Membrane Energization by Proton Pumps is Important for Compartmentalization of Drugs and Toxins: a New Type of Active Transport

where R is the gas constant, T is the temperature in K and F is the Faraday constant. Using pmf as a driving force, transporters accumulate substrates against a concentration gradient (Table 1). Several substrates/proton antiporters and uniporters have been identified in endomembrane systems (Njus et al. 1986; Kanner and Schuldiner, 1987; Johnson, 1988; Anraku, 1987; Schuldiner et al. 1995). In brain synaptic vesicles, at least four kinds of transport systems for neurotransmitters have been identified: the uptake of acetylcholine, monoamines or GABA/glycine may be catalyzed by substrate/proton antiporters (Njus et al. 1986; Kanner and Schuldiner, 1987; Marshall and Parsons, 1987; Haigh et al. 1994; Johnson, 1988; Sudhof and Jahn, 1991; Schuldiner et al. 1995), whereas the vesicular glutamate transporter seems to be a uniporter that transports glutamate anions electrophoretically (Maycox et al. 1988; Cidon and Shira, 1989; Shioi and Ueda, 1990; Moriyama et al. 1990). All these neurotransmitter transporters have been identified in microvesicles, the counterparts of brain synaptic vesicles in various endocrine cells (Annaert et al. 1993; Thomas-Reets et al. 1993; Erickson et al. 1994; Moriyama et al. 1992, 1995, 1996; Moriyama and Yamamoto, 1995a,b). These transport systems appear to use a chemiosmotic energy-coupling mechanism: a primary transporter (pump) energizes the membranes and a secondary transporter then drives the uphill transport of the substrates (Table 1) (see also Harvey and Slayman, 1994).

However, some systems do not belong to the transport categories mentioned above. For instance, exogenous phenazinemethosulfate incorporated into thylakoid membranes accelerates the rate of electron flow more than tenfold by catalyzing an oxidation and reduction cycle of endogenous electron donors and acceptors (Nelson et al. 1972; Trebst, 1974). Quinone derivatives in liposomal membranes transfer protons from extravesicular ascorbate into internal ferricyanate, resulting in the formation of a proton gradient (inside acidic). In these cases, the oxidation/reduction potential converts an electrochemical proton gradient without any membrane enzymes: that is, accumulation against a concentration gradient can occur without secondary or primary transporters. This type of transport has received little attention so far. However, I observed that a similar type of transport often occurs when the compartmentalization of various drugs and toxins is measured. In this article, I briefly discuss the significance of energy-dependent but transporter-independent drug transport for understanding the mode of action of drugs and toxins.

Most neuron blockers and local anesthetics are amphipathic amines. Since methylamine and aminoacridine derivatives, such as Acridine Orange, penetrate through membranes in a neutral form and are accumulated as a protonated form inside acidic compartments (Schuldiner et al. 1972; Dean et al. 1984) (Fig. 1A,B), similar accumulation could be expected for these psychomimetic drugs (Moriyama et al. 1991a, 1993b). Upon the addition of ATP, chromaffin granule membrane vesicles took up radiolabeled chlorpromazine, haloperidol and propranolol (Moriyama et al. 1993b). This uptake was sensitive to proton conductors such as 3,5-di-t-butyl-4-hydroxybenzylidene malononitrile (SF6847) and to V-ATPase inhibitors such as bafilomycin A1, indicating that an electrochemical proton gradient established by a V-ATPase drove the uptake. Ammonium ions and nigericin plus K⁺, which dissipate ΔpH, abolished the uptake, whereas valinomycin plus K⁺, a dissipator of ΔΨ? had no effect or was slightly stimulatory, indicating that ΔpH, but not ΔΨ, is a driving force for the uptake. The ATP-dependent uptake was insensitive to reserpine, tetrabenazine, imipramine (inhibitors of vesicular or plasma membrane monoamine transporter) (Njus et al. 1986; Johnson, 1988) and to vanadate and verapamil (inhibitors of drug-transporting P-glycoprotein) (Gottesman and Pastan, 1988; Tsuruo, 1991), suggesting that the accumulation process is diffusion across the lipid bilayer and that transporters specific for these drugs are not involved in this concentration process. Similar ATP-dependent and proton-conductor-sensitive accumulation of these compounds was observed in synaptic vesicles and reconstituted proteoliposomes containing only purified V-ATPase as a protein component. Furthermore, proteoliposomes containing purified F-ATPase (bacterial F₁F_o-ATP synthase) accumulated these neuron blockers using only ΔpH as a driving force (Fig. 2; Table 2). These results confirmed the hypothesis that the drugs are accumulated in an energy-dependent but transporter-independent manner (Moriyama et al. 1993b).

The energy-dependent but transporter-independent transport of drugs shows the following properties. (1) This process occurs against a concentration gradient and requires an energy supply through a primary proton pump. (2) This process is mediated by permeation through the lipid bilayer, and no specific transporters are involved. According to the classification of transport systems (Table 1) (Harvey and Slayman, 1994), this type of transport does not belong to any known transport category and represents a new type of ‘active transport’ (Table 1).

Accumulation of chlorpromazine in acidic compartments has been observed in cultured cells (M. Hayashi and Y. Moriyama, in preparation). Amphetamine derivatives (Sulzer and Rayport, 1990; Rudnick and Wall, 1992) or dopamine analogs (Johnson et al. 1982; Gasnier et al. 1986) are accumulated in acidic organelles in a similar manner. The energy-dependent but transporter-independent accumulation of amphipathic drugs into acidic organelles also occurs in vivo.

Since more than 2 units of ΔpH was formed at the steady state of energized chromaffin granules or synaptic vesicles, more than a 100-fold concentration gradient of neuron blockers and related drugs could be attained in cultured cells as well as in isolated membrane vesicles (Moriyama et al. 1993b). Thus, when administered at micromolar levels in the blood, these drugs could be concentrated more than 100-fold in synaptic vesicles. They could be secreted into the synaptic cleft through exocytosis upon nerve stimulation so as to cause effective nerve blocking due to binding to the receptors on postsynaptic membranes (Fig. 3). The concentrations of these drugs at presynaptic terminals may be important for effective nerve blocking. Therefore, synaptic vesicles may be one of the important target sites of psychomimetic drugs.

When excess amounts of neuron blockers or local anesthetics are present, the entry of these psychomimetic drugs may lower ΔpH and increase ΔΨ in synaptic vesicles, because ΔpH is converted to ΔΨ upon the accumulation of the protonated forms of drugs under ideal conditions (Fig. 3). Chlorpromazine strongly dissipated ΔpH but slightly increased ΔΨ without affecting V-ATPase activity (Fig. 4). Under these conditions, the uptake of neurotransmitters into synaptic vesicles must be selectively perturbed: the uptake of monoamines or GABA, which is driven by ΔpH, was decreased, while the uptake of L-glutamate, an excitatory neurotransmitter taken up by the ΔΨ, was stimulated to some extent (Moriyama et al. 1993b). All amphipathic neuron blockers and local anesthetics tested so far exhibited essentially the same tendency. The pharmacological significance of these phenomena is not known at present. However, the administration of excess amounts of local anesthetics or neuron blockers always causes ‘excitation’ as a side effect (Tucker and Mather, 1980; Runciman, 1987), which is consistent with the possible increase in the amount of glutamate released from presynaptic terminals due to increased glutamate uptake into synaptic vesicles. Ammonia is known to be a neurotoxin which affects glutaminergic neurotransmission under hyperammonemic conditions such as hepatic encephalopathy (Szerb and Butterworth, 1992). Enhanced glutamate release into the synaptic cleft, which is at least partly due to selective perturbation of transmitter uptake (Albrecht et al. 1994), is one mechanism for ammonium neurotoxicity (Moroni et al. 1983; Butterworth et al. 1991).

It is well known that daunomycin and some anti-neoplastic agents such as doxorubicin and vinblastine are accumulated in acidic compartments in vivo (Willingham et al. 1986) (Fig. 5A), this being one of the important pathways for the extrusion of anti-neoplastic agents from drug-resistant cells (Beck, 1987; Sehested et al. 1988; Marquardt and Center, 1991). However, P-glycoprotein, an active drug extrusion pump, was not detected in acidic organelles (Willingham et al. 1987), so the mechanism by which daunomycin is accumulated in acidic compartments is unclear. We showed that the accumulation of daunomycin and related anti-neoplastic agents is also energy-dependent but transporter-independent (Moriyama et al. 1994). These compounds are also amphipathic amines and are accumulated in chromaffin vesicles upon the addition of ATP (Fig. 5B). The uptake was insensitive to various transporter inhibitors and to inhibitors of P-glycoprotein, but sensitive to bafilomycin A1, concanamycins and various proton conductors such as SF6847, indicating that an electrochemical proton gradient established by V-ATPase drove the uptake (Moriyama et al. 1994). Dissipation of ΔpH by ammonium ions or nigericin plus K⁺ diminished the uptake, whereas treatment with valinomycin plus K⁺ to cancel the ΔΨ slightly stimulated the uptake, indicating that ΔpH was the sole driving force for the uptake. Similar ATP-dependent uptake can be reconstituted in proteoliposomes containing purified V-or F-ATPase. Daunomycin disappeared from acidic organelles in vivo when the internal pH of the acidic organelles became neutral after addition of bafilomycin A1, concanamycins or nigericin (Moriyama et al. 1994).

Ca²⁺ channel blockers and calmodulin inhibitors are known to increase the toxicity of daunomycin (Tsuruo, 1991). This effect could be explained by inhibition of P-glycoprotein by these compounds so as to increase the free concentration of daunomycin in the cytoplasm (Tsuruo, 1991). However, these compounds are also amphipathic in nature and dissipate ΔpH of acidic organelles so as to inhibit the accumulation of daunomycin in acidic organelles (Moriyama et al. 1994). The effect may also increase the free concentration of daunomycin in the cytoplasm, resulting in elevated cytotoxicity. Therefore, the amphipathic nature of these compounds may be one of the reasons why a variety of chemically unrelated compounds equally enhance the cytotoxicity of daunomycin.

The concept of energy-dependent but transporter-independent accumulation of drugs can be expanded to other organelles such as mitochondria. It is well known that mitochondria accumulate hydrophobic cations, such as Rhodamine 123 or JC-1, as a consequence of electrophoretic movement of cations across the mitochondrial membrane (Fig. 1C,D) (Johnson et al. 1980; Smiley et al. 1991). Rhodamine 123 and related compounds are used as anti-neoplastic agents and the specific accumulation of these compounds in mitochondria is important for their selective cytotoxicity (Modica-Napolitano and Aprille, 1987). Methylphenyltetrahydropyridine (MPTP) is a neurotoxin causing artificial parkinsonism (Langston et al. 1983; Heikkila et al. 1984). MPTP is oxidized in vivo to the active form 1-methyl-4-phenylpyridnium (MPP⁺) (Markey et al. 1984). MPP⁺ accumulates in mitochondria in neuronal or other cells and inhibits complex I of the respiratory chain, causing ATP depletion and cell death (Mizuno et al. 1987). This accumulation has been shown to be energy-dependent (Ramsay and Singer, 1986), but so far no transporters responsible for the uptake have been identified. We suggest that, like Rhodamine 123, MPP⁺ is also taken up in a transporter-independent manner. First, MPP⁺ is a relatively hydrophobic and permeant cation, which is taken up by chromaffin cells in a transporter-independent manner (Reinhard et al. 1990). Second, the involvement of tetraphenylboron (TPB⁻) significantly increased the uptake of MPP⁺, possibly due to increased hydrophobicity and reduction of the charge upon the formation of the MPP⁺/TPB⁻ complex (Aiuchi et al. 1988), as in the case of the tetramethylphenylphosphonium ion (TPMP⁺), an indicator of the membrane potential (Szmelcman and Adler, 1976). Third, the uptake of MPP⁺ was observed in liposomes when only an inside-negative membrane potential was imposed with K⁺/valinomycin (Y. Moriyama, unpublished observation).

It is noteworthy that MPP⁺ is recognized as a substrate by plasma membrane and vesicular monoamine transporters (Chiba et al. 1985; Darchen et al. 1988; Danielis and Reinhard, 1988; Reinhard et al. 1990; Moriyama et al. 1993a). This is the reason why MPP⁺ is retained for a long time in dopaminergic neurons and adrenal glands (Johannessen et al. 1986). Thus, I suggest that transporter-dependent transport of MPP⁺ is involved in the long-term and selective accumulation of MPP⁺ in dopaminergic neurons and, in turn, that transporter-independent but energy-dependent accumulation occurs in neuronal mitochondria and causes an energy crisis, resulting in cell death.

Some drugs or toxins are actively accumulated through lipid bilayer in organelles energized by proton pumps. This type of transport belongs neither to ‘simple diffusion’ nor to ‘facilitated diffusion’, because this process involves ‘uphill’ transport (Table 1). I would like to propose that this type of drug accumulation is termed ‘simple concentration’ (Table 1). Simple concentration is frequently seen upon administration of various kinds of drugs and toxins. After accumulation in the organelles or in restricted compartments, these drugs may exert their pharmacological effects and/or perturb the organellar functions, thereby exerting their toxicological effects. Therefore, ‘simple concentration’ is important for understanding modes of action of drugs and toxins. In this article, I have only discussed the accumulation of amphipathic drugs and hydrophobic cations. It is also possible that some acids, such as salicylic acid, may accumulate in alkaline environments (Oku, 1992) and that hydrophobic anions, such as suramin (an inhibitor of various enzymes, Moriyama and Nelson, 1988), are distributed among subcellular organelles by means of a similar mechanism.

I am very grateful to Professor M. Futai (Osaka University) and Professor N. Nelson (Tel Aviv University) for their encouragement throughout these studies and to Dr T. Yoshimori (Kansai Medical University), Dr Y. Tahara (Protein Research Institute), Mr H. Yamada and Miss M. Hayashi of this laboratory for taking the photographs.