Metabolic compartmentation of vertebrate glutamine synthetase: putative mitochondrial targeting signal in avian liver glutamine synthetase.

The evolution of uricoteley as a mechanism for hepatic ammonia detoxication in vertebrates required targeting of glutamine synthetase (GS) to liver mitochondria in the sauropsid line of descent leading to the squamate reptiles and archosaurs. Previous studies have shown that in birds and crocodilians, sole survivors of the archosaurian line, hepatic GS is translated without a transient, N-terminal targeting signal common to other mitochondrial matrix proteins. To identify a putative internal targeting sequence in the avian enzyme, the amino acid sequence of chicken liver GS was derived by a combination of sequencing of cloned cDNA, direct sequencing of mRNA, and sequencing of polymerase chain reaction (PCR) products amplified from reverse-transcribed mRNA. Analysis of the first 20 or so N-terminal amino acids of the derived sequence for the chicken enzyme shows that they are devoid of acidic amino acids, contain several hydroxy amino acids, and can be predicted to form a positively charged, amphipathic helix, all of which are characteristic properties of mitochondrial targeting signals. A comparison of the N-terminus of chicken GS with the N-termini of cytosolic mammalian GSs indicates that at least three amino acid replacements may have been responsible for converting the N-terminus of the cytosolic mammalian enzyme into a mitochondrial targeting signal. Two of these, His15 and Lys19, result in additional positive charges, as well as in changes in hydrophilicity. Both could have resulted from third-base-codon substitutions. A third replacement, Ala12, may contribute to the helicity of the N-terminus of the chicken enzyme. The N-terminus of the cytosolic chicken brain GS (positions 1-36) was found to be identical to that of the liver enzyme. The complete sequence of chicken retinal GS is also identical to that of the liver enzyme. GS is coded by a single gene in birds, so these sequence data suggest that, unlike the situation in other tissue-specific compartmental isozymes, differential targeting of avian GS to the mitochondrial or cytosolic compartments is not dependent on the sequence of the primary translation product of its mRNA but may involve some other tissue-specific factor(s).     

Glutamine in the Pathogenesis of Hepatic Encephalopathy: The Trojan Horse Hypothesis Revisited

Hepatic encephalopathy (HE) is major neuropsychiatric disorder occurring in patients with severe liver disease and ammonia is generally considered to represent the major toxin responsible for this condition. Ammonia in brain is chiefly metabolized (“detoxified”) to glutamine in astrocytes due to predominant localization of glutamine synthetase in these cells. While glutamine has long been considered innocuous, a deleterious role more recently has been attributed to this amino acid. This article reviews the mechanisms by which glutamine contributes to the pathogenesis of HE, how glutamine is transported into mitochondria and subsequently hydrolyzed leading to high levels of ammonia, the latter triggering oxidative and nitrative stress, the mitochondrial permeability transition and mitochondrial injury, a sequence of events we have collectively termed as the Trojan horse hypothesis of hepatic encephalopathy.     

Avian mitochondrial glutamine metabolism.

Intact avian liver mitochondria were shown to synthesize glutamine from glutamate in the absence of exogenous ATP and ammonia. With L-[U-14C]glutamate as the substrate, there was an approximate 1:1 stoichiometry between glutamate deaminated (as measured by the release of 14CO2 due to alpha-keto-[14C]glutarate oxidation) and glutamate amidated. With L-[15N]glutamate as the substrate, the isolated glutamine was shown by low and high resolution mass spectrometry of its phenylisothiocyanate derivative to contain 15N in both the alpha-amino and amide groups. Thus, for each mole of glutamate taken up, approximately 0.5 mol is deaminated and the other 0.5 mol serves as a substrate for glutamine synthetase previously localized in these mitochondria (Vorhaben, J. E., and Campbell, J. W. (1972) J. Biol. Chem. 247,2763). The permeability of L-glutamine to intact avian liver mitochondria was studied by a rapid centrifugation technique. Efflux as well as influx of L-glutamine were both rapid and appeared to occur by a passive, energy-independent process. These results indicate that the mitochondrial glutamine synthetase present in uricotelic species represents the primary ammonia detoxication reaction in that ammonia released intramitochondrially during amino acid catabolism is converted to glutamine for efflux to the cytosol where it may serve as a substrate for purine (uric acid) biosynthesis.     

Glutamine synthetase in liver of the American alligator, Alligator mississippiensis

Glutamine synthetase was shown to be localized in liver mitochondria of the American alligator, Alligator mississippiensis, by immunofluorescent staining of frozen liver sections and by the detection of enzymatic activity and immunoreactive protein in the mitochondrial fraction following subcellular fractionation of liver tissue by differential centrifugation. The primary translation product of alligator liver glutamine synthetase mRNA was shown to have an Mr = 45,000 which is similar if not identical in size to that of the mature subunit. This mRNA was found to be heterogeneous in size with a major form corresponding to 2.8-3.0 kb and a lesser form corresponding to around 2 kb. Both are in excess of the size required to code for the glutamine synthetase subunit. The synthesis and presumably the mitochondrial import of glutamine synthetase in alligator liver are thus very similar to the same processes in avian liver. Despite the excretion of a high percentage of nitrogen as ammonia, the demonstration of a mitochondrial glutamine synthetase indicates the alligator has the typical avian-type uricotelic ammonia-detoxification system in liver. This suggests that the transition to uricotelism occurred in the sauropsid line of evolution and has persisted through both the lepidosaurian (snakes, lizards) and archosaurian (dinosaurs, crocodilians, birds) lines.     

Changes in adenosine receptors during differentiation of 3T3-F442A cells to adipocytes.

We measured the amino acid concentrations in the afferent and efferent vessels of the liver in anaesthetized fed adult rats and in fed suckling rat pups. A much higher content of glutamine in the portal vein and the aorta than in hepatic veins suggests that this amino acid is actively taken up by the liver of fed suckling rat pups, conversely to what is found in adult rats. In an attempt to characterize further the mechanism(s) contributing to this enhanced glutamine uptake, we monitored the time course of 1 mM-glutamine transport into plasma-membrane vesicles purified from the livers of either adult or suckling rats. The concentrative Na+-dependent uptake of glutamine was lower in those vesicles obtained from pups than in those obtained from adult rats. Glutaminase and glutamine synthetase activities in livers from both experimental groups were also measured. Glutaminase and glutamine synthetase activities in suckling rats were about 3-fold higher and 2-fold lower respectively than those in adult rats. It is concluded that glutamine is a main nitrogen carrier to the liver in fed suckling rats. A high availability of this amino acid and an enzyme imbalance between glutamine-synthesizing and -degrading activities may account for the net uptake found in vivo.     

Submitochondrial localization and function of enzymes of glutamine metabolism in avian liver

Glutamine synthetase (EC 6.3.1.2) was localized within the matrix compartment of avian liver mitochondria. The submitochondrial localization of this enzyme was determined by the digitonin-Lubrol method of Schnaitman and Greenawalt (35). The matrix fraction contained over 74% of the glutamine synthetase activity and the major proportion of the matirx marker enzymes, malate dehydrogenase (71%), NADP-dependent isocitrate dehydrogenase (83%), and glutamate dehydrogenase (57%). The highest specific activities of these enzymes were also found in the matrix compartment. Oxidation of glutamine by avian liver mitochondria was substantially less than that of glutamate. Bromofuroate, an inhibitor of glutamate dehydrogenase, blocked oxidation of glutamate and of glutamine whereas aminoxyacetate, a transaminase inhibitor, had little or no effect with either substrate. These results indicate that glutamine metabolism is probably initiated by the conversion of glutamine to glutamate rather than to an alpha-keto acid. The localization of a glutaminase activity within avian liver mitochondria plus the absence of an active mitochondrial glutamine transaminase is consistent with the differential effects of the transaminase and glutamate dehydrogenase inhibitors. The high glutamine synthetase activity (40:1) suggests that mitochondrial catabolism of glutamine is minimal, freeing most of the glutamine synthesized for purine (uric acid) biosynthesis.     

Subcellular location of glutamine synthetase and urea cycle enzymes in liver of spiny dogfish (Squalus acanthias)

Glutamine synthetase and glutamine- and acetylglutamate-dependent carbamoyl-phosphate synthetase, both of which are present in high concentrations in liver of urea-retaining elasmobranchs, have been found to be located exclusively in the mitochondria in liver from the representative elasmobranch Squalus acanthias. This observation is consistent with the view that the function of this unique carbamoyl-phosphate synthetase is related to urea synthesis, and that the initial nitrogen-donating substrate for urea synthesis in these species is glutamine rather than ammonia. The urea cycle enzymes, ornithine carbamoyltransferase and arginase, are also located in the mitochondria, whereas argininosuccinate synthetase and argininosuccinate lyase are located in the cytosol. Glutamine synthetase and arginase are mitochondrial enzymes in uricotelic species, but are normally found in the cytoplasm in ureotelic species. the properties of the elasmobranch arginase, however, are characteristic of arginases from ureotelic species (e.g. the Km for arginine is 1.2 mM, and the enzyme has an Mr congruent to 100,000).     

Submitochondrial localization of arginase and other enzymes associated with urea synthesis and nitrogen metabolism, in liver of Squalus acanthias

The submitochondrial localization of the four mitochondrial enzymes associated with urea synthesis in liver of Squalus acanthias (spiny dogfish), a representative elasmobranch, was determined. Glutamine- and acetylglutamate-dependent carbamoyl-phosphate synthetase, ornithine carbamoyltransferase, glutamine synthetase, and arginase were all localized within the matrix of liver mitochondria. The subcellular and submitochondrial localization and activities of several related enzymes involved in nitrogen metabolism and gluconeogenesis in liver and dogfish are also reported. Pyruvate carboxylase and phosphoenolpyruvate carboxykinase were localized in the mitochondrial matrix. Synthesis of citrulline by isolated mitochondria from ornithine proceeds at a near optimal rate at ornithine concentrations as low as 0.08 mM. The same stoichiometry and rates of citrulline synthesis are observed when ornithine is replaced by arginine. The mitochondrial location of arginase does not appear to reflect a mechanism for regulating ornithine availability.     

Enzymology of Glutamine Metabolism Related to Senescence and Seed Development in the Pea (Pisum sativum L.)

The metabolism of glutamine in the leaf and subtended fruit of the aging pea (Pisum sativum L. cv. Burpeeana) has been studied in relation to changes in the protein, chlorophyll, and free amino acid content of each organ during ontogenesis. Glutamine synthetase [EC 6.3.1.2] activity was measured during development and senescence in each organ. Glutamate synthetase [EC 2.6.1.53] activity was followed in the pod and cotyledon during development and maturation. Maximal glutamine synthetase activity and free amino acid accumulation occurred together in the young leaf. Glutamine synthetase (in vitro) in leaf extracts greatly exceeded the requirement (in vivo) for reduced N in the organ. Glutamine synthetase activity, although declining in the senescing leaf, was sufficient (in vitro) to produce glutamine from all of the N released during protein hydrolysis (in vivo). Maximal glutamine synthetase activity in the pod was recorded 6 days after the peak accumulation of the free amino acids in this organ.

In the young pod, free amino acids accumulated as glutamate synthetase activity increased. Maximal pod glutamate synthetase activity occurred simultaneously with maximal leaf glutamine synthetase activity, but 6 days prior to the corresponding maximum of glutamine synthetase in the pod. Cotyledonary glutamate synthetase activity increased during the assimilatory phase of embryo growth which coincided with the loss of protein and free amino acids from the leaf and pod; maximal activity was recorded simultaneously with maximal pod glutamine synthetase.

We suggest that the activity of glutamine synthetase in the supply organs (leaf, pod) furnishes the translocated amide necessary for the N nutrition of the cotyledon. The subsequent activity of glutamate synthetase could provide a mechanism for the transfer of imported amide N to alpha amino N subsequently used in protein synthesis. In vitro measurements of enzyme activity indicate there was sufficient catalytic potential in vivo to accomplish these proposed roles.

     

Genetic basis for tissue isozymes of glutamine synthetase in elasmobranchs

Tissue-specific isozymes of glutamine synthetase are present in elasmobranchs. A larger isozyme occurs in tissues in which the enzyme is localized in mitochondria (liver, kidney) whereas a smaller form occurs in tissues in which it is cytosolic (brain, spleen, etc.). The nucleotide sequence of spiny dogfish shark (Squalus acanthias) liver glutamine synthetase mRNA, derived from its cDNA, shows there are two in-frame initiation codons (AUG) at the N-terminus which will account for the size differences between the two isozymes. Initiation at the up-stream and down-stream sites would yield peptides of 45,406 and 41,869 mol. wts. representing the precursor of the mitochondrial isozyme and the cytosolic isozyme, respectively. The additional N-terminal 29 amino acids present in the mitochondrial isozyme precursor contains two putative cleavage sites based on the Arg-X-(Phe,Ile,Leu) motif. The predicted two-step processing would remove 14 of the 29 N-terminal amino acids. These 14 amino acids can be predicted to form a very strong amphipathic mitochondrial targeting signal. Their removal would yield a mature peptide of 43,680 mol. wt. The calculated mol. wts. based on the derived amino acid sequence are therefore in good agreement with previous estimates of an approximately 1.5–2-kDa difference between the M_rs of the mitochondrial and cytosolic isozymes. A model for the evolution of the mitochondrial targeting of glutamine synthetase in vertebrates is proposed. Correspondence to: J.W. CampbellThe nucleotide sequence reported will appear in GenBank under accession number U04617     

IMP-GMP 5'-nucleotidase in reptiles: occurrence and tissue distribution in a crocodile and three species of lizards

IMP-hydrolyzing activity (which is reactive with goose anti-pig lung IMP-GMP 5'-nucleotidase (c-N-II: EC.3.1.3.5) serum) was detected in extracts from several tissues (liver, heart, kidney, spleen, stomach, lung and skeletal muscle) from constitutively uricotelic reptiles: a crocodile (Crocodylus siamensis), and three species of lizard (Furcifer oustaleti, Tupinambis rufescens and Varanus gouldi). The activities were markedly high in the livers: 3.0 units/g in the crocodile and 1.4-2.9 units/g in the lizards. These were similar to those previously reported for the livers from chicken and snakes (also constitutively uricotelic), and 4- to 10-fold higher than those in ammoniotelic or ureotelic vertebrates. These findings suggest that the high activity of IMP-GMP 5'-nucleotidase in the liver is a feature of constitutive uricotelism, and that the enzyme may participate in the production of uric acid as an end product of amino acid catabolism.     

The sleeper Bostrichthys sinensis (family Eleotridae) stores glutamine and reduces ammonia production during aerial exposure

Bostrichthys sinensis inhabits brackish water, living in the crevices of the river mouths of Shang Xi and Guangdong, China. In its natural habitat, it may encounter aerial exposure frequently during low tides, and it usually remains quiescent in the absence of water. Upon aerial exposure in the laboratory, the ammonia excretion rate decreased to one-fourth that of the submerged control. Although all the enzymes of the ornithine-urea cycle were detected in the liver of this fish, the activity of hepatic carbamoyl phosphate synthetase was too low for the cycle to be functioning. Indeed, ammonia accumulated in the tissues and was not converted to urea. Results indicate that ammonia produced through amino acid catabolism was detoxified to glutamine during the first 24 h of aerial exposure. The excess amount of glutamine stored in the muscle during this period couldaccount approximately for the reduction in ammonia equivalent excreted. There was indeed a significant increase in the activity of glutamine synthetase from the liver of specimens exposed to terrestrial conditions. In contrast to the production of alanine, formation of glutamine is energetically expensive. Since B. sinensis remained relatively inactive on land, the reduction in energy demand for muscular activity might provide it with the opportunity to exploit glutamine formation as a means to detoxify ammonia. After 72 h of aerial exposure, B. sinensis reduced internal ammonia production, possibly through reductions in proteolysis and amino acid catabolism, to avoid excessive accumulation of ammonia.     

Effect of different dietary proteins on plasma and liver fatty acid compositions in growing rats

The activities of key glutamine and urea cycle enzymes were assayed in liver homogenates from control and chronically acidotic rats and compared with citrulline and urea productions by isolated mitochondria and intact liver slices, respectively. Glutamine-dependent urea and citrulline synthesis were increased significantly in isolated mitochondria and in liver slices; the activities of carbamoyl phosphate synthetase and arginase were unchanged and increased, respectively. Glutamine was not a precursor in the carbamoyl phosphate synthetase system, suggesting that the glutamine effect is an indirect one and that glutamine requires prior hydrolysis. Increased mitochondrial citrulline synthesis was associated with enhanced oxygen consumption, suggesting glutamine acts both as a nitrogen and fuel source. Hepatic phosphate-dependent glutaminase was elevated by chronic acidosis. The results indicate that the acidosis-induced reduction in ureagenesis and reversal from glutamine uptake to release observed in vivo are not reflections of corresponding changes in the hepatic enzyme content. Rather, when available, glutamine readily supports ureagenesis, suggesting a close coupling of hepatic glutaminase flux with citrulline synthesis.     

Heterogeneous distribution of glutamine synthetase during rat liver development

Two days before birth, immunohistochemical detection of glutamine synthetase already reveals a heterogeneous distribution pattern related to the vascular architecture of the liver. Only a small number of hepatocytes in the vicinity of the efferent venules show relatively high staining intensity. Before that age, only megakaryocytes show intense staining, while liver parenchyma is only faintly stained. The developmental profile of glutamine synthetase activity shows two periods of increasing enzyme activity: one in the perinatal period and one in the second and third postnatal week. Both periods are correlated with high levels of circulating corticosteroid hormones. Although the relative number of intensely stained hepatocytes increases during the first rise in enzyme activity, the second rise is correlated with a decreasing number of glutamine synthetase-positive hepatocytes which, however, show a considerable increase in staining intensity. Carbamoylphosphate synthetase shows a homogeneous distribution pattern in the perinatal period. Conditions that lead during development to a relatively high level of glutamine synthetase expression in the pericentral compartment apparently originate before the appearance of conditions that lead to a relatively high level of carbamoylphosphate synthetase gene expression in the periportal compartment. Our results indicate that downstream localization of glutamine synthetase in liver acinus is essential from the perinatal period onwards, whereas reciprocal distribution of glutamine synthetase and carbamoylphosphate synthetase gene expression (that is found in adult rat liver) is not.     

Heterogeneous distribution of glutamine synthetase among rat liver parenchymal cells in situ and in primary culture.

The distribution of glutamine synthetase [L-glutamate: ammonia ligase (ADP-forming), EC 6.3.1.1)] among rat liver parenchymal cells in situ and in primary culture was investigated by indirect immunofluorescence using a specific antiserum. In intact liver, the enzyme was found to be localized exclusively within a very small population of the parenchymal cells surrounding the terminal hepatic venules. Other parts of the parenchyma including non-parenchymal cell types did not stain for this enzyme. Heterogeneity was preserved during isolation of liver parenchymal cells and persisted in cultured cells for at least 3 days. Despite alterations in enzyme activity due to the adaptation of the cells to the culture conditions or due to the hormonal stimulation of the enzyme activity, no change in the relative number of cells expressing this enzyme could be detected. This rather peculiar localization of glutamine synthetase demonstrates an interesting aspect of liver zonation and might have important implications for liver glutamine and, more generally, nitrogen metabolism. Furthermore, it raises the question of whether there might be a phenotypic difference among liver parenchymal cells.     

Amino acid neurotransmitters and their receptors in the brain synaptosomes of acute hepatic failure rats

Ammonia contents in the brain stem and prosencephalon markedly increased in a rat model of acute hepatic failure induced by partial hepatectomy following CCl4 intoxication. In hepatic failure rats, synaptosomal glutamic acid (excitatory amino acid neurotransmitter) contents decreased significantly in the prosencephalon, and GABA (inhibitory amino acid neurotransmitter) contents decreased significantly in the brain stem. The molar ratio of glutamic acid to glutamine significantly diminished in the brain stem. Glutamic acid decarboxylase activity in the synaptosomes and the binding of [3H]glutamic acid and [3H]GABA to synaptosomal membrane preparations were unchanged in acute hepatic failure rats. These results indicate than an insufficiency of both excitatory and inhibitory neurotransmitter amino acids is induced by high ammonia contents in the synaptosomes of the brain stem during acute hepatic failure.     

Postnatal development of amino acid metabolism enzymes in the liver and muscle of 'cafeteria' rats

The effect of feeding a high-energy highly palatable cafeteria diet on the liver and muscle ontogenesis of serine dehydratase, alanine transaminase, glutamine synthetase and adenylate deaminase during postnatal development of the rat has been studied. The results are in agreement with the lower amino acid utilization in cafeteria rats, both adults and during postnatal development. The feeding of excess energy coupled with high-quality protein resulted in changes in the ontogenesis of the studied enzymes that coincide with the development of protein synthesis and overall pup growth even before they had direct access to this rich diet, suggesting that cafeteria feeding already affects the amino acid metabolism of the pup through the dam's milk.     

N-terminal sequence of the rat liver beta-subunit in the mitochondrial ATPase-ATPsynthase

The N-terminal amino acid residues of the beta-subunit in the rat liver mitochondrial ATPase - ATPsynthase have been identified by direct microsequencing after electrophoresis of either purified F1 or F0F1. The mature rat liver beta-subunit begins by two alanine residues that precede the glutamine recently proposed as the first amino acid of the sequence (Boulet, D., Poirier, J. and C?té, C., 1989, Biochem. Biophys. Res. Commun. 159, 1184-1190). This result indicates that the proteolytic cleavage of the beta-subunit precursor may occur at the level of this first alanine. This may be important in the understanding of proteolytic processing events which lead to the assembly of the ATPase-ATPsynthase subunits during mitochondrial biogenesis.     

Expression of four glutamine synthetase genes in the early stages of development of rainbow trout (Oncorhynchus mykiss) in relationship to nitrogen excretion

The incorporation of ammonia into glutamine, catalyzed by glutamine synthetase, is thought to be important in the detoxification of ammonia in animals. During early fish development, ammonia is continuously formed as yolk proteins and amino acids are catabolized. We followed the changes in ammonia and urea-nitrogen content, ammonia and urea-nitrogen excretion, glutamine synthetase activity, and mRNA expression of four genes coding for glutamine synthetase (Onmy-GS01-GS04) over 3-80 days post fertilization and in adult liver and skeletal muscle of the rainbow trout (Oncorhynchus mykiss). Both ammonia and urea-nitrogen accumulate before hatching, although the rate of ammonia excretion is considerably higher relative to urea-nitrogen excretion. All four genes were expressed during early development, but only Onmy-GS01 and -GS02 were expressed at appreciable levels in adult liver, and expression was very low in muscle tissue. The high level of expression of Onmy-GS01 and -GS03 prior to hatching corresponded to a linear increase in glutamine synthetase activity. We propose that the induction of glutamine synthetase genes early in development and the subsequent formation of the active protein are preparatory for the increased capacity of the embryo to convert the toxic nitrogen end product, ammonia, into glutamine, which may then be utilized in the ornithine-urea cycle or other pathways.