DIVERSIDAD (H′) Y EQUITATIVIDAD (J′) DE LA DIETAS

For the H⫹ channel to sense local accumulation of protons (248), it must be physically near the source of protons. The analogous problem of how far Ca2⫹ can diffuse from the point source of a Ca2⫹channel into bulk cytoplasm has been addressed (964). The length constant ␭(roughly the average distance the free ion diffuses be- fore it is consumed by buffer) is given approximately by

␭⬇

冑

Bfreek

(7)

whereDHis the diffusion constant for H⫹ and is 1.08⫻

10⫺4cm2/s at 37°C (845), Bfree is the free buffer concen-

tration, which corresponds roughly with the buffering capacity of phagocyte cytoplasm, which has been esti- mated to be 15.6 –50 mM at physiological pHi(324, 940).

Stern (964) and others (950) have discussed the limita- tions of this calculation. The bimolecular rate constant of protonation of buffer (k) is assumed to be diffusion lim- ited at 1– 4 ⫻ 1010 M⫺1 䡠 s⫺1 (80, 287, 290, 768). Given these rough assumptions, ␭ is 2.3– 8.3 nm. Neglecting buffer diffusion and background proton concentration, the concentration of protons (cH) at distance r from a

source such as NADPH oxidase is given approximately by (E. Rı´os, personal communication)

cH⬇

H⫹generation rate 4␲DHr

e⫺r/␭ ₍₈₎ The exponential term incorporates the attenuation by buffer of the cH that would have resulted from diffusion

from the source alone.

Distances can be estimated from the previous esti- mates of 125–200 H⫹ channels/␮m2 and 2,100 –3,300 ac- tive NADPH oxidase complexes/␮m2of eosinophil mem- brane (see sect.VIH6). The following geometric analysis is by Ricardo Murphy. With the assumption of a uniform distribution in a square grid pattern, each H⫹ channel would be 70 –90 nm from its nearest neighbor, and each active NADPH oxidase complex would be 17–22 nm from

its neighbors. If the membrane area associated with each H⫹ channel (A ⫽ 5– 8 ⫻ 103_nm2_{) is approximated as a}

circle of radiusR⫽ 公(A/␲)⫽40 –50 nm, then on average, no NADPH oxidase molecule could be farther than this distance from a H⫹ channel. If H⫹channels and NADPH oxidase complexes were distributed uniformly, the aver- age distance of an NADPH oxidase molecule from the nearest H⫹channel, rmean, would be

rmean⫽

冕

冉

冊

冕

or 27–33 nm. The minimum estimated distance is thus 3␭, which fromEquation 8means that a H⫹ channel would sense⬍5% of the increased [H⫹] near the NADPH oxidase complex (assuming the effective radius of the NADPH oxidase complex isⱕ␭). In other phagocytes with a lower density of both H⫹channels and NADPH oxidase compo- nents, the distances would be greater. Thus, in the ab- sence of some structural link, such as colocalization of H⫹ channels and NADPH oxidase molecules in “lipid rafts” (480), it is unlikely that H⫹ channels would uni- formly sense highly localized changes in pH due to the activity of a single NADPH oxidase complex. These cal- culations ignore any enhancement that would result by processes that facilitate the movement of protons in the plane of the membrane surface (see sect.IID).

On the other hand, there is a large decrease in global pHi during the respiratory burst when Na⫹/H⫹ antiport

and H⫹channels are inhibited (430, 431, 738, 977), which demonstrates that protons are generated at a rate suffi- cient to overload the entire cytoplasmic buffering capac- ity. It is likely that during the respiratory burst, protons are concentrated near the membrane and that there is a proton gradient that dissipates toward the center of the cell. Direct evidence that this can occur is seen in perme- abilized-patch studies. Although the applied NH4⫹gradient

provides excellent control of pHiin the absence of a load

(248, 387), when the NADPH oxidase is activated under these conditions the observed Vrevof H⫹ currents shifts

⫺3.7 mV in neutrophils (248) and⫺5.8 mV or⫺5.3 mV in eosinophils stimulated with PMA (246) or AA (165), re- spectively. Eosinophils are the same size as neutrophils but have a more active NADPH oxidase. NADPH oxidase activity thus lowers pHiby⬃0.1 unit in spite of the NH4⫹

gradient (and in spite of application of occasional depo- larizing pulses to observe currents).

In summary, voltage-gated proton channels are likely too far from NADPH oxidase complexes to sense acute local pH changes, but they may respond to more global pHi changes that accumulate during sustained NADPH

oxidase activity.

VII. SUMMARY AND CONCLUSIONS

Protons are unique among cations in their tiny size, low free but enormous total concentration, reactivity with other molecules, and Grotthuss conduction mechanism. Proton channels have a unique conduction mechanism, the HBC, which in nonselective channels is a simple water wire, but in highly proton-selective channels includes at least one titratable amino acid residue. These protonation sites at once preclude the conduction of other cations, thus acting as “selectivity filters,” and provide a mecha- nism that mediates the interaction between pH and the molecular conformation and function of the channel. The key to proton selectivity is that the proton is conducted through a HBC as H⫹ rather than H3O⫹. Many different

molecules have proton channels that are of central im- portance to their function. Most of these channels are HBCs comprising water wires interrupted at crucial points by titratable amino acid residues. In several cases, proton conduction is enhanced by the presence of titrat- able groups at the entrance to the channel. Proton entry into (or exit from) a channel that occurs by direct Eigen- type proton transfer between buffer and a titratable group on the channel can increase the rate of proton transport far beyond that obtained with simple diffusion. Heavy metal cations often bind competitively with protons to titratable sites on proton channels, where they produce a variety of effects on molecular function.

Unique properties of voltage-gated proton channels, compared with other ion channels that are water-filled pores, include extraordinarily high selectivity, tiny unitary conductance, strong temperature and deuterium isotope effects on both conductance and gating kinetics, and insensitivity to block by steric occlusion. Many of these properties are manifestations of the HBC conduction mechanism. The gating of voltage-gated proton channels is regulated tightly by pH and voltage, with the result that under normal conditions, the channels open only when the electrochemical gradient is outward. The general function of these channels is therefore acid extrusion from cells. Their responsiveness to the voltage across the membrane in which they are located, with modulation by the local pH on both sides of the membrane, means that they are activated automatically when the need arises. Voltage-gated proton channels are expressed in many cells and appear to comprise at least four isoforms. The evidence for specific function is strongest in phagocytes, in which H⫹ channels extrude protons during the respi- ratory burst to compensate for electron extrusion by NADPH oxidase. Activation of H⫹ current during the respiratory burst is due to a combination of depolariza- tion, pH changes, and profound modulation of the prop- erties of the H⫹ channels. The functions of voltage-gated proton channels and NADPH oxidase are intimately inter- connected, but the bulk of evidence indicates that the

channel is a molecular entity distinct from any known NADPH oxidase component. Although much progress has been made, the most exciting discoveries lie in the future.

I thank Vladimir V. Cherny for nearly a decade of invalu- able collaboration and Tatiana Iastrebova for prodigious work organizing the references. For generously providing comments, suggestions, answers, preprints, and permission to use figures and unpublished data, I am indebted to Noam Agmon, Jennifer Bankers-Fulbright, Michael E. Barish, Howard C. Berg, Allan Bretag, Peter Brzezinski, Vladimir V. Cherny, Anatoly Cherny- shev, David W. Christianson, Andrew R. Cross, Sam Cukierman, Beverly L. Davidson, Audrey G. DeCoursey, Wesley F. DeCour- sey, J. Paul Devlin, Mary C. Dinauer, Claudia Eder, Shelagh Ferguson-Miller, Horst Fischer, Robert H. Fillingame, Miklo´s Geiszt, Mark E. Girvin, Sergio Grinstein, Menachem (Hemi) Gutman, Sharon Hammes-Schiffer, Ronald Haskell, Robert A. Lamb, Janos K. Lanyi, Eric V. Leaver, Tom L. Leto, Paul A. Loach, Vladislav S. Markin, Robert Meech, Denise A. Mills, Deri Morgan, Ricardo Murphy, John F. Nagle, William M. Nauseef, Mel Y. Okamura, Mark L. Paddock, Pamela A. Pappone, Law- rence H. Pinto, Eduardo Rı´os, Dirk Roos, Tom Schilling, Mark F. Schumaker, David N. Silverman, Clifford L. Slayman, Erik R. Swenson, John Tang, Justin Teissie, Roger C. Thomas, Mårten Wikstro¨m, Dixon J. Woodbury, and Colin A. Wraight. This re- view was truly a collective effort; many of these individuals patiently critiqued or in some cases essentially rewrote sections of this review.

This work was supported by National Heart, Lung, and Blood Institute Grants HL-52671 and HL-61437.

Address for reprint requests and other correspondence: T. DeCoursey, Dept. of Molecular Biophysics and Physiology, Rush Presbyterian St. Luke’s Medical Center, 1750 West Harrison, Chicago, IL 60612 (E-mail: [email protected]).