Control of Intracellular pH

Absorbance

Molecules with an absorbance, fluorescence excitation, and/or fluorescence emission spectrum sensitive to pH may be convenient probes for measuring pH _i . When exposed to light, pH-sensitive dye molecules may absorb some of the light as electrons make the transition to a higher-energy state. From the intensities of incident light (I _o ) and the light transmitted through the solution (I), we can compute the absorbance (A):

According to the Beer-Lambert law, A at a particular wavelength (A _λ ) is proportional to both the length of the light’s path through the solution ( l ) and the concentration of the dye (C):

where the proportionality constant ε _λ,pH is the dye’s extinction coefficient at a particular wavelength and pH. The wavelength dependence of ε _λ,pH defines the shape of the absorbance spectrum, and the pH dependence of ε _λ,pH defines how this shape is affected by changes in pH. However, A _λ depends not only upon pH, but also upon l and C. Although l and C are extremely difficult to ascertain in a live cell, we can obtain the absorbance data at two wavelengths (λ ₁ and λ ₂ ), and compute the absorbance ratio (A _λ1 /A _λ2 ):

Thus, because the l and C terms cancel out, the absorbance ratio in Equation 52.6 , depends only on pH. By choosing two wavelengths such that ε _λ1,pH /ε _λ2,pH varies considerably with pH, one can obtain sensitive pH measurements. Experimenters using fluorescein derivatives typically use the peak absorbance wavelength (~510 nm), and the isosbestic wavelength (~440 nm), where ε is insensitive to pH changes. Monitoring dye absorbance at the isosbestic wavelength is attractive because one can determine the extent of dye loss during the experiment, assuming l is constant.

An advantage of absorbance for measuring pH _i is that it tends to be extremely stable and sensitive. On the other hand, because absorbance is proportional to l and C, a relatively high intracellular dye concentration is required, even for thick preparations (e.g., renal tubules).

Fluorescence

After absorbing a photon, most molecules return to the ground state by gradually losing energy through a series of random collisions with other molecules (see ref. ). Some dyes, however, can lose a quantum of energy from an excited singlet state by emitting a photon (i.e., fluorescing). The intensity of emitted fluorescent light (I _emit ) can be measured with a photomultiplier tube (e.g., see ref. ) or an intensified [CCD] television camera, which provides imaging data (e.g., see ref. ). At most wavelengths, I _emit is sensitive to pH, but at all wavelengths, I _emit is sensitive to dye concentration as well as other parameters (e.g., position of cell in the incident light beam). Therefore, one usually employs a ratio technique to generate a parameter more uniquely related to pH. With the fluorescence-excitation ratio approach, commonly used with fluorescein dyes such as BCECF, one alternately excites at two wavelengths while monitoring I _emit at one wavelength. With the fluorescence-emission ratio approach, commonly used with some rhodamine dyes such as SNARF-1, one excites at one wavelength, while monitoring I _emit simultaneously at two wavelengths. By analogy with the absorbance ratio approach, one chooses three wavelengths to optimize the pH sensitivity of the ratio. A strength of the fluorescence-emission ratio approach is that the simultaneous capture of I _emit at two wavelengths improves temporal resolution (see below).

Fluorescence measurements offer the advantage of being extremely sensitive. Thus, it is possible to quantitate the fluorescence from small amounts of dye, even in a microdomain within a single cell. In addition, one can use fluorescence with two-dimensional imaging, confocal microscopy, and multi-photon microscopy. Because fluorescence is more sensitive than absorbance to the environment of the dye molecule, fluorescence measurements are in principle more prone to artifact.

Bleaching and Photodynamic Damage

Excessive illumination can photolyse (i.e., bleach) dye molecules, causing a progressive decrease in the concentration of native dye. Among the photolysis products may be free radicals that react with cellular components and injure the cell (“photodynamic damage”). Excessive illumination can also cause a time-dependent shift in the apparent intracellular calibration curve of the dye, presumably due to bleaching-induced generation of dye-related products with spectral characteristics slightly different than those of the parent compound.

Bleaching can be particularly problematic during experiments on single cells, where the number of dye molecules and thus the number of emitted photons is low. The problem is that the limiting signal-to-noise ratio (S/N) is proportional to the square root of the number of photons measured. One can increase the number of emitted photons by increasing the intensity of the excitation light source, but at the expense of exacerbating the bleaching and photodynamic damage. The damage can be minimized either by illuminating continuously with low-intensity light for longer, or by limiting the duty cycle of high-intensity exciting light. Thus, one must sometimes trade off pH _i resolution (i.e., S/N) against time resolution. In imaging experiments, where a cell may be represented by thousands of picture elements (pixels), the photon emission rate per pixel is exceedingly low. Here, one must trade off pH _i resolution, time resolution, and spatial resolution (i.e., the number of pixels that must be grouped to compute pH _i values).

More recent advances in fluorescence microscopy have helped to minimize the predicament of good pH _i vs. spatial/temporal resolution. For example, two-photon excitation laser scanning microscopy is a relatively new imaging technique that has allowed investigators to reduce photobleaching of ion-sensitive fluorescent dyes, even while acquiring high spatial and temporal resolution recordings. Multiphoton imaging technology is based on the general quantum principle that a molecule can exhibit fluorescence after absorbing two photons simultaneously when excited by high-intensity light of twice the wavelength necessary for single-photon absorption. Excitation at the higher wavelengths (and thus lower energies) reduces overall photobleaching of the dye. Two-photon microscopy has been particularly useful in measuring intracellular Ca ²⁺ transients with excellent spatial resolution in tissues and cellular microdomains (see ). The technique has also been used with pH-sensitive dyes. For example, intracellular pH has been measured with BCECF in microdomains of the epidermis, and with SNARF-4 in villous enterocytes in vivo .

Calibration

Choice of Dyes

Differential Dye Loading

One can exploit differential dye loading in preparations to target dyes to specific cells or locations. For example, intercalated cells in the cortical collecting tubule of the kidney incorporate BCECF-AM much more rapidly than the neighboring principal cells. In a somewhat different application, Chu et al. found that mouse colonic crypt cells exclude SNARF, which can then be used to measure extracellular pH changes in the intact epithelium. Also, a probenecid-sensitive organic-anion transporter can extrude BCECF from some cells including thyroid cells, and epithelial kidney and intestinal cells. In an elegant study, Harris et al. used BCECF to measure the pH of the lateral intracellular space (LIS) of MDCK cell monolayers by first loading the cells with BCECF, and subsequently allowing organic transporters to move the dye into the LIS.

BCECF generally stains the cytoplasm rather uniformly, although in some cells the nuclear region is more intense than peripheral areas. Working on Ehrlich ascites tumor cells, Thomas et al. found that 6-carboxyfluorescein is confined to the cytoplasm, but fluorescein is distributed in both the mitochondria and cytoplasm. Furthermore, Slayman et al. found that BCECF can accumulate in vacuoles of the fungus Neurospora when the cells are exposed to BCECF-AM.

One can evaluate a dye’s intracellular compartmentalization by monitoring the fluorescence loss elicited by selectively permeabilizing different compartments with detergents. For example, most of the BCECF loaded into rat hippocampal CA1 neurons appears to be in the cytoplasm because 0.01% saponin reduces the fluorescence signal by ~96%. Both absorbance- and fluorescence-derived estimates of pH _i can be influenced by dye in compartments other than the cytoplasm, depending upon each compartment's volume, dye concentration and pK, and pH.

Flux of the Neutral Weak Acid

When a cell is exposed to a neutral weak acid, HA enters and dissociates into A ⁻ and H ⁺ , thereby causing pH _i to fall:

This movement of the neutral weak acid (or a neutral weak base such as NH ₃ , discussed below) has been termed non-ionic diffusion. This process continues until [HA] _i equals [HA] _o . Subsequently, there should be no further change in pH _i , provided there is no flux of A ⁻ and no change in the rates of acid–base transporter activity. Although CO ₂ is often regarded as a weak acid, it is not an acid at all. Only after reacting with H ₂ O does CO ₂ yield the true weak acid, H ₂ CO ₃ , which then dissociates (as described above for the idealized weak acid HA):

These two reactions can be combined into a single one with an overall apparent equilibrium constant ( $K_a^{\prime }={[\text{HCO}}_3^{-}]$ [H ⁺ ]/[CO ₂ ]). The enzyme carbonic anhydrase (CA) greatly accelerates the formation of ${\mathrm{HCO}}_3^{-}$ from CO ₂ by catalyzing the reaction:

which generates H ⁺ by virtue of consuming OH ⁻ . The mechanism by which CA catalyzes this reaction is examined in more detail by Liljas et al. In the absence of CA, the reaction can also occur, particularly under alkaline conditions (i.e., pH >8).

In the example of Figure 52.3 , a cell is successively exposed to solutions equilibrated with 1, 2 and 5% CO ₂ (constant [ ${\mathrm{HCO}}_3^{-}$ ] _o =10 mM). Each time, CO ₂ produces a rapid and sustained fall of pH _i . The period during the CO ₂ exposure in which pH _i is relatively stable is termed the plateau phase . The magnitude of the CO ₂ -induced acidification is inversely related to the intracellular buffering power (β; see below). The magnitude of the pH _i decrease also increases with [CO ₂ ] _o . The degree of dissociation is governed by the relationship between pH _i and ${\mathrm{pK}}_a^{\prime }$ , which, in logarithmic form, is the familiar Henderson-Hasselbalch equation:

where ${\mathrm{pK}}_a^{\prime }$ is ~6.1 at 37°C. Therefore, if pH _i is 7.1, 10 molecules of incoming CO ₂ dissociate into H ⁺ plus ${\mathrm{HCO}}_3^{-}$ for each molecule of CO ₂ that remains CO ₂ . If the initial pH _i is only 6.1, then this ratio falls to 1:1, and fewer incoming CO ₂ molecules dissociate. There are two practical consequences of this rule. First, the lower the initial pH _i , the smaller the magnitude of acidification elicited by the subsequent exposure to the CO ₂ . Second, as one raises [CO ₂ ] _o by fixed increments, the magnitude of the acidification is not proportional to the successive [CO ₂ ] _o increment. Thus, in Figure 52.3 , the ΔpH _i produced by 2% CO ₂ is less than twice as large as that produced by 1% CO ₂ , and that produced by 5% CO ₂ is substantially less than five times that produced by 1% CO ₂ .

Although investigators have long believed that all gases penetrate all membranes simply by dissolving in the membrane lipid, Waisbren et al. demonstrated the first membrane with negligible permeability to a dissolved gas (i.e., CO ₂ and NH ₃ ). Moreover, Nakhoul et al. and Cooper and Boron identified the first gas channel—AQP1, which is permeable to CO ₂ . Later work showed that AQP1 is also permeable to NH ₃ and that rhesus (Rh) proteins can be permeable to both CO ₂ and NH ₃ . In human erythrocytes, AQP1 and the Rh complex, together, are responsible for ~90% of the CO ₂ permeability. Thus, at least two families of proteins can function as gas channels. Furthermore, just as ion channels display ion selectivity, the AQPs and Rh proteins display gas selectivity. Based on CO ₂ - or NH ₃ -mediated changes in the surface pH of oocytes expressing AQP or Rh-family members, the sequence of CO ₂ /NH ₃ selectivity is AQP4≅AQP5>AQP1>AmtB>RhAG.

Weak acids such as acetic acid, lactic acid, and DMO can also elicit intracellular acidifications, which have [HA] _o and pH _i dependencies similar to those described for CO ₂ . Aside from the dissolved gases discussed in the previous paragraph, many non-volatile neutral weak acids and bases may move through plasma membranes—at least in part—via protein pathways (e.g., channels, transporters) in addition to any traffic through the lipid phase of the membrane. In a general sense, all such fluxes can be thought of as non-ionic diffusion, regardless of the mechanism. CO ₂ is unusual, however, in that [CO ₂ ] _o is easily controlled, independent of [ ${\mathrm{HCO}}_3^{-}$ ] _o , either by the experimenter (who can equilibrate solutions with a known CO ₂ mixture) or by the intact organism (which can alter its external respiration). Of course, the price one pays for varying [CO ₂ ] at constant [ ${\mathrm{HCO}}_3^{-}$ ] is that—at equilibrium—pH must vary as well, as predicted by Equation 52.14 . In the laboratory, it is possible to overcome this limitation by using out-of-equilibrium CO ₂ / ${\mathrm{HCO}}_3^{-}$ solutions (see below) to change only one of the three parameters (pH, CO ₂ , or ${\mathrm{HCO}}_3^{-}$ ) at a time. For nonvolatile weak acids, the experimenter can only manipulate the total concentration of extracellular buffer ([TA] _o =[HA] _o +[A ⁻ ] _o ) and pH. When [TA] _o is fixed, changes in pH _o produce reciprocal alterations in [HA] _o and [A ⁻ ] _o , so that the magnitude of the HA-induced pH _i decrease is very pH _o sensitive.

Flux of the Neutral Weak Base

When a cell is exposed to a neutral weak base, B enters and associates with H ⁺ to form HB ⁺ , causing pH _i to rise:

This process continues until [B] _i equals [B] _o . Subsequently, there should be no further change in pH _i , provided there is no flux of HB ⁺ and no change in the rates of acid–base transporter activity. Removing external B reverses the reaction as written in Equation 52.15 , and B passively diffuses out of the cell. pH _i should then return to exactly its initial level. An example of the effect of the weak base NH ₃ is shown in Figure 52.5 . Applying 20 mM total ammonium (i.e., [NH ₃ ] _o +[ ${\mathrm{NH}}_4^{+}$ ] _o ) causes pH _i first to rise, and then stabilize (segment ab, pulse 1). Removing the NH ₃ / ${\mathrm{NH}}_4^{+}$ elicits a fall in pH _i to a value somewhat lower than the initial one (compare points a and c). The reason for this small pH _i undershoot will become clear below. As discussed above for HA-induced acidifications, the magnitude of the NH ₃ -induced alkalinization depends upon intracellular buffering power, [NH ₃ ] _o , and the degree to which entering NH ₃ is protonated. The last is governed by the relation:

where ${\mathrm{pK}}_a^{\prime }$ (~9.2 at 22°C) is the acid dissociation constant. Thus, at the initial pH _i of 7.3 (point a), approximately 99.4% of entering NH ₃ is protonated to form ${\mathrm{NH}}_4^{+}$ , whereas at a pH _i of 7.8 (point b), only approximately 98% is protonated. The dependence of the protonation of NH ₃ on the difference ( ${\mathrm{pK}}_a^{\prime }$ –pH _i ) has two consequences, analogous to the two discussed earlier for CO ₂ . First, the higher the initial pH _i , the smaller the magnitude of the alkalinization elicited by the subsequent exposure to the ${\mathrm{NH}}_4^{+}$ . Second, as one raises [NH ₃ ] _o by fixed increments, the magnitude of the alkalinization with each successive [NH ₃ ] _o increment does not increase in proportion. Thus, in Figure 52.5 , the alkalinization produced by 10 mM total ${\mathrm{NH}}_4^{+}$ is less than twice as great as that produced by 5 mM total ${\mathrm{NH}}_4^{+}$ .

In Figure 52.5 , the magnitude of the NH ₄ ⁺ -induced alkalinization actually depends more on [NH ₃ ] _o than on [total ${\mathrm{NH}}_4^{+}$ ] _o . Thus, the ${\mathrm{NH}}_4^{+}$ -induced alkalinizations will be identical for solutions in which [total ${\mathrm{NH}}_4^{+}$ ] and pH _o are varied reciprocally so as to keep [NH ₃ ] _o constant. Neutral weak bases such as lidocaine, procaine, and methylamine also elicit intracellular alkalinizations that have [B] _o and pH _i dependencies similar to those described for NH ₃ .

As noted above, the bacterial protein AmtB, an Rh homolog, functions as a gas channel permeable to NH ₃ . In addition, aquaporins 1, 4, 5, and AQP8 are permeable to NH ₃ .

It is possible to target the weak base NH ₃ and selectively raise the pH of a specific intracellular compartment. For example, Carraro-Lacroix et al. raised the pH of lysosomes by first endocytotically targeting urease to the lysosome compartment, and then exposing the cells to urea. The urease—now in the lysosomes—catalyzed the hydrolysis of the urea to produce one molecule of CO ₂ but two of the weak base NH ₃ . Both because of the 2:1 stoichiometry of generated NH ₃ :CO ₂ and the low initial lysosomal pH, the NH ₃ dominates and the net effect is the NH ₃ -driven consumption of H ⁺ to form ${\mathrm{NH}}_4^{+}$ , thereby selectively alkalinizing the lysosome.

Flux of the Cationic, Conjugate Weak Acid

When one exposes a cell to a weak base, the dominant effect on pH _i generally reflects the influx of the highly permeant neutral weak base (e.g., NH ₃ ). However, the flux of the conjugate weak acid (e.g., ${\mathrm{NH}}_4^{+}$ ) may produce pH _i decreases that are substantial or even dominant. The effects on pH _i of applying and withdrawing NH ₃ / ${\mathrm{NH}}_4^{+}$ can be separated into four steps. When a cell is exposed to NH ₃ / ${\mathrm{NH}}_4^{+}$ , the rapid influx of NH ₃ leads to an initial rise in pH _i ( Figure 52.6a ). However, this alkalinization is actually blunted by the simultaneous, although smaller, influx of ${\mathrm{NH}}_4^{+}$ . The most general mechanism of ${\mathrm{NH}}_4^{+}$ influx is the passive diffusion of ${\mathrm{NH}}_4^{+}$ , although the carrier-mediated transport of ${\mathrm{NH}}_4^{+}$ into cells can be substantial. For example, ${\mathrm{NH}}_4^{+}$ can enter cells by substituting for K ⁺ on the Na ⁺ K ⁺ pump and/or the Na ⁺ /K ⁺ /Cl ⁻ cotransporter. Although most of the entering ${\mathrm{NH}}_4^{+}$ remains ${\mathrm{NH}}_4^{+}$ , a small fraction (governed by the differences between ${\mathrm{pK}}_a^{\prime }$ and pH _i ) dissociates to form NH ₃ and H ⁺ . When external NH ₃ /NH ₄ ⁺ is removed, the NH ₄ ⁺ that previously entered, but failed to dissociate, now dissociates into NH ₃ (which rapidly leaves the cell) and H ⁺ (which is trapped within the cell). As a consequence, the final pH _i (point c in Figure 52.5 ) is slightly lower than the initial value (point a).

The acidifying effect of ${\mathrm{NH}}_4^{+}$ is much more evident when the NH ₃ / ${\mathrm{NH}}_4^{+}$ exposure lasts beyond the initial influx ( Figure 52.6a ) and equilibration of NH ₃ ( Figure 52.6b ). During an extended plateau phase, the continuing net influx of NH ₄ ⁺ produces a plateau-phase acidification, as illustrated in Figure 52.6c . In cells with a Cl ⁻ ${\mathrm{HCO}}_3^{-}$ exchanger (e.g., sheep cardiac Purkinje fibers), this plateau-phase acidification may be augmented by efflux of ${\mathrm{HCO}}_3^{-}$ in exchange for Cl ⁻ . When the external NH ₃ /NH ₄ ⁺ is eventually withdrawn, the pH _i undershoot is greatly exaggerated ( Figure 52.6d ). The magnitude of the undershoot reflects the previous influx of ${\mathrm{NH}}_4^{+}$ . The longer the exposure to NH ₃ / ${\mathrm{NH}}_4^{+}$ , the more extensive the plateau-phase acidification, and the larger the undershoot. This undershoot represents an acute intracellular acid load. Indeed, the ${\mathrm{NH}}_4^{+}$ -prepulse technique is widely used for acid loading cells in studies of pH _i regulation.

If the ${\mathrm{NH}}_4^{+}$ electrochemical gradient is reversed during the plateau phase, in a cell for which the dominant mechanism of ${\mathrm{NH}}_4^{+}$ entry is a passive flux, then pH _i rises during the plateau phase and the pH _i undershoot is converted to a shortfall.

Chemical Buffering

A weak base (B) of valence n reacts with H ⁺ to produce its conjugate weak acid (HB) of valence n+1:

This equilibrium is described by the relation:

where $K_a^{\prime }$ is the apparent acid dissociation constant. This relation can also be expressed in logarithmic form, whereupon it takes the form of Equation 52.16 . Starting with the definition of β ( Equation 52.17 ) and the statement of chemical equilibrium ( Equation 52.19 ), one can express β as a function of [H ⁺ ], $K_a^{\prime }$ , and total buffer concentration.

where 2.3 approximates ln(10), and [TB] is the concentration of total buffer (i.e., [TB]=[B ⁿ ]+[HB ⁿ⁺¹ ]). By obtaining the derivative of β with respect to [H ⁺ ] (i.e., d β/ d [H ⁺ ]) and setting this equal to zero, one can show that β is maximal when [H ⁺ ] equals $K_a^{\prime }$ . This maximal buffering power (β ^max ) is simply:

Although the pH of maximal buffering power is in general unique to the buffer, all buffers have the same molar buffering power (β ^max /[TB]), and the same dependence of β on the difference pH– ${\mathrm{pK}}_a^{\prime }$ . If more than one buffer is present, then β ^T is simply the sum of the individual buffering powers, each computed from Equation 52.20 .

The foregoing equation describing β was derived by taking the partial derivative of [B ⁿ ] with respect to pH at a fixed [TB]. That is, the analysis applies only to buffers in a closed system, in which [TB] is constant. Closed-system buffers include a solution of a nonvolatile buffer (e.g., inorganic phosphate) in a beaker, a solution of a volatile buffer (e.g., CO ₂ / ${\mathrm{HCO}}_3^{-}$ ) in a capped syringe, and an impermeant buffer (e.g., a protein) in the cytoplasm of a cell. In such a closed system, β for CO ₂ / ${\mathrm{HCO}}_3^{-}$ would be rather low at normal pH _i values, because pH _i (~7.0) is far higher than the ${\mathrm{pK}}_a^{\prime }$ of 6.1. However, cells are generally not a closed system with respect to CO ₂ / ${\mathrm{HCO}}_3^{-}$ . Rather, CO ₂ usually freely exchanges with the extracellular fluid (ECF) so that the total intracellular CO ₂ (i.e., [CO ₂ ] _i +[ ${\mathrm{HCO}}_3^{-}$ ] _i ) can change appreciably during intracellular acid–base disturbances. For example, when H ⁺ is added to cytoplasm containing CO ₂ / ${\mathrm{HCO}}_3^{-}$ , H ⁺ combines with ${\mathrm{HCO}}_3^{-}$ according to the reactions:

The cytoplasmic CO ₂ so formed is lost to the ECF. Thus, it is [CO ₂ ] _i that remains constant, while [ ${\mathrm{HCO}}_3^{-}$ ] _i and [total CO ₂ ] _i both fall. Because there is no buildup of CO ₂ within the cell, the extent of the reaction in Equation 52.22 is limited only by the availability of ${\mathrm{HCO}}_3^{-}$ . The amount of H ⁺ absorbed (i.e., buffered) by the CO ₂ / ${\mathrm{HCO}}_3^{-}$ system exactly equals the amount of ${\mathrm{HCO}}_3^{-}$ that disappears. That is:

where β ^{CO ₂} is the CO ₂ buffering power. When this equation, in differential form, is combined with the Henderson-Hasselbalch equation, it can be shown that:

when [CO ₂ ] is held constant. At very high pH values, when the equilibrium $H^{+}+{\mathrm{CO}}_3^{2-}\rightleftarrows {\mathrm{HCO}}_3^{-}$ cannot be ignored, the term 4.6 [ ${\mathrm{CO}}_3^{2-}$ ] must be added to the right-hand side of Equation 24 . This equation describes the buffering power of CO ₂ / ${\mathrm{HCO}}_3^{-}$ in an open system, such as a solution in a beaker equilibrated with gaseous CO ₂ . Because most cell membranes are highly permeable to CO ₂ , this gas rapidly equilibrates across membranes and stabilizes [CO ₂ ] _i , provided the ECF behaves as an infinite reservoir for CO ₂ . Thus, an isolated cell in vitro behaves as an open system for CO ₂ / ${\mathrm{HCO}}_3^{-}$ . Because the intact organism has mechanisms (e.g., alveolar ventilation in higher vertebrates) for stabilizing [CO ₂ ] in the ECF, CO ₂ / ${\mathrm{HCO}}_3^{-}$ behaves as an open-system buffer in vivo , both in the extra- and intracellular spaces.

In an open system, the CO ₂ / ${\mathrm{HCO}}_3^{-}$ buffer pair generally makes a substantial contribution to β ^T . For example, at a pH _i of 7.1, the buffering power of all non-CO ₂ buffers in rat renal mesangial cells is only about 10 mM, whereas the computed β ^{CO ₂} is nearly 29 mM when the cell is equilibrated with 5% CO ₂ . Thus, β ^{CO ₂} accounts for nearly 75% of the β ^T of 39 mM/pH unit.

It is essential to distinguish clearly between open- and closed-system buffers of a cell. The cell is a closed system for intracellular buffers that do not readily cross the cell membrane (e.g., inorganic phosphate and the imidazole groups of proteins). Thus, these buffers are influenced by the attributes of the cell (e.g., volume, temperature, and metabolic state), and not those of the extracellular fluid. These are termed intrinsic buffers . Biochemical and organellar buffering mechanisms ( vide infra ) are also intrinsic, and the total intrinsic buffering power (β ^I ) is the sum of biochemical, organellar, and intrinsic chemical buffering powers.

The cell is an open system when one member of the buffer pair readily crosses the cell membrane. The buffering power of such a buffer pair is very sensitive to extracellular conditions. As noted earlier, β ^{CO ₂} is proportional to [ ${\mathrm{HCO}}_3^{-}$ ] _i , which in turn is completely determined by pH _i and the extracellular P _{CO 2} (assuming that CO ₂ equilibrates across the cell membrane). The cell also behaves as an open system for buffers other than CO ₂ / ${\mathrm{HCO}}_3^{-}$ . These are generally conjugate pairs of which one member is a small, neutral molecule. Examples include NH ₃ / ${\mathrm{NH}}_4^{+}$ and acetic acid/acetate. Regardless of whether the charged species is a cation or an anion, the intracellular buffering power of an open-system buffer pair is always 2.3×[charged species] _i , provided the neutral species is equilibrated across the cell membrane. Thus, the intracellular buffering power of such a buffer pair is sensitive not only to the total amount of the buffer in the ECF (i.e., [TB] _o ), but also to pH _o . As with CO ₂ / ${\mathrm{HCO}}_3^{-}$ , these other open-system buffers can sometimes be the dominant component of β ^T . For example, β ^I in squid axons is ~9 mM. When these axons are exposed to a pH _o −7.7 NH ₃ / ${\mathrm{NH}}_4^{+}$ solution containing only 10 mM total ammonium, [ ${\mathrm{NH}}_4^{+}$ ] _i rises to 10 mM. The computed β ^{NH ₃} is 23 mM, nearly 70% of β ^T . Because the buffering power of such open-system buffers is so sensitive to factors external to the cell (e.g., pH _o ), these buffers are termed extrinsic . The total chemical buffering power of a cell is the sum of the buffering powers of the individual conjugate acid/base buffer pairs, be they intrinsic (closed system) or extrinsic (open system).

Biochemical Buffering

Because certain biochemical reactions consume or release H ⁺ and are pH _i sensitive, they can act as H ⁺ buffers. Examples include hydrolysis of ATP (which releases H ⁺ ) and hydrolysis of phosphocreatine (which consumes H ⁺ ). During the Cori cycle, the liver converts lactic acid into glucose, whereas skeletal muscle breaks down the glucose to produce ATP and more lactic acid. Also, H ⁺ buffering can arise when reactions induce a change in the ${\mathrm{pK}}_a^{\prime }$ of an ionizable group. A classic example is the buffering reactions of oxygenated versus deoxygenated hemoglobin.

A well studied example of biochemical buffering are reactions involving intermediary metabolites in rat brain. Cells that are acutely acid loaded with an increase of P _{CO 2} in the ECF respond by reducing intracellular levels of several acid metabolic intermediates (i.e., the acids of lactate, pyruvate, citrate, α-ketoglutarate, malate, glutamate, and aspartate), while raising those of glucose and glucose-6-phosphate. These observations are consistent with the hypothesis that reducing pH _i inhibits a step in the glycolytic pathway, possibly the phosphofructokinase reaction. Indeed, mouse muscle phosphofructokinase is markedly sensitive to pH changes ; reducing in-vitro pH from 7.2 to 7.1 produces more than a 90% inhibition. Rat-brain cells respond in the opposite fashion to acute intracellular alkali loads. Increases in pH _i induced by lowering the P _{CO 2} in the ECF lead to increased levels of lactate and pyruvate. Thus, the biochemical machinery of these cells seems to respond appropriately to pH _i changes, consuming H ⁺ in response to intracellular acid loads and releasing H ⁺ in response to alkali loads. The extent to which such biochemical reactions contribute to overall buffering power can be computed from changes in steady-state metabolite levels, provided that such changes are reciprocally linked to the production or consumption of neutral and/or readily diffusible molecules. For example, the consumption of one citrate molecule removes three H ⁺ , whereas malate removes two, and pyruvate, one. From data on rat brain, the buffering power of biochemical reactions amounts to about half that provided by all non-CO ₂ physicochemical buffers. Biochemical buffering power can be defined and quantitated in a manner analogous to chemical ( vide supra ) and organellar ( vide infra ) buffering powers [see Equation 52.17 ]. The biochemical buffering reaction can be written as:

where R is the reactant, P is the product, n is the valence, and q is the H ⁺ stoichiometry ( q >0). The biochemical buffering power is thus:

As noted above, biochemical and organellar ( vide infra ) buffering mechanisms, in addition to closed-system chemical buffers, comprise the intrinsic buffers.

Organellar Buffering

H ⁺ -transport systems have been identified or implicated in many intracellular organelles, including mitochodria, lysosomes, sarcoplasmic reticulum, Golgi network, endosomes, chromaffin granules, and zymogen granules. In addition, three of the cloned Na-H exchangers (NHE6, 7, and 9) are found in organellar membranes where they may contribute to organellar pH regulation. For reviews of vesicular H ⁺ pumps, see refs. It would be reasonable to expect decreases in cytoplasmic pH to stimulate H ⁺ uptake and/or inhibit H ⁺ extrusion by at least some organelles, and increases in pH _i to produce the opposite effects. Such buffering would lead to a net transfer of H ⁺ into these organelles following intracellular acid loads, and a net movement of H ⁺ into the cytoplasm following alkali loads. A transfer of acid or base across organellar membranes would constitute a de facto buffer mechanism for the bulk intracellular fluid. Although the extent to which such hypothetical organellar buffering mechanisms contribute to β ^T is not established, it can be inferred from published data that changes in extraorganellar pH can produce at least small changes in the intraorganellar pH of mitochondria, lysosomes, and sarcoplasmic reticulum vesicles. Organellar buffering power can be defined and quantitated in a manner analogous to that for physicochemical and biochemical buffering [see Equation 52.17 ]. The organellar buffering reaction is written:

where the parentheses refer to organellar contents. Organellar buffering power is thus:

Organellar buffers, along with the biochemical and closed-system chemical buffers, comprise β ^I .

Titration of Cell Homogenates

The easiest method for estimating β _i is to homogenize a tissue sample and titrate the homogenate. The slope of the pH titration curve is the buffering power of the homogenate. However, some disadvantages limit the accuracy of this technique. First, homogenization may disrupt cellular organelles and thus obscure the organellar buffering contribution. Second, homogenization-induced changes in metabolism must be prevented, for example by quick-freezing the sample before homogenization, and then performing the titration either at low temperature or in the presence of an inhibitor such as fluoride. Blocking metabolism will reduce biochemical buffering, and reducing temperature generally increases ${\mathrm{pK}}_a^{\prime }$ values, thereby altering chemical buffering.

Microinjection

The most straightforward approach for estimating β _i in an intact cell is to microinject a known amount of acid or base into a cell and monitor the resultant change in pH _i . The microinjection can be achieved either by iontophoresis or by pressure injection. This technique enables one to compute the total β _i (i.e., sum of physicochemical, biochemical, and organellar buffering) of the cell. For small neurons incubated in a CO ₂ -free solution, the iontophoresis and pressure-injection methods give β _i values of 10.8 mM/pH unit and 10.3 mM/pH unit, respectively. Disadvantages of this direct in-vitro titration approach include the necessity of using cells that (1) are large enough to withstand the microelectrode impalements, and (2) have a sufficiently compact geometry to permit rapid equilibration of the injected acid or base throughout the cell.

Weak-Acid and Base Methods

An approach that, at least in principle, is applicable to cells of all shapes and sizes, is measurement of the pH _i change produced by exposing the cell to a neutral weak acid HA ( $\mathrm{HA}\rightleftarrows H^{+}+A^{-}$ ), or weak base B ( $\mathrm{HB}\rightleftarrows H^{+}+B$ ). As discussed earlier, exposing a cell to a neutral weak acid leads to a decrease in pH _i as HA enters the cell and partially dissociates into H ⁺ and A ⁻ . The entry and dissociation halt once [HA] _i =[HA] _o . The extent of the dissociation depends on the weak acid’s ${\mathrm{pK}}_a^{\prime }$ , as well as the initial pH _i . Virtually all of the released H ⁺ is consumed by intracellular buffers ( Figure 52.8A and B ); the remaining fraction is responsible for the accompanying pH _i decrease. Although HA/A ⁻ is a buffer, it does not participate in buffering the acid load imposed by the intracellular dissociation of HA. Similarly, a B/BH ⁺ buffer system can not buffer changes in pH _i caused by the entry (or exit) of B. Because the dissociation of HA leads to the formation of one A ⁻ for each H ⁺ , the magnitude of the intracellular acid load is Δ[A ⁻ ] _i . This amount of added acid is the additive inverse of the amount of base added (ΔB) in the preceding definition of β:

Note that β _i is the non-HA/A ⁻ buffering power of the cell. It is the sum of chemical, biochemical, and organellar buffering. ΔpH _i is directly measured, and Δ[A ⁻ ] _i is calculated from final and initial values of pH _i and [HA] _o . In the simplest case, the initial [A ⁻ ] _i is zero, and final [A ⁻ ] _i is given by:

assuming [HA]=[HA] _i =[HA] _o . A commonly used weak acid for measuring β _i is CO ₂ , for which:

β _i is the non-CO ₂ / ${\mathrm{HCO}}_3^{-}$ or intrinsic buffering power, ${\beta }_i^I$ . Once ${\beta }_i^I$ is known, the total intracellular buffering power ( ${\beta }_i^T$ ) can be obtained by computing the open-system CO ₂ buffering power:

where ${\beta }_i^{{\mathrm{CO}}_2}$ , the CO ₂ buffering power, is given by 2.3×[ ${\mathrm{HCO}}_3^{-}$ ] _i , as outlined above.

There are as many as three major drawbacks of using the weak-acid method to measure β. First, [HA] _i may not exactly equal [HA] _o . In rapidly metabolizing cells for instance, [CO ₂ ] _i may exceed [CO ₂ ] _o . Second, the pH _i decrease elicited by the entry of HA may stimulate acid-extrusion mechanisms (i.e., pH _i -regulating systems described below). Thus, the actual ΔpH _i may be smaller than if all acid–base transport had been blocked ( Figure 52.7 ). By underestimating ΔpH _i , one will overestimate β _i . This error should in principle be eliminated by blocking the pH _i -regulating systems. One can also minimize the error by extrapolating the pH _i -versus-time curve back to a point at which acid extrusion is judged to have had no effect on pH _i . For example, in the hypothetical experiment of Figure 52.7 , one could extrapolate the pH _i recovery (i.e., bcdef) back to the time of the acid load. A third potential drawback of the weak-acid method is that the weak acid could indirectly alter pH _i by modifying cellular metabolism. Thus, it is critical that the weak acid affect pH _i only by entering and dissociating into H ⁺ and A ⁻ . Otherwise, any other effect on pH _i will lead to an error in the calculation of β _i .

The intrinsic intracellular buffering power can also be determined by exposing a cell to a neutral weak base, B ( Figure 52.8B ). As described in the preceding section, exposing a cell to B leads to a pH _i increase as most of the entering B combines with H ⁺ to form HB ⁺ . The entry of B and the alkali loading of the cell continues until [B] _i =[B] _o . Almost all of the entering H ⁺ that is converted to HB ⁺ is derived from cellular buffers. The minute amount that comes from the pool of free H ⁺ is responsible for the rise in pH _i . Thus, the change in [HB ⁺ ] _i is equivalent to the amount of strong base added to the intracellular fluid:

In the simplest case, the initial [HB ⁺ ] _i is zero, and the final [HB ⁺ ] _i is calculated from the statement of chemical equilibrium:

assuming that [B]=[B] _i =[B] _o . The most commonly used weak base is NH ₃ :

[ ${\mathrm{NH}}_4^{+}$ ] _i is determined in a manner analogous to that outlined earlier for [ ${\mathrm{HCO}}_3^{-}$ ] _i . Note that β _i in this case is the non-NH ₃ / ${\mathrm{NH}}_4^{+}$ buffering power (i.e., the buffering power of all buffers other than NH ₃ / ${\mathrm{NH}}_4^{+}$ ). If the NH ₃ titration is performed in the absence of CO ₂ / ${\mathrm{HCO}}_3^{-}$ , then the measured β _i is ${\beta }_i^I$ . If CO ₂ / ${\mathrm{HCO}}_3^{-}$ is present, then the measured β _i is ${\beta }_i^I+{\beta }_i^{{\mathrm{CO}}_2}={\beta }_i^T$ . An approach commonly used with ${\mathrm{NH}}_4^{+}$ is to compute β from the pH _i decrease caused by decreasing or withdrawing external NH ₄ ⁺ . During such experiments, one usually blocks acid-extrusion mechanisms (e.g., by removing external Na ⁺ ) that would likely counteract the acidification. By reducing [ ${\mathrm{NH}}_4^{+}$ ] _o in a stepwise fashion ( Figure 52.9a ), one can compute β _i as a function of pH _i ( Figure 52.9b ), as has been done in mesangial cells and gastric parietal cells. An alternate approach one can use with ${\mathrm{NH}}_4^{+}$ is to compute β _i from the pH _i increase that accompanies stepwise increases of external ${\mathrm{NH}}_4^{+}$ .

Introduction

As noted above, passive fluxes of H ⁺ , monovalent cationic acids (e.g., ${\mathrm{NH}}_4^{+}$ ), and monovalent anionic bases (e.g., ${\mathrm{HCO}}_3^{-}$ ) are governed by the electrochemical gradient for the ion in question, and normally tend to acidify the cell. In principle, such passive transport processes could be exploited for protecting the cell against alkaline loads. Nevertheless, such passive processes generally do not appear to have a substantial impact on pH _i . A notable exception discussed above is the GABA _A -activated Cl ⁻ channel at the crayfish neuromuscular junction and in cells from turtle cerebellum and rat hippocampus. This Cl ⁻ channel can conduct ${\mathrm{HCO}}_3^{-}$ and thus mediate a substantial ${\mathrm{HCO}}_3^{-}$ efflux that lowers pH _i . For the most part, however, major acid-loading transport pathways are ${\mathrm{HCO}}_3^{-}$ transporters—the most common ones being Cl ⁻ ${\mathrm{HCO}}_3^{-}$ exchange and Na ⁺ / ${\mathrm{HCO}}_3^{-}$ cotransport.

Cl ⁻ HCO ⁻ ₃ and Cl ⁻ OH ⁻ Exchangers

General Function and Molecular Identity

Found in a wide variety of animal-cell membranes, Cl ⁻ ${\mathrm{HCO}}_3^{-}$ exchangers normally couple the influx of Cl ⁻ to the efflux of ${\mathrm{HCO}}_3^{-}$ ( Figure 52.10a–h ). Na ⁺ -independent Cl ⁻ ${\mathrm{HCO}}_3^{-}$ exchangers (to distinguish them from the Na ⁺ -dependent Cl ⁻ ${\mathrm{HCO}}_3^{-}$ exchangers discussed later) are thought to serve two major functions in non-epithelial cells: regulation of intracellular pH (pH _i ), and regulation of intracellular [Cl ⁻ ]. The erythrocyte exchanger is known as the “band 3 protein” because of its position on SDS-polyacrylamide gels (for review, see ref. ) Band 3 has also been termed AE1 because it was the first “anion exchanger” to be cloned and sequenced ). As described in greater detail in Chapter 54 , the AE gene subfamily comprises at least three related genes AE1, AE2, and AE3. The AE gene subfamily is a branch of the bicarbonate-transporter (SLC4) family that includes the Na ⁺ -coupled ${\mathrm{HCO}}_3^{-}$ transporter subfamily (see Chapter 53 ). Several members of the SLC26 family can also mediate Cl ⁻ ${\mathrm{HCO}}_3^{-}$ exchange. However, unlike the AE family, which virtually always exhibits a Cl ⁻ : ${\mathrm{HCO}}_3^{-}$ stoichiometry of 1:1, the SLC26 family can exhibit stoichiometries of 1:2, 1:1, or 2:1 ; for review, see ref. .