Biogerontology

, Volume 9, Issue 3, pp 183–196

Forty percent and eighty percent methionine restriction decrease mitochondrial ROS generation and oxidative stress in rat liver

Authors

  • Pilar Caro
    • Departamento de Fisiología Animal-II, Facultad de Ciencias BiológicasComplutense University
  • José Gómez
    • Departamento de Fisiología Animal-II, Facultad de Ciencias BiológicasComplutense University
  • Mónica López-Torres
    • Departamento de Fisiología Animal-II, Facultad de Ciencias BiológicasComplutense University
  • Inés Sánchez
    • Departamento de Fisiología Animal-II, Facultad de Ciencias BiológicasComplutense University
  • Alba Naudí
    • Department of Experimental MedicineUniversity of Lleida-IRBLLEIDA
  • Mariona Jove
    • Department of Experimental MedicineUniversity of Lleida-IRBLLEIDA
  • Reinald Pamplona
    • Department of Experimental MedicineUniversity of Lleida-IRBLLEIDA
    • Departamento de Fisiología Animal-II, Facultad de Ciencias BiológicasComplutense University
Research Article

DOI: 10.1007/s10522-008-9130-1

Cite this article as:
Caro, P., Gómez, J., López-Torres, M. et al. Biogerontology (2008) 9: 183. doi:10.1007/s10522-008-9130-1

Abstract

Dietary restriction (DR) lowers mitochondrial reactive oxygen species (ROS) generation and oxidative damage and increases maximum longevity in rodents. Protein restriction (PR) or methionine restriction (MetR), but not lipid or carbohydrate restriction, also cause those kinds of changes. However, previous experiments of MetR were performed only at 80% MetR, and substituting dietary methionine with glutamate in the diet. In order to clarify if MetR can be responsible for the lowered ROS production and oxidative stress induced by standard (40%) DR, Wistar rats were subjected to 40% or 80% MetR without changing other dietary components. It was found that both 40% and 80% MetR decrease mitochondrial ROS generation and percent free radical leak in rat liver mitochondria, similarly to what has been previously observed in 40% PR and 40% DR. The concentration of complexes I and III, apoptosis inducing factor, oxidative damage to mitochondrial DNA, five different markers of protein oxidation, glycoxidation or lipoxidation and fatty acid unsaturation were also lowered. The results show that 40% isocaloric MetR is enough to decrease ROS production and oxidative stress in rat liver. This suggests that the lowered intake of methionine is responsible for the decrease in oxidative stress observed in DR.

Keywords

MitochondriaCaloric restrictionAgingOxygen radicalsLongevityProtein damageFatty acidsRespiratory complexes

Abbreviations

AIF

Apoptosis inducing factor

AASA

Aminoadipic semialdehyde in proteins

CEL

Carboxyethyl-lysine in proteins

CML

Carboxymethyl-lysine in proteins

DBI

Double bond index

DR

Dietary restriction

PI

Peroxidizability index

PR

Protein restriction

GSA

Glutamic semialdehyde in proteins

MDAL

Malondialdehyde-lysine in proteins

MetR

Methionine restriction

ROS

Reactive oxygen species

8-oxodG

8-Oxo-7,8-dihydro-2′deoxyguanosine

Introduction

The mitochondrial oxygen radical theory of aging is currently supported by both comparative and dietary restriction (DR) studies (Beckman and Ames 1998; Barja 2004). These investigations point to a main role of the mitochondrial rate of generation of reactive oxygen species (ROS) as the main source of oxidative molecular damage contributing to aging. It is known that the rate of mitochondrial ROS (mitROS) production is lower in long-lived than in short-lived animal species (Barja 2004; Robert et al. 2007; Lambert et al. 2007) and that DR consistently decreases that rate (reviewed in Gredilla and Barja 2005). In these two models, low rates of mitROS generation are accompanied by lower levels of oxidative damage to macromolecules including mitochondrial DNA (mtDNA), proteins and lipids (Barja 2004; Gredilla et al. 2001a; Pamplona et al. 2002a).

Recent studies from our laboratory have contributed to clarify what is the dietary factor responsible for the decrease in mitROS production and oxidative damage to mitochondrial macromolecules during DR. Thus, whereas neither lipid restriction (Sanz et al. 2006a) nor carbohydrate restriction (Sanz et al. 2006b) change mitROS production, protein restriction (PR) (Sanz et al. 2004) decreases mitROS generation and oxidative stress in rat liver in a way quantitatively and qualitatively similar to that induced by DR. These decreases were specifically related to the lowered protein ingestion and were not due to restriction of energy intake (Gómez et al. 2007). These results indicate that restriction of protein intake is responsible for the well-known decreases in mitROS production and oxidative stress that take place in DR. On the other hand, although it is classically believed that the anti-aging effect of DR is due to the decreased intake of calories themselves rather than to decreases in specific dietary components, recent findings question this consensus (Mair et al. 2005). Moreover, classic studies have repeatedly found that PR also increases maximum longevity in rats and mice (reviewed in Pamplona and Barja 2006), although the magnitude of these increases is around 30–40% that of DR, whereas neither carbohydrate (Khorakova et al. 1990) nor lipid restriction (Iwasaki et al. 1988; Shimokawa et al. 1996) seem to modify rodent longevity. Interestingly, it has been observed that methionine restriction (MetR) without energy restriction also increases maximum longevity in rats and mice (Richie Jr et al. 1994; Zimmerman et al. 2003; Miller et al. 2005) and delays or avoids many age-related changes in rodents (Miller et al. 2005; Malloy et al. 2006; Pavillard et al. 2006). Using the same dietary protocol with which it was demonstrated that MetR increases rodent longevity (80% MetR substituting methionine for glutamate in the diet, Richie Jr et al. 1994; Miller et al. 2005), we have recently found that MetR also decreases mitROS generation and oxidative stress in a very similar way to both PR and DR (Sanz et al. 2006c). This suggests that methionine could be the single dietary component responsible for those changes during 40% DR. However in those previous studies the decrease in mitROS production was demonstrated only at 80% MetR. In order to attribute to dietary methionine the decrease in mitROS generation at (40%) DR, it must be demonstrated that the decreases also occur at 40% MetR. In addition, the decrease in dietary methionine was compensated by adding glutamate to the diet (Sanz et al. 2006c). Thus, it is theoretically possible that the changes observed are due to glutamate supplementation instead of to MetR. In fact, comparing our previous MetR experiment (Sanz et al. 2006c) with those of PR (Sanz et al. 2004) or DR (Gredilla et al. 2001b), we found a discrepancy since mitochondrial oxygen consumption was increased by such kind of MetR (Sanz et al. 2006c) but not by the other two manipulations. It is possible that such increase in oxygen consumption is due to the glutamate dietary supplementation.

In order to overcome those limitations, in this investigation Wistar rats were subjected to both 40% and 80% MetR during a time period (6–7 weeks) that is known to decrease mitROS generation and oxidative stress in both PR (Sanz et al. 2004) and DR (Gredilla et al. 2001b). In addition, the dietary experiments were performed in a way that changed only dietary methionine, whereas all the other dietary components or calories were held at similar levels in the three dietary groups (controls, 40 and 80% MetR). In liver mitochondria of these methionine restricted rats we have measured the rates of mitROS generation, mitochondrial oxygen consumption in resting and phosphorylating states, percent free radical leak (FRL), the content of complexes I, II, III and IV of the respiratory chain, apoptosis inducing factor (AIF), the marker of mtDNA oxidative damage 8-oxo-7,8-dihydro-2′deoxyguanosine (8-oxodG), and five highly specific markers of protein oxidation [the specific protein carbonyls glutamic semialdehyde (GSA) and aminoadipic semialdehyde (AASA)], glycoxidation [carboxyethyl-lysine (CEL) and carboxymethyl-lysine (CML)] and lipoxidation [malondialdehyde-lysine (MDAL) and CML]. Since it is known that the degree of unsaturation of phospholipids can affect markers of protein lipoxidation, the full fatty acid composition of liver mitochondria was also measured.

Materials and methods

Animals and diets

Male Wistar rats of 250–300 g of body weight were caged individually and maintained in a 12:12 (light:dark) cycle at 22 ± 2°C and 50 ± 10% relative humidity. Semipurified diets prepared by MP Biochemicals (Irvine, CA, USA) were used. The detailed composition of the three diets is shown in Table 1. The composition of the 40% and 80% MetR diets was similar to that of the control diet except that l-methionine was present at 0.516% and 0.172%, which corresponds to amounts 40% and 80% lower than the l-methionine content of the control diet (0.86%). The % decrease in l-methionine in the 40% and 80% MetR diets was compensated with increases in all the rest of the dietary components in proportion to their presence in the diet. Since the % absolute decrease in l-methionine was small, with this procedure the % presence of all the rest of the dietary components was almost the same in the three experimental diets (Table 1). The control and the 40% MetR groups received each day the same amount of food that the 80% MetR animals had eaten as a mean the previous week (pair feeding). Daily visual inspection of the rat cages indicated that there were no differences in food spillage between control and MetR animals. After 6–7 weeks of dietary treatment the animals were sacrificed by decapitation. The liver was immediately processed to isolate mitochondria, which were used to measure mitochondrial respiration and H2O2 generation, and mitochondrial and liver samples were stored at −80°C for the assay of the rest of the biochemical parameters.
Table 1

Detailed composition of the three different diets used in this study (control, 40% and 80% methionine restricted)

Component

Control (AIN93G) (g/100 g)

40% MetR (g/100 g)

80% MetR (g/100 g)

l-Arginine

1.12

1.124

1.128

l-Lysine

1.44

1.445

1.450

l-Histidine

0.33

0.331

0.332

l-Leucine

1.11

1.114

1.118

l-Isoleucine

0.82

0.823

0.826

l-Valine

0.82

0.823

0.826

l-Threonine

0.82

0.823

0.826

l-Tryptophan

0.18

0.181

0.181

l-Methionine

0.86

0.516

0.172

l-Glutamic acid

2.70

2.709

2.719

l-Phenylalanine

1.16

1.164

1.168

l-Glycine

2.33

2.338

2.346

Dextrine

5.00

5.017

5.035

Corn starch

43.61

43.762

43.911

Sucrose

20.0

20.069

20.139

Cellulose

5.0

5.017

5.035

Choline bitartrate

0.2

0.201

0.201

Vitamin mix (AIN)

1.0

1.003

1.007

Mineral mix (AIN)

3.5

3.512

3.524

Corn oil

8.0

8.028

8.056

Total

100

100

100

Liver mitochondria isolation

Liver mitochondria were obtained from fresh tissue. The liver was rinsed and homogenized in 60 ml of isolation buffer (210 mM mannitol, 70 mM sucrose, 5 mM Hepes, 1 mM EDTA, pH 7.35). The nuclei and cell debris were removed by centrifugation at 1,000×g for 10 min. Supernatants were centrifuged at 10,000×g for 10 min and the resulting supernatants were eliminated. The pellets were resuspended in 40 ml of isolation buffer without EDTA and centrifuged at 1,000×g for 10 min. Mitochondria were obtained after centrifugation of the supernatants at 10,000×g for 10 min. After each centrifugation step any overlaying layer of fat was eliminated. The mitochondrial pellets were resuspended in 1 ml of isolation buffer without EDTA. All the above procedures were performed at 5°C. Mitochondrial protein was measured by the Biuret method. The final mitochondrial suspensions were maintained over ice and were immediately used for the measurements of oxygen consumption and H2O2 production.

Mitochondrial H2O2 generation

The rate of mitochondrial H2O2 production was assayed by measuring the increase in fluorescence (excitation at 312 nm, emission at 420 nm) due to oxidation of homovanillic acid by H2O2 in the presence of horseradish peroxidase essentially as described (Barja 2002; Sanz and Barja 2006). Reaction conditions were 0.25 mg of mitochondrial protein per milliliter, 6 U/ml of horseradish peroxidase, 0.1 mM homovanillic acid, 50 U/ml of superoxide dismutase and 2.5 mM pyruvate/2.5 mM malate, 2.5 mM glutamate/2.5 mM malate or 5 mM succinate + 2 μM rotenone as substrates, added at the end (to start the reaction) to the incubation buffer (145 mM KCl, 30 mM Hepes, 5 mM KH2PO4, 3 mM MgCl2, 0.1 mM EGTA, 0.1% albumin, pH 7.4) at 37°C, in a total volume of 1.5 ml. Unless otherwise stated, the assays with succinate as substrate were performed in the presence of rotenone in order to avoid the backwards flow of electrons to complex I. In some experiments rotenone (2 μM) or antimycin A (2 μM) were additionally included in the reaction mixture to assay maximum rates of complex I, or complex III H2O2 generation respectively. Duplicated samples were incubated for 15 min at 37°C. The reaction was stopped by transferring the samples to a cold bath and adding 0.5 ml of stop solution (2.0 M glycine, 2.2 M NaOH, 50 mM EDTA, pH 12), and the fluorescence was read in a LS50B Perkin-Elmer fluorometer. Known amounts of H2O2 generated in parallel by glucose oxidase with glucose as substrate were used as standards. Since the superoxide dismutase added in excess converts all O2•− excreted by mitochondria (if any) to H2O2, the measurements represent the total (O2•− + H2O2) rate of mitROS production.

Mitochondrial oxygen consumption

The rate of oxygen consumption of liver mitochondria was measured at 37°C in a water-thermostatized incubation chamber with a computer-controlled Clark-type O2 electrode (Oxygraph, Hansatech, UK) in 0.5 ml of the same incubation buffer used for the H2O2 measurements. The substrates used were complex I- (2.5 mM pyruvate/2.5 mM malate or 2.5 mM glutamate/2.5 mM malate) or complex II-linked (5 mM succinate + 2 μM rotenone). The assays were performed in the absence (state 4-resting) and in the presence (state 3-phosphorylating) of 500 μM ADP.

Mitochondrial free radical leak

H2O2 production and O2 consumption were measured in parallel in the same liver mitochondria under similar experimental conditions. This allowed the calculation of the fraction of electrons out of sequence which reduce O2 to ROS at the respiratory chain (the percent FRL) instead of reaching cytochrome oxidase to reduce O2 to water. Since two electrons are needed to reduce 1 mol of O2 to H2O2, whereas four electrons are transferred in the reduction of 1 mol of O2 to water, the percent FRL was calculated as the rate of H2O2 production divided by two times the rate of O2 consumption, and the result was multiplied by 100. The lower the FRL, the higher is the efficiency of the mitochondria in avoiding oxygen radical generation at the respiratory chain.

Measurement of mitochondrial complexes I, II, III and IV and AIF by western blot

The content of mitochondrial complexes I to IV and AIF was estimated using western blot analysis. Immunodetection was performed using a monoclonal antibody specific for the NDUFA9 subunit of complex I, the subunit 70 kDa of complex II, the subunit Core II of complex III and the subunit I of complex IV (1:1,000, 1:500, 1:1,000 and 1:1,000, respectively; Molecular Probes, Invitrogen Ltd., UK), and a polyclonal antibody for C-terminus (amino acids 593–613) of AIF (1:1,000; Sigma, Madrid, Spain). An antibody to porin (1:5,000, Molecular Probes, Invitrogen Ltd., UK) was also used in order to determine the proportion of complex I, II, III and IV and AIF referred to total mitochondrial mass. Appropriate peroxidase-coupled secondary antibodies and chemiluminescence HRP substrate (Millipore, MA, USA) were used for primary antibody detection. Signal quantification and recording was performed with a ChemiDoc Bio-Rad equipment (Bio-Rad Laboratories, Inc., Barcelona, Spain). Protein concentration was determined by the Bradford method. Data was expressed as arbitrary units.

Oxidation-derived protein damage markers measurements by GC/MS

Glutamic semialdehyde, AASA, CML, CEL and MDAL were determined as trifluoroacetic acid methyl esters derivatives in acid hydrolyzed delipidated and reduced mitochondrial protein samples by GC/MS using an isotope dilution method as previously described (Pamplona et al. 2005) using a 6890 Series II gas chromatograph (Agilent, Barcelona, Spain) with a 5973A Series Mass Selective Detector and a 7683 Series automatic injector, a 30-m × 0.25-mm × 0.25-μm HP-5MS column, and the described temperature program (Pamplona et al. 2005). Quantification was performed by external standardization using standard curves constructed from mixtures of deuterated and non-deuterated standards. Analyses were carried out by selected ion-monitoring GC/MS (SIM-GC/MS). The ions used were: lysine and [2H8]lysine, m/z 180 and 187, respectively; 5-hydroxy-2-aminovaleric acid and [2H5]5-hydroxy-2-aminovaleric acid (stable derivatives of GSA), m/z 280 and 285, respectively; 6-hydroxy-2-aminocaproic acid and [2H4]6-hydroxy-2-aminocaproic acid (stable derivatives of AASA), m/z 294 and 298, respectively; CML and [2H4]CML, m/z 392 and 396, respectively; CEL and [2H4]CEL, m/z 379 and 383, respectively; and MDAL and [2H8]MDAL, m/z 474 and 482, respectively. The amounts of product were expressed as the micromolar ratio of GSA, AASA, CML, CEL or MDAL/mole lysine.

Oxidative damage to mtDNA

Isolation of mtDNA was performed by the method of Latorre et al. (1986) adapted to mammals (Asunción et al. 1996). The isolated mtDNA was digested to deoxynucleoside level by incubation at 37°C with 5 U of nuclease P1 (in 20 μl of 20 mM sodium acetate, 10 mM ZnCl2, 15% glycerol, pH 4.8) for 30 min and 4 U of alkaline phosphatase (in 20 μl of 1 M Tris–HCl, pH 8.0) for 1 h (Loft and Poulsen 1999). Steady-state oxidative damage to mtDNA was estimated by measuring the level of 8-oxo-7,8-dihydro-2′deoxyguanosine (8-oxodG) referred to that of the non-oxidized base (deoxyguanosine, dG). 8-OxodG and dG were analyzed by HPLC with on line electrochemical and ultraviolet detection respectively. The nucleoside mixture was injected into a reverse-phase Beckman Ultrasphere ODS column (5 μm, 4.6 mm × 25 cm), and was eluted with a mobile phase containing 2.5% acetonitrile and 50 mM phosphate buffer pH 5.0. A Waters 510 pump at 1 ml/min was used. 8-OxodG was detected with an ESA Coulochem II electrochemical coulometric detector (ESA, Inc., Bedford, MA, USA) with a 5011 analytical cell run in the oxidative mode (400 mV/20 nA), and dG was detected with a Bio-Rad model 1806 UV detector at 254 nm. For quantification peak areas of dG standards and of three level calibration pure 8-oxodG standards (Sigma) were analyzed during each HPLC run. Comparison of areas of 8-oxodG standards injected with and without simultaneous injection of dG standards ensured that no oxidation of dG occurred during the chromatography.

Fatty acid analysis by GC/MS

Fatty acyl groups in brain lipids were analyzed as methyl esters derivatives by GC/MS as previously described (Pamplona et al. 2005). Separation was performed in a SP2330 capillary column (30 m × 0.25 mm × 0.20 μm) in a Hewlett Packard 6890 Series II gas chromatograph. A Hewlett Packard 5973A mass spectrometer was used as detector in the electron-impact mode. Identification of fatty acyl methyl esters was made by comparison with authentic standards and on the basis of mass spectra. Results are expressed as mol%. The following fatty acyl indices were also calculated: saturated fatty acids (SFA); unsaturated fatty acids (UFA); monounsaturated fatty acids (MUFA); polyunsaturated fatty acids from n-3 and n-6 series (PUFAn-3 and PUFAn-6); average chain length = {[(Σ%Total14 × 14) + (Σ%Total16 × 16) + (Σ%Total18 × 18) + (Σ%Total20 × 20) + (Σ%Total22 × 22)]/100}; double bond index (DBI) = [(1 × Σmol% monoenoic) + (2 × Σmol% dienoic) + (3 × Σmol% trienoic) + (4 × Σmol% tetraenoic) + (5 × Σmol% pentaenoic) + (6 × Σmol% hexaenoic)] and peroxidizability index (PI) = [(0.025 × Σmol% monoenoic) + (1 × Σmol% dienoic) + (2 × Σmol% trienoic) + (4 × Σmol% tetraenoic) + (6 × Σmol% pentaenoic) + (8 × Σmol% hexaenoic)].

Statistics

Data were analyzed by one way ANOVA. After the ANOVA, the Duncan test was performed for comparisons between pairs of groups. The minimum level of statistical significance was set at P < 0.05 in all the analyses.

Results

The mitochondrial rate of free radical generation with pyruvate/malate as substrate was significantly decreased both at 40% and 80% MetR compared to controls, and was also significantly lower at 80% than at 40% MetR (Fig. 1). With glutamate/malate both levels of MetR also decreased ROS generation and the level was again significantly lower in 80% than in 40% MetR (Fig. 1). When succinate was used as substrate both 40% and 80% MetR decreased mitROS production but no differences between both restricted groups were observed, and the magnitude of the decrease was much smaller than with the complex I-linked substrates (Fig. 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs10522-008-9130-1/MediaObjects/10522_2008_9130_Fig1_HTML.gif
Fig. 1

Rates of H2O2 generation of liver mitochondria in control, 40% and 80% methionine restricted rats in the presence of pyruvate/malate, glutamate/malate or succinate as substrates. Control values: 0.08 ± 0.01 (pyruvate/malate), 0.19 ± 0.02 (glutamate/malate), 0.27 ± 0.01 (succinate). The ANOVA resulted in significant differences with the three substrates (P < 0.001 with glu/mal and P < 0.01 with the other substrates). Asterisks represent significant differences: a* (compared to control group), b* (between 40% and 80% MetR groups); *P < 0.05, **P < 0.01. Values are mean ± SEM from ten different animals

The maximum rates of ROS generation were measured using appropriate combinations of substrates and inhibitors of the respiratory chain (Table 2). Full reduction of complex I was performed adding rotenone with pyruvate/malate or glutamate/malate (complex I-linked substrates). Although the differences did not reach statistical significance with pyruvate/malate + rotenone, when glutamate/malate + rotenone was used significant decreases in the 40% MetR and in the 80% MetR groups were observed in comparison with the controls (Table 2). No significant differences between groups were observed for maximum ROS production with succinate + antimycin A (full reduction of complex III).
Table 2

Maximum rates of H2O2 production (nanomoles H2O2/min mg protein) in the presence of respiratory chain inhibitors in liver mitochondria from control, 40% and 80% methionine restricted rats

Substrate

Control

40% MetR

80% MetR

Pyr/mal + rotenone

0.31 ± 0.02

0.28 ± 0.04

0.25 ± 0.03

Glu/mal + rotenone

0.44 ± 0.04

0.30 ± 0.03**

0.27 ± 0.03**

Succinate + antimycin A

1.93 ± 0.13

1.88 ± 0.15

1.83 ± 0.04

Values are mean ± SEM from ten different animals

Pyr/mal pyruvate/malate, Glu/mal glutamate/malate

** Significant differences compared to the control group (P < 0.01)

The rates of oxygen consumption of rat liver mitochondria are shown in Table 3. Oxygen consumption in state 3 (phosphorylating) was various fold higher than in state 4 (resting) with all the substrates used, indicating the tight coupling of the mitochondrial preparations. No significant differences between dietary groups were detected with any substrate in any mitochondrial state.
Table 3

Rates of oxygen consumption (nanomoles O2/min mg protein) of liver mitochondria from control, 40% and 80% methionine restricted rats

Substrate

Control

40% MetR

80% MetR

Pyr/mal (state 4)

5.4 ± 0.6

6.2 ± 0.8

5.6 ± 0.7

Pyr/mal (state 3)

17.1 ± 1.3

19.1 ± 1.8

15.6 ± 1.0

Glu/mal (state 4)

8.8 ± 0.8

9.1 ± 1.3

9.8 ± 1.1

Glu/mal (state 3)

94.4 ± 5.8

96.6 ± 6.0

98.0 ± 8.1

Succinate (state 4)

30.9 ± 2.4

30.7 ± 2.3

32.1 ± 2.6

Succinate (state 3)

120 ± 9.1

123 ± 9.6

122 ± 9.4

Values are mean ± SEM from ten different animals

Pyr/mal pyruvate/malate, Glu/mal glutamate/malate

The FRL with pyruvate/malate was significantly smaller in both MetR groups than in the controls (Fig. 2). The same was true with glutamate/malate and, in addition, with this substrate significantly lower FRL values were detected in 80% MetR than in 40% MetR. With succinate, however, no significant differences were found between dietary groups.
https://static-content.springer.com/image/art%3A10.1007%2Fs10522-008-9130-1/MediaObjects/10522_2008_9130_Fig2_HTML.gif
Fig. 2

Free radical leak (%) of liver mitochondria in control, 40% and 80% methionine restricted rats in the presence of pyruvate/malate, glutamate/malate or succinate as substrates. The ANOVA resulted in significant differences with the complex I-linked substrates pyruvate/malate (P < 0.02) and glutamate/malate (P < 0.0006) but not with succinate. Asterisks represent significant differences: a* (compared to control group), b* (between 40% and 80% MetR groups); *P < 0.05, **P < 0.01. Values are mean ± SEM from ten different animals

The concentration of the protein complexes of the respiratory chain and AIF are shown in Fig. 3. Significantly lower levels were found for all these parameters both in 40% MetR and in 80% MetR compared to the control group, with no significant differences detected between 40% and 80% MetR. The strongest decreases (70%) were observed for complex II, and the smallest for complex III (15–20%), and the other parameters decreased by 30–55%.
https://static-content.springer.com/image/art%3A10.1007%2Fs10522-008-9130-1/MediaObjects/10522_2008_9130_Fig3_HTML.gif
Fig. 3

Concentration of the respiratory chain protein complexes I to IV and AIF in liver mitochondria of control, 40% and 80% methionine restricted rats. Control values (ratio of complex I, II, III, IV or AIF/porin): 2.16 ± 0.08 (CX I), 0.39 ± 0.08 (CX II), 1.21 ± 0.02 (CX III), 1.34 ± 0.11 (CX IV), 0.90 ± 0.07 (AIF). Asterisks represent significant differences compared to the control group: a**P < 0.01, a***P < 0.001. Values are mean ± SEM from ten different animals

Since complex I and III contain the mitochondrial free radical generators, the values of ROS production shown in Fig. 1 were also referred to the levels of complex I and complex III proteins. ROS generation/complex I values were 0.057 ± 0.006 in controls, 0.032 ± 0.009 in 40% MetR and 0.020 ± 0.006 in 80% MetR. Both the ratio in 40% MetR (P < 0.04) and in 80% MetR (P < 0.07) were significantly lower than in controls, whereas no differences were found between 40% and 80% MetR (P < 0.30). ROS generation/complex III values were 0.102 ± 0.012 in controls, 0.044 in 40% MetR and 0.027 ± 0.008 in 80% MetR. In this case it was also found that the ratio in 40% MetR (P < 0.03) and in 80% MetR (P < 0.01) were significantly lower than in the controls, whereas no significant differences were found between 40 and 80% MetR (P < 0.28).

The five different specific markers of protein oxidation, glycoxidation and lipoxidation are shown in Fig. 4. GSA, AASA, CML and MDAL showed significantly lower values both in 40% and 80% MetR than in the controls. CEL was significantly lower in the 80% MetR group compared both to controls and to 40% MetR.
https://static-content.springer.com/image/art%3A10.1007%2Fs10522-008-9130-1/MediaObjects/10522_2008_9130_Fig4_HTML.gif
Fig. 4

Protein markers of oxidative, glycoxidative and lipoxidative damage in liver mitochondria of control, 40% and 80% methionine restricted rats. GSA glutamic semialdehyde; AASA aminoadipic semialdehyde, CEL carboxyethyl-lysine, CML carboxymethyl-lysine, MDAL malondialdehyde-lysine. Control values (μmoles/mol lysine): 4,461 ± 334 (GSA), 246 ± 17 (AASA), 300 ± 20 (CEL), 2,052 ± 169 (CML), 478 ± 26 (MDAL). Asterisks represent significant differences: a* (compared to control group), b* (between 40% and 80% MetR groups); the degree of significance was P < 0.001 in all cases. Values are mean ± SEM from ten different animals

Figure 5 shows the degree of oxidative damage in mtDNA. 8-OxodG levels were significantly lower both in 40% and 80% MetR rats compared to controls. No significant differences between 40% and 80% MetR groups were observed.
https://static-content.springer.com/image/art%3A10.1007%2Fs10522-008-9130-1/MediaObjects/10522_2008_9130_Fig5_HTML.gif
Fig. 5

Oxidative damage to mitochondrial DNA (8-oxodG) in liver of control, 40% and 80% methionine restricted rats. Asterisks represent statistically significant differences compared to the control group (a*P < 0.05). Values are mean ± SEM from eight to ten different animals

The full fatty acid composition of liver mitochondria and derived indexes are shown in Table 4. 40% MetR significantly increased 18:2n-6 and 22:5n-3 and decreased 18:3n-3 and 20:4n-6 compared to the controls. 80% MetR significantly increased 18:0, 18:2n-6 and 22:4n-6, and decreased 16:0, 16:1n-7, 18:1n-9, 18:3n-3, 20:3n-6 and 20:4n-6 compared to the control group, resulting in lowered MUFA, DBI and PI in 80% MetR than in the control group. Comparing 40% and 80% MetR, lower 16:0, 16:1n-7, 20:3n-6, 20:4n-6, higher 18:0 and 18:2n-6, and lower MUFA, DBI and PI were found in the 80% MetR group.
Table 4

Fatty acid composition of liver mitochondria from control, 40% and 80% methionine restricted rats

 

Control

40% MetR

80% MetR

14:0

0.41 ± 0.03

0.34 ± 0.04

0.39 ± 0.03

16:0

16.36 ± 0.24

16.22 ± 0.26

15.43 ± 0.23a**,b*

16:1n-7

0.70 ± 0.04

0.69 ± 0.06

0.43 ± 0.03a***,b***

18:0

20.20 ± 0.41

20.06 ± 0.46

22.07 ± 0.43a**,b**

18:1n-9

7.78 ± 0.30

7.61 ± 0.31

6.78 ± 0.35a*

18:2n-6

15.85 ± 0.23

17.82 ± 0.45a**

19.84 ± 0.48a***,b**

18:3n-3

0.25 ± 0.02

0.17 ± 0.01a**

0.20 ± 0.01a*

20:3n-6

0.18 ± 0.01

0.19 ± 0.01

0.14 ± 0.007b*

20:4n-6

32.07 ± 0.17

30.38 ± 0.15a***

27.73 ± 0.46a***,b***

22:4n-6

1.55 ± 0.10

1.71 ± 0.14

1.99 ± 0.13a*

22:5n-6

0.45 ± 0.02

0.50 ± 0.02

0.50 ± 0.03

22:5n-3

0.32 ± 0.02

0.44 ± 0.02a**

0.37 ± 0.02

22:6n-3

3.82 ± 0.10

3.82 ± 0.14

4.09 ± 0.08

ACL

18.53 ± 0.01

18.51 ± 0.01

18.50 ± 0.01

SFA

36.98 ± 0.51

36.63 ± 0.62

37.90 ± 0.55

UFA

63.01 ± 0.51

63.36 ± 0.62

62.09 ± 0.55

MUFA

8.49 ± 0.31

8.31 ± 0.36

7.21 ± 0.38a*,b*

PUFA

54.51 ± 0.25

55.05 ± 0.42

54.88 ± 0.53

PUFAn-6

50.11 ± 0.25

50.62 ± 0.49

50.21 ± 0.52

PUFAn-3

4.40 ± 0.10

4.43 ± 0.15

4.66 ± 0.08

DBI

202.85 ± 1.26

201.12 ± 1.15

195.77 ± 1.61a**,b**

PI

186.71 ± 1.14

183.43 ± 1.22

177.61 ± 1.83a***,b**

Values are mean ± SE from 10 different animals and are expressed as mol %

ACL acyl chain length, SFA saturated fatty acids, UFA unsaturated fatty acids, MUFA monounsaturated fatty acids, PUFA polyunsaturated fatty acids, DBI double bond index, PI peroxidizability index

aRepresents significant differences compared to controls, brepresents significant differences between 40% and 80% MetR groups; * P < 0.05, ** P < 0.01, *** P < 0.001

Discussion

In the present investigation it is shown for the first time that restricting the methionine dietary intake (by 80% or 40%), without changing other dietary components, decreases mitROS generation and percent FRL, lowers oxidative damage of mtDNA and proteins, and lowers the intrinsic sensitivity of cellular membranes to lipid peroxidation. For most parameters the decreases occurred from control to 40% MetR without further changes between 40% and 80% MetR. These results suggest that the decrease in methionine ingestion can be responsible for the same effects that have been previously observed in 40% PR and 40% DR studies, as well as for all (during PR) or part (during DR) of the increase in maximum longevity brought about by these two kinds of dietary manipulations.

Previous studies have consistently shown that DR decreases mitROS generation and oxidative damage (reviewed in Gredilla and Barja 2005), and that these changes can be reproduced by PR independently of caloric restriction (Sanz et al. 2004; Gómez et al. 2007; Ayala et al. 2007), but not by lipid (Sanz et al. 2006a) or carbohydrate (Sanz et al. 2006b) restriction. In a longevity study in rats it was found that that DR almost totally prevents rat nephropathy involving glomerular sclerosis, and that “DR without PR” also produced this although somewhat less than DR, meaning that a small effect of PR should be involved in that protection against renal disease (Masoro et al. 1989). In the same study no significant differences in maximum longevity were found between DR and “DR without PR”, which led the authors to conclude that PR was not affecting longevity in their experiment. However, all the studies known to us about the effect of PR on maximum longevity in laboratory rodents have been recently reviewed. Interestingly, it was found that in 16 out of 18 life-long survival studies PR increased maximum longevity in rats and mice (Pamplona and Barja 2006). The mean increase in maximum longevity taking into account the 16 positive studies was 19.6%, which is a longevity extension effect around 50% that of DR (which increases maximum longevity by 20–60%). In addition, the few available studies of which we are aware do not support that carbohydrate (Khorakova et al. 1990) or lipid restriction (Iwasaki et al. 1988; Shimokawa et al. 1996) affect rodent longevity. Therefore, around 50% of the effect of DR on rodent longevity seems to be due to restriction of dietary protein, whereas the other 50% would be due to other unknown dietary components or to the calories themselves.

Moreover, it is known that MetR also increases maximum longevity in rats (Richie Jr et al. 1994; Zimmerman et al. 2003) and mice (Miller et al. 2005) with an intensity generally consistent with that observed in PR. In addition, recent studies show that protein (Min and Tatar 2006) and methionine (Troen et al. 2007) restriction also increase longevity in Drosophila melanogaster. In agreement with its effects on maximum longevity, MetR also delays many age-related changes and degenerative diseases in rodents and produces various changes typical of DR. Thus, MetR in rats strongly decreases visceral fat mass, plasma insulin, IGF-1 and insulin response to glucose, and fully or mostly (by 75%) prevents the increases in blood triglycerydes and cholesterol respectively occurring in old control rats (Malloy et al. 2006). In mice, MetR slows down cataract development, decreases age-related changes in immune function, and lowers serum glucose, insulin, IGF-I and T4 levels (Miller et al. 2005). In sheep liver, MetR also lowers GH-induced IGF-I gene expression (Stubbs et al. 2002). Recent studies also show that MetR stops division of cancer cells (Pavillard et al. 2006) and inhibits colon carcinogenesis (Komninou et al. 2006).

On the other hand, many studies have shown that excess dietary methionine increases tissue oxidative stress (Mori and Hirayama 2000; Stefanello et al. 2005) and we have recently observed that it also increases mitROS production in rat liver (unpublished results). Chronic excessive methionine supplementation strongly stimulates artheriosclerotic changes in the vessel walls and damages tissues including brain and kidney. Recently, we have found that MetR, like DR and PR, also decreases mitROS production (Sanz et al. 2006c), which can help to explain the relationship between dietary methionine, oxidative stress and life extension. However, these last experiments were performed under the same conditions used in the studies that demonstrated the MetR-induced increase in maximum longevity: 80% MetR and substitution of methionine by glutamate in the diet, i.e., the experimental diets had a smaller methionine but also higher glutamate content (Richie Jr et al. 1994; Miller et al. 2005). If dietary methionine is responsible for the decrease in mitROS generation occurring in 40% DR and 40% PR, its effect should occur also at a level of restriction of 40% and without increasing dietary glutamate. This is demonstrated for the first time in the present investigation since both 40% and 80% MetR without changing any other dietary component decreased mitROS generation and FRL (which means that the efficiency in avoiding electron leak is increased). The decrease was somewhat stronger at 80% than at 40% when complex I-linked substrates were used, suggesting a dose-related methionine effect in this particular case. In addition, the only difference found between the effects of 80% MetR substituting dietary methionine for glutamate (Sanz et al. 2006c) and the effects of DR (reviewed in Gredilla and Barja 2005) and PR (Sanz et al. 2004) was that mitochondrial oxygen consumption was increased by the first but not by the other two manipulations. However, in the present study it is shown that when MetR is performed without substitution for dietary glutamate mitochondrial oxygen consumption is no longer increased. Thus, the increased oxygen consumption found in our previous study (Sanz et al. 2006c) was due to the higher glutamate instead of to the lower methionine content of the diet. In contrast, whereas the decreased ingestion of methionine is most likely the dietary factor responsible for the decrease in mitROS generation and FRL.

Concerning the possible mechanisms responsible for the decrease in mitROS production in MetR, various mechanisms have been proposed (Sanz et al. 2006c). Among them, in the present investigation we have studied the possible effects of variations in the amounts of the electron transport ROS generators. It is well-known that only two of the four multiprotein respiratory complexes, complex I (Barja and Herrero 1998; Genova et al. 2001) and complex III (Boveris et al. 1976), produce ROS in the mitochondria. Since complex I and complex III decreased in our study both in 40% and 80% MetR, these decreases can be responsible, at least in part, for the decreases in ROS generation. This is further supported by the magnitude of the quantitative decreases observed. The decrease in ROS production was much stronger with the complex I-linked substrates pyruvate/malate and glutamate/malate than with succinate plus rotenone (which generates ROS only at complex III). In agreement with this, the decrease in complex I protein content was much greater than the decrease in complex IIII amount both in 40% and 80% MetR. However, the decrease in ROS generation must be due also to different causes in addition to decreases in the amount of complexes I and III, because when the rate of ROS generation was calculated per unit of complex I or complex III, it was still lower in both degrees of MetR than in controls. Therefore MetR must also induce qualitative changes that lower ROS production. These can include lowering the degree of electronic reduction of the ROS generators or increasing their redox potential, changes that would agree with the lower FRL values found in MetR as well as in DR in previous studies (Gredilla et al. 2001a, b). On the other hand, the decrease in complex I content is also consistent with the fact that maximum ROS generation with glutamate/malate + rotenone was also decreased by MetR. Whereas basal ROS generation (with substrate alone) depends both on the concentration of the ROS generators and on their degree of electronic reduction, only the first factor is involved in the case of maximum ROS production. Concerning the reason for the decrease in complex I, AIF could be involved. It has been found that AIF, in addition to apoptotic functions, is also required for mitochondrial oxidative phosphorylation (Porter and Urbano 2006). In particular, AIF-deficient cells exhibit a reduced content of complex I (Vahsen et al. 2004), pointing to a role of AIF in the biogenesis or maintenance of this multipolypeptide complex, and mice with reduced AIF expression show a reduced expression of complex I subunits. Thus, it is possible that the decrease in AIF observed in 40% and 80% MetR can contribute to the decreases in complex I content.

Another possibility is that changes in ROS production during MetR could be due to possible variations in homocysteine levels. Raising dietary methionine could be detrimental due to its conversion to methionine cycle metabolites in vivo including homocysteine (Regina et al. 1993). Diets high in protein or methionine increase plasma homocysteine in rodents and humans (Verhoef et al. 2005; Velez-Carrasco et al. 2008). Homocysteine levels increase with age in humans and represent a risk factor for aging and free radical-associated degenerative diseases including at least atherosclerosis, cognitive decline, thrombosis, cancer, stroke, wasting, chronic kidney disease and Parkinson’s disease (Drögue 2001; Durand et al. 2001; Ninomiya et al. 2004). Homocysteine has a free thiol group that can be readily oxidized leading to the generation of protein mixed disulfides or disulfide bridges between different proteins or between subunits of the same protein. Addition of oxidized glutathione to mitochondrial complex I increases its superoxide radical generation rate although this effect has been demonstrated at supraphysiological concentrations (Taylor et al. 2003). Increases in oxidative damage due to homocysteine also result in PARP activation and p53 induction, leading to neuronal DNA damage and apoptosis (Kruman et al. 2000, 2002). Therefore, the decrease in mtROS during MetR could be due to a possible decrease in homocysteine concentration in mitochondrial matrix, a possibility that merits further investigation. On the other hand, the decrease in ROS generation could also be the result of a regulated response. Since methionine is a donor of methyl groups for different cellular reactions, changes in DNA methylation could also be partially the cause of the many changes and variations in gene expression described in DR (Weindruch 2003) including the decreases in respiratory complexes described in this investigation.

On the other hand, the decrease observed in all the four complexes of the respiratory chain at the two MetR levels studied parallels the results obtained after specific RNAi feeding in C. elegans (Dillin et al. 2002; Lee et al. 2003; Curran and Ruvkun 2007). These investigations have shown that inactivating genes coding for subunits of any of these four respiratory complexes in the adult animal increases the maximum life span of the worms. Thus, both kinds of manipulation extending life span, MetR and RNAi inactivation, have in common a marked decrease in the amounts of the four electron transport chain complexes.

In agreement with the decrease in ROS production observed in MetR rats, oxidative damage to mtDNA, estimated as the level of 8-oxodG, was also decreased both in 40% and in 80% MetR. Previous studies have shown the close correspondence between changes in mitROS production and 8-oxodG in mtDNA in DR, PR and 80% MetR, and when comparing animal species with different longevities (Barja 2004; Pamplona and Barja 2006). This correspondence could be due to the close vicinity or even contact between the sites of mitROS generation (especially those facing the matrix) and mtDNA. Thus, DR (Gredilla et al. 2001b), PR (Sanz et al. 2004) and MetR substituting it for dietary glutamate (Sanz et al. 2006c) decrease mitROS generation and 8-oxodG in mtDNA by a similar quantitative amount in rat liver. Furthermore, carbohydrate (Sanz et al. 2006b) or lipid restriction (Sanz et al. 2006a) do not change mitROS production, and, in agreement with this, they do not decrease 8-oxodG in mtDNA. The only discrepancy found in the present investigation between these two parameters was that 80% MetR led to lower levels of ROS generation than 40% MetR (with complex I- but not with complex II-linked substrates), whereas this was not observed in the case of 8-oxodG. In any case, the results of this study, together with many previous investigations, suggest that the decreases in oxidative damage to mtDNA can be involved in the increase in maximum longevity brought about by dietary, protein (reviewed in Pamplona and Barja 2006), and MetR (Richie Jr et al. 1994; Zimmerman et al. 2003; Miller et al. 2005).

In relation to protein oxidative modification, significant decreases in GSA, AASA, CEL, CML and MDAL were found in this investigation both in 40% and 80% MetR. These decreases in protein oxidation (GSA and AASA), glycoxidation (CEL and CML) and lipoxidation (CML and MDAL) also agree with the decrease in mitROS generation observed both in 40% and 80% MetR. Since decreases in these markers of protein damage have been found also in 40% DR (Pamplona et al. 2002a, b), these results further support the idea that the decrease in methionine ingestion is responsible for the decrease in mitochondrial oxidative stress observed in DR. Among the protein markers measured, some of them (CML and MDAL) are dependent on lipid peroxidation. Since lipid peroxidation increases strongly as a function of the number of double bonds per fatty acid, we measured the full fatty acid composition of liver mitochondria. We found that both 40% and 80% MetR significantly decrease the total number of double bonds. Thus, part of the decrease in CML and MDAL can be secondary to this decrease in fatty acid unsaturation. But this can not be the full explanation because the other three protein markers measured, which do not depend on lipid peroxidation (GSA, AASA and CEL), were also decreased by MetR. Thus, it seems that the decrease in mitROS generation induced by MetR leads to a generalized decrease in oxidation, glycoxidation and lipoxidation of mitochondrial proteins.

The decrease in fatty acid unsaturation in 40% and 80% MetR mitochondria was due to variations in various fatty acids, but the most important changes observed (quantitatively) were the decrease in the highly unsaturated 20:4n-6 (the fatty acid showing the strongest changes) and the increase in the much less unsaturated 18:2n-6. This is very similar to the main differences in fatty acid composition that have been observed when comparing long-lived to short-lived mammalian species: higher 18:2n-6 and lower 20:4n-6 and 22:6n-3 (reviewed in Pamplona et al. 2002c; Hulbert et al. 2007). These differences confer the membranes of long-lived animals a superior resistance to lipid peroxidation and lower lipoxidation-dependent damage to macromolecules. Thus, MetR mimics the changes induced by DR in mitochondrial oxidant generation, protein and mtDNA oxidative damage, and sensitivity of mitochondrial membranes to lipid peroxidation.

Reactive oxygen species production with succinate and with glutamate/malate + rotenone, FRL with pyruvate/malate, the amounts of the four respiratory complexes and AIF, as well as the markers of protein oxidation (GSA and AASA) and lipoxidation (CML and MDAL) and oxidative damage to mtDNA were decreased by both 40% and 80% MetR without showing significant differences between these two groups. This suggests than rather than a dose-related effect, a threshold of MetR near to 40% is enough to obtain most of the decrease in mitochondrial oxidative macromolecular damage in liver tissue. This is important due to various reasons: (a) if our results could be extrapolated to humans, it would mean that no strong (80%) restriction of methionine or dietary proteins would be needed to obtain the benefits of a decreased tissue oxidative stress and a slower aging rate; (b) all the previous studies of MetR, including those showing reduced incidence of degenerative diseases and decreases in detrimental biochemical changes (Miller et al. 2005; Malloy et al. 2006), increases in longevity (Richie Jr et al. 1994; Miller et al. 2005) and decreases in oxidative stress (Sanz et al. 2006c), have been performed up to now only at 80% MetR; but our study shows that at least the changes in oxidative stress can also be obtained with a much more moderate (40%) restriction of methionine, and also without any need to decrease the caloric intake. Therefore, we consider most important that future studies investigate the effects of MetR on molecular and physiological markers of health state, incidence of degenerative diseases, and maximum longevity, at a degree of MetR of 40%, and without changing any other dietary component.

In summary, the results of the present study strongly support the possibility that the reduced intake of dietary mehionine is the dietary factor responsible for the decrease in mitROS generation and oxidative stress and around half of the increase in maximum longevity that take place in DR. The fact that all the decreases in oxidative stress observed in this investigation are observed already at 40% MetR, without the need to further restrict it to 80%, further supports that the lowered methionine ingestion is responsible for those beneficial changes occurring during standard (40%) DR. The similarity of the magnitude of the decreases in mitROS production and oxidative stress in 40% DR (reviewed in Gredilla and Barja 2005), 40% PR (Sanz et al. 2004) and 40% MetR (this investigation) also supports that possibility. The results obtained can be relevant for possible extrapolation to humans, also because a large part of the MetR-induced decreases in growth rate and/or food consumption will be avoided by restricting methionine at 40% instead of at 80%.

Acknowledgments

This study was supported in part by I + D grants from the Spanish Ministry of Education and Science (BFU2006-14495/BFI), the Spanish Ministry of Health (ISCIII, Red de Envejecimiento y Fragilidad, RD06/0013/0012), the Generalitat of Catalunya (2005SGR00101) and “La Caixa” Foundation to R. Pamplona; and from the Ministry of Education and Science (BFU2005-02584) and from CAM/UCM groups (910521) to G. Barja; A. Naudi received a predoctoral fellowship from “La Caixa” Foundation. P. Caro and J. Gómez received predoctoral fellowships from the Ministry of Education and Science. We thank David Argiles for excellent technical assistance.

Copyright information

© Springer Science+Business Media B.V. 2008