Phospholipids as Ionophores

The ionophoretic capabilities of phospholipids have been examined by direct measurement in a Pressman cell of the phospholipid-mediated translocation of cations across an organic phase separating two aqueous phases. Cardiolipin and phosphatidic acid were the most active ionophores among the phospholipids tested, with activities comparable to that of X537A in respect to the translocation of divalent cations. both divalent and monovalent cations approximately equal rates. The ionophoretic activity of cardiolipin be modulated other phospholipids (stimulation), complexation with cytochrome c (inhibition), and by ruthenium red and lanthanum (inhibition). The rate of translocation of cations mediated by cardiolipin was independent of pH over a wide pH range (5.4 to 8.3). The same general pattern of properties observed for cardiolipin applied to phosphatidic acid except for stimulation by butacaine. Complexation of the ionophoretic capability of phospholipids 1 of magnitude. The the properties of a polyionophore. The of

The ionophoretic capabilities of phospholipids have been examined by direct measurement in a Pressman cell of the phospholipid-mediated translocation of cations across an organic phase separating two aqueous phases. Cardiolipin and phosphatidic acid were the most active ionophores among the phospholipids tested, with activities comparable to that of X537A in respect to the translocation of divalent cations. Cardiolipin translocates both divalent and monovalent cations at approximately equal rates. The ionophoretic activity of cardiolipin could be modulated by other phospholipids (inhibition), by butacaine (stimulation), by complexation with cytochrome c (inhibition), and by ruthenium red and lanthanum (inhibition). The rate of translocation of cations mediated by cardiolipin was independent of pH over a wide pH range (5.4 to 8.3). The same general pattern of properties observed for cardiolipin applied to phosphatidic acid except for stimulation by butacaine. Complexation of phospholipid mixtures, such as asolectin or mitochondrial lipid, with reduced cytochrome c, enhanced the ionophoretic capability of these phospholipids by 1 order of magnitude. The complex thus formed has the properties of a polyionophore.
The possible physiological significance of this enormous ionophoretic potential of phospholipids is examined.
The isolation of intrinsic ionophores from membranous organelles such as the mitochondrion has posed a formidable problem with respect to the relation of phospholipids to the ionophoretic capability of such organelles. The intrinsic ionophores occur in trace amounts and these have to be separated from about a lOOO-fold excess of phospholipid.
Blondin has devised isolative procedures which effectively eliminate phospholipids at the start of the isolation and has restricted his inquiry to molecular species other than phospholipids (1,2). But this begs the question whether phospholipids are ionophores in their own right and whether phospholipids must be considered in a total description of the ionophoretic potential of transducing organelles.
There are scattered reports in the literature that phospholipids have ionophoretic capabilities, but these studies have, for the most part, been suggestive, at best (3)(4)(5)(6)(7)(8)(9)(10). Rosano et al. (3,4) showed that phosphatidylcholine and phosphatidylethanolamine translocate monovalent cations across an organic bulk phase, an established criterion for ionophoretic activity (11). Harris and Farmer (8) found that a number of phospholipids stimulated transport of monovalent cations across a chloroform barrier in U-tubes over several days. Using a modification of the Schulman apparatus, Agate and Vishniac (10) found that phosphatidylserine was an ionophore for iron translocation across a pentanol phase. To our knowledge, no detailed study of phospholipid-mediated transport of magnesium and calcium has been carried out.
Ionophores interact with cations at a membrane barrier to form complexes which are soluble in the membrane and which can then traverse it. The ionophores must have the capability for interacting with cations, either in the two aqueous phases on each side of the membrane, or in the interface between the aqueous phase and the membrane. Moreover, ionophores must be capable of the conformational rearrangement required to permit entry of the cation into their interior space or exit of the cation from their interior space. Finally, they must contain the internal coordinating groups which can allow the cation to shed its hydration shell and transfer from the aqueous phase into the interior of the ionophore (12, 13).
The cells devised by Schulman and by Pressman simply substitute a bulk organic phase for the membrane barrier (14, 15). The ionophore must shuttle cations not across a 75-A barrier, but across one or more centimeters of bulk fluid. Operationally, the same maneuvers apply whichever barrier is used and there should be no difference except a kinetic one between the membrane barrier and the bulk phase barrier (16). Providing there is no barrier to translational mobility, the authentic ionophores work equally well in the Pressman cell as in a membrane,' although the amounts of ionophore required for measurement of activity may be vastly different in the two cases (11,19).
In the present study, we have examined the behavior of a considerable number of purified phospholipids in the Pressman cell and selected, for detailed study, the two most active phospholipids, namely cardiolipin and phosphatidic acid. was reduced with 0.014 ml of 1 M ascorbate and then mixed in order with 1.35 ml of an aqueous, sonicated suspension of 30 mM asolectin or mitochondrial phospholipid, 3.51 ml of water, 2.7 ml of ethanol, and 18 ml of heptane.
The mixture was blended on a Vortex mixer for 30 to 45 s. The phases were separated in a separatory funnel and the heptane layer collected. After a second extraction with 18 ml of heptane, the combined upper layers were evaporated to dryness under vacuum and redissolved in the chloroform medium used for the Pressman cell. Transfer to heptane was essentially quantitative. When lipid. c was prepared by interaction of reduced cytochrome c and a mixture of cardiolipin and lecithin, the procedure was modified as follows: 6.0 ml of a chloroform solution of the two phospholipids, each 1.0 mhr in concentration, were mixed with 4.8 ml of 0.15 mM cytochrome c (reduced first with 4 rnM tetramethylammonium ascorbate), and 12 ml of methanol.
This mixture forms a single phase. Phase separation was achieved by addition of 6 ml first of chloroform, and then of water. The chloroform layer containing the lipid. c complex was collected and used directly as the bulk phase in the Pressman cell.

RESULTS
Survey of Phospholipids as Ionophores-Under the standard conditions defined in the experimental section dealing with measurements in the Pressman cell, cardiolipin and phosphatidic acid were 1 to 3 orders of magnitude more efficient as ionophores for Cal+ transport than the rest of the phospholipids tested (Table I). All phospholipids, with the possible exception of egg yolk lecithin and sphingomyelin, showed significant ionophoretic activity under the conditions of this particular assay. In view of the unusually high ionophoretic activities of cardiolipin and phosphatidic acid, these two phospholipids were selected for detailed study. Cardiolipin and Phosphatic Acid as Zonophores-In Fig. 1, a comparison is made of the time course for Ca*+ transport by cardiolipin and X537A, one of the most efficient divalent metal ionophores yet described. The rate of transport by cardiolipin under steady state conditions at pH 8.3 is 40 to 50% of the rate for X537A (the concentrations of the two were identical). Clearly, cardiolipin is not a marginal ionophore. Two differences between cardiolipin and X537A exist. First, the induction period is more pronounced for cardiolipin than for X537A; second, the CaZ+ translocation rate with cardiolipin is insensi- or Rb+ (Table II). The transport rate for Ba*+ is about 70% of that of Car+ with both cations at the same concentration.
Moreover, the rates with either monovalent or divalent metal ion were near-maximal at about 1 mM concentration of the cation in the donor compartment (Table III). With phosphatidic acid as the ionophore, the maximal rate of translocation of Rb+ required somewhat higher concentrations of the cation in the donor compartment.
Although the translocation rates of cardiolipin with Ca*+ or Rb+ were about the same, competition experiments with equimolar concentrations of these cations revealed almost  exclusive transport of Ca*+ (>95% of the rate without Rb') (Table IV). In similar competition experiments, SrZ+ and Mg*+ depressed the rate of Cal+ transport by 40 and 70%, respectively. When ruthenium red was added to the donor compartment of the Pressman cell at a concentration equimolar with cardiolipin, the rate of Ca *+ transport was reduced to less than 5% of the rate in the absence of inhibitor (Table V). There was no visible detection of ruthenium red in the receiver compartment for up to 5 hours after the start, although pronounced color did develop in the organic phase. Interestingly, the ruthenium red that had interacted with cardiolipin turned brown, probably resulting from auto-oxidation of the reagent induced by the polyunsaturated acid residues in cardiolipin. The inhibition by ruthenium red was even more pronounced with phosphatidic acid. The effects of a set of other inhibitors and stimulants of Ca'+ transport in mitochondrial membranes on the cardiolipin-and on the phosphatidic acid-mediated transport of CaZ+ in the Pressman cell are summarized in Table V (6,27,28). Lanthanum chloride completely suppresses transport of Ca*+ and Rb+ (data for Rb+ not shown), whereas mercurials (fluorescein mercuric acetate and mersalyl) have no effect. Butacaine consistently doubled or tripled the rate of phospholipidmediated Ca*+ transport, whereas it inhibited the rate of Ca*+ transport by phosphatidic acid by 70%. In the absence of  Phospholipids can be shown to modulate the cardiolipinmediated transport of Cal+ under standard conditions (Table  VI). Thus, an equimolar mixture of cardiolipin and lecithin shows 60% of the transport rate with cardiolipin alone. Moreover, with the complex of cardiolipin, lecithin, and reduced cytochrome c, this rate is further reduced by 30 to 55%. Much the same effect of complexation with cytochrome c is observed in the cardiolipin-mediated transport of Mg*+. In the transport of Rb+, the 1ipid.c complex of cardiolipin and lecithin is at least 80% lower in efficiency than cardiolipin alone or the mixture of cardiolipin and lecithin. Different phospholipids show varying degrees of inhibition of the transport of Ca2+ by either cardiolipin or phosphatidic acid (Table VII). Phosphatidylinositol is the most efficient inhibitor.
The transport of Ca z+ from donor to receiver compartment is inefficient when there is no anion in the receiver compartment that can form tieht comnlexes with the Ca*+ (Table VTTI1. Citrate is the most efficient anion in this respect, followed by ADP and P,. Yet the transport of Rb+, which is as rapid as that of Ca*+, was found in separate experiments to require no such pulling force. If the transport of Rb+ is compared with the transport of Ca*+ in the absence of chelating agents, Rb+ transport would be more efficient than Ca*+ transport.

Cation/Cation
Exchange-Under standard conditions, the transport of Gag+ leads to acidification of the donor compartment. A pH drop of 1.0 to 1.5 units in this compartment was 1 _,_ observed over a 5-hour period when the donor compartment Phospholipids as Ionophores 1329  The increase in H+ concentration in the donor compartment was, however, lower than that expected if a CaZ+/2H+ exchange was the only process occurring. The explanation for this discrepancy was found to be the cardiolipin-mediated transport of tetramethylammonium ions in the opposite direction, i.e. from the receiver to the donor compartment.
The observed pH difference in the donor compartment was therefore the net result of two competing reactions: release of 2H+ by uptake of Cal+, and uptake of a proton by release of tetramethylammonium ion.
To confirm that this cation/cation exchange without participation of anions was the dominant process, the intercompartmental flux of anions was checked. Thus, in an experiment in which 440 ng atoms of Ca*+ were transported, only 3.6 ng atoms of s6C1-were found in the receiver compartment during the same time period. Much the same result was obtained when 5 mM [s2P]phosphate, at pH 5.4, was added to the donor compartment.
The transport of phosphate could not be detected. Formation of a neutral cardiolipin-Ca2+ adduct as the ionophoretic entity would obviate the necessity for movement of an anion.
Lipid~Cytochrone c as Zonophore-If comparison is made between the transport rate for Cal+ by asolectin with or without complexation by reduced cytochrome c, a lo-fold enhancement is observed ( Fig. 2A). A 4-to 5-fold enhancement is observed when mitochondrial lipid is formed in a complex with reduced cytochrome c (Fig. 2B). The interaction of cytochrome c with these phospholipid mixtures has significantly increased their ionophoretic potential for Ca*+. The asolectin.cytochrome c complex transports Rb+ at 10 mM 3 times as rapidly as asolectin alone (all other conditions as in the legend to Fig. 2A). At 1 mM concentration of the cation, the 1ipid.c complexes of asolectin, of mitochondrial lipids, and of cardiolipin plus lecithin, respectively, transport Ca*+ 10 times more rapidly than Rb+. As is true of the pure phospholipid systems, sF1-is not co-transported with Cal+. Lipid. c is a complex of a protein and 20 to 30 molecules of phospholipid (26). The molecular unit is therefore more appropriately defined as a polyionophore and the properties of the individual phospholipids in the complex are profoundly modified by the associative bonds between the acidic phospholipids and the basic protein.
Zonophore-mediated Partition of Ca*+ between Aqueous and Bulk Phases in Pressman Cell-A necessary, although not sufficient condition for transport is the requirement for ionophore-mediated partition of the cation into the organic phase. Phospholipids which fail to induce this partition are unable to function as ionophores and all phospholipids which are able to induce this partition under appropriate conditions can function as ionophores. The representative set of Ca*+ values in the organic phase for a variety of experimental conditions and for different phospholipids (summarized in Table IX) show that this condition is met. Even egg yolk phosphatidylcholine has a slight, but perceptible capability for effecting this partition. Under more appropriate experimental conditions, this phospholipid exhibits a transport capability for Ca*+, as well as for monovalent cations (3,4).
Along these lines, it should be noted that cardiolipin is ineffective as an ionophore in carbon tetrachloride by either At the concentrations used in these experiments, phospholipids show a lag period prior to achievement of the maximal rate. During the lag period with cardiolipin, the Cal+ concentration in the organic phase slowly builds to its maximal value (close to a 1:l molar ratio of Caz+: phospholipid) and remains constant thereafter. With phosphatidic acid, this molar ratio continues to increase, but at a much reduced rate (e.g. 150 nmol of Ca'+/ml after 1.5 hours, rising to 230 nmol/ml after 6.5 hours).
The limiting molar ratio of Caz+ to phospholipid for both cardiolipin and phosphatidic acid was determined by partition of Cal+ from an aqueous to an organic phase with variable concentrations of Cal+ in the aqueous phase and a fixed concentration of the phospholipid in the organic phase. The Bligh-Dyer procedure (see' "Experimental Procedures") was used to generate the aqueous and organic phases used in these partition experiments. The aqueous phase was supplemented with varying concentrations of '"CaCl,, buffered with either 25 mM TMA cacodylate at pH 5.4 or 25 mM TMA Tricine at pH 8.3. The molar ratio of Ca ?+ to cardiolipin in the chloroform layer was found to be 1:l regardless of pH as long as Ca2+ was in excess relative to cardiolipin.
With phosphatidic acid, this ratio was dependent on pH, being 1:1.7 when the pH of the aqueous phase was 5.4 and 1:l when the pH of the aqueous phase was 8.3. The variable stoichiometry of Ca*+ to phosphatidic acid in the organic phase during CaZ+ translocation is probably referable to the pH values used in the aqueous compartments and the transition from a 1:2 molar ratio to a 1:l molar ratio with time.
In contrast to the stoichiometric extraction of Ca*+ into the organic phase by these phospholipids, the Cl-concentration was less than 1% of the value for the Ca*+ concentration.

Rate of Cal+ Transport as Function of Concentra$on of
Phospholipid-The rate of Ca z+ transport is a straight line function of the cardiolipin concentration and of the square of the concentration of phosphatidic acid in the bulk phase (Fig.  3). These experiments were carried out in the small cell described under "Experimental Procedures." At the low phospholipid concentrations used in these experiments, the lag period was eliminated and no turbidity was observed in the organic phase during any run except with phosphatidic acid at 21 PM concentration.

DISCUSSION
The data which have been presented in this communication establish that cardiolipin and phosphatidic acid are highly efficient ionophores for divalent and monovalent cations. Implicit in the term "ionophore" is the assumption of a stoichiometric interaction between the cation and the molecular species designated by this term, and also of a separation of the cation from the anion with which it was associated in the aqueous phase. In both partition experiments and transport studies in the Pressman cell after steady state has been reached, these two conditions are met. One molecule of cardiolipin induces the transfer of one Cal+ from the aqueous to the organic phase without the simultaneous transfer of an anion; 1 molecule of phosphatidic acid induces the transfer of 1 atom ion of Caz+ at alkaline pH (pH 8.3) and 0.5 atom ion of Ca2+ at acid pH (pH 5.4) (see "Ionophore-mediated Partition of Cal+ between Aqueous and Bulk Phases in Pressman Cell"). Time (hr.9 FIG. 3. Rate of Cal+ translocation as a function of the phospholipid concentration. The small cell version having a capacity of 1.70 ml of organic phase and the standard conditions as described under "Experimental Procedures" was used. A, cardiolipin at 1.4, 2.8, 4.2, and 13 PM concentration in the organic phase; B, phosphatidic acid at 3.4, 8.5, and 21 pM concentration in the organic phase. "Ca (-1000 cpm/nmol) accumulation in the receiver compartment was determined as described in the legend for Fig. 1.
Moreover, the rate of translocation of Ca*+ is strictly proportional to the concentration of cardiolipin, as we would expect for a 1: 1 molar interaction of Ca2+ and cardiolipin, whereas the rate of Caz+ translocation is determined by the square of the concentration of phosphatidic acid, consistent with the 1:2 molar ratio for the interaction between Ca*+ and phosphatidic acid during transport (Fig. 3). In comparing the transport properties of X537A, an authentic ionophore for divalent metal ions (19), and those of cardiolipin and phosphatidic acid, we find no experimental basis for interpreting the transport properties of the phospholipids in terms other than those of ionophores.
Phospholipids tend to form polymeric structures (micelles) in solvents such as chloroform by virtue of their bimodal structure (29). There is no evidence presently available as to the state of the phospholipid when formed in a complex with monovalent or divalent cations. But whether the complexed phospholipids move as separate molecules, or as sets of complexed molecules would in no way affect the validity of the fundamental thesis that cardiolipin and phosphatidic acid can act as ionophores for cations. The simple linear relation between the rate of transport of Caz+ and the concentration of cardiolipin in the bulk phase would tend to argue against the postulate of polymeric arrays of the complexed phospholipid.
In the encapsulation of cations within classical ionophores, the hydration shell of the cation is shed in the course of its transit from the aqueous phase to the interior of the ionophore (12). It could be argued that the partial or complete shedding of the hydration shell by the cation is in fact the hallmark of an ionophore, and until this property is verified, the designation of any species as an ionophore may be premature. We completely agree with this definition of an ionophore. But we maintain that the very fact of the transfer of the cation from the aqueous phase to the interior ot the phospholipid provides the proof that the water hydration shell is being shed. There has to be a free energy drop to drive the transfer of the cation, and it is in fact the substitution of the electronic links between the cation and the coordinating groups in the phospholipid for the electronic links between the cation and the oxygen atoms of water which drives this transfer of the cation.
The sensitivity of cardiolipin or phosphatidic acid to the exact composition of ttz organic phase through which the cation must be transported deserves some comment. The initial rate of transport of cations by these two phospholipids is greatly enhanced in a chloroform/methanol/water mixture as compared to a chloroform/water mixture. In the former mixture, the lag phase in transport can be completely eliminated. The polarity of the bulk phase is a crucial factor for the transport of cations by ionophores generally (30). It would be a mistake to interpret this sensitivity of the transport properties of phospholipid to the composition and polarity of the bulk phase as a special idiosyncracy of phospholipid ionophores. That phospholipids such as cardiolipin and phosphatidic acid can function as ionophores is now an established fact. What remains to be evaluated is whether this ionophoric capability has physiological significance. Phospholipids can exist either as integral components of the bilayer continuum of biological membranes or as components of systems which are not in the bilayer modality.
The ionophoric capability of phospholipids would be physiologically relevant only when the phospholipids are not in the bilayer modality, a modality which would reduce to negligible proportions their translational mobility (31). But it would be relevant in all instances where the phospholipid would be capable of translational mobility. The lipid. c complex would be one such example and numerous other examples in which containment of proteins in lipid bilayers increases phospholipid mobility can be cited (32-34). The containment of phospholipids within membranespanning protein systems, analogous to the ionophores isolated by Blondin (1,2) would be another. The association of phospholipids with other molecules such as butacaine, an anaesthetic, which would lead to the formation of a complex with translational mobility of the phospholipid in the membrane phase would provide a third means for bypassing the constraints of the bilayer modality (6,35,36).
The present study is the first in a series dealing with the ionophoretic capability of phospholipids.
On the basis of other still unpublished studies in our laboratory, we are suggesting that phospholipids play a not inconsiderable role in the induction of ion movements and that this role is assumed by phospholipids which are not constrained within the bilayer modality of biological membranes.