Morphogenetic genes C and Nu3 overlap in bacteriophage lambda

Pseudomonas Phage MD8: Genetic Mosaicism and Challenges of Taxonomic Classification of Lambdoid Bacteriophages

Pseudomonas phage MD8 is a temperate phage isolated from the freshwater lake Baikal. The organisation of the MD8 genome resembles the genomes of lambdoid bacteriophages. However, MD8 gene and protein sequences have little in common with classified representatives of lambda-like phages. Analysis of phage genomes revealed a group of other Pseudomonas phages related to phage MD8 and the genomic layout of MD8-like phages indicated extensive gene exchange involving even the most conservative proteins and leading to a high degree of genomic mosaicism. Multiple horizontal transfers and mosaicism of the genome of MD8, related phages and other λ-like phages raise questions about the principles of taxonomic classification of the representatives of this voluminous phage group. Comparison and analysis of various bioinformatic approaches applied to λ-like phage genomes demonstrated different efficiency and contradictory results in the estimation of genomic similarity and relatedness. However, we were able to make suggestions for the possible origin of the MD8 genome and the basic principles for the taxonomic classification of lambdoid phages. The group comprising 26 MD8-related phages was proposed to classify as two close genera belonging to a big family of λ-like phages.

Functional domains of the HK97 capsid maturation protease and the mechanisms of protein encapsidation

Tailed double-stranded DNA bacteriophages and herpesviruses build capsids by co-assembling a major capsid protein with an internal scaffolding protein that then exits from the assembled structure either intact or after digestion in situ by a protease. In bacteriophage HK97, the 102-residue N-terminal delta domain of the major capsid protein is also removed by proteolysis after assembly and appears to perform the scaffolding function. We describe the HK97 protease that carries out these maturation cleavages. Insertion mutations at seven sites in the protease gene produced mutant proteins that assemble into proheads, and those in the N-terminal two-thirds were enzymatically inactive. Plasmid-expressed protease was rapidly cleaved in vivo but was stabilized by co-expression with the delta domain. Purified protease was found to be active during the assembly of proheads in vitro. Heterologous fusions to the intact protease or to C-terminal fragments targeted fusion proteins into proheads. We confirm that the catalytic activity resides in the N-terminal two-thirds of the protease polypeptide and suggest that the C-terminal one-fifth of the protein contains a capsid targeting signal. The implications of this arrangement are compared to capsid targeting systems in other phages, herpesviruses, and encapsulins.

The tripartite capsid gene of Salmonella phage Gifsy-2 yields a capsid assembly pathway engaging features from HK97 and lambda

Phage Gifsy-2, a lambdoid phage infecting Salmonella, has an unusually large composite gene coding for its major capsid protein (mcp) at the C-terminal end, a ClpP-like protease at the N-terminus, and a approximately 200 residue central domain of unknown function but which may have a scaffolding role. This combination of functions on a single coding region is more extensive than those observed in other phages such as HK97 (scaffold-capsid fusion) and lambda (protease-scaffold fusion). To study the structural phenotype of the unique Gifsy-2 capsid gene, we have purified Gifsy-2 particles and visualized capsids and procapsids by cryoelectron microscopy, determining structures to resolutions up to 12A. The capsids have lambdoid T=7 geometry and are well modeled with the atomic structures of HK97 mcp and phage lambda gpD decoration protein. Thus, the unique Gifsy-2 capsid protein gene yields a capsid maturation pathway engaging features from both phages HK97 and lambda.

In‐frame overlapping genes: the challenges for regulating gene expression

In-frame overlapping genes in phage, plasmid and bacterial genomes permit synthesis of more than one form of protein from the same gene. Having one gene entirely within another rather than two separate genes presumably precludes recombination events between the identical sequences. However, studies of such gene pairs indicate that the overlapping arrangement can make regulation of the genes more difficult. Here, we extend studies of in-frame overlapping genes II and X from filamentous phage f1 to determine if translational controls are required to regulate the gene properly. These genes encode proteins (pII and pX) with essential but opposing roles in phage DNA replication. They must be tightly regulated to maintain production of the proteins at relative steady state levels that permit continuous replication without killing the host. To determine why little or no pX appears to be made on the gene II/X mRNA, gene II translation was lowered by progressively deleting into the gene II initiator region. Increased pX translation resulted, suggesting that elongating ribosomes on the gene II mRNA interfere with internal initiation on the gene X ribosome binding site and limit gene X translation. As judged from systematically lowering the efficiency of suppression at a gene II amber codon upstream from the gene X start, the already modest level of gene II translation would have to be reduced by more than twofold to relieve all interference with internal initiation. Further downregulation of gene X expression proved to be required to maintain pX at levels relative to pII that are tolerated by the cell. Site-directed mutagenesis and nuclease mapping revealed that the gene X initiation site is sequestered in an extended RNA secondary structure that lowers gene X translation on the two mRNAs encoding it. The more general implications of the results for expression of in-frame overlapping genes are discussed.

Conserved translational frameshift in dsDNA bacteriophage tail assembly genes

A programmed translational frameshift similar to frameshifts in retroviral gag-pol genes and bacterial insertion elements was found to be strongly conserved in tail assembly genes of dsDNA phages and to be independent of sequence similarities. In bacteriophage lambda, this frameshift controls production of two proteins with overlapping sequences, gpG and gpGT, that are required for tail assembly. We developed bioinformatic approaches to identify analogous -1 frameshifting sites and experimentally confirmed our predictions for five additional phages. Clear evidence was also found for an unusual but analogous -2 frameshift in phage Mu. Frameshifting sites could be identified for most phages with contractile or noncontractile tails whose length is controlled by a tape measure protein. Phages from a broad spectrum of hosts spanning Eubacteria and Archaea appear to conserve this frameshift as a fundamental component of their tail assembly mechanisms, supporting the idea that their tail genes share a common, distant ancestry.

Translation at higher than an optimal level interferes with coupling at an intercistronic junction

In pairs of adjacent genes co-transcribed on bacterial polycistronic mRNAs, translation of the first coding region frequently functions as a positive factor to couple translation to the distal coding region. Coupling efficiencies vary over a wide range, but synthesis of both gene products at similar levels is common. We report the results of characterizing an unusual gene pair, in which only about 1% of the translational activity from the upstream gene is transmitted to the distal gene. The inefficient coupling was unexpected because the upstream gene is highly translated, the distal initiation site has weak but intrinsic ability to bind ribosomes, and the AUG is only two nucleotides beyond the stop codon for the upstream gene. The genes are those in the filamentous phage IKe genome, which encode the abundant single-stranded DNA binding protein (gene V) and the minor coat protein that caps one tip of the phage (gene VII). Here, we have used chimeras between the related phage IKe and f1 sequences to localize the region responsible for inefficient coupling. It mapped upstream from the intercistronic region containing the gene V stop codon and the gene VII initiation site, indicating that low coupling efficiency is associated with gene V. The basis for inefficient coupling emerged when coupling efficiency was found to increase as gene V translation was decreased below the high wild-type level. This was achieved by lowering the rate of elongation and by decreasing the efficiency of suppression at an amber codon within the gene. Increasing the strength of the Shine-Dalgarno interaction with 16S rRNA at the gene VII start also increased coupling efficiency substantially. In this gene pair, upstream translation thus functions in an unprecedented way as a negative factor to limit downstream expression. We interpret the results as evidence that translation in excess of an optimal level in an upstream gene interferes with coupling in the intercistronic junction.

Sequence analysis of the phage 21 genes for prohead assembly and head completion

Phage 21 is a temperate lambdoid coliphage, and its head-encoding genes, as well as those of phage lambda, are descended from a common ancestral phage. The head protein-encoding genes of phage 21 have been sequenced, confirming earlier genetic studies indicating that the head-encoding genes of 21 and lambda are analogous in location, size, and function. The phage 21 head-encoding genes identified (and their lambda analogues) include: 3(W), 4(B), 5(C), 6(Nu3), shp (D), 7(E), and 8(FII), respectively. An open reading frame, orf1, is analogous in position and shares some sequence identity with FI, a phage lambda gene involved in DNA packaging. The phage 21 major head protein, gp7, is predicted to have strong sequence identity (65%) with the lambda major capsid protein, gpE, including amino acids known to be important for capsid form determination. The nested genes 5/6 of phage 21 and C/Nu3 of lambda differ by several rearrangements including deletions and a triplication. The possibility that lambda genes C/Nu3 evolved from ancestal nested genes containing a triplication is discussed.

Translational options for the pir gene of plasmid R6K: multiple forms of the replication initiator protein pi

The autogenously controlled pir gene of plasmid R6K was believed to encode a single polypeptide that plays multiple roles in the plasmid's biology. We have isolated an opal (op) mutant at the 18th codon of the pir coding frame which does not totally abolish translation of pir mRNA. In extracts of cells containing this mutation two translational products (35 kDa and 30.2 kDa) have been detected. We propose that the 35-kDa polypeptide produced by the pir18 op mutation contains Trp substituted for Arg18 as the result of an opal readthrough. Translation, which results in the 30.2-kDa polypeptide, originates downstream from the UGA stop signal created by the mutation. Moreover, we realize now that the 30.2-kDa polypeptide is also produced in cells containing a wild-type (wt) pir gene. The shorter variant of the pi protein lacks replication initiation and inhibition functions, as well as autorepressor activity in vivo. We also show that an in-frame fusion of seven N-terminal codons of the trpE gene with a pir gene lacking the first two codons produces two polypeptides which replace the 35-kDa pi protein and are of similar molecular weight. Thus, at least three options exist in the translation of the wt pir mRNA. Start codons are most likely at codon positions 1, 6 or 7, and 36 or 38. Each of these five AUG codons is preceded by a consensus ribosome-binding site (RBS).

Bacteriophage P1 genes involved in the recognition and cleavage of the phage packaging site (pac)

The packaging of bacteriophage P1 DNA is initiated by cleavage of the viral DNA at a specific site, designated pac. The proteins necessary for that cleavage, and the genes that encode those proteins, are described in this report. By sequencing wild-type P1 DNA and DNA derived from various P1 amber mutants that are deficient in pac cleavage, two distinct genes, referred to as pacA and pacB, were identified. These genes appear to be coordinately transcribed with an upstream P1 gene that encodes a regulator of late P1 gene expression (gene 10). pacA is located upstream from pacB and contains the 161 base-pair pac cleavage site. The predicted sizes of the PacA and PacB proteins are 45 kDa and 56 kDa, respectively. These proteins have been identified on SDS-polyacrylamide gels using extracts derived from Escherichia coli cells that express these genes under the control of a bacteriophage T7 promoter. Extracts prepared from cells expressing both PacA and PacB are proficient for site-specific cleavage of the P1 packaging site, whereas those lacking either protein are not. However, the two defective extracts can complement each other to restore functional pac cleavage activity. Thus, PacA and PacB are two essential bacteriophage proteins required for recognition and cleavage of the P1 packaging site. PacB extracts also contain a second P1 protein that is encoded within the pacB gene. We have identified this protein on SDS-polyacrylamide gels and have shown that it is translated in the same reading frame as is PacB. Its role, if any, in pac cleavage is yet to be determined.

Structural and functional domains of E coli initiation factor IF2

Initiation of translation in prokaryotes requires the participation of at least three soluble proteins: the initiation factors IF1, IF2 and IF3. Initiation factor 2, which is one of the largest proteins involved in translation (97.3 kDa) has been shown to stimulate in vitro the binding of fMet-tRNA(fMet) to the 30S ribosomal subunit. After formation of 70S translation initiation complex, IF2 is believed to participate in GTP hydrolysis, thereby promoting its own release. Here we review evidence which indicates the functional importance of the different structural domains of IF2, emphasizing new information obtained by in vivo experiments.

Broad-host-range properties of plasmid RK2: importance of overlapping genes encoding the plasmid replication initiation protein TrfA

The trfA gene, encoding the essential replication initiation protein of the broad-host-range plasmid RK2, possesses an in-frame overlapping arrangement. This results in the production of TrfA proteins of 33 and 44 kDa, respectively. Utilizing deletion and site-specific mutagenesis to alter the trfA operon, we compared the replication of an RK2-origin plasmid in several distantly related gram-negative bacteria when supported by both TrfA-44 and TrfA-33, TrfA-33 alone, or TrfA-44/98L (a mutant form of the TrfA-44 protein) alone. TrfA-44/98L is identical to wild-type TrfA-44 with the exception of a single conservative amino acid alteration from methionine to leucine at codon 98; this alteration removes the translational start codon for the TrfA-33 protein. Copy number and stability were virtually identical for plasmids containing both TrfA-44 and TrfA-33 proteins or TrfA-44/98L alone in Pseudomonas aeruginosa and Agrobacterium tumefaciens, two unrelated bacteria in which TrfA-33 is poorly functional. This, along with recent in vitro studies comparing TrfA-44, TrfA-33, and TrfA-44/98L, suggests that the functional activity of TrfA-44 is not significantly affected by the 98L mutation. Analysis of minimal RK2 derivatives in certain gram-negative bacterial hosts suggests a role of the overlapping arrangement of trfA in facilitating the broad host range of RK2. RK2 derivatives encoding TrfA-44/98L alone demonstrated decreased copy number and stability in Escherichia coli and Azotobacter vinelandii when compared with derivatives specifying both TrfA-44 and TrfA-33. A strategy employing the trfA-44/98L mutant gene and in vivo homologous recombination was used to eliminate the internal translational start codon of trfA in the intact RK2 plasmid. The mutant intact RK2 plasmid produced only TrfA-44/98L. A small reduction in copy number and beta-lactamase expression resulted in E. coli, suggesting that overlapping trfA genes also enhance the efficiency of replication of the intact RK2 plasmid.

Cytomegalovirus assembly protein nested gene family: four 3'-coterminal transcripts encode four in-frame, overlapping proteins

The genomic region encoding the assembly protein of simian cytomegalovirus (CMV) strain Colburn has been cloned, sequenced, and found to be organized as a nested set of four in-frame, 3'-coterminal genes, each with its own TATA promoter element and translational start codon, and all using a single 3' polyadenylation signal. The 3' end of the longest open reading frame (1.770 bp) was identical to the 930-bp sequence coding for the assembly protein precursor, as determined from a cDNA clone. The assembly protein coding region of human CMV strain AD169 was similarly organized, suggesting that both viral genomes could give rise to four independently transcribed 3'-coterminal RNAs coding for four overlapping, in-frame, carboxy-coterminal proteins. These predictions were tested and confirmed. Four mRNAs corresponding in size and sequence to those predicted were identified in both human and simian CMV-infected cells by using transcript-specific antisense oligonucleotide probes in Northern (RNA blot) assays. The 5' ends of the three largest of these Colburn transcripts were determined by S1 nuclease protection assays and found to map between the anticipated TATA sequences and corresponding translational start codons. The four predicted overlapping proteins were identified by immunoassays in lysates of simian and human CMV-infected cells by using an antiserum specific for the carboxyl end of the assembly protein precursor. The structural relationship of both sets of proteins was verified by comparing their peptide patterns following protein cleavage at tryptophan residues by N-chlorosuccinimide. The similar organization of the homologous coding regions in other herpesviruses into at least two nested, in-frame, 3'-coterminal genes is discussed.

The 45-kilodalton protein of cytomegalovirus (Colburn) B-capsids is an amino-terminal extension form of the assembly protein

Intranuclear B-capsids from cytomegalovirus (strain Colburn)-infected cells contain an abundant 37-kDa assembly protein, thought to be involved in capsid formation, and three minor protein constituents (i.e., 45, 39, and 38 kDa) that are immunologically and structurally related to the assembly protein. In the experiments reported here, antisera produced against synthetic peptides were used in conjunction with chemical protein cleavage to examine the structural relationship of these proteins in more detail. Results of these experiments verify that the carboxyl end of the 39-kDa assembly protein precursor is lost during maturation and suggest that the 38-kDa protein may be a processing intermediate. It is shown that the 45-kDa protein is coterminal with the mature assembly protein at its carboxyl end but differs by a predicted 115-amino-acid extension at its amino terminus. In addition, evidence is presented that the 45-kDa protein has a 48-kDa precursor and a 47-kDa putative processing intermediate which have the same carboxy-terminal sequences and undergo the same maturational events as those of the assembly protein. A working model considering the structural relationship of these proteins is presented.

Translational efficiency of phi X174 lysis gene E is unaffected by upstream translation of the overlapping gene D reading frame

The lysis gene E of bacteriophage phi X174 is entirely embedded in gene D. Expression studies of genes D and E in Escherichia coli minicells and lysis times obtained in the presence or absence of D translation showed that the simultaneous expression of gene D does not affect protein E production. Thus, unlike other overlapping gene pairs, gene E expression is independent from the upstream translation of gene D. lacZ fusion studies and primer extension inhibition analysis (toeprinting) revealed an intrinsically weak E ribosome-binding site, which seems to be the major factor determining the low expression rate of the gene and thus proper scheduling of cell lysis.

Nucleotide sequences and expression in Escherichia coli of the in-phase overlapping Pseudomonas aeruginosa plcR genes

The translation products of chromosomal DNAs of Pseudomonas aeruginosa encoding phospholipase C (heat-labile hemolysin) have been examined in T7 promoter plasmid vectors and expressed in Escherichia coli cells. A plasmid carrying a 4.7-kilobase (kb) DNA fragment was found to encode the 80-kilodalton (kDa) phospholipase C as well as two more proteins with an apparent molecular mass of 26 and 19 kDa. Expression directed by this DNA fragment with various deletions suggested that the coding region for the two smaller proteins was contained in a 1-kb DNA region. Moreover, the size of both proteins was reduced by the same amount by an internal BglII-BglII DNA deletion, suggesting that they were translated from overlapping genes. Similar results were obtained with another independently cloned 6.1-kb Pseudomonas DNA, which in addition coded for a 31-kDa protein of opposite orientation. The nucleotide sequence of the 1-kb region above revealed an open reading frame with a signal sequence typical of secretory proteins and a potential in-phase internal translation initiation site. Pulse-chase and localization studies in E. coli showed that the 26-kDa protein was a precursor of a secreted periplasmic 23-kDa protein (PlcR1) while the 19-kDa protein (PlcR2) was mostly cytoplasmic. These results indicate the expression of Pseudomonas in-phase overlapping genes in E. coli.

Developmental expression of three proteins from the first gene of the RNA polymerase sigma 43 operon of Bacillus subtilis

The first gene of the Bacillus subtilis RNA polymerase sigma 43 operon, P23, has a protein-coding capacity of 23,000 daltons. Sequence analysis revealed three potential translational initiation sites within the same reading frame, which could encode proteins of 23,000 (P23), 19,000 (P19), and 9,000 (P9) daltons, respectively. An internal promoter (P3), which is expressed only during the sporulation stage, is located between the second and the third translational start sites. By protein fusion to the Escherichia coli beta-galactosidase gene, we showed that all three translational initiation sites of the P23 gene are used in vivo in both E. coli and B. subtilis, and regulation for differential expression of the three proteins during the development of B. subtilis is coupled to the transcriptional promoter switching mechanism. The physiological function of these multiple gene products is unknown and is currently under investigation.

Morphogenetic structures present in lysates of amber mutants of bacteriophage Mu

The Mu phage particle is structurally similar to that of the T-even phages, consisting of an icosahedral head and contractile tail. This study continues an analysis of the morphogenesis of the Mu phage particle by defining the structural defects resulting from mutations in specific Mu genes. Defective lysates produced by induction of 55 amber mutants, representing 24 essential genes, were examined in the electron microscope and categorized into eight classes based on the observed phage-related structures. (1) Mutations in genes lys, F and G, and some H mutations, did not cause a visible alteration in particle structure. (2) Mutants defective in genes A, B, and C produced no detectable phage structures, consistent with their lack of production of late RNA. (3) Extracts defective in genes L, M, Y, N, P, Q, V, W, and R contained only head structures, and these appeared normal. (4) K-defective mutants accumulated free heads as well as free tails which were longer than normal and variable in length. (5) Tails which appeared normal were the only structures found in T- and some I-defective extracts. (6) Free tails and empty heads accumulated in D-, E-, and some I- and H-defective extracts. These heads were as much as 16% smaller than normal heads. The heads found in some I amber lysates had a protruding neck-like structure and unusually thick shells suggestive of a scaffolding-like structure. (7) Defects in gene J resulted in the accumulation of unattached tails and full heads. (8) Previous analysis of lysates produced by inversion-defective gin mutants fixed in the G(+) orientation demonstrated that S and U mutants produced particles lacking tail fibers (F.J. Grundy and M.M. Howe (1984), Virology 134, 296-317). In these experiments with Gin+ phages S and U mutants produced apparently normal phage particles. Presumably the tail fiber defects were masked by the production of S' and U' proteins by G(-) phages in the population.

Gene X of bacteriophage f1 is required for phage DNA synthesis. Mutagenesis of in-frame overlapping genes

The gene II protein of bacteriophage f1 is a site-specific endonuclease required for initiation of phage viral strand DNA synthesis. Within gene II is another gene, X, encoding a protein of unknown function identical to the C-terminal 27% of the gene II protein, and separately translated from codon 300 (AUG) of gene II. By oligonucleotide mutagenesis, we constructed phage mutants in which this codon has been changed to UAG (amber) or UUG (leucine), and propagated them on cells carrying a cloned copy of gene X on a plasmid. The amber mutant makes no gene X protein, and cannot grow in the absence of the complementing plasmid; the leucine-inserting mutant can make gene X protein, and grows normally without the plasmid. Without gene X protein, phage DNA synthesis (particularly viral strand synthesis) is impaired. We discuss this finding in the context of other known in-frame overlapping genes (particularly genes A and A* of phage phi X174), many of which are also involved in the specific initiation of DNA synthesis, and suggest applications for the mutagenic strategy we employed.

Predicted secondary structure similarity in the absence of primary amino acid sequence homology: hepatitis B virus open reading frames

Proteins that are related evolutionarily may have diverged at the level of primary amino acid sequence while maintaining similar secondary structures. Computer analysis has been used to compare the open reading frames of the hepatitis B virus to those of the woodchuck hepatitis virus at the level of amino acid sequence, and to predict the relative hydrophilic character and the secondary structure of putative polypeptides. Similarity is seen at the levels of relative hydrophilicity and secondary structure, in the absence of sequence homology. These data reinforce the proposal that these open reading frames encode viral proteins. Computer analysis of this type can be more generally used to establish structural similarities between proteins that do not share obvious sequence homology as well as to assess whether an open reading frame is fortuitous or codes for a protein.

Plasmid RP4 encodes two forms of a DNA primase

The pri gene locus of the conjugative broad host range plasmid RP4 maps between coordinates 40.3 and 43.5 and encodes two antigenically related forms of a DNA primase with a molecular mass of 118 and 80 kDa (kilodalton). Genesis of these two products has been examined using Pri+-recombinant plasmids. As shown by deletion analysis, the primase polypeptides are tow separate translation products which arise from an in-phase overlapping gene arrangement. It is suggested that transcription of a set of RP4 genes including the pri gene starts at a promoter site within the Tra1 region. In vivo, RP4 mutant primase can apparently substitute for Escherichia coli primase as demonstrated by measuring suppression of the dnaG3 (ts) mutant.

Translational attenuation: the regulation of bacterial resistance to the macrolide-lincosamide-streptogramin B antibiotics

The regulation of ermC is described in detail as an example of regulation on the level of translation. ermC specifies a ribosomal RNA methylase which confers resistance to the macrolide-lincosamide-streptogramin B group of antibiotics. Synthesis of the ermC gene product is induced by erythromycin, a macrolide antibiotic. Stimulation of methylase synthesis is mediated by binding of erythromycin to an unmethylated ribosome. The translational attenuation model, supported by sequencing data and by mutational analysis, proposes that binding of erythromycin causes stalling of a ribosome during translation of a "leader peptide", resulting in isomerization of the ermC transcript from an inactive to an active conformer. The ermC system is analogous to the transcriptional attenuation systems described for certain biosynthetic operons. ermC is unique in that interaction with a small molecule inducer mediates regulation on the translational level. However, it is but one example of nontranscriptional -level control of protein synthesis. Other systems are discussed in which control is also exerted through alterations of RNA conformation and an attempt is made to understand ermC in this more general context. Finally, other positive examples of translational attenuation are presented.

The head genes of bacteriophage 21

Physical and genetic maps of the head genes of lambdoid phage 21 have been made and compared with the head gene map of lambda. Because 21 and lambda have partial sequence homology throughout the head genes it was expected that the head genes of 21 would be analogous to those of lambda. Eight head genes of 21 have been identified and it was found that each of the genes is analogous in position, structure, and/or function to a lambda head gene. Phage 21 genes analogous to the lambda D and FI genes were not identified by mutation. Complementation studies between phage 21 and lambda mutants indicate that only gpFII (the protein product of a gene is referred to as gp (gene product] is fully interchangeable, gpW and gpD are partially interchangeable, and the rest of the head morphogenetic proteins are phage specific. In analogy with phage lambda, it is found that the gpNu3 analog (gp6) of phage 21 is synthesized from the same reading frame as the gpC analog (gp5), resulting in a protein identical to the carboxy terminus of gp5.

Secretion of a lambda 2 immunoglobulin chain is prevented by a single amino acid substitution in its variable region

We have studied two derivatives of the IgA (lambda 2) secreting myeloma cell line MOPC315:MOPC315.26, which produces and secretes a lambda 2 light chain, and MOPC315.37, which produces but does not secrete the lambda 2 chain. It has been reported that the only alteration in the MOPC315-37 lambda 2 chain is located in the variable region (Mosmann and Williamson, (1980) Cell 20, 283-292). In order to determine the nature of this alteration, we cloned the fragment of the chromosome containing the rearranged lambda 2 gene from both the nonsecreting variant MOPC315-37 and the normal lambda 2-secreting parent MOPC315-26 and determined their nucleotide sequence. We found that the nucleotide sequences coding for the leader peptide and for the constant region of the lambda 2 chain were identical in the secretor and nonsecretor. The sequences of the variable region differed at a single base pair corresponding to the first nucleotide in the codon for amino acid number 15. MOPC315-26 has a G in this position creating the codon GGT which codes for glycine, and MOPC315-37 has a C in this position creating the codon CGT which codes for arginine. Thus, we have demonstrated that a single amino acid substitution of a neutral amino acid, glycine, for a positively charged amino acid, arginine, results in the failure of a protein to be secreted.

Overlapping genes at the DNA primase locus of the large plasmid ColI

The sog gene of the large plasmid ColIdrd-1 has previously been shown to encode a DNA primase and a smaller antigenically related polypeptide. Genesis of these two products has been examined using Sog+ recombinant plasmids. Effects of amber mutations, isolated after in vitro mutagenesis, and deletions into or within sog suggest that the smaller polypeptide is a separate translation product which is encoded by DNA specifying the C-terminal region of the larger protein. Under control of the lac promotor, synthesis of both polypeptides is reduced when transcription is repressed. These findings imply that transcription of sog yields a single transcript which is translated from two initiation sites.

Use of lambda pMu bacteriophages to isolate lambda specialized transducing bacteriophages carrying genes for bacterial chemotaxis

A general method for constructing lambda specialized transducing phages is described. The method, which is potentially applicable to any gene of Escherichia coli, is based on using Mu DNA homology to direct the integration of a lambda pMu phage near the genes whose transduction is desired. With this method we isolated a lambda transducing phage carrying all 10 genes in the che gene cluster (map location, 41.5 to 42.5 min). The products of the cheA and tar genes were identified by using transducing phages with amber mutations in these genes. It was established that tar codes for methyl-accepting chemotaxis protein II (molecular weight, 62,000) and that cheA codes for two polypeptides (molecular weights, 76,000 and 66,000). Possible origins of the two cheA polypeptides are discussed.

Overlapping genes at the cheA locus of Escherichia coli

The cheA locus of Escherichia coli, which is required for chemotactic behavior, encodes two polypeptide products designated p[cheA]L and p[cheA]S. The mode of synthesis of these two proteins was investigated by transferring various missense and nonsense mutations to a lambda transducing phage and observing the mutant cheA products made after infection of ultraviolet-irradiated host cells. Missense mutations had no effect on either the size or the relative amounts of the two cheA polypeptides. Most nonsense mutations caused premature translational termination of both cheA products, indicating that p[cheA]L and p[cheA]S must be translated from the same coding sequence in the same reading frame. Two exceptional nonsense alleles at the promoter-proximal end of cheA made an intact p[cheA]s but no detectable p[cheA]L. These findings show that the cheA locus may contain two different sites for initiation of translation. The synthesis of both proteins can be effected by the same promoter, but it is not yet clear whether both are translated from identical mRNA molecules. Complementation studies of cheA mutants provided evidence for two functional activities, one associated with the amino terminus of p[cheA]L and the other with the common portions of p[cheA]L and p[cheA]S. It is possible that each cheA product has a different function required for chemotaxis. The possible roles of these two products and the functional significance of bacterial genes with overlapping coding sequences are discussed.