Expression of Selenoproteins Is Maintained in Mice Carrying Mutations in SECp43, the tRNA Selenocysteine 1 Associated Protein (Trnau1ap)

Selenocysteine tRNA 1 associated protein (Trnau1ap) has been characterized as a tRNA[Ser]Sec-binding protein of 43 kDa, hence initially named SECp43. Previous studies reported its presence in complexes containing tRNA[Ser]Sec implying a role of SECp43 as a co-factor in selenoprotein expression. We generated two conditionally mutant mouse models targeting exons 3+4 and exons 7+8 eliminating parts of the first RNA recognition motif or of the tyrosine-rich domain, respectively. Constitutive inactivation of exons 3+4 of SECp43 apparently did not affect the mice or selenoprotein expression in several organs. Constitutive deletion of exons 7+8 was embryonic lethal. We therefore generated hepatocyte-specific Secp43 knockout mice and characterized selenoprotein expression in livers of mutant mice. We found no significant changes in the levels of 75Se-labelled hepatic proteins, selenoprotein levels as determined by Western blot analysis, enzymatic activity or selenoprotein mRNA abundance. The methylation pattern of tRNA[Ser]Sec remained unchanged. Truncated Secp43 Δ7,8mRNA increased in Secp43-mutant livers suggesting auto-regulation of Secp43 mRNA abundance. We found no signs of liver damage in Secp433-mutant mice, but neuron-specific deletion of exons 7+8 impaired motor performance, while not affecting cerebral selenoprotein expression or cerebellar development. These findings suggest that the targeted domains in the SECp43 protein are not essential for selenoprotein biosynthesis in hepatocytes and neurons. Whether the remaining second RNA recognition motif plays a role in selenoprotein biosynthesis and which other cellular process depends on SECp43 remains to be determined.


Introduction
Selenoproteins are proteins containing selenocysteine (Sec), the 21 st proteinogenic amino acid. Incorporation of Sec into protein requires the recoding of UGA codons in both eukaryotes and prokaryotes, although the details differ among eubacteria, archeae, and eukaryotes [1,2]. Selenoproteins occur in all three of the life kingdoms, Bacteria, Archaea and Eukaryota. Another expansion to the genetic code is pyrrolysine, the 22 nd proteinogenic amino acid, which has been found in only several archaea and eubacteria, including one human pathogen, and is incorporated into protein by recoding of UAG [3]. How exactly the expansion of the genetic code occurred by use of termination codons is an interesting phenomenon, because in both instances UGA and UAG also retained their functions as termination codons [4]. For example, in the case of Sec insertion into protein, the selenoprotein biosynthetic machinery encodes Sec insertion sequence (SECIS) elements, which are cis-acting RNA motifs, that occur immediately following the Sec UGA in eubacteria, or are positioned in the 3'-untranslated region in eukaryotes and archeae of the mRNA, to distinguish elongation from termination function [5]. SECIS element binding proteins interact directly or indirectly with the translation machinery dictating the insertion of Sec into the growing polypeptide chain, e.g., SECIS binding protein 2 (Secisbp2) in mammals. Sec-tRNA [Ser]Sec interacts with a specific translation elongation factor, designated SelB in bacteria and EF-Sec in archaea and eukaryotes, but not with the EF-Tu or EF-1α. The biosynthetic pathway of Sec initiates with the attachment of serine to tRNA [Ser]Sec by seryl-tRNA synthetase (Serrs) to form Ser-tRNA [Ser]Sec , which is further phosphorylated by the phosphoserine tRNA kinase (Pstk). Biosynthesis of Sec-tRNA [Ser]Sec is completed by selenophosphate synthetase 2 (Sephs2), a selenoprotein in many higher animals, that synthesizes the active donor of selenium, selenophosphate, and Sec synthase. This enzyme is designated SelA in bacteria [6], but the eukaryotic protein received several names, soluble liver antigen (SLA) [7], SecS [8], and Sepsecs [9], the latter being adopted as the systematic gene name. A hierarchy of Sec insertion has been described among selenoproteins which may correlate with the affinity of Secisbp2 to SECIS elements [10,11] and/or with the modification status of tRNA [Ser]Sec (reviewed in [12]). Transfer RNA [Ser]Sec exists in higher animals in two isoforms which differ by the occurrence of a 2'-O-hydroxymethyl group (Um34) on the ribosyl moiety of the highly modified base 34, methylcarbonylmethyl-uridine (mcmU 34 ). In the presence of amply available selenium (Se), mcmU 34 is further modified to Um34 [1]. Another modification which affects translation efficiency and possibly tRNA-protein interactions is isopentenylation of A 37 [13].
SECp43 was initially cloned in a degenerate PCR screen for RNA-binding proteins [14]. The protein harbors two RNA recognition motifs (RRM) and a C-terminal Tyr-rich domain of unknown function. Affinity purification of the native protein with an antiserum generated against recombinant protein co-purified a 90 nt RNA which was subsequently identified by direct RNA sequencing as tRNA [Ser]Sec [14]. It was thus proposed that SECp43 has a role in selenoprotein biosynthesis.
SECp43 further interacted with a 48 kDa protein [7], which was subsequently identified as Sepsecs. Knockdown of SECp43 and Sepsecs reduced selenoprotein expression in cultured NIH 3T3 cells [15]. Protein levels of glutathione peroxidase 1 (Gpx1) were reduced as measured by Western blot analysis and the fraction of methylcarbonylmethyl-uridine-Um34 (mcmU 34 m) tRNA [Ser]Sec compared to mcmU 34 tRNA [Ser]Sec was reduced in NIH 3T3 fibroblasts [15]. Finally, co-immunoprecipitation indicated binding of SECp43 to Sepsecs [15]. The interaction of SECp43 with purified, fully modified [ 75 Se]-Sec-tRNA [Ser]Sec bound to recombinant EF-Sec was examined [16]. Again, SECp43 did not interact with Sec tRNA [Ser]Sec , but reduced the electrophoretic mobility of a complex containing [ 75 Se]-Sec-tRNA [Ser]Sec and EF-Sec, albeit independent of GTP [16]. Based on the clear effects of SECp43 knockdown on selenoprotein expression in cultured cells, we decided to test the role of SECp43 in selenoprotein expression in vivo by gene targeting. Due to the organization of the genomic locus, two mutants were generated which led in both cases to in-frame deletions in the protein. Whereas constitutive deletion of exons 3 and 4 neither affected selenoprotein expression nor embryonic development, constitutive deletion of exons 7 and 8 was embryonic lethal. Because inactivation of the tRNA [Ser]Sec gene (Trsp) [17] or Secisbp2 [18] in mice likewise is embryonic lethal, we conditionally inactivated SECp43 in hepatocytes and neurons and extensively analyzed hepatic and cerebral selenoprotein expression.

Gene targeting and mice
The targeting vector for SECp43 Δ3,4 was made by inserting the genomic fragment spanning exons 3 and 4 (1.2 kb) between two loxP sites using Afl II and Mfe I restriction sites, followed by a loxP-and FRT-flanked neomycin phosphotransferase expression cassette. A 3.1 kb of 5'homologous sequences upstream and a 6.5 kb of 3'-homologous sequence downstream of the exons 3 and 4 were inserted before and after the loxP sites in the vector to guide the homologous integration (S1 Fig). Constitutive deletion of exons 3 and 4 (Secp43 Δ3,4 ) by crossing with EIIA-Cre mice did not affect the phenotype of mice in any recognizable way and did not reduce 75 Se incorporation in liver, kidney, testes, lung, heart, brain and plasma (S2 Fig). This mouse line was not further analyzed.
The targeting vector for Secp43 Δ7,8 was prepared using the same backbone vector as the Secp43 Δ3,4 knockout and contained a loxP flanked genomic fragment spanning exons 7 and 8 and adjacent intronic sequence followed by a loxP-and FRT-flanked neomycin phosphotransferase expression cassette. A 5.5 kb of 5'-homologous sequences upstream and a 6.4 kb of 3'homologous sequence downstream of the exons 7 and 8 were inserted before and after the loxP sites in the vector (Fig 1A) to guide the homologous integration. The linearized targeting vectors were electroporated into embryonic stem cells. Correct gene targeting was ascertained by Southern blotting with external probes as indicated in Fig 1A. The modified allele was transmitted through the germline and the selection cassette deleted by FLPe-mediated recombination. The resulting loxP-flanked conditional allele is designated Secp43 fl , the recombined deleted allele is designated Secp43 Δ7, 8 . The mice were backcrossed into a C57Bl/6 genetic background over several generations. Genotyping was done by PCR using the following primers, flSecp43fwd, 5'-catgtgggtcagggatcttc-3', and flSecp43rev, 5'-caaatgctaatcaaaagatgtca-3', which spanned the loxP-element in intron 8. The wild-type and floxed alleles yield products of 310 bp and 410 bp, respectively. Constitutive knockouts were produced using EIIA-Cre [19]. Hepatocyte-specific and neuron-specific knockouts were produced with Alb-Cre (albumin promoter-Cre) and Tα1-Cre (tubulin α1promoter-Cre) mice as described [18,20]. Mice were fed breeding diet (Ssniff, Soest, Germany) containing on average 0.2-0.3 ppm Se. 75 Se (specific activitỹ 1000 Ci/mmol) was obtained from the University of Missouri Research Reactor Center (Columbia, MO, USA) and metabolic labeling was performed as described [21]. Mice were handled in accordance with the National Institutes of Health Institutional Guidelines (NCI, NIH, Bethesda, MD, USA), and all mouse experiments were approved by the Animal Ethics Committee at the National Institutes of Health. Mutant mouse lines and other renewable research material are freely available to non-profit researchers upon request to the authors. The ARRIVE statement is found in S1 Text.

Selenium determination in serum
Se was determined by total reflection X-ray fluorescence (TXRF) using a Picofox S2 instrument (Bruker nano, Berlin, Germany) [22]. Gallium was used as an internal standard for quantification. Reference samples for serum (Sero, Billingstad, Norway) were included in all analyses and results were always within the reference range. All samples were measured in duplicate.

Liver transaminase determination
Alanine-aminotransferase (ALAT, GPT) and aspartate-aminotransferase (ASAT, GOT) measurements in serum were performed according to standard coupled tests involving 2-ketoglutarate and NADH plus alanine and lactate dehydrogenase (for ALAT) or aspartate and malate dehydrogenase (for ASAT). The decrease in A 340 was followed over three minutes and the activity was calculated as 1 unit = 1μmol/min from the slope of the linear part of the curve [18].

Selenoenzyme measurements
Gpx assays were carried out with tert-butyl hydroperoxide and thioredoxin reductase (Txnrd) activities were assessed with the DTNB assay in cytosolic fractions of tissue homogenates as described [23]. Protein concentrations were determined by the method of Bradford using IgG as a standard. Type I-deiodinase (Dio1) activity was determined as described with 125 I-rT 3 as substrate [24]. 15 μg of membrane fraction protein was used with 1 μM unlabeled rT 3 . The reaction time was 60 min and the reaction temperature 37°C. The assay was done in triplicate and repeated one additional time with similar results.

Antibodies and Western blot analysis
The antiserum directed towards Sepp (1:400) was generated in rabbits by immunization with a synthetic C-terminal peptide (ImmunoGlobe, Himmelstadt, Germany) and its specificity verified using wild-type and Sepp-knockout mice [22,25]. For serum Sepp quantification, 0.2 μl serum was applied per lane. 25-100 μg protein from the cytosolic fraction were separated on SDS/12% polyacrylamide gels. Antibodies against Sepk and Seps were from Sigma (ATLAS Prestige antibodies, rabbit polyclonal) and used at 1:500, 1:1000 dilution, respectively. Rabbit polyclonal antibodies against Gpx1 and Gpx4 were both from Abcam (Cambridge, UK) and were used at 1:2000 and 1:1000, respectively. The polyclonal rabbit Sepsecs antiserum (Sigma, München, Germany) was used at 1:1000 dilution. Chicken polyclonal antibodies directed towards SECp43 were kindly provided by Dr. Paula Grabowski (University of Pittsburgh) and used at a 1:1000 dilution. After electrotransfer, nitrocellulose membranes were stained with Ponceau Red, photographed and blocked with 5% BSA for 1 hour at 25°C. Rabbit polyclonal βactin antiserum (Sigma, Munich, Germany) was used as a loading control at 1:2000 dilution.

qPCR analysis
Total RNA was isolated from powdered mouse liver according to the TRIzol protocol (Invitrogen, Darmstadt, Germany). Samples were treated with RQ1 RNase-Free DNase (Promega, Madison, USA). cDNA was synthesized using the iScript cDNA synthesis kit (Biorad, München, Germany) according to the manufacturer's protocol. qPCR was performed using SYBR Green from Abgene (Thermo Scientific, Epsom, UK) on a Mastercyler epgradient S realplex (Eppendorf, Germany). Primers used for qPCR detection of selenoproteins genes were as described previously [18].

Isolation, fractionation and quantitation of tRNA
Total tRNA was isolated from mouse liver of control and Alb-Cre; SecP43 fl/fl mice, and the amounts of tRNA Sec were quantified following TBE-Urea gel electrophoresis and hybridization with a 32 P-labelled probe by Northern blot analysis using a PhosphorImager (Molecular Dynamics) [19]. Total tRNA from control and Alb-Cre; SecP43 fl/fl mice was aminoacylated with [ 14 C]serine with [ 3 H]serine, respectively, in the presence of rabbit reticulocyte synthetases. The resulting aminoacylated tRNA was fractionated on an RPC-5 column [26], first in the absence and then in the presence of Mg 2+ as given [19,21].

Immunohistochemistry
Mouse brains were immersion fixed in 4% paraformaldehyde in 0.1 M phosphate buffer (PB), pH 7.4. Over night fixation of the brains at 4°C was followed by cryoprotection in 30% sucrose in 0.1 M PB for 2 days, and the brains were frozen at -80°C. Sections were cut at 20 μm on a cryostat. Free floating sections were stained with polyclonal rabbit α-calbindin (Swant, Bellinzona, Switzerland) at a dilution of 1:10.000 and developed with horseradish peroxidase-anti rabbit conjugate [20]. Additional methods are found in S2 Text.

Statistics
For all computations, GraphPad Prism 6 for Mac OS X software was used for the tests indicated in the figure legends. Data are expressed as means ± standard error of the mean (S.E.M.). Statistical significance was determined and indicated as ÃÃ p < 0.01, or ÃÃÃ p < 0.001.

Genomic manipulation of the Tranu1ap locus in mice
The gene encoding SECp43, Trnau1ap, consists of 9 exons and is located on chromosome 1. The many small exons and their arrangement made it difficult to choose exons for gene targeting. We selected two conditional constructs, one of which deleted exons 3 and 4 eliminating portions of the first RRM (Secp43 Δ3,4 ), and the other that deleted exons 7 and 8 eliminating a portion of the Tyr-rich domain (Secp43 Δ7,8 ) ( Fig 1A). Both constructs had the drawback that the deletions were in-frame.
Constitutive Secp43 Δ3,4/ Δ3,4 mice showed no apparent phenotype and accordingly no change in selenoprotein expression. Construction and characterization of these mice are described in S1, S2 and S3 Figs. A Western blot using antiserum against SECp43 revealed a smaller band of SECp43 that was highly expressed in liver (S3 Fig) and likely corresponded to the product of an mRNA lacking exons 2 and 3.
To target the Tyr-rich domain, we introduced loxP elements flanking exon 7 and 8 leading to a Cre-responsive conditional allele, designated Secp43 fl (Fig 1A). Germline deletion of both exons resulted in embryonic lethality as observed among 192 offspring from heterozygous matings (Secp43 +/ Δ7, 8 x Secp43 +/ Δ7,8 ), wherein 69 Secp43 +/+ and 123 Secp43 +/ Δ7,8 mice were found, but none with a Secp43 Δ7,8 / Δ7,8 genotype. The numbers are compatible with Mendelian inheritance and embryonic lethality. Since we expected an essential function of SECp43 for selenoprotein expression and since Trsp-and Secisbp2-deficient mice also die early during embryonic development, we did not determine the exact phenotype of the homozygous knockout embryos, but proceeded to conditional inactivation of SECp43 in liver.
Hepatocyte-specific inactivation of Secp43 was achieved by Cre-mediated recombination using Alb-Cre transgenic mice. Western blot analysis of SecP43 revealed the loss of SECp43 in liver of Alb-Cre; Secp43 fl/fl mice, but not in kidney or testes, which were used as control tissues (S4 Fig). Quantitative deletion of exons 7+8 in Secp43 was demonstrated by qPCR using primers located in exons 7 and 8. Virtually no transcript containing exons 7 and 8 was detected in Alb-Cre; Secp43 fl/fl liver (Fig 1B). In contrast, when using primers located in exons 3 and 4, Secp43 Δ7,8 , mRNA levels were apparently increased in mutant liver (see below).
We then tested whether liver damage occurred in Secp43-mutant mice. Liver transaminase activities were determined in plasma from adult mice, but no significant increase in transaminase activity was noted in Secp43-mutant mice (Table 1). This result suggested that the mutation did not impair the integrity of the liver. Incorporation of Se into selenoproteins can be easily visualized by metabolic labeling of mice with 75 Se-selenite. Fig 2A shows an autoradiogram of liver, kidney, testis, and plasma proteins resolved by SDS-PAGE. Samples from a wild-type litter mate and an Alb-Cre; Secp43 fl/fl mouse were compared. 75 Se-labeling of selenoproteins in liver of Secp43-mutant and control mice was virtually identical. This finding indicated that at least the most abundant selenoproteins (Txnrd1, Gpx1, Gpx4, Dio1, and other, not unequivocally assigned, selenoproteins) incorporated normal amounts of selenium in Secp43 knockout hepatocytes. Plasma selenoproteins, selenoprotein P (Sepp) and Gpx3, incorporated similar amounts of 75 Se in both control and Secp43 mutant mice. An apparent size difference in Sepp was noted in an initial experiment between control and Secp43 mutant mice, but the difference was not apparent by Western blotting from additional mice that were further backcrossed into the C57Bl/6 genetic background ( Fig 2B). Accordingly, plasma selenium levels were not different between Alb-Cre; Secp43 fl/fl mice and their littermate controls (Table 2).

Hepatic selenoprotein expression in Secp43 Δ7,8 mutant mice
Sepp is the main plasma selenium carrier protein predominantly made by hepatocytes [25]. Biosynthesis of nascent Sepp polypeptide and increasingly glycosylated Sepp were not changed in livers from Secp43-deficient mutant as assessed by Western blotting (not shown). Expression of selenoproteins which can be assigned in 75 Se autoradiographs (Gpx1 and Gpx4) were also assessed by Western blot analysis. Again, the amounts of Gpx1 and Gpx4 were not reduced in Alb-Cre; Secp43 fl/fl liver. Likewise, levels of selenoproteins K and S (Sepk and Seps, respectively) were not changed (Fig 3A). Taken together, these results indicate that selenoprotein expression in hepatocytes was not impaired by loss of the essential Tyr-rich domain in SECp43.
We then tested enzymatic activities of selenoenzymes which can be readily assessed. Again, no differences were observed in the activities of cytosolic thioredoxin reductase (Txnrd), cytosolic glutathione peroxidase, and deiodinase 1 (Dio1) in livers from Secp43-mutant mice ( Fig  3B-3D).

mRNA levels of hepatic selenoproteins and selenoprotein biosynthesis factors
We speculated that even if SECp43 may not be an essential component of selenoprotein expression in liver, this protein may still gradually facilitate selenoprotein expression, and loss of SECp43 may be compensated by alterations in selenoprotein mRNA levels. We therefore assessed expression of several liver selenoprotein mRNAs by qPCR from Alb-Cre; Secp43 fl/fl and litter mate control mice. We did not detect any significant differences in mRNA levels encompassing both male and female mice (Fig 4A).
However, SECp43 has also been reported to interact with Sepsecs [14,15]. Thus, the deficiency in SECp43 may be compensated by an increased expression of Sepsecs and possibly  (5), male mutants (4), female controls (7), and female mutants (6). Although the mean plasma transaminase values appeared to increase, the differences were not significant (Student's t-test) and only values above 50 U/L would be regarded pathological. other proteins involved in selenoprotein biosynthesis. We therefore examined mRNA levels of Serrs, Pstk, Sepsecs, and Sephs2, which also is a selenoprotein. Very similar mRNA levels were observed for these enzymes (Fig 4B). Moreover, RNA-binding proteins which are known as positive and negative regulators of selenocysteine insertion, Secisbp2 and eIF4a3, respectively, were not changed (Fig 4B). Because of the reported interaction of SECp43 and Sepsecs, we tested Sepsecs expression by Western blotting. This analysis did not reveal any appreciable change in expression of this protein in Secp43 mutant liver (Fig 4C).

Methylation status and abundance of tRNA [Ser]Sec in liver
The ratio of mcmU 34 m to mcmU 34 tRNA [Ser]Sec isoforms responds to Se availability. The methylase involved has not been identified, but siRNA-mediated knockdown of Secp43 transcript has been shown to lower the fraction of mcmU 34 m in cultured cells [15]. We therefore assessed tRNA [Ser]Sec methylation patterns in livers from Alb-Cre; Secp43 fl/fl and control mice. The mcmU 34 m isoform is slightly more hydrophobic than the non-Um34 isoform and elutes slightly later from the reverse phase chromatographic column and a reduction of mcmU 34 m would be apparent as a reduction in level at the right-hand flank of the double peak. Virtually no change in the methylation pattern was observed in seryl-tRNA [Ser]Sec from Secp43-mutant compared to that from the corresponding control seryl-tRNA [Ser]Sec (Fig 5A). We also quantified the amount of tRNA [Ser]Sec in liver by Northern blotting and found that the levels of total tRNA [Ser]Sec were very similar in the livers of the two mouse lines (Fig 5B).

Neuron-specific Secp43 mutation
The lack of effect on selenoprotein expression in Secp43-mutant hepatocytes prompted us to ask whether this observation is specific for hepatocytes, whereas other cell types may require SECp43. In the brain, SECp43 expression starts during development and is maintained in mature neurons [28]. We therefore conditionally mutated SECp43 in neurons using Tα1-Cre transgenic mice which have been used before in our laboratory to inactivate tRNA [Ser]Sec [20,29]. Tα1-Cre; Secp43 fl/fl mice are born at the expected frequency and show no apparent phenotype. Efficient deletion of exons 7 and 8 in brains of mutant mice was verified by qPCR using primers specific for exons 7 and 8 ( Fig 6A). Selenoprotein expression was assessed by Western blotting in brain tissue from Tα1-Cre; Secp43 fl/fl and control mice (Fig 6B). We found that the expression of neuronal selenoproteins in Secp43-mutant mice compared to the corresponding control mice were very similar. Movement co-ordination is often impaired in selenoprotein-deficient mice [27,30,31]. The performance on the rotarod of Tα1-Cre; Secp43 fl/fl mice was significantly reduced compared to normal control animals indicating that the deletion of the Tyr-rich domain in SECp43 has indeed functional consequences (Fig 6C). Cerebellar development is sensitive to disruption of neuronal selenoprotein expression [20,32]. However, cerebellar foliation and calbindin expression appeared normal in Tα1-Cre; Secp43 fl/fl animals (Fig 6D).
Selenoprotein expression is, therefore, independent of the Secp43 mutation eliminating the Tyr-rich domain in neurons as well as in hepatocytes by the criteria used herein.  (5), female controls (7), and female mutants (6). No significant differences were found (Student's t-test). doi:10.1371/journal.pone.0127349.t002

Discussion
Evidence for a role of SECp43 on selenoprotein expression SECp43 co-purified with tRNA [Ser]Sec and was reported to interact with a 48 kDa protein which is now believed to represent Sepsecs [14]. It is hard to construct a case in which a lowabundance tRNA of 90 nt, when directly sequenced from the eluate of an affinity column loaded with SECp43, could be misidentified as tRNA [Ser]Sec because of a chance error. The following observation that knockdown of Secp43 in NIH 3T3 cells changed the methylation status of tRNA [Ser]Sec and reduced the abundance of several selenoproteins was in line with the expectation that SECp43 plays an important role in selenoprotein biosynthesis [15]. Together, these observations were the motivation for conducting the present study in which we wanted to determine the role of SECp43 in selenoprotein biosynthesis in vivo, i.e. in transgenic mice.

Attempts to target Secp43
Our initial attempt to inactivate Secp43 in mice was by targeting the first RRM by deletion of exons 3 and 4. As described in the supporting information, no effect on selenoprotein biosynthesis was noted in any organ tested and constitutive Secp43 Δ3,4/ Δ3,4 mice developed normally. Deletion of exons 7 and 8 in mice did not support embryonic development. We initially concluded that embryonic lethality is consistent with defective selenoprotein expression, since Trsp-and Secisbp2-KO mice also die early during embryonic development [17,18]. Studying the role of SECp43 in embryonic development was not the aim of the present study. Our goal was to assess the role of SECp43 in selenoprotein biosynthesis, and we therefore generated liver-specific Secp43 Δ7,8 -mutant mice.

Effect of Secp43 gene targeting on hepatic selenoprotein biosynthesis
Alb-Cre; Secp43 Δ7,8 -mutant mice are a good model to study selenoprotein biosynthesis, because hepatocytes express many selenoproteins, but can tolerate the complete abrogation of selenoprotein expression [24,25]. Lack of Txnrd1 expression induces Nrf2-dependent gene activation and compensatory expression of, e.g., glutathione-S-transferases [18,33,34]. Inactivation of Nrf2 combined with inactivation of tRNA [Ser]Sec leads to liver failure [35]. It is also known that inactivation of Txnrd1 causes liver degeneration when glutathione biosynthesis is impaired [36]. No liver damage was observed in Alb-Cre; Secp43 fl/fl mice as demonstrated by normal liver transaminases in plasma. Extensive analysis of selenoprotein expression in liver and plasma did not yield any indication that selenoprotein translation is impaired in Secp43-mutant liver. Metabolic labeling of selenoproteins with 75 Se, Western blot of selenoproteins, activity measurements of selenoenzymes, plasma Se content as a marker of liver Sepp production all remained normal in Alb-Cre; Secp43 fl/fl mice. Selenoprotein mRNAs are sometimes sensitive markers of reduced selenium bioavailability, but also their levels remained normal in Secp43-mutant livers. Moreover, there was no compensatory up-regulation of Sephs2 as e.g., in Sepp-deficient mice [37]. EIF4a3, a negative regulator of selenoprotein expression is regulated by Se availability in cells [38], but its mRNA level in liver was not changed in Secp43-mutant mice. Because no negative effect on selenoprotein expression was observed, we did not attempt to study expression of Nrf2 targets. Finally, neither tRNA [Ser]Sec abundance nor its methylation status was changed in Secp43-mutant liver. In summary, we found no phenotype with respect to selenoprotein expression despite a very extensive analysis.
Why is there no effect of Secp43 mutations on selenoprotein expression?
One possibility that we did not observe an effect of Secp43 deficiency on selenoprotein biosynthesis is that the targeted mutations we introduced did not disrupt the function of SECp43. Targeting the Secp43/Trnau1ap locus is indeed difficult. The Genbank database contains a transcript lacking exons 2 and 3 (NCBI XM_006539180.1) suggesting that exon skipping occurs in vivo. S3 Fig shows that, at least in liver, a smaller form of SECp43 is detected which is indistinguishable from SECp43 Δ3,4 . Since deletions of exons 2+3, 3+4, and 7+8 maintain the reading frame, it is possible that the truncated proteins retain some of their biological functions. While exons 2-4 encode the first RRM domain, exons 7+8 encode part of the Tyr-rich domain. Our Secp43 Δ7,8 resulting from the Secp43 fl/fl alleles removes 66 amino acids from a protein of 287 amino acids. This manipulation, in turn, leaves both RRM domains intact. Given the fact that the truncated Secp43 Δ7,8 mRNA is up-regulated rather than degraded, one may argue that our gene targeting strategy did not completely inactivate SECp43, but at least interferes with autoregulation of Secp43mRNA levels. The Tyr-rich region is completely conserved among six mammals (human, chimp, dog, cattle, mouse and rat) (S5 Fig). Of note, constitutive Secp43 Δ7,8 mice are embryonic lethal clearly indicating that the Tyr-rich domain is important for SECp43 function. This notion is further supported by our observation that movement co-ordination is impaired in neuron-specific Secp43-mutant mice and by the finding of developmental expression of SECp43 in the brain [28].
A second possibility is that SECp43 may not be required for selenoprotein biosynthesis in hepatocytes, but in other cell types. This possibility is not supported by our findings in neuronspecific Secp43-mutant mice. Tα1-Cre; Secp43 fl/fl mice have normal selenoprotein expression in brain. Since cerebellar development [20], cortical interneuron development [29], and striatal interneuron development [31] depend on selenoprotein biosynthesis, and even milder disruption of cerebral selenoprotein biosynthesis [27,31,39] causes clear neurological defects, we are confident that we have not missed a mere quantitative reduction of selenoproteins in neural tissues. The animals in our study were fed Se-replete diets. While it may be argued that a potential phenotype could be unmasked by dietary Se restriction, it is nevertheless clear that SECp43 is not essential for selenoprotein biosynthesis in hepatocytes and neurons.
A third possibility is that SECp43 has nothing to do with selenoprotein expression, but has an unrelated role in cells. In a systematic study on the conservation of the selenoprotein biosynthesis pathway in insects, Chapple & Guigo showed that several insect species which have abandoned selenoprotein expression during evolution have lost selenoprotein pathway genes, e.g., selenocysteine synthase or selenophosphate synthase 2 [40]. However, all species, whether they do or do not contain selenoproteins and selenoprotein biosynthesis factors, have retained Secp43 in their genomes implying that Secp43 may serve a conserved function independent of selenoprotein expression [40]. Another indication why Secp43 may not necessarily be associated with selenoprotein expression is the finding that parasitic plant nematodes of the Tylenchina clade have lost all selenocysteine incorporation factors, but retained functional Secp43 genes [41]. Finally, in a reconstituted system for in vitro selenoprotein translation, Gupta et al. noted that their wheat germ-based system does not contain SECp43, but was functional as long as Sec-tRNA [Ser]Sec , Secisbp2, EF-Sec, and mammalian ribosomes were added [42].