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'.

Translational initiation on structured messengers. Another role for the Shine-Dalgarno interaction

Translational efficiency in Escherichia coli is in part determined by the Shine-Dalgarno (SD) interaction, i.e. the base-pairing of the 3' end of 16S ribosomal RNA to a stretch of complementary nucleotides in the messenger, located just upstream of the initiation codon. Although a large number of mutations in SD sequences have been produced and analysed, it has so far not been possible to find a clear-cut quantitative relationship between the extent of the complementarity to the rRNA and translational efficiency. This is presumably due to a lack of information about the secondary structures of the messengers used, before and after mutagenesis. Such information is crucial, because intrastrand base-pairing of a ribosome binding site can have a profound influence on its translational efficiency. By site-directed mutagenesis, we have varied the extent of the SD complementarity in the coat-protein gene of bacteriophage MS2. The ribosome binding site of this gene is known to adopt a simple hairpin structure. Substitutions in the SD region were combined with other mutations, which altered the stability of the structure in a predictable way. We find that mutations reducing the SD complementarity by one or two nucleotides diminish translational efficiency only if ribosome binding is impaired by the structure of the messenger. In the absence of an inhibitory structure, these mutations have no effect. In other words, a strong SD interaction can compensate for a structured initiation region. This can be understood by considering translational initiation on a structured ribosome binding site as a competition between intramolecular base-pairing of the messenger and binding to a 30 S ribosomal subunit. A good SD complementarity provides the ribosome with an increased affinity for its binding site, and thereby enhances its ability to compete against the secondary structure. This function of the SD interaction closely parallels the RNA-unfolding capacity of ribosomal protein S1. By comparing the expression data from mutant and wild-type SD sequences, we have estimated the relative contribution of the SD base-pairs to ribosome-mRNA affinity. Quantitatively, this contribution corresponds quite well with the theoretical base-pairing stabilities of the wild-type and mutant SD interactions.

Separation of mutant and wild-type ribosomes based on differences in their anti Shine-Dalgarno sequence

We describe a system to isolate 30S ribosomal subunits which contain targeted mutations in their 16S rRNA. The mutations of interest should be present in so-called specialized 30S subunits which have an anti-Shine-Dalgarno sequence that is altered from 5' ACCUCC to 5' ACACAC. These plasmid-encoded specialized 30S subunits are separated from their chromosomally encoded wild-type counterparts by affinity chromatography that exploits the different Shine-Dalgarno complementarity. An oligonucleotide complementary to the 3' end of wild-type 16S rRNA and attached to a solid phase matrix retains the wild-type 30S subunits. The flow-through of the column contains close to 100% mutant 30S subunits. Toeprinting assays demonstrate that affinity column treatment does not cause significant loss of activity of the specialized particles in initiation complex formation, whereas elongation capacity as determined by poly(Phe) synthesis is only slightly decreased. The method described offers an advantage over total reconstitution from in vitro transcribed mutant 16S rRNA since our 30S subunits contain the naturally occurring base modifications in their 16S rRNA.

Translational initiation at the coat-protein gene of phage MS2: native upstream RNA relieves inhibition by local secondary structure

Maximal translation of the coat-protein gene from RNA bacteriophage MS2 requires a contiguous stretch of native MS2 RNA that extends hundreds of nucleotides upstream from the translational start site. Deletion of these upstream sequences from MS2 cDNA plasmids results in a 30-fold reduction of translational efficiency. By site-directed mutagenesis, we show that this low level of expression is caused by a hairpin structure centred around the initiation codon. When this hairpin is destabilized by the introduction of mismatches, expression from the truncated messenger increases 20-fold to almost the level of the full-length construct. Thus, the translational effect of hundreds of upstream nucleotides can be mimicked by a single substitution that destabilizes the structure. The same hairpin is also present in full-length MS2 RNA, but there it does not impair ribosome binding. Apparently, the upstream RNA somehow reduces the inhibitory effect of the structure on translational initiation. The upstream MS2 sequence does not stimulate translation when cloned in front of another gene, nor can unrelated RNA segments activate the coat-protein gene. Several possible mechanisms for the activation are discussed and a function in gene regulation of the phage is suggested.

Translational control by a long range RNA-RNA interaction; a basepair substitution analysis

One of the two mechanisms that regulate expression of the replicase cistron in the single stranded RNA coliphages is translational coupling. This mechanism prevents ribosomes from binding at the start of the replicase cistron unless the upstream coat cistron is being translated. Genetic analysis had identified a maximal region of 132 nucleotides in the coat gene over which ribosomes should pass to activate the replicase start. Subsequent deletion studies in our laboratory had further narrowed down the regulatory region to 12 nucleotides. Here, we identify a long-distance RNA-RNA interaction of 6 base pairs as the basis of the translational polarity. The 3' side of the complementarity region is located in the coat-replicase intercistronic region, some 20 nucleotides before the start codon of the replicase. The 5' side encodes amino acids 31 and 32 of the coat protein. Mutations that disrupt the long-range interaction abolish the translational coupling. Repair of basepairing by second site base substitutions restores translational coupling.

Differential response to frameshift signals in eukaryotic and prokaryotic translational systems

The genomic RNA of beet western yellows virus (BWYV) contains a potential translational frameshift signal in the overlap region of open reading frames ORF2 and ORF3. The signal, composed of a heptanucleotide slippery sequence and a downstream pseudoknot, is similar in appearance to those identified in retroviral RNAs. We have examined whether the proposed BWYV signal functions in frameshifting in three translational systems, i.c. in vitro in a reticulocyte lysate or a wheat germ extract and in vivo in E. coli. The efficiency of the signal in the eukaryotic system is low but significant, as it responds strongly to changes in either the slip sequence or the pseudoknot. In contrast, in E. coli there is hardly any response to the same changes. Replacing the slip sequence to the typical prokaryotic signal AAAAAAG yields more than 5% frameshift in E. coli. In this organism the frameshifting is highly sensitive to changes in the slip sequence but only slightly to disruption of the pseudoknot. The eukaryotic assay systems are barely sensitive to changes in either AAAAAAG or in the pseudoknot structure in this construct. We conclude that eukaryotic frameshift signals are not recognized by prokaryotes. On the other hand the typical prokaryotic slip sequence AAAAAAG does not lead to significant frameshifting in the eukaryote. In contrast to recent reports on the closely related potato leafroll virus (PLRV) we show that the frameshifting in BWYV is pseudoknot-dependent.

Secondary structure of the ribosome binding site determines translational efficiency: a quantitative analysis

We have quantitatively analyzed the relationship between translational efficiency and the mRNA secondary structure in the initiation region. The stability of a defined hairpin structure containing a ribosome binding site was varied over 12 kcal/mol (1 cal = 4.184 J) by site-directed mutagenesis and the effects on protein yields were analyzed in vivo. The results reveal a strict correlation between translational efficiency and the stability of the helix. An increase in its delta G0 of -1.4 kcal/mol (i.e., less than the difference between an A.U and a G.C pair) corresponds to the reduction by a factor of 10 in initiation rate. Accordingly, a single nucleotide substitution led to the decrease by a factor of 500 in expression because it turned a mismatch in the helix into a match. We find no evidence that exposure of only the Shine-Dalgarno region or the start codon preferentially favors recognition. Translational efficiency is strictly correlated with the fraction of mRNA molecules in which the ribosome binding site is unfolded, indicating that initiation is completely dependent on spontaneous unfolding of the entire initiation region. Ribosomes appear not to recognize nucleotides outside the Shine-Dalgarno region and the initiation codon.

Frameshift suppression at tandem AGA and AGG codons by cloned tRNA genes: assigning a codon to argU tRNA and T4 tRNA(Arg)

Arginine is coded for by CGN (N = G, A, U, C), AGA and AGG. In Escherichia coli there is little tRNA for AGA and AGG and the use of these codons is strongly avoided in virtually all genes. Recently, we demonstrated that the presence of tandem AGA or AGG codons in mRNA causes frameshifts with high frequency. Here, we show that phaseshifts can be suppressed when cells are transformed with the gene for tRNA(T4Arg) or E. coli tRNA(argU,Arg) demonstrating that such errors are the result of tRNA depletion. Bacteriophage T4 encoded tRNA(Arg) (anticodon UCU) corrects shifts at AGA-AGA but not at AGG-AGG, suggesting that this tRNA can only read AGA. Similarly, comparison of the translational efficiencies in an argU (Ts) mutant and in its isogenic wild type parent indicates that argU tRNA (anticodon UCU) reads AGA but not AGG. An argU (Ts) mutant barely reads through AGA-AGA at 42 degrees C but translation of AGG-AGG is hardly, if at all, affected. Overexpression of argU+ relaxes the codon specificity. The thermosensitive mutant in argU, previously called dnaY because it is defective in DNA replication, can be complemented for growth by the gene for tRNA(T4Arg). This implies that the sole function of the argU gene product is to sustain protein synthesis and that its role in replication is probably indirect.

Complete nucleotide sequence of the group I RNA bacteriophage fr

We report the complete nucleotide sequence of the group I RNA bacteriophage fr. The entire genome consists of 3575 nucleotides, six nucleotides more than the only other sequenced group I representative, MS2. The greatest divergence between these phages occurs in the 5' terminal region of the A gene, while the lysis-replicase gene overlap, the coat gene and the central region of the replicase gene are highly conserved. Overall sequence homology between fr and MS2 is 77%. Here, we present a general comparison between the two phages. In the accompanying paper we use phylogenetic sequence comparison between MS2 and fr to deduce the secondary structure at the 3' untranslated region.

Secondary structure at the 3' terminal region of RNA coliphages: comparison with tRNA

Secondary structure models for the 3' non-coding region of the four groups of coliphage RNA are proposed based on comparative sequence analysis and on previously published data on the sensitivity of nucleotides in MS2 RNA to chemical modification and enzymes. We report the following observations. (1) In contrast to the coding regions, the structure at the 3' terminus is characterized by stable regular helices. We note the occurrence of the loop sequences 5'-GUUCGC and 5'-CGAAAG, that are reported to confer exceptional stability to stem structures. These features are probably present to promote the segregation of mother and daughter strands during replication. (2) Comparison of homologous helices indicates that only those base pair substitutions are allowed that maintain the thermodynamic stability. (3) We have compared the structure of phage RNA with tRNA. Overall similarity is low, but one common element may exist. It is a quasi-continuous helix of 12 basepairs that could be the equivalent of the 12 basepair long coaxially stacked helix, formed by the T psi C arm and the aminoacyl acceptor arm in tRNA. As in tRNA, this structure element starts after the fourth nucleotide from the 3' end. (4) Phage RNA contains a large variable region of about 35 nucleotides bulging out from the quasi-continuous helix. We speculate that the variable loop in present-day tRNA could be the remnant of the variable region found in phage RNA. The variable region contains overlapping binding sites for the replicase enzyme and the maturation protein. This common binding site may serve as a switch from replication to packaging.

Scanning model for translational reinitiation in eubacteria

Premature termination of translation in eubacteria, like Escherichia coli, often leads to reinitiation at nearby start codons. Restarts also occur in response to termination at the end of natural coding regions, where they serve to enforce translational coupling between adjacent cistrons. Here, we present a model in which the terminated but not released ribosome reaches neighboring initiation codons by lateral diffusion along the mRNA. The model is based on the finding that introduction of an additional start codon between the termination and the reinitiation site consistently obstructs ribosomes to reach the authentic restart site. Instead, the ribosome now begins protein synthesis at this newly introduced AUG codon. This ribosomal scanning-like movement is bidirectional, has a radius of action of more than 40 nucleotides in the model system used, and activates the first encountered restart site. The ribosomal reach in the upstream direction is less than in the downstream one, probably due to dislodging by elongating ribosomes. The proposed model has parallels with the scanning mechanism postulated for eukaryotic translational initiation and reinitiation.

Secondary structure of the central region of bacteriophage MS2 RNA. Conservation and biological significance

The RNA of the Escherichia coli RNA phages is highly structured with 75% of the nucleotides estimated to take part in base-pairing. We have used enzymatic and chemical sensitivity of nucleotides, phylogenetic sequence comparison and the phenotypes of constructed mutants to develop a secondary structure model for the central region (900 nucleotides) of the group I phage MS2. The RNA folds into a number of, mostly irregular, helices and is further condensed by several long-distance interactions. There is substantial conservation of helices between the related groups I and II, attesting to the relevance of discrete RNA folding. In general, the secondary structure is thought to be needed to prevent annealing of plus and minus strand and to confer protection against RNase. Superimposed, however, are features required to regulate translation and replication. The MS2 RNA section studied here contains three translational start sites, as well as the binding sites for the coat protein and the replicase enzyme. Considering the density of helices along the RNA, it is not unexpected to find that all these sites lie in helical regions. This fact, however, does not mean that these sites are recognized as secondary structure elements by their interaction partners. This holds true only for the coat protein binding site. The other four sites function in the unfolded state and the stability of the helix in which they are contained serves to negatively control their accessibility. Mutations that stabilize helices containing ribosomal binding sites reduce their efficiency and vice versa. Comparison of homologous helices in different phage RNAs indicates that base substitutions have occurred in such a way that the thermodynamic stability of the helix is maintained. The evolution of individual helices shows several distinct size-reduction patterns. We have observed codon deletions from loop areas and shortening of hairpins by base-pair deletions from either the bottom, the middle or the top of stem structures. Evidence for the coaxial stacking of some helical segments is discussed.

Expression of the rat interferon-alpha 1 gene in Escherichia coli controlled by the secondary structure of the translation-initiation region

A synthetic ribosome-binding site (RBS) containing a 7-nucleotide-long Shine-Dalgarno (SD) sequence was placed ahead of the rat interferon (IFN)-alpha 1 coding region. The translational efficiency of this construct was extremely low. Structural probing of transcripts with RNases T1 and U2 combined with computer predictions revealed the presence of a stable hairpin in which the SD region was base-paired to codons 3, 4 and 5 of the IFN mRNA. Each mutation in this stem changing an A-U to an A.C or a G-C a G.U pair increased translational efficiency about fourfold and this effect could be reversed by a compensating stabilizing substitution in the other strand of the stem. We conclude that the strength of an RBS is to a major degree determined by its involvement in secondary structure. We also show that the negative effect of secondary structure on the efficiency of an RBS can be overcome by allowing upstream translation to terminate within the base-paired region. In our clones, termination-dependent restarts occur at a frequency comparable to that taking place in constructs containing destabilized hairpins.

Translational reinitiation in the presence and absence of a Shine and Dalgarno sequence

The process of translational reinitiation in Escherichia coli was studied in a two cistron system where expression of the downstream reporter gene was dependent on translation of an upstream reading frame. The dependence was almost absolute. Upstream translation increased expression of the downstream gene by two to three orders of magnitude. This large difference allowed us to quantitate restarts in a meaningful manner. In the absence of a Shine and Dalgarno (SD) region reinitiation occurred but its efficiency was about 10% of that found in the SD carrying counterpart. We discuss three ways by which translational coupling between neighboring cistrons can be enforced.

Translational regulation of the lysis gene in RNA bacteriophage fr requires a UUG initiation codon

Single nucleotide substitutions identify a UUG triplet as the initiation codon of the lysis gene in RNA bacteriophage fr. This initiation codon is non-functional in de novo initiation but is activated by translational termination at the overlapping coat gene. The UUG initiation codon is crucial for gene regulation in the phage, as it excludes uncontrolled access of ribosomes to the start of the lysis gene. Replacement of UUG by either GUG or AUG results in the loss of genetic control of the lysis gene. A model is presented in which initiation factor IF3 proofreads de novo initiation at UUG codons.

The cysteines in position 1 and 86 of rat interferon-alpha 1 are indispensable for antiviral activity

The human, bovine, murine and rat interferon (IFN)-alpha families contain 4 conserved cysteines located at positions 1, 29, 99 and 139 that are involved in disulfide bridges. Rat and murine IFN-alpha subspecies carry a fifth Cys (Cys-86) which is not conserved in bovine and human IFN-alpha subspecies except for human IFN-alpha 1. Changing Cys-86 in rat IFN-alpha 1 into Ser or Tyr virtually abolished antiviral activity. As shown by others, the substitution of Cys-86 to Ser in human IFN-alpha 1 had no pronounced effect on activity. This suggests that in contrast to human and bovine IFN-alpha, Cys-86 in rodent IFN-alpha plays a crucial role in receptor binding. Changing Cys-1 to Gly in rat IFN-alpha 1 also destroyed activity, in agreement with results obtained in the human IFN-alpha 1 system.

Nucleotide sequence from the ssRNA bacteriophage JP34 resolves the discrepancy between serological and biophysical classification

The nucleotide sequence of the coat and lysis genes of the single-stranded RNA bacteriophage JP34 is presented. Serological inactivation studies classified this phage as an intermediate between groups I and II. We show that the nucleotide similarity with group I is less than 45% but more than 95% for group II, classifying JP34 as a member of group II. The altered serotype of JP34 is most likely due to the change of three critical amino acids of the coat protein to residues present in group I phage MS2 at the homologous positions. Serological characterization of RNA bacteriophages is thus not unambiguous. Phylogenetic sequence comparison between JP34, GA, and MS2 confirms the existence of a conserved helix in the coat gene of group I and group II phages. We also show that the JP34 coat and lysis genes can be expressed in cDNA clones and that the translation of the lysis gene is coupled to coat gene translation analogous to the regulation found in the group I phages.

Induction of the autolytic system of Escherichia coli by specific insertion of bacteriophage MS2 lysis protein into the bacterial cell envelope

Bacterial lysis induced by the expression of the cloned lysis gene of the RNA bacteriophage MS2 in Escherichia coli was shown to be under the same regulatory control mechanisms as penicillin-induced lysis. It was controlled by the stringent response and showed the phenomenon of tolerance when E. coli was grown at pH 5. Changes in the fine structure of the murein were found to be the earliest physiological changes in the cell, taking place 10 min before the onset of cellular lysis and inhibition of murein synthesis. Both the average length of the glycan strands and, with a time lag, the degree of cross-linkage were altered, indicating that a lytic transglycosylase and a DD-endopeptidase had been triggered. After extensive separation of the membranes by isopycnic sucrose gradient centrifugation, the lysis protein was present predominantly in the cytoplasmic membrane and in a fraction of intermediate density and, to a lesser degree, in the outer membrane, irrespective of the conditions of growth. However, only under lysis-permissive conditions could a 17% increase in the number of adhesion sites between the inner and outer membranes be observed. Thus, a casual relationship between lysis and the formation of lysis protein-induced adhesion sites seems to exist.

Translation of the sequence AGG-AGG yields 50% ribosomal frameshift

We have inserted the sequence 5'-AAG-GAGGU-3', which is complementary to the 3' terminus of Escherichia coli 16S rRNA, in a reading frame and analyzed its effect on the accuracy and overall rate of translation in vivo. Translation over the sequence yields a 50% ribosomal frameshift if the reading phase is A-AGG-AGG-U. The other two possible frames do not give shifts. The introduction of a UAA stop codon before (UAA-AGG-AGG-U) but not after (A-AGG-AGG-UAA) the AGG codons abolishes the frameshift. The change in the reading phase occurs exclusively to the +1 direction. Efficient frameshifting is also induced by the sequence A-AGA-AGA-U. The arginine codons AGG and AGA are read by minor tRNA. Suppression of frameshifting takes place when a gene for minor tRNA(Arg) is introduced on a multicopy plasmid. We suggest that frameshifting during translation of the A-AGG-AGG-U sequence is due to the erroneous decoding of the tandem AGG codons and arises by depletion of tRNA(Arg). The complementarity of tandem AGG codons to the 3' terminus of 16S rRNA is a coincidence and apparently not related to the shift. Replacing the AGG-AGG sequence by the optimal arginine codons CGU-CGU does not increase the overall rate of translation.

Lysis induction of Escherichia coli by the cloned lysis protein of the phage MS2 depends on the presence of osmoregulatory membrane-derived oligosaccharides

Expression of the cloned lysis protein of phage MS2, which is sufficient to lyse wild type Escherichia coli, does not cause lysis of mutants lacking the osmoregulatory membrane-derived oligosaccharides (MDO). The lysis gene product normally found in the membrane fraction was not stably inserted into the membranes of a mdoA mutant; rather degradation and release from the membrane occurred. Gentle plasmolysis of the MDO-lacking mutant clearly showed an increased periplasmic space as compared to wild type cells. It is concluded that the MDOs play an important role in maintaining a proper arrangement of inner and outer membrane, a prerequisite for a functional insertion of the MS2 lysis protein.

A synthetic peptide corresponding to the C-terminal 25 residues of phage MS2 coded lysis protein dissipates the protonmotive force in Escherichia coli membrane vesicles by generating hydrophilic pores

The RNA phage MS2 encodes a protein, 75 amino acids long, that is necessary and sufficient for lysis of the host cell. DNA deletion analysis has shown that the lytic activity is confined to the C-terminal half of the protein. We have examined the effects of a synthetic peptide, covering the C-terminal 25 amino acids of the lysis protein, on the electrochemical potential, generated in Escherichia coli membrane vesicles and in liposomes reconstituted with cytochrome c oxidase. In all cases the peptide dissipates the electrochemical potential. The peptide also induces the release of carboxyfluorescein (376 daltons), but not of inuline (5500 daltons), from protein-free liposomes. The phenomena are observed at a lipid to peptide molar ratio of approximately 100:1. The possible connection between the dissipation of the proton-motive force and bacteriolysis is discussed.

Lysis gene of bacteriophage MS2 is activated by translation termination at the overlapping coat gene

The 3' boundary of the coat gene of the RNA bacteriophage MS2 lies 46 nucleotides downstream from the beginning of the lysis (L) cistron. The translation of both reading frames is coupled; the synthesis of the lysis protein does not occur unless translation of the overlapping coat gene takes place. In the preceding paper we showed that de novo initiation at the L gene is prevented by a hairpin structure that sequesters the ribosomal binding site. Here we examine how translation of the coat gene activates the L gene start site. The experiments show that the movement of ribosomes through the hairpin is in itself not sufficient to expose the lysis gene. Rather, the endpoint of translation is important. Termination at the natural end of the coat gene triggers the lysis response, but further downstream terminations do not. Activation of the L gene is suppressed when the stability of the lysis initiator hairpin is increased by mutations that create additional base-pairs. We assume that the ribosome, terminating at the coat reading frame, covers part of the lysis hairpin, thereby destabilizing the secondary structure. This may be sufficient to promote the binding of a vacant ribosome to the L gene start. Alternatively, the terminated but not yet released ribosome may reach the L gene start by random lateral movements along the mRNA and reinitiate there. The present findings are also discussed in relation to an earlier proposal for L gene activation.

Determination of the RNA secondary structure that regulates lysis gene expression in bacteriophage MS2

The lysis gene of the RNA bacteriophage MS2 is not expressed unless translation of the overlapping coat gene takes place. To understand the molecular basis for this translational coupling the RNA secondary structure around the lysis gene start was analyzed with structure-specific enzymes and chemicals. The existence of a hairpin between nucleotides 1636 and 1707 is in agreement with the structural mapping data and also with the conservation of base-pairing in the related M12 phage. In this hairpin, the G residues in the Shine and Dalgarno region and start codon are inaccessible to RNase T1, which is consistent with the fact that ribosomal access to the lysis gene is blocked when there is no coat gene translation. Deletions or point mutations that are predicted to destabilize the hairpin give rise to lysis protein synthesis that is independent of coat gene translation. Base substitutions that are not expected to weaken the helix do not lead to independent lysis gene expression. Finally, nucleotide changes that strengthen the hairpin lead neither to uncoupled nor to coupled synthesis of the lysis protein. Structural analysis of mutant MS2 RNA shows that small changes in the stability of the secondary structure lead to substantial differences in translation initiation. The function of the hairpin structure in coupling lysis gene to coat gene translation requires that its stability is kept within narrow limits.

The amino terminal half of the MS2-coded lysis protein is dispensable for function: implications for our understanding of coding region overlaps

We have asked whether genetic overlaps only evolve to provide extra coding capacity in genomes of restricted size. As a model system we have used the lysis gene of the RNA bacteriophage MS2. This gene overlaps with the distal part of the coat protein gene and with the proximal part of the replicase gene. Using recombinant DNA procedures we have determined whether either of the two overlaps codes for amino acids that are not essential for the function of the 75 amino acid long lysis protein. We find that the first 40 amino acids of the lysis protein are dispensable for function. Thus all of the genetic information essential to the synthesis of the active C-terminal peptide lies within the overlap with the replicase gene, whereas all dispensable residues are encoded in the overlap with the coat protein gene and in the intercistronic region. This suggests that the overlap with the coat protein gene is not required for extra coding capacity but serves to regulate the expression of the lysis gene. Comparative sequence analysis is consistent with this idea.