Physiology of V-ATPases

Protons migrate much faster than other ions through water, ice and water-lined membrane channels because they participate in hydrogen bonding and H⁺H₂O exchange. Similarly, hydrogen bonding enables protons with amino, carbonyl, phosphoryl and sulfonyl residues to influence critically the charge, conformation and stability of proteins. Therefore, it is not surprising that regulation of proton concentration, or pH, is an essential requirement in biological systems. It is no surprise either that enzymes which regulate proton concentration (i.e. proton pumps) should have evolved or that evolution should have used these enzymes further, for energization of biological membranes. At present there appear to be three classes of ATP-hydrolyzing proton pumps, or H⁺-ATPases, which were dubbed P-ATPases, F-ATPases and V-ATPases, by Pederson and Carafoli (1987). H⁺-translocating P-ATPases, as well as the Na⁺/K⁺-ATPase of plasma membranes and the Ca²⁺-ATPase of sarcoplasmic reticulum, form phosphoaspartyl intermediates and are inhibited by the phosphate analogue orthovanadate. F-ATPases are the proton-translocating ATP synthases of mitochondria, chloroplasts and bacterial plasma membranes and are inhibited by azide.

V-ATPases (V-type ATPases or vacuolar H⁺-ATPases) are the proton-translocating ATPases best known as acidifiers of cytoplasmic vesicles and vacuoles. During reaction, they separate protons from gegenions, thus producing an electrical potential difference, Δ ψ, across the membrane in which they reside. Δψ in turn can drive other ion movements, leading directly to acidification or alkalization and secondarily to salt transport and water movement. V-ATPases are inhibited by bafilomycin (in nanomolar concentrations) and N-ethylmaleimide (in micromolar concentrations).

In this paper the transport work that is accomplished when specific channels and porters are present in V-ATPase-energized membranes will be discussed using an intuitive thermodynamic analysis. Martin (1992) discusses similar cases using ionic circuit analysis. The two approaches provide complementary frameworks for understanding how the (output) compartment receiving protons from a V-ATPase can be rendered acidic, neutral or even alkaline, depending upon the selectivity of other carriers (porters) and channels present, upon the strength of the gegenions and upon the cellular and compartmental geometry. Encouragingly, both types of analysis lead to similar conclusions.

Biochemical studies on V-ATPases were initiated on vacuolar membranes of fungi and plants and on clathrin-coated vesicles and other components of endomembrane systems (reviewed by Sze, 1985; Mellman et al. 1986; Nelson, 1988, 1989; Forgac, 1989; Stone et al. 1989). However, V-ATPases were soon found to recycle between vesicles and the plasma membrane in renal epithelia (Gluck et al. 1992; Gluck, 1992). Increasingly they are being found as permanent residents of plasma membranes e.g. phagocytic cells (Grinstein et al. 1992), insect gastrointestinal and sensory epithelia (Wieczorek, 1992; Klein, 1992) and frog skin (Harvey, 1992). Unlike the Na⁺/K⁺-ATPase, which usually energizes basolateral membranes of ion-transporting epithelial cells, V-ATPases usually energize apical plasma membranes in renal and epidermal epithelia of vertebrates and in gastrointestinal and sensory epithelia of insects.

V-ATPases are ancient and ubiquitous, being closely related to the ATP synthases of archaebacteria (Ihara et al. 1992). They play a key role in the acidification of endomembrane-bordered compartments of all eukaryotic cells. In mammalian cells V-ATPases help make Golgi vesicles mildly acidic, clathrin-coated vesicles more acidic and lysosomes highly acidic (Nelson, 1992a,b; Forgac, 1992). They are involved in endocytosis and exocytosis, acidification being important in the addition of carbohydrate moieties, which act as recognition sites during cell sorting (Mellman, 1992; Klionsky et al. 1992). In yeasts and in other fungi such as Neurospora, V-ATPases acidify vacuoles used as storage sites for amino acids, Ca²⁺, carbohydrates, phosphate and hydrolases (Bowman et al. 1992; Stevens, 1992). In tonoplasts, the membranes surrounding the central vacuole of plant cells, they energize salt regulation and osmotic balance (Sze et al. 1992; Bertl and Slayman, 1992). V-ATPases energize accumulation of neurotransmitter amines by synaptic vesicles and chromaffin granules (Moriyama et al. 1992). In osteoclast membranes they acidify the extracellular fluid and lead to bone reabsorption, both normal and abnormal (as in osteoporosis; Chatterjee et al. 1992). In renal plasma membranes V-ATPases energize the acidification of urine and the reabsorption of bicarbonate by the blood that are essential for acid-base balance (Gluck, 1992; Gluck and Nelson, 1992; Brown et al. 1992).

Since protons are pumped into an acidic compartment in all of the above cases it would be tempting to conclude that proton transport leads only to acidification. However, a V-ATPase in the apical plasma membrane of phagocytic cells alkalizes the external medium (Grinstein et al. 1992); similarly, a V-ATPase in the apical plasma membrane of certain insect epithelial cells (Wieczorek et al. 1991) can produce a lumen pH exceeding 11 (Dow, 1984). The precise way in which a V-ATPase in the goblet cell apical membrane, GCAM, of lepidopteran midgut epithelial cells can alkalize the lumen is discussed by Wieczorek (1992), Klein (1992), Dow (1992), Wolfersberger (1992), Moffett and Koch (1992), Martin (1992) and Zeiske (1992).

However, a brief preview of the insect plasma membrane V-ATPase story may be helpful here. In goblet cell apical membranes of living lepidopteran midgut a very active V-ATPase leads to alkalization of the lumen, not acidification. This result comes about, first, because the transporting membrane is nearly impermeable to anions, which therefore cannot be driven by the membrane potential difference, Δ ψ (>180 mV), through the membrane in pursuit of protons, to maintain electroneutrality and, second, because a K⁺/nH⁺ antiporter in the same membrane rapidly exchanges protons for potassium ions. The large apical Δ ψ from proton pumping is also used to energize amino acid absorption and to regulate the concentrations of other ions (Dow, 1986). In Malpighian tubules and salivary glands massive K⁺ secretion is coupled to water flow, forming ‘primary urine’ and saliva, respectively (Maddrell and O’Donnell, 1992). In insect sensory sensilla electrogenic proton transport modulates the receptor potential and leads to a potassium-rich receptor lymph cavity (Thurm and Küppers, 1980; Klein, 1992).

V-ATPases are multimeric complexes with cytoplasmic V₁ and membrane-bound V₀ domains analogous to the F| and F₀ domains of F-ATPases (Nelson, 1992a; Gluck et al. 1992). Cytoplasmic ATP is hydrolyzed in the V₁ domain, forcing protons away from the cytoplasm through the V₀ domain and across the membrane. The large, hydrophilic V₁ domains protrude as particles on the cytoplasmic face of vacuolar and plasma membranes in both acidifying and alkalizing systems (Table 1). Their structure in fungal vacuoles is discussed by Bowman et al. (1992), in kidney plasma membranes and vesicles by Brown et al. (1992) and in insect plasma membranes by Klein (1992).

F₁-like particles in insect tissues were first reported on the cytoplasmic surface of apical plasma membranes in K⁺-transporting rectal epithelia (Gupta and Berridge, 1966) and in lepidopteran midgut (Anderson and Harvey, 1966). Subsequently, they have been reported on the apical plasma membranes of K⁺-secreting trichogen and tormogen cells of insect sensory sensilla (see references in Thurm and Küppers, 1980) and on many other K⁺-transporting plasma membranes of insect epithelia, including Malpighian tubules (see references in Harvey, 1980). Like F₁ particles, Vi particles are approximately 9–10nm in diameter and are always present on cytoplasmic surfaces of cation-transporting membranes. Unlike ATP-synthesizing F₁ particles, V₁ particles always couple ATP hydrolysis to the creation of cation electrochemical gradients. The term ‘portaseme’ was suggested to denote the transport particles visible in electron micrographs of insect cation-transporting membranes (Harvey et al. 1981).

V-ATPases pump protons without internal counterions and therefore are inherently electrogenic. Since they use ATP, they are also strongly oxygen-dependent in cells with low anaerobic phosphorylating capacity or low phosphagen stores. They energize membranes by transducing the energy from ATP hydrolysis into a proton current, which establishes an electrochemical gradient Δ_μH-TO emphasize the special nature of this electromotive force for protons distributed across biomembranes, Mitchell (1961) coined the term proton-motive force, or pmf. Just as a conventional electromotive force (emf) can drive a multiplicity of processes depending upon other components in an electric circuit, so a pmf can drive many processes depending upon other carriers and channels in biomembranes. Alone in a lipid bilayer, active V-ATPase would form a large voltage with very small changes in H⁺ concentration. In the presence of anion channels, however, the enzyme would drive a continuing acid flux; and in the presence of a C⁺/nH⁺ antiporter or symporter and appropriate gegenions (anions), the enzyme can drive a continuing alkalizing flux. Finally, the enzyme can generate or modulate a receptor Δ ψ and therefore influence neural signaling.

In discussing these diverse functions systematically, we will assume (1) that an electrogenic V-ATPase moves protons from the cytoplasm to the opposite side of the membrane, rendering the output side positive with respect to the cytoplasmic side; (2) that electroneutrality is preserved in bulk solutions; (3) that, in the absence of other pumps, channels, symporters or antiporters, the membrane is impermeable to all ions; (4) that a C⁺/nH⁺ antiporter can be driven by either the electrical (Δ ψ) or the chemical (RTΔpH) component of the pmf; and (5) that the H⁺/ATP and C⁺/H⁺ stoichiometries can be considered conveniently to be 3/1 and 2/3, respectively. [The actual value of n for H⁺/ATP is thought to be 3 by analogy with F-ATPases. For lepidopteran midgut the K⁺/H⁺ ratio is known only to be less than 1; if it were 2/3 then the K⁺/ATP ratio would be 2, which is the minimal value estimated from oxygen measurements (Harvey et al. 1967).]

Fig. 1A describes the primary membrane-energizing step. A V-ATPase is inserted into a membrane which is otherwise impermeable to all ions, including protons. The enzyme reaction merely separates H⁺ from A⁻ across the membrane and generates a positive membrane potential (Δψ) which provides capacitative energy for electrochemical work. The Δ ψ which develops instantaneously across the membrane capacitance equals the total pmf transduced by the V-ATPase. The AG for ATP hydrolysis is given by RT\nK_e/Q (where K_e is the equilibrium constant and Q is the products to reactants concentration ratio for ATP hydrolysis under actual conditions). This value is the total free energy available and sets an upper limit upon the pmf. The maximal pmf (in volts) can be estimated by dividing ΔG (in kJ mol⁻¹) by nzF (where n is the stoichiometric ratio of H⁺/ATP, z is the valence with sign and F is Faraday’s number). For example ΔG for ATP hydrolysis is estimated to be −44.1 kJ mol⁻¹ in lepidopteran midgut cells (Mandel et al. 1980); if n and z are both 1, then the pmf would be 440mV (Harvey et al. 1981). The protons charging the membrane capacitance can be expected to exchange instanteously with ions in the bulk solution; if a vesicle were filled with an aqueous nonelectrolyte (Fig. 1B), the H⁺/H⁺ exchange should have no effect on the pH of the bulk solution. The entire pmf should still appear as Δ ψ

However, real vacuoles and synaptic vesicles are likely to contain strong electrolytes, such as KC1. Instantaneous exchange of transported H⁺ for K⁺ could lead to an increase in the concentration of HCl in the vacuolar space (Fig. 1C). The maximum decrease in bulk-phase pH would depend upon the surface/volume ratio of the vesicle and the enclosed solute composition (Table 2) and would be diminished considerably by the buffering capacity of the surface membrane proteins.

In typical acidifying vacuolar membranes an anion conductance, normally associated with a chloride channel, is present in the membrane along with the V-ATPase (Fig. 2). Pumped protons are accompanied by anionic gegenions which cross the membrane via the channel, leading to a net proton-anion flux. The flux of a strong anion, such as chloride, would lead to net HC1 secretion. Much of the pmf would appear as a ΔpH with Δ ψ approaching zero and the output compartment becoming very acidic.

The pmf can also acidify the output compartment by exchanging a weak cation, the proton, for a strong cation, such as Na⁺ (Fig. 3). This combination amounts to a secondary Na⁺ pump and is thought to account for Na⁺ uptake from (low Na⁺) pond water by the frog skin (Harvey, 1992). The V-ATPase catalyzes separation of protons from anions across an anion-impermeable membrane. The resulting Δψ drives Na⁺ through a Na⁺ channel into the cell. Since the strong cation, Na⁺, is replaced by the weak cation, H⁺, the pond becomes ‘acidic’. Acidification of the kidney tubule lumen is thought to be accomplished by such a channel-mediated Na⁺/H⁺ interchange driven by the V-ATPase of the apical membrane (Gluck, 1992). In both frog skin and renal epithelia the V-ATPase and Na⁺ channel are located in separate cells.

In this theoretical case (Fig. 4A) the pmf drives electrophoretic C⁺/nH⁺ antiport. A V-ATPase is present in an ion-impermeable membrane containing a K⁺/nH⁺ antiporter that is responsive to the membrane potential. The Δ ψ drives protons (weak cations) back to the input side in exchange for strong cations, such as K⁺. K⁺ rather than H⁺ is separated from A⁻ across the membrane, yielding a Δ ψ nearly equivalent to the ΔG/n for K⁺/ATP (expressed in volts) of the ATP hydrolysis. The effect of the proton/cation antiport is to replace the very weak cation, H⁺, by a very strong cation, such as Na⁺ or K⁺, thereby replacing the conditions for acidification by conditions for alkalization.

Wieczorek et al. (1989) found that goblet cell apical membrane, GCAM, vesicles from Manduca sexta midgut became acidic (in vitro) in the absence of K⁺, implying a Cl⁻⁻conductance in the isolated vesicles. If the goblet cavity (output side of the GCAM in vivo) were to become acidic then the ΔpH could drive even an electroneutral K⁺/H⁺ antiport, which could recycle H⁺ to the cytoplasm and secrete K⁺ into the goblet cavity, thus accounting for the K⁺ ‘pumping’. However, Chao et al. (1991) ruled out an electroneutral antiport in the intact midgut on two grounds. (1) The chloride conductance of GCAM in vivo is very low, implying that HO could not be secreted. (2) The measured pH of the cytoplasm is 7.0, whereas that of the goblet cavity is a mildly alkaline 7.23, again implying that net HO is not being secreted into the cavity. Moreover, the Δ ψ across the apical membrane in isolated midgut can exceed 180mV, lumen positive (Dow and Peacock, 1989). Therefore, Chao et al. (1991) postulated an electrophoretic K⁺/nH⁺ antiport driven by the membrane potential. Wieczorek et al. (1991) identified and characterized the antiporter in isolated GCAM vesicles. They demonstrated that a V-ATPase in parallel with a C⁺/nH⁺ antiporter can render the output side of a proton-pumping membrane alkaline (Wieczorek, 1992; see also Grinstein et al. 1992).

The V-ATPase-K⁺/nH⁺ antiporter couple can also generate a higher Δ ψ than a V-ATPase alone. If the H+/ATP ratio for a V-ATPase is 3, then the maximal Δψ would be equivalent to AGI3. Coupled to an antiporter with a 2/3 K⁺/H⁺ ratio, the K+/ATP ratio would be 2 and the maximal A’P would be equivalent to ΔG/2. For example, in lepidopteran midgut AG is −44.1 kJ mol⁻¹, corresponding to approximately 450mV, which would yield a maximal Δ ψ of 150 mV for the V-ATPase operating alone but a maximal Δ ψ of 225 mV when coupled to the antiporter. Gluck (1992) argues that V-ATPases with H⁺/ATP stoichiometric ratios of 3 may have been selected for in the evolution of endomembranes, where high voltages might exceed dielectric limits, whereas P-ATPases with H⁺/ATP ratios of 1 may have been selected for in plant membranes. It appears that the midgut plasma membrane uses a V-ATPase yet generates a high voltage by adding a K⁺/nH⁺ antiporter to the V-ATPase-containing membrane. The V-ATPase/antiporter couple acts like a transformer, increasing the voltage across the membrane while decreasing the current (Martin, 1992).

In lepidopteran midgut and insect sensory sensilla a large transapical membrane potential is generated; the output compartment in each case is positive with respect to the cells and contains an extracellular sulfonated glycoprotein matrix. A role for this matrix is suggested by the finding that the pH of the goblet cavity is virtually neutral. In Fig. 4B, we assume that the V-ATPase catalyzes ATP hydrolysis, which separates 3H⁺ from 3A⁻ across GCAM, and that the 3H⁺ immediately exchange for 2K⁺via the electrophoretic antiporter. The 2K⁺ are neutralized by sulfated protein in the goblet cavity matrix. The matrix can be viewed as an extension of the membrane capacitance but charged with K⁺.

This case represents the apical region of an ideal isolated goblet cell in which a sulfated glycoprotein matrix receives transported potassium ions. X-ray microanalysis showed that the goblet cavity in isolated, open-circuit midgut has an elemental chlorine concentration corresponding to 19 mmol 1⁻¹ but an elemental sulfur concentration corresponding to 59 mmol 1⁻¹ (Dow et al. 1984). Assuming that the sulfur is present as sulfated proteoglycans, the strong potassium cations exchanged for the weak protons would be electrically balanced mainly by the strong sulfate anions, leaving the goblet cavity near neutrality and accounting for the pH of 7.23 recorded there by Chao et al. (1991). However, there would still be no explanation for the massive net K⁺ flux towards the lumen and the lumen pH of more than 10.5.

Fig. 5 is the current ‘standard’ model for midgut Δ ψ generation, K⁺ transport and lumen alkalization; it requires synergism between goblet and columnar cells. The weak carbonate anion, from the metabolism of both goblet and columnar cells, allows K⁺ to move from goblet matrix to lumen where K₂CO₃ accumulates and accounts for the high lumenal alkalinity. Dow’s (1984) suggestion that bicarbonate is secreted into the goblet cavity is ruled out by the pH of 7.23 recorded there; it seems likely instead that carbonate is secreted directly to the lumen.

V-ATPase activity in GCAM separates 3 protons from 3 cellular anions creating a membrane potential approaching the pmf calculated from is Δ G/n of ATP hydrolysis. The Δ ψ drives the 3 protons via the antiporter back to the cell in exchange for 2K⁺, leaving 2 anions behind in the cell, thereby maintaining a large Δ ψ The 2K⁺ associate with a sulfate group of the matrix protein. Simultaneously, metabolic CO₂ and H₂O form H₂CO₃, which, in the presence of carbonic anhydrase localized in CCAM and the goblet cell valve (Ridgway and Moffett, 1986), dissociates to 2H⁺ and CO₃²⁻. The CO₃²⁻ allows the 2K⁺ to leave the goblet matrix and move to the lumen. The strong cation, K⁺, associated with the very weak anion, CO₃²⁻ (pK₂=10.25 at 25°C), leads to the highly alkaline lumen. The AT is less than the pmf but remains high (Dow and Peacock, 1989; Moffett and Koch, 1988) because of a finite resistance to CO₃²⁻ movement so that it ‘lags behind’ the pumped K⁺ as they move by separate routes towards the lumen. The 2H⁺ from the H₂CO₃ electrically balance the 2A⁻ left behind in the basal extracellular space as 2K⁺ move into the cells replacing those antiported to the cavity. Since the cellular pH is near neutrality, a net increase must occur in hemolymph acidity.

The ratio of K⁺ transported to oxygen consumed can be as high as 2.0 in Hyalophora cecropia midgut (Harvey et al. 1967) so the ratio of K⁺ to carbonate could be 2/1 as suggested by this model. The mechanism of carbonate movement from cell to lumen is unknown, although carbonic anhydrase is present in the apical membranes (Ridgway and Moffett, 1986) and a bicarbonate/chloride antiporter has been postulated there (Chao et al. 1989). CO2 could cross the apical membrane by simple diffusion.

The midgut model has been helpful in explaining the mechanism of action of the caterpillar-specific delta endotoxin from Bacillus thuringiensis (Wolfersberger, 1992). In brief, the endotoxin binds to a receptor in columnar cell apical membrane, CCAM; a cation channel is formed, the Δ ψ is short-circuited and the membrane is de-energized. The proton barrier is thereby lost between pH 7 cells and pH 11 lumen and the cells become alkaline. F-ATPase activity, which requires mildly acidic cytoplasmic pH, is disrupted. The midgut cells break down and the caterpillar dies. The endotoxin is environmentally safe because the gastointestinal cells of birds and mammals do not possess a V-ATPase-energized GCAM in series with a Bt-receptor-containing CCAM, together forming a K⁺-impermeable apical membrane. Moreover, they do not possess the highly alkaline lumenal pH required to activate the endotoxin nor the high-affinity toxin-binding proteins.

Amino acid absorption in lepidopteran midgut is energized by the GCAM V-ATPase. Although the uptake is catalyzed by a K⁺/amino acid symporter, there is no K⁺ activity gradient between lumen and cells; instead the uptake is driven by the large Δ ψ, lumen positive, across the columnar cell apical membrane established by the electrically connected, V-ATPase-energized GCAM (Giordana et al. 1989).

What becomes of the K₂CO₃ in the midgut lumen and why is the hemolymph not acidic? The simplest hypothesis is that K⁺ is actively reabsorbed in the rectum, bringing the CO₃²⁻ along with it. Recycling of K⁺ is thought to keep the blood from becoming highly acidic; moreover, the slightly acid hemolymph pH allows CO₂ to be regenerated there, to re-enter the cells and to be excreted by the spiracles. Such acid-base shuttling is reminiscent of similar shuttling in the mammalian gastrointestinal tract. When HC1 is secreted to the stomach the hepatic portal blood becomes alkaline; conversely, when KHCO₃ is secreted in the pancreatic duct the hepatic portal blood returns to normality -the alkaline blood from the stomach never reaches the general circulation.

In a membrane in which the antiporter activity is small compared to the V-ATPase activity and the anion conductance is large, the pH can range from 7 to near zero depending upon the strength of the anion. However, if the antiporter activity is nearly as large as the V-ATPase activity then the pH can range from below 7 to near 14 depending upon the nature of the neutralizing cation or anion.

Whether the output compartment will be acidic, neutral or alkaline is conveniently deduced from the strong ion difference, SID, defined as the sum of the strong cation concentrations minus the sum of the strong anion concentrations (Stewart, 1981). If the SID is very positive, e.g. [K⁺] ≫ [Cl⁻], then the output solution is alkaline; if the SID is zero [K⁺]=[C1⁻] then the solution is neutral; finally, if the SID is very negative, e.g. [K⁺] ≪ [Cl⁻], then the solution is acidic (Table 3).

In Malpighian tubules and salivary glands of insects, massive fluxes of fluids (Maddrell and O’Donnell, 1992) are apparently energized by V-ATPases (Klein, 1992). This case is complex -an anion conductance must be present in the same membrane as a C⁺/nH⁺ antiporter to obtain the net salt movement required for osmotic water flow. The Δ ψ must be less than the V-ATPase pmf because part of the Δψ is shunted by the A⁻ flux.

V-ATPase energization of Malpighian tubule fluid transport can be explained by case 6. A net KC1 flux to the tubular lumen would set up a local osmotic gradient, enabling water to flow while shunting most of the Δ ψ. Since K⁺ is a strong cation and Cl⁻ is a strong anion, the pH would remain near 7.

Although the frog skin was the original model for Na⁺-transporting epithelia (Koefoed-Johnsen and Ussing, 1958), the mechanism by which Na⁺ could be taken up from dilute pond water remained unknown. It now appears that a V-ATPase is present in parallel with a Na⁺ channel in the apical (pond side) membrane (Harvey, 1992). Na⁺ is driven into the cells from the pond by the positive Δ ψ established by the V-ATPase; it is driven from the cells to the blood by the P-ATPase (Na⁺/K⁺-ATPase) in the basolateral membrane (Fig. 7). The transepithelial Δ ψ, TEΔψ, is the sum of the apically located V-ATPase Δ ψ and Na⁺ diffusion potential added to the basolaterally located P-ATPase Δ ψ and K⁺ diffusion potential (Fig. 7; modified from Harvey, 1992).

A similar mechanism accounts for the acidification of kidney tubular lumen and bicarbonate secretion to the blood (Gluck, 1992; Gluck and Nelson, 1992). Two cell types are involved. The V-ATPase is in the apical membrane of an A-type of intercalated cell whereas the Na⁺ channel is in the apical membrane of nearby cells. Weak protons are pumped to the lumen while the Δ ψ drives strong sodium ions from lumen to cell, accounting for the acidification of the lumen. The Na⁺ is pumped to the blood and HCO₃⁻ is exchanged for Cl⁻ across the basolateral membrane, accounting for NaHCO₃ absorption.

Many cases of secretion and absorption of ions and water by epithelial cells appear to be variations of a simple model -membrane energization (Fig. 8A) followed by ion transport (Fig. 8B). Membrane energization appears to be achieved by four mechanisms: (1) ATP synthesis using F-ATPases; ATP hydrolysis using (2) V-ATPases or (3) P-ATPases; and (4) assymetrically discharging metabolic gradients, e.g. CO₂. Ion transport can utilize (5) Cl⁻/HCO₃⁻ antiport, (6) K⁺ channels, (7) K⁺/H⁺ antiport, (8) anion channels, (9) Na⁺ channels and (10) cation/metabolite symport as well as numerous other transport proteins (Sze, 1992). Martin (1992) has quantified these complex relationships, using electrical circuit analysis. In conclusion, the electrogenicity of ion pumps under steady-state conditions -and particularly that of V-ATPases -is forcing us to rewrite membrane transport biophysics and physiology in terms of the interactions among pumps, channels and coupling transporters.

This review was supported in part by NIH research grant AI-22444. I am deeply indebted to Professor Dr Helmut Wieczorek, Dr Ulla Klein and Mr Ralph Graf for our ongoing collaboration regarding insect V-ATPases. I thank Dr Julian A. T. Dow, Dr Frank Martin, Dr Frans J. S. Novak, Dr Ranganath Parthasarathy, Dr Clifford L. Slayman and Dr Michael G. Wolfersberger for many stimulating discussions. Finally, I thank Mr Daniel J. Harvey and Ms Beth Gordesky-Gold for assistance with the artwork.