When developing a new pharmaceutical agent, tests are performed to determine its mode of action and whether it can be safely and efficiently delivered to the site where this action is carried out. Often, the site of action is at the cell surface, where the pharmaceutical agents (or "drugs") interact with proteins in the plasma membrane. The drugs may, for example, serve as substrates for enzymatic reactions, or initiate a biochemical cascade by binding to cell surface receptors.

The site of action may also be within the cell. In this event, the drug must traverse the outer plasma membrane and interact with components of the cytoplasm or with interior membranes, such as the nuclear membrane.

Alternatively, the site of action may be within specific membrane-bound organelles.

Many of the drugs that exert their actions by interacting with cellular membranes are difficult to study. These drugs are often amphipathic, having hydrophilic and hydrophobic regions, which cause them to form micelles above a certain concentration (the critical micellar concentration, or CMC) in a given fluid, e.g., a buffer. It is difficult to study drugs at concentrations below the CMC because these concentrations are typically in the subnanomolar range, where quantification is difficult and imprecise.

All biological membranes are similar in that they consist of lipid and protein molecules that are held together primarily by noncovalent interactions that impart a dynamic or fluid character to the membrane. The lipid molecules within the membrane are arranged as a

bilayer that forms a relatively impermeable barrier-to most water-soluble molecules. It is known that the rate with which molecules enter a membrane depends on the structure of the molecule, but there is no sure or simple way to predict, based on structural features, whether or not a molecule will pass through a membrane or interact with any of the components of the membrane.

In complex organisms, lipid bilayers severely restrict the entry of drugs into several specific organs or physiological systems, notably the brain. Thus, biological membranes are an important consideration in determining whether a pharmaceutical agent will be useful in the treatment of plants and animals, including humans, farm animals, and domestic pets.

Summarv of the Invention The invention features novel methods and compositions for examining the association of a test molecule with a membrane. The test molecule can associate with the membrane in a number of ways, for example, by entering and accumulating in the membrane, by crossing the membrane, by serving as a substrate for an enzymatic reaction that is catalyzed by proteins that are located within the membrane or that are localized to the inner or outer surfaces of the membrane, by binding to and modifying the activity of a receptor present in the membrane, or by altering the properties or function of the lipid molecules within the membrane. The method can be used, for example, to examine the rate with which the test molecule associates with or accumulates in the membrane, or the amount of the test molecule that associates with or accumulates in the membrane. This information is useful in determining whether a drug will effectively reach its site of action (as described further below). The method can be carried out by

examining the test molecule in a membranous system that includes a mimetic or a natural biological membrane (e.g., a lipid-based vesicle or a biological cell) and a reporter molecule. The reporter molecule provides a signal that changes in the presence of the test molecule.

Accordingly, the invention features a method for examining the association of a test molecule with a membrane by obtaining a membranous system that includes lipid molecules and a reporter molecule, applying the test molecule to the system, and measuring the signal generated by the reporter molecule. An alteration of the signal indicates that the test molecule associates with or has accumulated in the membrane. To determine the rate with which the test molecule associates with the membrane, the signal can be examined over time; the rate of change in the signal over time corresponds to the rate with which the test molecule associates with or accumulates in the membrane.

The test molecule can be a naturally occurring molecule or a synthetic molecule, and can be applied to the membranous system in a bodily fluid, such as saliva, whole blood, urine, or cerebrospinal fluid. The test molecule can also be administered to the membranous system with a carrier molecule.

The methods described above can be used to determine the movement of the test molecule anywhere within the membranous system. For example, one can assess movement of the test molecule: (1) from the inner bilayer of a membrane to the outer bilayer of the same membrane, (2) from the outer bilayer of a membrane to the inner bilayer of the same membrane, (3) from an aqueous medium to a hydrophobic medium, (4) from a first hydrophobic medium to a second hydrophobic medium, (5) from a first aqueous medium to a hydrophobic medium to a second aqueous medium.

Numerous molecules, including those listed in the detailed description of the invention, can serve as reporter molecules. Fluorophores, such as diphenylhexatriene and green fluorescent protein, are particularly useful. The reporter molecule can be or can include a lipid, protein, or carbohydrate.

The membranous system can include one or more lipids that may be present in the form of a bilayer or monolayer. Alternatively, the membranous system can include the plasma membrane of a biological cell, the membrane of an organelle, or a nuclear membrane of a biological cell. These membranes can be harvested from plants or animals or from plant or animal cells that have been placed in tissue culture. It is also within the scope of the invention to perform the method described above in vivo or in situ by allowing the membrane to remain intact within a plant or animal. Whenever a living cell is employed (as opposed to a membrane fraction or an artificial membrane), the membranous system can be formed by transfecting the cell with a construct that encodes a reporter molecule. The construct may also encode a heterologous molecule, such as a protein.

As described fully below the method can be used to assess the ability of a test molecule to cross or otherwise associate with particular physiological barriers. For example, if the membranous system includes a lipid monolayer or an alveolar cell, it can be used as a model of a membrane at the air-water interface of the lung; if the membranous system includes a capillary endothelial cell of the brain, it can be used as a model of the blood-brain barrier; if the membranous system includes a placental villi cell, it can be used as a model of the barrier between the maternal circulatory system and the fetal circulatory system; and if the

membranous system contains a dermal or epidermal cell, it can be used as a model of the skin.

The invention also features a method of determining whether a first molecule modulates the rate with which a second molecule associates with a membranous system. This method includes three steps: (1) applying the second molecule to the system and measuring the signal generated by the reporter molecule, (2) applying the second molecule to the system in the presence of the first molecule and measuring the signal generated by the reporter molecule, and (3) determining whether the signal generated in the first step differs from the signal generated in the second step. A difference in the signal indicates that the first molecule modulates the rate with which the second molecule associates with the membranous system.

A test molecule "associates with" a membrane if the test molecule enters the membrane, accumulates in the membrane, crosses the membrane, or otherwise interacts with any of the components of the membrane, which can include lipids, phospholipids, proteins, and carbohydrates.

The new methods have several advantages. For example, they are inexpensive to perform (in part because they require only micromolar quantities of a test molecule), they can be performed quickly and are amenable to robotics, they can be used to measure both extremely rapid and extremely slow rates of drug entry into a membrane (i.e., half-times of entry from milliseconds to many hours), they require no hazardous substances (such as radioactivity), and the components (for example, "fluorosomes") can be easily made and stored for long periods of time. Further, the new methods and compositions can be used to test the ability of molecules to traverse or otherwise interact with particular

membranes, such as those present within the blood-brain barrier, the barrier between maternal and fetal circulation,>and the skin.

The ability to rapidly and accurately determine the rate with which molecules enter or traverse cell membranes is important, not only because it contributes to understanding their mode of action, but because it can be used to improve the process by which these molecules are developed, modified, and used as pharmaceuticals.

For example, a determination can be made early in the course of drug screening as to which candidate molecules are likely to reach their sites of action. Similarly, the methods and compositions described herein will expedite screening of molecules that have been chemically modified in attempts to alter the features that previously limited their medical usefulness.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Although methods and materials similar or equivalent to those described herein can be used in the practice or testing of the present invention, the preferred methods and materials are described below. All publications, patent applications, patents, and other references mentioned herein are incorporated by reference in their entirety. In case of conflict, the present specification, including definitions, will control. In addition, the materials, methods, and examples are illustrative only and not intended to be limiting.

Other features and advantages of the invention will be apparent from the following detailed description, and from the claims.

Brief Description of the- Drawinqs Fig. 1 is a tracing depicting the intensity of fluorescence - (y axis) at various wavelengths (nM; x axis) when lipid-based vesicles containing a fluorescent reporter molecule (fluorosomes) were excited at 356 nM and their emission spectra scanned in a spectrophotometer.

Fig. 2 is a plot of the emission intensity (fluorescence; y axis) of unilamellar vesicles formed from egg phosphatidylcholine (100 HM) and diphenylhexatriene (DPH; 10 M). The excitation wavelength was 356 nM and fluorescence emission was measured at 433 nM. Progressive amounts of a test molecule, tetrahydrocannabinol (THC; x axis) were added to the cuvette from a 2.5 mM ethanolic THC solution.

Figs. 3A-3C are a series of tracings showing the results of typical stopped-flow spectrophotometric measurements made over time (seconds; x axis) on the THO analogues WIN-55,212-2 (Fig. 3A) and ASTHC (Fig. 3B), and the alcohol 1-octanol (Fig. 3C).

Figs. 4A-4K are a series of diagrams illustrating the chemical structure of exemplary test molecules.

These include many of the test molecules examined in the experiments described herein, i.e., A8-THC (Fig. 4A), CP-55,940 (Fig. 4B), WIN-55,212-2 (Fig. 4C), Anandamide (Fig. 4G), A8-THC-DMH (Fig. 4F), AM835 (Fig. 41), and chloramphenicol (Fig. 4K). In addition, the related molecules pravadoline (Fig. 4D), SR141716A (Fig. 4E) R-methanandamide (Fig. 4H), and diazepam (Fig. 4J) are also shown.

Detailed Description Described herein are methods and compositions for measuring the amount of accumulation as well as the rate with which a test molecule becomes associated with a

lipid-based membrane that contains a reporter molecule.

The membrane and reporter molecule constitute a "membranous -system, which can be constructed as follows.

The Membranous Svstem Either mimetic (i.e., artificial) or naturally occurring biological membranes can be used to perform the methods described herein. The membranes can be those of a plant or animal cell, or can be constructed to mimic membranes in plant or animal cells. The plant cell membrane can be the membrane of an edible crop (for example, a corn, wheat, carrot, potato, tomato, alfalfa, or onion plant), a plant that bears fruit (for example, an apple, pear, kiwi, banana, apricot, melon, or citrus plant), or a decorative plant (for example, a rose, iris, orchid, carnation, lily, or tulip). The animal cell membrane can be the membrane of a bacterial cell, a fungal cell (including a yeast cell), the cell of a higher eukaryote (for example, a worm, fly, amphibian, fish, or crustacean cell), or a mammalian cell (such a mouse, rat, rabbit, guinea pig, cat, dog, horse, pig, sheep, cow, chimpanzee, baboon, or human cell).

Furthermore, as discussed below, the membrane can be (or can mimic) that of a particular type of cell within a complex organism (for example, a hepatocyte, lymphocyte, dermal cell, neuron, fibroblast, glial cell, chondrocyte, osteocyte, renal cell, or muscle cell). In addition, the membrane can be (or can mimic) the plasma membrane of the cell or a membrane within the interior of the cell (for example, the mitochondrial or nuclear membrane).

The amphipathic nature of membrane lipid molecules causes them to assemble spontaneously into bilayers, even in simple artificial conditions. For example, when added to water. Methods for constructing mimetic membranes are

well known (for example, see Ceve and Marsh, "Phospholipid Bilayers: Physical Principles and Models," John Wiley ~Sons, New York, NY, 1987; Bangham, Models of Cell Membranes In "Cell Membranes: Biochemistry, Cell Biology, and Pathology, G. Weissmann and R. Claiborne, eds., Hospital Practice, New York, NY, 1975; and J. H.

Fendler, "Membrane Mimetic Chemistry," John Wiley & Sons, New York NY, 1982). Briefly, methods for constructing mimetic membranes include the following: solubilizing lipids in a detergent and subsequently removing the detergent by dialysis; sonicating a suspension of hydrated lipids; dissolving lipids in a solvent such as ether, injecting that solution into water, and subsequently evaporating the ether under vacuum; extruding a suspension of lipids through a French press or similar device; extruding a suspension of lipids through a porous filter; dissolving lipids in organic solvents, adding that solution to a buffer or to water, sonicating the mixture to form a gel, removing the solvent under vacuum, and adding water. The types and quantities of lipids that can be used to form mimetic membranes by the processes given above vary tremendously.

Any type of lipid or any combination of lipids can be used in subnanomolar to gram quantities. For convenience, the concentration of lipids used to construct a mimetic membrane is typically in the nanomolar to micromolar range.

All biological membranes are composed primarily of lipids. This is true for the plasma membrane and the internal membranes of eukaryotic cells, the plasma membrane of bacteria, and the membranes of plant cells.

Bacterial plasma membranes are often composed of one main type of phospholipid, while the plasma membranes of most eukaryotic cells contain a variety of phospholipids, primarily phosphatidylcholine, sphingomyelin,

phosphatidylserine, and phosphatidylethanolamine. -The approximate content of these lipids within many different types of cell membranes is known. For example, see the content of lipids in hepatocytes, erythrocytes, glial cells, mitochondria, the endoplasmic reticulum, and the bacterium E. coli, shown in Table 1 (adapted from "Molecular Biology of the Cell," 2nd ed., Alberts et al., Eds. Chapter 6, Garland Press, New York, NY, 1989).

Thus, it is possible to mimic a particular biological membrane by selecting the appropriate phospholipid or an appropriate combination of phospholipids.

TABLE 1 II Percentage of Total Lipid by Weight LIPID Hepato- Erythro- Myelin Mito- Endo- E. cyte cyte chond- plasmic coli rion Ret. Cholestrol 17 23 22 3 6 0 Phosphatidyl- 7 18 15 35 j 17 70 ethanol amine Phosphatidyl- 4 7 9 2 5 trace serine Phosphatidyl- 24 17 10 39 40 0 choline Sphingomyelin 19 18 8 0 5 | 0 Glycolipids 7 3 28 trace trace 0 Others 22 13 8 21 27 30 If necessary or desired, an even closer approximation of the membrane of a given cell can be made by incorporating particular membrane-associated proteins that are normally highly expressed by that cell into the mimetic membrane. For example, receptors for interleukins can be inserted into mimetic T cell membranes, receptors for neurotransmitters can be inserted into mimetic neuronal membranes, and chloride channels can be inserted into mimetic respiratory epithelial cell membranes.

Alternatively, the membranous system can include a natural biological membrane. Natural membranes can be obtained either from cells taken directly from a plant or animal, or from plant or animal cells that have been placed in tissue culture. Numerous methods for preparing membrane fractions from living cells are well known (for example, see J. L. Hall and A. L. Moore, "Isolation of Membranes and Organelles from Plant Cells," Academic Press, 1983; and R. B. Gennis, "Biomembranes," pp. 13-20, Springer-Verlag, New York, NY, 1989). Briefly, the cells are disrupted (e.g., by exerting osmotic pressure, by sonication, by homogenization with a Dounce or Potter- Elbehgen glass teflon tissue homogenizer, or by grinding in the presence of an abrasive substance such as sand, alumina, or glass beads), or nitrogen cavitation, and membranes are separated from other cellular components by standard techniques such as density gradient centrifugation (with, e.g., ficoll, metrizamide, percoll, sorbitol, or mannitol), phase-partitioning, continuous free flow electrophoresis, and affinity absorption. In the event membranes are being prepared from bacterial cells, the cell wall can be digested with a degradative carbohydrate-specific enzyme, such as lysozyme.

The use of natural biological membranes may be especially useful in determining whether a test molecule will associate with a particular type of cellular membrane in vivo. It is known that certain membranes effectively impede the access of pharmaceutical agents to their sites of action. It is difficult, for example, for many drugs to cross the blood-brain barrier or to be absorbed through the skin or the air-water interface in the lungs. There is also a great deal of concern whenever a drug is administered to an expectant mother.

In developing new drugs, it is important to understand

whether or not they have the ability to cross from the maternal circulation to the fetal circulation.

The methods and compositions described herein are useful in determining whether a drug will be able to cross the barriers described above (and any other physiological barrier). To make these determinations, one would simply carry out the methods, as described herein, using an appropriate membrane. For example, to determine whether a drug would cross the blood-brain barrier, the barrier between the maternal and fetal circulation, or the skin, one would simply harvest membranes from capillary endothelial cells of the brain, the placenta, and the dermis, respectively. Similarly, the ability to effectively administer a drug via the pulmonary system can be determined by harvesting an alveolar cell. Skilled artisans are well aware of the basic physiological functions of the cells within plants and animals and would be able to determine whether or not a cell could serve as a reasonable model of any given physiological barrier.

In addition, the methods of the invention can be performed on a natural membrane of a living cells, such as a cell of a plant or animal that has been placed in culture, or a cell that remains intact within the plant or animal (i.e., the membranous system can be constructed in vivo). Methods have been developed that allow direct and repeated visualization of particular cells within living animals. For example, a quantitative fluorescence-imaging technique has been used to study changes in the amount of fluorescently label acetylcholine receptors in living muscles over long periods of time (Turney et al., J. Neurosci. Meth.

64:199-208, 1996). This particular method compensates for spatial and temporal variations in image brightness due to the light source, microscope, and camera employed,

and was used to study a single neuromuscular junction in a living mouse by periodically viewing the same junction over three weeks time. For additional guidance on obtaining quantitative images of fluorescently labeled cells in situ, skilled artisans are directed to Wigston (J. Neurosci. 10:1753-1761, 1990), Wigston (J. Neurosci.

9:639-647, 1989), van Mier et al. (J. Neurosci. 14:5672- 5686, 1994), Balice-Gordon et al., (J. Neurosci. 13:834- 855, 1993), and Lichtman et al. (New Biol. 1:75-82, 1989) To perform the present method in vivo, one would simply place the subject (for example, a small plant, a portion of a plant, or an animal) under the lens of a microscope so that the cell membranes of interest are exposed (i.e., any biological material that covers the membranes, such as skin or viscera would be incised or retracted). The membranes can be associated with a reporter molecule at that time (see below). Association with lipophilic reporter molecules would be particularly straight-forward because these molecules integrate quickly into the exposed cell membrane(s).

If necessary, the reporter molecule can be administered before the cells are placed in a position to be viewed. This administration can be carried out physically (i.e., the reporter molecule can be brought into physical contact with the cellular membrane), or genetically. For example, it is possible to genetically modify living cells so that they express a fluorescent protein, for example, green fluorescent protein (GFP), as described in Chalfie et al., Science, 263:802-805, 1994 and Chalfie, Photochem. and Photobiol. 62:651-656, 1995) Several GFPs are known but the most extensively studied are those from the bioluminescent coelenterates Aequorea (A-GFP) and Renilla (R-GFP), which can be made

and will fluoresce in many cell types and cellular compartments. The first demonstrations of heterologous expression were in E. coli and the nematode C. elegans.

In C. elegans, a GFP cDNA was expressed as a reporter gene using a promoter utilized in only a few nerve cells.

The GFP fluorescence extended throughout the cytoplasm of the nerve cells, illuminating the axonal processes and the cell bodies. GFP has also been expressed as part of a chimeric polypeptide. The first chimeras were fusions of GFP with the product of the exuperantia (exu) gene of Drosophila, which is required for mRNA localization in the developing oocyte (Wang et al., Nature 369:400-403, 1994). Skilled artisans will be aware of numerous genetic constructs that can be used to target GFP to particular membranous compartments. For example, GFP fused to a G protein would be targeted to the inner face of the plasma membrane, while fusing it to the transmembrane domain of the G protein-coupled receptor itself would result in localization of GFP within the lipid bilayer.

Biological cells can be genetically modified in vitro or in vivo. Methods for transfecting cells, including those that produce stably transfected cell lines and transgenic animals are well known and widely practiced. If necessary, skilled artisans can refer to the following publications for guidance: Sambrook et al., "Molecular Cloning, A Laboratory Manual," 2nd Ed.

Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY 1989; Ruther et al., EMBO J. 2:1791, 1983; Inouye et al., Nucleic Acids Res. 13:3101-3109, 1985; Van Heeke et al., J. Biol. Chem. 264:5503-5509, 1989; U.S. Patent No. 4,873,191; Van der Putten et al., Proc. Natl. Acad.

Sci. USA, 82:6148, 1985; Thompson et al., Cell 56:313, 1989; Lo, Mol. Cell. Biol. 3:1803, 1983; Lasko et al., Proc. Natl. Acad. Sci. USA 89:6232, 1992; Hogan et al.,

"Manipulating the Mouse Embryo," Cold Spring Harbor- Press, Cold Spring Harbor, NY, 1986; and Palmiter et al., Cell 41:343,-1985).

Once the reporter molecule has come to reside within the membrane or has been bound to a component of the membrane (thus forming a membranous system in vivo) the test molecule is applied and changes in the signal generated by the reporter molecule (e.g., quenching) are recorded. The recording device may include a video camera (such as those used in the publications regarding in vivo imaging cited above), or any other device capable of detecting a change in the signal generated by the reporter molecule.

Using the methods and compositions described herein, one can determine whether a test molecule moves from one region of a membrane (either a mimetic or naturally occurring membrane) to another. For example, one could examine movement of a test molecule from a region in which a particular type of molecule is present in low abundance to a region where that molecule is highly abundant. It is known, for example, that particular regions of the membrane are rich in cholesterol. The determination could be made simply by using two types of reporter molecules (or more, if necessary) . One reporter molecule would be placed in regions that have one characteristic and the second reporter molecule would be placed in regions that have a different characteristic. A change in the signal generated by the first reporter molecule would indicate entry of the test molecule into the first region; similarly a change in the signal generated by the second reporter molecule would indicate entry of the test molecule into the second region.

As noted above, the membranous system includes a reporter molecule that generates a detectable signal that

can be measured to determine whether it changes over time or remains in a state of equilibrium. Fluorescent reporter molecules, which form membranous systems referred to herein as "fluorosomes" are described in the examples below, but any reporter molecule that generates a detectable signal that changes in response to a test molecule can be used. Useful reporter molecules are listed below, and a discussion of their incorporation into a membranous system follows.

Molecules that can be used as reporters include phospholipids and fluorescent molecules, which can be neutral or charged. Neutral fluorescent molecules useful in the invention include DPH (diphenylhexatriene), Nile Red, N-phenyl-l-napthalamine (NPN); derivatives of the BODIPYTM fluorophores; sterol analogues (for example, nitrobenzoxydiazole (NBD) cholesterol, NBD cholesterol ester, and pyrene steroids) ; environment sensitive molecules, for example, 6-propionyl-2-dimethylamino- napthaline and 6-dodecanoyl-2-dimethylaminonapthalene; bis-pyrene alkanes, aminonitrile, anthracene; the lipophilic Coumarin molecules; zwitterionic probes (for example, aminstyryl S-467 and Dansyl lysine), pyrenenonal, hexadecylpyrene, pyrenylglycerol, and dihexadecylamino NBD.

Useful anionic fluorescent molecules include fluorescent fatty acid analogues such as anthroyloxy fatty acids, BODIPYTM fatty acids, NBD, dansyl fatty acids; parinaric and DPH fatty acids; and pyrene and perylene fatty acids. Lipophilic fluorescein probes including acylaminofluoresceins and anilinonaphthalene sulfonates can also be used.

Cat ironic fluorescent molecules include the carbocyanines, such as those in the DISC, series; diakylaminostyrl molecules; single chain cationic fluorophores, such as octadecyl Rhodamine B; the blue

fluorescent ionic fluorphores, such as the pyridoxazole molecules; acridine orange analogues; the cationic DPH analogue TMA-DPH; and naphthalene based fluorescent probes.

Phospholipids with fluorescently labeled head groups or acyl chains are also useful. The detectable moieties used to label phospholipids can include biotin, BODIPY1M FL, BODIPYTM, caged fluorescein, dansyl (333/518), dinitrophenyl Fluorescein, Lisamine rhodamine, Maleimide, NBD, tetramethylrhodamine, pyrenesulfonyl, pyridyldithio, and Texas Red. The substances preferred for fluorescently labeling acyl chains are moieties such as DOH, NBD, perylene, pyrene, and BODIPYTM derivatives. A well known commercial source of fluorescent probes and markers is Molecular Probes, Inc. of Eugene, Oregon.

Alternatively, non-fluorescent molecules can serve as reporter molecules. For example, the membranous system can include spin-labeled phospholipids such as derivatives of N-oxyl,4'-dimethyloxazolidine.

Prior to insertion into a membrane, a reporter molecule can be dissolved in a solvent such as an alcohol, DMSO, or DMF, or it can be emulsified or suspended, for example in a buffered solution. These procedures can facilitate interaction between the reporter molecule and a lipid-based monolayer or bilayer.

Alternatively, the reporter molecules can be added to the membrane as a powder, paste, or gel. The reporter molecules can be added to the membranous system in a substantially pure form or bound to a carrier such as a protein (such as serum albumin) or a carbohydrate (for example, cyclodextrin). The way in which a reporter molecule should be prepared for addition to the membranous system will depend on the physical properties of the molecule and can be readily determined by skilled artisans. The primary physical property of the reporter

molecule to be considered in its preparation is solubility; reporter molecules that are soluble in alcohol wouLd be suspended in alcohol, and so forth.

Typically, the addition of a reporter molecule to a membrane is performed by injecting the solvent containing the reporter molecule into a membrane suspension that is being rapidly stirred. The reporter molecules can be inserted into a mimetic membrane at any time, i.e., before, during, or after membrane assembly.

In the event the membrane has not yet been assembled, the reporter molecules can be added to materials that are used to manufacture the membrane. For example, they can be cocrystallized with the lipids that are used to create the membrane in a mutually soluble organic solvent.

Alternatively, the reporter molecules can be added to the membrane preparation during or after its formation.

In the event a natural membrane is used according to the methods described herein, the reporter molecules can be added at various times. For example, the reporter molecule can be added to cells before or after they are harvested from culture or from a living plant or animal.

The techniques required to harvest cells are well known and the reporter molecule is added to membranous vesicles that are prepared from these cells as described herein.

Reporter molecules can be covalently linked or otherwise associated with moieties on either the exterior or interior face of the lipid bilayer. They can, for example, be coupled to a molecule that is anchored in the membrane, such as a sterol, and thereby serve as a reporter of activity only on one or the other face of a lipid bilayer or a particular membrane domain (i.e., a region of the membrane within the lateral plane of the membrane; an example of a membrane domain is a cholesterol-rich region). Although the membrane is fluid, the lipid molecules in synthetic bilayers very

rarely migrate from the monolayer on one side to that on the other. This process, called "flip-flop" is estimated to occur less than once a month for any individual lipid molecule. In cases where the rate of flip-flop may be unacceptably great, the reporter molecule can be anchored to a particular face of the bilayer by binding it to a protein that is localized to that face. Labeling one or the other bilayer of a mimetic or natural membrane with a reporter molecule can provide a basis for measuring entry of a test molecule into one leaf of the bilayer, movement across the bilayer ("flip-flop"), and movement out of the bilayer. Reporter molecules can also be trapped within the interior volume that is enclosed by a membrane.

These molecules would report the penetration of the test molecule through the membrane.

Test molecules can also be examined in membranous systems in which the lipid molecules form a monolayer (i.e., unilamellar membranous systems). It is well within the abilities of skilled artisans to construct unilamellar membranous systems. For example, unilamellar vesicles can be constructed by sonicating multilamellar vesicles, by reverse-phase evaporation, or using the French press extrusion technique.

Test Molecules and their Presentation to the Membranous System Test molecules can be applied to the membranous systems described herein in a variety of forms. They can, for example, be placed in a well-defined solution or suspension, or they can be presented in a natural bodily fluid such as blood, urine, saliva, or cerebrospinal fluid. In addition, the test molecules can be attached to a naturally occurring carrier molecule, such as serum albumin, or a synthetic carrier molecule, such as a cyclodextrin. Many pharmaceutical agents currently in use are administered with a carrier molecule. Thus,

examining the test molecule in this context provides a closer approximation of the way in which it may eventually be administered. Alternatively, test molecules can be presented as an emulsion, or in association with a micelle, a liposome, a natural membrane, or a derivative thereof.

The test molecule that is applied to the membranous system (as described below) is not tagged or labeled in any way; rather, measurements which indicate its presence are obtained by examining its effect on the reporter molecules in the membranous system. This arrangement underlies the ability to use the membranous system to examine association between the test molecule and virtually any given part of the membrane or, if an intact cell is used, with any given compartment within the cell. For example, if reporter molecules are introduced to the interior face of the membrane bilayer (for example, by attaching them to proteins that are expressed only on this face), alteration of the fluorescence signal provides a measure of the rate of entry of the test molecules into that bilayer. In a similar manner, if the reporter molecule is inside the cell, for example within the membrane of a mitochondrion or the nuclear membrane, a change in the signal it generates will reflect entry and accumulation of the test molecule in that membrane. Association of the test molecule with the outside or inside face of an internal membrane can also be measured by attaching the reporter molecule to a protein or other component that is expressed on one or the other face.

To attach the reporter molecule to a protein that localizes to a particular compartment of the cell or a particular face of a lipid bilayer, cells could be genetically modified to express a chimeric polypeptide by transfecting them with a construct (e.g., an expression

vector) that encodes the reporter molecule and the protein. An example of this type of construct, where proteins arewfused to GFP (green fluorescent protein) is described above. Cells could be transfected in vivo or in tissue culture. In the event a mimetic membrane is being used, the reporter molecule could be attached to a protein or phospholipid that will be incorporated within the membrane when it is constructed.

The rate with which the test molecule traverses or otherwise interacts with the membranous system can be determined by examining the rate with which the signal generated by the reporter molecule changes. Numerous devices, including the stopped-flow spectrophotometer used in the examples below, can be used to determine the intensity of the signal generated by the reporter molecule over time. For example, see Figs. 3A-3C, where the fluorescence emitted by a reporter molecule is quenched (y axis; "decreasing fluorescence") over time (x axis; seconds) as soon as a test molecule is brought into contact with the membranous system. When the signal generated by the reporter molecule no longer changes over time (see the point where the curve "flattens" in Figs.

3A-3C) the test molecule has reached a state of equilibrium (i.e., there is no net change in the amount of test molecule that is accumulating within or crossing the membrane).

Furthermore, once a state of equilibrium has been reached, it is a routine matter to determine, if desired, the amount of test molecule that has crossed or accumulated in the membrane (or otherwise interacted with components of the membrane, such as proteins or carbohydrates). One way to quantitate the accumulation of the test molecule within the membranous system is to centrifuge the entire contents of the system after a state of equilibrium is reached (so that the membranous

vesicles are pelleted) and to quantitate the amount- of the test molecule remaining within the supernatant. If 100 moles of test molecule were initially added, and 50 moles remain within the supernatant, then 50 moles must have entered or crossed the membranous vesicles. The quantitation can be performed by any standard method.

For example, the test molecule may be radiolabeled or analyzed by HPLC. Quantitating the test molecule in this way provides useful information about the solubility of the molecule in the membrane. This information can be used to determine whether a test molecule is likely to accumulate in tissues, perhaps to undesirably high levels, over time.

Virtually any molecule that is not completely water soluble can be studied in the membranous system described herein. The types of molecules that can serve as the test molecule include pharmaceutical agents used to treat humans or used for veterinary purposes. These molecules include barbiturates, aliphatic alcohols, anti-convulsants, muscle relaxants, narcotic analgesics, analgesics and antipyretics (for example, salicylates), cocaine, anticholinesterase agents, parasympathomimetic agents, sympathomimetic agents, antimuscarinic agents (for example, atropine), antihistamines, serotonin, digitalis and allied cardiac glycosides, antiarrhythmic drugs, antihypertensive agents, diuretics, oxytocics (for example, ergot, ergot alkaloids, and oxytocin), chemotherapeutics (including those for the treatment of cancer, parasitic, and microbial diseases), natural or synthetic hormones, and fat soluble vitamins. For applications to plant cells, any type of fertilizer or pesticide can be used as a test molecule.

The molecules tested in the examples below include alcohols, which can function as anesthetics, THC and THC analogues, which have multiple pharmacological

applications, and the antibacterial agent, chloramphenicol.

Pharmacologists are also concerned with interactions between pharmaceutical agents, and the methods and compositions described herein can be used to examine such interactions. For example, one can determine whether a first molecule is capable of modulating the rate with which a second molecule associates with a membranous system by examining the second molecule (exactly as described herein for any "test molecule") in the presence and absence of the first molecule. The method would be carried out by applying the first and second molecules to a membranous system and measuring the signal generated by the reporter molecule.

In an equivalent membranous system, the second molecule would be administered alone, and the signal generated by the reporter molecule measured. The two measurements can be obtained at different times, and the order in which they are obtained is irrelevant.

Once the two measurements are obtained, they are compared. If they are the same or substantially the same, one would conclude that the first molecule had no bearing on the ability of the second molecule to cross or otherwise interact with the membrane in the membranous system. However, if they were different, the first molecule would be said to modulate, by either increasing or decreasing, the rate with which the second molecule crosses or interacts with the membrane.

If desired, the influence of combinations of two or more "first" molecules can be tested for their ability to modulate a "test molecule" within a membranous system.

Any of the membranous systems described herein can be used to examine interactions between molecules. The examination can be carried out, for example, in a mimetic membrane consisting of a lipid-based monolayer or

bilayer, or in a naturally occurring membrane that is present in a plant or animal cell in vivo or in tissue culture.

Measurement and Monitoring Devices Any device that can measure the signal generated by a reporter molecule can be used to practice the invention. Examples include spectrophotometers and optical sensors. Skilled artisans will be aware of the variety of signals generated by reporter molecules and the devices capable of sensing them, such as detectors for fluorescent, infra-red, and visible light, and those utilizing the techniques of Raman, nuclear magnetic resonance, and electron-spin resonance.

Examples In the examples described below, a stopped-flow spectrophotometer in fluorescence mode was used to measure the rate of change in a signal generated by the fluorescent reporter molecule diphenylhexatriene (DPH; Sigma Chemical Co., St. Louis, MO) in a membranous system of fluorosomes. This reporter molecule generates a detectable fluorescent signal that is quenched by the test molecule. The spectrophotometer mixes test molecules with the membranous system and the resulting spectrum is presented as a functibn of time.

Example 1: Presaration of Fluorosomes Fluorosomes are membrane bound vesicles that contain a fluorescent reporter molecule. The fluorosomes used to test the molecules described below contained lipid and, as the reporter molecule, diphenylhexatriene.

These fluorosomes were constructed as follows. Sixty mg of lipid (egg phosphatidylcholine; Avanti Polar Lipids, Alabaster, AL) were added to 0.5 ml of buffer (10 mM

HEPES, 1 mM EDTA, pH 7.0) and agitated to form multilammelar vesicles (MLVs).

Large unilammelar vesicles (LWs) were then formed by passing the MLV suspension through a LiposoFast Extrusion Apparatus (Avestin Inc., Ottawa, Canada) containing 100 nM polycarbonate filters (as described by MacDonald et al., Biochim. Biophys. Acta 1061:297-303, 1991). Typically, 21 passages were performed. The resulting LW solution was brought to 75 ml total volume with buffer, which produced a 100 HM solution of lipid.

Generally, the volume of the solution containing the lipid was equal to the volume of the solution containing the test molecule and the two were mixed in a 1:1 ratio (described below).

Fluorosomes were formed by rapidly squirting a solution containing 10 yM of the fluorescent hydrocarbon diphenylhexatriene (DPH; Sigma Chemical Co., St. Louis, MO; in DMF) into a vigorously stirred suspension of Lugs.

The amount of DPH added was such that the final concentration of DPH was 10% of that of the phosphatidylcholine used to form the lipid portion of the vesicles. That is, generally, one in approximately ten molecules present in the membranous system was a reporter molecule.

The size of the fluorosomes was analyzed by photon correlation using a Coulter N4 submicron particle analyzer, with size distribution analysis and multiple scattering angle detection options (Coulter Electronics, Hileiah, FL) and the vesicles were found to be tightly centered around 1612 nM. This measurement indicates that the population of vesicles is largely uniform in size.

In addition, ULVs of this size closely approximate the bilayer curvature of naturally occurring plasma vesicles.

The fluorosomes created as described above were excited at 356 nM and their emission spectra scanned in

the Farand Spectrophotometer. The resulting curve is shown in Fig. 1. This spectra is typical of DPH in an organic milieu (MacDonald et al., Biochim. Biophys. Acta 1061:297-303, 1991). A peak from about 332 nM to 365 nM results from scatter of the excitation beam and the broad peak above 370 nM results from fluorescence. Therefore, a cut off filter excluding all light below 365 nM was used in the experiments using these fluorosomes.

Example 2: Preparation and Examination of Test Molecules A plot of the emission intensity was generated as progressive amounts of THC (tetrahydrocannibinol; 2.5 mM in ethanol) were added to the cuvette of a spectrophotometer containing fluorosomes prepared as described in Example 1. The solution was drawn up and down several times in a pipette so that it was well mixed after each addition of THC, and the fluorescence intensity was measured once the reading stabilized. The excitation wavelength was 356 nM and fluorescence emission was measured at 433 nM. The curve shown in Fig. 2 reveals a steady decline in the fluorescence signal generated by the reporter molecule with increasing concentrations of THC.

In addition, the rate with which various test molecules became associated with the fluorosomes prepared as described in Example 1 were determined. Solutions containing the test molecules were prepared by dissolving those molecules in a dispersing solvent (either ethanol or dimethylformamide (DMF)). The molecules were then suspended in buffer (10 mM HEPES, 1 mM EDTA, pH 7.0) so that the final concentration of the dispensing solvent was never more than 0.05% of the final total volume.

Generally, final concentrations of the test molecules ranged from 1.8 yM to 350 UM, and 10 ml of the solution

containing the test molecule was sufficient for multiple measurements.

To examine the test molecules, stopped-flow measurements were made with a S-4 Series stopped-flow spectrophotometer (Hi-Tech Scientific, Salisbury, England). The membranous vesicles associated with a reporter molecule (the "fluorosomes" prepared as described above) and a solution containing the test molecule were introduced into the reservoirs of two syringes. The test molecule was then added to the membranous system by triggering pneumatically driven pistons within each syringe to drive the two solutions from the syringes, at a ratio of 1:1, through feed coils and into a 40 yl quartz integral mixer/observation cell.

Both the observation cell and the feed coils were immersed in a thermostatic bath adjustable from -100 to 1000C + 0.10C. Excitation energy at a selected wavelength was delivered to the observation cell via fiber optics, and emission light was collected by fiber optics, passed through a Hi-Tech PF-3 cut-off filter, and delivered to the photomultiplier. The output was amplified by the SF-40C control unit and delivered to the data acquisition board (Hi-Tech Scientific) mounted in a Gateway 2000 P5-100 computer. The data were processed and analyzed by Hi-Tech Scientific's IS-2 Rapid Kinetics Suite. Static fluorescence measurements were made with a Farrand MK-2 Spectrofluorimeter (Farrand Optical Co., Valhalla, NY).

The results of typical stopped-flow spectrophotometric measurements of two hydrophobic drugs, the THC analogue WIN-55,212-2 and A8-THC, are shown in Figs. 3A and 3B, respectively. The measurement shown in Fig. 3A was made with a test solution of WIN at a concentration of 29 pM. When this concentration of test molecule is mixed with an equal volume of fluorosomes

consisting of 100 yM egg phosphatidylcholine, the final concentration of WIN-55,212-2 is 14.5 HM and the final concentration of lipid is 50 UM. The measurement shown in Fig. 3B was made with a test solution of A8-THC at a starting concentration of 3.6 yM, and final concentrations can be extrapolated in the same manner as for WIN-55,212-2. In addition to the experiments described above, the test molecules were examined at concentrations that varied from those given above by 10-fold. These concentration differences did not effect the result; the time course with which the signal generated by the reporter molecule changed was substantially identical at each concentration of test molecule.

In contrast to WIN-55,212-2 and A8-THC, the test substance shown in Fig. 3C is an amphipathic substance, 1-octanol. The starting concentration of 1-octanol was 768 zM. For this test molecule, the time course of entry into the bilayer was much longer than for the amphipathic molecules shown in Figs. 3A-3B.

Data obtained from curves generated as described above and shown as Figs. 3A-3C were obtained for numerous additional test molecules and are presented below as Table 2. In addition, experiments were carried out using a variety of bilayer impermeant molecules such as dextrose and lysozyme, and no association with the membrane was observed, as evidenced by lack of diminution in the fluorescence of the reporter molecule.

TABLE 2 Test Molecule Half-time of Entry (Seconds) AM356 1.57 Anandamide 3.17 AM245 4.71 WIN-55,212-2 5.96 AM835 10.60 SR1476 15.35 A8-THC 24.70 Ethanol 26.00 Pentanol 32.90 Methanol 51.80 Chloramphenicol 53.00 Isopropanol 68.60 AM411 94.00 CP-55,940 96.50 A8-THC-DMH 98.90 Octanol 129.00 The alcohols (ethanol, pentanol, methanol, isopropanol, and octanol) represent anesthetic agents; A8-THC, A8-THC-DMH, and the THC analogues (AM356, Anandamide, AM245, AM835, and AM411) have multiple functions (for example, as immunosuppressants, anti-convulsants, anti-inflammatory agents, hallucinogens, anti-emetics, and analgesics); and chloramphenicol has antibacterial properties. The structures of many of these test molecules can be seen in Figs 4A-4K.

Other Embodiments It is to be understood that while the invention has been described in conjunction with the detailed description thereof, that the foregoing description is intended to illustrate and not limit the scope of the invention, which is defined by the scope of the appended claims.

Other aspects, advantages, and modifications are within the scope of the following claims.