Mitochondria can synthesize a limited number of proteins encoded by mtDNA (mitochondrial DNA) by using their own biosynthetic machinery, whereas most of the proteins in mitochondria are imported from the cytosol. It could be hypothesized that the mitochondrial pool of amino acids follows the frequency of amino acids in mtDNA-encoded proteins or, alternatively, that the profile is the result of the participation of amino acids in pathways other than protein synthesis (e.g. haem biosynthesis and aminotransferase reactions). These hypotheses were tested by evaluating the pool of free amino acids and derivatives in highly-coupled purified liver mitochondria obtained from rats fed on a nutritionally adequate diet for growth. Our results indicated that the pool mainly reflects the amino acid composition of mtDNA-encoded proteins, suggesting that there is a post-translational control of protein synthesis. This conclusion was supported by the following findings: (i) correlation between the concentration of free amino acids in the matrix and the frequency of abundance of amino acids in mtDNA-encoded proteins; (ii) the similar ratios of essential-to-non-essential amino acids in mtDNA-encoded proteins and the mitochondrial pool of amino acids; and (iii), lack of a correlation between codon usage or tRNA levels and amino-acid concentrations. Quantitative information on the mammalian mitochondrial content of amino acids, such as that presented in the present study, along with functional studies, will help us to better understand the pathogenesis of mitochondrial diseases or the biochemical implications in mitochondrial metabolism.

INTRODUCTION

The mtDNA (mitochondrial DNA) encodes for 13 polypeptides, namely seven subunits of Complex I (ND1–ND6, including ND4L), cytochrome b, three subunits of Complex IV [COX (cytochrome c oxidase) I–III], and two subunits of Complex V (ATPase subunits 6 and 8 [1]), two rRNAs and 22 tRNAs. It can synthesize a few critical proteins by using its own biosynthetic machinery, but most of the proteins are imported from the cytosol. It could be hypothesized that the mitochondrial pool of amino acids is the result of protein synthesis of mtDNA-encoded proteins and degradation of mitochondrial proteins (regardless of the genome origin), in addition to those amino acids whose pathways are involved in other processes not related directly to protein synthesis (e.g. haem biosynthesis and aminotransferase reactions). Alternatively, the mitochondrial pool of free amino acids might be influenced more by one of these pathways in particular (e.g. following the profile of amino acids in mtDNA-encoded proteins). To test these hypotheses, reliable data on the pool of free amino acids in mitochondria should be available. However, very few publications have been presented, and most of them were performed using the technology available in the 1960s–1970s [24]. Thus the main goal of the present study was to provide a quantification of free amino acids and derivatives in highly-coupled purified liver mitochondria obtained from rats fed on a nutritionally adequate diet for growth. Furthermore, the putative differences between the cytosolic and mitochondrial pools of free amino acids were evaluated to ascertain if these are different pools to support different functions or if the mitochondrial pool is similar to the cytosolic one. In this regard, the occurrence of correlations between the levels of mitochondrial free amino acids and those of tRNA, amino-acid content in mitochondrial polypeptides or codon usage were also tested. Finally, comparisons between the mitochondrial, cytosolic and serum pool of amino acids were evaluated to ascertain the similarities and differences between these pools.

Quantitative information on the amino-acid content of mammalian mitochondria, such as that presented in this report, along with functional studies, will help us to understand better the pathogenesis of mitochondrial diseases or the biochemical implications in mitochondrial metabolism when protein deficiencies are encountered in diets.

EXPERIMENTAL

Chemicals and biochemicals

EDTA, EGTA, sodium succinate, mannitol, sucrose, Hepes and bovine serum albumin (fatty-acid free) were all purchased from Sigma. Tris, glycine, sodium chloride, and potassium chloride were purchased from Fisher Scientific. All other reagents were of analytical grade.

Animals and diet

The experiments described below were approved by the UC Davis Animal Care and Use Committee. Male Sprague–Dawley rats (6 weeks old) weighing 180–200 g were purchased from a commercial supplier (Harlan Sprague Dawley, Indianapolis, IN, U.S.A.), except where noted. All cages were kept in a vivarium maintained at 22±1°C, with a 12 h light/12 h dark cycle, following NIH (National Institutes of Health) guidelines for animal care and use. The lights were set to go out at 19:00 h. Cage maintenance was conducted during the light cycle.

Briefly, after arrival in our laboratory, rats were fed on an amino-acid-based diet. The diet was prepared in our laboratory using purified ingredients and free L-amino acids. The diet was amino-acid based with a full complement of all EAAs (essential amino acids) at ratios based on the egg protein (see Supplementary Tables S1A–S1C at http://www.bioscirep.org/bsr/028/bsr0280239add.htm). Over the testing period, the rats had free access to their food 24 h per day, except when cage maintenance was conducted. At day 9, the rats were removed from their cages and taken to the surgery room, where they were killed with CO2. Rats were killed between 09:00 h and 10:00 h. The livers were quickly removed from the rats, wet weights were recorded and rapidly homogeneized for mitochondria isolation or amino-acid analyses.

Isolation of mitochondria

Rat liver mitochondria were isolated as described previously [5,6]. Briefly, purified mitochondria from homogenized livers or hearts were collected by differential centrifugation followed by use of a Percoll gradient and multiple washes in MSHE buffer (220 mM D-Mannitol, 70 mM sucrose, 2 mM Hepes, 0.5 mM EGTA and 0.1% fatty-acid free bovine serum albumin) and 150 mM KCl. A faster procedure was designed based on a method published previously [7]. Essentially, 0.2–0.4 g of liver were homogenized in cold MSHE buffer, diluted with an equal volume of 36% (v/v) Percoll in MSHE buffer and applied directly on to a Percoll gradient over two pre-formed layers containing 36% (v/v) and 55% (v/v) Percoll. The gradients were centrifuged at 30700 g for 5 min. The third fraction from the top was collected, resuspended in MSHE buffer, centrifuged at 7300 g for 5 min and finally resuspended in 150 mM KCl.

Oxygen consumption

The oxygen uptake of mitochondria was measured using a Clark-type O2 electrode from Hansatech at 30°C [5,6]. Intact purified mitochondria were tested in 1 ml of reaction buffer containing 225 mM sucrose, 5 mM MgCl2, 20 mM KCl, 10 mM potassium phosphate and 20 mM Hepes/KOH (pH 7.4) in the presence of 1 mM malate and 10 mM glutamate (malate–glutamate) or 10 mM succinate as substrates and 0.45 mM ADP. The P/O ratios (moles of ATP synthesized per atom of oxygen consumed) were determined measuring the increase in the uptake of oxygen caused by the addition of a known amount of ADP, as described by Chance and Williams [8].

Amino-acid analyses

Tissue samples were immediately homogenized in ice-cold 3% (v/v) sulfosalicylic acid [10 ml: 1 g of liver wet weight; 10 ml: 1 g of mitochondrial protein; 1:1 (v/v) serum/sulfosalicylic acid] spiked with 2.45 μM L-γ-amino-n-butyric acid as the internal standard. Each sample was then centrifuged at 14000 g for 30 min at 4°C. All samples were processed for amino-acid analysis using a commercially available kit (AccQ Tag, Waters [9,10]). Amino acids were visualized using HPLC with fluorescence detection.

Statistical analyses

Results were analysed using the ANOVA test and are expressed as means±S.E.M. except where noted.

RESULTS AND DISCUSSION

Free amino acids in rat liver mitochondria

Male 6 weeks old Sprague–Dawley rats were fed on an amino-acid based diet containing all the nutrients required for adequate growth {diet with 0.87% threonine as described previously [11] (and see Supplementary Table S1)}. To determine the amino-acid levels present in liver mitochondria, and considering that these concentrations fluctuate significantly after or before a meal [12,13], it is important to indicate that the rats used in this study were killed between 09:00 h and 10:00 h, approx. 2–3 h after their first meal, and thus the amino-acid pool reported in this study should be taken as being representative of a postprandial period. Percoll-purified liver mitochondria showed negligible levels of contamination with other subcellular compartments and high coupling of the electron-transport pathway to oxidative phosphorylation [6]. The intactness and functionality of these mitochondria were assessed by evaluating the RCR (respiratory control ratio) and P/O values. The RCR with malate–glutamate was 10.7±0.1, and with succinate was 6.6±0.2. The P/O values with each substrate were 3.2±0.2 and 2.29±0.04, indicating that the three phosphorylation sites were operational.

The concentrations of free amino acids and derivatives found in these mitochondria are shown in Table 1. The total amount of amino acids found in the present study was, in general, within the values reported before for various species and organs (see Supplementary Table S2 at http://www.bioscirep.org/bsr/028/bsr0280239add.htm). The relative high concentration of amino acids reported in the present study (approx. 35 mM) and previous studies (see Supplementary Table S2) had raised the possibility that they might contribute significantly to the osmolarity of these organelles, whereas that of either serum (or plasma) or cytosol is constituted mainly by inorganic electrolytes. However, if the mitochondrial osmolarity has been calculated as 300–330 mM, where the sum of the concentrations of potassium, sodium, magnesium, chloride, orthophosphate and organic phosphate is between 150–180 mM, and the internal concentration of sucrose is approx. 150 mM (when isolating rat liver mitochondria with 0.25 M sucrose [14,15]), then the highest concentration of osmotically active amino acids would be 150–180 mM, approx. 20% of the observed experimental concentrations. The relatively high concentration of amino acids might also explain the low rate of radioactive label incorporation into mitochondrial proteins if cells have not been previously starved in methionine-free medium for an adequate period [16] because, as suggested previously [4], the relatively large size of the amino-acid pool leads to dilution of the isotopically labelled amino acids, resulting in an underestimation of the amount of radioactivity incorporated.

Table 1
Free amino acids in rat liver mitochondria

Results are means of 3 independent experiments with S.D.≤10%.

Amino acidConcentration (nmol/mg of protein)
 Alanine 2.63 
 Arginine 1.29 
 Aspartic acid 1.00 
 Asparagine 1.23 
 Cysteine 0.08 
 Glutamate 0.34 
 Glutamine 1.21 
 Glycine 1.23 
 Histidine 0.67 
 Isoleucine 1.89 
 Leucine 4.28 
 Lysine 2.55 
 Methionine 1.38 
 Phenylalanine 2.07 
 Proline 0.21 
 Serine 1.81 
 Threonine 1.33 
 Tryptophan 0.29 
 Tyrosine 1.28 
 Valine 1.48 
 Total 28.25 
Others  
 GSH 7.75 
 GSSG 0.73 
 Taurine 0.15 
 Ornithine 0.99 
 Citrulline 0.00 
 Anserine 0.00 
 Carnosine 0.00 
 Sarcosine 0.00 
 1-Methyl-histidine 0.00 
 3-Methyl-histidine 0.00 
 Hydroxyproline 0.00 
L-α-Aminobutyric acid 0.00 
TOTAL 37.87 
Amino acidConcentration (nmol/mg of protein)
 Alanine 2.63 
 Arginine 1.29 
 Aspartic acid 1.00 
 Asparagine 1.23 
 Cysteine 0.08 
 Glutamate 0.34 
 Glutamine 1.21 
 Glycine 1.23 
 Histidine 0.67 
 Isoleucine 1.89 
 Leucine 4.28 
 Lysine 2.55 
 Methionine 1.38 
 Phenylalanine 2.07 
 Proline 0.21 
 Serine 1.81 
 Threonine 1.33 
 Tryptophan 0.29 
 Tyrosine 1.28 
 Valine 1.48 
 Total 28.25 
Others  
 GSH 7.75 
 GSSG 0.73 
 Taurine 0.15 
 Ornithine 0.99 
 Citrulline 0.00 
 Anserine 0.00 
 Carnosine 0.00 
 Sarcosine 0.00 
 1-Methyl-histidine 0.00 
 3-Methyl-histidine 0.00 
 Hydroxyproline 0.00 
L-α-Aminobutyric acid 0.00 
TOTAL 37.87 

When the amino-acid profile for rat liver mitochondria was compared with others published previously (see Supplementary Table S3 at http://www.bioscirep.org/bsr/028/bsr0280239add.htm), the main difference was that, to our knowledge, this is the first study that reports values for asparagine, cysteine, proline, tryptophan and taurine, whereas these compounds were not detected before either because of the method used (e.g. asparagine detected as aspartic acid) or because the concentrations were below the detection limit [24]. In addition, we evaluated several other biologically relevant compounds whose concentrations were not sought (e.g. anserine, carnosine, 1-methyl-histidine and 3-methyl-histidine).

The amounts of EAAs and NEAAs (non-essential amino acids) in rat liver mitochondria were 159.4 and 135.7 nmol/mg of protein (Table 2), resulting in an EAA/NEAA ratio of 1.18. This ratio was distinct from that obtained for rat liver (0.44) and serum (0.57), suggesting that these pools are highly compartmentalized, and that a gradient of EAAs exists from blood into mitochondria. The experimental ratio of EAA/NEAA of rat liver mitochondria (1.18) was similar to that calculated from the results of Baird [2] (1.3), but quite different from those calculated from other previous studies (0.2 to 0.6) [3,4]. In fact, these latter ratios were more similar to rat liver EAA/NEAA ratios obtained in the present study (0.44) and previous work (0.12 [17,18] and 0.2 [19]), but below the values found for serum or plasma (0.57, the present study; 0.3–0.5 [20]; 0.65 [18]; 0.7 [19]; and 0.93 [17]). This indicated that, in these reports [4] (see Supplementary Table S3), a cross-contamination between mitochondria and cytosol had occurred during the isolation procedure.

Table 2
Pool of free amino acids and their derivatives in rat serum, liver and liver mitochondria

Samples from liver, serum and liver mitochondria were processed as described in the Experimental section. The values in bold are those whose concentrations were higher than or similar to those found in the mitochondrial matrix. Results are means of 3 determinations with S.D.≤10%. Serum values were within those published previously using either rat plasma [18,7678] or serum [79]; rat liver values were within those published previously [17,18,80] and comparable with those obtained with flash-frozen livers [81]. Numbers in parentheses show lysine and glycine only. The amino-acid concentrations indicated in this study were taken to be homogenously distributed throughout the mitochondrial matrix. If kinetic or functional studies were to be performed with these concentrations, other biological settings should be considered that might undermine the concept of ‘concentration of free amino acid’. For instance, some of these situations are substrate channelling (as described for the urea cycle [82,83]), microcompartmentation (as described for aspartic acid and glutamate [40,58]) and diffusion of metabolites. In the last term, concepts of ‘free water’, ‘bound water’ and ‘density of mitochondrial matrix’ become interconnected [84,85]. Thus mechanisms of metabolite transfer between enzymes will differ significantly from that usually assumed where dissociation and random diffusion of metabolite through the aqueous environment is responsible for transfer to the next enzyme site [86]. Furthermore, this situation is exacerbated in mitochondria, where essentially it seems to be difficult to visualize any ‘free water’ in the matrix when the relatively high protein [6], metabolite concentrations and their hydration shells are taken into account. To calculate the concentration of amino acids in mitochondria, an average value for mitochondrial matrix volume was obtained (0.83±0.04 μl/mg of protein) from values reported by different laboratories using various techniques. The volume of the cytoplasmic compartment was based on the cytoplasmic compartment occupying 72% of a cell volume of 4.9 pl [87]. The volume of the mitochondrial compartment was calculated from the percentage of the cytoplasmic compartment occupied by mitochondria (10.5% [87]) and the percentage of mitochondria occupied by the matrix (64% [88]). Using this approach, 4.8% of the tissue volume should be mitochondrial matrix volume (1 g of liver = 48 μl of mitochondrial matrix). Using our isolation procedure, from 1 g of liver we obtain on average 30 mg of mitochondrial protein (although if purified, approx. 15 mg). Then the matrix volume to mitochondrial protein ratio would be 1.6 μl/mg. Werkheiser and Bartley [14] showed that only 60% of intramitochondrial water was permeable to sucrose. Thus the corrected accessible water would be 0.96 μl/mg of mitochondrial protein. Others have reported a value of 0.8 μl/mg of protein [29,89], 0.74 μl/mg of protein [58] and 0.854 μl/mg of protein [40] using other approaches, in close agreement with the value found above. AA, aromatic amino acid; K,G,T; lysine, glycine and threonine.

Amino acids and derivativesSerum (mM)Liver (mM)Mitochondria (mM)
Essential    
 Histidine 0.09 0.57 0.80 
 Isoleucine 0.10 0.81 2.27 
 Leucine 0.15 1.57 5.16 
 Lysine 0.78 1.87 3.08 
 Methionine 0.06 0.50 1.66 
 Phenylalanine 0.07 0.67 2.49 
 Threonine 0.54 1.43 1.61 
 Tryptophan 0.02 0.15 0.35 
 Valine 0.21 1.46 1.79 
 Total EAA 2.02 9.03 19.21 
Non-essential    
 Alanine 0.63 4.27 3.17 
 α-Amino-butyrate 0.01 0.00 0.04 
 Anserine 0.00 0.00 0.00 
 Arginine* 0.19 0.01 1.55 
 Asparagine 0.23 0.46 1.48 
 Aspartic acid 0.02 2.02 1.20 
 Carnosine 0.00 0.00 0.00 
 Citrulline 0.10 0.02 0.00 
 Cysteine 0.005 0.04 0.07 
 Glutamine 0.73 0.57 1.46 
 Glutamate 0.12 3.13 0.41 
 Glycine 0.38 2.41 1.48 
 Hydroxyproline 0.06 0.31 0.00 
 1-Methyl-histidine 0.01 0.01 0.00 
 3-Methyl-histidine 0.00 0.00 0.00 
 Ornithine 0.06 1.24 1.19 
 Proline 0.19 1.14 0.25 
 Sarcosine 0.02 0.003 0.00 
 Serine 0.40 1.95 2.18 
 Taurine 0.21 2.38 0.18 
 Tyrosine 0.14 0.55 1.55 
 Total NEAA 3.51 20.51 16.21 
EAA/NEAA ratio* 0.57 0.44 1.18 
 Glycine/serine 0.95 1.24 0.68 
 Valine/glycine 0.55 0.61 1.21 
 Tyrosine/phenylalanine 2.0 0.80 0.60 
 AA 0.23 1.37 4.39 
 BCAA 0.46 3.84 9.22 
 AA/BCAA 0.50 0.36 0.48 
 K,G,T/BCAA 3.70 (2.52) 1.50 (1.11) 0.67 (0.49) 
Amino acids and derivativesSerum (mM)Liver (mM)Mitochondria (mM)
Essential    
 Histidine 0.09 0.57 0.80 
 Isoleucine 0.10 0.81 2.27 
 Leucine 0.15 1.57 5.16 
 Lysine 0.78 1.87 3.08 
 Methionine 0.06 0.50 1.66 
 Phenylalanine 0.07 0.67 2.49 
 Threonine 0.54 1.43 1.61 
 Tryptophan 0.02 0.15 0.35 
 Valine 0.21 1.46 1.79 
 Total EAA 2.02 9.03 19.21 
Non-essential    
 Alanine 0.63 4.27 3.17 
 α-Amino-butyrate 0.01 0.00 0.04 
 Anserine 0.00 0.00 0.00 
 Arginine* 0.19 0.01 1.55 
 Asparagine 0.23 0.46 1.48 
 Aspartic acid 0.02 2.02 1.20 
 Carnosine 0.00 0.00 0.00 
 Citrulline 0.10 0.02 0.00 
 Cysteine 0.005 0.04 0.07 
 Glutamine 0.73 0.57 1.46 
 Glutamate 0.12 3.13 0.41 
 Glycine 0.38 2.41 1.48 
 Hydroxyproline 0.06 0.31 0.00 
 1-Methyl-histidine 0.01 0.01 0.00 
 3-Methyl-histidine 0.00 0.00 0.00 
 Ornithine 0.06 1.24 1.19 
 Proline 0.19 1.14 0.25 
 Sarcosine 0.02 0.003 0.00 
 Serine 0.40 1.95 2.18 
 Taurine 0.21 2.38 0.18 
 Tyrosine 0.14 0.55 1.55 
 Total NEAA 3.51 20.51 16.21 
EAA/NEAA ratio* 0.57 0.44 1.18 
 Glycine/serine 0.95 1.24 0.68 
 Valine/glycine 0.55 0.61 1.21 
 Tyrosine/phenylalanine 2.0 0.80 0.60 
 AA 0.23 1.37 4.39 
 BCAA 0.46 3.84 9.22 
 AA/BCAA 0.50 0.36 0.48 
 K,G,T/BCAA 3.70 (2.52) 1.50 (1.11) 0.67 (0.49) 
*

To be able to compare EAA/NEAA ratios with values published previously, arginine has not been included as an essential amino acid, even though it is considered essential for growing rats [11]. If it were included, no significant differences would had been found in the EAA/NEAA ratios or in the final conclusions.

If outlier values are excluded (2-fold above or one-half of our values; see underlined values in Supplementary Table S3), our values were more similar to those reported by Baird [2]. However, this last report failed to measure concentrations for arginine, asparagine, cysteine, glutamine, proline, tryptophan, ornithine and taurine, and recorded relatively high (glutamate, lysine and serine) or low (methionine and tyrosine) concentrations for other amino acids.

Although the EAA/NEAA ratio calculated from [3] seemed to be closer to our value, a closer inspection of the distribution of sulfur-containing compounds indicated that most of cysteine and GSH were recovered as cystine, cysteic acid and GSSG. This is indicative of samples being exposed to oxidative stress. If this were the case, then the permeability of the mitochondrial membranes would be compromised, resulting in significant changes in metabolite concentrations. In contrast, almost 60% of the low-molecular-mass sulfur-containing compounds (GSH, GSSG, cysteine, methionine and taurine) in the present study was comprised of glutathionyl equivalents. The high ratio of [GSH]/[GSSG] found in the present study was comparable with results obtained previously by us [21] and others [22,23], indicating that these mitochondria were minimally exposed to oxidative stress.

Other significant reasons for the differences between the present study and previous studies could be related to model-specific differences (sex, age and strain [2426]) and time of sampling (circadian fluctuations of amino acids [27]).

It could be argued that the mitochondrial pool reported in the present study is the result of mitochondrial proteolysis that might have occurred during the isolation procedure. Several reasons suggest against this possibility. First, throughout the entire isolation procedure, EDTA and EGTA, metalloprotease inhibitors, were present at concentrations relevant to inhibit this activity. When, in addition to these, other proteolytic inhibitors ([4-(2-aminoethyl)benzenesulfonylfluoride, aprotinin, bestatin, E-64, PMSF and pepstatin A] were present throughout all steps of mitochondria isolation and purification, no changes in the quality and quantity of amino acids from mitochondria was obtained. Secondly, the isolation procedure was performed at 0–4°C (on ice, using cooled buffers, glassware, rotors and refrigerated centrifuge etc.) to minimize enzymatic activities, and the procedure was performed within 2 h to minimize the time of the handling of samples and their storage on ice. When a different procedure to isolate mitochondria was performed that allowed the isolation of mitochondria in less than 45 min, no significant changes in the amino-acid profile was obtained. Even if any unavoidable proteolysis or leakage were to occur during our procedure, it has been shown that incubating mitochondria (not purified and without proteolytic inhibitors) at 0°C for 4 h results in a total amino acid increase of 2–6 mM [28]. These increases are within the experimental error, and their contribution could be considered negligible to the pool of free amino acids. Thirdly, throughout the procedure, isolation buffers were supplemented with 0.1% fatty-acid free bovine serum albumin. This protein not only binds fatty acids which may uncouple mitochondria, but at the concentration used provided an alternative substrate for putative proteases, sparing mitochondrial proteins from degradation. Finally, if the amino acids present in the matrix were the result of proteolysis during the isolation procedure, then the ratio of EAA/NEAA should be similar to that of the most abundant mitochondrial proteins [e.g. CPS I (carbamoylphosphate synthetase I), glutamate dehydrogenase, α- and β-subunits of ATPase]. These ratios are (on average) 0.79, significantly different from the ratio determined for the mitochondrial pool (1.18), indicating that the values for these amino acids reflect their steady-state concentrations under the experimental conditions used in the present study.

Comparison between serum, liver and mitochondrial pools of free amino acids

It has been shown that some metabolites follow a non-uniform distribution between the mitochondrial and cytosolic compartments {e.g. 13 mM [ATP+ADP+AMP] compared with 8 mM [29]}, whereas others have similar concentrations (GSH, 7–10 mM [21,23]). Following this reasoning and considering several assumptions (see Table 2), the mitochondrial concentrations of the amino acids in the mitochondrial matrix water were calculated. We tested whether the mitochondrial pool of free amino acids was the same as that in the cytosol or serum, or if they constituted an independent, different pool. To this end, the serum and liver concentrations of free amino acids were evaluated in parallel with those in mitochondria (Table 2). Most amino acid and derivative concentrations were higher in mitochondria than in the cytosol or serum, with a few exceptions: citrulline, glutamate, hydroxyproline, 1-methyl-histidine, proline, sarcosine and taurine. Even if the amino acids are classified into EAAs and NEAAs, mitochondria still have the highest concentration of both and exhibited the highest ratio of EAA/NEAA (Table 2). However, considering the low Km of all amino acids for their respective aaRSs (aminoacyl-tRNA synthetases) (nM to low μM concentrations [30,31]), the amount of tRNAs (μM concentrations [32]) and amino acids (mM concentrations) in each compartment (see Supplementary Figure S1 at http://www.bioscirep.org/bsr/028/bsr0280239add.htm), no significant changes in the acylation of tRNAs would be expected in either compartment, resulting in a favoured protein synthesis as would be expected for a growing, well-fed animal. Moreover, given that codon misreading during amino-acid starvation presumably occurs because the amount of cognate charged tRNA drops to a very low level, increasing the likelihood of misreading or misincorporation by a nearly cognate tRNA [33], then the high levels of mitochondrial (and cytosolic) amino acids might also prevent drops in tRNAs, preventing misreading frequency and assuring the fidelity of protein synthesis.

Role of amino acids in mitochondria

The pool of free amino acids in mitochondria could be considered to be the result of the difference between the input (represented by the influx from the pool present in the liver cytosol, amino acids from mitochondrial protein degradation and amino acids formed by other processes) and the output (amino acids used for mitochondrial protein synthesis and other processes). If it is assumed that in a well-fed, young and growing animal, the rate of synthesis will be favoured over that of proteolysis, the contribution of mitochondrial protein degradation to the pool of mitochondrial free amino acids will be considered negligible. Thus if the original pool of mitochondrial amino acids is considered identical to that of liver cytosol, and from this, free amino acids (determined in this study) and amino acids used for mitochondrial protein synthesis are subtracted, it would be possible to evaluate the distribution and use of each amino acid in terms of functions other than protein synthesis (e.g. transamination reactions). Assuming that the main factor that determines the different molar abundance of mitochondrial Complexes is the rate of synthesis of their subunits [34], using the relative abundance of different Complexes from bovine heart mitochondria [35], the primary sequences of rat mtDNA-encoded subunits (from the Expasy website http://www.expasy.org), and setting as an upper limit the total amino-acid concentration, the distribution of each amino acid was estimated (Figure 1). In almost one-half of the cases (histidine, isoleucine, leucine, methionine, threonine, tryptophan, asparagine, cysteine and tyrosine), the addition of the concentration of the amino acid used for mitochondrial protein synthesis and that of the free amino acid accounted for the total amino-acid concentration, leaving negligible amounts for other functions. Conversely, amino acids involved in transamination or nitrogen metabolism (alanine, aspartic acid, glutamine, glutamate and glycine) presented a higher concentration destined for functions other than protein synthesis. In this category, amino acids whose catabolic pathway is mainly mitochondrial (lysine, threonine and glycine; indirectly serine to glycine) were also included.

Correlation of free amino-acid concentrations in mitochondria and amino-acid abundance in mtDNA-encoded proteins

Figure 1
Correlation of free amino-acid concentrations in mitochondria and amino-acid abundance in mtDNA-encoded proteins

The total pool of mitochondrial amino acids (black bars) was assumed to be identical to that of liver cytosol (from Table 2). Light grey bars, free amino acids (from Table 1). Mitochondrial protein synthesis was calculated as described above (white bars) and other functions of amino acis are also shown (dark grey bars). The amino acids are indicated by their three letter amino-acid code.

Figure 1
Correlation of free amino-acid concentrations in mitochondria and amino-acid abundance in mtDNA-encoded proteins

The total pool of mitochondrial amino acids (black bars) was assumed to be identical to that of liver cytosol (from Table 2). Light grey bars, free amino acids (from Table 1). Mitochondrial protein synthesis was calculated as described above (white bars) and other functions of amino acis are also shown (dark grey bars). The amino acids are indicated by their three letter amino-acid code.

If the main role of some matrix free amino acids were to support the synthesis of mtDNA-encoded proteins then, excluding those that have a main role in functions other than protein synthesis, a correlation between their levels and the amino-acid abundance in mtDNA-encoded proteins could be expected. The putative correlation between all free amino acid in the matrix and the amino-acid content in mtDNA-encoded proteins is shown in Figure 2. A positive slope was found, indicating that there was a trend between these variables (Figure 2), and analysis of the regression line indicated a correlation (r2) of 0.54 (n=20; Pearson's analysis resulted in a confidence interval of P<0.05). When amino acids known to be catabolized predominantly in rodent mitochondria and those involved mainly in functions other than protein synthesis were excluded from the analysis, the correlation coefficient was r2=0.93 (n=10; Pearson's analysis indicated a confidence interval of P<0.05), indicating a high correlation.

Dependence of free amino acids in mitochondria on the amino-acid content in mtDNA-encoded proteins

Figure 2
Dependence of free amino acids in mitochondria on the amino-acid content in mtDNA-encoded proteins

(A) All amino acids present. (B) Excluding amino acids whose concentration would mainly sustain protein synthesis (inferred from Figure 1). The amino acids are indicated by their three letter amino-acid code.

Figure 2
Dependence of free amino acids in mitochondria on the amino-acid content in mtDNA-encoded proteins

(A) All amino acids present. (B) Excluding amino acids whose concentration would mainly sustain protein synthesis (inferred from Figure 1). The amino acids are indicated by their three letter amino-acid code.

No amino-acid profile of a single polypeptide encoded by mtDNA was specifically more correlated to the amino-acid pool; instead, all polypeptides (except ATP8 with r2=0.58) had correlation coefficients of between 0.8 and 0.9.

It could be argued that amino acids that have other functions besides being substrates for protein synthesis might be at limiting concentrations, even if their concentrations are not among the lowest ones. However, several examples demonstrate that the mitochondrial concentrations are not limiting for protein synthesis (see above) or other functions. For example, arginine concentrations were 30-fold higher than the Km for arginine of nitric oxide synthase [5], confirming, as we indicated previously, [5,36] that arginine concentrations allow full saturation of this enzyme, and approx. 400-fold higher than that required for NAGS [NAG (N-acetylglutamate) synthase] activation [37]. In another example, glycine concentrations would allow 6% of the maximal activity of 5-amino-δ-levulinate synthase (Km for glycine=23 mM [38]) the first enzyme in the haem biosynthetic pathway, assuming a Michaelis–Menten behaviour and full saturation with the other substrate, succinyl-CoA. It has been reported that glutamate dehydrogenase appears to be saturated at intramitochondrial concentrations of approx. 3.7 nmol glutamate/mg of protein [39,40]. At the mitochondrial concentrations of glutamate, it would be expected that this enzyme would be operating at only 10% of its maximal activity, whereas a higher activity would be expected when an influx of glutamine or glutamate is directed to the liver (e.g. in early starvation).

Taurine, valine and lysine

Although the precise role for taurine is still not fully understood, it is known that it is not present in mtDNA-encoded proteins. Given that its free concentration is relatively small, a role in osmolarity control [4143], as has been proposed for this compound, seems unlikely (Table 1). Thus it seems that taurine is fulfilling another role in mitochondria. In this regard, a link between taurine deficiency and cardiomyopathies in cats has been established [44], and it has been found that taurine is present in modified uridine residues at the wobble position of mitochondrial tRNAs, namely 5-taurinomethyluridine in tRNALeu and 5-taurinomethyl-2-thiouridine in tRNALys [45,46]. The taurine wobble-modification deficiency has been shown to result in a translational defect that is likely to contribute significantly to mitochondrial dysfunction associated with mutations in mt-tRNALys or mt-tRNALeu [as in the case of MERRF (myoclonic epilepsy and ragged-red fibres) and MELAS (mitochondrial encephalopathy, lactic acidosis and stroke-like episodes)].

An unexpected result was obtained with valine. Although valine is quite abundant in mtDNA-encoded proteins, the concentration destined for other functions was higher than that estimated for protein abundance. This finding is in contrast with that found for the other BCAAs (branched-chain amino acids) (isoleucine and leucine), whose concentrations were mainly used to sustain mitochondrial proteins. This can be explained by the relatively lower (2-fold lower than those for the other two BCAAs; recalculated from [47]) second-order rate constant for valine catabolism by the BCAA metabolome.

One of the most abundant amino acids in mitochondria was found to be lysine. The concentration of lysine was not only higher than that required to support protein synthesis, but was also higher than for its catabolism. In rodents, lysine is predominantly catabolized in liver mitochondria by the subsequent action of two enzymes, LOR (lysine-α-oxoglutarate reductase) and SDH (saccharopine dehydrogenase) [48,49]. The concentration of lysine was 2-fold greater than the Km for purified LOR (Km for lysine=1.5 mM [5052]), indicating that its catabolism is not limited by lysine concentrations. It could also be inferred that the uptake of lysine is not the rate-limiting step for lysine oxidation, in contrast with a previous study using rats fed on various casein diets [27]. The LOR and SDH activities contribute not only to the general nitrogen balance in the organism (amino acid catabolism), but also to the controlled conversion of lysine into ketone bodies [5255]. These energetic compounds can be used as substrates for energy in situations of limited carbon supply. It has been found that the LOR and SDH activities increase in starved mice in comparison with well-fed animals [56]. Therefore under our experimental conditions, i.e. well-fed animals, it is expected that the oxogenesis pathway would be depressed, and as a consequence, the use of lysine for such a purpose would also be limited by other regulatory pathways, leading to higher steady-state concentrations of this amino acid. Two facts support this hypothesis: first, lysine (along with tyrosine and phenylalanine) plays a major role in hepatic oxogenesis [57]. Secondly, the catabolism of lysine in a well-fed animal (without the need for active oxogenesis) could be regulated at the level of LOR and SDH concentrations (long-term regulation) or by limiting the substrate for these enzymes (short-term regulation). Given that lysine concentrations and probably NADPH concentrations are higher than the Km required for full LOR activity [5052], it could be proposed that the availability of α-oxoglutarate required for this catabolic pathway is compromised, given that its Km (1 mM [5052]) is three orders of magnitude greater than its calculated mitochondrial concentration from the glutamate-oxaloacetate transaminase reaction (1.4 μM; this concentration was calculated using glutamate-oxaloacetate Keq=6.61, glutamate and aspartic acid concentrations from Table 1 and the oxaloacetate concentration from [58]) [40,58]. Assuming Michaelis–Menten kinetics for LOR, LOR activity would be <0.1% of its Vmax, even if all α-oxoglurate was available for use in this reaction.

Urea cycle activity: ornithine and citrulline

The concentration of ornithine was in the intermediate range of most amino acids, whereas the concentration of citrulline was undetectable. This might indicate that the synthesis of citrulline from ornithine at the level of ornithine transcarbamoylase was depressed. Since this enzyme utilizes ornithine and carbamoyl phosphate as substrates for the production of citrulline, it could be assumed that CPS I was not activated under these conditions. CPS I catalyses the first committed step in the urea cycle and is allosterically activated by NAG. NAG is synthesized from glutamate and acetyl-CoA by NAGS. When amino-acid breakdown increases, the concentration of glutamate increases as a result of transamination. Glutamate stimulates the synthesis of NAG. The resulting activation of CPS I increases the rate of urea production, facilitating the disposal of nitrogen as urea. Considering that the flux of glutamate is primarily committed to the glutamate dehydrogenase and glutamate-oxaloacetate transaminase reactions, the amount of glutamate available to NAGS was probably lower than the Km of glutamate for NAGS (3.7 mM [37]). This situation suggests that there is a limited amount of amino-acid breakdown, a situation that is expected for a growing well-fed animal.

Lack of correlation between tRNA concentrations, codon usage, concentration of mitochondrial transcripts and mRNA half-lives with free amino acids in mitochondria

A positive correlation has been found between tRNA in reticulocytes and the abundance of their respective amino acids in rabbit haemoglobin [59,60], suggesting that the tRNA content of reticulocytes is specialized for haemoglobin synthesis. In addition, Kurland and co-workers have shown that tRNA abundance can be correlated with the values of codon frequencies [6163]. Following this reasoning, we wanted to ascertain if there was a correlation between the levels of tRNA in mitochondria and the abundance of the respective amino acids in the mitochondrial matrix. We used the concentrations of tRNA from HeLa cell mitochondria [32] because the relative steady-state levels of the tRNAs examined in various species have rarely been found to vary by more than 10-fold and, for the majority of the tRNAs, this variation was less than 3-fold [6467]. No correlation was found between tRNA levels and the concentrations of free amino acids in mitochondria (Supplementary Figure S1; r2=0.003). Even if tRNAVal and tRNAPhe, the most abundant tRNAs, were excluded, the correlation did not improve significantly, indicating that the levels of tRNAs did not change with the concentrations of amino acids. Indeed, the concentrations of free amino acids were (on average) 30000-fold greater than those of tRNAs. This situation would result in full saturation of tRNAs with the corresponding amino acids, optimizing the rate of protein synthesis. Furthermore, considering that a severe imbalance of tRNAs can lead to alterations in translation, such as ribosomal frameshifting, amino-acid misincorporation and suppression of termination codons [33,68,69], the ratio of amino acids to tRNAs seems to be adjusted to optimize the accuracy of mitochondrial translation.

We also examined whether a correlation occurred between the concentration of free amino acids and the codon usage, the mRNA half-lives of mitochondrially-translated proteins (taken from [70]) or the concentrations of mitochondrial transcripts (see Supplementary Figures S2–S4 at http://www.bioscirep.org/bsr/028/bsr0280239add.htm). For the data set shown, there was no evidence of significant regression lines.

Lack of correlation between MesoH (mesohydrophobicity), GRAVY (grand average of hydropathicity) and H17 with free amino acids in mitochondria

It has been proposed that the absence of editing motifs in mitochondrial phenylalanine and leucine aaRSs leads to a lower fidelity during aminoacyl-tRNA synthesis in this compartment than in the cytosol [71]. The loss of the editing domain from mammalian mitochondrial leucine and phenylalanine aaRSs is in sharp contrast to the general pattern of synthetase evolution, where modules with new functions have been systematically integrated into existing proteins [72]. This also challenges the hypothesis of the origin of eukaryotes by the symbioses of two partners of well-defined groups, such as for example, Archea with α-Proteobacteria (see [72] for a complete discussion). This critical difference between mitochondrial and cytosolic aaRS evolution and function indicates that mitochondrial protein synthesis quality control might be focused at a co- or post-translational step because translation in this compartment could be inherently less accurate. Given the small number of proteins synthesized by mitochondria, reduced accuracy might be overcome by the existence of an efficient proteolytic system. Proteins accurately translated and folded would reach the inner membrane to become functional, whereas the rest would be degraded. In this regard, non-assembled COX subunit 2 from yeast is a substrate for the i-AAA (ATPase associated with various cellular activities) protease [73,74]. If this were the case, then mtDNA-encoded proteins rich in phenylalanine and leucine or, as an extension, rich in hydrophobic amino acids, would be more prone to proteolysis, and the matrix pool of amino acids should essentially reflect the composition of these mtDNA-encoded proteins according to their hydrophobic character. To test this hypothesis, we approached the problem using three different variables to assess hydrophobic content: one, GRAVY; two, MesoH, which is the average regional hydrophobicity over an extended region of amino acids, in this case, 100 amino acids; and three, H17, calculated as the maximum local hydrophobicity of a segment with a scanning window length of 17 amino acids that possess the highest mean hydrophobicity per residue. From each individual plot of free amino acids and amino-acid composition for each mtDNA-encoded subunit, a correlation coefficient was obtained. These coefficients from each mtDNA-encoded subunit were plotted against GRAVY, MesoH, and H17. None of these parameters showed a statistical significance, with r2=0.001, 0.002, and 0.54 for GRAVY, MesoH and H17 respectively. Thus these results indicated that hydrophobic proteins (evaluated by using these three different parameters) were not necessarily degraded at a higher rate than others. In agreement with this conclusion, it has been reported that approx. 10–15% (without ATP) and 40% (with ATP) of new polypeptides synthesized by isolated rat liver mitochondria are degraded within 1 hour of incubation [75], and that no single polypeptide is specifically degraded; instead all mtDNA-encoded proteins disappeared to a similar extent, in agreement with our conclusions.

Conclusions

The pool of free amino acids present in the mitochondrial matrix seems to be the result of proteolytic degradation of mtDNA-encoded proteins, rather than the fulfilment of an amino-acid profile to suit mitochondrial protein synthesis. This was inferred from the correlation between the amino-acid concentrations in the matrix and the amino-acid abundance in mtDNA-encoded proteins, the relatively high ratio of amino acids to RNA concentrations when compared with those in the cytosol, and the similar EAA/NEAA ratio between the matrix pool of amino acids and the amino-acid composition of mtDNA-encoded proteins. No correlation was found with tRNA levels or codon usage for protein synthesis, as determined from the DNA sequence and the rates of synthesis of the various polypeptides. The mitochondrial pool of amino acids appeared to be distinct from that of the cytosol and serum (individual values and EAA/NEAA ratios), reinforcing the concept that this pool is kept separate from those amino acids in the cytosol of liver and plasma (or serum), fulfilling a different role. Indeed, the EAA/NEAA ratio obtained from mitochondria (1.18) resembles more closely that found in rat red blood cells (0.8 [76]), cells which synthesized large quantities of a few proteins (globins) before losing all organelles.

If the main function of the amino-acid pool present in the mitochondrial matrix was to sustain mitochondrial protein synthesis and all tRNAs are saturated at the matrix concentrations of amino acids, then why is the profile of several free amino acids in mitochondria similar to that of the mtDNA-encoded proteins? It could be argued that the matrix pool reflects mainly either the catabolism of oxidatively modified mitochondrial proteins or that of the most abundant proteins. There are two main findings that do not support these options. One, given that mitochondrial proteins are exposed to higher oxidative and nitrative stress because these organelles constitute the main source of intracellular reactive oxygen and nitrogen species and that mitochondrial nitrated proteins have a significant faster turnover than the native unmodified proteins [6], then the matrix pool should reflect the composition of the most nitrated proteins, not necessarily the proteins encoded by mtDNA. The EAA/NEAA ratio of the most nitrated proteins (determined from our previous work [6]) was calculated to be between 0.6–0.8, significantly lower than that observed for the mitochondrial amino-acid pool. Two, the ratio of EAA/NEAA in the matrix did not reflect that of the most abundant proteins (e.g. CPS I, glutamate dehydrogenase and α- and β-subunits of ATPase), indicating that the mitochondrial pool is not the main result of catabolism of the most abundant proteins.

An alternative scenario could be envisioned if the translation of mitochondrial proteins were less accurate than cytosolic translation, their critical difference being their different origin in evolution (ancestral partners in endosymbiosis plus exchange of genetic material with others). Studies focused at understanding aaRS evolution indicate that mitochondrial protein synthesis quality control could be focused at the co-translational or post-translational steps: a misfolded protein is rapidly degraded by the mitochondrial proteolytic system because it is either not recognized by the corresponding chaperone (e.g. Oxa1) or by other subunits within a given Complex (e.g. COX subunits). Given the small number of proteins synthesized by mitochondria, the presence of an efficient proteolytic system might overcome the costly process of editing.

Abbreviations

     
  • aaRS

    aminoacyl-tRNA synthetase

  •  
  • BCAA

    branched-chain amino acid

  •  
  • COX

    cytochrome c oxidase

  •  
  • CPS I

    carbamoylphosphate synthetase I

  •  
  • EAA

    essential amino acid

  •  
  • GRAVY

    grand average of hydropathicity

  •  
  • LOR

    lysine-α-oxoglutarate reductase

  •  
  • MesoH

    mesohydrophobicity

  •  
  • MSHE buffer

    mannitol, sucrose, Hepes and EGTA buffer

  •  
  • mtDNA

    mitochondrial DNA

  •  
  • NAG

    N-acetylglutamate

  •  
  • NAGS

    N-acetylglutamate synthase

  •  
  • NEAA

    non-essential amino acid

  •  
  • RCR

    respiratory control ratio, SDH, saccharopine dehydrogenase

This study was supported by grants NIHES 005707 and 012691. We thank Professor Quinton Rogers for reading the manuscript and for providing us with excellent comments. We thank Mr Andrew Almendares for his technical assistance.

References

References
Attardi
 
G.
 
Animal mitochondrial DNA: an extreme example of genetic economy
Int. Rev. Cytol.
1985
, vol. 
93
 (pg. 
93
-
145
)
Baird
 
G. D.
 
The release of amino acids from rat-liver mitochondrial extract
Biochim. Biophys. Acta
1964
, vol. 
93
 (pg. 
293
-
303
)
Ferdinand
 
W.
Bartley
 
W.
Broomhead
 
V.
 
Amino acid production in isolated rat liver mitochondria
Biochem. J.
1973
, vol. 
134
 (pg. 
431
-
436
)
Truman
 
D. E.
Korner
 
A.
 
Incorporation of amino acids into the protein of isolated mitochondria
A search for optimum conditions and a relationship to oxidative phosphorylation
Biochem. J.
1962
, vol. 
83
 (pg. 
588
-
596
)
Giulivi
 
C.
Poderoso
 
J. J.
Boveris
 
A.
 
Production of nitric oxide by mitochondria
J. Biol. Chem.
1998
, vol. 
273
 (pg. 
11038
-
11043
)
Elfering
 
S. L.
Haynes
 
V. L.
Traaseth
 
N. J.
Ettl
 
A.
Giulivi
 
C.
 
Aspects, mechanism, and biological relevance of mitochondrial protein nitration sustained by mitochondrial nitric oxide synthase
Am. J. Physiol. Heart Circ. Physiol.
2004
, vol. 
286
 (pg. 
H22
-
H29
)
Sims
 
N. R.
 
Rapid isolation of metabolically active mitochondria from rat brain and subregions using Percoll density gradient centrifugation
J. Neurochem.
1990
, vol. 
55
 (pg. 
698
-
707
)
Chance
 
B.
Williams
 
G. R.
 
A simple and rapid assay of oxidative phosphorylation
Nature
1955
, vol. 
175
 (pg. 
1120
-
1121
)
Strydom
 
D. J.
Cohen
 
S. A.
 
Comparison of amino acid analyses by phenylisothiocyanate and 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate precolumn derivatization
Anal. Biochem.
1994
, vol. 
222
 (pg. 
19
-
28
)
Cohen
 
S. A.
De Antonis
 
K. M.
 
Applications of amino acid derivatization with 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate. Analysis of feed grains, intravenous solutions and glycoproteins
J. Chromatogr., A.
1994
, vol. 
661
 (pg. 
25
-
34
)
Institute for Laboratory Animal Research
Nutrient requirements of laboratory animals
1995
4th edn.
Washington DC
National Academy Press
(pg. 
11
-
59
)
Semon
 
B. A.
Leung
 
P. M.
Rogers
 
Q. R.
Gietzen
 
D. W.
 
Plasma and brain ammonia and amino acids in rats measured after feeding 75% casein or 28% egg white
J. Nutr.
1989
, vol. 
119
 (pg. 
1583
-
1592
)
Semon
 
B. A.
Leung
 
P. M.
Rogers
 
Q. R.
Gietzen
 
D. W.
 
Plasma ammonia, plasma, brain and liver amino acids and urea cycle enzyme activities in rats fed ammonium acetate
J. Nutr.
1989
, vol. 
119
 (pg. 
166
-
174
)
Werkheiser
 
W. C.
Bartley
 
W.
 
The study of steady-state concentrations of internal solutes of mitochondria by rapid centrifugal transfer to a fixation medium
Biochem. J.
1957
, vol. 
66
 (pg. 
79
-
91
)
Birt
 
L. M.
Bartley
 
W.
 
The pyridine nucleotide metabolism of mitochondria incubated with and without added substrates and metabolic inhibitors
Biochem. J.
1960
, vol. 
76
 (pg. 
427
-
438
)
Chomyn
 
A.
 
In vivo labeling and analysis of human mitochondrial translation products
Methods Enzymol.
1996
, vol. 
264
 (pg. 
197
-
211
)
Smirnov
 
V. Y.
Razvodovskii
 
Y. E.
Doroshenko
 
E. M.
 
Metabolic effects of mixture containing branched-chain amino acids and taurine during subchronic poisoning with barbiturates
Bull. Exp. Biol. Med.
2004
, vol. 
137
 (pg. 
443
-
446
)
Fashakin
 
J. B.
Furst
 
P.
 
Effect of dietary protein restriction on urinary nitrogen loss and tissue free amino acid patterns in mature rats
Nutr. Res.
1987
, vol. 
7
 (pg. 
285
-
298
)
Shear
 
L.
 
Internal redistribution of tissue protein synthesis in uremia
J. Clin. Invest.
1969
, vol. 
48
 (pg. 
1252
-
1257
)
Braham
 
J. E.
Bressani
 
R.
 
Effect of different levels of gossypol on transaminase activity, on nonessential to essential amino acid ratio, and on iron and nitrogen retention in rats
J. Nutr.
1975
, vol. 
105
 (pg. 
348
-
355
)
Steffen
 
M.
Sarkela
 
T. M.
Gybina
 
A. A.
Steele
 
T. W.
Trasseth
 
N. J.
Kuehl
 
D.
Giulivi
 
C.
 
Metabolism of S-nitrosoglutathione in intact mitochondria
Biochem. J.
2001
, vol. 
356
 (pg. 
395
-
402
)
Smith
 
C. V.
Jones
 
D. P.
Guenthner
 
T. M.
Lash
 
L. H.
Lauterburg
 
B. H.
 
Compartmentation of glutathione: implications for the study of toxicity and disease
Toxicol. Appl. Pharmacol.
1996
, vol. 
140
 (pg. 
1
-
12
)
Wahllander
 
A.
Soboll
 
S.
Sies
 
H.
Linke
 
I.
Muller
 
M.
 
Hepatic mitochondrial and cytosolic glutathione content and the subcellular distribution of GSH-S-transferases
FEBS Lett.
1979
, vol. 
97
 (pg. 
138
-
140
)
Hlavacova
 
N.
Bakos
 
J.
Jezova
 
D.
 
Differences in home cage behavior and endocrine parametres in rats of four strains
Endocr. Regul.
2006
, vol. 
40
 (pg. 
113
-
118
)
Walberer
 
M.
Stolz
 
E.
Muller
 
C.
Friedrich
 
C.
Rottger
 
C.
Blaes
 
F.
Kaps
 
M.
Fisher
 
M.
Bachmann
 
G.
Gerriets
 
T.
 
Experimental stroke: ischaemic lesion volume and oedema formation differ among rat strains (a comparison between Wistar and Sprague–Dawley rats using MRI)
Lab. Anim.
2006
, vol. 
40
 (pg. 
1
-
8
)
Geis
 
R. B.
Diel
 
P.
Degen
 
G. H.
Vollmer
 
G.
 
Effects of genistein on the expression of hepatic genes in two rat strains (Sprague–Dawley and Wistar)
Toxicol. Lett.
2005
, vol. 
157
 (pg. 
21
-
29
)
Blemings
 
K. P.
Crenshaw
 
T. D.
Benevenga
 
N. J.
 
Mitochondrial lysine uptake limits hepatic lysine oxidation in rats fed diets containing 5, 20 or 60% casein
J. Nutr.
1998
, vol. 
128
 (pg. 
2427
-
2434
)
Alberti
 
K. G.
Bartley
 
W.
 
The localization of proteolytic activity in rat liver mitochondria and its relation to mitochondrial swelling and aging
Biochem. J.
1969
, vol. 
111
 (pg. 
763
-
776
)
Soboll
 
S.
Sholz
 
R.
Heldt
 
H. W.
 
Subcellular metabolite concentrations. Dependence of mitochondrial and cytosolic ATP systems on the metabolic state of perfuse rat liver
Eur. J. Biochem.
1978
, vol. 
87
 (pg. 
377
-
390
)
Rogers
 
Q.
 
The nutritional and metabolic effects of amino acid imbalances
Protein Metab. Nutr.
1976
, vol. 
16
 (pg. 
279
-
301
)
Bonnefond
 
L.
Fender
 
A.
Rudinger-Thirion
 
J.
Giege
 
R.
Florentz
 
C.
Sissler
 
M.
 
Toward the full set of human mitochondrial aminoacyl-tRNA synthetases: characterization of AspRS and TyrRS
Biochemistry
2005
, vol. 
44
 (pg. 
4805
-
4816
)
King
 
M. P.
Attardi
 
G.
 
Post-transcriptional regulation of the steady-state levels of mitochondrial tRNAs in HeLa cells
J. Biol. Chem.
1993
, vol. 
268
 (pg. 
10228
-
10237
)
Ulrich
 
A. K.
Li
 
L. Y.
Parker
 
J.
 
Codon usage, transfer RNA availability and mistranslation in amino acid starved bacteria
Biochim. Biophys. Acta
1991
, vol. 
1089
 (pg. 
362
-
366
)
Chomyn
 
A
Attardi
 
G.
 
Papa
 
S.
Chance
 
B.
Ernster
 
L.
 
Heterogeneous efficiencies of mRNA translation in human mitochondria
Cytochrome Systems: molecular biology and bioenergetics
1987
New York
Plenum Press
(pg. 
145
-
152
)
Hatefi
 
Y.
Ragan
 
C. I.
Galante
 
Y. M.
 
Martonosi
 
A.
 
The enzymes and the enzyme complexes of the mitochondrial oxidative phosphorylation system
The Enzymes of Biological Membranes
1985
New York
Plenum Press
(pg. 
1
-
70
)
Giulivi
 
C.
 
Functional implications of nitric oxide produced by mitochondria in mitochondrial metabolism
Biochem. J.
1998
, vol. 
332
 (pg. 
673
-
679
)
Sonoda
 
T.
Tatibana
 
M.
 
Purification of N-acetyl-L-glutamate synthetase from rat liver mitochondria and substrate and activator specificity of the enzyme
J. Biol. Chem.
1983
, vol. 
258
 (pg. 
9839
-
9844
)
Gong
 
J.
Hunter
 
G. A.
Ferreira
 
G. C.
 
Aspartate-279 in aminolevulinate synthase affects enzyme catalysis through enhancing the function of the pyridoxal 5′-phosphate cofactor
Biochemistry
1998
, vol. 
37
 (pg. 
3509
-
3517
)
Kunin
 
A. S.
Tannen
 
R. L.
 
Regulation of glutamate metabolism by renal cortical mitochondria
Am. J. Physiol.
1979
, vol. 
237
 (pg. 
F55
-
F62
)
Schoolwerth
 
A. C.
LaNoue
 
K. F.
 
The role of microcompartmentation in the regulation of glutamate metabolism by rat kidney mitochondria
J. Biol. Chem.
1980
, vol. 
255
 (pg. 
3403
-
3411
)
Schaffer
 
S.
Solodushko
 
V.
Azuma
 
J.
 
Taurine-deficient cardiomyopathy: role of phospholipids, calcium and osmotic stress
Adv. Exp. Med. Biol.
2000
, vol. 
483
 (pg. 
57
-
69
)
Lourenco
 
R.
Camilo
 
M. E.
 
Taurine: a conditionally essential amino acid in humans? An overview in health and disease
Nutr. Hosp.
2002
, vol. 
17
 (pg. 
262
-
270
)
Militante
 
J. D.
Lombardini
 
J. B.
 
Taurine: evidence of physiological function in the retina
Nutr. Neurosci.
2002
, vol. 
5
 (pg. 
75
-
90
)
Pion
 
P. D.
Kittleson
 
M. D.
Rogers
 
Q. R.
Morris
 
J. G.
 
Myocardial failure in cats associated with low plasma taurine: a reversible cardiomyopathy
Science
1987
, vol. 
237
 (pg. 
764
-
768
)
Umeda
 
N.
Suzuki
 
T.
Yukawa
 
M.
Ohya
 
Y.
Shindo
 
H.
Watanabe
 
K.
Suzuki
 
T.
 
Mitochondria-specific RNA-modifying enzymes responsible for the biosynthesis of the wobble base in mitochondrial tRNAs. Implications for the molecular pathogenesis of human mitochondrial diseases
J. Biol. Chem.
2005
, vol. 
280
 (pg. 
1613
-
1624
)
Suzuki
 
T.
Suzuki
 
T.
Wada
 
T.
Saigo
 
K.
Watanabe
 
K.
 
Taurine as a constituent of mitochondrial tRNAs: new insights into the functions of taurine and human mitochondrial diseases
EMBO J.
2002
, vol. 
21
 (pg. 
6581
-
6589
)
Islam
 
M. M.
Wallin
 
R.
Wynn
 
R. M.
Conway
 
M.
Fujii
 
H.
Mobley
 
J. A.
Chuang
 
D. T.
Hutson
 
S. M.
 
A novel branched-chain amino acid metabolon: protein-protein interactions in a supramolecular complex
J. Biol. Chem.
2007
, vol. 
282
 (pg. 
11893
-
11903
)
Blemings
 
K. P.
Crenshaw
 
T. D.
Swick
 
R. W.
Benevenga
 
N. J.
 
Lysine-α-ketoglutarate reductase and saccharopine dehydrogenase are located only in the mitochondrial matrix in rat liver
J. Nutr.
1994
, vol. 
124
 (pg. 
1215
-
1221
)
Mihalik
 
S. J.
Rhead
 
W. J.
 
Species variation in organellar location and activity of L-pipecolic acid oxidation in mammals
J. Comp. Physiol. B, Biochem. Syst. Environ. Physiol.
1991
, vol. 
160
 (pg. 
671
-
676
)
Fjellstedt
 
T. A.
Robinson
 
J. C.
 
Properties of partially purified saccharopine dehydrogenase from human placenta
Arch Biochem Biophys.
1975
, vol. 
171
 (pg. 
191
-
196
)
Fjellstedt
 
T. A.
Robinson
 
J. C.
 
Purification and properties of L-lysine-α-ketoglutarate reductase from human placenta
Arch. Biochem. Biophys.
1975
, vol. 
168
 (pg. 
536
-
548
)
Hutzler
 
J.
Dancis
 
J.
 
Lysine-ketoglutarate reductase in human tissues
Biochim. Biophys. Acta
1975
, vol. 
377
 (pg. 
42
-
51
)
Hutzler
 
J.
Dancis
 
J.
 
Conversion of lysine to saccharopine by human tissues
Biochim. Biophys. Acta
1968
, vol. 
158
 (pg. 
62
-
69
)
Kemper
 
E. L.
Neto
 
G. C.
Papes
 
F.
Moraes
 
K. C.
Leite
 
A.
Arruda
 
P.
 
The role of opaque2 in the control of lysine-degrading activities in developing maize endosperm
Plant Cell
1999
, vol. 
11
 (pg. 
1981
-
1994
)
Noda
 
C.
Ichihara
 
A.
 
Control of ketogenesis from amino acids. IV. Tissue specificity in oxidation of leucine, tyrosine, and lysine
J. Biochem. (Tokyo)
1976
, vol. 
80
 (pg. 
1159
-
1164
)
Papes
 
F.
Kemper
 
E. L.
Cord-Neto
 
G.
Langone
 
F.
Arruda
 
P.
 
Lysine degradation through the saccharopine pathway in mammals: involvement of both bifunctional and monofunctional lysine-degrading enzymes in mouse
Biochem. J.
1999
, vol. 
344
 (pg. 
555
-
563
)
Bate
 
A. J.
Dickson
 
A. J.
 
Relative importance of physiological precursors for ketogenesis in the nutritional homeostasis of the development of the embryonic chick
Biochim. Biophys. Acta
1987
, vol. 
924
 (pg. 
7
-
12
)
Duszynski
 
J.
Mueller
 
G.
LaNoue
 
K.
 
Microcompartmentation of aspartate in rat liver mitochondria
J. Biol. Chem.
1978
, vol. 
253
 (pg. 
6149
-
6157
)
Litt
 
M.
Kabat
 
D.
 
Studies of transfer ribonucleic acids and of hemoglobin synthesis in sheep reticulocytes
J. Biol. Chem.
1972
, vol. 
247
 (pg. 
6659
-
6664
)
Smith
 
D. W.
McNamara
 
A. L.
 
Specialization of rabbit reticulocyte transfer RNA content for hemoglobin synthesis
Science
1971
, vol. 
171
 (pg. 
577
-
579
)
Dong
 
H.
Nilsson
 
L.
Kurland
 
C. G.
 
Co-variation of tRNA abundance and codon usage in Escherichia coli at different growth rates
J. Mol. Biol.
1996
, vol. 
260
 (pg. 
649
-
663
)
Emilsson
 
V.
Kurland
 
C. G.
 
Growth rate dependence of transfer RNA abundance in Escherichia coli
EMBO J.
1990
, vol. 
9
 (pg. 
4359
-
4366
)
Emilsson
 
V.
Naslund
 
A. K.
Kurland
 
C. G.
 
Growth-rate-dependent accumulation of twelve tRNA species in Escherichia coli
J. Mol. Biol.
1993
, vol. 
230
 (pg. 
483
-
491
)
Yamao
 
F.
Andachi
 
Y.
Muto
 
A.
Ikemura
 
T.
Osawa
 
S.
 
Levels of tRNAs in bacterial cells as affected by amino acid usage in proteins
Nucleic Acids Res.
1991
, vol. 
19
 (pg. 
6119
-
6122
)
Ikemura
 
T.
 
Correlation between the abundance of Escherichia coli transfer RNAs and the occurrence of the respective codons in its protein genes
J. Mol. Biol.
1981
, vol. 
146
 (pg. 
1
-
21
)
Ikemura
 
T.
 
Correlation between the abundance of yeast transfer RNAs and the occurrence of the respective codons in protein genes. Differences in synonymous codon choice patterns of yeast and Escherichia coli with reference to the abundance of isoaccepting transfer RNAs
J. Mol. Biol.
1982
, vol. 
158
 (pg. 
573
-
597
)
Ikemura
 
T.
Ozeki
 
H.
 
Codon usage and transfer RNA contents: organism-specific codon-choice patterns in reference to the isoacceptor contents
Cold Spring Harb. Symp. Quant. Biol.
1983
, vol. 
47 Pt 2
 (pg. 
1087
-
1097
)
Atkins
 
J. F.
Gesteland
 
R. F.
Reid
 
B. R.
Anderson
 
C. W.
 
Normal tRNAs promote ribosomal frameshifting
Cell
1979
, vol. 
18
 (pg. 
1119
-
1131
)
Weiss
 
W. A.
Friedberg
 
E. C.
 
Normal yeast tRNA(CAGGln) can suppress amber codons and is encoded by an essential gene
J. Mol. Biol.
1986
, vol. 
192
 (pg. 
725
-
735
)
Anderson
 
S.
Bankier
 
A. T.
Barrell
 
B. G.
de Bruijn
 
M. H.
Coulson
 
A. R.
Drouin
 
J.
Eperon
 
I. C.
Nierlich
 
D. P.
Roe
 
B. A.
Sanger
 
F.
, et al 
Sequence and organization of the human mitochondrial genome
Nature
1981
, vol. 
290
 (pg. 
457
-
465
)
Roy
 
H.
Ling
 
J.
Alfonzo
 
J.
Ibba
 
M.
 
Loss of editing activity during the evolution of mitochondrial phenylalanyl-tRNA synthetase
J. Biol. Chem.
2005
, vol. 
280
 (pg. 
38186
-
38192
)
Brindefalk
 
B.
Viklund
 
J.
Larsson
 
D.
Thollesson
 
M.
Andersson
 
S. G.
 
Origin and evolution of the mitochondrial aminoacyl-tRNA synthetases
Mol. Biol. Evol.
2007
, vol. 
24
 (pg. 
743
-
756
)
Nakai
 
T.
Yasuhara
 
T.
Fujiki
 
Y.
Ohashi
 
A.
 
Multiple genes, including a member of the AAA family, are essential for degradation of unassembled subunit 2 of cytochrome c oxidase in yeast mitochondria
Mol. Cell. Biol.
1995
, vol. 
15
 (pg. 
4441
-
4452
)
Weber
 
E. R.
Hanekamp
 
T.
Thorsness
 
P. E.
 
Biochemical and functional analysis of the YME1 gene product, an ATP and zinc-dependent mitochondrial protease from S. cerevisiae
Mol. Biol. Cell
1996
, vol. 
7
 (pg. 
307
-
317
)
Desautels
 
M.
Goldberg
 
A. L.
 
Liver mitochondria contain an ATP-dependent, vanadate-sensitive pathway for the degradation of proteins
Proc. Natl. Acad. Sci. U.S.A.
1982
, vol. 
79
 (pg. 
1869
-
1873
)
Filho
 
J. C.
Hazel
 
S. J.
Anderstam
 
B.
Bergstrom
 
J.
Lewitt
 
M.
Hall
 
K.
 
Effect of protein intake on plasma and erythrocyte free amino acids and serum IGF-I and IGFBP-1 levels in rats
Am. J. Physiol.
1999
, vol. 
277
 (pg. 
E693
-
E701
)
Bauerly
 
K. A.
Storms
 
D. H.
Harris
 
C. B.
Hajizadeh
 
S.
Sun
 
M. Y.
Cheung
 
C. P.
Satre
 
M. A.
Fascetti
 
A. J.
Tchaparian
 
E.
Rucker
 
R. B.
 
Pyrroloquinoline quinone nutritional status alters lysine metabolism and modulates mitochondrial DNA content in the mouse and rat. Biochim. Biophys
Acta
2006
, vol. 
1760
 (pg. 
1741
-
1748
)
Baum
 
J. I.
Layman
 
D. K.
Freund
 
G. G.
Rahn
 
K. A.
Nakamura
 
M. T.
Yudell
 
B. E.
 
A reduced carbohydrate, increased protein diet stabilizes glycemic control and minimizes adipose tissue glucose disposal in rats
J. Nutr.
2006
, vol. 
136
 (pg. 
1855
-
1861
)
Watanabe
 
A.
Higashi
 
T.
Sakata
 
T.
Nagashima
 
H.
 
Serum amino acid levels in patients with hepatocellular carcinoma
Cancer
1984
, vol. 
54
 (pg. 
1875
-
1882
)
Hagman
 
M.
Eriksson
 
T.
Kitson
 
K. E.
 
Similar effects of ethanol and tert-butanol on amino acid concentrations in rat serum and liver
Alcohol Clin. Exp. Res.
1993
, vol. 
17
 (pg. 
299
-
303
)
Soley
 
M.
Alemany
 
M.
 
A rapid method for the estimation of amino acid concentration in liver tissue
J. Biochem. Biophys. Methods
1980
, vol. 
2
 (pg. 
207
-
211
)
Cheung
 
C. W.
Cohen
 
N. S.
Raijman
 
L.
 
Channeling of urea cycle intermediates in situ in permeabilized hepatocytes
J. Biol. Chem.
1989
, vol. 
264
 (pg. 
4038
-
4044
)
Cohen
 
N. S.
Cheung
 
C. W.
Raijman
 
L.
 
Channeling of extramitochondrial ornithine to matrix ornithine transcarbamylase
J. Biol. Chem.
1987
, vol. 
262
 (pg. 
203
-
208
)
Srere
 
P. A.
 
Enzyme concentrations in tissues
Science
1967
, vol. 
158
 (pg. 
936
-
937
)
Srere
 
P. A.
Sumegi
 
B.
 
Organization of the mitochondrial matrix
Adv. Exp. Med. Biol.
1986
, vol. 
194
 (pg. 
13
-
25
)
Srivastava
 
D. K.
Bernhard
 
S. A.
 
Metabolite transfer via enzyme-enzyme complexes Science
1986
, vol. 
234
 (pg. 
1081
-
1086
)
Posakony
 
J. W.
England
 
J. M.
Attardi
 
G.
 
Mitochondrial growth and division during the cell cycle in HeLa cells
J. Cell Biol.
1977
, vol. 
74
 (pg. 
468
-
491
)
Schwerzmann
 
K.
Cruz-Orive
 
L. M.
Eggman
 
R.
Sanger
 
A.
Weibel
 
E. R.
 
Molecular architecture of the inner membrane of mitochondria from rat liver: a combined biochemical and stereological study
J. Cell Biol.
1986
, vol. 
102
 (pg. 
97
-
103
)
Klingenberg
 
M.
Pfaff
 
E.
 
Tager
 
J. M.
 
Structural and functional compartmentation in mitochondria
Regulation of Metabolic Processes in Mitochondria
1966
New York
Elsevier
(pg. 
180
-
201
)

Supplementary data