This invention relates to the formation of a membrane from peptide and lipid constituents, intended for use as a structural basis far further incorporation of biomolecules carrying out a function.

The problem hou tα make a non-fragile polar-nonpolar- pαlar membrane is one αf the most important ones in con¬ temporary biochemistry. The applications of such membranes, αnce stable ones can be mads, are numerous, and could be expected to speed up the development in fields of applied biochemistry in uhich the biological functions of molecu¬ lar components are used artificially to generate transmεmb- raneous electrical potentials, currents, and/or _. fluxes of lαu molecular weight compounds. One such example of a poten¬ tially useful traπstnembraneous gradient is the photo-mecha¬ nical energy conversion mediated by bactεriαrhαdopsin, in uhich case light is directly transformed into mechanical uork (l). Another category of examples uould be irj. vi o cαrrespαndances to cases uhen the plasma membrane receptors upon binding the appropriate ligands bring about, directly αr indirectly, a change of the trans-membraneous potentials. The latter category of examples uould include glucose re _¬ ceptors, glucose and amino acid transporters, membrane-bound irπmunαglαbuliπs uhich are part αf a trans-membrane poten¬ tial -modulating system, receptors for growth factors, various iαn channels and other carrier molecules. Obviously, these examples are of considerable clinical and medical importance. Furthermore, other applications of bioelectrici-

ty, biopotentials and/or biocurrertts, uhere biomoleculεs are used and uhere they have to be anchored, linked or bound to a pαlar-πoπpolar-polar (PNP) membraαe might be possible tα cαncieve in the future. From the above-menti¬ oned examples of biomαlecules the function of uhich is linked to a PNP-membranε, it is obvious that the formation of such membranes is an extremely important topic in bio¬ chemistry. Furthermore, it is obvious that any industrial applications αf these biαmαlecules, far example photomecha¬ nics photoαsmotic or photoelectric energy conversion using bacteriorhαdαpsiπ, glucose receptor units for automatic control of glucose levels in diabetic patients, or diag¬ nostic equipment based on. im uπo eceptors linked to a PNP- -membrane, uould depend on the type of PNP -membrane used. For these reasons, any improvement, in any respect, of the formation of PNP-membranes has considerable industrial impact, as far as the further development of applications αf PNP-membranes is concerned.

The fragility αf lipid bilayers is- ell knαun to re¬ searchers in the field and is primarily due to the fact that they mostly are stabilized irt the plane parallel to the surface of the membrane by ueak, or at least transient hydrαphobic interactions, and by transient hydrogen bonds between the polar residues of the lipid. These conditions permit lcng range diffusion of the individual lipid mole¬ cules above the phase transition point, when the lipid is "liquid" or "fluid". Many attempts have been made and pub¬ lished previously to "crαss-liπk" the lipid at its polar residues, binding it to polar, high molecular weight co -

pounds or surfaces. These attempts present the disadvan¬ tage that the short range interactions between the indivi¬ dual lipid molecules will have to adapt to any long range changes of the high molecular weight cross-linker or the surface. This means, for example, . that if the cross-linker αr the surface expands or contracts on the long range sca¬ le, the covaleπtly bound lipid will be farced to move in between its short range neighb.ours in the plane of the membrane and the lipid or any other molecules inserted into the membrane will be perturbed. This follows natural¬ ly from that the physico-chemical properties of the sur¬ face or the cross-linker and, in particular, their expan¬ sion coefficients given in per cent change of length or area per unit change αf any exterior conditions such as temperature, pH , ionic strength or composition of the buf¬ fer, may not. be the same as those of the lipid bilayers. Another disadvantage of using a long-range cross-linker αr surface external to the lipid bilayer is that any trans- membraneous events, such as fluxes and/or changes of elect¬ rical potential, will not easily be recorded, due to the presence of the extra diffusion barrier.

It is well known that many natural membranes are lipid bilayers and that the biological function of many molecular- components can be recovered by inserting them into lipid bilayers. In some cases, this requires that certain lipid species are present or that the lipid is "fluid". There are many known methods of forming lipid bilayers which can be subject to experimentation, such as the sonicatioπ αf phαsphαlipid, producing lipid vesicles, the formation of

planar "black iτiemhraπes" and the "patch clamp" technique -far* making a small area of bilayer membrane. As was menti¬ oned above, the disadvantage αf these techniques is the fragility αf lipid bilayers, uhich becomes manifest as soon as a large area of membrane is to be assembled. Anot¬ her disadvantage of pure lipid bilayers when considered far medical use in the body is the tendency αf such lipid to.bind to other components present there. It is well known, for example, that lipid binds to albumin and the various lipαprαteiπs present in plasma. Of course, this leads to that the lipid disappears from any structure where it would have been placed to form a membrane as sαon as that struc¬ ture is^ placed inside the body.

From what has been mentioned above, it is obvious that it is desirable to search for an alternative to the lipid bilayer as a thin compartmentalizing PNP-membrane. ^"

In contrast to the case of pure lipid bilayers, it has been suggested that biomembraπes would exist uhich are sta¬ bilized by lαng-raπgs covalent bonds and/or hydrogen bonds provided by amphiphilic "interphase peptides" which are shared by the hydrophilic and the hydrαphobic phases αf the boundary (2,3). The finding that certain peptides, no¬ tably poly-L-lysiπe, adheres to sαlid hydrαphobic surfaces CO provides experimental support that interphase peptides partitioning at the boundary between the hydrophilic and rrydrophobic phases exist, whatever their secondary or ter¬ tiary configuration may be. (Poly-I—lysiπe is a highly hydrophilic substance, yet it binds to hydrophobic surfa¬ ces). In the original theory of "interphase peptides" (2,3)

it has been suggested that they are stabilized by hydro- phαhic contact with other peptides and/αr lipid, uhich is irr accordance with the so called "hydrophαbic effect", and would thus seem to be thermodynamically possible. ("The hyd¬ rαphobic effect" is a term denoting the finding that polar and nonpolar solvents form separate phases). In addition, it has been suggested (2,3) that interphase peptides are stabilized by hydrogen bonds and salt bridges. These fea¬ tures will allow the membrane to have many properties simi¬ lar to those of lipid bilayers (3). That interphase pepti¬ des are thermodynamically stable or, more precisely, in accordance with the hydrophobic effect means that accor¬ ding to the original theory, conditions would be possible tα establish when they form spontaneously at a hydrαpho¬ bic - hydrophilic interphase. Since hydrophαbic interac¬ tions are predominantly van der Uaals interactions invol¬ ving fluctuating dipolε moments of almost equivalent non- polar groups of adjacent micelles, the formation of an in¬ terphase peptide at a hydrophobic surface αf a solid or at that αf a liquid would not be expected to be qualitatively different from each other. Hαuevsr, due to the thermal agi¬ tation within a nonpolar liquid, it can be anticipated that the formation of an interphase peptide at such a boun¬ dary would take longer time than at the boundary αf a non- polar solid.

From what has been said so far, briefly describing the state αf the art in the field of PNP-membranes and their industrial potential, it is obvious that present techni¬ ques need tα be improved if the possible applications of

such membranes are tα be realized.

The present invention describes a method to asse hle a core, structure αf a membrane based on interactions bet¬ ween interphase peptides and lipid, without any claim regarding the detailed applications of such a membrane. The method described will be passible to optimize. As far future applications αf the .method, many techniques have been published on hαu tα incorporate functional proteins into lipid moπolayεrs αr lipid bilayers, and on the con¬ ditions uhich are suitable for this. Since the membrane described here is partly based on lipid, it can be anti _¬ cipated that future attempts to incorporate functional proteins into it uould rely, at least partly, on previous work απ lipid mσnαlayers αr bilayers.

The interphase peptide is farmed from a solution of pσly-L-lysine (4-5 g/1), flaJ 15-30.000, αr specifically 20- 25.000) in 0.4 Pi KOH at approximately 0°C on top of which ethyl ether is layered. The aqueous phase conditions are essentially those given in previously published experi¬ ments (4), the only difference being that the solid hyd¬ rophobic surface has been replaced by a liquid one. Since interphase peptides are thermodynamically stable (3), these conditions uould not be critical, and other condi¬ tions suitable for the formation αf interphase peptides would be possible to establish by any skilled experimen¬ talist who has knowledge αf the chemical and physical properties of the peptide at various pH values and for all ranges αf molecular weights. The incubation at a ^* pprαxi ately 0 C is continued until a layer of inter-

phase peptide is detectable, for example by visual inspection. This may require that the incubation at 0°C is ^* continued far approximately a week, using the speci¬ fied conditions. It is advisable to interrupt the incu- hatiαn before the peptide is folded into the organic phase so that the membrane becomes rough on the nonpolar side. The conditions given are probably not the best far settling a:, stable interphase peptide. and would be possible to opti¬ mize by modulating tha ionic strength, the temperature, any other variable or the choice of organic solvent.

The experiments can be carried out- such that ^* one ml αf organic solvent is layered on top of one ml of aqueous phase in a cylindrical glass vial of 9 mm diameter having a flat bottom. On the bottom in the aqueous phase and below the phase boundary has been placed (before the solvent is added) a cylindrical tube the upper end onto which has been stretched and glued a porous supporting membrane and along one side αf which has been attached a long glass capillary extending tα the top of the glass vial to serve as: a handle. The. tube should not be made from teflon αr any other material uhich strongly adsorbs palylysine The purpose αf this devise is tα lift the interphase membrane from below after it has formed and it should be devised in such a way that there is no vigorous movement of liquid in the aqueous phase below the membrane when it is lifted from the surface αf the remaining liquid. It is desirable that the supporting membrane is smooth and that the pores in it are small since such conditions will decrease the probability that cracks appear in the

interphase membrane uhen it is lifted. It is possible to avoid that the liquid inside the tube drops out αf the tube uheπ it is lif ed, by making the lower end αf the tube more narrow than its average inner diameter.

The present invention dαes not relate to the type αf supporting membrane used and many choices are possible. A most simple and readily available choice is the transparent membrane uhich is used in the household for protecting food and is known under various commercial names such as "glad- wrap", "surround -wrap" or "handi-wrap" . This type αf memb¬ rane may be chosen far the experiments ana glued onto the above-mentioned tube with alkali-resistant glue after which it is perforated with a needle as shaxp-pointed as possib¬ le. Far further aptimation, it is desirable to select memb¬ ranes which are smooth and have smaller pores. If the in¬ terphase peptide is formed from an alkaline solution, the supporting membrane should be alkaliresistant. The purpose αf using porous supporting membranes is that trans-membra¬ neous events taking place across the interphase membrane will be passible to record using standard techniques.

The membrane or the devise uhich have been described are not optimal but merely a cαπveπiaπt way for any handy person to lift a planar interphase membrane from a surface or an interphase. It is desirable to make the devise smal¬ ler since this will decrease the probability of cracks in the interphase membrane, but on the other hand, this might require that the uhαlε process of lifting the membrane is performed using some standard micro-manipulator. As des¬ cribed above, the tube has a diameter of 5 mm and a height

αf 8 mm. One advantage of the devise that has been descri¬ bed is that the nonpolar part αf the interphase membrane is exposed and can be sealed to any other similarly for¬ med membrane, the nonpolar part of which is exposed. Due tα the "hydrophobic effect", the two membranes, one of which contains interphase peptide, will bind to each other, this being the essential feature αf the invention. Many other types αf devises can also be imagined, which would accomplish the same thing.

In addition tα pαly-L-lysiπe, it is cαncievable that other peptides having the property αf forming an inter¬ phase boundary exist αr will be discovered in the future. The present invention would be applicable irrespective of type αr molecular weight αf the interphase peptide, al¬ though in each particular case, the optimal experimental conditions uαuld be expectεd to be different. A necessary condition uould, of course, be that any significant amount αf interphase peptide is formed and that some of it stays at the boundary (αr the surface of the aqueous phase) du¬ ring subsequent experimental steps.

A significant advantage of using ether in the method described, as compared to many other organic solvents is that it easily can be evaporated, leaving the hydrophαbic part αf the interphase open to air. Another organic solvent which is not volatile could be expected tα be more diffi- ^" cult tα remove and might thicken the nonpolar part of any PNP-mεmbrane αf which it forms a constituent tα the extent that it would be impossible for inserted functional molecu-

lεs. tα carry out any vectorial trans-membraneous function. However, it uould be possible tα use other organic solvents if ^" they can be replaced with a volatile solvent after the interphase peptide has formed. This follows from the near equivalence αf hydrαphobic bonding, uhich involves fluctu¬ ating dipole moments (van der Uaals forces) irrespective αf the chemical composition 'of the compound. Therefore, it is; obvious that the basic mechanisms of the formation of a--- ther αdyπaπτ eally stable (cf. 3) interphase peptide are not altered if the choice of organic solvent is altered, as long as the solvent does not contain any amount of po¬ lar groups. Also the shape αf the solvent molecules will be α.f importance and it is advisable to check each solvent for its suitability tα bind the peptide.

According to the present invention, the interphase p-εptidε is incubated with lipid. In this context, lipid is regarded in its widest sense, as a group of amphiphi¬ lic elongated compounds composed of at lεast one polar group at one end and at least one nonpolar group at the apposite end, uhich definition may include, for example certain detergents.

The rationale far incubating the peptide uith lipid is that the spreading αf the interphase configuration in the plane of the boundary between the nonpolar and polar phases is restricted by the constraints αf the possible angles αf cαvaleπt bonding and it is expected that a memb¬ rane based solely on peptide would be porous. Furthermore, the binding blocks of the peptide are considerably larger

than those of lipid monomers in lipid bilayers, and tight fitting between the peptide building blocks down tα the atomic level can not be anticipated. In the original theory αf membranes based on interphase peptides (2, 3), the memb¬ ranes are stabilized not only by hydrαphobic interac¬ tions but also by the more specific hydrogen bonds and salt bridges. According tα the present invention, this is allowed for by incubating the peptide with lipid, which, according tα the original theory, stabilizes the membrane by interacting with the peptide in the nonpolar plane, a plane αf hydrogen bonding, and a plane of salt bridges ( _. 3) . None αf these interactions is a cαvalent bond, which allows a certain flexibility of movement of the compounds forming the membrane, similarly tα the case αf lipid bi¬ layers .

The lipid is added in the organic phase either in a few ul of chloroform - methanαl αr in a few ul αf ether or in any other suitable solvent. If a heterogene¬ ous mixture of lipid is used, many different kinds of interactions between the lipid and the protein will be passible and the chances are better that the pores bet _¬ ween the amino acid units are filled out. However, any type of amphiphilic lipid capable of settling at the pα- Iar-πoπpolar boundary will fill out the pores and improve the fitting between the various residues. It should bs remembered that there is competition for the boundary by the peptide and the lipid and if too much lipid is added, it may completely replace the peptide.

A heterσgeπic mixture of lipid can, for example, be pre¬ pared by dissolving one egg yolk to a total volume of 100 ΠTL in _. 1Q mi potassium phosphate buffer, pH 7.4, con¬ taining 0.15 Pi αf potassium chloride, mixing it with 100 ml of chloroform - methaπαl (2:1), then extracting 5Q ml of this mixture with 100 ml of chloroform and filte¬ ring the organic phase through several (5) layers of filter paper sα that it becomes clear.. The resulting clear organic phase can be used in the method described here. However, the extraction αf lipid from various raw materials has been well described in the literature and many other ways tα prepare lipid are known. In- particular, lipid fractions containing less chαlεsterol than the one speci¬ fied above can be obtained and used for the present inven¬ tion.

The lipid is added to the organic phase after the inter¬ phase peptide has formed αr while it is being formed, as long as it does not interfere with its formation. Using the conditions specified above, it is added to the organic phase after approximately 10 days of incubation with the paly—L—lysiπe, and the incubation at 0°C is then continued overnight _* Excess lipid is then removed by carefully re¬ placing the organic phase several times with pure ether.

The addition of lipid tα the organic phase is conveni¬ ent, and preferrable to adding it in the aqueous phase, in uhich case it may form micelles and may παt enter the bαun-- dary due tα stεric hindrance from the peptide or electro¬ static factors. However, whether the lipid is added from the aqueous phase or the organic phase is not critical for

the present invention as long as it does not dissolve the interphase peptide into the polar αr nonpolar phases.

Various types of lipid can be expected to fill out exis¬ ting pores more or less efficiently and the method has yet ^* to be optimized in this respect. To make a stable memb¬ rane, it is desirable to optimize the nonpolar interactions, the hydrαgεn bonding and the salt bridges within the memb¬ rane, as stated in the original theory (3). A way of doing this is tα select a heterogeneous mixture αf lipid and expect that the best fitted lipid will settle at the boundary and that the final proteolipid boundary will represent the thermodynamically most stable condition, excluding un-fit- tεd lipid species. Of course, once the interacting lipid species can be identified, they can be added specifically in well-defined amounts, but this will hardly affect the mechanisms -of the formation of the. proteolipid membrane, uhich is formed as a result αf thermodynamic farces.

After formation of the proteolipid membrane, excess organic solvent is evaporated or aspirated, leaving the interphase under the meniscus, and the proteolipid membrane is lifted put of the incubation vessel. This should be done in a cold room, prεferrably kεepiπg 0°C - - 2°C. In one realization αf the invention, the thus for¬ med proteolipid membrane is joined to a similarly formed proteolipid membrane which has been cαllectεd on another supporting membrane and another lifting devise of the same type, in such a way that the nonpolar parts of the proteolipid membranes are brought into contact.

It should be remembered that when the ether has evapora¬ ted-, one important factor in 'keeping the proteolipid memb- . raπe- together is lost and it is desirable tα join the two membranes as soon as possible. On the other hand, if there is still a significant amount αf organic solvent on the nonpolar side αf any αf the two proteolipid membranes, it uϋl be included in the PNP-membrane and make it thicker. If ^*" , subsequently, the organic solvent is equilibrated out αf the membrane, the latter may be perturbed.

An advantage αf the method of the invention which has been described is that the formed PNP- embrane will auto¬ matically be connected. to tuα aqueous phase chambers uhich " ^* can be modified far further experimentation and o-ptimation of- the invention. The two pieces αf tube would be possible tα prepare for various types of standard biochemical and biophysical experimentation. In particular, they can be made in such a uay that they seal tightly along the cir¬ cumference after the PNP- membrane has formed, using, for example, a clam.p mechanism.

However, many other methods tα join a peptide (-lipid) membrane to an assembly αf amphiphilic compounds, uhich is an essential feature αf the invention, can be imagined. The advantage αf a PNP-me braπe in which interphase pep¬ tides form an essential part, according tα the method described, as compared tα lipid bilayers, is that the membrane containing interphase peptides is stabilized in the plane parallel tα its surface by covalaπt bonds uhich extend continuously for a longer distance than the average

diameter of the lipid of lipid bilayers. This is a require¬ ment for the poly-L-lysine tα partition at the boundary of a nonpolar surface, as it is k.παuπ tα do (4) and is evident far the naked eye in the method described by the high vis¬ cosity of the interphase membrane based on peptide as oppo¬ sed tα that αf an interphase based on the described lipid iτrixture (added in chlorofαrm-methanol) only. In the farmer case, any irregularities _r particles at the interphase uill retain their approximate relative location even if the gl ss vial is gently agitated by hand. This clearly shows th t there is long range structure in the plane cf the membranε bordering the nαπpαlar phase when the peptide is there as opposed tα the case when only lipid has access tα the boundary between the polar and nonpolar phases. Uheπ the nonpolar phase is subsequently replaced with ano¬ ther membrane ^' such that the two membranes are joinεd by hydrophobic bonding, this state αf facts is, due to the near equivalence αf hydrαp ^"" hαbic banding irrespective of chemical compound, not qualitatively changed. The advanta¬ ge αf an interphase peptide providing long-range structure in the form αf cσvalent bonds εxtendiπg parallel tα the plane αf the membrane is, of course, that it will stabili¬ ze the uhαle membranε by providing permanent cαvalent bonds instead αf transient hydrαphobic bonding or hydrogen bon¬ ding. Above the melting point of the lipid, uhen the latter is. "fluid" or "liquid" and tends tα diffuse away in any lipid bilayer structure uhere it is a constituent, such stabilization by cαvalent bonds would be particularly

important for maintaining a structure in the plane of the membrane.

Thus, while the lipid bilayer is appreciated for its tight sealing perpendicular to the plane of the membrane, the proteolipid membrane described here provides stabili¬ zation parallel tα the plane of the membrane without any claim regarding the intactπess αf the membrane. The memb¬ rane described may contain pores and/or cracks the amount of uhich may vary depending on the skill of the experimen¬ talist. Therefore, the membrane should be regarded as a PNP-mεmbraπe uith a predominantly nonpolar core and suitab- lε for αptimation of its sεaling perpendicular to its pla¬ ne, as well as αf its stability parallel tα the membrane.

In addition tα joining the proteolipid membrane to another similarly formed membrane, it can be joined to a lipid mαπαlayer fαrmed..in a conventional way an the surface αf an aqueous solution and pressεd together to farm as high a density αf lipid per surface unit as pos¬ sible. The aqueous phase should have approximately the same iαπic strength and pH as that on the polar side αf the proteolipid membrane. The joining together of a pro¬ teolipid membrane as describεd uith a conventional lipid onolayer such that their nonpolar surfaces face each other follows the same principles (the hydrophobic effect) as does the joining together of two proteolipid membra¬ nes as describεd previously.

The PNP-me brane described in the method of the inven¬ tion may be subject to limited proteαlytic digestion

after its formation. Then, the composition of the buffer should be such that it does not destabilize the membrane.

References

1. R.C. Srivastava et. al . : Exp-erientia 40 (7), 773-775

(1984) Z. ■ E. Cerven: Upsala 3. Pled.Sci. 81, 193-200 (1976) 3.. E. Cerven: Upsala 3. Pied. Sci. 82, 167-181 (1977) 4. S.S. Broun : Piethods in Cell Biology. Vol 24, Pt. A

(εd: L. Uilsαn), pp 291-300, Academic Press, New York,

London, Paris, San Diego, San Fransisco, Sao Paulo,

Sydney, Tokyo, Toronto (1982)