The Selenoproteome of Clostridium sp. OhILAs: Characterization of Anaerobic Bacterial Selenoprotein Methionine Sulfoxide Reductase A

SUMMARY

Selenocysteine (Sec) is incorporated into proteins in response to UGA codons. This residue is frequently found at the catalytic sites of oxidoreductases. In the present study, we characterized the selenoproteome of an anaerobic bacterium, Clostridium sp. (also known as Alkaliphilus oremlandii) OhILA, and identified 13 selenoprotein genes, 5 of which have not been previously described. One of the detected selenoproteins was methionine sulfoxide reductase A (MsrA), an antioxidant enzyme that repairs oxidatively damaged methionines in a stereospecific manner. To date, little is known about MsrA from anaerobes. We characterized this selenoprotein MsrA which had a single Sec residue at the catalytic site but no cysteine (Cys) residues in the protein sequence. Its SECIS (Sec insertion sequence) element did not resemble those in Escherichia coli. Although with low translational efficiency, the expression of the Clostridium selenoprotein msrA gene in E. coli could be demonstrated by ⁷⁵Se metabolic labeling, immunoblot analyses, and enzyme assays, indicating that its SECIS element was recognized by the E. coli Sec insertion machinery. We found that the Sec-containing MsrA exhibited at least a 20-fold higher activity than its Cys mutant form, indicating a critical role of Sec in the catalytic activity of the enzyme. Furthermore, our data revealed that the Clostridium MsrA was inefficiently reducible by thioredoxin, which is a typical reducing agent for MsrA, suggesting the use of alternative electron donors in this anaerobic bacterium that directly act on the selenenic acid intermediate and do not require resolving Cys residues.

Introduction

Selenocysteine (Sec) is inserted into proteins when UGA (normally a stop codon) is recoded to serve as Sec codon ^1,2. This amino acid is typically found at the catalytic sites of oxidoreductase selenoproteins and is responsible for high catalytic activity of selenoenzymes as compared with their cysteine (Cys)-containing homologs ^3-6. Selenoproteins are found in all three domains of life: bacteria, archaea and eukaryotes. The glycine reductase A from Clostridium sticklandii was discovered as the first selenoprotein ^7,8. A complex machinery is required for translational insertion of Sec in response to UGA codon ^2,9,10. A stem-loop structure in selenoprotein mRNAs, called SECIS (Sec insertion sequence) element, is essential for Sec insertion in both prokaryotes and eukaryotes. SECIS elements are located immediately downstream of Sec-encoding UGA codons in bacteria, whereas they are present in 3′-untranslational regions in archaea and eukaryotes. In addition to this cis-acting factor, several trans-acting factors, such as Sec-specific elongation factor and tRNA^Sec, are required for Sec incorporation.

There has been significant progress recently in identifying selenoprotein genes in organisms from bacteria to mammals owing to both a dramatic increase in genomic sequence information and development of bioinformatics methods that allow searches for selenoprotein genes^11-16. These methods include searches for 1) SECIS elements; and 2) Sec/Cys pairs in predicted homologs.

Methionine residues are easily oxidized by various reactive oxygen species to methionine sulfoxides; however, the oxidized methionines can be reduced back to methionine by repair enzymes, the methionine sulfoxide reductases (Msrs) ¹⁷. Two distinct families of these enzymes, MsrA and MsrB, have evolved for stereospecific reduction of methionine sulfoxide residues in proteins. MsrA is specific for the reduction of the S-form of methionine sulfoxide, whereas MsrB is specific for the R-form. Msrs are present in most organisms in all three domains of life (they have been even found in anaerobic bacteria), but are absent in some hyperthermophiles and parasites ^18,19. Msrs play a pivotal role in the repair of oxidatively damaged proteins and defend cells from oxidative stress. Therefore, these proteins are viewed as protein repair and antioxidant enzymes ^20-22. These enzymes have been implicated in regulation of physiological and pathological processes such as aging and neurodegenerative diseases ^23,24.

In this study, we characterized the selenoproteome of an anaerobic gram-positive bacterium, Clostridium sp. (also known as Alkaliphilus oremlandii) OhILAs and found that this bacterium is predicted to be rich in selenoproteins and may contain at least 13 selenoproteins, including MsrA. We further experimentally characterized Clostridium selenoprotein MsrA. We found that its SECIS element on the mRNA is recognized by the Escherichia coli Sec insertion machinery, and that Sec plays a critical role in the catalytic activity of this selenoprotein.

Materials and Methods

Identification of selenoprotein genes in Clostridium sp. OhILAs

The complete genome of Clostridium sp. OhILAs (NC_009922) was downloaded from the NCBI ftp server (ftp://ftp.ncbi.nih.gov/genomes/Bacteria/). We used bSECISearch program developed previously ¹⁵ to search this genome for all candidate selenoprotein sequences containing in-frame UGA codons and bacterial SECIS (bSECIS) elements. The resulting dataset was then divided into homologs of previously known selenoproteins and selenoprotein candidates. All candidates were manually analyzed for the location of the UGA codon, occurrence of Sec- and Cys-containing homologs in Sec-utilizing or other organisms, and presence of SECIS elements in Sec-containing homologs. In parallel, independent BLAST homology searches for Sec-containing homologs of all known selenoprotein families were performed using default parameters. Finally, both known and candidate selenoprotein sets were generated. Multiple sequence alignments were performed using ClustalW program ²⁵ with default parameters and visualized with BoxShade program v3.21.

Cloning, expression, and purification of wild-type selenoprotein and its Cys mutant of Clostridium MsrA

A coding region of the selenoprotein msrA gene was amplified by PCR using Clostridium sp. OhILAs genomic DNA (kindly provided from Dr. John Stolz at the Duquesne University, Pittsburgh, USA) and cloned into NdeI/XhoI sites of pET21b (Novagen). To amplify the gene, forward (5′-GCGCCATATGGATACCAATCAGAAGTTG-3′) and reverse (5′-CGCGCTCGAGTTTAATTTCTTCGACGATAG-3′) primers were used. The resulting construct, designated pET-CLOS-MsrA, coded for full-length selenoprotein MsrA with a C-terminal His-tag (LEHHHHH). A SECIS element was found to be located immediately downstream of UGA Sec codon (see Figure 4). A Cys mutant form was also generated in which Sec16 was replaced with Cys by site-directed mutagenesis. The resulting construct was named pET-CLOS-MsrA/U16C.

Figure 4.

Structures of SECIS elements in E. coli fdhH (left) and Clostridium sp. OhILAs msrA (right). In the E. coli fdhH SECIS element, the minimal step-loop structure is boxed and the bulged U is shown in italics. The Clostridium msrA SECIS element lacks the bulged U. Conserved G in the apical loops is shown in red in both SECIS elements. Sec UGA codons are shown in blue.

To express selenoprotein msrA in E. coli, the plasmid pET-CLOS-MsrA was transformed into BL21(DE3) cells that also harbored a plasmid pSUABC ²⁶, which encodes E. coli selA, selB, and selC. The transformed cells were grown in LB media containing 4 μM sodium selenite, 100 μg/ml ampicillin, and 25 μg/ml chloramphenicol at 37°C with shaking until optical density at 600 nm reached 0.6∼0.8; then 0.2 mM IPTG was added and the cells were further cultured for 4 h at 30°C, harvested, and stored at −20°C until use. A typical expression level of the selenoprotein MsrA was 0.6 μg per mg crude protein.

The cell pellets were resuspended in the extraction buffer (50 mM sodium phosphate, pH 7.0, 300 mM NaCl, 10 mM imidazole, and 1 mM phenylmethylsulfonyl fluoride) and lysed by sonication. The selenoprotein MsrA was further purified using a Talon-metal affinity resin (Clontech) according to the manufacturer's protocol. The eluted proteins were dialyzed against 50 mM sodium phosphate, pH 7.5, and 50 mM NaCl. Typical yield of the purified selenoprotein MsrA was 3-5 μg from 250 ml culture broth, as estimated by Western blot assays. Although the protein was not pure, this preparation was sufficient to determine kinetic parameters of the enzyme.

To express the Cys mutant form (U16C), BL21(DE3) cells transformed with pET-CLOS-MsrA/U16C were cultivated in LB media containing 100 μg/ml ampicillin at 37°C with shaking until optical density at 600 nm reached 0.6∼0.8, 0.1 mM IPTG was added, and the cells were further cultured for 6 h at 30°C. The cells were harvested and stored at −20°C until use. Procedures for purification of this protein were as described above for the selenoprotein form. The purity of the Cys mutant protein was analyzed by SDS-PAGE and found to exceed 95%.

Metabolic labeling with ⁷⁵Se

To prepare cells expressing Clostridium selenoprotein MsrA with metabolically labeled ⁷⁵Se, E. coli BL21(DE3) cells were transformed with pET-CLOS-MsrA and grown at 37°C in 5 ml LB media containing ampicillin until optical density at 600 nm reached ∼0.6. Then, 0.05 mCi of freshly neutralized [⁷⁵Se]selenous acid (specific activity 1,000 Ci/mmol, University of Missouri Research Reactor, Columbia, USA) was added to the cell culture with 1 mM IPTG for induction. The cells were further grown at 37°C for 5 h, harvested, washed with PBS buffer and lysed. Cell extracts (40 μg of total protein) were applied to a 10% Bis-Tris gel (Invitrogen), electrophoresed, and transferred onto a PVDF membrane. The ⁷⁵Se radioactivity pattern on the membrane was visualized by using a PhosphorImager (GE Health Care).

Determination of protein concentration

Concentration of recombinant selenoprotein MsrA was determined by western blotting using His antibodies and Cys mutant as an internal standard, followed by quantifying the blot signals with a densitometer. Concentration of purified Cys mutant was determined by Bradford method using a Bio- Rad protein assay reagent (Bio-Rad) and bovine serum albumin as a standard.

MsrA enzyme assay and analysis of kinetics

MsrA activity was measured in the presence of dithiothreitol (DTT) or thioredoxin (Trx) as a reducing agent. In the DTT-dependent reaction, reaction mixture (100 μl) contained 50 mM sodium phosphate, pH 7.5, 50 mM NaCl, 20 mM DTT, 200 μM DABSyl-methionine-S-sulfoxide, and purified (or crude) proteins. In the Trx-dependent reaction, the reaction mixture (100 μl) contained 50 mM sodium phosphate, pH 7.5, 50 mM NaCl, 0.2 mM NADPH, 6.8 μM E. coli Trx (Sigma) or human Trx1 ²⁷, 0.8 μM human Trx reductase 1 ²⁷, 200 μM DABSyl-methionine-S-sulfoxide, and purified proteins. The reactions were carried out at 37°C for 30 min and analyzed by HPLC as described previously ²⁸. A full-length mouse MsrA was purified from E. coli as described previously ²⁹ and used for comparison.

K_m and k_cat values were determined from Lineweaver-Burk plots using 50-400 μM substrate in the DTT-dependent reaction.

Results and Discussion

Clostridium sp. OhILAs is a selenoprotein-rich organism

By searching the Clostridium sp. OhILAs genome against a set of all known selenoprotein genes, we identified 8 selenoproteins in Clostridium sp. OhILAs, including formate dehydrogenase alpha subunit (FdhA), which is the most widespread prokaryotic selenoprotein that might be responsible for maintaining the Sec utilization trait in sequenced bacteria ³⁰, and selenophosphate synthetase (SelD), another widespread selenoprotein which is a key component in prokaryotic selenoprotein biosynthesis. Other detected selenoproteins included proline reductase, HesB-like protein, glutaredoxin, glycine reductase selenoproteins A (GrdA) and B (GrdB), and MsrA. Previously, selenoprotein MsrA has only been characterized in a eukaryotic organism, Chlamydomonas reinhardtii ^31,32. The selenoproteome of Clostridium sp. OhILAs is shown in Table 1 and all detected selenoprotein sequences are included in the supplementary sequence data.

Table 1.

Known and new selenoproteins detected in Clostridium sp. OhILAs

Protein family	Accession no.	COG/Pfam	Occurrence	Sec position (protein length)
Known selenoproteins
FdhA*	YP_001513778	COG3383	1	349 (891)
SelD*	YP_001513274	COG0709	1	16 (346)
MsrA*	YP_001513481	COG0225	1	16 (209)
Proline reductase*	YP_001511596	pfam07355 (low homology)	1	151 (241)
HesB-like	YP_001511838 (partial sequence)	COG0316	1	32 (95)
Glutaredoxin*	YP_001513661	COG0695	1	13 (76)
GrdA*	YP_001512500	pfam04723	1	44 (158)
GrdB*	YP_001512501	pfam07355	1	350 (435)
New selenoprotein candidates
Putative anaerobic dehydrogenase*	YP_001511765	COG0243	1	257 (1021)
BFD-like (2Fe-2S)-binding domain protein	YP_001512630 (partial sequence)	COG2906 (low homology)	1	11 (162)
Split soret cytochrome c precursor*	YP_001512574 YP_001512575	pfam09719 (low homology)	2	83 (223) 81 (221)
Predicted NADH:ubiquinone oxidoreductase, subunit RnfC	YP_001511593 (partial sequence)	COG4656	1	389 (424)

^*annotated as selenoproteins in the current genome annotation

Next, we used bSECISearch to analyze Clostridium sp. OhILAs for full complement of selenoprotein genes. In addition to the 8 homologs of known selenoproteins, we detected 5 new selenoproteins that represented 4 protein families (Table 1). The mRNAs for all new selenoproteins contained predicted stable stem-loop structures downstream of Sec-encoding UGA codons that shared common features of bacterial SECIS elements ¹⁵, e.g., a stable stem with a small apical loop containing a G nucleotide located at a certain distance from the Sec UGA codon (Figure 1). No additional common features could be found in those Clostridium SECIS elements and all of them were predicted to lack a bulged U which is conserved in E. coli SECIS elements. Selenoprotein homologs of the new selenoproteins were also found in several other anaerobic bacteria, such as Desulfitobacterium hafniense, Alkaliphilus metalliredigens, and Clostridium botulinum (see Figure 2). In addition, conserved Cys-containing homologs could be detected for all new selenoproteins in other organisms. Multiple alignments of new selenoprotein families and their Cys homologs are shown in Figure 2 and highlight conserved Sec/Cys pairs in homologous sequences that support predicted redox functions of these proteins ³³.

Figure 1.

Predicted bacterial SECIS elements in new selenoprotein genes. Only sequences downstream of in-frame UGA codons are shown. In-frame UGA codons and conserved G nucleotides in the apical loop are shown in red.

Figure 2.

Multiple sequence alignments of new selenoproteins and their Cys homologs. The alignments show Sec-flanking regions in detected selenoproteins. Both selenoprotein sequences detected in Clostridium sp. OhILAs and their Sec/Cys-containing homologs present in indicated organisms are shown. Conserved residues are highlighted. Predicted Sec (U) and the corresponding Cys (C) residues in other homologs are shown in red and blue, respectively.

Although the functions of new selenoproteins are not clear, they either contained a domain of known function or were homologous to protein families with known functions. However, in some of them Sec was not part of the detected domains, suggesting the presence of functional redox-active sites in the locations outside of known domains. For example, Sec was located in the N-terminal region of BFD-like (2Fe-2S)-binding protein in which the BFD-like domain (COG2906) was present in the C-terminal region. It was also found in a CxxU (x, any amino acid residue; U, Sec) redox motif in the C-terminal region of a predicted NADH:ubiquinone oxidoreductase subunit RnfC, whereas the known RnfC domain (COG4656) was located in the N-terminal region. The other three new selenoproteins (putative anaerobic dehydrogenase and two cytochrome c precursor sequences) were annotated as putative selenoproteins in this genome although the criteria used in their detection/annotation are not clear. Since these sequences have passed the stringent criteria employed by bSECISearch in our study, they should be viewed as excellent candidate selenoproteins.

While this study was prepared for publication, the sequences of 3 new selenoproteins were released at GenBank database (putative anaerobic dehydrogenase, 158139457; split soret cytochrome c precursor 1, 158320067; split soret cytochrome c precursor 2, 158320068). It is interesting that the Sec positions differed in the predicted NADH:ubiquinone oxidoreductase subunit RnfC family between Clostridium sp. OhILAs and several strains of Clostridium botulinum (Figure 2). Taken together, our data suggested that novel Se-dependent pathways are present in this selenoprotein-rich organism. Further experimental verification is needed for the newly identified selenoproteins.

Clostridium selenoprotein MsrA sequence

Msr genes have been found in most organisms, even in anaerobes. The presence of Msr genes in anaerobic bacteria may be due to their roles in protein repair and antioxidant defense when transiently exposed to oxygen. To date, little is known about such anaerobe-based Msrs. We selected the anaerobic bacterial selenoprotein MsrA for further experimental examination. Interestingly, sequence context revealed lack of any Cys residues in this protein. Since Sec corresponded to the catalytic residue, this observation suggested the lack of resolving residues during catalysis (Figure 3). The catalytic mechanism of Cys-containing MsrA is well characterized and involves a sulfenic acid intermediate ^34-36. During the reaction, the catalytic Cys attacks sulfoxide moiety of methionine sulfoxide resulting in sulfenic acid on this residue with a concomitant release of product, methionine. Then, the sulfenic acid intermediate is rearranged into a disulfide bond by reacting with another Cys, known as resolving (or recycling) Cys, and finally the disulfide bond is reduced by a two-electron reductant. Trx is typically an in vivo reducing agent while DTT can be used in vitro. In the case of Sec-containing forms, the selenenic acid intermediate may form a selenenylsulfide bond with a resolving Cys; however, neither of the previously identified selenoprotein MsrAs possesses conserved Cys that could function as resolving residues ³¹. Furthermore, we previously reported that Chlamydomonas selenoprotein MsrA does not use a resolving Cys residue ³¹.

Figure 3.

Multiple sequence alignment of MsrAs. Sec (U) residues are highlighted in red in the catalytic GUFW(G/H) motif and the corresponding Cys residues are highlighted in blue. The known resolving Cys residues in the C-terminal region are highlighted in pink. Genbank accession numbers are as follows: Clostridium sp. OhILAs, 106894182; Chlamydomonas reinhardtii, 23452038; Syntrophobacter fumaroxidans, 71547479; E. coli, 1790665; Mus musculus, 31981013; Drosophila melanogaster, 23338220.

Expression of Clostridium selenoprotein MsrA in E. coli

The mechanism of Sec insertion has been well elucidated in E. coli ^3,37-39. The SECIS element in mRNA is essential for Sec insertion and is located immediately downstream of in-frame UGA codons in bacteria. Studies on E. coli SECIS elements revealed that only a minimal stem-loop is required for Sec insertion in which a GU sequence on the tip of apical loop and a bulged U present at the bottom of the stem are essential for Sec insertion (Figure 4, left) ^40-42. We recently reported a consensus bacterial SECIS structural model in which a conserved G is present in the apical loop, whereas the bulged U is an E. coli adaptation that is missing in most other bacterial SECIS elements ¹⁵. Consistent with this consensus, the SECIS element of Clostridium msrA had a G in the apical loop, but not the bulged U in the stem (Figure 4, right). The differences between E. coli SECIS elements and corresponding structures in other bacteria result in barriers in heterologous selenoprotein expression ⁴³.

To test whether Clostridium selenoprotein MsrA can be expressed in E. coli as recombinant protein, we metabolically labeled cells transformed with Clostridium msrA gene with ⁷⁵Se. As shown in Figure 5A, a 25 kDa radioactive band corresponding in size to the calculated molecular weight of recombinant Clostridium selenoprotein MsrA, was detected. We also verified the expression of this protein by western blot analysis (Figure 5B). In addition, enzyme assays revealed that crude extracts from cells expressing the Clostridium msrA gene had a 36-fold increased activity compared with that from cells containing an empty vector (Figure 5C). Thus, these data showed that Clostridium selenoprotein msrA could be expressed in E. coli, indicating that its SECIS element is recognized by the E. coli Sec insertion machinery. To estimate the efficiency in Sec insertion (UGA readthrough), we analyzed the amounts of wild-type and Cys mutant (U16C) forms of the enzyme in crude extracts by Western blotting using the purified U16C as an internal standard. It should be noted that most of U16C protein (>95%) expressed in BL21(DE3) cells was present in the soluble fraction. The yields of the wild-type selenoprotein and the U16C mutant were 0.6 and 380 μg per mg crude protein, respectively. Therefore, the translational efficiency of the selenoprotein MsrA was low, estimated to be 0.15% of that of its Cys variant. The activities of the wild-type selenoprotein and Cys mutant forms in the crude extracts as shown in Figure 5C agreed well with those of the purified forms, respectively (Table 2).

Figure 5.

Expression of Clostridium selenoprotein MsrA in E. coli. (A) ⁷⁵Se metabolic labeling. BL21(DE3) cells transformed with an empty vector (lane 1) or pET-CLOS-MsrA (lane 2) were metabolically labeled with ⁷⁵Se. Proteins were separated by SDS-PAGE, transferred onto a PVDF membrane, and visualized with a PhosphorImager. (B) Western blot analysis. Purified Cys mutant (U16C) and crude extracts from cells expressing Clostridium selenoprotein MsrA were analyzed using anti-His antibodies. Lane 1, purified U16C; lane 2, wild-type selenoprotein MsrA from cell extracts. (C) Enzyme assays. Crude extracts from cells containing an empty vector, pET-CLOS-MsrA (wild-type), or pET-CLOS-MsrA/U16C (U16C mutant) were prepared, respectively. A total of 200 μg crude protein from control and CLOS-MsrA extracts and 2 μg crude protein of the CLOS-MsrA/U16C extract were used in a DTT-dependent assay in the presence of 200 μM DABSyl-methionine-S-sulfoxide.

Table 2.

Specific activity and kinetic constants of Clostridium MsrA forms

MsrA forms	Specific activity (nmol/min/mg protein)	Km (mM)	kcat (s−1)	kcat/ Km (M−1 S−1)
WT	4,185 ± 46	10.2 ± 1.2	84 ± 5	8,235
U16C	464 ± 14	4.0 ± 0.2	3.8 ± 0.3	950

Purified proteins (90 ng WT and 1 μg U16C) were assayed in the DTT-dependent reaction. The substrate (DABSyl-methionine-S-sulfoxide) concentration used was 200 μM for determination of specific activity. K_m and k_cat values were determined from Lineweaver-Burk plots using 50-400 μM substrate in the presence of DTT. WT, wild-type selenoprotein; U16C, Sec to Cys mutant.

There have been several reports on heterologous expression of eukaryotic selenoproteins in E. coli^6,26,44,45. Either the entire SECIS element of E. coli fdhH (as in most studies) or a minimal SECIS stem-loop had to be introduced downstream of UGA codon to express eukaryotic selenoproteins. In E. coli, heterologous expression of bacterial selenoproteins that contain their own SECIS elements often fails due to SECIS structure barriers ⁴⁶. However, a few successful cases have been previously reported ^47,48. In our present study, irrespective of discrepancies in Clostridium msrA and E. coli SECIS elements, the Clostridium selenoprotein msrA gene could be expressed in E. coli. However, as discussed above, the translational efficiency was low (less than 0.2% compared to its Cys variant) even when the cells were co-expressed with selA, selB, and selC genes encoding selenocysteine synthase, SelB and tRNA^Sec, respectively. These results suggest a low functionality of the SECIS element of Clostridium msrA in the E. coli Sec insertion system.

Characterization of Clostridium selenoprotein MsrA

To investigate catalytic properties of Clostridium MsrA, we purified this protein (tagged at the C-terminus with His) from E. coli cells. As shown in Table 2, the specific activity of Clostridium selenoprotein MsrA was found to be 4,185 nmol/min/mg protein in a DTT-dependent reaction. This value was 9-fold higher than that of its Cys mutant (U16C). The k_cat value of the selenoprotein MsrA was 22-fold higher than that of its Cys mutant and the K_m value of the selenoprotein was 2.5-fold higher. Overall, these data show that selenium provides catalytic advantage to Clostridium MsrA. Importantly, this enzyme appears to be a highly efficient catalyst with kinetic properties comparable with those of Chlamydomonas MsrA characterized previously ³¹.

Trx is typically a natural reductant for regeneration of MsrA enzymes. To test if Trx can function as a reductant for Clostridium MsrA, we measured Trx-dependent activities of its natural Sec-containing and mutant Cys-containing forms using human Trx1 and Trx reductase 1, and compared the data with DTT-dependent activities of these MsrA forms. In addition, mouse MsrA served as control because this protein is known as a target of Trx. As shown in Table 3, the ratio of Trx- to DTT-dependent activities of mouse MsrA was 0.73, indicating that human Trx1 as well as DTT could efficiently reduce this protein. However, for Clostridium selenoprotein MsrA, this ratio was less than 0.01, suggesting that Trx1 was a poor reductant for this selenoprotein. Similarly, the Cys mutant could not be efficiently reduced by Trx1 (the ratio of Trx- to DTT-dependent activity was 0.03). To examine the possibility that human Trx couples poorly to the bacterial selenoprotein MsrA, we carried out the enzyme assay in the presence of bacterial (E. coli) Trx. The ratios of Trx- to DTT-dependent activities were consistent with the results for human Trx1; 0.93 for mouse MsrA, 0.016 for Clostridium selenoprotein MsrA, and 0.05 for its Cys mutant. Therefore, our data suggested that Trx is not a reducing agent in vivo for Clostridium MsrA and that unknown physiological reductant(s) must function in the reduction of this enzyme.

Table 3.

Ratios of Trx- to DTT-dependent activities

MsrA forms	Specific activity (nmol/min/mg protein)			Ratio of Trx/DTT activity
	DTT-dependent reaction	Trx-dependent reaction
		hTrx1	eTrx	hTrx1	eTrx
WT	4,185 ± 46	34 ± 5	70 ± 6	0.008	0.017
U16C	464 ± 14	14 ± 2	23 ± 2	0.03	0.05
Mouse MsrA	256 ± 24	188 ± 9	238 ± 8	0.73	0.93

Purified proteins (90 ng WT, 1 μg U16C, and 1 μg mouse MsrA) were used in the DTT- and Trx-dependent assays containing 200 μM DABSyl-methionine-S-sulfoxide. hTrx1, human cytosolic Trx; eTrx, E. coli Trx; WT, Clostridium selenoprotein MsrA; U16C, its Cys mutant.

Recent studies revealed that thionein (a low-molecular-mass Cys-rich protein) can function as a reducing agent for MsrA and MsrB ⁴⁹, and that selenium compounds such as selenocysteamine could reduce mammalian MsrBs ⁵⁰. Our previous study on Chlamydomonas selenoprotein MsrA and its Cys mutant revealed that the Trx-dependent activities of these enzyme forms were more than 20-fold lower than their DTT-dependent activities ³¹. Taken together and combined with the observation that known selenoprotein MsrAs lack conserved resolving Cys residues (or do not have any Cys as in Clostridium MsrA) ³¹, our data on Chlamydomonas and Clostridium MsrAs suggest that Trx is not a natural reducing agent for the selenenic acid intermediates of selenoprotein MsrAs.

Conclusion

This study shows that Clostridium sp. OhILAs, a gram-positive anaerobic bacterium, is predicted to be a selenoprotein-rich organism that may contain 13 selenoproteins, including 5 that have not been previously described. Interestingly, one of the detected selenoproteins was MsrA, which serves an antioxidant function by catalyzing the reduction of methionine-S-sulfoxide. Little is known about anaerobic bacterial MsrAs. We cloned and expressed this Clostridium selenoprotein in E. coli. Although its SECIS element did not resemble those in E. coli, we could verify the recombinant protein expression by ⁷⁵Se labeling, western blots, and activity assays. Thus, the SECIS element of Clostridium selenoprotein msrA is functional and recognized by the E. coli Sec insertion machinery, although the translational efficiency was found to be low. The enzyme assays and kinetic analyses revealed that the selenoprotein MsrA is much more active than its Cys mutant form, indicating a key contribution of selenium to the catalytic efficiency of this selenoenzyme. Our data also suggest that Trx, a general in vivo reducing agent for MsrA, is not a reductant for Clostridium selenoprotein MsrA. Further studies will be needed to characterize the detailed reductase steps including determination of the natural reducing agent for this selenoprotein MsrA.

Abbreviations

DTT

dithiothreitol

Msr

methionine sulfoxide reductase

Sec

selenocysteine

SECIS

selenocysteine insertion sequence

Trx

thioredoxin