| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Cellular and Molecular Physiology, College of Medicine, The Pennsylvania State University, Hershey, Pennsylvania 17033
Submitted 10 October 2001 ; accepted in final form 9 June 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
rhod 2; indo 1; fluo 3; fura 2; mitochondria; surface fluorescence
Although several methods have been described to load cells with these fluorescent probes for intracellular ion measurements, the most popular approach is the use of esters. Many fluorescent indicators are available as either acetoxymethyl esters (AMs) or acetate esters that are membrane permeable. Uptake of the dye is followed by hydrolysis of the esters by nonspecific intracellular esterases, leaving the anionic, non-permeable form of the dye trapped within the cell. The use of dual wavelength approaches with wavelength-shift dyes makes accurate quantitation of the intracellular ion species possible without requiring knowledge of the intracellular dye concentration (6).
The use of fura 2 and other indicators is complicated by fact that these dyes, to varying degrees, tend to load into subcellular organelles in addition to the cell cytosol. It has also been recognized that the pentaacetoxymethyl ester of fura 2 (fura 2-AM) does not completely hydrolyze in some preparations. In attempts to use these techniques for the measurement of intracellular free calcium in the intact rat heart, we observed an increase in the basal component of the tissue fluorescence upon dye loading that did not correspond to the anticipated level of calcium. In this report, we show that the AM forms of fura 2, indo 1, and fluo 3 are poorly hydrolyzed by the perfused rat heart. Separation of the free acid from the hydrolysis intermediates by HPLC demonstrated that the majority of these indicators remain in the tissue in their initial ester form. Of the dyes tested, the only calcium-sensitive dye found to be completely hydrolyzed to the free acid by the perfused rat heart was rhod 2. We also report that the distribution of these indicators between the cytosolic and mitochondrial compartment is variable and dependent on the particular fluorescent indicator.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
1 cm in diameter of the left ventricle was illuminated using the common end of the fiber-optic bundle. The emitted fluorescence was collected using the common end of the same fiberoptic bundle and was directed to the photomultiplier tube detector operating in the photon count mode, as described in detail previously (16). The excitation and emission wavelengths (center wavelength/bandpass; in nm) used for these probes were as follows: fura 2, 340/10 and 380/10 excitation and 500/30 emission; indo 1, 340/10 excitation and 405/10 emission; fluo 3, 500/15 excitation and 574/10 emission; and rhod 2, 532/10 excitation and 574/10 emission. Interference filters were obtained from either Omega Optical or Chroma Technology (Brattleboro, VT). Fluorescence measurements on tissue and mitochondrial extracts were made with the same apparatus and wavelengths except that the bifurcated light guide was directed toward a standard cuvette holder (model CV-1, C&L Instruments,).
HPLC separation of dye intermediates. Tissue extracts were assayed for fluorescent indicators after separation using a 10-cm, 3-µm particle size C-18 reverse-phase HPLC column (Microsorb-MV, Ranin). Fura 2-AM intermediates were separated using a gradient of acetonitrile (solvent B) and ammonium acetate plus 0.1 mM EGTA (solvent A). Solvent A was made by diluting glacial acetic acid [0.5% (vol/vol)] and concentrated ammonium hydroxide [0.3% (vol/vol)] in aqueous 0.1 mM EGTA (free acid). This solution had a pH of 4.2. We found that this procedure yielded more consistent retention times than mixing components based on the solution final pH. The gradient was linear from 10% to 50% B in 3 min and then linear from 10% to 70% B from 3 to 10 min. Indo 1 intermediates were separated using the same gradient and solvent system. Rhod 2-AM intermediates were also separated using the same gradient solvent system, but 1 mM calcium acetate was substituted for EGTA in solvent A. The same gradient system was used to separate fluo 3 intermediates except that 0.1 mM EGTA plus 1 mM calcium acetate was used in solvent A.
HPLC separations were performed using a Water Maxima 820 software system coupled to a Varian autoinjector (model 9010), an Applied Biosystems UV detector (model 759A), and an Applied Biosystems fluorescence detector (model 980). The excitation wavelengths for fura 2, rhod 2, indo 1, and fluo 3 fluorescence were 370, 546, 345, and 492 nm, respectively. Emission was detected using high-pass emission filters of >470, >570, >389, and >515, respectively. The absorbance of rhod 2 and fluo 3 was detected at 546 and 500 nm, respectively.
Isolation and extraction of mitochondria. Mitochondria were isolated from perfused rat hearts after the dye loading procedure described previously (14). After isolation, mitochondria were keep at 4°C in the isolation media at a protein concentration between 20 and 30 mg/ml until extracted for HPLC assay.
During the isolation process, a portion of the initial homogenate was withdrawn for tissue extraction of the dyes for HPLC analysis. These data were used to obtain the total amount of fluorescent indicator species in the heart. Mitochondria and homogenates were extracted by adding 15 ml methanol/g tissue. This mixture was sonicated twice for 30 s each (Vibra Cell, Sonics and Materials) and centrifuged at 16,000 g for 2 min to obtain a clear supernatant. The supernatant was used directly for HPLC analysis. We found that methanol was required for complete extraction (i.e., >95%) of the AM forms of the indicators. The recovery was confirmed by the addition of internal standards.
Assay of free acid forms. The free acid forms of the tested calcium indicators were measured by either HPLC separation or in vitro assay. HPLC separation was used to assay tissue extracts for the AM and free acid forms and hydrolysis intermediates (Table 1). More sensitivity, however, could be obtained using an in vitro assay specific for the free acid form (Table 2). This assay was performed by incubating either the tissue homogenate or mitochondria preparation in 50 mM MOPS plus 0.5% (g/g) Triton X-100 at pH 7.0. Typically, the equivalent of 3 mg heart tissue or 0.4 mg mitochondrial protein was solubilized per milliliter of this buffer. While the fluorescence was monitored at a calcium-sensitive wavelength, 1 mM EGTA was added to convert the indicator to the calcium-free form. CaCl2 (10 mM) was then added, followed by a known amount of the free acid form of the indicator. The difference in fluorescence between the calcium- and EGTA-containing solutions was recorded and calibrated in reference to the increase in fluorescence caused by the addition of the internal standard. The percentage of the indicator in mitochondria was calculated assuming 55 mg mitochondrial protein/g heart (8). Fluorescence wavelengths used for this assay were as described for cardiac surface fluorescence.
|
|
|
Isolated perfused rat hearts. Our Animal Resource Facility is accredited by American Association for Accreditation of Laboratory Animal Care International. Animal living conditions and all research protocols were conducted using protocols approved by our institutional review board. Isolated rat hearts were perfused at a constant aortic pressure of either 60 mmHg using the Langendorff procedure and a standard perfusion buffer as described elsewhere (16). In some experiments, hearts were perfused at an elevated perfusion pressure (110 mmHg) or with elevated levels of perfusate calcium (2.5 vs. 1.25 mM) to increase cardiac work. Loading of fluorescent dyes in perfused hearts was as described in RESULTS and figures. In some instances, hearts were electrically paced at 340 beats/min by a stimulator that was synchronized to the fluorescence data-acquisition system (C&L Instruments). The synchronization provided a consistent number of fluorescence measurements per cardiac cycle. This permitted a more detailed analysis of the fluorescence transients within the cardiac cycle using previously described signal averaging techniques (16).
Reagents and chemicals. Fura 2 and rhod 2 were obtained from Molecular Probes (Eugene, OR). Indo 1 and fluo 3 were obtained from Teflabs (Austin, TX). All other chemicals were purchased from Sigma (St. Louis, MO).
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
|
Procedures to calibrate the ratio fluorescence signal for quantitation of calcium levels have been previously reported (1). We found that application of this method yielded inconsistent results. Upon inspection of these data, it was noted that the basal fluorescence level of fura 2-loaded hearts (e.g., diastolic) was higher than expected based on reported levels of cytosolic calcium being between 50 and 100 nM (data not shown). Because the ester form of fura 2 is fluorescent but calcium insensitive, we explored the possibility that residual ester in the heart caused the elevated basal fluorescence.
To examine the hydrolysis of fura 2-AM by heart tissue, hydrolysis intermediates were separated by HPLC. Figure 2 illustrates the separation of standards generated by incubation of fura 2-AM with a homogenate of rat heart tissue. This procedure identified four major peaks that eluted between the two peaks that corresponded to those of the free acid and the parent pentaacetoxymethyl ester. Plots of the peak area as a function of the incubation time of the parent pentaester with the homogenate indicated that the number of residual ester groups solely determined the retention time (data not shown). During the hydrolysis process, however, there could be a total of 32 different compounds (e.g., 25) because the parent compound has five AM groups. It is likely that we observed only four intermediate peaks because of the reverse-phase mode of HPLC separation. Presumably, all the intermediates with the same number of residual ester groups coeluted. Of the possible 32 compounds, the separate number of unique compounds that could theoretically be in each peak would be 1, 5, 10, 10, 5, and 1, respectively, assuming this mode of separation. Similar separations for rhod 2, indo 1, and fluo 3 were developed. In all instances, the free acid of these dyes was the first form to elute and the AM ester was the last form. It is of interest that both rhod 2 and fluo 3 contain four AM groups, but yet graded hydrolysis by heart homogenate generated five intermediates that eluted between the free acid and AM forms. Presumably, the HPLC separation was able to resolve some intermediates that had the same number of total AM groups in the case of these dyes.
Table 1 shows the content of the fura 2 free acid form and fura 2-AM in heart tissue loaded under several conditions. Comparable data are also shown for indo 1, fluo 3, and rhod 2. It was surprising to find that with fura 2, indo 1, and fluo 3, the heart contained predominantly the parent AM form of these indicators. Under no conditions could we detect any intermediates in intact heart tissue. We could demonstrate the formation of these intermediates by limited timed exposures of fura 2-AM with homogenized heart tissue (Fig. 2), but these intermediates were not detected in intact hearts loaded with fura 2-AM. This observation suggests that, in the intact heart, the residual parent esters of these indicators reside in a compartment that is not accessible to the esterase(s) responsible for their hydrolysis. These data can explain the high background fluorescence we observed after fura 2-AM loading.
Unlike the other tested indicators, rhod 2 was completely converted to the acid form by heart tissue. Under no circumstances could we detect residual parent ester or any intermediates of rhod 2 in the whole heart tissue. It appeared that the rat heart has a higher avidity for rhod 2 than the other tested indicators. Figure 3 illustrates the loading process with rhod 2 and fura 2. Rhod 2 was taken up much more rapidly than other tested indicators. The heart poorly loaded with fluo 3. In fact, it was difficult to load hearts consistently with fluo 3 to levels significantly above initial background levels.
|
Loading into mitochondria. It is recognized that hydrolysis of these indicators can occur in both the cytosol and mitochondrial matrix to generate active free acid forms in these compartments. To examine this in per-fused rat hearts, hearts were loaded with rhod 2, indo 1, or fura 2, and mitochondria were isolated and analyzed for the free acid form. Table 2 illustrates these data. It is apparent that mitochondrial loading was significant for all three compounds, and mitochondria were the predominant subcellular compartment in the case of rhod 2.
It has been reported that rhod 2 could be preferentially loaded into mitochondria of adult rabbit cardiac myocytes by loading at 4°C (20). We found that loading at this temperature in the perfused heart had no effect on the relative mitochondrial loading of this indicator.
It is of interest that hearts could be loaded using considerably less rhod 2 than fura 2 for a comparable increase in fluorescence above initial baseline measurements. This can be attributed to the fact the baseline fluorescence of hearts is higher when excited using wavelengths in the ultraviolet range and that rhod 2 is intrinsically more fluorescent than fura 2.
In experiments with isolated mitochondria loaded with rhod 2 in this manner, it could be demonstrated that the active dye was in the matrix space. These experiments involved incubation of perfusion-loaded mitochondria and monitoring of the uptake of added calcium and calcium efflux upon subsequent addition of ruthenium red and sodium (data not shown). Mitochondria loaded with fura 2 in this manner behaved in a similar fashion to mitochondria loaded with fura 2 in vitro (23).
It should be pointed out that these data represent the amount of fluorescent active dye recovered in isolated mitochondria, and, therefore, this represents a minimal estimate of the mitochondrial dye content. It does not take into account any mitochondrially associated dye that may have leaked out of the matrix during the isolation process. These dyes are retained by mitochondria to varying extents. Mitochondria loaded in vitro by incubation with fura 2-AM are known to slowly leak the active free acid form (23). This leakage from the matrix, at least with rhod 2, presumably also occurs in the intact heart. In the absence of probenecid, a 50-min increase in the perfusion washout time was associated with a 22% decrease in the mitochondrial rhod 2 content. This was more evident in the presence of probenecid, where an increase in the duration of perfusion caused a 56% lower mitochondrial rhod 2 concentration.
The data presented in Table 2 indicate that the presence of probenecid has a marked affect on both the relative compartmentation of rhod 2 and the extent of rhod 2 loading. Probenecid, a nonspecific anion transport inhibitor, is often used to enhance retention of fura 2 in cells that exhibit rapid leakage of the indicator. In hearts, probenecid caused higher loading into both the cytosol and mitochondria.
Because the free acid forms of rhod 2 and fluo 3 are essentially nonfluorescent in the absence of calcium and when in the AM form, the concentration of the AM form of these indicators shown in Table 2 was measured using an absorbance detector. If calcium was added to the HPLC elution buffer, intermediates in the hydrolysis process that become fluorescence with calcium could be identified. Table 3 illustrates the fluorescence to the ultraviolet absorbance ratio of rhod 2 and hydrolysis intermediates separated by HPLC. These data indicate that rhod 2 does not acquire calcium-mediated fluorescence until there is only one remaining AM group. Assuming that loss of the AM does not alter the extinction coefficient of rhod 2, these data show that in the presence of calcium, rhod 2-AM fluorescence is increased 64-fold by complete hydrolysis.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
Incomplete fura 2-AM hydrolysis has been observed in cell preparations (13, 18, 19). In all these cases, residual fura 2-AM was identified by a HPLC approach similar to that used in the present study. In vascular smooth muscle cells, the AM was found to accumulate in the membrane fraction, whereas the soluble fraction contained only the free acid (18). Significant residual fura 2-AM was also found after perfusion loading of fura 2-AM in arterial strips (19). The data obtained in arterial strips are similar to our observations in the perfused heart. In all cases in which the residual ester was observed after loading, intermediates of the hydrolysis process were not found to accumulate, yet the tissue contained considerable quantities of the ester. In studies of a variety of cultured cell types, including rat hepatocytes, N1E-115 neuroblastoma cells, and human pulmonary artery endothelial cells, Oakes et al. (13) observed that these cells contained several metabolites of fura 2-AM as well as large amounts of residual fura 2-AM.
Problems arising from incomplete hydrolysis are dependent on the type of dye employed and the method used to quantitate the fluorescence signal. With the use of the ratio method of Grynkiewicz et al. (6) with the wavelength-shift dyes fura 2 and indo 1, incomplete hydrolysis does not affect quantitation so long as Rmax, Rmin, and 
[ = kd (sf2/sb2), where sf2 and sfb are the fluorescence intensity of the Ca2+-free and Ca2+-bound dye, respectively, at the excitation wavelength of 380 nm and kd is the dissociation constant for the Ca2+-dye complex] are obtained in the same preparation; this involves manipulating tissue calcium levels to generate the calcium-free form of the dye (Rmin) and the calcium-saturated form (Rmax) using the so-called "in vivo" calibration method. In the presence of residual ester, the drawback in this instance is a lower dynamic range of the indicator because the calcium-insensitive AM forms of these dyes fluoresce. More often, however, "in vitro" calibration methods are used because of their simplicity. In this instance, errors may be significant when the fluorescence signals are calibrated by comparison to in vitro measurements made using fura 2 free acid in solution.
With other calcium indicator dyes that are monitored with a single excitation and emission wavelength, such as rhod 2 and fluo 3, potential problems that can arise from accumulation of residual ester depend on the fluorescent properties of the ester. If the parent ester does not fluoresce appreciably, as in the case of rhod 2 and fluo 3, then the presence of residual ester has no effect on the calibration. With these dyes, however, calibration requires either "in vivo" methods or knowledge about the actual dye concentration. In some instances, use of dyes of this type has been coupled with a second dye to produce a ratio method (9, 22). The other noncalcium sensitive dye serves as a reference for the calcium-sensitive indicator, thus alleviating the need for measuring the amount of calcium-sensing dye in the preparation. In another more recent approach, Du et al. (3) and MacGowan et al. (10) described a method for the use of rhod 2 in the perfused mouse heart for calcium quantitation. The reflected absorbance and fluorescence spectra were determined and used to measure the fluorescence quantum yield of rhod 2 upon calcium binding. With the use of this ratio technique, motion artifacts are minimized, as are errors caused by loss of the indicator during the course of perfusion.
We determined that significant quantities of rhod 2, fura 2, and indo 1 free acid accumulate in the mitochondrial matrix. In the case of fura 2 and indo 1, this colocalization has been recognized for some time. In attempts to correct for this problem, two approaches have been used to remove the fluorescence contribution of the cytosolic component. Miyata et al. (12) described the use of Mn2+ to quench the fluorescence of the cytosolic component. In isolated and paced myocytes, the oscillatory contribution of the signal due to cytosolic indo 1 is lost upon quenching by Mn2+, leaving only the mitochondrial component. Use of this method is widespread, and it has been also used to correct for residual unhydrolyzed fura 2-AM in cell preparations (7). This method, however, has the potential problem of Mn2+ quenching mitochondrial indo 1 and fura 2. Mitochondria are known to take up Mn2+ by the same transport system used to transport Ca2+ (4). In fact, Zhou et al. (24) reported evidence for Mn2+ quenching of matrix indo 1. In another approach, Griffiths et al. (5) described incubation conditions after indo 1 loading that could be used to selectively promote time-dependent loss of the free acid from the cytosolic compartment.
Neither of these methods, however, addresses measurement of cytosolic calcium or of calcium in intact tissue preparations. The only methods that are generally accepted for measurement of cytosolic calcium of single cells with these indicators are by using the free acid directly either by microinjection of the free acid or by loading the free acid during the process of isolating myocytes (17).
The compartmentation of rhod 2 between the mitochondrial matrix and cytosol is controversial. Unlike other indicators, rhod 2-AM is a monovalent cation. For this reason, Minta et al. (11) proposed that it might selectively load into the matrix because the mitochondrial membrane potential could favor uptake of the cationic ester. Using myocytes isolated from the guinea pig, Del Nido and coworkers (2) concluded that rhod 2 loads only into the cytosolic compartment. These authors did not observe the punctate fluorescence pattern after rhod 2 loading that was typically observed after loading of tetramethylrhodamine ethyl ester (TMRE). TMRE is a fluorescent lipophilic cation that is taken up by mitochondria because of the membrane potential (15). TMRE fluorescence, but not that of rhod 2, dissipated upon application of the mitochondrial uncoupler FCCP (2). Similar results were observed in mouse hearts by MacGowan et al. (10). These authors loaded perfused mouse hearts with rhod 2, followed by fixation with a carbodiimide and visualization with double antibody staining. These authors did not observe the "punctate" staining pattern that would have been expected upon mitochondrial loading. On the other hand, Trollinger et al. (20) concluded that rhod 2 loads exclusively into the mitochondrial matrix of adult rabbit myocytes. These authors observed the clear punctate fluorescence appearance typical of mitochondrial loading. Possible reasons accounting for the differences between these studies remains elusive. Secondary indicators or fixing reagents, as in the study of MacGowan et al. (10), may not label all parts of the cell in a representative manner. Species differences may also account for the differences in mitochondrial loading of rhod 2 and other indicators. Our studies are based on the recovery of rhod 2 in the mitochondrial fraction after tissue loading and isolation of mitochondria. It could be argued that the uptake of either rhod 2 free acid and/or rhod 2-AM during the isolation process could have overestimated our measure of the mitochondrial dye content. However, these explanations are unlikely because we could not observe retention of rhod 2 free acid by mitochondria when added to the tissue homogenate during the isolation process (data not shown), and we were unable to detect rhod 2-AM after loading of the ester in the intact heart before isolation of mitochondria (Table 1).
In the intact heart, we found that loading of rhod 2 by mitochondria occurred to a greater extent than either fura 2 or indo 1 (Table 2). It should be noted that our method only accounts for the amount of indicator recovered after mitochondrial isolation, and thus it represents a minimum estimate of mitochondrial loading. In addition, uptake and hydrolysis of rhod 2 by the intact heart occurred much more quickly than with other indicators (Fig. 3). It is possible that the cationic ester also promoted cellular uptake either directly because of the sarcolemmal membrane potential or indirectly because of the mitochondrial membrane potential. We also noted that isolated rat heart mitochondria avidly load rhod 2 upon incubation with the AM form (data not shown).
In conclusion, our data with indo 1 and rhod 2 illustrate that in vitro calibration methods used to estimate cytosolic calcium in the perfused rat heart are plagued by complications arising from incomplete ester hydrolysis and subcompartmentation of the indicator. Data from other laboratories also indicate that this problem exists in the use of these indicators to measure calcium in a variety of isolated cell types. We found, however, that incomplete hydrolysis does not occur with rhod 2 in the intact rat heart. We conclude that rhod 2 is the preferred indicator for calcium measurements in the heart, although the fluorescence signal reflects a mixed signal arising from the cytosol and mitochondrial matrix.
|
DISCLOSURES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  O. Eerbeek, E. G. Mik, C. J. Zuurbier, M. v. t Loo, C. Donkersloot, and C. Ince Ratiometric intracellular calcium imaging in the isolated beating rat heart using indo-1 fluorescence J Appl Physiol, December 1, 2004; 97(6): 2042 - 2050. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
