Analysis of protein and peptide penetration into membranes by depth-dependent fluorescence quenching: theoretical considerations.

Depth-dependent fluorescence quenching in membranes is playing an increasingly important role in the determination of the low resolution structure of membrane proteins. This paper presents a graphical way of visualizing membrane quenching caused by lipid-attached bromines or spin labels with the help of a depth-dependent fluorescence quenching profile. Two methods are presently available to extract information on membrane penetration from quenching: the parallax method (PM; ) and distribution analysis (DA; A. S. Biophys. J. 64:290a (Abstr.); A. S. Methods Enzymol. 278:462-473). Analysis of various experimental and simulated data by these two methods is presented. The effects of uncertainty in the local concentration of quenching lipids (due to protein shielding or nonideality in lipid mixing), the existence of multiple conformations of membrane-bound protein, incomplete binding, and uncertainty in the fluorescence in nonquenching lipid are described. Regardless of the analytical form of the quenching profile (Gaussian function for DA or truncated parabola for PM), it has three primary characteristics: position on the depth scale, area, and width. The most important result, not surprisingly, is that one needs three fitting parameters to describe the quenching. This will keep the measures of the quenching profile independent of each other resulting in the reduction of systematic errors in depth determination. This can be achieved by using either DA or a suggested modification of the PM that introduces a third parameter related to quenching efficiency. Because DA utilizes a smooth fitting function, it offers an advantage for the analysis of deeply penetrating probes, where the effects of transleaflet quenching should be considered.     

Refining membrane penetration by a combination of steady-state and time-resolved depth-dependent fluorescence quenching

Accurate determination of the depth of membrane penetration of a fluorescent probe, attached to a lipid, protein, or other macromolecule of interest, using depth-dependent quenching methodology is complicated by thermal motion in the lipid bilayer. Here, we suggest that a combination of steady-state and time-resolved measurements can be used to generate a static quenching profile that reduces the contribution from transverse diffusion occurring during the excited-state lifetime. This procedure results in narrower quenching profiles, compared with those obtained by traditional measurements, and thus improves precision in determination of the underlying depth distribution of the probe.     

Stable reporter cell lines for peroxisome proliferator-activated receptor γ (PPARγ)-mediated modulation of gene expression

Phosphatidylinositol (PtdIns) is phosphorylated at D-3, D-4, and/or D-5 of the inositol ring to produce seven distinct lipid second messengers known as phosphoinositides (PIs). The PI level is temporally and spatially controlled at the cytosolic face of the cellular membrane. Effectors containing PI-binding domains (e.g., PH, PX, FYVE, ENTH, FERM) associate with specific PIs. This process is crucial for the localization of a variety of cell-signaling proteins, thereby regulating intracellular membrane trafficking, cell growth and survival, cytoskeletal organization, and so on. However, quantitative assessments of protein–PI interactions are generally difficult due to insolubility of PIs in aqueous solution. Here we incorporated PIs into a lipid–protein nanoscale bilayer (nanodisc), which is applied for studying the protein–PI interactions using pull-down binding assay, fluorescence polarization, and nuclear magnetic resonance studies, each facilitating fast, quantitative, and residue-specific evaluation of the protein–PI interactions. Therefore, the PI-incorporated nanodisc could be used as a versatile tool for studying the protein–lipid interactions by various biochemical and biophysical techniques.     

Fluorescence quenching in model membranes: phospholipid acyl chain distributions around small fluorophores

Fluorescence quenching in lipid bilayers is treated by a new approach based on calculation of the probability distribution of quenching and nonquenching acyl chains around a fluorophore. The effect of acyl lattice site dependence (i.e., correlations of phospholipid sister chain occupancy of neighbor sites) was modeled by use of Monte Carlo simulations of acyl chain occupancy. This explicit accounting of site occupancy correlation was found to fit observed quenching behavior better than did a model wherein phospholipid quenchers are considered to be independent. A key aspect of this approach is to evaluate the rate for quenching in a bilayer composed of pure quenching lipid. In order to evaluate this quenching rate, and also to provide a strong test of the calculated probability distributions, we synthesized lipids with both acyl chains labeled with a quenching moiety (Br or nitroxide), as well as the more usual single-chain quenchers. The fluorescence of tryptophan octyl ester (TOE), and of the 1,6-diphenyl-1,3,5-hexatriene (DPH) derivatives trimethylammonium-DPH (TMA-DPH) and 1-lauroyl-2-(DPH-propionyl)phosphatidylcholine (DPH-PC), was examined. We obtained consistent results with all the fluorophores and quenchers indicating that up to 18 neighboring acyl sites can contribute to quenching, corresponding to two shells of acyl sites on a hexagonal lattice. Calculated discrete distributions of fluorescence intensities were converted into fluorescence lifetimes and compared with Gaussian and Lorentzian continuous lifetime distributions. The fluorescence quenching theory presented here may be used to explain quantitatively the heterogeneity of fluorophore environments in multicomponent membranes.     

Partitioning of 2,6-Bis(1H-Benzimidazol-2-yl)pyridine fluorophore into a phospholipid bilayer: complementary use of fluorescence quenching studies and molecular dynamics simulations

Successful use of fluorescence sensing in elucidating the biophysical properties of lipid membranes requires knowledge of the distribution and location of an emitting molecule in the bilayer. We report here that 2,6-bis(1H-benzimidazol-2-yl)pyridine (BBP), which is almost non-fluorescent in aqueous solutions, reveals a strong emission enhancement in a hydrophobic environment of a phospholipid bilayer, making it interesting for fluorescence probing of water content in a lipid membrane. Comparing the fluorescence behavior of BBP in a wide variety of solvents with those in phospholipid vesicles, we suggest that the hydrogen bonding interactions between a BBP fluorophore and water molecules play a crucial role in the observed “light switch effect”. Therefore, the loss of water-induced fluorescence quenching inside a membrane are thought to be due to deep penetration of BBP into the hydrophobic, water-free region of a bilayer. Characterized by strong quenching by transition metal ions in solution, BBP also demonstrated significant shielding from the action of the quencher in the presence of phospholipid vesicles. We used the increase in fluorescence intensity, measured upon titration of probe molecules with lipid vesicles, to estimate the partition constant and the Gibbs free energy (ΔG) of transfer of BBP from aqueous buffer into a membrane. Partitioning BBP revealed strongly favorable ΔG, which depends only slightly on the lipid composition of a bilayer, varying in a range from − 6.5 to − 7.0 kcal/mol. To elucidate the binding interactions of the probe with a membrane on the molecular level, a distribution and favorable location of BBP in a POPC bilayer were modeled via atomistic molecular dynamics (MD) simulations using two different approaches: (i) free, diffusion-driven partitioning of the probe molecules into a bilayer and (ii) constrained umbrella sampling of a penetration profile of the dye molecule across a bilayer. Both of these MD approaches agreed with regard to the preferred location of a BBP fluorophore within the interfacial region of a bilayer, located between the hydrocarbon acyl tails and the initial portion of the lipid headgroups. MD simulations also revealed restricted permeability of water molecules into this region of a POPC bilayer, determining the strong fluorescence enhancement observed experimentally for the membrane-partitioned form of BBP.     

Transverse distribution of alpha-tocopherol in bilayer membranes studied by fluorescence quenching

The fluorescence properties of alpha-tocopherol in a range of solvents and in micelles and membrane vesicles have been measured. In solvents the fluorescence decay was fitted by a single exponential. In bilayer membranes of dipalmitoylphosphatidylcholine or egg phosphatidylcholine the fluorescence decay was more accurately fitted as a double exponential. This may indicate that alpha-tocopherol occupies two or more sites in such membranes. Depth-dependent quenching of alpha-tocopherol fluorescence by acrylamide and some doxyl stearates has also been studied. The results confirm that in gel-phase lipid the chromanol group has a transverse distribution close to the head-group region of the lipid. In fluid phase lipid in the presence of buffer the results indicate there is more penetration of the chromanol group into the bilayer.     

Average membrane penetration depth of tryptophan residues of the nicotinic acetylcholine receptor by the parallax method.

The membrane penetration depths of tryptophan residues in the nicotinic acetylcholine receptor from Torpedo californica have been analyzed in reconstituted membranes containing purified receptor and defined lipids. Dioleoylphosphatidylcholine and three spin-labeled phosphatidylcholines with the nitroxide group at three different positions on the fatty acyl chain were used for reconstitution of the receptor. The spin-labeled phospholipids serve as quenchers of tryptophan fluorescence. Differential quenching of the intrinsic fluorescence of the acetylcholine receptor by the spin-labeled phospholipids has been utilized to analyze the average membrane penetration depth of tryptophans by the parallax method [Chattopadhyay, A., & London, E. (1987) Biochemistry 26, 39-45]. Analyses of the quenching data indicate that the tryptophan residues on the average are at a shallow location (10.1 A from the center of the bilayer) in the membrane. In addition, the generally low levels of quenching imply that the majority of tryptophan residues are located in the putative extramembranous region of the receptor. These results are consistent with several proposed models for the tertiary structure of the acetylcholine receptor and are relevant to ongoing analyses of the overall conformation and orientation of the acetylcholine receptor in the membrane.     

Fluorescence studies of the secondary structure and orientation of a model ion channel peptide in phospholipid vesicles.

A 21-residue peptide of the sequence (LSSLLSL)3 forms ion channels when incorporated into planar lipid bilayer membranes of diphytanoylphosphatidylcholine (diPhy-PC). The frequency of channel openings increases with the applied voltage gradient. We investigated the molecular and structural mechanisms underlying this voltage dependence. A series of seven peptides, each containing a tryptophan substituted for a single residue in the middle heptad, was synthesized, purified, and incorporated into small, unilamellar, diPhy-PC vesicles. We measured circular dichroism, maximum fluorescence emission wave-lengths, and fluorescence quenching by both aqueous and lipid hydrocarbon-associated quenchers. Circular dichroism spectra and the observed sequence periodicity of all fluorescence and fluorescence quenching data are consistent with an alpha-helical peptide secondary structure. Energy transfer quenching measurements using N-terminally labeled (LSSLLSL)3 co-incorporated at lipid/peptide ratios greater than 100 into vesicles with one of the Trp-substituted peptides showed that the vesicle-associated peptide, in the absence of a voltage gradient across the bilayer, exists as an equilibrium mixture of monomers and dimers. Static fluorescence quenching measurements using different lipid-bound quenchers indicate that the helical axis of a representative lipid-associated peptide is, on average, oriented parallel to the surface of the membrane and located a few angstroms below the polar head group/hydrocarbon boundary. This surface orientation for the peptide is confirmed by the complementary sequence periodicity observed for Trp fluorescence emission wavelength shifts and collisional quenching by aqueous CsCl.(ABSTRACT TRUNCATED AT 250 WORDS)     

Insights into buforin II membrane translocation from molecular dynamics simulations

Buforin II is a histone-derived antimicrobial peptide that readily translocates across lipid membranes without causing significant membrane permeabilization. Previous studies showed that mutating the sole proline of buforin II dramatically decreases its translocation. As well, researchers have proposed that the peptide crosses membranes in a cooperative manner by forming transient toroidal pores. This paper reports molecular dynamics simulations designed to investigate the structure of buforin II upon membrane entry and evaluate whether the peptide is able to form toroidal pore structures. These simulations showed a relationship between protein–lipid interactions and increased structural deformations of the buforin N-terminal region promoted by proline. Moreover, simulations with multiple peptides show how buforin II can embed deeply into membranes and potentially form toroidal pores. Together, these simulations provide structural insight into the translocation process for buforin II in addition to providing more general insight into the role proline can play in antimicrobial peptides.     

Fluorescence of membrane-bound tryptophan octyl ester: a model for studying intrinsic fluorescence of protein-membrane interactions.

The fluorescence of a membrane-bound tryptophan derivative (tryptophan octyl ester, TOE) has been examined as a model for tryptophan fluorescence from proteins in membrane environments. The depth-dependent fluorescence quenching of TOE by brominated lipids was found to proceed via a dynamic mechanism with vertical fluctuations playing a central role in the process. The activation energy for the quenching was estimated to be 1.3 kcal/mole. The data were analyzed using the distribution analysis (DA) method, which extends the conventional parallax method to account more realistically for the transbilayer distributions of both probe and quencher and for possible variations in the probe's accessibility. DA provides a better fit than the parallax method to data collected with TOE in membranes formed of lipids brominated at either the 4,5, the 6,7, the 9,10, or the 11,12 positions of the sn-2 acyl chain. DA yields information on the fluorophore's most probable depth in the membrane, its conformational heterogeneity, and its accessibility to the lipid phase. Previously reported data on cytochrome b5 and melittin were reanalyzed together with data obtained with TOE. This new analysis demonstrates conformational heterogeneity in melittin and provides estimates of the freedom of motion and exposure to the lipid phase of membrane-embedded tryptophans of cytochrome b5.     

Use of pyrene as a luminescence indicator of the viscosity of model and biological membranes

The pyrene movement in a lipid bilayer has been shown to occur not only in the lateral but also transmembrane direction. Within the excited state lifetime the pyrene monomer elevates from the depth to the polar regions of the membrane and emits a luminescence photon. The excimer does not exhibit any marked transmembrane movement while luminescing from the hydrophobic regions. The luminescence quenching efficiency of monomers and excimers depends on the depth of quencher penetration into the membrane. In the lipid bilayer the pyrene luminescence is strongly quenched by molecular oxygen. The pyrene binding to membrane proteins protects it from quenching. A conclusion has been made that the carrying out estimations of membrane viscosity from pyrene luminescence require considerable correction.     

Joint determination by Brownian dynamics and fluorescence quenching of the in-depth location profile of biomolecules in membranes

The in-depth molar distribution function of fluorophores is revealed by a new methodology for fluorescence quenching data analysis in membranes. Brownian dynamics simulation was used to study the in-depth location profile of quenchers. A Lorentzian profile was reached. Since the Stern-Volmer equation is valid at every depth in the membrane for low quencher concentrations, the molar distribution of the fluorophore (also regarded as a Lorentzian) can be achieved. The average location and the broadness of the fluorophore distribution can be calculated. The importance of the knowledge of the location width is demonstrated and discussed, since this parameter reveals important conclusions on structural features of the interaction of membranes with probes and biomolecules (e.g., conformational freedom in proteins), as well as photophysical properties (e.g., differential fluorophore quantum yields). Subsequent use of this methodology by the reader does not, necessarily, involve the performance of simulations and is not limited to the use of Lorentzian function distributions.     

The transverse location of tryptophan residues in the purple membranes of Halobacterium halobium studied by fluorescence quenching and energy transfer

Fluorescence quenching by a series of spin-labelled fatty acids is used to map the transverse disposition of tryptophan residues in bacteriorhodopsin (the sole protein in the purple membranes of Halobacterium halobium). A new method of data analysis is employed which takes into account differences in the uptake of the quenchers into the membrane. Energy transfer from tryptophan to a set of $\text{n-(9-}\text{anthroyloxy)}$ fatty acids is used as a second technique to confirm the transverse map of tryptophan residues revealed by the quenching experiments. The relative efficiencies of quenching and energy transfer obtained experimentally are compared with those predicted on the basis of current models of bacteriorhodopsin structure. Most of the tryptophan fluorescence is located near the surface of the purple membrane. When the retinal chromophore of bacteriorhodopsin is removed, tryptophan residues deep in the membrane become fluorescent. These results indicate that the deeper residues transfer their energy to retinal in the native membrane. The retinal moiety is therefore located deep within the membrane rather than at the membrane surface.     

Isoform-specific membrane insertion of secretory phospholipase A2 and functional implications

Despite increasing evidence that the membrane-binding mode of interfacial enzymes including the depth of membrane insertion is crucial for their function, the membrane insertion of phospholipase A(2) (PLA(2)) enzymes has not been studied systematically. Here, we analyze the membrane insertion of human group IB PLA(2) (hIBPLA(2)) and compare it with that of a structurally homologous V3W mutant of human group IIA PLA(2) (V3W-hIIAPLA(2)) and with a structurally divergent group III bee venom PLA(2) (bvPLA(2)). Increasing the anionic charge of membranes results in a blue shift of the fluorescence of Trp(3) of hIBPLA(2), a decrease in quenching by acrylamide, and an increase in enzyme activity, reflecting an enhancement in the membrane binding of PLA(2). Fluorescence quenching by brominated lipids indicates significant penetration of Trp(3) into fluid POPC/POPG membranes but little insertion into the solid DPPC/DPPG membranes. Increased membrane fluidity also supports hIBPLA(2) activity, suggesting that membrane insertion of hIBPLA(2) is controlled by membrane fluidity and is necessary for the full activity of the enzyme. Trp fluorescence quenching of the V3W-hIIAPLA(2) and bvPLA(2) by water- and membrane-soluble quenchers indicates substantial membrane insertion of Trp(3) of V3W-hIIAPLA(2), similar to that found for hIBPLA(2), and no insertion of tryptophans of bvPLA(2). Our results provide evidence that (a) structurally similar group IB and IIA PLA(2)s, but not structurally diverse group III PLA(2), significantly penetrate into membranes; (b) membrane insertion is controlled by membrane fluidity and facilitates activation of IB and IIA PLA(2)s; and (c) structurally distinct PLA(2) isoforms may employ different tactics of substrate accession/product release during lipid hydrolysis.     

Structural Plasticity in the Topology of the Membrane-Interacting Domain of HIV-1 gp41

We use a number of computational and experimental approaches to investigate the membrane topology of the membrane-interacting C-terminal domain of the HIV-1 gp41 fusion protein. Several putative transmembrane regions are identified using hydrophobicity analysis based on the Wimley-White scales, including the membrane-proximal external region (MPER). The MPER region is an important target for neutralizing anti-HIV monoclonal antibodies and is believed to have an interfacial topology in the membrane. To assess the possibility of a transmembrane topology of MPER, we examined the membrane interactions of a peptide corresponding to a 22-residue stretch of the MPER sequence (residues 662–683) using fluorescence spectroscopy and oriented circular dichroism. In addition to the previously reported interfacial location, we identify a stable transmembrane conformation of the peptide in synthetic lipid bilayers. All-atom molecular dynamics simulations of the MPER-derived peptide in a lipid bilayer demonstrate a stable helical structure with an average tilt of 24 degrees, with the five tryptophan residues sampling different environments inside the hydrocarbon core of the lipid bilayer, consistent with the observed spectral properties of intrinsic fluorescence. The degree of lipid bilayer penetration obtained by computer simulation was verified using depth-dependent fluorescence quenching of a selectively attached fluorescence probe. Overall, our data indicate that the MPER sequence can have at least two stable conformations in the lipid bilayer, interfacial and transmembrane, and suggest a possibility that external perturbations can switch the topology during physiological functioning.     

Does diabetes mellitus affect diphenylhexatriene penetration into erythrocyte membrane ghosts?

Diphenylhexatriene transverse distribution has been studied in normal and diabetic erythrocyte membrane ghosts using fluorescence polarization and fluorescence quenching methods. Acrylamide quenched the fluorescence of diphenylhexatriene according to a dynamic mechanism in agreement with Stern-Volmer equation. Nonlinear least-squares analysis based on quenching results has shown greater accessibility of fluorophore to quencher molecules in diabetic ghosts (37.2 +/- 3.2% in normal vs. 67.5 +/- 6.4% in diabetic membranes). Steady-state fluorescence anisotropy measurements evidenced the lowered membrane lipid fluidity in diabetics (anisotropy values: 0.166 +/- 0.011 in normal subjects vs. 0.193 +/- 0.018 in diabetics). A model mechanism is proposed which attributes the lowered capacity of lipid bilayer in diabetes to the increased ordering and more compact structure of membrane phospholipids. The implications of the results for the resolving of steady-state anisotropy data are discussed.     

Chlorinated hydrocarbon-cell membrane interactions studied by the fluorescence quenching of carbazole-labeled phospholipids: Probe synthesis and characterization of the quenching methodology

In recognition of the need to understand better the interactions of the chlorinated hydrocarbon insecticides with cell membranes we investigated the use of fluorescence quenching of membrane-bound fluorophores by these chlorinated hydrocarbons. An extensive survey of potential fluorophores identified the N-alkyl derivatives of carbazole as being especially suitable fluorophores. The fluorescence emission of these derivatives is quenched by a wide variety of commonly-used chlorinated hydrocarbons. This quenching is collisional and does not result in significant photodecomposition.Four structurally distinct carbazole-labeled phospholipids were synthesized, and their structures were confirmed by 270 MHz proton NMR and by chromatographic and chemical means. The carbazole moiety of each labeled phospholipid should be localized at a different depth in lipid bilayer. However, water soluble quenchers indicate that the fluorophores are inaccessible to the aqueous phase, irrespective of their point of attachment to the phospholipids.When incorporated into lipid bilayers, the fluorescence lifetime of these carbazole-labeled phospholipids reveals the collisional frequency between the fluorophore and the chlorinated hydrocarbon. As a result quenching of membrane-bound fluorophores may be used to measure: (1) the diffusional rate of the chlorinated hydrocarbon in the bilayer; (2) the lipid-water partition coefficient; (3) the maximum binding capacity of the membrane for the chlorinated hydrocarbon. Examples of all these measurements are given, and the fluorometric results are confirmed by direct chemical analysis.     

Superquenching as a detector for microsphere-based flow cytometric assays.

BACKGROUND: Fluorescent conjugated polymers display high fluorescence quantum yields and enhanced sensitivity to quenching (superquenching) by oppositely charged quenchers through energy or electron transfer. Fluorescent polymers and their quenchers are used in bead-based biosensor applications where the polymers are coated on particles. In this work, we investigate a detection method that utilizes superquenching on microspheres, which can be used for flow cytometric assays. METHODS: Microspheres were coated with the fluorescent cationic polyelectrolyte poly(p-phenylene-ethynylene) (PPE), and its superquenching by 9,10-anthraquinone-2,6-disulfonic acid (AQS) was examined by fluorometric methods in presence and in absence of a barrier to superquenching in the form of an anionic lipid bilayer. RESULTS: Flow cytometry detected superquenching of PPE on microspheres (MS-PPE) by AQS where high levels of reduction in fluorescence were observed. Adding different concentrations of AQS to MS-PPE yielded a Stern-Volmer quenching constant of 0.8x10(6) M-1. While forming an anionic lipid bilayer around the MS-PPE acted as a barrier to superquenching by AQS, disrupting the lipid bilayer allowed superquenching to take place. CONCLUSIONS: The sensitivity of flow cytometry in detecting fluorescence of microspheres and the amplified quenching sensitivity of fluorescent conjugated polymers both offer advantages over other fluorometric methods and conventional quenching detection. This study used superquenching of fluorescent polymers as a new tool in flow cytometry, thus combining the advantages offered by both method and detector. In addition, we employed the formation and the disruption of a supported lipid bilayer in mediating superquenching to offer new biosensing applications.     

Bilayer interactions of pHLIP, a peptide that can deliver drugs and target tumors

The pH-dependent insertion of pHLIP across membranes is proving to be a useful property for targeting acidic tissues or tumors and delivering drugs attached to its C-terminus. It also serves as a model peptide for studies of protein insertion into membranes, so further elucidation of the insertion mechanism of pHLIP and its features is desirable. We examine how the peptide perturbs a model phosphatidylcholine membrane and how it associates with the lipid bilayer using an array of fluorescence techniques, including fluorescence anisotropy measurements of TMA-DPH anchored in bilayers, quenching of pHLIP fluorescence by brominated lipids and acrylamide, and measurements of energy transfer between aromatic residues of pHLIP and TMA-DPH. When pHLIP is bound to the surface of bilayers near neutral pH, the membrane integrity is preserved whereas the elastic properties of bilayers are changed as reported by an increase of membrane viscosity. When it is inserted, there is little perturbation of the lipids. The results also suggest that pHLIP can bind to the membrane surface in a shallow or a deep mode depending on the phase state of the lipids. Using parallax analysis, the change of the penetration depth of pHLIP was estimated to be 0.4 Å from the bilayer center and 2.8 Å from the membrane surface after the liquid-to-gel phase transition.     

Lipid specific penetration of melittin into phospholipid model membranes

The relative depth of penetration of melittin into egg phosphatidylcholine and bovine heart cardiolipin model membranes was investigated using fluorescence spectroscopy techniques. The tryptophan intrinsic fluorescence shift suggests a more hydrophobic surrounding of this residue in cardiolipin, while the accessibility for charged and uncharged aqueous quenchers is decreased in the cardiolipin system when compared with the phosphatidylcholine-bound situation. A lipid incorporated hydrophobic, collisional quencher and a resonance energy transfer acceptor on the other hand are more effective in quenching the tryptophan fluorescence of cardiolipin bound melittin. The combination of these results is interpreted as prove of a deeper positioning of the tryptophan containing part of the peptide molecule in the cardiolipin system in comparison with the situation in phosphatidylcholine. Models that take this difference into account are presented, which try to explain the opposite effect of melittin binding to the two lipid systems with respect to supramolecular structure, as reported in the preceding article (Batenburg, A.M., Hibbeln, J.C.L., Verkleij, A.J. and De Kruijff, B. (1987) Biochim. Biophys. Acta 903, 142-154).