INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

THE INTRACELLULAR FREE CONCENTRATION of Na⁺ ions ([Na⁺]_i) is an important determinant of cellular function. In neurons, the electrochemical gradient of Na⁺ across the plasma membrane plays a central role in determining excitability, and changes in [Na⁺]_i can modulate the activities of ion channels (40), Na⁺-coupled transporters and uptake mechanisms (4, 5, 20), and enzymes (12). Disturbances in neuronal intracellular Na⁺ homeostasis also play a role in pathophysiological events, including excitotoxic/anoxic injury (7, 9, 11, 39).

Although a variety of methods have been employed to estimate [Na⁺]_i, Na⁺-sensitive fluorescent dyes, especially Na⁺-binding benzofuran isophthalate (SBFI), are assuming an increased importance. Despite the many advantages of SBFI (see Refs. 26 and 32), a number of difficulties are associated with the calibration procedures required to convert experimentally derived SBFI ratio values into [Na⁺]_i. Thus in vitro calibration fails to take into account the spectral shifts that are introduced when the dye is present in the cytosol (3, 8, 15, 23, 28). On the other hand, the in situ techniques described to date do not favor the straightforward determination of separate values for all of the constants necessary to convert the ratio of emitted SBFI fluorescence signals into [Na⁺] values according to the standard equation of Grynkiewicz et al. (16)

(1)

where  = S_f2/S_b2 and is the ratio of the fluorescence of the free (unbound) dye (S_f2) to bound dye (S_b2) at the second excitation wavelength (₂^ex), K_d is the apparent dissociation constant of SBFI for Na⁺, R is the fluorescence ratio, R_min is the fluorescence ratio at [Na⁺] = 0 mM, and R_max is the fluorescence ratio at saturation. The commonly employed "three-point" calibration technique (18), for example, provides only a composite value for K_d; the accurate determination of a separate value for  (and, thus the K_d of SBFI for Na⁺) is precluded because a value for S_b2 cannot be derived. In the present study, therefore, we developed a method to determine, from data derived from full in situ calibrations, all of the constant parameters required to resolve Eq. 1. We also describe a one-point technique for the in situ calibration of SBFI ratios and examine the utility of this procedure to calibrate the changes in SBFI ratios evoked in neurons by veratridine or anoxia.

Another difficulty associated with SBFI is that, in common with many fluorophores, SBFI fluorescence is sensitive to changes in [H⁺] (13, 18, 28, 29, 35, 36). Although the full pH sensitivity of SBFI has been described in a cell-free in vitro system (26), the effects of changes in intracellular H⁺ concentration ([H⁺]_i) on measurements of [Na⁺]_i made with SBFI in situ remain relatively poorly defined. This represents a potential limitation to the accurate estimation of [Na⁺]_i with SBFI, in part because changes in [Na⁺]_i may give rise to changes in intracellular pH (pH_i), either directly (e.g., via alterations in Na⁺/H⁺ exchange) or indirectly [e.g., via alterations in Na⁺/Ca²⁺ exchange and subsequent changes in intracellular Ca²⁺ concentration ([Ca²⁺]_i)]. Therefore, we used the procedures developed in the first part of the study to assess the effect of changes in pH_i on the K_d of SBFI for Na⁺ in situ. We also characterized the [H⁺] sensitivity of SBFI in situ at different values of [Na⁺]_i and developed a method for correcting, if necessary, [Na⁺]_i values measured with SBFI for changes in pH_i.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Cell Culture

Solutions

Microspectrofluorometry

Measurements of [Na⁺]_i and pH_i were performed using the dual excitation ratio method, employing an imaging system (Atto Instruments, Rockville, MD) in conjunction with an Axiovert 10 epifluorescence microscope (Carl Zeiss Canada, Don Mills, ON). Full details have been provided previously (2, 14, 38). In brief, SBFI- or BCECF-loaded neurons were excited via a ×40 LD Achroplan objective with light provided by a 100-W Hg arc burner and band-pass filtered alternately at 334 and 380 nm (SBFI) or at 488 and 452 nm (BCECF). To reduce photobleaching of the fluorophores, the output of the ultraviolet (UV) light source was attenuated electronically, neutral density filters were placed in the light path, and a high-speed shutter was employed to limit UV exposure to the periods required for data acquisition. Fluorescence emissions, measured at 510 or 520 nm from neurons loaded with SBFI or BCECF, respectively, were detected by an intensified charge-coupled device camera (Atto Instruments) and collected from regions of interest placed on individual neuronal somata. Experiments were repeated on at least three (usually 5) different coverslips, each allowing collection of data from up to 99 individual neurons simultaneously. Raw emission intensity data at each excitation wavelength were corrected for background fluorescence before the calculation of a ratio; the intensity of background fluorescence was typically <15% of the total signal at any given excitation wavelength and remained constant during the course of a given experiment. Ratio pairs were acquired at 1- to 12-s intervals and analyzed off-line. The one-point high-[K⁺]/nigericin technique was employed to convert background-corrected BCECF emission intensity ratios (BI₄₈₈/BI₄₅₂) into pH_i values, as detailed previously (2, 6, 38). The procedures employed to calibrate the ratios of the emitted SBFI signals are detailed in RESULTS.

Data Analysis

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Calibration of SBFI Ratio Values

Determination of R_min, R_max, and K_d. Full calibrations were performed at room temperature in 23 neuronal populations by exposing neurons sequentially to pH 7.35 media containing 5 µM gramicidin D and eight different [Na⁺] values (range, 0-130 mM). Typical changes in BI₃₃₄ and BI₃₈₀ values are illustrated in Fig. 1A. Fluorescence emitted during excitation at 334 nm was essentially unaffected by changes in [Na⁺], whereas during excitation at 380 nm, emitted fluorescence intensity decreased as [Na⁺] increased. The lack of an effect of changes in [Na⁺] on fluorescence emitted in situ during excitation at 334 nm in epifluorescence systems has been reported and discussed previously (3, 13, 15, 21, 23, 24, 28, 29, 34, 35).

View larger version (25K): [in this window] [in a new window]   Fig. 1.   In situ full calibration of Na+-binding benzofuran isophthalate (SBFI) at room temperature and extracellular pH (pHo) 7.35. A: a full calibration experiment in which 10 neurons on a single coverslip were exposed to 5 µM gramicidin D-containing solutions at the [Na+] values shown above or below the traces. Displayed are changes in emitted fluorescence intensities during excitation at 334 nm (dotted line) and 380 nm (continuous line); the signals were corrected for background fluorescence. Whereas the BI380 value displayed stepwise changes in response to changes in [Na+], the BI334 value appeared insensitive. The trace is representative of 23 independent experiments. B: BI334/BI380 ratio values were computed from the data shown in A and normalized to unity at intracellular free Na+ concentration ([Na+]i) = 10 mM. The trace shows the stepwise changes in normalized BI334/BI380 ratio values (Rn) evoked by successive increases in [Na+]. C: curve a (solid line) is a plot of [Na+] vs. mean normalized BI334/BI380 ratio values (Rn) obtained from 4 calibration experiments of the type shown in B, conducted on sister cultures on the same day. The curve is the result of a three-parameter hyperbolic fit (Eq. 3) to the data points () indicated. Standard error bars are contained within each datum point. Curves b and c (dashed lines) are plots of [Na+] vs. mean normalized background-subtracted emission intensities during excitation at 380 nm. The curves are the result of three-parameter hyperbolic decay fits (Eq. 6) to the respective data points indicated. The data points employed for curve b () and curve c () were not drift corrected and were drift corrected, respectively (see text). In each case, standard error bars are contained within each datum point. D: Hanes plot of the calibration data shown in C (plot a). The continuous line is a linear least-squares regression fit to the data points indicated (r2 = 0.99). Rn(max) is drawn from the slope = 1/[Rn(max)  Rn(min)], and Kd is obtained from the intercept on the abscissa.

(2)

(3)

(4)

(5)

View this table: [in this window] [in a new window]   Table 1.   Values of Rn(min), Rn(max), and Kd determined by three-parameter hyperbolic fits to data points obtained in full calibrations

Determination of  and K_d. Neither Hanes plots of data derived from full calibrations nor the commonly employed three-point procedure (18) permits the straightforward determination of a separate value for  (and, thus a value for the K_d of SBFI for Na⁺). To determine , one needs to know S_f2 and S_b2, the intensities of fluorescence emissions when exciting the fluorophore at the second wavelength (₂^ex = 380 nm) at [Na⁺] = 0 mM and at saturating Na⁺, respectively. Although a value for S_f2 can be obtained experimentally, the determination of S_b2 requires very high [Na⁺] and is compromised by the sensitivity of SBFI to changes in ionic strength and other factors (3, 23, 26, 28). To derive S_b2, we employed the BI₃₈₀ value at [Na⁺] = 10 mM as a normalization factor. Normalized BI₃₈₀ values (S_n2) at different [Na⁺] are shown in Fig. 1C (plot b). The data points were fitted (r² >0.99) with a three-parameter hyperbolic decay equation having the form

(6)

where , , and  are constants. From Eq. 6 it can be seen that when [Na⁺] = 0, S_n2 = S_f2 = ( + ), and when [Na⁺] = , S_n2 = S_b2 = . Consequently

(7)

The values of the parameters for Eq. 6, obtained from the fit to the data points shown in Fig. 1C (plot b), were  (i.e., S_b2) = 0.41 ± 0.01,  = 0.94 ± 0.01, and  = 17.96 ± 0.97; thus S_f2 = 1.34 ± 0.03 and  (S_f2/S_b2) = 3.31 ± 0.03. As noted above, K_d = 51.85 ± 2.12 mM; thus the calculated K_d of SBFI for Na⁺ = 15.69 ± 0.15 mM.

View larger version (20K): [in this window] [in a new window]   Fig. 2.   Normalized and drift-corrected BI334 and BI380 values. Ten neurons on a single coverslip were exposed to 5 µM gramicidin D-containing solutions at the [Na+] values shown. Changes in BI334 and BI380 values were normalized to unity at [Na+] = 10 mM (dotted lines). The magnitude of the decline in BI334 values from unity at any given time point was then employed to correct (solid line) the normalized BI380 values for the drift that occurred over the course of the experiment.

One-point calibration. Full calibrations may be impractical at the end of an experiment, due to a loss of fluorescence signal with time. We therefore explored the possibility of applying a one-point procedure to calibrate SBFI ratios in situ. To illustrate the one-point technique, we measured the changes in SBFI ratios that occurred in hippocampal neurons at 37°C in response to 30 µM veratridine (n = 5, a total of 32 neurons) or anoxia (n = 12, a total of 91 neurons). At the end of an experiment, neurons were exposed to a single calibrating solution containing 5 µM gramicidin D and 10 mM Na⁺. The resulting BI₃₃₄/BI₃₈₀ value at Na⁺ = 10 mM for a given neuron in the sampled population was then used as a normalization factor for that neuron. After dividing experimentally derived BI₃₃₄/BI₃₈₀ values from a given neuron by the normalization factor for that neuron, each R_n was converted to [Na⁺]_i using Eq. 2 and the parameters [R_n(min), R_n(max), , and K_d] determined in a full calibration. Estimated in this manner, resting [Na⁺]_i before veratridine or anoxia was 9.8 ± 0.3 mM (n = 17), a value similar to that reported by others in hippocampal neurons (30, 34). During anoxia, [Na⁺]_i increased to 43.1 ± 2.8 mM (n = 12; Fig. 3A), a rise similar to that observed during energy deprivation or exposure to excitotoxins in a variety of mammalian central neurons (9, 10, 30). Veratridine (30 µM) evoked an increase in [Na⁺]_i that failed to recover toward resting values during the recording period (Fig. 3A).

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Changes in [Na+]i evoked by veratridine and/or anoxia. A: changes in [Na+]i induced by a 5-min exposure, indicated by the first bar above the traces, to either 30 µM veratridine () or an anoxic medium (continuous line) in 2 populations of neurons on different coverslips. Each experiment was performed at 37°C and pHo 7.35. For each selected neuron in the sampled population, a one-point in situ calibration of the SBFI ratio signal, indicated by the second bar above the traces, was performed with a pH 7.35 medium containing 5 µM gramicidin D and 10 mM Na+. The responses to veratridine and anoxia are means of data obtained simultaneously from 9 and 12 neurons, respectively. B and C: after a 5-min period of anoxia, indicated by the first bar above the traces, 5 neurons on a single coverslip were exposed sequentially to calibrating media (pH 7.35) containing 5 µM gramicidin D and 6, 10, or 40 mM Na+. Ratios of the BI334 and BI380 signals obtained from each neuron during the course of the experiment were then calibrated separately by the three-point (B) and one-point (C) methods, the former using the equations provided by Harootunian et al. (18) and the latter using Na+ = 10 mM as the normalization factor and values for Kd, Rn(min), and Rn(max) obtained from a full calibration (calibration 3, Table 1).

Effects of Changes in pH on [Na⁺]_i Measurements With SBFI In Situ

Effect of calibration media on pH_i. A change in pH_i during a calibration might affect the precision of the procedure. In addition, although SBFI ratios are often calibrated with a combination of gramicidin D, monensin, and ouabain (e.g., see Refs. 10, 13, 29, 34), it was originally suggested that gramicidin D alone might provide a more accurate calibration for estimating cytosolic [Na⁺] (18). Therefore, we assessed the effects on pH_i of exposing neurons to 5 µM gramicidin D alone or in combination with 10 µM monensin and 1 mM ouabain; in all cases, pH_o was 7.35 at 37°C.

View larger version (13K): [in this window] [in a new window]   Fig. 4.   Effect of calibrating media on intracellular pH (pHi) in neurons loaded with 2',7'-bis(2-carboxyethyl)-5(6)-carboxyfluorescein (BCECF). A: 2 different neurons on a single coverslip, one with resting pHi ~7.48 and the other with resting pHi ~6.95, during perfusion with standard pH 7.35 medium at 37°C, were exposed to a pH 7.35 medium containing 10 mM Na+ and 5 µM gramicidin D for the period indicated by the bar above the traces. The calibrating medium decreased the pHi of the neuron with a high initial resting pHi (continuous line) and increased the pHi of the neuron with a low initial resting pHi (). The subsequent addition of 10 µM nigericin failed to evoke an additional change in pHi. B: the changes in steady-state pHi (pHi) evoked in 22 different neurons by exposure to a pH 7.35, 10 mM Na+ medium containing 5 µM gramicidin D are plotted against the pHi values measured before application of the calibrating medium. The line shown is a linear least-squares regression fit to the data points indicated (r2 = 0.96). C: a neuron perfused with a standard pH 7.35 medium at 37°C was exposed to a pH 6.80 calibration medium containing 10 mM Na+ and 5 µM gramicidin D; pHi fell to a new steady-state value. The pHo was subsequently increased to 7.80, which caused pHi to increase to a new steady-state level. A one-point BCECF calibration with 10 µM nigericin was performed immediately after the end of the record shown. The trace is representative of responses observed in 51 neurons in 5 independent experiments.

Effects of changes in pH_i on the K_d of SBFI for Na⁺. To assess whether the K_d of SBFI for Na⁺ in situ is sensitive to changes in pH, full calibrations were performed at pH_o 6.80 and pH_o 7.80 (n = 8 in each case). The resulting normalized BI₃₃₄/BI₃₈₀ ratios and normalized drift-corrected fluorescence intensities (at ₂^ex = 380 nm) were then fitted by the appropriate equations (Eqs. 3 and 6, respectively). BI₃₃₄ values were essentially unaffected as pH_o increased from 6.80 to 7.80 (data not shown), as was the value for R_n(min) (0.75 ± 0.01 and 0.74 ± 0.03 at pH_o 6.80 and 7.80, respectively), whereas R_n(max) increased slightly from 2.24 ± 0.02 to 2.36 ± 0.05. Values for K_d also increased slightly, from 50.04 ± 1.88 mM to 54.67 ± 3.26 mM, as pH_o increased from 6.80 to 7.80; this change reflected an increase in , and K_d values obtained at pH_o 6.80 and 7.80 (17.95 ± 0.03 and 17.55 ± 0.20 mM, respectively) were, in both cases, not significantly different from the K_d value established at pH_o 7.35 (17.99 ± 0.31 mM; see above).

Effects of changes in pH_i on SBFI ratio values and [Na⁺]_i. Initially, we examined the effects on SBFI ratios of a series of calibrating media containing 5 µM gramicidin D and 0, 10, 40, or 130 mM Na⁺, at three different pH values (6.80, 7.35, and 7.80). Experiments were performed at room temperature (n = 6) and 37°C (n = 3); no differences were observed between results at the two temperatures, and the data were, therefore, pooled. As illustrated in Fig. 5A, R_n values at a given [Na⁺] were reduced at pH_o 6.80 and increased at pH_o 7.80, compared with pH_o 7.35, both effects increasing in magnitude as [Na⁺] was increased from 0 to 130 mM. Normalized ratios were converted into [Na⁺]_i using values for R_n(min), R_n(max), , and K_d determined at pH_o 7.35 in full calibrations, and the resulting traces of the effects of changes in pH on [Na⁺]_i at the different Na⁺ concentrations are shown in Fig. 5B. Acidification resulted in an apparent decrease and alkalinization in an apparent increase in [Na⁺]_i when [Na⁺]  10 mM.

View larger version (26K): [in this window] [in a new window]   Fig. 5.   [H+] sensitivity of SBFI in situ. A: neurons were superfused with a calibration medium containing 5 µM gramicidin D, and [Na+] was increased sequentially from 0 to 10 to 40 to 130 mM, as indicated by the bars beneath the trace. At each [Na+], pHo was changed from 7.35 to 6.80 to 7.80. BI334/BI380 ratio values for each cell in the sampled population were normalized to unity at [Na+] = 10 mM and pHo = 7.35. The trace is a mean of data obtained simultaneously from 18 neurons on a single coverslip. B: the normalized BI334/BI380 ratio values shown in A were transformed into [Na+] using the one-point procedure described in the text. C: apparent changes in [Na+] ([Na+]) at different absolute values of [Na+]i (10 mM, ; 40 mM, ; and 130 mM, ) are shown as a function of pHi. For each absolute [Na+], values of [Na+] were calculated according to Eq. 8 and were obtained from 9 experiments (3 conducted at 37°C and 6 conducted at room temperature) of the type illustrated in A; where not shown, standard error bars are contained within the datum point. The lines shown are linear least-squares regression fits to the data points indicated for each absolute value of [Na+] (r2 >0.99 in each case). D: [Na+](CF) as a function of [Na+]i at different pHi values. [Na+](CF) is the value of [Na+] (in mM) that is required to correct the apparent [Na+] obtained at the pHi values indicated for the pH sensitivity of SBFI in situ. The continuous lines are linear least-squares regression fits to data points obtained experimentally at pHi 6.85 () and 7.83 () for [Na+] = 10, 40, and 130 mM in each case (r2 >0.99 in each case). Each datum point represents the mean value obtained from 6 to 9 similar experiments; where missing, standard error bars lie within the point. Dotted lines show the values for [Na+](CF) that were calculated for pHi = 6.85 (), 7.10 (), 7.60 (), and 7.83 () by means of Eq. 9 and plotted as a function of absolute [Na+]i.

(8)

(9)

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The accurate in situ calibration of SBFI is important, not only because the spectral properties of the dye in situ differ markedly from those in vitro, but also because physiologically important changes in neuronal [Na⁺]_i may be small with respect to resting levels (4, 28). Although full SBFI calibrations have been fitted by a variety of different means (e.g., see Refs. 3, 13, 15, 19, 24, 29, 37), in this study we employed a three-parameter hyperbolic equation (Eq. 3) that not only provided a simple method for estimating R_min and K_d, but could also be transformed easily into the standard equation of Grynkiewicz et al. (16) to determine R_max. This is advantageous, given that the high ionic strength media required for the experimental determination of R_max may affect the characteristics of the fluorophore in situ (3, 23). Because 140 mM Na⁺ is not the saturation point for SBFI, the equation employed in the present study also obviates the approximation made when the ratio of fluorescence intensities emitted during excitation at 380 nm for [Na⁺] = 0 mM and ~140 mM are employed to derive values for  and thus the K_d of SBFI for Na⁺. We also took advantage of the insensitivity of the SBFI emission signal during excitation at 334 nm to changes in [Na⁺] in situ to correct the Na⁺-sensitive 380 nm signal for the drift that often occurs during the course of a full calibration experiment. The normalized drift-corrected BI₃₈₀ signal was fitted with a three-parameter hyperbolic decay equation (Eq. 6), a procedure that facilitated the determination of  and thus the K_d of SBFI for Na⁺. Although values for R_max and K_d obtained by the aforementioned methods were similar to those derived from Hanes plots of full calibration data, the equations employed in the present study offer the advantage over previously described methods (including Hanes plots and the three-point procedure) of allowing the straightforward in situ determination of separate values for  and K_d as well as R_min and R_max (i.e., individual values for all the constant parameters of the standard equation). In this way, potential errors introduced by the use of K_d values obtained in vitro for the calibration of signals from experiments in intact cells can be avoided, and the effects of experimental maneuvers, such as changes in pH, on K_d values can easily be determined.

Values for R_min, R_max, , and K_d estimated from three-parameter hyperbolic fits to the data points obtained in full calibrations were highly reproducible. However, a full calibration employing 8 concentrations of Na⁺ at the end of an experiment may be impractical (e.g., see Ref. 13). A frequently employed alternative is the three-point method introduced by Harootunian et al. (18). Nevertheless, marked loss of signal may occur even during this less protracted procedure under some circumstances (e.g., following anoxia in mammalian neurons; unpublished observations). Given these limitations, we examined the possibility of applying a one-point technique to calibrate SBFI ratio values in situ. In this procedure, a full calibration curve is constructed (Fig. 1C, curve a), and the curve is constrained to pass through the points BI₃₃₄/BI₃₈₀ = 1.0, [Na⁺] = 10 mM. The advantage of this normalization step is that it permits a one-point calibration for each cell studied. At the end of every experiment, cell(s) are exposed to a medium containing 10 mM Na⁺ and ionophore(s), and BI₃₃₄/BI₃₈₀ values from the entire experiment for a given cell are divided by the BI₃₃₄/BI₃₈₀ value at [Na⁺] = 10 mM for that cell; the normalized BI₃₃₄/BI₃₈₀ values are then used to calculate [Na⁺]_i, utilizing Eq. 2 and the appropriate fitted calibration parameters. The latter are derived from full in situ calibration experiments, which are required only when the cell type under study or optical equipment is changed. Not only does a one-point calibration offer the advantages of being simpler and faster to perform at the end of an experiment than a full or a three-point calibration, but also the accuracy of the method appears at least equivalent to that of a three-point procedure. Thus changes in [Na⁺]_i evoked by anoxia in rat hippocampal neurons could be estimated as precisely by the one-point technique as by the three-point procedure (Fig. 3, B and C). The utility of the one-point calibration procedure is also illustrated by the data presented in Fig. 5B, where the solution employed for the one-point calibration contained 10 mM Na⁺ at pH 7.35. From this figure, it is apparent that the calculated values of [Na⁺]_i at pH_o 7.35 closely approximate the values of [Na⁺]_o employed during the course of the experiment. Thus although only a single calibration point was employed, the normalized BI₃₃₄/BI₃₈₀ values obtained at values of [Na⁺]_o other than 10 mM were accurately transformed into appropriate values of [Na⁺]_i. These points having been noted, it is nevertheless important to state that absolute values for [Na⁺]_i derived via any calibration procedure should be held with caution.

The changes in pH_i that may occur not only in response to changes in [Na⁺]_i but also to the ionophores employed in calibration procedures represent a potential confound to estimates of [Na⁺]_i made with SBFI. In the present study, the application of gramicidin D alone was found to equilibrate pH_i and pH_o, suggesting that neither monensin nor nigericin are required in calibrating media to abolish transmembrane H⁺ gradients. On the other hand, given the sensitivity of SBFI ratio measurements to [H⁺]_i in situ (see below), the fact that the ionophore(s) employed in SBFI calibration procedures cause pH_i to equal pH_o reinforce the suggestion (23, 35) that media employed to calibrate SBFI ratio signals should be titrated to the normal mean resting pH_i of the cell type under study.

Consistent with the reported apparent negative logarithm of the acidic dissociation constant (6.09) of SBFI in vitro (26; also see Ref. 17), changes in pH in the range 6.8-7.8 exerted minimal effects on the K_d of SBFI for Na⁺ measured in situ. With regard to the [H⁺] sensitivity of SBFI ratio measurements at a constant [Na⁺], acidification resulted in an apparent decrease and alkalinization in an apparent increase in [Na⁺]_i when [Na⁺] 10 mM, results that are in broad agreement with those presented previously (13, 28, 29, 33, 34, 36). Although Harootunian et al. (18) failed to observe significant changes in SBFI ratios when the intracellular compartment of REF52 cells was alkalinized by 0.4 pH units, Rose and Ransom (34, 35), for example, found that at [Na⁺] = 20 mM, 0.4-pH unit changes induced apparent changes in [Na⁺] of 3.0 ± 0.6 mM and 6.1 ± 1.5 mM in hippocampal neurons and astrocytes, respectively (acidification and alkalinization evoking decreases and increases, respectively, in apparent [Na⁺]). Similarly, Nett and Deitmer (29) observed an apparent 0.77 mM decrease in [Na⁺]_i for a 0.3-pH unit reduction in the pH of the calibrating solution at [Na⁺] = 10 mM in leech giant glial cells. The greater magnitude of the apparent [Na⁺] changes evoked by alterations in pH reported by Rose and Ransom (34, 35), compared with Nett and Deitmer (29), may in part reflect the observation made in the present study that the magnitude of the effect of a change in pH on apparent [Na⁺]_i measured with SBFI is dependent upon the absolute value of [Na⁺]_i. However, for a given [Na⁺], the relationship between the change in pH_i and the change in apparent [Na⁺]_i is reasonably linear (Fig. 5C), and a simple procedure was developed to correct for pH-induced changes in apparent [Na⁺] measured with SBFI. The correction procedure described may be of use in experiments in which [Na⁺]_i and pH_i are measured concurrently, under which conditions [Na⁺]_i could be corrected for pH_i on a region-by-region or pixel-by-pixel basis (as described for simultaneous measurements of [Ca²⁺]_i and pH_i; see Refs. 25 and 27) or in studies where SBFI fluorescence measurements are made at a known pH_i (e.g., following the application of an H⁺ ionophore or during combined patch-clamp and fluorescence ratio imaging, in which the patch pipette solution contains a high concentration of H⁺ buffer). Nevertheless, together, the results of the present study indicate that the effects of changes in pH_i on neuronal [Na⁺]_i values estimated with SBFI are relatively small and are unlikely to affect the interpretation of results under most experimental conditions.

In summary, we developed and tested simplified procedures for calibrating SBFI ratio measurements in situ. Values for all of the SBFI calibration constants are drawn from in situ measurements by means of three-parameter hyperbolic equations, and, once the parameters of the standard equation have been determined, SBFI ratios measured during the course of an experiment are transformed into [Na⁺]_i values by a one-point procedure applied at the end of the experiment. The results also demonstrate that the in situ pH sensitivity of SBFI ratio measurements is related linearly to [Na⁺]_i. Although SBFI is only weakly sensitive to changes in pH in the range pH 6.8-7.8 (compared, for example, with fura 2 or indo 1; see Refs. 1, 25, and 31), correction factors can be employed, if required, to correct SBFI-derived measurements of apparent [Na⁺]_i for pH_i.