High-sensitivity Titration Calorimetry Research Articles

The effect of binding of spider-derived antimicrobial peptides, oxyopinins, on lipid membranes

Oxyopinins (Oxki1 and Oxki2) are antimicrobial peptides isolated from the crude venom of the wolf spider Oxyopes kitabensis. The effect of oxyopinins on lipid bilayers was investigated using high-sensitivity titration calorimetry and 31P solid-state NMR spectroscopy. High-sensitivity titration calorimetry experiments showed that the binding of oxyopinins was exothermic, and the binding enthalpies (Δ H) to 1-palmitoyl-2-oleoyl- sn-glycero-3-phosphatidylcholine (POPC) small unilamellar vesicles (SUVs) were − 18.1 kcal/mol and − 15.0 kcal/mol for Oxki1 and Oxki2, respectively, and peptide partition coefficient ( K p) was found to be 3.9 × 10 3 M − 1 . 31P NMR spectra of 1,2-dielaidoyl- sn-glycero-3-phosphoethanolamine (DEPE) membranes in the presence of oxyopinins indicated that they induced a positive curvature in lipid bilayers. The induced positive curvature was stronger in the presence of Oxki2 than in the presence of Oxki1. 31P NMR spectra of phosphaditylcholine (PC) membranes in the presence of Oxki2 showed that Oxki2 produced micellization of membranes at low peptide concentrations, but unsaturated PC membranes or acidic phospholipids prevented micellization from occurring. Furthermore, 31P NMR spectra using membrane lipids from E. coli suggested that Oxki1 was more disruptive to bacterial membranes than Oxki2. These results strongly correlate to the known biological activity of the oxyopinins.

540 POSTER Action of immunotherapy with interleukin-2 on innate immunity cells in peripheral blood and in tumoral tissue of pancreatic adenocarcinoma patients

The thermodynamic parameters of interactions between six variants of the anti-idiotypic monoclonal antibody (mAb) CGP 60686 produced by the hybridoma MK 2–23 with an idiotypic mAb and five different anti-anti-idiotypic mAb were studied with high sensitivity titration calorimetry. CGP 60686 recognizes an epitope in the antigen-combining region of the human high-molecular-weight-melanoma-associated antigen (HMW-MAA)-specific mouse mAb CGP 76873 produced by the hybridoma 763.74. The five HMW-MAA-specific anti-anti-idiotypic mAbs GH 464, GH 518. GH 149, GH 386 and GH 586 were generated from mice immunization with mAb CGP 60686. All interactions between the anti-idiotypic mAb and the idiotypic mAb or the anti-anti-idiotypic mAb showed large exothermic binding enthalpies between −15 and −23 kcal/mol and binding affinities larger than 6 × 109 M−1. Four of the five anti-anti-idiotypic mAbs tested exhibited significantly higher binding enthalpies for the interaction with the anti-idiotypic than the idiotypic mAbs. Replacement of either the heavy or the light chain variable region of the anti-idiotypic mAbs with an unrelated variable region abolished the idiotype to anti-idiotype interaction. Thus, both the heavy and the light chain variable region of the anti-idiotypic mAbs are required for binding to the idiotype. The values of the binding enthalpy showed only small variations when binding of the idiotypic mAb CGP 76873 to four variants of the anti-idiotypic mAb CGP 60686 with different immunoglobulin constant regions, but identical variable regions were compared. Furthermore. Fab fragments of the idiotypic mAbs showed almost the same binding enthalpy per binding site as the whole IgG molecules. Immunoglobulin constant regions thus had little influence on the idiotype to anti-idiotype interactions. Taken together, the observed thermodynamic parameters suggest that the idiotype to anti-idiotype interactions studied here are enthalpy-driven processes with only minor entropic contributions. High sensitivity titration calorimetry was used to monitor protein-protein interactions within an anti-idiotypic antibody cascade. It was found that the direct measurement of the interaction enthalpy allowed a quantitative characterization of the binding processes studied.

Enthalpy of captopril-angiotensin I-converting enzyme binding

High-sensitivity titration calorimetry is used to measure changes in enthalpy, heat capacity and protonation for the binding of captopril to the angiotensin I-converting enzyme (ACE; EC 3.4.15.1). The affinity of ACE to captopril is high and changes slightly with the pH, because the number of protons linked to binding is low. The determination of the enthalpy change at different pH values suggests that the protonated group in the captopril-ACE complex exhibits a heat protonation of approximately −30 kJ/mol. This value agrees with the protonation of an imidazole group. The residues which may become protonated in the complex could be two histidines existing in two active sites, which are joined to the amino acids coordinated to Zn 2+. Calorimetric measurements indicate that captopril binds to two sites in the monomer of ACE, this binding being enthalpically unfavorable and being dominated by a large positive entropy change. Thus, binding is favored by both electrostatic and hydrophobic interactions. The temperature dependence of the free energy of binding Δ G° is weak because of the enthalpy-entropy compensation caused by a large heat capacity change, Δ C p=−4.3±0.1 kJ/K/mol of monomeric ACE. The strong favorable binding entropy and the negative Δ C p indicate both a large contribution to binding due to hydrophobic effects, which seem to originate from dehydration of the ligand-protein interface, and slight conformational changes in the vicinity of the active sites.

Magainin 2 amide interaction with lipid membranes: calorimetric detection of peptide binding and pore formation.

The interaction of the antibiotic magainin 2 amide (M2a) with lipid bilayers was investigated with high-sensitivity titration calorimetry. The enthalpy of transfer of the cationic M2a to negatively charged small unilamellar vesicles composed of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) and 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoglycerol (POPG) (75:25, mol/mol) was measured as delta H = -17.0 +/- 1 kcal/mol of peptide. The adsorption isotherm was determined by injecting lipid vesicles into peptide solutions at low peptide concentrations (cPo < 7 microM). The apparent partition coefficient was Kapp approximately 1.2 x 10(4) M-1 at a peptide equilibrium concentration of 1 microM but decreased with increasing peptide concentration. The hydrophobic partitioning of M2a into the lipid membrane is modulated by electrostatic effects that arise from the attraction of the positively charged peptide to the negatively charged membrane. Using the Gouy-Chapman theory to correct for electrostatic attraction, the experimental binding isotherms can be explained with an intrinsic (hydrophobic) partition coefficient of K = 55 +/- 5 M-1 and an effective peptide charge of z = 3.7-3.8. The free energy of binding is delta G = -4.8 kcal/mol. At peptide concentrations cPo > approximately 7 microM, a second effect comes into play, and the titration enthalpies can no longer be explained exclusively by peptide partitioning. The first few injections produce enthalpies of reaction which are distinctly smaller than expected from a pure partition equilibrium, followed by a series of injections with reaction heats larger than expected. After subtracting the enthalpic contribution due to partitioning, the residual enthalpies are endothermic for the first few injections, and exothermic for the consecutive steps. Furthermore, the endothermic excess heat is compensated exactly by the exothermic excess heat; i.e., the excess heat consumed in the first part of the titration experiment is returned during the second part. Endothermic excess enthalpies are observed for total molar peptide-to-lipid ratios of P/L > approximately 3.0%, whereas exothermic excess heats were seen for 0.7% < P/L < 3.0%. Below P/L < approximately 0.7%, the binding follows the partition equilibrium. Based on earlier spectroscopic evidence, it is suggested that magainin 2 amide binds to the lipid membrane and forms pores at high peptide-to-lipid ratio, this process being characterized by an endothermic reaction enthalpy. Pore formation is reversed with increasing lipid concentration, and the peptide pores disintegrate. The limiting peptide-to-lipid ratio deduced from titration calorimetry for M2a pore formation is in excellent agreement with spectroscopic methods. The enthalpy of pore formation amounts to delta H = +6.2 +/- 1.6 kcal/mol peptide or delta H approximately 25-45 kcal/mol pore if the pore is comprised of 4-7 peptide molecules.

Interaction of octyl-beta-thioglucopyranoside with lipid membranes

Octyl-beta-thioglucopyranoside (octyl thioglucoside, OTG) is a nonionic surfactant used for the purification, reconstitution, and crystallization of membrane proteins. The thermodynamic properties of the OTG-membrane partition equilibrium are not known and have been investigated here with high-sensitivity titration calorimetry. The critical concentration for inducing the bilayer <==> micelle transition was determined as cD* = 7.3 mM by 90 degree light scattering. All thermodynamic studies were performed well below this limit. Sonified, unilamellar lipid vesicles composed of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) with and without cholesterol were employed in the titration calorimetry experiments, and the temperature was varied between 28 degrees C and 45 degrees C. Depending on the surfactant concentration in the membrane, the partition enthalpy was found to be exothermic or endothermic, leading to unusual titration patterns. A quantitative interpretation of all titration curves was possible with the following model: 1) The partitioning of OTG into the membrane follows a simple partition law, i.e., Xb = Kc(D,f), where Xb denotes the molar amount of detergent bound per mole of lipid and c(D,f) is the detergent concentration in bulk solution. 2) The partition enthalpy for the transfer of OTG from the aqueous phase to the membrane depends linearly on the mole fraction, R, of detergent in the membrane. All calorimetric OTG titration curves can be characterized quantitatively by using a composition-dependent partition enthalpy of the form deltaHD(R) = -0.08 + 1.7 R (kcal/mol) (at 28 degrees C). At low OTG concentrations (R < or = 0.05) the reaction enthalpy is exothermic; it becomes distinctly endothermic as more and more surfactant is incorporated into the membrane. OTG has a partition constant of 240 M(-1) and is more hydrophobic than its oxygen-containing analog, octyl-beta-D-glucopyranoside (OG). Including a third nonionic amphiphile, octa(ethyleneoxide) dodecylether (C12EO8), an empirical relation can be established between the Gibbs energies of membrane partitioning, deltaGp, and micelle formation, deltaGmic, with deltaGp = 1.398 + 0.647 deltaGmic (kcal/mol). The partition constant of OTG is practically independent of temperature and of the cholesterol content of the membrane. In contrast, the partition enthalpy shows a strong temperature dependence. The molar specific heat capacity of the transfer of OTG from the aqueous phase to the membrane is deltaCp = -98 cal/(mol x K). The OTG partition enthalpy is also dependent on the cholesterol content of the membrane. It increases by approximately 1 kcal/mol at 50 mol% cholesterol. As the partition constant remains unchanged, the increase in enthalpy is compensated for by a corresponding increase in entropy, presumably caused by a restructuring of the membrane hydration layer.

Titration calorimetry study of an anti-idiotypic antibody cascade in a human melanoma-associated antigen system

The thermodynamic parameters of interactions between six variants of the anti-idiotypic monoclonal antibody (mAb) CGP 60686 produced by the hybridoma MK 2–23 with an idiotypic mAb and five different anti-anti-idiotypic mAb were studied with high sensitivity titration calorimetry. CGP 60686 recognizes an epitope in the antigen-combining region of the human high-molecular-weight-melanoma-associated antigen (HMW-MAA)-specific mouse mAb CGP 76873 produced by the hybridoma 763.74. The five HMW-MAA-specific anti-anti-idiotypic mAbs GH 464, GH 518. GH 149, GH 386 and GH 586 were generated from mice immunization with mAb CGP 60686. All interactions between the anti-idiotypic mAb and the idiotypic mAb or the anti-anti-idiotypic mAb showed large exothermic binding enthalpies between −15 and −23 kcal/mol and binding affinities larger than 6 × 10 9 M −1. Four of the five anti-anti-idiotypic mAbs tested exhibited significantly higher binding enthalpies for the interaction with the anti-idiotypic than the idiotypic mAbs. Replacement of either the heavy or the light chain variable region of the anti-idiotypic mAbs with an unrelated variable region abolished the idiotype to anti-idiotype interaction. Thus, both the heavy and the light chain variable region of the anti-idiotypic mAbs are required for binding to the idiotype. The values of the binding enthalpy showed only small variations when binding of the idiotypic mAb CGP 76873 to four variants of the anti-idiotypic mAb CGP 60686 with different immunoglobulin constant regions, but identical variable regions were compared. Furthermore. Fab fragments of the idiotypic mAbs showed almost the same binding enthalpy per binding site as the whole IgG molecules. Immunoglobulin constant regions thus had little influence on the idiotype to anti-idiotype interactions. Taken together, the observed thermodynamic parameters suggest that the idiotype to anti-idiotype interactions studied here are enthalpy-driven processes with only minor entropic contributions. High sensitivity titration calorimetry was used to monitor protein-protein interactions within an anti-idiotypic antibody cascade. It was found that the direct measurement of the interaction enthalpy allowed a quantitative characterization of the binding processes studied.

Thermodynamics of interaction of octyl glucoside with phosphatidylcholine vesicles: partitioning and solubilization as studied by high sensitivity titration calorimetry

The interaction of the surfactant octyl glucoside (OG) with dimyristoylphosphatidylcholine (DMPC), dipalmitoylphosphatidylcholine (DPPC), distearoylphosphatidylcholine (DSPC), and soy bean phosphatidylcholine (soy bean PC) was studied using high-sensitivity titration calorimetry. We determined the partition coefficient of OG between water and lipid bilayers and the transfer enthalpy of the surfactant by addition of lipid vesicles to OG monomers or vice versa. Comparison with the micellization enthalpy of the surfactant gives information on differences in the hydrophobic environment of OG in a liquid-crystalline bilayer or a micelle. The average partition coefficient P in mole fraction units for x e≈0.12–0.2 decreases slightly from 4152 at 27°C to 3479 at 70°C for DMPC and from 4260 to 3879 for soy bean PC, respectively. The transfer enthalpy Δ H T of OG into lipid vesicles is positive at 27°C and negative at 70°C. Its temperature dependence is larger for the incorporation of OG into DMPC than into soy bean PC vesicles. It is concluded that OG in DMPC vesicles is better shielded from water than in soy bean PC vesicles or in micelles. Titration calorimetry was also used to determine the phase boundaries of the coexistence region of mixed vesicles and mixed micelles in the systems OG/DMPC, OG/DPPC, OG/DSPC, and OG/soy bean PC vesicles at 70°C in the liquid-crystalline phase. DMPC and soy bean PC solubilization was also studied at 27°C to investigate the effect of temperature. The effective surfactant to lipid ratios at saturation, R e sat, for all PCs studied are in the range between 1.33–1.72 and the ratios at complete solubilization, R e sol, are between 1.79–3.06. At 70°C, the R e sat values decrease with increasing chain length of the saturated PC. The ratios depend also slightly on temperature and the degree of unsaturation of the fatty acyl chains. For the OG/soy bean PC system, the coexistence range for mixed vesicles and mixed micelles is larger than for the corresponding PCs with saturated chains.

Vesicle−Micelle Transformation of Phosphatidylcholine/Octyl-β-d-glucopyranoside Mixtures As Detected with Titration Calorimetry

The vesicle−micelle transition of lipid bilayers induced by the addition of the nonionic detergent octyl-β-d-glucopyranoside (OG) was studied with high-sensitivity titration calorimetry and spectroscopic methods. Sonified phospholipid vesicles composed of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) were titrated into OG solutions with concentrations between 15 and 22 mM. The critical OG concentration for bilayer micellization was determined as c*D = 15.4 ± 0.3 mM whereas the critical micellar concentration of pure OG was ccmc = 23.7 mM (in buffer). All calorimetric titrations were performed below the cmc of pure OG solutions. The titration curves can be explained by a superposition of three processes, namely (i) a partitioning of surfactant molecules into the lipid membrane, (ii) a membrane micellization, and (iii) a membrane re-formation. The basic process is a partitioning of OG molecules into the phospholipid vesicles which occurs during the whole titration experiment. It can be described according to Xb = KcD,f where Xb is the degree of binding (effective surfactant-to-lipid ratio) and cD,f is the equilibrium detergent concentration free in solution. The thermodynamic parameters were determined as K = 88 ± 3 M-1 and ΔH°D = 1.7 ± 0.4 kcal/mol for the transfer of OG from the aqueous phase to the membrane (15 mM ≤ c°D ≤ 22 mM). OG partitioning entails a bilayer micellization in the initial phase of the lipid-into-detergent titration experiment. Bilayer micellization is an endothermic process with ΔHmic = +1.85 ± 0.1 kcal/mol lipid. With continuous addition of lipid, bilayer micellization comes to a halt and is replaced by the reverse process. The demicellization process is exothermic with ΔHdemic = −1.84 ± 0.1 kcal/mol lipid. Micellization and demicellization are thus symmetrical processes. Once the free OG concentration falls below the critical limit of c*D = 15.4 mM, the titration curves follow the predictions of the partition model. A quantitative model is proposed to describe the excess heat, Δ , of the bilayer ⇄ micelle equilibrium. To a first approximation, Δ follows the symmetrical mixing behavior of two closely related nonpolar liquids. Phosphorus-31 nuclear magnetic resonance and right-angle light scattering were used to monitor the structural changes of the OG/POPC mixtures. They confirm a phase boundary between the bilayer phase and the bilayer/micelle coexistence phase at c*D = 15.4 ± 0.3 mM and Xb = 1.36 ± 0.04. The phase boundary between the bilayer/micelle coexistence phase and the pure micellar phase is found for Xb ≥ 2.8. The calorimetric titrations can be used to construct a rather precise phase diagram of the POPC/OG system as the discontinuities of the titration curves define the critical concentrations of detergent and lipid at the phase boundaries.

Heat changes in lipid membranes under sudden osmotic stress.

Closed lipid vesicles act as osmometers increasing or decreasing their volume under the influence of osmotic gradients. The enthalpy changes accompanying membrane compression or expansion have not been measured yet, and first results obtained with high-sensitivity titration calorimetry are reported here. Phospholipid vesicles suspended in and in equilibrium with an electrolyte or nonelectrolyte with a defined initial concentration of c(i), were injected into a solution with a final concentration of c(f), and the heat changes were monitored with a titration microcalorimeter. Osmotic compression (delta c = c(f) - c(i) > 0) produced an exothermic heat change with deltaH approximately -500 +/- 100 cal/mol and osmotic expansion (delta c < 0) an endothermic heat change with deltaH approximately 1000 +/- 200 cal/mol; both results normalized to a concentration gradient of delta c = 1 M NaCl. The heats of compression and expansion varied linearly with the lipid content and the size of the osmotic gradient but were independent of the vesicle size. The cubic thermal expansion coefficient alpha(v) which equals (1/V)(deltaV/deltaT)p could be derived and was found to be 1.25 x 10(-3) and 2.5 x 10(-3) K(-1) for the compressed and expanded bilayer vesicles, respectively. The entropy changes associated with compression and expansion could be estimated. Compression of the membrane led to a negative entropy change and increased the hydrocarbon chain order. Expansion of the membrane was accompanied by a positive entropy change which can be explained, in part, by more disordered hydrocarbon chains. Vesicle expansion and compression thus appear to be asymmetric as far as the thermodynamic driving force is concerned.

Paclitaxel Partitioning into Lipid Bilayers

Paclitaxel (taxol) is a diterpenoid anticancer drug with a new mechanism of cytostatic action. It is under investigation in clinical trials for treatment of various types of human cancer. A major difficulty in developing paclitaxel as a chemotherapeutic agent is its poor water solubility. In order to improve the bioavailability of paclitaxel, novel vehicle systems such as mixed micelles or liposome-based formulations are being developed. In this study we determined the partition coefficient of paclitaxel partitioning into small unilamellar lipid vesicles composed of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine using two different methods, namely high-sensitivity titration calorimetry and fluorescence spectroscopy. We measured a partition coefficient of Kp ≈ 9500 MT–1, a partition enthalpy of ∆H= –25 ± 3 kcal mol–1 and a free energy of binding of ∆G= –7.9 kcal mol–1. The binding reaction is enthalpy-driven, which can be explained by van der Waals interactions between the hydrophobic drug and the hydrophobic region of the lipid bilayer. The large AH value leads to a strong temperature dependence of the partition equilibrium. A temperature increase of 10 °C reduces the paclitaxel solubility in the lipid phase by a factor of 4.

Binding of substance P agonists to lipid membranes and to the neurokinin-1 receptor.

Three new analogues of the neuropeptide substance P (SP) were synthesized. The C-terminal message segment was made more hydrophilic in (Arg9)SP or more hydrophobic in (Nle9)SP. In (AcPro2, Arg9)SP the charge at the N-terminal address segment was reduced, while that of the message segment was increased. The rationale underlying these substitutions was to correlate the physical-chemical properties of the SP-analogues, in particular their lipid-induced conformation and membrane-binding affinity, with receptor binding and functional activity. In solution, all three analogues exhibited random coil conformations as evidenced by circular dichroism spectroscopy. Addition of SDS micelles induced partially alpha-helical structures. The same structure was also produced by negatively charged lipid vesicles for (AcPro2, Arg9)SP and (Arg9)SP whereas both alpha-helix-like structures and beta-sheet structures were observed for SP and (Nle9)SP. The measurement of the Gibbs adsorption isotherms and monolayer expansion studies provided quantitative data on the surface area requirement and on the membrane penetration area of the SP analogues. The thermodynamic parameters for lipid binding were determined with monolayer expansion for measurements and high-sensitivity titration calorimetry. The apparent binding constants, Kapp, for membranes containing 100% POPG were of the order of 10(3)- 10(5) M(-1). The binding was due to electrostatic attraction of the cationic peptides to the negatively charged membrane surface. The intrinsic (hydrophobic) binding constants, obtained after correcting for electrostatic effects, were much smaller with Kp=10+/- 1 M(-1) for (Arg9)SP, 9 +/- 1 M(-1) for (AcPro2, Arg9)SP, and 39 +/- 3 M(-1) for (Nle9)SP. The measurement of the binding affinities to the NK-1 receptor and of the in vitro activities showed that all three peptides behaved as agonists. Their binding affinity to the neurokinin-1 receptor decreased with the size of the side chains at position 9 of the amino acid sequence but was independent of the cationic charge of the peptides. The fact that even the highly charged (Arg9)SP has agonistic activity provides evidence that the binding epitope at the receptor is in a rather hydrophilic environment. This finding is in agreement with the low hydrophobic binding constants and the weak penetration of the three peptides into negatively charged membranes. It argues against a membrane mediated receptor mechanism and suggests that the agonist approaches the receptor binding, site from the aqueous phase.

Binding of hisactophilin I and II to lipid membranes is controlled by a pH-dependent myristoyl-histidine switch.

The interaction of the two N-terminally myristoylated isoforms of Dictyostelium hisactophilin with lipid model membranes was investigated by means of the monolayer expansion method and high-sensitivity titration calorimetry. The two isoforms, hisactophilin I and hisactophilin II, were found to insert with their N-terminal myristoyl residue into an electrically neutral POPC monolayer corresponding in its lateral packing density to that of a lipid bilayer. The partition coefficient for this insertion process was Kp = (1.1 +/- 0.2) x 10(4) M-1. The area requirement of the protein in the lipid membrane was estimated as 44 +/- 6 A2 which corresponds to the cross sectional area of the myristoyl moiety with an additional small contribution from amino acid side chains. The interaction of hisactophilin I (hisactophilin II) with negatively charged membrane surfaces is modulated in a pH-dependent manner by charged amino acid residues clustered around the myristoyl moiety. The electrostatic binding site consists of three lysine (one arginine and two lysine), seven (nine) histidine, and four (four) glutamic acid residues and has an isoelectric point of 6.9 (7.1). For small unilamellar POPC/POPG (75/25 mole/mole) vesicles, an apparent binding constant, K(app) = (8 +/- 1) x 10(5) M-1, was measured at pH 6.0 by means of high-sensitivity titration calorimetry. Electrostatic interactions hence increase the binding constant by about 2 orders of magnitude compared to hydrophobic binding alone. With increasing pH, the electrostatic attraction decreases and turns into an electrostatic repulsion at pH > 7.0 +/- 0.1. The area occupied by the cluster of charged residues constituting the membrane binding region was 280 +/- 20 A2 as derived from monolayer measurements in close agreement with molecular modeling data derived from the NMR structure of hisactophilin I [Habazettl et al. (1992) Nature 359, 855-858].

Self-association of β-Amyloid Peptide (1–40) in Solution and Binding to Lipid Membranes

The β-amyloid peptide (βAP), a 39 to 43 residue peptide, is the major component of Alzheimer plaques. Using circular dichroism spectroscopy, titration calorimetry, and analytical, ultracentrifugation we have analyzed the self-association of βAP(1–40) in aqueous solution and the binding of βAP(1–40) to negatively charged lipid vesicles. The CD spectra of both aggregation and membrane binding are characterized by an iso-dichroic point at 212 nm, indicating a simple two-state equilibrium for both cases. In aqueous solution βAP(1–40) exhibits a reversible, concentration-dependent random coil ⇄ β-structure transition which can be described by a cooperative aggregation model with an association constant ofs= 1.05 x 104M−1and a nucleation parameter of σ = 0.012. A similar conformational change is observed upon addition of lipid. At a given peptide concentration, the addition of negatively charged, small unilamellar vesicles also induces a conformational change from a random coil conformation to a conformation with 40 to 60% β-structure. The binding isotherm can be measured with high sensitivity titration calorimetry. It is approximately linear in the initial binding phase and exhibits an apparent saturation behaviour. The apparent binding constant decreases with concentration from Kapp≈ 2100 M−1at low concentration to 700 M−1at the highest concentration measured. Peptide penetration into the lipid membrane and peptide aggregation at the membrane surface are proposed as possible mechanisms to explain the lipid-induced random coil ⇄ β-structure transition.

Thermodynamics of Micelle Formation as a Function of Temperature: A High Sensitivity Titration Calorimetry Study

Summary and Conclusions Titration calorimetry can be used as a routine method for the determination of all thermodynamic quantities for the micellization of surfactant^.'.^^^^-'^ Particularly, the availability of power compensated microtitration calorimeters with high sensitivity has improved the speed and precision of the method.I6 The cmc and the micellization enthalpy can be determined from a single experiment. The temperature dependence of these quantities is easily accessible in the temperature range between 0 and 80 “C. From the temperature dependence of the demicellization enthalpy, udemic, the heat capacity change, ACp,demic, can be directly determined and thus information on the change in hydrophobic contacts upon demicellization can be estimated. The temperature at which the cmc minimum occurs can be determined with great precision from the temperature where AHdemic is zero. The shape of the titration curves also contains information on the aggregation number. A simulation of the titration curves for deoxycholate was attempted using a simple mass action model with a low aggregation number and the association constant and three enthalpic terms as adjustable parameters. The calculations support previous findings that cholates and deoxycholates form micelles with low aggregation numbers.

Enthalpy of Antibody-Cytochrome c Binding

High-sensitivity titration calorimetry is used to measure changes in enthalpy, heat capacity, and protonation for binding of two monoclonal antibodies (MAbs) to topologically distinct surfaces of cytochrome c. MAb 2B5 binds near the exposed heme crevice in a reaction involving proton uptake, while there is no change in protonation for MAb 5F8 binding to the opposite side of the molecule. Both antibodies have association rate constants with the activation enthalpy and viscosity dependence expected of diffusion-limited reactions [Raman et al. (1992) Biochemistry 31, 10370-10379], and bind with high affinity (delta Gzerob = -12.6 kcal mol-1 for MAb 2B5 and -13.9 kcal mol-1 for MAb 5F8, at pH 7, 25 degrees C). At 25 degrees C, the equilibrium enthalpy and entropy contributions to the free energy of binding are negative for both antibodies (delta Hzerob = -21.0 kcal mol-1, delta Szerob = -28.2 cal mol-1 K-1 for MAb 2B5; and delta Hzerob = -21.7 kcal mol-1, delta Szerob = 26.3 cal mol-1 K-1 for MAb 5F8). The enthalpy of MAb 2B5-cytochrome c association exhibits a marked temperature dependence (delta Cp = -580 cal mol-1 K-1), but the enthalpy for MAb 5F8 binding is much less dependent on temperature (delta Cp = -172 cal mol-1 K-1). The large differences in delta Cp for binding of the two antibodies suggest corresponding differences in the mode of binding, or in the molecular surfaces buried in the binding reactions. In particular, factors other than hydrophobic effects may be significant contributors to the thermodynamics of antibody-cytochrome c binding, especially when delta Cp is small (MAb 5F8).

Domain formation induced by lipid-ion and lipid-peptide interactions.

High sensitivity titration calorimetry was performed for metal ions such as calcium and lanthanum and for different types of Alzheimer peptides. Ca2+ adsorbs to mixed phosphatidylcholine (PC)/phosphatidylglycerol (PG) membranes with an endothermic reaction enthalpy of delta H approximately +0.1 kcal/mol. La3+ binds to sonified PC vesicles with a reaction enthalpy of delta H approximately + 1.8 kcal/mol. The binding constants are of the order of 10 M-1 for Ca2+ and 4 x 10(3) M-1 for La3+. The role of lipids in the random coil<-->beta-sheet equilibrium of different types of Alzheimer model peptides was investigated with circular dichroism (CD) and high sensitivity titration calorimetry. Alzheimer peptide beta AP(1-40)OH and several fragments of this peptide undergo a concentration-dependent, co-operative random coil<-->beta-sheet transition in solution which can be described by a linear association model with a nucleation parameter sigma approximately 0.2-0.01 and a growth parameter s approximately 10(4) M-1. Addition of sonified lipid vesicles containing negatively charged lipids shifts the equilibrium towards the beta-sheet conformation. This can be explained by an aggregation phenomenon at the lipid/water interphase. The cationic peptides are attracted to the negatively charged membrane surface causing a local increase in peptide concentration. The high peptide concentration, together with the ordering of the peptide molecules on the membrane surface, facilitates beta-sheet formation, constituting the first experimental evidence for the induction of beta-sheet formation via the membrane surface. The binding of Alzheimer peptide fragments to the lipid membrane is accompanied by an exothermic heat of reaction with delta H in the range -2 - -8 kcal/mol.

Alzheimer beta-amyloid peptide 25-35: electrostatic interactions with phospholipid membranes.

The role of lipids in the aggregation of three Alzheimer model peptides was investigated with circular dichroism spectroscopy and high-sensitivity titration calorimetry under conditions of low ionic strength. In solution, the peptides beta AP(25-35)OH and beta AP(25-35Nle)NH2 exhibit a reversible random-coil<-->beta-sheet (or beta-structured aggregate) transition. Addition of lipid vesicles containing negatively charged lipids shifts the random-coil<-->beta-sheet equilibrium almost completely toward beta-sheet structure, which can be explained by the specific conditions created at the membrane surface: the cationic peptides are attracted to the negatively charged membrane, and the increase in peptide concentration together with the partial alignment of the peptide molecules then facilitates beta-sheet formation. The third peptide, beta AP-(25-35)NH2, also binds to the lipid membrane but was found to adopt an essentially random-coil structure, both with and without lipids. A quantitative characterization of the binding equilibrium was possible with high-sensitivity titration calorimetry. All three peptides exhibited exothermic binding enthalpies which varied between delta H approximately -2 kcal/mol for beta AP(25-35)OH and -8 kcal/mol for beta AP(25-35)NH2. The apparent binding constants, calculated with bulk concentrations, were large and varied between 500 and 5 x 10(4) M-1, depending on the experimental conditions. However, after correction for electrostatic charge effects using the Gouy-Chapman theory, the intrinsic binding constants were found to be constant and much smaller with K approximately 2-10 M-1.(ABSTRACT TRUNCATED AT 250 WORDS)

Adsorption of Ca 2+ and La 3+ to bilayer membranes: Measurement of the adsorption enthalpy and binding constant with titration calorimetry

The adsorption of Ca 2+ and La 3+ ions to the surface of lipid bilayer membranes was studied with high sensitivity titration calorimetry. Ca 2+ adsorbs to mixed phosphatidylcholine/phosphatidylglycerol membranes with a reaction enthalpy of Δ H ≈ 0.1–0.2 kcal/mol. La 3+ binds to sonified phosphatidylcholine vesicles with a reaction enthalpy of Δ H ≈ + 1.8 kcal/mol. Adsorption of La 3+ to phosphatidylcholine bilayers imparts a net positive charge to the membrane surface which makes the binding of further La 3+ increasingly more difficult. From the decreasing amplitudes in the calorimetric titration experiment a La 3+ adorption constant of K ≈ (4.1 ± 1.1)· 10 3 M −1 was evaluated. Electrostatic effects were corrected for by means of the Gouy-Chapman theory. The adsorption constant of Ca 2+ was determined previously as K ≈ 10–20 M −1 using the same binding model. Since the reaction enthalpies of Ca 2+ and La 3+ adsorption are endothermie, the adsorption of both metal ions to the membrane surface is driven by a distinct change in entropy.

Binding of the calcium antagonist flunarizine to phosphatidylcholine bilayers: charge effects and thermodynamics

We have examined the partitioning/transfer of the Ca2+ antagonist flunarizine from the aqueous phase into phospholipid bilayers. We show that the binding of the cationic amphiphilic drug flunarizine to phospholipid bilayers displays traditional linear concentration-dependent characteristics once unmasked of electrostatic effects. The coefficient for the binding/partitioning of flunarizine to phosphatidylcholine was found to be 28700 M-1, supporting the notion that this drug may be particularly membrane-active. The thermodynamics of the partitioning/transfer process have also been studied using high-sensitivity titration calorimetry. Binding was found to be predominantly enthalpy-driven with only a small entropic contribution; delta H = -22.1 kJ.mol-1 (-5.3 kcal.mol-1) at 27 degrees C. This is in conflict with established ideas of entropy-driven partitioning of drugs into phospholipid membranes as a result of the 'hydrophobic effect'. The strong enthalpic nature of binding is interpreted as being indicative of strong lipophilic interactions between the drug and the phospholipid phase.

Peptide binding to lipid bilayers. Nonclassical hydrophobic effect and membrane-induced pK shifts.

The binding of the cyclic peptide (+)-D-Phe1-Cys2-Phe3-D-Trp4-(+)-Lys5-Thr6- Cys7-Thr(ol)8, a somatostatin analogue (SMS 201-995), and the potential-sensitive dye 2-(p-toluidinyl)naphthalene-6-sulfonate (TNS) to lipid membranes was investigated with high-sensitivity titration calorimetry. The binding enthalpy of the peptide was found to vary dramatically with the vesicle size. For highly curved vesicles with a diameter of d congruent to 30 nm, the binding reaction was enthalpy-driven with delta H congruent to -7.0 +/- 0.3 kcal/mol; for large vesicles with more tightly packed lipids, the binding reaction became endothermic with delta H congruent to +1.0 +/- 0.3 kcal/mol and was entropy-driven. In contrast, the free energy of binding was almost independent of the vesicle size. The thermodynamic analysis suggests that the observed enthalpy-entropy compensation of about 8 kcal/mol can be related to a change in the internal tension of the bilayer and is brought about by an entropy increase of the lipid matrix. The "entropy potential" of the membrane may have its molecular origin in the excitation of the hydrocarbon chains to a more disordered configuration and may play a more important role in membrane partition equilibria than the classical hydrophobic effect. The binding of the peptide to the membrane surface induced a pK shift of the peptide terminal amino group. Neutral membranes were found to destabilize the NH3+ group, leading to a decrease in pK; negatively charged membranes, generated an apparent increase in pK due to the increase in proton concentration near the membrane surface. No pK shifts were seen for TNS. Titration calorimetry combined with the Gouy-Chapman theory can be used to determine both the reaction enthalpy and the binding constant of the membrane-binding equilibrium.