QUEUOSINE DEFICIENCY IN EUKARYOTES COMPROMISES TYROSINE PRODUCTION THROUGH INCREASED TETRAHYDROBIOPTERIN OXIDATION

tautomerisation of quinonoid dihydrobiopterin caused by a defect at the DHPR recycling step, (iii) non-enzymatic oxidation of BH4 or (iv) competition between BH2 and dihydrofolate for reduction by the DHFR enzyme.

Queuosine is a modified pyrrolopyrimidine nucleoside found in the anticodon loop of transfer RNA acceptors for the amino acids tyrosine, asparagine, aspartic acid, and histidine. Since it is exclusively synthesised by bacteria, higher eukaryotes must salvage queuosine or its nucleobase queuine from food and the gut microflora. Previously, animals made deficient in queuine died within 18 days of withdrawing tyrosine-a non-essential amino acid-from the diet [Marks T, Farkas WR (1997) Biochem Biophys Res Commun 230:233-7]. Here we show that human HepG2 cells deficient in queuine and mice made deficient in queuosine modified transfer RNA, by disruption of the tRNA guanine transglycosylae (TGT) enzyme, are compromised in their ability to produce tyrosine from phenylalanine. This has similarities to the disease phenylketonuria, which arises from mutation in the enzyme phenylalanine hydroxylase or from a decrease in the supply of its cofactor tetrahydrobiopterin (BH4). Immunoblot and kinetic analysis of liver from TGT deficient animals indicate normal expression and activity of phenylalanine hydroxylase. By contrast, BH4 levels are significantly decreased in the plasma and both plasma and urine show a clear elevation in dihydrobiopterin, an oxidation product of BH4, despite normal activity of the salvage enzyme dihydrofolate reductase. Our data suggest that queuosine modification limits BH4 oxidation in vivo and thereby potentially impacts on numerous physiological processes in eukaryotes.
Bacteria and humans have co-evolved for millenia and many examples exist of how various symbiotic and commensal partnerships contribute to human health and nutrition ranging from the metabolism of complex carbohydrates to the provision of vital micronutrients (1). Queuosine is an example of a micronutrient, synthesised exclusively by bacteria but which, for poorly defined reasons, is utilised by almost all eukaryotic species with the exception of the baker's yeast, S. cerevisiae (2).
Bacterial queuosine biosynthesis occurs in two stages. Firstly, a series of five enzymatic steps convert guanosine triphosphate nucleoside (GTP) to the soluble 7-aminomethyl-7-deazaguanine (preQ 1 ) molecule. Subsequently, preQ 1 is inserted into the wobble position of tRNA containing a GUN consensus sequence (tyr, asp, asn, his) by means of the single enzyme species, tRNA guanine transglycosylase (TGT) and further remodeled in situ to queuosine (3). Eukaryotes must acquire queuosine or its free nucleobase, queuine, from food and the gut-microflora. Curiously, both cytosolic and mitochondrial tRNA species are modified by queuosine (2). The eukaryotic enzyme that performs this reaction, queuine tRNA ribosyltransferase, has recently been identified as a heterodimeric complex, consisting of the eukaryotic homologue of the catalytic TGT subunit and a related protein called queuine tRNA ribosyltransferase domain containing 1 (QTRTD1), both of which localize to the mitochondria (4,5).
Studies on germ-free (axenic) mice maintained on a chemically defined diet, provided clear evidence that eukaryotes are non-autotrophic for queuosine biosynthesis (6). Unchallenged, these animals appear normal. However, withdrawal of tyrosine from the diet resulted in symptoms of squinting, stiffness, lethargy, convulsion, and ultimately death after 18 days (7). Readministration of either chemically synthesised queuine or tyrosine alone prevented the symptoms; the latter result suggesting that tyrosine uptake and utilisation is unaffected by queuine status. It has been long established that tyrosine is a nonessential amino acid in higher eukaryotes as it can be synthesized from phenylalanine by the action of the phenylalanine hydroxylase (PAH) enzyme. It has been suggested therefore that the absence of queuine may affect the translation of the PAH enzyme leading to a dietary dependency on tyrosine supply (2,8).
In humans, tyrosine production occurs principally in the liver and kidney correlating with the expression of the PAH enzyme (9). Deficiency in PAH leads to the disease phenylketonuria, characterized by increased blood levels of phenylalanine (referred to hyperphenylalaninemia) and reduced levels of tyrosine. In performing its reaction PAH requires molecular oxygen and BH4 cofactor (Fig. 1). BH4 is produced from GTP by the enzymes GTP cyclohydrolase I, 6-pyruvoyltetrahydropterin synthase and sepiapterin reductase. In generating BH4, other intermediary reactions at the sepiapterin reductase step are performed by carbonyl reductase and member proteins of the aldo-keto reductase family (10). The BH4 cofactor may also be recycled by the activity of two enzymes, pterin-4a-carbinolamine dehydratase and dihydropteridine reductase, which has particular importance for tyrosine biosynthesis in the liver. Deficiency in any of the BH4 enzymes, with the exception of sepiapterin reductase, causes hyperphenylalaninemia. Recent studies have shown that BH4 is also highly susceptible to autooxidation in vivo producing the metabolite 7,8 dihydrobiopterin (BH2), whose accumulation is limited by the enzyme dihydrofolate reductase (DHFR) through reduction of BH2 to BH4 (11).
By extrapolation from the current data, the tyrosine dependency of queuine deficient animals may relate to the loss of PAH activity or altered BH4 cofactor supply. In the present study, we show that queuosine modification of tRNA, or hypothetically another unknown RNA substrate of the queuine tRNA ribosyltranferase enzyme, as opposed to free queuine base, is required for normal tyrosine production in eukaryotes. Decreased BH4 levels, concomitant with a marked accumulation of BH2 suggests that oxidation of BH4 cofactor underlies the defect.

Animals:
Mice were bred and housed under specific pathogen-free conditions. Procedures were performed on mice at 6-8 weeks of age, unless otherwise stated, according to regulations and guidelines of the Ethics Committee, Trinity College Dublin and the Irish Department of Health. Tyrosine free chemically defined diet (AIN-76A purified diet) was obtained from Harlan Teklad.

Radiolabelled phenylalanine hydroxylase assay:
HepG2 cells were grown in serum-free (SF) medium, SF medium supplemented with queuine (300nM) or in DMEM with 10% FBS. Intracellular tyrosine and phenylalanine were depleted by incubation in SF medium not containing tyrosine or phenylalanine (Lonza; custom synthesis) and supplemented with cyclohexamide (20µg/ml) with constant shaking at 70 rpm for 1 hour. Subsequently, cells were washed and incubated in KRHL buffer (120 mM NaCl, 4.8 mM KCl, 10 mM D-glucose, 2.5 mM CaCl 2 , 1.3 mM MgSO 4 , 2.5 mM Hepes, pH 7) containing 20 µM, 60 µM and 100 µM L-[2,6-3 H]-Phenylalanine (54 Ci/mmol; Amersham) and 20 µg/ml cyclohexamide for 30 minutes with constant shaking. Duplicate wells for each experimental point were used to determine cell protein content. Cells were solubilised with M-PER solution (Pierce) and the protein concentration determined by Bradford assay (Bio-Rad). After 30 minutes, medium was removed, cells were washed rapidly with ice-cold PBS and lysed in ice-cold 10% trichloroacetic acid (100 µl per well). Cells were scraped from the dish and the lysate centrifuged at 16,000 x g for 15 minutes. The supernatant was stored at -70 o C until analysed. Intracellular phenyalanine and tyrosine were measured by HPLC on a Zorbax 300SB-C18 column (Agilent) pre-equilibrated with Mobile Phase Buffer (100 mM sodium phosphate pH 1.9, 300 µM octyl sodium sulphate, 500 µM of EDTA and 6% HPLC grade Methanol) and run at 1 ml/min. Cell extracts (50 µl) were spiked with cold tyrosine (0.1 mM final) and phenylalanine (0.5 mM final) and detected by fluorescence with an excitation of 258 nm and emission of 288 nm. Samples (1 ml) were collected and radiolabelled amino acids evaluated by scintillation counting.
Phenylalanine metabolism in mouse: L-Phenylalanine (40 mg/ml) was dissolved in 0.9% saline solution containing 7.5 mM NaOH and injected i.p. at 1 mg per gram body weight. Blood was collected from the ventral caudal artery in K 2 -EDTA tubes (36 µg per ml of blood) and promptly centrifuged at 2,000 x g for 10 minutes at 4 o C. Samples were deproteinized by adding 9 x volumes of 1.11 M HClO 4 , the sample allowed stand for 10 minutes, then centrifuged at 16,000 x g for 6 minutes at 4 o C. An aliquot of the supernatant was treated with 0.091 volumes of 10 M KOH, incubated on ice for 10 minutes and then centrifuged at 16,100 x g for 6 minutes at 4 o C. The supernatant was filtered through a 0.22µm ultrafree-MC filter unit (Millipore) at 16,100 x g for 1 min at 4 o C. HPLC separations were performed at ambient temperature on a Zorbax 300SB-C18 column (Agilent) as described above. Samples (65 µl per run) had a pH value less than 7.0. Tyrosine and phenylalanine were detected by fluorescence with an excitation wavelength of 258 nm and an emission wavelength of 288 nm and the levels quantified by integration of the peak area using EZStart 7.3SPI Software (Shimadzu).
Tetrahydrobiopterin measurement in plasma and urine: Blood (150 µl) was collected from the tail ventral caudal artery in tubes containing K 2 EDTA (36 µg per ml of blood) and 0.1% (w/v) DTT and plasma separated by centrifugation at 2,000 x g for 10 minutes at room temperature. Urine samples were collected and kept at 4 o C and out of sunlight. Total protein was determined by the Bradford Assay (BioRad). Prior to biopterin measurements urine samples were diluted to 30 µg/ml and plasma samples to 3 mg/ml. BH4 levels were measured by HPLC analysis after iodine oxidation in acidic or alkaline conditions as described (15) on a Spherisorb ODS1 C18 column (particle size 5 µm; 250 x 4.6 mm; Waters) by an isocratic gradient in water with 5% (v/v) methanol at a flow rate of 0.6 ml/min.

LC-MS analysis of tRNA.
Queuosine nucleoside content of bulk tRNA (16) was determined by liquid chromatography-tandem mass spectrometry according to published protocols (17). DHFR activity assay: Liver (1 gram) was homogenized in 3 ml of ice-cold homogenization buffer (0.2M Tris-HCl, pH 7.6 at 4 °C containing 10mM DTT and protease inhibitors) and centrifuged at 100,000 x g for 60 minutes at 4 °C. The standard spectrophotometric assay was used to measure DHFR activity (18).

Queuine deficiency in HepG2 cells compromises phenylalanine to tyrosine conversion:
The original queuine and tyrosine depletion study was performed on an outbred Swiss mouse strain, raising the possibility that the effect is unique to mice or to the genetic background of the animals used. To rule out this possibility we examined the effect of queuine deficiency on tyrosine production in HepG2 cells-a human hepatoma cell line functionally capable of synthesizing tyrosine from phenylalanine and containing the necessary enzymes for BH4 synthesis and recycling (19,20).
Cell growth medium supplemented with 10% foetal bovine serum (FBS) contains approximately 1 to 2 x 10 -8 M queuine (21), allowing for full modification of tRNA. By contrast, horse serum (HS) is essentially queuine free (22) providing a means to deplete queuine from cells. Unfortunately, FBS and HS influence PAH activity in ways unrelated to queuine (23). Therefore, for the purpose of this study HepG2 cells were also grown in serum-free (SF) medium in the absence or presence of chemically synthesized queuine. The guanine incorporation assay (4) was used to evaluate the queuosine status of tRNA ( Fig. 2A). High levels of guanine incorporation occurred in tRNA extracted from HepG2 cells grown in SF or HS-containing medium (signifying a depletion of the queuosine modification in tRNA), whereas guanine incorporation was low for tRNA extracted from cells grown in SF medium supplemented with queuine or in FBS-containing medium. The results indicate that SF medium can deplete queuosine modified tRNA (Q-tRNA), and by inference, queuine levels in cells. The results of the enzymatic assay were confirmed by LC-MS analysis ( Supplementary Fig. 1).
To measure tyrosine production in vivo, HepG2 cells were depleted of tyrosine and phenylalanine and subsequently incubated with increasing concentrations of 14 C-phenylalanine. Intracellular phenylalanine (Fig. 2B) and tyrosine (Fig. 2C) were analysed by reversed phase HPLC. Our results show that queuine status does not influence phenylalanine uptake, which is linear with respect to the concentration supplied. However, the ability to produce tyrosine is reduced by 15-40% depending on the amount of phenylalanine administered. The lack of a dose response for tyrosine synthesis in the experiment is considered to arise from a low capacity of HepG2 cells for phenylalanine hydroxylation, which is saturated even at the lowest concentration of 14 Cphenylalanine used.
In an attempt to distinguish whether the effect on tyrosine formation arises from the depletion of queuine base or queuosine modification of (t)RNA, a known inhibitor of the queuine tRNA ribosyltransferase activity, 7-methylguanine (24), was used to treat HepG2 cells (Fig. 2D). Taking the queuosine status of tRNA of cells grown in serum free medium as being fully unmodified (39.15 pmol [ 14 C] guanine insertion/A260 tRNA) and those grown in serum free medium in the presence of queuine as being fully modified (4.54 pmol [ 14 C] guanine insertion/A260 tRNA), the administration of 7-methylguanine resulted tRNA being approximately 19% unmodified with respect to queuosine (11.16 pmol [ 14 C] guanine insertion/ A260 tRNA). This was concomitant with a significant reduction in tyrosine formation compared to queuine sufficient controls. These results suggest that the status of queuosine modified (t)RNA, as opposed to the levels of queuine nucleobase, influences the ability of HepG2 cells to produce tyrosine.
Qtrt1 gene-trap mice are deficient in queuosine modified tRNA: In order to create mice lacking the TGT enzyme, a gene-trap insertion strategy was employed. The ES cell line FHCRC-GT-S12-11A1 contains an integrated copy of the ROSAFARY vector in the Qtrt1 gene, which encodes the TGT subunit. The ROSAFARY insert was putatively mapped to intron 2 creating the allele Qtrt1 Gt(FHCRC-GT-S12-11A1)Sor , subsequently abbreviated as Qtrt1 Gt .
Attempts to confirm the suspected position of the gene-trap cassette were unsuccessful and a nested PCR approach was used to map the region spanning exon 1-4. Our results placed the cassette in exon 3 of the Qtrt1 gene (Fig 3A). Subsequent cloning and sequencing of the 5'and 3' flanking regions of the cassette unequivocally mapped the cassette to position 21,216,900 on chromosome 9; dividing exon 3 of Qtrt1 in approximately two parts. ES cells were microinjected into blastocysts, giving rise to 12 male mice exhibiting 50-98% chimerism, two of which achieved germline transmission on multiple occasions.
The ROSAFARY vector is designed such that a promoter trap module (SAβgeo * pA) with an artificial adenoviral splice acceptor acts as the 3' terminal exon to create a β-galactosidaseneomycin (β-geo) fusion marker with any upstream exons. In addition, a poly-A trap module containing the hygromycin resistance gene (PGKhygSD) with a downstream splice donor site forms a fusion transcript with any downstream exons (25). The placement of the cassette in exon 3 of Qtrt1 would necessitate that this exon is skipped during splicing as a read through transcript would encounter multiple stop codons precluding the production of the β-geo protein. This was not considered possible since β-geo was used to select for G418 antibiotic resistance during ES cell screening (25). Analysis of the transcript from the Qtrt1 Gt allele by RTPCR demonstrated that, as expected, exon 3 is not produced (Fig. 3B) and the TGT protein is made in two fragments; a Met1-Thr104::β-geo fusion and a hygromycin::Gly151-Thr403 fusion (Fig. 3A). Importantly, the loss of exon 3 would remove an essential active site aspartate (Asp102; according to mouse TGT numbering) and a serine residue involved in substrate recognition (Ser103) negating the possibility of catalytic activity (4,5).
To obtain mice lacking active TGT protein heterozygous animals (Qtrt1 Gt/+ ) were intercrossed. Southern blotting analysis of the 5'-and 3'-ROSAFARY insertion sites and the β-geo cassette region are consistent with a single gene-trap insertion in exon 3 of the Qtrt1 locus (Fig. 3C). As explained above, disruption of the Qtrt1 gene, which encodes the catalytic subunit of the eukaryotic queuine tRNA ribosyltransferase complex would be expected to be functionally incapable of Q-tRNA formation. Confirmation of the Q-tRNA status of animals was made by LC-MS analysis of bulk tRNA extracted from the liver of six-week old Qtrt1 +/+ , Qtrt1 Gt/+ and Qtrt1 Gt/Gt mice (Fig. 3D). As expected, wild-type animals contained Q-tRNA. However, Q-tRNA could not be detected in either heterozygous or homozygous animals. That heterozygous mice failed to produce detectable levels of Q-tRNA may be explained by the fact animals are born germ-free and without Q-tRNA leading to the possibility that the single normal Qtrt1 allele is haploinsufficient or that the fusion constructs are acting in a dominant negative manner to sequester away limited amounts of QTRTD1 from active TGT protein. Irrespectively, analysis of tRNA from the liver of older heterozygous animals, at sixteen weeks of age, revealed that appreciable levels of Q-tRNA had been produced (Supplementary Fig. 2).
Genotype analysis of 36 litters (229 pups) from heterozygous intercrossing (Supplementary Fig. 3) showed that TGT deficiency does not influence viability or sex bias (Supplementary Table 1). In addition, breeding of homozygous animals revealed that both males and females have normal fecundity (Supplementary Table 2), concurring with the lack of an obvious phenotype in queuine deficient fly (26), worm (27), and mouse (28).

TGT disruption in mouse decreases tyrosine production from phenylalanine: The studies on
HepG2 cells suggested that queuine tRNA ribosyltransferase inactivation negatively impacts tyrosine biosynthesis. The generation of TGT deficient mice provided a means to explore this effect in whole animals. Phenylalanine was injected into the peritoneum of 6-8 week old mice (1 mg per gram body weight) that had been maintained on a normal diet. At various time points, blood samples were collected from the ventral caudal artery and plasma tyrosine analysed by HPLC (Fig. 4A). Mice of all three genotypes were found to produce equivalent amounts of tyrosine within the first hour. However, subsequent to this, a sharp decline in tyrosine production by Qtrt1 Gt/+ and Qtrt1 Gt/Gt mice was observed. This contrasts to the sustained production of tyrosine in wild-type animals, which only begun to decline from two hours onwards as serum phenylalanine from the inital peritoneal bolus became exhausted. The increased ability of wild-type mice to produce tyrosine relative to hetero-and homo-zygous genetrap mice is readily apparent from the conversion ratio of phenylalanine to tyrosine in plasma (Fig.  4B). Similar results were obtained using animals that had been fasted for twenty-four hours prior to phenylalanine loading (Supplementary Fig. 4).
To determine whether TGT deficiency phenocopies the tyrosine dependency of axenic, queuine-deficient mice, two male animals of each genotype Qtrt1 Gt/+ , Qtrt1 Gt/+ and Qtrt1 Gt/Gt were maintained on a chemically defined tyrosine free diet (AIN-76A based) for two months. No lethality occurred and no overt physical or behavioural abnormalities could be visually ascribed. This result suggests that either the germfree status of the original study presented additional compounding factors, that it was the lack of free queuine which was responsible for the phenotype, and/or that bacteria in the gut of TGT deficient mice can supply sufficient amino acid to permit survival. In this regard, studies show that the intestinal microbiota of adult humans may provide 1-20% of circulating plasma lysine and threonine (1).

TGT inactivation does not affect PAH expression or activity in mouse liver:
A decrease in the ability to metabolise phenylalanine to tyrosine is a characteristic of the disease phenylketonuria which can arise from defects in the expression or activity of the PAH enzyme. The position of queuosine within the wobble position of the anticodon could potentially affect PAH translation as the absence of queuosine was shown to drastically decrease translation of virF mRNA in the pathogen Shigella flexneri (29) and to induce +1 frameshifting in bacteria (30). To explore a possible effect on PAH translation, antisera was raised to recombinant PAH and liver samples analyzed by immunoblot assay (Fig. 5A). No observable difference in PAH expression was detected across wild-type and gene-trap animals ruling out a defect in translation.
A variety of other mechanisms are known to regulate PAH activity in vivo, including its activation by phenylalanine and phosphorylation, in addition to its allosteric inhibition by BH4 cofactor (31,32). Previous reports suggest that queuine can enhance the phosphorylation of unspecified cytosolic proteins (22) whereas in other cases a decrease in phosphorylation was observed (33,34,35). In addition, notable similarities exist between the structure of queuine and tetrahydrobiopterin-both being derived from GTP. Indeed, biopterins are known inhibitors of the queuine ribosyltransferase activity in vitro (36) and in vivo (24). As such, a reciprocal relationship between the queuosine modification and PAH could be envisaged by, for example, counterbalancing the allosteric inhibition of PAH by BH4.
Liver cytosolic fractions were examined for PAH activity in all three genotypes revealing that queuine has no impact on the specific activity of tyrosine formation; all being approximately 50 mU.mg −1 across each of the genotypes (Fig. 5B). The results clearly show that the ex vivo activity of the PAH enzyme is not affected by the queuosine status of (t)RNA.
Conceivably, the absence of TGT could lead to the production of an inhibitory metabolite for the PAH reaction which would be diluted out in the standard spectrophotometic assay. It is known for example that 7-biopterin-formed by spontaneous re-arrangement of 4a-hydroxy-tetrahydrobiopterin during BH4 recycling-can competitively inhibit the PAH enzyme (37). To rule out this possibility, tyrosine production by cytosolic liver extracts diluted only marginally (1:20) by the addition of buffer, catalase (170 ng/µl) and 6R-BH4 cofactor (100 µM) was performed (Fig. 5C). The results show that tyrosine production remained unchanged between wild-type and homozygous gene-trap mice over a one hour period. It can therefore be concluded that neither the expression of PAH protein nor its ex vivo activity is affected by the queuosine status of (t)RNA and further that queuosine deficiency does not lead to the accumulation of inhibitory metabolites in the liver.
TGT disruption results in decreased plasma BH4 and elevated BH2 in plasma and urine: Given that loss of TGT has no impact on PAH an analysis of pterins in plasma and urine was carried out following differential oxidation with iodine under acidic and basic conditions (representative figures of these results are presented in Supplementary Fig. 5).
In the plasma of heterozygous and homozygous gene-trap animals, no significant changes occurred in the total biopterin levels relative to wild-type animals (Fig. 6A). However, the levels of BH4 were decreased by ~30% (Fig.  6B) concomitant with an increase in the oxidised biopterin, dihydrobiopterin (Fig. 6C). In urine, there was an increase of ~20% in total biopterin in gene-trap mice relative to wild-type animals (Fig.  6D). Although no significant changes were seen in BH4 levels (Fig. 6E) an increase of ~40% in BH2 was observed (Fig. 6F). These results could be explained either by increased auto-oxidation of BH4, changes in the activity of the sepiapterin reductase and dihydropterin reductase enzymes or alternatively that the activity of the salvage pathway for BH2, through the dihydrofolate reductase enzyme is suboptimal.
TGT inactivation does not affect the activity of dihydrofolate reductase in liver: Previously, reports have shown that mice treated with methotrexate-a potent inhibitor of dihydrofolate reductase (DHFR)-experience increased levels of BH2 coupled to a decrease in endogenous BH4 in liver, kidney and blood (15). Interestingly, studies show that exogenously administered BH4 is primarily oxidised to BH2 in the body through an ill-defined mechanism before being taken up by tissue and reduced back to BH4 (38), underlining the importance of DHFR in limiting the accumulation of oxidised biopterin.
In light of the clear accumulation of BH2 in the plasma and urine of Qtrt1 gene-trap mice the levels of DHFR were examined. DHFR activity was measured spectrophotometrically in liver homogenate using 7,8-dihydrofolate as substrate (Fig. 7A). There was no detectable difference in activity across each of the three genotypes ruling out the possibility of a defect in the BH2 salvage pathway and instead pointing to increased production of oxidised BH4 as being a principle defect in Qtrt1 gene-trap mice (Fig. 7B).

DISCUSSION
Previously, it has been shown that animals deficient in the bacterial derived queuine molecule require dietary tyrosine for their survival (7). In the current study we demonstrate that both human HepG2 cells made deficient in queuine and transgenic animals incapable of forming queuosine modified (t)RNA (Qtrt1 gene-trap mice) exhibit a decreased ability to produce tyrosine from phenylalanine; the former result suggesting the defect is cell autonomous and relevant to humans and the latter result indicating that the defect arises from the absence of the queuosine modification in RNA-either tRNA or hypothetically another unknown RNA substrate of the queuine tRNA ribosyltransferase complex-as opposed to the free queuine nucleobase. The relatively mild defect seen in the catabolism of phenylalanine to tyrosine would not be expected to present as hyperphenylalanemia except under circumstances of unusually high phenylalanine intake. Therefore, that mutations to the queuine pathway may be of relevance to phenylketonuria in humans is doubtful.
Qtrt1 gene-trap animals, similar to queuine deficient mice (28), appeared normal, displaying similar viability and fecundity to wild-type littermates. This observation is in agreement with earlier studies on various other queuine deprived eukaryotic species including dictostyelium (39), fly (26) and worm (27). It may be concluded that queuine or queuosine modified (t)RNA does not impact on the development, growth or reproduction of eukaryotic organisms under laboratory conditions.
At variance with the lack of an overt phenotype in queuine deficient animals the physiological manifestation of animals' co-deficient in queuine and tyrosine was dramatic and included symptoms of lethargy, laboured breathing, convulsion and death after only eighteen days (7). Although TGT deficient animals had decreased ability to produce tyrosine none of the aforementioned symptoms presented when these animals were placed on a tyrosine free diet. A number of explanations can be envisaged. Firstly, the germfree status of the animals in the queuine depletion study coupled to the administration of a synthetic liquid diet may have compounded the severity of tyrosine deprivation. This situation contrasts with our study where animals were fed a chow based chemically defined diet and would be expected to have a normal gut flora. Secondly, and related to the above point, queuine deficiency and TGT disruption may not be equivalent. Queuine deficient mice have no queuine or queuosine modified (t)RNA whereas Qtrt1 gene-trap mice are almost certainly unabated in their ability to harvest queuine from the gut and transport it into the cell. Thirdly, it is conceivable that Qtrt1 gene-trap animals acquire sufficient tyrosine from the gut miroflora as studies using 15 N-labelled microbial amino acids determined that up to 20% of circulating lysine and threonine can be derived from the intestinal microbiota (1).
At the outset of the study it was considered that queuine status may influence the activity of the PAH protein, which is required for the catabolism of phenylalanine to tyrosine. Our results show that neither the expression nor ex vivo activity of the PAH protein is affected by TGT disruption. Rather, the levels of the essential PAH cofactor BH4 is significantly decreased in plasma concomitant with an accumulation of BH2, the inactive oxidised product of BH4. It is important to stress that the Qtrt1 gene-trap mice are capable of producing significant amounts of tyrosine but this ability is lost over time following a phenylalanine challenge, presumably due to diminished BH4 levels and accumulating BH2-a known competitive inhibitor of PAH from in vitro studies (40). That a 30% decrease in plasma BH4 levels and a twofold increase in BH2, as observed in our study, could impinge on tyrosine formation may be appreciated from the fact that the concentration of BH4 in mouse liver (21 pmol/mg protein) is only half that of PAH (40 pmol PAH subunit/mg protein) (41) and that under normal conditions the levels of BH4 are sub-saturating (5-10 µM) with respect to the PAH K m value for BH4 cofactor (25 µM) (42). Thus, even small changes in cofactor supply could negatively impact on the production of tyrosine from phenylalanine.
Despite the increased levels of BH2 seen in Qtrt1 gene-trap mice, the activity of the DHFR enzyme, responsible for salvaging oxidised cofactor, remained unaffected. This result suggests that DHFR is incapable of maintaining tetrahydrobiopterin cofactor in a reduced state under queuosine deficient conditions. The ineffectiveness of the DHFR enzyme in this regard may be due to the low levels of BH2 generated in liver-given that we observe only a concentration of 1µM in plasma-and the fact that the K m of the DHFR enzyme for BH2, at 6.42 µM, is very much higher than the other principle DHFR substrate 7,8 dihydrofolate, which has a K m of 0.17 µM (43). Although the present study has not considered how queuosine deficiency may relate to changes in folate metabolism, one could envisage based on the arguments above that an elevation in 7,8 dihydrofolate could directly compete with BH2 for reduction by DHFR and, as such, the problems seen with tetrahydrobiopterin metabolism in the current study could be secondary to changes in folate metabolism.
The impact of Qtrt1 disruption on brain function may differ significantly from that of liver, where the concentration of DHFR is low (44). Should a similar accumulation of BH2 occur it could potentially limit the production of numerous biogenic amino neurotransmitters whose production also under the control of BH4 through the activity of tyrosine hydroxylase and tryptophan hydroxylase (Fig.1).
BH4 is also required for the activity of each of the nitric oxide synthase isoforms, inducible (iNOS), neural (nNOS) and endothelial (eNOS) and participates in several steps in nitric oxide generation through stabilising the active dimeric form of the enzyme, acting as an electron donor during oxygen activation and functioning in electron recapture prior to nitric oxide release (45). Numerous studies indicate that BH4 is highly susceptible to auto-oxidation to BH2 in vivo (15,37) and that a low BH4/BH2 ratio, as seen in Qtrt1 gene-trap mice, results in the uncoupling of eNOS leading to superoxide formation and endothelial dysfunction (46,11). Indeed, increased BH4 has proven to be protective in several experimental disease models by its ability to reduce blood pressure (47), decrease atherosclerosis (48), and prevent diabetic complications (49). It remains to be determined whether queuine has a protective role in any of these disease processes.
The underlying cause of BH4 depletion and BH2 accumulation in TGT deficient mice is at present uncertain. In vitro, oxygen and peroxinitrite can oxidise BH4 to quinonoid dihydrobiopterin which readily rearranges to BH2 (50,51) and in vivo it has been shown that BH2 rapidly forms in the circulation following administration of BH4 (15). The authors consider changes to intracellular homeostasis as the most probable explanation for the increased oxidation of tetrahydrobiopterin as queuine deficiency has previously been suggested to affect the activity of a number of antioxidant systems (52,53) and to influence the metabolic state of the cell (33,52,54). Other explanations for the altered BH4/BH2 ratio include inadequate BH4 recycling at the dihydropterin reductase step-responsible for quinonoid BH2 reduction to BH4-or increased production of BH2 through the sepiapterin reductase/carbonyl reductase step. A further cause for elevated BH2 may relate to elevated intracellular dihydrofolate since DHFR has a K m for dihydrofolate that is more than an order of magnitude lower than that of dihydrobiopterin, as described earlier (43).
It is reasonable to assume that the queuosine modification of tRNA may subtly affect a number of biological processes through broad changes in the protein translation profile. In this regard, it has previously been shown using molecular simulations that queuosine helps to confine the dynamic movement of the anticodon (55) and studies using two histidine isoacceptors from Drosophila showed that queuosine can limit the bias that exists among synonymous codon usage (56). It is hoped that further physiological analysis of TGT deficient mice will provide further insight into the role of this intriguing bacterial derived micronutrient.
dihydropteridine reductase (DHPR) to regenerate reduced cofactor. Disorders that affect the biosynthesis or regeneration of BH4 may be detected by quantitative analysis of neopterin, sepiapterin, 7-biopterin and BH2 as indicated. Adapted from (38). . tRNA that has been modified by queuosine is unable to accept 14 C-Guanine into the anticodon loop by the E. coli TGT enzyme reaction. The control reaction contained tRNA from SF grown cells but did not contain TGT enzyme. (B) The influence of queuine on phenylalanine uptake and (C) tyrosine production by HepG2 cells. Cells, grown in the indicated medium, were placed in a phenylalanine and tyrosine deficient medium for one hour. Subsequently, 14 Cphenylalanine was added for 30 minutes. Cells were lysed and radiolabelled amino acids analyzed by separation on a C18 HPLC column followed by scintillation counting. (D) The effect of 7-methylguanine on tyrosine biosynthesis by HepG2 cells. Cells were cultivated in either SF medium alone (SF), SF medium supplemented with queuine (SF + Queuine) or SF medium pre-treated with 7-methylguanine (7MG) for 24 hours prior to, and after the addition of queuine (SF + Queuine + 7MG). Cells were incubated with increasing concentrations of 14 C phenylalanine for 1 hour (rather than the 30 minutes used earlier) and intracellular phenylalanine and tyrosine levels measured as described above.  (1mg/g body weight) was administered i.p. to eight individual wild-type (+/+), heterozygous (+/Gt) and homozygous (Gt/Gt) animals and at the times indicated blood was withdrawn from the ventral caudal artery. Plasma phenylalanine and tyrosine were separated by HPLC on a reversed-phase C18 column and detected by their absorbance at 206 nm. Each point was measured in triplicate. (B) The ratio of plasma tyrosine to phenylalanine (conversion ratio) in animals at various times after the administration of phenylalanine.   Under conditions of full queuosine modification of (t)RNA only basal levels of BH2 are produced and are cleared by DHFR. However, when (t)RNA is unmodified by queuosine, BH2 production is enhanced, despite normal DHFR activity, either through (i) increased production of BH2 by SR/CR at the pyruvoyl tetrahydrobiopterin to BH4 reaction step, (ii) increased production and tautomerisation of quinonoid dihydrobiopterin caused by a defect at the DHPR recycling step, (iii) non-enzymatic oxidation of BH4 or (iv) competition between BH2 and dihydrofolate for reduction by the DHFR enzyme.
by guest on July 25, 2018