Recombinant production of the p10CKS1At protein from Arabidopsis thaliana and 13C and 15N double-isotopic enrichment for NMR studies.

The CKS1At gene product, p10CKS1At from Arabidopsis thaliana, is a member of the cyclin-dependent kinase subunit (CKS) family of small proteins. These proteins bind the cyclin-dependent kinase (CDK)/cyclin complexes and play an essential, but still not precisely known role in cell cycle progression. To solve the structure of p10CKS1At, a protocol was needed to produce the quantity of protein large enough for nuclear magnetic resonance (NMR) spectroscopy. The first attempt to express CKS1At in Escherichia coli under the control of the T7 promoter was not successful. E. coli BL21(DE3) cotransformed with the CKS1At gene and the E. coli argU gene that encoded the arginine acceptor tRNAUCU produced a sufficient amount of p10CKS1At to start the structural study by NMR. Replacement of four rare codons in the CKS1At gene sequence, including a tandem arginine, by highly used codons in E. coli, restored also a high expression of the recombinant protein. Double-isotopic enrichment by 13C and 15N is reported that will facilitate the NMR study. Isotopically labeled p10CKS1At was purified to yield as much as 55 mg from 1 liter of minimal media by a two-step chromatographic procedure. Preliminary results of NMR spectroscopy demonstrate that a full structural analysis using triple-resonance spectra is feasible for the labeled p10CKS1At protein.     

Construct for high-level expression and low misincorporation of lysine for arginine during expression of pET-encoded eukaryotic proteins in Escherichia coli

The arginine codon AGA is rarely used in E. coli but is common in eukaryotic genes. Prior studies have shown that the low level of tRNA(UCUArg) can lead to low expression and misincorporation of lysine for arginine, during expression of genes containing AGA codons in E. coli. The chloramphenicol-selectable plasmid pJY2 is designed to facilitate the expression of such genes cloned into pET vectors: it encodes T7 lysozyme (to depress constitutive expression of the cloned gene) and tRNA(UCUArg) (to suppress lysine misincorporation at AGA codons). Using pJY2, we observed robust and translationally faithful expression of mutant ubiquitin genes in which 14% (11 out of 76) of the total codons were AGA. Competent BL21(DE3)pJY2 cells can be used to suppress lysine misincorporation and achieve high-level expression of pET-encoded target genes without modification of AGA codons in the target gene sequence.     

Mistranslational errors associated with the rare arginine codon CGG in Escherichia coli

In Escherichia coli, CGG is a rare arginine codon occurring at a frequency of 0.54% in all E. coli mRNAs or 9.8% when an arginine residue is encoded for. When present in high numbers or in clusters in highly expressed recombinant mRNA, rare codons can cause expression problems compromising product yield and translational fidelity. The coding region for an N-terminally polyhistidine tagged p27 protease domain from Herpes Simplex Virus 2 (HSV-2) contains 11 of these rare arginine codons, with 3 occurring in tandem near the C-terminus of the protein. When expressed in E. coli, the majority of the recombinant material produced had an apparent molecular mass of 31 kDa by SDS-PAGE gels or 3 kDa higher than predicted. Detailed biochemical analysis was performed on chemical and enzymatic digests of the protein and peptide fragments were characterized by Edman and MS/MS sequencing approaches. Two major species were isolated comprising +1 frameshift events at both the second and third CGG codons in the triplet cluster. Translation proceeded in the missense frame to the next termination codon. In addition, significant levels of glutamine misincorporating for arginine were discovered, suggesting second base misreading of CGG as CAG. Coexpression of the argX gene, which encodes the cognate tRNA for CGG codons, largely eliminated both the frameshift and misincorporation events, and increased expression levels of authentic product by up to 7-fold. We conclude that supplementation of the rare arginyl tRNA(CGG) levels by coexpression of the argX gene can largely alleviate the CGG codon bias present in E. coli, allowing for efficient and accurate translation of heterologous gene products.     

Improvement of human interferon HUIFNalpha2 and HCV core protein expression levels in Escherichia coli but not of HUIFNalpha8 by using the tRNA(AGA/AGG)

High-level expression from one particular heterologous gene in Escherichia coli generally requires the optimization of codon usage. Genes encoding for Hepatitis C virus core protein (HCcAg), human interferon alpha2 and 8 subtypes (HUIFNalpha2 and HUIFNalpha8) show a high content of AGA/AGG codons. These are encoded by the product of the dnaY gene in E. coli. The proteins used in this work have a high therapeutic value and were used as models for studying the effects of these rare codons on the efficiency of heterologous gene expression in E. coli. Expression plasmids were constructed to express any of these proteins and the dnaY gene product simultaneously in E. coli. After dnaY gene expression, HCcAg, and HUIFNalpha2 expression levels increased 5 and 3 times, respectively. However, HUIFNalpha8 expression was barely detected either supplying or not the additional dnaY gene product. These results suggest that the high frequency of AGA/AGG codons present in the HCcAg and HUIFNalpha2 genes could be one of the factors limiting its expression in E. coli. Nevertheless, for HUIFNalpha8 it seems that other factors prevail upon the lack of dnaY product. Data presented here for HCcAg and HUIFNalpha2 expressions proved the value of this approach to obtain therapeutic proteins in E. coli.     

Expression and characterization of the HMG-CoA reductase of the thermophilic archaeon Sulfolobus solfataricus

The thermostable class I HMG-CoA reductase of Sulfolobus solfataricus offers potential for industrial applications and for the initiation of crystallization trials of a biosynthetic HMG-CoA reductase. However, of the 15 arginine codons of the hmgA gene that encodes S. solfataricus HMG-CoA reductase, 14 (93%) are AGA or AGG, the arginine codons used least frequently by Escherichia coli. The presence of these rare codons in tandem or in the first 20 codons of a gene can complicate expression of that gene in E. coli. Problems include premature chain termination and misincorporation of lysine for arginine. We therefore sought to improve the expression and subsequent yield of S. solfataricus HMG-CoA reductase by expanding the pool size of tRNA(AGA,AGG), the tRNA that recognizes these two rare codons. Coexpression of the S. solfataricus hmgA gene with the argU gene that encodes tRNA(AGA,AGG) resulted in an over 10-fold increase in enzyme yield. This has provided significantly greater quantities of purified enzyme for potential industrial applications and for crystallographic characterization of a stable class I HMG-CoA reductase. It has, in addition, facilitated determination of kinetic parameters and of pH optima for all four catalyzed reactions, for determination of the K(i) for inhibition by the statin drug mevinolin, and for comparison of the properties of the HMG-CoA reductase of this thermophilic archaeon to those of other class I HMG-CoA reductases.     

Functional characterization of arginine 30, lysine 40, and arginine 62 in Tn5 transposase

Three N-terminal basic residues of Tn5 transposase, which are associated with proteolytic cleavages by Escherichia coli proteinases, were mutated to glutamine residues with the goal of producing more stable transposase molecules. Mutation of either arginine 30 or arginine 62 to glutamine produced transposase molecules that were more stable toward E. coli proteinases than the parent hyperactive Tn5 transposase, however, they were inactive in vivo. In vitro analysis revealed these mutants were inactive, because both Arg(30) and Arg(62) are required for formation of the paired ends complexes when the transposon is attached to the donor backbone. These results suggest Arg(30) and Arg(62) play critical roles in DNA binding and/or synaptic complex formation. Mutation of lysine 40 to glutamine did not increase the overall stability of the transposase to E. coli proteinases. This mutant transposase was only about 1% as active as the parent hyperactive transposase in vivo; however, it retained nearly full activity in vitro. These results suggest that lysine 40 is important for a step in the transposition mechanism that is bypassed in the in vitro assay system, such as the removal of the transposase molecule from DNA following strand transfer.     

High-level expression of recombinant genes in Escherichia coli is dependent on the availability of the dnaY gene product

We have observed that proteins, such as human tissue-type plasminogen activator, pro-urokinase or gp41 of human immunodeficiency virus, which have a high content of rare codons in their respective genes, are not readily expressed in Escherichia coli. Furthermore induction of these heterologous genes leads to growth inhibition and plasmid instability. Supplementation with tRNA(AGA/AGG(Arg)) by cotransfection with the dnaY gene, which supplies this minor tRNA, resulted in high-level production with greatly improved cell viability and plasmid stability.     

Codon preference reflects mistranslational constraints: a proposal.

Following the observation of lysine for arginine misincorporation at the poor choice codon arg-AGA, a comparison of codon usage patterns for highly expressed mRNA's in E. coli provides a basis for the proposal that the major codon preference is subject to mistranslational constraints. In addition, the codons are utilized, as well as arranged, to provide a hydropathically conservative amino acid as the most probable replacement resulting from a mistranslational event.     

A relationship between asparagine synthetase A and aspartyl tRNA synthetase.

A highly conserved protein motif characteristic of Class II aminoacyl tRNA synthetases was found to align with a region of Escherichia coli asparagine synthetase A. The alignment was most striking for aspartyl tRNA synthetase, an enzyme with catalytic similarities to asparagine synthetase. To test whether this sequence reflects a conserved function, site-directed mutagenesis was used to replace the codon for Arg298 of asparagine synthetase A, which aligns with an invariant arginine in the Class II aminoacyl tRNA synthetases. The resulting genes were expressed in E. coli, and the gene products were assayed for asparagine synthetase activity in vitro. Every substitution of Arg298, even to a lysine, resulted in a loss of asparagine synthetase activity. Directed random mutagenesis was then used to create a variety of codon changes which resulted in amino acid substitutions within the conserved motif surrounding Arg298. Of the 15 mutant enzymes with amino acid substitutions yielding soluble enzyme, 13 with changes within the conserved region were found to have lost activity. These results are consistent with the possibility that asparagine synthetase A, one of the two unrelated asparagine synthetases in E. coli, evolved from an ancestral aminoacyl tRNA synthetase.     

An in vivo approach to identifying sequence context of 8-oxoguanine mutagenesis

Base substitution mutations are not distributed randomly in that most are located at a few specific hotspots sites. We have been studying 7,8-dihydro-8-oxoguanine mutagenesis in Escherichia coli in the supF gene carried in a plasmid. Among hotspots, guanine within the 5'-AGA-3' located in the anticodon site was susceptible to the induction of G:C-->T:A transversion. In this study, we constructed variants of the supF gene in which the hotspot 5'-AGA-3' was modified to 5'-AGT-3', 5'-AGG-3' and 5'-AGC-3' to determine the influence of 3' neighboring base on G:C-->T:A mutational activity. Using these variant supF genes propagated in a 7,8-dihydro-8-oxoguanine repair-deficient host, we found that guanine within 5'-AGA-3' and 5'-AGG-3' produce G:C-->T:A, but guanine within 5'-AGT-3' and 5'-AGC-3' reduce the formation of G:C-->T:A. These changes were thus due to the effect of sequence context on the efficiency of mutation formation at the sites of 7,8-dihydro-8-oxoguanine. We also observed a longer range base-pair effect on hotspot formation.     

Expression and Characterization of the HMG-CoA Reductase of the Thermophilic Archaeon Sulfolobus solfataricus

The thermostable class I HMG-CoA reductase of Sulfolobus solfataricus offers potential for industrial applications and for the initiation of crystallization trials of a biosynthetic HMG-CoA reductase. However, of the 15 arginine codons of the hmgA gene that encodes S. solfataricus HMG-CoA reductase, 14 (93%) are AGA or AGG, the arginine codons used least frequently by Escherichia coli. The presence of these rare codons in tandem or in the first 20 codons of a gene can complicate expression of that gene in E. coli. Problems include premature chain termination and misincorporation of lysine for arginine. We therefore sought to improve the expression and subsequent yield of S. solfataricus HMG-CoA reductase by expanding the pool size of tRNA^AGA,AGG, the tRNA that recognizes these two rare codons. Coexpression of the S. solfataricus hmgA gene with the argU gene that encodes tRNA^AGA,AGG resulted in an over 10-fold increase in enzyme yield. This has provided significantly greater quantities of purified enzyme for potential industrial applications and for crystallographic characterization of a stable class I HMG-CoA reductase. It has, in addition, facilitated determination of kinetic parameters and of pH optima for all four catalyzed reactions, for determination of the K_i for inhibition by the statin drug mevinolin, and for comparison of the properties of the HMG-CoA reductase of this thermophilic archaeon to those of other class I HMG-CoA reductases.     

Phosphonate biosynthesis: molecular cloning of the gene for phosphoenolpyruvate mutase from Tetrahymena pyriformis and overexpression of the gene product in Escherichia coli.

The phosphoenolpyruvate mutase gene from Tetrahymena pyriformis has been cloned and overexpressed in Escherichia coli. To our knowledge, this is the first Tetrahymena gene to be expressed in E. coli, a task made more complicated by the idiosyncratic codon usage by Tetrahymena. The N-terminal amino acid sequence of phosphoenolpyruvate mutase purified from T. pyriformis has been used to generate a precise oligonucleotide probe for the gene, using in vitro amplification from total genomic DNA by the polymerase chain reaction. Use of this precise probe and oligo(T) as primers for in vitro amplification from a T. pyriformis cDNA library has allowed the cloning of the mutase gene. A similar amplification strategy from genomic DNA yielded the genomic sequence, which contains three introns. The sequence of the DNA that encodes 10 amino acids upstream of the N-terminal sequence of the isolated protein was found by oligonucleotide hybridization to a subgenomic library. These 10 N-terminal amino acids are cleanly removed in Tetrahymena in vivo. The full mutase gene sequence codes for a protein of 300 amino acids, and it includes two amber (TAG) codons in the open reading frame. In Tetrahymena, TAG codes for glutamine. When the two amber codons are each changed to a glutamine codon (CAG) that is recognized by E. coli and the gene is placed behind a promoter driven by the T7 RNA polymerase, expression in E. coli is observed. The mutase gene also contains a large number of arginine AGA codons, a codon that is very rarely used by E. coli. Cotransformation with a plasmid carrying the dnaY gene [which encodes tRNA(Arg)(AGA)] results in more than 4-fold higher expression. The mutase then comprises about 25% of the total soluble cell protein in E. coli transformants. The mutase gene bears significant similarity to one other gene in the available data bases, that of carboxyphosphonoenolpyruvate mutase from Streptomyces hygroscopicus, an enzyme that catalyzes a closely related transformation. Due to the large evolutionary distance between Tetrahymena and Streptomyces, this similarity can be interpreted as the first persuasive evidence that the biosynthesis of phosphonates is an ancient metabolic process.     

Can tRNAs act as antisense RNA? The case of mutA and dnaQ

Many mutator genes have been characterized in E. coli, but the realization that mutA, the most recent mutator pathway described, encodes for a missense suppressor glycine tRNA caused a real surprise. The connection between expression of mutA and a 10 times increase in the spontaneous mutation rate is not readily explainable. The first attempt to describe the mechanism of action suggested a direct mistranslation of one subunit of polymerase III (PolIII) and the ideal candidate was the epsilon subunit carrying the 3'-->5' exonuclease activity. This subunit increases PolIII accuracy about 100 times. However, such direct mistranslation of epsilon was later ruled out when it became clear that all mutA cells express an error-prone form of PolIII. This result could not be reconciled with the very low level of mistranslation (1%) caused by mutA. But there is no need to invoke amino acid misincorporation in epsilon to destroy its activity. On the contrary, I suggest a new way to regulate epsilon amount, based on the reinterpretation of the mutA pathway through the new and puzzling observation that several tRNAs (including mutA which encodes for a glycine missense suppressor tRNA) are complementary to the 5' end of dnaQ mRNA. Accordingly, I propose that uncharged tRNAs can act as antisense RNAs, decreasing translation of dnaQ and possibly other genes. This could represent a new regulatory function for tRNAs and of course gives a direct and unrecognized link between starvation and mutation rate.     

Arginine-agmatine antiporter in extreme acid resistance in Escherichia coli

The process of arginine-dependent extreme acid resistance (XAR) is one of several decarboxylase-antiporter systems that protects Escherichia coli and possibly other enteric bacteria from exposure to the strong acid environment of the stomach. Arginine-dependent acid resistance depends on an intracellular proton-utilizing arginine alpha-decarboxylase and a membrane transport protein necessary for delivering arginine to and removing agmatine, its decarboxylation product, from the cytoplasm. The arginine system afforded significant protection to wild-type E. coli cells in our acid shock experiments. The gene coding for the transport protein is identified here as a putative membrane protein of unknown function, YjdE, which we now name adiC. Strains from which this gene is deleted fail to mount arginine-dependent XAR, and they cannot perform coupled transport of arginine and agmatine. Homologues of this gene are found in other bacteria in close proximity to homologues of the arginine decarboxylase in a gene arrangement pattern similar to that in E coli. Evidence for a lysine-dependent XAR system in E. coli is also presented. The protection by lysine, however, is milder than that by arginine.     

The increase by spermidine of fidelity of protamine synthesis in a wheat-germ cell-free system

The influence of spermidine on the fidelity of natural mRNA-directed protein synthesis has been investigated. With protamine mRNA as a template for protamine synthesis, misincorporation of lysine, histidine, threonine and cysteine for arginine was measured in the presence and absence of spermidine. It was found that misincorporation of these four amino acids in the presence of spermidine was less than or nearly equal to that occurring in the absence of spermidine; however, incorporation of arginine was stimulated greatly by spermidine. These results clearly show that spermidine induced an increase of fidelity in protamine synthesis. The increase of fidelity in the presence of spermidine occurred mainly at the level of binding of aminoacyl-tRNA to ribosomes. The frequency of misreading the 5' base of the codon (misincorporation of cysteine) was greater than that of the middle base of the codon (misincorporation of histidine), but spermidine reduction of misreading was more marked at the middle base of the codon. Misincorporation of lysine (misreading of G to A residue at the middle base of the codon) was greater than that of threonine (misreading of G to C residue), but spermidine reduction of misreading was more marked in the misincorporation of threonine. It was deduced from these results that spermidine inhibited low-frequency misreading more effectively than high-frequency misreading.