SciELO - Scientific Electronic Library Online

vol.33 número2Free radical chemistry in biological systemsAntioxidant flavonols from fruits, vegetables and beverages: measurements and bioavailability índice de autoresíndice de assuntospesquisa de artigos
Home Pagelista alfabética de periódicos  

Biological Research

versão impressa ISSN 0716-9760

Biol. Res. v.33 n.2 Santiago  2000 

A comparison of methods employed to evaluate
antioxidant capabilities.


1Faculty of Biological Sciences, Catholic University of Chile, Santiago, Chile
2Faculty of Chemistry and Biology, University of Santiago, Santiago, Chile
3 Faculty of Sciences, University of Chile, Santiago, Chile


Thrce ditferent methodologies frequently employed to evaluate the indexes that report the antioxidant capabilities of pure compounds and/or complex mixtures of antioxidants are applied to a series of mono- and polyphenols, as well as to two wine (red and white) samples. These methodologies are based on the bleaching of a stable radical, the effect of the additive upon luminol chemiluminescence induced by peroxyl radicals, and the effect of the additive upon the bleaching of the fluorescence from a dye molecule. Widely ditferent responses are obtained from the different methodologies. These differences are interpreted in terms of the different factors (stoichiametric factors and/or reactivities) that determines the indexes evaluated by these different methodologies.

Key words: ABTS radical bleaching; antioxidants; c-phycocyanin bleaching; total antioxidant reactivity; total reactive antioxidant potential.


There is consensus that free radical reactions are relevant in many normal and pathological processes of living cells. The damage inflicted on a tissue or the degradation of valuable biomolecules by free radicals will depend upon the free radical and its production rate at the considered site (if the process is extremely fast, as in hydroxyl radical reactions) and/ or be proportional to the steady state free radical concentration at the locus of the event. This steady state concentration will be determined by chemical reactions and/ or diffusion processes, and in a general way will be given by

[Radicals] ss = Rp / k

where Rp is the rate of incorporation of free radicals to the volume considered, and k is the pseudo first order rate constant comprising all the processes that remove the radical. In order to control the steady state concentration of radical, aerobic organisms have developed a series of strategies aimed at decreasing Rp, (preventive antioxidants) and increasing the removal rate constant k (free radical scavengers). Furthermore, living organisms have developed a third line of defense comprised of repair mechanisms able to eliminate and/or repair damaged lipids, proteins and nucleic acids. The number of defense mechanisms emphasizes the importance that controlling the free radicals has for the normal development of aerobic organisms.

Scavengers (or antioxidant) molecules are entities able to transform an active free radical that can damage biomolecules in an inactive species. The presence of a scavenger compound (XH) will increase the removal rate of the active radicals (k) since

k = k° + kXH [XH]

where k° is the removal rate constant in absence of XH, and kXH is the specific rate constant for the interaction of the considered radical and the added scavenger. The introduction of k in Eqn. (1) shows that the steady state concentration of free radicals, and hence the damage associated with them, can be controlled by increasing the number and/or concentration of the scavengers. The effect elicited by a given compound will then be determined by its concentration and reactivity (measured by kXH ) to the considered radicals. A complex mixture of antioxidants, such as that found in a biological fluid (plasma), beverage or plant extracts, will have an effect in controlling the free radical steady concentration given by a pseudo constant k defined by
k = kº, + Si ki [XH]i

where the summation takes place over all the compounds with capacity to trap the considered radicals. This indicates that the effect of the mixture will depend on the quantity and quality (through ki) of the antioxidants present. Furthermore, it also depends upon the radical considered.

In principle, there are several approaches to measuring the capacity of a given compound or a complex mixture to act as a protection against the damage due to free radicals:

i) To measure the time required to deplete the antioxidant(s) when radicals are produced at a known rate. This type of approach computes the Total Radical Antioxidant Potential (TRAP) (Wayner et al., 1985). It computes stoichiometric factors a number of reactive groups (in a polyfunctional compound or in a mixture of compounds), but is unrelated to the quality of the compounds (as far as their ability to produce some sort of induction times). In principle, it has little relation to the type of radical being trapped. This type of procedure was originally proposed by Wayner et al. (1985), employing 2,2’ azobis(2-amidinopropane) (AAPH) as a free radical source and the oxygen consumption associated with the oxidation of a lipid dispersion as a measure of the rate of the oxidation process. This methodology has been also applied to other radical sources, other substrates and/or other properties of the system to monitor the rate of the oxidation process (Lissi et al., 1995; Uotila et al., 1994; Candy et al., 1990; Whitehead et al., 1992; Ghiselli et al., 1995, Winston et al., 1998). The main drawback of this type of methodology is that they only quantified stoichiometric factors and concentrations.

ii) To monitor the effect of the additive(s) on a property directly related to the steady state concentration of free radicals. These procedures give parameters related to the value of k in Eqn. (3) and are a measure of the Total Antioxidant Reactivity (TAR) of the added compounds (Lissi et al., 1995). This type of approach has been applied using AAPH as a free radical source and the luminescence from luminol as a parameter related to the steady state free radical concentration. The main drawback of this type of methodology is that they are extremely dependent upon the radicals being trapped. On the other hand, this type of index is the most relevant with respect to the additive’s ability to reduce the steady state concentration of radicals, and hence its potential damaging capability.

iii) To directly monitor the decay of a free radical (Romay et al., 1 996a). For most radicals, this requires a pulsed production of the radicals and fast techniques for the detection of the radical decay. This can be done by employing stable radicals of relatively long lifetimes, such as DPPH, galvinoxyl or the radical derived from ABTS. This type of methodology can give information regarding the quantity and reactivity of the added compounds. The main drawback is that the kinetics of the process is usually complex (Campos and Lissi, 1997; Aliaga and Lissi, l999), the answer markedly depends upon the stable free radical employed (Lissi et al., 1999) and that these radicals have characteristics of little biological interest.

iv) To integrate a property related to the substrate concentration until it is completely consumed (ORAC: Oxygen radical absorbance capacity) (Cao et al., 1993), to integrate it for a fixed time (Winston et al., 1998), or to measure it at a fixed reaction time (Miller et al., 1993). In these systems the index measured is a composite of quantities and reactivities and should provide information that is in some way intermediate between that of TRAP and TAR indexes. One drawback of this procedure is that mixtures with completely different consumption profiles can yield the same value of the measured parameter.

The foregoing discussion illustrates that different methodologies can provide completely different responses with respect to the antioxidant capacity of a pure compound or complex mixture, and this should be borne in mind when comparing results obtained by different methodologies. It also shows that a more meaningful answer could be obtained if they were applied in parallel procedures with estimations of either quantities or quantities plus reactivities. Comparison of both indexes could provide an estimation of the average reactivity of the incorporated compound(s).

In the present study, we present a comparison of antioxidative indexes evaluated by six procedures. The indexes obtained were based on I) The quenching of luminol chemiluminescence induced by the peroxyl radicals produced in the thermolysis of 2,2'-azobis(2 amidinopropane); II) The bleaching of preformed radicals derived from the oxidation of ABTS; and III) The protection of the bleaching of the fluorescence from c-phycocyanin when a solution of the protein is exposed to the peroxyl radicals generated in the thermolysis of AAPH. These methodologies were applied to a series of polyphenols and to the mixtures present in white and red wine. Trolox was employed as the reference antioxidant in all the systems considered.


I) Quenching of luminol chemiluminescence associated with its oxidation by peroxyl radicals. The experimental method employed was similar to that described previously (Lissi et al., 1995; Campos et al., 1996; Romay et al., 1 996b). When luminol is incubated in the presence of the free radical source, a steady chemiluminescence is observed that can be directly related to the rate of luminol oxidation. The addition of free radical scavengers reduces this intensity. High (micromolar) concentrations of the antioxidants almost completely quench the luminescence. The time to recover the luminescence (to a given percentage of the initial value) is related to the total amount of antioxidants present in the added sample. Comparison with the induction times elicited by Trolox allows an evaluation of the added sample TRAP. Sub-micromolar concentrations of the additive only produce an instantaneous decrease in intensity that is linearly related to the added antioxidant concentration. Comparison of this decrease with that elicited by submicromolar Trolox concentrations allows an evaluation of the TAR of the added antioxidant(s).

II) Bleaching of pre-formed ABTS derived radical cations. The methodology employed was similar to that described previously (Romay et al., 1996a; Campos et al., 1996; Campos and Lissi, 1996). The main difference was the method employed to produce the radical cation. In the present study, they were produced in the oxidation of ABTS promoted by MnO2. The addition of a few mg of MnO2 to a 100 micromolar solution of ABTS instantaneously produces the blue color characteristic of the radical cation. After centrifugation and filtering to remove the excess oxide, the absorbance of the solution remains stable for several hours. The incorporation of additives (either the pure antioxidant compounds or the wine samples) produces a decrease in absorbance (measured at 734) that is directly related to their capacity to bleach the pre-formed radical cation. For most compounds, the bleaching does not occur instantaneously, but is almost complete after 15 minutes reaction. We have evaluated the "pront" (measured ca 10 sec after addition) and total bleaching (measured after 15 minutes). A comparison of these values with that obtained instantaneously after Trolox addition provides an estimation of the total amount of compounds that react instantaneously and that react slowly with the pre-formed radical.

III) Protection of c-phycocyanin fluorescence bleaching. The methodology applied follows that described by Cao et al. (1993), with two modifications. First, c-phycocyanin was employed as the target molecule in place of b-phycoherytrin (Atanasiu et al., 1998). Second, to obtain ORAC values the fluorescence vs. time plot was integrated up to 80% bleaching of the initial fluorescence. This was performed to avoid excessively long reaction times. Furthermore, we also measured, both in absence and the presence of the antioxidant(s), the initial slope of a log (F1°/FI) vs. t plot (where F1° is the fluorescence intensity immediately after the AAPH addition, and F1 is the fluorescence at time t. Measurements were performed at 37 °C employing ca. 0.01 mg/ mL c-phycocyanin and 10 mM AAPH.


The following indexes were defined as a measure of the antioxidant action of the pure compounds and wines, in terms of equivalent micromolar concentrations of Trolox:

I) From the quenching of the luminol luminescence, TRAP and TAR values were obtained employing the following formulas:

TRAPphenol = (tphenol/tphenol ) (Trolox) / (Phenol)
TRAPwine = (twine / tTrolox) (Trolox) f
TARphenol = [(Iº/I)phenol/(Iº/I)Trolox] (Trolox)/(Phenol)
TARwine = [ (I°/l)wine / (I°/I)Trolox] f (Trolox)


t is the induction time, defined as the time required to recover 20% of the initial light intensity after almost total quenching by the addition of an excess of Trolox, phenol or wine.

I°/l is the ratio between the luminescence prior to (1°) and immediately after (I) the addition of a small amount of Trolox, phenol or wine. (Trolox) and (Phenol) are the micromolar concentrations producing the measured effect; and f is a dilution factor equal to the ratio between the total volume of the AAPH/luminol solution and the added volume of wine.

II) Fast reacting equivalents (FRE) and total reacting equivalents (TRE) were obtained from the bleaching of preformed ABTS-derived radical cations:

(FRE)Phenol = [(DºA)phenol/ DºA)Trolox)] (Trolox)/(Phenol)
(FRE)wine = [(DºA)phenol(DºA)Trolox] f (Trolox)
(TRE)Phenol = [D15A))phenol/ D15A)Trolox] (Trolox)/(Phenol)
(TRE)wine = [(D15A))phenol/ (D15A)Trolox] f (Trolox)


Dº is the decrease in absorbance (734 nm) measured immediately after the scavenger addition.D15 is the decrease in absorbance (734 nm) measured after 15 minutes addition and f is a dilution factor.

III) The following indexes were obtained from the effect of the additives on the bleaching of c-phycocyanin fluorescence:

0RAC phenol = [(Areaphenol - Area control)/(Area Trolox -
Area control)] (Trolox)/(Phenol)
0RAC wine = [(Area wine -Area control)/(Area Trolox -
Area control] f (Trolox)
TAR Phenol = {[(Slope control/Slope phenol) - 1 ]
/[(Slope control/Slope Trolox)- 1 ]} (Trolox)/(Phenol)
TAR wine = {[(Slope control/Slope wine) - 1 ]
/ [(Slope control/Slope Trolox) - 1 ] } f (Trolox)


Area represents- the area under the fluorescence vs. time plot, integrated between t=0 and the time at which 80 % of the initial fluorescence intensity was bleached.

Slope is the initial slope of a log (F1°/F1) vs time plot. The values obtained for several mono and polyphenols and for a sample of red and white wines are shown in Table I.


Experimentally determined indexes obtained by the different procedures





1 1.0 1.0 1.0 1.0 1.0 1.0


1   0.5 0.005 0.004 0.0013 -

Caffeic acid

2 0.9 1.9 1.0 1.9 1.7 3.4
Gallic acid 3 2.8


2.0 1.7 1 .3 1.1
Catechin 4 3.1 3.3 0.7 1.5 1.9 9.3
Rutin 4 1.5 4.3 1.0 2.3 2.0 2.7
Quercitin 5 2.4 4.9 3.3 3.5 1.0 3.5
Red wine - 18000 34000 7000 1 1000 8400 18000
White wine - 2500 4300 800 1200 1200 2000

n: umber of hydroxyl groups per molecule.


Before discussing the results presented in Table I, the meaning of each calculated indexes must be clarified. The TRAP, TAR and ORAC values have been discussed previously (Lissi et al., 1995; Cao et al., 1993; Romay et al., 1996b). The TRAP of a pure compound is only a measure of how many radicals each molecule can trap. In fact, this index corresponds to the ratio between the number of radicals trapped per molecule of additive and the number trapped per each Trolox molecule. It can then be expected that it will be related to the number of reactive phenol groups present in the compound. In plant extracts or beverages such as wine, the TRAP value corresponds to the Trolox concentration that produces the same induction time as the wine. In this sense, it should be related to the total concentration of reactive phenol groups in the beverage. The TRAP index is then closely related to the total bleaching of the ABTS radical cation (the TRE index). This index can be expected to be a measure of the total number of "titriable" phenol groups in the compound (TRE phenol) or in the wine sample (TRE wine) (Campodónico et al., 1998).

It is important to note that this method titrates all phenolic groups, regardless of their reactivity (Campos and Lissi, 1997). The soichiometric coefficient, which corresponds to the number of ABTS radicals removed by each titrated phenol group, can vary between one and two (Campos and Lissi, 1997). On the other hand, the TRAP index tends to compute only those groups whose reactivity is enough to completely deplete the luminol luminescence at the added concentrations. In this sense, it can be expected that mixtures of antioxidants of widely different reactivity would give higher values of TRE than TRAP. This explains the differences observed between these two indexes both for red and white wines. The TARphenol, TAR’phenol, and FREphenol indexes are somehow related to the average reactivity of the phenolic groups of the added compound. However, it is to be expected that rather large differences between indexes based on trapping peroxyl radicals (TAR and TAR’) and that based on the scavenging of the ABTS derived radicals, due to the complexity of the reactions involving stable radicals (Campos and Lissi, 1997; Lissi et al., 1999; Aliaga and Lissi, 1999). The same considerations hold when these indexes are applied to the evaluation of the total content of antioxidants in complex samples.

The meaning of the ORAC index is less straightforward. For very reactive compounds, it tends to quantify the number of groups and thus is similar to TRAP indexes. For less reactive groups, it tends to be strongly influenced by the reactivity of the groups and therefore should be related to TAR indexes. For intermediate compounds and/or for complex mixtures of compounds (or when several groups of different reactivities are present), the meaning of the index is intermediate between those of the TRAP and TAR indexes.

There is another noticeable difference between the two types of indexes, those measured instantaneously, such as TAR, TAR’ and FRE and those integrated over a longer time period (ORAC, TRAP and TRE). While the former measures the reactivity of the initially-added compound (or the added mixture), the latter can be influenced by the reactions of the products formed in the reaction of the most reactive groups and/or compounds present in the added sample.

Analysis of the data presented in Table I suggests the following conclusions:

a) TRE values correlate with the number of phenolic groups (r2 = 0.949, slope 1.025 ± 0.106, n = 7). This index is extremely insensitive to the reactivity of the compounds. Even tyrosine, a compound with very low reactivity, tends to give a TRE value per reactive group that is similar to the other compounds. The low reactivity of this compound is emphasized by the fact that FRE and TRAP values could not be determined. This supports previous results that indicate that methods based on the bleaching of ABTS radical cations almost constitute an alternative means of phenol groups titration (Campos and Lissi, 1997;Campodónico et al., 1998). Using this methodology, some insight can be gained with respect to the reactivity of the phenol groups by measuring the instantaneous bleaching of the ABTS radical absorbance. On the other hand, ORAC, TAR and TAR’ indexes are more closely related to the reactivity of the groups (see the low values obtained for tyrosine).

b) Although TAR’ and ORAC values are somehow related (r2= 0.703, slope 0.82 ± 0.24, n = 7), there are noticeable differences between the two. For example, in catechin, TAR’ is less than one, while the ORAC index is greater than one. This can be explained in terms of the presence of more groups of smaller reactivity in catechin (compared with Trolox).

c) TRAP values, although determined by the number of groups and stoichiometric factors, are poorly correlated (r2 = 0.21, slope 0.92 ± 0.91, n = 6) with the number of phenolic groups in the molecule.

d) All the methods indicate that in caffeic acid, the reaction of the first group does not preclude the antioxidant action of the second group. This conclusion cannot be extended to gallic acid.

e) There is not a good correlation between TAR and TAR’ values. This is rather surprising, as both indexes should be determined by the reactivity of the compounds toward peroxyl radicals. However, since the TAR value can be influenced by the trapping of the luminol derived radicals (Lissi et al., 1992), the lack of correlation could be related to differences in the radicals being trapped in both methodologies.

We consider that the most direct evaluation of the reactivity of a compound and/or the value of the pseudo first order constant k (from Eqn 3) in wine samples can be derived from the values of the TRAP’ index, in which the most active polyphenol is quercitin. In fact, in this assay, quercitin was 660 times more efficient than tyrosine. This is in agreement with the structures of the compounds evaluated in the present work done by Lien et al., 1999.

f) As expected, there are large differences in the antioxidant capacity of wine as measured by the different methodologies. The values for red wine range from 34000 (TRE) to 7000 (TRAP’) and from 4300 (TRE) to 850 (TRAP’) for white wine. A comparison of values would indicate that the phenol groups titrated by the TRE, ORAC and TRAP methodologies are, on average, slightly less reactive than the phenol group in Trolox.

g) Regardless of the employed methodology, red wine has approximately 7 times more antioxidants than white wines. This implies that the average reactivity of the phenol groups is similar in both types of beverage. The differences between the antioxidant indexes of the two types of wine is therefore due to the quantity and not to the quality of the antioxidants present in both beverages.

It is then relevant to ask which index of antioxidant activity of a given compound or complex mixture is most relevant. There is no single answer to this fundamental question. If we are interested in protecting a sample for a long period of time during which the added antioxidant can be consumed, the indexes related to the total amount of reactive antioxidants (TRE, ORAC or TRAP) will be more appropriate. The same applies for the situation of strong oxidative stress when the antioxidants should be preserved to minimize the damage. If we are interested in knowing how a given concentration of antioxidants would protect other valuable molecules, then indexes emphasizing reactivity (TAR or TAR’) should be preferred. In any case, we hold that this type of evaluation should only be considered as a rough estimate of the capacities of antioxidants and that multiple indexes should be employed in an attempt to separate quantities (and stoichiometric factors) from reactivities, since both types of information are necessary to asses the value and extent of the antioxidant action.

Corresponding Author: Dr. Eduardo Lissi. Departamento de Química, Facultad de Química y Biología, Universidad de Santiago de Chile. Casilla 40, Correo 33, Santiago, Chile. FAX: 56-2-6812108, E-mail:

Received: December 30, 1999. Accepted: December 30, 1999


This study was supported by the PUC-PMC98 program (Molecular Basis of Chronic Diseases Program, Catholic University of Chile) and by DICYT (University of Santiago).


ALLIAGA C, LISSI EA (1998) Reaction of ABTS derived radicals with hydroperoxides. Kinetics and Mechanism. Int J Chem Kinet 30: 565-570         [ Links ]

ATANASIU RL STEA D, MATEESCU MA, VERGELY C, DALLOZ F, BRIOT F, MAUPOIL V, NADEAU R, ROCHETTE L, (1998) Direct evidence of caeruloplasmin antioxidant properties. Mol Cell Biochem 189: 127- 135         [ Links ]

CAMPODONICO P, BARBIERI E, PIZARRO M, SOTOMAYOR CP, LISSI EA (1998) A comparison between total phenol content of wines and their TRAP values measured by the bleaching of ABTS radical cations. Bol Soc Chil Quim 43: 281-285         [ Links ]

CAMPOS AM, ESCOBAR J, LISSI EA (1996) The total reactive antioxidant potential (TRAP) and total antioxidant activity (TAR) of Ilex paraguayensis extracts and red wine. J. Braz. Chem. Soc. 7: 43-49         [ Links ]

CAMPOS AM, LISSI, EA (1996) Total antioxidant potential of chilean wines. Nutrition Res. 16: 385-389         [ Links ]

CAMPOS AM, LISSI, EA (1997). Kinetics of the reaction between 2,2'-azinobis (3-ethylbenzothiazoline-6-sulfonic acid (ABTS) derived radical cations and phenols. Int J Chem Kinetics 29: 219-224         [ Links ]

CANDY TEG, HODGSON M, JONES, P (1990) Kinetics and mechanisms of a chemiluminescent clock reaction based on the horseradish peroxidase catalyzed oxidation of luminol by hydrogen peroxide. J Chem Soc Perkin Trans 2: 1385-1388         [ Links ]

CAO G, ALESSIO HM, CUTLER RG (1993) Oxygen radical absorbance capacity assay for antioxidants. Free Rad Biol Med 14: 303-31 1         [ Links ]

CELEDON C, FUENTES M, LISSI EA (1999) Global modification promoted by different radical sources. Role of heme group. Biol Res 32: R-197         [ Links ]

GHISELLI A, SERAFINI M, MIAIANI G, AZZINI E (1995). A fluorescence-based method of measuring total plasma antioxidant capability. Free Radical Biol Med 18: 29-36         [ Links ]

LIEN EJ, REN S, BUL H-H, WANG R (1999) Quantitative structure-activity relationship analysis of phenolic antioxidants. Free Rad Biol Med 26: 285-294         [ Links ]

LISSI EA, PASCUAL C, DEL CASTILLO ML (1992) Luminol luminescence induced by 2,2’ azobis (2 amidinopropane) thermolysis. Free Radical Res. Comms 17:299-306         [ Links ]

LISSI EA, SALIM-HANNA, M., PASCUAL C, DEL CASTILLO MD (1995) Evaluation of total antioxidant potential (TRAP) and total antioxidant reactivity (TAR) from luminol enhanced chemiluminescence measurements. Free Rad Biol Med 18: 153-158         [ Links ]

LISSI EA, MODAK B, TORRES R, ESCOBAR J, URZUA A (1999) Total antioxidant potential of resinous exudate from Heliptropium Species, and a comparison of the ABTS and DPPH methods. Free Rad Res 30: 471-477         [ Links ]

MILLER NJ, RICE-EVANS C, DAVIES MJ, GOPINATHAN V, MILNER A (1993) A novel method for measuring antioxidant capacity and its application to monitoring the antioxidant status in premature neonates. Clin. Sci 84: 407-412         [ Links ]

ROMAY C, DEL CASTILLO MC, PASCUAL C, CAMPOS AM, ESCOBAR J, LISSI EA (1996b) Evaluation of the total content of antioxidants in complex mixtures. Ciencia e Cultura 48: 86-95         [ Links ]

ROMAY C, PASCUAL C, LISSI EA (1996a) The reaction between ABTS radical cations and antioxidants and its use to evaluate the antioxidant status of serum samples. Braz J Med Biol Res 29: 175-185         [ Links ]

UOTILA JT, KIRKKOLA A-L, RORARIU M., TUIMALA RJ, METSA-KETELA T (1994) The total peroxyl radical-trapping ability of plasma and cerobrospinal fluid in normal and preclamptic parturients. Free Rad Biol Med 16: 581-590         [ Links ]

WAYNER DD, BURTON GW, INGOLD KU, LOCKE S (1985) Quantitative measurement of the total peroxyl radical-trapping antioxidant capability of human blood plasma by controlled lipid peroxidation. FEBS Lett 187: 33-37         [ Links ]

WHITEHEAD TP, THORPE GHG, MAXWELL, SRJ (1992) Enhanced chemiluminescent assay for antioxidant capacity in biological fluids. Anal Chim Acta 266: 265-277         [ Links ]

WINSTON GW, REGOLI F, DUGAS AJ, FONG .IH, BLANCHARD KA (1998) A rapid gas chromatographic assay for determining oxyradical scavenging capacity of antioxidants and biological fluids. Free Rad Biol Med. 24: 480-491         [ Links ]