Optimized bacterial production of nonglycosylated human transferrin and its half-molecules.

Transferrin is a glycoprotein functioning in iron transport in higher eukaryotes, and consists of two highly homologous domains. To study the function of the glycan residues attached exclusively to the C-terminal domain, we have constructed a plasmid allowing production of nonglycosylated human transferrin in Escherichia coli. By molecular biological and genetic techniques, production was stepped up to 60 mg/l. Similar plasmids were constructed for production of the two half-transferrins. The recombinant proteins accumulate in inclusion-body-like aggregates, where they appear to bind iron without causing bacteriostasis. Proteins active in iron binding have been purified from these inclusion bodies.

CCC.UGA: a new site of ribosomal frameshifting in Escherichia coli.

To activate expression of a human transferrin (Tf)-encoding cDNA in Escherichia coli by translational coupling, it was placed in an expression plasmid downstream from a 5'-terminal fragment from the replicase (R)-encoding gene of bacteriophage MS2. The resulting construct was found to produce, besides the desired Tf, a protein with the mobility of a fusion product (RTf) of the N-terminal R fragment and Tf. Analysis of available mutants showed that this fusion results from +1 ribosomal frameshifting at the end of the R reading frame. This region contains the sequence, CCC.UGA, suggesting that before termination occurs, tRNA(Pro) may dislodge from the CCC codon and reassociate with the +1 triplet CCU. By further site-directed mutagenesis, we demonstrate that both the CCC codon and the termination codon are indeed required for the observed 2-4% frameshifting. When either triplet is changed, the frequency of frameshifting drops to 0.3% or less. These results classify CCC.UGA as a new '+1 shifty stop'.

Is there a special function for U.G basepairs in ribosomal RNA?

U.G basepairs are well-established elements of RNA structure. The geometry of this pair is different, however, from classical Watson-Crick basepairs. This leads to an unusual stacking of the basepair: overlap with the basepair at the 5' side of the U (and the 3' side of the G) is strong (stacked) while it is weak with the basepair on the other side (destacked). The closure of an RNA helix by a U.G pair will be energetically unfavourable when the U residue occupies the 5' end. In transfer RNA there is a strong selection against a 'destacked' U.G pair at helix ends. In the 16S rRNA model of Escherichia coli there are 72 U.G pairs of which 36 or 22 occupy a helix end, depending on how such an end is defined. There is a slight preference for 'stacked' U.G's in these positions. It is remarkable, however, that of 13 very conserved U.G pairs in the 16S (-like) rRNA, 7 occur at helix ends and that 5 of these have the 'destacked' configuration. It is suggested that these pairs, if they exist at all in a hydrogen-bounded form, are stabilized by co-axial stacking with other helices or by interaction with a protein.

Site-specific mutation of the conserved m6(2)A m6(2)A residues of E. coli 16S ribosomal RNA. Effects on ribosome function and activity of the ksgA methyltransferase.

In vitro synthesis of mutant 16S RNA and reconstitution with ribosomal proteins into a mutant 30S ribosome was used to make all possible single base changes at the universally conserved A1518 and A1519 residues. All of the mutant RNAs could be assembled into a ribosomal subunit which sedimented at 30 S and did not lack any of the ribosomal proteins. A series of in vitro tests of protein synthesis ability showed that all of the mutants had some activity. The amount varied according to the assay and mutant, but was never less than 30% and was generally above 50%. Therefore, neither the conserved A1518 nor A1519 residues are essential for ribosome function. The mutant ribosomes could also be methylated by the ksgA methyltransferase to 70-120% of the expected amount. Thus, neither of the A residues is required for methylation of the other, ruling out any obligate order of methylation of A1518 and A1519.

Conformational and thermodynamic effects of naturally occurring base methylations in a ribosomal RNA hairpin of Bacillus stearothermophilus.

The 3'-terminal colicin fragments of 16S ribosomal RNA were isolated from Bacillus stearothermophilus and from its kasugamycin-resistant (ksgA) derivative lacking N6-dimethylation of the two adjacent adenosines in a hairpin loop. The fragment from the ksgA strain still contains a naturally occurring N2-methylguanosine in the loop. An RNA molecule resembling the B. stearothermophilus colicin fragment but without modified nucleosides was synthesized in vitro using a DNA template and bacteriophage T7 RNA polymerase. Proton-NMR spectra of the RNAs were recorded at 500 MHz. The imino-proton resonances of base-paired G and U residues could be assigned on the basis of previous NMR studies of the colicin fragment of Escherichia coli and by a combination of methylation-induced shifts and thermal melting of base pairs. The assignments were partly confirmed by NOE measurements. Adenosine dimethylation in the loop has a distinct conformational effect on the base pairs adjoining the loop. The thermal denaturation melting curve of the enzymatically synthesized RNA fragment was also determined and the transition midpoint (tm) was found to be 73 degrees C at 15 mM Na+. A comparison with previously determined thermodynamic parameters for various colicin fragments demonstrates that base methylations in the loop lead to a relatively strong destabilization of the hairpin helix. In terms of free energy the positive contribution of the methylations are in the order of the deletion of one base pair from the stem. Other data show that recently published free-energy parameters do not apply for certain RNA hairpins.

Autogenous regulation of the Escherichia coli ksgA gene at the level of translation.

Various plasmids that contain the Escherichia coli ksgA gene, which encodes a 16S rRNA adenosine dimethyltransferase (methylase), were constructed. In one of these plasmids, the DNA encoding the N-terminal part of the methylase was fused to the lacZ gene, and in another construct, the ksgA gene contained a deletion which resulted in a truncated version of the methylase. When a cell contained one plasmid directing the synthesis of the intact, active methylase and another plasmid encoding the methylase-beta-galactosidase protein, production of the latter product became strongly reduced. Likewise, synthesis of the truncated version of the methylase was diminished when the cell at the same time contained a plasmid producing the complete enzyme. These results were partly substantiated by in vitro experiments with a coupled transcription-translation assay system. By using a recently developed gel electrophoresis system for measuring protein-nucleic acid interactions, a specific binding of the ksgA methylase with its own mRNA could be established. Our results demonstrate that the expression of the ksgA gene can be, at least partly, autogenously controlled at the level of translation.

Increased kasugamycin sensitivity in Escherichia coli caused by the presence of an inducible erythromycin resistance (erm) gene of Streptococcus pyogenes.

An inducible erythromycin resistance gene (erm) of Streptococcus pyogenes was introduced into Escherichia coli by transformation with a plasmid. The recipient E. coli cells were either kasugamycin sensitive (wildtype) or kasugamycin resistant (ksgA). The MIC values of erythromycin increased from 150 micrograms/ml to greater than 3000 micrograms/ml for E. coli. An extract of transformed cells, particularly a high-salt ribosomal wash, contained an enzyme that was able to methylate 23S rRNA from untransformed cells in vitro; however, 23S rRNA from transformed cells was not a substrate for methylation by such an extract. 165 rRNA and 30S ribosomal subunits of either the wild type or a kasugamycin resistant (ksgA) mutant were not methylated in vitro. Transformation of E. coli by the erm-containing plasmid led to a reduction of the MIC values for kasugamycin. This happened in wild-type as well as in ksgA cells. However, in vitro experiments with purified ksgA encoded methylase demonstrated that also in erm transformed E. coli, the ksgA encoded enzyme was active in wild-type, but not in ksgA cells. It was also shown by in vitro experiments that ribosomes from erm ksgA cells have become sensitive to kasugamycin. Our experiments show that in vivo methylation of 23S rRNA, presumably of the adenosine at position 2058, leads to enhanced resistance to erythromycin and to reduced resistance to kasugamycin. This, together with previous data, argues for a close proximity of the two sites on the ribosome that are substrates for adenosine dimethylation.

The 3' terminal colicin fragment of Escherichia coli 16S ribosomal RNA. Conformational details revealed by enzymic and chemical probing.

The conformation of the colicin fragment of E. coli 16S rRNA was probed with various nucleases and with the adenosine-specific reagent diethylpyrocarbonate (DEP). The results confirm the presence of a stable central hairpin in the colicin fragment and a weaker additional secondary structure involving the regions 5' and 3' to this hairpin. By monitoring DEP accessibility at various stages of heat-denaturation sequential unfolding of individual base pairs was followed. In agreement with previous results it could be shown that dimethylation of the two adjacent adenosines in the hairpin loop (a feature in virtually all ribosomes) leads to a destabilization of the hairpin helix. Accessibilities of G residues, involved in the weaker additional secondary structure is anomalous. One G residue is sensitive to the single strand specific RNase T1 and insensitive to DEP, while the situation is reversed for the adjoining G residue. The strong reaction of the latter G-residue with DEP is unusual and indicates a very special conformation.

Characterization of the ksgA gene of Escherichia coli determining kasugamycin sensitivity.

In the plasmid pUC8ksgA7, the coding region of the ksgA gene is preceded by the lac promoter (Plac) and a small open reading frame (ORF). This ORF of 15 codons is composed of nucleotides derived from the lacZ gene, a multiple cloning site and the ksgA gene itself. The reading frame begins with the ATG initiation codon of lacZ and ends a few nucleotides beyond the ATG start codon of ksgA. The ksgA gene is not preceded by a Shine-Dalgarno (SD) signal. Cells transformed with pUC8ksgA7 produce active methylase, the product of the ksgA gene. Introduction of an in-phase TAA stop codon in the small ORF abolishes methylase production in transformed cells. On the plasmid pUC8ksgA5, which contains the entire ksgA region, the promoter of the ksgA gene was found to reside in a 380 base pair Bgl1-Pvu2 restriction fragment, partly overlapping the ksgA gene, by two independent methods. Cloning of this fragment in front of the galK gene in plasmid pKO1 stimulates galactokinase activity in transformants and its insertion into the expression vector pKL203 makes beta-galactosidase synthesis independent of the presence of Plac. The sequence of the Bgl1-Pvu2 fragment was determined and a putative promoter sequence identified. An SD signal could not be distinguished at a proper distance upstream from the ksgA start codon. Instead, an ORF of 13 codons starting with ATG in tandem with an SD signal and ending 4 codons ahead of the ksgA gene was identified. This suggests that translation of the ORF is required for expression of the ksgA gene.(ABSTRACT TRUNCATED AT 250 WORDS)

Influence of the codon following the AUG initiation codon on the expression of a modified lacZ gene in Escherichia coli.

In a lacZ expression vector (pMC1403Plac), all 64 codons were introduced immediately 3' from the AUG initiation codon. The expression of the second codon variants was measured by immunoprecipitation of the plasmid-coded fusion proteins. A 15-fold difference in expression was found among the codon variants. No distinct correlation could be made with the level of tRNA corresponding to the codons and large differences were observed between synonymous codons that use the same tRNA. Therefore the effect of the second codon is likely to be due to the influence of its composing nucleotides, presumably on the structure of the ribosomal binding site. An analysis of the known sequences of a large number of Escherichia coli genes shows that the use of codons in the second position deviates strongly from the overall codon usage in E. coli. It is proposed that codon selection at the second position is not based on requirements of the gene product (a protein) but is determined by factors governing gene regulation at the initiation step of translation.

Secondary structure as primary determinant of the efficiency of ribosomal binding sites in Escherichia coli.

Using a previously described vector (pKL203) we fused several heterologous ribosomal binding sites (RBSs) to the lacZ gene of E. coli and then studied the variation in expression of the fusions. The RBSs originated from bacteriophage Q beta and MS2 genes and the E. coli genes for elongation factor EF-Tu A and B and ribosomal protein L11 (rplK). The synthesis of the lacZ fusion proteins was measured by an immuno precipitation method and found to vary at least 100-fold. Lac-specific mRNA synthesis follows the variation in protein production. It appears that there is a correlation between the efficiency of an RBS to function in the expression of the fused gene and the lack of secondary structure, involving the Shine and Dalgarno nucleotides (SDnts) and/or the initiation codon. This efficiency is context dependent. The sequence of the SD nts and the length and sequence of the spacer region up to the initiation codon alone are not able to explain our results. Deletion mutations, created in the phage Q beta replicase RBS, reveal a complex pattern of control of expression, probably involving the use of a "false" initiation site.

Circular dichroism and 500-MHz proton magnetic resonance studies of the interaction of Escherichia coli translational initiation factor 3 protein with the 16S ribosomal RNA 3' cloacin fragment.

The RNA helix destabilizing properties of Escherichia coli initiation factor 3 protein (IF3), and its affinity for an evolutionarily conserved sequence at the 3' end of 16S rRNA, led us to examine the details of the protein-nucleic acid interactions upon IF3 binding to the 49-nucleotide 3'-terminal cloacin DF13 fragment of 16S rRNA by studying the circular dichroism (CD) and proton magnetic resonance spectra of the RNA, the protein, and their complex. In a physiological tris(hydroxymethyl)aminomethane buffer, where the interaction is primarily nonionic and sequence specific, addition of IF3 decreases the RNA 268-nm CD peak hyperbolically by 19% to an end point of about one IF3 per RNA strand. The titration curve is best fit by an association constant of (1.80 +/- 0.05) X 10(7) M-1, within the range estimated by a nuclease mapping study of the same system [Wickstrom, E. (1983) Nucleic Acids Res. 11, 2035-2052]. In a low-salt phosphate buffer without Mg2+, where the interaction is primarily ionic and nonspecific, titration with IF3 decreases the peak CD sigmoidally by 35% to an end point of two IF3 per strand. The titration curve is best fit by an intrinsic association constant of (1.7 +/- 0.7) X 10(6) M-1 for each IF3 and a cooperativity constant of 33 +/- 6. In a physiological phosphate buffer lacking Mg2+, the dispersion of aromatic proton magnetic resonance peaks and upfield-shifted methyl proton resonances indicates a high degree of secondary and tertiary structure in the protein. In an equimolar mixture of IF3 and RNA cloacin fragment, several changes in identifiable IF3 and RNA resonances are observed.(ABSTRACT TRUNCATED AT 250 WORDS)

Formylmethionyl-tRNA binding to 30 S ribosomes programmed with homopolynucleotides and the effect of translational initiation factor 3.

Binding of the polynucleotides poly(U), poly(X) and poly(dT) to 30 S ribosomes of Escherichia coli triggers IF2-dependent binding of initiator-tRNA (fMet-tRNA) to these particles. Poly(A) and poly(C) are inactive. A minimum chain-length of approximately 100 residues in poly(U) is required for full activity in fMet-tRNA binding, although much shorter polymers bind tightly to 30 S particles and do stimulate the binding of acPhe-tRNA. The stimulation of fMet-tRNA binding to 30 S ribosomes is strongly reduced under conditions where the polynucleotides adopt secondary structure. Complexes containing fMet-tRNA and the non-cognate codon UUU or XXX are destabilized by IF3, whereas the formation of such a complex containing an AUG codon is slightly enhanced by the factor. Consistent with previous observations, it was found that all model initiation complexes containing acPhe-tRNA are strongly destabilized by IF3, even when the cognate codon (UUU) is present. Our results suggest that IF3 counteracts 'unnatural' initiation events in vitro and suggest a regulatory role for this factor in vivo.

Effects of GUG and AUG initiation codons on the expression of lacZ in Escherichia coli.

We have replaced the ribosomal binding site (RBS) of the lacZ gene of E. coli by those of the maturation (A) gene of phage MS2 and that of the tufA gene. Both RBSs contain a GUG initiation codon. The expression with the tufA RBS is at least 25-fold higher than with the phage RBS. Changing the GUG into AUG results in a 3-fold increase in expression in both cases. In general, higher expression is accompanied by an increase of lac-specific mRNA. It is argued that this is a consequence of the more efficient translation of the mRNA.

Nucleotide sequence of the ksgA gene of Escherichia coli: comparison of methyltransferases effecting dimethylation of adenosine in ribosomal RNA.

The ksgA gene of Escherichia coli encodes a methyltransferase (MeT) that specifically dimethylates two adjacent adenosines near the 3' end of 16S RNA in the 30S particle. Its inactivation leads to kasugamycin (Ksg) resistance. Several plasmids were constructed with inserts which complemented chromosomal ksgA mutations. One of these inserts was sequenced and found to contain an open reading frame (ORF) sufficient to code for the previously identified 30-kDal MeT. We have compared the amino acid (aa) sequence of the ksgA-encoded enzyme with three published sequences of MeT involved in dimethylation of an adenosine residue in 23S RNA and rendering the organisms resistant to the MLS antibiotics. The homologous patches in the sequences of all four enzymes suggest that those might correspond to contact points for the common substrates, e.g., for the adenosine residue(s) and S-adenosylmethionine (SAM).

Effects of heterologous ribosomal binding sites on the transcription and translation of the lacZ gene of Escherichia coli.

A vector (pKL203) was constructed which contains the promoter-operator region of the lacZ gene and the major part of the coding sequence of the lac operon. The lacZ translation initiation signals [Shine-Dalgarno (SD) sequence and AUG codon] were deleted, and in their place a synthetic linker sequence was inserted, providing single restriction sites for SmaI and BamHI. With this vector constructions were made in which initiation signals of other prokaryotic genes (phage MS2 maturation protein, phage Q beta A2 gene and tufB gene) were fused to the lacZ gene, giving rise to various fusion proteins. The introduction of N-terminal amino acids (aa) in beta-galactosidase (beta-gal) which differ from the wild-type aa invariably leads to an enzyme with a strongly reduced thermostability as compared to the wild-type enzyme. Therefore an immunoprecipitation method was used to measure the amount of fusion protein. It was found that these amounts varied strongly from one construction to another. Concomitant determinations of the amounts of lac-operon-specific mRNA showed an unexpectedly large variation among the clones. No strict correlation could be found between the level of lac mRNA and beta-gal production. Per molecule of lac mRNA, translation appears to be most efficient when the homologous lacZ initiation signal is present.

Partial methylation of two adjacent adenosines in ribosomes from Euglena gracilis chloroplasts suggests evolutionary loss of an intermediate stage in the methyl-transfer reaction.

Bacterial, cytoplasmic and organellar ribosomes from a wide phylogenetic spectrum of organisms have a characteristic m6(2)Am6(2)A structure near the 3' end of the RNA of the small ribosomal subunit (SSU). We have studied one of the few exceptions to this extremely conserved post-transcriptionally modified sequence, i.e. dimethylation of only one of the two A's in chloroplasts from Euglena gracilis. It was established that only the A closest to the 5' end is dimethylated, the other one being unmodified. The methylation reaction was studied in vitro using ribosomes from a kasugamycin resistant mutant (ksgA) of Escherichia coli and purified methyl-transferase. Using limited amounts of the methyl donor S-adenosylmethionine (SAM) a partial level of methylation (50% of control) was attained. It is shown that in this case the 3' proximal A is dimethylated while the other is not. This suggests that dimethylation takes place in two successive stages. Apparently in E. gracilis chloroplasts the first stage of methylation does not occur.

Increased translational fidelity caused by the antibiotic kasugamycin and ribosomal ambiguity in mutants harbouring the ksgA gene.

The aminoglycoside kasugamycin, which has previously been shown to inhibit initiation of protein biosynthesis in vitro, also affects translational accuracy in vitro. This is deduced from the observation that the drug decreases the incorporation of histidine relative to alanine into the coat protein of phage MS2, the gene of which is devoid of histidine codons. The read-through of the MS2 coat cistron, due to frameshifts in vitro, is also suppressed by the antibiotic. In contrast, streptomycin enhances histidine incorporation and read-through in this system. The effects of kasugamycin take place at concentrations that do not inhibit coat protein biosynthesis. Kasugamycin-resistant mutants (ksgA) lacking dimethylation of two adjacent adenosines in 16 S ribosomal RNA, show an increased leakiness of nonsense and frameshift mutants (in the absence of antibiotic). They are therefore phenotypically similar to previously described ribosomal ambiguity mutants (ram).

Phylogeny of the conserved 3' terminal structure of the RNA of small ribosomal subunits.

The strongest conserved part of the RNA of small ribosomal subunits is probably located near the 3' end. This paper reviews the primary and secondary structures of some 40 sequenced 3' termini and tries to classify these structures according to common features and differences. The regions under consideration contain at the 5' side an almost universal, supposedly single-stranded stretch of nucleotides with the sequence--AAGUCGUAACAAGGU--. This is followed by a stem-loop structure. The stem always contains 9 basepairs (including U-G pairs) and no mismatches or bulged nucleotides. The loop of the hairpin is either (m2)GGm62Am62A (bacteria, chloroplasts and mitochondria) or UGm62Am62A (cytoplasm). The hairpin is, in most cases, followed at the 3' side by--GGAUCA--. Next to it bacteria and chloroplasts contain the so-called "Shine and Dalgarno" sequence --CCUCC--. The stem region of the hairpin contains a conserved A-U U-G junction. The two basepairs between this junction and the loop are either of type 1 (G-C G-C) or type 2 (C-G C-G). Classification according to type links certain bacteria with mitochondria of yeast and plants and others with chloroplasts and with animal mitochondria.