N-independent leftward transcription in coliphage lambda: deletions, insertions and new promoters bypassing termination functions

Natural RNA Polymerase Aptamers Regulate Transcription in E. coli

In search for RNA signals that modulate transcription via direct interaction with RNA polymerase (RNAP), we deep sequenced an E. coli genomic library enriched for RNAP-binding RNAs. Many natural RNAP-binding aptamers, termed RAPs, were mapped to the genome. Over 60% of E. coli genes carry RAPs in their mRNA. Combining in vitro and in vivo approaches, we characterized a subset of inhibitory RAPs (iRAPs) that promote Rho-dependent transcription termination. A representative iRAP within the coding region of the essential gene, nadD, greatly reduces its transcriptional output in stationary phase and under oxidative stress, demonstrating that iRAPs control gene expression in response to changing environment. The mechanism of iRAPs involves active uncoupling of transcription and translation, making nascent RNA accessible to Rho. iRAPs encoded in the antisense strand also promote gene expression by reducing transcriptional interference. In essence, our work uncovers a broad class of cis-acting RNA signals that globally control bacterial transcription.

Learning from the Leaders: Gene Regulation by the Transcription Termination Factor Rho

The RNA helicase Rho triggers 20-30% of transcription termination events in bacteria. While Rho is associated with most transcription elongation complexes, it only promotes termination of a subset. Recent studies of individual Rho-dependent terminators located within the 5' leader regions of bacterial mRNAs have identified novel mechanisms that govern Rho target specificity and have revealed unanticipated physiological functions for Rho. In particular, the multistep nature of Rho-dependent termination enables regulatory input from determinants beyond the sequence of the Rho loading site, and allows a given Rho-dependent terminator to respond to multiple signals. Further, the unique position of Rho as a sensor of cellular translation has been exploited to regulate the transcription of genes required for protein synthesis, including those specifying Mg(2+) transporters.

Role of an RNase III binding site in transcription termination at lambda nutL by HK022 Nun protein

The phage HK022 Nun protein excludes phage lambda by binding nascent lambda pL and pR transcripts at nutL and nutR, respectively, and inducing transcription termination just downstream of these sites. Termination is more efficient at nutL than at nutR. One difference between nutL and nutR is the presence of RNase III processing sites (rIII) located immediately promoter distal to lambda nutL. We found that deletion of rIII dramatically reduced Nun transcription arrest in vitro but had little effect on termination in vivo. However, consistent with the in vitro results, overexpression of a transcript carrying nutL and rIII efficiently titrated Nun, allowing lambda to grow on a strain that expressed Nun, whereas a transcript carrying only nutL or nutL-rIII with nucleotides 97 to 141 deleted was ineffective. Rnc70, an RNase III mutant that binds but does not cleave rIII, also prevented Nun-mediated lambda exclusion. We propose that rIII enhances the on-rate of Nun at nutL, stimulating Nun-mediated arrest in vitro. We have shown that a specific element in rIII, i.e., box C (G89GUGUGUG), strongly enhances arrest on rIII+ templates. Nun-rIII interactions do not stimulate Nun termination in vivo, presumably because formation of the Nun-nutL complex is normally not rate-limiting in the cell. In contrast to Nun, N is not occluded by Rnc70 and is not efficiently titrated by a nutL-rIII transcript.

Bacteriophage lambda repressor allelic modulation of the Rex exclusion phenotype

The sensitivity of delta red-gam delta ren mutants of bacteriophage lambda to Rex exclusion by lambda rexA+ rexB+ lysogens is modulated by the prophage cI repressor allele. We show the following: (i) lambda spi156 delta nin5 forms plaques on a cI+-rexA+-rexB+ lysogen with 10(5)-fold higher efficiency than on cI[Ts]-rexA+-rexB+ derivatives. (ii) The cI[Ts]857 allele augmentation of Rex exclusion is recessive to cI+. (iii) The cI857-mediated increase in Rex exclusion activity involves the participation of a genetic element mapping outside of cI-rexA-rexB.

Promoter resurrection by activators--a minireview

Frequently, in nature, defective promoters can be resurrected by activator proteins in response to cellular demands. The activators bind to nearby DNA sites for action. Various protein-protein and DNA-protein contacts involving activators, RNA polymerase, and different segments of DNA in and around a defective promoter form a DNA-multiprotein complex (cage) which enhances transcription.

Vectors lambda 200g and lambda 200c: two useful derivatives of lambda 2001

We describe the construction and uses of two new phage lambda replacement vectors. Both are modifications of lambda 2001 which allows selection of cloned inserts by the Spi- phenotype, following the replacement of a gam+ filler fragment. Vector lambda 200g has a second gam gene with an amber mutation in its right arm which enables the selection of clones with inserts in the normal way in Su0 host bacteria and the maintenance of their genetic stability by subsequent growth in Su+ strains. Vector lambda 200c is designed for complementation studies. It contains a cat gene and the attP site of lambda. Recombinants can be inserted into the bacterial chromosome or into an episome with the aid of appropriate helpers.

Bacteriophage lambda P gene shows host killing which is not dependent on lambda DNA replication

Bacteriophage lambda, having a mutation replacing glycine by glutamic acid at the 48th codon of cro, kills the host under N- conditions; we call this the hk mutation. In lambda N-N-cl-hk phage-infected bacteria, the late gene R is expressed to a significant level, phage DNA synthesis occurs with better efficiency, and the Cro activity is around 20% less, all compared to those in lambda N-N-cl-hk(+)-infected bacteria. Segments of lambda DNA from the left of pR to the right of tR2, carrying cro, cII, O, P, and the genes of the nin5 region from the above hk and hk+ phages, were cloned in pBR322. Studies with these plasmids and their derivatives having one or more of the lambda genes deleted indicate that the hk mutation is lethal only when a functional P gene is also present. When expression of P from pR is elevated, due to the deletion of tR1, host killing also occurs without the hk mutation. We conclude that the higher levels of P protein, produced either (1) when cro has the hk mutation or (2) when tR1 is deleted, are lethal to the host. We also show that due to the hk mutation, the Cro protein becomes partially defective in its negative regulation at pR, resulting in the expression of P to a lethal level even in the absence of N protein-mediated antitermination. This P protein-induced host killing depends neither on lambda DNA replication nor on any other gene functions of the phage.

Thermodynamic and enzymological characterization of the interaction between transcription termination factor rho and lambda cro mRNA

Termination of transcription at tR1, the rho-dependent terminator between genes cro and cII of bacteriophage lambda, is mediated by interactions between rho protein and an RNA sequence element called rut. We show, using a filter retention assay technique, that rho protein binds with about 10-fold lower affinity to variants of cro RNA lacking both parts of rut or to normal cro RNA having one or the other part of rut bound to a complementary DNA oligonucleotide than it binds to unmodified cro RNA. These same variant and modified forms are nearly devoid of the strong rho ATPase cofactor activity of cro RNA. Estimates of binding energies of the rho-cro RNA interaction under different conditions reveal that termination function correlates with about 12.6 kcal of binding energy, of which two-thirds is due to nonelectrostatic interactions. The rut segment is shown to contribute about 1 kcal, nearly all to nonelectrostatic interactions. KCl is found to be more effective than potassium glutamate as a competitive counterion, and a decrease in 1.4 kcal of binding energy due to counterion competition correlates with a loss of termination and ATPase activities. In sum, the results indicate that the rut sequence contributes substantially to the overall binding affinity, that ionic interactions are also important, and that mere binding of rho to RNA is not sufficient for rho ATPase activation.

Rho-dependent transcription termination in the tyrT operon of Escherichia coli

A 178-bp repeat sequence comprises the distal end of the tyrT operon and also encodes the signal for Rho-dependent transcription termination. It is shown here that Rho-dependent transcription termination is highly efficient for each of the individually cloned first and second repeats. The presence of either repeat results in a termination efficiency of 88% (12% readthrough). When an individual DNA repeat is shortened, starting from the promoter-proximal end, so that only 56 bp remain upstream from the termination site, Rho-dependent termination activity is reduced approximately three-fold (to give 38% readthrough) and with further shortening, so that only 15 bp remain upstream, Rho-dependent termination is reduced an additional 1.8-fold (to give 59% readthrough). Finally, it is shown that the presence or absence of the tyrT terminator upstream from the lambda tR1 terminator does not affect the efficiency of transcription at termination lambda tr1.

Nucleotide sequence homologies in control regions of prokaryotic genomes

Functional recognition sites for several regulatory factors, including RNA polymerase, cyclic adenosine monophosphate receptor protein and ribosomes, do not always have strong consensus nucleotide sequence homology, yet they are capable of biological activity. Using the computer, other nucleotide sequences can be found that have equal or significantly greater consensus homology, but whose biological function has not been characterized. This analysis shows that no arbitrary 'cutoff score' can successfully distinguish active recognition sites from uncharacterized homologies, due to the great natural diversity in the strength and conservation of functional sites. It also predicts that the strong 'cryptic' homologies presented here are of two types: some might already have a biological function which has so far not been detected, whereas certain single-point mutations might be able to confer activity upon the others by correcting a key structural defect.

Modification enhancement by the restriction alleviation protein (Ral) of bacteriophage lambda

The product of the lambda ral gene alleviates restriction and enhances modification by the Escherichia coli K-12 restriction and modification system. An open reading frame (orf) located between genes N and Ea10 has been assigned to the ral gene. We have cloned this orf in a plasmid where its transcription is controlled by a thermolabile lambda repressor. Inactivation of the lambda repressor caused a 1000-fold reduction in K-specific restriction of unmodified lambda phage and a 100-fold increase in modification. In minicells transformed with ral+ plasmids, derepression resulted in the appearance of a polypeptide with a lower mobility than that predicted for a protein encoded by the orf attributed to ral; in a transcription and translation system in vitro DNA from a ral+ plasmid encoded a polypeptide with the same mobility. This polypeptide was absent when the plasmid DNA carried a mutant ral gene. The nucleotide sequence of this mutant gene defined two base changes, one of which inactivates the initiation codon of the orf. The K restriction endonuclease, which is also a K-specific methylase, is encoded by three genes designated hsdR, hsdM and hsdS, although the hsdR polypeptide is not essential for the methylase activity. We show that Ral enhances modification in a host strain lacking the entire hsdR gene, and lambda phages carrying the hsdM and S genes modify their own DNA inefficiently in the absence of Ral, despite the fact that derivatives of these phages provide efficient amplification of the K-specific methylase. Our data support a model in which, as a consequence of the interaction of Ral with either the hsdM or the hsdS polypeptide, the conformation of the enzyme is changed and the efficiency of methylation of unmodified target sites is enhanced. It has been postulated that Ral counteracts Rho, but in our experiments Ral did not relieve transcriptional polarity.

Transcription termination and processing sites in the bacteriophage lambda pL operon

S1 nuclease mapping was performed on transcripts from the major leftward operon of the bacteriophage lambda in order to locate the 3' ends of stable RNA species produced in vivo. The analysis was carried out on RNA purified from either an induced lambda prophage or bacteria carrying a plasmid containing a large segment of lambda including the intact PL operon through the bet gene. The S1 nuclease mapping was performed on transcripts produced in the presence and the absence of the N antitermination function, and in the presence and the absence of either the RNase III processing enzyme or the Rho factor. The results of this work indicate that the intercistronic region between the N and ral genes of lambda contains three sites at which transcripts end under N-Rho+ conditions (positions on the lambda sequence: 34,826, 34,558 and 34,393). The distal two correspond to the two sites previously described in this region as tL1 (on both sides of the BamHI site). In the region between ral and Ea10, we mapped the 3' ends of three species of RNA. The 3' end of one species was found to be located 90 nucleotides proximal to tL2a, at 34,000 in the lambda sequence. The terminator at this site may be partially N-resistant. In an RNase III deficient host, an additional RNA species is formed. The 3' end of this RNA species is located at tL2a (33,910 on the lambda sequence). In the presence of the antitermination N gene product, the readthrough transcripts are processed to form a 3' end at position 33,980 on the lambda sequence. These results suggest that elongation of transcription of the lambda PL operon is reduced gradually by clusters of termination located between genes and that the expression of the terminated products is further controlled by processing of the mRNA.

RNA sequence and secondary structure requirements for rho-dependent transcription termination

The interaction of E. coli termination factor rho with the nascent RNA transcript appears to be a central feature of the rho-dependent transcription termination process. Based on in vitro studies of the rho-dependent termination of the transcript initiated at the PR promoter of bacteriophage lambda, and on earlier studies, Morgan, Bear and von Hippel (J. Biol. Chem. 258, 9565-9574, 1983) proposed a model defining the features of a potential binding site for rho protein on transcripts subject to rho-dependent termination. This model suggested that an effective rho binding site on a nascent RNA transcript should be: (i) greater than 70-80 nucleotide residues in length; (ii) essentially unencumbered with stable secondary structure; (iii) relatively sequence non-specific; and (iv) located within a few hundred nucleotide residues upstream of the potential rho-dependent terminus. In this paper we examine the sequences and secondary structures of several transcripts that exhibit rho-dependent termination to test this hypothesis further. Unstructured regions of approximately the expected size and location were found on all the transcripts examined. Though several short specific sequence elements were found to occur in a very similar arrangement on the lambda PR- and lambda PL-initiated transcripts of lambda phage, no such elements of sequence regularity were found on any of the other rho-dependent transcripts. The results of the sequence comparisons reported here strongly support the generality of the "unstructured binding site" hypothesis for rho-dependent termination.

Conservation of genome form but not sequence in the transcription antitermination determinants of bacteriophages lambda, phi 21 and P22

Comparisons are made among DNA sequences upstream from terminators in both leftwards and rightwards early operons of related coliphages lambda, phi 21 and P22. These sequences include both left and right determinants of response to phage-coded antitermination proteins, "N", as well as the N structural genes themselves. Despite almost total disparity of DNA sequence, the three genomes can be discerned to include the same elements in the same order and spacing: downstream from the early left promoter are sequentially a site of recognition for host nusA protein, a dyad symmetry "nut" essential for N function in lambda, overlapping sites for processing of the transcript by RNAase III and then the N structural genes; downstream from the cro gene on the right are sites of nusA recognition and nut dyad symmetries homologous to those on the left. Because the N proteins of lambda, phi 21 and P22 do not for the most part complement each other, a specific site of N recognition has been postulated for each N-responding operon. The nut dyad symmetry qualifies as such a site, since the loop of the left dyad in lambda is marked by mutations that block N function leftwards, and since DNA sequences here show close homology between the loops of left and right dyads for each phage, but less if not little homology for different phages.

Studies on polylysogens containing lambda N-cI- prophages. II. Role of high multiplicities in lysogen formation by lambda N-cI- phage

Results of the experiments presented in this paper show that lambda N-cI- phage can lysogenize a nonpermissive host Escherichia coli when it infects at very high multiplicities (around 100), and lambda N-cI-cII- and lambda cIII-N-cI- lysogenize poorly at similar high multiplicities. The latter two phages lysogenize with appreciable frequency when either lambda N-cI- or lambda int-cN-cI-cII- is used as helper. The phages, lambda N-cI-, lambda N-cI-cII-, and lambda cIII-N-cI- can lysogenize also at relatively low m.o.i. of 20 in presence of the above lambda int-c helper, and the lambda int-cN-cI-cII- phage alone forms converted lysogens at an m.o.i. as low as 12. All these results suggest that the establishment of prophage integration by lambda N-cI- is positively regulated, like lambda N+cI+ phage, by the cII/cIII-promoted expression of the int gene of lambda, and under the N- condition, high multiplicities are needed to provide optimum levels of cII and cIII products, especially the latter.

Sequence changes in coliphage lambda mutants affecting the nutL antitermination site and termination by tL1 and tL2

The 17-bp sequence designated nutL is required for the N-mediated antitermination of transcription in the major leftward operon of coliphage lambda. The single-stranded sequence can be folded into a hairpin structure. Ten independently isolated spontaneous lambda nutL- mutants have changes that affect the same nucleotide, located in the loop of the hairpin structure, changing the guanine to adenine, thymine or cytosine. Another mutant (lambda nutL3), selected by a different means, has a deletion of one GC base pair and thus eliminates one C in the stem of the hairpin structure, destabilizing it -11.2 to -2.2 kcal/mol. True reversions of the nutL point mutations restore the guanine. The second-site revertant lambda ninL99 was found to have a deletion of 417 bp between the tL1 terminator and the N gene, removing bases +523 to +939 (counted from SL = +1). This deletion include codons for the six carboxy-terminal amino acids of gene N product, but the fusion allows continuation of translation for 53 additional amino acid residues beyond the truncated N gene before reaching a nonsense codon. The fused N product is active.

Synthesis of the nutL DNA segments and analysis of antitermination and termination functions in coliphage lambda

The nut antiterminator sequence, when present between a promoter and a terminator, permits the N-mediated antitermination of transcription in phage lambda. The efficiency of nutL was determined by assaying the activity of gene galK placed on a plasmid downstream from the promoter, nutL and terminator modules. As a reference and an estimate of the plasmid copy number, we have used an improved and very reproducible assay for bla activity. Sequences consisting of the 17-bp nutL core flanked by two HindIII cohesive sites were synthesized by the phosphite coupling method, and cloned in proper orientation between the Pp promoter of pBR322 and lambda gene N followed by the tL1 terminator on a galK-expression plasmid. The antitermination efficiencies for two synthetic 17-bp nutL sequences, one wild type and one point mutant at the base of the nutL stem, are similar but substantially reduced in comparison with the native 25-bp nutL sequence cloned at the same site in the otherwise identical galK-expression plasmid. Multiple tandem insertions of the synthetic 17-bp nutL segment successively increase antitermination efficiency, but also to levels below those of comparable plasmids carrying multiple copies of the native 25-bp nutL sequence. Thus, several specific base pairs in the flanking sequences appear to be important for the efficient nut function. In an inverted orientation the 17-bp nutL sequence has lost its antitermination function. It also lost the termination activity exhibited by inversion of the longer 25-bp and 74-bp native nutL sequences.

Amplification of maize ribulose bisphosphate carboxylase large subunit synthesis in E. coli by transcriptional fusion with the lambda N operon

The maize chloroplast gene coding for the large subunit of ribulose bisphosphate carboxylase (3-phospho-D-glycerate carboxy-lyase (dimerizing), EC 4.1.1.39) has been placed under the transcriptional control of the bacteriophage lambda promoter PL, by fusion with the lambda N operon located on a multicopy plasmid. Transcription from PL was repressed at 32 degrees C by the presence in the E. coli chromosome of a cIts gene that specifies a temperature-sensitive repressor. After inactivation of the repressor at 45 degrees C unmoderated transcription of the chloroplast gene occurred from the PL promoter. Translation was probably initiated from a chloroplast Shine-Dalgarno sequence located five nucleotides from the N-terminal methionine initiation codon to yield a polypeptide the same size as that synthesised in maize. This direct translation results in a level of expression of the chloroplast gene corresponding to approximately 2% of the total E. coli cell protein as ribulose bisphosphate carboxylase large subunits. Transcriptional fusions with the lambda N operon should provide a generally applicable, simple method for the amplification and regulation of chloroplast gene expression in E. coli.

Transcription termination: sequence and function of the rho-independent tL3 terminator in the major leftward operon of bacteriophage lambda

For cloning, assaying the function and sequencing terminators, we have constructed the pD12 plasmid, in which the late promotor p'R of phage lambda controls the expression of the galK gene of the pK03 plasmid of McKenney et al. (1981). The lambda tL3 terminator region was cloned in this plasmid between the promoter and the galK gene, and found to be 90-94% effective in preventing galactokinase expression in both rho+ and rho- hosts. Is is also active in vitro, both in the presence or absence of the rho factor. The termination point is located at 4320 bp to the left of the SL startpoint of the PL-RNA, just downstream of gene exo. We have sequenced 356 bp of the hitherto uncharted lambda DNA to the right of the TaqI cut, which in turn is 110 bp to the right of the b522 deletion at 63.9% lambda. The tL3 terminator has several features common to other rho-independent termination sequences, including an 81% G+C-rich region of 2X8-bp symmetry ("stem") with a 5-bp intervening "loop", partially overlapping and followed by a sequence transcribed into the pyrimidine-rich CCUUUCU-OH 3' terminus of the RNA. The termination point that follows the last U was determined by the S1 mapping technique.

Antitermination and termination functions of the cloned nutL, N, and tL1 modules of coliphage lambda

A plasmid containing a pp-galK operon was constructed to assay galK expression as a measure of transcriptional regulation by the cloned nutL, N and tL1 modules in a Rho+, Nus+ Escherichia coli host. Insertion of tL1 and rut, both carried on a lambda fragment located between the bamHI site and gene N, between the promoter and galK reduced expression of galK by 80--90%. Gene N alone, when controlled by pp, stimulated galK expression by about two-fold. Cloning of both gene N and nutL in the proper orientation resulted in about a 60% decrease in the tL1 termination. Additional tandem nutL sites increased efficiency of antitermination. A shortened nutL segment, containing only 25 bp of the original genomic nutL sequence, was found to have nearly equal ability to bring about antitermination. If the orientation of the nutL fragment is reversed, antitermination is abolished and the insert now displays a termination function. Termination efficiency is 60 to 73% for one to four tandem nutL modules in reverse orientation. Similarly, the inverted N module acts as a terminator, with 67% efficiency for one and 90% for two tandem inserts. Termination by inverted nutL or N fragments is probably unrelated to their normal functions, but indicates a fortuitous presence of a terminator sequence in the inverted orientation.

Novel bacteriophage lambda cloning vector

A simple method for generating phage collections representing eukaryotic genomes has been developed by using a novel bacteriophage lambda vector, lambda 1059. The phage is a BamHI substitution vector that accommodates DNA fragments 6-24 kilobases long. Production of recombinants in lambda 1059 requires deletion of the lambda red and gamma genes. The recombinants are therefore spi- and may be separated from the spi+ vector phages by plating on strains lysogenic for bacteriophage P2. Random fragments suitable for insertion into lambda 1059 are obtained by partial digestion of high molecular weight eukaryotic DNA with Sau3a. This restriction enzyme cleaves at the sequence G-A-T-C and leaves a 5'-tetranucleotide "sticky end." Because G-A-T-C extensions are also produced by BamHI cleavage, these fragments may be annealed directly to BamHI-cleaved lambda 1059. By using these methods, a set of clones covering the entire Caenorhabditis elegans genome was constructed. DNA segments which include the unc-54 myosin heavy chain gene have been isolated from this collection.

A comprehensive molecular map of bacteriophage lambda

Physical and genetic mapping of deletion mutations has been correlated with the available molecular sizes of the lambda gene products and the DNA base sequence to construct a comprehensive molecular map of the phage lambda genome. The physical length of the DNA making up the left arm from the cos site through gene J is not sufficient to account in a nonoverlapping manner for all the proteins of the sizes reported to be coded, especially in the Nu1--C region. In the right arm all the coding capacity has not been accounted for, and it appears to be oversaturated only in the gam-ral region. The positions of several IS and Tn elements, and of restriction endonuclease cleavage sites are specified.

The site controlling the specificity of N action is outside the promoter-operator region: a triple hybrid phage lambda N21 imm434nin5

A short interval of homology between imm lambda, imm434 and imm21 DNAs was identified near the leftward promoter-operator region. This homology, denoted Hs, was revealed by electron microscopic examination of lambda imm lambda/lambda imm21 and lambda imm434/lambda imm21 heteroduplexes, and permitted us to construct a special lambda hybrid (lambda hyB) which contains the N region of phage 21 and the adjacent imm region from phage 434. This triple hybrid, labmda N21 imm434nin5, was analysed by genetic, transcriptional and electronic micrographic techniques. Its leftward and rightward promoter-operator regions are of phage 434 specificity and are controlled by the 434 repressor. Surprisingly, the N21 gene of lambda hyB was found to be defective, perhaps to preserve the viability of the hybrid. Its leftward N-recognition system (nutL) is of phage 21 specificity since it responds only to the N21 function in complementation tests, as measured by antitermination of leftward transcription initiated at the pL promotor in the imm434 region. We conclude, therefore, that the pLoL region of 434 contains no information for the specificity of N antitermination. Both lambda imm21 and lambda hyB were found to be missing the tL1 terminator function (see also Salstrom and Szybalski, 1978b). In these phages, the tL2 terminator was found to be only 60% effective under N21 conditions, and therefore expression of their red-gam genes is sufficient to endow the lambda hyB and lambda N21- imm21nin5 phages with the Fec+ phenotype.