The efficiency of selenocysteine incorporation is regulated by translation initiation factors

Abstract

Selenocysteine (Sec) incorporation is an essential process required for the production of at least 25 human selenoproteins. This unique amino acid is co-translationally incorporated at specific UGA codons that normally serve as termination signals. Recoding from stop to Sec involves a cis-acting Sec insertion sequence element in the 3' untranslated region of selenoprotein mRNAs as well as Sec insertion sequence binding protein 2, Sec-tRNA(Sec), and the Sec-specific elongation factor, eEFSec. The interplay between recoding and termination at Sec codons has served as a focal point in researching the mechanism of Sec insertion, but the role of translation initiation has not been addressed. In this report, we show that the cricket paralysis virus intergenic internal ribosome entry site is able to support Sec incorporation, thus providing evidence that the canonical functions of translation initiation factors are not required. Additionally, we show that neither a 5' cap nor a 3' poly(A) tail enhances Sec incorporation. Interestingly, however, the presence of the internal ribosome entry site significantly decreases Sec incorporation efficiency, suggesting a role for translation initiation in regulating the efficiency of UGA recoding.

Fig. 1

Constructs used to test Sec incorporation for the requirement of canonical translation initiation factors. The monocistronic firefly luciferase (FLuc) reporters (FLuc-Sec/wt or mut) have an in-frame UGA codon at position 258 and a wild-type (wt) or mutant (mut) GPx4 SECIS element. The position of the poly(A) tract for poly(A)+ variants is shown for FLuc-Sec/wt, which is annotated as FLuc-Sec/wt/p(A) in the text. All dual luciferase constructs have 5′ Renilla luciferase cistron (RLuc) and a 3′ FLuc cistron. The intercistronic region consists of highly structured RNA (ΔEMCV) and either contain (IRES) or lack (ΔIRES) the CrPV intergenic IRES. The dual luciferase constructs containing or lacking the cricket paralysis virus intergenic IRES (CrPV IRES) were generously provided by Dr. Eric Jan (University of British Columbia) and are described by Wilson et al. Variations of the dual luciferase reporters were generated such that FLuc contains a Sec codon at position 258 and a wild-type or mutant GPx4 SECIS element in the 3′ UTR.

Fig. 2

5′ cap and 3′ poly(A) tail are not required for Sec incorporation. (a) SDS-PAGE gel of [³⁵S]methionine labeled in vitro translated capped and uncapped FLuc-Sec/wt mRNAs with or without a poly(A) tail, as indicated, in the presence or absence of 120 nM recombinant CT-SBP2. The mock reaction lacks both mRNA and CT-SBP2. SDS-PAGE gels were visualized by Phosphor-Imaging and quantitated with ImageQuant software (GE Healthcare). (b) The Sec incorporation efficiencies for reactions containing CT-SBP2 were calculated as percent readthrough=FLuc/(FLuc+pre-Sec peptide)×100 after correction for the number of methionine residues per peptide. The asterisk denotes that the +cap pA value is significantly less than the +cap by Student’s t test (p=0.009). Data are mean±standard error of the mean, n=3.

Fig. 3

Sec incorporation does not require translation initiation factors. (a) Representative SDS-PAGE gel of [³⁵S] methionine labeled in vitro translation reactions of the indicated uncapped mRNAs in the presence or absence of 120 nM CT-SBP2. Translation products are indicated by the arrows. (b) Sec incorporation efficiencies for uncapped FLuc-Sec/wt and RLuc/IRES/FLuc-Sec/wt in the presence of 120 nM CT-SBP2 were calculated as percent readthrough=FLuc/(FLuc+pre-Sec peptide)×100 after correction for the number of methionine residues per peptide. Data are mean ±standard error of the mean, n=3. (c) FLuc/RLuc activity ratios (F/R) for the same reactions as in (a). Dark and light bars represent the absence or presence of 120 nM CT-SBP2, respectively. Data are mean±standard error of the mean, n=3, except for RLuc/IRES/FLuc-Sec/mut where n=2.

Fig. 4

Inefficient translation initiation does not reduce Sec incorporation. (a) Representative SDS-PAGE gel of [³⁵S]methionine labeled in vitro translation reactions of capped FLuc-Sec/wt mRNA in the presence of 120 nM CT-SBP2 and 0–1 mM m⁷G(5′)ppp(5′)G cap analog. Reactions supplemented with cap analog contained 0.8 mol MgCl₂ per mole of cap. (b) Quantitation of total translation in the presence of cap analog as measured by band densitometry (arbitrary units). (c) Quantitation of Sec incorporation efficiency, as described for previous figures, in the presence of cap analog. Data are mean± standard error of the mean, n=3.