RNA-dependent conversion of phosphoserine forms selenocysteine in eukaryotes and archaea

Abstract

The trace element selenium is found in proteins as selenocysteine (Sec), the 21st amino acid to participate in ribosome-mediated translation. The substrate for ribosomal protein synthesis is selenocysteinyl-tRNA^Sec. Its biosynthesis from seryl-tRNA^Sec has been established for bacteria, but the mechanism of conversion from Ser-tRNA^Sec remained unresolved for archaea and eukarya. Here, we provide evidence for a different route present in these domains of life that requires the tRNA^Sec-dependent conversion of O-phosphoserine (Sep) to Sec. In this two-step pathway, O-phosphoseryl-tRNA^Sec kinase (PSTK) converts Ser-tRNA^Sec to Sep-tRNA^Sec. This misacylated tRNA is the obligatory precursor for a Sep-tRNA:Sec-tRNA synthase (SepSecS); this protein was previously annotated as SLA/LP. The human and archaeal SepSecS genes complement in vivo an Escherichia coli Sec synthase (SelA) deletion strain. Furthermore, purified recombinant SepSecS converts Sep-tRNA^Sec into Sec-tRNA^Sec in vitro in the presence of sodium selenite and purified recombinant E. coli selenophosphate synthetase (SelD). Phylogenetic arguments suggest that Sec decoding was present in the last universal common ancestor. SepSecS and PSTK coevolved with the archaeal and eukaryotic lineages, but the history of PSTK is marked by several horizontal gene transfer events, including transfer to non-Sec-decoding Cyanobacteria and fungi.

Amino acids enter protein synthesis as aminoacyl-tRNAs (aa-tRNAs). The identity of an amino acid inserted into the nascent polypeptide is determined by two factors: the interaction of the aa-tRNA anticodon with an appropriate mRNA codon, and the correct pairing of amino acid and tRNA anticodon in the aa-tRNA. Thus, faithful protein synthesis requires the presence in the cell of a full set of correctly aminoacylated tRNAs (1). Two routes to aa-tRNA synthesis exist (2): (i) direct acylation of the amino acid onto its cognate tRNA catalyzed by an aa-tRNA synthetase, and (ii) tRNA-dependent modification of a noncognate amino acid attached to tRNA. Although the majority of aa-tRNAs are made by direct aminoacylation, the indirect pathway is also widely used, e.g., for Gln-tRNA^Gln and Asn-tRNA^Asn formation (3).

The biosynthetic route of aa-tRNA formation is understood for all of the currently known cotranslationally inserted amino acids. The sole exception is the case of selenocysteine (Sec) in eukarya and archaea (e.g., refs. 4 and 5). The bacterial case was solved in the 1990s when genetic and biochemical studies with Escherichia coli revealed that bacterial Sec-tRNA^Sec was synthesized by an indirect tRNA-dependent amino acid transformation mechanism (6). Sec is cotranslationally inserted into proteins in response to the codon UGA; thus the tRNA^Sec species has the corresponding anticodon UCA (7). This tRNA is misacylated with serine by E. coli seryl-tRNA synthetase (SerRS) to form Ser-tRNA^Sec (8). Then a pyridoxal phosphate (PLP)-dependent Sec synthase (SelA) catalyzes the Ser-tRNA^Sec → Sec-tRNA^Sec conversion using selenophosphate as selenium donor (Fig. 1Upper). For archaea and eukarya our understanding is still fragmentary. Serylation of tRNA^Sec by eukaryal (9–12) and archaeal (13–15) SerRS enzymes has been well established. However, the demonstration of an analogous reaction to the bacterial pathway converting Ser-tRNA^Sec → Sec-tRNA^Sec has remained elusive. In particular, the archaeal protein MJ0158, annotated as the ortholog to bacterial SelA, did not support in vitro Sec formation (15).

Fig. 1.

tRNA-dependent amino acid transformations leading to Sec and cysteine. (Upper) The SelA route is the bacterial pathway (6). The PSTK/SepSecS is the archaeal/eukaryal route. (Lower) The SepRS/Sep-tRNA:Cys-tRNA synthase (SepCysS) pathway operates in methanogens to synthesize cysteine (16).

Recently, a pathway for cysteine formation was discovered in methanogens based on a tRNA-dependent conversion of O-phosphoserine (Sep) to cysteine (16). This route involves the attachment of Sep to tRNA^Cys by SepRS and the subsequent Sep-tRNA^Cys → Cys-tRNA^Cys transformation by Sep-tRNA:Cys-tRNA synthase, a PLP-dependent enzyme using a still unknown sulfur donor (Fig. 1 Lower). The implications of this pathway for Sec formation were obvious (16), especially because a phosphoseryl-tRNA^Sec kinase (PSTK) had just been identified (17). Its activity was characterized in vitro for the murine (17) and the Methanocaldococcus jannaschii (15) enzyme; this kinase phosphorylates specifically Ser-tRNA^Sec to Sep-tRNA^Sec. This finding would lend credence to early reports of the existence of mammalian Sep-tRNA formed from Ser-tRNA (18, 19). Because the initial search for a PLP-dependent SelA activity as manifested by E. coli SelA was unsuccessful (15), we considered other proteins. The most promising candidate was the human protein SLA/LP (soluble liver antigen/liver pancreas), which forms a ribonucleoprotein antigenic complex with tRNA^Sec in patients suffering from an autoimmune chronic hepatitis (20). This protein was classified computationally as a PLP-dependent serine hydroxymethyltransferase that might have SelA function (21). To lend further credibility, this protein (22, 23) has very well conserved archaeal orthologs but only in Methanocaldococcus jannaschii (MJ0610), Methanococcus maripaludis (MMP0595), Methanopyrus kandleri (MK0672), Methanococcus voltae, Methanothermococcus thermolithotrophicus, and Methanococcus vannielii; these are the known archaea that use Sec. Considering all of these data, it appeared probable that this protein was the missing Sep-tRNA:Sec-tRNA synthase (SepSecS) (Fig. 1 Upper).

Here, we report the identification of the human SLA/LP and Methanococcus maripaludis MMP0595 proteins as SepSecS enzymes (encoded by spcS) that convert Sep-tRNA^Sec → Sec-tRNA^Sec.

Results

SepSecS Rescues Selenoprotein Biosynthesis in an E. coli selA Deletion Strain.

When grown anaerobically, E. coli produces the selenium-dependent formate dehydrogenase H (FDH_H). Its activity enables the cells to reduce benzyl viologen (BV) in the presence of formate, which is usually observed in agar overlay plates under anaerobic conditions (24). We therefore constructed an E. coli selA deletion strain (JS1) in which selenoprotein production is abolished. Complementation of the JS1 strain with archaeal SepSecS genes from Methanococcus maripaludis (MMP0595) and Methanocaldococcus jannaschii (MJ0610) and with the human homolog (Q9HD40) allowed us to test the ability of these genes to restore selenoprotein biosynthesis as detected by FDH_H activity. The E. coli cells were grown under anaerobic conditions for 24 h and overlaid with top agar containing BV and formate. Blue-colored cells indicated that BV was reduced and therefore FDH_H activity was present, whereas cells without FDH_H activity remained colorless. Neither the archaeal nor the human SepSecS genes were able to complement the E. coli JS1 strain (Fig. 2). However, complementation was achieved when the SepSecS genes were cotransformed with the gene encoding Methanocaldococcus jannaschii PSTK; the latter gene alone was not able to restore Sec synthesis in the E. coli JS1 strain. Confirmation of these results was with MacConkey nitrate agar plates where cells with inactive formate dehydrogenase N form dark red colonies (data not shown). These data imply that Sec formation was achieved by a two-step process: PSTK phosphorylates the endogenous Ser-tRNA^Sec to Sep-tRNA^Sec, and then this misacylated aa-tRNA species is converted to Sec-tRNA^Sec by SepSecS. Thus, the human and archaeal SepSecS enzymes are active in E. coli. The data also demonstrate that Ser-tRNA^Sec, present in E. coli JS1, cannot be converted to Sec-tRNA^Sec, and therefore is not a substrate for SepSecS.

Fig. 2.

SepSecS genes restore FDH_H activity in an E. coli selA deletion strain. The E. coli JS1 strains complemented with the indicated SepSecS genes were grown anaerobically on glucose minimal medium plates supplemented with 0.01 mM IPTG at 30°C for 2 days. FDH_H activity was observed by overlaying top agar containing formate and BV. Colonies with active FDH_H reduced BV to a blue color. The SepSecS genes were from Methanocaldococcus jannaschii (MJ), Methanococcus maripaludis (MMP), human (HS), and E. coli selA (EC). K284T denotes a mutant human SepSecS where the Lys residue critical to PLP binding is changed to Thr.

The bacterial SelA is a PLP-dependent enzyme with a critical lysine residue engaged in PLP binding (25). As the archaeal and eukaryal SepSecS proteins also have PLP binding domains (21), we introduced Lys → Ala mutations to replace the critical lysine position in these enzymes (Methanocaldococcus jannaschii SepSecS K277A, Methanococcus maripaludis SepSecS K278A, and human SepSecS K284A). Even in the presence of PSTK, the mutant SepSecS genes were no longer able to complement the bacterial selA deletion (the result for the human SepSecS mutant is shown in Fig. 2). This finding strengthens the notion that the SepSecS proteins need PLP to carry out the Sep → Sec conversion.

⁷⁵Selenium Incorporation into E. coli Proteins.

To directly follow selenoprotein production in E. coli, we labeled with ⁷⁵Se the E. coli JS1 transformants (with SepSecS genes). These JS1 transformant strains were grown anaerobically in the presence of [⁷⁵Se]selenite and formate. Subsequently, the cell extracts were separated on a polyacrylamide gel and the radioactively labeled FDH_H was visualized by autoradiography. In agreement with the BV assay results, ⁷⁵Se-labeled FDH_H was detectable in the presence of archaeal SepSecS only in the strains also transformed with PSTK, whereas no labeling occurred when PSTK was absent (Fig. 3). The relatively lower labeling efficiency compared with the positive control, which was the transformant with E. coli WT selA, may be caused by the inefficient expression of the archaeal enzyme and the heterologous Sec biosynthesis system. The second labeled band with lower molecular weight was likely caused by the degradation of FDH_H considering the cells were harvested in stationary phase.

Fig. 3.

⁷⁵Se incorporation into the E. coli selenoprotein FDH_H. Cells were grown in the presence of [⁷⁵Se]selenite in TGYEP medium (see Materials and Methods) supplemented with 0.05 mM IPTG anaerobically at 37°C for 24 h. Cell extracts were separated by 10% SDS/PAGE, and the formation of ⁷⁵Se-containing FDH_H was followed by autoradiography. The E. coli strain JS1 was complemented with E. coli SelA (lane 1), empty plasmids (lane 2), Methanocaldococcus jannaschii SepSecS (lane 3), Methanocaldococcus jannaschii SepSecS with PSTK (lane 4), and Methanocaldococcus jannaschii SepSecS, PSTK with Methanococcus maripaludis tRNA^Sec (lane 5).

In a different experiment we investigated the effect of a homologous SepSecS:tRNA^Sec system on selenium incorporation by coexpressing archaeal tRNA^Sec in the E. coli JS1 strain. However, no significant increase in FDH_H production was observed, which suggests that the endogenous E.coli tRNA^Sec is a good substrate for archaeal SepSecS. As the secondary structure of tRNA^Sec differs considerably among tRNA species from the different domains, SepSecS might display a relaxed recognition of the tRNA moiety of the Sep-tRNA substrate.

SepSecS Converts Sep-tRNA^Sec to Sec-tRNA^Sec in Vitro.

To prove that SepSecS is a SelA that catalyzes a tRNA-dependent Sep → Sec synthesis we analyzed the conversion of [¹⁴C]Sep-tRNA to [¹⁴C]Sec-tRNA. To this aim Methanococcus maripaludis tRNA^Sec was acylated with [¹⁴C]serine by using pure Methanococcus maripaludis SerRS (26). Phosphorylation was performed with pure Methanocaldococcus jannaschii PSTK. The isolated tRNA was then incubated under anaerobic conditions with purified recombinant SepSecS and E. coli selenophosphate synthetase (SelD) in the presence of selenite. After the reaction aa-tRNA was deacylated, and the amino acids were identified by TLC in two systems (one is shown in Fig. 4). The data show that SepSecS from both Methanococcus maripaludis (Fig. 4, lane 3) and human (Fig. 4, lane 4) were able to form Sec-tRNA^Sec from Sep-tRNA^Sec, but not from Ser-tRNA^Sec (Fig. 4, lane 5). Thus, Sep-tRNA is the crucial precursor for Sec-tRNA formation in archaea and eukarya.

Fig. 4.

Conversion of in vitro-synthesized Sep-tRNA^Sec to Sec-tRNA^Sec. Phosphorimages of TLC separation of [¹⁴C]Sep and [¹⁴C]Sec recovered from the aa-tRNAs of the SepSecS activity assay (see Materials and Methods). Sec was analyzed in its oxidized form as selenocysteic acid (Secya). Lane 1, Ser marker; lane 2, Sep marker; lane 3, Sep-tRNA^Sec with Methanococcus maripaludis SepSecS; lane 4, Sep-tRNA^Sec with human SepSecS; lane 5, Ser-tRNA^Sec with Methanococcus maripaludis SepSecS; lane 6, reaction of lane 3 with selenite omitted.

Evolution of Two Sec Biosynthesis Pathways.

Although genetic coding of Sec is limited in organismal distribution, selenoproteins are found in all domains of life. Phylogenetic evidence demonstrates that the core Sec-decoding genes (tRNA^Sec, SelB, and SelD), and thus translation of genetically encoded Sec, is an ancient process that was already in existence at the time of the last universal common ancestor (27–29). SelA is of bacterial origin and was vertically inherited in that domain (30). Phylogenetic analysis of SepSecS and PSTK reveals distinct archaeal and eukaryal versions of these enzymes. This pathway can be traced back (at least) to the evolutionary split between the archaeal and eukaryal sister lineages (Fig. 5). SepSecS is confined to Sec decoding archaea and eukarya. Its evolutionary history is consistent with vertical inheritance and gene loss in non-Sec-decoding lineages (Fig. 5A).

Fig. 5.

Phylogenies of SepSecS (A) and PSTK (B). Organism names are color-coded according to the domain of life: Eukarya (green), Archaea (blue), and Bacteria (red). Non-Sec-decoding organisms are labeled in bold. Scale bar shows 0.1 changes per site. Only bootstrap values <90 are shown.

More interesting is the complex picture presented in the PSTK phylogeny (Fig. 5B). Like SepSecS, PSTK shows an initial deep divide between a strictly eukaryal group on the one hand and an archaeal group on the other. The archaeal PSTK appears to have been horizontally transferred to several of the Sec-decoding eukarya. The unexpected association between Oryza sativa and Strongylocertrotus purpuratus implies multiple horizontal gene transfer events, whereas a single transfer to mammals early in their evolution accounts for the clustering of these organisms within the archaeal genre. An additional horizontal gene transfer of the archaeal version to non-Sec-decoding Cyanobacteria is also suggested. The eukaryal PSTK is clearly discernable in a distinct group of non-Sec-decoding fungi that lack all other Sec biosynthesis components. These fungi form a coherent cluster in the eukaryal group, implying vertical inheritance. Retention of PSTK in this group suggests that it was recruited to perform a different function. Some eukarya (e.g., Caenorhabditis and Xenopus) retained both their native eukaryal PSTK and the acquired archaeal version, whereas in other lineages, including ancestors of mammals and trypanosomes, the archaeal version displaced its eukaryal counterpart. Most surprisingly, Plasmodium falciparum is Sec-decoding (30, 31), yet neither the PSTK nor SelA gene can be found, which contributes to the idea that a third pathway for Sec formation may exist in some organisms.

In summary, the last universal common ancestor genetically encoded Sec, and whether one or both pathways for converting Ser-tRNA^Sec → Sec-tRNA^Sec were present in the common ancestor remains to be seen.

Discussion

SepSecS has been suggested to be involved in Sec synthesis in eukarya since the early 1990s (20). In the current work, in vivo and in vitro results revealed that SepSecS is indeed a SelA analog able to catalyze the Sec-tRNA^Sec formation. However, in contrast to bacterial SelA that converts Ser-tRNA^Sec directly to Sec-tRNA^Sec, SepSecS acts on the phosphorylated intermediate, Sep-tRNA^Sec (Fig. 1). Therefore archaeal and eukaryal Sec biosynthesis requires three enzymes, SerRS, PSTK, and SepSecS, whereas bacteria accomplish the same task without PSTK. Both SelA and SepSecS are PLP-dependent enzymes. From a chemical standpoint a Sep → Sec conversion is desirable, as Sep would provide a better leaving group (phosphate) than Ser (water) for replacement with selenium. The detailed reaction mechanism of SepSecS should be subject to future structural and biochemical investigations. Sep-tRNA is reported to be more stable than Ser-tRNA (17), which is known to deacylate faster than any other aa-tRNA species (32). This fact would greatly improve the overall efficiency of the selenocysteinylation reaction and the rate of production of Sec-tRNA in all organisms that possess an extended selenoproteome. Most selenoprotein-containing bacteria only have one to three selenoproteins that are expressed under specific environmental conditions and a nonessential tRNA^Sec (33). In contrast, humans have an essential tRNA^Sec and an extended selenoproteome whose 25 members have been variously implicated in health and disease (4, 34).

Phylogenetic analyses showed that SelA is of bacterial origin, whereas SepSecS is confined to archaea and eukarya, and PSTK, principally an archaeal and eukaryal enzyme, has been horizontally transferred to the Cyanobacteria. tRNA^Sec has distinct secondary structure features: the long extra arm for recognition by SerRS, the elongated acceptor stem for SelB binding (35), and still unknown recognition sites for PSTK and SepSecS. Considering the number of proteins that tRNA^Sec interacts with, the clear division of bacterial and archaeal/eukaryal protein lineages and the discernable differences in tRNA^Sec structures in the three domains of life, it is interesting to see that E. coli tRNA^Sec can be recognized by archaeal and human SepSecS at least in vivo (35). Moreover, although the biological archaeal and mammalian selenium donors have not been characterized (6), our results indicate that selenophosphate provided by E. coli SelD can be used by SepSecS both in vitro and in vivo.

Are there other pathways? The absence of both PSTK and SelA genes in P. falciparum brings out interesting questions: Is there a third pathway or enzyme to make Sec? And why is Sec the only amino acid among the 22 naturally cotranslationally inserted protein building blocks that has not developed its own aa-tRNA synthetase?

Materials and methods

General.

Oligonucleotide synthesis and DNA sequencing was performed by the Keck Foundation Biotechnology Resource Laboratory at Yale University. [¹⁴C]serine (163 mCi/mmol) was obtained from Amersham Pharmacia Biosciences (Piscataway, NJ). [⁷⁵Se]selenite was purchased from the University of Missouri Research Reactor Facility (Columbia, MO).

Bacterial Strains and Plasmids.

The selA deletion strain JS1 was constructed by replacing the gene with kanamycin cassette in BW25113 E. coli strain as described (36). The T7 polymerase gene was inserted in the genome by P1 transduction. The selC deletion strain FM460 (37) was obtained from the Yale E. coli Stock Center. SepSecS genes from Methanococcus maripaludis (MMP0595) and Methanocaldococcus jannaschii (MJ0610) were amplified from genomic DNA, and the human SepSecS gene (GenBank accession no. BX648976) was amplified from an EST clone [clone ID: DKFZp686J1361, obtained from the RZPD German Resource Center for Genome Research, Berlin, Germany (38)] followed by cloning into the pET15b vector. E. coli selD was cloned into pET15b vector. Methanococcus maripaludis tRNA^Sec and Methanocaldococcus jannaschii PSTK gene were cloned into the pACYC vector individually or together. Human SepSecS K284T was generated by using the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA) according to the manufacturer's directions.

Protein Expression and Purification.

Hum5an SepSecS, Methanococcus maripaludis SepSecS, and E. coli SelD were transformed into BL21(DE3) cd+ strain. Cells were grown to A₆₀₀ = 1.0, and protein production was induced by the addition of 0.5 mM isopropyl β-d-thiogalactoside (IPTG). The cells were grown for 20 h at 15°C and then spun down and resuspended in 50 mM Hepes (pH 8.0), 300 mM NaCl, 10 mM imidazole, 3 mM DTT, 10 μM PLP, and protease inhibitor mix (Roche, Indianapolis, IN). After sonication, cell lysates were applied to Ni-NTA resin (Qiagen, Valencia, CA). The resin was washed, and the His₆-tagged proteins were eluted according to the manufacturer's manual. Proteins were then thoroughly dialyzed against 50 mM Hepes (pH 8.0), 300 mM NaCl, 10 mM DTT, 10 μM PLP, and 50% glycerol.

Complementation of an E. coli ΔselA Strain.

Methanocaldococcus jannaschii, Methanococcus maripaludis, or human SepSecS were transformed into a ΔselA E. coli JS1 strain, with or without the Methanocaldococcus jannaschii PSTK gene. The deletion strain complemented with its own selA gene served as positive control. E. coli ΔselA strain plus PSTK, SepSecS, or both genes was plated on glucose-minimum medium agar plates with 0.01 mM IPTG and grown anaerobically for 2 days at 30°C. The plates then were overlaid with agar containing 1 mg/ml BV, 0.25 M sodium formate, and 25 mM KH₂PO₄ adjusted to pH 7.0. The appearance of blue/purple color is the indication of active FDH_H (24). The same experiments were also carried out for SepSecS mutants. Formate dehydrogenase N activity was tested with MacConkey nitrate medium plates.

Metabolic Labeling with [⁷⁵Se]selenite.

E. coli cells were grown aerobically overnight and diluted (1:50) under anaerobic conditions in 5 ml of TGYEP medium (0.5% glucose, 1% tryptone, 0.5% yeast extract, 1.2% K₂HPO₄, 0.3% KH₂PO₄, 0.1% formate, 1 μM Na₂MoO₄, pH adjusted to 6.5) (39) with 1 μCi [⁷⁵Se]selenite and 0.05 mM IPTG added. The cells were further grown anaerobically at 37°C for another 24 h and harvested. Cell lysates were prepared and subjected to SDS/PAGE, followed by autoradiography.

Preparation of tRNA Substrates.

The Methanococcus maripaludis tRNA^Sec gene was cloned into pUC18 with T7 promoter and terminator. The plasmid was transformed together with a pACYC vector encoding T7 RNA polymerase gene to the E. coli selC deletion strain (FM460). tRNA^Sec was overexpressed and purified with biotinylated DNA oligonucleotides as described (40). Purified tRNA^Sec (10 μM) was serylated in the presence of [¹⁴C]Ser (200 μM) and Methanococcus maripaludis SerRS (6 μM) in reaction buffer (100 mM Hepes, pH 7.0/10 mM KCl/10 mM magnesium acetate/1 mM DTT/0.1 mg/ml BSA) at 37°C for 1 h. To prepare Sep-tRNA^Sec, 3 μM PSTK was added to the reaction. Aminoacylated tRNA products were purified by phenol extraction and ethanol precipitation followed by G-25 desalting column application.

In Vitro Sep → Sec Conversion.

Purified recombinant SepSecS was incubated in reaction buffer (100 mM Hepes, pH 7.0/300 mM KCl/10 mM MgCl₂) with Sep-tRNA^Sec or Ser-tRNA^Sec (10 μM), DTT (1 mM), Na₂SeO₃ (250 μM), and purified recombinant E. coli SelD (100 μM) at 37°C for 30 min. The oxygen in the buffer was removed, and the reaction was carried out in the anaerobic chamber. After incubation, proteins were removed by phenol extraction. The tRNAs were purified with G-25 column and ethanol precipitation to remove small molecules and salts. Purified tRNA products were deacylated in 20 mM NaOH at room temperature for 10 min. The released amino acids were oxidized with performic acid (16) and subjected to TLC analysis [system 1: 85% ethanol; system 2: butanol/acetic acid/water (4:1:1)].

Phylogenetic Analysis.

Sequences, taken from the National Center for Biotechnology Information nonredundant database or the ERGO database, were aligned by using Multiseq in VMD 1.8.5 (41). The most parsimonious trees were generated with PAUP4b10 (42). Gaps were counted as missing data and a parsimony cost matrix, based on the Blosum45 substitution matrix (43), was used. Of the 1,000 most parsimonious trees, the topology with the maximum likelihood was chosen and branch lengths were optimized with PHYML v.2.4.4 (44). The JTT+Γ model with eight rate categories was used for amino acid substitution, and adjustable parameters were derived from maximum-likelihood estimates. Bootstrap values were computed by using the re-estimation of log likelihoods (RELL) method in PROTML from the Molphy 3.2 package (45).

Abbreviations

aa-tRNA

aminoacyl-tRNA

BV

benzyl viologen

FDH_H

formate dehydrogenase H

IPTG

isopropyl β-d-thiogalactoside

PLP

pyridoxal phosphate

PSTK

phosphoseryl-tRNA^Sec kinase

Sec

selenocysteine

SelA

selenocysteine synthase

SelB

elongation factor SelB

SelD

selenophosphate synthetase

Sep

O-phosphoserine

SepSecS

Sep-tRNA:Sec-tRNA synthase

SerRS

seryl-tRNA synthetase.