Near-membrane iminocoumarin-based low affinity fluorescent Ca(2+) indicators

Two new potential near-membrane iminocoumarin-based fluorescent Ca(2+) indicators were synthesized and the spectral profiles of their free and Ca(2+) bound forms were studied. The probes incorporate in their BAPTA-related structures, the 3-(benzimidazolyl)iminocoumarin or the 3-(benzothiazolyl)iminocoumarin moiety, substituted at the imino nitrogen with an n-dodecyl lipophilic chain. The compounds are excited with visible light and have Ca(2+) dissociation constant values of 5.50 and 4.49 microM, respectively, the highest reported to date in the literature. Fluorescence spectra studies indicated a clear shift in their excitation wavelength maxima upon Ca(2+) binding along with changes in fluorescence intensity that enable the compounds to be used as ratiometric near-membrane, low Ca(2+) affinity probes.     

Measurements of the free luminal ER Ca(2+) concentration with targeted "cameleon" fluorescent proteins

The free ER Ca(2+) concentration, [Ca(2+)](ER), is a key parameter that determines both the spatio-temporal pattern of Ca(2+) signals as well as the activity of ER-resident enzymes. Obtaining accurate, time-resolved measurements of the Ca(2+) activity within the ER is thus critical for our understanding of cell signaling. Such measurements, however, are particularly challenging given the highly dynamic nature of Ca(2+) signals, the complex architecture of the ER, and the difficulty of addressing probes specifically into the ER lumen. Prompted by these challenges, a number of ingenious approaches have been developed over the last years to measure ER Ca(2+) by optical means. The two main strategies used to date are Ca(2+)-sensitive synthetic dyes trapped into organelles and genetically encoded probes, based either on the photoprotein aequorin or on the green fluorescent protein (GFP). The GFP-based Ca(2+) indicators comprise the camgaroo and pericam probes based on a circularly permutated GFP, and the cameleon probes, which rely on the fluorescence resonance energy transfer (FRET) between two GFP mutants of different colors. Each approach offers unique advantages and suffers from specific drawbacks. In this review, we will discuss the advantages and pitfalls of using the genetically encoded "cameleon" Ca(2+) indicators for ER Ca(2+) measurements.     

Iminocoumarin-based low affinity fluorescent Ca2+ indicators excited with visible light.

A series of iminocoumarin-based fluorescent Ca2+ indicators were synthesized and the spectral profiles of their free and Ca2+ bound forms were studied. The newly-synthesized compounds incorporate the Ca2+ chelating structure of BAPTA. The chromophore moieties are iminocoumarins substituted at the 3-position with benzothiazolyl, benzoxazolyl and benzimidazolyl groups. These compounds are excited with visible light and their Ca2+ dissociation constants range from 5.4 to 27.5 microM. Fluorescence spectra studies of these probes indicated a clear shift in their excitation wavelength maxima upon Ca2+ binding along with changes in fluorescence intensity that enable the compounds to be used as low Ca2+ affinity, visible excitable probes.     

A high signal-to-noise Ca(2+) probe composed of a single green fluorescent protein

Recently, several groups have developed green fluorescent protein (GFP)-based Ca(2+) probes. When applied in cells, however, these probes are difficult to use because of a low signal-to-noise ratio. Here we report the development of a high-affinity Ca(2+) probe composed of a single GFP (named G-CaMP). G-CaMP showed an apparent K(d) for Ca(2+) of 235 nM. Association kinetics of Ca(2+) binding were faster at higher Ca(2+) concentrations, with time constants decreasing from 230 ms at 0.2 microM Ca(2+) to 2.5 ms at 1 microM Ca(2+). Dissociation kinetics (tau approximately 200 ms) are independent of Ca(2+) concentrations. In HEK-293 cells and mouse myotubes expressing G-CaMP, large fluorescent changes were observed in response to application of drugs or electrical stimulations. G-CaMP will be a useful tool for visualizing intracellular Ca2+ in living cells. Mutational analysis, together with previous structural information, suggests the residues that may alter the fluorescence of GFP.     

Imaging Ca2+ concentration changes at the secretory vesicle surface with a recombinant targeted cameleon.

Regulated exocytosis involves the Ca(2+)-triggered fusion of secretory vesicles with the plasma membrane, by activation of vesicle membrane Ca(2+)-binding proteins [1]. The Ca(2+)-binding sites of these proteins are likely to lie within 30 nm of the vesicle surface, a domain in which changes in Ca2+ concentration cannot be resolved by conventional fluorescence microscopy. A fluorescent indicator for Ca2+ called a yellow 'cameleon' (Ycam2) - comprising a fusion between a cyan-emitting mutant of the green fluorescent protein (GFP), calmodulin, the calmodulin-binding peptide M13 and an enhanced yellow-emitting GFP - which is targetable to specific intracellular locations, has been described [2]. Here, we generated a fusion between phogrin, a protein that is localised to secretory granule membranes [3], and Ycam2 (phogrin-Ycam2) to monitor changes in Ca2+ concentration ([Ca2+]) at the secretory vesicle surface ([Ca2+]gd) through alterations in fluorescence resonance energy transfer (FRET) between the linked cyan and yellow fluorescent proteins (CFP and YFP, respectively) in Ycam2. In both neuroendocrine PC12 and MIN6 pancreatic beta cells, apparent resting values of cytosolic [Ca2+] and [Ca2+](gd) were similar throughout the cell. In MIN6 cells following the activation of Ca2+ influx, the minority of vesicles that were within approximately 1 microm of the plasma membrane underwent increases in [Ca2+](gd) that were significantly greater than those experienced by deeper vesicles, and greater than the apparent cytosolic [Ca2+] change. The ability to image both global and compartmentalised [Ca2+] changes with recombinant targeted cameleons should extend the usefulness of these new Ca2+ probes.     

Cellular calcium imaging: so, what's new?

The ability to uncover the fine details of intracellular Ca(2+) signals has improved remarkably in recent years, largely as a result of developments in methods for reporting Ca(2+), coupled with great improvements in measurement instrumentation. The cell biologist wishing to image intracellular Ca(2+) has a range of options to consider. These include the use of photoproteins, commercially available fluorescent indicators or the new generation of fluorescent protein Ca(2+) probes. Molecular biology and biophysics are now joining forces to bring major advances in the art of deciphering the complexity of spatiotemporal Ca(2+) signals.     

Blind spectral decomposition of single-cell fluorescence by parallel factor analysis

Simultaneous measurement of multiple signaling molecules is essential to investigate their relations and interactions in living cells. Although a wide variety of fluorescent probes are currently available, the number of probes that can be applied simultaneously is often limited by the overlaps among their fluorescence spectra. We developed the experimental system to measure and analyze many overlapping fluorescent components in single cells. It is based on the recording of two-dimensional single-cell fluorescence spectra and on the blind spectral decomposition of fluorescence data by method of parallel factor analysis. Because this method does not require any preknowledge about the shapes of individual component spectra, it can be applied to the specimens that contain fluorescent components with unknown spectra. By examining the performance using the mixture solutions of fluorescent indicators, it was confirmed that >10 largely overlapping spectral components could be easily separated. The effectiveness in the physiological experiments was proven in the applications to the temporal analysis of intracellular Ca(2+) concentration and pH, as well as the intrinsic fluorescent components, in single mouse oocytes.     

Properties of tri- and tetracarboxylate Ca2+ indicators in frog skeletal muscle fibers.

Recently a number of lower-affinity fluorescent Ca2+ indicators have become available with principal absorbance bands at visible wavelengths. This article evaluates these indicators, as well as two shorter wavelength indicators, mag-fura-5 and mag-indo-1, for their suitability as rapid Ca2+ indicators in frog skeletal muscle fibers. With three lower-affinity tricarboxylate indicators (mag-fura-5, mag-indo-1, and magnesium orange), the change in fluorescence in response to an action potential (delta F) appeared to track the myoplasmic Ca2+ transient (delta[Ca2+]) without delay. With three lower-affinity tetracarboxylate indicators (BTC, calcium-orange-5N, and calcium-green-5N) and one tricarboxylate indicator (magnesium green), delta F responded to delta[Ca2+] with a small delay. Unfortunately, with the tetracarboxylate indicators, other problems were detected that appear to limit their usefulness as reliable Ca2+ indicators. Surprisingly, delta F from mag-fura-red, another tricarboxylate indicator, was biphasic (with 480 nm excitation), a feature that also greatly limits its usefulness. With several of the indicators, estimates were obtained for the myoplasmic value of KD, Ca (the indicator's dissociation constant for Ca2+) and found to be elevated severalfold in comparison with the value measured in a simple salt solution. These and other problems related to the quantitative use of Ca2+ indicators in the intracellular environment are evaluated and discussed.     

Calcium microdomains at presynaptic active zones of vertebrate hair cells unmasked by stochastic deconvolution

Signal transduction by auditory and vestibular hair cells involves an impressive ensemble of finely tuned control mechanisms, strictly dependent on the local intracellular Ca(2+) concentration ([Ca(2+)](i)). The study of Ca(2+) dynamics in hair cells typically combines Ca(2+)-sensitive fluorescent indicators (dyes), patch clamp and optical microscopy to produce images of the patterns of fluorescence of a Ca(2+) indicator following various stimulation protocols. Here we describe a novel method that combines electrophysiological recordings, fluorescence imaging and numerical simulations to effectively deconvolve Ca(2+) signals within cytoplasmic microdomains that would otherwise remain inaccessible to direct observation. The method relies on the comparison of experimental data with virtual signals derived from a Monte Carlo reaction-diffusion model based on a realistic reconstruction of the relevant cell boundaries in three dimensions. The model comprises Ca(2+) entry at individual presynaptic active zones followed by diffusion, buffering, extrusion and release of Ca(2+). Our results indicate that changes of the hair cell [Ca(2+)](i) during synaptic transmission are primarily controlled by the Ca(2+) endogenous buffers both at short (<1mu) and at long (tens of microns) distances from the active zones. We provide quantitative estimates of concentration and kinetics of the hair cell endogenous Ca(2+) buffers and Ca(2+)-ATPases. We finally show that experimental fluorescence data collected during Ca(2+) influx are not interpreted correctly if the [Ca(2+)](i) is estimated by assuming that Ca(2+) equilibrates instantly with its reactants. In our opinion, this approach is of potentially general interest as it can be easily adapted to the study of Ca(2+) dynamics in diverse biological systems.     

Subplasma membrane Ca2+ signals

Ca(2+) may selectively activate various processes in part by the cell's ability to localize changes in the concentration of the ion to specific subcellular sites. Interestingly, these Ca(2+) signals begin most often at the plasma membrane space so that understanding subplasma membrane signals is central to an appreciation of local signaling. Several experimental procedures have been developed to study Ca(2+) signals near the plasma membrane, but probably the most prevalent involve the use of fluorescent Ca(2+) indicators and fall into two general approaches. In the first, the Ca(2+) indicators themselves are specifically targeted to the subplasma membrane space to measure Ca(2+) only there. Alternatively, the indicators are allowed to be dispersed throughout the cytoplasm, but the fluorescence emanating from the Ca(2+) signals at the subplasma membrane space is selectively measured using high resolution imaging procedures. Although the targeted indicators offer an immediate appeal because of selectivity and ease of use, their limited dynamic range and slow response to changes in Ca(2+) are a shortcoming. Use of targeted indicators is also largely restricted to cultured cells. High resolution imaging applied with rapidly responding small molecule Ca(2+) indicators can be used in all cells and offers significant improvements in dynamic range and speed of response of the indicator. The approach is technically difficult, however, and realistic calibration of signals is not possible. In this review, a brief overview of local subplasma membrane Ca(2+) signals and methods for their measurement is provided.     

Structural dynamics of the Ca2(+)-ATPase of sarcoplasmic reticulum. Temperature profiles of fluorescence polarization and intramolecular energy transfer

The temperature dependence of fluorescence polarization and F?rster-type resonance energy transfer (FRET) was analyzed in the Ca2(+)-ATPase of sarcoplasmic reticulum using protein tryptophan and site-specific fluorescence indicators such as 5-[2-[iodoacetyl)amino)ethyl]aminonaphthalene-1-sulfonic acid (IAEDANS), fluorescein 5'-isothiocyanate (FITC), 2',3'-O-(2,4,3-trinitrophenyl)adenosine monophosphate (TNP-AMP) or lanthanides (Pr3+, Nd3+) as probes. The normalized energy transfer efficiency between AEDANS bound at cysteine-670 and -674 and FITC bound at lysine-515 increases with increasing temperature in the range of 10-37 degrees C, indicating the existence of a relatively flexible structure in the region of the ATPase molecule that links the AEDANS to the FITC site. These observations are consistent with the theory of Somogyi, Matko, Papp, Hevessy, Welch and Damjanovich (Biochemistry 23 (1984) 3403-3411) that thermally induced structural fluctuations increase the energy transfer. Structural fluctuations were also evident in the energy transfer between FITC linked to the nucleotide-binding domain and Nd3+ bound at the putative Ca2+ sites. By contrast the normalized energy transfer efficiency between AEDANS and Pr3+ was relatively insensitive to temperature, suggesting that the region between cysteine-670 and the putative Ca2+ site monitored by the AEDANS-Pr3+ pair is relatively rigid. A combination of the energy transfer data with the structural information derived from analysis of Ca2(+)-ATPase crystals yields a structural model, in which the location of the AEDANS-, FITC- and Ca2+ sites are tentatively identified.     

A ryanodine fluorescent derivative reveals the presence of high-affinity ryanodine binding sites in the Golgi complex of rat sympathetic neurons, with possible functional roles in intracellular Ca(2+) signaling

The plant alkaloid ryanodine (Ry) is a high-affinity modulator of ryanodine receptor (RyR) Ca(2+) release channels. Although these channels are present in a variety of cell types, their functional role in nerve cells is still puzzling. Here, a monosubstituted fluorescent Ry analogue, B-FL-X Ry, was used to reveal the distribution of RyRs in cultured rat sympathetic neurons. B-FL-X Ry competitively inhibited the binding of [3H]Ry to rabbit skeletal muscle SR membranes, with an IC(50) of 150 nM, compared to 7 nM of unlabeled Ry. Binding of B-FL-X Ry to the cytoplasm of sympathetic neurons is saturable, reversible and of high affinity. The pharmacology of B-FL-X Ry showed marked differences with unlabeled Ry, which are partially explained by its lower affinity: (1) use-dependent reversible inhibition of caffeine-induced intracellular Ca(2+) release; (2) diminished voltage-gated Ca(2+) influx, due to a positive shift in the activation of voltage gated Ca(2+) currents. B-FL-X Ry-stained sympathetic neurons, viewed under confocal microscopy, showed conspicuous labeling of crescent-shaped structures pertaining to the Golgi complex, a conclusion supported by experiments showing co-localization with Golgi-specific fluorescent probes and the breaking up of crescent-shaped staining after treatment with drugs that disassemble Golgi complex. The presence of RyRs to the Golgi could be confirmed with specific anti-RyR(2) antibodies, but evidence of caffeine-induced Ca(2+) release from this organelle could not be obtained using fast confocal microscopy. Rather, an apparent decrease of the cytosolic Ca(2+) signal was detected close to this organelle. In spite of that, short-term incubation with brefeldin A (BFA) suppressed the fast component of caffeine-induced Ca(2+) release, and the Ca(2+) release process lasted longer and appeared less organized. These observations, which suggest a possible role of the Golgi complex in Ca(2+) homeostasis and signaling in nerve cells, could be relevant to reports involving derangement of the Golgi complex as a probable cause of some forms of progressive neuronal degeneration, such as Alzheimer's disease and amyotrophic lateral sclerosis.     

A pyrimidopyrimidoindole nucleoside (dC PPI): photophysical properties and thermal stability of the modified DNA duplexes

A new fluorescent deoxycytidine analog, 10-(2-deoxy-beta -D-ribofuranosyl)-pyrimido[4',5' :4,5]-pyrimido[1,6-a]indole-6,9(7H)-dione (dC(PPI)) was synthesized. Its fluorescent properties were studied in detail. It was found that this fluorescent nucleoside dC(PPI) could be used as a fluorescent label for DNA probes with minimal disturbance of their overall structure.     

Protease-activated receptor-2 increases exocytosis via multiple signal transduction pathways in pancreatic duct epithelial cells

Protease-activated receptor-2 (PAR-2) is activated when trypsin cleaves its NH(2) terminus to expose a tethered ligand. We previously demonstrated that PAR-2 activates ion channels in pancreatic duct epithelial cells (PDEC). Using real-time optical fluorescent probes, cyan fluorescence protein-Epac1-yellow fluorescence protein for cAMP, PH(PLC-delta1)-enhanced green fluorescent protein for phosphatidylinositol 4,5-bisphosphate, and protein kinase Cgamma (PKCgamma)-C1-yellow fluorescence protein for diacylglycerol, we now define the signaling pathways mediating PAR-2 effect in dog PDEC. Although PAR-2 activation does not stimulate a cAMP increase, it induces phospholipase C to hydrolyze phosphatidylinositol 4,5-bisphosphate into inositol 1,4,5-trisphosphate and diacylglycerol. Intracellular Ca(2+) mobilization from inositol 1,4,5-trisphosphate-sensitive Ca(2+) stores and a subsequent Ca(2+) influx through store-operated Ca(2+) channels cause a biphasic increase in intracellular Ca(2+) concentration ([Ca(2+)](i)), measured with Indo-1 dye. Single-cell amperometry demonstrated that this increase in [Ca(2+)](i) in turn causes a biphasic increase in exocytosis. A protein kinase assay revealed that trypsin also activates PKC isozymes to stimulate additional exocytosis. Paralleling the increased exocytosis, mucin secretion from PDEC was also induced by trypsin or the PAR-2 activating peptide. Consistent with the serosal localization of PAR-2, 1 microm luminal trypsin did not induce exocytosis in polarized PDEC monolayers; on the other hand, 10 microm trypsin at 37 degrees C damaged the epithelial barrier sufficiently so that it could reach and activate the serosal PAR-2 to stimulate exocytosis. Thus, in PDEC, PAR-2 activation increases [Ca(2+)](i) and activates PKC to stimulate exocytosis and mucin secretion. These functions may mediate the reported protective role of PAR-2 in different models of pancreatitis.     

A new approach to the synthesis of APTRA indicators

We have devised a general synthesis of Mg(2+) indicators which is based on the aminophenol triacetic acid (APTRA) structure. The key step is a palladium-catalyzed coupling reaction of a precursor of the APTRA ligand with a fluorescent group. This strategy resulted in new ratioable fluorescent APTRA indicators and the finding that the fluorescence response of these indicators is different for Mg(2+) and Ca(2+) in some cases. We believe that this represents a generally useful approach for combining fluorophore and chelator functionalities.     

Enhanced dynamic range in a genetically encoded Ca2+ sensor

Genetically encoded fluorescence resonance energy transfer (FRET) indicators are powerful tools for real-time detection of second messenger molecules and activation of signal proteins. However, these fluorescent protein-based sensors typically display marginal FRET efficiency. To improve their FRET efficiency for optical imaging and screening, we developed a number of fluorescent protein mutants based on cyan fluorescent protein (CFP) and yellow fluorescent protein (YFP). To improve FRET ratios, which were initially within a narrow dynamic range, we used DNA shuffling to develop a new FRET pair called 3xCFP/Venus. The optimized 3xCFP/Venus pair exhibited higher FRET ratios than CyPet/YPet, which has one of the greatest dynamic ranges of protein-based FRET pairs. We converted this FRET pair to a Ca(2+) FRET indicators using circular permutation Venus (cpVenus) linked with 3xCFP to form 3xCFP/cpVenus, which displayed an ～11-fold change in dynamic range in response to Ca(2+) binding. The enhanced dynamic range for Ca(2+) concentration detection using 3xCFP/cpVenus was confirmed in PC12 cells using previously established indicators (TN-XXL, ECFP/cpCitrine). To our knowledge, this FRET pair displays the largest dynamic range so far among genetically-encoded sensors, and can be used for sensitive FRET detection.     

Use of co-loaded Fluo-3 and Fura Red fluorescent indicators for studying the cytosolic Ca(2+)concentrations distribution in living plant tissue

A method for visualisation of cytosolic [Ca(2+)] distribution was applied to living plant tissue. A mixture of the fluorescent probes Fluo-3 and Fura Red was used. The emitted fluorescence was scanned simultaneously in two channels with a laser-scanning confocal microscope and rationing was performed. The homogeneity of the Fluo-3/Fura Red concentration ratio throughout the tissue after AM-ester loading was proven. In vitro calibration permitted conversion of Fluo-3/Fura Red fluorescence ratios to [Ca(2+)] values. Apparent K(D)of 286 nM, R(min)of 0.43 and R(max)of 18 were calculated. The in vivo determination of extreme ratio values was performed by permeabilizing the plasmalemma for Ca(2+)with a ionophore and manipulating the extracellular [Ca(2+)]. The resultant R(minv)of 1.33 and R(maxv)of 2.69 for vegetative apices, and R(mini)of 1.26 and R(maxi)of 3.45 for apices induced to flowering, suggested incomplete equalization of extra- and intracellular Ca(2+)levels in these experiments. In Chenopodium rubrum, the cytosolic [Ca(2+)] patterns of apical tissue obtained using Fluo-3 and Fura Red were significantly different between vegetative apices and apices after photoperiodic flower induction. This methodological approach may also be helpful for studying cytosolic [Ca(2+)] distribution in other living plant tissues.     

Quantitative estimate of mitochondrial [Ca2+] in stimulated motor nerve terminals

Peak values reported for mitochondrial matrix [Ca(2+)] following stimulation have ranged from micromolar to near-millimolar in various cells. Measurements using fluorescent indicators have traditionally used high-affinity dyes such as rhod-2, whose fluorescence would be expected to saturate if matrix [Ca(2+)] approaches millimolar levels. To avoid this potential problem, we loaded lizard motor terminal mitochondria with the low-affinity indicator rhod-5N (K(d) approximately 320 microM). During trains of action potentials at 50Hz, matrix fluorescence transients (measured as F/F(rest)) increased to a plateau level that was maintained throughout the stimulus train. This plateau of matrix [Ca(2+)] occurred in spite of evidence that Ca(2+) continued to enter the terminal and continued to be sequestered by mitochondria. When the stimulation frequency was increased, or when Ca(2+) entry per action potential was increased with the K(+) channel blocker 3,4-diaminopyridine (3,4-DAP), or reduced by lowering bath [Ca(2+)], the rate of rise of matrix [Ca(2+)] changed, but the plateau amplitude remained constant. Calculations demonstrated that the F/F(rest) measured at this plateau corresponded to a matrix [Ca(2+)] of approximately 1 microM. The high K(d) of rhod-5N ensures that this value is not a result of dye saturation, but rather reflects a powerful Ca(2+) buffering mechanism within the matrix of these mitochondria.     

In the neuronal cell line SH-SY5Y, oxidative stress-induced free radical overproduction causes cell death without any participation of intracellular Ca(2+) increase

Adding the membrane-permeant oxidant tert-butylhydroperoxide (t-BOOH) to the incubation medium, in SH-SY5Y human neuroblastoma cells, induced a marked and progressive concentration-dependent (300, 500 and 1000 microM) increase of free radical production, as evaluated by the fluorescent probe 2',7'-dichlorodihydrofluorescein diacetate (DCFH-DA) and of the intracellular Ca(2+) ion concentrations [Ca(2+)](i). The removal of extracellular Ca(2+) ions did not prevent t-BOOH-induced [Ca(2+)](i) elevation, whereas the intracellular Ca(2+) ion chelator 1,2-bis(o-aminophenoxy) ethane-N,N, N',N'-tetraacetic acid (BAPTA) (10 microM) was shown to be effective. Both t-BOOH-induced free radical formation and the [Ca(2+)](i) increase were completely prevented by the peroxyl scavenger alpha-tocopherol (50 microM). t-BOOH induced a time-dependent SH-SY5Y cell injury, monitored by a 3-(4,5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT) assay (approximately 25% at 1 h, 50% at 3 h, 80% at 5 h) and by fluorescein diacetate (FDA)-propidium iodide (PI) fluorescent staining. The entity of t-BOOH-induced cell damage was the same both in the absence and in the presence of the intracellular Ca(2+) ion chelator BAPTA. By contrast, the peroxyl scavenger alpha-tocopherol (50 microM) completely prevented cell injury due to oxidative stress. Finally, superoxide dismutase (SOD) (500 ng/ml) caused a 30% reduction of t-BOOH-induced 2', 7'-dichlorofluorescein (DCF) fluorescence, whereas it did not modify the extent of cell injury produced by the oxidant. Collectively, the results of the present study demonstrated that in SH-SY5Y human neuroblastoma cells, the rise of [Ca(2+)](i) which occurs during oxidative stress is not involved in cell injury. Therefore, oxidative stress-induced cell death may be exclusively attributed to free radical overproduction.