I. Ion Channels in Human THP-1 Monocytes

The THP-1 human monocytic leukemia cell line is a useful model of macrophage differentiation. Patch clamp methods were used to identify five types of ion channels in undifferentiated THP-1 monocytes. (i) Delayed rectifier K⁺ current, I _DR, was activated by depolarization to potentials positive to −50 mV, inactivated with a time constant of several hundred msec, and recovered from inactivation with a time constant ∼21 sec. I _DR was inhibited by 4-aminopyridine (4-AP), tetraethylammonium (TEA⁺), and potently by charybdotoxin (ChTX). (ii) Ca-activated K⁺ current (I _SK) dominated whole-cell currents in cells studied with 3–10 μm [Ca²⁺] _i . I _SK was at most weakly voltage-dependent, with reduced conductance at large positive potentials, and was inhibited by ChTX and weakly by TEA⁺, Cs⁺, and Ba²⁺, but not 4-AP or apamin. Block by Cs⁺ and Ba²⁺ was enhanced by hyperpolarization. (iii) Nonselective cation current, I _cat, appeared at voltages above +20 mV. Little time-dependence was observed, and a panel of channel blockers was without effect. (iv) Chloride current, I _Cl, was present early in experiments, but disappeared with time. (v) Voltage-activated H⁺ selective current is described in detail in a companion paper (DeCoursey & Cherny, 1996. J. Membrane Biol. 152:2). The ion channels in THP-1 cells are compared with channels described in other macrophage-related cells. Profound changes in ion channel expression that occur during differentiation of THP-1 cells are described in a companion paper (DeCoursey et al., 1996. J. Membrane Biol. 152:2). Received: 19 September 1995/Revised: 14 March 1996     

III. Ion Channel Expression in PMA-differentiated Human THP-1 Macrophages

Ion channel expression was studied in THP-1 human monocytic leukemia cells induced to differentiate into macrophage-like cells by exposure to the phorbol ester, phorbol 12-myristate 13-acetate (PMA). Inactivating delayed rectifier K⁺ currents, I _DR, present in almost all undifferentiated THP-1 monocytes, were absent from PMA-differentiated macrophages. Two K⁺ channels were observed in THP-1 cells only after differentiation into macrophages, an inwardly rectifying K⁺ channel (I _IR) and a Ca²⁺-activated maxi-K channel (I _BK). I _IR was a classical inward rectifier, conducting large inward currents negative to E _K and very small outward currents. I _IR was blocked in a voltage-dependent manner by Cs⁺, Na⁺, and Ba²⁺, block increasing with hyperpolarization. Block by Na⁺ and Ba²⁺ was time-dependent, whereas Cs⁺ block was too fast to resolve. Rb⁺ was sparingly permeant. In cell-attached patches with high [K⁺] in the pipette, the single I _IR channel conductance was ∼30 pS and no outward current could be detected. I _BK channels were observed in cell-attached or inside-out patches and in whole-cell configuration. In cell-attached patches the conductance was ∼200–250 pS and at potentials positive to ∼100 mV a negative slope conductance of the unitary current was observed, suggesting block by intracellular Na⁺. I _BK was activated at large positive potentials in cell-attached patches; in inside-out patches the voltage-activation relationship was shifted to more negative potentials by increased [Ca²⁺]. Macroscopic I _BK was blocked by external TEA⁺ with half block at 0.35 mm. THP-1 cells were found to contain mRNA for Kv1.3 and IRK1. Levels of mRNA coding for these K⁺ channels were studied by competitive PCR (polymerase chain reaction), and were found to change upon differentiation in the same direction as did channel expression: IRK1 mRNA increased at least 5-fold, and Kv1.3 mRNA decreased on average 7-fold. Possible functional correlates of the changes in ion channel expression during differentiation of THP-1 cells are discussed. Received: 19 September 1995/Revised: 14 March 1996     

Fluid Flow Modulates Calcium Entry and Activates Membrane Currents in Cultured Human Aortic Endothelial Cells

Human aortic endothelial cells (HAEC) respond to flow with Ca²⁺ entry, activation of a nonselective cation channel, activation of a chloride channel, and activation of a calcium-activated potassium channel. Conversely, human capillary endothelial cells were unaffected by similar flow rates. In HAEC the flow induced cytosolic free calcium increase ([Ca²⁺] _i ) and the ionic currents associated with it were sustained for up to 15 min after perfusion was stopped. In the absence of extracellular Ca²⁺, fluid flow was unable to evoke the [Ca²⁺] _i increase or the increase in membrane currents but the response could be restored by addition of extracellular Ca²⁺. Surprisingly, the flow response was inhibited in 50% of the cells by inhibitors of nitric oxide production. The results suggest that the sustained flow response in HAEC may be partially mediated by nitric oxide production and release. Received: 29 January 1999/Revised: 2 June 1999     

pH-Induced Proton Permeability Changes of Plasma Membrane Vesicles

In vivo studies with leaf cells of aquatic plant species such as Elodea nuttallii revealed the proton permeability and conductance of the plasma membrane to be strongly pH dependent. The question was posed if similar pH dependent permeability changes also occur in isolated plasma membrane vesicles. Here we report the use of acridine orange to quantify passive proton fluxes. Right-side out vesicles were exposed to pH jumps. From the decay of the applied ΔpH the proton fluxes and proton permeability coefficients (P_H+) were calculated. As in the intact Elodea plasma membrane, the proton permeability of the vesicle membrane is pH sensitive, an effect of internal pH as well as external pH on P_H+ was observed. Under near symmetric conditions, i.e., zero electrical potential and zero ΔpH, P_H+ increased from 65 × 10⁻⁸ at pH 8.5 to 10⁻¹ m/sec at pH 11 and the conductance from 13 × 10⁻⁶ to 30 × 10⁻⁴ S/m². At a constant pH _i of 8 and a pH _o going from 8.5 to 11, P_H+ increased more than tenfold from 2 to 26 × 10⁻⁶ m/sec. The calculated values of P_H+ were several orders of magnitude lower than those obtained from studies on intact leaves. Apparently, in plasma membrane purified vesicles the transport system responsible for the observed high proton permeability in vivo is either (partly) inactive or lost during the procedure of vesicle preparation. The residue proton permeability is in agreement with values found for liposome or planar lipid bilayer membranes, suggesting that it reflects an intrinsic permeability of the phospholipid bilayer to protons. Possible implications of these findings for transport studies on similar vesicle systems are discussed. Received: 5 April 1995/Revised: 28 March 1996     

Membrane Potassium Channels and Human Bladder Tumor Cells: II. Growth Properties

These experiments were done to determine the effect of glibenclamide and diazoxide on the growth of human bladder carcinoma (HTB-9) cells in vitro. Cell growth was assayed by cell counts, protein accumulation, and ³H-thymidine uptake. Glibenclamide added at 75 and 150 μm for 48 hr reduced cell proliferation. Dose-inhibition curves showed that glibenclamide added for 48 hr reduced cell growth at concentrations as low as 1 μm (IC₅₀= 73 μm) when growth was assayed in the absence of added serum. This μM-effect on cell growth was in agreement with the dose range in which glibenclamide decreased open probability of membrane K_ATP channels. Addition of glibenclamide for 48 hr also altered the distribution of cells within stages of the cell cycle as determined by flow cytometry using 10⁻⁵ m bromodeoxyuridine. Glibenclamide (100 μm) increased the percentage of cells in G₀/G₁ from 33.6% (vehicle control) to 38.3% (P < 0.05), and it reduced the percentage of cells in S phase from 38.3% to 30.6%. On the other hand, diazoxide, which opens membrane K_ATP channels in HTB-9 cells, stimulated growth measured by protein accumulation, but it did not increase the cell number. We conclude that the sulfonylurea receptor and the corresponding membrane K_ATP channel are involved in mechanisms controlling HTB-9 cell growth. However, K_ATP is not rate-limiting among the signaling mechanisms or molecular switches that regulate the cell cycle. Received: 12 June 1997/Revised: 21 October 1997     

Influence of Membrane Physical State on the Lysosomal Proton Permeability

Influence of membrane physical state on the proton permeability of isolated lysosomes was assessed by measuring the membrane potential with 3,3′-dipropylthiadicarbocyanine iodide and monitoring their proton leakage with p-nitrophenol. Changes in the membrane order were examined by the steady-state fluorescence anisotropy of 1,6-diphenyl-1,3,5-hexatriene. Both the membrane potential and proton leakage increased with fluidizing the lysosomal membranes by benzyl alcohol and decreased with rigidifying the membranes by cholesteryl hemisuccinate. The proton permeability increased to the maximum of 42% by the benzyl alcohol treatment and decreased to the minimum of 38.1% by the cholesteryl hemisuccinate treatment. Treating the lysosomes with protonophore CCCP increased the proton permeability by 58%. The effects of the membrane fluidization and rigidification can be reversed by rigidifying the fluidized membranes and fluidizing the rigidified membranes, respectively. The results indicate that the proton permeability of lysosomes increased and decreased with increasing and decreasing their membrane fluidity, respectively. Moreover, the lysosomal proton permeability did not alter further if the changes, either an increase or a decrease, in the fluidity exceeded some amount. The results suggest that the proton permeability of lysosomes can be modulated finitely by the alterations in their membrane physical state. Received: 27 September 1999 / Revised: 27 December 1999     

Effects of NEM on Voltage-activated Chloride Conductance in Toad Skin

The regulation of the voltage-activated chloride current conductance (G _Cl ) in toad skin was investigated by the use of the SH reagents N-ethylmaleimide (NEM) and p-chloro-mercuricbenzenesulfonic acid PCMBS. This anion pathway is controlled by a voltage-sensitive gating regulator. Mucosal application of NEM decreased the voltage-activation in a time and concentration dependent manner, half-maximal inhibition being exerted at a concentration of 30 μm within 20 min. At concentrations higher than 100 μm, the voltage-activated G _Cl was near-completely and irreversibly inhibited in less than 10 min. Resting, deactivated conductance was essentially unaffected. NEM had no effect on active sodium transport (measured as I _sc ) under conditions, which fully dissipated the voltage-activated G _Cl . After complete inhibition of the voltage-activated G _Cl with NEM, chloride conductance could still be stimulated by CPT-cAMP as in control tissues. Under these conditions, NEM at concentrations above 1 mm decreased G _Cl reversibly. Mucosal application of PCMBS at 500 μm inhibited the activated conductance by 35%, which was slightly reversible. Inhibition of voltage-activated G _Cl , which was observed after mucosal addition of the membrane-impermeable NEM analogue, eosin-5-maleimide, was completely reversible after washout. This suggests that the binding site for the maleimide is not accessible from the external face of the apical membrane. Brief application of NEM at lower concentrations (1–3 min, ≤100 μm) led to partial inhibition of G _Cl , followed by occasionally complete recovery upon washout of NEM. Recovery of voltage-activated G _Cl was progressively attenuated and eventually disappeared after subsequent brief applications of NEM. This could reflect recruitment of permeation/control sites from a finite pool. The data are discussed in the frame of a working model for the voltage-activated Cl⁻-pathway, that contains two principle components, i.e., an anion-selective permeation path which is controlled by regulatory protein(s). Received: 18 December 1996/Revised: 28 April 1997     

Two Different Conductances Contribute to the Anion Currents in Coffea arabica Protoplasts

The anion conductance of the plasma membrane of Coffea arabica protoplasts was isolated and characterized using the whole-cell patch clamp technique. Voltage pulse protocols revealed two components: a voltage-gated conductance (G _s ) and a voltage-independent one (G _l ). G _s is activated upon depolarization (e-fold activation every +36 mV) with time constants of 1 sec and 5 sec at all potentials. G _l and G _s also differ by their kinetic and biophysical properties. In bi-ionic conditions the current associated with G _s shows strong outward rectification and its permeability sequence is F⁻ > NO₃ ⁻ > Cl⁻. In the same conditions the current associated with G _l does not rectify and its permeability sequence is F⁻≫ NO₃ ⁻= Cl⁻. Furthermore, at potentials over +50 mV G _s , but not G _l , increases with a time constant of several minutes. Finally the gating of G _s is affected by stretch of the membrane, which leads to an increased activation and a reduced voltage sensitivity. Anion conductances similar to the ones described here have been found in many plant preparations but G _l -type components have been generally interpreted as the background activation of the slow voltage-gated channels (corresponding to G _s ). We show that in coffee protoplasts G _l and G _s are kinetically and biophysically distinct, suggesting that they correspond to two different molecular entities. Received: 25 November 1996/Revised: 9 April 1997     

Effect of Adenosine on Na+ and Cl− Currents in A6 Monolayers. Receptor Localization and Messenger Involvement

The effect of adenosine regulation on sodium and chloride transport was examined in cultured A₆ renal epithelial cells. Adenosine and its analogue N⁶-cyclopentyladenosine (CPA) had different effects on short-circuit current (I _sc) depending on the side of addition. Basolateral CPA addition induced an approximately threefold increase of the I _sc that reached a maximum effect 20 min after addition and was completely inhibited by preincubation with either an A₂ selective antagonist, CSC, or the sodium channel blocker, amiloride. Apical CPA addition induced a biphasic I _sc response characterized by a rapid fourfold transient increase over its baseline followed by a decline and a plateau phase that were amiloride insensitive. The A₁ adenosine antagonist, CPX, completely prevented this response. This I _sc response to apical CPA was also strongly reduced in Cl⁻-free media and was significantly inhibited either by basolateral bumetanide or apical DPC preincubation. Only basolateral CPA addition was able to induce an increase in cAMP level. CPA, added to cells in suspension, caused a rapid rise in [Ca²⁺] _i that was antagonized by CPX, not affected by CSC and prevented by thapsigargin preincubation. These data suggest that basolateral CPA regulates active sodium transport via A₂ adenosine receptors stimulating adenylate cyclase while apical CPA regulates Cl⁻ secretion via A₁ receptor-mediated changes in [Ca²⁺] _i .     

Single-Channel Activities of the Human Epithelial Ca2+ Transport Proteins CaT1 and CaT2

The human epithelial channels, CaT1 and CaT2, were expressed in oocytes, and their single-channel characteristics were compared. In the presence of Na⁺ and K⁺ as charge carriers in the pipette solutions, channel activities were observed only when the the extracellular sides of the patches were exposed to nominally Ca²⁺- and Mg²⁺-free solutions. In patches of both CaT1- and CaT2-expressing oocytes, multiple channel openings were observed, but the current levels were higher in CaT2-expressing oocytes, particularly at more negative voltages. With K⁺ as a charge carrier in patches of CaT1-expressing oocytes, the channel activity was low at −10 to −60 mV, but increased dramatically at more negative potentials. This voltage dependence was observed in the presence of both Na⁺ and K⁺. The channel activity with Na⁺, however, was higher at all potentials. Differences between the voltage dependencies for the two cations were also observed in CaT2-expressing oocytes, but the channel activities were higher than those in CaT1-expressing oocytes, particularly in the presence of Na⁺. We also found that low concentrations of extracellular Mg²⁺ (5–50 μm) elicited a strong inhibitory action on the CaT channels. Activation of the CaT1 and CaT2 channels by hyperpolarization and other factors may promote increased Ca²⁺ entry that participates in stimulation of intestinal absorption and renal reabsorption and/or other Ca²⁺ transport mechanisms in epithelial cells. Received: 8 March 2001/Revised: 24 July 2001     

Mechanically Activated Currents in Chick Heart Cells

As predicted from stretch-induced changes of rate and rhythm in the heart, acutely isolated embryonic chick heart cells exhibit whole-cell mechanosensitive currents. These currents were evoked by pressing on cells with a fire polished micropipette and measured through a perforated patch using a second pipette. The currents were carried by Na⁺ and K⁺ but not Cl⁻, and were independent of external Ca²⁺. The currents had linear I/V curves reversing at −16 mV and were completely blocked by Gd³⁺≥ 30 μm and Grammostola spatulata venom at a dilution of 1:1000. Approximately 20% of cells showed time dependent inactivation. In contrast to direct mechanical stimulation, hypotonic volume stress produced an increase in conductance for anions rather than cations—the two stimuli are not equivalent. The cells had two types of stretch-activated ion channels (SACs): a 21 pS nonspecific cation-selective reversing at −2 mV and a 90 pS K⁺ selective reversing at −70 mV in normal saline. The activity of SACs was strongly correlated with the presence of whole-cell currents. Both the whole-cell currents and SACs were blocked by Gd³⁺ and by Grammostola spatulata spider venom. Mechanical stimulation of spontaneously active cells increased the beating rate and this effect was blocked by Gd³⁺. We conclude that physiologically active mechanosensitive currents arise from stretch activated ion channels. Received: 8 April 1996/Revised: 8 August 1996     

Histamine Increases [Ca2+] in and Activates Ca-K and Nonselective Cation Currents in Cultured Human Capillary Endothelial Cells

We characterized the effects of histamine on intracellular Ca²⁺ and activation of ionic currents in human capillary endothelial cells. Histamine produced both a transient and sustained increase in intracellular Ca²⁺. The transient response was mediated largely through intracellular Ca²⁺ release and the sustained response was due to extracellular Ca²⁺ entry. The increase in intracellular Ca²⁺ by histamine was not affected by the H2 blocker cimetidine. But was entirely blocked by the H1 antagonist diphenhydramine showing that the histamine response in these cells is mediated through the H1 receptor. A transient ionic current is coactivated with the histamine-induced increase in intracellular Ca²⁺ and this current has several properties of a nonselective, Ca²⁺ permeable, cation channel (NSC). The magnitude of the NSC current does not strictly correlate with intracellular Ca²⁺ levels. A Ca²⁺-activated K⁺ current (BKCA) is activated by the increase in intracellular Ca²⁺ and this current is blocked by the selective BKCA blocker iberiotoxin. Received: 16 June 1999/Revised: 22 September 1999     

Full-Length L1 Elements Have Arisen Recently in the Same 1-kb Region of the Gorilla and Human Genomes

New copies of the mammalian retrotransposon L1 arise in the germline at an undetermined rate. Each new L1 copy appears at a specific evolutionary time point that can be estimated by phylogenetic analysis. In humans, the active L1 sequence L1.2 resides at the genomic locus LRE1. Here we analyzed the region surrounding the LRE1 locus in humans and gorillas to determine the evolutionary history of the region and to estimate the age of L1.2. We found that the region was composed of an ancient L1, L1Hs-Lrg, which was significantly divergent from all other L1 sequences available in the databases. We also determined that L1.2 was absent from the gorilla genome and arose in humans after the divergence of gorilla and human lineages. In the gorilla LRE1 region, we discovered a different full-length L1 element, L1Gg-1, which was allelic and present at a high gene frequency in gorillas but absent from other primates. We determined the nucleotide sequence of L1Gg-1 and found that it was 98% identical to L1.2, suggesting a close relationship between active L1s in gorillas and humans. Received: 28 December 1997 / Accepted: 20 March 1998     

The Proton Permeability of Liposomes Made from Mitochondrial Inner Membrane Phospholipids: Comparison with Isolated Mitochondria

Unilamellar liposomes with native phospholipid fatty acid composition were prepared from rat liver mitochondrial inner membrane phospholipids by extrusion in medium containing 50 mm potassium. They were diluted into low potassium medium to establish a transmembrane potassium gradient. A known membrane potential was imposed by addition of valinomycin, and proton flux into liposomes was measured. Valinomycin in the range 10 pm–1nm was sufficient to fully establish membrane potential. Valinomycin concentrations above 3 nm catalyzed additional proton flux and were avoided. At 300 pm valinomycin, proton flux depended nonlinearly on membrane potential. At 160 mV membrane potential the flux was 30 nmol H⁺/min/mg phospholipid—approximately 5% of the proton leak flux under comparable conditions in isolated mitochondria, indicating that leak pathways through bulk phospholipid bilayer account for only a small proportion of total mitochondrial proton leak. Received: 5 August 1996/Revised: 1 October 1996     

Proton Inhibition of Sodium Channels: Mechanism of Gating Shifts and Reduced Conductance

Extracellular acidosis affects both permeation and gating of the expressed rat skeletal muscle Na⁺ channel (μ1). Reduction of the extracellular pH produced a progressive decrease in the maximal whole-cell conductance and a depolarizing shift in the whole-cell current-voltage relationship. A smaller depolarizing shift in the steady-state inactivation curve was observed. The pK of the reduction of maximal conductance was 6.1 over the pH range studied. An upper limit estimate of the pK of the shift of the half-activation voltage was 6.1. The relative reduction in the maximal whole-cell conductance did not change with higher [Na⁺] _o . The conductance of single fenvalerate-modified Na⁺ channels was reduced by extracellular protons. Although the single-channel conductance increased with higher [Na⁺] _o , the maximal conductances at pH 7.6, 7.0 and 6.0 did not converge at [Na⁺] _o up to 280 mm, inconsistent with a simple electrostatic effect. A model incorporating both Na⁺ and H⁺ binding in the pore and cation binding to a Gouy-Chapman surface charge provided a robust fit to the single-channel conductance data with an estimated surface charge density of 1e⁻/439?². Neither surface charge nor proton block alone suffices to explain the effects of extracellular acidosis on Na⁺ channel permeation; both effects play major roles in mediating the response to extracellular pH. Received: 14 May 1996/Revised: 19 September 1996     

Induction of Ca2+-Activated K+ Current and Transient Outward Currents in Human Capillary Endothelial Cells

Human capillary endothelial cells (HCEC) in normal media contain noninactivating outwardly rectifying chloride currents, TEA-sensitive delayed rectifier K⁺ currents and an inward rectifier K⁺ current. Two additional ionic currents are induced in HCEC when the media are allowed to become conditioned: A Ca²⁺-activated K⁺ current (BKCA) that is sensitive to iberiotoxin is induced in 23.5% of the cells, a transient 4-AP-sensitive K⁺ current (A current) is induced in 24.7% of the cells, and in 22.3% of the cells both the transient and BKCA currents are coinduced. The EC₅₀ for Ca²⁺ activation of the BKCA current in HCEC from conditioned media is 213 nM. RNA message for BKCA (hSlo clone) is undetecable after PCR amplification in control cells but is seen in those from conditioned cells. The induction of BKCA current is not blocked by conditioning with inhibitors of nitric oxide synthase, cyclo-oxgenase or lypo-oxygenase pathways. Apparently the characteristics of human endothelial cells are highly malleable and can be easily modified by their local environment. Received: 21 May 1998/Revised: 23 September 1998     

A Long-Term Blockade of L-Type Calcium Currents Upregulates the Number of Ca2+ Channels in Skeletal Muscle

The effects of a long-term blockade of L-type Ca²⁺ channels on membrane currents and on the number of dihydropyridine binding sites were investigated in skeletal muscle fibers. Ca²⁺ currents (I _Ca) and intramembrane charge movement were monitored using a voltage-clamp technique. The peak amplitude of I _Ca increased by more than 40% in fibers that were previously incubated for 24 hr in solutions containing the organic Ca²⁺ channel blocker nifedipine or in Ca²⁺-free conditions. A similar incubation period with Cd²⁺, an inorganic blocker, produced a moderate increase of 20% in peak I _Ca. The maximum mobilized charge (Q _max) increased by 50% in fibers preincubated in Ca²⁺-free solutions or in the presence of Cd²⁺. Microsomal preparations from frog skeletal muscle were isolated by differential centrifugation. Preincubation with Cd²⁺ prior to the isolation of the microsomal fraction doubled the number of ³H-PN200-110 binding sites and produced a similar increase in the values of the dissociation constant. The increase in the number of binding sites is consistent with the increase in the peak amplitude of I _Ca as well as with the increase in Q _max. Received: 31 August 1998/Revised: 7 December 1998     

Relationships Between Na+/Glucose Cotransporter (SGLT1) Currents and Fluxes

The relationships between currents generated by the rabbit Na⁺/glucose cotransporter (SGLT1) and the fluxes of Na⁺ and sugar were investigated using Xenopus laevis oocytes expressing SGLT1. In individual voltage-clamped oocytes we measured: (i) the current evoked by 10 mmαMG and the ²²Na⁺ uptake at 10 mm Na⁺; (ii) the currents evoked by 50 to 500 μm [¹⁴C]αMG and the [¹⁴C]αMG uptakes at 100 mm Na⁺; and (iii) phlorizin-sensitive leak currents in the absence of sugar and ²²Na⁺ uptakes at 10 mm Na⁺. We demonstrate that the SGLT1 leak currents are Na⁺ currents, and that the sugar-evoked currents are directly proportional to both αMG and Na⁺ uptakes. The Na⁺/αMG coupling coefficients were estimated to be 1.6 at −70 mV and 1.9 at −110 mV. This suggests that the rabbit SGLT1 Na⁺/αMG stoichiometry for sugar uptake is 2 under fully saturating, zero-trans conditions. Coupling coefficients of less than 2 are expected under nonsaturating conditions due to uncoupled Na⁺ fluxes (slippage). The similarity between the Na⁺ Hill coefficients and the coupling coefficients suggests strong cooperativity between the two Na⁺ binding sites. Received: 6 October 1997/Revised: 5 December 1997     

Evolution of the Cytochrome c Oxidase Proton Pump

The superfamily of quinol and cytochrome c terminal oxidase complexes is related by a homologous subunit containing six positionally conserved histidines that ligate a low-spin heme and a heme–copper dioxygen activating and reduction center. On the basis of the structural similarities of these enzymes, it has been postulated that all members of this superfamily catalyze proton translocation by similar mechanisms and that the Cu_A center found in most cytochrome c oxidase complexes serves merely as an electron conduit shuttling electrons from ferrocytochrome c into the hydrophobic core of the enzyme. The recent characterization of cytochrome c oxidase complexes and structurally similar cytochrome c:nitric oxide oxidoreductase complexes without Cu_A centers has strengthened this view. However, recent experimental evidence has shown that there are two ubiquinone(ol) binding sites on the Escherichia coli cytochrome bo ₃ complex in dynamic equilibrium with the ubiquinone(ol) pool, thereby strengthening the argument for a Q(H₂)-loop mechanism of proton translocation [Musser SM et al. (1997) Biochemistry 36:894–902]. In addition, a number of reports suggest that a Q(H₂)-loop or another alternate proton translocation mechanism distinct from the mitochondrial aa ₃ -type proton pump functions in Sulfolobus acidocaldarius terminal oxidase complexes. The possibility that a primitive quinol oxidase complex evolved to yield two separate complexes, the cytochrome bc ₁ and cytochrome c oxidase complexes, is explored here. This idea is the basis for an evolutionary tree constructed using the notion that respiratory complexity and efficiency progressively increased throughout the evolutionary process. The analysis suggests that oxygenic respiration is quite an old process and, in fact, predates nitrogenic respiration as well as reaction-center photosynthesis. Received: 11 June 1997 / Accepted: 30 October 1997