Nucleotide sequence of a ssRNA phage from Acinetobacter: kinship to coliphages

The complete nucleotide sequence of ssRNA phage AP205 propagating in Acinetobacter species is reported. The RNA has three large ORFs, which code for the following homologues of the RNA coliphage proteins: the maturation, coat and replicase proteins. Their gene order is the same as that in coliphages. RNA coliphages or Leviviridae fall into two genera: the alloleviviruses, like Q(beta), which have a coat read-through protein, and the leviviruses, like MS2, which do not have this coat protein extension. AP205 has no read-through protein and may therefore be classified as a levivirus. A major digression from the known leviviruses is the apparent absence of a lysis gene in AP205 at the usual position, overlapping the coat and replicase proteins. Instead, two small ORFs are present at the 5' terminus, preceding the maturation gene. One of these might encode a lysis protein. The other is of unknown function. Other new features concern the 3'-terminal sequence. In all ssRNA coliphages, there are always three cytosine residues at the 3' end, but in AP205, there is only a single terminal cytosine. Distantly related viruses, like AP205 and the coliphages, do not have significant sequence identity; yet, important secondary structural features of the RNA are conserved. This is shown here for the 3' UTR and the replicase-operator hairpin. Interestingly, although AP205 has the genetic map of a levivirus, its 3' UTR has the length and RNA secondary structure of an allolevivirus. Sharing features with both MS2 and Q(beta) suggests that, in an evolutionary sense, AP205 should be placed between Q(beta) and MS2. A phylogenetic tree for the ssRNA phages is presented.

Translational control by delayed RNA folding: identification of the kinetic trap

The maturation or A-protein gene of single-stranded RNA phage MS2 is preceded by a 130-nt long untranslated leader. When MS2 RNA folding is at equilibrium, the gene is untranslatable because the leader adopts a well-defined cloverleaf structure in which the Shine-Dalgarno (SD) sequence of the maturation gene is taken up in long-distance base pairing with an upstream complementary sequence (UCS). Synthesis of the A-protein takes place transiently while the RNA is synthesized from the minus strand. This requires that formation of the inhibitory cloverleaf is slow. In vitro, the folding delay was on the order of minutes. Here, we present evidence that this postponed folding is caused by the formation of a metastable intermediate. This intermediate is a small local hairpin that contains the UCS in its loop, thereby preventing or slowing down its pairing with the SD sequence. Mutants in which the small hairpin could not be formed made no detectable amounts of A-protein and were barely viable. Apparently, here the cloverleaf formed quicker than ribosomes could bind. On the other hand, mutants in which the small intermediary hairpin was stabilized produced more A-protein than wild type and were viable. One hardly growing mutant that could not form the metastable hairpin and did not make detectable amounts of A-protein was evolved. The emerging pseudo-revertant had selected two second site repressor mutations that allowed reconstruction of a variant of the metastable intermediate. The pseudo-revertant had also regained the capacity to produce the A-protein.

A long-range pseudoknot in Qbeta RNA is essential for replication

A puzzling aspect of replication of bacteriophage Qbeta RNA has always been that replicase binds at an internal segment, the M-site, some 1450 nt away from the 3' end. Here, we report on the existence of a long-range pseudoknot, base-pairing eight nt in the loop of the 3' terminal hairpin to a single-stranded interdomain sequence located about 1200 nt upstream, close to the internal replicase binding site. Introduction of a single mismatch into this pseudoknot is sufficient to abolish replication, but the inhibition is fully reversed by a second-site substitution that restores the pairing. The pseudoknot is part of an elaborate structure that seems to hold the 3' end in a fixed position vis a vis the replicase binding site. Our results imply that the shape of the RNA confers the functonality. We discuss the possible relevance of our findings for replication of other viral RNAs.

A long-range interaction in Qbeta RNA that bridges the thousand nucleotides between the M-site and the 3' end is required for replication

The genome of the positive strand RNA bacteriophage Qbeta folds into a number of structural domains, defined by long-distance interactions. The RNA within each domain is ordered in arrays of three- and four-way junctions that confer rigidity to the chain. One such domain, RD2, is about 1,000-nt long and covers most of the replicase gene. Its downstream border is the 3' untranslated region, whereas upstream the major binding site for Qbeta replicase, the M-site, is located. Replication of Qbeta RNA has always been puzzling because the binding site for the enzyme lies some 1,500-nt away from the 3' terminus. We present evidence that the long-range interaction defining RD2 exists and positions the 3' terminus in the vicinity of the replicase binding site. The model is based on several observations. First, mutations destabilizing the long-range interaction are virtually lethal to the phage, whereas base pair substitutions have little effect. Secondly, in vitro analysis shows that destabilizing the long-range pairing abolishes replication of the plus strand. Thirdly, passaging of nearly inactive mutant phages results in the selection of second-site suppressor mutations that restore both long-range base pairing and replication. The data are interpreted to mean that the 3D organization of this part of Qbeta RNA is essential to its replication. We propose that, when replicase is bound to the internal recognition site, the 3' terminus of the template is juxtaposed to the enzyme's active site.

Long-range translational coupling in single-stranded RNA bacteriophages: an evolutionary analysis

In coliphage MS2 RNA a long-distance interaction (LDI) between an internal segment of the upstream coat gene and the start region of the replicase gene prevents initiation of replicase synthesis in the absence of coat gene translation. Elongating ribosomes break up the repressor LDI and thus activate the hidden initiation site. Expression studies on partial MS2 cDNA clones identified base pairing between 1427-1433 and 1738-1744, the so-called Min Jou (MJ) interaction, as the molecular basis for the long-range coupling mechanism. Here, we examine the biological significance of this interaction for the control of replicase gene translation. The LDI was disrupted by mutations in the 3'-side and the evolutionary adaptation was monitored upon phage passaging. Two categories of pseudorevertants emerged. The first type had restored the MJ interaction but not necessarily the native sequence. The pseudorevertants of the second type acquired a compensatory substitution some 80 nt downstream of the MJ interaction that stabilizes an adjacent LDI. In one examined case we confirmed that the second site mutations had restored coat-replicase translational coupling. Our results show the importance of translational control for fitness of the phage. They also reveal that the structure that buries the replicase start extends to structure elements bordering the MJ interaction.

Base complementarity in helix 2 of the central pseudoknot in 16S rRNA is essential for ribosome functioning

Helix 2 of the central pseudoknot structure in Escherichia coli 16S rRNA is formed by a long-distance interaction between nt 17-19 and 918-916, resulting in three base pairs: U17-A918, C18-G917and A19-U916. Previous work has shown that disruption of the central base pair abolishes ribosomal activity. We have mutated the first and last base pairs and tested the mutants for their translational activity in vivo , using a specialized ribosome system. Mutations that disrupt Watson-Crick base pairing result in strongly impaired translational activity. An exception is the mutation U916-->G, creating an A.G pair, which shows almost no decrease in activity. Mutations that maintain base complementarity have little or no impact on translational efficiency. Some of the introduced base pair substitutions substantially alter the stability of helix 2, but this does not influence ribosome functioning, neither at 42 nor at 28 degrees C. Therefore, our results do not support models in which the pseudoknot is periodically disrupted. Rather, the central pseudoknot structure is suggested to function as a permanent structural element necessary for proper organization in the center of the 30S subunit.

Rescue of the RNA phage genome from RNase III cleavage

The secondary structure of the RNA from the single-stranded RNA bacteriophages, like MS2 and Qb, has evolved to serve a variety of functions such as controlling gene expression, exposing binding sites for the replicase and capsid proteins, allowing strand separation and so forth. On the other hand, all of these foldings have to perform in bacterial cells in which various RNA splitting enzymes are present. We therefore examined whether phage RNA structure is under selective pressure by host RNases. Here we show this to be true for RNase III. A fully double-stranded hairpin of 17 bp, which is an RNase III target, was inserted into a non-coding region of the MS2 RNA genome. In an RNase III-host these phages survived but in wild-type bacteria they did not. Here the stem underwent Darwinian evolution to a structure that was no longer a substrate for RNase III. This was achieved in three different ways: (i) the perfect stem was maintained but shortened by removing all or most of the insert; (ii) the stem acquired suppressor mutations that replaced Watson-Crick base pairs by mismatches; (iii) the stem acquired small deletions or insertions that created bulges. These insertions consist of short stretches of non-templated A or U residues. Their origin is ascribed to polyadenylation at the site of the RNase III cut (in the + or - strand) either by Escherichia coli poly(A) polymerase or by idling MS2 replicase.

RNA folding kinetics regulates translation of phage MS2 maturation gene

The gene for the maturation protein of the single-stranded RNA coliphage MS2 is preceded by an untranslated leader of 130 nt, which folds into a cloverleaf, i.e., three stem-loop structures enclosed by a long distance interaction (LDI). This LDI prevents translation because its 3' moiety contains the Shine-Dalgarno sequence of the maturation gene. Previously, several observations suggested that folding of the cloverleaf is kinetically delayed, providing a time window for ribosomes to access the RNA. Here we present direct evidence for this model. In vitro experiments show that ribosome binding to the maturation gene is faster than refolding of the denatured cloverleaf. This folding delay appears related to special properties of the leader sequence. We have replaced the three stem-loop structures by a single five nt loop. This change does not affect the equilibrium structure of the LDI. Nevertheless, in this construct, the folding delay has virtually disappeared, suggesting that now the RNA folds faster than ribosomes can bind. Perturbation of the cloverleaf by an insertion makes the maturation start permanently accessible. A pseudorevertant that evolved from an infectious clone carrying the insertion had overcome this defect. It showed a wild-type folding delay before closing down the maturation gene. This experiment reveals the biological significance of retarded cloverleaf formation.

Molecular cloning and analysis of the gene encoding the thermostable penicillin G acylase from Alcaligenes faecalis

Alcaligenes faecalis penicillin G acylase is more stable than the Escherichia coli enzyme. The activity of the A. faecalis enzyme was not affected by incubation at 50 degrees C for 20 min, whereas more than 50% of the E. coli enzyme was irreversibly inactivated by the same treatment. To study the molecular basis of this higher stability, the A. faecalis enzyme was isolated and its gene was cloned and sequenced. The gene encodes a polypeptide that is characteristic of periplasmic penicillin G acylase (signal peptide-alpha subunit-spacer-beta subunit). Purification, N-terminal amino acid analysis, and molecular mass determination of the penicillin G acylase showed that the alpha and beta subunits have molecular masses of 23.0 and 62.7 kDa, respectively. The length of the spacer is 37 amino acids. Amino acid sequence alignment demonstrated significant homology with the penicillin G acylase from E. coli A unique feature of the A. faecalis enzyme is the presence of two cysteines that form a disulfide bridge. The stability of the A. faecalis penicillin G acylase, but not that of the E. coli enzyme, which has no cysteines, was decreased by a reductant. Thus, the improved thermostability is attributed to the presence of the disulfide bridge.

Rapid evolution of translational control mechanisms in RNA genomes

We have introduced 13 base substitutions into the coat protein gene of RNA bacteriophage MS2. The mutations, which are clustered ahead of the overlapping lysis cistron, do not change the amino acid sequence of the coat protein, but they disrupt a local hairpin, which is needed to control translation of the lysis gene. The mutations decreased the phage titer by four orders of magnitude but, upon passaging, the virus accumulated suppressor mutations that raised the fitness to almost wild-type level. Analysis of the pseudorevertants showed that the disruption of the local hairpin, controlling expression of the lysis gene, had apparently been so complete that its restoration by chance mutations could not be achieved. Instead, alternative foldings initiated by the starting mutations were further stabilized and optimized. Strikingly, in the pseudorevertants analyzed, translational control of the lysis gene had been restored. This feat was accomplished by, on average, four suppressor mutations that generally occurred at codon wobble positions. We also introduced 11 mutations in a hairpin more upstream in the coat protein gene and not implicated in lysis control. Here the titer dropped by three logs, but pseudorevertants with a fitness close to wild-type were soon generated. These pseudorevertants again were the result of the optimization of alternative foldings induced by the mutations. The transition of the secondary structure from wild-type to pseudorevertant could be visualized by structure probing. Our study shows that the folding of the RNA is an important phenotypic property of RNA viruses. However, its distortion can easily be overcome by optimizing alternative base-pairings. These new structures are not qualitatively equivalent to the original one, since they do not successfully compete with the wild-type.

Acetylation of ribosomal protein S5 affected by defects in the central pseudoknot in 16S ribosomal RNA?

We have analyzed the ribosomal protein profile of Escherichia coli 30S subunits with the mutation C18A in the central pseudoknot of their 16S ribosomal RNA. This mutation was shown to inhibit translational activity in vivo and to affect ribosome stability in vitro. The majority of the mutant 30S particles were present as free subunits in which a reproducible decrease in amount of proteins S1, S2, S18 and S21 was observed. The protein gels also showed the appearance of a satellite band next to S5. This band reacted with anti-S5 antibodies and had a slightly increased positive charge. The simplest interpretation of these findings, also considering published data, is that the satellite band is S5 with a non-acetylated N-terminal alanine. Underacetylation of S5 due to mutations in the 16S rRNA implies that the modification is performed on the ribosome.

[Vancomycin-resistant enterococci . Work Group Hospital Infection Epidemiology]

Since 1995 patients with infections with vancomycin-resistant enterococci (VRE) have been reported in Rotterdam, the Netherlands. From prevalence studies in the Netherlands at the end of 1995 and the start of 1996, it appeared that 2% of the patients studied in and outside hospitals had VRE. Restrictive antibiotic prescription in human and veterinary medicine is indicated in order to postpone the problem.

Evolutionary reconstruction of a hairpin deleted from the genome of an RNA virus

The intercistronic region between the maturation and coat-protein genes of RNA phage MS2 contains important regulatory and structural information. The sequence participates in two adjacent stem-loop structures, one of which, the coat-initiator hairpin, controls coat-gene translation and is thus under strong selection pressure. We have removed 19 out of the 23 nucleotides constituting the intercistronic region, thereby destroying the capacity of the phage to build the two hairpins. The deletion lowered coat-protein yield more than 1000-fold, and the titer of the infectious clone carrying the deletion dropped 10 orders of magnitude as compared with the wild type. Two types of revertants were recovered. One had, in two steps, recruited 18 new nucleotides that served to rebuild the two hairpins and the lost Shine-Dalgarno sequence. The other type had deleted an additional six nucleotides, which allowed the reconstruction of the Shine-Dalgarno sequence and the initiator hairpin, albeit by sacrificing the remnants of the other stem-loop. The results visualize the immense genetic repertoire created by, what appears as, random RNA recombination. It would seem that in this genetic ensemble every possible new RNA combination is represented.

The central pseudoknot in 16S ribosomal RNA is needed for ribosome stability but is not essential for 30S initiation complex formation

To examine the function of the central pseudoknot in 16S rRNA, we have studied Escherichia coli 30S subunits with the A18 mutation in this structure element. Previously, this mutation, which changes the central base pair of helix 2, C18--G917, to an A18xG917 mismatch, was shown to inhibit translation in vivo and a defect in initiation was suggested. Here, we find that the mutant 30S particles are impaired in forming 70S tight couples and predominantly accumulate as free 30S subunits. Formation of a 30S initiation complex, as measured by toeprinting, was almost as efficient for mutant 30S subunits, derived from the tight couple fraction, as for the wild-type control. However, the A18 mutation has a profound effect on the overall stability of the subunit. The mutant ribosomes were inactivated by affinity chromatography and high salt treatment, due to easy loss of ribosomal proteins. Accordingly, the particles could be reactivated by partial in vitro reconstitution with 30S ribosomal proteins. Mutant 30S subunits from the free subunit fraction were already inactive upon isolation, but could also be reactivated by reconstitution. Apparently, the inactivity in initiation of these mutant 30S subunits is, at least in part, also due to the lack of essential ribosomal proteins. We conclude that disruption of helix 2 of the central pseudoknot by itself does not affect the formation of a 30S initiation complex. We suggest that the in vivo translational defect of the mutant ribosomes is caused by their inability to form 70S initiation complexes.

Secondary structure model for the first three domains of Q beta RNA. Control of A-protein synthesis

We present a secondary structure model for the first 860 nucleotides of Q beta RNA. The model is supported by phylogenetic comparison, nuclease S1 structure probing and computer prediction using energy minimization and a Monte Carlo approach. To provide the necessary data for the comparative analysis we have sequenced the single-stranded RNA coliphages MX1, M11 and NL95. Together with the known sequences of Q beta and SP, this yields five sequences with sufficient sequence diversity to be useful for the analysis. The part of the Q beta genome examined contains the 60 nucleotide 5' untranslated region and the first 800 nucleotide of the maturation protein gene. The RNA adopts a highly ordered structure in which all hairpins are held in place by a network of long-distance interactions, which form three-way and four-way junctions. Only the 5'-terminal hairpin is unrestrained, while connected by a few single-stranded nucleotides to the body of the RNA. The start region of the A-protein gene, which is part of the network of long-distance interactions, is base-paired to three non-contiguous downstream sequences. As a result, translation is expected to be progressively quenched when the length of the nascent chains increases. This feature explains the previous observation that A-protein synthesis on Q beta RNA can start only on short nascent strands. Translational control of the A protein in the distantly related phage MS2 was recently shown to be controlled by the kinetics of RNA folding. This basic difference and its possible biological purpose can be explained by the different RNA folding pathways in Q beta and MS2. Interestingly, due to the presence of G-U pairs, structure prediction for the minus strand differs in some aspects from that for the plus strand. More specifically, there is a minus-strand specific, long-distance interaction bordering the minus-strand equivalent of the 5'-terminal hairpin. This interaction extends at the expense of the lower part of the terminal helix, thereby exposing the terminal C residues at which replication starts. This long-distance interaction, which was recently shown to be required for minus-strand replication, is strongly supported by our comparative data.

An oligonucleotide hybridization assay for the identification and enumeration of F-specific RNA phages in surface water

F-specific RNA phages can be used as model organisms for enteric viruses to monitor the effectiveness of sewage treatment, and to assess the potential contamination of surface water with these viruses. In this paper a method is described which identifies RNA phages quantitatively by a plaque hybridization assay. Oligonucleotide probes were developed that can assign phages to their phylogenetic subgroups. Such a distinction is important, since some subgroups preferentially occur in sewage of human origin, while others tend to be associated with animal wastewater. The method has been tested on a large number of isolates and represents an improvement in time and reliability over the previously used serological classification.

Random removal of inserts from an RNA genome: selection against single-stranded RNA

We have monitored the evolution of insertions in two MS2 RNA regions of known secondary structure where coding pressure is negligible or absent. Base changes and shortening of the inserts proceed until the excessive nucleotides can be accommodated in the original structure. The stems of hairpins can be dramatically extended but the loops cannot, revealing natural selection against single-stranded RNA. The 3' end of the MS2 A-protein gene forms a small hairpin with an XbaI sequence in the loop. This site was used to insert XbaI fragments of various sizes. Phages produced by these MS2 cDNA clones were not wild type, nor had they retained the full insert. Instead, every revertant phage had trimmed the insert in a different way to leave a four- to seven-membered loop to the now extended stem. Similar results were obtained with inserts in the 5' untranslated region. The great number of different revertants obtained from a single starting mutant as well as sequence inspection of the crossover points suggest that the removal of redundant RNA occurs randomly. The only common feature among all revertants appears the potential to form a hairpin with a short loop, suggesting that single-stranded RNA negatively affects the viability of the phage. To test this hypothesis, we introduced XbaI fragments of 34 nucleotides that could form either a long stem with a small loop or a short stem with a large loop (26 nucleotides). The base-paired inserts were perfectly maintained for many generations, whereas the unpaired versions were quickly trimmed back to reduce the size of the loop. These data confirm that single-stranded RNA adversely affects phage fitness and is strongly selected against. The repair of the RNA genome that we describe here appears as the result of random recombination. Of the plethora of recombinants, only those able to adopt a base-paired structure survive. The frequency with which our inserts are removed seems higher than measured by others for small inserts in a reading frame in Q beta RNA. To account for this higher frequency, we suggest models in which the single-stranded nature of our inserts induces random recombination at the site of the insertion.

Genotyping male-specific RNA coliphages by hybridization with oligonucleotide probes

F-specific (F+) RNA coliphages are prevalent in sewage and other fecal wastes of humans and animals. There are four antigenically distinct serogroups of F+ RNA coliphages, and those predominating in humans (groups II and III) differ from those predominating in animals (groups I and IV). Hence, it may be possible to distinguish between human and animal wastes by serotyping F+ RNA coliphage isolates. Because serotyping is laborious and requires scarce antiserum reagents, we investigated genotyping using synthetic oligonucleotide probes as an alternative approach to distinguishing the four groups of F+ RNA coliphages. Oligoprobes I, II, III, IV, A, and B were selected to detect group I, II, III, IV, I plus II, and III plus IV phages, respectively. Methods for phage transfer from zones of lysis on a host cell lawn to candidate membrane filters and fixation of genomic nucleic acid on the membranes were optimized. The oligoprobes, which were end labeled with digoxigenin, were applied in DNA-RNA hybridization, and hybrids were observed by colorimetric, immunoenzymatic detection. Of 203 isolates of F+ RNA coliphages from environmental samples of water, wastes, and shellfish, 99.5 and 96.6% could be classified into each group by serotyping and genotyping, respectively. Probes A and B correctly identified 100% of the isolates. On the basis of these results, this method for genotyping F+ RNA coliphages appears to be practical and reliable for typing isolates in field samples.

Stabilised secondary structure at a ribosomal binding site enhances translational repression in E. coli

The expression of the gene encoding Escherichia coli threonyl-tRNA synthetase is negatively autoregulated at the translational level. The negative feedback is due to the binding of the synthetase to an operator site on its own mRNA located upstream of the initiation codon. The present work describes the characterisation of operator mutants that have the rare property of enhancing repression. These mutations cause (1) a low basal level of expression, (2) a temperature-dependent expression, and (3) an increased capacity of the synthetase to repress its own expression at low temperature. Surprisingly, this enhancement of repression is not explained by an increase of affinity of the mutant operators for the enzyme but by the formation, at low temperature, of a few supplementary base-pairs between the ribosomal binding site and a normally single-stranded domain of the operator. Although this additional base-pairing only slightly inhibits ribosome binding in the absence of repressor, simple thermodynamic considerations indicate that this is sufficient to increase repression. This increase is explained by the competition between the ribosome and repressor for overlapping regions of the mRNA. When the ribosomal binding site is base-paired, the ribosome cannot bind while the repressor can, giving the repressor the advantage in the competition. Thus, the existence of an open versus base-paired equilibrium in a ribosomal binding site of a translational operator amplifies the magnitude of control. This molecular amplification device might be an essential component of translational control considering the low free repressor/ribosome ratio of the low affinity of translational repressors for their target operators.

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.

Secondary structure model for the last two domains of single-stranded RNA phage Q beta

We have determined the nucleotide sequence of three positive single-stranded RNA coliphages and have used this information, together with the known sequences of the related phages Q beta and SP, to construct a secondary structure model for the two distal domains of Q beta RNA. The 3' terminal domain, which is about 100 nucleotides long, contains most of the 3' untranslated region and folds into four short, regular hairpins. The adjacent 3' replicase domain contains about 1100 nucleotides. Hairpins in this protein-coding domain are much longer and more irregular than in the 3' untranslated region. Both domains are defined by long-distance interactions. The secondary structure is not a collection of hairpin structures connected by single-stranded regions. Rather, the RNA stretches between the stem-loop structures are all involved in an extensive array of long-distance interactions that contract the molecule to a rigid structure in which all hairpins are predicted to have a fixed position with respect to each other. A general feature of the model is that helices tend to be organized in four-way junctions with little or no unpaired nucleotides between them. As a result, there is a potential for coaxial stacking of adjacent stems. The essential features of the model are supported by the S1 nuclease cleavage pattern. Viral RNA sequences are strongly constrained by their coding function. As a result, structural evolution by simple base-pair substitution is not always possible, as this usually requires the juxtaposition of the codon wobble positions in stems. Rather, we often observe co-ordinate base substitutions that maintain the stem, but tend to change the position at which bulges or internal loops are found. Structures that differ in this way are apparently equally fit. Also, the relative position of hairpin loops can shift several nucleotides through an alignment based on maximal sequence identity i.e. amino acid homology. The fact that these structural irregularities do not occur at the 3' untranslated region suggests indeed that the coding function of the RNA constrains the secondary structure. Hairpins with the stable tetraloop motif GNRA and UNCG or their complement are over-represented. This suggests their involvement in segregation of plus and minus strand. The genome of the coliphages contains a well-defined high affinity binding site for the coat protein, which serves to suppress replicase translation and also acts as a nucleation point in capsid formation. Close to the 3' end we find an additional conserved helix that meets the described consensus criteria for coat-protein binding.

Translational control of maturation-protein synthesis in phage MS2: a role for the kinetics of RNA folding?

The gene for the maturation (A) protein of the single-stranded RNA coliphage MS2 is preceded by an untranslated leader of 130 nt. Secondary structure of the leader was deduced by phylogenetic comparison and by probing with enzymes and chemicals. The RNA folds into a cloverleaf, i.e., three stem-loop structures enclosed by a long-distance interaction (LDI). This LDI is essential for translational control. Its 3'moiety contains the Shine-Dalgarno region of the A-protein gene, whereas its complement is located 80 nt upstream, i.e., about 30 nt from the 5'-terminus of the RNA chain. Mutational analysis shows that this base pairing represses expression of the A-protein gene. We present a model in which translational starts can only take place on nonequilibrated RNA, in which base pairing between the complementary regions has not yet taken place. We suggest that this pairing is kinetically delayed by the intervening sequence, which contains the three hairpins of the cloverleaf. The model is mainly based on the observation that reducing the length of the intervening sequence reduces expression, whereas increasing the length has the opposite effect. In addition, further stabilization of the LDI by a stronger base pair does not lead to a decrease in A-protein synthesis. Such a decrease is predicted to occur if translation would be controlled by the equilibrium structure of the leader RNA. These and other observations fit a kinetic model of translational control by RNA folding.

Coevolution of RNA helix stability and Shine-Dalgarno complementarity in a translational start region

The initiation region of the coat-protein gene of RNA bacteriophage MS2 adopts a well-defined hairpin structure with the start codon occupying the loop position, while the Shine-Dalgarno (SD) sequence is part of the stem. In a previous study, we introduced mutations in this hairpin that changed its thermodynamic stability. The resulting phages evolved to regain the wild-type stability by second-site compensatory substitutions. Neither the original nor the suppressor mutations were in the SD region. In the present analysis, we have made changes in the SD region that shorten or extend its complementarity to the 3' end of 16S rRNA and monitored their evolution to a stable pseudorevertant species. Phages in which the SD complementarity was decreased evolved an initiator hairpin of lower stability than wild type while those in which the complementarity was extended evolved a hairpin with an increased stability. We conclude that weaker SD sequences still allow maximal translation if the secondary structure of the ribosome-landing site is destabilized accordingly. Alternatively, translation-initiation regions with a stronger secondary structure still allow maximal expression, if the SD complementarity is extended. These findings support a previously published model in which the SD interaction helps the ribosome to melt the structure in a translation-initiation region.

Control of translation by mRNA secondary structure in Escherichia coli. A quantitative analysis of literature data

Translational efficiency in Escherichia coli is strongly controlled by the secondary structure of the mRNA in the translational initiation region. We have previously shown that protein production from the coat-protein gene of RNA bacteriophage MS2 is directly related to the fraction of mRNA molecules in which the ribosome binding site is unfolded. This fraction is dictated by the free energy (delta Gf0) of the local secondary structure. We now present a similar analysis of published data on four other ribosome binding sites. The results conform quantitatively to the same relationship as found for the MS2 coat-protein gene. The efficiency of translation is determined by the overall stability of the structure at the ribosome binding site, whether the initiation codon itself is base-paired or not. Structures weaker than -6 kcal/mol usually do not reduce translational efficiency. Below this threshold, all systems show a tenfold decrease in expression for every -1.4 kcal/mol, as predicted from theory.

Leeway and constraints in the forced evolution of a regulatory RNA helix

The start of the coat protein gene of RNA phage MS2 adopts a well-defined hairpin structure of 12 bp (including one mismatch) in which the start codon occupies the loop position. An earlier expression study using partial MS2 cDNA clones had indicated that the stability of this hairpin is important for gene expression. For every -1.4 kcal/mol increase in stability a 10-fold reduction in coat protein was obtained. Destabilizations beyond the wild-type value did not affect expression. These results suggested that the hairpin was tuned in the sense that it has the highest stability still compatible with maximal ribosome loading. Employing an infectious MS2 cDNA clone, we have now tested the prediction that the delta G 0 of the coat protein initiator helix is set at a precise value. We have introduced stabilizing and destabilizing mutations into this hairpin in the intact phage and monitored their evolution to viable species. By compensatory mutations, both types of mutants quickly revert along various pathways to wild-type stability, but not to wild-type sequence. As a rule the second-site mutations do not change the encoded amino acids or the Shine-Dalgarno sequence. The return of too strong hairpins to wild-type stability can be understood from the need to produce adequate supplies of coat protein. The return of unstable hairpins to wild-type stability is not self-evident and is presently not understood. The revertants provide an evolutionary landscape of slightly suboptimal phages, that were stable at least for the duration of the experiment (approximately 20 infection cycles).(ABSTRACT TRUNCATED AT 250 WORDS)