The role of V-ATPase in neuronal and endocrine systems

Synaptic vesicles play an important role in neuronal transmission in nerve termini. They accumulate high concentrations of neurotransmitters and, upon receiving the proper signal, secrete them by exocytosis. Similarly, secretory granules in endocrine cells accumulate biogenic amines and secrete them by exocytosis (Ueda, 1986; Johnson, 1988; Njus et al. 1987). We have been interested in the bioenergetics of the accumulation process, which is an initial step in neuronal transmission or biogenic amine secretion. Our results are consistent with a chemiosmotic energy coupling mechanism for the accumulation: a V-ATPase forms an electrochemical gradient of protons across vesicular membranes and transmitters are taken up through specific transporters coupled with this gradient (Fig. 1). Here, we discuss our recent results supporting this model.

V-ATPases purified from adrenal chromaffin granules (Moriyama and Nelson, 1987a) and brain clathrin-coated vesicles (Xie and Stone, 1986) have essentially the same subunit structures (115, 72, 57, 41, 39, 34, 33, 20 and 16×10³M_r subunits with a stoichiometry of 1, 3, 3, 1, 1, 1, 1, 1 and 5–6) (Arai et al. 1988). Cold treatment in the presence of MgATP and appropriate salts (e.g. 0.1 moll⁻¹ NaNO₃) caused rapid inactivation accompanied by release of a group of hydrophilic subunits (subunit A, B, C,D and E of 72, 57,41, 34 and 33 × 10³M_r respectively) but four polypeptides (115, 39,20 and 16×10³M_r) remained in the membrane (Moriyama and Nelson, 1989a). Subunit A contains an ATP binding site (Moriyama and Nelson, 1987b) and is homologous to the subunit of F₁-ATPase (Zimniak el al. 1988). Subunit c (16×10³M_r) binds N,N′-dicyclohexylcarbodiimide (DCCD) and is homologous to the c subunit of the F_o portion of F-ATPase (Mandel et al. 1988; Sutton and Apps, 1981). These findings indicate that the enzyme is composed of two different sectors (catalytic and hydrophobic membrane sectors). All V-ATPases from plants and animals show similar properties and essentially the same subunit structures (Moriyama and Nelson, 1989b,c). This conclusion was confirmed by analyses of the primary structures of the subunits (for reviews, see Nelson, 1989;Forgac, 1989; Futai et al. 1990).

Immunoeiectron microscopy, using proteoliposomes reconstituted with purified chromaffin granule V-ATPase, demonstrated the presence of particles of 10nm diameter on the membrane: these particles could be removed by cold treatment and were recognized by antibodies against the hydrophilic sector of chromaffin granule ATPase (Moriyama et al. 1991c). Thus, this particle is the hydrophilic sector of V-ATPase and is similar to F| of F-ATPase.

Using the immunoblotting technique Moriyama and Nelson (1990) and Cidon and Shira (1989) demonstrated that brain synaptic vesicles possess V-ATPase. The enzyme is a major constituent of the vesicles (approximately 20% of the total membrane protein) (Moriyama and Futai, 1990a) (Fig. 2). This result also suggests a high density of the ATPase in the brain, as synaptic vesicles are abundant in the brain. In fact, the levels of mRNAs encoding subunits of V-ATPase were higher in the brain than in other tissues (Nelson et al. 1990; Wang et al. 1988; Hirsch et al. 1988). Is the V-ATPase of synaptic vesicles structurally similar to those in other organelles, or are different isoforms expressed in different organelles? To answer this question, we cloned a cerebellar cDNA encoding a I6× 10³M_r subunit, a major intrinsic membrane protein of V-ATPase (Hanada et al. 1991). The deduced amino acid sequence showed that this brain polypeptide (155 amino acids resides; relative molecular mass, 15 808) was highly hydrophobic and homologous to subunits from bovine adrenal medulla, the Torpedo marmorata electric lobe, Drosophila and yeast. Glu-139, thought to be essential for proton transport, was also conserved as a potential DCCD binding site.

We then cloned four different 16×10³M_r genes including pseudogenes from a human genomic bank using the cDNA as a probe (Hasebe et al. 1992). One gene corresponded to the gene encoding the 16×10³M_r subunit expressed in HeLa cells and had a coding sequence separated by two introns. The second intron was located in the DNA segment coding for a loop region between the second and third transmembrane helices, suggesting that the 16× 10³M_r subunit with four transmembrane helices had evolved by gene duplication. We are analyzing whether other genes are actually transcribed. The primary sequence deduced from the second clone had a termination codon behind the third transmembrane helix. Possible translation products from other clones had no putative acidic residues essential for proton transport. We do not yet have evidence for organelle-specific differences in V-ATPases. However, detailed studies on genes from different sources may eventually answer this question.

Synaptic vesicles contain transporters for monoamines, acetylcholine, anionic glutamate and γ-aminobutyrate (GABA) as well as zwitterionic glycine (Ueda, 1986; Johnson, 1988; Njus et al. 1987; Hell et al. 1988; Kish et al. 1989). A monoamine transporter that is sensitive to reserpine and tetrabenzine is also present in various secretory granules. We concluded that V-ATPase couples the uptakes of these transmitters to ATP hydrolysis, because selective inhibition of the ATPase by bafilomycin Ai and cold treatment or treatment with sodium carbonate diminished the uptakes (Moriyama and Futai, 1990a).

The electrochemical gradient of protons established by V-ATPase consists of a proton gradient and a membrane potential. The uptakes of monoamines and GABA are sensitive to nigericin plus K⁺ or ammonium (a dissipator of the proton gradient) but are not affected by valinomycin plus K⁺ (a dissipator of the membrane potential), indicating that these uptakes are coupled to a proton gradient. In contrast, glutamate uptake is coupled to a membrane potential (inside positive), because it is sensitive to valinomycin plus K⁺ or a high concentration of Cl⁻ but insensitive to nigericin plus K⁺ and slightly accelerated by ammonium (Ueda, 1986; Moriyama et al. 1990).

Purification of the transporter from synaptic vesicles has been hampered by the difficulty of measuring its activity after its solubilization from the membranes. We are able to measure the activity of transporters by following uptake into liposomes upon application of an artificial driving force, but this procedure was not suitable as a routine assay. However, the transport activity was assayed easily by the addition of ATP after the solubilized transporters had been incorporated into liposomes along with F-ATPase. We established a rapid and efficient purification procedure for F-ATPase from overproducing Escherichia coli cells (Moriyama et al. 1991b). The monoamine transporter was solubilized from chromaffin membranes after removal of V-ATPase by cold treatment. It was mixed with purified F₁F₀ and azolectine liposomes, freeze-thawed and diluted with an appropriate buffer. The resulting proteoliposomes exhibited reserpine-sensitive, ATP-dependent monoamine uptake, which was sensitive to sodium azide (an inhibitor of F-ATPase) but resistant to bafilomycin A₁ and N-ethylmaleimide (inhibitors of V-ATPase), indicating that the driving force for the uptake was supplied by the F-ATPase. The glutamate transporter could be reconstituted by a similar method (H.-L. Tsi, Y. Moriyama and M. Futai, unpublished observation). It should be noted that limited numbers of proteins derived from chromaffin vesicles were incorporated into liposomes, suggesting that the transporters were partially purified. Thus, the reconstitution systems using F1F0 will be useful for identifying and purifying the transporters.

Neuronal transmission is perturbed by reagents that abolish the accumulation of transmitters in vesicles. These reagents can be classified as inhibitors of various steps in the formation and utilization of the electrochemical gradient in the vesicles. One class of such compounds interacts directly with the transporters and inhibits their activities. Reserpine, an example of a specific inhibitor of vesicular transport, binds to the monoamine transporter from inside the vesicles and blocks adrenergic neurons (Shore and Norn, 1971). Another class of inhibitors dissipates the electrochemical gradient of protons (e.g. CCCP and ammonium), resulting in indirect inhibition of the uptake (Johnson, 1988). We found that some neuron blocking reagents show both types of inhibition.

Vesamicol (AH5183) and bromocryptine were proposed as inhibitors of the acetylcholine transporter (Marshall and Parsons, 1987) and a-blocker (Fuxe et al. 1978), respectively. However, they also inhibited dopamine transport in synaptic vesicles and chromaffin granules (Moriyama et al. 1991c). The inhibition was not due to the direct interaction of these compounds with the monoamine transporter but to dissipation of the electrochemical gradient across secretory vesicles by their uncoupling effect. Thus, treatment of vesicles with these reagents caused non-specific inhibition of transmitter uptake by the synaptic vesicles without affecting the V-ATPase activity. The uncoupling may be due to protonation/deprotonation of the compounds as these compounds are hydrophobic tertiary amines. Various kinds of neuron blocking agents and local anesthetics are hydrophobic amines. Some neuron blockers and local anesthetics, including chloropromazine, strongly inhibit the formation of a proton gradient in secretory vesicles, thereby blocking the accumulation of monoamines (Y. Moriyama, R. Suzuki, H.-L. Tsui and M. Futai, unpublished observation) (Fig. 3). Thus, all of these compounds can be expected to show uncoupling effects similar to those of vesamicol and bromocryptine and their dissipation of the proton gradient may, at least in part, be the reason for the neuron blocking action.

False transmitters, another type of reagent interacting with the transporters, are recognized as substrates by the transporters and accumulate in the vesicles. MPP⁺, a highly bioactive metabolic product of MPTP, is a neurotoxin that causes degradation of dopaminergic neurons in man and experimental animals (artificial Parkinsonism) (Langston et al. 1983; Heikkila et al. 1984). We found that this neurotoxin is the substrate of the vesicular monoamine transporter.

Upon addition of ATP, MPP+ was actively accumulated in chromaffin membrane vesicles and synaptic vesicles, leading to a concentration gradient of about 100-fold (Y. Moriyama, K. Amakatsu and M. Futai, in preparation). The uptake was sensitive to bafilomycin Ai and reserpine, indicating that it was mediated by V-ATPase and the monoamine transporter. These results suggest that MPP⁺ accumulates in these organelles in vivo by this mechanism. Furthermore, addition of MPP⁺ inhibited dopamine uptake but did not affect the uptake of glutamate or GABA. These results are consistent with the hypothesis that MPP⁺ is toxic to dopaminergic neurons but has little effect on other neurons. The compartmentation of MPP⁺ may also be important for understanding the mechanism of long-term toxicity of the compound and will provide one molecular basis for development of a Parkinsonian syndrome.

In this article, we have described the structure and function of V-ATPase in synaptic vesicles. The electrochemical gradient of protons established by the ATPase is used as a driving force for the concentration of transmitters inside the organelles. Various drugs (e.g. neuron blockers, local anesthetics) or toxins (e.g. MPP⁺) may affect the uptake into the synaptic vesicles and so perturb neuronal transmission. Thus, V-ATPase has key physiological and pharmacological roles in the nervous and endocrine systems.

This work was supported by grants from the Ministry of Education, Science and Culture of Japan, the Human Frontier Science Program and Naito Fundation.