3...... .mJfiC...k€§!£a:. 2...]: Q, :1 .. $96 .V 1 a . 1.32)}. . , 1‘ . ..i x 9! )hr. 1.. 1:54;. .. 1 .O .. .31.. a... a... n. t x. «H L; tutu! In: :J!vtf);7 1.le l. 1 LI .33... xii; - $333. 3542!}: .3 , :3 .3.n..vx i2iixit: ‘ é . 2 .: .35-: 332;... , V .2 < i 2...: .3. it} i a a... .3 I6. .I. .t’ln [:1 ID'lutw (-1525)? 1 t..!:x.. Io : 31!. .1. .. i...» l .1» v : ,. it? 1... 3:! a . N"; . s. ... ' v I > Vt'fljvpuyyu} 155:5“. . (I535 (D. '1 MICHISGAN III III II III IIIIIIIIII IIIIII IIIIIIIIIIIII 1421 7248 This is to certify that the dissertation entitled Studies on the dynamics of lipid alkyl chains in biological membranes subjected to environmental stress presented by Luc R. Bérubé has been accepted towards fulfillment of the requirements for “Mb degreein Emma/us?” Major prbé‘ssor {not III/IL L Date «Llflllflf MSU is an Affirmative Action/Equal Opportunity Institution 0— 12771 LIBRARY Michigan State Unlverslty PLACE IN RETURN BOX to romovo thIo checkout from your record. TO AVOID FINES return on or botoro duo duo. DATE DUE DATE DUE DATE DUE MSU IoAnNflrmotIvoActIon/Equd OppomJnItyIMItwon STUDIES ON THE DYNAMICS OF LIPID ALKYL CHAINS IN BIOLOGICAL MEMBRANES SUBJECTED TO ENVIRONMENTAL STRESS. By Luc Raoul Bérubé A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Biochemistry 1995 ABSTRACT STUDIES ON THE DYNAMICS OF LIPID ALKY L CHAINS IN BIOLOGICAL MEMBRANES SUBJECTED TO ENVIRONMENTAL STRESS. By Luc Raoul Bérubé a,m-bifunctional long chain fatty acids were identified in the membrane of the Gram positive anaerobic eubacterium Sarcina ventriculi when this organism was grown at high pH (9.7). Under the same growth conditions, the presence of acetal lipids was also detected. Since lipids in Sarcina ventriculi are modified in response to environmental stress, we investigated the dynamics of lipid alkyl chains as a mean to verify the principle of membrane homeoviscous adaptation. Using 1H NMR spin-lattice relaxation time (T1), it was found that the energy of activation (E3) of gauche-trans isomerism in lipid alkyl chains is identical in lipid extracts of cells grown at pH 7.0 (< 10% long chain fatty acids) or pH 3.0 (> 40% long chain fatty acids). Using FT-IR, it was found that the degree of alkyl chain order is similar in both lipid extracts as assessed by the temperature dependence of the symmetric methylene stretching mode. However, analysis of the scissoring and wagging regions of the infrared spectrum revealed that the alkyl chain rotational conformers accessible to pH 7.0 and pH 3.0 lipid extracts are different. Furthermore, the differential scanning calorimetry (DSC) thermogram of the two lipid extracts were significantly different. The analytical tools developed to calculate E3 of gauche-trans isomerism in lipid alkyl chains in Sarcina ventriculi allowed us to estimate E3 in model phospholipid membranes containing lipopolysaccharides (LPS), which are responsible for bacterial pathogenesis in mammals. It was found that non-endotoxically active LPS from Hhodobacter sphaeroides did not change Ea while endotoxically-active LPS from Sarcina ventriculi elevated Ea, a result that supports the contention that LPS's exert their cellular effects by modifying the dynamics of target membranes. Overall the results presented in this thesis suggest that Ea of gauche-trans isomerism in lipid alkyl chains is afundamental physical parameter of lipid membranes that must be conserved in bacteria subjected to environmental stress and that its disruption may lead to abnormal cellular responses. This thesis is dedicated to my parents. iv ACKNOWLEDGEMENTS I would like to thank the following people: First, my thesis advisor Dr. Rawle Hollingsworth who opened my mind to a more physical way of looking at cellular phenomena. Second, my committee members, Drs. Loran Bieber, Estelle McGroarty, Rosetta Reusch, Jack Holland and Frank Dazzo for helping me being critical about my work. Finally, the members of the Hollingsworth's lab over the years who are, literally, too numerous to mention all. But in particular, Odette, Sophie and myself would like to thank Jean, Ben and Zachary for their unconditional friendship. TABLE OF CONTENTS LIST OF TABLES viii LIST OF FIGURES Ix ABBRE VIA TIONS x I CHAPTER 1. Literature review 1 Introduction 1 Physical foundations of the molecular organization and dynamics of membranes 3 Correlation between membrane physical state and physiological functions 1 4 Homeoviscous adaptation 22 CHAPTER 2. Synthesis of very long a,a>-bifunctional alkyl species and acetal linkages between lipid head groups in the membrane of Sarcina ventriculi grown at alkaline pH. 3 1 Introduction 31 Material and methods 34 Results and discussion 36 CHAPTER 3. Membrane lipid alkyl chain motlonal dynamics Is conserved In Sarcina ventriculi despite pH induced adaptative structural modifications including tail to tail coupling 5 2 Introduction 5 2 Materials and methods 5 7 Results and discussion 59 CHAPTER 4. The effect of lipid head to head and tail to tail cross-linking on the thermal behavior of the membrane of Sarcina ventriculi: A differential scanning calorimetry and FT-lR study. 86 Introduction 8 6 vi Materials and methods 88 Results and discussion 90 CHAPTER 5. Endotoxic lipopolysaccharide from Salmonella typhimurium modifies lipid alkyl chain dynamics in model membranes but non-endotoxic lipopolysaccharide from Rhodopseudomonas sphaeroides does not 1 1 5 Introduction 1 15 Materials and methods 121 Results and discussion 123 CONCLUSION 1 4 5 Relevance of the present studies to microbial ecology 150 BIBLIOGRAPHY 1 5 4 LIST OF TABLES TABLE 1. PARAMETERS FOR THE THEORETICAL FITTING OF T1 DATA. viii LIST OF FIGURES FIGURE 1. STRUCTURE OF SOME BACTERIAL LIPIDS. 4 FIGURE 2. SCHEMATIC DIAGRAM OF THE RATE OF MOTIONS OF THE DIFFERENT LIPID MOTIONS AND THE TECHNIQUES USED TO MEASURE THEM 7 FIGURE 3. LIPID POLYMORPHISM IN THE LIQUID CRYSTALLINE STATE 12 FIGURE 4. INTERDEPENCE BETWEEN FUNCTIONAL ASPECTS OF MEMBRANES AND THEIR MOLECULAR ORGANIZATION 1 5 FIGURE 5. STRUCTURES OF BIFUNCTIONAL LONG CHAIN LIPIDS 24 FIGURE 6. GAS CHROMATOGRAM OF FATTY ACID METHYL ESTERS DERIVATIZED FROM TOTAL LIPID EXTRACTS OF SARCINA VENTRICULI IN CULTURE AT PH 9.7. 37 FIGURE 7. MASS SPECTRA OF THE FATTY ACID METHYL ESTERS DERIVATIZED FROM TOTAL LIPID EXTRACTS OF SARCINA VENTRICULI IN CULTURE AT PH 9.7. 39 FIGURE 8. PROPOSED MECHANISM FOR TAIL TO TAIL COUPLING OF ALKYL CHAINS. 4 1 FIGURE 9. PROPOSED MECHANISM FOR THE FORMATION OF ACETAL LIPIDS. 43 FIGURE 10. TWO-DIMENSIONAL PROTON-CARBON HETERONUCLEAR MULTIOUANTUM COHERENCE (HMOC) NMR SPECTRUM OF THE TOTAL LIPIDS FROM SARCINA VENTRICULI GROWN AT PH 9.7. 47 FIGURE 11. TWO-DIMENSIONAL THIN LAYER CHROMATOGRAPHY OF LIPIDS FROM SARCINA VENTRICULI. 49 FIGURE 12. THIN LAYER CHROMATOGRAPHY OF LIPID EXTRACTS FROM CELLS OF SARCINA VENTRICULI GROWN AT PH 7.0 OR 3.0 BEFORE AND AFTER NMR SPECTROSCOPY. 60 FIGURE 13. GAS CHROMATOGRAM OF FATTY ACID METHYL ESTERS DERIVATIZED FROM TOTAL LIPID EXTRACTS OF SARCINA VENTRICULI. 6 2 FIGURE 14. SCHEMATIC DIAGRAM OF A MEMBRANE ARRANGEMENT FROM S. VENTRICULI GROWN AT PH 7.0 OR 3.0. 65 FIGURE 15. 500 MHZ 1H NMR OF TOTAL LIPID EXTRACT FROM ix S. VENTRICULI. 6 7 FIGURE 16. 500 MHZ T1 SPIN-LATTICE RELAXATION TIME AS A FUNCTION OF INVERSE TEMPERATURE OF ALKYL CHAIN PROTONS. 69 FIGURE 17. 300 MHZ T1 SPIN-LATTICE RELAXATION TIME FOR METHYLENE PROTONS OF ALKYL CHAINS AS A FUNCTION OF INVERSE TEMPERATURE. 72 FIGURE 18. FREQUENCY DEPENDENCE OF T1 AT 37°C. 74 FIGURE 19. THEORETICAL FIT OF THE METHYLENE PROTON T1 VALUES. 84 FIGURE 20. INFRARED ABSORBANCE SPECTRA OF THE CARBON-HYDROGEN STRETCHING MODES REGION. 91 FIGURE 21. TEMPERATURE DEPENDENCE OF THE FREQUENCY OF THE CH2 STRETCHING BANDS. 93 FIGURE 22. DSC THERMOGRAMS OF TOTAL LIPID EXTRACTS FROM CELLS OF S. VENTRICULI. 96 FIGURE 23. INFRARED ABSORBANCE SPECTRA OF THE CARBONYL STRETCH MODE REGION. 1 00 FIGURE 24. BANDWIDTH OF THE CH2 SYMMETRIC STRETCHING MODE BAND AS A FUNCTION OF TEMPERATURE. 104 FIGURE 25. INFRARED ABSORBANCE SPECTRA OF CH2 BENDING MODES REGION. 1 0 7 FIGURE 26. INFRARED ABSORBANCE SPECTRA OF CH2 WAGGING MODES REGION. 1 09 FIGURE 27. DIFFERENCE SPECTRUM OF PH 7.0 AND 3.0 LIPIDS IN THE REGION COVERING THE BENDING AND WAGGING MODES. 1 12 FIGURE 26. STRUCTURES OF LIPID A. 116 FIGURE 29. 500 MHZ 1H NMR OF PC:PE:PS PHOSPHOLIPID MIXTURE. 124 FIGURE 30. SPIN-LATTICE RELAXATION TIME (T1) OF METHYLENE PROTONS AS A FUNCTION OF INVERSE TEMPERATURE. 126 FIGURE 31. THEORETICAL FIT OF THE METHYLENE T1 DATA. 128 FIGURE 32. MODEL FOR THE MOLECULAR ARRANGEMENT OF LPS MOLECULES IN LIPID BILAYERS. 133 FIGURE 33. SPIN-LATTICE RELAXATION TIME (T1) OF METHYL PROTONS AS A FUNCTION OF INVERSE TEMPERATURE. 136 FIGURE 34. THEORETICAL FIT OF THE METHYL T1 DATA. 138 FIGURE 35. ENVELOPE CURVES FOR THE METHYLENE PROTONS T1 DATA.141 FT-IR LPS PBS ABBREVIATIONS Nuclear magnetic resonance Hetronuclear multiquantum coherence Fourier transform infrared spectroscopy Differential scanning calorimetry Lipopolysaccharide Phosphate buffer saline CHAPTER 1 Literature review Introduction The fluid mosaic model of biological lipid bilayer as proposed by Singer and Nicholson (1) laid the conceptual foundation for our present day understanding of membrane architecture. The model pr0poses that the membrane is a fluid two dimensional lipid bilayer matrix containing proteins in which both lipids and proteins can diffuse laterally in the plane of the membrane. Singer and Nicholson's model suggests that the distribution of lipid and protein molecules in the membrane lack any significant degree of lateral order. Since then it has been shown that the molecular distribution in membranes is in fact heterogeneous with a high degree of organization (2). Lipids and proteins are laterally segregated in domains that exhibit different composition and physical properties (phase separation) (2). Yet another level of heterogeneity is provided by the asymmetric distribution of both lipids and proteins between the two leaflets of the bilayer (3). The molecular composition and organization of lipids and proteins are determined by the physical and physiological state of the membrane. The importance of an in- depth understanding of the molecular and supramolecular properties 2 of lipids is made clear when one realizes that these properties can be correlated with functional adaptation in bacterial membranes. In this review of the literature, the physical basis for the organization of lipids in membranes will be reviewed first. Then, the relationship between the physical state of the membrane and its physiological functions will be examined. Finally, the principle of homeoviscous adaptation and its significance for bacterial physiology will be discussed. These different aspects of membrane structure and function provide the conceptual framework for the experimental work presented in this thesis. 3 Physical foundations of the molecular organization and dynamics of membranes Biological lipids are a special case of the more general class of liquid crystals which have been described as the fourth state of matter (4). Therefore, lipids have mechanical properties that are liquid-like while their optical and electronic properties are similar to those of crystals. Accordingly, membrane properties can be adequately described by the degree of organization and by the rates of movement of lipids. Full characterization of a lipid system thus requires knowledge about the chemical structures of individual lipid molecules, their conformations and conformational dynamics, and also the intermolecular interactions between lipid molecules which form the basis for their colligative properties. Bacterial lipid species include glycerophospholipids, glycoglycerolipids, plasmalogens, and sterols. Unusual lipids have been identified in archaebacteria such as tetraether lipids (5) and are believed to play a crucial role in the tolerance of these organisms to extreme environments. The structures of some of these lipids are shown in figure 1. The motions accessible to a lipid molecule are intramolecular motions and motions involving the molecule as a whole. lntramolecular reorientational motions include torsion oscillations around single bonds, trans-gauche isomerization (as a result of Figure 1. Structure of some bacterial lipids O ‘CH; WM b1 0' . O \ I ’ CH...o_P_o_x W I I O O O \C‘H; O CH ’ \cua_o_x W O FIGURE 1 PHOSPHOGLYCEROLIPIDS X = ETHANOLAMINE GLYCEROL G LYCOCLYCEROLIPIDS X = CARBOHYDRATES PLASMALOG ENS HOPANE 6 carbon-carbon rotation in the alkyl chain) and free rotation of the terminal methyl group of the alkyl chain and of other functional groups in the polar, hydrophilic region of the lipid. The carbon- carbon single bonds of the alkyl chain allow the possibility for one trans and two gauche conformers (rotational conformers) at each position of a saturated chain. The distribution of these rotational conformers provide a measure of the instantaneous state of molecular order of the alkyl chains. Motions of the molecule as a whole include rotation around the long molecular axis (director) and time dependent orientation of the director for axial diffusion (wobbling). If a lipid bilayer is considered, other motions involving the lipid molecule as a whole must be defined including lateral diffusion in the plane of the bilayer and 180° rotation around an axis parallel to the membrane plane that results in a lipid molecule going from one leaflet of the bilayer to the other (”flip-flop“). The time scale of these motions along with the techniques used to measure them are summarized in figure 2. The motions described above have their origin in the forces acting on lipid molecules. The conformations of lipid molecules and intermolecular forces between lipid molecules are interdependent. Thus, intramolecular forces will give rise to specific conformations that will preferentially exist depending on the molecular surroundings. Conversely, intermolecular forces can be modified by cooperative conformational changes of single lipid molecules to give rise to a new lattice arrangement or phase. The contributions from intra and intermolecular forces can be described in terms of a single Figure 2. Schematic diagram of the rate of motions of the different lipid motions and the techniques used to measure them. References for the various techniques are as follows: Raman/IR (8), (9), ESR (10), Fluorescence (11), NVR (12,13). N mmDOE Eozmxsme msz“ I I. 3 . 0 oocoomoBEEIll W m mmm II. 0 529551 me... one . ~49 «m: mm: 9m: 3“: 9.09.3 Sconce.“— _ . _ I _ . _ . _ . ._ c2353 N:oI w. Setmsom. mcmtocoaaw TIL 0 Eva. 93. Sons 5:22 an: I u .:o_m:=_u .823 I. m S .32.... ea: Tlll 9 chain in an effective molecular field (molecular field theory) (6,7). In this model, the total energy per chain is given by Elf} = EIIIint + Eifldisp + PAlf} (1) where EIfiint is the internal energy of chain configuration {f}. EIIIdisp is the dispersive or van der Waals interaction of the chain with its neighbors and PM“ contains all other interactions such as steric repulsions, electrostatic interactions and hydrophobic effects. To obtain the internal energy of the membrane one can perform a statistical-mechanical averaging over all configurations {f}. This model describes the orientational ordering when the lipid system is in the fluid state. Lipid bilayers are capable of undergoing a number of phase transitions that can be brought about by changes in the energetic balance described in the molecular field theory (equation 1). As a result, lipids will change their orientational ordering in a more or less cooperative fashion. In general, fully hydrated bilayers composed of single lipid species undergo a well defined thermal Iamellar phase transition in which the lipid alkyl chains change from an ordered (gel) state to a fluid (liquid crystalline) state. In addition to being sensitive to temperature, lipid phases are also sensitive to any factor capable of modifying intermolecular forces between lipid molecules. Ion concentration, water content, pH, pressure and the presence of proteins are such physiologically relevant factors capable of triggering isothermal transitions 10 (14,15,16,17). Under certain circumstances lipids can also form non-Iamellar phases such as hexagonal and cubic phases (see figure 3 for examples). The shape-structure concept has been suggested as the molecular basis for lipid polymorphism (7,18). The model states that bilayer (lamellar)-preferring lipids are roughly cylindrical whereas hexagonal structure-preferring lipids are cone-shaped with the polar head group at the apex of the cone. Certain lipids with smaller head groups therefore preferentially form inverted hexagonal structures or cubic phases. However, increase in the effective volume occupied by alkyl chains of lipids that otherwise tend to form bilayer structures may induce the formation of hexagonal phases. The formation of non-Iamellar phases has been suggested to be biologically significant (20,21). However, direct observation of such phases in cell membranes is difficult and their occurrence has been indirectly inferred by the presence of lipids that are known to be prone to form non—Iamellar phases in artificial systems (18). Indirect evidence that non-Iamellar structures are important also comes from the observation of membrane phenomena in which the lipid bilayer must be at least transiently disrupted. Examples include membrane fusion and endocytosis/exocytosis. Several phases may coexist within the bilayer and result in the lipids being organized laterally into compositionally and functionally specific domains (phase separation). These domains exist on different time and length scales (22,23,24). Lipid domains originate from non-ideal (non-homogeneous) mixing of structurally different lipids as a result of the intermolecular forces at 11 thermodynamic equilibrium. The delicate balance between these forces can be shifted to a different equilibrium by changes in pH, temperature, pressure, ion concentration and by the presence of drugs metabolites, exogenous lipids and proteins (2). 12 Figure 3. Lipid polymorphism in the liquid crystalline state. A) Lamellar phase, B) hexagonal inverted phase (H||), C) hexagonal phase (H|) and D) cubic phase. Adapted from reference (19). B HGURE 3 14 Correlation between membrane physical state and physiological functions The central dogma in membrane physiology is that there exists an intimate coupling between the chemical composition, physical state and organization of lipid domains and the functions of membranes (including enzyme activity within the membrane) (25,26,27,28,29). Figure 4 summarizes the interdependence between the functional aspect and the molecular organization of membranes. As early as 1962, Luzzati and Husson (30) have postulated that membrane functions are optimal in the liquid crystalline phase and that therefore at least some of the lipids must be in that phase to support growth of living organisms. With the introduction of the fluid mosaic model, it became clear that the functions of integral membrane proteins might be modulated by the physical state of the lipids. This hypothesis has since been supported by studies showing a correlation between discontinuities in energies of activation (or activities) of membrane bound enzymes and the main (gel to liquid crystal) phase transition (31). Furthermore, the existence of proteins sensitive to lateral lipid packing such as mechanosensitive ion channels (32.33.34) also provides indirect evidence for the role of lipid fluidity in regulating membrane enzymes activities. Despite the above mentioned examples, relatively little data exist demonstrating a direct, quantitative correlation between membrane 15 Figure 4. lnterdepence between functional aspects of membranes and their molecular organization. Adapted from reference (26). 16 COMPOSITION 4—» PHYSICAL STATE I l FUNCTION 4—— ORGANIZATION FIGURE 4 l7 enzyme activity and the "fluidity” of liquid-crystalline lipids. Moreover, it has proposed that fluidity of membrane phospholipids plays only a minor role in regulating the functions of membrane proteins (35). In fact, Sylvius et al. (36) have been able to demonstrate that the osmoregulatory (Na+, Mgz+)-ATPase in Acholeplasma Iaidlawii is not affected by changes in the fluidity of the membrane so long as the lipids remain in the liquid-crystalline phase. However, the ATPase is inactivated when lipids enter the gel phase. Based on these results and on the observation of discontinuities in Arrhenius plots of membrane enzyme activity at the phase transition, these authors proposed that the phase state of membrane lipids is more important than the fluidity of liquid crystalline membrane lipids in regulating membrane enzymes activities. Implicit in their proposal is the fact that the fluidity of the gel phase is incompatible with optimal enzyme activity. This contention has been refuted by Zakim et al. (37) who argued that the differences in energies of activation of enzyme activity above and below the break points in Arrhenius plots are much larger than the calculated energies involved in the mechanical properties (viscosity) of the membrane. This suggests that the discontinuities in energy of activation of enzyme activity at the phase transition are independent of the viscosity per se but may result from discontinuous changes in other physical properties of the membrane at the phase transition. As discussed above, biomembranes exist as a dynamic and precisely controlled mixture of lipid domains of different phases. There is 18 new evidence that these domains can regulate membrane associated enzyme activity. In a very elegant study, Grainger et al (38) , using fluorescence microscopy. have shown that phospholipase A2 (PLA2) targets and hydrolyzes preferentially solid phase domains of L-a- dipalmitoylphosphatidylcholine. Furthermore, they showed that after a critical extent of monolayer hydrolysis the enzyme aggregates in proteinaceous domains leading to enzyme inactivation. Again using PLA2 Sen et al (39) have provided evidence that the peak in enzymatic activity corresponded with the onset of appearance of non-bilayer structures (hexagonal or H" phase). They interpreted the results as the enzymatic activity being sensitive to the lipid molecular packing stress at the onset of the transition. Another study has shown that once the transition from bilayer to hexagonal phase is complete the activity of PLA2 is loss (40). Similarly, protein kinase C (PKC) activation can be promoted by the presence of H|| forming lipids (41), a result that was also interpreted as in terms of the molecular packing properties at the bilayer surface. Similarly, Gruner has proposed that the presence of non-Iamellar forming lipids results in the buildup of elastic strain in the membrane that can modulate enzyme activity (42). Of course, the time dependent modification of domain organization represents an important mean of controlling enzyme activities, The spatio-temporal modification of lipid domains has been explicitly described using percolation theory. Using binary and ternary lipid mixtures, it was shown that lateral diffusion of a fluorescent probe suddenly increases at the percolation threshold (43). The percolation 19 threshold is the point at which previously disconnected domains (for example the gel phase domains) are now continuous as a result of an increase in the number of lipids in that phase. Lipid percolation has been postulated to modulate the activity of a pancreatic lipase (PL). This lipase was shown to be sensitive to the relative proportion of lipids (which dictate the relative amount of gel and fluid phase) in a binary lipid system (44). Because integral proteins are preferentially soluble in specific lipid domains (45), modulation of the domains architecture, leading to a percolation threshold for example, can directly regulate the interaction between previously separated proteins. In a theoretical paper, Melo et al (46) have shown that indeed such compartmentalization of reactants can affect the reaction yield. The importance of the physical state of lipids is not limited to its modulation of enzyme activity. The physical state of lipids per se can directly mediate important membrane functions. Perhaps the best studied phenomenon in this regard is membrane permeability to water and ions as well as small organic molecules. Membranes exhibit a permeability maximum to ions at the gel to liquid crystalline phase transition which may be explained by critical fluctuations in lateral compressibility (density) of the membrane at or close to the phase transition (47,48). Alternatively, mismatch in lipid molecular packing at the interface of coexisting gel and liquid crystalline domains which results in the variation of the interfacial area can also lead to increase permeability (49). Modification of passive transmembrane permeation caused by lateral density 20 fluctuations and interface formation is not restricted to small positive ions but can also apply to molecules such as water and TEMPO-choline (a spin-label cation) (50,51) The possibility has been suggested that increase in flip-flop rates at the chain melting phase transition could also result in increase permeability (52). Electrical properties of membranes have also been shown to be sensitive to the physical state of lipids. In a study by Yagisawa et al. (53,54) it was shown that self-sustained oscillation of the electric potential of lipid bilayer membranes is induced by the gel to liquid crystalline phase transition. The transitions themselves are triggered by the repetitive adsorption and desorbtion of protons at the interface between ionic solutions and the polar head group of the lipid bilayer. The degree of disorder in fatty acid monolayers has been correlated with the electron transfer efficiency (all trans chains having a higher efficiency) (55), a very important phenomenon known to involve proteins as well as lipids. The proton gradient across the membrane contributes to the proton motive force. Transport of protons across the bilayer usually takes place through transport enzymes (pumps). It has been shown that protons may move laterally via an H-bond network formed by polar head groups of lipids providing an efficient way to reach the spatially separated pumps (56). In turn, the H-bond network depends critically on the physical state of lipids which therefore can regulate the proton motive force (57). 21 The physical state of lipids can also play a direct role in intracellular signaling through modulation of cations concentration and spatial distribution. Cations can bind to anionic lipids and influence the thermodynamic equilibrium of membranes thus modulating lipid packing, phase and domains (58). Less often appreciated is the fact that this modulation is reciprocal. That is, modification of lipid packing, phase or domains by temperature, pressure and pH for example, can affect the spatial organization and local concentrations of cations by shifting the thermodynamic equilibrium resulting in the redistribution of anionic lipids. This can provide a very efficient and precise way of regulating intracellular responses that are cation sensitive. in this respect, the role of Ca2+ in cellular signaling has been extensively studied (59). Lipids can also influence cytosolic molecules by acting directly on proteins of the cytoskeletal apparatus (60) which themselves can be tightly coupled to intramolecular signaling. So far, I have discussed the physical foundations for the molecular organization and dynamics of lipids and how the physical state of membrane can affect its physiological functions. In the remainder of this literature review, I will discuss the mechanisms by which cells (bacteria) modify the structures of membrane lipids in order to compensate for the disruption of the thermodynamic equilibrium of the membrane by various environmental stresses. 22 Homeoviscous adaptation From the previous discussion it can be appreciated that the membrane can function as a physicochemical sensor that can detect alterations in its environment such as temperature, pH, pressure etc. and transduce them into meaningful functional changes. In addition, bacteria are capable of modifying the chemical composition and molecular organization of their membranes so as to maintain their physical properties at a level compatible with optimal function. This adaptability of bacterial membranes is crucial for the survival of organisms which can be subjected to large variations in the environment of their habitat. In fact, modifications of bacterial membrane lipid composition in response to changes in environmental conditions such as temperature (61,62) and pressure (63) is a well established phenomenon. The nature of the changes in lipid composition depends both on the nature of the environmental perturbation and on the organism under consideration (64). A common adaptative response to a variation in growth temperature is a change in the degree of unsaturation of membrane lipid alkyl chains. Other types of changes include modification of acyl chain length and branching and cyclization of fatty acyl chains as well as modifications of lipid head groups. All these modifications are capable of altering membrane fluidity. In addition to these commonly observed modifications in acyl chains, the occurrence of long chain bifunctional lipids (of up to 36 carbons) has been documented in a number of eubacteria (65,66,67). 23 Furthermore, the amount of these unusual lipids in the membrane is modulated by the growth conditions indicating that they represent yet another mean by which membranes can adjust their physical properties. The structure of these lipids has been solved for at least two of these organisms, namely Sarcina ventriculi (68) and Butyrivibrio sp. (65,69). The structures were found to be very similar (figure 5) and reminiscent of the tetraether lipids of archaebacteria. These long chain lipids could represent a lipid structure that is particularly well adapted to extreme environment such as very low (high) pH, high temperatures and extremes in salinity. The changes in lipid chemical composition in response to environmental pressure are well characterized but the mechanisms (and especially the cellular sensors) responsible for these changes are still poorly understood. Essentially two mechanisms have been proposed. The first mechanism involves temperature sensitive enzymes responsible for chain elongation and unsaturation. The second mechanism suggests that the primary sensor of temperature is the physical state of the membrane which in turn can activate proteins involved in lipid synthesis and modifications. Evidences exist in support of both mechanisms. For example, in organisms using the anaerobic pathway of fatty acid synthesis (anaerobic and some facultative anaerobic bacteria) fatty acid synthetase produces both saturated and unsaturated fatty acids. The enzyme that catalyses the elongation of the chain, B-ketoacyl-ACP synthase 24 Figure 5. Structures of bifunctional long chain lipids.A) From Butyrivibrio spp. (71), B) from S. ventriculi (72) and C) from Methanococcus jannaschii (5). 25 A caps HOH O-CH; OH OH CH3 CH3 (CH2)13CHs CH (CH) rho-comm),,iu-LH-(CHQBCOb—CH .3- 13\/\0é82 on I cnzo—r—ocnzcuoncn2 II C3H7CO B H C cazou HZC—o I —CH HC—O 0—04, I cnzon FIGURE 5 26 exists in two forms. The type II enzyme is more efficient at elongating unsaturated fatty acids and is more temperature labile than type I. Therefore, at low temperatures, the activity of type II is increased relative to type I and more unsaturated fatty acids are synthesized. This type of temperature control of lipid composition has been well documented in B. ammoniagenes (70). In contrast, the fatty acid syntethase of organisms using the aerobic pathway of fatty acid synthesis only produces saturated fatty acids. In these organisms, desaturation is catalyzed by a membrane bound desaturase. There is some indications that the activity of this enzyme can be modulated by the physical state of the membrane (73). In M. cryophilus, membrane fluidity can also be regulated by a membrane bound enzyme that has been suggested to be sensitive to the physical state of the membrane (74). This enzyme is an elongase capable of modifying the ratio of alkyl chains with 16 carbons to that of 18 carbons upon temperature changes. More recently, it has become increasingly obvious that the physical state of membrane can directly control the enzymatic activity responsible for lipid chemical modifications. Vigh et al. (75) have shown that the catalytic hydrogenation of membrane fatty acids activates the transcription of the desaturase gene which lead to the re-synthesis of unsaturated fatty acids. This elegant experiment conclusively demonstrated that desaturase system is sensitive to the physical state of the membrane since the effect of hydrogenation was localized in the membrane. Similarly, fatty acid desaturase activity in Tetrahymena was found to depend on membrane fluidity but not on the cell temperature (76). 27 The control of lipid composition by the physical state of the membrane also extends to the synthesis of long chain lipids. in Sarcina ventriculi, it was shown that the amount of these long chain lipids can be modulated by the presence of membrane fluidizing agents such as ethanol and butanol (68). In the fatty acid auxotroph Butyrivibrio sp., the presence of long chain lipids correlated with the incorporation of fatty acids with different degree of unsaturation (69). In the archaebacterium Methanacoccus jannaschii the proportion of tetraether lipids was increased in response to increasing temperatures (77). In all cases the formation of long chain lipids has been suggested to occur by tail to tail coupling of preexisting lipids presumably already present in the membrane. These results strongly suggest that the physical state of the membrane is the cellular sensor responsible for triggering the synthesis of long chain lipids. The fact that membrane physical state can serve as the cellular sensor triggering changes in lipid composition is not surprising since in cases in which only membrane fluidity is affected, this represent the only mechanism capable of triggering the appropriate changes in lipid composition. Furthermore, this mechanism is self-regulated. The first physical measurements on the dynamics of lipids following adaptative changes were made with the membrane of E. coli using electron spin resonance spectroscopy (ESR). The rate of motion of the bulk lipids was measured by the correlation time of the ESR probe and was shown to remain unchanged at different growth temperatures despite changes in the lipid composition (78). The 28 term homeoviscous adaptation was introduced to describe this phenomenon. Despite evidences supporting homeoviscous adaptation, it was observed that some organisms only partially compensate changes in membrane fluidity brought about by changes in growth conditions (79,80). Cossins and Sinensky (80) have introduced the concept of homeoviscous efficacy to quantitatively describe fluidity adaptation. The coefficient of homeoviscous efficacy represents the ratio of the change in membrane order to the difference in growth conditions (difference of temperature or pH or ion concentration etc.). Thus for a complete adaptation that is, if there is no difference in the fluidity before and after the perturbation then the homeoviscous efficacy factor is 1.0. A partial fluidity adaptation leads to a value between 0 and 1.0. Using this scale it was shown that the homeoviscous efficacy of prokaryotes membranes varies between 0.2 and 1.0. The wide range of homeoviscous efficacy ratios can be explained in part by the technique and/or the molecular probe used. Different techniques and probes may yield different fluidity values because they may be sensitive to different type of motions. lmplicitly, this suggests that modifications of membranes chemical composition in response to environmental perturbations may result in the preservation of specific types of lipid motions at the expense of others. Which lipid motions are preserved may depend on the nature of the perturbation and on the organism under consideration. From the observation of large variations in homeoviscous efficacy it can be argued that fine tuning of the fluidity of lipids in the liquid crystalline phase may not be crucial for the survival of 29 microorganisms. As discussed previously, some authors have suggested that changes in the chemical composition of membranes is aimed at preserving the optimal lipid phase. This proposition is supported by experiments showing that the temperature of the gel to liquid crystalline phase transition is shifted to lower temperatures when an organism is adapted for growth at lower temperature. In the eukaryote Tetrahymena, for example, the extent of phase separation (proportion of gel vs liquid crystalline) was found to remain constant in cells grown at 15°C or 39°C. However, the homeoviscous efficacy was estimated to be only 0.25 (76). Modification of lipid composition may also be regulated so as to increase or decrease the proportion non-bilayer forming lipids. As already mentioned these unusual lipid phases may play an important role in the physiology of membranes. Stabilization of the Iamellar phase in Clostridium butiricum under growth conditions that would otherwise favor the hexagonal phase, is achieved by an increased in the relative abundance of the glycerol acetal of plasmenylethanolamine (81). This unusual lipid favors the stabilization of the Iamellar phase by virtue of its larger polar head which tend to give a more cylindrical shape to the lipid molecule (see above for explanation on the molecular shape hypothesis of lipid phase). The control of lipid phase in Acholeplasma Iaidlawii is achieved by a similar mechanism this time by varying the ratio of monoglucosydiglyceride to that of diglucosyldiglyceride (82). 30 The importance of changes in lipid composition and its effect on their dynamics during bacterial adaptation to environmental stress will be studied in this thesis using 8. ventriculi as a model. It is relevant to study this organism since as described above, it is able to synthesize unusual long chain lipid similar to those found in other eubacteria and some archaebacteria. Therefore, the results reported here may have broad significance. In a second study, the effect of lipids dynamic on lipid mediated cell signalling will also be assessed in a lipopolysaccharide- phospholipids model system. Lipopolysaccharides can disrupt normal cellular functions when incorporated in the membranes of mammalian cells. This model system will provide an example of the correlation between membrane physical state and its physiological funcflons. CHAPTER 2 Synthesis of very long moo-bifunctional alkyl species and acetal linkages between lipid head groups in the membrane of Sarcina ventriculi grown at alkaline pH Introduction Structural modification of membrane lipids is a common adaptative phenomenon in response to changes in environmental conditions which can affect membrane fluidity (see chapter 1). It has been proposed that such adaptations can restore membrane dynamics to level compatible with optimal membrane functions (78). Typical lipid alkyl chain modifications include alteration of alkyl chain length, degree of unsaturation, cyclization and changes in lipid composition (nature of head group) (73). There is still some controversy as to the identity of the cellular mechanism responsible for detecting environmental perturbations. A large part of the literature on this subject supports the involvement of proteins as being the critical cellular component responsible for sensing external perturbations and transducing them either directly or through transcription into lipid changes that can restore membrane 31 32 fluidity to functional levels (83). More recently, however, it has become increasingly clear that membrane physical state plays a central role as the sensor of environmental perturbations (75). Thus, changes in the fluidity of the lipid bilayer can activate (inactivate) membrane associated proteins that will modify the membrane chemical composition to restore fluidity to an optimal level. Unusual lipid modifications have been uncovered in the gram- positive, strictly anaerobic eubacterium Sarcina ventriculi .This organism has been shown to synthesize very long (32—36 carbons) chain bifunctional lipids that span the entire bilayer (68,84). Strong chemical evidence show that the long chain lipids are formed by the tail to tail coupling of preexisting fatty acids from opposite sides of the lipid bilayer, possibly by an enzyme catalyzed radical mechanism (84). Such long chain lipids have also been observed in other organisms (85.66.67). In S. ventriculi, the synthesis of these long chain lipids is promoted by changes in cell culture conditions such as depression of pH, increase in temperature and the presence of organic solvents (68). In addition, the presence of acetal linkages between lipid head groups was observed when S. ventriculi was grown at low pH (72). Previous studies have focused on the characterization of lipids from Sarcina ventriculi grown at pH 7.0 or 3.0 (68,72) . However, it is known that this organism can grow over a wide range of pH (2.0 to 10.0). Lipid adaptations at high pH have remained unexplored although the formation of spores at pH 8.0 or greater has been reported (86). We therefore studied the effect of 33 high pH on the synthesis of long, transmembrane lipids and acetal linkages in S. ventriculi. 34 Material and methods Organism and culture conditions: S.ventriculi JK was cultivated as described previously (87). Growth under pH control was performed using a 12 liter Microferm fermentor (New Brunswick Scientific, Edison, N.J.). The fermentor was equipped with a pH electrode and the pH was adjusted by addition of either 5M NaOH or HCl. Cells were harvested at 4 °C at mid-exponential phase and washed with distilled water and stored at -20°C. For cells in culture at pH 9.7, the cells were first grown at pH 7.0 to mid exponential phase and then the pH was increased to 10.0 and the culture was maintained at this pH for 3 more hours. The final pH was measured to be 9.7 at the time of harvesting. Total fatty acid analysis: Fatty acids analyses were performed on whole cells as described previously (68) with slight modifications. Briefly, approximately 5 mg (wet weight) of cells was suspended in 0.3 ml of chloroform and 1.5 ml of 5% methanolic HCl solution and heated for 24 h at 72 °C. The mixture was sonicated (5 min) every 8 h. The samples were dried under nitrogen and partitioned between water and chloroform. The organic phase was filtered through glass wool. The fatty acid methyl esters so prepared were subjected to gas chromatography analysis on a 25 M J&W Scientific DBl capillary column using helium as the carrier gas and atemperature program of 150 °C initial temperature, 0.0 min. hold time, and a rate of 3°C/min. to a final temperature of 300°C. This temperature was held for 70 min. Gas chromatography-mass 35 spectrometry (GC-MS) analysis was performed using a Jeol JMS- AX505H spectrometer interfaced with Hewlett-Packard 5890A gas chromatograph. Llpld extraction: Lipids were extracted following the procedure of Jung et al (68). Briefly, lipids from approximately 50 g wet weight of cells were extracted at 45 °C with 400 ml of a mixture of chloroform/methanol/water (15:32, by vol.) for 2 hr., followed by 200 ml of chloroform/methanol (5:1, v/v). The organic layer was taken to dryness in a rotary evaporator, dissolved in 1 ml of chloroform and store under N2 at -20 °C. NMR: proton-carbon heteronuclear multiquantum coherence (HMQC) NMR spectrum of pH 9.7 lipids was obtained at a proton frequency of 500 MHz (125 MHz for 13) in a perdeuterated solvent system composed of chloroform / methanol / pyridine and 40% DCI in 020 (10:2:1:1). 36 Results and discussion The gas chromatogram profile of fatty acid methyl esters obtained from cells of Sarcina ventriculi in culture at pH 9.7 is shown in figure 6 and exhibits two families of peaks. The family of peaks eluting early corresponds to typical membrane fatty acyl components ranging from 14 to 18 carbons. The second family of peaks appears at much longer retention times. Previous studies on cells grown at pH 3.0 have shown that these fatty acids consists of a,co-dicarboxilic acid dimethyl esters of 32 to 36 carbons that are essentially absent from lipids of cells grown at pH 7.0 (68,84). In order to better characterize the chemical structure of the fatty acids with long retention times obtained from cells grown at pH 9.7, we analyzed them by gas chromatography-mass spectrometry (60- MS). The mass spectra of the major peaks from the GCchromatogram are shown in figure 7 and are essentially identical to those obtained from cells grown at pH 3.0 or at high temperature (84). The corresponding structures are shown in inset. In addition to promoting the synthesis of long chain lipids, lowering the pH of cultures from 7.0 to 3.0 was also shown to promote the formation of acetal lipids and acetal linkages between adjacent lipids (72). Some of the lipids containing acetal linkages have been identified by electrospray mass spectrometry and NMR spectroscopy as the glucose acetal of phosphatidyl glycerol plasmalogen (GluAPG), glycerol acetal of phosphatidyl glycerol plasmalogen (GAPG) and head to head coupled glycolipids (72). These acetal linkages were proposed to arise from the nucleophilic attack of a hydroxyl group on 37 Figure 6. Gas chromatogram of fatty acid methyl esters derivatlzed from total lipid extracts of Sarcina ventriculi in culture at pH 9.7. Cells were cultured at pH 7.0 until they reached mid exponential phase and then the pH was shifted to 9.7 and the culture was allowed to grow for 3 more hours. The major components are A. 017:1 fatty aldehyde B. C1620 carboxilic acid methyl ester C. 013:1 carboxylic acid methyl ester D. 013:0 carboxylic acid methyl ester 1. 032:1 m-formylmethylester 2. 032:0 a,m-dicarboxylic acid dimethyl ester 3. C34:1 a,m-dicarboxylic acid dimethyl ester 4. 036:2 a,co-dicarboxylic dimethyl ester. 38 V o mun—DOE 558. 9.52 328:3 mczu cchmEm ~_ . _ I 10108180 asuodsal 39 Figure 7. Mass spectra of the fatty acid methyl esters derivatized from total lipid extracts of Sarcina ventriculi in culture at pH 9.7. the corresponding structures are shown in inset. A) peak 2, B) peak 3 and C) peak 4 from figure 6 GOJDQJCU’D 4|<-:r‘b-—OZJ 011390.3ch “Crawl—NZ] OODDQDCFD fl(-r¢D—IIZJ 1813- S- 60- 40‘ 28‘ 0 .... W N 0 . _r a“ 5 B h 2°C 4 B I' 355 . I r I r 41 .. .IIIIHII ”I'JIII'I' I I ”I I I ._.Ill .41 631.7 188 EDD 38% 488 588 BB 788 "/2 133 g5"? 69 55 ea. ,, . 83 u or, :II 97 W m. a“ W 60‘ . 1 48‘ - S 297 2 3 2 s 5013 ' 29‘ 235 l 3 4 514 - 1 3T. DJ I' II I l" I J . TAI III IlIlj..:i9uLn.I IJ'.l ‘ . 180 208 388 408 , 530 6130 7GB I‘l/Z 995 FIGURE 7 H/Z 41 Figure 8. Proposed mechanism for tail to tail coupling of alkyl chains. Radical mechanism involving the enzyme catalyzed abstraction of a hydrogen from each chain followed by the coupling of the radicals. 42 CH. H ..' CH, H FIGURE 8 43 Figure 9. Proposed mechanism for the formation of acetal lipids. OH HO HO OH HO + OH 0 i o H+ or Metal ion 0 II 0 O { (CH2),, (CH2)n < (CH2)n Y OH HO HO HO H OH (CHZ)" O O O 0 HC 0 (CH2)n (CH2)n (EH2). < FIGURE 9 45 the enol ether function of a plasmalogen (see figure 9). The catalysis has been suggested to involve either a proton (acid catalysis) or a metal ion. Evidence of acetal groups in lipids from pH 3.0 cells came mainly from the observation of a signal at 104 ppm in the 13C spectrum which correlated with a triplet at 4.20 ppm, in a 2- dimensional proton-carbon correlated multiquantum coherence (HMQC) NMR spectrum (72). This combination of chemical shifts was not present in the spectrum of lipids from pH 7.0 cells. The 13C spectrum of total lipid extracts from cells in culture at pH 9.7 exhibited a signal at 104 ppm, analogous to the signal of the total lipid extract of pH 3.0 lipids, which correlated in the l-MQC NVR spectrum with a proton signal at around 4.4 ppm (figure 10). The similarity between the spectra of pH 9.7 and 3.0 lipids also includes two signals in the 13C spectrum appearing between 80 and 100 ppm that are absent in the spectrum of pH 7.0 lipids. The signal at 92 ppm was previously assigned to the anomeric carbon of a glucosyl residue in which the reducing end is free. The signal at 82 ppm is attributed to a carbon atom involved in an ether linkage (72). These results strongly suggest that acetal groups are also present in lipids of cells maintained in culture at pH 9.7 and have important implications for the mechanism of acetal formation. We can first deduce that the catalysis does not necessarily involves a proton (that is, it is not necessarily acid catalyzed) since acetals are observed at both pH 3.0 and pH 9.7. Secondly, the fact that the presence of acetals is not the fortuitous result of chemical conditions favoring the reaction implies that they are probably 46 synthesized as part of a membrane adaptative mechanism. This is consistent with our proposal that the formation of acetals in S. ventriculi membrane is catalyzed by an enzyme (possibly with assistance from a metal ion) responsive to changes in the physical state of the lipids. Changes in the physical state of the membrane in S. ventriculi in response to pH variations could possibly be explained by the protonated state of some of the lipid species. The membrane of S. ventriculi has been shown to contain mainly phosphatidylglycerol (PG), diacylglycerol (D6) and monoglucosyl diacylglycerol (MGDG) (72) (at low pH the relative proportion of phospholipids is decreased (68), figure 11). At neutral pH PG, for which the pK of the phosphate is around 3, exists in the deprotonated state (one negative charge). This negative charge allows PG to participate in hydrogen bonding with other lipids such as DG or MGDG via the hydroxyl groups of these lipids. On lowering the pH, PG becomes protonated and the hydrogen bond network can be disrupted, leading to perturbation of the membrane (88). At high pH little or no changes are expected in the state of protonation of PG, OS or MGDG. However, changes in the protons concentration in the bulk solution can have profound effects on membrane physical state. For example, it has been shown that gel to liquid crystalline phase transition can be driven by concentration gradients of H+ across the membrane (53,54). Whatever the nature (as yet unknown) of the mechanisms leading to perturbation of membrane order/dynamics, it is clear that a common adaptation mechanism is involved in response to different types of 47 Figure 10. Two-dimensional proton-carbon heteronuclear multiquantum coherence (l-IMQC) NMR spectrum of the total lipids from Sarcina ventriculi grown at pH 9.7. 48 2&9 E ow mm or mm on ma om mm co." med n p p P —I b b h w —. b P PL —. p b h P — IF - h b — n p — P — b P p b mI-I P PIbI _ b h h .- — b n P .P — b o m. .4 .9 mm .. . I a . . o 0 o . 0 «AW wlv v 0. o mom v .e .. . o To.v @. cm W o . m . w 00 m o . mlw.m. ac .. w e o m 0 win 1%” 0 . Wig . .fiwa waxwfiwnc Qau as warm Wave 0 o .oODemm: m. oe.oomnwmu. o . . bwemfi i «a ages} gasses FIGURE 10 49 Figure 11. Two-dimensional thin layer chromatography of lipids from sarcina ventriculi. (Adapted from reference (68)). A) pH 7.0 B) pH 3.0. A) (a,b,c,j) and B) (a,b,k) phospholipids; A) (e,f,g,h,i) and B) (c,d,f,g,h,i,j) glycolipids. Note the decrease in the proportion of phospholipids at acidic pH. The major lipids have been identified as phosphatidylglycerol (PG), monoglucosyldiacylglycerol (MGDG) and diacylglycerol (DG) (72). 50 FIGURE 11 51 environmental stress. Acetal lipids were also identified in Clostridium butyricum (81), (89) and are similar to those identified in Sarcina ventriculi. Goldfine has also proposed that acetal lipids are synthesized in the membrane in response to changes in its physical state. More specifically, he proposed that enzymes involved in lipid synthesis may be affected by the presence, in the membrane, of microdomains of non-Iamellar lipids. These transient non- Iamellar microdomains would enhance the access of water-soluble precursors to the enzymes involved in lipid inter conversion. Whether this type of enzyme activity modulation takes place in the membrane of Sarcina ventriculi is not known. In conclusion, we have shown that long chain a,m-dicarboxilic acids dimethyl esters are synthesized in S. ventriculi maintained in culture at pH 9.7 and that they are identical to those found in cells grown at pH 3.0. Furthermore, strong IWIR evidence suggests the presence of acetal linkages in cells cultured at high pH similar to those found in cells grown at low pH. We suggest that these structural changes are the result of an adaptation mechanism that compensate for physical perturbation of the membrane. CHAPTER 3 Membrane lipid alkyl chain motional dynamics is conserved in Sarcina ventriculi despite pH induced adaptative structural modifications including tail to tail coupling. introduction It is generally recognized that to be functional, biological membranes need to be in the liquid crystalline phase. This is characterized by a complex mixture of domains the physical and chemical properties of which are regulated by a large variety of interlinked chemical processes. Environmental changes may, however, shift the thermodynamic balance toward phases that are biologically non functional. Membranes of bacteria, therefore, have evolved chemical mechanisms enabling them to maintain a dynamical state compatible with membrane functions in the face of adverse environmental changes. Membrane adaptative processes in bacteria are usually very dynamic, occuring even as the environmental conditions are being changed. Adaptative mechanisms that are triggered in response to temperature variations range from minor ones, such as modification of the chain length distribution of fatty acid residues and changes in the degree of fatty acid unsaturation 52 53 (90,91). Recently a physical interpretation of the significance of fatty acid heterogeneity in bacterial membranes has led to the idea that the lipid chains act as baths for a high density of energy states of magnitude kT. It was further proposed that this distribution of states varied to reflect the Maxwellian distribution of kinetic energies characteristic to that temperature (92). Sometimes, theenvironmental perturbations are too large to be met by simple changes in fatty acid chain length and require more dramatic processes such as synthesis of cyclohexane-containing fatty acids and synthesis of hopanes (93,94). One especially dramatic adaptative response has recently been uncovered in Sarcina ventriculi.The discovery of this response began with the characterization of unusual, very long a,m-bifunctional fatty acids in this organism when it was subjected to increasing temperature or treatment with organic solvent or depression of pH (68). It has subsequently been demonstrated using combinatorial arguments (95) and chemical arguments (84) that these a,co- bifunctional fatty acids might be formed by rapid, dynamic, tail to tail coupling of existing fatty acids from opposite sides of the membrane bilayer (figure 1). More recently, a cell free activity capable of taking foreign, exogenously added fatty acids which were incorporated into membrane vesicles and further incorporating them (tail to tail), into existing membrane lipids has been demonstrated (72). This process has some general significance among eubacteria since very long a,m-bifunctional fatty acids have been found in several other eubacteria inculding Thermoanaerobacter ethanolicus 54 (67), Butyrivibrio sp. (65.96.85) and Thermatoga maritime (66). in Butyrivibrio, tail to tail coupling has also been suggested (65). It has also been demonstrated that in archaebacteria, the proportion of tetraether lipids, now thought to be formed by trans-membrane tail to tail coupling of normal diethers (97,98), is increased with increasing temperature (77). From the standpoint of molecular dynamics, it seems intuitively reasonable that one way of offsetting the increased membrane motion caused by (say) a temperature increase or the addition of an organic solvent or the lowering of pH (which decreases membrane organization by reducing the negative charge of the headgroups thus reducing the extent of their cross-linking with cations) is to tie the tails of fatty acid groups from opposite side of the membrane together which should just balance the magnitude of the perturbation. This is , in fact, a statement of the principle of homeoviscous adaptability (78) in molecular terms. The obvious question to ask is “what motional (dynamic) aspect of membrane has been conserved during this adaptative response and how can this be measured?" One good candidate for such a measurement is trans- gauche isomerism in the alkyl chains of the membrane lipids. This is reasonable since one outcome of this striking adaptative response should be to conserve the average motional (dynamic) properties of the bulk membrane lipid chains thus ensuring that the catalytic properties of membrane proteins (which are controlled by viscosity) and the frequencies of channel openings and closings are not impaired. 55 Nuclear magnetic resonance (NMR) spectroscopy is an excellent practical tool for measuring the dynamic properties of molecules of all sizes. One especially useful NMR parameter is the spin-lattice or the longitudinal relaxation time or T1. This parameter measures the rate of reestablishment of the equilibrium magnetization of the spins (nuclei) in question after it has been disturbed by the absorption of a radiofrequency pulse. The rate of recovery of magnetization is a measure of the motional freedom of the groups in question (99). In any NMR experiment, the measuring pulse or the preceeding pulses cause a departure from the Boltzmann distribution of spin states. Before the next acquisition, the Boltzmann distribution population of spin states must be reestablished. For spin 1/2 nuclei (such as protons) the return to this distribution occurs exponentially with a rate constant of 1/T1 where T1 is the spin-lattice (or longitudinal) relaxation time. The fundamental origin of this relaxation phenomenon is energy exchange by dipole- dipole interactions or exchanges, the frequency of which is determined by the extent of motion of nuclei in question. Information on the timescale of the motion of the relaxing nuclei can also be obtained from a determination of the probability that the vector separating two dipoles which are generating relaxing oscillating fields remains oriented relative to some fixed direction. This probability decreases exponentially with time with a time constant of magnitude to. This value, To. is also called the molecular correlation time and is a good measure of the timescale of motion of the relaxing nuclei and can be related to T1. Hence, the longitudinal 56 relaxation time for a spin 1/2 nucleus or group of nuclei is related to the correlation time (Tc) of their molecular motion by the relation: 1/T1 =Tc/{1+(onc)2} (1) where coo is the Larmor frequency in rad/sec. It follows from equation 1 that for T1 to be a minimum, the timescale of molecular motion (rc) must be matched to the spectrometer frequency such that the condition more = 0.7 is met at a particular temperature. At this condition, the relaxation time is a minimum and the relaxation mechanism is most efficient. Temperature giving rise to motional timescales greater or less than to lead to a reduction in relaxation efficiency and to an increase of T1. For appropriate spectrometer frequencies, therefore, a plot of T1 vs temperature (or 1/temperature) passes through a minimum. It is then possible to determine to at this temperature since it is approximately equal to 1/mo. The theory tested in this study is that after a perturbation that leads to adaptative tail to tail coupling (along with any other headgroup modifications) the motional dynamics of the bulk methylene group is still unchanged at any temperature resulting in Tc being the same at the same temperature in both cases. 57 Materials and methods Organism and culture conditions: S.ventriculi JK was cultivated as described previously (87). Growth under pH control was performed using a 12 liter Microferm fermentor (New Brunswick Scientific, Edison, N.J.). The fermentor was equipped with a pH electrode and the pH was adjusted by addition of either 5M NaOH or HCL. Cells were harvested at midexponential phase and washed with distilled water and stored at -20°C. Lipid extraction: Lipids were extracted according to the procedure of Jung et al. (68). (See chapter 2 for details) Total fatty acid analysis: Fatty acids analyses were performed on whole cells as described previously (68) with slight modifications. Briefly, approximately 5 mg (wet weight) of cells was suspended in 0.3 ml of chloroform and 1.5 ml of 5% methanolic HCI solution and heated for 24 h at 72 °C. The mixture was sonicated (5 min) every 8 h. The samples were dried under nitrogen and partitioned between water and chloroform. The organic phase was filtered through glass wool. The fatty acid methyl esters so prepared were subjected to gas chromatography analysis on a 25 M J&W Scientific DB1 capillary column using helium as the carier gas and a temperature program of 150 °C initial temperature, 0.0 min hold time, and a rate of 3 °C/min to a final temperature of 300 °C. This temperature was held for 70 min. 58 T1 measurements: Proton T1 measurements were performed at 300, 400 and 500 MHz by the inversion recovery technique using a n- t-n/2 pulse sequence. Lipids for T1 measurements were prepared by drying 5 mg of total lipid extract under nitrogen gas followed by evacuation under low pressure for at least 8 h to ensure complete removal of chloroform. The lipids were then resuspended in 0.7 ml of 10 mM phosphate buffered saline (PBS) at pH 7.0 or 3.0. The following salts were added: KCI 2.5 mM, MgCl2 6H20 0.5 mM and NaCl 0.1mM. The buffer was made up in deuterated water. The resulting suspension was sealed under nitrogen and heated to 55 °C for 2 min and cooled to room temperature. This cycle was repeated 5 times to ensure good hydration. The sample was then transferred to a 5 mm hMR tube and sealed under nitrogen gas. To ensure that the above procedure did not result in significant degradation of the lipid preparation, a thin layer chromatography analysis was performed using the lipids after they were extracted from the cells and compared with the same lipids after the NMR T1 experiments. Lipids were separated by TLC using chloroform/methanol/amonia/water (3.3:1.0:0.1:0.05, by volume). The analysis was performed on silica gel plates (Merck) and the spots were made visible by spraying with 50% ethanolsulfuric acid and heating to 120°C. There was essentially no difference between lipids after extraction and the same lipids after NMR spectroscopy (figure 12). 59 Results and discussion The lipid composition of cells of Sarcina ventriculi has previously been determined (see chapter 2, figure 11). Glycolipids and phospholipids are present at both pH but the phospholipid content decreases substantially at low pH. The major lipids were identified as being monoglucosyl diacylglycerol, phosphatidylglycerol and diacylglycerol. Also, the percentage of plasmalogens was estimated by NMR to be approximately 25% (with 75% of ester lipids) (Lee and Hollingsworth, unpublished results). The presence and high proportion (>50%) of very long bifunctional fatty acids in the cells of Sarcina ventriculi grown at pH 3.0 has been well documented (68,84) and is reproduced here for completeness (figure 13) by gas chromatographic analysis. These fatty acids were shown by mass spectrometry in the last cited works to possess alkyl chains of up to 36 carbons. The proportion of long chain lipids relative to regular chain lipids (14 to 18 carbons) in cells grown at pH 3.0 was estimated by integrating the peaks area of the GCanalysis and found to be in excess of 50 % In contrast cells grown at pH 7.0 have none or only a small proportion of long chain lipids (< 10 %). Freeze fracture electron microscopy of membranes from cells grown at pH 3.0 did not show any evidence of concave/convex fracture between the bilayer leaflets suggesting that the long chain lipids span the entire membrane thickness (Jung and Hollingsworth, unpublished results). The resulting lipid arrangement is, therefore, that of a bipolar monolayer. A schematic Figure 12. Thin layer chromatography of lipid extracts from cells of Sarcina ventriculi grown at pH 7.0 or 3.0 before and after NMR spectroscopy. Lanes 1 and 3, pH 7.0 and 3.0 respectively before NMR spectroscopy and lanes 2 and 4 pH 7.0 and 3.0 respectively after NMR spectroscopy. 61 Origin > j—l N 1.).) A FIGURE 12 62 Figure 13. Gas chromatogram of fatty acid methyl esters derivatized from total lipid extracts of Sarcina ventriculi. A) Fatty acid methyl esters from cells grown at pH 7.0 and B) from cells grown at pH 3.0. The major components are: 1. 014:0 carboxylic acid methyl ester 2. 017:1 fatty aldehyde 3. 016:0 carboxilic acid methyl ester 4. 013:1 carboxylic acid methyl ester 5. 013:0 carboxylic acid methyl ester A. 033:1 m-formylmethyl ester B. 032:0 a,co-dicarboxylic acid dimethyl ester 0. 035:1 m-formylmethylester D. 034:1 a,m-dicarboxylic acid dimethyl ester E. 033:2 a,m-dicarboxylic dimethyl ester. 63 LILILIL - - - . .LLIIIILIL Retention time FIGURE 13 64 representation of a membrane system containing long chain lipids is shown in figure 14. The 500 MHz proton MIR spectra of lipid extracts from cells of Sarcina ventriculi grown at pH 7.0 or 3.0 in aqueous suspensions (in buffer at pH 7.0 and 3.0 respectively) are shown in figure 15A and 158 respectively. The broad lines, which result from dipolar coupling, are typical of aqueous lipid suspensions forming large multilamellar liposomes (10). The strong signal at approximately 1.5 ppm was assigned to the bulk of the methylene protons of the acyl chain and the signal at 1.0 ppm to the terminal methyl protons. In the spectrum of pH 3.0 lipids the signal at 0.85 ppm was assigned to the branched methyl groups in the long transmembrane lipids and the signal at 1.81 ppm was assigned to the methylene protons Bto the carbonyl group. Because the methylene and terminal methyl resonances were well resolved, it was possible to probe the motion of both the terminal part and the bulk of the alkyl chain by measuring T1 relaxation times for the methyl and the methylene protons. Typically, the recovery of the equilibrium magnetization as a function of time t after perturbation by a 1: pulse followed an exponential behavior as shown in figure 16A. T1 values for methyl and methylene protons obtained at 500 MHz were plotted as a function of temperature and are shown in figure 163. A global T1 minimum was observed for the methylene protons of lipids from cells grown at pH 7.0 and 3.0 and for the methyl protons of lipids from cells grown at pH 3.0 when the pH of the sample was the same as the pH at which the cells were grown. These global minima appeared approximately at the same temperature (around 15 °C). As 65 Figure 14. Schematic diagram of a membrane arrangement from S. ventriculi grown at pH 7.0 or 3.0. Note the long chain lipids at pH 3.0 with tail to tail coupling represented by wavy lines. 67 Figure 15. 500 MHz 1H NMR of total lipid extract from S. ventriculi. A) From cells grown at pH 7.0 and B) 3.0 in aqueous suspension (in PBS at pH 7.0 and 3.0 respectively). Spectra were obtained at 40°C. 68 v111111111111111111r111 1 IliTIIIIIfi—TIII111I 111TTXIITTTTIT1171W1TIW1111111I11‘IT‘ 2.9 1.8 1.6 1.4 1.2 1.0 0.8 0.6 Ppm FIGURE 15 69 Figure 16. 500 MHz T1 spin-lattice relaxation time as a function of inverse temperature of alkyl chain protons. A) Typical recovery of the magnetization M(z) of the methylene groups as a function of time after inversion. B) 500 MHz T1 spin-lattice relaxation time for methylene and methyl protons of alkyl chains as a function of inverse temperature. Squares; methylene protons and circles; methyl protons. Open symbol represent lipid extract obtained from cells of S. ventriculi grown at pH 7.0 in suspension in aqueous buffer at pH 7.0. Closed symbols represent lipid extract from cells of S. ventriculi grown at pH 3.0 in suspension in aqueous buffer at pH 3.0 and open symbols with dots represent lipid extract from cells of S. ventriculi grown at pH 7.0 but in suspension in aqueous buffer at pH 5.0. Error bars are shown only for the closed squares and represent the standard error from the regression analysis on the magnetization M(z) vs time (from the pulse sequence) used to determine T1. Typically the error bars did not exceed the size of the symbols. Each curve corresponds to one experiment representative of several repeats. The data were fitted with a cubic spline curve. The absolute values for T1 are within 10% from sample to sample. However the position of the local and global minima did not change. 70 M(z) _l 1 I l u l I o l a I 0.0 0.51.01} 2.0 2.5 3.0 3.5 4.0 1.5 5.0 r (s) 0.5 - J_ I ,l l l l 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7 1H x1000 1K") FIGURE 16 71 we stated earlier at the temperature of the minimum, Tc can be evaluated and at the proton frequency of 500 MHz used here ((Do = 2 it x 500 MHz), To is approximately equal to 10'10 8. T1 was also measured at 300 MHz (figure 17). Only the methylene proton relaxation times are reported here. At this frequency no global minimum was observed. The absence of such a minimum at lower field indicated that the one observed in the 500 MHz experiment was not due to a freezing phenomenon or other such artefact. The minimum is expected to appear at a lower temperature at lower field strength. A plot of 1/T1 vs (0'2 (for fields corresponding to 300, 400 and 500 MHz for protons) at any particular temperature did give a straight line as is expected for nuclei undergoing relaxation by a purely exponential process (figure 18). The temperature dependence T1 curves at 500 MHz also showed evidence of the presence of unresolved local minima. These could be attributed either to separate domains which had the same correlation time (but at different temperatures) or to a dynamic shifting of membrane structure as a function of temperature such that the correlation time passes through that same value over a range of temperatures. Both possibilities are especially interesting since it means that at any given temperature, there will always be domains such that critical membrane proteins can always be in the same dynamic environment. The presence of these local minima at different temperatures (suggesting domains with the same correlation time at different temperatures) also means that local temperature fluctuations will be dynamically buffered since the 72 Figure 17. 300 MHz T1 spin-lattice relaxation time for methylene protons of alkyl chains as a function of inverse temperature. Open symbols represent lipid extract obtained from cells of S. ventriculi grown at pH 7.0 in suspension in aqueous buffer at pH 7.0. Closed symbols represent lipid extract from cells of S. ventriculi grown at pH 3.0 in suspension in aqueous buffer at pH 3.0. 0.65 0.6 0.55 _ (SA) 1:045 0.4 0.35 0.3 73 0.5 _ l l l I l' l 3 3.1 3.2 3.3 3.4‘ 3.5 336 3.7 1/T x1000 (K'1l FIGURE 17 74 Figure 18. Frequency dependence of T1 at 37°C. 1/T1(s“) 2.2 _ 2.1 75 TjITjII l I l l l I 0.5‘ 1 1.5 2 2.5 3 1/(w02X10'19) FIGURE 18 3.5 4 76 same dynamic state is reproduced several times along the changing temperature profile. The fact that the methyl groups in lipids obtained from cells grown at pH 3.0 and measured at pH 3.0 also have similar dynamics to the bulk methylene groups is not too surprising since their motion will be constrained by neighboring tail to tail coupled alkyl chains with very little mobility. As expected, the behavior of the terminal methyl groups of the lipid chains of cells grown at pH 7.0, thus containing few transmembrane lipid species, was markedly different. There was no global minimum but there was several pronounced local minima with much longer relaxation times indicating relatively free unrestricted motion. It is our hypothesis that a particular membrane lipid chemistry resulting from bacterial adaptation will support optimum membrane dynamics only in environmental conditions equivalent to those under which adaptation occured. We therefore predicted that by suspending lipid extracts in buffer at a different pH than that at which cells were grown, lipid alkyl chains dynamics should be perturbed. Indeed, measurements at pH 5.0 of the temperature dependence of T1 relaxation times of lipids isolated from cells cultured at pH 7.0 resulted in a very different profile to that obtained when the measurements were made at pH 7.0 (figure 168). Resuspension of pH 7.0 lipids at pH 3.0 was not successful and resulted in precipitation of the lipids. Since pH can hardly affect the motion of alkyl chains directly, this is most likely being transmitted through the protonation and charge state of the headgroup. This is reasonable since the lipid headgroup charge state determines the extent of their 77 solvation, hydrogen bonding and their interaction with and crosslinking by metal cations (100). This, in turn, determines the stability of the membrane structure and the extent to which individual lipid molecules can diffuse in the plane of the membrane. The extent of lateral motion and the space between lipid headgroups do affect the packing density of the hydrocarbon chains (15). In Sarcina ventriculi the major phospholipid has been identified as phosphatidylglycerol (PG) (Lee and Hollingsworth unpublished results). It has been suggested that this phospholipid, which has a pr04 of 2.9 (52) can participate in intermolecular hydrogen bonding through its phosphate group resulting in the stabilization of the lipid system (88). Thus, at pH 7.0 PG is able to function as a hydrogen acceptor. Potential hydrogen donors in Sarcina ventriculi include monoglucosyl diacylglycerol, diacyl glycerol and partially protonated PG molecules. However, at lower pH PG becomes protonated and the hydrogen bonds network can be disrupted. This phenomenon has been shown to result in a decrease in the temperature of the phase transition in phosphatidic acid (88). Structural modifications involving the preservation of alkyl chain dynamics should, therefore, involve some lipid headgroup contributions. In earlier work (72) it was demonstrated that the adaptation of S. ventriculi to low pH also involves important headgroup structural modifications. These modifications included the head to head coupling of lipid molecules via acetal linkages. This mechanism leads to the formation of a family of structures in which various headgroup components are modified by acetalization to other components. 78 In order to better understand the nature of the molecular motions giving rise to T1 relaxation in our lipid systems we used NVFl relaxation theory together with the model proposed by Petersen and Chan (101) which provides a framework to explain T1 behavior of protons in lipid alkyl chains. For spin 1/2 nuclei, relaxation by dipole-dipole interaction is usually dominant. It has been demonstrated that the methylene protons relax by interacting with other methylene protons in the same alkyl chain instead of with those in neighboring chains (102). lntramolecular dipolar spin- lattice relaxation for a two spin system with equal spins can be described (103) by: 1/T1={2(7)4(h/21C)2I(I+1)/5l‘6}{(12c/1+102m02)+(4Tc/1+4’tc2m02)} (2) where 'y is the magnetogyric ratio, h is the Planck constant, I is the quantum number and r is the distance separating the two interacting nuclear spins. If relaxation is an exponential process, as we show in figure 15A then the relaxation frequency (v) can be related to the energy barrier by the relationship v= vo exp(-Ea/kT) (3) where Ea is the energy of activation characteristic for the relaxation process and k is the Boltzmann canstant. This allows one to write the customary relationship: to = To exp(Ea/kT) (4) the activation process in membrane alkyl chain relaxation is known to be governed by gauche-trans isomerization and, for normal bilayer 79 membranes, the value of E3 has been estimated to be 3-4 kcal/mole (101). Hence, knowing E3 and To at any one temperature, it was possible to evaluate the constant To in equation (4). These values could then be used to evaluate Tc at any other temperature by substitution into the same equation. Once “Co (at any given temperature) is known, one can then calculate T1 from equation (2). This would, however, require evaluation of the term 274(h/21t)2l(l+1)/5r6 in that equation. The only variable in this term is r which is the distance separating the relaxing dipoles in the same chain and is very temperature insensitive. The entire term can, therefore, be evaluated and treated as being independent of temperature. Since the plot of T1 vs 1/T passes through a minimum at a field strength corresponding to 500 MHz, (00, T1 and “Co are all known under this special condition. The term in r6 could then be calculated. The calculation from equation (2) of T1 at any temperature, using Tc values calculated from equation (4) with a value for Ea of 4.6 kcal/mol (obtained by trial but based on the original estimate of 3-4 kcal/mol by Petersen and Chan) was then possible. We evaluated T1 as a function of temperature for the methylene protons of pH 7.0 and 3.0 lipids (table 1). The calculated curves (figure 19) followed the envelope of the measured curves and displayed a minimum at the same position indicating that these unusual membranes had the classical relaxation energetics. The T1 values at the minima for the two curves were not exactly the same but this probably reflects the fact that the membranes are structurally different. The fact that they passed through a minimum 80 Table 1. Parameters for the theoretical fitting of T1 data. The error on T1 values was estimated to be between 5- 1 0%. 81 Table 1: Parameters for the theoretical fitting {of T1 data (a) Temperature to T1 (pH 3,0) T1 (pH 7.0) (Kl . (x 10-10 s) (S). (S) 278 2.97 0.59 0.67 283 2.53 0.57 0.65 288 . 2.19 0.56 0.63 293 1.91 0.56 0.63 298 1.67 0.57 0.64 303 1.47‘ 0.58 0.65 308 1.30 0.60 0.67 313 1.15 0.63 , 0.71 318 1.03 0.66 0.75 323 0.92 0.70 0.79 , 328 0.82 ~ 0.75 0.85 (a) The error on T1 values washestimated to be between 540%. 82 at the same temperature, however, was more important and confirmed that the correlation times for their relaxation are the same at that temperature (figure 19). These calculations also confirmed that even these highly cross linked membrane systems (at pH 3.0) had the same dynamics and activation energy as the classical membrane systems. The Petersen-Chan model predicts a “Co of approximately 10'10 to 10'11 s, in very good agreement with our estimates. A similar “Co of approximately 10'10 s for gauche-trans isomerism in deuterated lipids has been observed by Meier ef al. (104) and Seelig et al. (105). The fact that both the activation energy and Tc are very similar for membrane lipid alkyl chains of pH 7.0 and pH 3.0 cells strongly suggest that the average number of gauche conformers and their rate of propagation along the chain axis (106), (101) is the primary aspect of membrane dynamics that is being conserved during this adaptation. The results obtained in the present study indicate that, despite dramatic changes in the structure of the membrane lipids, the molecular motion detected by spin-lattice relaxation time measurements on membrane alkyl chains of Sarcina ventriculi is conserved on changing the growth pH from 7.0 to 3.0. At 500 MHz, it was possible to estimate the correlation time (12c) of the relaxation process at the temperature of the global minimum. Membrane lipids are shown to be dynamic with respect to their chemical composition. This composition ultimately ensures that the molecular motions of the lipid molecules remain constant with changing environmental parameters in order to support function. This is in agreement with 83 the theory that the thermal energy trapped in collective vibrational modes of membranes are dynamically tuned to the magnitude of the thermal energy bath, I'kT" (92). 84 Figure 19. Theoretical fit of the methylene proton T1 values. Using equation 2 and values from table 1. Lines represent the calculated values and symbols are the experimental data points reproduced from figure 16. pH 3.0 (closed symbols) and 7.0 (open symbols) lipids. 0.5" 85 3 3.1 3.2 3.3 3.4 3.5 3.6. 3.7 1/T x1000 1K") FIGURE 19 CHAPTER 4 The effect of lipid head to head and tail to tail cross-linking on the thermal behavior of the membrane of Sarcina ventriculi: A differential scanning calorimetry and FT-IR study. Introduction In the two previous chapters, we have shown that the energy of activation of gauche-trans isomerism in lipid alkyl chains of Sarcina ventriculi is conserved on changing the pH of culture from 7.0 to 3.0. This result is remarkable considering the extent of lipid structure modifications upon adaptation to low pH (68,72). Since the probability distribution of alkyl chain rotational states is likely to be affected by the presence of transmembrane linkages and acetal linkages between lipid polar head groups, it seems reasonable that despite the similar energetic for gauche-trans isomerization in the two membrane system, the actual populations of such conformers should be different. The degree of disorder and the nature of rotational conformers in alkyl chains can be monitored by infrared spectroscopy. Extensive studies have shown that the region of the C-H stretching modes 86 87 (2850-2960 cm“) is sensitive to the degree of disorder in lipid alkyl chains (8,107). Furthermore, the work of Snyder (108,109,110) has helped establish the CH2 bending and wagging modes (1400- 1500 and 1300-1400 cm'1 respectively) as a diagnostic region for determining the presence of specific gauche conformers. We have used these regions to characterize the rotational states accessible to the lipid alkyl chains of S. ventriculi grown either at pH 7.0 or 3.0. The fluidity of lipid membranes encompasses several domains of freedom, the most important of which is probably the temperature at which the hydrocarbon chains are fluid. In fact, biological membranes normally function in the fluid state. There are other aspects of fluidity that would involve changes in domain structure and composition and the relative lateral motion of lipid groups in the membrane plane. The temperature at which the lipid chains melt reflects the energetic of gauche-trans isomerizations and kink diffusion. Since our earlier NMR studies indicate that the energetic of these processes is conserved in membranes of cells from bacteria grown at pH 3.0 instead of at pH 7.0, it is expected that the temperature of melting of membranes from these two systems would be similar. This was examined by differential scanning calorimetry. 88 Materlals and methods Organism and culture conditions: 8. ventriculi JK was cultivated as described previously (87). Growth under pH control was performed using a 12 liters Microferm fermentor (New Brunswick Scientific, Edison, NJ). The fermentor was equipped with a pH electrode and the pH was adjusted (under feedback from the sensor) by the automatic addition of either 5 M NaOH or HCI. Cells were harvested at mid exponential phase and washed with distilled water and stored at - 20° C. Lipids preparation: Total lipids were extracted following the procedure of Jung et al. (68) See chapter 2 for details. FT-lR spectroscopy: Lipids for infrared measurements were prepared by drying 5 mg of total lipid extract under nitrogen gas followed by evacuation under low pressure for at least 8 h to ensure complete removal of chloroform. The lipids were then resuspended in 3 ml of 10 mM phosphate buffered saline (PBS) at pH 7.0 or 3.0. The following salts were added: KCI 2.5 mM, MgCl2 6HzO 0.5 mM and NaCl 0.1mM. The resulting suspension was sealed under nitrogen and heated to 55° C for 2 min and cooled to room temperature. This cycle was repeated 5 times to ensure good hydration. The sample was then transferred to a jacketed ATR (attenuated total reflectance) cell. The temperature was adjusted by circulating water in the jacket with the use of a temperature-controlled circulating bath. The temperature was also monitored by a RTD (resistive temperature 89 device) probe in thermal contact with the internal face of the jacket. Two hundred scans were acquired at each temperature with a resolution of 2 cm“. When necessary, smoothing of the data was performed using a 9 points smoothing algorithm. Differential scanning calorimetry: DSC was performed on a MC- 2 Microcal instrument. The lipids were prepared as described for FT- lFi measurements and were injected into the cell at a concentration of 2.5 mg/ml. A scan rate of 60° C /h was used. 90 Results and discussion The infrared spectra of aqueous suspensions of lipids from Sarcina ventriculi grown at either pH 7.0 or 3.0 and measured at the corresponding pH values were recorded between 6 and 50°C. We first analyzed the carbon-hydrogen stretching modes region (figure 20) for which the assignments of the peaks are well established (107). The bands at around 2850 and 2923 cm'1 correspond to the methylene symmetric and asymmetric stretching modes respectively. Shoulders at around 2875 and 2960 cm'1 have been assigned to the methyl symmetric and asymmetric stretching respectively. The temperature dependence of the frequency of the carbon-hydrogen stretching modes is an indication of the degree of order of the lipid alkyl chain and can thus be used to follow phase transitions (8). In figure 21, the temperature dependence of the symmetric and asymmetric methylene stretch frequencies for pH 7.0 and pH 3.0 lipids is shown. In both lipid systems, the symmetric and asymmetric modes increase sharply at around 8-10°C. The frequency of the symmetric stretching band increases from 2852.2 to 2855.3 cm-1 for pH 3.0 lipids and from 2851.5 to 2855 cm-1 for pH 7.0 lipids. The asymmetric band increases from 2922.3 to 2926 cm‘1 for pH 3.0 lipids and from 2921.2 to 2926.8 for pH 7.0 lipids. The transitions were rather sharp spanning only a few degrees. The results indicate that the lipids of pH 3.0 grown cells were more disordered than those of pH 7.0 grown cells below the phase transition. It has been sugested that the presence of the methyl branching (as is the case for the long chain lipids of Sarcina 91 Figure 20. Infrared absorbance spectra of the carbon- hydrogen stretching modes region. A) Total lipid extract from cells of S. ventriculi grown at pH 7.0 and B) pH 3.0. Absorbance l l l 1 3000.0 2975.0 2950.0 2925.0 2900.0 92 1 2875.0 2850.0 2825.0 1 2800.0 1 2800.0 J L 2575.0 2630.0 2825.0 4 .l 1 3000.0 2975.0 2950.0 2925.0 2900.0 Wavenumber (cm'1) FIGURE 20 93 Figure 21. Temperature dependence of the frequency of the CH2 stretching bands. A) Asymmetric stretching and B) symmetric stretching. Open symbols, pH 7.0 lipids, closed symbols, pH 3.0 lipids. Wavenumber (cm-1) 94 2923 A 2925‘ ’ - 2924‘ 2922~ ' 2920 I T I I I 7 0 1o 20 30 40 so 2856 B 6 5 “‘ ./' - , 2" . V- 2855i / ' ’ v I 2354- I 2853‘ l 2852‘ 2351 . . 1 . . ~ . - o 10 20 30 40 50 Temperature (00) FIGURE 21 95 ventriculi grown at pH 3.0, see chapter 1) can interfer with lipid packing and therefore introduce disorder in the alkyl chain (71). The thermal behavior of lipid extracts from cells grown at pH 7.0 or 3.0 is shown in figure 22. The pH 7.0 lipids exhibited several endothermic transitions. The onset temperature of the initial transition is approximately 12°C with a peak at around 25°C. The onset temperature of this transition was slightly higher than the onset temperature of the transition observed using the carbon- hydrogen stretching frequencies. in contrast with pH 7.0 lipids, the pH 3.0 lipids exhibited only one major endothermic transition with an onset temperature of 12°C with a peak around 30°C. The onset temperature of this transition (12°C) was again higher than the onset temperature observed using the temperature dependence of CH2 stretching frequency and was also broader perhaps as a result of the more rapid heating rate used in the DSC experiments (60°C/hr compare to 10°C/hr in the FT-IR experiments). Interestingly, the thermal behavior of the pH 3.0 lipids was similar to that observed for membranes of Butyrivibrio spp. which also contains long chain lipids (71). At first, it may seem surprising that the degree of disorder in the alkyl chains above the phase transition, as assessed by the methylene symmetric stretching frequencies, is similar in pH 3.0 and 7.0 lipids considering the presence of long chain lipids and acetal linkages between head groups in pH 3.0 lipids. However, we recall from chapter 3 that the energy of activation for gauche-trans 96 Figure 22. DSC thermograms of total lipid extracts from cells of S. ventriculi. A) Cells grown at pH 7.0 and B) pH 3.0. ———>— . Endothermic 97 4 1 Y I V I 20 4o 60 Temperature (°C) FIGURE 22 80 98 isomerism is identical in both pH 7.0 and 3.0 lipids. Therefore, one would expect that the degree of disorder in the alkyl chain (as measured by the methylene stretch frequencies which, in fact, is a measure of the amount of gauche conformers) and the temperature of the melting (DSC) transition to be similar for both type of lipid systems. However, at higher temperatures, we observed important differences in the DSC thermogram of pH 7.0 and 3.0 lipids. These changes were only weakly correlated with minor variations in the methylene stretch frequencies at similar temperatures. Consequently, explanations for the differences observed in the DSC thermograms were sought in the nature of the interactions between lipid head groups. As far as the contribution from the polar head is concerned, we can discuss lipid phase transitions in term of the free energy change at the phase transition, AGpol which is explicitly described in the following equation (15). AGpol = AGhyd + AGeI + AGbond + (1 ) where AGhyd is the contribution from hydration, AGe| is from electrostatic and AGbond is the contribution from interlipid bonds such as hydrogen bonding for example. Hydration is the dominant term followed by electrostatic and bond terms (52). The free energy difference between anhydrous and hydrated lipids is usually negative, indicating that the transition temperature of the former would be more elevated (that is, the system is more stable). The 99 thermogram of pH 7.0 lipids displays a broad transition around 40°C followed by several weaker ones at higher temperatures. However, pH 3.0 lipids exhibit no transitions between 35 and 50°C and relatively weak transitions are observed between 50 and 60°C. This difference could result from a difference in the degree of hydration and/or hydrogen bonding in the polar head group region in pH 7.0 and 3.0 lipids. It is possible to estimate the degree of hydrogen bonding at the polar-apolar interface by the analysis of the lipids ester carbonyl stretch mode. This band is located at around 1730 cm'1 and the actual value of the frequency is sensitive to hydrogen bonding (111). The ester carbonyl stretch bands for pH 7.0 and 3.0 lipids are fairly broad and are probably the convolution of two or more underlying bands (figure 23). At 10°C, the pH 7.0 lipids exhibited a symmetric band centered at around 1717 cm“. At higher temperatures, a broad shoulder appeared on the high frequency side of this band shifting the peak maximum to higher frequencies. Work on phospholipids has conclusively demonstrated that the carbonyl stretch band is composed of at least two underlying bands. The high frequency component was assigned to the free carbonyl groups while the lower frequency component was assigned to carbonyl groups involved in hydrogen bonding (111). We therefore attributed the presence of the high frequency shoulder in pH 7.0 lipids at high temperatures to a reduction in hydrogen bonding. A similar shift to higher frequencies was observed for pH 3.0 lipids but a close analysis of the spectra revealed that the relative contribution from the higher frequency component to the overall band was more important than for pH 7.0 lipids suggesting that pH 3.0 lipids at high 100 Figure 23. Infrared absorbance spectra of the carbonyl stretch mode region. A) Total lipid extract from cells of S. ventriculi grown at pH 7.0 and B) pH 3.0. i Absorbance 101 A 40°C 10°C 3770.0 3757.5 1745.0 1752.5 3720.0 1707.5 3505.0 1552.5 1570.0 T B 40 0c 10°C 3770.0 {757.5 {745.0 2752.5 {720.0 {707.5 3505.0 {552.5 3570.0 Wavenumber (cm-1) FIGURE 23 102 temperatures are almost completely devoided of hydrogen bonding in the region of the carbonyl while some hydrogen bonding persists in pH 7.0 lipids. Eventhough monoglucosyl diacylglycerol, which is proportionally more abundant in pH 3.0 lipids, has been shown to take up less water than phospholipids (112), it is not possible, from our results, to conclusively attribute the near absence of hydrogen bonding involving the ester carbonyl in pH 3.0 lipids solely to dehydration. In fact, apart from water, the hydrogen donor to the ester carbonyl could be hydroxyl groups from phosphatidylglycerol, diacyl glycerol or monoglucosyl diacylglycerol, the major lipid molecules in S.ventriculi. For example, it has been suggested that phosphatidylglycerol can form hydrogen bonds between the glycerol OH and the ester carbonyl in the alkyl chain (111). in addition to hydrogen bonding in the carbonyl region, the phase transitions could also be influenced by hydrogen bonding between polar heads, the composition of which is different in pH 7.0 and 3.0 lipids (see chapter 2). In particular, hydrogen bonding between molecules of monoglucosyl diacylglecerol (which is proportionally more abundant in pH 3.0 lipids) has been suggested to give rise to relatively high transition temperatures (57) which could partly explain the absence of major transitions in those lipids between 35- 50°C. Furthermore, the nature of hydrogen bondings involving phosphatidylglycerol (pr04 = 2.9) can be drastically modified at pH's near pro4 (88). Regardless of the nature of the hydrogen donor 103 contributing to hydrogen bonding with the ester carbonyl or the nature of hydrogen bonding between polar head groups, it is clear that pH 7.0 and 3.0 lipids are different in that respect, a fact that may contribute to the increase stability of pH 3.0 lipids at high temperature. Analysis of the width of the CH2 symmetric stretch band revealed that it decreases with increasing temperature for both pH 7.0 and 3.0 lipids (figure 24). A decrease in the width of this band is associated with a lower degree of mobility of the alkyl chains and, in phospholipids, correlates well with the gel to liquid crystalline phase transition (8). Our result is somewhat surprising considering the increase in alkyl chain disorder over the same temperature range as assessed by the frequency of the methylene stretch mode. This may indicate that the transition observed in pH 7.0 and 3.0 lipids is not a gel to liquid crystalline but some other type of phase transition. Equation 1 suggests that part of the changes in the free energy of phase transitions contributed by the polar head groups are due to the nature of bonds between lipids. Previous studies in our laboratory have documented the formation of acetal linkages between adjacent lipids on changing the pH of growth from 7.0 to 3.0 (72). Some of these lipids have been characterized and identified as the glucose acetal of phosphatidylglycerol plasmalogen (GluAPG), glycerol acetal of phosphatidylglycerol plasmalogen (GAPG) and head to head coupled glycolipids. Such linkages can profoundly affect the nature and 104 Figure 24. Bandwidth of the CH2 symmetric stretching mode band as a function of temperature. The width was measured at 75 % of the total intensity of the peak. Open symbols pH 7.0 lipids and closed symbols pH 3.0 lipids. BANDWIDTH (cm-1) 105 15 14‘ 13‘ 12“ 11' 10" Temperature (°C) FIGURE 24 6O 106 stability of lipid phases. These acetal lipids are very similar to those identified in Clostridium butyricum (89). In this organism it was shown that a relative increase in the concentration of the glycerol acetal of plasmenylethanolamine (GAPlaE) can stabilize the Iamellar phase by increasing significantly the temperature at which the transition from Iamellar to inverted hexagonal takes place (81). The main reason given to explain this stabilization is that the additional glycerol residue increases the size of the head group and according to the shape-structure concept of lipid polymorphism (see chapter 1) will stabilize the Iamellar phase. The same explanation could obviously apply to our lipid systems and explain why the less stable pH 7.0 lipids undergo several transitions at temperatures above the main transition. The very different chemical structure and molecular organization of lipids from cells grown at pH 7.0 or 3.0 prompted us to ask whether such changes would influence the number and nature of the configurations (rotational states) accessible to lipid alkyl chains. To address this question we looked at the methylene bending and the methylene wagging modes (figures 25 and 26). These modes are sensitive to the packing and to the specific configurations of the alkyl chains (110,113,114). The central band around 1468 cm'1 in the CH2 bending region is due to the methylene out-of-phase mode (109). Methylene bending frequency at around 1468 cm'1 has been found to be characteristic of disordered liquid crystalline state of lipids (115). In that region, the bands around 1440, 1455 and 1460 cm'1 have been assigned to gauche conformers. The band at 1440 107 Figure 25. Infrared absorbance spectra of CH2 bending modes region. A) Total lipid extract from cells of S. ventriculi grown at pH 7.0 and B) pH 3.0. Absorbance 1500.0 3457.5 3475.0 3452.5 108 ¥ 1 3450.0 3457.5 3425.0 3432.5 3400.0 40°C 10 °C 3500.0 3407.5 3475.0 3452.5 3450.0 3437.5 3425.0 3432.5 3400.0 Wavenumber (cm‘1) FIGURE 25 109 Figure 26. Infrared absorbance spectra of CH2 wagging modes region. A) Total lipid extract from cells of S. ventriculi grown at pH 7.0 and B) pH 3.0. 110 _l ——>— Absorbance ‘ s - + : : . ..400.0 3503.2 3552.5 3575.7 3555.0 3555.2 3547.5 3555.7 3550.0 3 3400.0 3503.2 3552.5 3575.7 3555.0 3555.2 3547.5 3555.7 3550.0 Wavenumber (cm-1) FIGURE 26 111 cm"1 has been tentatively assigned to a methylene group adjoining two gauche conformers. The shoulder at around 1460 cm'1 has been attributed to methylene groups adjoining a trans and a gauche bond (109). two additional bands at 1447 (mainly observed in pH 7.0 lipids) and 1420 cm"1 were observed but no definitive assignment was made for these bands. The abundance of these conformers in this region of the infrared spectrum changes for both pH 7.0 and 3.0 lipids as the temperature is increased (figure 25). This is expected since the number of gauche conformers increases with temperature. More interestingly, the relative abundance of these conformers was found to be different at a given temperature when pH 7.0 and 3.0 lipids were compared. This difference is made more evident in the difference spectrum that covers the scissoring and the wagging regions (figure 27) Analysis of the methylene wagging regions also revealed differences in the nature of the rotational states that are accessible to the alkyl chains of the two lipid systems. The same general increase of gauche conformers with increasing temperature for pH 3.0 and 7.0 lipids was also observed in the wagging region. In this region, specific frequencies have been associated with specific sequences of gauche-trans conformers. The strong band around 1380 cm'1 is due to the methyl umbrella vibration, and the bands at 1368, 1353 and 1341 cm'1 arise from kink (gtg' + gtg), double gauche (gg) and end gauche (eg) respectively (110). Qualitative analysis of the relative intensity of the bands revealed that thekink band at 1368 cm'1 was more intense at low temperatures in pH 3.0 lipids than in 112 Figure 27. Difference spectrum of pH 7.0 and 3.0 lipids in the region covering the bending and wagging modes. Subtraction was performed using spectra obtained at 40 °C. 113 ”.mflfln - V.dund . U.NDNH . 0.n01« hm Mazur— S.-Eov 250E552“; h.hflv« 5.0VV4 0.«h?« - - m.fl0?« 0.0«fla «.Dnnn aoueqxosqv +— 114 pH 7.0 lipids, an observation consistent with the higher degree of disorder in pH 3.0 lipids as assessed by the methylene symmetric stretch frequency (figure 21). Also, the band at 1341 cm"1 (eg) appeared to be more intense (relative to other bands in the same spectrum) in pH 7.0 lipids compare to the same band in pH 3.0 lipids, which would be consistent with the presence of transmembrane alkyl chains rendering the formation of this rotational isomer in pH 3.0 lipids less likely. The difference spectrum again revealed that the conformers accessible to pH 7.0 and 3.0 lipids in the wagging region are different. The analysis of the scissoring and wagging regions suggests that the probability of the rotational states of alkyl chains in a particular lipid system at a given temperature are different. In conclusion, we have shown that the lipids from Sarcina ventriculi grown at two different pH (pH 3.0 and 7.0) exhibit different conformations of the alkyl chains as well as the polar head region. This can be rationalized easily when one considers the presence of long chain lipids and acetal linkages in pH 3.0 lipids. The important point is that despite the different conformations accessible to thetwo lipid systems, the rate and energy of activation of gauche- trans isomerism is conserved (chapter 3). The lipid systems here studied (pH 7.0 and 3.0) can be likened to a statistical thermodynamic system in which the energy of the systems are the same but the distribution of that energy amongst the (vibrational) energy levels is different. CHAPTER 5 Endotoxic lipopolysaccharide from Salmonella typhimurium modifies lipid alkyl chain dynamics in model membranes but non- endotoxic lipopolysaccharide from Fihodopseudomonas sphaeroides does not Introduction Lipopolysaccharides (LPS) are found in the outer membrane of gram- negative bacteria where they participate in the assembly of membrane proteins and serve as a hydrophobic barrier. LPS's play a central role in microbial pathogenesis (116). They have been implicated as being responsible for triggering the host biological responses to bacterial infections. These responses involve (principally but not exclusively) the activation of macrophages by LPS which then release mediator molecules such as tumor necrosis factor, interleukins, nitric oxide, prostaglandins etc (117,116). These mediator molecules have harmful as well as beneficial effects for the host organism. Chemically, two domains can be identified in LPS. The first domain, termed lipid A, is the least variable part within the LPS molecule among different species. In Salmonella typhimurium it is 115 116 Figure 28. Structures of lipid A. A) Salmonella typhimurium and B) Fihodopseudomonas sphaeroides. Adapted from reference (118) 117 0‘1 OH HO \CH2 0 HO\ O Ho O NH O "Ho—Pi-O HO O ONHO 0 oNHo-Pi-o :0 O O OH ° 0 OH O OH OH OH 0 OH ‘0 :0 =0 ‘0 IR.) ; (Rs) (6.) (R4) 10.1 (“21 (32,) (R2) A 3 FIGURE 28 118 characterized by the presence of a 1,4'-biphosphorylated B-1,6- linked glucosamine disaccharide substituted with amide and ester- linked 3-hydroxy fatty acids carrying two acyl chains. The fatty acyl chains are between 12 and 16 carbons long. The second domain, a polysaccharide, is glycosylated to the 6 position of the non-reducing glucosamine of the lipid A through an eight carbon sugar, 3-deoxy-D- manno-2-octulosonic acid (KDO). This domain consists of a non repeating oligosaccharide which forms the core domain and a distinct oligosaccharide linked to the core and termed the O-antigen domain. The core is usually composed of less than 10 sugar residues whereas the O-antigen may contain up to 40 repeats of 3 sugar each. The whole molecule is known as LPS (figure 28). The molecular phenomena leading to biological responses triggered by LPS are largely unknown. It is known however, that the lipid A portion of LPS molecules is necessary and sufficient for cell activation (119,120), a fact that must be accounted for by any model explaining the molecular mechanism of biological response. It has been suggested by many authors that LPS is capable of binding to a variety of both membrane and serum proteins (121,116,117). Evidence mainly comes from studies employing photoactivatable LPS probes as well as binding studies on cell extracts. Several proteins exhibiting a wide range of molecular weights have thus been identified (122). The diversity of proteins capable of binding LPS may reflect the spectrum of biological responses triggered by LPS. Despite these findings, there is no direct evidences linking LPS binding to mechanisms of cell activation (such as second messengers). 119 The amphipatic character of LPS raises the possibility that the interaction with the cell could take place through intercalation of the lipid A part into the membrane. This is supported by the fact that anti-lipid A antibodies used to probe LPS structure are unable to access lipid A specific epitopes (123) and these antibodies are inefficient at blocking the effects of LPS. Also Jackson and coworkers demonstrated that interaction with macrophages did not change the correlation time of a spin-label incorporated in the sugar residues of LPS (124). Therefore, it is likely that LPS must undergo some kind of hydrophobic rearrangements in order to expose its lipid A portion which is required for biological activity. In support of this hypothesis, LPS have been shown to intercalate spontaneously into membrane bilayers (125). These authors have suggested that the intercalation of LPS in membrane bilayers takes place in two steps. The first step involves an electrostatic interaction between the charged polysaccharide head group of LPS and charged species at the surface of cells (the polar head of membrane phospholipids, carbohydrates or proteins). The electrostatic interaction may confer specificity by allowing only certain cells to recognized LPS. This primary interaction would be followed by intercalation of the hydrophobic part of LPS into the lipid bilayer. Other studies have also shown that LPS may be capable of spontaneously intercalate in the membrane bilayer (126,127). This transfer of LPS into the host cell membrane may lead to important changes in the physico- chemical properties of the membrane capable of modulating the activity of membrane proteins. In fact, studies with both artificial membrane systems and whole cells have demonstrated that LPS's are 120 capable of modifying the motional properties of phospholipids (126,128,125). In chapter 3, we have demonstrated that the energy of activation for gauche-trans isomerism in lipid alkyl chains, a fundamental physical property of the membrane, is conserved in Sarcina ventriculi despite environmentally triggered changes in lipid composition. This suggested to us that this energy of activation might be an important parameter in the control of membrane functions. Therefore, in the present chapter, we investigated the effect of LPS on gauche-trans isomerism dynamics of model membranes. Furthermore, we compared the effect of LPS from Rhodopseudomonas sphaeroides and Salmonella typhimurium which have very different lipid A structures (figure 27). H. sphaeroides LPS is non-toxic and provides a unique opportunity to correlate lipid A structure with biological activity. We present further evidences that LPS may exert their biological activities by intercalating into and modifying membrane physical properties. 121 Materials and methods Preparation of lipid dispersion. Phospholipids and Salmonella LPS were purchased from Sigma chemical Co. (St-Louis). Phosphatidylcholine (PC) from bovine heart, phosphatidylethanolamine (PE) from bovine liver and phosphatidylserine (PS) from bovine brain were mixed in the following proportion (1.2 mg PC, 0.8 mg PE, 0.3 mg P8) in their original solvent (chloroform) and dried under nitrogen and evacuated at low pressure for at least 2 hours to remove all traces of solvent. LPS were prepared as a stock solution (4 mg/mL) in D20. The dry mixture of phospholipids was resuspended in D20 with buffer (KH2PO4 1M and Na2HPO4 8 mM) and LPS was added, except for controls. The resultant mixture was then sonicated for 30 s and heated to 65°C. This cycle was repeated 3 times. Following the last cycle the mixture was evaporated at 50 °C under low pressure by rotary evaporation for 20 minutes. The dry lipids were resuspended in 0.6 ml 020 with buffer and vortexed until complete suspension was obtained. Finally the following salts were added; CaCl21mM, KCI 2.6 mM, MgCl2 0.5 mM and NaCl 137 mM. The mixture was heated again to 65°C for 10 min and maintained at 4°C overnight before T1 measurements. NMR measurements. 1H NMR measurements were performed on a Varian VXR-SOOS spectrometer operating at 500 MHz. T1 relaxation times were obtained using the 180°-t-90°-AQ pulse sequence. 122 Thirty seconds was allowed between acquisitions for full recovery of the magnetization. 123 Results and discussion The 500 MHz 1H NVIR spectrum of the aqueous suspension of phosphatidylcholine (PC), phophatidylethanolamine (PE) and phosphatidylserine (PS) (2.5 : 1.5 :1) (w :w :w) is shown in figure 29. The signals at 1.1 and 1.5 ppm were assigned to the terminal methyl and methylene protons of alkyl chains respectively. The signal at 3.4 ppm was assigned to the choline methyl protons of PC. The broad lines are typical of aqueous suspension of lipids and are due to dipolar coupling. The spin-lattice relaxation times (T1) as a function of temperature for the methylene protons are shown in figure 30. Also shown on this figure are the T1 curves of the PC : PE : PS mixture to which was added 10 % (w/w) LPS from either Salmonella typhimurium or Fihodopseudomonas sphaeroides. A global minimum was observed for all curves. The importance of a global minimum lies in the fact that it is possible to evaluate the correlation time (Tc) for the motion giving rise to relaxation. At the temperature of the minimum To = 0.7/ 000 where (00 is the Larmor frequency in radiant. At the frequency used in our experiments (500 MHz), To is equal to 2.2 X 10'10 s. The temperature at which the global minima were observed is approximately 20°C. By using the analytical procedure described in chapter 3, it was possible to obtain a theoretical fit of the experimental data (figure 31). From this analysis we can derive the energy of activation (Ea) for gauche-trans isomerism in the alkyl chains for the phospholipid 124 Figure 29. 500 MHz 1H NMR of PC:PE:PS phospholipid mixture. 125 FIGURE 29 P?!“ 0.5 .0 3 126 Figure 30. Spin-lattice relaxation time (T1) of methylene protons as a function of inverse temperature. Circles, PC:PE:PS, squares, PC:PE:PS with 10% w/w R. sphaeroides LPS and diamonds, PC:PE:PS with 10% w/ w S. typhimurium LPS. Data were fitted with a cubic spline curve. T1(S) 1.1_ 0.6 ' 127 0.9 0.8 0.7 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7 1/1 x 1000 (16‘) FIGURE 30 128 Figure 31. Theoretical fit of the methylene T1 data. Bottom curve, PC:PE:PS with 10% w/w S. typhimurium LPS, top curves, PC:PE:PS and PC:PE:PS with 10% w/w Fi. sphaeroides. Note that the top curves are superimposed. Symbols are reproduced from figure 30. 129 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7 1/7 x1000 (K‘) FIGURE 31 130 mixture with or without LPS. The Ea for PC : PE : PS and PC : PE: PS with R. sphaeroides LPS was essentially identical (5 kcal/mole). However, in the presence of S. typhimurium LPS, the Ea is increased to 6.0 kcal/mole. This analysis immediately yielded an interesting result. That is, R. sphaeroides LPS, which is endotoxically inactive (129), does not change the energy of activation of gauche-trans isomerism in our model membrane lipid system. In contrast, LPS from S. typhimurium for which endotoxicity is well documented (117), increases the Ea. These results are in good agreement with the hypothesis that LPS exert their cellular effect by modifying the membrane dynamics (130,131). The absence of effect of R sphaeroides LPS on Ea could possibly be explained by the inability of this LPS to intercalate in the phopholipid mixture. However, close inspection of the T1 vs temperature curves for both methylene and terminal methyl protons from the phospholipid mixture with and without Fi. sphaeroides LPS reveals that the temperatures at which local minima occur are different. This indicates that Rsphaeroides LPS is capable of modifying the dynamics of lipid domains (see below) suggesting that the LPS molecules interact with phospholipids. It was not possible to exclude the possibility that Fi sphaeroides LPS do not intercalate into the hydrophobic part of the phospholipid bilayer and that the changes in the T1 curves are the result of interactions at the level of the polar heads only. An increase in the energy of activation of gauche-trans isomerism implies that the probability of gauche conformers is reduced. This can be easily understood in terms of an energy diagram for the 131 rotational isomerism about a carbon-carbon bond in an alkyl chain. The trans configuration possesses a lower energy than the gauche conformer, the two being separated by an energy barrier Ea. According to the theory of the reaction rates, a higher energy barrier will translate into a lower concentration of the product which, in our case, is represented by the gauche conformer. Therefore, the presence of S. typhimurium but not H. sphaeroides, LPS increases the average alkyl chain order in phospholipids. This result is in excellent agreement with those of MacKey et al. (132) who showed an increase in the alkyl chain order of dipalmitoylphosphatidylcholine molecules in the presence of E.coli LPS using 2H NMR. Even though the exact nature of the interaction(s) between LPS and phospholipids is not known, it is possible to rationalize the above results with the available data on the physical properties of both molecules. The most straightforward explanation for the increase in the E3 of gauche-trans isomerism of phospholipid alkyl chains by S. typhimurium LPS is the difference in the intrinsic alkyl chains fluidity of the two molecules. The phospholipids mixture used in our experiments contains alkyl chains that are unsaturated and non homogeneous in length. The gel to liquid-crystalline phase transition in such mixtures, although not measured directly in the present study, is typically low (20°C or less) (133). In fact, results from this laboratory have shown that the 1H NMR spectrum of gel phase lipids is infinitely broad (due to dipolar interactions). However, we observed relatively sharp peaks in the 1H NMR spectrum of our 132 phospholipids mixture at temperatures as low as 5°C suggesting that the melting temperature is even lower. In contrast, our own determination, by DSCof the gel to liquid-crystalline transition of S. typhimurium LPS measured in conditions similar to those used in T1 experiments, yielded a value of around 35°C (data not shown). Others have found similar values for phase transitions in LPS (134). Therefore, one would expect the incorporation of S. typhimurium LPS in our phospholipids mixture to increase the average alkyl chain order. This is indeed what is observed. The inability of R sphaeroides LPS to affect the Ea can be likewise explained by the fact that non-endotoxically active LPS, which possess shorter and unsaturated chains exhibit lower transition temperature than their endotoxically active counterparts (135). Perhaps a more subtle explanation for the effect of S. typhimurium LPS can be related to the ability of these molecules to form a supramolecular hexagonal lattice (136) (figure 32). In such a lattice, each hexagonal unit is formed by 6 molecules of lipid A. Several hexagonal units can form extended hexagonal arrays by Van der Waals interactions between their alkyl chains and by electrostatic interactions mediated by divalent cations bridges between phosphate groups. Incorporation of LPS in phospholipids could lead to a new equilibrium of the Van der Waals and electrostatic interactions resulting in atwo dimensional arrangement capable of affecting the dynamics of phospholipids alkyl chains. For example, it is reasonable to envision that a redistribution of divalent cations could modify the effective area occupied by a phospholipid molecule 133 Figure 32. Model for the molecular arrangement of LPS molecules in lipid bilayers. A) Computer generated picture of the molecular arrangement of lipid A molecules. B) Possible organization of LPS in the phospholipid bilayer. 134 .. 5...... .2351... 83$ . awe...“ 3a..» a...“ 25.. 0.3.5. 5.3...“ 03¢». 00%....” 5%...» 5,3... 25.25.53»... 3.4.4. 4.3...» «new. 5.5.... 3.5.4. 0.03.325 2.4.2 .fiw Hanan an...“ fififwfik 340% «flow 5.4%... New...“ 7% . 5%...." 5.5.4... $.55 "3m? mw£&&%¢3%0.fl 5.4.4. Wkfi $.05" 5V¢$¥R #494. «new... fifiaufimfl kinaréma 25.4.24». 85.». 3.4.4. 53.... 3.4...." 2.32 .3. fin... 333585233 5.5.4» 5%..“ 35.... $0....” 5.9... 5%....“ . 4.2.. . my“? 3.3”... .3. 5.5.4.4 5%.... 2.4% 5.2.4. 023...... 52.5... 23...... 5.5...." Wm.» .25... $553... ”£53245 swans?“ .25.». 5.94.3.5»... 5...... .2... 55.94.... 8.35. 8.4% 5%.... 23¢ 3.4.4 5.4.4. 55.4.55...“ 5.9...» B 0 0 0 0 00w0.0.0.0m00 000 0.0.0. 00 0 0.00 0 0.0 0 0.00 .00m0.00 0 0 0O 0 6.6 O 00.0 0 .000. o poo 666 O 8 FIGURE 32 135 which, in turn, is known to affect the degree of order of alkyl chains (52) The formation of hexagonal units strongly depends on the chemical structure of lipid A (136,137). Although there is no data describing the supramolecular arrangement of Fl. sphaeroides LPS, its different chemical composition would probably result in favoring a different arrangement. In order to obtain dynamic information from different part of the phospholipid molecules, we also performed a T1 analysis on the terminal methyl protons (figure 33). T1 relaxation of terminal methyl protons arises from the rotation of the methyl group (104). Analysis of the methyl protons T1 data yielded an E3 of 5.3 kcal/mole for the phospholipid mixture alone and 5.5 kcal/mole for the phospholipid mixture with Fl. sphaeroides LPS (figure 34). The difference is not significative since we estimated the error on Eato be around 5%. The evaluation of the terminal methyl protons Eafor S. typhimurium was unreliable due to the absence of data below 15°C as a result of broadening of the 1H NMR signal for these protons which made the peak indistinguishable and rendered the evaluation of T1 impossible. However, the broadening of the terminal methyl protons peak is consistent with the fact that S. typhimurium LPS increases the order (reduces the rate of motions) of phospholipid alkyl chain. 136 Figure 33. Spin-lattice relaxation time (T1) of methyl protons as a function of inverse temperature. Circles, PC:PE:PS, squares, PC:PE:PS with 10% w/w Fl. sphaeroides LPS and diamonds, PC:PE:PS with 10% WM S. typhimurium LPS. Data were fitted with a cubic spline curve. T, (s) 1.1_ 0.8: 0.7: 0.6 ’ 137 0.9: 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7 117 x 1000 (K1) FIGURE 33 138 Figure 34. Theoretical fit of the methyl T1 data. - - -, PC:PE:PS, , PC:PE:PS with 10% w/w Fl. sphaeroides and — -— - -- , PC:PE:PS with 10% WM S. typhimurium. Symbols are reproduced from figure 33. (S) 139 1.3 _ l l l l l l 1.2 r ‘. : 1l1: ..‘ 1 0.9 g- 0.82— “0. '8 a —: : \:Q, ‘80 6 : - \‘ . " " .2 0.7 5- 53;~§-;g2/ : l * l l l l l 6 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7 1/T X1000 ”(1) FIGURE 34 140 In addition to the global minima, several local minima were observed in all T1 vs temperature curves. Similar T1 behavior was also observed for lipid extracts from Sarcina ventriculi (chapter 3). The local minima were attributed either to separate domains which had the same correlation time (but at different temperatures) or to a dynamic shifting of membrane domains as a function of temperature. A better understanding of the dynamics of the lipid system can be gained by a more detailed analysis of the signification of T1. Each T1 value represents a weighted average of all alkyl chains from the lipid mixture at a particular temperature. Thus, in the case where several domains are present, it is likely that To (and consequently T1) will differ for alkyl chains from different domains. In such a situation the lipid mixture as a whole will exhibit deviations from the ideal exponential behavior of To, hence the local minima and maxima. The difference in T1 values between successive maxima and minima can be better visualized by an envelope curve. In such a depiction of the T1 data, the top curve of the envelope joins successive maxima and the bottom curve of the envelope joins successive minima. The magnitude of the difference between the top and bottom curve of an envelope (the width of the envelope) is a reflection of the difference in the dynamics (Tc) of gauche-trans isomerism between domains. Envelope curves for methylene protons T1 are shown in figure 35. It is clear that the envelopes for the PC:PE:PS mixture and the mixtures containing LPS are very different. LPS from S. typhimurium increases the width of the envelope. This indicates that S. typhimurium LPS may accentuate the difference in the dynamics of gauche-trans isomerism between 141 Figure 35. Envelope. curves for the methylene protons T1 data. The filled areas represent the area between the top curve that joins successive maxima and the bottom curve that joins successive minima. Shaded area with horizontal lines, PC:PE:PS, black area, PC:PE:PS with 10% R. sphaeroides LPS and shaded area with oblique lines, PC:PE:PS with 10% S. typhimurium LPS. 142 d-—l—fi-— -—_——___— __————fl~—_ 30.6 3.6 3.5 3.4 3.3 3.2 3.1 1/T x 1000 (K-1) FIGURE 35 143 domains. In other words, S. typhimurium LPS seems to preferentially affect certain domains. In contrast, R. sphaeroides LPS was found to reduce the width of the envelope curve. Following our reasoning, this would imply that R. sphaeroides LPS reduces the disparity in the dynamics of alkyl chains between domains perhaps by making the lipid mixture more homogeneous. The analysis of the T1 data by assessing the width of the envelope curve is attractive but needs to be confirmed by other techniques that are sensitive to the difference in alkyl chains motions between lipid domains. The above results are consistent with the proposal that LPS exert their biological activity by perturbing membrane dynamics. It has been proposed however, that biological activity is dictated by the supramolecular structure of the hydrophobic part of LPS termed lipid A (135). These authors have established a strong positive correlation between the tendency of lipid A to form inverted structures and its biological activity. In their study, those lipid A with a tendency to form Iamellar structures have little biological activities. In the same study, it was pointed out that the supramolecular structure is more important in determining biological activity than the fluidity of lipid A. However, it must be kept in mind that this study was done with lipid A and not LPS which may not have the same tendency to form inverted structures. Furthermore, in the model presented by Brandenburg et al. it is not specified whether the supramolecular structure of LPS has a direct influence on the dynamics of membrane lipids or if it simply promotes the fusion of LPS with target membranes. The realization 144 that the supramolecular structure of LPS can play a role in determining its biological activity is undoubtedly important. However, we feel that the crucial question remains what will be the effect of LPS on the physical properties of the target membrane once incorporated into it. Modification of the Ea of gauche-trans isomerism could be such a property affected by LPS that would result in the disruption (or not in the case of non-endotoxicaly active LPS) of normal membrane functions. CONCLUSION The most important finding of this thesis is the fact that the energy of activation for gauche-trans isomerism (chain rotational isomerism) is conserved in the lipid alkyl chains of Sarcina ventriculi despite important changes in lipid composition and structure as a result of adaptation to low pH. The importance of this finding stems from the fact that gauche-trans isomerism is the molecular basis for the fluidity of lipid bilayers in the liquid crystalline phase. Indeed, only those theoretical models which treat chain rotational isomerism in a realistic way give a satisfactory description of the fluid phase (138,52,6,139). Furthermore, we obtained proof that changes in lipid composition alter alkyl chain packing by revealing (Chapter 4) that the relative proportion (is. the relative probability) of rotational conformers is different in the alkyl chains of lipids from Sarcina ventriculi grown at pH 7 (regular chain length lipids) or pH 3 (very long chain lipids). As mentioned in the introduction, there exist many examples of deviation from the principle homeoviscous adaptation. This can be attributed in part to the use of molecular probes to measure membrane fluidity. Each probe is sensitive to particular lipid motions and the type of motions they are sensitive to may depend on their partitioning in the membrane. Accordingly, such measurement 145 146 may have the undesirable property of depending on lipid composition. Furthermore, there is always the possibility that the probe itself will disturb the thermodynamic equilibrium of the lipid system. Failure to observe homeoviscous adaptation may also simply be because viscosity is not conserved upon changes in lipid composition as a result of adaptation to changing environment. In other words, preserving membrane viscosity may not be the goal of changes in lipid composition. In fact, the term viscosity encompasses all the possible lipid motions (lateral diffusion, flip-flop, carbon-carbon rotation etc.) (140) and it is perfectly conceivable that some or all of these motions are not conserved when dramatic structural changes take place in the membrane. What is the goal, then, of modifications of lipid composition in bacteria? Our results suggest that the energy of activation of carbon-carbon rotation in lipid alkyl chains may be a more fundamental physical property dictating the nature of lipid changes during adaptation. In our experimental system, this property was shown to be independent of lipid composition and organization. In addition, the energy of activation and correlation times of gauche- trans isomerism in our lipid system were derived using a totally non-invasive method (NMR). The general validity of these conclusions for other organisms remains to be explored. Other important results obtained in this thesis pertain to the specific mechanisms of lipid adaptation in S. ventriculi namely the synthesis of long chain lipids and the formation of acetal lipids. The 147 synthesis of long chain lipids is not unique to S. ventriculi but is also observed in a number of eubacteria. The mechanism for the synthesis is becoming clearer. There already exists indirect evidence that the synthesis is occurring in the membrane by an enzyme mediated radical mechanism (84). In chapter 2, more evidence was obtained that this is indeed the case. Furthermore, our experiments also suggest that the physical state of the membrane is the cellular sensor capable of modulating membrane enzyme activity. It is not excluded that environmental perturbations, especially temperature, may act directly on cytosolic enzymes involved in lipid biosynthesis, but the physical state of the membrane is a logical point of control since it is responsive to all environmental perturbations and in such, constitutes the perfect feedback system capable of integrating and transmitting any changes in membrane fluidity to membrane enzymes. Recently, strong experimental evidence has been obtained in support of the physical state of membrane being the cellular sensor of environmental perturbations (141,75). The presence of acetal lipids is also not unique to S. ventriculi. They have been detected in several species of clostridia (89). In Clostridium butyricum the phosphatidylglycerol (PG) acetal and glycerol acetal (GA) of plasmenylethanolamine (PlaE) have been identified. In addition, pulse-chase studies have shown that PIaE is a precursor of GAPIaE (142). This suggests that the formation of GAPIaE occurs in the membrane and that its synthesis is modulated by the physical state of the membrane. In this regard, Goldfine has 148 proposed a model in which the chemical modifications of lipids in the membrane is facilitated by the formation of microdomains of non-Iamellar lipids (81). In S. ventriculi, the formation of acetal linkages is also believed to occur in the membrane in response to changes in its physical properties (72). In chapter 2 we proposed a mechanism for their formation. These two important lipid modifications (formation of long chains and acetal linkages) as we have shown, contribute to the conservation of the energy of activation of gauche-tans isomerism. However, exactly how they do so is still unclear. One aspect of this question pertains to the effect of these changes on the lipid phase(s). Several authors have suggested that the conservation of the (Iamellar) phase is the driving force behind membrane adaptation (see chapter 1). The nature of the phase(s) adopted by lipids in S. ventriculi is unknown. However, the long chain bifunctional lipids of S. ventriculi (grown at pH 3) share some similarities with tetraether lipids found in archaebacteria for which a number of studies have described their phase behavior. Both type of lipids span the bilayer thickness and they possess asymmetry with regard to the polar heads on a given lipid molecule (143,144). Tetraether lipids has been shown to adopt a wide variety of phases near physiological conditions (145,146). Some of these phases are non-Iamellar and models have been proposed to illustrate how these phases may be important for bacterial physiology. The presence of long chain lipids in S. ventriculi could similarly give rise to non-Iamellar phases and raises the possibility that changes in lipid composition in this 149 organism would result in the disruption of lipid phases. However, the synthesis of long chain lipids in S. ventriculi is accompanied by the formation of acetal lipids which, in C. butyricum, have been suggested to stabilize Iamellar phases (81). These observations illustrate the importance of investigating the nature of the phases adopted by the lipids of S. ventriculi. It must be kept in mind that adaptation of microorganisms to changes in membrane fluidity can only be truly adaptative if it influences the competitive fitness of individuals and species and therefore is strictly meaningful only from an ecological perspective (80). However, as pointed out by Cossins and Sinensky (80) experiments designed at testing the effect of the loss of an adaptation mechanism on the competitive fitness of an organism are very difficult to perform. The difficulty is compounded by the fact that organisms living in the same environment have evolved completely different genotypes with regard to their lipid composition and adaptation to changing environments, making it difficult to assess the adaptability of organism only by looking at lipid structures. Therefore, there is a need to better understand the fundamental physical and thermodynamic basis of membrane adaptation. 150 Relevance of the present studies to microbial ecology The environment of the natural habitats of bacteria are rarely, if ever, constant. Considerable changes in temperature, pressure, salinity, pH etc. are rather common. Consequently, bacteria must adapt to survive. In particular, the membrane(s) which is at the interface between the environment and the cell interior, exhibits a remarkable adaptability capacity and important changes in lipid composition are often observed during stress. Sarcina ventriculi, is a good example of a bacterium subjected to dramatic changes in the environment of its habitats. This organism has been found in the stomach, garden soil, sand and acid peat bog (147,148,149). It is known that S. ventriculi can grow over a wide range of pH (2 to 10) which makes this organism well suited for survival in the gastrointestinal tract where changes in pH are important. Acclimatization to such variations in pH includes the synthesis of long chain lipids that stabilizes the dynamics of alkyl chains. Other organisms also found in the stomach include the facultative anaerobes streptococci and lactobacilli (150). Interestingly, Iactobacillus heterohiochii, an alcoholophilic spoilage bacterium of Japanese rice wine (Sake) has been found to synthesize long chain (up to 030) lipids (151,152) which indicates that lactobacilli could exploit the synthesis of long chain lipids to survive the changes in pH along the gastrointestinal tract in a fashion similar to S.ventriculi. Furthermore, recent results in our laboratory have shown that Helicobacter pylori , which is also found in the stomach, 151 is capable of synthesizing long chain lipids (Mindock and Hollingsworth, unpublished results). From the same genus as Sarcina ventriculi, Sarcina maxima, which is found in the outer coat of cereal grains, is also capable of growing over a wide range of pH (153). Its close phylogenetic relationship with S. ventriculi suggests that this organism may also synthesize long chain lipids in response to pH stress. However, the presence of long chain lipids is not limited to bacteria tolerant to changes in pH. Long chain dicarboxilic acids were found in the anaerobic eubacterium Thermatoga maritima which lives in geothermally heated sea floors in Italy and the Azores (66). Long chain lipids were also identified in Thermoanaerobacter ethanolicus 39E (67) which possesses an optimum growth temperature of around 67-69 °C and from Clostridium thermohydrosulfuricum. This latter organism can grow at temperatures up to 78 00 yet it has been found in a number of habitats that differ widely in temperature (from soil sample of moderate temperature, from a sewage plant in Georgia and from hot springs) (154). The membranes of the above organisms are likely to be stabilized by the presence of the long chain lipids at high temperatures. This possibility is substantiated by the fact that the membranes of archaebacteria living at high temperatures contain tetraether lipids (long C40 lipids, see chapter 1) which are believed to stabilize the dynamics of alkyl chains. For example, Methanacoccus jannaschii, an archaebacterium found in deep sea hydrothermal vents, can grow over a wide range of temperatures (47 to 75 °C) and adapts to changes in temperatures by modifying the 152 content of tetraether lipids in its membrane (higher proportion of tetraether lipids at higher temperatures) (77). Further evidence supporting the role of tetraether lipids in stabilizing membranes at high temperatures comes from the observation that Methanobacterium spp., which has an optimum growth temperature of around 65 0C contains almost two times less tetraether lipids than Methanothermus spp. which possesses an optimum growth temperature of around 85 to 90 °C (64). Bifunctional long chain lipids, very similar to those of S. ventriculi, have been identified in the fatty acid auxotroph anaerobic eubacterium Butyrivibrio spp. which is found in the rumen (155,65,96,156,85). This organism is capable of modifying the proportion of long chain lipids in its membrane when fed fatty acids differing in length and in degree of unsaturation (69,71). This adaptation mechanism probably enables Butyrivibrio to feed on different fatty acids produced in the rumen. The synthesis of long chain lipids represents only one particular mechanism available to bacteria to maintain the fluidity of their membranes within a range compatible with life. Other mechanisms and examples have been discussed in the introduction. From the above discussion, it is clear that the presence of long chain, transmembrane lipids is not limited to S. ventriculi, neither is it limited to organisms facing extremes of pH in their habitat but rather they are found in the archae domain as well as in the bacteria and in organisms facing changes in temperature, salt, pressure, carbon source etc. These observations strongly suggest that the 153 synthesis of long chain lipids is an important and wide spread mechanism of membrane adaptation to environmental stress and is highly relevant to the ecology of bacteria. BIBLIOGRAPHY BIBLIOGRAPHY Singer, 8. J. and Nicholson, G. L. (1972) The fluid mosaic model of the stucture of cell membranes.Science 175: 720-731. Welti, R. and Glaser, M. (1994) Lipid domains in model and biological membranes.Chem. Phys. Lipids 73: 121-137. Op den Kamp, J. A. F. (1979) Lipid asymmetry in membranes.Ann. Rev. Biochem. 48: 47-71. deGennes, P. G. (1974) in The physics of Iquid crystals Clarendon Press, Oxford. De Rosa, M., Gambacorta, A. and Gliozzi, A. (1986) Structure, biosynthesis, and physicochemical properties of archaebacterial lipids.MicrobioI. Rev. 50: 70-80. Marcelja, S. (1974) Chain ordering in liquid crystals. ll. structure of bilayer membranes.Biochim. Biophys. Acta 367: 165-176. lsraelachvili, J. N., Marcelja, S. and Horn, R. G. (1980) Physical principles of membrane organization.0. Rev. Biophys. 13: 121- 200. 154 10. 11. 12. 13. 14. 15. 16. 155 Mantsch, H. H. and McElhaney, R. N. (1991) Phospholipid phase transitions in model and biological membranes as studied by infrared spectroscopy.Chem. Phys. Lipids 57: 213-226. Bunow, M. R. and Levin, I. W. (1977) Comment on the carbon- hydrogen stretching vibrational Raman spectra of phospholipids.Biochim. Biophys. Acta 487: 388-394. Knowles, P. F. and Marsh, D. (1991) Magnetic resonance of membranes.Biochem. J. 274: 625-641. Lakowicz, J. R. (1983) in Principles of fluorescence spectroscopy. Plenum Press, New-York. Seelig, A. and Seelig, J. (1974) The dynamic structure of fatty acyl chains in a phospholipid bilayer measured by deuterium magnetic resonance.Biochemistry 13: 4839-4845. Davis, J. H., Jeffrey, K. R. and Bloom, M. (1978) Spin-lattice relaxation as a function of chain position in perdeuterated potassium palmitate.J. Magn. Res. 29: 191-199. Chong, P. L.-G. and Weber, G. (1983) Pressure dependence of 1,6-diphenyI-1,3,5-hexatriene fluorescence in si ngle- component phosphatidylcholine liposomes.Biochemistry. 22: 5544-5550. Cevc, G. (1987) How membrane chain melting properties are regulated by the polar surface of the lipid bilayer.Biochemisfry 26: 6305-6310. Parente, R. A. and Lentz, B. R. (1984) Phase behavior of large unilamellar vesicles composed of synthetic phospholipids.Biochemistry 23: 2353-2362. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 156 Trauble, H. and Eibl, H. (1974) Electrostatic effects on lipid phase transitions: Membrane structure and ionic environment.Proc. Natl. Acad. Sci. USA 71: 214-219. de Kruijff, B. (1987) Polymorphic regulation of membrane lipid compostion.Nature 329: 587-588. Gruner, S. M. (1992) in The structure of biological membranes. (Yeagle, P ed.), pp. 211-250, CRC Press, Boca Raton. Cullis, P. R., Hope, M. J., de Kruijff, B., Verkleij, A. J. and Tilcock, C. P. S. (1985) in Phospholipids and cellular regulation. (Kuo, J. F., eds.), pp. 1-59, CRC press, Boca Raton. Cullis, P. R., Hope, M. J. and Tilcock, C. P. S. (1986) Lipid polymorphism and the roles of lipids in membranes.Chem. Phys. Lipids 40: 127-144. McConnell, H. M., Tamm, L. K. and Weis, R. M. (1984) Periodic structures in lipid monolayer phase transition.Proc. Natl. Acad. Sci. USA 81: 3249-3253. Weis, R. M. and McConnell, H. M. (1984) Two-dimensional chiral crystals of phopholipids.Nature 310: 47-49. Peters, R. and Beck, K. (1983) Translational diffusion in phospholipid monolayers measured by fluorescence microphotolysis.Proc. Natl. Acad. Sci. USA 80: 7183-7187. Kinnunen, P. K. J. (1991) On the principle of functional ordering in biological membranes.Chem. Phys. Lipids 57: 375-399. Kinnunen, P. K. J., Koiv, A., Lehtonen, J. Y. A., Rytomaa, M. and Mustonen, P. (1994) Lipid dynamics and peripheral interactions of proteins with membrane surfaces.Chem. Phys. Lipids 73: 181-207. 27. 28. 29. 30. 31. 32. 33. 34. 35. 157 Reiss-Husson, F. and Luzzati, V. (1967) in Thermobiology (Rose, A. H., ed.), pp. 87-107, Academic press, New-York. Chapman, D. (1967) in Thermobiology (Rose, A. H., ed.), pp. 123- 146, Academic press, New-York. Sackmann, E. (1984) in Biological membranes (Chapman, D., ed.), pp. 105-143, Academic Press, New-york, NY. Luzzati, V. and Husson, F. (1962) The structure of the liquid- crystalline phases of lipid-water systems.J. Cell Biol. 12: 207-219. George, R., Lewis, R. N. A. H., Mahajan, S. and McElhaney, R. N (1989) Studies on the purified, lipid reconstituted (Na+ + Mgz+)-ATPase from Acholeplasma Iaidlawii B membranes.J. Biol. Chem. 264: 11598-11604. Martinac, B., Buechner, M., Delcour, A. H., Adler, J. and Kung, C. (1987) Pressure-sensitive ion channel in Escherichia coli.Proc. Natl. Acad. Sci. USA 84: 2297-2301. Morris, C. E. and W.J., S. (1989) Stretch-inactivated ion channels coexist with stretch-activated ion channels.Science 243: 807-809. Morris, C. E. (1990) Mechanosensitive lon channels.J. Membrane Biol. 113: 93-107. Lee, A. G. (1991) Lipids and their effects on membrane proteins: Evidence against a role for fluidity.Prog. Lipid Res. 30: 323-348. 36. 37. 38. 39. 40. 41. 42. 43. 158 Silvius, J. R., Mak, N. and McElhaney, R. N. (1980) in Membrane fluidity: Biophysical techniques and cellular regulation. (Kates, M. and Kuksis, A., eds.), pp. 213-222, Humana Press, Clifton, NJ. Zakim, D., Kavecansky, J. and Scarlata, S. (1992) Are membrane enzymes regulated by the viscosity of the membrane environment.Biochemistry 31: 11589-11594. Grainger, D. W., Reichert, A., Ringsdorf, H. and Salesse, C. (1989) An enzyme caught in action: direct imaging of hydrolytic function and domain formation of phospholipase A2 in phosphatidylcholine monolayers.FEBS Letters 252: 73-82. Sen, A., lsac, T. V. and Hui, S. W. (1991) Bilayer packing stress and defects in Mixed dilinoleoylphosphatidylethanolamine and palmitoyloleoylphosphatidylcholine and their susceptibility to phospholipase A2.Biochemistry 30: 4516-4521. Dawson, R. M. C., Irvine, R. F., Bray, J. and Quinn, P. J. (1984) Long-chain unsaturated diacylglycerols cause a perturbation in the structure of phospholipid bilayers rendering them susceptible to phospholipase attack.Biochem. Biophys. Res. Comm. 125: 836-842. Epand, R. M. and Bottega, R. (1987) Biochemistry 26: 1820- 1825. Gruner, S. M. (1985) Ann. Rev. Biophys. Biophys. Chem. 14: 211. Almeida, P. F. F., Vaz, W. L. C. and Thompson, T. E. (1993) Percolation and diffusion in three-component lipid bilayers: Effect of cholesterol on an equimolar mixture of two phosphatidylcholine.Biophys. J. 64: 399-412. 44. 45. 46. 47. 48. 49. 50. 51. 159 Muderhwa, J. M. and Brockman, H. L. (1992) Lateral lipid distribution is a major regulator of lipase activity.J. Biol. Chem. 267: 24184-24192. Jacobson, K., Sheets, E. D. and Simson, R. (1995) Revisiting the fluid mosaic model of membranes.Science 268: 1441-1442. Melo, E. C. C., Lourtie, I. M. G., Sankaram, M. 8., Thompson, T. E and Vaz, W. L. C. (1992) Effects 0 domain connection and disconnection on the yields of in-plane bimolecular reactions in membranes.Biophys. J. 63: 1506-1512. lpsen, J. H., Jorgensen, K. and Mouritsen, O. G. (1990) Density fluctuations in saturated phospholipid bilayers increase as the acyl chain length decreases.Biophys. J. 58: 1099-1107. Georgallas, A., MacArthur, J. D., Ma, X. P., Nguyen, C. V., Palmer, G. R., Singer, M. A. and Tse, M. Y. (1987) The diffusion of small ions through phospholipid bilayers.J. Chem. Phys. 86: 7218- 7226. Kanehisa, M. l. and Tsong, T. Y. (1978) Cluster model of lipid phase transitions with application to passive permeation of molecules and structure relaxations in lipid bilayers.J. Amer. Chem. Soc. 100: 424-432. Marsh, D., Watts, A. and Knowles, P. F. (1976) Evidence for phase boundary lipid. Permeability of Tempo-choline into dimyristoylphosphatidylcholine vesicles at the phase transition.Biochemistry 15: 3570-3578. Carruthers, A. and Melchior, D. L. (1983) Studies of the relationship between bilayer water permeability and bilayer physical state.Biochemistry 22: 5797-5807. 52. 53. 54. 55. 56. 57. 58. 59. 160 Cevc, G. and Marsh, D. (1987) in Cell biology: A series of monographs (Bittar, E. E., ed.) Vol. 5, pp. 441, John Wiley & Son, New-York, NY. Yagisawa, K., Naito, M., Gondaira, K.-I. and Kambara, T. (1993) A model for self-sustained potential oscillation of lipid bilayer membranes induced by the gel-liquid crystal phase transition.Biophys. J. 64: 1461-1475. Zuckermannn, M. J. (1993) Self-sustained pote ntial oscillations and the main phase transition of lipid bilayers.Biophys. J. 64: 1369-1370. Haran, A., Waldeck, D. H., Naaman, R., Moons, E. and Cahen, D. (1994) The dependence of electron transfer efficiency on the conformational order in organic monolayers.Science 263: 948-950. Teissié, J., Prats, M., LeMassu, A., Stewart, L. C. and Kates, M (1990) Lateral proton conduction in monolayers of phospholipids from extreme halophiles.Biochemistry 29: 59- 65. Boggs, J. M. (1984) in Membrane fluidity. (Kates, M. and Manson, L. A., eds.), pp. 3-53, Plenum Press, New-york, NY. Knoll, W., Schmidt, G., Rotzer, H., Henkel, T., Pfeiffer, W., Sackmann, E., Mittler-Neher, S. and Spinke, J. (1991) Lateral order in binary lipid alloys and its coupling to membrane functions.Chem. Phys. Lipids 57: 363-374. Tsien, R. W. and Tsien, R. Y. (1990) Calcium channels, stores and oscillations.Ann. Rev. Cell Biol. 6: 715-760. 60. 61. 62. 63. 64. 65. 66. 67. 161 Gicquaud, C. (1993) Actin conformation is drastically altered by direct interaction with membrane lipids: A differential scanning calorimetry study.Biochemistry 32: 11873-11877. De Mendonza, D. and Cronan, J. E. (1983) Thermal regulation of membrane lipid fluidity in bacteria.Trends Biochem. Sci. 49- 52. Sinensky, M. (1971) Temperature control of phospholipid biosynthesis in Escherichia coli.J. Bacterial. 106: 449-455. Delong, E. F. and Yayanos, A. A. (1985) Adaptation of the membrane lipids of a deep-sea bacterium to changes in hydrostatic pressure.Science 228: 1101-1103. Russell, N. J. and Fukunaga, N. (1990) A comaprison of thermal adaptation of membrane lipids in psychrophilic and thermophilic bacteria.FEMS Microbial. Rev. 75: 171-182. Clarke, N. G., Hazlewood, G. P. and Dawson, R. M. C. (1980) Structure of diabolic acid-containing phospholipids isolated from Butyrivibrio sp.Biochem J. 191: 561-569. Huber, R., Langworthy, T. A., Konig, H., Thomm, M., Woese, C. R., Sleytr, U. B. and Stetter, K. O. (1986) Thermatoga maritima sp. nov. represents a new genus of unique extremely thermophiliceubacteria growing up to 90°C.Arch. Microbial. 144: 324-333. Jung, 8., Zeikus, J. G. and Hollingsworth, R. l. (1994) A new family of very long chain a,w-dicarboxilic acids is a major structural fatty acyl component of the membrane lipids of Thermoanaerobacter ethanolicus 39E.J. Lipid Res. 35: 1057- 1065. 68. 69. 70. 71. 72. 73. 74. 75. 162 Jung, 8., Lowe, S. E., Hollingsworth, R. l. and Zeikus, J. G. (1993) Sarcina ventriculi synthesizes very long chain dicarboxylic acids in response to different forms of environmental stress.J. Biol. Chem. 268: 2828-2835. Hauser, H., Hazelwood, G. P. and Dawson, R. M. C. (1979) Membrane fluidity of a fatty acid auxotroph grown with palmitic acid.Nature 279: 536-538. Kawaguchi, A. and Seyama, Y. (1984) in Membrane fluidity (Kates, M. and Manson, L. A., eds.), pp. 279-302, Plenum Press, New-York. Hauser, H., Hazlewood, G. P. and Dawson, R. M. C. (1985) Characterization of membrane lipids of a general fatty acid auxotrophic bacterium by electron spin resonance spectroscopy and differential scanning calorimetry.Bichemistry 24: 5247- 5253. Lee, J., Jung, 8., Lowe, S., Zeikus, J. G. and Hollingsworth, R. l. (1995) A bilayer to cross-linked bipolar monolayer membrane transition, a mechanism and formula for microbial adaptation to extremes. (in press) Russell, N. J. (1984) mechanisms of thermal adaptation in bacteria: blueprints for survival.Trends Biochem. Sci. 9: 108— 112. Russell, N. J. (1984) in Membrane fluidity (Kates, M. and Manson L.A. eds.), pp. 329-347, Plenum Press, New-York. Vigh, L., Los, D. A., Horvath, l. and Murata, N. (1993) The primary signal in the biological perception of temperature: Pd- catalyzed hydrogenation of membrane lipids stimulated the expression of the desA gene in Synechacystis PC06803.Prac. Natl. Acad. Sci. USA 90: 9090-9094. 76. 77. 78. 79. 80. 81. 82. 163 Martin, C. E., Hiramitsu, K., Kitajima, Y., Nozawa, Y., Skriver, L. and Thompson, G. A. (1976) Molecular control of membrane properties during temperature acclimatation. Fatty acid desaturase regulation of membrane fluidity in acclimating Tetrahymena cells.Biachemistry 15: 5218-5227. Sprott, G. D., Meloche, M. and Richards, J. C. (1991) Proportions of diether, macrocyclic diether, and tetraether lipids in methanococcus jannaschii grown at different temperatures.J. Bacterial. 173: 3907-3910. Sinensky, M. (1974) Homeoviscous adaptation- A homeostatic process that regulates the viscosity of membrane lipids in Escherichia coli.Proc. Nat. Acad. Sci. USA 71: 522-525. Parola, A. H., lbdah, M., Gill, D. and Zaritsky, A. (1990) Deviation from homeoviscous adaptation in Escherichia coli membranes.Biophys. J. 57: 621-626. Cossins, A. R. and Sinensky, M. (1984) in Physiology of membrane fluidity (Shinitzky, M., eds.), pp. 1-20, CRC Press Inc, Boca Raton. Goldfine, H., Johnston, N. C., Mattai, J. and Shipley, G. G. (1987) Regulation of bilayer stability in Clostridium butyricum: Studies on the polymorphic phase behavior of the ether lipids.Biachemistry 26: 2814-2822. Wieslander, A., Rilfors, L. and Lindblom, G. (1986) Metabolic changes of membrane lipid composition in Acholeplasma Iaidlawii by hydrocarbons, alcohols and detergents: Arguments for effects on lipid packing.Biochemistry 25: 7511-7517. 83. 84. 85. 86. 87. 88. 89. 90. 164 Fulco, A. J. and Fujii, D. K. (1980) in Membrane fluidity: Biophysical techniques and cellular regulation. (Kates, M. and Kuksis, A., eds.), pp. 77-98, Humana Press, clifton, NJ. Jung, S. and Hollingsworth, R. l. (1994) Structures and stereochemistry of the very long bifunctional alkyl species in the membrane of Sarcina ventriculi indicate that they are formed by tail-to-tail coupling of normal fatty acids.J. Lipid Res. 35: 1932-1945. Klein, R. A., Hazlewood, G. P., Kemp, P. and Dawson, R. M. C. (1979) A new series of long-chain dicarboxylic acids with vicinal dimethyl branching found as major components of the lipids of Butyrivibrio spp.Biachem J. 183: 691-700. Lowe, S. E., Pankratz, H. S. and Zeikus, J. G. (1989) Influence of pH extremes on sporulation and ultrastructure of Sarcina ventriculi.J. Bacteriol. 171: 3775-3781. Goodwin, S. and Zeikus, J. G. (1987) Physiological adaptations of anaerobic bacteria to low pH: Metabolic control of proton motive force in Sarcina ventriculi.J. Bacterial. 169: 2150- 2157. Eibl, H. (1983) in Membrane fluidity in biology. (Aloia, R. C., ed.), pp. 217-236, Academic Press, New-York. Johnston, N. C. and Goldfine, H. (1988) Isolation and characterization of a novel four chain ether lipid from Clostridium butyricum: the phosphatidylglycerol acetal of a plasmenylethanolamine.Biachim. Biophys. Acta 961: 1-12. Cronan, J. E. (1975) Thermal regulation of the membrane lipid composition of Escherichia coli.J. Biol. Chem. 250: 7074- 7077. 91. 92. 93. 94. 95. 96. 97. 98. 99. 165 Ingram, L. O. (1976) Adaptation of membrane lipids to alcohols.J. Bacterial. 125: 670-678. Hollingsworth, R. l. (1995) Membrane lipid alkyl chain heterogeneity is a molecular transform of the available thermal energy spectrum.J. Phys. Chem. 99: 3406-3410. Mycke, B., Narjes, F. and Michaelis, W. (1987) Bacteriohopanetetrol from chemical degradation of an oil shale kerogen.Nature 326: 179-181. Ourisson, G. (1987) Bigger and better hopanoids.Nature 326: 126-127. Jung, S. and Hollingsworth, R. I. (1995) A mathematical models explaining the molecular weights and distribution of very long chain dicarboxylic acids formed during the adaptative response of Sarcina ventriculi. J. Theor. Biol. Hazlewood, G. and Dawson, R. M. C. (1979) Characteristics of a lipolytic and fatty acid-requiring Butyrivibrio sp. isolated from the ovine rumen.J. Gen. Microbial. 112: 15-27. Kushwaha, S. C., Kates, M., Sprott, G. D. and I.C.P., S. (1981) Novel polar lipids from the methanogen Methanaspirillum Hungatei GP1.Biachim. Biophys. Acta 664: 156-173. Nishihara, M., Morii, H. and Koga, Y. (1989) Heptads of polar ether lipids of an Archaebacterium, Methanobacterium thermoautotrophicum: Structure and biosynthetic relationship.Biochemistry 28: 95-102. Abragam, A. (1961) in Principles of nuclear magnetism (Birman, J., Edwards, S. F., LLewellyn Smith, C. H. and Rees, M., eds.) pp. 599, Clarendon Press, Oxford. 100. 101. 102. 103. 104. 105. 106. 107. 166 Cevc, G. (1990) Biachim. Biophys. Acta 1031: 311-382. Petersen, N. O. and Chan, 8. I. (1977) More on the motional state of lipid bilayer membranes: Interpretation of order parameters obtained from nuclear magnetic resonance experiments.Biachemistry 16: 2657-2667. Kroon, P. A., Kainosho, M. and Chan, 8. l. (1976) Proton magnetic resonance studies of lipid bilayer membranes. Experimental determination of inter- and intramolecular nuclear relaxation rates in sonicated phosphatidylcholine bilayer vesicles.Biachim. Biophys. Acta 433: 282-293. James, T. L. (1975) in Nuclear magnetic resonance in biochemistry. pp. 298-347, Academic press, New-York. Meier, P., Ohmes, E. and Kothe, G. (1986) Multipulse dynamic nuclear magnetic resonance of phospholipid membranes.J. Chem. Phys. 85: 3598-3614. Seelig, J., Tamm, L., Hymel, L. and Fleischer, S. (1981) Deuterium and phosphorous nuclear magnetic resonance and fluorescence depolarization studies of functional reconstituted sarcoplasmic reticulum membrane vesicles.Biachemistry 20: 3922-3932. Horwitz, A. F., Horsley, W. J. and Klein, M. P. (1972) Magnetic resonance studies on membrane and model membrane systems: Proton magnetic relaxation rates in sonicated lecithin dispersions.Prac. Natl. Acad. Sci. USA 69: 590-593. Asher, l. M. and Levin, I. W. (1977) Effects of temperature and molecular interactions on the vibrational infrared spectra of phospholipid vesicles.Biachim. Biophys. Acta 468: 63-72. 108. 109. 110. 111. 112. 113. 114. 115. 167 Snyder, R. G. (1967) Vibrational study of the chain conformation of the lipid n-paraffins and molten polyethylene.J. Chem. Phys. 47: 1316-1360. Snyder, R. G. (1960) Vibrational spectra of crystalline n- paraffins. Part I. methylene rocking and wagging modes.J. Mal. Spect. 4: 411-434. Maroncelli, M., Ping Qi, S., Strauss, H. L. and Snyder, R. G. (1982) Nonplanar conformers and the phase behavior of solid n- alkanes.J. Am. Chem. Soc. 104: 6237-6247. Blume, A., Hiibner, W. and Messner, G. (1988) Fourier transform infrared spectroscopy of 13‘C=O-labeled phospholpids hydrogen bonding to carbonyl groups.Biochemistry 27: 8239-8249. Wieslander, A., Ulmius, J., Lindblom, G. and Fontell, K. (1978) Water binding and phase structures for different Acholeplasma Iaidlawii membrane lipids studied by deuteron nuclear magnetic resonance and X-ray diffraction.Biachim. Biophys. Acta 512: 241-253. Casal, H. L. and McElhaney, R. N. (1990) Qunatitative determination of hydrocarbon chain conformational order in bilayers of saturated phosphatidylcholines of various chain lengths by fourier transform infrared spectroscopy.Biochemistry 29: 5423-5427. Senak, L., Davies, M. A. and Mendelsohn, R. (1991) A quantitative IR study of hydrocarbon chain conformation in alkanes and phospholipids: CH2 wagging modes in disordered bilayer and H” phases.J. Phys. Chem. 95: 2565-2571. Mantsch, H. H., Madec, 0., Lewis, R. N. A. H. and McElhaney, R. N (1985) Thermotropic phase behavior of model membranes composed of phosphatidylcholines containing iso-branched 116. 117. 118. 119. 120. 121. 122. 168 fatty acids. 2. Infrared and 31F IWIR spectroscopic studies.Biochemistry 24: 2440-2446. Morrison, D. C. (1989) The case for specific lipopolisaccharide receptors expressed on mammalian cells.Micrabial Pathogenesis. 7: 389-398. Raetz, C. R. H., Ulevitch, R. J., Wright, S. D., Sibley, C. H., Ding, A. and Nathan, C. F. (1991) Gram-negative endotoxin: an extraordinary lipid with profound effects on eukaryotic signal transduction.FASEB J. 5: 2652-2660. Qreshi, N.,Takayama, K. and Kurtz, R. (1991) Diphophoryl lipid A obtained from the nontoxic lipopolysaccharide of Rhadapseudamanas sphaeroides is an endotoxin antagonist in miceJnfect. Immun. 59: 441-444. Rietschel, E. T., Brade, H., Westphal, O. and Shiba, T. (1987) Lipid A, The endotoxic center of bacterial lipopolysaccharides: Relation of chemical structure to biological activity.Progr. Clin. Biol. Res. 231: 25-53. Galanos, C., Liideritz, O., Rietschel, E. T., Westphal, O., Brade, H., Brade, L., Freudenberg, M., Schade, U., lmoto, M., Yoshimura, H., Kusumoto, S. and Shiba, T. (1985) Synthetic and natural Escherichia caIi free lipid A express identical endotoxic activities.Eur. J. Biochem. 148: 1-5. Morrison, D. C. and Rudbach, J. A. (1981) in Contemporary topics in molecular immunology. (Maudy, W. J. and lnman, F. P., eds.) Vol. 8, pp. 187-218, Plenum, New-York. Morrison, D. C., Lei, M.-G., Kirikae, T. and Chen, T.-Y. (1993) Endotoxin receptors on mammalian cells.lmmunabialagy 187: 212-226. 123. 124. 125. 126. 127. 128. 129. 130. 169 Chia, J. K. S., Pollack, M., Guelde, G., Koles, N. L., Miller, M. and Evans, M. E. (1989) Lipopolysaccharide (LPS)-reactive monoclonal antibodies fail to inhibit LPS-induced tumor necrosis factor secretion by mouse-derived macrophages.J. Infect. Dis. 159: 872-880. Jackson, S. K., James, P. E., Rowlands, C. C. and Mile, B. (1992) Binding of endotoxin to macrophages; interactions of spin- labelled saccharide residues.Biachim. Biophys. Acta 1135: 165-170. Price, R. M. and Jacobs, D. M. (1986) Fluorescent detection of lipopolysaccharide interactions with model membranes.Biochem. Biophys. Acta 859: 26-32. Larsen, N. E., Enelow, R. l., Simons, E. R. and Sullivan, R. (1985) Effect of bacterial endotoxin on the transmembrane electrical potential and plasma membrane fluidity of human monocytes.Biach. Biaph. Acta 815: 1-8. Benedetto, D. A., Shands, J. W. and Shah, D. O. (1973) The interaction of bacterial lipopolysaccharide. with phospholipid bilayers and monolayers.Biochim. Biophys. Acta 298: 145-157. Liu, M.-S., Toyoshi, O. and Snelgrove, N. E. (1982) Changes in phase transition temperature of phospholipids induced by endotoxin.Biachem. Biophys. Acta 710: 248-251. Strittmatter, W., Weckeser, J., Salimath, P. V. and Galanos, C. (1983) Nontoxic lipopolisaccharide from Rhadapseudamanas sphaeroides ATCC 17023.J. Bacteriol. 155: 153-158. Hollingsworth, R. l. and Lill-Elghanian, D. (1990) in Cellular and molecular aspects of endotoxin reactions. (A. Nowotny, J. J. S., E.J. Ziegler, eds.), pp. 73-84, Elsevier, 131. 132. 133. 134. 135. 136. 137. 170 Larsen, N. E., Enelow, R. l., Simons, E. R. and Sullivan, R. (1985) Effect of bacterial endotoxin on the transmembrane electrical potential and plasma membrane fluidity of human monocytes.Biachim. Biophys. Acta 815: 1-8. Mackay, A. L., Nichol, C. P., Weeks, G. and Davis, J. H. (1984) A proton and deuterium nuclear magnetic resonance study of orientational order in aqueous dispersions of lipopolysaccharide and lipopolysaccharide/dipalmitoylphosphatidylcholine mixtures.Biachem. Biophys. Acta 774: 181-187. Lewis, R. N. A. H. and McElhaney, R. N. (1992) in The structure of biological membranes. (Yeagle, P., ed.), pp. 73-155, CRC Press, Boca Raton. Seydel, U. and Brandenburg, K. (1992) in Bacterial endotoxic lipopolysaccharides (Morrison, D. C. and Ryan, J. L., eds.), pp. 225-250, CRC Press, Boca Raton. Brandenburg, K., Mayer, H., Koch, M. H. J., Weckesser, J., Rietschel, E. T. and Seydel, U. (1993) Influence of the supramolecular sructure of free lipid A on its biological activity.Eur. J. Biochem. 218: 555-563. Wang, Y. and Hollingsworth, R. l. (1995) An IWIR Spectroscopy and molecular mechanics study of the molecular basis for the supramolecular structure of lipopolysaccharides.Submitted Kastowsky, M., Gutberlet, T. and Bradaczek, H. (1992) Molecular modelling of the three-dimensional structure and conformational flexibility of bacterial lipopolysaccharide.J. Bacteriol. 174: 4798-4806. 138. 139. 140. 141. 142. 143. 144. 145. 146. 147. 171 Caillé, A., Pink, D., De Verteuil, F. and Zuckermann, M. J. (1980) Theoretical models for quasi-two-dimensional mesomorphic monolayers and membrane bilayers.Can. J. Phys. 58: 581-611. Nagle, J. F. (1980) Theory of the main lipid bilayer phase transition.Ann. Rev. Phys. Chem. 31: 157-195. Smith, I. C. P. (1985) in Nuclear magnetic resonance of liquid crystals (Emsley, J. W., eds.), pp. 533-566, D. Reidel, Dordrecht. Maresca, B. and Cossins, A. R. (1993) Fatty feedback and fluidity.Nature 365: 606-607. Koga, Y. and Goldfine, H. (1984) J. Bacterial. 159: 597-604. Gliozzi, A., Rolandi, R., De Rosa, M. and Gambacorta, A. (1983) Monolayer black membranes from bipolar lipids of archaebacteria and their temperature-induced structural changes.J. Membrane Biol. 75: 45-56. Gliozzi, A., Bruno, 8., Basak, T. K., De Rosa, M. and Gambacorta, A. (1986) Organization and dynamics of polar lipids from Sulfa/abus salfataricus in bulk phases and in monolayer membranes.System. Appl. Microbial. 7: 266-271. Gulik, A., Luzzati, V., DeRosa, M. and Gambacorta, A. (1985) Structure and polymorphism of bipolar isopranyl ether lipids from Archaebacteria.J. Mal. Biol. 182: 131-149. Luzzati, V. and Gulik, A. (1986) Structure and polymorphism of tetraether lipids from Sulfa/abus salfataricus: ll. Conjectures regarding biological significance.System. Appl. Microbial. 7: 262-265. Smit, J. (1933) The biology of the sarcinae.J. Pathal. Bacteriol. 112: 618-621. 148. 149. 150. 151. 152. 153. 154. 155. 172 Canale-Parola, E. and Wolfe, R. S. (1960) Studies on Sarcina ventriculi. I. Shock culture methods.J. Bacterial. 79: 857-859. Goodwin, S. and Zeikus, J. G. (1987) Ecophysiological adaptation of anaerobic bacteria to low pH: Analysis of anaerobic digestion in acidic bog sediments.Appl. Environ. Microbial. 53: 57-64. Lowe, S. E., Jain, M. K. and Zeikus, J. G. (1993) Biology, ecology, and biotechnological applications of anaerobic bacteria adapted to environmental stresses in temperature, pH, salinity or substrate.Micrabia/. Rev. 57: 451-509. Uchida, K. (1974) Lipids of alcoholophilic lactobacilli. occurence of polar lipids with unusually long acyl chains in Lactabacillus heterahiachii.Biachim. Biophys. Acta 369: 146- 155. Uchida, K. (1974) Occurence of saturated and mono-unsaturated fatty acids with unusually-long-chains (Cgo-C3o) in Lactabacillus heterahiachii, an alcoholophilic bacterium.Biochim. Biophys. Acta 348: 86-93. Canale-Parola, E. (1986) in Bergey's manual of systematic bacteriology (Sneath, P. H. A., Mair, N. S., Sharpe, M. E. and Holt, J. G., eds.), pp. 1100-1103, The Williams and Wilkins Co, Baltimore. Wiegel, J., Ljungdahl, L. G. and Rawson, J. R. (1979) Isolation from soil and properties of the extreme thermophile Clostridium thermahydrosulfuricum.J. Bacterial. 139: 800- 810. Clarke, N. G., Hazlewood, G. P. and Dawson, R. M. C. (1976) Novel lipids of Butyrivibrio spp. Chem. Phys. Lipids 17: 222-232. 173 156. Hazlewood, G. P., Clarke, N. G. and Dawson, R. M. C. (1980) Complex lipids of a lipolytic and general-fatty-acid-requiring Butyrivibrio sp isolated from the ovine rumen.Bichem J. 191: 555-560. 111011an STATE UNIV. LIBRARIES 1|111111111111111111111”I"IIII11111111111111 31293014217248