A programme for the construction of a lambda phage

The lambdoid bacteriophages are a family of viruses which multiply in Escherichia coli. Infection by a lambdoid phage can either cause death of the host cell by lysis or can lead to the stable integration of the viral DNA as a prophage into the host chromosome, with survival of a lysogenic cell.

The lytic and lysogenic responses both require the co-ordinated activities of several functions, encoded by both viral and host genes. In lytic infection, DNA replication, genome packaging and cell lysis proceed sequentially. In the lysogenic mode, the synthesis of the phage integration system and the repressor protein must be synchronized. In both cases, functions that might interfere with the chosen pathway must be excluded. Thus the lambdoid phage infection follows a simple, properly co-ordinated developmental plan.

While no one would claim that phage infection provides a valid model system for the study of development, the present, detailed understanding of the lambda developmental programme might have some valuable lessons for developmental biologists working on more complex systems. This review will present a simple account of the key steps in the lytic and lysogenic pathways, with emphasis on the control mechanisms governing co-ordination and commitment to a particular pathway.

The genome of bacteriophage lambda contains approximately fifty genes, about half of which are essential for lytic growth. The DNA is packaged into the head of the mature phage particle as a non-permuted, linear, duplex molecule with single-stranded, 5′ projections of 12 nucleotides at each end. These mutually cohesive termini assure rapid circularization of the genome following infection.

The genetic and physical map of the lambda chromosome (Fig. 1) shows marked clustering of genes of related function. The essential genes concerned with head (A-F) and tail (Z-J) formation and assembly are contiguous within the left-hand third of the genome. The region of the map between J and att, the phage attachment site, contains no essential genes (Hendrix, 1971). Genes to the right of att govern site-specific (int and xis) and generalized (red) recombination of phage DNA. None of the genes between att and N is essential for lytic growth of lambda on normal hosts, though they may affect growth on certain mutant host strains. The product of gene N is the early regulatory protein that is normally necessary to activate transcription of most other phage genes. The cl gene codes for the lambda repressor, the regulatory protein that switches off transcription of prophage genes in the lysogenic state. The presence of the lambda repressor makes a λ-lysogenic cell immune to superinfection by another lambda phage and is responsible for the characteristic turbidity of lambda plaques. Lambda carries an additional regulatory gene, cro, that encodes a repressor-like, DNA-binding protein. The cro gene-product functions as an antagonist to the cI-product, acting to prevent lysogenic development and to promote lytic growth (Gussin et al. 1983). The O and P genes are required for replication of λDNA (Brooks, 1965; Joyner, Isaacs, Echols & Sly, 1966), the Q gene product for activation of late transcription (Dove, 1966; Couturier, Dambly & Thomas, 1973) and the S, R and Rz gene products for lysis of host cells (Harris, Mount, Fuerst & Siminovitch, 1967;.Young et al. 1979).

The infection of an E. coli cell by lambda is initiated when a phage particle adsorbs, via its tail fibre, to a specific receptor on the cell surface (Schwartz, 1976; Thirion & Hofnung, 1972). The linear phage genome is injected through the cell membrane into the cytoplasm (Mackay & Bode, 1976), where it rapidly circularizes via its cohesive termini and is covalently sealed by host DNA ligase.

Transcription of lambda genes by the host’s RNA polymerase proceeds left-wards from promoter pL through gene N and rightwards from promoter PR through the cro gene (Fig. 2). Most of these initial transcripts terminate t_L1 at and t_R1 immediately beyond genes N and cro respectively. The leftward transcript is translated to yield the N protein and the early rightward transcript produces the cro protein. Both of these gene products have important regulatory functions that influence subsequent λ development.

The N protein exerts its controlling effect by influencing RNA polymerase to ignore transcription-termination signals (Adhya, Gottesman & de Crombrugghe, 1974; Franklin, 1974; Segawa & Imamoto, 1974). In the presence of A protein, transcription initiated at pt and PR is elongated through the terminators íLi and íRi and through a series of ‘delayed early’ genes whose products regulate both the lytic and the lysogenic responses. The infecting phage genome shortly thereafter becomes committed to one pathway or the other.

The apparent specificity of the N protein for transcripts initiating at PL and PR is not a consequence of the promoters themselves, but of nucleotide sequences, termed N-utilization (‘nut’) sites, located between the promoters and the first transcription termination signals (Salstrom & Szybalski, 1978; de Crombrugghe, Mudryj, Di Lauro & Gottesman, 1979). The two nut sites, which have 16 out of 17 identical nucleotides, show hyphenated dyad symmetry, consistent with the formation of a stable stem-loop structure in the RNA or the DNA (Rosenberg et al. 1978). Though the nut sequences have been shown to be sufficient to allow the A-protein to influence RNA polymerase to ignore subsequent termination sequences (de Crombrugghe et al. 1979), the precise mechanism by which this is achieved has not been elucidated.

During the phase of uncommitted growth, phage DNA replication is activated by transcription in the vicinity of the unique origin within the O gene (Dove et al. 1969; Dove, Inokuhi & Stevens, 1971) and by the products of the genes O and P (Ogawa & Tomizawa, 1968). The early phase of replication, giving monomeric circular molecules, ceases after a few minutes. At this stage the genome either stops replicating and enters the lysogenic phase or switches to rolling-circle replication as it begins the lytic cycle.

During lytic growth, the production of mature phage particles requires the synthesis of concatemeric phage DNA, composed of covalently joined, tandemly repeated unit copies of the lambda chromosome (Szpirer & Brachet, 1970; Stahl et al. 1972; Feiss & Margulies, 1973). The rolling-circle mode of replication is believed to provide the mechanism for the production of the concatemeric molecules (Eisen, Pereira de Silva & Jacob, 1969; Gilbert & Dressier, 1969), which are the ideal substrate for DNA encapsidation (Skalka, 1977). The packaging of lambda DNA into phage heads requires the unit chromosome to be bounded by cos sites (Emmons, 1974; Feiss & Campbell, 1974); a monomeric circular lambda chromosome cannot be packaged in vivo (Szpirer & Brachet, 1970).

The onset of rolling-circle replication is facilitated by lambda’s gam gene product, an inhibitor of the host’s recB, C-encoded exonuclease V (Unger & Clark, 1972; Unger, Echols & Clark, 1972; Enquist & Skalka, 1973). In the absence of the gam-product this nuclease attacks both the rolling circles themselves and an intermediate required for their formation, thus inhibiting the synthesis of maturable phage DNA.

An alternative route to packageable DNA can be provided by recombination between two monomeric circular molecules to produce a circular dimer containing two cos sites. Both the red systems of the phage and the rec system of the host can catalyse such recombinational dimerization (Stahl et al. 1972; Enquist & Skalka, 1973), but the latter does so very inefficiently with wild-type lambda DNA (Stahl, Crasemann & Stahl, 1975).

While ‘delayed early’ leftward transcription gives rise to gam gene expression, similar transcription rightwards facilitates expression of the Q gene. Note that the expression of the Q gene will necessarily be delayed by some 2 ·5 min, the time required for RNA polymerase to reach the gene from its promoter PR, 7000 base-pairs away. The product of Q activates transcription initiated at P′R to traverse the late genes, whose products are responsible for DNA encapsidation and cell lysis. The Q protein acts as an antiterminator, employing a Q-utilization Çquf) site in a manner formally analogous to the N protein, allowing transcripts initiated at P′R to proceed through the termination sequence, Z′R into the large late gene region (Forbes & Herskowitz, 1982).

Delayed early transcription begins to diminish after about 5 min of infection, dropping to a very low level by 10 min (Szybalski et al. 1970). This diminution is caused by the binding of the cro gene-product to the OL and OR operator sequences, interfering with the initiation of transcription at pL and PR (Fig. 3). The kinetics of action of the cro-product are explained by the observation that the active form of the DNA-binding protein is a dimer, and that relatively high concentrations are needed to block the RNA polymerase-binding sites at pt and PR .

From about 10 min until the end of infection, transcription from P′R activated by the Q protein is predominant and expresses all the late genes. The products of the A and Nul genes aggregate to form the terminase, which binds to the concatemeric substrate DNA at the cos sites. Several other late gene products are assembled into a capsid prohead, while others are independently assembled to form the tail fibre. The prohead attaches to a DNA-terminase complex, and a unit length of the phage chromosome is taken up until the next terminase complex is reached. The terminase then cleaves the DNA to yield a mature, DNA-filled head. The final stage of capsid assembly is the addition of a tail fibre to the mature phage head.

Coincident with the production of mature virions is the continuing expression of the lysis genes. About 60min after infection initiated, the accumulated products of the S, R and Rz genes are sufficient to lyse the host cell, releasing a burst of progeny phage.

The establishment of the lysogenic state involves the coordinated expression of two key genes, the cl gene that encodes the phage repressor and the int gene, whose product catalyses the site-specific integration event. Transcription of cl and int is synchronized by the action of a positive effector, the ell gene-product, which stimulates the binding of RNA polymerase to the separate promoters, pre and pint (Ho & Rosenberg, 1982) (Fig. 4). The cll-product is metabolically unstable, due to the action of the host’s HflA protein, but is protected by the lambda cIII product (Hoyt et al. 1982). The controlling ell and cIII genes are both expressed mainly by delayed early transcription from PR and pt during the uncommitted phase of phage development.

The cI-encoded repressor is the effector of the switch into lysogenic development to the exclusion of lytic growth. The protein binds to the phage DNA at the OL and OR operators, with initial preference for regions OLI and ORI . These operator sub-sites are located within the RNA polymerase-binding sites and the bound repressor prevents further transcription from initiating at PL and PR . The repressor, in blocking transcription of the N, gam and red genes from pt and the O, P and Q genes fromPR, effectively stops phage DNA replication, late protein synthesis and host cell lysis.

The repressed, circular phage DNA is integrated into the host chromosome by a site-specific recombination event catalysed by the mi-protein (integrase). The int gene is transcribed from p_int, under the influence of cII protein, and from PL, driven by the action of N-protein. However, an interesting post-transcriptional mechanism, ‘retroregulation’, ensures that the latter transcription is non-productive during phage infection (Fig. 5). This phenomenon depends on a exacting regulatory element, sib, located downstream from int, beyond the phage attachment site (Mascarenhas, Kelley & Campbell, 1981; Guarneros, Montanez, Hernandez & Court, 1982). The sib sequence displays extensive dyad symmetry, enabling the elongated RNA transcript to form a hairpin structure that is believed to be a preferential substrate for RNAse III. The action of this enzyme on the RNA could create a 3′ end, processive 3′ to 5′ degradation from which destroys the adjacent int region of the transcript (Guarneros et al. 1982).

Transcription from the cII-dependent promoter, p_int, terminates at a rho- dependent terminator, t_int, located within the sib region (Schmeissner, Court, McKenney & Rosenberg, 1981), and fails to complete the RNAse Ill-sensitive structure. This regulatory mechanism ensures that only those int transcripts initiated at p_int are effective as integrase messenger, and the requirement for ell protein ensures coordination with repressor synthesis.

In the prophage of a λ-lysogenic cell the sib site has been removed from proximity to the int gene by the integrative recombination event involving the phage attachment site. On induction of a λ-lysogenic cell the adjacent int and xis genes, whose products are both required to bring about excision of the prophage DNA, are coordinately expressed via transcription from PL

The alternative modes of expressing int during infection and induction thus ensures that the appropriate protein is produced to drive the recombination event in the required direction.

Host proteins are also important in modifying the lysogenic response. The E. coli HflA protein decreases the amount of active cII-product in the infected cell, but is antagonized by the phage cIII-product. The synthesis of the HflA protein itself is repressed by the host cyclic AMP-activated catabolite repression system (Belfort & Wulff, 1974). The integrative host factor (IHF), a dimeric protein comprised of the products of the E. coli him A and himD genes (Miller & Nash, 1981; Nash & Robertson, 1981) has a dual function in lysogenic development. It is a DNA-binding protein that is essential for integrative recombination (Nash & Robertson, 1981) and promotes the synthesis of the phage ell product (Hoyut et al. 1982).

Overall, a single regulatory protein, the ell gene-product, coordinates the lysogenic response. Not only does the ell protein synchronize the expression of the cl and int genes, but it also acts as the receptor protein for transmission of environmental signals via the host IHF and HflA control proteins.

Within the phage-infected cell, an individual lambda genome can enter either the lytic or the lysogenic pathway. Although these two pathways have their initial steps in common, a decision between them is made at an early stage of phage development.

The particular pathway that ensues depends upon whether the cro-protein or the repressor wins the competition for the operator sites, particularly OR . The cro-product is able to lock the genome into the lytic pathway, inhibiting the expression of CII and cIII and ultimately preventing repressor synthesis. The repressor ensures the lysogenic response by preventing expression of cro itself, and also of the O, P, and Q genes whose products are required for DNA replication and late gene expression.

The delicately balanced competition between repressor and cro protein is influenced by several factors, the level of CII protein being a crucial one. The CII product is required to activate transcription of the CI gene from PRE, and its activity is influenced by several host- and phage-coded products. When the activity of the CII-product is high, the lysogenic response is favoured; when it is low, the lytic response is enhanced. Thus, for example, carbon-starved cells, which have reduced levels of HflA protein, favour activity of the CII protein and show elevated frequencies of lysogenization (Belfort & Wulff, 1974).

In the lysogenic cell, the prophage state is maintained by the binding of repressor to the operator sites, OL and OR, blocking transcription from the two major early promoters, PL and PR .

Each of the operators OL and OR contains three repressor-binding sites (Ptashne et al. 1976) (Fig. 6). These are similar in sequence and contain axes of two-fold hyphenated symmetry. The three sites in OR overlap both PR, the major early rightward promoter, and OR, the promoter from which the cI gene is transcribed in the prophage. At low concentrations, repressor binds preferentially, as a dimer, at ORI : such binding potentiates subsequent binding of repressor to OR2 (Johnson, Meyer & Ptashne, 1979). Binding to 0R3 shows reduced affinity and is only achieved at high concentrations of repressor.

The binding of repressor dimers′at OR1 and OR2 stimulates the binding of the E. coli RNA polymerase to/?RM, by providing extra contact sites at this position, thereby enhancing transcription of the cl gene (Reichardt & Kaiser, 1971; Meyer, Maurer & Ptashne, 1980). The repressor, in blocking transcription from PR, is also acting to stimulate its own synthesis (Fig. 7). At high intracellular concentrations repressor will bind to OR3 and inhibit its own synthesis, thus providing an effective homeostatic mechanism.

When the DNA of a λ -lysogenic cell is damaged, for example by ultraviolet irradiation, thymine starvation or a chemical carcinogen, the prophage is induced into the lytic cycle. The induction process is triggered by the proteolytic cleavage of the repressor, catalysed by the activated product of the cellular rec A gene (Roberts, Roberts & Craig, 1978; Craig & Roberts, 1981). Destruction of the repressor leads first to transcription fromPR, with production of the cro gene product, and slightly later to transcription from pL. (Repressor binds more tightly at OL than at OR .) The cro-protein, also active as a dimer, has high affinity for the sites within OR, though in the order OR3 >OR2 = OR1 . Binding of croprotein to OR3 blocks access of RNA polymerase to PRM, eliminating the maintenance mode of repressor synthesis (Fig. 7). This action of the cro-product effectively prevents recovery of repressor synthesis and locks the induced prophage into the lytic cycle.

Lambda has evolved an effective system for monitoring the environment in the infected cell and using this information in the simple decision between lysis and lysogeny. The alternative responses branch from a common pathway, the ultimate direction being governed by the products of the cro and cl genes. These two DNA-binding proteins compete for the same region of the phage DNA and repress each other’s synthesis. The balance between the cl and cro products is modulated by the ell protein, whose activity is sensitive to the cell’s metabolic state via its interaction with the host HflA gene-product.

Lambda’s relatively simple control circuits reveal great variety in the mechanisms by which gene expression is controlled. Genes can have alternative modes of expression; control proteins can affect either initiation or termination of transcription and can act at multiple sites to coordinate the expression of separate genes, and the same protein can be both an activator and a repressor of transcription. It is generally the case that regulatory interactions between competing pathways are reinforced, to prevent possible interference, and that many regulatory proteins are metabolically unstable, allowing rapid switching should circumstances change. It will be interesting to see how many of these themes are rediscovered in examining the mechanisms of determination and commitment in more complex developmental systems.