Sec (selenocysteine) is biosynthesized on its tRNA and incorporated into selenium-containing proteins (selenoproteins) as the 21st amino acid residue. Selenoprotein synthesis is dependent on Sec tRNA and the expression of this class of proteins can be modulated by altering Sec tRNA expression. The gene encoding Sec tRNA (Trsp) is a single-copy gene and its targeted removal in liver demonstrated that selenoproteins are essential for proper function wherein their absence leads to necrosis and hepatocellular degeneration. In the present study, we found that the complete loss of selenoproteins in liver was compensated for by an enhanced expression of several phase II response genes and their corresponding gene products. The replacement of selenoprotein synthesis in mice carrying mutant Trsp transgenes, wherein housekeeping, but not stress-related selenoproteins are expressed, led to normal expression of phase II response genes. Thus the present study provides evidence for a functional link between housekeeping selenoproteins and phase II enzymes.
- gene expression
- selenocysteine (Sec) tRNA
- Trsp knockout
Several trace elements have important roles in human health and their over-abundance or reduced levels result in severe health problems. Selenium is one such essential micronutrient with antioxidant properties, whose deficiency has been associated with several disorders . Selenium is incorporated into proteins (selenoproteins) as the amino acid Sec (selenocysteine) and its biological function is believed to be exerted in large part by these proteins . To date, 25 selenoprotein genes have been identified in the human genome and 24 in the mouse genome . The incorporation of Sec into proteins is a unique process in that it uses the stop codon, UGA, to decode this amino acid and involves a distinctive tRNA, designated tRNA[Ser]Sec. A number of other cis- and trans-acting factors are also required that form a complex with Sec-tRNA[Ser]Sec mediating the co-translational incorporation of Sec into protein [2,4,5].
Higher vertebrates have two Sec-tRNA[Ser]Sec isoforms that differ from each other by a single methyl group on the 2′-O-hydroxyribosyl moiety at position 34 . This methyl group is designated Um34. Both isoforms also contain the base, 5′-methylcarboxylmethyluracil (mcm5U), at position 34. Since both isoforms contain mcm5U, but only one of them contains Um34, they are designated mcm5U (i.e. the isoform lacking Um34) and 5′-methylcarboxymethyl-2′-O-methyluridine (mcm5Um; i.e. the isoform containing Um34). Sec-tRNA[Ser]Sec has three additional modified bases, pseudouridine at position 55, 1-methyladenosine at position 58 and N6-isopentenyladenosine (i6A) at position 37. The addition of Um34 is the last step in the maturation of Sec-tRNA[Ser]Sec and this step is stringently dependent on the prior synthesis of all base modifications . In addition, Um34 synthesis is influenced by selenium status, and selenium deficiency leads to an enrichment of mcm5U as compared with mcm5Um, whereas selenium adequacy reverses this ratio . The levels of the two isoforms modulate expression of different selenoproteins wherein some selenoproteins [e.g. GPx (glutathione peroxidase) 1 and 3 that function largely as stress-related proteins] are preferentially expressed in the presence of mcm5Um, whereas others [e.g. TR (thioredoxin reductase) 1 and 3 that function as essential housekeeping proteins] are preferentially expressed in the presence of mcm5U [7,8].
The gene encoding Sec-tRNA[Ser]Sec (Trsp) is present in single copy and its expression is essential for the synthesis of all selenoproteins. Selenoproteins are the only known class of proteins in eukaryotes whose expression is regulated by a single tRNA, and manipulating the expression of Trsp in mice modulates selenoprotein synthesis. Since removal of Trsp is embryonic lethal [9,10], the conditional knockout of Trsp  gave rise to several useful models for studying the role of selenium and selenoproteins in development and health (reviewed in ). In one of these models, we targeted the removal of Trsp in hepatocytes that demonstrated an essential role of selenoproteins in proper liver function . Additionally, we rescued Trsp null mice with transgenic mice carrying a mutant Trsp transgene [7,8]. In one mutant Trsp transgene, A37 was changed to G [7,13], which resulted in loss of both i6A and Um34 . We also produced a second transgenic mouse, where T34 was changed to A in the mutant transgene  and the resulting tRNA gene product also lacked Um34. Transgenic mice carrying mutant Trsp transgenes were used to replace selenoprotein synthesis in mice lacking Trsp in hepatocytes by matings between these two mouse lines as described previously . Introduction of either the A34 or G37 mutant transgenes into the liver Trsp-knockout mice selectively replaced selenoproteins involved in housekeeping functions, but not those involved in stress-related functions [7,8]. Furthermore, the number of gene copies of the mutant G37 transgene varied from 2 in one of the transgenic mouse lines we developed to 16 in the other transgenic mouse line.
In the present study, a comparative analysis of gene expression in the liver of the Trsp-knockout mice, designated ΔTrsp herein, and the A34 and G37 transgenic mice with gene expression of wild-type mice, was carried out using microarrays. These studies showed that the loss of selenoproteins in Trsp-knockout mice was associated with an enhanced expression of several phase II response genes and their corresponding enzymes. Phase II response genes are enzymes involved in detoxification as well as protection against oxidative stress. Interestingly, replacement of housekeeping selenoproteins in A34 or G37 transgenic mice resulted in the levels of phase II enzymes returning to normal. Taken together, the results suggest a functional association between housekeeping selenoproteins and phase II enzymes, wherein the loss of function of some housekeeping selenoproteins may be compensated for by phase II enzymes in the liver of the knockout mouse.
NuPage polyacrylamide gels, PVDF membranes, See-Blue Plus2 protein markers, TRIzol® and Superscript II RT (reverse transcriptase) were purchased from Invitrogen. SuperSignal West Dura extended duration substrate was from Pierce, and Cy3 and Cy5 mono-reactive dyes were from GE Healthcare. GSTA [GST (glutathione transferase) Alpha], GSTM (GST Mu) and EPHX1 (epoxide hydrolase 1) antibodies were obtained from Detroit R&D, and HMOX1 (haem oxygenase 1) antibodies, anti-mouse and anti-rabbit HRP (horseradish peroxidase)-conjugated secondary antibodies were from Santa Cruz Biotechnology. AOX1 (aldehyde oxidase 1) antibodies were from BD Biosciences and β-actin antibodies and anti-goat HRP-conjugated secondary antibodies were from Abcam. Primers used for real-time PCR were obtained from Sigma–Genosys. All other reagents were of the highest grade available and were obtained commercially.
Mouse lines and genotyping
The mice analysed in the present study were all males, 6–8 weeks of age in a B6/FVB genetic background and were fed a selenium-sufficient diet. Each mouse line used in the present study, preparation of the mutation carried in the A34 and G37 Trsp transgenes, and the manner in which these mouse lines were generated are described in detail elsewhere [7,8], and their genotypes and designations are summarized in Table 1. The care of animals was in accordance with the National Institutes of Health institutional guidelines under the expert direction of Dr Kyle Stump (NCI, National Institutes of Health, Bethesda, MD, U.S.A.). DNA was extracted from mouse tail clippings and the genotype determined by PCR with the appropriate primers as described previously [7,12].
Total RNA from liver of wild-type (Trsp), liver knockout (ΔTrsp) and transgenic (A34, G37L and G37H; for an explanation of these see Table 1) mice was isolated using TRIzol® reagent according to the manufacturer's protocol (Invitrogen) and labelled using the Fairplay® II microarray labelling kit (Stratagene). For indirect labelling of RNA, 15 μg of both control and experimental RNA was used to generate cDNA, using aminoallyl dNTP mix according to the manufacturer's protocol. The resulting cDNA was purified using a MinElute column (Qiagen) and eluted from the column with 10 μl of elution buffer (provided in the MinElute kit) and dried using a speed-vac for 15 min. Samples were next coupled to 5 μl of 2× coupling buffer (provided in the Fairplay®II microarray labelling kit) and 5 μl of monofunctional dye and incubated at room temperature (22 °C) in the dark for 30 min. Following incubation, the labelled cDNA was purified using a MinElute column and eluted with 10 μl of elution buffer.
Mouse oligonucleotide glass arrays, containing 70-mer oligonucleotides (printed on Corning epoxide slides), were obtained from the NCI Microarray Facility, Frederick, MD, U.S.A. Each slide in these oligonucleotide arrays has 48 blocks containing 28 rows and 28 columns, each with 36960 oligonucleotide spots with a spacing of 155 μm.
Slides were pre-hybridized for 1 h at 42 °C with 40 μl of pre-hybridization buffer [5×SSC (1×SSC is 0.15 M NaCl/0.015 M sodium citrate), 1% BSA and 0.1% SDS]. Pre-hybridization solution was removed by plunging the slides, first into deionized water and then into propan-2-ol, for 2 min each. The slides were air-dried prior to hybridization. For hybridization, the Cy3- and Cy5-labelled cDNAs were combined and mixed with 1 μl of COT-1 DNA, preheated at 100 °C for 1 min to denature the targets and snap cooled on ice. This mixture was added to 20 μl of 2× F-hybridization buffer (50% formamide, 10× SSC and 0.2% SDS) and pre-warmed at 42 °C. The total cDNA/hybridization solution mixture was loaded on to each pre-hybridized slide and covered with an M Series Lifterslip (Erie Scientific). The slides were placed in hybridization chambers and incubated overnight at 42 °C. The humidity in each chamber was maintained by the addition of 20 μl of 3×SSC solution. Post-hybridization washing included 5 min in 2×SSC+0.1% SDS, 5 min in 1×SSC and 5 min in 0.2×SSC, after which the slides were dried by centrifugation (44 g for 5 min at 22 °C).
Data processing and analysis
Microarray slides were scanned for each fluoroprobe at 10 μm using a Genepix® 4000B scanner and analysed with GenePix Pro 3.0 software (Axon Instruments). Scanned images were exported as Tiff files to GenePix Pro 3.0 software for analysis. For data analysis, data files (in gpr format) and images (in jpeg format) were imported into the microarray database (mAdb) and analysed using software tools provided by the National Cancer Institute, Center for Cancer Research in collaboration with the National Institutes of Health, Center for Information Technology, Bioinformatics and Molecular Analysis Section. Transcripts whose expression level varied at least 2-fold in ΔTrsp mice as compared with Trsp mice in more than 50% of the experiments with a P value≤0.05 were selected and the corresponding transcript levels were then analysed in A34, G37L and G37H transgenic mice relative to Trsp mice. A hierarchical clustering analysis was performed on genes in the resultant analysis. Grouping of genes into different biological functions was performed using the David database (http://david.abcc.ncifcrf.gov) and/or the mAdb software.
Q-PCR (quantitative real-time PCR)
Two-step Q-PCR was performed to validate the relative expression of genes, using the primer sequences outlined in Table 2. For each sample, 2 μg of total RNA was reverse transcribed to synthesize first strand cDNA using SuperScript II RT enzyme and random primers. The resulting cDNA was diluted, and in combination with 500 nM of each primer, iQ™ SYBR green supermix (Bio-Rad Laboratories) and DNA Engine Opticon® 2 Real-Time PCR Detection System (MJ Research), used for transcript quantification. The PCR reaction had an initial denaturation of 5 min at 95 °C, followed by 40 cycles consisting of 20 s at 94 °C, 20 s at 55 °C and 30 s at 72 °C. The reactions were carried out in triplicate and the specificity of the primers was verified by melting curve analysis. RNA levels were normalized to β-glucuronidase (Gusb) and expression levels were compared with those of wild-type mice.
Protein extracts prepared from liver of Trsp, ΔTrsp, A34, G37L and G37H mice were electrophoresed on 10% polyacrylamide gels, transferred on to PVDF membranes and immunoblotted with antibodies against GSTA (1:10000 dilution), GSTM (1:10000 dilution), EPHX1 (1:10000 dilution), HMOX1 (1:500 dilution), AOX1 (1:250 dilution) and β-actin (1:1000 dilution). Anti-goat HRP-conjugated secondary antibody (1:40000) was used for GSTA, GSTM, EPHX1 and β-actin, whereas anti-rabbit HRP-conjugated secondary antibody (1:25000) was used for HMOX1, and anti-mouse HRP-conjugated secondary antibody (1:30000) was used for AOX1. Following the attachment of the secondary antibody, membranes were washed with TBS (Tris-buffered saline, 20 mM Tris/HCl, pH 7.5 and 150 mM NaCl) containing 0.1% Tween 20, incubated in SuperSignal West Dura Extended Duration Substrate and exposed to X-ray film.
Gene expression profile in liver of ΔTrsp mice
The overall gene expression profile associated with the conditional knockout of Trsp in mouse liver (ΔTrsp) and in liver of mice following the selective replacement of selenoproteins with mutated Trsp transgenes was examined (Figure 1). An analysis of gene expression in livers from Trsp, ΔTrsp, A34, G37L and G37H mice showed that the loss of Trsp was associated with altered levels of some mRNAs, reflected through changes in gene expression and/or mRNA stability. Initially, gene expression in ΔTrsp mice was compared with that in Trsp mice and genes displaying a greater than 2-fold change in the microarray analysis with a P value≤0.05 were selected. These genes were then segregated as up-regulated (Table 3) or down-regulated (Table 4) and ordered by their pattern of gene expression by hierarchical clustering (Figure 1). Transcripts up-regulated in ΔTrsp mice are shown in Figure 1(A) and those down-regulated are shown in Figure 1(B), along with the relative expression of these transcripts in A34, G37L and G37H transgenic mice. Following filtering, genes up-regulated in ΔTrsp mice were grouped under six major hierarchical clusters, whereas those down-regulated were grouped into five major hierarchical clusters.
Genes elevated or repressed in ΔTrsp mice in comparison with Trsp mice were grouped according to their functions, whereas the transcript levels of corresponding genes in A34, G37L and G37H transgenic mice were analysed relative to Trsp and are represented in Tables 3 and 4. Genes elevated in ΔTrsp (Table 3) were mainly involved in detoxification, stress response, xenobiotic metabolism, intracellular communication, cellular transport and cell growth and differentiation. Some genes that were significantly up-regulated in ΔTrsp mice are involved in detoxification and xenobiotic metabolism and include epoxide hydrolase 1 (Ephx1), carboxylesterase 1 and 2 (Ces1, Ces2), cytochrome P450, family 2, subfamily a, polypeptide 5 (Cyp2a5), members of the glutathione transferase family (Gst), haem oxygenase 1 (Hmox1) and aldehyde oxidase 1 (Aox1). Interestingly, the levels of expression of these genes in A34, G37L and G37H mice were similar to Trsp.
Genes that were repressed in ΔTrsp mice compared with Trsp mice were grouped in a similar manner to those manifesting enhanced expression and their relative transcript levels were measured in transgenic mice (Table 4). Although most of the genes down-regulated in ΔTrsp mice had diverse or unknown functions, some of them could be grouped as being involved in transcription, intracellular communication and cellular transport.
Q-PCR validation of elevated genes
The expression levels of 22 genes elevated in ΔTrsp mice were verified by Q-PCR (Figure 2) and were in excellent agreement with the microarray analysis. Expression of the corresponding transcripts in transgenic mice was similar to that in Trsp mice. The genes analysed by Q-PCR were grouped according to their function (Table 3) as metabolism (Figure 2A), defence stress and detoxification (Figure 2B), intracellular communication/signal transduction (Figure 2C) and cell cycle/growth and differentiation (Figure 2D).
Expression of phase II enzymes
Protein expression profiles from the five mouse lines appeared similar in liver samples, as observed on Coomassie-Blue-stained gels, with the exception of a prominently enriched band of approx. 25 kDa in ΔTrsp mice (Figure 3A, indicated by the arrow). This observation was also noted in a previous study and the elevated band was sequenced and identified as GST . As expected, the mRNA levels of the Gst isoforms were also increased (Figure 2). To verify that the induced mRNA levels also gave rise to a consequential increase in the corresponding protein levels, we analysed the amounts of several phase II enzymes by Western blotting (Figure 3B). Indeed, a marked increase in two of the GST isoforms, GSTA and GSTM, was observed in ΔTrsp mice as compared with Trsp. Furthermore, several other Phase II proteins, EPHX1, HMOX1 and AOX1, were increased in ΔTrsp mice compared with Trsp. The increase in the amounts of these proteins in ΔTrsp mice paralleled their induced mRNA levels. Most interestingly, the protein levels of these enzymes in the transgenic mice were virtually the same as in Trsp mice, providing strong evidence that the link between enhanced Phase II protein expression and loss of selenoprotein expression is due to the absence of one or more housekeeping selenoproteins. These observations are further considered below. β-Actin was examined by Western blotting as a control protein and its level was unaffected in the five mouse lines (see lowest panel in Figure 3B).
Biochemical and in silico studies have identified 25 selenoprotein genes in humans and 24 in mice [3,14]. The functions of many of these selenoproteins have not been identified and most that have been characterized serve as oxidoreductases associated with various metabolic pathways, e.g. free radical scavenging, maintenance of intracellular redox status and repair of oxidized methionine residues [15,16]. Our previous studies have shown that selective knockout of Trsp in mouse hepatocytes resulted in the virtual absence of selenoproteins in liver and a pronounced reduction in selenium levels, even though the low molecular mass selenocompounds were little affected . These results demonstrated that selenoproteins are essential for proper liver function and their absence causes severe necrosis and hepatocellular degeneration, accompanied by necrosis of peritoneal and retroperitoneal fat . Subsequently, we replaced the selenoprotein population in this knockout mouse with either one of two mutant transgenes that produce tRNA gene products lacking i6A and Um34, or mcm5U and Um34 respectively, demonstrating that, although most of the selenoproteins were absent or diminished in the knockout mice, some were selectively replaced in the transgenic mice . These replaced selenoproteins were housekeeping selenoproteins which are essential for liver function . To assess the consequences of selenoprotein loss in ΔTrsp mice and their subsequent partial replacement with mutant transgenes, we examined gene expression in Trsp, ΔTrsp, A34, G37L and G37H transgenic mice by microarray analysis. These analyses showed an elevated expression of several members of the phase II enzyme family in ΔTrsp mice. This change was validated through Q-PCR and Western blotting of the corresponding proteins. Several major phase II response genes that were up-regulated in ΔTrsp mice included Gsta1, Gsta2, Gsta4, Gstm1, Gstm2, Gstm3, Cyp2a5, Ephx1, Hmox1 and Aox1.
Phase II enzymes conjugate xenobiotics or Phase I products to small donor molecules, such as glutathione, making them water soluble and easily excretable from the body, thus assisting in chemoprotection and detoxification . They can be induced in animals by (i) chemical compounds which can react with a sulfhydryl group; (ii) regulation of common promoter elements [e.g. ARE (antioxidant responsive element)]; and (iii) reactions leading to catalysis of electrophiles and ROS (reactive oxygen species) (reviewed in ). Induction of phase II enzymes in tissues has been shown to protect against carcinogens . GST isoenzymes conjugate electrophilic compounds to glutathione, thus preventing their interaction with DNA , whereas EPHX1 is a bifunctional protein, that metabolizes polycyclic aromatic hydrocarbons  and mediates sodium-dependent uptake of bile acids . HMOX1 is a cytoprotective enzyme, which degrades haem to biliverdin, which is further reduced to bilirubin , with both biliverdin and bilirubin acting as antioxidants . CYP2A5 metabolizes toxic xenobiotic compounds, such as nitrosamines and aflatoxins [24,25], takes part in the degradation of bilirubin  and is induced during hepatic pathogenesis . AOX1 is a molybdenum-containing flavoprotein which plays an important role in ethanol-induced hepatic lipoperoxidation . The expression of AOX1 may determine the susceptibility of liver cells to some pharmacological agents and the levels of ROS produced under certain pathophysiological conditions . The effect of dietary selenium on hepatic chemoprotective enzymes or xenobiotic enzymes in rodents have been extensively studied over the last few decades and results indicated a role of this element in the regulation of several phase II enzymes, including GST [30,31], EPHX  and HMOX1 . Deficiency of selenium has been associated with an increase in these enzymes in rodents [30–33]. Our results indicate a similar elevation in the levels of phase II enzymes in Trsp knockout mice, suggesting this phenomenon to be a result of loss of selenoproteins rather than a reduction in dietary selenium. Interestingly, the levels of phase II response genes were normal when housekeeping selenoproteins were replaced in transgenic mice, providing strong evidence of their up-regulation being a consequence of deficiency in housekeeping selenoproteins. In contrast, reduced expression of stress-related selenoproteins, such as GPx1 or SELR, had no role in the up-regulation of phase II enzymes. An earlier study reported that inhibition of TR by aurothioglucose leads to induction of hepatic HMOX1 activity . These investigators postulated that the lack of TR, or a TR-related reaction, induces hepatic HMOX1. GPx and GST are both responsible for detoxification of xenobiotic electrophiles by the addition of reduced glutathione (GSH) and possess similar enzyme folds in the GSH-binding site. Earlier studies have demonstrated that the Alpha-class GST isoenzymes also exhibit selenium-independent GPx activity in rodents [35,36], and these isoenzymes are very effective at reducing hydroperoxides, thus providing protection against membrane lipid peroxidation . In the liver of ΔTrsp mice, the elevated levels of GST might functionally compensate for GPx and/or another selenoprotein(s) that might also be involved in detoxification.
The present study shows that an interplay exists between the loss of one or more housekeeping selenoproteins and enrichment in members of the phase II response protein class. The fact that several members of the phase II protein class manifesting a wide variety of functions are up-regulated suggests that several members of the housekeeping selenoprotein class are likely to be involved in this interplay. Thus our results provide strong evidence of a functional link between housekeeping selenoproteins and phase II enzymes.
This work was supported by the Intramural Research Program of the Center for Cancer Research, National Cancer Institute, National Institutes of Health (D. L. H.), and by National Institutes of Health grants (V. N. G.).
Abbreviations: AOX, aldehyde oxidase; EPHX, epoxide hydrolase; GPx, glutathione peroxidase; GST, glutathione transferase; GSTA, GST Alpha; GSTM, GST Mu; HMOX, haem oxygenase; HRP, horseradish peroxidase; i6A, N6-isopentenyladenosine; mcm5U, 5′-methylcarboxylmethyluracil; mcm5Um, 5′-methylcarboxymethyl-2′-O-methyluridine; Q-PCR, quantitative real-time PCR; ROS, reactive oxygen species; RT, reverse transcriptase; Sec, selenocysteine; TR, thioredoxin reductase; Um34, 2′-O-methylribose at position 34 in selenocysteine tRNA
- © The Authors Journal compilation © 2008 Biochemical Society