This Article Abstract Full Text (PDF) Supplemental material Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints and Permissions Copyright Information Books from ASM Press MicrobeWorld Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Michalowski, C. B. Articles by Little, J. W. Search for Related Content PubMed PubMed Citation Articles by Michalowski, C. B. Articles by Little, J. W.

Sequence Tolerance of the Phage P_RM Promoter: Implications for Evolution of Gene Regulatory Circuitry

Christine B. Michalowski,¹ Megan D. Short,^1, and John W. Little^1,2*

Department of Biochemistry and Molecular Biophysics,¹ Department of Molecular and Cellular Biology, University of Arizona, Tucson, Arizona²

Received 18 February 2004/ Accepted 1 September 2004

	ABSTRACT

Top Abstract Introduction Materials and Methods Results Discussion References

 
Much of the gene regulatory circuitry of phage centers on a complex region called the O_R region. This 100-bp region is densely packed with regulatory sites, including two promoters and three repressor-binding sites. The dense packing of this region is likely to impose severe constraints on its ability to change during evolution, raising the question of how the specific arrangement of sites and their exact sequences could evolve to their present form. Here we ask whether the sequence of a cis-acting site can be widely varied while retaining its function; if it can, evolution could proceed by a larger number of paths. To help address this question, we developed a cloning vector that allowed us to clone fragments spanning the O_R region. By using this vector, we carried out intensive mutagenesis of the P_RM promoter, which drives expression of CI repressor and is activated by CI itself. We made a pool of fragments in which 8 of the 12 positions in the –35 and –10 regions were randomized and cloned this pool into the vector, making a pool of P_RM variant phage. About 10% of the P_RM variants were able to lysogenize, suggesting that the regulatory circuitry is compatible with a wide range of P_RM sequences. Analysis of several of these phages indicated a range of behaviors in prophage induction. Several isolates had induction properties similar to those of the wild type, and their promoters resembled the wild type in their responses to CI. We term this property of different sequences allowing roughly equivalent function "sequence tolerance " and discuss its role in the evolution of gene regulatory circuitry.

	INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

 
Complex gene regulatory circuits can have a large number of interlocking components. This degree of interconnectivity raises two issues. First, how did these circuits evolve? Second, how can we understand the behavior of these existing circuits and predict their behavior in the face of small changes in parameters such as promoter strength? For these and other reasons, we have been analyzing the behavior of the regulatory circuitry of phage in the intact system. This system is probably the best-understood complex circuit (38, 39). Most, if not all, of the regulatory interactions have been identified, and most of these are well characterized at the mechanistic level. Previous analysis of this circuit generally has been carried out in uncoupled systems (such as the use of reporter genes and fusions with the lac promoter), an approach necessary to disentangle the causality of this system. With a circuit diagram in hand, it is now possible to return to the intact circuit and ask how particular changes affect the overall operation of the system.

Many of the critical interactions in the circuit center on a complex regulatory region termed the O_R region (Fig. 1C). This 100-bp region is densely packed with cis-acting sites, including two promoters and three sites to which both the CI and Cro repressors can bind (38). In addition, the promoters and repressor-binding sites overlap extensively. CI and Cro regulate the expression of these promoters, and their actions determine the overall behavior of the regulatory circuitry.

View larger version (22K): [in this window] [in a new window]   FIG. 1. Structure of the JL387 vector and sequence of PRM. All maps are to scale, as indicated. (A) Comparison with l+. The indicated BsrGI and XhoI sites in JL351 were introduced; that for BsrGI is the only BsrGI site to the right of the XhoI site. The letter X in the JL387 map indicates the site of the XhoI site removed by filling in. (B) Maps of cI and cro. The NsiI and BglII sites shown were used for routine subcloning of the OR region and are present in both and JL351. (C) Close-up of the OR region showing the locations of the promoters and operators and those of the new XhoI and BsrGI sites. Cro and CI both bind to sites OR1, OR2, and OR3. JL387 contains v1 and v3 mutations (not shown; see text) not present in JL351 or the wild type. (D) Sequence of the PRM promoter, inverted from the usual order of the map. The locations of OR3 and the end of OR2 are shown in bold; the –35 and –10 regions of PRM are underlined, and the NNNN sequences represent positions that were randomized (Fig. S3 of the supplemental material contains more details). Above the sequence is that of a consensus promoter. Positions at which PRM matches it are underlined.

It is of interest to ask how this particular arrangement and spacing of cis-acting sites benefits the phage, how it arose during the course of evolution, and how the sequences of the sites were refined during the course of evolution. We chose to address the last question by asking whether one of these cis-acting sites, the P_RM promoter, could tolerate substantial genetic changes and still allow relatively normal behavior of the intact circuitry.

Although unstimulated P_RM is a relatively weak promoter, it differs from the consensus ⁷⁰-specific promoter in only 4 positions out of 12 (Fig. 1D). Is its resemblance to the consensus important, or could it tolerate substantial changes and still preserve its function? Previous work and our unpublished studies have identified several point mutations in P_RM that weaken or strengthen it, and these mutations affect the behavior of the circuitry (12, 31, 42, 43; our unpublished work), suggesting that its sequence is important. It is unclear whether more extensive changes in the promoter would destroy the operation of the circuitry. For instance, promoter mutations might also affect other processes operating in this complicated regulatory region. In this work, we have tested the effects of extensive changes in P_RM on the operation of the circuitry.

In order to facilitate this and similar studies, it would be useful to be able to clone variants of the O_R region into a wild-type background. This would allow one, for example, to use intensive site- or region-directed mutagenesis to create large pools of recombinant molecules and then to introduce these into intact genomes and assess their effects by using the powerful screens and selections afforded by genetics. This approach could be extended to combinatorial approaches in which many combinations are made and tested simultaneously. We describe here the isolation of a cloning vector that allows such manipulations in the O_R regulatory region and use it to carry out extensive mutagenesis of P_RM. We find that sequences differing markedly from the wild-type sequence can confer nearly normal behavior on the circuitry.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

 
Reagents and media. Restriction enzymes and T4 DNA ligase were from New England Biolabs (Beverly, Mass.), Boehringer-Mannheim, and Promega Corp. (Madison, Wis.). Taq DNA polymerase was from Roche. Polymyxin B and mitomycin C were from Sigma. Oligonucleotides were from Midland Certified Reagent Corp. (Midland, Tex.). Shrimp alkaline phosphatase was from United States Biochemicals. Streptavidin-coated paramagnetic beads (product CM01N), in a 1% (wt/wt) slurry, were from Bangs Laboratories (Fishers, Ind.); 5-bromo-4-chloro-3-indolyl-ß-D-galactopyranoside (X-Gal) was from Research Products International Corp. The DNA isolation kit was from QIAGEN (Valencia, Calif.). The Phagemaker packaging kit was from Novagen (Madison, Wis.). TE was 10 mM Tris-HCl (pH 8.0) plus 1 mM EDTA; BTE was TE plus 0.1 M NaHCO₃; TZ8 (7) was 100 mM Tris-HCl (pH 8.0) plus 1 mM MgSO₄ plus 10 mM KCl. LB and tryptone broth were as previously described (27) and supplemented with antibiotics as appropriate at the levels previously described (27), except that lysogens of JL351 and the P_RM variants were grown with 10 µg of kanamycin per ml and were not fully resistant to 30 µg/ml. LBGM and LBMM were LB with 1 mM MgSO₄ and either 0.2% glucose or 0.2% maltose, respectively (27). DNA sequencing was carried out as previously described (27). For each MD phage (see Table 2), DNA sequencing was done on a PCR product extending from the distal end of cro to the end of cI; sequencing was done from the end of cro through about the first third of cI, spanning the entire cloned region.

Construction of cloning vector JL387 and preparation of vector arms. For a description of the construction of cloning vector JL387 and the preparation of vector arms, see Fig. S1 and S2 of the supplemental material.

Preparation of doped P_RM pools. A collection of DNA fragments bearing randomized bases in the eight positions shown by the NNNN sequences in Fig. 1D was prepared as outlined in Fig. S3 of the supplemental material.

Cloning in JL387. Vector arms in equimolar amounts at a total concentration of 75 µg/ml were mixed with a twofold molar excess of the doped P_RM pool, treated with DNA ligase overnight at 16°C in a total volume of 10 µl, and packaged with a commercial packaging kit (Novagen). From 0.75 µg of input DNA we recovered only 20,000 plaque-forming phage, far fewer than expected from the stated efficiency of this kit (10⁷ plaques per µg of ligation mixture), and infer that ligation or packaging was inefficient.

Subcloning of P_RM variants. The vector pJWL502 (from Andrea Watson) was the same as pJWL344 (Fig. S1C of the supplemental material), except that it lacked the SpeI site; also, the SalI site had been removed by cutting, filling in, and ligation; hence, the HincII site within O_R was unique. pJWL502 was cut with NsiI and HincII as a backbone; inserts were a PCR fragment (with JL163 as the template) from the NsiI site to a BsiHKAI site at codon 2 of cI and BsiHKAI-HincII fragments from PCR products of various P_RM variants. The resulting plasmids lacked the XhoI and BsrGI sites while retaining the variant P_RM and were used in two ways. First, the P_RM variants were crossed into a wild-type background by a cross with JL161; recombinants were identified as turbid-plaque formers, verified by DNA sequencing, and termed MDWx. Second, to make protein fusions of these promoters with lacZ, the plasmids were used as PCR templates with the primers GGGGGGGATCCTTAAGGCGACGTGC and CCCCCGAATTCAACCTCCTTAGTAC, followed by digestion with EcoRI and BamHI and cloning into pRS414. The resulting protein fusions contained the P_RM variants and the first 18 amino acids of CI fused to lacZ. Codons in cI are numbered beginning not with the N-terminal Met, since this is removed posttranslationally (45, 46), but with the following Ser.

Modification of Simons vectors. We modified the Simons vector RS45 (50) in three ways (detailed in the supplemental material). First, we isolated a Kan^r derivative, allowing selection of lysogens. Second, we deleted the distal part of the lac operon, from the middle of lacY to the end of the operon (7). Finally, we made another derivative in which a variant of the O_L region was placed downstream of lacZ. These steps created a pair of phages, JL611 (shown in Fig. S4 of the supplemental material) and JL628, without and with the O_L variant, respectively. These phages were crossed with derivatives of pRS414 bearing various P_RM variants (see above) to create reporter phages used in the assay described below. In derivatives of JL628 with P_RM::lacZ protein fusions, the O_L site lay 3.6 kb from O_R, slightly closer than in the phages of Dodd et al. (7) (3.8 kb) and more distant than in (2.46 kb).

Reporter assays. ß-Galactosidase assays were performed as previously described (7), with several modifications, with the Molecular Devices Spectra MAX 340PC plate reader. Overnight cultures were diluted 1:6,000 with shaking in tubes in 3 ml of LBGM without and with graded amounts of isopropyl-ß-D-thiogalactopyranoside (IPTG) at 37°C to 1 x 10⁸ to 3 x 10⁸ cells per ml. Cell density was measured by determining optical density at 600 nm (OD₆₀₀) with a plate reader (250 µl of culture per well); 50 µl of culture was added to 150 µl of TZ8 assay buffer (TZ8 plus 2.7 µl of ß-mercaptoethanol per ml and 2.5 µl of 2% polymyxin B per ml). Assays were run in triplicate in flat-bottom 96-well plates (Evergreen Scientific) at 28°C for 1 or 5 h after addition of 50 µl of 0.4% ONPG (o-nitrophenyl-ß-D-galactopyranoside) dissolved in TZ8. Units were defined by the following formula: 1,000 x (OD₄₂₀/min)/(OD₆₀₀ x 1.5 x volume of culture in milliliters). Units defined in this way are roughly comparable to Miller units (32).

For each P_RM variant, assays for reporter gene function were carried out with two pairs of strains. In each pair, the host carried either pA3B2, which provided a low level of CI (JL6740), or pJWL486, a derivative of pACYC184 (like pA3B2) but lacking cI (JL6853). The pairs differed in whether the fusion had the O_L variant distal to lacZ. In all cases, single lysogens were used (37). Strains with pA3B2 carried pJWL615 as a source of Lac repressor; various levels of CI were supplied in these strains by adding graded amounts of IPTG.

Distinguishing vector from recombinants. Most recombinants formed clear plaques. These were distinguished from the JL387 vector, which bears v1 and v3 and also forms clear plaques, as follows. The v1 and v3 mutations are present in vir, which can grow in a lysogen. vir also contains a third mutation, v2, in O_L1, allowing expression of the P_L operon. CI binds poorly to the mutant operators and cannot repress the lytic genes of vir. Since the only essential gene product from the P_L operon is the N protein, v1 v3 should be virulent if N is provided in trans, as in strain JL5434. and a cI mutant could not form plaques on JL5434, but v1 v3 and JL387 formed small, clear plaques. Hence, clear plaques formed on JL2497 were tested individually for the ability to plate on JL5434. Alternatively, phage were plated on a 10:1 mixture of JL2497 and JL5434; v1 v3 and JL387 formed clear plaques, while , cI, and recombinants formed turbid plaques.

Prophage induction. Experiments were carried out as previously described (27), with two exceptions. First, after that study we changed the brand of UV lamp used for irradiation and found that the set point for the wild type changed from 5 to 12 J/m², a difference we attribute to changes in the spectral properties of the lamp relative to those of the meter (International Light model IL1400A) used to measure its output. Second, phage titers were determined with lawns made from exponentially growing cells (JL2497 was grown to 2 x 10⁸/ml in LBMM, centrifuged, and resuspended in 1/10 volume of TMG [27]),rather than the starved overnight lawns used in the previous studies. This change gave higher and more reproducible titers.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

 
Approach. Although has been widely used as a cloning vector, it has seldom been used in this way for analyzing itself. As described below, we developed a vector that allows us to clone fragments carrying only the O_R region and a small portion of cI. We then used this vector to carry out cassette mutagenesis of the P_RM promoter and analyzed progeny with properties resembling those of wild-type .

We judged the resemblance of the behavior of P_RM variants to that of the wild type by several criteria. First was the ability to grow lytically, which is not a demanding criterion since P_RM has no role in lytic growth. Second was the ability to establish and maintain a stable lysogenic state. Expression of CI from P_RM is required for maintenance of this state. Third was the ability to undergo prophage induction, or the "genetic switch," in which the regulatory state changes from lysogenic to lytic upon induction of the host SOS system (26, 41).

Design of the cloning vector. The cloning vector used, termed JL387 (Fig. 1), has several useful features. First, it carries a kanamycin resistance marker replacing a portion of the b2 region, allowing selection of lysogens. Second, it has two new restriction sites flanking the O_R region (Fig. 1), both arising by single-base changes: a unique, silent XhoI site located at codons 12 to 13 of cI, and a BsrGI site at the start of the P_R transcript. These sites were used in preparing vector DNA for cloning (Fig. S2 of the supplemental material). Third, to distinguish the vector from recombinants bearing inserts in the O_R region, it contains two mutations in the O_R operators, v1 in O_R2 and v3 in O_R1. These mutations prevent lysogeny (16), so the cloning vector formed clear plaques. Recombinants forming turbid plaques could easily be recognized by screening for this phenotype. In addition, we developed genetic tests to distinguish recombinants forming clear plaques from the vector (see Materials and Methods).

JL351, which differs from the cloning vector only in that it lacks mutations in the O_R operators, was used as a wild-type control for the behavior of recombinants. This phage differed phenotypically from wild-type in two ways. First, it formed plaques that were slightly less turbid than the wild type. Second, in prophage induction experiments (not shown), lysogens of this phage were induced at UV doses roughly half of those of the wild type (see Discussion). While these differences can limit the usefulness of this vector, as described below, its use has allowed us to gain initial insights into regulation that can then be explored in a more systematic way.

Doping of the P_RM promoter. We made a pool of recombinant DNA molecules in which 8 of the 12 bases in P_RM were randomized (Fig. 1D; detailed in the supplemental material). We avoided the first two positions of the –35 and –10 regions because these overlap O_R2 and O_R3, respectively (Fig. 1D), and changes in these positions would also affect CI and Cro binding. This approach should allow 4⁸ different P_RM variants. This recombinant pool was cloned into JL387, yielding about 20,000 plaque-forming phage; hence, not all possible P_RM variants were recovered. The resulting pool of packaged phage was then characterized.

Wild-type forms turbid plaques because lysogens arise in the plaque and continue to grow. By contrast, many mutants, including those defective in CI, unable to bind CI (such as the cloning vector), or unable to make CI due to a defect in P_RM, form clear plaques. Most of the phage in the pool formed clear plaques, either because they had an O_R region from the vector or because P_RM was nonfunctional. To assess the fraction of phage containing an insert, individual clear plaques formed on JL2497 were tested on JL5434, which allows phage containing the O_R region from the vector, but not those containing an insert, to make plaques (see Materials and Methods); about half of the clear plaques tested were recombinants. On JL2497, about 5% of the total phage formed turbid plaques, indicating that roughly 10% of the lytically competent P_RM variants could form turbid plaques. The analysis below suggests that recombinants that are not lytically competent are likely to be rare. In any case, the ability of many isolates to form turbid plaques suggested that these phages might be able to lysogenize and hence that their P_RM promoter was functional in the context of the circuitry.

Properties of P_RM variants. We purified and further characterized 61 turbid-plaque-forming phage, picking isolates with a range of phenotypes (Table 2). The degree of turbidity varied substantially, from barely turbid to more turbid than JL351. A few isolates formed plaques more uniformly turbid across the plaque than those of JL351 or wild-type , for which the center of the plaque was more turbid than the periphery.

From all but seven of the isolates, we could readily isolate lysogens by streaking plaques on kanamycin plates. Among those that lysogenized, six isolates gave small numbers of lysogens but the colonies were healthy; most of these isolates also formed plaques less turbid than those of the wild type. Nine other isolates conferred a phenotype we term the unstable-lysogen phenotype; most of the Kan^r colonies of these isolates were small, with a few large ones interspersed. Extensive analysis of another phage, prm240, which contains a very weak P_RM promoter and confers this same phenotype, suggests that the small colonies are single lysogens, which are barely stable (unpublished data), while the larger colonies are multiple lysogens. We infer, but have not confirmed directly, that the present isolates that confer the unstable-lysogen phenotype likewise had very weak P_RM promoters and that the stability of lysogens was improved by multiple lysogeny.

One might expect that all of the variants forming turbid plaques would be able to form stable lysogens, but this was not the case. Lysogens arising in a plaque might not be stable in isolation, because the conditions in a plaque differ from those affecting single cells streaked on a plate. Lysogens in a plaque are frequently superinfected by phage. A phage with a weak P_RM could follow the lysogenic pathway after infection, since CI is expressed during the establishment phase from the much stronger P_RE promoter. Hence, lysogens can arise as a plaque grows. CI is then expressed from P_RM, which might be too weak to maintain lysogeny in a single copy. In a plaque, superinfecting phage would also express P_RM, leading to higher CI levels through a gene dosage effect and maintenance of lysogeny. In contrast, when cells are isolated by streaking, superinfection no longer occurs and unstable lysogens could switch to the lytic state. In other studies, we have also isolated numerous derivatives of O_R323 (27) that form turbid plaques but not stable lysogens (27; our unpublished data), and the prm240 phage with the unstable-lysogen phenotype described above forms plaques indistinguishable from those of the wild type.

Sequence analysis of the P_RM variants revealed a wide range of P_RM sequences (Table 2). Each isolate had a different sequence; MD29 was the same as the wild type. These sequences are further analyzed in the Discussion. We conclude from these findings that a very wide range of P_RM sequences is compatible with at least some functioning of the regulatory circuitry.

Prophage induction. All of these phage isolates were able to grow lytically (as expected, since they were isolated as plaques and since P_RM has no role in lytic growth), and most could form stable lysogens, indicating that these aspects of the regulatory circuitry were functional. As a more stringent test for normal behavior of the regulatory circuitry, we asked whether the P_RM variants could undergo prophage induction. In this process, often termed the genetic switch, a lysogen is induced to switch from the lysogenic state to the lytic state by induction of the host SOS regulatory system (26, 41). Treatment with DNA-damaging agents, such as UV irradiation, activates the co-protease activity of RecA (24), which mediates cleavage and inactivation of CI, leading to prophage induction.

In wild-type , prophage induction has several characteristics, each of which can be assessed in the mutants. First, it gives a burst size of about 100 phage per induced cell, comparable to that seen after infection. Second, it exhibits threshold behavior. Induction is inefficient at low levels of DNA damage (such as that which occurs with low doses of UV light) but abruptly becomes efficient at higher doses. Finally, and related, this threshold behavior has a particular set point, which is the dose at which induction becomes efficient. Mutants are known that have changes in any one or combinations of these characteristics (27; our unpublished data).

To identify P_RM variants resembling the wild type, we carried out spot tests (not shown) on plates with low and various levels of mitomycin C, an SOS-inducing agent. With increasing mitomycin C concentrations, phage spots became progressively less turbid. Among the isolates that appeared in this test to be as hard as or harder to induce than JL351, nine were selected for further characterization. Because JL351 differed somewhat from the wild type in prophage induction, we cloned the variant P_RM regions into a wild-type background (see Materials and Methods), giving variants termed MDWx.

Prophage induction experiments were performed (Fig. 2) with single lysogens of these variants. A variety of dose responses was evident. Several isolates showed curves similar to that of the wild type. Some were harder to induce, some gave a small burst size, and mutant MDW14 was scarcely induced at all. We conclude that several different P_RM variants can be found that allow relatively normal prophage induction.

View larger version (24K): [in this window] [in a new window]   FIG. 2. UV induction of selected MDW phages. Nine variants of PRM were subcloned into a wild-type (WT) background (see text), making phages termed MDW for MD wild type. UV induction curves were obtained for single lysogens as described in Materials and Methods. Results of typical experiments are shown. The data in each panel were obtained in a single experiment.

With this system, we analyzed promoter strengths for these nine P_RM variants. All of those similar to the wild type were stimulated by CI (Fig. 3A); however, all five variants differed somewhat from the wild type either in their basal (unstimulated) levels, in the maximum level seen in the presence of CI, or in both features. One variant, MDW28, had about the same stimulated level but a higher basal level, three others were stimulated to somewhat lower levels, and MDW49 was about half as strong as the wild type, although this result is hard to interpret since MDW49 has an extra base pair at the end of the P_RM –10 region. We conclude that variants similar but not identical to wild-type P_RM in this uncoupled assay were able to substitute effectively for wild-type P_RM in the context of the intact circuitry.

View larger version (28K): [in this window] [in a new window]   FIG. 3. Promoter activity and response to various levels of CI. Single lysogens of phages bearing PRM::lacZ protein fusions were prepared. For each, two host strains were used. The first, JL6853, contained no CI; the second, JL6740, contained pA3B2 (51), which is regulated by the Lac repressor and makes low levels of CI. Cells were grown in the presence of the indicated levels of IPTG, and then ß-galactosidase levels were measured as described in Materials and Methods. Results of a typical experiment are shown. (A and B) Activity in the absence of OL. (C and D) Activity in the presence of an OL variant located distal to the lacZ reporter. WT, wild type.

P_RM is also under negative autoregulation (7, 30), an effect mediated in large part by looping between CI dimers bound at the O_R region and those bound at O_L (7). We tested the same isolates in a reporter gene construct carrying a version of O_L distal to lacZ. We found that the variant promoters also were negatively autoregulated (Fig. 3C and D), that the magnitude of this effect was similar to that seen with the wild type, and that their relative strengths were similar to those seen in the absence of O_L.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

 
We first discuss the cloning vector used in these studies. We then analyze the sequences of the P_RM variants and discuss the mechanisms by which this promoter is expressed. Finally, we consider the implications of our findings for the evolution of complex regulatory circuitry.

Cloning vector. Use of JL387 allowed us to carry out localized combinatorial mutagenesis of the O_R region. The same strategy can be used with any restriction site that is unique to the right of the XhoI site (Fig. 1). For instance, we have used an NsiI site located beyond cro to carry out extensive modification of the region between this site and the XhoI site (1).

A major advantage of this vector is that it allows us to identify variants with a desired behavior in the context of an intact regulatory circuit. In this way, promising candidates for further analysis can readily be identified and then subjected to more intensive analysis, much as is done for rapid screens such as two-hybrid systems or microarrays.

As noted above, the regulatory behavior of JL351 (the O_R⁺ version of this vector) was not completely wild type. The silent XhoI site in cI (Fig. 1) is responsible for this, at least in part. A phage carrying the XhoI site but not the BsrGI site resembled JL351 in prophage induction and plaque turbidity, while a phage lacking both sites had a set point and turbidity like those of wild-type (data not shown). Also, in the presence of the XhoI site the strength of wild-type P_RM was reduced by about 50%, with or without O_L (Fig. 3B and D). Perhaps these effects result from changes in the translation efficiency of the leaderless cI mRNA (e.g., see reference 49) or in its stability. In addition, in other versions of the vector, we found (see supplemental material) that minor changes at the start of the P_R transcript also affected the circuitry, since phages with the XhoI site in cI and an SpeI site or a Bsu36I site instead of the BsrGI site at the start of the P_R transcript formed plaques less turbid than those of the XhoI-bearing phage. The O_R region appears to be extremely sensitive to several seemingly innocuous changes, which makes our findings obtained with P_RM variants even more unexpected (see below).

P_RM variants. Our data indicate that many P_RM sequences are compatible with relatively normal regulatory behavior. The ease with which we obtained variants with behavior close to that of the wild type strongly suggests that many possible variants in this region could substitute adequately for wild-type P_RM.

We tabulated the changes both for the variants that could form stable lysogens and for those that could not (Table 3). Among the lysogenizing variants, no clear patterns emerged, except that the last position of the –10 region was usually a T. At all other positions, no base was present in more than half of the isolates. The most frequently occurring base matched the consensus at five positions out of the eight that we mutagenized and matched the wild-type P_RM sequence at four positions. Possibly the –35 region was able to vary more than the –10 region. Among the much smaller set of variants forming turbid plaques but not stable lysogens, again no clear patterns were evident.

P_RM variants with very strong promoters might not have been identified in our screen. In other work (J.W.L., unpublished data), we have made and studied a phage carrying the up promoter mutation prmup-1 (a change of the –35 region from TAGATA to TAGACA), which makes P_RM 10-fold stronger in the absence of CI and four- to fivefold stronger than the stimulated wild type in the presence of CI (31). prmup-1 has a low plating efficiency and forms very turbid plaques that are heterogeneous in size; perhaps singly infected cells often follow the lysogenic pathway or cannot complete the lytic pathway. Hence, isolates with promoters as strong as or stronger than the prmup-1 promoter might not plate efficiently and might not have been isolated in our screen.

Because P_RM is intrinsically a rather weak promoter, it is not surprising that many variants could provide a promoter with roughly equivalent strength. More surprising is that it is very difficult to predict from the sequences of the P_RM variants whether they will create a functional promoter in the context of the regulatory circuitry. The P_RM sequences from many of the phage that can lysogenize bear no recognizable resemblance to the consensus site (see reference 33).

We observed changes in response to CI in some of the P_RM variants (Fig. 3 and Table4). At least two related explanations are possible for this finding. First, expression of P_RM is stimulated by CI bound to O_R2. Activation involves a contact between residues in the DNA-binding domain of CI and the ⁷⁰ subunit of RNA polymerase (19, 21, 22, 35). This contact takes place close to the DNA. Subtle, sequence-dependent structural variations in different –35 regions might lead to local variations in the way ⁷⁰ interacts with the –35 region, influencing the strength of this interaction, as previously suggested (18). Second, it is known (13, 19) that CI stimulates the isomerization of the closed complex to the open complex (k_f), rather than the binding step (K_B). Since this transition involves several steps (2, 5), a mutation could alter the response to an activator that stimulates k_f by changing which step in open-complex formation is rate limiting.

Although several P_RM variants had properties similar to those of the wild type in laboratory experiments, the wild-type promoter sequence may be more fit in nature, since several different wild isolates of phages with lambda immunity have the same P_RM sequence as that of . These include HK97 (20) and 933H (36); in addition, we have sequenced the O_R regions of HK106, HK244, HK542, HK544, CL707, and CL715 (D. Wert and J. W. Little, unpublished data). Interpretation of these data is complicated by the likelihood that these phages are not reproductively isolated from each other, but they did contain a few other changes in and near the O_R region, suggesting that gene flow among them is not exceedingly rapid.

Do the changes in our mutants affect other aspects of the circuitry? Preliminary evidence obtained with P_R::lacZ operon fusions indicates that in the nine mutants analyzed in detail, the strength of P_R was reduced to about 70 to 90% of its wild-type value. This finding is consistent with previous evidence (44) suggesting that a weak UP element for P_R may overlap the –35 region of P_RM. P_R is slightly stronger when the region from –66 to –42 is present than in its absence (44), although this region does not have this effect in a different context. An AT-rich segment within this region matches fairly well with the upstream part of a consensus UP element (44), and a P_RM mutation, prm116, in the –35 region that improves this match is known to increase the strength of P_R (9). The region changed here in the –35 region of P_RM is included in that AT-rich segment, and the match is weakened in the nine variants analyzed.

It is unclear whether other aspects of regulation have also been subtly altered in some variants. If so, perhaps we have chosen phages for which the balance of forces resembles those in the wild type, and restoring balance in this way is relatively straightforward. Evaluation of this possibility must await more detailed analysis of functional interactions in this region. However, comparison between the set points for prophage induction and the promoter strengths provides some evidence for other subtle changes in O_R function, as we now discuss.

Prophage induction and promoter strength. Our evidence suggests a relationship between P_RM strength and the set point for prophage induction. Table 4 and Fig. 4 summarize the data from Fig. 2 and 3. For MD2 and MD19, the dashed lines in Fig. 4 indicate the uncertainty in estimating the set point (Table 4). Considering the entire set of mutants, there is a clear correlation between the value of the set point and the strength of stimulated P_RM.

View larger version (13K): [in this window] [in a new window]   FIG. 4. Correlation between set points and maximal promoter activities. Data from Table 4 were plotted to illustrate a correlation between the set point for UV induction and the maximal promoter strength observed in the presence (A) or absence (B) of the OL site located distal to lacZ. WT, wild type. For MD2 and MD19, dashed lines indicate uncertainty in estimating the set point.

This model implies that one could predict the set point from promoter strength; that is, some function of the promoter strength would give the set point and the set point would increase monotonically as promoter strength increased. It is not straightforward to determine what this function is, for several reasons. First, it is hard to predict the CI levels expected for each promoter, owing to negative feedback, conferred mostly by looping with O_L. Second, it is not clear how negative feedback and positive feedback change as CI levels drop. Third, the rate of RecA-mediated cleavage almost certainly varies with CI concentrations, both because only CI monomers are efficiently cleaved (4, 6, 11) and because the in vivo level of monomers is probably below the K_m for this reaction (11).

Whatever the function relating promoter strength and the set point may be, our data are not completely consistent with the expectation of a monotonic function. Considering the wild type and the group of mutants with set points between 10 and 20 J/m² (Fig. 4), there is little correlation between promoter strength and the set point. The most likely explanation for this disparity is that some of the changes in the mutants affect other aspects of O_R function (see above). We conclude that the strength of P_RM plays an important role in controlling the set point, particularly with large changes in P_RM strength, but that it is not the only determinant.

The model discussed above suggests that other forces contributing to the set point are the rate of CI cleavage and the CI level at which switching becomes likely. In turn, the latter value is likely controlled by many factors. Some of these, such as cooperative CI binding and looping or efficiency of P_RM mRNA translation, are unlikely to be affected by changes in P_RM. Others, such as the affinity of CI for its operators or the affinity of Cro for O_R3, might be modestly affected by changes adjacent to the operators. In addition, our mutations slightly reduce the strength of P_R, although it is uncertain that this would have marked effects on the set point.

The term "robustness" refers to a property that describes the relative insensitivity of a system to changes in certain parameters, such as promoter strength and the affinity of DNA-binding proteins for their targets. An alternative term is "parameter sensitivity" (15, 47, 48), which quantifies how much an output (such as the set point) varies with changes in an input (such as promoter strength). The mathematical treatment of this property has not been extended to bistable switches (M. A. Savageau, cited in reference 27), so that a quantitative treatment of our findings is not possible at present. In addition, it is uncertain that only one input parameter has changed markedly. Qualitatively, however, the set point is not highly resistant to changes in promoter strength; a twofold change in promoter strength usually results in about a twofold change in the set point.

Evolution of gene regulatory circuitry. We have previously suggested (27) that robustness has played an important role in the evolution of complex circuits in two ways. First, it allows an initial circuit to take a wide variety of forms, provided that the circuit offers a selective advantage (the lysogenic state in the case of ). Second, as this initial circuit is refined in the later stages of evolution, robustness likewise allows a wider range of pathways for refinement.

An important aspect of this refinement process is evolution of the cis-acting sites to which RNA polymerase and regulatory proteins bind. Evolution of cis-acting sites likely involves two types of changes: changing the relative arrangements and locations of these sites, including spacing between sites, and refining the exact sequences of these sites. The present findings bear primarily on the latter issue.

We find that the sequence of P_RM can adopt a wide variety of forms compatible with a functional regulatory circuit. One might use the term "sequence robustness" to describe this property, but the term "robustness" is increasingly being used in different ways, and this would further confuse its meaning. Instead, we offer the term "sequence tolerance" to describe this feature. Sequence tolerance is in a sense complementary to robustness, in that sequence tolerance allows different sequences to confer the same or similar parameter values on the system, rather than making the system tolerant of changes in parameter values as robustness does.

The functional consequences of sequence tolerance are also likely to be complementary to those of robustness, for both the establishment and refinement stages of the evolutionary pathway described above. Sequence tolerance further expands the range of acceptable initial solutions to the regulatory problem. Moreover, as these initial circuits are refined during a subsequent stage of evolution, it allows this refinement process to follow a wider range of pathways. Operating together, parameter robustness and sequence tolerance greatly expand the range of accessible pathways beyond those allowed by either feature alone.

A promoter like P_RM is a favorable case for sequence tolerance, since P_RM is a weak promoter and multiple sequences can likely provide comparable strength. In contrast, strong promoters like P_R and P_L are less likely to allow this feature. Use of a weak promoter imposes additional constraints on a bistable system, however. For such systems to operate, the forces stabilizing each epigenetic state must be in balance (e.g., see references 1, 10, and 55). If a system is not in balance, one state can take over even when the other has been established. For instance, two regulatory proteins might stabilize two alternative states; if the two proteins had equal affinities for DNA and equal dimer dissociation constants but were present in markedly unequal amounts, the action of the more abundant one would eventually lead to establishment of the state it stabilized. Accordingly, a protein whose expression is driven by a weak promoter might use other mechanisms, such as tighter DNA binding or dimerization, cooperativity, or increased stability, to compensate for its lower levels.

Our set of variants had a wide variety of phenotypes. This is also likely to be important for evolution, in two ways. First, we do not study (or its host) in its natural environment. This environment almost certainly affords a wide variety of selective pressures at different times and places. It is likely that different P_RM variants would be advantageous under different conditions. Environmental variability would offer a further expanded set of pathways by which the circuitry could evolve to its present state.

Second, lambdoid phages have a "mosaic" or modular organization, and it is thought that modules assort rather freely among lambdoid phages during evolution. Hence, the immunity region of is likely to have evolved at various times in different genome contexts that were subjected to differing selective pressures. For instance, recent studies with phages H-19B and 933W, which express toxins from the late lytic pR' promoter, show that these phages have a much lower set point for prophage induction (28), and exhibit far higher levels of spontaneous induction, than does . It was hypothesized that this low set point confers a selective advantage on the lysogen: If a small fraction of the bacterial population undergoes induction, the resulting toxin causes diarrhea in the vertebrate host, dispersing the surviving bacterial cells in the environment and promoting their spread. In contrast, we have argued that the set point of is optimized to allow prophage induction to occur efficiently only at doses of DNA damage likely to kill the host cell.

Shifting selective pressures might facilitate the evolution of P_RM sequences. If an immunity region with a P_RM with strength comparable to that of became associated with a toxin-producing late region, selective pressure for a weakerP_RM might result. In a later step, loss of the toxin genes would change the selective pressures, favoring a stronger P_RM, but possibly one with a different sequence. We suggest that variable contexts provide yet another way to expand greatly the number of possible pathways for evolution of cis-acting sites such as P_RM.

	ACKNOWLEDGMENTS

 
This work was supported by grant GM24178 from the National Institutes of Health.

We are grateful to Stanley Brown, Piero Bianco, Don Court, and Gary Gussin for helpful discussions; Tarlochan Dhillon, Ian Dodd, David Friedman, and Shota Atsumi for communicating unpublished results; Tarlochan Dhillon and Bob Weisberg for strains; Ian Dodd, Andrea Watson, and Naomi Franklin for plasmids; and Carol Dieckmann, Andrea Watson, and Gary Gussin for helpful comments on the manuscript. We thank Alex Blumenkron, Kantad Supamit, and Louise Lin for excellent technical assistance.

	FOOTNOTES

 
* Corresponding author. Mailing address: Department of Molecular and Cellular Biology, University of Arizona, Tucson, AZ 85721. Phone: (520) 621-5629. Fax: (520) 621-3709. E-mail: jlittle{at}email.arizona.edu.

Supplemental material for this article may be found at http://jb.asm.org/.

Present address: Department of Pediatrics, University of Colorado Health Sciences Center, Denver, CO 80262.

	REFERENCES

Top Abstract Introduction Materials and Methods Results Discussion References

 

1. Atsumi, S., and J. W. Little. 2004. Regulatory circuit design and evolution using phage . Genes Dev. 18:2086-2094.[Abstract/Free Full Text]
2. Buc, H., and W. R. McClure. 1985. Kinetics of open complex formation between Escherichia coli RNA polymerase and the lac UV5 promoter. Evidence for a sequential mechanism involving three steps. Biochemistry 24:2712-2723.[CrossRef][Medline]
3. Churchward, G., D. Belin, and Y. Nagamine. 1984. A pSC101-derived plasmid which shows no sequence homology to other commonly used cloning vectors. Gene 31:165-171.[CrossRef][Medline]
4. Cohen, S., B. J. Knoll, J. W. Little, and D. W. Mount. 1981. Preferential cleavage of phage lambda repressor monomers by recA protease. Nature 294:182-184.[CrossRef][Medline]
5. Craig, M. L., O. V. Tsodikov, K. L. McQuade, P. E. Schlax, Jr., M. W. Capp, R. M. Saecker, and M. T. Record, Jr. 1998. DNA footprints of the two kinetically significant intermediates in formation of an RNA polymerase-promoter open complex: evidence that interactions with start site and downstream DNA induce sequential conformational changes in polymerase and DNA. J. Mol. Biol. 283:741-756.[CrossRef][Medline]
6. Crowl, R. M., R. P. Boyce, and H. Echols.1981 . Repressor cleavage as a prophage induction mechanism: hypersensitivity of a mutant cI protein to RecA-mediated proteolysis. J. Mol. Biol. 152:815-820.[CrossRef][Medline]
7. Dodd, I. B., A. J. Perkins, D. Tsemitsidis, and J. B. Egan. 2001. Octamerization of lambda CI repressor is needed for effective repression of P_RM and efficient switching from lysogeny.Genes Dev. 15:3013-3022.[Abstract/Free Full Text]
8. Dodd, I. B., K. E. Shearwin, A. J. Perkins, T. Burr, A. Hochschild, and J. B. Egan. 2004. Cooperativity in long-range gene regulation by the CI repressor. Genes Dev. 18:344-354.[Abstract/Free Full Text]
9. Fong, R. S. C., S. Woody, and G. N. Gussin.1994 . Direct and indirect effects of mutations in lambda P_RM on open complex formation at the divergent P_R promoter. J. Mol. Biol. 240:119-126.[CrossRef][Medline]
10. Gardner, T. S., C. R. Cantor, and J. J. Collins. 2000. Construction of a genetic toggle switch in Escherichia coli. Nature 403:339-342.[CrossRef][Medline]
11. Gimble, F. S., and R. T. Sauer. 1989. Lambda repressor mutants that are better substrates for RecA-mediated cleavage. J. Mol. Biol. 206:29-39.[CrossRef][Medline]
12. Gussin, G. N., S. Brown, J. Ferm, and K. Matz. 1987. New mutations in the P_RM promoter of bacteriophage lambda. Gene 54:291-297.[CrossRef][Medline]
13. Hawley, D. K., and W. R. McClure. 1982. Mechanism of activation of transcription initiation from the lambda P_RM promoter. J. Mol. Biol. 157:493-525.[CrossRef][Medline]
14. Herskowitz, I., and D. Hagen. 1980. The lysis-lysogeny decision of phage : explicit programming and responsiveness. Annu. Rev. Genet. 14:399-445.[CrossRef][Medline]
15. Hlavacek, W. S., and M. A. Savageau. 1995. Subunit structure of regulator proteins influences the design of gene circuitry: analysis of perfectly coupled and completely uncoupled circuits. J. Mol. Biol. 248:739-755.[CrossRef][Medline]
16. Hopkins, N., and M. Ptashne. 1971. Genetics of virulence, p.571 -574. In A. D. Hershey (ed.), The bacteriophage lambda. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.
17. Huynh, T. V., R. A. Young, and R. W. Davis.1985 . Constructing and screening cDNA libraries using gt10 and gt11, p.49 -80. In D. Glover (ed.), DNA cloning: a practical approach, vol. I. IRL Press, Inc., Boca Raton, Fla.
18. Hwang, J. J., and G. N. Gussin. 1988. Interactions between Escherichia coli RNA polymerase and lambda repressor. Mutations in P_RM affect repression of P_R. J. Mol. Biol. 200:735-739.[CrossRef][Medline]
19. Jain, D., B. E. Nickels, A. Hochschild, and S. A. Darst. 2004. Structure of a ternary transcription activation complex. Mol. Cell 13:45-53.[CrossRef][Medline]
20. Juhala, R. J., M. E. Ford, R. L. Duda, A. Youlton, G. F. Hatfull, and R. W. Hendrix.2000 . Genomic sequences of bacteriophages HK97 and HK022: pervasive genetic mosaicism in the lambdoid bacteriophages. J. Mol. Biol. 299:27-51.[CrossRef][Medline]
21. Kuldell, N., and A. Hochschild. 1994. Amino acid substitutions in the –35 recognition motif of ⁷⁰ that result in defects in phage lambda repressor-stimulated transcription.J. Bacteriol. 176:2991-2998.[Abstract/Free Full Text]
22. Li, M., H. Moyle, and M. M. Susskind. 1994. Target of the transcriptional activation function of phage lambda cI protein. Science 263:75-77.[Abstract/Free Full Text]
23. Little, J. W. 1984. Autodigestion of lexA and phage lambda repressors. Proc. Natl. Acad. Sci. USA 81:1375-1379.[Abstract/Free Full Text]
24. Little, J. W. 1993. LexA cleavage and other self-processing reactions. J. Bacteriol. 175:4943-4950.[Free Full Text]
25. Little, J. W. 1983. The SOS regulatory system: control of its state by the level of RecA protease. J. Mol. Biol. 167:791-808.[CrossRef][Medline]
26. Little, J. W., and D. W. Mount. 1982. The SOS regulatory system of Escherichia coli. Cell 29:11-22.[CrossRef][Medline]
27. Little, J. W., D. P. Shepley, and D. W. Wert.1999 . Robustness of a gene regulatory circuit. EMBO J. 18:4299-4307.[CrossRef][Medline]
28. Livny, J., and D. I. Friedman. 2004. Characterizing spontaneous induction of Stx encoding phages using a selectable reporter system. Mol. Microbiol. 51:1691-1704.[CrossRef][Medline]
29. Mao, C., and J. W. Little. 1998. Mutations affecting cooperative DNA binding of phage HK022 CI repressor.J. Mol. Biol. 279:31-48.[CrossRef][Medline]
30. Maurer, R., B. J. Meyer, and M. Ptashne. 1980. Gene regulation at the right operator (O_R) of bacteriophage . I. O_R3 and autogenous negative control by repressor. J. Mol. Biol. 139:147-161.[CrossRef][Medline]
31. Meyer, B. J., R. Maurer, and M. Ptashne. 1980. Gene regulation at the right operator (O_R) of bacteriophage . II. O_R1, O_R2, and O_R3: their roles in mediating the effects of repressor and cro. J. Mol. Biol. 139:163-194.[CrossRef][Medline]
32. Miller, J. H. 1972. Experiments in molecular genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.
33. Mulligan, M. E., and W. R. McClure. 1986. Analysis of the occurrence of promoter-sites in DNA. Nucleic Acids Res. 14:109-126.[Abstract/Free Full Text]
34. Mustard, J. A., and J. W. Little. 2000. Analysis of Escherichia coli RecA interactions with LexA, CI, and UmuD by site-directed mutagenesis of recA. J. Bacteriol. 182:1659-1670.[Abstract/Free Full Text]
35. Nickels, B. E., S. L. Dove, K. S. Murakami, S. A. Darst, and A. Hochschild. 2002. Protein-protein and protein-DNA interactions of ⁷⁰ region 4 involved in transcription activation by cI. J. Mol. Biol. 324:17-34.[CrossRef][Medline]
36. Perna, N. T., G. Plunkett III, V. Burland, B. Mau, J. D. Glasner, D. J. Rose, G. F. Mayhew, P. S. Evans, J. Gregor, H. A. Kirkpatrick, G. Posfai, J. Hackett, S. Klink, A. Boutin, Y. Shao, L. Miller, E. J. Grotbeck, N. W. Davis, A. Lim, E. T. Dimalanta, K. D. Potamousis, J. Apodaca, T. S. Anantharaman, J. Lin, G. Yen, D. C. Schwartz, R. A. Welch, and F. R. Blattner. 2001. Genome sequence of enterohaemorrhagic Escherichia coli O157:H7.Nature 409:529-533.[CrossRef][Medline]
37. Powell, B. S., D. L. Court, Y. Nakamura, M. P. Rivas, and C. L. Turnbough, Jr. 1994. Rapid confirmation of single copy lambda prophage integration by PCR. Nucleic Acids Res. 22:5765-5766.[Free Full Text]
38. Ptashne, M. 1992. A genetic switch: phage and higher organisms. Cell Press and Blackwell Scientific Publications, Cambridge, Mass.
39. Ptashne, M. 2004. A genetic switch: phage lambda revisited. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.
40. Revet, B., B. Von Wilcken-Bergmann, H. Bessert, A. Barker, and B. Müller-Hill. 1999. Four dimers of lambda repressor bound to two suitably spaced pairs of lambda operators form octamers and DNA loops over large distances. Curr. Biol. 9:151-154.[CrossRef][Medline]
41. Roberts, J. W., and R. Devoret. 1983. Lysogenic induction, p. 123-144. In R. W. Hendrix, J. W. Roberts, F. W. Stahl, and R. A. Weisberg (ed.), Lambda II. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.
42. Rosen, E. D., and G. N. Gussin. 1979. Clustering of Prm mutations of bacteriophage lambda in the region between 33 and 40 nucleotides from the cI transcription start point.Virology 98:393-410.[CrossRef][Medline]
43. Rosen, E. D., J. L. Hartley, K. Matz, B. P. Nichols, K. M. Young, J. E. Donelson, and G. N. Gussin. 1980. DNA sequence analysis of prm mutations of coliphage lambda. Gene 11:197-205.[CrossRef][Medline]
44. Ross, W., S. E. Aiyar, J. Salomon, and R. L. Gourse.1998 . Escherichia coli promoters with UP elements of different strengths: modular structure of bacterial promoters. J. Bacteriol. 180:5375-5383.[Abstract/Free Full Text]
45. Sauer, R. T. 1978. DNA sequence of the bacteriophage cI gene. Nature 276:301-302.[CrossRef][Medline]
46. Sauer, R. T., and R. Anderegg. 1978. Primary structure of the lambda repressor. Biochemistry 17:1092-1100.[CrossRef][Medline]
47. Savageau, M. A. 1974. Comparison of classical and autogenous systems of regulation in inducible operons.Nature 252:546-549.[CrossRef][Medline]
48. Savageau, M. A. 1971. Parameter sensitivity as a criterion for evaluating and comparing the performance of biochemical systems. Nature 229:542-544.[CrossRef][Medline]
49. Shean, C. S., and M. E. Gottesman. 1992. Translation of the prophage lambda cI transcript.Cell 70:513-522.[CrossRef][Medline]
50. Simons, R.