JOURNAL OF CELLULAR PHYSIOLOGY 151596-603 (1992)
Quick and Accurate Method to Convert
BCECF Fluorescence to pH,: Calibration in
Three Different Types of Cell Preparations
M A R I L Y N R. JAMES-KRACKE
Department of Pharmacology, School of Mcdionc, University of Mksouri-Columbia,
Columbia, Missouri 652 12
A rapid, easy, and accurate method for converting the fluorescence of BCECF to
pH, as an alternative to the nigericin method, is described. The ratio of the
fluorescence intensities for BCECF can be converted to pH between 4 and 9 by a
formula similar to the one used to calculate [Ca2'Ii from the fluorescence offuraz.
The formula is inverted because H i binding to BCECF causes a decrease in
fluorescence, whereas Ca2+ binding to fura2 causes an increase in fluorescence.
The ratio of the fluorescence intensities i s a sigmoidal function of the [H'] between p H 4 and 9 with an essentially linear mid region from pH 6 to 8. This
calibration procedure in cells is similar to thc popular method for fura2 where
ionomycin, Ca",
and an alkaline EGTA solution are added in succession to
change the intracellular pCa from 4 to 9. For BCECF in cells, a protonophore,
FCCP or CCCP, is added and the cells are titrated with acid to an intracellular pH
of 4 and then back to pH 9 with base by observing the gradual change in
fluorescence as it asymptotically reaches its limiting minimum and maximum
values. This method does not require changing the medium to one with high KCI
to depolarize the membrane potential nor does the proton concentration need to
be equilibrated across the plasma membrane. The technique can be used to
calibrate BCECF in sheetsof cells. as well as suspensions of cells over a wide range
o 1YY:! wilcy-Liss, Inc
of p H sensitivities.
The intracellular pH (pH,) and the intracellular
free calcium concentration ([Ca2+Ii)are important regulators of activity in most cells (Grinstein and Cohen,
1987; Wray, 1988). Changes in [Ca2+l,are often linked
to changes in pH, and vice versa (Grinstein and Goetz,
1985). Studies of [Ca2+liand pHi have accelerated rapidly since the development of the ion sensitive fluorescent dyes, fura2 (Grynkiewicz et al., 1985) and BCECF
(Rink et al., 1982).The ease of calibration (Grynkiewicz
et al., 1985; Rink et al., 1982) and loading (Tsicn, 1981)
and the advantages of the ratio method for studying
cellular samples (Tsien and Poenie, 1986) have largely
been responsible for the popularity of these dyes.
The only major difference in the procedures to monitor fura2 and BCECF thus far has been the approach to
the calibration (Rink et a]., 1982; Thomas et al., 1979;
Moolenaar et al., 1983). For BCECF, a Kt/H+ exchanger, nigericin, is used to equilibrate the pH, and
extracellular pH (pH,) in a depolarizing high K+ medium. The pH, is changed t o three values over the linear range of its pH sensitivity. Difficulties with this
technique include cell volume changes due to Na+ loss,
very slow transmembrane pH equilibration with
nigericin in some cell types (Negulescu and Machen,
1990),the lack of complete equilibration of pH, and pH,
(Eisner et al., 1989a; Restrepo et al., 1990), and artifacts due to hypercontraction of depolarized myocytes
Q 1992 WILEY-LISS, INC
(Borzak et al., 1990).More importantly, this calibration
approach is not suitable for studies of pH, activation
of the Na+/H+ exchange since much of the region to
be tested is below the linear range (Restrepo et al.,
1990).
We have shown that a similar calibration procedure
and a1 orithm used for conversionof fura2 fluorescence
to [Ca 1, can be used for conversion of BCECF fluorescence in cells to pH,. This quick and easy approach is
demonstrated in three types of cell preparations including sheets of cells. The protonophore FCCP or CCCP is
added before the pH, is adjusted to 4 and then 9 by the
addition of acid and base, respectively. This is analogous to adding ionomycin to cells loaded with fura2 or
quin2 and then Ca2+and an alkaline EGTA solution in
succession t o change intracellular pCa from 4 to 9
(Tsien et al., 1982). This type of calibration procedure,
using a n algorithm suitable for the sigmoidal calibration curve, allows BCECF to be used over a wider range
of its sensitivities. It has the advantage that the medium does not need to be changed to high KC1 which
saves time and avoids the loss of cells from suspensions.
if+
Received .June 19,1991; accepted January 30,1992
597
CONVERSION OF BCECF FLUORESCENCE TO pH,
A.
B.
T
0
<--add
0
acafia odd---
200
400
)(
---add
500
M.b-----1
MKI
1000
time (sec)
c.
'OT
a
1
...............
(--add a d o odd---
)(
----odd
..........
Trb ban----)
2
0
200
400
600
lo00
time (aec)
4
S
a
i
s
pH electrode
0
10
Fig. 1. The calibration procedure for BCECF fluorescence and its
conversion to pH. A Fluorescence of BCECF (0.25 p M ) a t 490 and 439
nm a t various pHs starting in HEPES buffered Tyrode's solution and
after addition of 3 M acetic acid (1 p1added 5 times and 5 p1 added 5
times: final concentration after 10 additions was 30 mM) and 2.5 M
Tris base (pH 10) (5 pl added 10 times and 25 p1 added 8 times: total
final concentration was 250 mM) to titrate the pH from 7.4 to 4.44 and
back to 9.06. B: The ratio of the 490 and 439 nm fluorescence shown in
A. Values for F,,,, 439/Facrd
439), Ldx,
and L,,are given. C: The
conversion of the ratios to pH using equation 3. A dotted line is drawn
across pH 7.4 and passes through the initial pH. D: The correlation of
pH measured by the pH electrode (abscissa) to the pH calculated from
BCECF fluorescence (ordinate) by equation 3, method 1 (@) and by
equation 2, method 2 (01. A dashed line is drawn through the theorctical line for equivalence. Inset: The change in fluorescence ratio of
BCECF by method 1 ( 0 ) and method 2 (0)a s a function of pH dctermined by pH electrode.
EXPERIMENTAL PROCEDURES
The fluorescence of BCECF can be converted to proton activity which is approximately equal to proton
concentration by the formula
molecules. The Henderson-Hasselbalch equation is a
logarithmic transformation of the Michaelis-Menten
equation for the dissociation of protons from acids. In
theory
where K,, the acid dissociation constant, is 107 nM
(pK, = 6.97) (Rink et al., 1982). R is the ratio of the
fluorescence obtained when BCECF is excited at 490
and 439 nm when emission is measured at 530 nm.
R,
is the ratio of these fluorescences when the fluorescence is at its maximum under alkaline conditions.
R,, is the ratio of the fluorescence at pH 4. Fbase439/
Facid439 is the ratio of the fluorescences at 439 nm
under the basic and acidic conditions to obtain R,, and
R,,,. This formula and the related formula for [Ca"],
(Grynkiewiczet al., 1985)are derived from the familiar
Michaelis-Menten equation for the association of two
Experimentally this approach is feasible because R is
fairly constant over two broad ranges (for example, in
our fluorometers, R,,
9 to 10 from pH 8.5 to 9.5 and
k,,,
1to 1.5 frompH 4 to 5) (Fig. lD, inset). The actual
La,
and R,, values may vary depending on the wavelengths chosen and whether the detection of light has
been adjusted to be equal a t the two wavelengths in the
particular fluorometer. (G,,- R)/(R - R,,) is equivalent to [acidl/[basel in the Henderson-Hasselbalch
equation. Since the fluorescence changes of BCECF
at wavelength 439 nm is relatively insensitive to
pH changes (Rink et al., 1982) from pH 4 to 9, the
598
JAMES-KRACKE
Fbae 439/Facid
439 should be approximately equal to 1.
Then as a close approximation for the ratio method:
(3)
The ratio file can be converted to pH using the math
functions within the spectrophotometer software or by
outputting the ratio file to, for example, a spreadsheet
program or program in Basic. However, equation 2 may
be more accurate in fluorometers that do not yield a
439 unless a n “apparent”
value near 1 for Fhase439/FacLd
pK, value is substituted to balance out this term (Graberet al., 1986).
A good approximation of pH, from the fluorescence in
BCECF loaded cells can be determined using this algorithm by treating the cells with a protonophore (FCCP)
(Hladky and Rink, 1982) and titrating the outside pH to
4 and 9. The protonophore is used to promote the flow of
protons across the membrane but complete equilibration of pH, with pH, is not necessary. Extracellular
BCECF was kept to a minimum by washing the cells
just before monitoring (Fig. 2) or by frequent solution
changes in sheets of cells (Fig. 4).The free acid form of
BCECF does not leak out of resealed ghosts at any
appreciable rate (Fig. 3).
Fluorescence measurements
Fluorescence was measured on a Spex (Edison, NJ)
C M l T l l I spectrofluorometer (Fig. 1, method 1, and
Figs. 2, 3, 4)or in a Perkin Elmer MPF-66 (Norwalk,
CT) (Fig. 1, method 2). Similar pH, values are obtained
using either equation 2 or 3 when the light is balanced
equally to and from each monochromator (lamp
aligned, equal slits, and the same photomultiplier used
to detect the emission a t both excitation wavelengths
and each emission level is ratioed to a reference detector to correct for slight variations in lamp intensities
and lamp alignment). Each fluorometer has a stirred
temperature controlled turret. The emission at 530 nm
(excitation 490 and 439 nm) was measured using 5 nm
bandpass slits. The autof luorescence level (including
scatter) of the cells was subtracted before ratios were
calculated.
Calibration solutions
The fluorescence of BCECF in calibration solutions
was approached in two ways. Method 1is demonstrated
in Figure 1. The final results from method 2 were very
similar and they are compared to those obtained by
method 1in Figure 1D.
In method 1,0.25 FM BCECF was monitored in 3 ml
of HEPES buffered Tyrode’s solution containing 136
mM NaC1, 2.6 mM KCI, 1.8 mM CaCl,, 1 mM MgCl,,
0.36 mM NaH,P04, 5.55 mM glucose, and 5 mM
HEPES, pH 7.4. A tiny combination pH microelectrode
(World Precision Instruments, Inc., New Haven, CT)
was fixed in the corner of the cuvette and pH was recorded during a series of additions of 3 M acetic acid (pH
4)and 2.5 M Tris base (pH 10) to titrate the pH from 7.4
to 4 and back to 9. Similar results were obtained by
adding small aliquots of 1 M HC1 and 1 M NaOH in
calibration solutions (method 2) or to cells (Figs. 2,3,4).
In method 2, BCECF (0.25 pM) in 50 ml of a saline
solution containing 145 mM KC1,5 mM NaCl, 0.01 mM
1.4E67
1.2E6-1 .OE6 --
6.OE5--
T
i
FCCP
4.OE5-2.0E5--439nrn
0.0
<--HCI;-)
<---NaOH---)
7
0
300
‘0
300
time ( s e c )
600
900
600
900
1
.-
I
Q
time (sec)
Fig. 2. Calibration of pH in erythroid progenitor cells obtained from
the spleen of mice infected with the murine leukemic Friend virus.
Cells were prepared by the sedimentation method of Sawyer et al.
(1987). BCECF was loaded into the cells by incubating them in a
HEPES buffered balanced salt solution with 1 pM BCECF/am for 1h
at 37°C and washed with the same saline and incubated for another
hour. A: Fluorescence measurements are shown to demonstrate the
calibration procedure in cells. FCCI’ (5 p M ) was added and 12 additions of2 p1 of 1M HCI (total 24 p1) until the fluorescence a t 490 and
439 nm converged. Then 12 additions of5 p1of 1M NaOH (total 60 p1)
made the fluorescences diverge as far as possible, indicating the response of the dye over the pH range from 4 to 9 in cells. B: The
conversion of the fluorescence to pH units by equation 3 indicates the
resting pH level was 7.06. The arrow indicates the addition of approximately 100 nglml of phospholipase A, as a contaminant in a thrombin
preparation from United States Biochemicals (Park et al., 1990). Inset: The ratio of fluorescences from A as a function of the pH determined in B at many points to demonstrate the sigmoidal relationship
between ratio and p H in cells.
EDTA, and 5 mM HEPES, initial pH 6, was warmed to
37°C in a shaker water bath. A pH electrode (Corning
combination EX-L electrode and model 150 ion meter)
on a swinging arm monitored the buffer in the shaker
water bath. The pH of this BCECF solution was titrated
in steps of 0.25 pH units up to pH 8.0 with 1M NaOH,
and the ratio of the fluorescence (4901439 nm) of a 2 ml
aliquot was measured a t each step. A portion of the
remaining sample was titrated to pH 4 with HC1 and
another to pH 9.15 with NaOH and the ratio at each pH
extreme was measured. The ratios were converted to
pH by equation 2.
Calibration in cells
To demonstrate the effectiveness of this approach,
examples of measurements of pH in 3 different types of
599
CONVERSION OF BCECF FLUORESCENCE TO pH,
10
9
8
.1
7
Q
6
5
.
... . . .
FCCP full range calibration
nigericin linear range calibration
4
0
100
200
300
400
500
600
700
800
900
time ( s e c )
Fig, 3. Calibration of pH in red cell ghosts. Human red cell ghosts
were prepared to compare the nigericin calibration method to the
FCCP full range algorithm method. After hypotonic lysis, passive
resealed ghosts were loaded with the Gee acid form of the dye (25 p M
BCECF) and 145 mM KCl before resealing rapidly at 37"C, pH 7.4
(James-Kracke, 1992). Nigericin (10 pM) was added at 60 sec and the
pH, of the medium (7.4)was measured (a pause of at least 5 min occurs
at each of these pH adjustment steps between 0 and 400 sec). Then the
pH was titrated to 7.0,6.5,back to 7.0, and up to 7.5. FCCP (2 pM)was
added at 400 sec and then small aliquots of 0.1 M HC1 were added a t
approximately 15 sec intervals followed by aliquots of 0.1 M KOH
beginning at 620 to 660 sec to calibrate the dye. R,,,, and R,,i,, F,, 43L1/
F,,,, 439 were 9.68 i ,052, 1.04 t ,0057, and 1.13 * ,038, respectively.
The pII was monitored in medium huffered to pH 7.4 with 5 mM
HEPES containing 145 mM KCl and 5 mM NaCl. There are 4 solid
lines which represcnt the conversion of the fluorescence to pH, by the
FCCP full range method and 4 dotted lines for the conversion of the
same 4 runs by the nigericin 3 point calibration method. Inset: The
calibration curve to determine the pK, of BCECF in situ (6.7) assuming that nigericin equilibrates pH,, with pH,.
preparations are given. The calibration procedure is
effective whether the cells are in sheets or suspensions.
The cell preparations for demonstration are erythroid
progenitor cells in suspension (Fig. 2) (Park et al.,
1990), resealed white ghosts prepared from human red
cells in suspension (Fig. 3) (James-Kracke, 1992), or
contractile smooth muscle cells as a thin sheet on a
frame mounted in a cuvette (Fig. 4) (James-Kracke and
Bozoky, 1989). Brief descriptions of the preparations of
these cells are outlined in the figure legends. These
examples demonstrate that BCECF behaves the same
in cells as in calibration solutions whether loaded directly as the free acid as in the case of red cell ghosts or
as the acetoxymethyl ester used for erythroprogenitor
cells o r smooth muscle cells. Note that intermediate dye
forms of BCECFiam are not apparent and R,,
in cells
in calibration solutions unlike the beis equal to R,,
havior of fura2/am in some cells where intermediate
forms persist (Scanlon et al., 1987; Moore et al., 1990).
Also note that the K, of the BCECF determined in cells
(Fig. 3) is usually similar (Liston et al., 1991; Restrepo
et al., 1988) but not precisely equal (Chaillet and Boron, 1985; Boyarsky et al., 1988) to the K, determined
in calibration solutions (Fig. 1)(Boyarsky et al., 1988).
However, the K, value determined in cells is more than
adequate for measurement of relative changes in pH,
(see Discussion).
Materials
BCECF and BCECFiam were purchased from Molecular Probes (Eugene, OR). FCCP which is carbonyl cyanide 4-(trifluoro-methoxy~phenylhydrazonewas purchased from Aldrich Chemical Co. (Milwaukee, WI).
Tris (Trizma base), acetic acid, HCl and NaOH, carbachol, nigericin, DMSO, and all physiological salts were
purchased from Sigma Chemical Co. (St. Louis, MO).
Thrombin was purchased from United States Biochemical Corp. (Cleveland, OH).
RESULTS
Calibration in solutions
Since 439 nm is a n isosbestic point, the fluorescence
a t 439 nm changes very little a s acid and base are added
to drive the pH to 4 and then 9 (Fig. 1A). The fluorescence a t 490 nm asymptotically approaches the fluorescence a t 439 nm as the pH declines t o 4.This, in part,
depends on the calibration of the fluorometer (see Discussion). Below pH 4, the fluorescence a t 490 nm declines below the fluorescence measured at 439 nm due
to titration of additional groups in BCECF with a lower
pK, value than the phenolic group with a pK, a t 6.97
(Rink et al., 1982). During additions of base, the fluorescence a t 490 nm increased to a maximum and then
declined slightly due t o dilution. However, dilution de-
600
JAMES-KRACKE
7.3
7.2
,
r
1
i
I
7.1
._
I
Q
7.0
6.9
I
6.8
-7.6
:
0
:
600
1200
time (sec)
Fig. 4. Calibration of pH in smooth muscle strips. The intestine of a
13-day-old chicken embryo was cut helically and mounted on a plexiglass frame in a cuvette (James-Krackeand Bozoky, 1989). Autofluorescence was measured and the strip was loaded with 1 pM
BCECFiam for 1 h and washed for 1 h in HEPES buffered Tyrode's
solution. Autofluorescence values were subtracted before ratios and
pH were calculated. Carbachol (100 pM) caused alkalinization which
was reversible t o the resting pH of 7.04 upon washing with Tyrode's
solution. The pH began to spontaneously oscillate during this recovery
phase. Depolarization in high K medium also caused alkalinization
which was less sustained. After returning to normal Tyrode's solution,
FCCP (5 pM) caused acidification to pH 6.8. To the 3 ml of Tyrode's
buffer in the cuvette, 60 pI of 1 M HCl and 90 p1 of 1M NaOH were
added to first acidify the pH, to 4 and then alkalinize the pH, to 9,
respectively. The pH range from 6.7 to 7.3 is shown to illustrate the
effect of FCCP on an appropriate scale. R,, and R,,, in this preparation were 9.12 and 0.615.
creased the fluorescence relatively more at 439 nm indicating that the decline in fluorescence at 490 nm due
to dilution was partly offset by increased fluorescence
as the pH was increased. Consequently, the ratio increased to a maximum (Fig. 1B). This demonstrates
how the ratio method cancels many artifacts. Care
should be taken not to exceed pH 9.5 because precipitates of CaPO,, CaC03, or Ca(OH),, may form depending on the composition of the medium. This calibration
approach may not be suitable in bicarbonate buffered
medium.
The ratio of the fluorescence intensities (490/439nm)
approaches 1during titration to pH 4 and increases to a
maximum of 9.1 after additions of base to pH 9 to 9.5
(Fig. 1B). When these ratios were converted to pH by
equation 3, the starting fluorescence in our physiological buffer converted to a pH of 7.4, as would be expected
(dotted line, Fig. 1C). The ratios are a sigmoidal function of pH with the inflection point indicating a pK, of
6.97 (Fig. l D , inset). The pH values calculated by equation 2 or 3 closely conformed to the theoretical line with
a slope of 1 when plotted versus the pH measured by
electrodes (Fig. 1D). The correlation is best over the
physiological range from pH 6 to 8 or 1 log unit above
and below the pK, for BCECF. The calculated values
give good approximations over the entire range, but
errors of 0.2 to 0.4 units may occur near pH 4 and 9.
However, these errors are much smaller than those
obtained using the nigericin method based on the slope
of the linear range (see Fig. 3).
3,4). Starting from neutral pH, the dye is titrated with
acid (acetic acid or HC1) until the fluorescence at 490
nm asymptotically approaches the fluorescence a t 439
nm. Then the pH is titrated back to pH 9 by adding
small aliquots of 1M NaOH or 2.5 M Tris and observing
the increase of the fluorescence at 490 nm until it rises
no further with similar additions of base. By monitoring the fluorescence changes until they reach the extremes, the method is not dependent on the equilibration of pH across the membrane and does not require an
electrode in the cuvette. The dye is monitored to its
maximum and minimum values within the cell even
though the external pH may be more acidic or alkaline.
Because of this, it is not necessary to depolarize the
membrane potential to equilibrate the protons across
the membrane as in the nigericin calibration procedure. It has been shown that nigericin does not completely equilibrate pH, with pH, (Restrepo et al., 1990)
and the difference can be as much as 0.15 pH units.
In cells, the protonophore FCCP (Hladky and Rink,
1982) allows the equilibration of protons across the
membrane according to their concentration gradient
and membrane potential. Although 5 pM FCCP might
be expected t o allow protons to leave because the extracellular [H+] is less than the intracellular [H'], Ht
actually enter due to the negative membrane potential.
In some cells, or after some treatments, little change in
pH, is observed upon addition of FCCP (Figs. 2, 3) indicating that the membrane potential difference is small
or that protons are already at equilibrium across the
membrane. In smooth muscle cells, with a membrane
potential of at least -60 mV, FCCP causes the cells to
acidifv
" (Fin.- 4).When the outside DH is acidified to 4.
protons quickly enter due to the >,000-fold inwardly
Calibration in cells
The in vitro calibration method in Figure 1 was designed to resemble the techniaue used in cells (Figs. 2.
Y
,
CONVERSION OF BCECF FLUORESCENCE TO pH,
directed proton gradient. This influx of cations possibly
depolarizes the cells and hinders H + entry which may
explain why R,,, is usually slightly above 1 (i.e., 1.051.2).
This error (R,,, 1vs. 1.2) is small (typically changing
the calculated pH, value by 0.01-0.04 over the range
pH 7.7 to 6.6, respectively) and should not be compensated by lowering the outside pH further as greater
acidity may damage cells before R,, is determined and
requires more base to be added to reach R,,.
The error
(e.g., as a n extreme example, 9
for variations in R,,
vs. 10) will convert to pH values differing by 0.08 to .38
units for pH 6.5 and 7.5, respectively.
In our experience, i t is better to determine R,,, before R,,
because cells seem to tolerate a n acid stress
better than a n alkaline stress to pH 9 but either way
can be done and others have preferred to determine
R,, first (Eisner et al., 198913).
Cells in suspension can be titrated t o pH 4 with either HCl or acetic acid and to pH 9 with 1 M NaOH
or 2.5 M Tris (pH 10). However, for a sheet of cells on
a frame or coverslip or single cells, acetic acid is
less appropriate since it extracts Ca2+ (e.g., acetic acid
is used to extract Ca for atomic absorption spectrophotometry measurements) and causes some weakening of the cell attachments within the layer and possibly loss of cells from the beam before Rma,can be determined. Adding small amounts of 1M HC1 works just
as well but it is possible to go below pH 4 if the volume
of acid to be added is not determined carefully. Likewise titration with NaOH or KOH (Fig. 2) works a s
well a s Tris but additions of base which raise the pH
above 9.5 should be avoided to prevent tissue disintegration and the formation of Ca precipitates in the medium. However, if the correct volumes of 1M HC1 and 1
M NaOH to adjust the pH of the buffered cell medium to
4 and 9 are determined with a pH meter, then these
volumes can be used as a first approximation when
working with cells. Since the cells themselves add buffering capacity, the exact amount of acid and base
should be determined in cell samples after adding a
protonophore. While adding acid and then base in small
aliquots, the fluorescence at 490 nm is observed to
approach and then diverge from the fluorescence at
439 nm, respectively (Figs. 2, 3). This should be done
fairly quickly (it is possible to do i t in under 60 sec) to
avoid leakage of the protonated carboxyfluorescein
dyes which occurs over 1.5-2 h (Barbet et al., 1984).
Once the correct amount of acid and base are known,
then these can be added in 2 steps a t the end of each
measurement of cells, to determine &,
R,,,, and
439. As a n example, if these ratio values
Fbese 439/Facld
are typically and consistently 9 to 10, 1 and 1, respectively, (Figs. 3, 4)this ensures that the dye has been
effectively deesterified in the cells and that no quenching or aberrant fluorescent substances have been
Kln,and
added. Since we have observed that R,,,
Fbase439/Facld
439 are fairly constant in our fluorometers
in many different types of cells, a n approximate calculation of pH, by the algorithm (equation 3) can be
done without necessarily calibrating the dye in each
sample. However, since this procedure takes very little
time, we routinely check the responsiveness of the
60 1
dye and the effect of experimental agents in each
sample.
DISCUSSION
The calibration of pH in cells loaded with BCECF can
be determined by a formula derived from the Henderson-Hasselbalch equation using minimum and maximum fluorescence ratios a t pH 4 and 9. This conversion
to pH can be checked in every cell sample quickly, after
the cells have been tested for their biological responsiveness, by adding the protonophore FCCP or CCCP
and then HC1 followed by NaOH. Graber et al. (1986)
reviewed the standard method for BCECF calibration
which uses nigericin in high K' depolarizing medium
with adjustment of the pH to three values from 6.5 to 8
(Thomas et al., 1979). Because nigericin exchanges K +
for H+, the extracellular and intracellular K' concentrations must also be set equal at the expense of decreasing the extracellular Na' concentration. The consequent depolarization and reversal of the Na+
gradient adds uncertainty about whether the intracellular and extracellular pHs have equilibrated. In some
cells, equilibration with nigericin can take up to 30 min
(Negulescu and Machen, 1990) and therefore, the calibration procedure can be very time consuming (Eisner
et al., 1989a).
The ionophore problem was circumvented in A431
cells treated with ouabain for 2 h since this causes the
Na+/H+antiporter to be the only mediator of pH equilibration across the plasma membrane (Rothenberg
et al., 1983; Eidelman and Cabantchik, 1989). This approach may be appropriate in other cells with a Na+/H+
exchanger (Eidelman and Cabantchik, 1989).However,
the hours of treatment by ouabain makes this approach
less than ideal for calibration of each aliquot of cells
unless Na+ pump inhibition or Na+ loading of cells
happens to be part of the experimental protocol as in
the original study of this type (Rothenberg et al., 1983).
Another problem created by the nigericin approach
in high K + medium is hypercontracture of muscle cells.
A calibration procedure using pyruvate and ATP in
high K C medium with nigericin, ionomycin, EDTA,
and FCCP to equilibrate the protons across the membrane was used to avoid hypercontracture while extracellular pH was titrated with HC1 or KOH to 6.54,7.27,
and 7.75 (Borzak et al., 1990). Hypercontracture does
not appear to be a problem using our technique for
smooth muscle sheets since high potassium medium is
not used, and at pH 4,Ca binding to contractile filaments would be limited. EGTA could be added after
addition of NaOH to prevent contracture in the alkaline range.
Another approach for absolute calibration of fluorescent pH, indicators is a modification of a method used to
calibrate intracellular pH electrodes (Szatkowski and
Thomas, 1985).For this null point method, a weak acid
(butyrate) and base (trimethylamine) are added in concentrations determined to cause equal deflection of fluorescence in the acidic and alkaline direction. This not
only cirvumvents the problems of the nigericin method
but allows determination of the buffering capacity of
the cells (Eisner et al., 1989a).However. i t is limited to
the linear range of the dye. This method, while very
602
JAMES-KRACKE
accurate, seems more complicated to perform and calculate routinely on each cell sample.
Within the last year, there has been at least one
study in which the investigators have successfully
studied pH, using BCECF in the nonlinear range from
pH 5 to 6-for example, to determine that the pH for
half maximal stimulation of NafiH+ exchange is 5.87
in promyelocytic leukemia (HL-60) cells (Restrepo
et al., 1990). The nigericin calibration method was extended into the nonlinear range by measuring pH at
many steps with an electrode. The fluorescence in this
range was fitted by an equation for the single wavelength method that essentially is a rearrangement of
equation 1. However, this equation was not used to
calculate pH, in the cells. The calibration relied, in
part, on subsequent release of the dye by digitonin after
the nigericin calibration. This approach is appropriate
for cells in suspension but cannot be used for single
cells or sheets of cells.
Recently a similar equation (with rearrangement of
terms) was used by Graber et al. (1986) to determine
the pK, of BCECF in calibration solutions. They omitted the term log Fbase439/Facld
439 in order to obtain
greater agreement between the calculated pK, and the
visual estimate of pK, from graphs of the ratio of the
fluorescence versus pH. Paradiso et al. (1987) omitted
this term because, at the isosbestic wavelength, the
need for this term is obviated (Eisner et al., 198913).
Most calibrations of pH, in cells have been done with
the nigericin procedure in high K medium over the
linear range (Graber et al., 1986) while in calibration
solutions, this algorithm (similar to eqn 2 or 3) has been
used to determine the pK, over the entire range of pH
sensitivities. However, this algorithm has been used in
studies of gastric oxyntic cells (Paradiso et al., 19871,
isolated cardiac myocytes (Eisner et al., 1989131, and
smooth muscle cells from the bladder (Liston et al.,
1991), respectively. Solutions containing 140 mM KC1
and 10 pM nigericin at pH 9.2, 4.3, and 7.0 were perfused over single cells attached to a support in a fluorescence ratio microscope system and pH, was calculated by an equation similar to equation 3 (using 439
nm as an isosbestic point) or equation 2 (using 430 nm
for the pH insensitive wavelength). Bright et al. (1989)
published another formula for converting fluorescence
to pH, (eqn 4b in their review) which requires knowing
the slope of the linear portion of the curve. In our hands,
the data in Figure 3 were not converted to logical values by their formula, nor was this method as quick.
Each step of the nigericin calibration procedure usually requires 5 min for a total of 15 rnin as nigericin is
slow to equilibrate pH, and pH, (Eisner et al., 1989a)
and may take up to 30 min at each step in some cells
(Negulescu and Machen, 1990). Changing to high K
medium also requires solution changes, which can be
done easily in attached cells but risks the loss of cells in
suspensions after centrifugation and causes contraction of muscle cells. Our method avoids these two problems and is quicker (less than 5 min per cell sample)
and just as accurate without requiring extracellular pH
electrodes. For a 3 point pH calibration curve on each
tissue sample, equilibration at each pH can be time
consuming. For this reason it has been reported that a 1
point calibration in the linear range is sufficient when
the data is normalized to a ratio of 1at pH 7 (Boyarsky
et al., 1988). But in the nonlinear range, the time for
calibration at many points per cell sample restricts the
number of experimental factors that can be tested on
the cell aliquots before their viability declines. In any
case, the accuracy is dependent upon the assumption
that nigericin equilibrates pH, with pH,.
Calibration by the method described here by the full
range FCCP calibration can be conveniently checked in
each aliquot of cells rather than in a few representative
samples. There is no need to equilibrate the membrane
potential or monitor the pH with an electrode simultaneously because the intracellular dye fluorescence can
be monitored to observe the diminishing changes in
fluorescence as acid and base are added as the endpoints (R,,,, and R,,) are approached. This procedure
is equivalent t o adding ionomycin and saturating concentrations of Ca2+ and EGTA under alkaline condiand R,,,.
tions in fura2 loaded cells to determine R,,
This calibration method helps to determine that the
phenolic oxygen of BCECF has been deesterified but
does not indicate whether the carboxylic acids are deesterified and it is the latter which may make the dye not
leak. Intermediate forms of BCECFiam which remain
in some cells with low esterase activity are best detected by checking for dye leakage in the extracellular
fluid. However, intermediates are less of a problem for
BCECF because BCECFiam itself is not fluorescent
(Kaunitz, 1988) and it seems to be a better substrate for
intracellular esterases than fura2/am (Bright et al.,
1989).For example, in smooth muscle cells, the R,
for
BCECF is equal to R,
in calibration experiments,
whereas R,,,,, for fura2 in these cells is not equal t o the
R,, in calibration curves of the free acid.
The choice of equation 3, rather than equation 2,
should be made cautiously after reviewing the simplifying assumptions. For most dual excitation fluorometers
with calibrated monochromators, aligned light path,
and equivalent sensitivity of light detection for both
wavelengths, equation 3 should yield results equivalent to equation 2. Otherwise the fluorescence intensity
at 490 nm will approach the intensity at 439 nm a t pH 4
but not asymptotically. Also one must have a preparation of cells in which extracellular BCECF is negligible.
This approach also assumes that the pK, of BCECF is
equivalent in cells and calibration solutions. This point
has been questioned. Some investigators report that
the apparent pK, in situ is 0.2 units above or below the
pK, determined in vitro and they have attributed these
differences to changes in the dye properties induced by
cytosolic factors (Boyarsky et al., 1988; Negulescu and
Machen, 1990). But this explanation is unlikely in
white ghosts where the cytosol was replaced by calibration-like solutions. Also the difference is not constant at
all pHs but greatest at pH 6.5. Therefore, at least in red
cell ghost in HEPES buffered medium, nigericin in
high K medium probably does not equilibrate pH, with
pH,. But if the pK, was found to be different in the
cytosol of a particular cell, that value could be substituted in equation 2 or 3. Our approach also assumes
that FCCP induces a transmembrane flow of protons
large enough to drive the pH, to values suitable to determine R,
and R,,,. Although this is true for the examples shown here, it may not be true in all cell types.
CONVERSION OF BCECF FLUORESCENCE TO pH,
One must be cautious when applying this approach to
sheets of cells. It is best not to overshoot pH 4 or pH 9,
since the tissue will begin to disintegrate and the dye
will be lost from the sheet of cells. Figure 4 demonstrates that it is possible to use this approach without
losing dye from the cells in the sheet. However, titrating the pH to 4 will decrease the negative charge of the
carboxyl groups and make the dye more permeant.
Acidification was shown to release carboxyfluorescein
from liposomes at pH 4 after 1.5-2 h (Barbet et al.,
1984). The acidification to pH 4 should be kept a s brief
as possible (typically 1-2 min needed at most). Over
this time frame, dye leakage from the sheets of cells
are fairly conwas not detectable. Since Rmin and R,,
stant over a broad range, the pH need not be adjusted
precisely to 4 and 9.
In conclusion, the ratio of BCECF fluorescence in
cells is a sigmoidal function of pH. R,, and Rmincan be
determined by adding a protonophore like FCCP and
adding acid and base in succession to drive the pH to 4
and then 9. The algorithm described for converting
BCECF fluorescence to pH is very accurate over the
linear central portion of the curve from pH 6 to 8 but
also allows approximate calibration of BCECF over the
entire range from pH 4 to 9.
ACKNOWLEDGMENTS
This project was begun at S.U.N.Y. Health Science
Center in Syracuse and continued a t UMC. The technical assistance of J. Chai, I. Bozoky, and P.A. Eichen is
appreciated. Dr. D.J. Park prepared the erythroprogenitor cells. This project was supported by the New York
Affiliates of the American Heart AssociationiMidHudson and Finger Lakes regions, the National American Heart Association, the NIH (AR35435), and the
Research Foundation of the State University of New
York.
LITERATURE CITED
Barbet, J.: Machy, P.: Truneh, A. and Leserman, L.E. (1984) Weak
acid-induced release of liposome-encapsulated carboxyfluoroescein.
Biochim. Biophys. Acta, 772t347-356.
Borzak, S.,
Kelly, R.A., Kramer, B.K., Matoba, Y., Marsh, J.D., and
Reers, M. (1990) In situ calibration of fura-2 and BCECF fluorescence in adult rat ventricular myocytes. Am. J.Physiol., 259tH973H981.
Boyarsky, G., Ganz, M.B., Sterzel, R.B., and Boron, W.F. (1988) pH
regulation in single glomerular mesangial cells. I. Acid extrusion in
absence and presence of HCO,?-. Am. J. Physiol., 255:C84PC856.
Bright, G.R., Fisher, G.W.; Rogowska, J., and Taylor, D.L. (1989)
Fluorescence ratio imaging microscopy. Methods Cell Biol., 30:
157-192.
Chaillet, J.R., and Boron, W.P. (1985) Intracellular calibration of a
pH-sensitive dye in isolated perfused salamander proximal tubules.
J. Gen. Physiol., 86t765-794.
Eidelman, O., and Cabantchik, Z.I. (1989) Continuous monitoring of
transoort bv fluorescence on cells and vesicles. Biochim. Bionhvs.
Acta,'988:3i9334.
Eisner, D.A., Kenning, N.A., O'Neill, S.C., Pocock, G., Richards, C.D.,
and Valdeolmillos,-M. (1989a) A novel method for absolute calibration of intracellular pH indicators. Pflugers Arch., 413553458.
Eisner, D.A., Nichols, C.G., ONeill, S.C., Smith, G.L., and Vadeolmillos, M.V. (1989h) The effects of metabolic inhibition on intracellular
calcium and pH in isolated rat ventricular cells. J. Physiol.,
411:393%418.
Graber, M.L., DiLillo, D.C., Friedman, B.L., and Pastoriza-Munoz, E.
(1986) Characteristics of fluoroprobes for measuring intracellular
pH. Anal. Biochem., 156t202-212.
I
"
603
Grinstein, S., and Cohen, S. (1987) Cytoplasmic [Ca''] and intracellular pH in lymphocytes: Role of membrane potential and volumeactivated Na'iH' exchange. J. Gen. Physiol., 89t185-213.
Grinstein, S., and Goetz, J.D. (1985) Control of free cytoplasmic calcium by intracellular pH in rat lymphocytes. Biochim. Biophys.
Acta, 819:267-270.
Grynkiewicz, G., Poenie, M., and Tsien, R.Y. (1985) A new generation
of Ca2 indicators with greatly improved fluorescence properties. J.
Biol. Chem., 260:3440-3450.
Hladky, S.B., and Rink, T.J. (1982) The use of ion transporters, pH
measurements and light scattering with red blood cells. In: Red Cell
Membranes: A Methodological Approach. J.C. Ellory and J.D.
Young, eds. Academic Press, London, pp. 335357.
James-Kracke, M.R. (19Y2) Calmodulin activation of the Ca2' pump
revealed by fluorescent chclator dyes in human red blood cell ghost.
J. Gen. Physiol., 99t41-62.
James-Kracke, M.R., and Bonoky, I. (1989) Development of oscillating
contractions and acetylcholine receptors in chicken embryonic
smooth muscle. Physiologist, 32:148.
Kaunitz, J.D. (1988) Preparation and characterization of viable epithelial cells from rabbit distal colon. Am. J. Physiol., 254tG50S
G512.
Liston,T.G.,Palfrey,E.L.H.,Reimhach,S.J.,andFry,C.H.
(1991)The
effects of pH changes on human and ferret detrusor muscle function.
J . Physiol., 432:l-21.
Moolenaar, W.H., Tsien, R.Y., van der Saag, P.T., and de Laat, S.W.
(1983)Na+iH+exchange and cytoplasmic pH in the action of growth
factors in human fibroblasts. Nature, 3043345-648,
Moore, E.D., Becker, P.L., Fogarty, K.E., Williams, D.A., and Fay, F.S.
(1990)Ca2+imaging in single living cells: Theoretical and practical
issues. Cell Calcium, 11:157-179.
Negulescu, P.A., and Machen, T.E. (1990) Intracellular ion activities
and membrane transport in parietal cells measured with fluorescent dyes. Methods Enzymol., 192:38-81.
Parddiso, A.M., Tsien, R.Y., and Machen, T.E. (1987) Digital image
processing of intracellular pH in gastric oxyntic and chief cells.
Nature. 325t447-450.
Park, D.J., James-Kracke, M., and Kim, H.D. (1990) Concomitant
changes in cytoplasmic calcium and intracellular pH induced by
thrombin in erythroprogenitor cells. Blood. 76:112a.
Restrepo, D., Kozody, D.J., Spinelli, L.J., and Knauf, P.A. (1988) pH
homeostasis in promyelocytic leukemic HL60 cells. J. Gen. Physiol.,
92t489-507.
Restrepo, D., Cho, D.S., and Kron, M.J. (1990) Essential activation of
N a ' / H + exchange by lH+li in HL-60 cells. Am. J . Physiol., 259:
C490-C502.
Rink, T.J., Tsien, R.Y., and Pozzan, T. (1982)Cytoplasmic pH and free
Mg2+in lymphocytes. J. Cell Biol., 95t189-196.
Rothenberg, P., Glaser, L., Schlesinger, P., and Cassel, D. (1983) Activation of Na'/H+ exchange by epidermal growth factor elevates
intracellular pH in A431 cells. J. Biol. Chem., 258t12644-12653.
Sawyer, S.T., Koury, M.J., and Bondurant, M.C. (1987) Large-scale
procurement of erythropoictin-responsiveerythroid cells: Assay for
biological activity of erythropoietin. Methods Enzymol., 147:340353.
Scanlon, M., Williams, D.A., and Fay, F.S. (1987) A Ca2'-insensitive
form of fura-2 associated with polymorphonuclear leukocytes: Assessment and accurate Ca"+ measurement, J . Biol. Chem.,
262r6308-6312.
Szatkowski, M.S., andThomas, R.C. (1985) Calculation of steady-state
pH, from pH, changes caused by weak acids and hases in snail
neurones. J. Physiol., 371t153P.
Thomas, J.A., Buchsbaum, R.N., Zimniak, A., and Racker, E. (1979)
Intracellular pH measurements in Ehrlich Ascites tumor cells utilizing spectroscopic probes generated in situ. Biochemistry,
18:2210-2218.
Tsien, R.Y. (1981) A non-disruptive technique for loading calcium
buffers and indicators into cells. Nature, 290527-528.
Tsien, R.Y., and Poenie, M. (1986) Fluorescence ratio imaging: A new
window into intracellular ionic signaling. Trends Biochem. Sci.,
11:450455.
Tsien, R.Y., Pozzan, T., and Rink, T.J. (1982) Calcium homeostasis in
intact lymphocytes; cytoplasmic free calcium monitored with a new
intracellularly trapped fluorescent indicator. J. Cell Biol., 94t325334.
Wray, S. (1988) Smooth muscle intracellular pH: Measurement, regulation, and function. Am. J. Physiol., 254tC213-CZ25.
+