SP B 59 80 on lipid dynamics is modulated by the degree of lipid saturation in LS mimics, much work is left if we want to have a better understanding of SP B 59 80 interaction with lipids particularly study of the high resolution structures and dyna mics of SP B 59 80 peptide in the DPPC:POPG and POPC:POPG lipid bilayers The solution NMR structures of SP B 59 80 in methanol and DPC micelles discussed in this dissertation provide an important model for how the peptide might structura l l y adapt in a lipid environment, which prompted us to create a model of its interaction with DPPC lipid (Fig. 5 1) However, o ther structural and dynamics measurements of the peptide in DPPC:POPG and POPC:POPG lipid bilayers are needed to establish the true mechanism for the peptide interaction with and remodeling of lipid assemblies D irect measurements of t he actual depth of SP B 59 90 penetration into lipid bilayer s via power saturation electron paramagnetic resonance (EPR) measurements could confirm our findings at amino acid resolution. Experiments of this type would complement our existing deuterium data, confirming conclusions made from our determination of relative order parameters in the acyl chains of the individual lipid species. Backbone measurements of SP B 59 90 i n the DPPC:POPG and POPC:POPG lipid bilayer s via solid state NMR will provide high resolution structures of the peptide in these bilayers. This would expand our understanding of its mechanism of interaction with lipids which would ultimately allow underst anding of how lipid trafficking happens in lung surfactant.
166 A model of how SP B 59 80 could be involved in direct ing DPPC adsorption to the air water interface and facilitating tight packing interactions between the bilayers of tubular myelin has been prop osed in Fig. 5 1 based on the structures of SP B 59 80 in methanol and DPC micelles obtained in this dissertation. Figure 5 1 Molecular model s of SP B 59 80 directing specific DPPC adsorption to the air water interface and faci litating tight packing interactions between the bilayers of tubular myelin P hospholipids (mainly DPPC) are transferred f rom tubular myelin bilayers to the monolayer at the air water interface to efficiently cover and stabilize this interface for adequate respiration. The t wo structures of SP B 59 80 in DPC micelle are positioned at the fusion neck to demonstrate how they can stabilize the neck and participate in DPPC lipid trafficking. The s tructure of SP B 59 80 in methanol is placed at the bilayer interfac e to show how it can be involved into interaction between the two bilayers to facilitate the ir tight packing in tubular myelin.
167 APPENDIX SCRIPTS FOR STRUCTUR E CALCULATIONS AND R EFINEMENT S cript modified from anneal_nordc.py example script in XPLOR NIH eginput directory, (173) which was used for initial structure calculation : xplor.requireVer sion("2.24") # # slow cooling protocol in torsion angle space for protein G. Uses # NOE, RDC, J coupling restraints. # # this script performs annealing from an extended structure. # It is faster than the original anneal.py # # CDS 2009/07/24 # # this c hecks for typos on the command line. User customized arguments can # also be specified. # xplor.parseArguments() # filename for output structures. This string must contain the STRUCTURE # literal so that each calculated structure has a unique name. The S CRIPT # literal is replaced by this filename (or stdin if redirected using <), # but it is optional. # outFilename = "SPB5980_SCRIPT_STRUCTURE.sa" numberOfStructures=100 #usually you want to create at least 20 # protocol module has many high level help er functions. # import protocol protocol.initRandomSeed() #set random seed by time command = xplor.command # generate PSF data from sequence and initialize the correct parameters. # from psfGen import seqToPSF seqToPSF('spbc_sequence_xplor.seq', sta rtResid=59) # generate random extended initial structure with correct covalent geometry # protocol.genExtendedStructure() xplor.command("write struct output=spbc.psf end") protocol.writePDB("spbc_extended.pdb")
168 # # a PotList contains a list of potential terms. This is used to specify which # terms are active during refinement. # from potList import PotList potList = PotList() from xplorPot import XplorPot # parameters to ramp up during the simulated annealing protocol # from simulationTools import MultRa mp, StaticRamp, InitialParams rampedParams= highTempParams= # compare atomic Cartesian rmsd with a reference structure # backbone and heavy atom RMSDs will be printed in the output # structure files # #from posDiffPotTools import create_PosDiffPot #refRMSD = create_PosDiffPot("refRMSD","name CA or name C or name N", # pdbFile='sa_5.pdb', # cmpSel="not name H*") # set up NOE potential noe=PotList('noe') potList.append(noe) from noePotTools import create_NOEPot for (name,scale,file) in [('all',1,"spbc_TCEP_hocd3_HeiVolNote_heightsorted_duplicates_commented_cl assified_noindex.tbl"), #add entries for additional tables ]: pot = create_NOEPot(name ,file) # pot.setPotType("soft") # if you think there may be bad NOEs pot.setScale(scale) noe.append(pot) rampedParams.append( MultRamp(2,30, "noe.setScale( VALUE )") ) # set up J coupling with Karplus coefficients #from jCoupPotTools import create_JCoupPot #jCoup = create_JCoupPot("jcoup","jna_coup.tbl", # A=6.98,B= 1.38,C=1.72,phase= 60.0) #potList.append(jCoup) # Set up dihedral angles #from xplorPot import XplorPot #dihedralRestraintFilename="ubiquitin_1g6j_noA_dihe .tbl" #protocol.initDihedrals(dihedralRestraintFilename, #useDefaults=False # by default, symmetric sidechain # restraints are included
169 # ) #potList.append( XplorPot(' CDIH') ) #highTempParams.append( StaticRamp("potList['CDIH'].setScale(10)") ) #rampedParams.append( StaticRamp("potList['CDIH'].setScale(200)") ) # set custom values of threshold values for violation calculation # #potList['CDIH'].setThreshold( 5 ) # gy ration volume term # #from gyrPotTools import create_GyrPot #gyr = create_GyrPot("Vgyr", # "resid 1:56") # selection should exclude disordered tails #potList.append(gyr) #rampedParams.append( MultRamp(.002,1,"gyr.setScale(VALUE)") ) # hbda distance/angle bb hbond term # #protocol.initHBDA('hbda.tbl') #potList.append( XplorPot('HBDA') ) #Rama torsion angle database # protocol.initRamaDatabase() potList.append( XplorPot('RAMA') ) rampedParams.append( MultRamp(.002,1,"potList['RAMA'].se tScale(VALUE)") ) # # setup parameters for atom atom repulsive term. (van der Waals like term) # potList.append( XplorPot('VDW') ) rampedParams.append( StaticRamp("protocol.initNBond()") ) rampedParams.append( MultRamp(0.9,0.8, "command('param nbonds repel VALUE end end')") ) rampedParams.append( MultRamp(.004,4, "command('param nbonds rcon VALUE end end')") ) # nonbonded interaction only between CA atoms highTempParams.append( StaticRamp("""proto col.initNBond(cutnb=100, rcon=0.004, tolerance=45, repel=1.2, onlyCA=1)""") ) potList.append( XplorPot("BOND") ) potList.append( XplorPot("ANGL") ) potList['ANGL'].setThreshold( 5 ) rampedParams.append( MultRamp(0.4,1,"potList['ANGL'].setScale(VALUE)") ) potList.append( XplorPot("IMPR" ) )
170 potList['IMPR'].setThreshold( 5 ) rampedParams.append( MultRamp(0.1,1,"potList['IMPR'].setScale(VALUE)") ) # Give atoms uniform weights, configure bath/molecule friction coeff. # protocol.massSetup() # IVM setup # the IVM is used for perfo rming dynamics and minimization in torsion angle # space, and in Cartesian space. # from ivm import IVM dyn = IVM() # configure ivm topology for torsion angle dynamics # protocol.torsionTopology(dyn) # minc used for final cartesian minimization # minc = IVM() protocol.initMinimize(minc) protocol.cartesianTopology(minc) # object which performs simulated annealing # from simulationTools import AnnealIVM init_t = 3500. # Need high temp and slow annealing to converge cool = AnnealIVM(initTemp =init _t, finalTemp=25, tempStep =12.5, ivm=dyn, rampedParams = rampedParams) def calcOneStructure(loopInfo): """ this function calculates a single structure, performs analysis on the structure, and then writes out a pdb file, with remarks. """ # generate a new structure with randomized torsion angles # from monteCarlo import randomizeTorsions randomizeTorsions(dyn) protocol.fixupCovalentGeom(maxIters=100,useVDW =1)
171 # set torsion angles from restraints # from torsionTools import setTorsionsFromTable # setTorsionsFromTable(dihedralRestraintFilename) protocol.writePDB(loopInfo.filename()+".init") # calc. initial tensor orientation # # initialize parameters for high temp dynamics. InitialParams( rampedParams ) # high temp dynamics setup only need to specify parameters which # differfrom initial values in rampedParams InitialParams( highTempParams ) # high temp dynamics # protocol.initDynamics(dyn, potList=potList, # potential terms to use bathTemp=init_t, initVelocities=1, finalTime=200, # stops at 800ps or 8000 steps numSteps=2000, # whichever comes first printInterval=200) #was 100 dyn.setETolerance( init_t/100 ) #used to det. stepsize. default: t/1000 dyn.run() # initialize para meters for cooling loop InitialParams( rampedParams ) # initialize integrator for simulated annealing # protocol.initDynamics(dyn, potList=potList, numSteps=100, #at each temp: 100 steps or finalTime=.2 # .2ps, whichever is less printInterval=100) # perform simulated annealing # cool.run() # final torsion angle minimization # protocol.initMinimize(dyn, printInterval=50) dyn.run() # final all atomic degrees of freedom minimization #
172 protocol.initMinimize(minc, potList=potList, dEPred=10) minc.run() #do analysis and write structure loopInfo.writeStructure(potList) pass from simulationTools import StructureLoop, FinalParams StructureLoop(numStructures=numberOfStructures, pdbTemplate=outFilename, structLoopAction=calcOneStructure, genViolationStats=1, averageTopFraction=0.1, #report stats on best 50% of structs averageContext=FinalParams(rampedParams), # averageCrossTerms=refRMSD, averageSortPots=[potList['BOND'],potList['ANGL'],potList['IMPR'], noe], averagePotList=potList).run()
173 S cript modified from refine _nordc.py example script in XPLOR NIH eginput directory (173) which was used for refinement of ensemble of SP B 59 80 structures in methanol. xplor.require Version("2.18") # # slow cooling protocol in torsion angle space for protein G. Uses # NOE, J coupling restraints. # # this version refines from a reasonable model structure. # # CDS 2005/05/10 # (opts,args) = xplor.parseArguments(["quick"]) # check fo r command line typos quick=False for opt in opts: if opt=="quick": #specify quick to just test that the script runs quick=True pass pass outFilename = "SPB5980_SCRIPT_STRUCTURE.sa" numberOfStructures=3000 if quick: numb erOfStructures=3 pass # protocol module has many high level helper functions. # import protocol protocol.initRandomSeed(8870649) #explicitly set random seed # # annealing settings # command = xplor.command protocol.initParams("protein") # genera te PSF data from sequence and initialize the correct parameters. # #from psfGen import seqToPSF #seqToPSF('1g6j_sequence.seq') #protocol.initStruct("g_new.psf") # or from file
174 # generate a random extended structure with correct covalent geometry # save s the generated structure in the indicated file for faster startup # next time. # #protocol.genExtendedStructure("gb1_extended_%d.pdb" % # protocol.initialRandomSeed()) # or read an existing model # protocol.loadPDB("SPBC_ref ine_nordc_95.sa") xplor.simulation.deleteAtoms("not known") protocol.fixupCovalentGeom(maxIters=100,useVDW=1) # # a PotList contains a list of potential terms. This is used to specify which # terms are active during refinement. # from potList import PotL ist potList = PotList() # parameters to ramp up during the simulated annealing protocol # from simulationTools import MultRamp, StaticRamp, InitialParams rampedParams= highTempParams= # compare atomic Cartesian rmsd with a reference structure # bac kbone and heavy atom RMSDs will be printed in the output # structure files # #from posDiffPotTools import create_PosDiffPot #refRMSD = create_PosDiffPot("refRMSD","name CA or name C or name N", # pdbFile='g_xray.pdb', # cmpSel="not name H*") # set up NOE potential noe=PotList('noe') potList.append(noe) from noePotTools import create_NOEPot for (name,scale,file) in [('all',1,"spbc_TCEP_hocd3_HeiVolNote_heightsorted_duplicates_commented_cl assified_noind ex11ambi.tbl"), #add entries for additional tables ]: pot = create_NOEPot(name,file) pot.setPotType("soft") # if you think there may be bad NOEs pot.setScale(scale) noe.append(pot) ramped Params.append( MultRamp(2,30, "noe.setScale( VALUE )") )
175 # set up J coupling with Karplus coefficients from jCoupPotTools import create_JCoupPot jCoup = create_JCoupPot("jcoup","jna_coup_spbc_in_methanol9.tbl", A=6.98,B= 1.38,C=1 .72,phase= 60.0) rampedParams.append( MultRamp(1,3, "jCoup.setScale( VALUE )") ) potList.append(jCoup) # Set up dihedral angles from xplorPot import XplorPot #protocol.initDihedrals("ubiquitin_1g6j_noA_dihe.tbl", # #useDefaults=False # by default, symmetric sidechain # # restraints are included # ) #potList.append( XplorPot('CDIH') ) #highTempParams.append( StaticRamp("potList['CDIH'].setScale(10)") ) #rampedParams.appen d( StaticRamp("potList['CDIH'].setScale(200)") ) # set custom values of threshold values for violation calculation # #potList['CDIH'].setThreshold( 5 ) #5 degrees is the default value, though # gyration volume term # # gyration volume term # #from gyr PotTools import create_GyrPot #gyr = create_GyrPot("Vgyr", # "resid 1:56") # selection should exclude disordered tails #potList.append(gyr) #rampedParams.append( MultRamp(.002,1,"gyr.setScale(VALUE)") ) # hbda distance/angle bb hbond term # #protocol.initHBDA('hbda.tbl') #potList.append( XplorPot('HBDA') ) # hbdb distance/angle bb hbond term # #protocol.initHBDB() #potList.append( XplorPot('HBDB') ) #Rama torsion angle database # from torsionDBPotTools import create_TorsionDBPot to rsionDB = create_TorsionDBPot('torsionDB') potList.append( torsionDB ) rampedParams.append( MultRamp(.002,2,"torsionDB.setScale(VALUE)") ) # # setup parameters for atom atom repulsive term. (van der Waals like term) #
176 potList.append( XplorPot('VDW') ) ram pedParams.append( StaticRamp("protocol.initNBond()") ) rampedParams.append( MultRamp(0.9,0.8, "command('param nbonds repel VALUE end end')") ) rampedParams.append( MultRamp(.004,4, "command('param nbonds rcon VALUE end end')") ) # nonbonded interaction only between CA atoms highTempParams.append( StaticRamp("""protocol.initNBond(cutnb=100, rcon=0.004, tolerance=45, repel=1.2, onlyCA=1)""") ) potList.append( XplorPot("BOND") ) potList.append( XplorPot("ANGL") ) potList['ANGL'].s etThreshold( 5 ) rampedParams.append( MultRamp(0.4,1,"potList['ANGL'].setScale(VALUE)") ) potList.append( XplorPot("IMPR") ) potList['IMPR'].setThreshold( 5 ) rampedParams.append( MultRamp(0.1,1,"potList['IMPR'].setScale(VALUE)") ) # Give atoms un iform weights, except for the anisotropy axis # protocol.massSetup() # IVM setup # the IVM is used for performing dynamics and minimization in torsion angle # space, and in Cartesian space. # from ivm import IVM dyn = IVM() # reset ivm topology for torsion angle dynamics # dyn.reset() protocol.torsionTopology(dyn) # minc used for final cartesian minimization # minc = IVM() protocol.initMinimize(minc) protocol.cartesianTopology(minc)
177 # object which performs simulated annealing # from simulation Tools import AnnealIVM init_t = 1500. # Need high temp and slow annealing to converge cool = AnnealIVM(initTemp =init_t, finalTemp=25, tempStep =3.25, ivm=dyn, rampedParams = rampedPa rams) def accept(potList): """ return True if current structure meets acceptance criteria """ if potList['noe'].violations()>0: return False if potList['CDIH'].violations()>0: return False if potList['BOND'].violati ons()>0: return False if potList['ANGL'].violations()>0: return False if potList['IMPR'].violations()>1: return False return True def calcOneStructure(loopInfo): """ this function calculates a single structure, performs analysis on the structure, and then writes out a pdb file, with remarks. """ # initialize parameters for high temp dynamics. InitialParams( rampedParams ) # high temp dynamics setup only need to specify parameters which # differfrom initial values in rampedParams InitialParams( highTempParams ) # high temp dynamics # protocol.initDynamics(dyn, potList=potList, # potential terms to use bathTemp=init_t, initVelocities=1, finalTime=10, # stops at 10ps or 5000 steps numSteps=5000, # whichever comes first printInterval=100) dyn.setETolerance( ini t_t/100 ) #used to det. stepsize. default: t/1000 dyn.run() # initialize parameters for cooling loop
178 InitialParams( rampedParams ) # initialize integrator for simulated annealing # protocol.initDynamics(dyn, potList=potList, numSteps=100, #at each temp: 100 steps or finalTime=.2 # .2ps, whichever is less printInterval=100) # perform simulated annealing # cool.run() # final torsion angle minimization # protocol.initMinimize(dyn, printInterval=50) dyn.run() # final all atom minimization # protocol.initMinimize(minc, potList=potList, dEPred=10) minc.run() #do analysis and write structure loopInfo.writeStructure(potList) pass from simulationTools import StructureLoop, FinalParams StructureLoop(numStruc tures=numberOfStructures, pdbTemplate=outFilename, calcMissingStructs=True, structLoopAction=calcOneStructure, genViolationStats=1, averagePotList=potList, averageSortPots= [potList['BOND'],potList['ANGL'],potList['IMPR'],jCoup, noe], # averageCrossTerms=refRMSD, averageTopFraction=0.005, #report only on best 50% of structs # averageAccept=accept, #only use structures which pass accept() averageContext=FinalParams(rampedParams), averageFilename="SPB5980_SCRIPT_ave.pdb", #generate regularized ave structure averageFitSel="name CA", averageCompSel="n ot resname ANI and not name H*" ).run()
179 S cript modified from refine _nordc.py example script in XPLOR NIH eginput directory (173) which was used for refinement of ensemble of SP B 59 80 structures in DPC micelles. xplor.requireVersion("2.18") # # slow cooling protocol in torsion angle space for protein G. Uses # NOE, J coupling restrain ts. # # this version refines from a reasonable model structure. # # CDS 2005/05/10 # (opts,args) = xplor.parseArguments(["quick"]) # check for command line typos quick=False for opt in opts: if opt=="quick": #specify quick to just test that the script runs quick=True pass pass outFilename = "2ISPB5980_SCRIPT_STRUCTURE.sa" numberOfStructures=600 if quick: numberOfStructures=3 pass # protocol module has many high level helper functions. # import protocol protocol.in itRandomSeed(3421) #explicitly set random seed # # annealing settings # command = xplor.command protocol.initParams("protein") # generate PSF data from sequence and initialize the correct parameters. # #from psfGen import seqToPSF #seqToPSF('1g6j_seq uence.seq') #protocol.initStruct("g_new.psf") # or from file
180 # generate a random extended structure with correct covalent geometry # saves the generated structure in the indicated file for faster startup # next time. # #protocol.genExtendedStructure(" gb1_extended_%d.pdb" % # protocol.initialRandomSeed()) # or read an existing model # protocol.loadPDB("2ISPBC_refine_nordc_108.sa") xplor.simulation.deleteAtoms("not known") protocol.fixupCovalentGeom(maxIters=100,useVDW=1) # # a PotList contains a list of potential terms. This is used to specify which # terms are active during refinement. # from potList import PotList potList = PotList() # parameters to ramp up during the simulated annealing protocol # from simulationTools import MultRamp, StaticRamp, InitialParams rampedParams= highTempParams= # compare atomic Cartesian rmsd with a reference structure # backbone and heavy atom RMSDs will be printed in the output # structure files # #from posDiffPotTools import creat e_PosDiffPot #refRMSD = create_PosDiffPot("refRMSD","name CA or name C or name N", # pdbFile='g_xray.pdb', # cmpSel="not name H*") # set up NOE potential noe=PotList('noe') potList.append(noe) from no ePotTools import create_NOEPot for (name,scale,file) in [('all',1,"15Nspbc_noesy3D_dpc_1Dec11_04042012_noe_sorted2_ambi17.tbl"), #add entries for additional tables ]: pot = create_NOEPot(name,file) pot.setPotType("soft") # if you think there may be bad NOEs pot.setScale(scale) noe.append(pot) rampedParams.append( MultRamp(2,30, "noe.setScale( VALUE )") )
181 # set up J coupling with Karplus coefficients #from jCoupPotTools import create_JCo upPot #jCoup = create_JCoupPot("jcoup","jna_coup_spbc_in_methanol7.tbl", # A=6.98,B= 1.38,C=1.72,phase= 60.0) #potList.append(jCoup) # Set up dihedral angles from xplorPot import XplorPot #protocol.initDihedrals("ubiquitin_1g6j_noA_ dihe.tbl", # #useDefaults=False # by default, symmetric sidechain # # restraints are included # ) #potList.append( XplorPot('CDIH') ) #highTempParams.append( StaticRamp( "potList['CDIH'].setScale(10)") ) #rampedParams.append( StaticRamp("potList['CDIH'].setScale(200)") ) # set custom values of threshold values for violation calculation # #potList['CDIH'].setThreshold( 5 ) #5 degrees is the default value, though # gyrati on volume term # # gyration volume term # #from gyrPotTools import create_GyrPot #gyr = create_GyrPot("Vgyr", # "resid 1:56") # selection should exclude disordered tails #potList.append(gyr) #rampedParams.append( MultRamp(.002,1,"gyr.s etScale(VALUE)") ) # hbda distance/angle bb hbond term # #protocol.initHBDA('hbda.tbl') #potList.append( XplorPot('HBDA') ) # hbdb distance/angle bb hbond term # #protocol.initHBDB() #potList.append( XplorPot('HBDB') ) #Rama torsion angle database # from torsionDBPotTools import create_TorsionDBPot torsionDB = create_TorsionDBPot('torsionDB') potList.append( torsionDB ) rampedParams.append( MultRamp(.002,2,"torsionDB.setScale(VALUE)") ) # # setup parameters for atom atom repulsive term. (van der Waa ls like term) # potList.append( XplorPot('VDW') )
182 rampedParams.append( StaticRamp("protocol.initNBond()") ) rampedParams.append( MultRamp(0.9,0.8, "command('param nbonds repel VALUE end end')") ) rampedParams.append( MultRamp( .004,4, "command('param nbonds rcon VALUE end end')") ) # nonbonded interaction only between CA atoms highTempParams.append( StaticRamp("""protocol.initNBond(cutnb=100, r con=0.004, tolerance=45, repel=1.2, onlyCA=1)""") ) potList.append( XplorPot("BOND") ) potList.append( XplorPot("ANGL") ) potList['ANGL'].setThreshold( 5 ) rampedParams.append( MultRamp(0.4,1,"potList['ANGL'].setScale(VALUE)") ) potList.append( XplorPot("IMPR") ) potList['IMPR'].setThreshold( 5 ) rampedParams.append( MultRamp(0.1,1,"potList[ 'IMPR'].setScale(VALUE)") ) # Give atoms uniform weights, except for the anisotropy axis # protocol.massSetup() # IVM setup # the IVM is used for performing dynamics and minimization in torsion angle # space, and in Cartesian space. # from i vm import IVM dyn = IVM() # reset ivm topology for torsion angle dynamics # dyn.reset() protocol.torsionTopology(dyn) # minc used for final cartesian minimization # minc = IVM() protocol.initMinimize(minc) protocol.cartesianTopology(minc) # object which performs simulated annealing
183 # from simulationTools import AnnealIVM init_t = 2000. # Need high temp and slow annealing to converge cool = AnnealIVM(initTemp =init_t, finalTemp=25, tempStep =3.25, ivm=dyn, rampedParams = rampedParams) def accept(potList): """ return True if current structure meets acceptance criteria """ if potList['noe'].violations()>0: return False if potList['CDIH'].violations()>0: return False if potList['BOND'].violations()>0: return False if potList['ANGL'].violations()>0: return False if potList['IMPR'].violations()>1: return False return True def calcOneStructure(loopInfo): """ this function calculates a single structure, performs analysis on the structure, and then writes out a pdb file, with remarks. """ # initialize parameters for high temp dynamics. InitialParams( rampedParams ) # high temp dynami cs setup only need to specify parameters which # differfrom initial values in rampedParams InitialParams( highTempParams ) # high temp dynamics # protocol.initDynamics(dyn, potList=potList, # potential ter ms to use bathTemp=init_t, initVelocities=1, finalTime=10, # stops at 10ps or 5000 steps numSteps=5000, # whichever comes first printInterval=100) dyn.setETolerance( init_t/100 ) #used to det. stepsize. default: t/1000 dyn.run() # initialize parameters for cooling loop InitialParams( rampedParams )
184 # initialize integrator for simulated annealing # protocol.initDynamics(dyn, potList=potList, numSteps=100, #at each temp: 100 steps or finalTime=.2 # .2ps, whichever is less printI nterval=100) # perform simulated annealing # cool.run() # final torsion angle minimization # protocol.initMinimize(dyn, printInterval=50) dyn.run() # final all atom minimization # protocol.initMinimize(minc, potList=potList, dEPred=10) minc.run() #do analysis and write structure loopInfo.writeStructure(potList) pass from simulationTools impo rt StructureLoop, FinalParams StructureLoop(numStructures=numberOfStructures, pdbTemplate=outFilename, #calcMissingStructs=True, structLoopAction=calcOneStructure, genViolationStats=1, a veragePotList=potList, averageSortPots=[potList['BOND'],potList['ANGL'],potList['IMPR'], noe], # averageCrossTerms=refRMSD, averageTopFraction=0.017, #report only on best 50% of struct s # averageAccept=accept, #only use structures which pass accept() averageContext=FinalParams(rampedParams), averageFilename="2ISPB5980_SCRIPT_ave.pdb", #generate regularized ave structure average FitSel="name CA", averageCompSel="not resname ANI and not name H*" ).run()
185 LIST OF REFERENCES 1. Daniels, C. B., and Orgeig, S. (2003) Pulmonary surfactant: the key to the evolution of air breathing, News in Physiolo gical Sciences 18 151 157. 2. Shields, T. W., LoCicero, J., Ponn, R. B., and Rusch, V. W. (2004) General thoracic surgery: scientific foundations and clinical practice 3. Serrano, A. G., and Perez Gil, J. (2006) Protein lipid interactions and surface act ivity in the pulmonary surfactant system, Chemistry and Physics of Lipids 141 105 118. 4. Perez Gil, J. (2002) Molecular interactions in pulmonary surfactant films, Biology of the Neonate 81 6 15. 5. Goerke, J. (1998) Pulmonary surfactant: functions and molecular composition, Biochimica Et Biophysica Acta Molecular Basis of Disease 1408 79 89. 6. Veldhuizen, R., Nag, K., Orgeig, S., and Possmayer, F. (1998) The role of lipids in pulmonary surfactant, Biochimica Et Biophysica Acta Molecular Basis of Dis ease 1408 90 108. 7. Goerke, J. (1998) Pulmonary surfactant: functions and molecular composition, Biochimica Et Biophysica Acta Molecular Basis of Disease 1408 79 89. 8. Pison, U. (2005) Interfacial properties of pulmonary surfactant layers, Adv. Colloid Interface Sci. 117 33 58. 9. De la Serna, J. B., Perez Gil, J., Simonsen, A. C., and Bagatolli, L. A. (2004) Cholesterol rules: direct observation of the coexistence of two flui d phases in native pulmonary surfactant membranes at physiological temperatures, J. Biol. Chem. 279 40715 40722. 10. Notter, R. H. (2000) Lung Surfactants: Basic science and clinical applications Vol. 149, CRC Press. 11. Munford, R. S., Sheppard, P., and O'Harat, P. J. (1995) Saposin like proteins (SAPLIP) carry out diverse functions on a common backbone structure, J. Lipid Res. 36 1653 1663. 12. Kingma, P. S., and Whitsett, J. A. (2006) In defense of the lung: surfactant protein A and surfactant protein D, Curr. Opin. Pharmacol. 6 277 283. 13. Hawgood, S., and Poulain, F. R. (2001) The pulmonary collectins and surfactant metabolism, Annu. Rev. Physiol. 63 495 519.
186 14. Sarin, V. K., Gupta, S., Leung, T. K., Taylor, V. E., Ohning, B. L., Whitsett, J. A., and Fox, J. L. (1990) Biophysical and biological activity of a synthetic 8.7 kDa hydrophobic pulmonary surfactant protein SP B, Proc. Natl. Acad. Sci. U. S. A. 87 2633 2637. 15. Johansson, J., and Curstedt, T. (1997) Molecular structures and interactions of pulmonary surfactant components, Eur. J. Biochem. 244 675 693. 16. Sarin, V. K., Gupta, S., Leung, T. K., Taylor, V. E., Ohning, B. L., Whitsett, J. A., and Fox, J. L. (1990) Biophysical and biological activity of a synthetic 8.7 Kda hydrophobic pulmo nary surfactant protein SP B, Proceedings of the National Academy of Sciences of the United States of America 87 2633 2637. 17. Revak, S. D., Merritt, T. A., Hallman, M., Heldt, G., Lapolla, R. J., Hoey, K., Houghten, R. A., and Cochrane, C. G. (1991) The use of synthetic peptides in the formation of biophysically and biologically active pulmonary surfactants, Pediatric Research 29 460 465. 18. Johansson, J., and Curstedt, T. (1997) Molecular structures and interactions of pulmonary surfactant components, European Journal of Biochemistry 244 675 693 19. Power, J. H. T., Doyle, I. R., Davidson, K., and Nicholas, T. E. (1999) Ultrastructural and protein analysis of surfactant in the Australian lungfish Neoceratodus forsteri: Evidence for conservation of co mposition for 300 million years, Journal of Experimental Biology 202 2543 2550. 20. Beck, D. C., Na, C. L., Whitsett, J. A., and Weaver, T. E. (2000) Ablation of a critical surfactant protein B intramolecular disulfide bond in transgenic mice, Journal of Biological Chemistry 275 3371 3376. 21. Zaltash, S., Palmblad, M., Curstedt, T., Johansson, J., and Persson, B. (2000) Pulmonary surfactant protein B: a structural model and a functional analogue, Biochimica Et Biophysica Acta Biomembranes 1466 179 186. 22. Munford, R. S., Sheppard, P. O., and Ohara, P. J. (1995) Saposin like proteins (SAPLIP) carry out diverse functions on a common backbone structure, Journal of Lipid Research 36 1653 1663. 23. Vandenbussche, G., Clercx, A., Clercx, M., Curstedt, T., Jo hansson, J., Jornvall, H., and Ruysschaert, J. M. (1992) Secondary structure and orientation of the surfactant protein SP B in a lipid environment a fourier transform infrared spectroscopy study, Biochemistry 31 9169 9176.
187 24. Cruz, A., Casals, C., and Perezgil, J. (1995) Conformational flexibility of pulmonary surfactant proteins SP B and SP C, studied in aqueous organic solvents, Biochimica Et Biophysica Acta Lipids and Lipid Metabolism 1255 68 76. 25. Wuestneck, N., Wuestneck, R., Perez Gil, J., and Pison, U. (2003) Effects of oligomerization and secondary structure on the surface behavior of pulmonary surfactant proteins SP B and SP C, Biophysical Journal 84 1940 1949. 26. Perez Gil, J., Casals, C., and Marsh, D. (1995) Interactions of hydrophobic l ung surfactant proteins SP B and SP C with dipalmitoylphosphatidylcholine and dipalmitoylphosphatidylglycerol bilayers studied by electron spin resonance spectroscopy, Biochemistry 34 3964 3971. 27. Longo, M. L., Bisagno, A. M., Zasadzinski, J. A. N., Bru ni, R., and Waring, A. J. (1993) A function of lung surfactant protein SP B, Science 261 453 456. 28. Perez Gil, J., Casals, C., and Marsh, D. (1995) Interactions of hydrophobic lung surfactant proteins SP B and SP C with dipalmitoylphosphatidylcholine an d dipalmitoylphosphatidylglycerol bilayers studied by electron spin resonance spectroscopy, Biochemistry 34 3964 3971. 29. Cruz, A., Worthman, L. A., Serrano, A. G., Casals, C., Keough, K. M. W., Gil, J. (2000) Microstructure and dynamic surface properties of surfactant protein SP B/dipalmitoylphosphatidylcholine interfacial films spread from lipid protein bilayers, Eur. Biophys. J. 29 204 213. 30. Ross, M., Krol, S., Janshoff, A., and Galla, H. J. (2002) Kinetics of phospholipid insertion into monolayers containing the lung surfactant proteins SP B or SP C, Eur. Biophys. J. 31 52 61. 31. Oosterlaken Dijksterhuis, M. A., Haagsman, H. P., Van Golde, L. M. G., and Demel, R. A. (1991) Characterization of lipid insertion into monomolecular layers me diated by lung surfactant proteins SP B and SP C, Biochemistry 30 10965 10971. 32. Vandenbussche, G., Clercx, A., Clercx, M., Curstedt, T., Johansson, J., Jornvall, H., and Ruysschaer, J. M. (1992) Secondary structure and orientation of the surfactant pro tein SP B in a lipid environment. A Fourier transform infrared spectroscopy study, Biochemistry 31 9169 9176. 33. Vandenbussche, G., Clercx, A., Curstedt, T., Johansson, J., Jornvall, H., and Ruysschaert, J. M. (1992) Structure and orientation of the surf actant associated protein C in a lipid bilayer, Eur. J. Biochem. 203 201 209.
188 34. Nag, K., Taneva, S. G., Perez Gil, J., Cruz, A., and Keough, K. M. (1997) Combinations of fluorescently labeled pulmonary surfactant proteins SP B and SP C in phospholipid films., Biophys. J. 72 2638 2650. 35. Whitsett, J. A., Nogee, L. M., Weaver, T. E., and Horowitz, A. D. (1995) Human surfactant protein B structure, function, regulation, and genetic disease, Physiological Reviews 75 749 757. 36. Nogee, L. M., Garnier G., Dietz, H. C., Singer, L., Murphy, A. M., Demello, D. E., and Colten, H. R. (1994) A mutation in the surfactant protein B gene responsible for fatal neonatal respiratory disease in multiple kindreds, Journal of Clinical Investigation 93 1860 1863. 37 Clark, J. C., Wert, S. E., Bachurski, C. J., Stahlman, M. T., Stripp, B. R., Weaver, T. E., and Whitsett, J. A. (1995) Targeted disruption of the surfactant protein B gene disrupts surfactant homeostasis, causing respiratory failure in newborn mice, Proc eedings of the National Academy of Sciences of the United States of America 92 7794 7798. 38. Robertson, B., Kobayashi, T., Ganzuka, M., Grossmann, G., Li, W. Z., and Suzuki, Y. (1991) Experimental neonatal respiratory failure induced by a monoclonal anti body to the hydrophobic surfactant associated protein SP B, Pediatric Research 30 239 243 39. Foster, C. D., Zhang, P. X., Gonzales, L. W., and Guttentag, S. H. (2003) In vitro surfactant protein B deficiency inhibits lamellar body formation, American Jo urnal of Respiratory Cell and Molecular Biology 29 259 266. 40. Stahlman, M. T., Gray, M. P., Falconieri, M. W., Whitsett, J. A., and Weaver, T. E. (2000) Lamellar body formation in normal and surfactant protein B deficient fetal mice, Laboratory Investig ation 80 395 403. 41. Perez Gil, J. (2008) Structure of pulmonary surfactant membranes and films: The role of proteins and lipid protein interactions, Biochimica Et Biophysica Acta Biomembranes 1778 1676 1695. 42. Schram, V., and Hall, S. B. (2001) Therm odynamic effects of the hydrophobic surfactant proteins on the early adsorption of pulmonary surfactant, Biophys. J. 81 1536 1546. 43. J., and Janshoff, A. (2000) Formation of three dimensional protei n lipid aggregates in monolayer films induced by surfactant protein B, Biophys. J. 79 904 918. 44. Hawco, M. W., Coolbear, K. P., Davis, P. J., and Keough, K. M. W. (1981) Exclusion of fluid lipid during compression of monolayers of mixtures of dipalmitoy lphosphatidylcholine with some other phosphatidylcholines, Biochim. Biophys. Acta Biomembranes 646 185 187.
189 45. Discher, B. M., Maloney, K. M., Schief, W. R., Grainger, D. W., Vogel, V., and Hall, S. B. (1996) Lateral phase separation in interfacial fil ms of pulmonary surfactant, Biophys. J. 71 2583 2590. 46. Gil, J. (2008) Structure of pulmonary surfactant membranes and films: The role of proteins and lipid protein interactions, Biochim. Biophys. Acta Biomembranes 1778 1676 1695. 47. Serrano, Gil, J. (2006) Protein lipid interactions and surface activity in the pulmonary surfactant system, Chemistry and Physics of Lipids 141 105 118. 48. Bishop, M. L., Fody, E. P., and Schoeff, L. E. (2004) Clinical Chemistry 49. Lewis, J. F. and Veldhuizen, R. (2003) The role of exogenous surfactant in the treatment of acute lung injury, Annu. Rev. Physiol. 65 613 642. 50. Zuo, Y. Y., Veldhuizen, R. A. W., Neumann, A. W., Petersen, N. O., and Possmayer, F. (2008) Current perspectives in pul monary surfactant Inhibition, enhancement and evaluation, Biochim. Biophys. Acta Biomembranes 1778 1947 1977. 51. Revak, S. D., Merritt, T. A., Hallman, M., Heldt, G., La Polla, R. J., Hoey, K., Houghten, R. A., and Cochrane, C. G. (1991) The use of s ynthetic peptides in the formation of biophysically and biologically active pulmonary surfactants, Pediatr. Res. 29 460 465. 52. Waring, A., Taeusch, W., Bruni, R., Amirkhanian, J., Fan, B., Stevens, R., and Young, J. (1989) Synthetic amphipathic sequence s of surfactant protein B mimic several physicochemical and in vivo properties of native pulmonary surfactant proteins, Pept. Res. 2 308 313. 53. Mingarro, I., Lukovic, D., Vilar, M., and Perez Gil, J. (2008) Synthetic pulmonary surfactant preparations: N ew developments and future trends, Current Medicinal Chemistry 15 393 403 54. Seurynck, S. L., Patch, J. A., and Barron, A. E. (2005) Simple, helical peptoid analogs of lung surfactant protein B, Chemistry & Biology 12 77 88. 55. Cochrane, C. G., and Re vak, S. D. (1991) Pulmonary surfactant protein B (SP B) structure function relationships, Science 254 566 568. 56. Waring, A., Taeusch, W., Bruni, R., Amirkhanian, J., Fan, B., Stevens, R., and Young, J. (1989) Synthetic amphipathic sequences of surfact ant protein B mimic several physicochemical and in vivo properties of native pulmonary surfactant proteins, Peptide Research 2 308 313.
190 57. Waring, A. J., Walther, F. J., Gordon, L. M., Hernandez Juviel, J. M., Hong, T., Sherman, M. A., Alonso, C., Alig, T., Braun, A., Bacon, D., and Zasadzinski, J. A. (2005) The role of charged amphipathic helices in the structure and function of surfactant protein B, Journal of Peptide Research 66 364 374. 58. Bruni, R., Taeusch, H. W., and Waring, A. J. (1991) Surfacta nt protein B lipid interactions of synthetic peptides representing the amino terminal amphipathic domain, Proceedings of the National Academy of Sciences of the United States of America 88 7451 7455. 59. Kang, J. H., Lee, M. K., Kim, K. L., and Hahm, K. S. (1996) The relationships between biophysical activity and the secondary structure of synthetic peptides from the pulmonary surfactant protein SP B, Biochemistry and Molecular Biology International 40 617 627. 60. Booth, V., Waring, A. J., Walther, F. J., and Keough, K. M. W. (2004) NMR structures of the C terminal segment of surfactant protein B in detergent micelles and hexafluoro 2 propanol, Biochemistry 43 15187 15194. 61. Sarker, M., Waring, A. J., Walther, F. J., Keough, K. M. W., and Booth, V. ( 2007) Structure of mini B, a functional fragment of surfactant protein B, in detergent micelles, Biochemistry 46 11047 11056. 62. Morrow, M. R., Stewart, J., Taneva, S., Dico, A., and Keough, K. M. W. (2004) Perturbation of DPPC bilayers by high concentra tions of pulmonary surfactant protein SP B, European Biophysics Journal with Biophysics Letters 33 285 290 63. Morrow, M. R., Perezgil, J., Simatos, G., Boland, C., Stewart, J., Absolom, D., Sarin, V., and Keough, K. M. W. (1993) Pulmonary surfactant ass ociated protein SP B has little effect on acyl chains in dipalmitoylphosphatidylcholine dispersions, Biochemistry 32 4397 4402. 64. Cruz, A., Casals, C., Plasencia, I., Marsh, D., and Perez Gil, J. (1998) Depth profiles of pulmonary surfactant protein B i n phosphatidylcholine bilayers, studied by fluorescence and electron spin resonance spectroscopy, Biochemistry 37 9488 9496. 65. Walther, F. J., Hernandez Juviel, J. M., Gordon, L. M., Sherman, M. A., and Waring, A. J. (2002) Dimeric surfactant protein B peptide SP B1 25 in neonatal and acute respiratory distress syndrome, Experimental Lung Research 28 623 640.
191 66. Veldhuizen, E. J. A., Waring, A. J., Walther, F. J., Batenburg, J. J., van Golde, L. M. G., and Haagsman, H. P. (2000) Dimeric N terminal segm ent of human surfactant protein B (dSP B1 25) has enhanced surface properties compared to monomeric SP B1 25, Biophysical Journal 79 377 384. 67. Tanaka, Y., Takei, T., Aiba, T., Masuda, K., Kiuchi, A., and Fujiwara, T. (1986) Development of synthetic lun g surfactants, Journal of Lipid Research 27 475 485. 68. Antharam, V. C., Farver, R. S., Kuznetsova, A., Sippel, K. H., Mills, F. D., Elliott, D. W., Sternin, E., and Long, J. R. (2008) Interactions of the C terminus of lung surfactant protein B with lipi d bilayers are modulated by acyl chain saturation, Biochimica Et Biophysica Acta Biomembranes 1778 2544 2554. 69. Farver, R. S., Mills, F. D., Antharam, V. C., Chebukati, J. N., Fanucci, G. E., and Long, J. R. (2010) Lipid polymorphism induced by surfacta nt peptide SP B1 25, Biophysical Journal 99 1773 1782. 70. Waring, A. J., Walther, F. J., Gordon, L. M., Hernandez Juviel, J. M., Hong, T., Sherman, M. A., Alonso, C., Alig, T., Braun, A., D., B., and Zasadzinski, J. A. (2005) The role of charged amphipat hic helices in the structure and function of surfactant protein B, J. Pept. Res. 66 364 374. 71. Cochrane, C. G., Revak, S. D., Merritt, A., Heldt, G. P., Hallman, M., Cunningham, M. D., Easa, D., Pramanik, A., Edwards, D. K., and Alberts, M. S. (1996) Th e efficacy and safety of KL(4) surfactant in infants with respiratory distress syndrome, American Journal of Respiratory and Critical Care Medicine 153 404 410. 72. Cochrane, C. G., Revak, S. D., Merritt, T. A., Schraufstatter, I. U., Hoch, R. C., Henders on, C., Andersson, S., Takamori, H., and Oades, Z. G. (1998) Bronchoalveolar lavage with KL4 Surfactant in models of meconium aspiration syndrome, Pediatric Research 44 705 715. 73. Wiswell, T. E., Smith, R. M., Katz, L. B., Mastroianni, L., Wong, D. Y., Willms, D., Heard, S., Wilson, M., Hite, R. D., Anzueto, A., Revak, S. D., and Cochrane, C. G. (1999) Bronchopulmonary segmental lavage with surfaxin (KL4 surfactant) for acute respiratory distress syndrome, American Journal of Respiratory and Critical Car e Medicine 160 1188 1195. 74. Sinha, S. K., Lacaze Masmonteil, T., Soler, A. V. I., Wiswell, T. E., Gadzinowski, J., Hajdu, J., Bernstein, G., d'Agostino, R., and Dist, S. T. A. R. (2005) A multicenter, randomized, controlled trial of lucinactant versus p oractant alfa among very premature infants at high risk for respiratory distress syndrome, Pediatrics 115 1030 1038.
192 75. Ghodrat, M. (2006) Lung surfactants, American Journal of Health System Pharmacy 63 1504 1521. 76. Cochrane, C., Revak, S., Merritt, T ., Heldt, G., Hallman, M., Cunningham, M., Easa, D., Pramanik, A., Edwards, D., and Alberts, M. (1996) The efficacy and safety of KL4 surfactant in preterm infants with respiratory distress syndrome, Am. J. Respir. Crit. Care Med. 153 404 410. 77. Gustafs son, M., Vandenbussche, G., Curstedt, T., Ruysschaert, J. M., and Johansson, J. (1996) The 21 residue surfactant peptide (LysLeu4)4Lys(KL4) is a t helix with a mixed nonpolar/polar surface, Febs Letters 384 185 188. 78. Bulheller, B. M., Rodger, A., and Hirst, J. D. (2007) Circular and linear dichroism of proteins, Physical Chemistry Chemical Physics 9 2020 2035. 79. Kelly, S. M., Je ss, T. J., and Price, N. C. (2005) How to study proteins by circular dichroism, Biochimica Et Biophysica Acta Proteins and Proteomics 1751 119 139 80. Adler, A. J., Greenfield, N. J., and Fasman, G. D. (1973) Circular dichroism and optical rotatory dispe rsion of proteins and polypeptides, Methods in enzymology 27 675 735 81. Holzwarth, G., Doty, P., and Gratzer, W. B. (1962) Optical activity of polypeptides in far ultraviolet, Journal of the American Chemical Society 84 3194 3196 82. Brown, M. F., Lo pe Piedrafita, L., Martinez, G. V., and Petrache, H. I. (2006) Solid state deuterium NMR spectroscopy of membranes Modern Magnetic Resonance 249 260. 83. Seddon, J. M. (1990) Structure of the inverted hexagonal (HII) phase, and non lamellar phase transit ions of lipids, Biochimica Et Biophysica Acta 1031 1 69. 84. Koynova, R., and Caffrey, M. (2002) An index of lipid phase diagrams, Chemistry and Physics of Lipids 115 107 219 85. Saupe, A. (1968) Recent results in field of liquid crystals, Angewandte Ch emie International Edition 7 97 112 86. Castro, V. (2005) "Seeing" lipid membranes by solidstate NMR, In From Cells to Proteins: Imaging Nature across Dimensions (Evangelista, V., Barsanti, L., Passarelli, V., and Gualtieri, P., Eds.), pp 451 458, Spring er, Po Box 17, 3300 Aa Dordrecht, Netherlands.
193 87. Auger, M. (2000) Biological membrane structure by solid state NMR, Current issues in molecular biology 2 119 124. 88. Phosphorus 31 NMR Spectroscopy Springer Berlin Heidelberg, Berlin, Heidelberg. 89. Seelig, J. (1978) 31P nuclear magnetic resonance and the head group structure of phospholipids in membranes, Biochim Biophys Acta 515 105 140. 90. Smith, I. C. P., and Ekiel, I. H. (1984) Phosphorus 31 NMR: Principles and applications El sevier Inc. 91. Dubinnyi, M. A., Lesovoy, D. M., Dubovskii, P. V., Chupin, V. V., and Arseniev, A. S. (2006) Modeling of P 31 NMR spectra of magnetically oriented phospholipid liposomes: A new analytical solution, Solid State Nuclear Magnetic Resonance 29 305 311. 92. Levitt, M. H. (2008) Spin dynamics: Basics of nuclear magnetic resonance 2 ed., Wiley. 93. Thayer, A. M., and Kohler, S. J. (1981) P 31 nuclear magnetic resonance spectra characteristic of hexagonal and isotropic phospholipid phases generate d from phosphatidylethanolamine in the bilayer phase, Biochemistry 20 6831 6834. 94. Burnett, L. J., and Muller, B. H. (1971) Deuteron quadrupole coupling constants in 3 solid deuterated paraffin hydrocarbons C2D6, C4D10, C6D14, Journal of Chemical Physic s 55 5829 5831 95. Seelig, J. (1977) Deuterium magnetic resonance theory and application to lipid membranes, Quarterly Reviews of Biophysics 10 353 418. 96. Dufourc, E. J., Smith, I. C. P., and Jarrell, H. C. (1983) A H 2 NMR analysis of dihydrostercu loyl containing lipids in model membranes structural effects of a cyclopropane ring, Chemistry and Physics of Lipids 33 153 177. 97. Vaz, N. A. P., and Doane, J. W. (1980) NMR measurements of chain ordering in some liquid crystalline lamellar phases, Ph ysical Review A 22 2238 2244. 98. Brown, M. F., Lope Piedrafita, L., Martinez, G. V., and Petrache, H. I. (2006) Solid state deuterium NMR spectroscopy of membranes, Modern Magnetic Resonance 249 260.
194 99. Hemminga, M. A., and Poile, T. W. (1989) Biologic al applications of solid state NMR., In In: Dynamic properties of biomolecular assemblies, S.E. Harding & A.J. Rowe (eds.). Royal Soc. Chemistry, Cambridge, UK (1989) 74 89. 100. Schafer, H., Madler, B., and Sternin, E. (1998) Determination of orientationa l order parameters from H 2 NMR spectra of magnetically partially oriented lipid bilayers, Biophysical Journal 74 1007 1014 101. Sternin, E., Schafer, H., Polozov, I. V., and Gawrisch, K. (2001) Simultaneous determination of orientational and order param eter distributions from NMR spectra of partially oriented model membranes, Journal of Magnetic Resonance 149 110 113. 102. Sternin, E., Bloom, M., and Mackay, A. L. (1983) De Pake Ing of NMR spectra, Journal of Magnetic Resonance 55 274 282. 103. Petrach e, H. I., Dodd, S. W., and Brown, M. F. (2000) Area per lipid and acyl length distributions in fluid phosphatidylcholines determined by H 2 NMR spectroscopy, Biophysical Journal 79 3172 3192. 104. Wuthrich, K. (1986) NMR of proteins and nucleic acids Joh n Wiley & Sons, Inc., USA. 105. Wuthrich, K. (1990) Protein structure determination in solution by NMR spectroscopy, Journal of Biological Chemistry 265 22059 22062. 106. Higman, V. A. (2012) Protein NMR (a practical guide). Double resonance backbone assi gnment. 107. Girvin, M. E. (2012) Introduction to 3D triple resonance experiments (lecture notes). 108. Karplus, M. (1959) Contact electron spin coupling of nuclear magnetic moments, Journal of Chemical Physics 30 11 15. 109. Karplus, M. (1963) Vicinal pr oton coupling in nuclear magnetic resonance, Journal of the American Chemical Society 85 2870 2871 110. Shen, Y., Delaglio, F., Cornilescu, G., and Bax, A. (2009) TALOS plus : a hybrid method for predicting protein backbone torsion angles from NMR chemic al shifts, Journal of Biomolecular N MR 44 213 223. 111. Schwieters, C. D., Kuszewski, J. J., Tjandra, N., and Clore, G. M. (2003) The Xplor NIH NMR molecular structure determination package, Journal of Magnetic Resonance 160 65 73.
195 112. Schwieters, C. D. Kuszewski, J. J., and Clore, G. M. (2006) Using Xplor NIH for NMR molecular structure determination, Progress in Nuclear Magnetic Resonance Spectroscopy 48 47 62. 113. Nilges, M., Gronenborn, A. M., Brunger, A. T., and Clore, G. M. (1988) Determination of 3 dimensional structures of proteins by simulated annealing with interproton distance restraints application to crambin, potato carboxypeptidase inhibitor and barley serine proteinase inhibitor 2, Protein Engineering 2 27 38. 114. Laskowski, R. A., M acarthur, M. W., Moss, D. S., and Thornton, J. M. (1993) PROCHECK a program to check the stereochemical quality of protein structures, Journal of Applied Crystallography 26 283 291. 115. Merrifield, R. B. (1963) Solid phase peptide synthesis. I. The syn thesis of a tetrapeptide, J. Am. Chem. Soc. 85 2149 2154. 116. Chan, W. C., and White, P. D. (2000) Fmoc Solid Phase Peptide Synthesis: A Practical Approach 1st ed., Oxford University Press, USA. 117. Grant, G. A. (2002) Synthetic peptides: a user's guid e Oxford University Press US A 118. Albericio, F. (2000) Solid phase synthesis: a practical guide 1st ed., CRC Press, Boca Raton. 119. Li, Y. (2009) Carrier proteins for fusion expression of antimicrobial peptides in Escherichia coli, Biotechnology and Applied Biochemistry 54 1 9. 120. Kuliopulos, A., and Walsh, C. T. (1994) Production, purification, and cleavage of tandem repeats of recombinant peptides, Journal of the American Chemical Society 116 4599 4607. 121. Rodriguez, J. C., Wong, L., and Jenni ngs, P. A. (2003) The solvent in CNBr cleavage reactions determines the fragmentation efficiency of ketosteroid isomerase fusion proteins used in the production of recombinant peptides, Protein Expression and Purification 28 224 231. 122. Novagen, I nc (1 994) pET Peptide Expression System 31. 123. Novagen, I nc (1999) pET system manual, 8th ed. 124. Aguilar, M. I. Methods in molecular biology HPLC of peptides and proteins: methods and protocols Vol. 251, Humana Press Inc., Totowa, NJ.
196 126. Terzi, E., Ho lzemann, G., and Seelig, J. (1997) Interaction of Alzheimer beta amyloid peptide(1 40) with lipid membranes, Biochemistry 36 14845 14852. 127. Ramamoorthy, A., Thennarasu, S., Lee, D. K., Tan, A. M., and Maloy, L. (2006) Solid state NMR investigation of t he membrane disrupting mechanism of antimicrobial peptides MSI 78 and MSI 594 derived from magainin 2 and melittin, Biophysical Journal 91 206 216. 128. Wieprecht, T., Apostolov, O., Beyermann, M., and Seelig, J. (1999) Thermodynamics of the alpha helix c oil transition of amphipathic peptides in a membrane environment: Implications for the peptide membrane binding equilibrium, Journal of Molecular Biology 294 785 794. 129. Veldhuizen, R., Nag, K., Orgeig, S., and Possmayer, F. (1998) The role of lipids in pulmonary surfactant, Biochimica Et Biophysica Acta Molecular Basis of Disease 1408 90 108. 130. Chen, P. S., Toribara, T. Y., and Warner, H. (1956) Microdetermination of phosphorus, Analytical Chemistry 28 1756 1758. 131. Seelig, A., and Seelig, J. (19 74) Dynamic structure of fatty acyl chains in a phospholipid bilayer measured by deuterium magnetic resonance, Biochemistry 13 4839 4845. 132. Antharam, V. C., Elliott, D. W., Mills, F. D., Farver, R. S., Sternin, E., and Long, J. R. (2009) Penetration de pth of surfactant peptide KL4 into membranes is determined by fatty acid saturation, Biophysical Journal 96 4085 4098. 133. Chapman, R., Kulp, J. L., Patgiri, A., Kallenbach, N. R., Bracken, C., and Arora, P. S. (2008) Trapping a folding intermediate of t he alpha helix: Stabilization of the pi helix, Biochemistry 47 4189 4195. 134. Andrade, M. A., Chacon, P., Merelo, J. J., and Moran, F. (1993) Evaluation of secondary structure of proteins from UV circular dichroism spectra using an unsupervised learning neural network, Protein Engineering 6 383 390. 135. Qiu, X. X., Mirau, P. A., and Pidgeon, C. (1993) Magnetically induced orientation of phosphatidylcholine membranes, Biochimica Et Biophysica Acta 1147 59 72. 136. Seelig, J., Borle, F., and Cross, T. A. (1985) Magnetic ordering of phospholipid membranes, Biochimica Et Biophysica Acta 814 195 198.
197 137. Saenz, A., Canadas, O., Bagatolli, L. A., Johnson, M. E., and Casals, C. (2006) Physical properties and surface activity of surfactant like membranes cont aining the cationic and hydrophobic peptide KL4, Febs Journal 273 2515 2527 138. Ma, J. W., Koppenol, S., Yu, H. U., and Zografi, G. (1998) Effects of a cationic and hydrophobic peptide, KL4, on model lung surfactant lipid monolayers, Biophysical Journal 74 1899 1907. 139. Bunow, M. R., and Levin, I. W. (1977) Raman spectra and vibrational assignments for deuterated membrane lipids 1,2 dipalmitoyl phosphatidylcholine d9 and 1,2 dipalmitoyl phosphatidylcholine d62, Biochimica Et Biophysica Acta 489 191 206. 140. Seelig, J., and Seelig, A. (1980) Lipid conformation in model membranes and biological membranes, Quarterly Reviews of Biophysics 13 19 61. 141. Henzler Wildman, K. A., Martinez, G. V., Brown, M. F., and Ramamoorthy, A. (2004) Perturbation of t he hydrophobic core of lipid bilayers by the human antimicrobial peptide LL 37, Biochemistry 43 8459 8469. 142. Porcelli, F., Buck, B., Lee, D. K., Hallock, K. J., Ramamoorthy, A., and Veglia, G. (2004) Structure and orientation of pardaxin determined by NMR experiments in model membranes, Journal of Biological Chemistry 279 45815 45823. 143. Abu Baker, S., Qi, X. Y., and Lorigan, G. A. (2007) Investigating the interaction of saposin c with POPS and POPC phospholipids: A solid state NMR spectroscopic stud y, Biophysical Journal 93 3480 3490. 144. Abu Baker, S., Qi, X. Y., Newstadt, J., and Lorigan, G. A. (2005) Structural changes in a binary mixed phospholipid bilayer of DOPG and DOPS upon saposin C interaction at acidic pH utilizing P 31 and H 2 solid sta te NMR spectroscopy, Biochimica Et Biophysica Acta Biomembranes 1717 58 66. 145. Huster, D., Dietrich, U., Gutberlet, T., Gawrisch, K., and Arnold, K. (2000) Lipid matrix properties in cationic membranes interacting with anionic polyelectrolytes: A solid state NMR approach, Langmuir 16 9225 9232. 146. Laroche, G., Dufourc, E. J., Pezolet, M., and Dufourcq, J. (1990) Coupled changes between lipid order and polypeptide conformation at the membrane surface a H 2 NMR and raman study of polylysine phosphatid ic acid systems, Biochemistry 29 6460 6465. 147. Stockton, G. W., and Smith, I. C. P. (1976) Deuterium nuclear magnetic resonance study of condensing effect of cholesterol on egg phosphatidylcholine bilayer membranes .1. Perdeuterated fatty acid probes, C hemistry and Physics of Lipids 17 251 263.
198 148. Oldfield, E., Meadows, M., Rice, D., and Jacobs, R. (1978) Spectroscopic studies of specifically deuterium labeled membrane systems nuclear magnetic resonance investigation of effects of cholesterol in mod el systems, Biochemistry 17 2727 2740. 149. Nezil, F. A., and Bloom, M. (1992) Combined influence of cholesterol and synthetic amphiphillic peptides upon bilayer thickness in model membranes, Biophysical Journal 61 1176 1183. 150. de Planque, M. R. R., G reathouse, D. V., Koeppe, R. E., Schafer, H., Marsh, D., and Killian, J. A. (1998) Influence of lipid/peptide hydrophobic mismatch on the thickness of diacylphosphatidylcholine bilayers. A H 2 NMR and ESR study using designed transmembrane alpha helical pe ptides and gramicidin A, Biochemistry 37 9333 9345. 151. Vogt, B., Ducarme, P., Schinzel, S., Brasseur, R., and Bechinger, B. (2000) The topology of lysine containing amphipathic peptides in bilayers by circular dichroism, solid state NMR, and molecular m odeling, Biophysical Journal 79 2644 2656. 152. Perkins, W. R., Dause, R. B., Parente, R. A., Minchey, S. R., Neuman, K. C., Gruner, S. M., Taraschi, T. F., and Janoff, A. S. (1996) Role of lipid polymorphism in pulmonary surfactant, Science 273 330 332. 153. Moya, F. R., Gadzinowski, J., Bancalari, E., Salinas, V., Kopelman, B., Bancalari, A., Kornacka, M. K., Merritt, T. A., Segal, R., Schaber, C. J., Tsai, H., Massaro, J., d'Agostino, R., and Int Surfaxin Collaborative Study, G. (2005) A multicenter, r andomized, masked, comparison trial of lucinactant, colfosceril palmitate, and beractant for the prevention of respiratory distress syndrome among very preterm infants, Pediatrics 115 1018 1029. 154. Mills, F. D., Antharam, V. C., Ganesh, O. K., Elliott, D. W., McNeill, S. A., and Long, J. R. (2008) The helical structure of surfactant peptide KL(4) when bound to POPC : POPG lipid vesicles, Biochemistry 47 8292 8300 155. Ryan, M. A., Akinbi, H. T., Serrano, A. G., Perez Gil, J., Wu, H. X., McCormack, F. X ., and Weaver, T. E. (2006) Antimicrobial activity of native and synthetic surfactant protein B peptides, Journal of Immunology 176 416 425. 156. Weaver, T. E., and Conkright, J. J. (2001) Functions of surfactant proteins B and C, Annual Review of Physiol ogy 63 555 578. 157. Bechinger, B. (2005) Detergent like properties of magainin antibiotic peptides: A P 31 solid state NMR spectroscopy study, Biochimica Et Biophysica Acta Biomembranes 1712 101 108.
199 158. Ye, Q., and Biltonen, R. L. (1994) Differential scanning and dynamic calorimetric studies of cooperative phase transitions in phospholipid bilayer membranes, Sub cellular B iochemistry 23 121 160. 159. Cochrane, C. G., and Revak, S. D. (1994) Protein phospholipid interactions in pulmonary surfactant, Ch est 105 S57 S62. 160. Chapman, R., Kulp, J. L., Patgiri, A., Kallenbach, N. R., Bracken, C., and Helix: Helix, Biochemistry 47 4189 4195. 161. Antharam, V. C., Elliott, D W., Mills, F. D., Farver, R. S., Sternin, E., and Long, J. R. (2009) The penetration depth of surfactant peptide KL4 into membranes is determined by fatty acid saturation, Biophys. J. 96, 4085 4098 162. Long, J. R. Mehta, A. K., Meng, Y., Roitberg, A. E., Antharam V C ., Ganesh, O. K., and Farver, R. S. (200 9 ) The structural plasticity of surfactant peptide KL4 in lipid membranes, Biophys. J. 96, 159 163. Warschawski, D. E., Arnold, A. A., Beaugrand, M., Gravel, A., Chartrand, E., and Marcotte, I. (20 11) Choosing membrane mimetics for NMR structural studies of transmembrane proteins, Biochimica Et Biophysica Acta Biomembranes 1808 1957 1974 164. King, D. S., Fields, C. G., and Fields, G. B. (1990) A cleavage method which minimizes side reactions foll owing Fmoc solid phase peptide synthesis, International Journal of Peptide and Protein Research 36 255 266. 165. Samuel Landtiser, M., Zachariah, C., Williams, C. R., Edison, A. S., and Long, J. R. (2005) Uniform isotopic labeling of proteins through expr ession in E.Coli using IPTG induction, In Current Protocols in Protein Science pp 26.23.28 26.23.29. 166. Piotto, M., Saudek, V., and Sklenar, V. (1992) Gradient tailored excitation for single quantum NMR spectroscopy of aqueous solutions, Journal of Biom olecular N MR 2 661 665. 167. Sklenar, V., Piotto, M., Leppik, R., and Saudek, V. (1993) Gradient tailored water suppression for H 1 N 15 HSQC experiments optimized to retain full sensitivity, Journal of Magnetic Resonance Series A 102 241 245. 168. Mario n, D., Ikura, M., Tschudin, R., and Bax, A. (1989) Rapid recording of 2D NMR spectra without phase cycling application to the study of hydrogen exchange in proteins, Journal of Magnetic Resonance 85 393 399
200 169. Bax, A., and Davis, D. G. (1985) MLEV 17 based two dimentional homonuclear magnetization transfer spectroscopy, Journal of Magnetic Resonance 65 355 360. 170. Shaka, A. J., Lee, C. J., and Pines, A. (1988) Iterative schemes for bilinear operators application to spin decoupling Journal of Mag netic Resonance 77 274 293. 171. Delaglio, F., Grzesiek, S., Vuister, G. W., Zhu, G., Pfeifer, J., and Bax, A. (1995) NMRPIPE a multidimensional spectral processing system based on unix pipes, Journal of Biomolecular N MR 6 277 293. 172. Goddard, T. D., and Kneller, D. G. (2008) SPARKY 3, University of California, San Francisco. 173. Schwieters, C. D., and Clore, G. M. (2001) The VMD XPLOR visualization package for NMR structure refinement, Journal of Magnetic Resonance 149 239 244. 174. Morris, A. L., Macarthur, M. W., Hutchinson, E. G., and Thornton, J. M. (1992) Stereochemical quality of protein structure coordinates, Proteins Structure Function and Genetics 12 345 364. 175. Chen, V. B., Arendall, W. B., Headd, J. J., Keedy, D. A., Immormino, R. M., Kapral, G. J., Murray, L. W., Richardson, J. S., and Richardson, D. C. (2010) MolProbity: all atom structure validation for macromolecular crystallography, Acta Crystallographica Section D Biological Crystallography 66 12 21. 176. Greenfield, N., and Fasm an, G. D. (1969) Computed circular dichroism spectra for evaluation of protein conformation, Biochemistry 8 4108 4116. 177. Pastore, A., and Saudek, V. (1990) The relationship between chemical shift and secondary structure in proteins, Journal of Magnetic Resonance 90 165 176. 178. Wishart, D. S., Sykes, B. D., and Richards, F. M. (1992) The chemical shift index a fast and simple method for the assignment of protein secondary structure through NMR spectroscopy, Biochemistry 31 1647 1651. 179. Smith, L. J., Bolin, K. A., Schwalbe, H., MacArthur, M. W., Thornton, J. M., and Dobson, C. M. (1996) Analysis of main chain torsion angles in proteins: Prediction of NMR coupling constants for native and random coil conformations, Journal of Molecular Biology 255 494 506.
201 180. Millhauser, G. L., Stenland, C. J., Hanson, P., Bolin, K. A., and vandeVen, F. J. M. (1997) Estimating the relative populations of 3(10) helix and alpha helix in Ala rich peptides: A hydrogen exchange and high field NMR study, Journal of Mol ecular Biology 267 963 974. 181. Dyson, H. J., Rance, M., Houghten, R. A., Wright, P. E., and Lerner, R. A. (1988) Folding of immunogenic peptide fragments of proteins in water solution .2. The nascent helix, Journal of Molecular Biology 201 201 217. 182 Nemethy, G., Phillips, D. C., Leach, S. J., and Scheraga, H. A. (1967) A second right handed helical structure with parameters of Pauling Corey alpha helix, Nature 214 363 365 183. Richardson, J. S. (1981) The anatomy and taxonomy of protein structure, Advances in protein chemistry 34 167 339. 184. Barlow, D. J., and Thornton, J. M. (1988) Helix geometry in proteins, Journal of Molecular Biology 201 601 619. 185. Sikorska, E., Greber, K., Rodziewicz Motowidlo, S., Szultka, L., Lukasiak, J., and Kamysz W. (2009) Synthesis and antimicrobial activity of truncated fragments and analogs of citropin 1.1: The solution structure of the SDS micelle bound citropin like peptides, Journal of Structural Biology 168 250 258. 186. Zamoon, J., Mascioni, A., Thomas, D. D., and Veglia, G. (2003) NMR solution structure and topological orientation of monomeric phospholamban in dodecylphosphocholine micelles, Biophysical Journal 85 2589 2598. 187. Lamberth, S., Schmid, H., Muenchbach, M., Vorherr, T., Krebs, J., Carafoli E., and Griesinger, C. (2000) NMR solution structure of phospholamban, Helvetica Chimica Acta 83 2141 2152. 188. Pollesello, P., Annila, A., and Ovaska, M. (1999) Structure of the 1 36 amino terminal fragment of human phospholamban by nuclear magnetic r esonance and modeling of the phospholamban pentamer, Biophysical Journal 76 1784 1795. 189. Bhunia, A., Domadia, P. N., Torres, J., Hallock, K. J., Ramamoorthy, A., and Bhattacharjya, S. (2010) NMR structure of pardaxin, a pore forming antimicrobial pepti de, in lipopolysaccharide micelles, Journal of Biological Chemistry 285 3883 3895. 190. Bhunia, A., Ramamoorthy, A., and Bhattacharjya, S. (2009) Helical hairpin structure of a potent antimicrobial peptide MSI 594 in lipopolysaccharide micelles by NMR spe ctroscopy, Chemistry a European Journal 15 2036 2040.
202 191. Domadia, P. N., Bhunia, A., Ramamoorthy, A., and Bhattacharjya, S. (2010) Structure, interactions, and antibacterial activities of MSI 594 derived mutant peptide MSI 594F5A in lipopolysaccharide m icelles: role of the helical hairpin conformation in outer membrane permeabilization, Journal of the American Chemical Society 132 18417 18428. 192. Bruhn, H. (2005) A short guided tour through functional and structural features of saposin like proteins, Biochemical Journal 389 249 257. 193. Liepinsh, E., Andersson, M., Ruysschaert, J. M., and Otting, G. (1997) Saposin fold revealed by the NMR structure of NK lysin, Nature Structural Biology 4 793 795. 194. Anderson, D. H., Sawaya, M. R., Cascio, D., Ern st, W., Modlin, R., Krensky, A., and Eisenberg, D. (2003) Granulysin crystal structure and a structure derived lytic mechanism, Journal of Molecular Biology 325 355 365. 195. de Alba, E., Weiler, S., and Tjandra, N. (2003) Solution structure of human sapo sin c: pH dependent interaction with phospholipid vesicles, Biochemistry 42 14729 14740. 196. Poulain, F. R., Allen, L., Williams, M. C., Hamilton, R. L., and Hawgood, S. (1992) Effects of surfactant apolipoproteins on liposome structure implications fo r tubular myelin formation, American Journal of Physiology 262 L730 L739. 197. Oosterlakendijksterhuis, M. A., Haagsman, H. P., Vangolde, L. M. G., and Demel, R. A. (1991) Characterization of lipid insertion into monomolecular layers mediated by lung surf actant proteins SP B and SP C, Biochemistry 30 10965 10971. 198. Ryan, M. A., Qi, X. Y., Serrano, A. G., Ikegami, M., Perez Gil, J., Johansson, J., and Weaver, T. E. (2005) Mapping and analysis of the lytic and fusogenic domains of surfactant protein B, B iochemistry 44 861 872 199. Lipp, M. M., Lee, K. Y. C., Zasadzinski, J. A., and Waring, A. J. (1996) Phase and morphology changes in lipid monolayers induced by SP B protein and its amino terminal peptide, Science 273 1196 1199. 200. Ding, J. Q., Takamo to, D. Y., von Nahmen, A., Lipp, M. M., Lee, K. Y. C., Waring, A. J., and Zasadzinski, J. A. (2001) Effects of lung surfactant proteins, SP B and SP C, and palmitic acid on monolayer stability, Biophysical Journal 80 2262 2272. 201. Takamoto, D. Y., Lipp, M. M., von Nahmen, A., Lee, K. Y. C., Waring, A. J., and Zasadzinski, J. A. (2001) Interaction of lung surfactant proteins with anionic phospholipids, Biophysical Journal 81 153 169.
203 202. Schram, V., and Hall, S. B. (2001) Thermodynamic effects of the hy drophobic surfactant proteins on the early adsorption of pulmonary surfactant, Biophysical Journal 81 1536 1546. 203. Lipp, M. M., Lee, K. Y. C., Takamoto, D. Y., Zasadzinski, J. A., and Waring, A. J. (1998) Coexistence of buckled and flat monolayers, Phy sical Review Letters 81 1650 1653. 204. Ding, J. Q., Doudevski, I., Warriner, H. E., Alig, T., and Zasadzinski, J. A. (2003) Nanostructure changes in lung surfactant monolayers induced by interactions between palmitoyloleoylphosphatidylglycerol and surfac tant protein B, Langmuir 19 1539 1550. 205. Krol, S., Ross, M., Sieber, M., Kunneke, S., Galla, H. J., and Janshoff, A. (2000) Formation of three dimensional protein lipid aggregates in monolayer films induced by surfactant protein B, Biophysical Journal 79 904 918 206. Hawgood, S., Derrick, M., and Poulain, F. (1998) Structure and properties of surfactant protein B, Biochimica Et Biophysica Acta Molecular Basis of Disease 1408 150 160 207. Hawco, M. W., Coolbear, K. P., Davis, P. J., and Keough, K. M. W. (1981) Exclusion of fluid lipid during compression of monolayers of mixtures of dipalmitoylphosphatidylcholine with some other phosphatidylcholines, Biochimica Et Biophysica Acta 646 185 187 208. Discher, B. M., Maloney, K. M., Schief, W. R., Grainge r, D. W., Vogel, V., and Hall, S. B. (1996) Lateral phase separation in interfacial films of pulmonary surfactant, Biophysical Journal 71 2583 2590. 209. Suzuki, Y., Fujita, Y., and Kogishi, K. (1989) Reconstitution of tubular myelin from synthetic lipids and proteins associated with pig pulmonary surfactant, American Review of Respiratory Disease 140 75 81. 210. Williams, M. C., Hawgood, S., and Hamilton, R. L. (1991) Changes in lipid structure produced by surfactant proteins SP A, SP B, and SP C, Americ an Journal of Respiratory Cell and Molecular Biology 5 41 50. 211. Merutka, G., Dyson, H. J., and Wright, P. E. (1995) Random coil H 1 chemical shifts obtained as a function of temperature and trifluoroethanol concentration for the peptide series GGXGG, J ournal of Biomolecular N MR 5 14 24 212. Wishart, D. S., and Nip, A. M. (1998) Protein chemical shift analysis: a practical guide, Biochemistry and Cell Biology Biochimie Et Biologie Cellulaire 76 153 163.
204 213. Tieleman, D. P., van der Spoel, D., and Ber endsen, H. J. C. (2000) Molecular dynamics simulations of dodecylphosphocholine micelles at three different aggregate sizes: Micellar structure and chain relaxation, Journal of Physical Chemistry B 104 6380 6388.
205 BIOGRAPHICAL SKETCH Anna Kuznetsova wa s born in 1984 in Stupino Moscow region, Russia. Sh e received h er chemistry from Moscow State University in 200 6 In 2007, Anna Kuznetsova joined the graduate program in chemistry department at the University of Florida. Sh e received h er Ph.D. from the University of Florida in the spring of 201 3 Anna Kuznetsova is the author of several scientific papers. She was also a TA in General Chemistry (I, II) and Biochemistry Laboratory classes.