On the role of Cro in λ prophage induction

Results

The formation of an O _L–CI–O _R regulatory complex containing a DNA loop has been demonstrated to increase repression of P _RM and P _R (8, 10, 11). This additional level of regulation is required for the prophage's compensatory response to low doses of DNA damage, which sets the threshold level of DNA damage to which the prophage responds by induction (11). Repressor bound to O _L can also increase repression of the P _R promoter when artificially placed 3,600 bp downstream of the P _R transcription start site, suggesting that the formation of an O _L–CI–O _R complex might also affect transcription from the early lytic promoters (9). To study the effects of the O _L–CI–O _R DNA loop on regulation of the early lytic promoters, we constructed P _R::lacZ fusions by inserting part of the immunity region of λ between two reporter genes at the lac locus of E. coli as described in Supporting Text and Table 4, which are published as supporting information on the PNAS web site, and shown in Fig. 1 C and D. The constructs fuse the cII gene with lacZ, rendering the expression of β-galactosidase under P _R control. P _R is one of the two early lytic promoters that are repressed in the lysogen but activated upon induction; thus, production of β-galactosidase in the fusion strains reflects induction of the lytic pathway.

Strains NC416 and NC417 contain the part of the λ immunity region ranging from the AUG initiation codon of gene N gene on the left to the 30th codon of cII on the right. These strains retain both the left and the right operator regions (Fig. 1D). In strains NC414 and NC415, only the right operator region is retained; the strains carry the λ immunity region from the AUG initiation codon of the rexA gene on the left to the 30th codon of cII on the right (Fig. 1C). The AUG codon of N or rexA is fused to the 2nd codon of the luc reporter gene, while the 30th codon of cII is fused to the 7th codon of lacZ. In strains NC414 and NC415, luciferase is expressed from P _RM, whereas in NC416 and NC417, the reporter is under the control of the lytic P _L promoter. The fusions also vary with respect to the cro gene. NC414 and NC416 carry the wild-type cro gene, whereas NC415 and NC417 carry a nonfunctional missense cro27 allele (12). Together, the four strains represent the four possible combinations of the or ΔOL and cro ⁺ or cro ^– genotypes.

The O_L–CI–O_R Complex Increases Repression of P_R in the Native Prophage Configuration. The expression of β-galactosidase in log phase NC414, NC415, NC416, and NC417 cultures growing at 30°C under repressed conditions is shown in Table 1. It is evident that the presence of the O _L region increases repression of P _R when the immunity region is located in the chromosome in its native configuration confirming the results obtained in multicopy systems (8–10). Note that our O _L-deletion also removes the rex genes between O _L and cI. We addressed the possibility that the decreased repression in the O _L-deletion is not because of the absence of rex and demonstrated that the presence or absence of rex genes in the background does not measurably affect repression of P _R at temperatures ranging from 30°C to 37°C (data not shown). In the absence of functional Cro, the β-galactosidase activity is ≈4-fold higher in the absence of the left operator region than in its presence, reflecting that O _L increases the repression of P _R 4-fold. The 4-fold increase in P _R activity with O _L deleted is essentially the same as when CI is supplied at a constant level in trans (8–10), showing that O _L is increasing the efficiency of P _R repression at the normal lysogenic concentration of CI by >4-fold. We noted that in the absence of the left operator region, P _R activity is decreased ≈30% in the cro ^– mutant (NC415) as compared to cro ⁺ (NC414). This observation could either reflect a uniformly elevated expression from P _R in the cro ⁺ cells in comparison with the cro ^– cells, or it could reflect that a fraction of the cro ⁺ population has switched to the “lytic” state where P _R is not repressed by CI. The latter possibility is supported by the observation that, when the four strains are grown on MacConkey lactose agar plates, red colonies appear more frequently among NC414 (O _L ^– cro ⁺) cells than among NC415, NC416, and NC417 cells (Fig. 2). The red color indicates that these cells are producing higher levels of β-galactosidase, which metabolizes lactose. However, when both O _L and O _R regions are present, we were unable to detect any difference between cro ⁺ and cro ^– in P _R expression (Table 1, compare NC416 and NC417). This observation could indicate that the interaction between O _L and O _R prevents Cro from switching the prophage from the lysogenic to the lytic state. To examine this hypothesis further, we studied the activity of P _R at low repressor concentrations. The cI allele present in our reporter strains is the cI857 allele, which encodes a temperature-sensitive repressor protein (22). The cI857 allele behaves as wild-type cI at 30°C, whereas there is essentially no repressor activity at temperatures above 40°C. Thus, a range of repressor activities can be obtained by growing the strains at a range of temperatures between the permissive and restrictive temperatures.

Role of Cro in the Switch Process? The expression of lacZ as a function of growth temperature for NC414, NC415, NC41,6 and NC417 is shown in Fig. 3. If cro is critical for the switch (14), we expect that derepression of P _R will occur at lower temperatures in the cro ⁺ than in the cro ^– cells, suggesting that Cro can sway the switch from lysogenic to lytic growth at intermediate concentrations of CI. We obtained that result for the strains that do not carr y the left operator region: NC414(cro ⁺) and NC415(cro ^–) (Fig. 3). However, the expected pattern was not seen in the strains that carry both the left and the right operator regions. Although the role of rex genes, if any, in the switch process remains to be investigated in strains containing both operator regions, no difference in β-galactosidase levels were observed between NC416(cro ⁺) and NC417(cro ^–) cells. In fact, the derepression of P _R as a function of growth temperature appears identical for NC415, NC416, and NC417, suggesting that the DNA loop between the left and right operator region renders the switch insensitive to the presence of the cro ⁺ gene.

The Effect of Cro Mediated Repression of P_R Transcription in the Presence and Absence of O_L. We investigated whether the left operator region had any influence on the regulation of P _R in the absence of functional CI. We inactivated CI by shifting growing cultures from 30°C to 42°C and followed the kinetics of P _R derepression by monitoring β-galactosidase activity at various times after the temperature switch. The activities of β-galactosidase as a function of time after induction for NC414, NC415, NC416, and NC417 are shown in Fig. 4. For NC414 and NC416, which carry cro ⁺, it can be seen that ≈10 min after induction, Cro presumably has accumulated to a level where it binds O _R1 and/or O _R2 and represses P _R, causing the cro ⁺ curves to diverge from the cro ^– curves (NC415 and NC417). It took equally long for Cro to accumulate to a concentration where it negatively autoregulates P _R in NC414 (ΔO _L) and NC416 (), and the degree of Cro-mediated repression of P _R was similar in the two strains (Table 2). Therefore, we saw no indication that the left operator region participated in regulation of P _R by Cro when CI was absent as expected (8), consistent with the hypothesis that O _L participates in regulation at O _R through the CI-mediated DNA loop (10).

Repression by Cro. It was suggested earlier on that the primary role of Cro in prophage induction is to repress the early lytic promoters, P _R and P _L, as well as the maintenance promoter P _RM. The relationship between CI and Cro is that CI is a strong repressor specialized for complete turnoff of lytic functions, whereas Cro is a weak repressor functioning in a partial turn down of the early lytic promoters to allow progression into the late lytic phase (23, 24). Our unique divergent reporter constructs allowed us to measure the degree of repression of P _R, P _L, and P _RM by CI and Cro in this study. To measure the degree of repression of P _L by Cro, we compared the luciferase activity of NC416 (N::luc, cro ⁺) and NC417 (N::luc, cro ^–) at 42°C, where CI857 is inactive (Table 3). Likewise, to measure the degree of repression of P _RM by Cro, we compared luciferase in NC414 (rexA::luc, cro ⁺) and NC415 (rexA::luc, cro ^–) at 42°C (Table 3). The degree of repression of P _L and P _R by CI in the presence of both the left and right operator regions was estimated by comparing the activity of the respective reporter genes in NC417(P _R::lacZ, P _L::luc, cro ^–) grown at 30°C, where CI857 is active, with NC417 grown at 42°C, where CI857 is inactive (Table 3). The luciferase protein is inherently heat sensitive and forms inactive aggregates after temperature shifts, but enzyme activity is restored with the aid of host chaperones after a few minutes at the high temperature (25). We observed full restoration of luciferase activity ≈15 min after shifting from 30°C to 42°C in a reporter containing wild-type CI (not heat-sensitive; strain WA5) (data not shown). Luciferase activities reported here were measured 60 min after the temperature shift. We observed that, in the absence of CI, Cro represses the P _L promoter ≈1.7-fold (Table 3), although a stronger repression was reported when a fusion in which transcription terminators were present between P _L and a reporter gene (galK) was used (26). In the current bidirectional reporter setup, Cro repressed P _RM promoter ≈2.5-fold. (Table 3).

Discussion

Repression of P_R by the O_L–CI–O_R Complex in its Native Configuration. Tight repression of P _R is believed to be vital for the maintenance of lysogeny, because otherwise Cro would be produced, bind to O _R 3, and repress P _RM, turning off cI transcription (27). Mathematical modeling studies of the lysis–lysogeny switch have concluded that CI bound at O _R cannot repress Cro production sufficiently to stably maintain the lysogen unless there are additional levels of repression of P _R (16, 28). We have confirmed that an interaction between the left and right operator region increases CI-mediated repression of P _R 4-fold. Thus, our data are consistent with that of Revet et al. (9) and Dodd et al. (8) in suggesting that the DNA loop between O _L and O _R constitutes the missing level of regulation. Our data also support the model of Dodd et al. (8) to explain that the means by which the presence of O _L allows tighter repression of P _R at physiological CI concentrations. The extra level of cooperativity added by the formation of a CI octamer increases the binding affinity of CI to the involved operator sites, thus lowering the critical concentration of CI needed to maintain lysogeny.

O_L Contributes to Repression of Cro Synthesis. According to the classic model, Cro's major role during induction is to bind O _R 3 and thereby inhibit cI expression from P _RM (14, 27). If a Cro-mediated reduction in cI expression were important for effective induction of the early lytic promoters, we would expect a cI857 cro ^– prophage grown at intermediate temperatures to contain more repressor than the corresponding cro ⁺ prophage, and thus require inactivation of a larger fraction of the repressor proteins to effectively derepress P _R. Instead, we observe the same amount of derepression of P _R at the same growth temperature in cro ⁺ (NC416) and cro ^– (NC417), indicating that the same fraction of CI must be denatured to derepress P _R in the two strains (Fig. 3). It is possible that differences in derepression temperature exist, which we cannot detect with our assay. The fraction of CI inactivated is not a linear function of the growth temperature, and it is likely that the concentration of active CI varies dramatically within a limited range of growth temperatures (22).

When O _L–CI–O _R complex formation is prevented by the absence of O _L, P _R is less repressed, more Cro is produced, P _RM is more repressed, and less CI is made, so it is easier to switch to the lytic mode. In contrast, in the presence of O _L, P _R is tightly repressed, and so Cro production is blocked, thus maintaining the lysogenic mode. These results suggest that Cro may not play any role in the switch from lysogenic to lytic state of λ and are in agreement with the recent findings of Dodd and colleagues (8, 10), who showed that Cro may have a lesser role in prophage induction than previously perceived.

Cro as a Weak Repressor of Early Lytic Functions. The primary role of Cro in prophage induction may be to repress P _R and P _L after inactivation of CI (23, 24). In support of this hypothesis, a λcI₈₅₇ cro ^– phage is capable of forming plaques when grown at 37°C, although no plaques are observed at 30°C or 42°C (24). Apparently, efficient plaque formation requires partial repression of the early lytic promoters and the repression can be exerted either by Cro or can, in the absence of Cro, be substituted by partially active CI. The Gibbs free energies of CI and Cro dimers binding to the operator regions (29, 30) support a stronger repression of P _R and P _L by CI than by Cro, and this has also been demonstrated for P _R in vivo (31). Our double reporter constructs allowed us to measure the degree of repression of P _R and P _L by Cro and CI all in the same genetic background, allowing a direct comparison of the obtained values. As expected, we observe a much higher degree of repression of both lytic promoters by CI (≈400-fold) in the absence of Cro than by Cro (≈1.7-fold) in the absence of CI, consistent with the idea that CI acts as a strong repressor of lytic functions and Cro acts as a weak repressor of the early lytic promoters. Because Cro plays little, if any, role in effecting the lysogenic to lytic epigenetic switch, as discussed above, and because single-copy cro partly represses P _RM and the lytic promoters, we conclude that this ability of Cro to partly repress P _RM and/or the lytic promoters is not needed in inducing the “thermolabile lysogen.” Formally, it remains possible that Cro repression of one or more of these promoters is needed for induction of the complete prophage.