Utilization of non-AUG initiation codons in a flow cytometric method for efficient selection of recombinant cell lines

Here we describe a method that couples flow cytometric detection with the attenuated translation of a reporter protein to enable efficient selection of CHO clones producing high levels of recombinant proteins. In this system, a small cell surface reporter protein is expressed from an upstream open reading frame utilizing a non-AUG initiation (alternate start) codon. Due to the low translation initiation efficiency of this alternate start codon, the majority of translation initiation events occur at the first AUG of the downstream open reading frame encoding the recombinant protein of interest. While translation of the reporter is significantly reduced, the levels are sufficient for detection using flow cytometric methods and, in turn, predictive of protein expression from the gene of interest since both ORFs are translated from the same mRNA. Using this system, CHO cells have been sorted to obtain enriched pools producing significantly higher levels of recombinant proteins than the starting cell population and clones with significantly better productivity than those generated from limiting dilution cloning. This method also serves as an effective screening tool during clone expansion to enable resources to be focused solely on clones with both high and stable expression.     

A synthetic arabinose-inducible promoter confers high levels of recombinant protein expression in hyperthermophilic archaeon Sulfolobus islandicus

Despite major progresses in genetic studies of hyperthermophilic archaea, recombinant protein production in these organisms always suffers from low yields and a robust expression system is still in great demand. Here we report a versatile vector that confers high levels of protein expression in Sulfolobus islandicus, a hyperthermophilic crenarchaeon. Two expression vectors, pSeSD and pEXA, harboring 11 unique restriction sites were constructed. They contain coding sequences of two hexahistidine (6×His) peptide tags and those coding for two protease sites, the latter of which make it possible to remove the peptide tags from expressed recombinant proteins. While pEXA employed an araS promoter for protein expression, pSeSD utilized P(araS-SD), an araS derivative promoter carrying an engineered ribosome-binding site (RBS; a Shine-Dalgarno [SD] sequence). We found that P(araS-SD) directed high levels of target gene expression. More strikingly, N-terminal amino acid sequencing of recombinant proteins unraveled that the protein synthesized from pEXA-N-lacS lacked the designed 6×His tag and that translation initiation did not start at the ATG codon of the fusion gene. Instead, it started at multiple sites downstream of the 6×His codons. Intriguingly, inserting an RBS site upstream of the ATG codon regained the expression of the 6×His tag, as shown with pSeSD-N-lacS. These results have yielded novel insight into the archaeal translation mechanism. The crenarchaeon Sulfolobus can utilize N-terminal coding sequences of proteins to specify translation initiation in the absence of an RBS site.     

Pharmacological-based translational induction of transgene expression in mammalian cells

In the quest for the development of pharmacological switches that control gene expression, no system has been reported that regulates at the translational level. To permit small-molecule control of transgene translation, we have constructed a farnesyl transferase inhibitor-responsive translation initiation factor. This artificial protein is a three-component chimaera consisting of the ribosome recruitment core of the eIF4G1 eukaryotic translation initiation factor, the RNA-binding domain of the R17 bacteriophage coat protein and the plasma membrane localization CAAX motif of farnesylated H-Ras. This membrane-delocalized translation factor is inactive unless liberated in the cytosol. Farnesyl transferase inhibitor FTI-277 prevents the membrane association of the CAAX motif and thus increases the cytoplasmic levels of the eIF4G fusion protein, which is then capable of inducing translation of the second cistron of a bicistronic messenger RNA containing an R17-binding site in its intercistronic space. Such direct translational control by farnesyl transferase inhibitors provides a system for fast, graded and reversible regulation of transgene expression.     

Escherichia coli translation initiation factor 3 discriminates the initiation codon in vivo

In a genetic selection designed to isolate Escherichia coli mutations that increase expression of the IS 10 transposase gene ( tnp ), we unexpectedly obtained viable mutants defective in translation initiation factor 3 (IF3). Several lines of evidence led us to conclude that transposase expression, per se , was not increased. Rather, these mutations appear to increase expression of the tnp'–'lacZ gene fusions used in this screen, by increasing translation initiation at downstream, atypical initiation codons. To test this hypothesis we undertook a systematic analysis of start codon requirements and measured the effects of IF3 mutations on initiation from various start codons. Beginning with an efficient translation initiation site, we varied the AUG start codon to all possible codons that differed from AUG by one nucleotide. These potential start codons fall into distinct classes with regard to translation efficiency in vivo : Class I codons (AUG, GUG, and UUG) support efficient translation; Class IIA codons (CUG, AUU, AUC, AUA, and ACG) support translation at levels only 1–3% that of AUG; and Class IIB codons (AGG and AAG) permit levels of translation too low for reliable quantification. Importantly, the IF3 mutations had no effect on translation from Class I codons, but they increased translation from Class II codons 3–5-fold, and this same effect was seen in other gene contexts. Therefore, IF3 is generally able to discriminate between efficient and inefficient codons in vivo , consistent with earlier in vitro observations. We discuss these observations as they relate to IF3 autoregulation and the mechanism of IF3 function.     

Escherichia coli gpt gene provides dominant selection for vaccinia virus open reading frame expression vectors.

Mycophenolic acid, an inhibitor of purine metabolism, was shown to block the replication of vaccinia virus in normal cell lines. This observation led to the development of a dominant one-step plaque selection system, based on expression of the Escherichia coli gpt gene, for the isolation of recombinant vaccinia viruses. Synthesis of xanthine-guanine phosphoribosyltransferase enabled only the recombinant viruses to form large plaques in a selective medium containing mycophenolic acid, xanthine, and hypoxanthine. To utilize the selection system efficiently, we constructed a series of plasmids that contain the E. coli gpt gene and allow insertion of foreign genes into multiple unique restriction endonuclease sites in all three reading frames between the translation initiation codon of a strong late promoter and synthetic translation termination sequences. The selection-expression cassette is flanked by vaccinia virus DNA that directs homologous recombination into the virus genome. The new vectors allow high-level expression of complete or partial open reading frames and rapid construction of recombinant viruses by facilitating the cloning steps and by simplifying their isolation. The system was tested by cloning the E. coli beta-galactosidase gene; in 24 h, this enzyme accounted for approximately 3.5% of the total infected-cell protein.     

eIF4AII is dispensable for miRNA-mediated gene silencing

MicroRNAs (miRNAs) are short noncoding RNAs that regulate gene expression through partial complementary base-pairing to the 3′ untranslated region (UTR) of target mRNAs. Inhibition of translation initiation has been identified as an early event of miRNA-mediated gene repression, but the underlying mechanistic details of this process are not well understood. Recently, eukaryotic initiation factor (eIF) 4AII was identified as a critical modulator of miRNA activity with depletion of this factor alleviating miRNA-mediated gene repression. Using the CRISPR/Cas9-editing system, we generated a novel cell line in which expression of eIF4AII was eliminated. The absence of eIF4AII does not affect cell viability, proliferation, or global mRNA translation. Importantly, we show that eIF4AII is dispensable for miRNA-mediated gene silencing.     

A novel, high stringency selection system allows screening of few clones for high protein expression

To obtain highly productive mammalian cell lines, often large numbers of clones need to be screened. This is largely due to low selection stringencies, creating many, but low protein producing clones. To remedy this problem, a novel, very stringent selection system was designed, to create few, but high protein producing clones. In essence, a selection marker with a startcodon that confers attenuated translation initiation frequency was placed upstream of the gene of interest with a startcodon that confers optimal translation initiation. From the transcribed bicistronic mRNA, the selection marker is translated at a low frequency, and the protein of interest at a high frequency. This selection system is so stringent that clones form only rarely. However, application of anti-repressor elements, which increase promoter activity, did induce the formation of clones that expressed proteins at high levels. When combined with anti-repressor elements, this novel selection system can be a valuable tool to rapidly create few, but highly productive mammalian cell lines.     

Binding of the bacteriophage T4 regA protein to mRNA targets: an initiator AUG is required.

Bacteriophage T4 regA protein translationally represses the synthesis of a subset of early phage-induced proteins. The protein binds to the translation initiation site of at least two mRNAs and prevents formation of the initiation complex. We show here that the protein binds to the translation initiation sites of other regA-sensitive mRNAs. Analysis of mRNA binding by filtration and nuclease protection assays shows that AUG is necessary but not sufficient for specific binding of regA protein to its mRNA targets. Anticipating the need for large quantities of regA protein for structural studies to further define the regA protein-RNA ligand interaction, we also report cloning the regA gene into a T4 overexpression system. The expression of regA protein in uninfected E. coli is lethal, so in our system regA driven by a strong T7 promoter is sequestered in a T4 phage until 'induction' by phage infection is desired. We have replaced the regA sensitive wild-type ribosome binding site with a strong insensitive ribosome binding site at an optimal distance from the regA initiation codon for maximizing expression. We have obtained large amounts of regA protein.     

Translation at higher than an optimal level interferes with coupling at an intercistronic junction

In pairs of adjacent genes co-transcribed on bacterial polycistronic mRNAs, translation of the first coding region frequently functions as a positive factor to couple translation to the distal coding region. Coupling efficiencies vary over a wide range, but synthesis of both gene products at similar levels is common. We report the results of characterizing an unusual gene pair, in which only about 1% of the translational activity from the upstream gene is transmitted to the distal gene. The inefficient coupling was unexpected because the upstream gene is highly translated, the distal initiation site has weak but intrinsic ability to bind ribosomes, and the AUG is only two nucleotides beyond the stop codon for the upstream gene. The genes are those in the filamentous phage IKe genome, which encode the abundant single-stranded DNA binding protein (gene V) and the minor coat protein that caps one tip of the phage (gene VII). Here, we have used chimeras between the related phage IKe and f1 sequences to localize the region responsible for inefficient coupling. It mapped upstream from the intercistronic region containing the gene V stop codon and the gene VII initiation site, indicating that low coupling efficiency is associated with gene V. The basis for inefficient coupling emerged when coupling efficiency was found to increase as gene V translation was decreased below the high wild-type level. This was achieved by lowering the rate of elongation and by decreasing the efficiency of suppression at an amber codon within the gene. Increasing the strength of the Shine-Dalgarno interaction with 16S rRNA at the gene VII start also increased coupling efficiency substantially. In this gene pair, upstream translation thus functions in an unprecedented way as a negative factor to limit downstream expression. We interpret the results as evidence that translation in excess of an optimal level in an upstream gene interferes with coupling in the intercistronic junction.     

Role of ribosome recycling factor (RRF) in translational coupling

RNA phage GA coat and lysis protein expression are translationally coupled through an overlapping termination and initiation codon UAAUG. Essential for this coupling are the proximity of the termination codon of the upstream coat gene to the initiation codon of the lysis gene (either a <3 nucleotide separation or physical closeness through a possible hairpin structure) but not the Shine-Dalgarno sequence. This suggests that the ribosomes completing the coat gene translation are exclusively responsible for translation of the lysis gene. Inactivation of ribosome recycling factor (RRF), which normally releases ribosomes at the termination codon, did not influence the expression of the reporter gene fused to the lysis gene. This suggests the possibility that RRF may not release ribosomes from the junction UAAUG. However, RRF is essential for correct ribosomal recognition of the AUG codon as the initiation site for the lysis gene.     

Two alternative translation mechanisms are responsible for the expression of the HCV ARFP/F/core+1 coding open reading frame

HCV-1 produces a novel protein, known as ARFP, F, or core+1. This protein is encoded by an open reading frame (ORF) that overlaps the core gene in the +1 frame (core+1 ORF). In vitro this protein is produced by a ribosomal frameshift mechanism. However, similar studies failed to detect the ARFP/F/core+1 protein in the HCV-1a (H) isolate. To clarify this issue and to elucidate the functions of this protein, we examined the expression of the core+1 ORF by the HCV-1 and HCV-1a (H) isolates in vivo, in transfected cells. For this purpose, we carried out luciferase (LUC) tagging experiments combined with site-directed mutagenesis studies. Our results showed that the core+1-LUC chimeric protein was efficiently produced in vivo by both isolates. More importantly, neither changes in the specific 10-A residue region of HCV-1 (codons 8-11), the proposed frameshift site for the production of the ARFP/F/core+1 protein in vitro, nor the alteration of the ATG start site of the HCV polyprotein to a stop codon significantly affected the in vivo expression of the core+1 ORF. Furthermore, we showed that efficient translation initiation of the core+1 ORF is mediated by internal initiation codon(s) within the core/core+1-coding sequence, located between nucleotides 583 and 606. Collectively, our data suggest the existence of an alternative translation initiation mechanism that may result in the synthesis of a shorter form of the core+1 protein in transfected cells.