A new technique for probing inter-membrane interactions

It remains an impor tant and challenging problem to characterize the interactions between biomembranes. In vitro, nonspecific interactions can lead to membrane-bound states, fusion, and budding (cf. Lipowsky, 1995; Chernomord ik and Zimmerberg, 1995; and references therein) . In addition, living cells possess a variety of protein-based mechanisms for performing and regulating such tasks as endocytosis, exocytosis, fusion, and adhesion. Whether or not cells utilize nonproteinbased mechanisms, the underlying (nonspecific) interactions will be present and must be elucidated in the course of unraveling the complexities of the cell membrane. The general features of membrane interactions are well understood. The free energy governing the conformations of single bilayers contains curvature, stretching, and entropic contributions. In termembrane interactions are mediated by the ambient solvent and include Van der Waals (fluctuating dipoles), electrostatics (charge), hydration (water ordering) , and Helfrich entropic repulsion (excluded volume interaction between fluctuating membranes) . Dissolved ions affect the electrostatic interaction through screening, condensation, and (as we see below) cross-bridges between bilayers. Scattering and force measurement techniques have gone far in characterizing these interactions (cf. Israelachvili and Wennerstr6m, 1996; and Safinya, 1989), but it has proved difficult to explore the detailed structure in the region of contact between membranes. New approaches are needed. An impressive step in this direction is described in the article by Niles et al. (1996) in this issue of the Journal of GeneralPhysiology. These investigators use resonance energy transfer (RET) between donor and acceptor fluorophores to gain information on the distribution o f membrane separation. The donor is incorporated into the membrane of a vesicle that contains the negatively charged phosphatidyl serine (in addition to o ther lipid components) ; the acceptor is incorporated into a planar bilayer that also contains phosphatidyl serine. To optimize delivery, vesicles are ejected from a pipette that is placed near the planar membrane. At solvent calcium concentrations o f a few millimolar or greater, vesicles are bound to the planar membrane, whereas in the absence of divalent cations, vesicles are unbound. The acceptor f luorophores are excited by the donors with a probability that falls off with the sixth power of their separation. The characteristic length at which the probability of resonance transfer is 0.5 is ~ 3 nm. Nevertheless, for separations below ~35 rim, RET fluorescence provides a sensitive measure of distance. RET thus produces a two-dimensional map of intermembrane separation in a relatively noninvasive manner; no solid probes or substrates are needed. The local distance sensitivity is limited by the intensity calibration (>0.5 nm for distances >-1.0 rim); the lateral distance resolution is limited by diffraction and imaging (~0.5 I~m). Niles et al. find that increasing [Ca 2+] up to ~ t 5 mM increases the area of contact and decreases the distance between membranes monotonically, presumably as a result of Debye screening. Beyond ~15 mM [Ca2+], however, the contact area decreases; the decrease is ascribed to charge reversal. Throughout , punctate regions, which are areas of close membrane apposition (<2 nm), are observed, with the density of punctate sites increasing with increasing [Ca2+]. They suggest that these sites correspond to calcium cross-bridges between phosphatidyl serine headgroups in the apposed membranes. From the intensity autocorrelat ion function of the fluorescence images, Niles et al. also find an apparent order in the ar rangement of punctate sites. Surprisingly, the autocorrelat ion function is not rotationally symmetric. These punctate sites are reminiscent of other mobile junct ional contacts between membranes such as gapjunct ion channels and ligand receptor -media ted membrane attachments (Chiruvolu et al. 1994). A theoretical analysis shows that there is a strong tendency for such junct ional contacts to aggregate (R. Bruinsma et al., 1994), and it is possible that o rdered structures will arise. An alternative interpretat ion is that the pipette ejection of vesicles might produce strong lubrication forces and a hydrodynamic instability, which could result in order ing of the punctate sites. It is more difficult