A Ca2’-insensitive Form of Fura-2 Associated with Polymorphonuclear Leukocytes ASSESSMENT AND ACCURATE Ca2+ MEASUREMENT*

The new, fluorescent Ca2+ indicator, fura-2, prom- ises to expand our understanding of the role of subcellular changes in Ca2+ underlying cell function. During an investigation of the role of Ca2+ in the polarization response of human polymorphonuclear leukocytes to formyl-methionyl-leucyl-phenylalanine, we found that fura-2 trapped by cells incubated with the acetoxy-methyl ester of fura-2, F2-AM, yielded measurements of Ca2+ that were depressed at rest and during the response to formyl-methionyl-leucyl-phenylalanine. Fura-2, trapped by the cells, exhibited a spectrum in the presence of saturating Ca2+ that differed from that of fura-2 free acid. We have shown that the cellular fluorescence can be spectrally decomposed into two components: one with Ca2+ sensitivity identical to fully deesterified fura-2, and another which is Ca2+-’ msen- sitive. The Ca2+-insensitive component appears to be more fluorescent than F2-AM as well as spectrally different from F2-AM. The insensitive form probably results from incomplete deesterification of F2-AM by the cells. In order to accurately measure Ca2+ in polymorphonuclear leukocytes, it is imperative to check for the presence of Ca2+-insensitive fluorescence. The contribution of Ca2+-insensitive fura-2 fluorescence can be assessed routinely from spectral data obtained

* This work was supported in part by National Institutes of Health Grant HL 14523 and by a grant from the Muscular Dystrophy Association. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ Supported by an American Heart Association (Massachusetts Affiliate) Fellowship.
To whom correspondence should be addressed Dept. of Physiology, University of Massachussetts Medical School, 55 Lake Ave., N., Worcester, MA 01605. in several cell types to examine local Ca2+ changes in small regions of the cell in a nonperturbing manner. Ca2+ changes following contractile activation of smooth muscle (3) and during mitosis of Pt K2 epithelial cells (4) are among processes successfully examined in detail using fura-2. On the other hand, there have been reports in the literature of methodological difficulty associated with the use of fura-2, suggesting that, in certain cell types, assumptions inherent in the use of fura-2 are not valid. For instance, some cells appear to be deficient in the esterase required to cleave F2-AM and, as a consequence, microinjection of these cell types with fura-2 free acid is necessary to obtain a reliable Ca2+ signal (5-7). In other cell types, the fluorescent response of cell-associated fura-2 to Ca2+ appears to be different from that of free acid, requiring calibration of cellular fluorescence with high viscosity gelatin solutions in order to mimic the intracellular environment (2). A recent report proposes that a modified form of F2-AM may become partially bound to isolated vesicles of skeletal muscle (8). Prior to this, we had reported in preliminary form on the presence of a Ca2+-insensitive yet highly fluorescent form of fura-2 associated with polymorphonuclear leukocytes (PMNs)' incubated with F2-AM (9,10).
In this communication, we propose that the identity of the Ca2+-insensitive fluorescence associated with PMNs represents intermediates of the deesterification of F2-AM. The presence of this fluorescence may also account for problems associated with the use of fura-2 in other cell types. Since the presence of the insensitive form of fura-2 precludes the use of the standard procedure for calculation of [Ca"'] (l), which assumes only two intracellular forms of fura-2, we suggest a method which allows for accurate measurement of [Ca2+] in all cells where intermediates of the intracellular deesterification of F2-AM, or F2-AM itself, are present. This method involves in situ calibration of cell-associated fura-2 by permeabilization of the cells to Ca2+ using the Ca2+ ionophore, ionomycin, with the assumption that the concentration of ionomycin is sufficient to equilibrate intracellular and extracellular [Ca2+]. The calibration step not only checks for the presence of Ca2+-insensitive fura-2 and leads to correct Ca2+ measurement in spite of the insensitive dye, but it also provides a means for characterizing the insensitive moiety.

MATERIALS AND METHODS
Preparation of Human Polymorphonuclear Leukocytes-Whole blood was obtained by venipuncture of healthy donors and sedimented for 30 min over 6% dextran (Pharmacia P-L Biochemicals). The The abbreviations used are: PMNs, polymorphonuclear leukocytes; fMLP, formyl-methionyl-leucyl-phenylalanine; HBSS, Hanks' balanced salt solution; EGTA, [ethylenebis(oxyethylenenitrilo)]tetraacetic acid. white cell-enriched supernatant was centrifuged at 1200 rpm for 15 min in a Sorvall RC2-B centrifuge at 4 "C. The pellet was resuspended in a small volume of Hanks' balanced salt solution (GIBCO); remaining red cells were removed by hypotonic lysis in 9.5 ml of distilled water for 30 s, followed by addition of 0.5 ml of 18% NaCl to restore normal osmolarity. After centrifugation at 1200 rpm for 10 min at 4 "C, the pellet was resuspended in Hanks' balanced salt solution containing 1% Knox gelatin (HBSS-gel), pH 7.0-7.2. The preparation contained between lo6 and lo7 PMNs/ml. The cells were kept at room temperature during the course of the experiment and remained viable for 6 h after collection.
Fura-2 Loading-Fura-2 and its acetoxy-methyl ester, F2-AM, were obtained from Molecular Probes, Eugene, OR. PMNs were loaded with fura-2 by addition of 1 FM F2-AM to the cell suspension (range, 2-6 x lo6 cells/ml) for 1 h in a 37 "C incubator, 90% 02, 10% CO,. After incubation, the cells were centrifuged at 1200 rpm for 10 min to remove extracellular dye. Cells were postincubated at 37 "C for at least 30 min. The intracellular fura-2 concentration was calculated according to the method of Williams and Fay (17). The cell volume was taken as 5.25 X liters/cell. Cell density was determined with a hemocytometer and ranged from 2 X lo6 to 6 X 10' cells/ml. Spectrofluorimetric Studies-Spectrofluorimetric experiments were conducted on a Perkin-Elmer MPF-3 spectrofluorimeter in the laboratory of Dr. David Wolf, Worcester Foundation for Experimental Biology, Shrewsbury, MA, or on a Spex CM-1 spectrofluorimeter operated in ratio mode with rhodamine B quantum counters. Slitwidths were 4 nm in the Perkin-Elmer spectrofluorimeter and 11 nm in the Spex CM-1. Cells were maintained at 37 "C by a circulating water bath and continually stirred by a small magnetic stirrer to ensure rapid and complete dispersal of agonists. Fluorescence was corrected for volume dilution after addition of agonist. Since autofluorescence represents approximately 10% of the total fluorescence of cells loaded with F2-AM and since it is slightly greater at 340 nm than at 380 nm, the fluorescence of unloaded cells was subtracted from an equivalent density of cells loaded with fura-2 to obtain fluorescent signals that were solely representative of intracellular fura-2. The autofluorescence of PMNs was virtually unchanged upon addition of fMLP and after all manipulations. Thus, in all experiments, the fluorescence of untreated and unloaded cells was used for subtraction of autofluorescence. Excitation spectra were collected at 5-nm intervals from 300 to 450 nm; emission was fixed at 510 nm. For a general review of the principles and techniques of fluorescence, see Ref. 16.
Single Cell Studies-After the loading procedure, some cells were attached to a glass coverslip in a 37 "C incubator for 15 min. The coverslip was then positioned in a Zigmond chamber (11) containing HBSS-gel in one well and lo-' M formyl-methionyl-leucyl-phenylalanine (fMLP) (Sigma) in the other. Measurements of resting [Ca'+] were made in the absence of a gradient of fMLP by placing HBSSgel in both wells of the chamber. The fluorescence of PMNs loaded with fura-2 was quantified using digital imaging microscopy (12).
Procedure for [Caz+/ Calculation-The standard procedure (1) for calculating [Ca"] from dual wavelength measurements of fura-2 fluorescence involves the use of the following equation.
[Ca"+l = -Rmin)/(Rrnax -I?)].@ R is the ratio of fluorescence of the sample at 340 and 380 nm; RmaX and Rmi. represent the ratios for fura-2 free acid at the same wavelengths in the presence of saturating Ca2+ and in nominally zero Ca2+, respectively. The limiting ratios of fura-2, Rmi,, and R,,,, are determined for a sample of fura-2 free acid introduced into the spectrofluorimeter or on the stage of the fluorescent microscope under the assumption that intracellular fura-2 possesses properties identical to fura-2 in solution. @ is the ratio of fluorescence of fura-2 at 380 nm in zero and saturating Ca2+; Kd is the dissociation constant of fura-2 for Ca2+, assumed to be 224 nM at 37 "C. An important assumption in this procedure for [Ca*+] calculation is that there are only two forms of fura-2 associated with the cell: fura-2 bound to and free of Ca2+ (1).

RESULTS AND DISCUSSION
As a first step in using the new Ca2+-sensitive dye fura-2 to determine subcellular changes in [Ca"] underlying the polarization of PMNs to fMLP, we investigated the ability of fura-2 to detect changes in [Ca'+] in a population of PMNs in suspension. Cells were incubated with F2-AM as described under "Materials and Methods," washed, and postincubated at room temperature for 30 min. The postincubation step was included with the assumption, later shown not to be valid, that it would facilitate complete cleavage of all acetoxy-methyl groups. PMNs loaded with fura-2 were then exposed to lo-' M fMLP, a concentration of chemoattractant known to yield maximal polarization (11). The time course of the [Ca'+] change, calculated using the standard procedure for ratiometric [Ca'+] measurements (I), is shown for a typical experiment in Fig. 1. The magnitude and the time course of the polarizing effect of fMLP (not shown) was similar to that reported by other investigators (13). In contrast, although increases in [Ca"] in response to fMLP were apparent from the change in cellular fluorescence of fura-2, the resting level of [Ca2+], as well as the magnitude and duration of the Ca2+ transient, were both depressed relative to values reported using the Ca2+ indicator quin2 (14, 15). [Ca"] rose from 35 to 75 nM within 10 s of addition of fMLP and returned to prestimulus levels within 20 s; control cells maintained a basal [Ca"] of 50 nM. These differences were unexpected as the cellular Ca2+ buffering directly added by fura-2 (125 p~) should be much less than that contributed by the millimolar intracellular concentrations of quin2 generally used (14,15) and, as a consequence, less attenuation of Ca2+ transients would be expected.
There are two possible explanations for the diminished Ca2+ transient observed in Fig. 1. The first is that an assumption inherent in the use of fura-2 fluorescence to calculate [Ca'+] is erroneous. Alternatively, the Ca2+ transient might appear to be attenuated if the response to fMLP of individual cells within the suspension was asynchronous. Fluorescence measurements, using digital imaging microscopy, of single cells loaded with fura-2 suggested that the latter possibility is unlikely as the mean It is necessary, therefore, to consider the possibility that, under these circumstances, in PMNs loaded with fura-2, at least one assumption underlying the traditional, ratiometric measurement of [Ca2+] using dual excitation wavelengths (340 and 380 nm) is in error (see "Materials and Methods"). The central premise in the ratiometric method is that the fluorescence of the cell represents the weighted average of only two fluorescent species within the cell, Ca2+-bound fura-2 (F2-

Ca2+-insensitive Form of Fura-2 in PMNs
Ca") and unbound fura-2 (F2), which possess properties identical to fura-2 free acid in solution (Fig. 2 A ) . There were several observations made during the course of experimentation which raised questions regarding this assumption. We found that experimental values of the ratio of fluorescence a t 340 and 380 nm ( R ) were occasionally less than Rmin, the predicted value for fura-2 in Ca2+-free solution, and led to calculation of apparent negative [Ca"] values. Moreover, after addition of 5 p~ ionomycin, cellular values of R increased but not to R, , . , the predicted value of fura-2 in Ca2+saturating solution. Since ionomycin is a potent and selective Ca2+ ionophore, the cellular Ca2+ gradient should be dissi- Caz+-free solution was obtained through the addition of 3.6 mM EGTA (pH 8.5) to the normal Ca2+-containing medium (1.8 mM Ca") bathing the cells (final free [Ca"] less than 5 nM). Although cellular data were corrected for autofluorescence by the subtraction of the fluorescence of an equivalent density of unloaded PMNs, it is apparent that cellular fura-2 fluorescence is not spectrally identical to that of fura-2 free acid in solution. There is apparently a fluorescent species present other than Ca2+-bound (F2-Ca") and Ca2+-free (F2) fura-2. C, difference spectra, obtained by the subtraction of the Ca*+-free from the Ca2+-saturated spectrum of fura-2 in solution (from A ) and in PMNs (from B ) , indicate that the Caz+-sensitive characteristics of intracellular fura-2 (A) are remarkably similar to those of fura-2 free acid (0). pated, allowing for the saturation of intracellular fura-2 with Ca2+; however, R after ionomycin addition to cells in suspension was always considerably lower than RmaX ( R = 5.06 & .45, n = 6; RmaX = 11.18 +. .82, n = 6). Total cell lysis by Triton X-100 in the presence of 1.8 mM Ca2+ resulted in R values which still remained lower than R,,, ( R = 5.11 & 0.98, n = 4; R,,, = 12.36 +-2.36, n = 4); however, values of R after Triton X-100 lysis were not significantly different from values obtained after ionomycin treatment, suggesting that ionomycin had effectively equilibrated intracellular and extracellular Ca2+, thereby fully saturating fura-2 with Ca2+. Furthermore, the similarity between R values obtained after ionomycin permeabilization and Triton X-100 lysis suggests that the intracellular environment has little effect on the behavior of cell-associated fura-2.
The differences between experimental and predicted R values for the Ca2+-saturating condition indicate that PMNs incubated with F2-AM contain a fluorescent species other than fura-2 acid (F2 or F2-Ca2+). The species may represent additional Ca2+-sensitive forms with spectral characteristics that differ from those of free acid or, alternatively, F2-AM or a form of F2-AM that has been converted into a moiety that is also Ca2+-insensitive. These possibilities were investigated by determining the spectral characteristics of the Ca2+-sensitive species associated with PMNs. This process involved subtraction of the excitation spectrum of ionomycin-treated cells (Fig. 2B) obtained in the absence of Ca2+ from that in the presence of saturating Ca2+. This difference spectrum for ionomycin-treated cells is compared to that of fura-2 free acid in solution in Fig. 2C. The two curves are nearly identical, suggesting that the only Ca2+-sensitive species associated with PMNs is free acid. Although, the Ca2+-sensitive spectra are extremely similar, the excitation spectra of cell-associated fura-2 and free acid in high and low Ca2+ environments are clearly different (see Fig. 2, A and B). In the presence of saturating Ca2+, there is higher fluorescence at 380 nm associated with the ionomycin-treated cells than with free acid. The increased fluorescence intensity at 380 nm is consonant with a value of R (after ionomycin treatment) that is dimin-

ished in comparison with
Rmax. Since the Ca2+-difference spectra indicate that the Ca2+-sensitive component associated with the PMNs appears to be spectrally identical to fura-2 free acid, the observed difference between the excitation spectra of ionomycin-treated cells and free acid must be due to an additional fluorescent component that is Ca2+-insensitive. The spectral characteristics of the Ca2+-insensitive form of fura-2 were dissected from the spectra of ionomycin-treated cells obtained in the presence of Ca2+-saturating and Ca2+free solutions (see Fig. 2B The value of z is determined from the fluorescences of-a known amount of free acid in solution, in Ca'+-free and Ca2+saturating buffers. The spectral characteristics of the Ca"-insensitive species are plotted in Fig. 3A, along with the excitation spectrum of F2-AM, the starting material used to load the PMNs. It is clear that the Ca2+-insensitive fluorescence trapped by the cells is not F2-AM but rather a form of fura-2 which is more fluorescent than F2-AM. Furthermore, the Ca2+-insensitive species associated with the cells exhibits a broad excitation spectrum from 300 to 430 nm, clearly different from F2-AM whose spectral characteristics exhibit a peak a t 380 nm with relatively little contribution below 340 nm. We considered the possibility that the intracellular environment changes the spectral characteristics of F2-AM, leading to the spectrum displayed by the unknown component in Fig. 3A, but found that environmental effects were not the cause of the observed spectrum. Specifically, addition of F2-AM to unloaded cells treated with ionomycin resulted in spectral characteristics which were indistinguishable from those of F2-AM in solution. In Fig. 3B, the spectral characteristics of the Ca2+insensitive fluorescence of cells from another experiment, loaded identically to those exhibited in Fig. 3A, are shown. The Ca"-difference spectra after ionomycin treatment (not shown) indicate that the Ca"-sensitive species is identical to fura-2 free acid. The Ca2+-insensitive species in this experiment is different from that illustrated in Fig. 3A as well as from F2-AM. In this experiment, the unknown species associated with the cells exhibits at least two peaks of fluorescence: one at 380 nm and the other at 430 nm. The difference between the spectral characteristics of the unknown species illustrated in Fig. 3, A and B suggests that, although in both cases the unknown species is not F2-AM, the spectral characteristics of the Ca2+-insensitive moiety varies among PMNs from different donors in spite of the fact that the incubation conditions are identical in all experiments.
We believe that the Ca'+-insensitive species associated with PMNs is incompletely deesterified F2-AM.
We base this hypothesis on the results of the in vitro hydrolysis of F2-AM by bovine esterase. The intermediates of this hydrolysis show progressive increases in fluorescence intensity which precede the acquisition of Ca2+ sensitivity. Ca"-difference spectra, obtained at various times during the hydrolysis, indicate that the Ca2+-sensitive species is always identical to fura-2 and that the absolute concentration of this species appears to increase progressively with time. In addition, spectra of the Ca*+-insensitive intermediates, at the same time points, exhibit maxima at 400 nm, generally similar to the in vivo situation shown in Fig. 3B.
In our experiments, there appear to be at least two causes for incomplete deesterification of F2-AM. In many cases, the problem is specific to cell type. In quiescent BALB/c 3T3 cells, no significant deesterification takes place (6); whereas, in PMNs, deesterification does not always continue to completion. We suggest that quiescent BALB/c 3T3 cells do not possess the esterase specific for the intracellular deesterification of F2-AM and that, in PMNs, F2-AM may not remain in contact with intracellular esterase for a sufficient period of time to ensure complete deesterification. In this situation, the cell-associated fura-2 moieties, containing from one to four acetoxy-methyl esters on the Ca'+-binding portion of the molecule, would not be expected to be Ca2+-sensitive. Incomplete deesterification was also occasionally apparent in smooth muscle cells, a cell type in which deesterification generally proceeds to completion (12). The source of Ca2+insensitive fluorescence in smooth muscle cells was found to depend on the lot of F2-AM used, and the problem could be circumvented by switching to a new lot of dye, suggesting that a t times the inability of cells to fully deesterify "F2-AM" may have been due to impurities in the dye.
In any situation in which there is a form of fura-2 associated with cells in addition to F2 and F2-Ca'', direct application of the standard procedure for [Ca'+] measurement, which includes the parameters, Rmi,, R,,, and p for the free acid, leads to erroneous results. We have developed a method to account for Ca2+-insensitive fluorescence and to accurately measure [Ca"] within PMNs. This method is applicable to any cell type which exhibits intracellular, Ca'+-insensitive fluorescence. The method requires treatment of cells with ionomycin, in the presence of Ca2+-free and Ca'+-saturating buffers, to generate RL,, and R L , which reflect the cell-associated, minimum and maximum values of R, and 0'. RLi,,, RL,, and p' are then used in the standard equation to measure cellular [Ca'+]. An assumption in the method is that the concentration of ionomycin used is sufficient to equilibrate intracellular and extracellular [Ca"]. Equilibration is attained when the value of R is unchanged by further ionomycin addition, working near the Kd for fura-2.
We verified the use of this modified method by determining the spectral characteristics of the Ca2+-insensitive fluorescence associated with PMNs, as described earlier, and subtracting them from the total cell fluorescence (measured prior to treatment with ionomycin) at selected wavelengths (i.e. 340 and 380 nm). The resultant Ca2+-sensitive fluorescences at these wavelengths can then be used according to the standard procedure of [Ca'+] calculation (l), as described under "Materials and Methods." The results of either method are identical. An advantage to the second method is that it assesses the forms of fura-2 fluorescence associated with the cell and allows for use of the appropriate procedure for calculating [ CaZ+].

Ca2+-insensitiue For
Using R,&, R i , and p' in place of R,,,, R,,,, and p, we have recalculated [Ca'+] measurements from the experiment described earlier (Fig. 1). The results are plotted in Fig. 4 and now indicate significant changes in [Ca2+] upon stimulation with fMLP. Within 12 s of fMLP stimulation, [Ca"] increased severalfold and then fell to a level higher than the resting level for the remainder of the experiment. The time course of Ca2+ change is representative of more than 20 experiments using lo-' M fMLP; the results are independent of the amount of Ca2+-insensitive fura-2 within PMNs.
The most striking aspect of the work described here is the pitfall of making assumptions about the intracellular behavior of fura-2, based solely on theoretical grounds. In order to use fura-2 in any cell type, and with each new lot of F2-AM, the simplest way to test for cell-associated Ca2+-insensitive fluorescence is by an in situ calibration of cell-associated dye with 'm of Fura-2 in PMNs ionomycin, in the presence of Ca2+-saturating and Ca2+-free buffers. If these R values do not match those of free acid in Ca2+-saturating and Ca2+-free buffer, then the assumption of only two intracellular forms of the dye, F2-Ca2+ and F2, is not valid and the use of the standard procedure for [Ca"] measurement, using R,,, and Rmi,, is not possible. Having identified the nature of the Ca2+ insensitivity in PMNs loaded with fura-2, and with methods for valid [Ca"] measurement, we are now able to use fura-2 as a Ca2+ indicator in PMNs. The results of experiments on changes in [Ca'+] after fMLP stimulation will be described in a future paper.