UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Studies on the transport of metal ions and modified amino acids and peptides into liposomes in response… Chakrabarti, Ajoy C. 1992

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

Item Metadata

Download

Media
831-ubc_1993_spring_phd_chakrabarti_ajoy.pdf [ 5.06MB ]
Metadata
JSON: 831-1.0086350.json
JSON-LD: 831-1.0086350-ld.json
RDF/XML (Pretty): 831-1.0086350-rdf.xml
RDF/JSON: 831-1.0086350-rdf.json
Turtle: 831-1.0086350-turtle.txt
N-Triples: 831-1.0086350-rdf-ntriples.txt
Original Record: 831-1.0086350-source.json
Full Text
831-1.0086350-fulltext.txt
Citation
831-1.0086350.ris

Full Text

We accept this thesis as conformingSTUDIES ON THE TRANSPORT OF METAL IONS AND MODIFIEDAMINO ACIDS AND PEPTIDES INTO LIPOSOMES IN RESPONSE TOTRANSMEMBRANE pH GRADIENTSbyAJOY C. CHAKRABARTIB.Sc.(Hons) Carleton University, 1986M.Sc. Carleton University, 1988A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIESDepartment of BiochemistryTHE UNMERSITY OF BRITISH COLUMBIAOctober. 1992© Ajoy C. Chakrabarti. 1992In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(SignaturBiochemistryDepartment of ^The University of British ColumbiaVancouver, CanadaDate December 10, 1992DE-6 (2/88)ABSTRACTA universal characteristic of cells is the membrane boundarythat separates the interior cytoplasm from the external environment.This thesis examines the net transport of molecules such as certainmetal ions, modified amino acids and peptides across the membranesof model (liposome) systems in response to transmembrane pHgradients. This net transport arises from the membrane permeablenature of the neutral form of the molecule or ion complex. Forcomparison, permeability studies were conducted for several types ofamino acids which cannot readily adopt a neutral form.The first section demonstrates that iron (primarily in the form ofFe+2) and barium can be accumulated into EPC and DSPC-cholesterol (55:45 mole %) large unilamellar vesicles (LUV's) inresponse to a transmembrane pH gradient (interior acidic) and in thepresence of the ionophore A23187. It is shown that the maximallyloaded Fe- and Ba-containing LUVs exhibit increased densities in thata significant fraction of the maximally loaded LUV's can be pelleted bylow speed centrifugation and the Ba+2-loaded systems can be directlyvisualized by cryo-electron microscopy.The second area of investigation concerned the uptake ofderivatives of lysine and a pentapeptide (Ala-Met-Leu-Trp-Ala), inwhich the C-terminal carboxyl functions have been converted tomethyl esters or amides, in response to transmembrane pH gradientsin LUV systems. It is shown that the presence of a pH gradient(interior acidic) results in the rapid and efficient accumulation ofthese basic amino acid and peptide derivatives into LUVs in a mannerconsistent with permeation of the neutral (deprotonated) form. Theiipermeability coefficient of the neutral form of lysine methyl ester was2.1 X 10"2 cm.s"1. It is suggested that this property may have generalimplications for mechanisms of transbilayer translocation of peptides,such as signal sequences, which exhibit weak base characteristics.The third part of the thesis concerned the permeabilityproperties of neutral, hydrophobic, polar and charged amino acids,which did not have weak base properties. The rates of efflux ofglycine, lysine, phenylalanine, serine and tryptophan were determinedafter they were passively entrapped in EPC and DMPC LUVs. Thepermeability coefficients for the neutral, polar and charged aminoacids were approximately-110-12 cm s for EPC vesicles, while thosefor DMPC vesicles were approximately 10-11 cm s"1. The LUVs were10-100 times more permeable to the hydrophobic amino acids.Variations in pH had only minor effects on the permeabilitycoefficients. Permeation rates for the amino acids studied were 108slower than those of the modified amino acids indicating that differentmechanisms of efflux, including permeation of the neutral or chargedforms of the amino acids or transient defects, may be responsible.The final area of study involved examining the influence ofhydrophobicity and charge distribution on the uptake of modifiedpeptides, which were weak bases. Here, the ability of small (2-3 aminoacid) peptides, composed exclusively of basic (lysine) and hydrophobic(tryptophan) amino acids, to accumulate into LUV systems wasinvestigated. In the case of the dipeptides Trp-Lys-Amide and Lys-Trp-Amide, remarkable differences in the rate constants associated withnet transport were observed. In EPC:cholesterol LUV systemsexhibiting a ApH of 3 units (pH1=4.0; pH0=7.0), for example, the rateconstant for the uptake of the Lys-Trp-Amide was some 5 X 103 fasterthan for the Trp-Lys-Amide. This difference could not be attributed tochanges in the membrane-water partition coefficients. Related effectswere observed for the tripeptides composed of one lysine and twotryptophan residues. It is concluded that different chargedistributions in short peptides of identical amino acid compositioncan profoundly influence the ability of these groups to associate withand permeate across lipid bilayers.ivTABLE OF CONTENTSABSTRACT^ iiTABLE OF CONTENTS^ vLIST OF TABLES viiiLIST OF FIGURES^ ixABBREVIATIONS xiiACKNOWLEDGEMENTS^ xv1.^INTRODUCTION1.1 Biological Membranes as Permeability Barriers ^ 11.2 Structures of Membranes ^ 21.3 The Structures of Lipids 51.4 Properties of Lipids ^ 51.5 Transport across Biological Membranes^ 91.6 Preparation of Model Membrane Systems 101.6.1 Multilamellar Vesicles ^ 121.6.2 Small Unilamellar Vesicles 121.6.3 Large Unilamellar Vesicles ^ 141.6.4 Monolayers ^ 161.6.5 Planar Lipid Membranes ^ 161.7 Uses of Liposomal Model Membrane Systems ^ 161.8 Transmembrane Electrical Potentials andpH Gradients ^ 181.9.1 Transport of Nonionic Solutes ^ 211.9.2 Transport of Ions ^ 241.9.3 Lipid Factors which influence Ion Transport ^271.9.4 Transport using lonophores ^ 281.9.5 Transport of Signal Sequences and other Peptides ^29vvi1.10 Summary^ 332. GENERATION AND CHARACTERIZATION OF IRON ANDBARIUM LOADED LIPOSOMES2.1 Introduction ^ 352.2 Materials and Methods ^ 362.3 Results ^ 402.4 Discussion 533. UPTAKE OF BASIC AMINO ACIDS AND PEPTIDES INTOLIPOSOMES IN RESPONSE TO TRANSMEMBRANE pHGRADIENTS3.1 Introduction ^ 553.2 Materials and Methods ^ 573.3 Results ^ 613.4 Discussion 764. PERMEABILITY OF LIPID BILAYERS TO AMINO ACIDS4.1 Introduction ^ 794.2 Materials and Methods ^ 814.3 Results ^ 864.4 Discussion 925. INFLUENCE OF CHARGE, CHARGE DISTRIBUTION ANDHYDROPHOBICITY ON THE TRANSPORT OF SHORT MODELPEPTIDES INTO LIPOSOMES IN RESPONSE TOTRANSMEMBRANE pH GRADIENTS5.1 Introduction ^ 975.2 Materials and Methods ^ 995.3 Results ^ 1045.4 Discussion 116vii6. SUMMARY^ 1217. REFERENCES 126LIST OF TABLES1.1. Names and structures of some common fatty acids^62.1. Effect of liposome size on iron entrapment^452.2. Gravimetric properties of EPC LUVs loaded with iron^462.3. Gravimetric properties of EPC LUVs loaded with barium(100-nm pore size)^ 484.1. Permeability coefficients (P) and rate constants (k) for the effluxof glycine, lysine and serine from 200 nm EPC vesiclesat various pH values^ 884.2. Permeability coefficients (P) and rate constants (k) for the effluxof glycine, lysine and serine from 200 nm DMPC vesicles atvarious pH values^ 894.3. Partition coefficients for some of the amino acids studied (fromLeo et al., 1971)^ 955.1. Partition coefficients (K') and activation energies (Ea) for the di-and tri- peptides^ 113viiiLIST OF FIGURES1.1. Structural model of a biological membrane^ 31.2. The structure of a phospholipid and cholesterol ^ 71.3. Diagrams of the different phases that can be adopted bylipids 81.4. Various forms of transmembrane transport utilized by cells.... 111.5. Structural representations of the common classes of modelmembrane systems^ 131.6. The relationship between Alp and ApH^ 191.7. Redistribution of weak bases in response to transmembranepH gradients^ 221.8. Permeability coefficients of lipid bilayers to various types ofmolecules^ 251.9. Structure of A23187 (modified from Ma et al., 1990)^301.10. A diagrammatic representation of protein translocation acrossmembranes^ 322.1. Time course of uptake for Fe+2 and Fe+3 into 100 nm EPCvesicles bearing a transmembrane pH gradient (pHi=4.0;pH0=7.5)^ 412.2. Influence of A23187 concentrations on the accumulation ofFe+2 in 100 nm EPC vesicles bearing a transmembrane pHgradient (pH1=4.0; pH0=7.5)^ 42ix2.3. A. Effect of varying the external iron concentration and internalcitrate concentration on the amount of iron accumulated by 100nm EPC vesicles and B. Effect of varying external ironconcentrations and internal citrate concentrations on thetrapping efficiency^ 442.4. Time course of loading Fe+2 into (A) EPC LUVs and (B)DSPC:cholesterol LUVs. Part (C) illustrates the retention ofbarium following loading into DSPC:cholesterol LUVs^492.5. Cryo-electron micrograph of 100 nm DSPC:cholesterol vesiclescontaining barium^ 523.1. Time course of uptake of lysine methyl ester into 100 nm EPCvesicles^ 623.2. Time course of uptake of lysine methyl ester into 100 nm EPCvesicles for different external pH values^ 643.3. Arrhenius plot of the rate constants (k) for lysine methyl esteruptake^ 663.4. A) Time course of uptake of the pentapeptide (Ala-Met-Leu-Trp-Ala methyl ester) into 100 nm EPC vesicles^693.4. B) Time course of uptake of the pentapeptide into 100 nm EPCvesicles exhibiting a valinomycin induced K+ membranepotential with high and low internal buffering capacity^703.5. Time course of uptake of the hydrophobic pentapeptide into100 nm EPC vesicles for different external pH values^743.6. Arrhenius plot of the rate constants (k) for pentapeptideuptake^ 744.1. A) Time course of release of serine from 200 nm DMPC vesiclesat various pH values and B) calculation of rate constants ^87x4.2. A) Time course of the efflux of tryptophan from 200 nm EPCliposomes at pH 6.0 measured by the quenching of tryptophanfluorescence and B) time course of influx of phenylalaninemonitored by changes in light scattering^ 915.1. Structures of the model peptides that were synthesized ^ 1005.2. Time course of uptake of Lys-Trp-Amide and Trp-Lys-Amideinto 100 nm EPC vesicles exhibiting a pH gradient (pHi. 4.0;pHo= 7.5)  1055.3. Time course of uptake of Lys-Trp-Amide and Trp-Lys-Amideinto 100 nm EPC:cholesterol (55:45; mol:mol) exhibiting atransmembrane pH gradient^ 1075.4. Arrhenius plot of the rate constants (k) for Lys-Trp-Amideuptake^ 1085.5. Arrhenius plot of the rate constants (k) for Trp-Lys-Amideuptake^ 1105.6. Time course of uptake of tri-peptides into 100 nm EPC vesiclesexhibiting a transmembrane pH gradient^ 114xiABBREVIATIONS USEDApH^Transmembrane pH gradientAlp^Transmembrane electrochemical potentialXem^Emission wavelengthXex^Excitation wavelengthA Area of the LUV membraneBaC12^Barium chlorideCHES^2-(N-cyclohexylamino)ethane-sulfonic acidCtl^ControlEa^Activation energyEPPS^N-(2-Hydroxyethyl)piperazine-N'-3-propanesulfonicacidFeC13^Ferric chlorideFe504^Ferrous sulfateFATMLVs Frozen and thawed MLVsH+^ProtonHA^Neutral amineH2A+^Protonated amineHBS^HEPES-buffered salineHEPES^[4-(2-Hydoxyethy1)1-piperazine ethanesulfonic acidIM^Inner monolayerRate constantDissociation constantK'^Membrane-water partition coefficient of the chargedspeciesLUVs^Large unilamellar vesiclesLUVET^Large unilamellar vesicles by extrusion techniquesxiiLys^LysineMES^2-(N-Morpholino)ethanesulfonic acidMLVs^Multilamellar vesiclesNA^Not applicableNMR^Nuclear magnetic resonanceOM^Outer monolayerEffective permeability coefficientpK^Negative log of the dissociation constantPm^Membrane permeability coefficientPhospholipidsEPC^(Egg) PC derived from hen egg yolkDMPC^DimyristoylphosphatidylcholineDSPC^DistearoylphosphatidylcholinePC^PhosphatidylcholinePE^PhosphatidylethanolaminePG^PhosphatidylglycerolPI^PhosphatidylinositolPS^PhosphatidylserinePL^PhospholipidPLT^Pelletpsi^Pounds per square inchQELS^Quasi-elastic light scatteringSEM^Standard error of the meanSPN^SupernatantSUVs^Small unilamellar vesiclest1/2^Half-time for transportTPP+^TetraphenylphosphoniumxivTe^Gel to liquid-crystalline temperatureTLC^Thin layer chromatographyTNBS^Trinitrobenzenesulfonic acidTricine^N-tris(hydroxymethyl)-methylglycineTris^2 -Amino-2(hydroxymethyl)propane- 1.3 -diolTrp^TryptophanUV^UltravioletACKNOWLEDGEMENTSCullis Lab: Pieter (special thanks for taking all the abuse, for lettingme do my own thing and, "as a final point of interest", fordramatically improving all of my papers), Nellie, Nancy, Richard(for critical abilities beyond the call of duty & for the cabin onGaliano), Kim, Torn M., Jeff V., Michel, Myrna, Barb, Shane,Troy, Austin, Conrad (or is it "Rad-Con"),....Storey Lab (Carleton U.; "my lab away from home") - Ken & Jan - Foracting as my "surrogate" scientific parents; Steve, Jeannie,...Inger (my companion-to-be for the India trip) and Capt. DAKDeamer Lab (UC Davis)- Dave (my "supervisor-to-be"), Ann, John - Forall the encouragement, support and friendship.Housemates: Diane, Leanne, Laura (Kits) - For being great friends andgiving me insight into female psychology; Kate (Davis) ; Kristin(East Van.) - For some wild parties with Oceanography.Misc. Friends - Yardenah (extra-special thanks for all the support,caring and fun times; you made writing up easier but leavingmore difficult...), Anya & Karen, Katherine & Jihan, John, all ofmy "Soccer Saturday" buddies.... A final thank you to all thoseunmentioned persons in Biochem., Oceanography, etc., whoenriched my life through their presence and participation in it.This work was supported financially by a University GraduateFellowship (UGF); a GREAT Award from the B.C. Science Council; aNASA Planetary Biology Internship (PBI) and a PPMATE (Pieter PaysMe At The End).This thesis is dedicated to the memory of Gene Roddenberry (1991),who inspired many in my generation to "boldly go where no one hasgone before....""Make us choose the harder right instead of the easier wrong, andnever be content with a half truth when the whole truth can be won."- West Point cadet prayer"Money won is twice as sweet as money earned."- Paul Newman as pool hustler "Fast Eddie" Felson in the film "TheColour of Money"."I'll be back...."- Arnold Schwarzenneger in the film "The Terminator"1CHAPTER 1 - INTRODUCTION1.1 Biological Membranes as Permeability Barriers A universal characteristic of cells is the membrane boundary whichseparates the interior cytoplasm from the exterior environment. Thisboundary structure has two basic components: the lipid bilayer whichprovides a barrier to the free diffusion of solutes; while specialized integralmembrane proteins permit specific solutes to pass, either by passivediffusion (channels or carriers) or by enzyme-catalysed active transport.This permeability barrier allows for the formation of concentration gradientsof metabolites and ions across biological membranes. Transmembrane iongradients (such as K+, Na+, Ca+2 and H+ gradients) have essential roles inprocesses such as energy transduction and metabolite translocation. Thework presented in this thesis examines the transport of metal ions, aminoacids, modified amino acids and peptides into liposomes (model membranesystems) in response to transmembrane pH gradients in the absence of theproteins usually associated with such transport.This chapter provides background knowledge concerning thecharacteristics of lipids and membranes with regard to the studiesundertaken in this thesis. The structures and properties of lipids andmembranes will be briefly described. The methods of production andcharacteristics of model membrane systems will then be discussed.Measurement of transmembrane ion gradients will be examined. Finally,transport processes across lipid membranes will be briefly reviewed, withparticular emphasis on transport involving ionophores, nonionic solutes,electrolytes, peptides and signal sequences.21.2 Structure of Membranes The fluid mosaic model developed by Singer and Nicholson in 1972 isstill regarded as the standard model for the structure of biologicalmembranes (see Fig. 1.1; Singer and Nicholson, 1972). The fluid aspectrelates to the fluid nature of the lipid bilayer, which allows rapid lateraldiffusion of membrane components, whereas the mosaic aspect refers toproteins extending into and through the bilayer. In mammalian systems,phospholipids are the major lipid species found in membranes. They areamphipathic molecules that will spontaneously form ordered structures,such as bilayers, in aqueous media. The phospholipids are therefore themolecules that are most responsible for establishing membrane structure.Proteins, in the form of integral (generally membrane-spanning proteins)and peripheral (surface-associated) proteins, serve many functions such asenergy transducers, enzymes and receptors for transmembrane signals. Inaddition, oligosaccharides (carbohydrates linked to lipids and proteins) actas receptors for various ligands, as well as being involved in cell attachmentand recognition (Kleinfeld, 1987).The key element to membrane structure is that membranes arecooperative structures; they are held together by noncovalent interactions.The lipids comprising the membrane will spontaneously organize so as tominimize the number of exposed hydrocarbon chains. Van der Waalsattractive forces between hydrocarbon chains are the major factorsfavouring this organization. There are two significant attributes that arisefrom this tendency: lipid bilayers tend to form compartments and lipidbilayers tend to be self-sealing (to leave no hydrocarbon chains exposed).An important feature of biological membranes is that they areasymmetric with respect to structure and function. Membrane proteinsCvtoplasm Integralprotein Peripheral proteinsExteriorPhospholipidPolarheadFatty acyltailsOligosaccharideIntegralproteinGlycoprotein^GlycolipidPeripheralproteinsPhospholipidbilayer3Figure 1.1. Structural model of a biological membrane (taken from Darnellet al., 1986).4maintain an absolute asymmetry (due to the thermodynamic barriers ofprotein 'flipping), while membrane lipids exhibit somewhat less absoluteasymmetry (except for glycolipids). Lipid asymmetry has been demonstratedin human erythrocytes, bacterial outer membranes and many otherbiological membranes (Op den Kamp, 1979; Houslay and Stanley, 1982;Gennis, 1989). Choline containing lipids are predominantly located on theouter monolayer of plasma membranes such as the erythrocyte membrane,while aminophospholipids are predominantly on the inner monolayer (Opden Kamp, 1979). Lipid asymmetry has been found to be important inprocesses such as lipid exchange and membrane fusion (Eastman et al.,1992). The absolute asymmetry of membrane proteins allows for thevectorial transport of solutes (Houslay and Stanley, 1982).Transmembrane electrical potentials (4) can be generated acrossbiological membranes by the vectorial transport of an ion, which is notcompensated by the movement of a counter-ion. This process is generallylinked to the consumption of energy in living systems by processes such asATP hydrolysis, substrate oxidation and photosynthesis. Transmembraneion gradients, such as pH gradients, can be created in a similar manner.Transmembrane pH gradients have been established as being present invarious subcellular organelles, such as the endoplasmic reticulum (insideacidic; Rees-Jones and Al-Awqati, 1984; Thevenod et al., 1989), chloroplasts(inside acidic; Rottenburg and Grunwald, 1972), lysosomes, endosomes(inside acidic; Alberts et al., 1989) as well as across procaryotic membranes(outside acidic; Padan et al., 1976; Driessen, 1992).51.3 The structure of lipids There exist four main classes of membrane lipids: phospholipids,glycolipids, sphingolipids and the sterols. The phospholipids are the mostabundant in eukaryotic membranes. Figure 1.2 and Table 1.1 give thenames and structures of the common phospholipids. All these molecules areamphiphilic in nature due to the presence of both a hydrophilic head groupand a hydrophobic tail. The "hydrophobicity" of the molecule is determinedby the length of the acyl chain.The glycolipids are sugar-containing lipids, such as cerebrosides.Most glycolipids are sphingolipids and have sphingosine, an amino alcohol,as their backbone. However, sphingomyelin, a common mammaliansphingolipid, is not a glycolipid.Another class of membrane lipids is the sterols, of which cholesterol isthe most abundant in mammals. Cholesterol has a large planar, apolarregion and a small polar region consisting of a hydroxyl group. The nonpolarregion is associated with the nonpolar hydrocarbons of the bilayer, while thehydroxyl function is associated with the ester carbonyls of the lipids (at themembrane-water interface; Worcester and Franks, 1976; Gennis, 1989).1.4 Properties of lipids Lipid polymorphism refers to the many possible structures that canbe adopted by lipids upon hydration. These structures include micelles,bilayers and hexagonal phases (Cullis and deKruijff, 1979; Fig. 1.3). Thelipids in biological membranes generally adopt bilayer configurations. Non-bilayer structures are generally thought to occur as transitory intermediatesin the membrane fusion process (Cullis and deKruijff, 1979) and thepresence of non-bilayer lipids in a bilayer environment has been proposed toTable 1.1. Names and structures of some common fatty acids.6# of C Atoms^Structural FormulaSaturated fatty acids 14^CH3(CH2) 12CO2H16^CH3(CH2)1 4CO2H18^CH3(CH2)16CO2H20^CH3(CH2)1 8CO2HNamemyristic acidpalmitic acidstearic acidarachidic acidUnsaturated fatty acids 16 CH3(CH2)5CH=CH(CH2)7CO2H18 CH3(CH2)7CH=CH(CH2)7CO2H18 CH3(CH2)4(CH=CHCH2)2(CH2)6CO2H18 CH3CH2(CH=CHCH2)3(CH2)6CO2H20 CH3(CH2)4(CH=CHCH2)4(CH2)2CO2Hpalmitoleic acidoleic acidlinoleic acidlinolenic acidarachidonic acid—CH2CH2tIH3aNH3—CR, —CRCOO ——CH2CH(OH)CH2OH— CH2HC OH-CH2CH3CH3CH,HO7Figure 1.2. The structure of a phospholipid and commonly occurring headgroups and the structure of cholesterol (taken from Cullis and Hope, 1985).PhosOnatolcrlooneCH3H—C—CH2—CH2—CH2—C—CH3CCH3^ H3'EInanolamine(Phosonatlaylethanoianne)SerInePhosonalitlylserine)Glycerol(Pt,osonallaylglycerollGIycerotConosohalidylolycerol or carceoliolniMyo.lnositol(PhosphatrOyltnositol)Cholesterol8Figure 1.3. Diagrams of the different phases that can be adopted by lipids(taken from Cullis and Hope. 1985).LIPIDLYSOPHOSPHOLINDSDETERGENTSPHASEw-gt-77i.:MOLECULARSHAPE-- v --MICELLAR INVERTED CONE..,PHOSPHATIDYLCHOLINESPHINGOMYELINPHOSPHATIDYLSERINE !!!!!!!!PHOPHATIDYLINOS1TOLPHOSPHATIDYLGLYCEROL 1?)PHOSPHATIDIC ACIDCARDIOLIPINDIGALACTOSYLDIGLYCER IDEBILAYER CYLINDRICAL, ,..........s,' .,PHOSPHATIDYLETHANOLAMINE(UNSATURATED)CARDIOLIPIN — Ca 2+,‘,,,/.,, ;PHOSPHATIDIC ACID— Ca 2+ I(pH<6.0)PHOSPHATIDIC^ACID:41;11W4ij^,-■„ah, / ----^----(pH<3.0)PH6SPHATIDYLSERINE(pH <4.0)MONOGALACTOSYLDIGLYCER IDEHEXAGONAL (H11) CONE,^ _ _9modulate surface properties (Epand, 1990). This thesis concerns the barrierproperties of bilayer systems and therefore the properties of lipid bilayerswill be discussed in more detail.Lipid molecules in a bilayer can exist in gel (ordered and rigid) orliquid-crystalline (disordered and fluid) states. The key difference betweenthese two states is that the gel state is a crystalline state where all the C-Cbonds are in the trans conformation, while the liquid-crystalline state is afluid state where some are in the gauche conformation. Lateral diffusionoccurs slowly, if at all, in the gel state but is rapid in the liquid-crystallinestate. The transition between these two states occurs at a temperaturedefined as Tc. the melting or crystallization temperature. Decreasing thelength of the fatty acyl chains and increasing their degree of unsaturationlowers this transition temperature. The nature of the polar head group alsohas a significant effect on Te. An example of this is thatphosphatidylethanolamine (PE)-based lipids exhibit Te values 20 °C higherthan that of the corresponding phosphatidylcholines (Cullis and Hope,1985). Cholesterol also has a great influence on the phase adopted by lipids.At concentrations below 30 mol% it broadens the gel-liquid crystallinephase transition and progressively lowers the enthalpy of the transition. Athigher concentrations the transition is no longer observable (Chapman,1975).1.5 Transport across Biological Membranes Lipid bilayers are generally impermeable to most polar or chargedmolecules. However, cells must be able to transport nutrients inside andexcrete waste products, as well as regulate intracellular ion concentrations.Specialized transmembrane proteins (membrane transport proteins) are10responsible for the transport of specific molecules or families of molecules(including ions, sugars and amino acids). There exist two broad types oftransport processes: passive and active.Passive transport refers the movement of molecules acrossmembranes primarily in response to concentration gradients. There existthree possible mechanisms by which such transport can take place: simplediffusion; channel-mediated diffusion and carrier-mediated diffusion (seeFigure 1.4). Carrier proteins act by binding a specific solute and thenundergoing a conformational change in order to transfer the solute acrossthe membrane. Channel proteins function by creating water-filled channelsthat traverse the membrane. These channels can allow ions of theappropriate size to pass through the bilayer when they are open (Alberts etal., 1989). Active transport involves the use of carrier proteins which act tomove certain molecules against their electrochemical gradient via the inputof energy. The type of transport process that is examined in this workinvolves simple diffusion across model membrane (protein-free) lipid bilayersystems.1.6 Preparation of Model Membrane Systems The term "liposome" is a generic term referring to an aqueousdispersion of bilayer forming lipids. These "model membrane systems" maybe large (> 1 vm diameter) or small (< 50 nm diameter) and may consist ofmany concentric bilayers or a single lipid bilayer. Liposomal modelmembrane systems can be prepared using a variety of techniques (forexamples see Hope et al., 1986). Various model membrane systems havediffering sizes and degrees of lamellarity resulting in particular permeabilityproperties. A short evaluation of each of the commonly used types ofFigure 1.4. Diagram of the various forms of transmembrane transportutilized by cells (taken from Alberts et al. 1989).1 112liposomes will be presented, with comments concerning their utility formeasuring flux across membranes (See Fig. 1.5 for representations of eachof the common classes of model membrane systems).1.6.1 Multilamellar vesiclesMultilamellar vesicles (MLVs) can be created by simply dispersinglipid in aqueous buffer. These vesicles typically have diameters in excess of1000 nm, are heterogeneous in size and exhibit unequal solute distributionsacross the lamellae. Repeated freezing and thawing of an MLV preparation(termed Frozen and Thawed MLVs, or FATMLVs) results in an equilibriumdistribution of entrapped solute (Mayer et al., 1985). The advantages ofusing MLVs are that they are simple to create, relatively large in size andcan be formed from any bilayer forming lipid. The major disadvantages arethat they are multilamellar and heterogeneous in size. This multilamellaritymakes examining transport processes using MLV systems difficult, sinceonly a small amount (as little as 10%) of the lipid is present in the outermonolayer and diffusion must therefore occur across several lamellae (Hopeet al., 1986). Several techniques have been developed to make unilamellarmembrane systems to overcome the disadvantages of MLVs.1.6.2 Small Unilamellar VesiclesSmall unilamellar vesicles (termed SUVs) can be formed by thesonication of MLVs. The size of the resulting vesicles (25-50 nm diameter) isdependent on the lipid composition. The French press technique can also beused to create SUVs (Barenholz et al., 1979). The principal advantages ofSUV systems are that they are easy to prepare and are relativelyhomogeneous in size. The main disadvantages of SUVs are that the vesiclesFigure 1.5. Structural representations of the common classes of modelmembrane systems (taken from Culls et al., 1989).1313UV'sLUV'sMLV's0.2 - 0.5 pm0.05 — 1 prnDiametetr!^M14are unstable (tend to fuse), have small trapped volumes (< 0.2 [11/molephospholipid) and there exists the potential to generate lipid oxidationproducts during preparation. In addition, the permeability coefficients forion transport across SUVs are typically 1-2 orders of magnitude smallerthan for the corresponding large unilamellar systems (Deamer andBramhall, 1986; Perkins and Cafiso, 1986). The explanation for thisanomalous observation is unknown. Several procedures have beendeveloped to overcome the disadvantages associated with SUVs by creationof unilamellar vesicles of larger sizes.1.6.3 Large Unilamellar VesiclesLarge unilamellar vesicles (LUVs) can be created by the followingtechniques: ethanol injection (Kremer et al., 1977); ether infusion (Deamerand Bangham, 1976); reverse phase evaporation (Szoka andPapahadjopoulos, 1978); detergent dialysis (Mimms et al., 1981) andmedium pressure extrusion (Hope et al., 1985). The first three techniquesinvolve dispersing the lipids in an organic solvent (ethanol or ether, forexample). The lipid is then hydrated by injection of the lipid-organic solventmixture into buffer. The organic solvent can then be removed by severaldifferent techniques, such as dialysis. Column chromatography is oftenemployed as an additional step to remove any residual organic solvents.Detergent dialysis techniques involve the detergent-induced solubilization oflipid into micelles. The detergent is then removed by dialysis. LUVs preparedusing these techniques all have average diameters of between 50-200 nmand trapped volumes of 1-3111/ptmole phospholipid (Szoka andPapahadjopoulos, 1980). The major disadvantages of these techniques arethat they are tedious, different lipid species may have different solubilities in15solvents used and the resulting vesicles may contain residual organicsolvents or detergents capable of altering the permeability properties of theresulting vesicles.The best technique for making large unilamellar vesicles involvesrepeatedly passing MLVs through filters of decreasing defined pore sizeunder medium pressure. The resulting vesicles are called LUVETs (LargeUnilamellar Vesicles by Extrusion Techniques). This technique wasdeveloped by Cullis and co-workers following the initial observation byOlson et al. (1979) that greater size homogeneity could be obtained forreverse phase vesicles following low pressure extrusion throughpolycarbonate filters of a defined pore size. The resulting 100 nm LUVETs(following passage through 100 nm pore size filters) have been establishedas being unilamellar by freeze-fracture electron microscopy, NMR and quasielastic light scattering techniques (Hope et al., 1985; for a more extensivediscussion of extrusion techniques see Hope et al., 1986). The majoradvantages of this technique include its speed, the avoidance ofsolvents/detergents, the high trapping efficiencies obtained and thehomogeneous nature of the resulting vesicle populations. The onlysignificant disadvantage of this technique is that homogeneous unilamellarvesicles of sizes larger than 200 nm cannot be created. Larger vesicles canbe produced but as filter pore size increases, an increasing proportion of thevesicles will be multilamellar and the vesicle population will become moreheterogeneous. Extrusion techniques (using 100 or 200 nm pore size filters)were used to produce the vesicles used for all experiments presented in thisthesis.161.6.4. MonolayersPhospholipids will form an oriented monolayer at an air-waterinterface such that polar parts are in contact with the water and thehydrocarbon tails are extended into the air. These systems have been usedto study lipid packing, protein-lipid interactions and enzymology at lipid-water interfaces (Gennis, 1989). However, these systems are of limited valuefor studying transmembrane transport processes because transport acrossmembranes between aqueous solutions cannot be achieved.1.6.5. Planar Lipid MembranesPlanar lipid membranes are created by applying a mixture of lipid inorganic solvent to an aperture separating two aqueous compartments(Fettiplace et al., 1974). The solvent collects at the perimeter of the aperture,creating a lipid bilayer across the aperture. The major advantages of planarmembranes are that electrodes can easily be introduced on each side of themembrane, solutions can be changed quickly and thus sensitive, accuratemeasurements of ion gradients can be achieved. These bilayers havetherefore proven useful for studying pores, channels and transporters. Themajor disadvantages of this technique include the small amount of lipidpresent, permeability changes caused by residual organic solvent andpossible artifacts created at the lipid-aperture interface (Gennis, 1989).1.7. Uses of Liposomal Model Membrane Systems Model membranes are useful for determining information about thestructure and function of complex biological membranes because theindividual components of the membranes and various associated proteinscan be studied in isolation. The internal and external environments of the17model membranes can also be altered with ease in order to determine theeffect of specific factors on membrane properties.Pure lipid systems of defined composition have proven useful forpermeability experiments (for an excellent review see Deamer and Bramhall,1986) as outlined in Chapter 4. In addition, pure lipid systems can also beemployed for the measurement of lateral diffusion of lipids in membranes(Vaz et al., 1982) and for studies of membrane fusion (Eastman et al., 1992).Reconstitution of integral membrane proteins into liposomes isparticularly important in terms of understanding membrane proteinfunction and building model systems of biological membranes. Manydifferent membrane proteins have been reconstituted into liposomal systems(see, for example, Racker, 1973 or Racker, 1985). The potential informationthat can be acquired using this technique includes: identification of reactionmechanisms of membrane enzymes; determining the structural propertiesof membrane proteins; and evaluating interactions between differentmembrane components (Madden, 1986).Liposomes are also useful as drug delivery vehicles (see Mayer et al.,1986a; Mayer et al., 1988; Cullis et al., 1989; Madden et al., 1990; Culls etal., 1991). The primary advantages of using liposomes for drug delivery arethe reduced toxicity observed when compared to the free drug (particularlyfor the anti-cancer drugs), as well as the increased efficacy observed againstcertain diseases and tumours (Ostro and Cullis, 1989). Drugs can bepassively trapped within liposomes simply by preparing the vesicles in asolution containing the compound of interest. The major problems with thisprocedure are the low efficiencies of entrapment and the potential forleakage during storage.18Drugs which are lipophilic weak bases can be accumulated into LUVsin response to transmembrane pH gradients. This uptake process occurs asoutlined for weak bases in Section 1.7 (Nichols and Deamer, 1976; Mayer etal., 1986a; Mayer et al., 1986c; Chakrabarti et al., 1992). This procedurecan be used for a variety of compounds which are lipophilic weak acids orbases. Two of the liposomal drugs prepared via this technique which arecurrently in preliminary study or clinical trials are vincristine anddoxorubicin (Mayer et al., 1989). The main advantages of this techniqueinclude retarding drug leakage, the high drug:lipid ratios possible and thehighly efficient trapping, which can approach 100%.1.8 Transmembrane Electrical Potentials and pH Gradients Transmembrane electrical potentials (Alp) are created by the movementof a charged ion or electron across a bilayer which is not compensated forby the movement of a counter-ion. Alp is generally measured by determiningthe ratio of the inside and outside concentrations of a permeable ion atequilibrium (Hope et al., 1985; Cafiso and Hubbell, 1978; Rottenberg,1979). A transmembrane electrical potential would be expected to drive themovement of a permeable ion so that the electrochemical potential for theion would be zero at equilibrium. Fig. 1.6 illustrates the generation of a Alpin an LUV containing a K+ buffer placed in an external medium containing aNa+ buffer. Valinomycin is an ionophore which preferentially transports K+out of the LUV interior. This uncompensated charge transfer leads to abuild-up of an electrical potential, hindering further net outward movementof K+. As H+ ions are usually the most membrane-permeable cation species,they will redistribute in response to a transmembrane electrical potentialuntil such time as the19Figure 1.6. The relationship between Alp and epH at electrochemicalequilibrium. Addition of valinomycin to vesicles containing K+ results in 1{±efflux. creating an electrical potential across the membrane. H+ ions willredistribute in response to this electrical potential until the chemicalpotential of the protons equals that of the transmembrane electricalpotential. resulting in the formation of a transmembrane pH gradient.+ValinomycinA= -59 logf[K+]in gielout IA p Heq= logl[e]in /[K+] out I20transmembrane electrical potential equals the chemical potential of theprotons. A transmembrane electrical potential will therefore drive theformation of a transmembrane pH gradient and vice versa (Fig. 1.6).Membrane potentials can be calculated as follows from the Nernst equation:= (RT/nF) In (Pion[]; +^)^(1)(PionRD + ...)This in turn can be simplified to:Aip = (RT/nF) ln (Ci/Co)^ (2)where Aip is the membrane potential (V), C is the concentration of the mostmembrane-permeable ion (K+, in this case), the subscripts i and 0 refer tothe inside and outside of the vesicle, respectively, R is the gas constant (calmo1-1 K-1, T is the temperature (K), F is the Faraday constant (2.3 X 104 calV-1 mol-1), n is the charge of the ion and Pim„ is the permeability coefficientfor the ion.Many biological compounds have proton donating (acidic) or protonaccepting (basic) groups. The dissociation constant (Ka) for a weak base canbe described as follows:Ka = [HI [131/[BH+1^ (3)where [W] is the hydrogen ion concentration, [13] is the concentration of theneutral form of the weak base and [BH+lis the concentration of theproton ated base. The pKa is the negative log of the base dissociationconstant. The Henderson-Hasselbach equation relates the pH of themedium to the relative concentrations of the ionized and protonated weakbases as follows:pH = pKa + logal31/[BH-FJ)^ (4)Analogous equations can be utilized to describe the behavior of weak acids.21The distribution of weak acids and bases can be used to measuretransmembrane pH gradients, and, conversely, pH gradients can be used todrive the net accumulation of weak bases into liposomal systems. Assumingthat the dissociation constant (Ka) remains the same for a weak base onboth sides of the membrane, it follows that:Ka = [H+]1[131i/[BH+]1 = [H÷J01B)0/[BH+]0^(5)where the subscripts i and 0 refer to inside and out. It is generally acceptedthat the neutral form of weak acids and bases is by far the most membrane-permeable species (Addanki et al., 1968; Rottenberg, 1972). Thus, atequilibrium, the concentration of the neutral species will be the same onboth sides of the membrane. The resulting distribution of a weak base(assuming that pKa»pH0»pHi) will reflect the pH gradient as follows (seeFig. 1.7):[BH+J1/O3H+10 = [H]/[H10^ (6)Thus, measurement of the concentration of the exterior and interiorconcentrations of weak base can give a measure of the proton gradient.Equally important, from the point of view of this thesis, is the correspondingfact that the presence of a pH gradient will drive the net uptake of weakbases. A pH gradient of 3 units (acidic interior), for example, could beexpected to result in a 1000-fold higher concentration of weak base insidethe LUV than outside, at equilibrium.1.9.1. Transport of nonionic solutesThe net flux of a non-ionic solute across a membrane can bedescribed by:J = P Am (Ci - Co)^(7)where J is the flux in moles cm2 sec-1, P is the permeability coefficient in cm22Figure 1.7. Redistribution of weak bases in response to transmembrane pHgradients; where D represents the drug/weak base of interest. InOut< I1+ +DH DH> vlDD++H++HAt equilibrium, if [D]i^[D]o[DH].^[W].[DH+],23sec-1, Am is the membrane area in cm2, C is the solute concentration andthe subscripts i and 0 refer to inside and out. The driving force for the netflux of a non-ionic solute across a membrane is the chemical potential of thesolute.Permeability measurements are typically made by monitoring the netrelease of solute from vesicles following passive entrapment. It can be shownthat the net diffusion of a compound out of an LUV with constant volumewill conform to the following equations based on Fick's law (assuming a firstorder process; Cullis and Hope, 1985):dN• = - P Am (Ci - C0)^ (8)dtdC• = - P A Ci (9)dt^Viwhere Ni is the number of molecules inside the vesicle, Vi is the internalaqueous volume and assuming that C1>> C0. A solution to this equation isgiven byC(t) = Ci(0) e-kt^(10)where k (P Am)/Vi. This can be rewritten asIn {C1(t)/C1(0)} = -ktThe permeability coefficient for the efflux of passively entrapped solute canbe calculated from the relation:P = kVi/Am^ (12)where k is the rate constant. This rate constant can be derived from theslope of a plot of In I[C(eq) - C(t)]/C(eq)} versus time (where C(eq) is thesolute concentration at equilibrium and C(t) is the solute concentration attime t; see equation 11).Two models have been developed to account for the transport ofsolutes across model membranes (for a review see Deamer and Bramhall,241986). The solubility-diffusion model states that the solute dissolves into thebilayer and then subsequently diffuses across (Finkelstein and Cass, 1968).The transient defect model stipulates that transient defects or pores in themembrane are responsible for solute transport (Carruthers and Melchior,1983). There is evidence to support both these models and it has beensuggested as a result that both may play a role in membrane permeability(Deamer and Bramhall, 1986).The rate limiting step for transbilayer transport of nonionic solutes isthought to be the partitioning of the molecule into the membrane. Partitioncoefficients (Kr). which are a measure of the partitioning of small moleculesbetween the aqueous phase and membrane, can be defined as follows:Kp = Cb/Cf^ (13)where Cb and Cf are the concentrations of ligand bound to the membraneand free in solution (Germ's. 1989).There exists a strong correlation between the permeability coefficientof a molecule and its hydrophobicity (as defined by its octanol:waterpartition coefficient, for example). This relationship is known as Overton'srule (Walter and Gutnecht, 1986; Gennis, 1989). Other factors, such asincreasing molecular size and hydrogen bonding capacity, have also beenfound to impede transbilayer movement (McElhaney, 1986). Factors whichincrease the order (degree of motion of the phospholipids present) of thehydrocarbon chains will generally decrease the permeability coefficients fortransport of nonionic solutes.1.9.2. Transport of ionsTypical permeability coefficients for the transbilayer movement ofsmall ions range from 10-10 to 10-14 cm s" I (see Fig. 1.8). The energeticFigure 1.8. Permeability coefficients of lipid bilayers to various types ofmolecules (taken from Alberts et al.. 1989).251 0-210-6H204--- IndoleGlycerol, Urea4-- Tryptophan4-- Glucose10-1226barrier to ions is therefore much greater than that for nonionic solutes. Ingeneral, the net flux of ions across protein-free membranes is quite low.As discussed in Section 1.8, net flux of an ion out of an LUV inresponse to a chemical potential will result in the creation of an electricalpotential (Alp), in the absence of any compensating charge movement. Thiselectrical potential will create reduced net flux of the ion out of the LUV. Thenet flux of the ion would be expected to be zero at equilibrium, where thetransmembrane chemical potential for the ion is equal to and opposite tothat of the established electrical potential. Permeability measurements forions must therefore take into account this inherent difficulty withmeasuring electrolyte transport (see Fig. 1.6).The Born energy barrier is the major barrier to the movement ofelectrolytes through membranes (Parsegian, 1969). This barrier isapproximately 40 kcal mol-1 for a monovalent ion (diameter of 0.2 nm)moving from a medium with a high dielectric constant (80; buffered aqueoussolution) to a medium with a low dielectric constant (2; interior of amembrane). Increasing the radius of the ion or increasing the dielectricconstant of the bilayer decreases the energy barrier to solute transport,which results in increased net flux of the ion (Dilger and McLaughlin, 1979).Flewelling and Hubbell (1986a; 1986b) have suggested that transbilayertransport of electrolytes may be influenced by other factors in addition tothe Born energy, such as the image energy, the dipole energy and a neutralenergy term (encompassing all the nonelectrical interactions that anelectrolyte has within a membrane).The actual mechanism of electrolyte transport is not well understood,but the solubility-diffusion and transient defect models have both beenimplicated as being potentially responsible for electrolyte transport (Deamer27and Bramhall, 1986). At present, further investigation is required todetermine which model is more suitable.1.9.3. Lipid factors which influence ion transportFactors which increase hydrocarbon chain order (such as lowtemperature, increased saturation and higher cholesterol content) usuallydecrease the transbilayer transport of nonionic solutes and electrolytes.However, lipid bilayers have been shown to be particularly permeable attheir phase transition temperature (Deamer and Bramhall, 1986). It hasbeen suggested that regions of gel phase and liquid-crystalline lipid co-existat or close to the Te of a pure lipid. This coexistence could lead to ahydrated defect being formed.Head group structure can influence liposome permeability. Acidiclipids (such as phosphatidic acid, PA, phosphatidylglycerol, PG, andphosphatidylserine, PS) impart a negative surface charge to the membrane.The presence of such a negative surface potential (negatively charged groupson the membrane surface) would increase the localized concentrations ofcations and, conversely, decrease the concentration of anions due to chargerepulsion near the membrane. This could serve to increase the flux ofcations and decrease the flux of anions across such membranes. In the caseof weak bases, the negative surface potential would be expected to increasethe local concentration of the positively charged species. However, thenegative surface charge would also be expected to increase the localized H+concentrations (decreasing the pH near the membrane). This, in turn, couldserve to reduce the flux of weak bases by decreasing the proportion presentin the neutral form.281.9.4. Transport using ionophoresIonophores are molecules which facilitate the transport of ions acrossmembranes and are therefore examples of carrier-mediated transport(Pressman, 1973; Gennis, 1989). Ionophores generally possess antibioticcapabilities. Antibiotic ionophores usually function by creating a cyclicstructure which has a central hydrophilic cavity and a very hydrophobicouter surface. The ion is held in the cavity while the ionophore:ion complexdiffuses across the membrane and is released when the complex reaches theother side due to concentration differences and/or the presence of chelatingagents (Pressman, 1973; Gennis, 1989).Ionophores have been utilized to entrap metal ions of interest, such asindium, calcium and gadolinium, within lipid vesicles (Mauk and Gamble,1979; Veiro and Cullis, 1990; Tilcock et al., 1990). Earlier work (Mauk andGamble, 1979) relied upon the presence of a passively entrapped chelator toobtain metal ion accumulation. Uptake was thus limited by the levels ofchelator that could be passively entrapped within the liposome.Developments upon this ionophore-mediated entrapment technique outlinedhere (see Veiro and Culls, 1990 and Chapter 2) offer a significant advantageover passive entrapment procedures because of the high encapsulationefficiencies possible. One goal of such entrapment is to maximally loadmetal ions into the liposomes, which can result in the creation of electrondense liposomes to aid in their visualization using electron microscopy orliposomes which are dense enough to allow for their separation to beachieved through gravimetric techniques.A23187 is a carboxylic ionophore which is capable of binding divalentcations (for an early review see Pressman, 1973; see also Gennis, 1989). Ithas previously been shown that A23187 can transport Ca+2 into LUV29systems (Veiro and Cullis, 1990). The transport is believed to occur via theformation of a neutral 2:1 A23187:Ca+2 complex (see Fig. 1.9; Gennis,1989).Valinomycin is a cyclic antibiotic ionophore which forms 1:1complexes with monovalent cations, particularly K+. The complex acts as ahydrophobic ion by having the IC+ located at the center of the complex.Valinomycin results in the formation of K+ diffusion potentials (membranepotentials; 6,11)) in vesicles where the initial internal and external K+concentrations are different (See Fig. 1.5; Pressman, 1973; Gennis, 1989).1.9.5. Transport of signal sequences and other peptidesPeptide-membrane interactions are essential to many cellular events,such as toxin entry and the transport of newly synthesized proteins (Parkeret al., 1990; Gierasch, 1989). Transmembrane pH gradients and other iongradients have been shown to play a significant role in several of theseprocesses (Skerjanc, 1990). However, the exact mechanism by which theseion gradients influence these processes is still largely obscure (Gierasch,1989; Rapoport, 1990).In 1975, the signal hypothesis was first proposed by Blobel andDobberstein (1975). The basic premise of the signal hypothesis is that thesignal for the translocation of nascently synthesized polypeptides across theendoplasmic reticulum membrane is a "signal sequence" near the amino-terminus (N-terminus) of the polypeptide. The signal hypothesis and signalsequences have been an area of intense investigation since that time (forrecent reviews see Verner and Schatz, 1988; Gierasch, 1989; Rapoport,1990; Skerjanc. 1990). A diagrammatic representation of some of the stepsH_0^°Figure 1.9. Structure of A23187 (modified from Ma et al., 1990).Free Acid30A23187 complex with Metal Ion (Me+2)31involved in protein trans location for two types of peptides across themembrane of the endoplasmic reticulum is shown in Fig. 1.10.Subsequent research has shown that signal sequences are short (13-36 residue) peptides attached to the N-terminal (amino terminal) of allsecreted endoplasmic reticulum and lysosomal proteins, as well as manytransmembrane proteins. Signal sequences consist of a 7-13 residuehydrophobic core flanked by several relatively hydrophilic residues, usuallyincluding one or more basic residues near the N-terminus. This centralnonpolar core is essential for signal sequence translocation. If thehydrophobicity of this region is decreased (for example by the addition of acharged residue) protein transport is inhibited. Other than obeying thesegeneral criteria there is much variety in the actual amino acid compositionof signal sequences (Reithmeier, 1985; Gierasch, 1989).The exact mechanism by which signal sequences and thepolypeptides to which they are attached translocate across membranes isstill uncertain. A variety of accessory proteins have been established asbeing involved in peptide translocation across both eucaryotic (endoplasmicreticulum) and procaryotic membranes (Stryer, 1988; Rapoport, 1990).However, it has also been established that translocation can occur in theabsence of membrane proteins (Rapoport. 1990). In addition, signal peptidescan spontaneously insert into monolayers and the degree of insertion intolipid bilayers corresponds to their ability to translocate (Tamm, 1986;Rapoport, 1990). Another hypothesis states that signal sequences may be"pulled" across the bilayer by the presence of electrochemical potential. Apreliminary binding step would be followed by translocation in response tothe electrochemical potential (inside negative; Skerjanc, 1990). It is ofinterest to note that transport of proteins across mitochondial membraneslipidbilayerspecifically boundsignal peptide(start-transfer peptide)• •LUMENY.`4-600HNH2mature solubleproteinTRANSLOCATIONcleavage site""- NH2 --'"^ NH2^ NH2I . ..LEAVAGEOFSTART-TRANSFERPEPTIDE6HOOCTRANSLOCATIONCOMPLETEDCYTOSOLN H2ribosomeCOOHstop-transferpeptide— ..--- CYTOSOLNH2^s- - -^,-., v,I 1----^---Ir=t1COOH"LUMENcleavage sitespecifically boundstart-transfer peptide(B)--).TRANSLOCATION^CLEAVAGESTOPS OF START-TRANSFERPEPTIDEmaturetransmembrane protein(A)-"INTlipidbilayerTRANSLOCATION32Figure 1.10. A diagrammatic representation of protein translocation acrossthe membrane of the endoplasmic reticulum for (A) soluble proteins and (B)transmembrane proteins (taken from Alberts et al., 1989).ribosome33occurs in the absence of accessory proteins provided that a membranepotential is maintained, despite having an alkaline interior (Skerjanc, 1990).Hexagonal lipid phases present in membranes could also be potentiallyinvolved in assisting protein translocation by providing a hydrophilicenvironment through which the polypeptide can pass, although this mustbe regarded as speculative (Rapoport, 1990).As indicated in Section 1.7, transmembrane pH gradients can causethe accumulation of weak bases into LUVs with an acidic interior. Thisapplies to a large variety of lipophilic weak bases (Madden et al., 1990). Inthis regard the fact that signal sequences exhibit similar weak basecharacteristics is of interest. The study of the translocation of amino acidsand peptides, modified to become exclusively weak bases, into LUV systemsconstitutes the bulk of this thesis. The specific questions that wereaddressed include: 1) what is the mechanism by which these peptides aretransported in response to transmembrane pH gradients; 2) how does thedegree of hydrophobicity and charge affect such transport; and 3) do aminoacids, modified amino acids and model peptides respond differently to pHgradients in model membrane systems?1.10 SummaryThis section has attempted to summarize key features responsible forthe general permeability and transport characteristics of membranes. Modelmembrane systems are useful because the individual components ofmembranes can be studied in isolation and because of their ease ofmanipulation. LUVs produced by extrusion techniques are the modelmembrane system of choice. Solutes can translocate across membranes bydiffusing across the membrane or through transient defects in the34membrane. Ionophores facilitate the transport of ions across membranes byforming neutral complexes and, in conjunction with transmembrane pHgradients, can result in high interior ion concentrations being achieved.LUVs with an acidic interior can rapidly and efficiently accumulate lipophilicweak bases via translocation of the neutral (deprotonated) form of the weakbase. Signal sequences and some other peptides are lipophilic weak basesand may therefore be capable of transport into LUVs in response totransmembrane pH gradients.In Chapter 2 of this thesis the ability of transmembrane pH gradientsand A23187 to facilitate uptake of Ba+2 and Fe+2 into LUV systems ischaracterized. Chapter 3 investigates the uptake of basic amino acids andpeptides into liposomes in response to transmembrane pH gradients.Chapter 4 examines the permeability of lipid bilayers to various unmodifiedamino acids. Finally. Chapter 5 investigates the influence of charge, chargedistribution and hydrophobicity on the transport of small model peptidesinto liposomes in response to transmembrane pH gradients.35CHAPTER 2 - GENERATION AND CHARACTERIZATION OF IRON ANDBARIUM LOADED LIPOSOMES 2.1 INTRODUCTION Previous work from this laboratory has shown that rapid and efficientaccumulation of Ca+2 into large unilamellar vesicles (LUVs) in response totransmembrane pH gradients (inside acidic) can be achieved in the presenceof the Ca+2-ionophore A23187 (Veiro and Culls, 1990). The uptake of Ca+2was shown to be dependent upon the buffering capacity of the internalcitrate buffer, indicating a Ca+2-2H+ exchange process mediated by thepresence of the ionophore (Veiro and Cullis, 1990). Among otherapplications this procedure offers interesting possibilities for increasing thedensity of LUVs by exchanging protons for cations of much higher atomic ormolecular weight. In the present work we have investigated the potential ofthe A23187-mediated ApH loading technique to accumulate iron and bariuminto LUV systems. It is shown that rapid and efficient loading of Fe+2 andBa+2 can be achieved by this method. The resulting systems displayproperties consistent with increased densities, as evidenced by theirpelleting properties after low speed centrifugation and visualization byelectron microscopy. These systems may be of utility in proceduresrequiring separation of LUVs from aqueous media and visualization in vitroor in vivo.362.2 MATERIALS AND METHODS Egg phosphatidylcholine (EPC) and distearoylphosphatidylcholine(DSPC) were obtained from Avanti Polar Lipids (Birmingham, Alabama).14C-methylamine was purchased from New England Nuclear. The ionophoreA23187 was obtained from Calbiochem (Calgary). All other chemicals usedwere purchased from Sigma Chemical Co (St. Louis).To produce EPC vesicles, multilamellar vesicles (MLVs) were firstproduced by hydrating 50 mg EPC (100 mole %) in 1.0 ml citrate buffer (pH4.0). The MLVs were frozen in liquid nitrogen and thawed in warm water forfive freeze-thaw cycles. This treatment increased the trapped volume of thevesicles and promoted equilibrium solute distribution. Extrusion of thefrozen and thawed MLVs (FATMLVs) through two stacked polycarbonatefilters (Nuclepore) (100 nm or 200 nm pore size) was done ten times usingan Extruder obtained from Lipex Biomembranes Inc. (Vancouver, Canada).In accordance with previous results (Mayer et al., 1986) the resulting largeunilamellar vesicles (LUVs) were 108 nm (100 nm pore size) or 214 (200 nmpore size) in diameter, as calculated by quasielastic light scatteringtechniques employing a NICOMP particle sizer.DSPC:cholesterol (55:45; mol:mol) vesicles were made by dissolvingappropriate amounts of both compounds in chloroform. The chloroform wasthen removed under a stream of nitrogen and by subsequent incubationunder reduced pressure. After hydration to form MLVs, theDSPC:cholesterol mixture was extruded at 65 °C.In order to establish a transmembrane pH gradient, the LUVs in pH4.0 buffer were passed down a 10 cm Sephadex G-50 (50-150) columnpreviously equilibrated with 150 mM NaC1, 20 mM Hepes (pH 7.5) (Hepesbuffered saline. HBS). The A23187 was dispensed from a stock chloroform37solution of 2.51.tg/m1 in chloroform:ethanol (4:1). The chloroform wassubsequently removed under a stream of nitrogen and 1.1 ml of buffer (pH7.5) containing 150 mM NaC1, 20 mM Hepes (and 10 mM ascorbate for theiron experiments) was added. To this, 0.150 ml of either 10 mM FeSO4 or 10mM BaSO4 was added. Uptake was initiated by the addition of 0.250 ml ofLUVs prepared as indicated above (pH 7.5 outside, pH 4.0 inside) to thismixture. To assay uptake, 0.100 ml aliquots were removed at appropriatetimes by passage through 1.0 ml Sephadex G-50 columns (pre-spun). Thesecolumns were spun for one minute at 2500 r.p.m. in a clinical centrifuge toremove untrapped iron or barium. Use of the spin columns precluded theuse of sampling times shorter than 5 minutes. All experiments wereconducted at 23 °C unless stated otherwise. The final phospholipidconcentrations of the reactions mixtures were generally in the range of 3-5mM.In order to establish which oxidation state of iron (Fe+2 or Fe+3) wastransported, conditions appropriate to a particular oxidation state wereemployed. Fe+2 samples were maintained in the reduced state by theaddition of ascorbate (10 mM) to the buffers used and by gassing of thebuffers with N2 gas to minimize oxidation. Fe+3 samples were obtained fromFeC13.The magnitude of the pH gradients present was measured using 14C-methylamine at a concentration of 1 tiCi/ml. The amount of this probeimported into the LUVs was determined via liquid scintillation countingafter separating untrapped material employing a spin column procedure(Mayer et al., 1988). The transmembrane pH gradient was calculated usingthe relationship: ApH = log {[Methylamine]in/fMethylaminejout}, assuming a38trapped volume of 1.5 0/timol for the systems extruded through 100 nmpore size filters (Hope et al., 1985).Centrifugation (15,000 X g) of the iron- or barium-liposomes wasperformed at 4 °C using an Eppendorf micro-centrifuge. Barium content ofthe liposomes was determined by using Inductively Coupled Plasma (ICP)-Mass Spectroscopy (Elemental Research Inc., North Vancouver, Canada).Iron was assayed spectrophotometrically using bathophenanthrolinesulfonate (Pippard and Stray, 1982). In this assay 0.250 ml of 3.3 M sodiumacetate (pH 4.7), 0.100 ml of 5.0% Triton X-100, 0.050 ml of 0.2% (w/v)bathophenanthroline sulfonate in 1.0 % thioglycolic acid and 0.550 ml ofwater were added to 0.050 ml of the sample to be assayed (total volume =1.000 ml). The reaction was allowed to proceed for at least 15 minutesbefore the absorbance was measured at 535 nm using a Shimadzu UV-160spectrophotometer.Phospholipid concentrations were calculated by a modification of themethod of Fiske and Subbarow (Fiske and Subbarow, 1925). All data aremeans, n=3 + SEM (standard error of the mean). Where error bars are notshown on figures. the SEM values were within the dimensions of the symbolused.Electron dense liposomes containing Ba+2 were prepared using amodification of techniques used to load Fe+2/Ba+2 into liposomes. A driedDSPC:cholesterol (55:45; mol:mol) film was hydrated in 600 mM citratebuffer adjusted to pH 3.0 with arginine at a concentration of 25 mg/mlbuffer. LUVs were then prepared as described previously. A 500 [t1 of theresulting 100 nm vesicles was added to a Sephadex G-25 column and theexternal volume was exchanged by elution with 150 mM NaC1, 20 mMHepes (adjusted to pH 7.4) to produce vesicles with a transmembrane pH39gradient (pH 3.0 in; pH 7.4 out). BaC12 (24 ml of a 500 mM stock in water)was added to the ionophore (12 [tg of A23187) and vortexed. A 1.0 ml aliquotof vesicles was added to give a final barium concentration of 23 mM. Thesuspension was vortexed vigorously and placed at 60 °C for 2.5 hours. TheBa+2 loaded liposomes were stored at 4 °C until observed via cryo-electronmicroscopy.For cryo-electron microscopy, a thin film was formed by applying adrop of the vesicles suspension to a 700 mesh gold grid with a Pasteurpipet, followed by blotting from behind with Watman #50 filter paper. Thethin films were then vitrified by plunging the grids into liquid ethane kept atthe temperature of liquid nitrogen in a Reichart Jung Universal CryoFixation system. The grid was then removed from the ethane in a liquidnitrogen-chilled environment and blotted from the edge with filter paper toremove any excess ethane remaining on the grid. The grid was kept in liquidnitrogen within the fixation unit until the cryotransfer device was in place.The grid was immediately transferred to a specimen holder within theciyotransfer unit. At no time was the grid exposed to the temperature of theambient air and care was taken to dry and cool all tools before handling thegrid.The grid was transferred to a Zeiss 10C STEM electron microscopeequipped with a Gatan 126 cold stage. The stage and anticontaminator werekept at 120K and 110-119K, respectively, with liquid nitrogen. Regions ofthin vitreous ice were observed with an acceleration voltage of 60-85kV.Care was taken not to damage the sample with high beam intensities. Thevesicles were readily observed in the vitreous layer. The cryo-electronmicroscopy was performed by Dr. J.J. Wheeler.402.3 RESULTS The first set of experiments was designed to demonstrate the A23187-mediated uptake of iron into EPC LUVs in response to a pH gradient (insideacidic) and to determine the specificity for the Fe+2 or Fe+3 species. Asshown in Fig. 2.1, iron presented as FeSO4, and maintained in the ferrousstate by the presence of ascorbate (10 mM), was readily accumulated by theLUVs to achieve high interior concentrations of approximately 250nmolnlmol phospholipid under the experimental conditions employed.Assuming an interior trapped volume of 1.51A1/[tmole phospholipid, thiscorresponds to an interior concentration of 167 mM, as compared to the(initial) exterior Fe+2 concentration of 2 mM. As also shown in Fig. 2.1, Fe+3(presented as FeC13) was accumulated to only background levels indicatingthat A23187 was effectively unable to transport the ferric form.In order to determine the appropriate concentration of A23187 forFe+2 transport, a variety of A23187 concentrations were employed anduptake monitored as shown in Fig. 2.2. A convenient time course for uptakewas observed at 2.5 [tg/m1 A23187, corresponding to an initial rate of Fe+2transport of 30 nmol/min trnol phospholipid. This may be compared to arate of approximately 4 nmol Ca+2/min itmol phospholipid in the presenceof 0.1 tg/m1 A23187 observed under the same experimental conditions(Veiro and Cullis, 1990), and reflects the less efficient transport of Fe+2 bythe ionophore.As indicated in the Introduction one of the objectives of thisinvestigation was to achieve high encapsulation levels of iron, resulting inLUV systems which, due to their higher electron density could be readilypelleted by centrifugation, or which due to their higher electron densitycould be more readily visualized by electron microscopy techniques. It is41Figure 2.1. Time course of the accumulation of Fe+2 (0) and Fe+3 (0) at 23°C into EPC LUVs experiencing a transmembrane pH gradient, inside acidic(pH1=4.0; pH0=7.5) in the presence of A23187 (101Ag/m1). The initialexternal iron concentration was 2.0 mM. The EPC LUVs (3 mM) wereprepared in 300 mM citrate (pH 4.0) and the untrapped (exterior) buffer wasexchanged for 150 mM NaC1, 20 mM Hepes (pH 7.5). The ferrous (Fe+2)species was provided as FeSO4 and was maintained in the reduced state bythe presence of ascorbate (10 mM). Control vesicles exhibiting no pHgradient (pHi=pH0=7.5, A and pHi=pH0=4.0, 0 ) were also tested.4.^4^^ ED^C.!^0 0-0.5 1.0Time (Hours)1.5I.i)2.07 .300E.-6.230o.U30.0Q. 200cs-6E4...a100ti..03cri-6E 50c150-T01 r0300042Figure 2.2. Influence of the A23187 concentration on the accumulation ofFe+2 into EPC LUVs (100 nm diameter) experiencing a transmembrane pHgradient, inside acidic (pH1=4.0; pH0=7.5). The Fe÷2 was kept in thereduced state by the presence of ascorbate (10 mM). For other conditionssee Fig. 2.1 and Methods. The A23187 concentrations employed were 0.5la g/m1(0); 1.0 g/m1 (A); 2.5 tig/m1 (El); 5.0 ptg/m1(v); 10 IA g/m1 (*).10^20^30^40^50^60Time (minutes)43likely that Fe+2 uptake proceeds via a Fe+2-2H+ exchange, and thus higherinternal buffer concentrations should result in higher levels of accumulatedFe+2. Entrapment levels were therefore monitored for internal citrateconcentrations up to 1M for a variety of external Fe+2 concentrations. Asshown in Fig. 2.3A, the amount of accumulated Fe+2 increases as theconcentration of entrapped citrate increases (though only slightly beyond100 mM) until a plateau region is reached, which is dependent on the initialexterior Fe+2 concentrations. As indicated in Fig. 2.3B, the point at whichthe plateauing effect is observed coincides with trapping efficiencies of 80%or more, and thus reflects the lack of available exterior Fe+2 to import. Themaximum uptake levels achieved (250 nmolhimol phospholipid) were for the1M entrapped citrate and 10 mM (initial) exterior Fe+2 concentrations.As shown elsewhere (Mayer et al., 1986) the trapped volume of LUVsgenerated by extrusion increases significantly as the vesicle size increases.Increased uptake (on an iron/phospholipid basis) could therefore beexpected for larger systems. due to an increased buffering capacity. Thiseffect is shown in Table 2.1. The small increase in accumulated Fe+2 per[imol lipid observed for diameters above 200 nm is primarily due to theincreasingly multilamellar character of such liposomes (Mayer et al., 1986).The gravimetric properties of the 100- and 200-nm systems, as reflected bypelleting properties in centrifugation (15 000 g, 20 min) are summarized inTable 2.2. It may be observed that the increased densities of the 100-nmFe+2-loaded systems are not sufficient to result in significantly greaterpelleting on low speed centrifugation. however, the 200-nm diameter Fe+2-loaded systems do exhibit significantly increased proportions in the pelletfollowing centrifugation, indicating a heterogeneity in Fe+2 accumulation.300 ^-t,a 250–.co°313- 200 • "1076 150–E4.-...,.100 -Li^ AOil^ov " • • ... • . . . . . -I^i^t^1200 400 600 800Citrate Concentration (mM)• 6 1. -....... .. . . &vv 00AAa01000CI77^100 .^,^41^•^.A..^80^-......Ir'll------604020200^400^600^800Citrate Concentration (mM):-45 50–^0-0Ec44Figure 2.3(A). Effect of varying the external iron concentration and internalcitrate concentration on the amount of iron accumulated by 100 nm EPCvesicles with an acidic interior (pli1=4.0; p1-10.7.5). Iron concentrations of 1(0), 2 (46. ), 5 ( 0) and 10 (y) MM were used. The concentration of A23187used was 51.1g/m1 and the temperature was 23 °C.Figure 2.3(B). Effect of varying external iron concentrations and internalcitrate concentrations on the trapping efficiency of 100 nm EPC vesicles.The filled symbols represent the iron concentrations used in Fig. 2.3(A).45Table 2.1. Effect of Liposome Size on Iron EntrapmentFilter Pore Size (nm)Liposome Diameter^Iron Uptake (nm)^(nmoles Fe/Rmole phospholipid) 50 nm^73 + 23 nm^ 176 + 4100 nm^93 + 31 nm 181 + 2200 nm^164 + 72 nm^347 + 15400 nm^244 + 97 nm 387 + 2The mean vesicle diameter was determined after loading 300 mM citratevesicles with Fe+2 (2 mM initial external concentration) for 2 hours at 23 °Cin the presence of 10 pg/m1 A23187 (n=3).46Table 2.2. - Gravimetric Properties of EPC LUVs Loaded with IronFilter Pore OEL,S size Lipid Iron Iron:Lipid ratio(nm) (nm) (%) (%) nmol Fe+2timol PLSPN 100 123 97.3 + 3.0 98.2 + 3.5 271 + 50PLT 100 190 2.7 + 1.0 1.8 + 0.6 487 + 80SPN (Ctl) 100 106 100 NAPLT (Ctl) 100 0 NA -SPN 200 214 53.5 + 5.9 34.6 + 3.6_ 630 + 30PLT 200 214 46.5 + 5.3 65.4 + 7.3 1370 + 85SPN (Ctl) 200 214 85.0 + 3.6 NA -PLT (Ctl) 200 214 15.0 ±2.6 NAUptake conditions: initial citrate concentration of 1.0 M.; initial externaliron concentration 10.0 mM; accumulation time 2 h at 23 °C; concentrationof A23187 10 pg/ml. Control vesicles underwent no iron loading. The resultspresented are the means of three separate experiments. PL indicatesphospholipid; SPN indicates supernatant; PLT indicates pellet formedfollowing centrifugation at 15 000 X g for 20 minutes (n=3); Ctl indicatescontrols. NA, not applicable.47The 200-nm diameter EPC LUVs loaded with Fe+2 identified in Table2.2 constitute a significant step towards satisfying one of the initial aims ofthis investigation, namely to prepare LUV systems which can be readilyisolated by low-speed centrifugation. It is of interest, however, to extendthese studies to elements of higher atomic weight, which could exhibitimproved density characteristics. In this regard, it was found that A23187-mediated, pH gradient dependent loading of barium into EPC LUVs could bereadily achieved (results not shown). The larger atomic weight of barium(137.3) compared to iron (55.8) would be expected to result in more denseLUVs , assuming comparable levels of loading. This was observed, as shownin Table 2.3 for the 100-nm diameter EPC system, where Ba+2 loaded LUVsexhibit improved pelleting properties upon low speed centrifugation incomparison to the 100-nm diameter Fe+2-loaded systems (Table 2.2).A final area of investigation concerned the influence of accumulatedFe+2 or Ba+2 on the electron density of the LUV systems as visualized bycryo-electron microscopy. The stability of the LUVs containing Fe+2 or Ba+2was of concern in these experiments given possible lytic events that couldoccur prior to or during the vitrification process. In this regard, the presenceof long chain saturated lipids and cholesterol is well known to increase thestability and decrease leakage from liposomal systems. This is illustrated inFig. 2.4 for entrapped Fe+2 and Ba+2. As shown in Fig. 2.4A, over a 24 htimecourse (23 °C), considerable leakage of entrapped Fe+2, following therapid initial uptake is observed for EPC LUVs. In contrast, little leakage isobserved over a 24 h time course at 65 °C (chosen to ensure that thevesicles were in the liquid-crystalline phase) following uptake of Fe+2 intoDSPC:cholesterol (55:45. mol:mol) LUVs (Fig. 2.4B). Similar stability is alsoobserved for DSPC:cholesterol LUVs loaded with Ba+2. as illustrated in48Table 2.3. Gravimetric Properties of EPC LUVs Loaded with barium (100-nm pore size)Supernatant PelletTotal Lipid 79.4% + 5.7% 20.6% + 3.8%Total Lipid - Controlsa 93.8% + 2.2% 6.2% + 0.6%Total Barium 61.1% + 5.1% 38.9% + 13.7%nmoles Bahtmole Lipid 402 + 60 984 + 344Particle Diameter 110 + 30 130 + 36a 100 nm vesicles that contained no barium.Ba+2 uptake conditions: 2 h at 60 °C with 20 mM external barium, 50Itg/ml A23187 (n=3).49Figure 2.4. Time course of loading Fe+2 into (A) 100 nm EPC LUVs and (B)100 nm DSPC:cholesterol (55:45; mol:mol) LUVs. The LUVs were preparedby extrusion through filters with 100 nm pore size and exhibited a pHgradient (pH1=4.0; pH0=7.5). The internal buffer concentration was 300 mMcitrate and the external Fe+2 concentration was 2.0 mM. The experimentaltemperature was 65 °C for the DSPC:cholesterol LUVs and 23 °C for theEPC LUVs. Part (C) illustrates the retention of barium following loading into100 nm DSPC:cholesterol (55:45; mol:mol) LUVs at 65 °C. The internalbuffer concentration was 1.0 M citrate and the external bariumconcentration was 10 mM.^ 3.5A.-3.03.-2.5-13-2.0124)a.-1.5 •. 1.0 •0 0.5 •400-350 - +41300-250 -1200 -0150 -100 -50 -0..... 600 --o-8. 500 -li.c^00,0. 400- •fao.co.0 .300-q. 200--.,0^/a3 100 -co00^ctCSg^0^0 4^8^12^16Time (Hours)20-3.00- 2.5- 2.0-1.5-1.0-0.50.02404 0.08 12^16Time (Hours)20 240I^300 ^  3.0....0^250 -11%8.^o;-o^%\0^0li.- 2.5 0o.1----e2^0^ 0^c.co200 -I0•.^- 2.0 -aca. 150 - / \0^-,x0ao0- 1.5 ca'3 ...„.............^0.I i 00 - /0r • •re7•• • .. . .^I^' ' " 1.0• D• ^.—.u.^50 -#• •a-0.5 I75•0Time 12^1616^20^240.0 ......g^0^4^8•••0 700 •^ 3.5C.505 1Fig. 2.4C. Cryo-electron microscopy studies on the Fe+ 2-loaded systemsrevealed only slightly improved contrast resulting from a uniform darkeningof the LUV interior (Fig. 2.5(B)). However, the Ba+ 2-loaded systemsexhibited remarkably improved contrast as illustrated in Fig. 2.5(A). It islikely that the electron dense particles observed entrapped within the LUVscorrespond to crystalline barium-citrate precipitates.52Figure 2.5A. Cryo-electron micrograph of 100 nm DSPC:cholesterol (55:45;mol:mol) vesicles containing barium which was loaded using A23187 (12 [tg)and a transmembrane pH gradient (pH1=3.0; pH0=7.4) at 60 °C for 2.5 h(see Methods). The internal buffer concentration was 600 mM citrate andthe external barium concentration was 12 mM. The barium was visualizedby intensity contrast, while the membranes were visualized by phasecontrast. The scale bar represents 500 nm. The entrapped bariumconcentration was 212 nmol Ba+2 per !_tmol phospholipid.2.58. Cryo-electron micrograph of iron-loaded DSPC:cholesterol vesicles.532.4 DISCUSSION The studies reported here establish the ability of A23187 to mediateApH dependent loading of Fe+2 and Ba+2 into LUVs with an acidic interior.Features of interest concern the mechanism of the A23187 loading process,the physical state of the loaded Fe+2 and Ba+2 and the utility of the denseliposomes that can thus be generated. With respect to the ability of A23187to mediate the transbilayer transport of iron, it was first reported by Younget al. (Young et al., 1975) that A23817 could facilitate the transbilayertransport of iron in model (liposome) membranes and red blood cells. Theresults presented here show that A23187 transports the ferrous form andthat extremely high levels of Fe+2 entrapment (170 mM) can be achievedemploying the pH gradient approach. The ability of A23187 to transportBa+2 has been reported previously, where the affinity of A23187 for Ba+2 isat least 100-fold less than that observed for Ca+2 (Pfeiffer et al., 1974). Thiscorresponds to the reduced ability of A23187 to transport Ba+2 ascompared to Fe+2 (Fig. 2.4). Even though the affinity of A23187 for Ba+2 issmall, substantial amounts can be loaded into liposomes (330 mM interiorconcentration) with a transmembrane pH gradient (inside acidic).With regard to mechanism, it is likely that the A23187-mediatedprocess involves the exchange of Fe+2 or Ba+2 for two H+ ions, as observedfor Ca+2 (Veiro and Cullis, 1990). What is less clear is the form of theaccumulated material. Experiments with barium-citrate solutions indicatedthat precipitates will form at concentrations well below those observed forthe barium-loaded liposomes. Thus, the observation of electron denseregions in the Ba+2-loaded systems suggests the formation of barium citrateprecipitates.54LUVs with enhanced densities as evidenced by gravimetric propertiesand higher electron densities giving rise to greater electron microscopycontrast have a variety of important potential uses. The ability to isolateLUVs from serum components in vitro, for example, would be ofconsiderable utility in identifying serum proteins which associate stronglywith liposomes and which may play a role in the clearance process in vivo(Bronte and Juliano, 1986). Alternatively, enhanced electron microscopiccontrast could allow the direct detection of extracellular and intracellularLUVs in vitro and in vivo. Such an ability would be of considerable utility inassaying the targeting of liposomes, as well as determining their integrityand metabolic fate.The degree of success in attaining these two goals has varied. Thedevelopment of a viable system for isolating LUVs from living systems willdepend upon creating a stable homogeneous population of dense liposomes,based upon the techniques described here. Greater success has beenachieved with regard to creating electron dense liposomes. Barium-loadedliposomes have been visualized from liver tissue using cryo-electronmicroscopy following administration to mice (J. Wheeler, personalcommunication).In summary, the results of this work establish the A23187-mediatedloading procedure as a relatively general method for exchanging protonsentrapped in LUVs for cations of considerably higher molecular weight. Theenhanced gravimetric densities of these systems has potential applicationsin LUV isolation from complex media, whereas the enhanced electrondensity can result in the direct visualization of LUVs employing electronmicroscopy.55CHAPTER 3 - UPTAKE OF BASIC AMINO ACIDS AND PEPTIDES INTOLIPOSOMES IN RESPONSE TO TRANSMEMBRANE pH GRADIENTS 3.1 INTRODUCTION The ability of the neutral forms of certain weak acids and bases topermeate through membranes has long been recognized. Early studies inthis area included that of Jacobs (1940) on red cells; they were followed bystudies using mitochondria (Chappell and Crofts, 1966) and chloroplasts(Crofts, 1967) to demonstrate the rapid transbilayer movement of theneutral form of amines, such as ammonia, and carboxylic acids, such asacetate. These observations have led to the use of weak acids and bases asprobes for pH gradients (ApH) in cells and organelles (Rottenburg, 1979) andhave also stimulated studies demonstrating the uptake of weak bases(Deamer et al., 1972; Nichols and Deamer, 1976) into simple liposomalsystems exhibiting a ApH.More recently, it has been shown that a variety of amino-groupcontaining drugs can be accumulated into large unilamellar vesicles (LUVs)with an acidic interior (Madden et al., 1990) and that the transbilayerdistributions of phospholipids, such as phosphatidylglycerol (PG;Redelmeier et al. 1990) or phosphatidic acid (PA; Eastman et al., 1991), inLUVs can also be readily modulated by transbilayer ApHs. All theseprocesses rely on the fact that the neutral forms of the transportedcompounds are considerably more permeable than the charged forms,which can result in the net transport of weak bases into acidiccompartments and weak acids into basic compartments. The concentrationgradients that can be achieved at equilibrium are considerable. Under56appropriate conditions, a lipophilic amine, for example, will redistributeacross a membrane to reflect the proton gradient. As indicated in Chapter 1,a ApH of three units (inside acidic) can, therefore, result in a 1000-foldhigher concentration of the amine inside than outside.In this chapter, this principle is tested for those amino acids andpeptides in which the carboxyl groups have been modified to the methylester or amide forms. It is shown that such compounds are readilyaccumulated into LUVs with an acidic interior, and a detailed kineticanalysis indicates that transport proceeds via the neutral form. This workthus provides a different expansion of the work presented in Chapter 2 inthat it involves the study of charged compounds traversing lipid bilayers ina neutral form.573.2 MATERIALS AND METHODS Egg phosphatidylcholine (EPC) was obtained from Avanti Polar Lipids(Birmingham, ALA). 14C-methylamine and 3H- tetraphenylphosphoniumbromide were purchased from New England Nuclear (Boston, MA). Lysinemethyl ester and all other chemicals used were purchased from SigmaChemical Co (St. Louis, MO). The hydrophobic pentapeptide (Ala-Met-Leu-Trp-Ala; de Kroon et al., 1989) was synthesized as a carboxyl amide or freeacid by use of solid phase methods in the lab of Dr. I. Clark-Lewis (Clark-Lewis and Kent, 1989). The carboxylic acid form was converted to a methylester as described by de Kroon et al. (1989). Vesicles were produced asoutlined previously (see Section 2.2).To generate the transmembrane pH gradient, the LUVs in the pH 4.0media were passed down a 10 cm Sephadex 0-50 (050-150) columnpreviously equilibrated with 150 mM NaC1, 20 mM HEPES, pH 7.5 (HEPESbuffered saline, HBS). Uptake of the lysine methyl ester and pentapeptideswas performed by first dissolving them in 1 ml HBS media, to which 0.25 mlLUVs (final lipid concentration of 1-5 mM) exhibiting a ApH (pHo = 7.5, pHi= 4.0) were added. Entrapment levels were monitored using 0.1 ml aliquots,which were removed at selected times from this incubation mixture andpassed through 1.0 ml Sephadex 0-50 columns (pre-spun) by centrifugationfor 1 minute at 2500 rpm to remove exterior (untrapped) material. Allexperiments were conducted at 20 °C unless otherwise indicated.Transmembrane Na+/K+ (external/internal) gradients were createdusing a modification of the technique of Hope et al. (1985). Vesicles wereprepared in 150 mM K2SO4, 300 mM HEPES (pH 7.5) after which theexternal K+-buffer was replaced by 150 mM Na2SO4, 300 mM HEPES (pH7.5) by passing the vesicles down a 10 ml Sephadex G-25 column. In some58cases (see Results) a HEPES concentration of 20 mM in the internal andexternal buffers was employed. Uptake of the hydrophobic pentapeptide inresponse to Alp was carried out at 20 °C by dissolving the pentapeptide inthe Nat-buffer and adding EPC vesicles bearing a Nat/Kt gradient (finallipid concentration 2-3 mM) and 1 RM valinomycin. Aliquots were taken atindicated intervals and untrapped material removed via passage through1.0 ml Sephadex G-50 columns (pre-spun) as indicated above.Lysine methyl ester concentrations were determined by a modificationof the technique employed by Hope and Cullis (1987), employing TNBS(trinitrobenzenesulfonic acid) to label the primary amino groups of lysinemethyl ester. The buffer used for the labelling was 100 mM NaHCO3, 50mM H3B03, pH 10.0. A reference cuvette containing 2.5 ml of buffer, pH10.0, was placed in the reference beam. The sample cuvette contained 2.5ml of buffer, pH 10.0, with 0.5 mM TNBS. 50 iLt1 aliquots of vesiclescontaining lysine methyl ester were then added. The resulting change inabsorbance was measured at 420 nm after incubation for 1 hour (in thedark). 200 [il, 0.5% Triton X-100 was added to both cuvettes to solubilizethe vesicles and thus expose all primary amino groups to the TNBS. Theabsorbance in the presence of detergent was taken to represent 100%labelling (Hope and Cullis, 1987).The amount of hydrophobic pentapeptide trapped was quantified bymeasuring the tryptophan fluorescence (kex = 280 nm, kern = 354 nm)employing a Perkin-Elmer LS 50 Luminescence Spectrometer (The Perkin-Elmer Corp., Norwalk. CT), in 0.5% (w/v) sodium cholate containing buffer(300 mM HEPES, 150 mM Na2SO4; pH 7.5). The sample volumes used wereadjusted so that the resulting emission intensity was directly proportional tothe amount of peptide present.59The magnitudes of the pH gradients and membrane potentials presentwere measured using 14C-methylamine and 3H-tetraphenylphosphoniumbromide (3H-TPP), respectively, as indicated elsewhere (Hope et al., 1985;Madden et al., 1990). The concentrations used were 1 [ICl/mi. The amount ofprobe accumulated was determined via liquid scintillation counting afterremoving untrapped label. Transmembrane pH gradients could then becalculated using the relationship ApH = log{[methylamine]in/[methylaminelout} as indicated in Mayer et al. (1988).Membrane potentials were calculated similarly for 3H-TPP using therelationship Av = -59 log WCPP)in/(TPP1out} (see Hope et al., 1985).Phospholipid concentrations were determined by a modification of themethod of Fiske and Subbarow (1925). Typical phospholipid concentrationswere approximately 3 mM. All data are means, n=3 + SEM (standard error ofthe mean). Where error bars are not shown on figures, the SEM values werewithin the dimensions of the symbols used.The kinetics of the uptake process were analyzed assuming that onlythe neutral (deprotonated) form of the amino acid or peptide could moveacross the LUV bilayer. As developed more completely elsewhere (Cullis etal., manuscript in preparation; see also Section 1.9.1), the accumulationprocess should then obey the relation[A(t)11^fA(eq)li (1 - et)^ (1)where [A(t)} i is the interior concentration of the amino acid or peptide attime t, [A(eq)Ji is the equilibrium interior concentration at t=infinity and k isthe rate constant associated with the uptake process.We consider the analysis for the lysine methyl ester, which containstwo ionizable (primary) amino functions with corresponding dissociationconstants K1 and K2. Under the assumption that only the neutral (fully60deprotonated) form is able to translocate the membrane, it can be shownthatk = (P Am/Vo) (1 + [H}0/K1 + (1-11]02/K1K2)-1^(2)where P is the permeability coefficient for the neutral form, Am is the area ofthe LUV membrane, Vo is the aqueous volume of the lipid dispersion and(Who is the exterior proton concentration. Thus if K1, K2 « [H4]0, weobtaink = P Am K1 K2 /V0 [H+]o2^ (3)Alternatively, if the peptide contains only one amino function, as is the casefor the pentapeptide investigated here, we obtaink = P Am K1 /V0 (11-110^ (4)It should be noted that P is an effective permeability coefficient which isrelated to the actual membrane permeability coefficient, Pm, by the relationPm = KP where K is the membrane-water partition coefficient of the charged(protonated) species of the amine. Also, this is an initial rate analysis andassumes that the concentration of the neutral amine on the outer surface ofthe LUVs is much greater than that on the inner surface. The rate constant,k, can be calculated from the slope of a plot of ln {([A(eq)]1 - [A(t)]i)/[A(eq)lilversus t, and a subsequent plot of log k versus the external pH shouldexhibit a slope of two if the uptake actually proceeds via the neutral form forthe lysine methyl ester, or a slope of one for the pentapeptide methyl ester.The rate constant (k) was determined by applying a linear least-squareanalysis to the uptake data using a commercially available plotting program(Sigma-Plot, Jandel Scientific, Corte Madera, CA; 1986), employing k and[A(eq)]i as variables.613.3 RESULTS The first set of experiments was aimed at demonstrating that a basicamino acid derivative could respond to transmembrane pH gradients in amanner similar to that observed for other amines (Mayer et al. 1986;Madden et al. 1990). As shown in Figure 3.1 lysine methyl ester was rapidlyand efficiently accumulated into EPC LUVs with an acidic interior (pHi =4.0, pHo = 7.5). Maximal entrapment levels were reached within 5-10minutes. A corresponding decrease in the pH gradient using MCmethylamine was observed (the gradient dropped from 3.2 to 1 pH units).The maximal concentrations entrapped under these conditions wereapproximately 65 nmoles lysine methyl ester/mole phospholipid(corresponding to an interior concentration of 43 mM). This high level ofinternalized lysine methyl ester was maintained for at least 24 hr with nomeasurable leakage. It should be noted that the drop in the pH gradientupon accumulation of lysine methyl ester is larger than can be accountedfor by the uptake of lysine. It is possible that this results from reducedtrapped volumes that can be observed at high citrate concentrations. Thecontrol vesicles exhibiting no pH gradient (7.5/7.5; 4.0/4.0) showed littleuptake of the lysine methyl ester (10% or less of that observed for thevesicles with a pH gradient). Lysine (unmodified) was not accumulated invesicles bearing a pH gradient (data not shown). This may be attributed tothe presence of the negatively charged carboxylic acid group which would beexpected to inhibit transbilayer diffusion (see, for example. Gutnecht andWalter, 1981). Control experiments employing TNBS revealed that lysinemethyl ester accumulated into these LUVs did not react with the TNBS.Labelling occurred only after the vesicles had been solubilized with Triton X-100. Hence it was concluded that the primary amino groups of90807060 —50 —40 — •0.0^0.262Figure 3.1. Time course of uptake of lysine methyl ester into 100 nm EPCvesicles (3.8 mM) bearing a 7.5/4.0 (external/internal) pH gradient (0 ).Control vesicles with no pH gradient (7.5/7.5 A and 4.0/4.0 0) were alsotested. Uptake was conducted at 20 °C. The initial external concentration oflysine methyl ester was 1.3 mM.3.50— 3.00—2.500—2.00—1.50•^•- 1.008  60.50^0.000.4^0.6^0.8^1.0Time (Hours)63accumulated lysine methyl ester were not accessible to the TNBS reagent(eg. not surface associated) in the intact vesicles.The next set of experiments was designed to clarify the mechanism ofuptake, specifically to show that the lysine methyl ester was accumulated inthe neutral form. As outlined in the Materials and Methods, if this is thecase, the slope of a plot of log k versus pHo should exhibit a slope of two. Asshown in Figure 3.2A, the initial rate of uptake of the lysine methyl esterinto these LUVs increased considerably as the exterior pH was raised from6.40 to 7.25. The absolute value of ApH was maintained at 3.5 units byvarying the pH of the internal (citrate) buffer used and the external (HBS)buffer in unison. Rate constants (k) for the different pHo values wereextracted from this data employing the plots of Fig. 3.2B, and a subsequentplot of log k versus pHo (Fig. 3.2c) reveals a straight line with a slope of1.85. This supports a mechanism involving the transport of the neutral(deprotonated) form of the lysine methyl ester.Previous studies from this laboratory have shown that acidicphospholipids, such as phosphatidylglycerol (PG) and phosphatidic acid (PA)can also experience transmembrane transport in response to pH gradients,again involving transport of the neutral form (Hope et al., 1989; Redelmeieret al., 1990). Such transport exhibits high activation energies, in the rangeof 30 kcal/mol. As shown in Fig. 3.3A, the rate of uptake of the lysinemethyl ester also increased considerably as the temperature is raised from25 to 45°C for an EPC system where pHo= 6.5 (20 mM MES, 150 mM NaC1)and pHi = 3.0. An analysis of the rate constants (Fig. 3.3B) and subsequentcalculation of the activation energy (Ea) from the Arrhenius plot of Fig. 3.3Crevealed an Ea of 36 kcal/mol.64Figure 3.2A. Time course of uptake of lysine methyl ester into 100 nm EPCvesicles (approximately 1.2 mM) for different external pH values. The pHvalues of the samples (external/internal) were 6.4/2.9 (G ), 6.7/3.2 ( V ),7.0/3.5 ( CI ) and 7.25/3.75 ( A ). The initial external concentration of lysinemethyl ester used was 1.8 mM.3.2B. Plot of In {aA(eq)Ji - [A(t)10/[A(eq)]i} versus time; where [A(t)ii is theinterior concentration of the accumulated amine at time t and [A(eq)]i is theinterior concentration at equilibrium. For details see Materials and Methods.The slope of the lines gives the rate constant (k) for the transbilayertransport of lysine methyl ester.3.2C. Plot of log k versus external pH. The slope of this line is 1.85.In [A(eCI — [AM)tgeqnl1.Li, b Liinmolee lysIne methyl esterhamole phoephollpld^5.1^5-0^ch de 6 4 14 b al 4.1.3 o•66Figure 3.3A. Time course of uptake of lysine methyl ester into 100 nm EPCvesicles (2.2 mM) bearing a 6.5/3.0 (external/internal) pH gradient. Uptakewas conducted at 20 °C (0), 25 °C (A), 30°C ( 0) 35 °C (v ) and 40 °C (0).The external concentration of lysine methyl ester was 1.8 mM.3.3B. Plot of in {flA(eq)Ji - (A(t)Ji)/fA(eq)Jil versus t; where [A(t)11 and [A(eq)]ihave the same meanings as indicated in the legend to Fig. 3.2.3.3C. Arrhenius plot of the rate constants (k) for lysine methyl ester uptake.The activation energy calculated from the slope of this plot is Ea = 36kcal/mol.pictotidootid opunlitspre 'Anew ouistA 11111101.UUq^ci^o ri•^0^r-i7 7Jg.5q q 0 n q°^7 7 r4^r)^1^1^1i^gb.y4 ^)1 Mil — IRb•Yv] 1 ul68In a previous study de Kroon et al. (1989) have shown that the methylester of a hydrophobic, basic pentapeptide (Ala-Met-Leu-Trp-Ala-(Me0)) canbe accumulated into LUVs exhibiting a membrane potential (Alp). Asemphasized above, weak bases generally permeate through bilayermembranes in the neutral (deprotonated) form. In this regard, the presenceof a Alp (inside negative) results in the induction of a pH gradient (insideacidic) at electrochemical equilibrium (Cafiso and Hubbel, 1983). It istherefore likely that the behaviour observed by de Kroon et al. (1989)reflects uptake in response to an induced ApH than the Alp per se. As shownin Figure 3.4a, the pentapeptide methyl ester employed by de Kroon et al.(1989) is very rapidly accumulated into LUVs exhibiting a ApH (pHo= 7.5,pHi = 4.0), supporting this conclusion. As in the case of the lysine methylester, a drop in the residual pH gradient was also observed, correspondingto protonation of the pentapeptide amino function after traversal of themembrane in the neutral form. The maximal concentrations entrapped wereapproximately 35 nmoles peptide/mole phospholipid. This level ofinternalized peptide was maintained for at least 24 hr with no leakage (datanot shown). In contrast, uptake in response to Alp revealed slower uptake(Fig. 3.4B) which was dependent on the internal buffering capacity.Specifically, the uptake was much slower at high internal bufferconcentrations (300 mM HEPES, pH 7.5) than for LUVs with lower internalbuffer concentrations (20 mM HEPES, pH 7.5). The slower uptake for thesystem containing 300 mM HEPES may be attributed to a smaller inducedpH gradient due to the high internal buffering capacity. Subsequentexperiments revealed that the amide form of the pentapeptide exhibitedsimilar uptake behaviour as the methyl ester form in response to both Alpand ApH (results not shown).69Figure 3.4A. Time course of uptake of the pentapeptide (Ala-Met-Leu-Trp-Ala methyl ester) into 100 nm EPC vesicles (6.4 mM) bearing a 7.5/4.0(external/internal) pH gradient (0 ). Control vesicles with no pH gradient(7.5/7.5 A and 4.0/4.0 0 ) were also tested. Uptake was conducted at 20°C. The initial external peptide concentration was 0.4 mM.3.500-0/^-0 0A0-3.000 -2.50-2.00.- 1 50•- 1.00-0.50- iiel.•"'s0140 ^0.0042Time (Hours)3703.4B. Time course of uptake of the pentapeptide into 100 nm EPC vesicles(6.6 mM) exhibiting a valinomycin induced K+ membrane potential withhigh (300 mM HEPES) internal buffering capacity ( 0 ) and low (20 mMHEPES) internal buffering capacity ( A ). Control vesicles with novalinomycin present generated background levels of peptide uptake ( 0 ).The filled symbols ( A • ) represent the measured Alp values correspondingto the high and low buffering concentrations used. Uptake was conducted at20 °C. The initial external peptide concentration was 0.5 mM.—160•—140o.—120.."'S014004..0 351372.t 30'6EC8-0-ro0^11^-------\.AL^ 0—100 <A —80 •—60 1A •—400^0^020I I^02^3 4Time (Hours)71Subsequently experiments were undertaken to verify that thepentapeptide was indeed accumulated in the neutral form, as observed forlysine methyl ester. As outlined previously, the slope of a plot of log kversus pHo should exhibit a slope of one if there is only one basic titratablegroup present (as is the case for the hydrophobic pentapeptide). Figure 3.5Areveals that the rate of uptake of the pentapeptide increased significantly asthe exterior pH was raised from pH 6.50 to 7.75. Again, the absolute valueof ApH was maintained at 3.5 units by varying internal and external pHvalues in unison. A plot of log k versus pHo reveals a straight line with aslope of 0.85 + 0.10 (Fig. 3.5C). This strongly supports a mechanisminvolving transport of the neutral (deprotonated) form of the pentapeptide.Experiments were undertaken to determine the activation energy forthe transport of the pentapeptide. As shown in Fig. 3.6A, raising thetemperature from 23.5 °C to 49 °C for an EPC system where pHo = 6.0 (20mM MES, 150 mM NaC1) and pHi = 3.0 (300 mM citrate) resulted in adramatic increase in the rate of transport. An analysis of the rate constants(Fig. 3.6B) and subsequent calculation of the activation energy from theArrhenius plot of Fig. 3.6C revealed an activation energy of 30.6 kcal/mol.72Figure 3.5A. Time course of uptake of the hydrophobic pentapeptide into100 nm EPC vesicles (approximately 3.3 mM) for different external pHvalues. The pH values of the samples (external/internal) were 6.50/3.00( G ), 6.75/3.25 ( v ), 7.00/3.50 ( 0), 7.25/3.75 ( A ), 7.50/4.00 (0) and7.75/4.25 ( • ). The initial external concentration of pentapeptide used was0.42 mM.3.5B. Plot of in {([A(eq)li - [A(t)]0/[ii(eq)11l versus t; where [A(t)11 is the interiorconcentration of the accumulated amine at time t and [A(eq)li is the interiorconcentration at equilibrium. For details see Materials and Methods. Theslope of the resulting straight lines gives the rate constant (k) fortransbilayer transport of the hydrophobic pentapeptide.3.5C. Plot of log k versus external pH. The slope of this line is 0.97.17C1:17,.co. 40-0e• 30-I. 25-.,..20--0=a.• 15-o.0t° 10-toa.5-0e75^00.00.0..Yt735. 45EA1.0^2.0lime (Hours3.0^40...v.-,c -0.8-^— 1.0 I^I^I^I^I1^2^3^4^5Time (Hours)^3.5^-4.0-- 4.5 -log k -1°-- 5.5--6.0^7.00^7.25^7.50^7.75External pH6.50^6.7574Figure 3.6A. Time course of uptake of the pentapeptide into 100 nm EPCvesicles (4.5 mM) bearing a 6.0/3.0 (external/internal) pH gradient. Uptakewas conducted at 23.517) 45 °C (),^0°C (0), 30.5 °C (A), 35 OC (0), 40 °C (and 49.5 °C (•). The external concentration of the pentapeptide was 0.85mM.3.6B. Plot of in WA(eq)Ji - [A(t)]0/[A(eq)11l versus t; where [A(t)] i and [A(eq)]ihave the same meanings as indicated in the legend to Fig. 3.2.3.6C. Arrhenius plot of the rate constants (k) for pentapeptide uptake. Theactivation energy calculated from the slope of this plot is Ea = 30.6kcal/mol.35150 •10^Er•,11),77 r-7-00 ^3025^•^V/•20 • /V0000 1.0^2.0^3.0Time (Hours- 1.0-- 1.5-- 2.0--2.5-1..- 3.0^00 1.0^2.0^3.0Time (Hours)-7.5- 8.0 --8.5 --9.0 -- 9.5 -6.05.04.0 6.0In k-10.0-- 10.5-- 11.0-- 11.5-- 12.03.10^3.15^3.20^3.25^3.301/T (X 10 -3)3.35^3.4075763.4 DISCUSSION The results of this report establish that basic amino acids andpeptides. in which the C-terminal carboxyl functionalities have beenmodified to form methyl esters or amides, can be rapidly and efficientlyaccumulated into LUVs in response to transmembrane pH gradients (insideacidic). Two points of interest concern the mechanism of uptake and theimplications for transbilayer translocation of peptides and proteins in vivo.With regard to mechanism, previous work on a variety of lipophilic amino-containing drugs (Mayer et al. 1986; Madden et al. 1990) yielded uptakebehaviour consistent with transport of the neutral (deprotonated) form ofthe amine. In the case of the lysine methyl ester investigated here, thedependence of the rate constant k on the external pH, which is consistentwith an inverse square relationship between the uptake rates and theexternal proton concentration, also provides strong evidence for transport ofthe neutral (deprotonated) form. The rapid uptake of the methyl esterderivative employed by de Kroon et al. (1989) in response to ApH observedhere, as well as the much slower uptake observed in response to Ai at highinternal buffer concentrations (which may be attributed to reduced uptakein response smaller pH gradients induced by the K+ diffusion potential), andthe linear dependence of the rate constant k on the external pH are also allconsistent with accumulation of the neutral form and indicate that uptakein response to Alp per se does not occur.The rapid uptake of the lysine methyl ester in response to ApH atexternal pH values which are considerably below the pKs of the associatedprimary amines clearly suggests that the neutral form must be highlymembrane permeable. In this regard, the analysis in the Materials andMethods section indicates that the permeability coefficient P of the neutral77form obeys the relationP = k Vo [1-14]02/ Am K1K2.^ (5)From the data of Fig. 3.2B at 20°C and pHo = 7.0, k can be calculated as1.86 X 10-4 sec-1. The pKi of the N-terminus amino function can beestimated as 9.20, whereas the pK2 of the lysine side chain is 10.80 (Stryer,1988). Thus, assuming an area per molecule for EPC of 70 A°2, it can becalculated that P = 1.31 X 10-2 cm sec-1, corresponding to extremely rapidtransbilayer movement of the neutral form.The activation energies observed for the uptake of lysine methyl esterand the pentapeptide are of interest, particularly with respect to thoseobserved for PG and PA, which exhibit high activation energies in the rangeof 30 kcal/mole (Eastman et al., 1991). Such activation energies arecomparable with those observed for the pentapeptide (30.6 kcal/mol) butlower than those observed for the lysine methyl ester (36 kcal/mol).Activation energies associated with diffusion across membranes have beenrationalized on the basis of the need to break and reform hydrogen bonds(Stein, 1967). Specifically, the activation energies associated with transfer ofa molecule into the membrane hydrocarbon can be estimated according tothe number of hydrogen bonds (e.g. with water) that must be broken toenter the hydrocarbon, less the number of bonds that reform in thehydrocarbon. Thus, for the lysine methyl ester the number of effectivehydrogen bonds to be broken (see Stein, 1967) would be four for the twoprimary amines and 0.5 for the methyl ester. Assuming a value of 5kcal/mole for the breaking of each hydrogen bond, this gives an activationenergy of approximately 23 kcal/mole, which is considerably lower thanobserved here. However, in the case of PG and PA, similar analyses indicateactivation energies of 40 and 28 kcal/mole, respectively. It is likely that the78value for PG would be reduced by an ability to form intramolecularhydrogen bonds when in the hydrocarbon. Clearly, the net activation energyfor the pentapeptide is difficult to estimate given the considerable potentialfor forming intramolecular hydrogen bonds on insertion into thehydrocarbon.This work has implications for mechanisms whereby peptides caninsert into and translocate across membranes in vivo. Certain polypeptidescan translocate across membranes in the absence of membrane proteins(Rapoport, 1990) by a mechanism that is not yet understood. Furthermore,signal sequences associated with translocated proteins exhibit lipophilicweak base character (Gierasch, 1989) and the mechanism whereby suchsequences move across membranes during protein biosynthesis are notunderstood (Verner and Schatz, 1988; Gierasch, 1989). It has beensuggested that signal sequences may be effectively pulled across bilayers bythe electrochemical potential present (Skerjanc, 1990). The work presentedhere indicates that transmembrane pH gradients could fulfill such a role bytrapping signal sequences in the lumen of the endoplasmic reticulum.Endoplasmic reticulum membranes possess an inwardly directed protonpump which can generate pH gradients (interior acidic) of up to 2 units, asdemonstrated in microsomal systems (Rees-Jones and Al-Awqati, 1984;Thevenod et al., 1989). Net translocation of signal sequences and somepeptides in response to transmembrane pH gradients in vivo is a clearpossibility.CHAPTER 4 - PERMEABILITY OF LIPID BILAYERS TO AMINO ACIDS4.1 INTRODUCTION A universal characteristic of cells is the membrane boundary thatseparates the interior cytoplasm from the exterior environment. The lipidbilayer provides a barrier to the free diffusion of solutes, while specializedintegral proteins permit specific solutes to pass, either by passive diffusion(channels and carriers) or enzyme-catalysed active transport. The generalbarrier properties of the bilayer have been well-established by previousstudies, particularly with respect to ionic solutes (sodium, potassium,chloride, hydrogen ion), polar solutes (water, glucose) and small neutralmolecules (glycerol, urea) (Deamer and Bramhall, 1986). Surprisingly littlework has been done on more complex ionic solutes such as the amino acids,clearly one of the most important solute species involved in cell function.In this chapter, we will describe both direct and indirect permeabilitymeasurements for several amino acids. In particular, we have addressed thefollowing questions:1. What are the permeability coefficients for the primary amino acidclasses, including neutral, polar, hydrophobic, and charged species?2. What is the effect of pH and phospholipid chain length on permeationrates?3. How do the permeabilities compare with measurements for other ionicsolutes, particularly phosphate, potassium and chloride?4. What mechanism best explains the measured permeation rates? That is,are the permeation rates controlled by standard models involvingpartitioning and Born energy barriers, or are they better explained in termsof bilayer fluctuations and transient defects?79These concerns are directly relevant to the permeation of lipid bilayers bycertain peptides, for example during insertion and translocation of signalsequences (Gierasch, 1989). The permeability coefficients and possiblemechanisms of transbilayer movement of the amino acids studied in thischapter are also of interest when compared to those of the modified aminoacids studied in Chapter 3.804.2 MATERIALS AND METHODS: Materials:All amino acids, chemicals and buffers used were obtained fromSigma Chemical Co. (St. Louis, Mo.). Egg phosphatidylcholine (EPC) anddimyristoyl phosphatidylcholine (DMPC) were obtained from Avanti Polar-Lipids (Birmingham, Ala.).Preparation of Lipid Vesicles:LUVs were prepared as described previously (see Section 2.2), exceptfor 50 mM amino acid being incorporated into the hydration (interior) buffer.Also, a custom-built extrusion device modelled upon one produced by LipexBiomembranes Inc. (Vancouver, Canada) as described by Hope et al. (1985)was used to produce 200 as opposed to 100 nm (pore size) vesicles.Measuring Efflux of Amino Acids from the Liposomes: To remove external solutes the LUVs containing the passively-trappedamino acid were then passed down a 10 cm Sephadex G-50 (G50-150)column previously equilibrated with buffer of the appropriate pH. Thecomposition of the buffers used was 50 mM buffer and 150 mM NaCl. Thefollowing buffers were used: citrate (pH 2-4); MES (pH 5-6); HEPES (pH 7-8)and Tricine (pH 9).The amino acid-containing LUVs (1.5 ml) were subsequently placed in16 mm dialysis tubing (SpectraporTM). This tubing has a molecular weightcut-off of approximately 12000-14000, so free amino acids lost from theliposomes were able to diffuse out of the tubing while the liposomesremained inside. The loaded dialysis tubing was then placed in a beakercontaining 100 ml of buffer solution of the appropriate pH which was stirredcontinuously. Samples (1.0 ml) were taken at appropriate time points fromthe external buffer solution to monitor amino acid efflux from the liposomes.81Control experiments with free amino acid solutions (50 mM) entrapped indialysis tubing revealed that essentially all the amino acid diffused out ofthe tubing within 10 minutes. With the exception of the hydrophobic aminoacids, half-times of amino acid permeation were in the range of one hour, sothat the diffusion barrier of the dialysis membrane was negligible. Allexperiments were conducted at 20-22 °C for the EPC liposomes and 30-32°C for the DMPC liposomes to ensure that both types of vesicles were in theliquid-crystalline phase. Samples were stored at 4 °C until amino acidconcentrations were determined.Fluorescamine labeling of Amino Acids: Fluorescamine (Sigma Chemical Co., St. Louis MO) was used tofluorescently label the primary amino groups of the amino acids forquantitative analyses. A 0.5 ml aliquot of the amino acid-containing sampleto be measured was added to 1.0 ml of sodium borate buffer (200 mM, pH9.0). The sample was vortexed vigorously while 0.5 ml of fluorescamine (at aconcentration of 20 mg/ 100 ml in acetone) was added. Mixing wascontinued for several seconds following fluorescamine addition to ensurelabeling of all amino acid present.The concentrations of the resulting fluorescently-labeled amino acidswere determined by measuring their fluorescence (excitation wavelength =390 nm; emission wavelength = 480 nm) using an SLM 8000CSpectrofluorimeter. Standard curves were created for each amino acid(generally in the range of 0-25 nmoles). It was established under theseconditions that the emission intensity was directly proportional to theamount of amino acid present.82Measurement of hydrophobic amino acid effluxPreliminary experiments revealed that the efflux of tryptophan andphenylalanine from the liposomes was too rapid to be measured using thedialysis tubing protocol. A modification of a technique using KI as anaqueous quencher was employed to allow for the on-line monitoring oftryptophan efflux (de Kroon et al., 1990). A 50 ml aliquot of vesicles wasadded to 2.0 ml of 1.0 M KI (containing 0.25 mM Na2S203 as anantioxidant) in a cuvette. The vesicles were prepared in a solutioncontaining 50 mM tryptophan, 50 mM buffer and 900 mM NaCl. Thevesicles were used immediately after the extrusion procedure was completedand were not passed down a column. Tryptophan fluorescence (excitationwavelength = 280 nm; emission wavelength = 360 nm) was monitoredcontinuously for a period of five minutes following addition of the vesicles.Any free tryptophan present at the start of the experiment was immediatelyquenched by the KI. The efflux of tryptophan was then measured as adecrease in the fluorescence intensity due to quenching of the inherenttryptophan fluorescence upon exposure to the KI present in the externalsolution. Control experiments measuring the efflux of 14C-glucoseestablished that 1M KI had no significant effects on membrane permeabilityover time courses of up to 3 hours in the EPC systems used.Phenylalanine permeability was measured using light scatteringchanges related to osmotically-driven water flux (Cohen and Bangham,1972). Liposomes were initially prepared in a low osmotic strength buffer(2.0 mM buffer, 6.0 mM NaC1). A 200 1.il aliquot of these vesicles was thenadded to 1.8 ml of high osmotic strength buffer containing 100 mMphenylalanine, 50 mM buffer and 150 mM NaCl. Light scattering wasmonitored by measuring the light scattering at 550 nm over a 5 minute83interval. An initial increase in scattering was observed, corresponding toosmotic shrinkage of the liposomes in response to the initial osmoticgradient. A point of minimum volume was reached, followed by a decreasein scattering as influx of the amino acid occurred. The maximum rate oflight scattering decrease following the minimum was used as a measure ofthe rate of influx of phenylalanine (Cohen and Bangham, 1972). Controlexperiments indicated that osmotic gradients of the magnitude employedhere did not result in increased membrane permeability to solutes, such asMC-glucose.Calculation of Rate Constants and Permeability Coefficients: The kinetics of the efflux process were calculated using an initial rateanalysis, which assumes that the concentration of amino acid inside theliposome is much greater than that on the outside (except forphenylalanine). As shown in Section 1.9.1, the efflux process should obeythe relationA(t)1 = A(0)1e-kt^(1)which can be rewritten as belowA(t)ex = (Vi/Vex) (A(0)1 - A(t)1)^(2)orA(t)ex = A(eq)ex (1 e t)^ (3)where A(t) is the concentration of amino acid at time t, A(0) is theconcentration at time 0. A(eq) is the equilibrium concentration at t = infinity,the subscripts i and ex represent interior and exterior and k is the rateconstant associated with the uptake process. Thus, a plot of ln{([A(eq flex -[A(t)]ex)/A[(eq)Jexl versus t should give a slope of -k, the rate constant. Rateconstants were determined by applying a linear least-square analysis to84efflux data using a commercially available graphing program (Cricket GraphVersion 1.2, Malvern, PA).Permeability coefficients (P) were calculated using the expressionP= (Vo/Am) k^ (4)where P is expressed in cm.s"1, k is the rate constant, Vo is the aqueousvolume of the lipid dispersion and Am is the area of the LUV membrane (seesection 1.7.1). This expression can be readily simplified to:13. (r/3)k^ (5)where r is the radius of the vesicle. This was assumed to be 100 nm for the200 nm vesicles used in this work. Errors for the P values calculated wereapproximately 5 X 10" 12 cm.s-1 for the EPC vesicles and 5 X 10-11 cm.s"1for the DMPC vesicles.854.3 RESULTS Several representative amino acids were chosen to be studied. Theseamino acids were: glycine (no side chain); serine (aliphatic hydroxyl sidechain); lysine (charged (basic) side chain); phenylalanine and tryptophan(aromatic (hydrophobic) side chains). The first set of experiments wasdesigned to determine whether pH had an effect on the permeabilitycoefficient of the amino acids. Figure 4.1 illustrates a typical set of effluxdata obtained for serine from 200 nm DMPC vesicles. Figure 4.1A showsthat the time course of amino acid efflux from the liposomes wasexponential and reached equilibrium after approximately 2 hours. Figure4.1B reveals how rate constants for the efflux process were derived (seematerials and methods for further details) from a linear transformation ofthe data in Figure 4.1A. Table 4.1 shows that pH had only a minor effect onthe rate constants and permeability coefficients of the charged and aliphaticamino acids, with all rate constants being in the range of 104 s"1 and withthe permeability coefficients being in the range of 10-12 cm.s"1.The next set of experiments was aimed at finding out whether alteringlipid chain length would have an effect on the permeability of amino acidsfor liposomes of the same size (approximately 200 nm). Previous work hasshown that liposomes made from C12 and larger amphiphiles have reducedmembrane permeability (Hargreaves and Deamer, 1978). DMPC (14 C) waschosen as the lipid for comparison with EPC (18 C) since it was the shortestchain lipid that still formed stable liposomes (provided a sufficientpermeability barrier for these studies to be undertaken; unpublishedresults). Control experiments established that the differences intemperature between the EPC and DMPC systems (approximately 10 °C)would only account for a 1.5-fold increase in permeation rates. Table 4.286201816• 14Ci 12m101864200.0x...,i xI A -0.6.-.^)41 1cr^—0.9•rc—1.5c^—1.230^60^90^120 150 180 210 240Time (minutes)Figure 4.1A. Time course of release of serine from 200 nm DMPC vesicles atvarious pH values. The pH values of the samples were 2.0 (0), 4.0 (0), 6.0 (A)and 8.0 (v). The experiment was conducted at 31 0C and the initial serineconcentration was 50 mM.4.1B. Plot of ln{(fA(eq)1ex - (A(t)}ex)/[A(eq)lexl versus time; where [A(eq)]ex isthe exterior concentration of amino acid at equilibrium and [A.(t)]ex is theexterior concentration of amino acid at time t. For details see Materials andMethods. The slopes of the lines gives the rate constant (k) for thetransbilayer diffusion of serine.870^600^1200 1800 2400 3000 3600Time (sec.)Table 4.1. Permeability Coefficients (P) and Rate constants (k) for the effluxof glycine, lysine and serine from 200 nm EPC vesicles at various pH values.ED_ P (X 10-12 cm s-1)^k (X 10-4 s-1)Glycine3 7.3 2.24 8.3 2.55 4.0 1.26 5.0 1.57 5.3 1.68 5.0 1.59 4.7 1.4Lysine2 9.0 2.74 2.0 0.67 3.7 1.18 6.7 2.09 4.0 1.2Serine2 0.7 0.23 1.3 0.44 4.7 1.45 8.3 2.56 10.7 3.27 3.3 1.08 4.3 1.39 4.0 1.288Table 4.2. Permeability Coefficients (P) and Rate constants (k) for the effluxof glycine. lysine and serine from 200 nm DMPC vesicles at various pHvalues.P (X 10-11 ems-1) k (X 10-4 s-1)Glycine3 2.6 7.94 1.9 5.76 1.6 4.78 1.7 5.2Lysine2 2.3 7.04 1.8 5.46 1.8 5.38 1.7 5.2Serine2 3.3 10.04 1.2 3.66 1.1 3.38 0.8 2.589shows that permeability coefficients and rate constants for efflux of thecharged and aliphatic amino acids for the DMPC liposomes were in thesame range as those observed for the EPC liposomes (approximately 10-11cm.s-1 and 104 s"1).A key observation of the previous experiments was that thehydrophobic amino acids (phenylalanine and tryptophan) both had muchhigher rates of efflux than any of the other types of amino acid examined.Both of these amino acids exhibited transbilayer movement which was toorapid to be monitored using the dialysis tubing protocol. The permeabilitycoefficients of phenylalanine and tryptophan were therefore determined viathe use of on-line fluorescence methods, as outlined in the Materials andMethods section. Figures 4.2A and 4.2B show typical time courses fortryptophan and phenylalanine efflux from EPC vesicles at pH 6.0 asmonitored by changes in fluorescence quenching (tryptophan) or lightscattering (phenylalanine). Permeability coefficients calculated from thisdata were 4.1 X 10-10 cm.s-1 for tryptophan and 2.5 X 10-10 cm.s"1 forphenylalanine, two orders of magnitude greater than those observed for theother amino acids (average permeability coefficient of 5.4 X 10-12 cm.s-1).90Figure 4.2A. Time course of the efflux of tryptophan from 200 nm EPCliposomes at pH 6.0 measured by the quenching of tryptophan fluorescence(arbitrary fluoresence units; excitation wavelength = 280 nm; emissionwavelength = 360 nm). The experiment was conducted at 22 °C and theinitial tryptophan concentration was 50 mM.4.2B. Time course of influx of phenylalanine into 200 nm EPC liposomes atpH 6.0 monitored by changes in light scattering (excitation and emissionwavelengths = 550 nm). The experiment was conducted at 22 °C and theinitial phenylalanine concentration was 45 mM. The light scattering startsfrom the maximum point of vesicle shrinkage.ATRYPTOPHAN20^40TIME (SECONDS)PHENYLALANINE9120^40TIME (SECONDS)4.4 DISCUSSION The results of this chapter establish that amino acid permeation oflipid bilayers is independent of the pH. The length of the lipid chain didhave a modest effect on permeability, with shorter chain lipids (DMPC) beingslightly more permeable, as might be expected. Lipid bilayers were observedto be much more permeable to hydrophobic amino acids under theconditions used here. Earlier research by Naoi et al. (1977) concluded thatlipid composition modulated amino acid permeability. However, rather thancomparing pure lipid systems of different chain lengths, those investigatorsused a variety of lipids mixed with egg phosphatidylcholine. Therefore longerchain lipids were always present and it was difficult to draw conclusionsabout the effect of varying hydrocarbon chain length on amino acidpermeability. Our results indicate that decreasing hydrocarbon chain lengthfrom 18 to 14 carbons increased amino acid permeation rates byapproximately 5-fold. This increase in permeability may also be attributed todifferences in chain kinking between the EPC and DMPC vesicles (presenceof one or more double bonds in the EPC vesicles) and the heterogeneity inchain length of the EPC vesicles.Previous research measured the permeation rates of certain modifiedamino acids and peptides in response to transmembrane pH gradients(Chakrabarti et al., 1992; see Chapter 3 for additional details). Lysinemethyl ester was used, because it would be neutral at higher pH values andpositively charged at lower pH values. Hence, the effect of ionic charge onpermeation rates could be determined. It was demonstrated that the neutralform of the modified amino acid and peptides was the form that wastranslocated across the bilayer. This translocation was very rapid (P = 2.1 X10-2 cm.s-1) for the lysine methyl ester (Section 3.4; Chakrabarti et al.,921992). In the work presented here, the transbilayer movement for theunaltered amino acid lysine is many orders of magnitude slower (average P= 5.1 X 10-12 cm.s-1) than that of the lysine methyl ester. Our results alsoindicate that amino acid flux is essentially independent of the pH, whichcontrasts with the dramatic pH effects observed for transport of lysinemethyl ester (Sections 3.3 and 3.4; Chakrabarti et al, 1992). This would beexpected for a zwitterion, which would remain charged regardless of pH.Previous research has established that the presence of charged groupsrestricts transbilayer diffusion, when compared to the neutral form(Gutknecht and Walter, 1981).The differences in permeability coefficients for the neutral, polar andcharged amino acids, together with the observation that transport isunaffected by pH, indicates that a mechanism other than partitioning islikely responsible for the observed efflux. It seems probable that the efflux ofamino acids occurs because of transient hydrated defects or "leaks" in themembrane (Hagle and Scott, 1978). These transient defects were firstdescribed by Nagle and Scott as cavities which allow small ions to enter andsubsequently pass through the bilayer without having to overcome the Bornenergy required for a charged molecule to "dissolve" in the low dielectricmembrane interior (Parsegian, 1969). This conclusion is supported by thefact that the permeability coefficients for monovalent cations such as Na+and K+ are in the range of 10-12-10-13 cm.s-1, remarkably similar to thosepresented here for several amino acids and phosphate. Transient defectshave been implicated as the mechanism by which these monovalent cationspermeate lipid bilayers (Deamer and Bramhall, 1986; Nagle and Scott,1978).93Our results confirm the observation by Naoi et al. (1977) thatpermeability is dependent on the structure of the amino acid . Theyobserved that bilayers were approximately 25-30 times more permeable tohydrophobic amino acids than charged amino acids such as lysine. Howevertheir technique involved the indirect determination of amino acid fluxthrough measurement of oxidation rates via the use of passively entrappedD-amino acid oxidase (Naoi et al., 1977). Results presented here using amore direct approach indicate that bilayer permeability to hydrophobicamino acids may be up to 100 times higher when compared with otheramino acids. The permeability differences between amino acids appears tobe related to differences in their partition coefficients between the lipid andwater phase. Table 4.3 gives previously published partition coefficient datafor most of the amino acids that were examined in this study (Leo et al.,1971). Both phenylalanine and tryptophan partition into the organic phase(roughly equivalent to the lipid bilayer) 10-100 times more readily than anyof the aliphatic or charged amino acids. It follows that the translocation ofhydrophobic amino acids may involve a modest amount of partitioning intothe membrane interior, as well as translocation through transient defects.For comparison, a truly neutral solute such as lysine methyl estertranslocates entirely by partitioning, and can therefore permeate at rates upto 108 faster than the hydrophobic amino acids. The latter must bringcharged groups into the hydrophobic phase and therefore still faces aformidable Born energy barrier.Another possible explanation for the permeability coefficientspresented here is that it is the neutral form of the amino acids that istranslocating across the bilayer. Only a very small proportion of the aminoacids present would be in the neutral form (the doubly charged [+/-] is in94Table 4.3. Partition coefficients for some of the amino acids studied (fromLeo et al., 1971).Amino Acid^ Log P Octanol Tryptophan -1.04Phenylalanine^ -1.43Lysine^ -2.82Glycine -2.26 (average)95equilibrium with the single charged forms [+ or -], which in turn are also inequilibrium with the neutral form). The small size of the neutral populationcould account for the much slower permeation rates observed whencompared to modified amino acids, such as lysine methyl ester (which arepredominantly in the neutral form at basic pH values; see Sections 3.3 and3.4 or Chakrabarti et al.. 1992). This possibility is currently underinvestigation.This work is also pertinent to the mechanism by whichpeptides/amino acids insert into and translocate across lipid bilayers invivo. Simple diffusion of amino acids is clearly too slow to permit sufficientprotein translocation rates to allow for cell growth. It follows that a differentmechanism must be in place so that potentially charged amino acid sidechains are able to overcome the Born energy barrier. One possibility is thattransmembrane pH gradients are involved, so that peptide signal sequenceswould be produced on the alkaline side of the membrane. These would thenbe able to permeate as the neutral form, with a net accumulation on theacidic side of the membrane (see Section 3.4 for additional details).9697CHAPTER 5 - INFLUENCE OF CHARGE, CHARGE DISTRIBUTION AND HYDROPHOBICITY ON THE TRANSPORT OF SHORT MODEL PEPTIDES INTO LIPOSOMES IN RESPONSE TO TRANSMEMBRANE pH GRADIENTS5.1 INTRODUCTION Translocation of weak acids and bases in response to transmembranepH gradients has been previously demonstrated for amine uptake inchloroplasts (Crofts, 1967) and in liposomal systems for fluorescent aminesused as pH indicators (Deamer et al., 1972), biogenic amines such asepinephrine (Schuldiner et al., 1978), various drugs (Madden et al., 1990)and acidic phospholipids (Hope et al., 1989; Redelmeier et al., 1990;Eastman et al., 1991). This phenomenon, which has been attributed torapid transbilayer movement of the neutral form of the weak acid or base,can result in large transbilayer concentration gradients. For example, a pHgradient (acidic interior) serves to produce a considerable net transport ofweak bases by trapping the neutral membrane-permeable species in theircharged (protonated) form after they traverse the bilayer. Thus, the weakbase accumulates in the acidic compartment of the liposome. Raising theexternal pH will increase the proportion of exterior molecules in the neutralform and hence the rate and extent of transport. It is straightforward toshow that lipophilic amines can be accumulated into LUVs with an acidicinterior to achieve inside/ outside concentration ratios which correspond tothe inside/outside concentration of protons.Previous work in this thesis has shown that the presence oftransmembrane pH gradients across lipid bilayers can cause the rapid andefficient transbilayer movement of amino acids and peptides in which thecarboxyl functions were modified to create amides or methyl esters98(Chakrabarti et al., 1992; Chapter 3). It was further established that thesecompounds permeated as the neutral species. In this chapter, these studieswere extended to specifically designed di- and tri-peptides composedexclusively of basic (lysine, Lys) and hydrophobic (tryptophan, Trp) aminoacids. Six peptides were synthesized to determine the influence of charge,charge distribution and hydrophobicity on transbilayer movement into LUVsystems. It is shown that while most of these peptides can be accumulatedinto EPC and EPC:cholesterol (55:45; mol %) LUVs with an acidic interior,the rates of transport are extremely sensitive to the number and location ofbasic functions within the molecule. Peptides with their charged groupsconcentrated at the N-terminal generally exhibited faster rates of transportand lower activation energies than peptides with their charged groups moreequally distributed throughout the molecule.995.2 MATERIALS AND METHODS Materials were obtained as outlined previously (see Section 3.2). Thepeptides used were kindly prepared in the lab of Dr. I. Clark-Lewis. Thestructures of the peptides employed are given in Figure 5.1, with thelocation of charged residues being indicated by "+" symbols. Vesicles wereproduced as outlined previously in Section 2.2.EPC:cholesterol (55:45; mol:mol) vesicles were created by dissolvingappropriate amounts of both compounds separately in chloroform, mixingthe solutions and drying down the resulting mixture using N2 gas followedby incubation under reduced pressure for several hours. The 100 nmvesicles were then prepared as described previously (Section 2.2).In order to generate the transmembrane pH gradient, the LUVs in thepH 4.0 media were passed down a 10 cm Sephadex G-50 (G50-150) columnpreviously equilibrated with 150 mM NaCl, 20 mM HEPES (pH 7.5) (HEPESbuffered saline, HBS). Tricine (pH 8.5 and 9.0) and MES (pH 5.5) were alsoused for some experiments as the external buffer at a concentration of 20mM. Uptake of the peptides was performed by first dissolving them in theFIBS media (1 ml), to which the LUVs (0.25 ml, final lipid concentration of 1to 5 mM) exhibiting a ApH (pHo = 7.5, pHi = 4.0; unless otherwise indicated)were added. Entrapment levels were monitored employing aliquots (0.1 ml)which were removed at selected times from this incubation mixture andpassed through 1.0 ml Sephadex G-50 columns (pre-spun) by centrifugationfor one minute at 2500 rpm to remove exterior (untrapped) material. Allexperiments were conducted at 20 °C unless otherwise indicated.The amount of peptide trapped was quantified by measurement oflysine concentrations using TNBS (trinitrobenzenesulfonic acid) (Hope andTryptophan-CH -COCYNH, •^ NH, •Lysine3--CH -COO-NH, •NH,- ---CH,IH211 H2^CHI 2^i H26H -CONN-- CH -CONH -CHINH3 •^ 600NH,Trp-Lys-Trp-AmideNH, • -- CH,CH,CH,C142CH -CONH-&--CONH -4HNH, *Lys-Trp-Trp-AmideNH, • --TH.,-CONH -4H-CONH -CHNH, •^ 600NH,Trp-Trp-Lys-AnaideFigure 5.1. Structures of the model peptides that were synthesized. The 4"sign indicates the presence of a positively charged amino group.100NH, • - CH,:11216142,6H,-CONH -NH, •COONH,Lys-Trp-Amide Trp-Lys-AmideNH, •^CH,12°NI6H -CONH-6H-CONHNIH3 •^ 4j.:00N112Lys-Trp-Lys-Amide101Culls, 1987) or by measuring tryptophan fluorescence (de Kroon et al.,1989) as outlined in Section 3.2.The magnitude of the pH gradients present were measured using[14C]methylamine as indicated elsewhere (Hope et al., 1985; Madden et al.,1990). The concentration used was l[tCi/ml. The amount of probeaccumulated was determined via liquid scintillation counting after removinguntrapped label. Transmembrane pH gradients could then be calculatedusing the relationship ApH = log {[methylamine]m/[methylamine]out} asindicated in Mayer et al. (1988).Phospholipid concentrations were determined by a modification of themethod of Fiske and Subbarow (1925). Typical phospholipid concentrationswere approximately 3 mM.Kinetic analysis of peptide translocation: Under the assumption that only the neutral form of the peptidetraverses the membrane (see, for example, Harrigan, 1992 or Harrigan etal., 1992b) it follows thatdipiotOt_ _p Am npiorrl _ iphM)dt^Vo[1]where [P]o" is the total exterior concentration of the peptide (includingcharged, uncharged, free and membrane-bound species), Pm is themembrane permeability coefficient of the neutral form, Am is the area of themembrane and [Plom and [P]im are the concentrations of the neutral formsof the peptide in the outer and inner monolayers, respectively.As shown elsewhere (Harrigan et al., 1992b) for a compoundcontaining one amino group. the concentration [P]0m in the outermonolayer can be written102^K ^1p1 tot^[2][H+]0where K is the membrane-water partition coefficient for the charged form ofthe peptide, Kd is the dissociation constant of the amino group and [H+10 isthe exterior proton concentration. From Eq. 1. we then obtain[P(t)]i = [P]ieci (1 - exp(-kt))^ [3]where the rate constant k can be written ask P^Kd^ [4]VoIn the case where the peptide has two amino groups it can be shownthatk _ P A niC Kdi_Kda^ [51Vo^[Fr-J0-2where K, Kd2 are the dissociation constants for each of the two aminogroups. This can be readily extended to a peptide containing three aminogroups whereK Kdl Kd2 K01,3^ [6]Vo^[HTToiThe rate constant (k) was determined by applying a linear least-squareanalysis to the uptake data using a commercially available plotting program(Sigma-Plot, Jandel Scientific, 1986).Determination of the membrane-water partition coefficient K': Partition coefficients were determined using an equilibrium filtercentrifugation technique employing a CentrifreeTM micropartition system(Amicon Div., W.R. Grace and Co., Danvers, MA). A 100 p.1 aliquot of vesicles(pH 7.5; 150 mM NaC1, 20 mM HEPES, referred to as HBS) was added to900 vtl of 2 mM peptide solution (pH 7.5, HBS) in the CentrifreeTMapparatus. The mixture was vortexed and allowed to stand for 15 minutes.It was then centrifuged for 10 minutes at 3000 r.p.m. The liposome-103associated ("supernatant") peptide solution and "free" peptide solution wereseparated and stored at 4°C until analyzed. Peptide concentrations for bothsolutions were obtained and phospholipid concentrations for the"supernatant" were calculated as outlined previously. The membrane-waterpartition coefficient (K') was then obtained using the relationK' = [Peptide]/[Peptide]0t^ ('71assuming an aqueous volume of 2.8 til/limole phospholipid and assumingthat 27% of volume is lipid (for 100 nm vesicles).pH titration of peptides:Titrations were performed in the lab of Dr. A.G. Mauk by M.R. Maukusing a Radiometer ABU93 Triburette equipped with three 1 ml burettesand a SAM90 sample station housed within an aluminum Faraday cage.The autoburette was computer-controlled for the titration and dataacquisition. For further details refer to Mauk et al. (1991).1045.3 RESULTS The first set of experiments was aimed at exploring the transportproperties of certain dipeptides with the same amino acid composition butdifferent sequence order. The two peptides synthesized were Lys-Trp-amideand Trp-Lys-amide. Incubation of these peptides at 20 °C with EPC LUVs(100 nm diameter) exhibiting a transbilayer pH gradient, inside acidic (pHo=7.5, pHi= 4.0), revealed markedly different transport phenomena. As shownin Fig. 5.2, the Lys-Trp-amide was accumulated effectively instantaneouslyinto the LUVs, whereas the Trp-Lys-amide exhibited much slower uptakekinetics. A corresponding decrease in the pH gradient, as measured by[14C]methylamine, was observed for the Lys-Trp-amide (the gradientdropped from 3.2 to 1.6 pH units). This decrease, which resulted fromprotonation of the amino functions of the accumulated Lys-Trp-amide, wasnot observed for the Trp-Lys-amide corresponding to the fact that little or nouptake was observed for this peptide under these conditions.Three approaches were used to reduce the rate of uptake of the Lys-Trp-amide so that the rate constants associated with the transbilayermovement of the two dipeptides could be compared. The first of thesemeasures involved lowering the external pH. As indicated in the Methods,the rate constant of translocation for a compound such as Lys-Trp-amide,which contains two amino functions, will decrease as the square of theexternal proton concentration for pH values less than the pK of the aminogroups. Thus, a decrease in the external pH from 7.5 to 5.5, for example,should decrease the rate of uptake by a factor of 104. However, even at anexternal pH of 5.5, the uptake of the Lys-Trp-amide was still too rapid toallow for an accurate kinetic analysis of rates (results not shown).105Figure 5.2. Time course of uptake of Lys-Trp-amide a and Trp-Lys-amide(0) into 100 nm EPC vesicles (3.7 mM) exhibiting a pH gradient (pHi= 4.0;pHo. 7.5). Uptake was conducted at 20 °C and the initial externalconcentrations of the peptides were 0.16 mM (Lys-Trp-amide) and 0.30 mM(Trp-Lys-amide). The residual pH gradients measured for the uptake of eachpeptide are represented by the filled symbols....-k.0I 5003.522..a. 40 II.c^.0a. LOco0 30 \.c0.^\•^•, ............ ............_______• 2Time (Hours)3.012.52.01.51.00.50.04106A second approach was to incorporate cholesterol into the LUVs,which would also be expected to reduce rates of transbilayer transport. Asshown in Fig. 5.3, the kinetics associated with the uptake of Lys-Trp-amideinto EPC:cholesterol (55:45; mol:mol) LUVs, at an exterior pH of 5.5 andtemperature of 21 °C can readily be measured. However, the uptake of theTrp-Lys-amide was much too slow to be measured under these conditions.In order to increase the rate of uptake of the Trp-Lys-amide for theEPC:cholesterol system increased exterior pH values and temperatures wereemployed. As shown previously (Chakrabarti et al., 1992), the rate oftranslocation of a similar molecule (lysine methyl ester) can be dramaticallyaffected by changes in temperature. As shown in Fig. 5.3, at an exterior pHof 8.0 and temperature of 55 °C the kinetics associated with the uptake ofTrp-Lys-amide into EPC:cholesterol (55:45; mol:mol) vesicles could bereadily determined.It is of interest to further characterize the temperature dependence ofthe uptake of these dipeptides. Related molecules can exhibit highactivation energies, in the range of 31-36 kcal/mol (Chakrabarti et al.,1992). As shown in Figure 5.4A, the uptake of Lys-Trp-amide increasedconsiderably as the temperature was raised from 9.5°C to 30.5°C for anEPC:cholesterol (55:45; mol:mol) LUV system where pHo= 5.5 (20 mM MESand 150 mM NaCl) and pHi= 3.0. An analysis of the rate constants (Fig.5.4B) and subsequent calculation of the activation energy from theArrhenius plot of Fig. 5.4C revealed an activation energy of 23.7 kcal/mol. Aset of data analogous to that obtained for Lys-Trp-amide was also generatedfor Trp-Lys-amide in an EPC:cholesterol (55:45; mol:mol) system, wherepHo= 8.0 (20 mM HEPES and 150 mM NaC1) and pHi= 4.0, over thetemperature range of 40-55°C (Fig. 5.5). The resulting activation energy3025201510500.0 1.0^2.0^3.0^4.0Time (Hours)6.05.0107Figure 5.3. Time course of uptake of Lys-Trp-amide 0 and Trp-Lys-amide(0) into 100 nm EPC:cholesterol (55:45; mol:mol; 1.7 mM for the Lys-Trp-amide and 4.9 mM for the Trp-Lys-amide) exhibiting a pH gradient (pHi=3.0, pHo= 5.5, 21°C for the Lys-Trp-amide; pHi= 4.0, p1-10= 8.0, 55°C forthe Trp-Lys-amide). The initial external peptide concentration was 0.43 mM.108Figure 5.4A. Time course of uptake of Lys-Trp-amide into 100 nmEPC:Cholesterol vesicles (55:45; mol:mol; 1.7 mM) bearing a 5.5/3.0(external/internal) pH gradient. Uptake was conducted at 9.5 °C (0), 15 °C(A), 21 °C (0), 26 °C (V) and 30.5 °C (0 ). The external concentration ofLys-Trp-amide was 0.43 mM.5.4B. Plot of In {([A(eq)]i - [A(t)]0/[A(eq)h} versus t; where [A(t)l is the interiorconcentration of the accumulated amine at time t and (A(eq)li is the interiorconcentration at equilibrium.5.4C. Arrhenius plot of the rate constants (k) for Lys-Trp-amide uptake. Theactivation energy calculated from the slope of this plot is Ea = 23.7kcal/mol.0.5^1.0^1.5^2.0^2.5Time (Hours)0.0—0.5—1.03025201510'.77 I—1.5-- 2.0^I"<c—a.)^—2.5—3.0C—3.5——4.0—6.0—7.0In k —8.0• —9.00 1 2Time (Hours)—10.03.25 3.35^3.451/T (X 10 —3)3.55109110Figure 5.5A. Time course of uptake of Trp-Lys-amide into 100 nmEPC:Cholesterol vesicles (55:45; mol:mol; 4.9 mM) bearing a 8.0/4.0(external/internal) pH gradient. Uptake was conducted at 40 °C (0), 45 °C(A), 50 °C (0) and 55 °C (v). The external concentration of Trp-Lys-amidewas 0.42 mM.5.5B. Plot of in {UA(eq)li - [A(t)]0/A(eq)J1l versus t; where [A(t)11 and [A(eq)lihave the same meanings as indicated in the legend to Fig. 5.4B.5.5C. Arrhenius plot of the rate constants (k) for Trp-Lys-amide uptake. Theactivation energy calculated from the slope of this plot is Ea = 29.0kcal/mol.61^2^3^4Time (Hours)000^1.0^2.0^3.0Time (Hours5.04.0 60.••••••■0.0—0.2—0.6—rm.(1.) —0.8 —- 1.■JCD —1.0--1.2--1.4--1.5 ^0In k3.20—12.0 ^3.05 3.10^3.151/T (X 10 —3 )111112calculated for Trp-Lys-amide was 29.0 kcal/mol (Fig. 5.5C).Subsequent experiments revealed that the activation energy oftranslocation for Trp-Lys-amide in EPC vesicles was 26.8 kcal/mol at pH7.5 (Table 5.1). It should also be noted that the activation energies weresomewhat dependent upon the lipid system used, with EPC:cholesterol(55:45; mol:mol) vesicles giving higher activation energies than EPC alone(Table 5.1).As indicated under Methods, the rate of transport should beproportional to the membrane-water partition coefficient of the peptide. Inorder to determine whether such partitioning affects could explain theextremely large differences in uptake rates, partition coefficients weredetermined for Lys-Trp-amide (74.6 for EPC) and Trp-Lys-amide (85.5 forEPC) employing the centrifugation method outlined in Materials andMethods (see Table 5.1). The similarity of these values indicates thatpartitioning is not responsible for the differences observed in uptake rates.Another possibility was that the proximity of the amino groups in theLys-Trp-amide might result in alterations of their pK values. This, in turn,could account for the differences observed in uptake rates. In order todetermine whether the pK values of the Lys-Trp-amide were significantlydifferent from those of Trp-Lys-amide, pH titrations were performed on thesetwo peptides. The titration curves for both peptides were essentiallyidentical (data not shown). The pK of the NH2 terminus amino function wasfound to be 7.4, while that of the lysine side was 10.4 for both peptides.In order to further characterize the influence of charge distribution onthe transbilayer movement of peptides, a series of tripeptides weresynthesized based upon the Lys/Trp combinations. The peptides that weresynthesized were Lys-Trp-Lys-amide, Trp-Lys-Trp-amide, Lys-Trp-Trp-amideTable 5.1. Partition Coefficients (K') and Activation Energies (Ea) for thepeptidesPeptide K' Ea (kcal/mol)Lys-Trp-amide 74.6 (23.7)Trp-Lys-amide 85.5 26.8 (29.0)Lys-Trp-Trp-amide 105.9 19.6Trp-Trp-Lys-amide 111.0 33.4Trp-Lys-Trp-amide 84.1 14.9Partition coefficients and activation energies were calculated as described inthe Materials and Methods. Activation energies given in brackets are thosecalculated for EPC:cholesterol (55:45; mol:mol) vesicles. The activationenergy for the Lys-Trp-amide could not be determined in EPC LUVs (seeResults).113200 ^180—160—140—120—100—80—60—40—A20—SO^00^0.50ATime (Hours)1.0^1.5^2.0^2.5^3.0114Figure 5.6. Time course of uptake of Lys-Trp-Trp-amide (0), Trp-Trp-Lys-amide (A), Trp-Lys-Trp-amide (0) and Lys-Trp-Lys-amide (0) into 100 nmEPC vesicles (4.1 mM) exhibiting a pH gradient (pHi= 5.0; pHo= 8.5; exceptfor the Lys-Trp-Lys-amide where pHo= 10.0). Uptake was conducted at 45°C and the initial external peptide concentration was 1.67 mM.115and Trp-Trp-Lys-amide (Fig. 5.1). The ability of these peptides to beaccumulated at 45 °C into EPC LUVs (100 nm diameter) experiencing atransbilayer pH gradient, inside acidic (pHo= 8.5; pHi= 5.0), were quitevaried (Fig. 5.6). The Trp-Lys-Trp-amide (k = 2.93 X 10-4 s"1) and Lys-Trp-Trp-amide (k = 2.83 X 10-4 s"1) exhibited the most rapid uptake followed bythe Trp-Trp-Lys-amide (k = 1.89 X 10"4 s" IL). The Lys-Trp-Lys-amide showedno appreciable accumulation, even after several hours at high temperature(50 °C) and high exterior pH (pHo = 10.0; Fig. 5.6).The three combinations of one lysine and two tryptophan residuesrevealed significantly different rates of accumulation, depending upon thedistribution of the two charged residues (amino terminus and lysine sidechain) within the molecule. Rate constants for the translocation of thesethree peptides were derived and activation energies associated with uptakeinto EPC systems were calculated from the resulting Arrhenius plots (Table5.1). The activation energies for the transbilayer movement of these peptidesranged from 14.9 kcal/mol to 33.4 kcal/mol (Table 5.1), remarkabledifferences for molecules that were composed of the same three aminoacids. Partition coefficients were also determined for these three moleculesand were found to range from 111.0 for Trp-Trp-Lys-amide to 105.9 for Lys-Trp-Trp-amide and 84.1 for Trp-Lys-Trp-amide (Table 5.1).1165.4 DISCUSSIONThe results of this report establish that basic peptides of identicalamino acid composition which vary only in the sequence of the amino acidspresent can show markedly different abilities to permeate throughmembranes under identical conditions of temperature and pH. The primaryinterest of this work concerns the influence of charge, charge distributionand the degree of hydrophobicity of the peptides on the transbilayermovement of these molecules.As indicated in Section 5.1 (Materials and Methods), the ability ofweak bases to translocate across lipid bilayers, which relies on permeationof the neutral form and can result in net accumulation into LUVs with anacidic interior, is strongly dependent on the external pH. This dependence,together with the activation energy (Ea) associated with permeation of theneutral form can be expressed in a generalized rate equation for thepeptides containing two amino functions according to-Ea/RTI (TI/T - 1) 2(pH - pHI)k (T, pH) = k(TI, pHi) exp^10^[8]where k (T1, pHI) is the rate constant measured at a particular temperature(T1) and pH (pH1).Employing the data of Fig. 5.3 for EPC:cholesterol LUVs we note thatkKw (21 °C, 5.5) = 5.8 X 10-4 s-1, whereas kwK (55 °C, 8.0) = 6.4 X 10-5 s-1. where the subscripts Kw and wK indicates Lys-Trp-amide and Trp-Lys-amide, respectively. Thus, given the activation energies EaKw = 23.7kcal/mol and Ea'K = 29.0 kcal/mol, we obtain the generalized rateequations for the Lys-Trp-amide and the Trp-Lys-amide dipeptides askKw (T, pH) = 5.8 X 10-4 X 102(PH-5.5) exp {40.6 (294/T) -^1)} [9]kwK (T. pH) = 6.4 X 10-5 X 102(PH-8.0) exp {44.5 (328/T) -^1)} [10]117Thus at 20°C (pH = 7.0) kKw = 0.67 s" ft1 • - 1 /2 = 1 s), whereas kwK = 1.3 X10"4 s"1 (t1/2 = 1.5 h). The very large difference in rate constants (the ratiois 5.2 X 103) cannot be attributed to changes in the partition coefficientsbecause the partition coefficient of the Lys-Trp-amide is actually slightlylower than that of the Trp-Lys-amide. This difference also cannot be causedby changes in the pK values of the amino groups of the peptides since thesewere identical. Instead other unidentified factors, including perhapsconformational differences between the two peptides, must be responsiblefor the significant differences observed in rate constants.The activation energies observed for the uptake of the variouspeptides studied here are of interest. Activation energies associated with thetransfer of molecules from aqueous media into the membrane have beenestimated according to the number of hydrogen bonds that must be brokento enter the membrane less the number created once inside (Stein, 1967;Walter and Gutknecht; 1986). Calculation of activation energies in thismanner becomes more difficult as the size of the molecule increases, giventhe problem of estimating formation of intramolecular hydrogen bonds onceinside the membrane. Previous work has shown that lysine methyl esterexhibited a high activation energy in the range of 36 kcal/mole (see Section3.3; Chakrabarti et al.. 1992). Addition of a tryptophan residue to the lysinemolecule appears to lower the activation energy by at least 9 kcal/mole(Table 5.1; Sections 3.3 and 3.4; Chakrabarti et al., 1992), which may arisedue to the increased hydrophobicity of the resulting molecule. In the workpresented here on di- and tripeptides (composed of one lysine and twotryptophan residues), we show that peptides of identical amino acidcomposition can have different activation energies depending upon thedistribution of charges within the molecule (Table 5.1), indicating that118amino acid composition per se is insufficient to accurately predict activationenergies for transbilayer translocation. Depending on the location of thelysine residue (at the amino terminus, in the middle or at the amide-modified carboxyl terminus), the activation energies for the transbilayermovement of the tripeptides range from 15-33 kcal/mole (Table 5.1). Thepeptides with the lowest activation energies had the lysine residue in the N-terminal or middle position (Table 5.1). Again, the concentration of chargedresidues away from the modified end of the molecule (the amide-modifiedcarboxyl group) creates a more hydrophobic C-terminus for both peptides.This amphipathic quality is presumably related to the lowered activationenergies observed when compared to the Trp-Trp-Lys-amide peptide (seeTable 5.1).The negligible transport of the Lys-Trp-Lys-amide may be attributedto the hydrophilic nature of this molecule, as well as the improbability of theformation of the fully deprotonated net neutral form. Use of the Kyte-Doolittle hydrophobicity scales (where positive values indicatehydrophobicity and negative values indicate hydrophilicity) to determine thedegree of hydrophobicity of the peptides studied in this work indicated thatthe dipeptides (Lys-Trp- or Trp-Lys-amides) were hydrophilic molecules (-4.8on the Kyte-Doolittle scale); while the tripeptides composed of one lysineand two tryptophan residues were slightly more hydrophilic on average (-5.7on the Kyte-Doolittle scale; Branden and Tooze. 1991). However, thetripeptides composed of two lysines and one tryptophan are very hydrophilicmolecules (-8.7 on the Kyte-Doolittle scale). Further, Lys-Trp-Lys-amidepossesses 3 amino functions, which all must be deprotonated in order forthe molecule to translocate across the bilayer in a neutral form. Previouswork has shown that it is the neutral form of weak acids and bases thatpermeate lipid bilayers and that the presence of a single charged residuecan result in permeabilities up to 1010 slower than that observed for theneutral species (see Chapter 4 for additional details; Gutknecht and Walter,1981; Chakrabarti et al., 1992; Chakrabarti and Deamer, 1992). The rate oftransport of Lys-Trp-Lys-amide would decrease corresponding to the inverseof the cube of the external proton concentration at pH values less than thepK of the amino groups (pH 7.4 for the NH2 terminus and pH 10.4 for theside chain of lysine). Hence, the negligible transport of the Lys-Trp-Lys-amide can also be attributed to the very small fraction in the neutral form atpH values of 9 or lower.A final point of interest concerns the implications that this work hasregarding the mechanisms by which peptides insert into and translocateacross membranes in vivo. Charged amino acids, such as lysine, have beenimplicated in the localization of secreted and membrane proteins (Boyd andBeckwith, 1990). This localization may be due to interactions of the basicamino acids with the acidic head groups of phospholipids or due tointeractions with the electrochemical gradient across membranes (Boyd andBeckwith, 1990; Skerjanc, 1990). Previous work (Chapters 3 and 4;Chakrabarti et al., 1992) and the work presented here indicate thattransmembrane pH gradients could drive the transbilayer movement andanchoring of peptides or signal sequences containing lysine residues.In summary, this investigation shows that the ability of peptides topartition into and permeate through lipid bilayers is determined not only bythe peptide charge and amino acid composition but also by the sequence ofthe amino acids that are present. Peptides that have their basic functionssequestered toward one end of the molecule exhibit dramatically improved119abilities to translocate across membranes. This ability also correlates tosome extent with lower activation energies associated with transport.120CHAPTER 6. SUMMARYThis thesis has investigated the transport of metal ions and modified(weak base) amino acids and peptides into liposomes (model membranesystems) in response to transmembrane pH gradients. Studies were alsoperformed to examine the permeability of unaltered (zwitterionic) aminoacids to lipid bilayers. A general theme of this thesis has concerned theeffects of rapid transbilayer movement of neutral entities, be they cation-ionophore complexes or weak bases in the neutral form.It was shown in Chapter 2 that under appropriate conditions oftemperature, internal buffering capacity and Ca+2-ionophore A23187concentration, iron (primarily in the form of Fe+2) and barium could beaccumulated into EPC and DSPC-cholesterol (55:45 mole %) LUVs inresponse to a transmembrane pH gradient. The DSPC-cholesterol LUVsystems exhibited superior retention properties. It is shown that themaximally loaded Fe- and Ba-containing LUVs exhibit increased densities inthat a large fraction of the maximally loaded LUVs can be pelleted by lowspeed centrifugation and the Ba+2-loaded systems can be directly visualizedby cryo-electron microscopy. The significance of this work lies in thepotential utility of liposomes loaded with metal ions. The two specificapplications studied, creating dense and electron dense liposomes, weresuccessful to varying degrees. Building upon the system outlined here mayresult in the creation of a more homogeneous population of denseliposomes, which would be important for their use in separation andvisualization procedures.Chapter 3 involved studying the uptake of derivatives of lysine and apentapeptide (Ala-Met-Leu-Trp-Ala), in which the C-terminal carboxyl121122functions have been converted to methyl esters or amides, in response totransmembrane pH gradients in LUV systems. It was shown that thepresence of a pH gradient (interior acidic) results in the rapid and efficientaccumulation of these weakly basic amino acid and peptide derivatives intoLUVs in a manner consistent with permeation of the neutral (deprotonated)form. The significance of this work is that this property may have generalimplications for the mechanism of transbilayer translocation of peptides,such as signal sequences, which exhibit weak base characteristics. Theprecise mechanism by which certain proteins permeate lipid bilayers (in thepresence or absence of membrane proteins) is still unclear. The workpresented here indicates that for peptides which are weakly basic (1 or 2charged groups), the neutral form of the molecule may be able to translocateacross the bilayer in response to transmembrane pH gradients. Furtherwork will be required in order to discover whether signal sequences andother larger peptides will behave in a similar manner to the amino acids andshort peptides studied so far.Chapter 4 dealt with determining the permeability coefficients forneutral, hydrophobic, polar and charged amino acids. The rates of efflux ofglycine, lysine, phenylalanine, serine and tryptophan were determined afterthey were passively entrapped EPC and DMPC LUVs. The permeabilitycoefficients were approximately 10-12 cm s"1 for the EPC vesicles, whilethose for the DMPC vesicles were approximately 10-11 cm s-1. Decreasinglipid chain length increased permeability slightly, while variations in pH hadonly minor effects on the permeability coefficients. The hydrophobic aminoacids were 10-100 times more permeable than any of the other forms,corresponding to their increased octanol:water partition coefficients.Permeation of the hydrophobic amino acids may have been enhanced by123partitioning into the hydrocarbon region of the bilayer prior to passageacoss the membrane. These observations suggest that permeation rates forthe amino acids studied may involve partitioning into the membrane andpassage across, either in the charged or neutral form, as observed for themodified amino acids and peptides studied in Chapter 3. The small size ofthe neutral population present or the Born energy barrier faced by acharged molecule translocating across a membrane could both account forthe slow permeation rates observed. Alternatively, permeation rates may becontrolled by bilayer fluctuations and transient defects. The differences inpermeability coefficients between a charged molecule, lysine (10-12 cmand the same molecule that was able to translocate in the neutral form,lysine methyl ester (10-2 cm s-1), support these conclusions.Chapter 5 examined the influence of hydrophobicity and chargedistribution on the uptake of modified peptides. Here, the ability of small (2-3 amino acid) peptides, composed exclusively of basic (lysine) andhydrophobic (tryptophan) amino acids, to accumulate into LUV systems wasinvestigated. Rates of transport, activation energies for transbilayermovement and partition coefficients were calculated for the followingpeptides: Trp-Lys-Amide; Lys-Trp-Amide; Trp-Lys-Trp-Amide; Trp-Trp-Lys-Amide; and Lys-Trp-Trp-Amide. The peptide Lys-Trp-Lys-Amide was alsosynthesized but was only accumulated to a slight degree under extremeconditions of temperature and pH. The sequence of the amino acids inpeptides of identical amino acid composition was observed to have profoundeffects on the rates of transport and activation energies, but only minoreffects on the partition coefficients. Peptides with their charged groupsconcentrated at the N-terminal of the molecule generally exhibited muchfaster rates of accumulation than those with their charged groups more124evenly dispersed. The creation of a hydrophobic end at the C-terminal of thepeptides may therefore serve to enhance permeation through the bilayer,resulting in increased transport rates. This work thus builds upon theobservations and conclusions made in Chapter 3. The significance of thiswork lies in the observation that charge and charge distribution can greatlyinfluence the transbilayer movement of small peptides in response totransmembrane pH gradients. An understanding of these parameters mayprove useful for deciding why certain peptides permeate lipid bilayers, whileothers permeate at very different rates or not at all.There exist several areas of investigation that could lead from theresearch presented in this thesis. Metal ions of higher atomic weights thaniron or barium (such as lanthanum) could be loaded to create liposomeswith the potential for enhanced separation and visualization. Metalradioisotopes (such as 67Ga3+) could also be loaded by means of thecombination of ionophores and transmembrane pH gradients presentedhere to create vesicles with very high specific activities for use inbiodistribution studies and possible therapeutic applications.In the case of amino acid permeation, there is still some doubt as tothe actual mechanism whereby unmodified amino acids permeate lipidbilayers (Chapter 4). The possibility exists that it is the neutral form of theamino acids that is translocating across the bilayer (see Section 4.4). Byemploying a pH titration experiment analogous to the type presented inChapter 3 (see Fig. 3.2), it may be possible to discern efflux of the neutralsub-population for certain amino acids, such as histidine. However, thesensitivity of the techniques available to monitor transbilayer flux of aminoacids is insufficient to allow for detection of efflux of the putative smallneutral population present.125Other future experiments pertain to the work begun in Chapter 3,followed by Chapter 5; namely, understanding the process of peptidetranslocation across lipid bilayers in response to transmembrane pHgradients. Establishing an explanation for the vastly different rates ofuptake observed for Lys-Trp-amide versus Trp-Lys-amide is quite important.Other as yet unidentified factors, such as subtle differences in theconformation of these two peptides, may account for some of the differencesobserved. Research into how factors such as charge and hydrophobicityaffect the transport of longer peptides will be of interest. Synthesis of signalsequences or membrane-spanning peptides and subsequent transportstudies would be of particular interest. Transport of a signal sequence intoLUVs in response to transmembrane pH gradients would go a long waytowards proving the suggestion presented in Chapter 3 that such gradientsmay play a role in protein translocation in vivo. Studies of membrane-spanning peptides (approximately 20-24 residues in length, with chargedgroups on either end) would allow for experiments examining the orientationof the peptide in the bilayer, including potentially "flipping" the membrane-spanning peptide by changing the pH gradient present. Other variations onthis line of research could involve the synthesis of even longer peptides thatcould "loop" multiple times into and out of the bilayer (see Fig. 1.10, forexample).CHAPTER 7. REFERENCESAddanki, S., Cahill, F.D. and Sotos, J.F. (1968) J. Biol. Chem. 243, 2337-2343.Alberts, B., Bray, D., Lewis, J., Raff, M., Roberts, K. and Watson, J.D.(1989) Molecular Biology of the Cell (2nd edition) Garland PublishingInc., New York, pp. 300-340.Barenholz, Y., Amselem, S. and Lichtenberg, D. (1979) FEBS Lett. 99, 210-213Blobel, G. and Dobberstein, B. (1975) J. Cell Biol. 67, 835-851.Boyd, D. and Beckwith, J. (1990) Cell 62, 1031-1033.Branden, C. and Tooze, J. (1991) in Introduction to Protein Structure, pp.201-215, Garland Publishers Inc., New York.Bonte, F. and Juliano, R.L. (1986) Chem. Phys. Lipids 40, 359-372.Cafiso, D.S. and Hubbell, W.L. (1978a) Biochemistry 17, 187-195.Cafiso, D.S. and Hubbell, W.L. (1978b) Biochemistry 17, 3871-3788.Cafiso, D.S. and Hubbell, W.L. (1983) Biophys. J. 44, 49-57.Carruthers, A. and Melchior, D.L. (1983) Biochemistry 22, 5797-5807.Chakrabarti, A.C., Clark-Lewis, I., Harrigan, P.R. and Cullis, P.R. (1992)Biophys. J. 61, 228-234.Chakrabarti, A.C. and Deamer, D.W. (1992) Biochim. Biophys. Acta, inpress.Chapman, D. (1975) Q. Rev. Biophys. 8, 185-235.Chappel, J.B. and Crofts, A.R. (1966) in Regulation of Metabolic Processes inMitochondria (Tager, J.M., Papa, S., Quagliariello, E. and Slater, E.C.,editors) pp. 293-316, Elsevier Publishing Co., Amsterdam.126127Clark-Lewis, I. and Kent, S. (1989) in Receptor Biochemistry andMethodology: The use of HPLC in protein purification andcharacterization (Kerlavage, A.R., Ed.) pp 43-93, A.R. Liss Inc., NewYork.Cohen, B.E. and Bangham, A.D. (1972) Nature 236, 173-174.Crofts, A.R. (1967) J. Biol. Chem. 242, 3352-3359.Cullis, P.R. and de Kruijff, B. (1979) Biochim. Biophys. Acta 559, 399-420.Cullis, P.R. and Hope, M.J. (1985) in Biochemistry of Lipids and Membranes(Vance, D.E. and Vance, G.E., ed) Benjamin/Cummings PublishingCo., Menlo Park, CA, pp. 25-70.Cullis, P.R., Mayer L.D., Bally. M.B., Madden, T.D., and Hope, M.J. (1989)Adv Drug Del. Rev. 3, 267-282.Culls, P.R., Bally, M.B., Madden, T.D., Mayer L.D. and Hope, M.J. (1991)Trends in Biotech. 9, 268-272.Darnell, J., Lodish, H. and Baltimore, D. (1986) Molecular Cell Biology,Scientific American Books, New York, pp. 911-983.de Kroon, A. I. P. M., de Gier, J. and, de Kruijff, B. (1989) Biochim. Biophys.Acta. 981, 371-373.de Kroon, A. I. P. M., Soekarjo, J., de Gier, J. and, de Kruijff, B. (1990)Biochemistry 29, 8229-8240.Deamer, D,W., Prince, R.C. and Croft, A.R. (1972) Biochim. Biophys. Acta274, 323-335.Deamer, D.W. and Bangham, A.D. (1976) Biochim. Biophys. Acta 443, 629-634.Deamer, D.W. and Bramhall, J. (1986) Chem. Phys. Lipids 40, 167-188.Dilger, J.P. and McLaughlin, S. (1979) Science 206, 1196-1198.Driessen, A.J.M. (1992) Trends in Biochem. Sci. 17, 219-223.128Eastman, S.J., Hope, M.J. and Culls, P.R. (1991) Biochemistry 30, 1740-1745.Eastman, S.J., Hope, M.J., Wong, K.F. and Cullis, P.R. (1992) Biochemistry31, 4262-4268.Epand, R.M. (1990) Biochem. Cell Biol. 68, 17-23.Fettiplace, R., Gordon, L.G.M., Hladky, S.B., Requena, J.. Zingshem, H.P.and Haydon, P.A. (1974) in Methods in Membr. Biology, Vol. 4 (Korn,E.D., ed.) New York, pp. 1-75.Finkelstein, A. and Cass, A. (1968) J. Gen Physiol. 52, 145s-172s.Fiske, C.H. and Subbarow, Y. (1925) J. Biol. Chem. 66, 375-400.Flewelling, R.F. and Hubbell, W.L. (1986a) Biophys. J. 49, 531-540.Flewelling, R.F. and Hubbell, W.L. (1986b) Biophys. J. 49, 541-552.Frederik, P.M., Stuart, M.C.A., Bomans, P.H.H., Busing, W.M., Burger,K.N.J. and Verkleij, A.J. (1991) J. Microsc. 161, 253-262.Gennis, R.B. (1989) Biomembranes, Molecular Structure and Function,Springer-Verlag, New York.Gierasch, L.M. (1989) Biochemistry 28, 923-930.Gutnecht, J. and Walter, A. (1981) Biochem. Biophys. Acta 649, 149-154.Hargreaves, W.R. and Deamer, D.W. (1978) Biochemistry 17, 3759-3768.Harrigan, P.R. (1992) Transmembrane pH gradients in liposomes: drug-vesicle interactions and proton flux. Ph.D. Thesis, The University ofBiritish Columbia.Harrigan, P.R., Hope, M.J., Redelmeier, T.E. and Cullis, P.R. (1992a)Biophys. J., in press.Harrigan, P.R., Wong, K.F., Redelmeier, T.E., Wheeler, J.J. and Culls, P.R.(1992b) submitted to Biochemistry.129Hope, M.J., Bally, M.B., Webb, G., and Cullis, P.R. (1985) Biochim. Biophys.Acta 812, 55-65.Hope, M.J., Bally, M.B., Mayer, L.D., Janoff, A.S. and Cullis, P.R. (1986)Chem. Phys. Lipids 40, 89-108.Hope, M. J. and Cullis, P.R. (1987) J. Biol. Chem. 262, 4360-4366.Hope, M. J., Redelmeier,T. E., Wong, K. F., Rodrigueza, W. and Cull's, P. R.(1989) Biochemistry. 28, 4181-4187.Hong, K., Freind, D.S., Glabe, C.G. and Papahadjopoulos, D. (1983)Biochim. Biophys. Acta 732, 320-323.Houslay, M.D. and Stanley, K.K. (1982) Dynamics of Biological Membranes,John Wiley and Sons. Toronto.Hwang, K.J. and Mauk, M.R. (1978) Proc. Natl. Acad. Sci. USA 74, 4991-4995.Hwang, K.J., Merriam, J.E., Beaumier, P.L. and Luk, K.S. (1982) Biochim.Biophys. Acta 812, 101-109.Jacobs, M.H. (1940) Cold Spring Harbour Symp. Quant. Biol. 8, 30-39.Kleinfeld, A.M. (1987) in Current Topics in Memebranes and Transport, Vol.29, Academic Press Inc., pp.1-27.Kremer, J.M.H., Esker, M.W.J., Pathmamanohovan, C. and Wiersema, P.H.(1977) Biochemistry 16, 3932-3935.Leo, A., Hansch, C. and Elkins, D. (1971) Chem. Reviews 71, 525-616.Ma, L.D., Magin. R.L. and Dunn, F. (1990) Biochim. Biophys. Acta 1022, 17-26.Madden, T.D. (1986) Chem. Phys. Lipids 40, 207-226.Madden, T. D., Harrigan, P. R., Tai, L. C. L.. Bally, M. B., Mayer, L. D.,Redelmeier, T. E., Loughrey, H. C., Tilcock, C.P.S., Reinish, L. W. andCullis, P. R. (1990) Chem. Phys. Lipids. 53, 37-46.130Mathews, C.K. and van Ho1de, K.E. (1990) Biochemistry,Benjamin/Cummings Publishing Company, Inc., Don Mills, Ontario,Chapter 9.Mauk, M.R. and Gamble, R.C. (1979) Anal. Biochem. 94, 302-307.Mauk, M.R., Barker, P.D. and Mauk, A.G. (1991) Biochemistry 30, 9873-9881.Mayer, L.D., Hope, M.J., Cullis, P.R. and Janoff. A.S. (1985) Biochim.Biophys. Acta 817, 193-196.Mayer, L.D., Bally, M.B. and Cullis, P.R., (1986a) Biochirn. Biophys. Acta857, 123-126.Mayer, L.D., Hope. M.J. and Cullis, P.R., (1986b) Biochim. Biophys. Acta858, 161-168.Mayer, L.D., Bally, M.B., Hope, M.J. and Cullis, P.R. (1986c) Chem. Phys.Lipids 40, 333-345.Mayer, L.D., Wong, K.F., Menon, K., Chong, C., Harrigan, R.H. and Cullis,P.R. (1988) Biochemistry 27, 2053-2060.Mayer, L.D., Tai, L.C.L., Ko, D.S.C., Masin, D., Ginsberg, R.S., Cullis, P.R.and Bally, M.B. (1989) Cancer Res. 49, 5922-5930.McElhaney, R.N. (1986) Biochem. Cell Biol. 64, 58-65.Mimms, L.T., Guido. Z., Nozaki, Y., Tanford, C. and Reynolds, J.A. (1981)Biochemistry 20, 833-840.Nagle, J.F. and Scott, H.L. (1978) Biochim. Biophys. Acta 513, 236-243.Naoi, M., Naoi, M.. Shimizu, T., Malviya, A.N. and Yagi, K. (1977) Biochim.Biophys. Acta 471. 305-310.Nichols, J.W. and Deamer, D.W. (1976) Biochirn. Biophys. Acta 455, 269-271.131Olson, F., Hunt, C.A., Szoka, E.C., Vail, W.J. and Papahadjopoulos, D.(1979) Biochim. Biophys. Acta 557, 9-23.Op den Kamp, J.A.F. (1979) Ann. Rev. Biochem. 48, 47-71.Ostro, M.J. and Cullis, P.R. (1989) Am. J. Hosp. Pharm. 46, 1576-1587.Padan, E., Zilberstein, D. and Rottenburg, H. (1976) Eur. J. Biochem. 63,533-540.Parker, M.W., Tucker, A.D., Tsernoglou, D. and Pattus, F. (1990) TrendsBloc/tern. Sci. 15, 126-129.Parsegian, V.A. (1969) Nature 221, 844-846.Perkins, W.R., and Cafiso, D.S. (1986) Biochemistry 25, 2270-2276.Pfeiffer, D.R., Reed, P.W. and Lardy, H.A. (1974) Biochemistry, 13, 4007-4014.Pick, U. (1981) Arch. Biochem. Biophys. 212, 186-194.Pippard, M.J. and Stray, S. (1982) Amer. J. Clin. Path. 77, 324-327.Pressman, B.C. (1973) Federation Proc. 32, 1698-1703.Racker, E. (1973) J. Menibr. Biol. 10, 221-233.Racker, E. (1985) Reconstitution of Transporters, Receptors and PathologicalStates, Academic Press, Orlando, FloridaRapoport, T.A. (1990) Trends Biochem. Sci. 15, 355-358.Redelmeier, T. E.. Hope, M. J. and Cullis, P. R. (1990) Biochemistry. 29,3046-3053.Rees-Jones, R. and Al-Awqati. Q. (1984) Biochemistry 23, 2236-2240.Reithmeier, R.A.F. (1985) in Biochemistry of Lipids and Membranes (Vance,D.E. and Vance. G.E., ed) Benjamin/Cummings Publishing Co.,Menlo Park, CA, pp. 503-558.Rottenberg, H., Grunwald, T. and Avron, M. (1972) Eur. J. Biochem. 25, 54-63.132Rottenberg, H. and Grunwald, T. (1972) Eur. J. Biochem. 25, 71-74.Rottenberg, H. (1979) Methods Enzymol. 55, 547-569.Schuldiner, S. Fishkes, H. and Kanner, B.T. (1978) Proc. Natl. Acad. Sci.75, 3713-3716.Shastri, B.P., Sankaram, M.B. and Easwaran, K.R.K. (1987) Biochemistry26, 4925-4930.Singer, S.J. and Nicholson, G.L. (1972) Science 175, 720-731.Skerjanc, I. (1990) Biochem. Cell Biol. 68, 9-16.Stein, W.D. (1967) in The Movement of Molecules Across Cell Membranespp.73-90, Academic Press, New York.Stryer, L. (1988) Biochemistry, W.H. Freeman and Co., New York, Chapter12.Szoka, F. and Papahadjopoulos, D. (1978) Proc. Natl. Acad. Sci. 79, 4194-4198.Szoka, F. and Papahadjopoulos, D. (1980) Ann. Rev. Biophys. Bioeng . 9,467-508.Tamm, L.K. (1986) Biochemistry 25, 7470-7476.Thevenod, F., Kemmer, T.P., Christian, A.L. and Schulz, I. (1989) J. Membr.Biol. 107, 263-275.Tilcock, C.P.S., Ahkong, Q.F. and Parr, M. (1991) Invest. Radio!. 26, 242-247.Vaz, W.L.C., Derzko, I., and Jacobsen, K.A. (1982) Cell Surf Rev. 8, 84-128.Veiro, J.A. and Hunt, G.R.A. (1985) Chem.-Biol. Interactions 54, 337-348.Veiro, J.A. and Cullis, P.R. (1990) Biochim. Biophys. Acta 1025, 109-115.Verner, K. and Schatz, G. (1988) Science 241, 1307-1313.Walter, A. and Gutknecht, J. (1986) J. Membr. Biol. 90, 207-217.Worcester, D.L. and Franks, N.P. (1976) J. Mol. Biol. 100, 359-378.133Young, S.P., Baker, E., Gomperts, B.D. and Huehns, E.R. (1975) in Proteinsof Iron Storage and Transport in Biochemistry and Medicine. (R.R.Crichton, ed.). North-Holland/American Elsevier, pp. 417-426.Young, S.P. and Gomperts, B.D. (1977) Biochim. Biophys. Acta 469, 281-291.

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0086350/manifest

Comment

Related Items