High-temperature cultivation and 5' mRNA optimization are key factors for the efficient overexpression of thermostable Deinococcus geothermalis purine nucleoside phosphorylase in Escherichia coli

Overexpression of genes from thermophiles in Escherichia coli is an attractive approach towards the large-scale production of thermostable biocatalysts. However, various factors can challenge efficient heterologous protein expression--one example is the formation of stable 5' mRNA secondary structures that can impede an efficient translation initiation. In this work, we describe the expression optimization of purine nucleoside phosphorylase from the thermophilic microbe Deinococcus geothermalis in E. coli. Poor expression levels caused by stable secondary 5' mRNA structure formation were addressed by two different approaches: (i) increasing the cultivation temperature above the range used typically for recombinant protein expression and (ii) optimizing the 5' mRNA sequence for reduced secondary structures in the translation initiation region. The increase of the cultivation temperature from 30°C to 42°C allowed a more than 10-fold increase of activity per cell and optimizing the 5' mRNA gene sequence further increased the activity per cell 1.7-fold at 42°C. Thus, the combination of high-temperature cultivation and 5' sequence optimization is described as an effective approach to overcome poor expression levels resulting from stable secondary 5' mRNA structure formation. We suggest that this method is especially suitable for improving the expression of proteins derived from thermophiles in E. coli.     

A novel bicistronic vector for overexpressing Mycobacterium tuberculosis proteins in Escherichia coli

A putative DNA glycosylase encoded by the Rv3297 gene (MtuNei2) has been identified in Mycobacterium tuberculosis. Our efforts to express this gene in Escherichia coli either by supplementing tRNAs for rare codons or optimizing the gene with preferred codons for E. coli resulted in little or no expression. On the other hand, high-level expression was observed using a bicistronic expression vector in which the target gene was translationally coupled to an upstream leader sequence. Further comparison of the predicted mRNA secondary structures supported the hypothesis that mRNA secondary structure(s) surrounding the translation initiation region (TIR), rather than codon usage, played the dominant role in influencing translation efficiency, although manipulation of codon usage or tRNA supplementation did further enhance expression in the bicistronic vector. Addition of a cleavable N-terminal tag also facilitated gene expression in E. coli, possibly through a similar mechanism. However, since cleavage of N-terminal tags is determined by the amino acid at the P₁′ position downstream of the protease recognition sequence and results in the addition of an extra amino acid in front of the N-terminus of the protein, this strategy is not particularly amenable to Fpg/Nei family DNA glycosylases which carry the catalytic proline residue at the P₁′ position and require a free N-terminus. On the other hand, the bicistronic vector constructed here is potentially valuable particularly when expressing proteins from G/C rich organisms and when the proteins carry proline residues at the N-terminus in their native form. Thus the bicistronic expression system can be used to improve translation efficiency of mRNAs and achieve high-level expression of mycobacterial genes in E. coli.     

Shine-Dalgarno-like sequences are not required for translation of chloroplast mRNAs in Chlamydomonas reinhardtii chloroplasts or in Escherichia coli

Initiation of translation in Escherichia coli and related eubacteria involves well-defined interactions between a conserved Shine-Dalgarno (SD) sequence immediately upstream of the initiation codon in the mRNA leader and an equally conserved anti-SD sequence at the 3′ end of the 16S rRNA. SD-like sequences found in the leaders of many, but not all, mRNAs from cyanobacteria and chloroplasts are hypervariable in location, size, and base composition compared to those in E. coli, while anti-SD sequences in the respective 16S rRNAs remain highly conserved. We have examined the function of the SD-like sequences found in the leaders of four chloroplast genes of the green alga Chlamydomonas reinhardtii using replacement mutagenesis to eliminate complementarity with the anti-SD sequences and insertion of canonical SD sequences (GGAGG) at positions −9 to −5 relative to the initiation codon. Promoter-leader regions of the atpB, atpE, rps4, and rps7 genes representing the diversity of chloroplast SD-like sequences were fused to aadA and uidA reporter genes encoding spectinomycin resistance and GUS activity respectively. Analysis of chloroplast transformants of C. reinhardtii and transformants of E. coli carrying the wild-type and mutant reporter constructs revealed that mutagenic replacement of the putative SD sequences had no effect on the expression of either the aadA or uidA reporter genes. Chloroplast transformants with the canonical SD sequence also showed no differences in reporter gene expression, whereas expression of the reporter genes was increased by 10 to 30% in the E. coli transformants. Collectively our results suggest that even though SD-dependent initiation predominates in E. coli, this bacterium also has the capacity to initiate translation by an SD-independent mechanism. In contrast, plant chloroplasts, and very probably their cyanobacterial ancestors, appear to have adopted the SD-independent mechanism for translational initiation of most mRNAs. Received: 8 July 1997 / Accepted: 9 September 1997     

Viral and chloroplastic signals essential for initiation and efficiency of translation in Agrobacterium tumefaciens

The construction of high-level protein expression vectors using the CaMV 35S promoter in concert with highly efficient translation initiation signals for Agrobacterium tumefaciens is a relatively less explored field compared to that of Escherichia coli. In the current study, we experimentally investigated the capacity of the CaMV 35S promoter to direct GFP gene expression in A. tumefaciens in the context of different viral and chloroplastic translation initiation signals. GFP expression and concomitant translational efficiency was monitored by confocal microscopy and Western blot analysis. Among all of the constructs, the highest level of translation was observed for the construct containing the phage T7 translation initiation region followed by the chloroplastic Rubisco Large Subunit (rbcL) 58-nucleotide 5′ leader region including its SD-like sequence (GGGAGGG). Replacing the SD-like (GGGAGGG) with non SD-like (TTTATTT) or replacing the remaining 52 nucleotides of rbcL with nonspecific sequence completely abolished translation. In addition, this 58 nucleotide region of rbcL serves as a translational enhancer in plants when located within the 5′ UTR of mRNA corresponding to GFP. Other constructs, including those containing sequences upstream of the coat proteins of Alfalfa Mosaic Virus, or the GAGG sequence of T4 phage or the chloroplastic atpI and/or PsbA 5′ UTR sequence, supported low levels of GFP expression or none at all. From these studies, we propose that we have created high expression vectors in A. tumefaciens and/or plants which contain the CaMV 35S promoter, followed by the translationally strong T7 SD plus RBS translation initiation region or the rbcL 58-nucleotide 5′ leader region upstream of the gene of interest.     

Sequence analysis of rice rubi3 promoter gene expression cassettes for improved transgene expression

In a construct containing a GUS reporter gene driven by the 5′ regulatory elements from rubi3, expression was enhanced 4-fold when a 20-nucleotide (nt) GUS 5′ untranslated sequence was replaced with 9 nt sequences derived from rubi3′s second exon. The roles of the sequences immediately upstream from the GUS translation initiation codon, and their significance in gene expression, were investigated. Sequence analysis suggests that complementarity between sequences immediately 5′ of a translation initiation codon and the rice 17S rRNA may be responsible for the reduction in protein levels from constructs containing the GUS leader sequence. The results demonstrate an affect sequences immediately upstream from transgenic coding sequences have on expression, and when using the rubi3 5′ regulatory sequence in particular.     

Factors regulating cryIVB expression in the cyanobacterium Synechococcus PCC 7942

The expression of the larvicidal Bacillus thuringiensis subsp. israelensis cryIVB gene in cyanobacteria has been suggested to be an effective means of controlling mosquito populations. Using a variety of cryIVB constructs, in this study we have examined the effect of Synechococcus PCC 7942 culture age on intracellular toxin levels and have attempted to determine the mechanisms by which cryIVB gene expression is regulated. The data suggest that specific degradation of the cryIVB mRNA limits toxin production; however, the addition of cyanobacterial 3 untranslated DNA sequences to the cryIVB gene did not improve mRNA stability or toxin levels. An analysis of the cryIVB sequence and comparison of codon usage patterns with highly expressed cyanobacterial genes suggest that inefficient translation and intragenic ribosomal binding sites impede protein synthesis and result in rapid turnover of the toxin mRNA.     

Ribosome-binding sites on chloroplastrbcL andpsbA mRNAs and light-induced initiation of D1 translation

Chloroplast ribosome-binding sites were identified on the plastidrbcL andpsbA mRNAs using toeprint analysis. TherbcL translation initiation domain is highly conserved and contains a prokaryotic Shine-Dalgarno (SD) sequence (GGAGG) located 4 to 12 nucleotides upstream of the initiator AUG. Toeprint analysis ofrbcL mRNA associated with plastid polysomes revealed strong toeprint signals 15 nucleotides downstream from the AUG indicating ribosome binding at the translation initiation site.Escherichia coli 30S ribosomes generated similar toeprint signals when mixed withrbcL mRNA in the presence of initiator tRNA. These results indicate that plastid SD sequences are functional in chloroplast translation initiation. ThepsbA initiator region lacks a SD sequence within 12 nucleotides of the initiator AUG. However, toeprint analysis of soluble and membrane polysome-associatedpsbA mRNA revealed ribosomes bound to the initiator region.E. coli 30S ribosomes did not associate with thepsbA translation initiation region.E. coli and chloroplast ribosomes bind to an upstream region which contains a conserved SD-like sequence. Therefore, translation initiation onpsbA mRNA may involve the transient binding of chloroplast ribosomes to this upstream SD-like sequence followed by scanning to localize the initiator AUG. Illumination 8-day-old dark-grown barley seedlings caused an increase in polysome-associatedpsbA mRNA and the abundance of initiation complexes bound topsbA mRNA. These results demonstrate that light modulates D1 translation initiation in plastids of older dark-grown barley seedlings.     

Structure, evolution and expression of the mitochondrial ADP/ATP translocator gene from Chlamydomonas reinhardtii

Summary The first AUG in the Chlamydomonas reinhardtii ADP/ATP translocator (CRANT) mRNA initiates an open reading frame (ORF) which is very similar (51–79% amino acid identity) to other ANT proteins. In contrast to higher plants, no evidence for a long amino-terminal extension was obtained. The 5 non-transcribed region of the single-copy CRANT gene contains sequence motifs present in other C. reinhardtii nuclear genes. Four introns, whose positions are not conserved in other ANT genes, interrupt the protein coding region. A short heat shock specifically reduces CRANT mRNA levels. CRANT mRNA levels were unaffected by a mutation in photosynthesis. In a dark/light regime CRANT mRNA levels are high in the dark phase and low in the early light phase. Data on translation initiation sites, splice junctions and the codon preferences of C. reinhardtii nuclear genes were compiled. With the exception of two rare codons, ACA and GGA, the CRANT gene exhibits the biased codon usage of C. reinhardtii nuclear genes that are highly expressed during normal vegetative growth.     

Adaptive response to cold temperatures and characterization of cspA in Salmonella typhimurium LT2

Salmonella typhimurium is a major foodborne microbial pathogen which primarily contaminates poultry products causing salmonellosis in humans. S. typhimurium LT2 cultures, when transferred from 37 °C to 5 °C or 10 °C, showed an initial lag period in growth with an approximate generation time of 10–25 h. Western blot assay using E. coli CS7.4 antibody and analysis of radiolabeled total cellular proteins from S. typhimurium cultures after exposure to 10 °C or 5 °C showed elevated expression of a major cold shock protein, CS7.4. Identification of a decreased level of CS7.4 at 37 °C suggests that the expression of this protein may require a large temperature downshift. Putative regulatory protein binding segment on the 5-untranslated region referred as Fragment 7 in S. typhimurium exhibited a 90.6% and a 56.25% nucleotide sequence identity when compared with the Fragment 7 of E. coli and S. enteritidis, respectively. The differences in the nucleotide sequence within the Fragment 7 between S. typhimurium and S. enteritidis may explain the differential expression of CspA at 37 °C. The nucleotide sequence of the open reading frame of S. typhimurium cspA gene showed a single base difference at 816 bp position from a G to a C which altered the amino acid residue from a glycine to an alanine. In addition to CspA, an elevated expression of a 105 kDa, and decreased expression of 6 proteins were evidenced when cultures of S. typhimurium were exposed to 10 °C or 5 °C. Differential expression of the CspA and other proteins in S. typhimurium following exposure to cold temperatures suggest that adaptation and continued growth and survival at cold temperatures in this pathogen may be aided by these cold-responsive proteins.     

The sequence surrounding the translation initiation codon of the pea plastocyanin gene increases translational efficiency of a reporter gene

The 5-upstream region of the pea plastocyanin gene (petE) directed 5–10-fold higher levels of -glucuronidase (GUS) activity than the cauliflower mosaic virus 35S promoter in transgenic tobacco plants, although the levels of GUS mRNA were similar. The sequence (AAAAAUGG) around the translation initiation codon of petE enhanced translation of the GUS mRNA 10-fold compared to translation from the GUS translation initiation codon in transgenic tobacco plants and transfected protoplasts.     

The translation of recombinant proteins in E. coli can be improved by in silico generating and screening random libraries of a −70/+96 mRNA region with respect to the translation initiation codon

Recombinant protein translation in Escherichia coli may be limited by stable (i.e. low free energy) secondary structures in the mRNA translation initiation region. To circumvent this issue, we have set-up a computer tool called ‘ExEnSo’ (Expression Enhancer Software) that generates a random library of 8192 sequences, calculates the free energy of secondary structures of each sequence in the −70/+96 region (base 1 is the translation initiation codon), and then selects the sequence having the highest free energy. The software uses this ‘optimized’ sequence to create a 5′ primer that can be used in PCR experiments to amplify the coding sequence of interest prior to sub-cloning into a prokaryotic expression vector. In this article, we report how ExEnSo was set-up and the results obtained with nine coding sequences with low expression levels in E. coli. The free energy of the −70/+96 region of all these coding sequences was increased compared to the non-optimized sequences. Moreover, the protein expression of eight out of nine of these coding sequences was increased in E. coli, indicating a good correlation between in silico and in vivo results. ExEnSo is available as a free online tool.     

The wild-type allele of tonB in Escherichia coli is dominant over the tonB1 allele, encoding TonBQ160K, which suppresses the btuB451 mutation

The entire coding sequence of the tonB gene, except for nine codons at the 3 end, was deleted from the chromosome of Escherichia coli. Introduction of the btuB451 suppressor mutant tonB1 into the chromosome of such a tonB deletion strain showed that the tonB1 allele was active as a suppressor in a single copy at 37° C and 42° C but not at 28° C. No temperature dependence was seen when FepA- or FhuA-dependent activities of the tonB1 gene product (TonB_Q160K) were tested. The btuB451 suppressor activity of tonB1 was inhibited by the simultaneous presence within the cells of the tonB ⁺ allele on a multicopy plasmid. This represents the first case of dominance among different tonB alleles. Inhibition of suppression was abolished by overexpression of the btuB451-encoded receptor protein. Competition for binding of TonB⁺ and TonB_Q150K to ExbB was excluded as the cause of dominance. Based on our data we conclude that competition for binding of TonB ⁺ and TonB_Q160K to the btuB451 gene product is the reason for the observed dominance. The implications of these findings for the mechanism of btuB451 suppression by tonB1 are discussed.     

The virB operon of Agrobacterium tumefaciens pTiC58 encodes 11 open reading frames

Summary Agrobacterium tumefaciens genetically transforms plant cells by transferring a copy of its T-DNA to the plant where it is integrated and stably maintained. In the presence of wounded plant cells this process is activated and mediated by the products of the vir genes which are grouped into six distinct loci. The largest is the virB locus spanning 9.5 kb. Transposon mutagenesis studies have shown that virB gene products are required for virulence but their functions remain largely unknown. To provide information relevant to understanding the function of VirB polypeptides, the nucleotide sequence of the virB operon from a nopaline plasmid, pTiC58, is presented here. Eleven open reading frames (ORFs) are predicted from this sequence. The predicted sizes of 10 of the 11 VirB polypeptides are verified by specific expression in Escherichia coli. Only the product of the smallest ORF potentially encoding a 5.8 kDa polypeptide has not been detected. The initiation of translation of five virB ORFs occurs at codons that overlap the termination codons of the ORF immediately upstream; thus, translational coupling may be an important mechanism for efficient translation of the large virB polycistronic mRNA. Based on hydropathy plot analysis nine of the virB ORFs encode proteins that may interact with membranes; these data support the earlier hypothesis (Engstromm et al. 1987) that virB gene products may form a membrane pore or channel to mediate exit of the T-DNA copy (T-strands) from Agrobacterium into the plant cell. A comparison of the two published octopine virB sequences with the nopaline sequence presented here is made.     

High-level expression of the xylanase from <Emphasis Type="Italic">Thermomyces lanuginosus</Emphasis> in <Emphasis Type="Italic">Escherichia coli</Emphasis>

According to the amino acid sequence, a codon-optimized xylanase gene (xynA1) from Thermomyces lanuginosus DSM 5826 was synthesized to construct the expression vector pHsh-xynA1. After optimization of the mRNA secondary structure in the translational initiation region of pHsh-xynA1, free energy of the 70 nt was changed from −6.56 to −4.96 cal/mol, and the spacing between AUG and the Shine-Dalgarno sequence was decreased from 15 to 8 nt. The expression level was increased from 1.3 to 13% of total cell protein. A maximum xylanase activity of 47.1 U/mL was obtained from cellular extract. The recombinant enzyme was purified 21.5-fold from the cellular extract of Escherichia coli by heat treatment, DEAE-Sepharose FF column and t-Butyl-HIC column. The optimal temperature and pH were 65 °C and pH 6.0, respectively. The purified enzyme was stable for 30 min over the pH range of 5.0–8.0 at 60 °C, and had a half-life of 3 h at 65 °C.     

Gel shift selection of translation enhancer sequences using messenger RNA display

We designed a new approach for selection of translation enhancer sequences that enables efficient protein synthesis in cell-free systems. The selection is based on a gel shift assay of a messenger RNA (mRNA)–protein fusion product that is synthesized in a cell-free translation system using an mRNA display method. A library of randomized 20-nt-long sequences, with all possible combinations of the four nucleotides, upstream of a coding region was screened by successive rounds of screening in which the translation time of the succeeding round was reduced compared with the previous round. An efficient translation enhancer sequence capable of more rapid initiation of cell-free protein synthesis, with a minimal translation time of 5 min, than a natural longer enhancer sequence (Xenopus β-globin 5′UTR) was selected using rabbit reticulocyte extract as a model cell-free translation system. Furthermore, a successful screening of cap-independent translation enhancer sequence and a significant sequence similarity of the selected candidates validated the efficiency of the combined mRNA display and gel shift assay method for the rapid development of advanced cell-free translation systems.     

Isolation of a cDNA fragment coding for Chlamydomonas reinhardtii ferredoxin and expression of the recombinant protein in Escherichia coli

A cDNA clone coding for mature C. reinhardtii ferredoxin has been isolated from a cDNA library using PCR and two oligonucleotide primers based on the N- and C-termini of the protein's amino acid sequence. The nucleotidic sequence of the PCR fragment (299 bp) agreed well with the amino acid sequence since a single conservative substitution (Thr-7 to Ser) could be deduced. The PCR fragment was inserted into the expression vector pTrc 99A, using the incorporated NcoI and BamHI restriction sites and the construction used to transform E. coli (DH5α F′). After subsequent large scale expression and purification of the recombinant protein, biochemical and biophysical analysis have indicated that the product isolated from E. coli is homologous to native ferredoxin isolated from green algae.