Neurochemistry International 55 (2009) 76–84

A novel glutaminase isoform in mammalian tissues

Vanessa de la Rosa, Jose A. Campos-Sandoval 1, Mercedes Martın-Rufian, Carolina Cardona,Jose M. Mates, Juan A. Segura, Francisco J. Alonso, Javier Marquez *

Departamento de Biologıa Molecular y Bioquımica, Laboratorio de Quımica de Proteınas, Facultad de Ciencias, Universidad de Malaga, 29071 Malaga, Spain

A R T I C L E I N F O

Article history:

Received 1 December 2008

Received in revised form 24 February 2009

Accepted 27 February 2009

Available online 9 March 2009

Keywords:

Glutamine

Glutamate

Neurone

Mammalian isoenzymes

Glutaminase genes

Transcript variants

A B S T R A C T

The synthesis of neurotransmitter glutamate in brain is mainly carried out by glutaminase enzymes. This

synthesis must be exquisitely regulated because of its harmful potential giving rise to excitotoxic

damage. It is noteworthy that two glutaminase isozymes coded by different genes are expressed in the

brain of mammals. The need for two genes and two isozymes to support the single process of glutamate

synthesis is unexplained, and identifying the role of each glutaminase is an important factor in

understanding glutamate-mediated neurotransmission. Multiple transcripts for glutaminase genes and

simultaneous expression of glutaminase isoforms have been reported in mammalian tissues and cells.

The recent discovery of protein interacting partners widens the possibilities of regulatory mechanisms

controlling these biosynthetic enzymes. The expression of distinct isozymes and binding partners may

represent the biochemical and molecular basis to achieve fine-tuning control of glutamate synthesis in

different cell types or developmental states. In this review, we will briefly summarize recent works on

glutaminase proteins in mammals, with particular emphasis on brain studies. We present convergent

evidence supporting the existence of a novel glutaminase isozyme in mammalian tissues.

� 2009 Elsevier Ltd. All rights reserved.

Contents lists available at ScienceDirect

Neurochemistry International

journa l homepage: www.e lsev ier .com/ locate /neuint

1. Introduction

The enzymes which catalyze cleavage of the side chain amidegroup in glutamine (Gln) leaving a g-glutamyl acyl chain andammonium ions and its subsequent transfer to a specific substrateare called glutamine amidotransferases (EC 2.4.2). In contrast,phosphate-activated glutaminase (EC 3.5.1.2; GA) is a Glnamidohydrolase, a true hydrolytic enzyme because the acyl andammonium acceptor is water and not other molecule. Krebs in hispioneer work on amino acids metabolism discovered GA whilestudying the enzymic hydrolysis of Gln in animal tissues (Krebs,1935). Later on, Errera and Greenstein (1949) first used the termphosphate-activated GA (originally named glutaminase I) todescribe the enzymatic activity in kidney, liver, brain and spleenextracts that was greatly increased by added phosphate. Inmammals, the enzyme needs inorganic phosphate for in vitro

activity but it is yet unknown whether phosphate is the trueactivator in vivo.

* Corresponding author. Tel.: +34 952 132024; fax: +34 952 132041.

E-mail address: marquez@uma.es (J. Marquez).1 Present address: Marie Curie Laboratory for Membrane Proteins, Biology

Department, National University of Ireland, Maynooth, Co. Kildare, Ireland.

Abbreviations: BPTES, bis-2-(5-phenylacetamido-1,2,4-thiadiazol-2-yl)ethyl sul-

fide; C/EBP, CAAT-enhancer binding protein; CRE, cAMP-responsive element;

DON, 6-diazo-5-oxo-norleucine; HNF, hepatocyte nuclear factor; PDZ, PSD95/Dlg/

ZO1 domains; PMN, polymorphonuclear neutrophils; RREB, Ras-responsive

element binding protein; UTR, untranslated region.

0197-0186/$ – see front matter � 2009 Elsevier Ltd. All rights reserved.

doi:10.1016/j.neuint.2009.02.021

Some of the main physiological functions of GA include renalammoniagenesis, nitrogen supply for urea biosynthesis in the liver,synthesis of the excitatory neurotransmitter glutamate (Glu) in thebrain, and energy supply for the bioenergetics of many normal andtransformed cell types (Kovacevic and McGivan, 1983; Curthoysand Watford, 1995). In brain, although several different precursorshave been proposed for the synthesis of transmitter Glu, Gln isconsidered the most important source through GA reaction(Kvamme, 1984; Nicklas et al., 1987). Besides being the majorexcitatory neurotransmitter in the CNS (Fonnum, 1984), Glu fulfilsmany other crucial roles in synaptogenesis, synaptic plasticity,pathogenesis of neuropsychiatric diseases (Conti and Weinberg,1999 and references therein), synthesis of g-aminobutyric acid(GABA), and brain energy metabolism (Erecinska and Silver, 1990).

The role of two key genes of Gln metabolism, glutaminesynthetase (GS, EC 6.3.1.2) and GA, has also focused considerablyattention in tumor biology, because Gln behaves as a centralmetabolite for growth and proliferation (Mates et al., 2002). Thecatabolism of Gln has been linked to neoplastic transformation(Kovacevic and McGivan, 1983). The high rate of glutaminolysisobserved in a wide variety of tumors would be essential tomaintain their proliferative capacity (Souba, 1993). Thus, GAoverexpression seems to be a hallmark exhibited by many tumors(Aledo et al., 1994). Studies on experimental and human tumorslooking at changes in enzymatic activity and relative mRNA levelsof both GS and GA revealed a similar pattern repeated in manycases: a knock-down or repression of GS expression along with an

V. de la Rosa et al. / Neurochemistry International 55 (2009) 76–84 77

overexpression of GA (Matsuno and Goto, 1992; Medina et al.,1992; Gebhardt and Williams, 1995).

In conclusion, it can be stated that glutaminase is expressed inmost mammalian tissues and cancer cells, but the regulation of itsorgan- and tumor-specific expression is largely unknown. There-fore, an essential step towards studying the regulation of GA inmammals is the characterization of its tissue-specific isozymepattern of expression, to which particular attention is given in thisshort review.

2. Mammalian glutaminase genes and transcripts

In humans, the GA family consists of two main members whichare encoded by separate genes in different chromosomes: the Gls

gene, located in chromosome 2, encodes isoforms known as kidney

Fig. 1. Schematic diagram showing the mRNA transcripts derived from Gls and Gls2 mam

boxes. The introns are shown as lines. The 50- and 30-UTR regions are separated from thei

the transcripts. The names of identified transcripts are also indicated.

(K-type) glutaminases, and the Gls2 gene, located on chromosome12, codes for liver (L-type) isozymes (Aledo et al., 2000).Orthologous genes have been described in other mammalianspecies, like mouse and rat, for Gls (Mock et al., 1989) and Gls2(Chung-Bok et al., 1997).

2.1. Gls gene and transcripts

The human Gls gene spans 82 kb. By comparison with availablehuman cDNAs, the gene was split into 19 exons (Porter et al., 2002).At least two different transcripts arise from this gene: the KGAmRNA formed by joining exons 1–14 and 16–19, and thealternative spliced transcript named glutaminase C (GAC) mRNAwhich uses only the first 15 exons, omitting exons 16–19 (Elgadiet al., 1999; Porter et al., 2002) (Fig. 1). The K-type cDNA named

malian glutaminase genes. Exons are indicated by numbers and depicted as empty

r respective exons by a vertical black line in the genes and shown as dotted boxes in

V. de la Rosa et al. / Neurochemistry International 55 (2009) 76–8478

GAC was originally isolated from a HT-29 human colon cDNAlibrary (Elgadi et al., 1999) but it has been reported to be presentalso in rat kidney and porcine kidney cell line (Porter et al., 2002).In human tissues, GAC mRNA is expressed predominantly incardiac muscle and pancreas, appreciably in placenta, kidney, andlung but not in liver and brain (Elgadi et al., 1999). Thus, multiplemammalian species express two distinct classes of K-type mRNAsthat differ in their 30-terminal coding region.

KGA mRNA was found to be ubiquitous in most nonhepatichuman tissues (Aledo et al., 2000). K-type cDNAs have been clonedfrom different mammalian tissues including kidney (Shapiro et al.,1991), brain (Nagase et al., 1998) and colon cancer cells (Elgadiet al., 1999). Differential expression of multiple K-type mRNAshave been detected in LLC-PK1-FBPase+ porcine proximal tubule-like cells (Porter et al., 1995), rat lymphocytes (Sarantos et al.,1993), mouse splenocytes (Aledo et al., 1998), Ehrlich ascitestumor cells (Aledo et al., 1994) and human kidney (Elgadi et al.,1999). The length of these transcripts is in the range of 3.0–6.0 kband can be produced by use of alternative polyadenylation sites(Porter et al., 2002) or by alternative splicing (Elgadi et al., 1999).The differential pattern of expression of these transcripts seems tobe tissue-specific and has been ascribed to alterations indevelopmental, proliferative and metabolic states of cells, as wellas to specific changes induced by tumors in the host tissues (Porteret al., 2002; Aledo et al., 1998).

In kidney, the KGA enzyme is strongly induced by metabolicacidosis: a post-transcriptional mechanism accounts for itsenhanced expression due to increased stability of the KGA mRNA(Laterza et al., 1997). This increase in renal Gln catabolism is part ofan adaptative mechanism to counterbalance the systemic acidosis.The presence of a pH-response element (AU-rich sequences) in the30-untranslated region (30-UTR) of the KGA mRNA is responsible ofits selective stabilization through binding of the protein j-crystallin/NADPH-quinone reductase (Tang and Curthoys, 2001).However, in tissues other than kidney, expression of the Gls gene islikely to be regulated at the level of transcription. For example, insmall intestinal mucosal cells dexamethasone increased KGAmRNA and specific activity. The increase in message preceded theincrease in activity, consistent with de novo RNA synthesisfollowed by protein synthesis. Glucocorticoids may accelerateintestinal Gln utilization by increasing glutaminase expression, anadaptive response that could provide more energy for mucosalcells in stress states (Sarantos et al., 1992). The primary role ofglutaminase in the intestine is to regulate enterocyte metabolism.This is because Gln is the main respiratory fuel of enterocytes andGA catalyzes the rate-limiting step of Gln degradation. Glutamine-enriched parenteral nutrition increases KGA activity in intestinalcells by promoting the accumulation of KGA mRNA (Kong et al.,2000).

In proliferative cells and tumors transcriptional regulation ofthe Gls gene seems to be operative and associated with cell growthand proliferation. Glutamine is the principal fuel used bylymphocytes and a fast rise of KGA mRNA levels was observedafter mitogenic challenge with endotoxin (Sarantos et al., 1993).Again, the increase in message preceded the increase in activityconsistent with gene transcription prior to enzyme biosynthesis.Cytokines like interleukin-1, interleukin-6, tumor necrosis factor-a, and g-interferon decreased KGA activity, protein content andmRNA levels in cultured human fibroblast; nuclear run-on assaysshowed a decrease in the amount of glutaminase transcriptsconsistent with transcriptional regulation. The authors relate thesefindings to the effect of different pro-inflammatory cytokinesimpairing fibroblast proliferation (Sarantos et al., 1994). In Ehrlichascites tumor cells, a long-term regulation for tumor KGAexpression during tumor development was deduced, in such away that maximum activity and mRNA levels were found in the

exponential growth phase when compared with the stationaryphase of growth (Aledo et al., 1994). Although GA activity iselevated in transformed cells, maximum expression occurs whenthe cells are actively dividing during the exponential phase ofgrowth, coincident with their maximum nitrogen requirement.

With regard to mammalian brain regulation, it is noteworthythat the rat KGA as well as the human KGA and GAC cDNAs containvariable CAG trinucleotide repeats in their 50 ends (Shapiro et al.,1991; Elgadi et al., 1999). However, the CAG repeats are located inthe 50-UTR region of the human K-type mRNAs, while in the ratthey are located within the coding region. CAG repeat-lengthpolymorphisms are shown to be the cause of various neurode-generative diseases, including Huntington’s disease, spinocere-bellar ataxias, and myotonic dystrophy (Singer, 1998). Thetrinucleotide CAG repeats encode polyglutamine stretches thatmay give rise to protein aggregation and damage of neural tissue.The functional significance of these CAG repeats in GA, if any, isunknown. Whether the CAG repeat has an impact in either of thetwo GA genes affecting its expression or contributing to aneurological disorder remain to be determined.

2.2. Gls2 gene and transcripts

The human Gls2 gene, located on chromosome 12, has a lengthbigger than 18 kb and splits into 18 exons (Aledo et al., 2000;Perez-Gomez et al., 2003) (Fig. 1). Apart from the additional exonpresent in the human Gls gene, the main differences in the codingsequences of both genes are located at exons 1 and 18. Exon 1shares 62.5% similarity, but it codes for 129 amino acids in KGA andonly for 61 amino acids in human glutaminase L, accounting for the67 extra amino acids of KGA protein at the N-terminal end. Thesequences encoded by exon 1 contain the signals involved inmitochondrial targeting and translocation processes (Shapiro et al.,1991; Gomez-Fabre et al., 2000). Interestingly, exon 1 of humanglutaminase L contains an LXXLL signature motif, which mightexplain the nuclear localization recently demonstrated forglutaminase L in mammalian brain (Olalla et al., 2002). Likewise,exon 18, which codes for the C-terminal region of both proteins,shows the lowest sequence similarity (29.4%). This region of thehuman L-type GA protein has been demonstrated recently to beinvolved in the recognition of PDZ (PSD95/Dlg/ZO1 domains)-interaction modules (Olalla et al., 2001; see also paper by Marquezet al. in this issue). Therefore, the most significant differencesbetween human Gls and Gls2 exons are located in regions involvedwith organelle targeting and protein–protein interactions, whichmay help to explain their differential function and regulation.Conversely, exons 3–17 of both mRNA transcripts have the samelength and show a high sequence similarity.

L-type transcripts derived from the Gls2 gene were originallythought to be present in adult liver tissue and absent inextrahepatic tissues (Smith and Watford, 1990; Curthoys andWatford, 1995). This restricted pattern of expression was generallyaccepted until quite recently, when results from our laboratoryshowed clear evidence for a wider presence in mammalian tissues:expression also occurs in extrahepatic tissues like brain, pancreasand breast cancer cells (Gomez-Fabre et al., 2000). L-type cDNAshave been cloned from rat liver (Smith and Watford, 1990) andhuman breast cancer cells (Gomez-Fabre et al., 2000).

Glutamine is a major substrate for hepatic gluconeogenesis andurea synthesis (Watford, 1993). Increased expression of hepatic GAhas been reported in diabetes, starvation, or on feeding a high-protein diet. An enhanced rate of gene transcription was shown tobe the mechanism responsible for these adaptive changes, at leastin the case of high-protein diets and starvation (Watford et al.,1994). This up-regulation of GA expression and Gln catabolism hasbeen related to increased gluconeogenesis from Gln and to a

V. de la Rosa et al. / Neurochemistry International 55 (2009) 76–84 79

heightened utilization of Gln into the urea cycle for removal ofammonia excess (Nurjhan et al., 1995; Watford, 1993). In contrastto KGA, hepatic GA is not affected by changes in acid–base status.Analysis of the rat hepatic GA promoter revealed a cAMP-responsive element (CRE) that may explain the gene’s respon-siveness to low insulin and/or high glucagon levels (Chung-Boket al., 1997). The proximal promoter of the rat Gls2 gene lacks afunctional TATA box, but contains recognition elements HNF-1,HNF-5 and CAAT-enhancer binding protein (C/EBP) that may beimportant for its basal expression in liver (Chung-Bok et al., 1997).

The first complete sequence of a human L-type glutaminase wasdeduced from breast cancer cDNA cloned in our laboratory(Gomez-Fabre et al., 2000). The human full-length L-type GAcDNA isolated from ZR-75 breast cancer cells is 2408-nt long with a1806-base open reading frame. Translation of the cDNA results in apredicted protein of 602 residues with a molecular mass of66,309 Da. Sequence comparisons showed that it was highlysimilar (89% identity) to the rat liver GA cDNA. With regard to thededuced amino acid sequence, human L-type GA shares aconsiderable degree of identity (94%) with the rat liver enzymebut the human enzyme extends over 67 residues at the N-terminalend. In contrast, the human L-type GA showed only 68.5% identitywith the rat kidney KGA cDNA, a percentage similar to that foundbetween rat liver and kidney GA cDNA species (Chung-Bok et al.,1997).

As mentioned before, the human Gls2 gene was determined tospan more than 18.1 kb and consists of 18 exons and 17 introns.DNA coding sequences perfectly matched the cDNA sequencecloned from breast cancer cells except at six nucleotide positionsthat were lately identified as Taq polymerase errors. The exon/intron organization of the human Gls2 gene was determined usingthe sequence of the ZR-75 breast cancer cDNA clone. Accordingly,the exon/intron boundaries for the rat Gls2 gene may be predictedby using genomic clones in the rat genome database and the cDNApreviously cloned for the rat liver LGA transcript. Nucleotide andamino acid sequences of human and rat L-type transcripts share aremarkably high level of identity (89% and 94%, respectively). Thepositions interrupted in the human and rat Gls2 genes differslightly, thus the exon/intron organization of both genes closelyresemble each other (Fig. 2) (Perez-Gomez et al., 2003). Never-theless, it is noteworthy that rat LGA is 67 amino acids shorter thanZR-75 L-type GA at the N-terminus; the rat LGA protein lacks all theamino acids encoded by exon 1 and the first six amino acids of exon2 (Fig. 1). Therefore the existence of alternative spliced forms of theGls2 gene seems very likely, as happens with the Gls gene. Basedupon sequence differences and additional experimental evidencesthat will be discussed later, we conclude that human L-type GAfrom ZR-75 breast cancer cells represents a novel mRNA speciesencoding a further GA isoform termed GAB.

The core promoter regions of both the human and rat Gls2 genesdo not contain canonical TATA boxes, but do have G + C-richdomains (Perez-Gomez et al., 2003; Chung-Bok et al., 1997). Thischaracteristic is also found in the human KGA promoter (Porteret al., 2002). Notwithstanding the high sequence identity (89%)between the coding regions of human and rat Gls2 genes (Gomez-Fabre et al., 2000), their proximal promoter regions did not showsignificant sequence homology. Unlike rat Gls2, human Gls2 has acanonical CAAT box and RREBs, though it lacks a HNF-5 site presentin rat Gls2. Mutagenesis and transient transfections clearlydemonstrated that two CAAT boxes play a crucial role in thetranscriptional regulation of the human Gls2 gene, both in cells ofliver origin (HepG2) and in MCF7 breast cancer cells (Perez-Gomezet al., 2003). CAAT-enhancer binding proteins (C/EBPs) control cellgrowth and differentiation, causing growth arrest and inducingcellular differentiation in several adipocyte, granulocyte andkeratinocyte lineages (Darlington et al., 1998). Furthermore,

HepG2 hepatoma cells express significantly lower levels of C/EBPa and C/EBPb than those found in normal terminallydifferentiated hepatocytes (Friedman et al., 1989). Thus conversionof hepatocytes into proliferating hepatoma cells might requirestrong downregulation of C/EBPa and C/EBPb expression.

The GAB enzyme from ZR-75 human breast cancer cells alsoshowed a long-term regulation depending on the cell proliferationstate: maximal activities were found at the beginning of theexponential growth phase, with a remarkable decrease at thestationary phase of growth when cell confluence was achieved.There was also a fivefold increase in the relative levels of GABmRNA at the beginning of the exponential growth phase incomparison with those shown when proliferation had alreadyceased (Gomez-Fabre et al., 2000). These results agree with thosepreviously reported for Ehrlich ascitic tumor cells and indicate along-term regulation of GA in tumors by differential geneexpression. Thus, GA overexpression seems to be a phenotypiccharacteristic exhibited by many experimental and human tumors.

3. Mammalian glutaminase enzymes

Krebs in his early studies on Gln catabolism in mammals,already distinguished between a liver-type and a brain-typeenzymes (Krebs, 1935), although this last isoform was latelynamed kidney-type enzyme (Kovacevic and McGivan, 1983).Besides differences in molecular structures, the distinct kineticbehavior has been a hallmark frequently used to distinguishbetween GA isoforms. These enzymes are intricately regulated by anumber of low-molecular mass effectors: the main kineticdifferences have been observed in the dependence of the activatorinorganic phosphate (Pi) – low for L-type, high for K-type – therelative affinity for the substrate Gln – higher in K- than in L-types– and the inhibitory effect of Glu, a unique characteristic reportedonly for K-type isozymes (Kovacevic and McGivan, 1983; Curthoysand Watford, 1995). Both isoforms have been traditionallyconsidered as mitochondrial enzymes (Curthoys and Watford,1995).

3.1. Molecular structures and kinetics properties

KGA protein has been purified from pig kidney (Kvamme et al.,1970) and brain (Svenneby et al., 1973; Nimmo and Tipton, 1980),from rat kidney (Curthoys et al., 1976a) and brain (Haser et al.,1985; Kaneko et al., 1987), from cow brain (Chiu and Boeker, 1979)and from Ehrlich ascitic tumor cells (Quesada et al., 1988; Seguraet al., 1995). These enzymes are not activated by ammonia butinhibited by Glu, have a relatively low K0.5 for Gln, a high K0.5 for Pi,and their antibodies showed crossed reactions, but not with theLGA isoenzyme (Curthoys et al., 1976b). Purified native KGA in theabsence of polyvalent anions is an inactive protomer with anapparent relative molecular mass (Mr) ranging from 90,000 to137,000 (see purification references cited above). This so-calledTris-form of the enzyme dimerizes by addition of phosphate (P-form) and undergoes a reversible polymerization into a highmolecular mass aggregate in phosphate/borate buffer (PB-form)correlated with enzyme activation (Svenneby et al., 1973;Curthoys et al., 1976a).

Rat renal and brain KGA contain two different peptides of Mr66,000 and 68,000 that are present in the ratio 3:1 and areproduced in vivo from a common precursor (Haser et al., 1985;Perera et al., 1990). In rat kidney, the two mature KGA subunitsforming the tetramer are produced by mitochondrial processing ofa common 74-kDa precursor (Srinivasan et al., 1995). In contrast,KGA from pig brain has been reported to contain a singlepolypeptide of Mr 64,000 (Svenneby et al., 1973) or 73,000(Nimmo and Tipton, 1980). These discrepancies can be ascribed to

Fig. 2. Comparison of the deduced amino acid sequences and of the exon/intron organization of the human Gls2 gene with those of rat Gls2 (GenBank # J05499, NM_138904; Rat

Genome Database, Medical College of Wisconsin, Genome assembly 3.1, Chromosome 7, Position 1484396–1495283) and human Gls (Porter et al., 2002) genes. Arrowheads

indicate the positions at which introns interrupt exons (adapted from Perez-Gomez et al., 2005. Biochem. J. 386, 535–542). Abbreviations: hGAB, novel human L-type glutaminase

isozyme (long transcript of human Gls2 gene); rLGA, classical rat L-type glutaminase isozyme (short transcript of rat Gls2 gene); hKGA, human K-type glutaminase.

V. de la Rosa et al. / Neurochemistry International 55 (2009) 76–8480

V. de la Rosa et al. / Neurochemistry International 55 (2009) 76–84 81

species-specific differences or, alternatively, they could beexplained by proteolytic degradation and the use of differentmolecular mass standards for calibration of the SDS gels (Quesadaet al., 1988; Segura et al., 1995).

On the other hand, the enzyme LGA from rat liver was partiallypurified by Patel and McGivan (1984), and purification tohomogeneity was reported by Heini et al. (1987) and Smith andWatford (1988). In contrast to KGA, the liver LGA enzyme is notinhibited by Glu, has a higher Km for Gln, is fully activated at lowerconcentrations of Pi and shows activation by ammonia. Rat liverLGA possesses a unique subunit with a Mr of 58,000, but a greatvariability has been reported for the molecular mass of the nativeprotein, with values ranging from 162,000 to 170,500 Da,estimated by sucrose gradient ultracentrifugation (Heini et al.,1987; Smith and Watford, 1988), to more than 300,000 Da byestimations using gel filtration chromatography (Patel andMcGivan, 1984; Smith and Watford, 1988). This considerablediscrepancy was also observed in the estimates of native Mr for ratkidney KGA from gel filtration data and from sucrose gradientanalysis (Curthoys et al., 1976a). An abnormal partial specificvolume, aggregation of the enzyme and protein inactivation as aresult of dilution have been argued to explain this poor correlation(Heini et al., 1987; Smith and Watford, 1988; Segura et al., 1995).The molecular data obtained from sucrose gradient analysissuggest that rat liver glutaminase is a trimer composed of threeidentical subunits (Heini et al., 1987). In contrast to KGA, no signsfor a similar aggregation of the liver LGA enzyme were found evenat high Pi concentrations (Heini et al., 1987; Smith and Watford,1988).

The human GAB mRNA has been heterologously expressed inprokaryotic (Campos et al., 2003) and baculovirus (Campos-Sandoval et al., 2007) systems. The first heterologous overproducersystem for a mammalian GA was described for human GABisoenzyme in Escherichia coli. By employing several geneticconstructs in different vectors, the expression of human GAB wasoptimized: full-length GAB accounted for about 25% of total E. coli

protein but was devoid of catalytic activity (Campos et al., 2003).Then, we studied expression in a eukaryotic system: the human GABisoform expressed in baculovirus yielded functional recombinantenzyme in Sf9 insect cells (Campos-Sandoval et al., 2007). Insectsystems constitute a potent eukaryotic expression system with theadvantage that insect cells possess similar post-translationalmodifications to mammalian cells. This was the first directdemonstration that the cDNA of human GAB, originally clonedfrom ZR-75 breast cancer cells (Gomez-Fabre et al., 2000), encodesan active GA enzyme. The molecular and kinetic properties of purerecombinant GAB will be detailed below, as further experimentalevidence for the existence of a new L-type GA isoform.

A regulatory function was early indicated for GA based on itsactivation or inhibition by a wide variety of endogenous andexogenous effectors. In particulate preparations from brain(homogenates and synaptosomes) stimulatory effectors include:Pi, sulfate, chloride, carboxylic acids, nucleotide triphosphates andriboflavin phosphate, all active in mM concentrations (Erecinskaand Silver, 1990). Other compounds like CoA, short- and long-chainacyl-CoA, and the dye bromothymol blue activates GA at mMconcentrations but become inhibitors in mM concentrations(Kvamme et al., 2000). Interpretations of these data should bedone with caution, due to difficulty in obtaining reproducibleresults when using tissue preparations or crude cell extracts and tomethodological differences. For example, calcium is a potentiallyimportant candidate for GA regulation because its concentrationschange with brain function. Ca2+ activates GA in mitochondria,brain synaptosomes, brain slices and homogenates, but does notact on purified enzyme, indicating that its effect is indirect(Erecinska and Silver, 1990; Kvamme et al., 2000).

Nowadays, it has been shown that at least two GA isoforms,with different kinetic and regulatory properties, are expressed inmammalian brain (Olalla et al., 2002). Therefore, the assays done inbrain and other tissues and cells where co-expression occursshould be carefully reevaluated to avoid mixed findings comingfrom measuring distinct GA isozymes at the same time (Perez-Gomez et al., 2005). We described the simultaneous expression ofL-type and K-type mRNA transcripts in human brain (Gomez-Fabreet al., 2000; Aledo et al., 2000). The regional distribution of both GAtranscripts in human brain showed a similar pattern of expression:they were ubiquitously expressed in all regions of the brainexamined, with the strongest signal in cerebral cortex. Further-more, expression of K- and L-type transcripts in brain was alsodemonstrated in other mammalian species like cow, mouse, rabbitand rat. The co-expression was verified at the protein level bybiochemical and immunological approaches employing isozyme-specific antibodies (Olalla et al., 2002). Simultaneous expression ofboth K- and L-type GA isozymes in the same cell type is morefrequent than previously thought: apart from neurones, it has beenfound in human colorectal tumor cells (Turner and McGivan,2003), human hepatoma HepG2 cells, medullar blood mono-nuclear cells from patients suffering from leukaemia, KU812Fhuman myeloid cells and human breast cancer cells MCF7 and ZR-75 (Perez-Gomez et al., 2005).

Phosphate is the most prominent stimulator of GA. Purifiedbrain KGA is an allosteric enzyme and its Pi activation curve wasfound to be sigmoid (Haser et al., 1985; Kvamme et al., 2000). Theabruptness of the change shown by the activation curve meansthat KGA is very sensitive to changes in the level of Pi. Maineffectors of GA activity are, in fact, compounds that alter the Piactivation (Kvamme et al., 2000). Whether Pi is the truephysiological stimulator of GA in brain remains to be determined,but considering its brain concentration and the fact that it can berapidly altered during neuronal activity, its candidature as animportant physiological regulator of brain GA in vivo has beenpostulated (Erecinska and Silver, 1990). Main inhibitors of brain GAinclude Glu, ammonium ions, protons, and cAMP and cGMP at mMconcentrations (Krebs, 1935; Erecinska and Silver, 1990; Kvammeet al., 2000). Other non-specific inhibitors are compounds thatreact with thiol groups as well as Gln analogs like 6-diazo-5-oxo-norleucine (DON) (Campos et al., 1998; Curthoys and Watford,1995). A recently discovered KGA-specific inhibitor is BPTES [bis-2-(5-phenylacetamido-1,2,4-thiadiazol-2-yl)ethyl sulfide] whichprevents the formation of large phosphate-induced oligomers(Robinson et al., 2007).

Glutamate is a competitive inhibitor and the relative con-centrations of Gln and Glu in glutamatergic terminals (Kvammeet al., 2000) suggest that GA can be strongly inhibited in nerve cellsas in their terminals, in agreement with early experiments donewith synaptosomes (Erecinska and Silver, 1990, and referencestherein). In fact, the in vivo GA activity in intact brain ofhyperammonemic rats measured by 15N nuclear magneticresonance is about 1% of the reported in vitro activity measuredin rat brain homogenates (Kanamori and Ross, 1995). Theseresults suggest that in intact brain GA activity is maintained at alow level by a short-term regulation of its key effectors, the mostlikely ones being Pi and Glu.

4. Evidence supporting GAB as a novel glutaminase isoform

GAB and LGA enzymes differ markedly in their molecular mass,kinetic characteristics, regulatory properties, tissue distributionand subcellular localization. We favour the use of the term GAB forthe long transcript of the Gls2 gene, originally isolated from breastcancer cells, leaving the acronym LGA for the classical liver isoformencoded by a short transcript of the same gene.

V. de la Rosa et al. / Neurochemistry International 55 (2009) 76–8482

4.1. Molecular structure

Human GAB was isolated as a cDNA clone from ZR-75 breastcancer cells with an ORF of 1806 nucleotides encoding a protein of602 amino acids. As mentioned before, this canonical longtranscript of the Gls2 gene is formed by joining the full 18 exonsof the human gene. Human GAB is 67 amino acids longer than ratliver LGA protein (Fig. 2). Comparison of GAB and LGA transcriptsshows that LGA lacks all the amino acids encoded by exon 1 and thefirst six amino acids of exon 2 of the GAB transcript. Evidence of theexistence of other long transcript (GAB) variants has been obtainedin mouse tissues. We have recently cloned by RT-PCR a Gls2transcript variant of 2398 bp from mouse brain poly(A+)mRNA(GenBank # EU770627) showing a high identity (83%) with humanGAB cDNA. In addition, data obtained from immunoblot analysis ofmouse tissues using anti-GAB-specific antibodies strongly supportthe existence of transcript variants encoding L-type proteins ofdifferent lengths: L-type isozyme was revealed in mouse brain asbands of higher molecular mass than those seen in mouse liver thatwere of the same range of size (approx. 58 kDa) as purified rat liverLGA (Campos et al., 2003). Finally, the existence in the database ofother expressed sequence tags and mRNAs isolated from rat liverwhich are longer than the original one (e.g. GenBank # BC104712and # BC089776) strongly support the view of GAB and LGA beingalternative spliced forms of the Gls2 gene.

4.2. Kinetic properties

The existence of a novel L-type GA isoform was also suggestedfrom kinetic data in different cell models. We found striking kineticdifferences between the L-type isoforms GAB and LGA. First,subcellular fractionation and Western blot analysis with anti-GABantibodies revealed that rat brain L-type GA was enriched in nuclei(see next section) where it was catalytically active (Olalla et al.,2002). Nuclear GA exhibited a kinetic behavior that resembles thatof the LGA enzyme with regard to the low phosphate concentrationrequirement; however, nuclear GA showed a strong and unex-pected inhibition by Glu, a property that is absent in the LGAenzyme. On the other hand, nuclear GA activity cannot be assignedto KGA, which was absent from the nuclear preparations (Olallaet al., 2002).

In human colorectal cancer cells meaningful results have beenrecently reported supporting GAB as a novel isoform. The slow-growing adenoma-derived cell line AA/C1 and the rapidlyproliferating carcinoma cell line HT29 co-expressed GAC andGAB mRNAs, whereas Western blotting showed the presence, in

Fig. 3. Kinetic properties of purified recombinant human GAB. Left panel: dependence on

wide range of Gln concentrations at two phosphate concentrations: 5 mM (&) and 150 m

correspond to the Vmax calculated for each curve. A clear sigmoidal dependence was found

index of 2.7 was deduced. Right panel: glutamate inhibition. The GA activity of pure GAB

and 60 mM (&). In both cases, the phosphate concentration was 5 mM (adapted from

both cell lines, of GA isoforms with an molecular mass of 63 kDa,intermediate between that of KGA and LGA (Turner and McGivan,2003). In both cell lines GA activity was inhibited by Glu, and thepresence of ammonia did not affect GA activity in either cell type,under conditions where activation of LGA was observed. Theauthors concluded that there was no evidence for the presence ofan enzyme with L-type GA kinetic properties in colorectal tumorcells, despite the abundant GAB mRNA found in these cells.

It is always complicated to withdrawn clear-cut conclusionsfrom whole-cell activity assays because the kinetic results may besomehow obscured by the presence of contaminants or interferingsubstances in such crude extracts. In order to obtain a clearerunderstanding of the differences between L-type isoforms, it isnecessary to study them in isolation. We recently described theoverexpression and purification of GAB to apparent homogeneityand the determination of some of its molecular and kineticproperties (Campos-Sandoval et al., 2007). Purified human GAB isan allosteric enzyme: measurements of the Gln binding kineticsyielded a sigmoidal curve with a Hill index of 2.7 and S0.5 values of32 and 64 mM for high and low Pi concentrations, respectively(Fig. 3). Whereas the protein showed a low Pi dependence typicalfor L-type GA, the enzyme was unexpectedly inhibited by Glu, akinetic characteristic exclusive of K-type isozymes. At low Piconcentration (5 mM) and suboptimal Gln concentration (20 mM)the IC50 for Glu was 50 mM (Fig. 3). This data is in the range of Kivalues reported for Glu competitive inhibition of K-type enzymesat high Pi concentrations.

Finally, the dependence of purified GAB activity on ammoniawas studied: only a slight activation effect was seen (20% increaseover control values without ammonia added) (Campos-Sandovalet al., 2007). This is another distinctive feature of human GABcompared with classical liver LGA isoforms, because hepatic LGAhas an absolute requirement for activation by ammonia (Patel andMcGivan, 1984). Therefore, the characterization of pure humanGAB indicated that it truly represents a novel kind of GA isoformwith distinctive kinetic properties. In contrast to GAB, the putativeK-type GAC isoform has not been heterologously expressed orpurified from native sources. The only evidence suggesting thatGAC splice transcript encodes a functional GA protein came fromthe high GA activity measured in TSE breast cancer cells thatexpress only the human GAC mRNA (Elgadi et al., 1999).

4.3. Subcellular locations

A crucial difference between GAB and the rest of GA isoforms sofar described is that GAB has been found in extramitochondrial

glutamine concentrations. The GA activity of pure GAB protein was assayed over a

M (*). The values are presented as percentage of enzymatic activity, where the 100%

. (Inset) A representative Hill plot of the same data for high and low phosphate. A Hill

protein was assayed in the presence of Glu at two Gln concentrations: 20 mM (*)

Campos-Sandoval et al., 2007. Int. J. Biochem. Cell Biol. 39, 765–773).

V. de la Rosa et al. / Neurochemistry International 55 (2009) 76–84 83

localizations, while KGA and LGA have been always exclusivelyconfined in mitochondria (Erecinska and Silver, 1990; Curthoysand Watford, 1995). The first report of an extramitochondriallocalization for a mammalian GA enzyme came from immunocy-tochemistry studies employing anti-GAB antibodies in rat andmonkey brain (Olalla et al., 2002). In cerebral cortex, hippocampus,cerebellum and striatum the immunolocalization revealed an L-type GA immunostaining concentrated in neuronal nuclei. Thenuclear role of this L-type GA has yet to be determined, but thisnovel location has been linked to a potential function as atranscriptional coregulator (Marquez et al., 2006; Szeliga et al.,2008).

It is noteworthy that recombinant functional GAB hasconfirmed its capacity of targeting to the nucleus in Sf9 insectcells. Subcellular fractionation demonstrated that recombinanthuman GAB was targeted to both mitochondria and nucleus, and inboth locations the protein was catalytically active (Campos-Sandoval et al., 2007). This nuclear GAB cannot be ascribed tomitochondrial contamination of nuclear fraction because neithercytochrome c oxidase activity nor Hsp60 protein – used asmitochondrial markers – were detected on isolated nuclearfraction. These data constitute the first demonstration of GABbeing located in two different compartments in a single cell type. Inaddition, the nuclear localization supports previous immunocy-tochemistry studies in mammalian brain which revealed segrega-tion of an L-type enzyme in the nucleus of neurones (Olalla et al.,2002).

Whereas GAB seems to have structural determinants needed formitochondrial targeting (Gomez-Fabre et al., 2000), it does notpossess a discernible classical nuclear localization signal. However,human GAB has other sequence motifs and conserved modulesthat may be essential for its nuclear import (Marquez et al., 2006).For example, a PDZ-protein recognition motif would be implicatedin GAB specific targeting to selective cellular locations (Olalla et al.,2008; see also paper by Marquez et al. in this issue). A secondextramitochondrial localization of GAB, apart from cell nuclei, hasbeen found in human polymorphonuclear neutrophils (PMNs).Using GAB-specific antibodies in combination with flow cytometryand confocal microscopy, L-type GA was shown to be present onthe surface of human PMN and was released into cell culturemedium upon stimulation (Castell et al., 2004). The presence of L-type GA in these leukocytes was related with their bactericidalaction through a Gln-dependent mechanism of superoxideproduction. These results point towards a cell-specific subcellularlocation of GAB increasing the number of potential roles thisprotein may fulfil.

5. Concluding remarks

There are key issues about the functional operation of theglutamatergic transmission that remain unresolved. An importantquestion, for example, is which factors regulate in vivo GA activity,and hence Glu supply, during neuronal stimulation. Anotherincomplete issue is the full description and comprehension of brainGlu metabolism, whose degree of complexity is further increasedby its exquisite and complex degree of anatomic compartmenta-tion. The co-expression of different GA isozymes in brain displaysregional, cellular and subcellular variations that may have aprofound impact on those key issues. Even more, multipletranscripts from each of the two GA genes can be selectivelyexpressed which further complicates the pattern of regulation ofGA expression in mammalian tissues.

The molecular portrait of GA isoform expression in mammaliantissues involves multiple transcripts even in a single cell type. Theabundance of a particular GA mRNA species may significantlychange depending upon the tissue type or the developmental or

metabolic state of the tissue. The physiological meaning of theexistence of different species of mRNA coding for GA is notunderstood. However, it is tempting to speculate that eachtranscript may represent a specific target for different stimuli,the overall GA expression being the balance between these stimuli.We have shown robust experimental evidences for a novel GAenzyme cloned from ZR-75 breast cancer cells. The data imply thatthis GAB protein represents a new L-type GA isoform, differentfrom the classical liver isozyme, and being able to performadditional functions depending on cell type and status.

Acknowledgements

This work was supported by Excellence Grant CVI-1543 fromthe regional Andalusian government (Junta de Andalucia), GrantSAF2007-61953 from the Ministry of Education and Science ofSpain, and Grant RD06/1012 of the RTA RETICS network from theSpanish Health Institute Carlos III.

References

Aledo, J.C., Segura, J.A., Barbero, L.G., Marquez, J., 1998. Early differential expressionof two glutaminase mRNAs in mouse spleen after tumor implantation. CancerLett. 133, 95–99.

Aledo, J.C., Segura, J.A., Medina, M.A., Alonso, F.J., Nunez de Castro, I., Marquez, J.,1994. Phosphate-activated glutaminase expression during tumor development.FEBS Lett. 341, 39–42.

Aledo, J.C., Gomez-Fabre, P.M., Olalla, L., Marquez, J., 2000. Identification of twohuman glutaminase loci and tissue-specific expression of the two related genes.Mamm. Genome 11, 1107–1110.

Campos, J.A., Aledo, J.C., del Castillo-Olivares, A., del Valle, A.E., Nunez de Castro, I.,Marquez, J., 1998. Involvement of essential cysteine and histidine residues inthe activity of isolated glutaminase from tumour cells. Biochim. Biophys. Acta1429, 275–283.

Campos, J.A., Aledo, J.C., Segura, J.A., Alonso, F.J., Gomez-Fabre, P.M., Nunez deCastro, I., Marquez, J., 2003. Expression of recombinant human L-glutaminase inEscherichia coli: polyclonal antibodies production and immunological analysisof mouse tissues. Biochim. Biophys. Acta 1648, 17–23.

Campos-Sandoval, J.A., Lopez de la Oliva, A.R., Lobo, C., Segura, J.A., Mates, J.M.,Alonso, F.J., Marquez, J., 2007. Expression of functional human glutaminase inbaculovirus system: affinity purification, kinetic and molecular characteriza-tion. Int. J. Biochem. Cell Biol. 39, 765–773.

Castell, L., Vance, C., Abbott, R., Marquez, J., Eggleton, P., 2004. Granule localizationof glutaminase in human neutrophils and the consequence of glutamineutilization for neutrophil activity. J. Biol. Chem. 279, 13305–13310.

Chiu, J.F., Boeker, E.A., 1979. Cow brain glutaminase: partial purification andmechanism of action. Arch. Biochem. Biophys. 196, 493–500.

Chung-Bok, M.-I., Vincent, N., Jhala, U., Watford, M., 1997. Rat hepatic glutaminase:identification of the full coding sequence and characterization of a functionalpromoter. Biochem. J. 324, 193–200.

Conti, F., Weinberg, R.J., 1999. Shaping excitation at glutamatergic synapses. TrendsNeurosci. 22, 451–458.

Curthoys, N.P., Kuhlenschmidt, T., Godfrey, S.S., 1976a. Regulation of renal ammo-niagenesis. Purification and characterization of phosphate-dependent glutami-nase from rat kidney. Arch. Biochem. Biophys. 174, 82–89.

Curthoys, N.P., Kuhlenschmidt, T., Godfrey, S.S., Weiss, R.F., 1976b. Phosphate-dependent glutaminase from rat kidney. Cause of increased activity in responseto acidosis and identity with glutaminase from other tissues. Arch. Biochem.Biophys. 172, 162–167.

Curthoys, N.P., Watford, M., 1995. Regulation of glutaminase activity and glutaminemetabolism. Annu. Rev. Nutr. 15, 133–159.

Darlington, G.J., Ross, S.E., MacDougald, O.A., 1998. The role of C/EBP genes inadipocyte differentiation. J. Biol. Chem. 273, 30057–30060.

Elgadi, K.M., Meguid, R.A., Qian, M., Souba, W.W., Abcouwer, S.F., 1999. Cloning andanalysis of unique human glutaminase isoforms generated by tissue-specificalternative splicing. Physiol. Genomics 1, 51–62.

Erecinska, M., Silver, I.A., 1990. Metabolism and role of glutamate in mammalianbrain. Prog. Neurobiol. 35, 245–296.

Errera, M., Greenstein, J.P., 1949. Phosphate-activated glutaminase in kidney andother tissues. J. Biol. Chem. 178, 495–502.

Fonnum, F., 1984. Glutamate: a neurotransmitter in mammalian brain. J. Neuro-chem. 42, 1–11.

Friedman, A.D., Landschulz, W.H., McKnight, S.L., 1989. CCAAT/enhancer bindingprotein activates the promoter of the serum albumin gene in cultured hepatomacells. Genes Dev. 3, 1314–1322.

Gebhardt, R., Williams, G.M., 1995. Glutamine synthetase and hepatocarcinogen-esis. Carcinogenesis 16, 1673–1681.

Gomez-Fabre, P.M., Aledo, J.C., del Castillo-Olivares, A., Alonso, F.J., Nunez de Castro,I., Campos, J.A., Marquez, J., 2000. Molecular cloning, sequencing and expressionstudies of the human breast cancer cell glutaminase. Biochem. J. 345, 365–375.

V. de la Rosa et al. / Neurochemistry International 55 (2009) 76–8484

Haser, W.G., Shapiro, R.A., Curthoys, N.P., 1985. Comparison of the phosphate-dependent glutaminase obtained from rat brain and kidney. Biochem. J. 229,399–408.

Heini, H.G., Gebhardt, R., Mecke, D., 1987. Purification and characterization of ratliver glutaminase. Eur. J. Biochem. 162, 541–546.

Kanamori, K., Ross, B.D., 1995. In vivo activity of glutaminase in the brain ofhyperammonaemic rats measured by 15N nuclear magnetic resonance. Bio-chem. J. 305, 329–336.

Kaneko, T., Urade, Y., Watanabe, Y., Mizuno, N., 1987. Production, characterization,and immunohistochemical application of monoclonal antibodies to glutami-nase purified from rat brain. J. Neurosci. 7, 302–309.

Kong, S.E., Hall, J.C., Cooper, D., McCauley, R.D., 2000. Glutamine-enriched parent-eral nutrition regulates the activity and expression of intestinal glutaminase.Biochim. Biophys. Acta 1, 67–75.

Kovacevic, Z., McGivan, J.D., 1983. Mitochondrial metabolism of glutamine andglutamate and its physiological significance. Physiol. Rev. 63, 547–605.

Krebs, H.A., 1935. Metabolism of amino acids. IV. The synthesis of glutamine fromglutamic acid and ammonia and the enzymic hydrolysis of glutamine in animaltissues. Biochem. J. 29, 1951–1969.

Kvamme, E., 1984. Enzymes of cerebral glutamine metabolism. In: Haussinger, D.,Sies, H. (Eds.), Glutamine Metabolism in Mammalian Tissues. Springer Verlag,Berlin, pp. 32–48.

Kvamme, E., Roberg, B., Torgner, I.Aa., 2000. Phosphate activated glutaminase andmitochondrial glutamine transport in the brain. Neurochem. Res. 25, 1407–1419.

Kvamme, E., Tveit, B., Svenneby, G., 1970. Glutaminase from pig renal cortex. I.Purification and general properties. J. Biol. Chem. 245, 1871–1877.

Laterza, O.F., Hansen, W.R., Taylor, L., Curthoys, N.P., 1997. Identification of an mRNA-binding protein and the specific elements that may mediate the pH-responsiveinduction of renal glutaminase mRNA. J. Biol. Chem. 272, 22481–22488.

Marquez, J., Lopez de la Oliva, A.R., Mates, J.M., Segura, J.A., Alonso, F.J., 2006.Glutaminase: a multifaceted protein not only involved in generating glutamate.Neurochem. Int. 48, 465–471.

Mates, J.M., Perez-Gomez, C., Nunez de Castro, I., Asenjo, M., Marquez, J., 2002.Glutamine and its relationship with intracellular redox status, oxidative stressand cell proliferation/death. Int. J. Biochem. Cell Biol. 34, 439–458.

Matsuno, T., Goto, I., 1992. Glutaminase and glutamine synthetase activities inhuman cirrhotic liver and hepatocellular carcinoma. Cancer Res. 52, 1192–1194.

Medina, M.A., Sanchez-Jimenez, F., Marquez, J., Quesada, A.R., Nunez de Castro, I.,1992. Relevance of glutamine metabolism to tumor cell growth. Mol. Cell.Biochem. 113, 1–15.

Mock, B., Kozak, C., Seldin, M.F., Ruff, N., D’Hoostelaere, L., Szpirer, C., Levan, G.,Seuanez, H., O’Brien, S., Banner, C., 1989. A glutaminase (gls) gene maps tomouse chromosome 1, rat chromosome 9, and human chromosome 2. Geno-mics 5, 291–297.

Nagase, T., Ishikawa, K., Suyama, M., Kikuno, R., Hirosawa, M., Miyajima, N., Tanaka,A., Kotani, H., Nombra, N., Ohara, O., 1998. Prediction of the coding sequences ofunidentified human genes. XII. The complete sequences of 100 new cDNAclones from brain which code for large proteins in vitro. DNA Res. 5, 355–364.

Nicklas, W.J., Zeevalk, G., Hyndman, A., 1987. Interactions between neurons and gliain glutamate/glutamine compartmentation. Biochem. Soc. Trans. 15, 208–210.

Nimmo, G.A., Tipton, K.F., 1980. Purification of soluble glutaminase from pig brain.Biochem. Pharmacol. 29, 359–367.

Nurjhan, N., Bucci, A., Perriello, G., Stumvoll, M., Dailey, G., Bier, D.M., Toft, I., Jenssen,T.G., Gerich, J.E., 1995. Glutamine: a major gluconeogenic precursor and vehiclefor interorgan carbon transport in man. J. Clin. Investig. 95, 272–277.

Olalla, L., Aledo, J.C., Bannenberg, G., Marquez, J., 2001. The C-terminus of humanglutaminase L mediates association with PDZ domain-containing proteins. FEBSLett. 488, 116–122.

Olalla, L., Gutierrez, A., Campos, J.A., Khan, Z.U., Alonso, F., Segura, J.A., Marquez, J.,Aledo, J.C., 2002. Nuclear localization of L-glutaminase in mammalian brain. J.Biol. Chem. 277, 38939–38944.

Olalla, L., Gutierrez, A., Jimenez, A.J., Lopez-Tellez, J.F., Khan, Z.U., Perez, J., Alonso,F.J., de la Rosa, V., Campos-Sandoval, J.A., Segura, J.A., Aledo, J.C., Marquez, J.,

2008. Expression of Scaffolding PDZ Protein GIP (Glutaminase-Interacting-Protein) in Mammalian Brain. J. Neurosci. Res. 86, 281–292.

Patel, M., McGivan, J.D., 1984. Partial purification and properties of rat liverglutaminase. Biochem. J. 220, 583–590.

Perera, S.Y., Chen, T.C., Curthoys, N.P., 1990. Biosynthesis and processing of renalmitochondrial glutaminase in cultured proximal tubular epithelial cells and inisolated mitochondria. J. Biol. Chem. 265, 17764–17770.

Perez-Gomez, C., Mates, J.M., Gomez-Fabre, P.M., del Castillo-Olivares, A., Alonso,F.J., Marquez, J., 2003. Genomic organization and transcriptional analysis of thehuman L-glutaminase gene. Biochem. J. 370, 771–784.

Perez-Gomez, C., Campos-Sandoval, J.A., Alonso, F.J., Segura, J.A., Manzanares, E.,Ruiz-Sanchez, P., Gonzalez, M.E., Marquez, J., Mates, J.M., 2005. Co-expression ofglutaminase K and L isoenzymes in human tumour cells. Biochem. J. 386, 535–542.

Porter, D., Hansen, W.R., Taylor, L., Curthoys, N.P., 1995. Differential expression ofmultiple glutaminase mRNAs in LLC-PK1-F+ cells. Am. J. Physiol. 269, F363–373.

Porter, L.D., Ibrahim, H., Taylor, L., Curthoys, N.P., 2002. Complexity and speciesvariation of the kidney-type glutaminase gene. Physiol. Genomics 9, 57–66.

Quesada, A.R., Sanchez-Jimenez, F., Perez-Rodrıguez, J., Marquez, J., Medina, M.A.,Nunez de Castro, I., 1988. Purification of phosphate-dependent glutaminasefrom isolated mitochondria of Ehrlich ascites-tumour cells. Biochem. J. 255,1031–1036.

Robinson, M.M., McBryant, S.J., Tsukamoto, T., Rojas, C., Ferraris, D.V., Hamilton, S.K.,Hansen, J.C., Curthoys, N.P., 2007. Novel mechanism of inhibition of rat kidney-type glutaminase by bis-2-(5-phenylacetamido-1,2,4-thiadiazol-2-yl)ethyl sul-fide (BPTES). Biochem J. 406, 407–414.

Sarantos, P., Abouhamze, A., Abcouwer, S., Chakrabarti, R., Copeland, E.M., Souba,W.W., 1994. Cytokines decrease glutaminase expression in human fibroblasts.Surgery 116, 276–283.

Sarantos, P., Abouhamze, A., Souba, W.W., 1992. Glucocorticoids regulate intestinalglutaminase expression. Surgery 112, 278–283.

Sarantos, P., Ockert, K., Souba, W.W., 1993. Endotoxin stimulates lymphocyteglutaminase expression. Arch. Surg. 128, 920–924.

Segura, J.A., Aledo, J.C., Gomez-Biedma, S., Nunez de Castro, I., Marquez, J., 1995.Tumor glutaminase purification. Protein Expr. Purif. 6, 343–351.

Shapiro, R.A., Farrell, L., Srinivasan, M., Curthoys, N.P., 1991. Isolation, character-ization and in vitro expression of a cDNA that encodes the kidney isoenzyme ofthe mitochondrial glutaminase. J. Biol. Chem. 266, 18792–18796.

Singer, R.H., 1998. Triplet-repeat transcripts: a role for DNA in disease. Science 280,696–697.

Smith, E.M., Watford, M., 1988. Rat hepatic glutaminase: purification and immu-nochemical characterization. Arch. Biochem. Biophys. 260, 740–751.

Smith, E.M., Watford, M., 1990. Molecular cloning of a cDNA for rat hepaticglutaminase. Sequence similarity to kidney-type glutaminase. J. Biol. Chem.265, 10631–10636.

Souba, W.W., 1993. Glutamine and cancer. Ann. Surg. 218, 715–728.Srinivasan, M., Kalousek, F., Curthoys, N.P., 1995. In vitro characterization of the

mitochondrial processing and the potential function of the 68-kDa subunit ofrenal glutaminase. J. Biol. Chem. 270, 1185–1190.

Szeliga, M., Obara-Michlewska, M., Matyja, E., Lazarczyk, M., Lobo, C., Hilgier, W.,Alonso, F., Marquez, J., Albrecht, J., 2008. Transfection with liver-type glutami-nase (LGA) cDNA alters gene expression and reduces viability, migration andproliferation of T98G glioma cells. Glia., ahead of print, PMID: 19062176.

Svenneby, G., Torgner, I.Aa., Kvamme, E., 1973. Purification of phosphate-dependentpig brain glutaminase. J. Neurochem. 20, 1217–1224.

Tang, A., Curthoys, N.P., 2001. Identification of zeta-crystallin/NADPH:quinonereductase as a renal glutaminase mRNA pH response element-binding protein.J. Biol. Chem. 276, 21375–21380.

Turner, A., McGivan, J.D., 2003. Glutaminase isoform expression in cell lines derivedfrom human colorectal adenomas and carcinomas. Biochem. J. 370, 403–408.

Watford, M., 1993. Hepatic glutaminase expression: relationship to kidney-typeglutaminase and to the urea cycle. FASEB J. 7, 1468–1474.

Watford, M., Vincent, N., Zhan, Z., Fannelli, J., Kowalski, T., Kovacevic, Z., 1994.Transcriptional control of rat hepatic glutaminase expression by dietary proteinlevel and starvation. J. Nutr. 124, 493–499.