Molecular basis of pyelonephritis-associated pili phase variation in Escherichia coli

The expression of pyelonephritis-associated pili (Pap) in Escherichia coli is under a phase variation control mechanism in which individual cells alternate between pili+ (ON) and pili- (OFF) states. This occurs through a process involving DNA methylation by deoxyadenosine methylase (Dam). Methylation of two GATC sites (GATC-I and GATC-II) within the pap regulatory region is differentially inhibited in phase ON and phase OFF cells. The GATC-I site of phase ON cells is nonmethylated and GATC-II site is fully methylated. Conversely, in phase OFF cells the GATC-I site is fully methylated whereas the GATC-II site is nonmethylated. Two transcription activators, Lrp and PapI, are required for this specific methylation inhibition. Low resolution DNA footprint analyses using nonmethylated pap DNA indicated that Lrp binds near the GATC-II, whereas PapI does not bind specifically to pap regulatory region. However, the addition of Lrp and PapI together resulted in an additional footprint around the GATC-I site, indicating that both Lrp and PapI are required for binding to the GATC-I region. To define the role of Dam methylation in pap gene regulation, the GATC-I and GATC-II sites were mutated so that they could not be methylated, and the effects of these mutations on Pap phase variation were examined. The results indicated that methylation of GATC-I blocks formation of the phase ON state by inhibiting PapI-dependent Lrp binding to this DNA region. In contrast, methylation of GATC-II is required for the phase OFF to ON transition. Evidence suggests that this occurs by the inhibition of Lrp to sites overlapping the papBA promoter, which may occlude RNA polymerase. The Lrp binding sites in the pap regulatory region were further defined by methylation protection analysis. Six Lrp binding sites were found, each separated by about three helical turns of DNA. Lrp bound with highest affinity to three sites (1, 2, and 3) proximal to the papBA promoter. A mutational analysis indicated that the binding of Lrp to sites 2 and 3 inhibits pap transcription, which is consistent with the fact that Lrp binding site 3 is located between the -35 and -10 RNA polymerase binding region of papBA promoter. The addition of PapI decreased the affinity of Lrp for sites 1, 2, and 3 and increased its affinity for the distal Lrp binding sites 4 and 5. Mutations within Lrp binding sites 4 and 5 shut off pap transcription, indicating that the binding of Lrp to this pap region activated transcription. The pap GATC-I and GATC-II sites are located within Lrp binding sites 5 and 2, respectively, providing a mechanism by which Dam controls Lrp binding and Pap phase variation. A model for Pap phase variation is presented based on these results..

THE MOLECULAR BASIS OF PYELONEPHRITIS-ASSOCIATED PILI PHASE VARIATION IN ESCHERICHIA COLI by Xiangwu Nou A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Experimental Pathology Department of Pathology The University of Utah June 1996 Copyright © Xiangwu Nou 1996 All Rights Reserved THE UNIVERSITY OF UTAH GRADUATE SCHOOL SUPERVISORY COMMITTEE APPROVAL of a dissertation submitted by XiangwuNou This dissertation has been read by each member of the following supervisory committee and by majority vote has been found to be satisfactory. Chair: David A. Low John R. Roth Lf/.)· (1/'- r- If'-Ii ! .I B~bara JI raves !/ l J //,/;'//L I i i ... · ! ) ~ ~'(/i -:-- /~'olin H. Wels / , THE UNIVERSITY OF UTAH GRADUATE SCHOOL FINAL READING APPROVAL To the Graduate Council of the University of Utah: I have read the dissertation of Xiangwu Nou in its final form and have found that (1) its format. citations. and bibliographic style are consistent and acceptable; (2) its illustrative materials including figures, tables, and charts are in place; and (3) the final manuscript is satisfactory to the supervisory committee and is ready for submission to The Graduate School. / I t'; (2. ""/,:;:{. Date David A. Low Chair, Supervisory Committee ;~ Walter Stevens Chair!Dean Approved for the Graduate Council Ann W. Hart Dean of TIle Graduate School ABSTRACT The expression of pyelonephritis-associated pili (Pap) in Escherichia coli is under a phase variation control mechanism in which individual cells alternate between piJi+ (ON) and pili- (OFF) states. This occurs through a process involving DNA methylation by deoxyadenosine methylase (Dam). Methylation of two GATC sites (GATC-I and GATC-II) within the pap regulatory region is differentially inhibited in phase ON and phase OFF cells. The GATC-I site of phase ON cells is nonmethylated and GATC-II site is fully methylated. Conversely, in phase OFF cells the GATC-I site is fully methylated whereas the GATC-II site is nonmethylated. Two transcription activators, Lrp and PapI, are required for this specific methylation inhibition. Low resolution DNA footprint analyses using nonmethylated pap DNA indicated that Lrp binds near the GATC-II, whereas PapI does not bind specifically to pap regulatory region. However, the addition of Lrp and PapI together resulted in an additional footprint around the GA TC-I site, indicating that both Lrp and PapI are required for binding to the GATC-I region. To define the role of Dam methylation in pap gene regulation, the GATC-I and GATC-II sites were mutated so that they could not be methylated, and the effects of these mutations on Pap phase variation were examined. The results indicated that methylation of GATC-I blocks formation of the phase ON state by inhibiting PapI-dependent Lrp binding to this DNA region. In contrast, methylation of GATC-II is required for the phase OFF to ON transition. Evidence suggests that this occurs by the inhibition of Lrp to sites overlapping the papBA promoter, which may occlude RNA polymerase. The Lrp binding sites in the pap regulatory region were further defined by methylation protection analysis. Six Lrp binding sites were found, each separated by about three helical turns of DNA. Lrp bound with highest affinity to three sites (1, 2, and 3) proximal to the papBA promoter. A mutational analysis indicated that the binding of Lrp to sites 2 and 3 inhibits pap transcription, which is consistent with the fact that Lrp binding site 3 is located between the -35 and -10 RNA polymerase binding region of papBA promoter. The addition of PapI decreased the affinity of Lrp for sites 1, 2, and 3 and increased its affinity for the distal Lrp binding sites 4 and 5. Mutations within Lrp binding sites 4 and 5 shut off pap transcription, indicating that the binding of Lrp to this pap region activated transcription. The pap GATC-I and GATC-II sites are located within Lrp binding sites 5 and 2, respectively, providing a mechanism by which Dam controls Lrp binding and Pap phase variation. A model for Pap phase variation is presented based on these results. v To Baihua, Aurora, and Angelina TABLE OF CONTENTS ABSTRACT...... ... ......... ........................... ........... ...... ..... .... ..... .... ..... iv LIST OF TABLES.............................................................................. ix LIST OF FIGURES............................................................................ x ACKNOWLEDGMENTS.................................................................... xii Chapter I. INTRODUCTION........................................................................ 1 Overview.................................................................................. 2 The Pap Pili............................................................................... 3 DNA Methylation ......................................................................... 13 Lrp and Gene Regulation........... ............... ..... ...................... ........... 18 Pap Pili Phase Variation ................................................................. 26 Introduction to Work in This Thesis................................................... 34 References ................................................................................. 35 II. REGULATION OF PYELONEPHRITIS-ASSOCIATED PILI PHASE­VARIATION IN ESCHERICHIA COll: BINDING OF THE PAPI AND THE LRP REGULATORY PROTEINS IS CONTROLLED BY DNA METHyLATION ............................................................. 48 Summary .................................................................................. 49 Introduction ............................................................................... 49 Results ..................................................................................... 50 Discussion ................................................................................. 53 Experimental Procedures ................................................................ 55 Acknowledgments......................... .............................................. 56 References ................................................................................. 56 III. METHYLATION PATTERNS IN PAP REGULATORY DNA CONTROL PYELONEPHRITIS-ASSOCIATED PILI PHASE VARIATION IN E. COLI ... .................................................. 58 Summary .................................................................................. 59 Introduction ............................................................................... 59 Results ..................................................................................... 61 Discussion ................................................................................. 66 Experimental Procedures................................................................ 68 Acknowledgments......................... .............................................. 69 References..... ............................................. ...... ................... ..... 69 IV. DIFFERENTIAL BINDING OF LRP TO TWO SETS OF PAP DNA BINDING SITES MEDIATED BY PAPI REGULATES PAP PHASE VARIATION IN ESCHERICHIA COLI .................................... 71 Introduction............................................................................... 72 Results ..................................................................................... 73 Discussion ................................................................................. 76 Materials and Methods................................................................... 83 Acknowledgments....................................................................... 83 References ................................................................................. 83 V. SUMMARY AND DISCUSSION ..................................................... 85 The Role of Lrp in Pap Pili Phase Variation........................................... 86 The Role of PapI in Pap Pili Phase Variation........................................ 102 The Role of DNA Methylation in Pap Pili Phase Variation ......................... 107 The Roles of Other Regulatory Factors in Pap Pili Phase Variation...... ......... 110 A Working Model for Pap Pili Phase Variation ...................................... 112 Implications for Other Pili Phase Variation Systems................................ 119 Concluding Remarks .................................................................... 121 References................................................................................ 122 viii LIST OF TABLES Table Page 1-1 Lrp regulon................................................................................ 21 2-1 The pap! and lrp genes are required for transcription initiated at the papBA pilin promoter ................................................................... 50 3-1 E. coli strains, plasmids, and bacteriophage used in this study ..................... 60 3-2 The effects of Lrp and Darn on pap transcription in E. coli strains containing wild-type or mutant pap regulatory DNAs............................... 62 3-3 The effects of PapI overproduction on methylation protection of the GATC1028 site and pap transcription ................................................... 65 4-1 Phenotypes of pap regulatory mutants.. .................. ..... ................. ....... 80 4-2 Effects of Lrp and PapJ on the phenotypes of pap regulatory mutants............. 80 4-3 Effects of Darn methylase on the phenotypes of pap regulatory mutants.......... 80 4-4 Escherichia coli strains, plasrnids and bacteriophage used in this study.. ..... .... 82 LIST OF FIGURES Figure Page 1-1 Organization of the pap genes........................................................... 7 1-2 Mechanisms of pili phase variation in bacteria................ ....................... 29 2-1 Map of plasmid pDAL336.... .... ..... ..... ..... ....... ....... ...... ...... ... .......... 50 2-2 Gel retardation analysis................. ................................................ 51 2-3 DN ase I footprint analysis of fully methylated and non-methylated pap DNAs ................................................................................. 52 2-4 Summary of DNase I footprint results using the pap regulatory region ............. 53 2-5 DNase I footprint analysis of hemimethylated pap DNAs ........................... 54 3-1 The pap regulatory region ............................................................... 61 3-2 Analysis of LacZ phenotypes ofE. coli strain MC4100 containing single-copy wild-type and regulatory mutant pap'lacZ fusions ..................... 61 3-3 Analysis of the GATC methylation states of pap DNA sequences isolated from wild-type and regulatory mutant E. coli strains......... .............. 63 3-4 DNase I footprint analysis of wild-type and mutant pap DNAs ..................... 64 3-5 Pap phase variation model............................................................... 67 4-1 The localization of Lrp binding sites in the pap regulatory region by methylation protection analysis ......................................................... 74 4-2 The location of Lrp binding sites in the pap regulatory region...................... 75 4-3 A quantitative DNase I footprint analysis of nonmethylated wild-type and Lrp binding site mutant pap DNAs ................................................ 77 4-4 Summary of the quantitative footprint analysis of Lrp binding to wild-type and mutant pap DNAs ........................................................ 79 4-5 Pap phase variation modeL.............................................................. 81 5-1 Comparison of Lrp footprints in pap, ilvIH, lysU and ompC regulatory regions ..................................................................................... 92 5-2 Proposed DNA conformations in phase ON and phase OFF cells................. 113 5-3 Model of Pap pili phase variation............................. ............ ... ........... 116 xi ACKNOWLEDGMENTS Fonnost, I am deeply indebted to my thesis advisor Dr. David Low. As a respected teacher, David patiently guided me throughout my work of this thesis with insights and enthusiasm. His guidance and instructions have been key to work presented in this thesis. Past and current members of the Low laboratory: Bruce Braaten, Larry Blyn, Chris White-Ziegler, Majan van der Wood, Linda Kaltenbach, Brad Nicholen, Nate Wyand, and Brad Hale, all deserve my profond thanks for providing helps in various experiments, in English language, and in preparing this thesis, as well as for provinding encouragements and friendship and for sharing their insights and wisdom. I especially want to thank Dr. Chris Hill and Felix Vajdos for providing me with purified Lrp. I would like to thank my supervisory committee members for their support and suggestions. I especially wish to thank Dr. John Roth for bringing my attention to the Molecular Biology Program at The University of Utah I wish to thank Blackwell Science Ltd., Cell Press, and Oxford University Press for granting me permission to use copyrighted materials, which constitute Chapters II, III, and IV, of this thesis. CHAPTER I INTRODUCTION 2 Overview Urinary tract infection (UTI) is a common health concern, especially among women. About one tenth of men and one fifth of women experience episodes of UTI during their lifetimes (74, 82, 96). UTI can be caused by a variety of pathogens, including bacteria, yeast and viruses. However, the most common etiologic agents of UTI are uropathogenic Escherichia coli strains (72, 130, 144). It is estimated that 60-85% of human UTI are caused by E. coli strains. The source of UTI-causing strains is believed to be the opportunistic pathogens in the fecal flora (46, 141). Those bacterial strains are capable of, upon contamination, colonizing and invading the lower urinary tract, causing cystitis, and further ascending to the upper urinary tract, causing pyelonephritis. In some cases, they invade the deep tissue, causing urosepsis (72, 124, 130, 139). Uropathogenic and nonpathogenic E. coli strains differ in their abilities to colonize and to invade host tissues. Bacterial cells from a pathogenic strain can adhere to host epithelial cells by way of specific types of pili (also known as fimbriae )(130, 139). Such pili are usually a complex surface structure composed of specific pili subunits and adhesin molecules with different specificity to the components of host cell surface (72, 84). Most E. coli clinical isolates from pyelonephritis patients express a distinct type of pilus that enable the bacterial cells to attach to the epithelial cells lining the urinary tract (72). These pili are therefore referred to as pyelonephritis-associated pili (Pap) (68, 118). Production of pili confers to the bacterial cell the advantage to adhere tightly to host epithelial cell surfaces, a necessary step for colonization in environments like the urinary tract, where bacterial cells not attaching to the host cells are readily removed (72). Nevertheless, the presence of pili can sometime also be disadvantageous for a bacterial cell. First, the tight adhesion to epithelial cells mediated by pili reduces the mobility of the bacterial cells (pili are not structures of locomotion), which restricts the bacterial cells from moving to new locations. Second, the proteinacious pili may be an easy target for the host 3 immunological system, rendering piliated bacterial cells more vulnerable to eradication by the host (132). Third, the production of the pili also consumes a large quantity of energy. Uropathogenic E. coli have developed a mechanism that allows them to maintain the advantages of pili production and avoid its disadvantages by altering the states of pili expression. This phenomenon is known as phase variation, by which bacterial cells alternate between phase ON (active expression) and phase OFF (no expression) states. In some other systems, such as Neisseria gonorrhoeae pili expression, the bacterial cells produce different types of pili with different antigenic characteristics, a phenomenon known as antigenic variation. Antigenic variation as well as phase variation occurs by different mechanisms in different cases, though in most cases reversible rearrangement of DNA is often involved (132). Understanding the mechanisms underlying different types of antigenic and phase variation is therefore important for understanding bacterial pathogenesis. The Pap Pili Characterization of Pap Pili A uropathogenic E. coli strain can produce different types of pili, including Type 1 pili, Pap pili, S pili, and others (28, 72,81). A given type of pilus is characterized by its morphology and adhesion properties but can exist as different serotypes. Pap pili produced by uropathogenic E. coli strains are hairlike proteinacious surface structures composed of a rigid shaft and a flexible tip (84). The shaft is composed of about one thousand major pilin monomers, which are arranged as a tightly packed right-handed helix that is 7 nm wide and 0.2-2 J..Lm long with a 2.0-2.5 nm central helical hole (84, 85). The major pilin subunits account for 99.9% of the pilus mass and determine the antigenic specificity of the pilus (58, 104). The tip is composed of a variety of minor pilin subunits, which form a thin open helical structure (2 nm wide) termed the fimbrillum (84, 85). The tip determine the adhering specificity of the pilus (93, 103). A bacterial cell may produce hundreds to thousands of Pap pili on its surface. 4 Unlike the more common Type 1 pili, which bind to mannose residues of epithelial cells and therefore the binding to cells is inhibited by the presence of mannose (134), Pap pili can agglutinate human or guinea pig erythrocytes in the presence of mannose (145). This mannose-resistant hemagglutination is due to the presence of an adhesin in the pili with a different receptor specificity. E. coli strains expressing Pap pili, as well as purified pili from those strains, bind to erythrocytes expressing P blood group antigens (75, 76, 80), which are a family of globoseries glycolipid containing Gal«(xJ-4)Gal~ moiety (Gal- Gal), including globotriaosylceramise (Gb03) and globoside (Gb04). Several lines of experiments confirmed that Pap pili adhesin recognized Gal-Gal disaccharide moiety of glycolipids on the cell surface (72). First, E. coli strains expressing Pap pili were shown to bind to cells carrying such glycolipids naturally or to cells or latex beads treated with such glycolipids (31, 89, 98, 142). Second, adherence of strains expressing Pap pili to cells carrying Gal-Gal containing glycolipids or treated with Gal-Gal containing glycolipids was blocked by substances containing the Gal-Gal moiety (75, 89, 142). Third, strains expressing Pap pili were shown to adhere to glycolipids containing Gal-Gal but not to those without Gal-Gal (17). In humans the globoseries glycolipids Gb03 and Gb04 are found predominantly within the kidney, the ureter, the red blood cells, and less abundantly within the gastrointestinal tract (24,84). This provides an explanation for why Pap piliated E. coli strains colonize and invade the human urinary tract, causing UTI, whereas other tissues are relatively not infected. Recently Pap pili were found also to bind to immobilized fibronectin (157), providing a mechanism for Pap piliated bacterial cells to interact with the extracellular matrix (59). Pap and UTI A number of epidemiological studies have established the importance of the expression of Pap pili in bacteria caused urinary tract infections (72, 139). Marild et al. (106) showed that about 90% of children with first-time acute pyelonephritis carried E. coli expressing Pap pili. Sanderberg et al. (135) showed that the majority of adult women of 5 various ages with acute pyelonephritis carried Pap-piliated E. coli as causing agents. In the case of uncomplicated urosepsis, Johnson et at. (73) showed that up to 100% of the blood isolates expressed Pap pili. Overall, about 70% of E. coli isolates from pyelonephritis patients and 36% from cystitis patients expressed Pap pili, whereas only about 24% from asymptomatic bacteriuria patients and 19% from fecal flora of healthy individuals expressed Pap pili (72). However, in individual patients compromised by pregnancy, anatomic abnormalities, or instrumentation, the frequency of clinical isolates expressing Pap pili was substantially lower comparing with individual patients without such complicating factors (32, 72, 73, 99). These studies indicate that although nonpiliated strains are capable of infecting the lower urinary tract and upper urinary tract of the compromised individuals, Pap piliated strains are more likely to reach to the upper urinary tract of healthy individuals and cause pyelonephritis. The importance of Pap pili in uropathogenesis was also demonstrated by using mice and monkeys as animal models (72). In both cases, inoculation with strains expressing Pap pili led to colonization of the urinary tract by the bacteria and to urinary tract infections (37, 54, ll8, 131). Administration of substances containing Gal-Gal along with the bacteria inoculum protected the animals from urinary tract colonization by the bacteria (140). Immunization with purified Pap pili proteins also protected the animals from urinary tract infection following subsequent challenges with the E. coli strains expressing homologous Pap pili (55, 118, 131). Urinary tract infections by Pap piliated E. coli cause local and systematic inflammatory responses which are responsible for the symptoms. Strains expressing Pap pili, either living or killed, elicit inflammatory responses when inoculated into mice (95). This inflammatory response is blocked by coadministration of a Gal-Gal containing substance (95). This observation suggests that the role of Pap pili in uropathogenesis is not restricted to allow the bacterial cells to adhere to the uroepithelial cells. It has been found that Pap piliated E. coli cells stimulate epithelial cells to release interleukin-6 and 6 interleukin-8 in vivo and in vitro (2, 3, 62, 63). The cytokines, in tum, activate the host inflammatory responses (139). The pap Operon The genes encoding the Pap pili are located on the chromosome as a cluster in the uropathogenic E. coli strains. It is believed that the uropathogenic E. coli strains have gained virulence determinant genes as a consequence of mobile element transposition and that different types of pili among a variety of enteric bacteria may have common origins (86). Eleven pap genes in the cluster have been identified and the function of the products studied (144). As shown in Figure 1-1, the pap genes are transcribed divergently, with papJ being transcribed leftward as a monocistronic unit and the other 10 pap genes rightward as a polycistronic unit. The products of papJ and papB are positive regulatory proteins for the transcription of the pap operons (43). The papA gene encodes for the major pilin subunit of the Pap pili (9). PapA is the most heterogeneous gene product among different Pap pili producing isolates and determines the antigenic specificity of the Pap pili. However, PapA is not necessary for Gal-Gal adhesion. Mutants in papA gene produce cell adhesin which mediates Gal-Gal adhesion in rough strains but not in smooth strains (93, 150) suggesting that the pili shaft is important to extend the adhesin away from cell surface when the cell is coated with polysaccharides. The genes papK, papE, papF, papG, and papH code for the minor components of Pap pili. PapG is the adhesin, present on the very tip of a pilus and determining the specific adhesion to the Gal-Gal moiety of glycolipids on the surfaces of epithelial cells (85). Mutants in the papG gene are unable to mediate mannose-resistant hemagglutination. The assembly of Pap pili starts with the assembly of the tip, and the subunits are sequentially added at the base so that the pilus "grows" from the base (69). PapF is a linker between the adhesin and the tip in the pilus. Its incorporation initiates the assembly of the pilus (71). 7 Figure 1-1. Organization of the pap genes. Eleven pap genes are organized as two divergently transcribed operons. The function of each gene product is indicated. The papI­papB intergenic region is the pap regulatory region. Regulator Regulator Major Pilin Anchor Assembly Periplasmic chaperone ? Outer Membrane Usher 8 Flexibility Attachment Major Tip Component Adhesin Adaptor Initiator Adaptor Initiator ~ Transcriptional regulatory protein ~ Pilin protein ~ Protein required for pili assembly 9 PapE subunits are then added to PapF. PapE is the most abundant minor pilin subunit and the only one known to form homopolymers. The presence of a string of PapE is believed to confer flexibility to the tip of the pilus (85). PapE also mediates the fibronectin-dependent adhesion (158). The incorporation of PapK into the pilus tip then initiates the assembly of PapA polymer (71), forming the long rigid shaft of the pilus structure. PapH serves both as a stopper of PapA polymerization and a molecular anchor for the pilus to the membrane. Mutants in papH produce nonanchored long pili and PapH overproducers produce shorter pili.(8). The products of papC and papD are not present in the pilus but are required for Pap pili assembly (84). PapD is a periplasmic molecular chaperone that binds to the major and minor pilin subunits to prevent premature folding (94). Interaction with PapD also prevent the major and minor pilin subunits from being degraded by the periplasmic DegP proteinase (84). PapC, called outer membrane usher, is a specific outer membrane protein that determines the site and the order of pilus assembly (36). It may also serve as a deck and anchor for the assembly of Pap pili. The function of PapJ has not been well defined, though it has been suggested to facilitate the assembly of the PapA subunits (84). Genetic Regulation of Pap Expression The expression of Pap pili genes is affected by a variety of environmental factors, including carbon source, nitrogen source, temperature, and phase of growth (46). Generally, growth on rich medium inhibits the expression of pili genes and growth on minimal medium stimulates the expression. Similarly, growth at body temperature (37°C) induces the production of Pap pili whereas growth at room temperature (23°C) shuts off the pili production. It has been proposed that bacterial cells sense such environmental signals to determine whether they are inside a host (160). The genetic basis underlying such responses to environmental signals is complex. Many proteins, both operon specific factors encoded by the pap genes and global regulatory factors, have been shown to be involved in this process. The regulation of Pap pili expression occurs at both transcriptional and 10 posttranscriptionallevels (102). In addition, the regulation of pap transcription involves a phase variation mechanism. Transcriptional regulation Both PapB and PapI are required for the transcription of the papBA operon (6). PapB is also autoregulated. Overexpression of PapB in trans leads to the down regulation of the transcription from the papBA operon (43). PapB has been shown to specifically bind to two sites in the intercistronic regulatory region as well as a site in the coding region of papB. Of those three binding sites, the most upstream one, centered at about -240 of the papBA transcription initiation point and about -90 of that of pap/, has the highest affinity. A second binding site overlaps with the -10 region of the papBA promoter and a third one in the papB coding sequence at +100. Assuming the binding of PapB to the lower affinity sites at -10 and +100 interferes with the binding of RNA polymerase, this may explain the autoregulation of PapB. PapI does not bind to DNA, and recent evidence from our laboratory suggests it functions by affecting the DNA binding specificity of a major regulatory protein, Lrp (see further discussion later). The global regulator CAP is also required for transcription and has been shown to bind in a cAMP-dependent manner to the regulatory region centered at 215.5 bp upstream of the papBA transcription initiation point and 115.5 bp upstream of the pap/ transcription initiation point, adjacent to the PapB high affinity binding site (52). Mutations introduced into this CAP binding site shut off the transcription from both operons. The involvement of CAP in transcription activation has been well studied. Depending on the properties of the promoter and the location of the binding site relative to the RNA polymerase binding site, CAP either directly interacts with RNA polymerase or induces DNA conformational change by bending DNA (25). The facts that the CAP binding site is far away from that of the RNA polymerase and that other regulatory factors are involved in the transcriptional activation suggest that papBA promoter is a class III CAP-activated promoter, in which direct CAP-RNA polymerase interaction has not been directly observed (25, 128). It has 1 1 been proposed that CAP, by interacting with PapB upon binding to the regulatory region, induces DNA conformational change that favors the formation of an open complex at the promoter regions (52). The chromatin-associated protein H-NS may also playa regulatory role in the transcription of pap operons. The transcription of the papBA operon, when located on a multicopy plasmid, is stimulated twofold by mutation in drdX, a gene identical to hns. Moreover, cAMP-CAP and PapB are no longer required for papBA transcription in such mutants (53). It has been proposed that H-NS represses an intrinsically active papBA promoter and that cAMP-CAP and PapB are required as antirepressors to antagonize the silencing by H-NS (44). Recent work in this laboratory using chromosomally located single copy papBA operon indicates, however, that Pap phase variation occurs in an hns mutant background (146). This finding argues against the hypothesis that H-NS plays a central role in the regulation of papBA transcription. Another chromosomal locus that is involved in the transcriptional regulation of pap operons was identified by screening for mutations giving a locked OFF phenotype following Tn-lO mutagenesis (20). This gene, originally named mbf for Methylation Blocking Factor, was later found to be identical to lrp (22), the gene encoding a global regulatory protein that has been implicated in the regulation of dozens of genes or operons (26). Lrp is absolutely required for the transcription of the papBA operon. Recent evidence suggests that Lrp playa central role in the regulation of papBA expression regulation and Pap pili phase variation (21). Posttranscriptional regulation In a single pilus, the shaft is composed of about a thousand of PapA major pilin subunits and the fimbrillum is composed of minor pilins present at much lower numbers (84). An obvious question then is how cells maintain the balance of those subunits. Such differential gene expression from a single polycistronic operon can be achieved by transcriptional attenuation or termination, by transcription from an internal promoter, or by 1 2 differential degradation of processed mRNA. A stem-loop terminator sequence is found between papA and papH genes. It is postulated that the transcription from the papBA promoter is attenuated at this sequence and only a small proportion of the transcription events can proceed passing this sequence and generate mRNA for the papH and downstream genes (6, 7, 116). Therefore, the predominant species of mRNA produced by the transcription from the papBA promoter is a dicistronic mRNA encompassing papB and papA. However, probing a northern blot with papA sequence revealed two mRNA species, one the expected size of the dicistronic mRNA of papB and papA and a smaller one that was not detected by probing with papB sequence (7), This is because the dicistronic mRNA is further processed so that a shorter mRNA species encompassing only the papA sequence is produced, while the part encompassing papB is rapidly degraded. This mRNA processing is dependent on the activity of RNase E (115), a RNase involved in the processing of a variety of mRNA precursors (111, 113, 114). The processed papA monocistronic mRN A is translation ally active and markedly more stable than the dicistronic mRNA, with a half life of about 27 min in contrast to the 2.5 min half life of the papB papA dicistronic mRNA. It is believed that the papA mRNA is stabilized by the stem-loop terminator structure at the end of papA and another stem-loop structure formed at the head ofthe processed papA mRNA (ll5). The production of Pap pili is shut off when cells are grown at 26°C or lower temperature. This thermoregulation is abolished by mutations in the gene rim] (159), which encodes an N-terminal acetylase of ribosomal protein 5S (162). In such mutants, transcription from papBA promoter at lower temperature occurs as well as at 37°C, as determined by ~-galactosidase activity of papBA-IacZYA fusion and by Northern blotting assays (160). The involvement ofRimJ suggests that posttranscriptional events are, at least partly, responsible for Pap pili thermoregulation. It is hypothesized that the pap mRNA is destabilized at low temperature so that it is quickly degraded and ribosome or RimJ is involved in such a process (160). However, recent observations from our laboratory 1 3 suggest that H-NS may play an important role in thermoregulation of pap and that RimJ may affect the function of H-NS (C. White-Zigler and D. Low, unpublished data). Phase variation The expression of Pap pili is also controlled by a phase variation mechanism at the transcriptional level. This phase variation control involves differential DNA methylation and interaction with the global regulatory protein Lrp (15, 21). This will be further discussed in a later section. DNA Methylation Deoxyladenosine Methylase (Dam) Deoxyladenosine methylase (Dam), also known as DNA adenine rnethyltransferase, is a 32 KDa protein that methylates the adenine at the N6 position in a DNA sequence containing 5'-GATC-3' (107, 108). Dam is present in a variety of prokaryotic organisms, including Escherichia coli, Salmonella typhimurium, Streptococcus pneumoniae, and Vibrio cholerae (10, 107). In E. coli, mutation or overexpression of the dam gene leads to a hypermutable phenotype and uncoordinated DNA replication (l08). The function of Dam has been shown to be involved in various cellular processes, such as mismatch repair, DNA replication initiation and transcriptional regulation (108). Mismatch Repair Mutants of the dam gene or Dam overproducer strains exhibit elevated frequencies of point mutations (64, 109). This is because Dam is an integral part of the methyl-directed mismatch repair system of bacterial cells. Mismatches occurring during DNA replication are first corrected by the proofreading function of the DNA polymerase. Those that are missed by the proofreading machinery are recognized by a mismatch repair system that includes MutS, MutL, and MutH, which subsequently replaces one of the mismatching nucleotides with one that matches (87, 112). Normally, DNA methylation trails DNA replication by a few minutes. Sometimes this lag can be up to approximately 10 min or one third of the cell 14 cycle due to the limited level of Dam in the cells. As a result, the newly synthesized strand is undennethylated immediately following the passage of DNA replication fork, leading to the hemimethylated GATC sites (108). The protein MutH cleaves DNA 5' to the G in the nonmethylated strand of a hemimethylated GATC site adjacent to a mismatch recognized by MutS and MutL (87, 112). The mismatch repair system is therefore able to discriminate the newly replicated strand from the old template strand. A stretch of DNA up to several Kb on the newly replicated, and therefore nonmethylated, strand is then excised and resynthesized by DNA polymerase III (87). Using A phages containing mismatches in specific genes, it was shown that the repair of the mismatch following phage infection was biased in favor of the methylated strand when hemimethylated phage DNA was used, whereas the repair was random when nonmethylated phage DNA was used (123). Similar results were obtained using plasmid DNA carrying mismatches in a defined in vitro system (87). This is in agreement with the early observation that in dam mutants the mutation rate is much higher then in the Wild-type. In the case of Dam overproduction, the lag between DNA replication and DNA methylation is greatly reduced, resulting in completely methylated DNA before the mismatches can be corrected. The methyl-directed mismatch repair cannot correct the mismatches when both strands are methylated, because a methylated GATC site is not recognized by the repair system (112). Coordination of DNA Replication Initiation In a rapidly growing population of bacterial cells, a new round of DNA replication occurs before the end of the previous one. As a result, multiple replication initiation sites (orie) exist at the same time. In wild-type cells, the initiation at these sites is coordinated so that all the sites are used. Therefore, there are even numbers of initiation sites at any given time during the cell cycle (137). However, the initiation in the dam mutant cells is not coordinated so that odd numbers of initiation sites are present at any time (19). This finding suggests that Darn methylase is involved in the coordination of DNA replication initiation. 15 The distribution of GATe sequences in a DNA molecule is not even or random. There are 11 GATe sites in the 245 bp minimal oriC of E. coli, of which 8 are phylogenically conserved (163). Hemimethylated (but not completely methylated or nonmethylated) DNA containing oriC attaches to the cytoplasmic membrane in vivo and in vitro (119). Fully methylated plasmid DNA with oriC as a replication origin transforms dam mutants very poorly, and only hemimethylated plasmid DNA is recovered following such transformation. Mutants in dam can be transformed normally by nonmethylated plasmids with oriC as replication origin. In contrast, the wild-type cells are transformed normally by either methylated or nonmethylated plasmids with oriC as replication origin (133). These observations indicate that hemimethylated oriC sites are not effective for initiation of DNA replication in E. coli. The GATe sites in the oriC region become hemimethylated after initiation and the passage of the replication fork. It has been suggested that such hemimethylated oriC regions are subsequently sequestered by binding to cytoplasmic membrane and are not accessible to methylation by Dam methylase or the DNA replication machinery for a new round of DNA replication (27). The sequestered oriC are released later, become methylated, and are subject to DNA replication initiation (27). In dam mutant cells, the oriC sites cannot be hemimethylated and therefore presumably cannot be sequestered. These unsequestered oriC sites are available for replication initiation at any time, resulting in uncoordinated replication initiation. DnaA is a protein that determines the initiation of DNA replication (97). Interestingly, there are six GATe sites in the promoter region of dnaA, and the expression of dnaA is stimulated by DNA methylation (23, 83). The dnaA promoter may be sequestered after the passage of replication fork by the same mechanism as that of oriC (27). Therefore, DNA methylation can coordinate the initiation of DNA replication by linking the expression of the key factor in DNA replication initiation with the availability of the replication origin. 16 Transcriptional Modulation DNA methylation has not been recognized as a general mechanism controlling gene expression in prokaryotes. There are only a few known cases in which the expression of a gene is affected by DNA methylation by Dam. In most of these cases, a mutation in the dam gene has a modest two- to sixfold effect on the transcription of the genes in question, either positively or negatively (25, 128). Besides the positive regulation of dnaA by DNA methylation, the best known genes regulated by DNA methylation are those encoding transposon transposases. Expression of phage Mu mom gene, of which the promoter is downstream of a cluster of GATe sites, is reduced by 20-fold in a dam mutant host cell (60). The mechanism by which DNA methylation stimulates gene transcription is not completely understood. It is possible that the methylation at or near the promoter region causes DNA conformational change that precludes the binding of a negative transcriptional regulator. One such repressor gene, momR, was identified by screening mutants that render the transcription of mom independent of DNA methylation (18). The protein MomR, which is identical to the oxidative stress regulatory protein OxyR, binds to the mom regulatory region encompassing the cluster of GATe sites when the DNA is nonmethylated. Binding of MomR is not observed when DNA containing the mom regulatory region is methylated (18). DNA methylation negatively affects the transcription of sulA, trpA, trpS, glnS and the transposase genes of some transposons, including Tn-lO, Tn-5 and Tn-903 (11, 138). GATe sites are present within the promoter regions of all of these genes. The expression of those genes is stimulated in dam mutants by two- to tenfold. DNA methylation at the promoter region is believed to interfere by steric hindrance with the access of RNA polymerase to the promoter, causing reduced transcription (11, 138). Alternatively, the repression of these genes by DNA methylation can also be due to reduced affinity for the binding of a positive transcriptional regulatory protein. 1 7 In all of the systems described above, the methylation states of the GATC sequences are not regulated. Those GATC sequences are either fully methylated during most of the cell cycle or hemimethylated immediately following DNA replication. As a result, no variable patterns of DNA methylation corresponding to environmental signals are observed. Although such DNA methylation sensitive expression ties many cellular activities to the cell cycle, it is not an effective means for cells to respond to environmental signals. Nonmethylated GATC Sites It was believed that all the GA TC sites in E. coli chromosome were completely methylated, except for a short period of temporal hemimethylation immediately following the DNA replication (108). However, using more sophisticated techniques, it has recently been shown that about 0.2% of the estimated 18,000 GATC sites in E. coli K-12 chromosome are completely nonmethylated (129). This finding raises the question why those GATC sites are resistant to methylation. One possibility is that some of these sites are located in stretches of non-B-form DNA sequences that are resistant to methylation or protected by some other types of conformational steric hindrance. Another possibility is that these sites are protected by binding of proteins that are involved in transcriptional regulation or chromosomal organization (129). The fact that the methylation states of many of these GATC sites depend on environmental signals, including nutrition and temperature, suggests that the methylation states are coupled with the transcriptional activity of some genes (56, 129). Some of the nonmethy lated sites have been identified by cloning the flanking sequences. Of the nine nonmethy lated GA TC containing sequences cloned by Wang and Church (152), seven were shown to match genes in E. coli database. All the seven nonmethylated GATC sites are located within the 5' noncoding regions of specific genes, suggesting the involvement of binding of transcription regulatory proteins in methylation protection of those GATC sites. Though some of the sequences containing nonmethylated 1 8 GA TC sites exhibit strong consensus to the CAP binding sequence, protection by the binding of other proteins is also suggested. Hale et al. (56) have similarly cloned 60 sequences containing nonmethylated GA TC sites. An incomplete analysis of the 60 clones identified 10 different sequences that were located to the 5' noncoding regions of some known genes and some unidentified open reading frames. More importantly, the methylation states of these nonmethylated GATC sites were shown to be affected by stages of growth, the presence or absence of leucine, and the presence or absence of the global regulatory protein, Lrp (leucine responsive regulatory protein). Lrp and Gene Regulation Leucine Responsive Regulatory Protein. Lrp was first identified by mutations that affected the transport of branched-chain amino acids (4) and later independently identified by researchers studying different gene regulation systems (5, 20, 70, 92). Purified Lrp is a 164 amino acid, 18.8 KDa protein with a pI of 9.3 that exists as a dimer in solution. It is moderately abundant, about 3,000 dimers per cell or about 0.1 % of total cellular protein when cells are grown on a glucose­based minimal medium, as determined by antibody titration (161). Genes encoding Lrp in Salmonella typhimurium, Klebsiella aerogenes and Enterobacter aerogenes have also been cloned and sequenced. They are about 90% identical to their E. coli counterpart and among themselves on the DNA sequence level. The amino acid sequences encoded by the lrp genes of the four different enteric bacteria are almost identical, differing only at two positions. Almost all of the mutations on the DNA sequence level are located at the third "wobble" position of codons, resulting in no change at the amino acid sequence level (45). The extremely high degree of conservation of Lrp among the four genera suggests that Lrp is a very important protein and is highly adapted for its function in enteric bacteria. 19 Lrp does not show extensive homology to any known protein, except to AsnC (79), Pseudomonas putida BkdR (105), and Zymomonas mobilis Grp (120), which are all involved in the metabolism of amino acids. AsnC is a positive transcriptional regulator of the asparagine synthetase encoding gene asnA; BkdR is a regulatory protein of the genes involved in the branched-chain amino acids degradation, and Grp is a protein involved in the regulation of glutamate uptake, respectively. Lrp shares 25% identity and at least 50% close similarity with AsnC, 36.5% identity and 55.8% similarity with BkdR, and 33% identity and 52% similarity to Grp, suggesting that they are evolution ally related. A plasmid carrying the Zrp gene is capable of partially complementing bkd mutations in regulating the bkdAIA2B-ZpdVoperon (105). A plasmid carrying the grp gene has been shown to partially complement Zrp mutations, as determined by activation of ilvIH transcription (120). Recently, the whole genome of Haemophilus injluenzae was sequenced. Two Zrp­like genes were found, which encode proteins carrying 77.2% identity and 86.7% similarity and 29.6% identity and 52.6% similarity to E. coli Lrp. However, the functions of these two gene products have not been reported. A helix-tum-helix motif has been identified within Lrp (26, 121), suggesting a DNA-binding protein and consistent with its role as a transcription regulatory protein. Mutational analyses using the ilvIH system suggest that Lrp has three functional domains. Mutations affecting DNA binding occur within the N-terminal third of the protein, coincidental to the proposed helix-tum-helix motif, suggesting a DNA binding domain. Mutations within the C-terminal third of the protein affect only the responsiveness to leucine, suggesting a leucine responding domain. Mutations within the middle part of the protein result in reduced transcription without affecting DNA binding, suggesting a transcription activation domain (26, 121). The expression of Zrp is autoregulated by Lrp. Wang et aZ. (156) showed that the expression of Lrp was stimulated threefold in a Zrp mutant and reduced tenfold in a Lrp overproducing strain. This autoregulation is not affected by leucine. The lrp gene is 20 maximally expressed when cells are grown on minimal medium. The expression is reduced by four- to lO-fold when the cells are grown in rich medium. However, this effect of rich medium on lrp expression is not mediated by Lrp, as the same effect is observed in lrp mutants (90). Therefore, there is another mechanism which regulates the expression of Lrp in responding to growth conditions. Lrp Regulon The gene encoding Lrp has been independently identified by a number of groups studying different gene regulation systems. These include the biosynthesis and the degradation of amino acids, the transportation of amino acid and peptide, the synthesis of tRNA, and the production of pili (26). Conceivably Lrp regulates the transcription of a large number of genes involved in different aspects of cellular activities. Systematic searches for Lrp regulated genes have identified a large number of genes or gene products of which the expression is regulated by Lrp (40, 90, 143). In one of these studies polypeptides whose production is affected by the presence or absence of Lrp were identified by two-dimensional gel electrophoresis. It has been shown that the expression of at least 30 polypeptides is affected by Lrp (40). Another study shows that in at least 66 clones carrying random A placMu, the expression of lacZ is affected by Lrp, suggesting that a large number of genes with these random insertions are regulated by Lrp (90). However, these studies do not distinguish genes directly or indirectly regulated by Lrp. Table 1-1 is an incomplete list of members of Lrp regulon identified in various studies. Lrp-regulated genes (operons) are involved in a variety of aspects of cell activities (26). As shown in Table 1-1, these include, but are not limited to, biosynthesis of amino acids (glnALG, gltBDF, ilvIH, leuABCD, and serA) (40, 41, 90, 122), catabolism of amino acids (sdaA, tdh, gcv, and glyA) (90, 92), transport of amino acids and oligopeptides (livJ, livKHMGF, oppABCD, and ompC) (5, 40, 42, 57), catabolism of sugars (malT, malEFG, malKlamBmalM, CP8, [recently shown to be an allele of gitD]) 21 Table 1-1. Lrp regulon. An incomplete list of genes or operons that are regulated by Lrp. A plus or minus sign in the "Effect of Lrp" column indicates that Lrp positively or negatively regulate the corresponding gene or operon. A plus sign in the "Effect of Leucine" column indicates that leucine is required for the function of Lrp in regulating the corresponding gene or operon, a minus sign indicates that leucine inhibits the function of Lrp; an equal sign indicates that leucine has little or no effect on the function of Lrp, a star indicates that the effect of leucine is not reported or is not clear. 22 Effect Effect Class Operon Function of of Ref. Lrp leucine Amino acid glnALG Glutamine biosynthesis + * (40) biosynthesis gltBDF Glutamate biosynthesis + (40,41) ilvlH Branched chain amino + (122) acid biosynthesis serA Serine biosynthesisis + (92, 125) leuABCD Leucine biosynthesis + (90) Amino acid sdaA Serine degradation (92) degradation tdh Thereonine degradation (92) gcv Glycine degradation + = (90) Generating 1 C units glyA Serine to glycine * (30) transfer Amini acid /iv] High affinity branched + (57) transport chain amino acid transport livKHMGF High affinity branched + (57) chain amino acid transport sdaC Serine transport + + (136, 143) oppABCD oligopeptide transport * (5) ompC general transport = (42) Sugar malT Regulating maltose + * (143) degradation usage malEFG Maltose uptake + * (143) maiKlamB Maltose uptake + * (143) CP8 (gitD) Pentose degradation + * (30, 143) Pili papBA Pap pili + = (22) formation JaeBC K88pili * (65,66) JanABC K99 pili + (22) sJaBA Spili + = (147, 148) daa F1845 pili + = (147, 148) JimBE Type 1 pili phase + + (13) variation Others lysU Lysyl-tRNA synthesis (77, 90) pnt DADPH synthesis + (30,50) osmY periplasmic protein, (88) Function unknown bp Transcriptional = (90, 156) regulation 23 (90, 143), production of pili (papBA, jaeBC,fanABC, sjaBA, daaA, andjimA) (12, 13, 15, 22, 66, 67, 147), and others (lysU [48, 91]; pnt [30] and osmY [88]). There are also a number of unidentified genes and proteins which appear to be regulated by Lrp (40, 90). The most striking aspect of the Lrp regulon is not only the large number of its members but also the diverse patterns of the effects of Lrp on the expression of those genes. Lrp can be either a positive or a negative regulator of different operons, and leucine can either positively or negatively affect the function of Lrp on a particular regulatory system. Yet in some other cases, leucine has no effect on the function of Lrp. Therefore, there are six modes of regulation by Lrp (26). (i) Lrp activates transcription, and the activation is inhibited by leucine, such as the regulation of the ilvIH operon (122); (ii) Lrp activates transcription, and the activation is dependent or stimulated by leucine, such as in the case of regulation ofjimB andjimE (13); (iii) Lrp activates transcription, and leucine has little or no effect on the activation, such as in the case of the regulation of papBA (22). (iv) Lrp represses transcription, and the repression is relieved by leucine, such as in the case of the regulation of sdaA (92); (v) Lrp represses transcription, and the repression requires leucine, such as in the case of the regulation of liv] and livKHMGF (57); and (vi) Lrp represses transcription, and leucine has little or no effect on the repression, such as in the case of the regulation of ompC (42). It has been proposed that Lrp coordinates the response of bacterial cells to conditions of feast or famine (26). Lrp functions to regulate metabolic pathways in response to the availability of amino acids and nitrogen bases in the environments. In general, when cells are grown in a low nutrient environment, genes required for the biosynthesis of amino acids are activated by Lrp. and genes required for the degradation of amino acids are repressed by Lrp. Now cells are adjusted to a famine. Conversely, when cells are grown in a nutrient rich environment, the activation of amino acid biosynthesis genes and the repression of amino acid degradation genes are inhibited by the elevated level of leucine. This feast vs. famine response may also account for the transcriptional 24 activation of pili genes by Lrp, because special types of pili are often required for colonization in specific nutrient-deficient environments. Besides its function as transcription regulatory protein for a variety of operons, Lrp may also have a function as a chromosomal organizer (30). This hypothesis is based on the observation that Lrp binds not only to the regulatory region of operons under its regulation but also to some other sequences. For example, Lrp has been reported to bind to 4 of the 10 HinFI fragments of the plasmid pBR322 (30), though the relative affinity of Lrp to those fragments in comparison with other Lrp-bound DNA sequences is not reported. Taking the binding of Lrp to pBR322 as an indication, there might be a huge number of low affinity Lrp binding sites on the chromosome. Lrp binding to those hypothetical sites would significantly affect the local structure and overall conformation of the chromosome, with a concomitant effect on the transcription of some operons. Interaction of Lrp with DNA Molecules Lrp binds specifically to DNA fragments containing the regulatory regions of the genes it regulates (47, 49, 66, 117, 122, 126, 153). In the case of ilvIH, a 331 bp region upstream of the transcription initiation point is sufficient for both transcription initiation and leucine-mediated repression of the operon (61). In vitro, Lrp binding to the ilvIH regulatory region is reduced by leucine, but not by isoleucine or valine (126). Lrp binds specifically to at least six sites over a 200 bp region in the regulatory region, as assayed by MPE footprinting (153). The binding to those sites is highly cooperative. The two sites with the highest affinity for Lrp binding, sites 2 and 4, contain dyad symmetrical sequence 5'-AGAATtttATTCT-3' and 5'-AGGATtttATCGT-3', respectively. The other sites, which have much lower affinity for Lrp binding, have sequences that conform to half of the symmetric sites. At low concentration, Lrp binds cooperatively to the upstream binding sites. At higher concentration, Lrp also cooperatively binds to the four downstream binding sites. It is the binding to these downstream sites that is required for Lrp activation of the ilvIH operon, since mutations in the upstream sites reduces the transcription slightly and 25 mutations in each of the downstream sites, except the one most proximal to the transcription initiation point, reduce the transcription significantly (153). The stoichiometry of Lrp binding to ilvIH regulatory region has been determined, corresponding to a dimer binding to each site (154). Salmonella typhimurium has a cryptic ilvIH operon that is transcribed, but the translation is prematurely terminated due to a nonsense mutation near the start codon (127). The upstream region of S. typhimurium ilvIH operon is less than 60% identical to that of E. coli but still bound specifically by purified Lrp (155). MPE footprinting also revealed six Lrp binding sites, of which three were similar to those from E. coli. A consensus sequence, AgaATTTT ATtcT, has been proposed as Lrp binding sequence, based on the analysis of the 12 Lrp binding sites identified in E. coli and S. typhimurium ilvIH regulatory regions (155). An Lrp binding consensus sequence has also been drawn by selecting from a pool of random sequences the ones bound by Lrp. A pool of 50 bp DNA fragments with the middle 20 bp randomized was used as probe in Lrp mediated gel retardation. DNA molecules bound by Lrp were purified from the gel and amplified by PCR and subjected to a new round of selection. After a few rounds of selection-amplification, the selected sequences were cloned and sequenced. Comparison of the cloned sequences revealed a consensus sequence of YAGHA W ATTWTDCTR, where Y = C or T, H = no G, W = A or T, D = no C, R = A or C (29). This consensus sequence is similar to the one drawn by comparing the Lrp binding sites of E. coli and S. typhimurium ilvIH regulatory regions. Surprisingly, the Lrp bound sequences selected both in the presence and in the absence of leucine are very similar, suggesting leucine does not affect the specificity of Lrp binding to DNA (29). Another Lrp consensus sequence, TTTATTCtNaAT, has also been proposed based on the comparison of the regulatory regions of 11 Lrp regulated genes (125). However, further analysis has not been done on most of those sequences, and whether the sequences 26 corresponding to the proposed consensus sequence are bound by Lrp remains to be determined. It is possible that neither of the proposed consensus sequence accounts for the true Lrp-recognizing motif in all the systems regulated by Lrp, because such sequences do not exist in the regulatory regions of all the genes regulated by Lrp, and in some cases, such sequences, when present in the regulatory region, footprinting analyses fail to detect Lrp binding (30). Lrp has also been shown to bend DNA containing Lrp binding sites from the ilvIH regulatory region. Circular permutation experiments showed that with a single Lrp binding site, the high affinity site number 4 of the ilvIH Lrp binding sites, Lrp binding induced a bending of 550 . When two adjacent sites were tested, the DNA was bent at least 1300 • Considering the extensive footprinting caused by Lrp binding, it is most likely that the binding of Lrp induces a DNA conformational change such that a nucleoprotein complex is formed, either by DNA looping or wrapping around the protein molecules (154). However, binding of Lrp to the regulatory region alone is not sufficient for its function as transcription activator, according to the functional analyses of a number of Zrp activation mutants. The Lrp molecules produced by these activation mutants bind to the regulatory region of ilvIH normally but do not activate transcription (121). These same mutants Lrp also failed to activate papBA transcription, although the binding patterns to the pap regulatory region were similar to that of the wild-type Lrp (146). It has been proposed that Lrp interacts either directly with RNA polymerase or with another transcription activator, such as CAP, to activate transcription. Pap Pili Phase Variation E. coli isolates from patients with urinary tract infections produce Pap pili of different serotypes, which can be distinguished by specific antibodies to the major pilin subunit, PapA (33, 34). A number of pap operons have been cloned (9, 35, 149). In some cases a single clinical isolate expresses multiple antigenic types of Pap pili simultaneously. For example, a serotype 06 strain, C1212, contains two pap sequences that encode two 27 antigenically distinct Pap pili, antigenic type F71 and antigenic type F72 (100, 149). The type F71 pili are composed of pilin monomers of 17 KDa (Pap-17), whereas type F72 pili contain 21 KDa (Pap-21) pilin monomers. Although the majority of the cell (about 85%) in a single colony express the Pap-21, only a small portion, about 5%, of cells in the same colony express Pap-17. About another 10% of cells express neither Pap-17 nor Pap- 21(101). The differential expression state of Pap pili among the cells in a single colony is due to phase variation, by which individual cells alternate the states of pili expression. The expression of Pap-17 becomes constitutive when the operon is cloned on a multicopy plasmid and transformed into E. coli K-12, suggesting that the presence of the excessive copies of pap genes abolishes the regulation of the expression. The expression of Pap-17 is restored to the wild-type level when the operon is subcloned on a single copy plasmid (101). Similarly, in another uropathogenic isolate, A55, which contains only the chromosome encoded Pap pili similar to Pap-17, about 6% of cells express the Pap pili (101). These observations indicate that the phase variation of Pap pili expression is delicately regulated and it can only be observed when the genes are present at single or very low copies. The phase variation of Pap pili, examplified by the expression of Pap-17, was further studied by subcloning the pap sequence into a phage A.. vector so that the papBA promoter controls the transcription of lacZYA. The A.. construct containing the pap regulatory sequence and the lacZYA fusion was then integrated into the chromosome of E. coli K12 strain MC4100 at the att site. The expression of LacZYA in this strain is regulated in the same way as the expression of the Pap pili (16). Both Lac+ and Lac- colonies are observed at frequencies reflecting the piliated and nonpiliated cells from a parental colony. The expression of a Lac+ phenotype is also repressed by growing in rich medium or in glucose based minimal medium. The expression state of LacZY A is heritable as phase ON (Lac+) colonies give rise to predominantly phase ON colonies and phase OFF (Lac-) give rise to predominantly phase OFF colonies. The frequencies of Pap pili phase variation of 28 cells grown on M9~glycerol minimal medium at 37°C are determined by plating thousands of cells from a single colony, assuming each parental Lac+ cell gives rise to a Lac+ colony and each Lac- cell gives rise to a Lac- colony. The phase ON to OFF switch occurs at a frequency of 2.6 X 10-2 per cell per generation and from phase OFF to phase ON at a frequency of 1.57 X 10-4 . Mechanisms of Pili Phase Variation in Bacteria A number of mechanisms of genetic variation (phase variation and antigenic variation) have been described (132) (see Figure 1-2). Type 1 pili phase variation occurs by RecA-independent site-specific recombination. This process requires the activities of FimB, FimE and IHF (39, 78). DNA sequencing indicates that a 314 bp DNA fragment, which contains the promoter region of the major pilinfimA gene, is differentially oriented in phase ON and phase OFF cells. In one orientation the promoter faces towards the fimA gene so that transcription can occur. In the other orientation the promoter faces away from fimA gene so that transcription is precluded (1,38). In contrast, the phase and antigenic variations of N. gonorrhoeae occur by RecA­dependent general homologous recombination (110). On the chromosome there are a pilE locus, where the major pilin gene pU is expressed, and pilS loci, where a number of silent copies of the pit gene encoding pilins of different antigenic specificity are located. The copies of the pi! genes at the pitS locus are silent because they lack the amino terminal portion of the gene including the promoter. However, those silent copies of the pit gene can replace the copy at the pilE locus by aRecA-dependent unreciprocal recombination, leading to the expression of a pilin with different antigenic specificity (110). As N. gonorrhoeae cells are readily transformed by DNA they pick up in the medium, the donor DNA to the pilE locus can also come from sibling cells undergoing autolysis. (51). When the replacing copy of pi! gene contains a nonsense mutation so that no pilin is expressed or a misense mutation so that the expressed pilin cannot undergo proper modification and assembly, no 29 Figure 1-2. Mechanisms of pili phase variation in bacteria. (A). RecA-independent site­specific recombination of E. coli Type 1 pili. (B). RecA-dependent nonreciprocal recombination of N. gonorrhoeae pili. (C). Short repeats variation of H. influenza pili. (D). Differential DNA methylation of E. coli Pap pili. 30 A. RecA-independent site specific recombination. --E. coli Type 1 pili fimAp ~ ON ~ tlmE I l,tlmB ... -I ilmL\ ~ ! t ~ OFF ~fimEI l,tlmB ... ... 1 fimA ~ B. RecA-dependent un reciprocal recombination. --N. gono"hoeae pili pilEp pilE ~ pilS-l pilS-2 ON ~I-S' X x~H ~ Pitli ! t ~ pilEp pilE pilS-l pilS-2 OFF ,X ~ No Pili 3 1 C. Short repeats variation. --H. injluenzae pili -35 hifBp -10 ~ ON ~ h~fA 1 k lOxTA I .. ~ 1 hitB OS .. ~ -10 hifAp t -35 -35 ~ hifBp -10 ~ OFF ~ h(fA 1 .. 9xTA I .. ~ 1 hifB OS .. ~ -10 hifAp -35 D. Differential DNA methylation --E. coli Pap pili r ~ papBAp ON ~ pap! 1 0 1 papB OS ~ t H3 r papBAp~ OFF ?papI 1 0 1 papB OS H3 32 pilus is produced, leading to a phase OFF state (110). Another common mechanism of pili phase variation is the use of short repeated sequence. Such repeated sequences can be found in the coding or the regulatory region. Mispairing can happen during DNA replication or DNA recombination, resulting in deletion or insertion of the repeating units, which in turn changes the coding frame or the promoter activity, and ultimately the expression state ofthe pili (132). This mechanism is employed by the phase variation system of Haemphilus inj1uenzae pili. There are IOTA repeats between the -10 and -35 sequences of the overlapping promoters of the divergently transcribed pili genes hifA and hifB, which confers the maximal expression of the pili genes. When mispairing occurs during DNA replication so that one TA repeat is deleted or inserted, the expression of the pili genes is turned off or significantly reduced, resulting in a phase OFF state (151). Mechanism of Pap Pili Phase Variation Whereas most phase variation systems, such as that of E. coli Type 1 pili, N. gonorrhoeae pili, and H. inj1uenzae pili, involve DNA sequence rearrangements or alteration, Pap pili phase variation does not. The pap regulatory region of DNA from populations of phase ON and phase OFF cells was sequenced following PCR amplifications. No changes at the DNA sequence level were detected in the phase ON and phase OFF populations. Moreover, Pap pili phase variation occurs independent of RecA. These and other results strongly indicate that Pap pili phase variation occurs by a mechanism that does not involve DNA rearrangement (16). Interestingly, DNA sequence analysis of pap revealed two GATC sites, which are the Dam methylase targeting sequence, locate in two nearly perfect (24/27 bp identical) inverted repeats in the regulatory region of the pap DNA, 102 bp apart (15). DNA from phase ON and phase OFF populations were analyzed by Southern blot following digestion by a number of restriction enzymes. Though the restriction patterns by most restriction enzymes did not change, different patterns were observed when Dam methylation sensitive 33 enzymes were used. Using restriction enzymes MboI, which cut only at nonmethylated GATC, DpnI, which cut only at methylated GATC, and Sau3AI, which cuts at both methylated and nonmethylated GATC sequence, it was established that the GATC proximal to pap/ (denoted as GATCI028 or GATC-I) is nonmethylated in the phase ON cells and methylated in the phase OFF cells, and conversely, the GATC proximal to papB (GATC1130 or GATC-II) is methylated in the phase ON cells and nonmethylated in the phase OFF cells (15). The differential DNA methylation patterns displayed by DNA from Pap phase ON and phase OFF cells suggest DNA methylation by Dam methylase is involved in regulating Pap phase variation. Examination of Pap phase variation in a Dam- background showed Dam methylase is indeed required for phase variation. In a mutant that carries a Tn-9 insertion in the dam gene, no phase OFF to phase ON transition was observed and the transcription from the papBA promoter was at the same level of the phase OFF population (15). However, complementation with plasmid carrying the dam gene did not restore the transcription and the phase variation. Nevertheless, replacing the mutant dam gene with a wild-type copy by transduction did restore the transcription and the phase variation to the wild-type level (15). These results indicate that high levels of Dam methylase also inhibit transcription from the papBA promoter and phase variation. Indeed, in the strain carrying the dam gene on a plasmid, the Dam level was fourfold higher and both GATC sites in the pap regulatory region were methylated (15). The finding of nonmethylated GATC sites in the Pap regulatory region on the E. coli chromosome was unusual, because it was believed that virtually all the GATC sites on E. coli chromosome were methylated, except for a short period of hemimethylation immediately following DNA replication (l08). However, a GATC site can remain nonmethylated if it is protected from methylation by the binding of a protein. Mutational analysis showed that PapI, PapB and CAP were not required for the protection of GATC-II from methylation, indicating another factor was involved in the protection of the GATC 34 sites from being methylated (20). Mutagenesis using mini Tn-lO identified a chromosomal locus unlinked to pap responsible for the observed protection. Mutation of this gene, mbf (for methylation blocking factor), resulted in the methylation of both GATC sites in the pap regulatory region. Also, the transcription of pap is shut off in mbf mutants (20). Mbf seems solely responsible for the methylation protection at GATC-II. However, both Mbf and PapI were required for the protection of GATC-I from being methylated (20). The gene mbfhas been shown to be identical to lrp (22). A model was proposed for the DNA methylation- and Lrp-dependent Pap phase variation (14). According to this model, the phase variation is determined by the interaction of the two GATC sites with Dam and Lrp or Lrp/Papl. For example, in a phase ON cell GATC-I is nonmethylated and GATC-II hemimethylated following DNA replication. Now Dam and Lrp or Lrp/PapI can compete for the two GATC sites. If Dam methylates GATC­II and Lrp/PapI bind to GATC-I, the cell remains phase ON. Conversely, if Dam methylates GATC-I and Lrp binds to GATC-II, the cell becomes phase OFF. A phase OFF to phase ON transition can be explained by a similar process. Introduction to Work in This Thesis Previous work has established that Pap pili phase variation is controlled by a mechanism involving the differential DNA methylation at two GATC sites in the regulatory region of the pap operons (14). This differential DNA methylation is dependent on the function of a global transcription regulatory protein, Lrp, which is hypothesized to protect the GATC sites from being methylated by binding to these sequences. A pap encoded protein, PapI, is also required for the methylation protection of the GATC-I from being methylated (20). Evidence will be presented in this thesis to show that (i) Lrp binds specifically to the pap regulatory region, (ii) interaction of PapI with Lrp alters the binding of Lrp to pap DNA, (iii) methylation states of the DNA greatly affect its interaction with Lrp and with Lrp/PapI, and (iv) cooperative binding of Lrp and Lrp/PapI to different sites 35 in the regulatory region helps to regulate Pap pili phase variation. The roles of Lrp, PapI, and DNA methylation in Pap pili phase variation will be discussed and a mechanistic model for Pap pili phase variation will be presented based on the experimental observations and some speculations in comparison with other gene regulation systems. References 1. Abraham, J. M., C. S. Freitag, J. R. Clements, and B. I. Eisenstein. 1985. An invertible element of DNA controls phase variation of type 1 fimbriae of Escherichia coli. Proc. Natl. Acad. Sci. USA 82: 5724-5727. 2. Agace, W., H. Hedges, M. Ceska, and C. Svanborg. 1992. Interleukin-8 and the neutrophil response in mucosal Gram-negative infections. J. Clin. Invest. 92: 780. 3. Agace, W., S. Hedges, J. Anderson, U. Anderson, M. Ceska, and C. Svanborg. 1993. Selective cytokine production by epithelial cells following exposure to E. coli. Infect. Immun. 61: 602. 4. Anderson, J. J., S. C. Quay, and D. L. Oxender. 1976. Mapping of two loci affecting the regulation of branched-chain amino acid transport in Escherichia coli K- 12. J. Bacteriol. 126: 80··90. 5. Andrews, J. C. and S. A. Short. 1986. opp-lac fusions and transcriptional regulation of the Escherichia coli trp-linked oligopeptide pelTIlease. J. Bacteriol. 165: 432-434 6. Baga, M., M. Goransson, S. Normark, and B. E. Uhlin. 1985. Transcriptional activation of a pap pilus virulence operon from uropathogenic Escherichia coli. EMBO J. 4: 3887-3893. 7. Baga, M., M. Goransson, S. Normark, and B. E. Uhlin. 1988. Processed mRNA with differential stability in the regulation of E. coli gene expression. Cell 52: 197-206. 8. Baga, M., M. Norgren, and S. Normark. 1987. Biogenesis of E. coli Pap-pili: PapH, a minor pilin subunit involved in cell anchoring and length modulation. Cell 49: 241-251. 9. Baga, S., S. Normark, J. Hardy, P. O'Hanley, D. Lark, O. Olsson, G. Schoolnik, and S. Falkow. 1984. Nucleotide sequence of the papA gene encoding the Pap pilus subunit of human uropathogenic Escherichia coli. J. BacterioL 157: 330. 10. Bandyopadhyay, R. and J. Das. 1994. The DNA adenine methyltransferase­encoding gene (dam) of Vibrio cholerae. Gene 140: 67-71. 11. Barras, F. and M. G. Marinus. 1989. The great GATC: DNA methylation in E. coli. Trends Genet. 5: 139-143. 36 12. Bilge, S. S., C. R. Clausen, W. Lau, and S. L. Moseley. 1989. Molecular characterization of a fimbrial adhesin, F1845, mediating diffuse adherence of diarrhea-associated Escherichia coli to Hep-2 cells. J. Bacteriol. 171: 4281-4289. 13. Blomfield, I. C., P. J. Calie, K. J. Eberhardt, M. S. McClain, and B. I. Eisenstein. 1993. Lrp stimulates phase variation of type 1 fimbriation in Escherichia coli K-12. J. Bacteriol. 175: 27-36. 14. Blyn, L. B 1991. The Regulation of Pap Pilin Phase Variation. Ph. D. Thesis. The University of Utah. 15. Blyn, L. B., B. A. Braaten, and D. A. Low. 1990. Regulation of pap pilin phase variation by a mechanism involving differential Dam methylation states. EMBO J. 9: 4045-4054. 16. Blyn, L. B., B. A. Braaten, C. A. White-Ziegler, D. A. Rolfson, and D. A. Low. 1989. Phase-variation of pyelonephritis-associated pili in Escherichia coli: Evidence for transcriptional regulation. EMBO J. 8: 613-620. 17. Bock, K., M. E. Breimer, A. Brignole, G. C. Hannsson, K. Karlsson, A., G. Larsson, H. Leffler, B. Samuelsson, E., N. Stromberg, C. Svanborg Eden, and J. Thurin. 1985. Specificity of binding of a strain of uropathogenic Escherichia coli to Gala.I-4Gal-containing glycosphingollipids. J. BioI. Chern. 260: 8545-8551. 18. Bolker, M. and R. Kahmann. 1989. The Escherichia coli regulatory protein OxyR discriminates between methylated and unmethylated states of the phage Mu mom promoter. EMBO J. 8: 2403-2410. 19. Boye, E. and A. Lobner-Olesen. 1990. The role of dam methyltransferase in the control of DNA replication in E. coli. Cell 62: 981-989. 20. Braaten, B. A., L. B. Blyn, B. S. Skinner, and D. A. Low. 1991. Evidence for a methylation-blocking factor (mbf) locus involved in pap pilus expression and phase variation in Escherichia coli. J. Bacteriol. 173: 1789-1800. 21. Braaten, B. A., X. Nou, L. S. Kaltenbach, and D. A. Low. 1994. Methylation patterns in pap regulatory DNA control pyelonephritis-associated pili phase variation in Escherichia coli. Cell 76: 577-588. 22. Braaten, B. A., J. V. Platko, M. W. van der Woude, B. H. Simons, F. K. de Graaf, J. M. Calvo, and D. A. Low. 1992. Leucine-responsive regulatory protein (Lrp) controls the expression of both the pap and fan pili operons in Escherichia coli. Proc. Nat!. Acad. Sci. USA 89: 4250-4254. 23. Braun, R. E. and A. Wright. 1986. DNA methylation differentially enhances the expression of one of the two E. coli dnaA promoters in vivo and in vitro. Mol. Gen. Genet. 202: 246-250. 24. Breimer, M. E. and K. A. Karlsson. 1983. Chemical and immunological identification of glycolipid-based blood group ABH and Lewis antigen in human kidney. Biochim. Biophys. Acta 755: 170. 37 25. Busby, S., A. Kolb, and H. Bug. 1995. The E. coli cyclic AMP receptor protein. p. 177-191.In F. Eckstein and D. M. J. Lilley (ed), Nucleic Acids and Molecular Biology, Vol. 9. Springer-Verlag, Berlin. 26. Calvo, J. M. and R. G. Matthews. 1994. The Leucine-responsive regulatory protein, a global regulator of metabolism in Escherichia coli. Microbiol. Rev. 58: 466-490. 27. Campbell, J. L. and N. Kleckner. 1990. E. coli oriC and the dnaA gene promoter are sequestered from dam methy ltransferase following the passage of the chromosomal replication fork. Cell 62: 967-979. 28. Clegg, S. and G. F. Gerlach. 1987. Enterobacterial fimbriae. J. Bacteriol. 169: 934-938. 29. Cui, Y., Q. Wang, G. D. Strormo, and J. M. Calvo. 1995. A consensus sequence for binding of Lrp to DNA. J. Bacteriol. 177: 2872-4880. 30. D'Ari, R., R. T. Lin, and E. B. Newman. 1993. The leucine-responsive regulatory protein: more than a regulator. Trends Biochem. Sci. 18: 260-263. 31. de Man, P., B. Cedergren, S. Enerback, A. C. Larsson, H. Leffler, L. A. L., B. Nilsson, and C. Svanborg~Eden. 1987. Receptor-specific agglutination tests for detection of bacteria that bind globoderies glycolipids. J. CHn. Microbiol. 25: 401-406. 32. de Man, P., I. Claeson, M. Johanson, U. Jodal, and C. Svanborg~ Eden. 1989. Bacterial attachment as a predictor of renal abnormalities in boys with urinary tract infection. J. Pediatr. 115: 915-922. 33. de Ree, J. M., P. Schwillens, and J. F. van den Borsh. 1986. Monoclonal antibodies for serotyping the P fimbriae of uropathogenic Escherichia coli. J. Clin. Microbiol. 24: 121-125. 34. de Ree, J. M. and J. F. van den Boesh. 1987. Serological response to the P fimbriae of uropathogenic Escherichia coli in pyelonephritis. Infect. Immun. 55: 2204-2207. 35. Denich, K., B. B. Blyn, A. Craiu, B. Braaten, J. Hardy, D. A. Low, and P. D. O'Hanley. 1991. DNA sequences of three papA genes from uropathogenic Escherichia coli: evidence of structural and serological conservation. Infect. Immun. 59: 3849-3858. 36. Dodson, K. W., F. Jacob~Dubuisson, R. T. Striker, and S. J. Hultgren. 1993. Outer membrane PapC usher discriminately recognizes periplasmic chaperone-pilus subunit complexes. Proc. NatL Acad. Sci. USA 90: 3670. 37. Dominque, G. J., R. Laucirica, P. Baglia, S. Covington, J. A. Robledo, and S. C. Li. 1988. Virulence of wild-type Escherichia coli uroisolates in experimental pyelonephritis. Kidney Int. 34: 761-765. 38 38. Dorman, C. J. and C. F. Higgins. 1987. Fimbrial phase variation in Escherichia coli: Dependence on integration host factor and homologies with other site-specific recombinases. J. Bacteriol. 169: 3840-3843. 39. Eisenstein, B. I., D. S. Sweet, V. Vaughn, and D. I. Friedman. 1987. Integration host factor is required for the DNA inversion that controls phase variation in Escherichia coli. Proc. Natl. Acad. Sci. USA 84: 6506-6510. 40. Ernsting, B. R., M. R. Atkinson, A. J. Ninfa, and R. G. Matthews. 1992. Characterization of the regulon controlled by the leucine-responsive regulatory protein in Escherichia coli. J. Bacteriol. 174: 1109-1118. 41. Ernsting, B. R., J. W. Denninger, R. W. Blumenthal, and R. G. Matthews. 1993. Regulation of the gltBDF operon of Escherichia coli: how is a leucine-insensitive operon regulated by the leucine-responsive regulatory protein? J. Bacteriol. 175: 7160-7169. 42. Ferario, M., B. R. Ernsting, D. W. Borst, D. E. Wiese II, R. M. Blumenthal, and R. G. Matthews. 1995. The leucine-responsive regulatory protein of Escherichia coli negatively regulates transcription of ompC and micF and positively regulates translation of ompF. J. Bacteriol. 177: 103-113. 43. Forsman, K., M. Gorannson, and B. E. Uhlin. 1989. Autoregulation and multiple DNA interactions by a transcriptional regulatory protein in E. coli pili biogenesis. EMBO J. 8: 1271-1277. 44. Forsman, K., B. Sonden, M. Goransson, and B. E. Uhlin. 1992. Antirepression function in Escherichia coli for the cAMP-cAMP receptor protein transcriptional activator. Proc. Natl. Acad. Sci. USA 89: 9880-9884. 45. Friedberg, D., J. V. Platko, B. Tyler, and J. M. Calvo. 1995. The amino acid sequence of Lrp is highly conserved in four enteric microorganisms. J. Bacteriol. 177: 1624-1626. 46. Gaastra, W. and F. K. De Graaf. 1982. Host-Specific Fimbrial Adhesins of Noninvasive Enterotoxigenic Escherichia coli Strains. Microbiol. Rev. 129-161. 47. GaIly, D., T. Rucker, and I. Blomfield. 1994. The leucine-responsive regulatory protein binds to the fim switch to control phase variation of type 1 fimbrial expression in Escherichia coli K-12. J. Bacteriol. 176: 5665-5672. 48. Gazeau, M., F. Delort, P. Dessen, S. Blanquet, and P. Plateau. 1992. Escherichia coli leucine-responsive regulatory protein Lrp controls lysyl-tRNA synthetase expression. FEBS Lett. 300: 254-258. 49. Gazeau, M., F. Delort, M. Fromant, P. Dessen, S. Blanquet, and P. Plateau. 1994. Structure-function relationship of the Lrp-binding region upstream of lysU in Escherichia coli. J. Mol. BioI. 241: 378-389. 50. Gerolimatos, B. and R. L. Hansen. 1979. Repression of Escherichia coli pyridine nucleotide transhydrogenase by leucine. J. Bacteriol. 134: 394-400. 39 51. Gibbs, C. P., B. Y. Remann, E. Schultz, A. Kaufmann, and T. F. Meyer. 1989. Reassortment of pilin genes in Neisseria gonorrhoeae occurs by two distinct mechanisms. Nature (London) 338: 651-652. 52. Goransson, M., P. Forsman, P. Nilsson, and B. E. Uhlin. 1989. Upstream activating sequences that are shared by two divergently transcribed operons mediate cAMP-CRP regulation of pilus-adhesin in Escherichia coli. Mol. Microbiol. 3: 1557-1565. 53. Goransson, M., B. Sonden, P. Nilsson, B. Dagberg, K. Forsman, K. Emanuelsson, and B. E. Uhlin. 1990. Transcriptional silencing and thermoregulation of gene expression in Escherichia coli. Nature (London) 344: 682-685. 54. Hagberg, L., R. Hull, S. Hull, S. Falkow, R. Freter, and C. Svanborg-Eden. 1983. Contribution of adhesion to bacterial persistence in the mouse urinary tract. Infect. Immun. 40: 265-272. 55. Hagberg, L., H. Leffler, and C. Svanborg Eden. 1985. Non-antibiotic prevention of urinary tract infection. Infection 13: S 196-S200. 56. Hale, W. B., M. W. van der Woode, and D. A. Low. 1994. Analysis of nonmethylated GATC sites in the Escherichia coli chromosome and identification of sites that are differentially methylated in response to environmental stimuli. J. Bacterio!' 176: 3438-3441. 57. Haney, S. A., J. V. Platko, D. L. Ox ender, and J. M. Calvo. 1992. Lrp, a leucine-responsive protein regulates branched chain amino acid transport genes in Escherichia coli. 1. Bacteriol. 174: 108-115. 58. Hanley, J., I. E. Salit, and T. Hofmann. 1985. Immunochemical characterization of P pili from invasive Escherichia coli. Infect. Immun. 49: 581- 586. 59. Hasty, D. L., H. S. Courtney, E. V. Sokurenko, and I. Ofek. 1994. Bacteria-extracellular interactions., p. 197-211. In P. Klemm (ed.), Fimbriae: Adhesion, Genetics, Biogenesis and Vaccines. CRC Press, Boca Raton. 60. Hattman, S. 1982. DNA methyltransferase-dependent transcription of the phage Mu mom gene. Proc. Natl. Acad. Sci. USA 79: 5518-5521. 61. Haughn, G. W., C. H. Squires, M. DeFelice, C. T. Lago, and J. M. Calvo. 1985. Unusual organization of the ilvIH promoter of Escherichia coli. J Bacteriol. 163: 186-198. 62. Hedges, S., P. Anderson, G. Lidin-Jason, P. de Man, and C. Svanborg. 1991. Interleukin-6 response to deliberate colonization of the human urinary tract with Gram-negative bacteria. Infect. Immun. 59: 421. 63. Hedges, S., M. Svensson, and C. Svanborg. 1992. Interleukin-6 response of epithelia11ines to bacterial stimulation in vitro. Infect. Immun. 60: 1295. 64. Herman, G. E. and P. Modrich. 1981. Escherichia coli K-12 clones that overproduce dam methylase are hypermutable. J. Bacteriol. 145: 367-380. 40 65. Huisman, T. T. 1995. Regulation of Expression of the K88 and K99 Fimbrial Operons in Escherichia coli. Ph. D. Thesis. Vrije University, The Netherland. 66. Huisman, T. T., D. Bakker, P. Klaasen, and F. K. de Graaf. 1994. Leucine-responsive regulatory protein, IS 1 insertions, and the negative regulator FaeJA control the expression of the fae (K88) operon in Escherichia coli. Mol. Microbiol. 11: 525-536. 67. Huisman, T. T. and F. K. de Graf. 1995. Negative control of fae (K88) expression by the 'global' regulator Lrp is modulated by the 'local' regulator FaeA and affected by DNA methylation. Mol. Microbiol. 16: 943-953. 68. Hull, R. A., R. E. Gill, B. Hsu, B. Minshew, and S. Falkow. 1981. Construction and expression of recombinant plasmids encoding type 1 or D­mannose- resistant pili from a urinary tract infection Escherichia coli isolate. Infect. Immun. 33: 933-938. 69. Hultgren, S. J. and S. Normark. 1991. Biogenesis of the bacterial pilus. Curro Opin. Genet. DeveL 1: 313-318. 70. Ito, K., K. Kawakami, and Y. Nakamura. 1993. Multiple control of Escherichia coli lysyl-tRNA synthetase expression involves a transcriptional repressor and a translational enhance element. Proc. Natl. Acad. Sci. USA 90: 302- 306. 71. Jacob·Dubuisson, F., J. Heuser, K. Dodson, S. Normark, and S. Hultgren. 1993. Initiation of assembly and association of the structural elements of a bacterial pilus depends on two specialized tip proteins. EMBO 1. 12: 837. 72. Johnson, J. R. 1991. Virulence factors in Escherichia coli urinary tract infection. Clin. MicrobioL Rev 4: 80-128. 73. Johnson, J. R., P. L. Roberts, and W. E. Stamm. 1987. Fimbriae and other virulence factors in Escherichia coli urosepsis: Association with patients' characteristics. J. Infect. Dis. 156: 225. 74. Johnson, J. R. and W. E. Stamm. 1989. Urinary tract infections in women: diagnosis and therapy. Ann. Intern. Med. 111: 906-917. 75. Kallenius, G., R. Mollby, B. Svenson, J. Winberg, and H. Hultberg. 1980. Identification of a carbohydrate receptor recognized by uropathogenic Escherichia coli. Infection 8: S288-S293. 76. Kallenius, G., R. Mollby, S. B. Svenson, J. Winberg, A. Lundblad, S. Svensson, and B. Cedergren. 1980. The pk antigen as receptor for the haemagglutinin of pyelonephritic Escherichia coli. FEMS MicrobioL Lett. 7: 297- 302. 77. Kawakami, K., K. Ito, and Y. Nakamura. 1992. differential regulation of two genes encoding lysyl-tRNA synthetase in Escherichia coli: lysU constitutive mutants compensate for a lysS null mutation. MoL Micobiol. 6: 1739-1745. 41 78. Klemm, P. 1986. Two regulatory Jim genes, JimB andJimE, control the phase variation of type 1 fimbriae in Escherichia coli. EMBO. 1. 5: 1389-1393. 79. Kolling, R. and H. Lother. 1985. AsnC, an autogenously regulated activator of asparagine synthetase A transcription in Escherichia coli. J. Bacteriol. 164: 310- 315. 80. Korhonen, T. K., V. Vaisanen, H. Saxen, H. Hultber, and S. B. Svenson. 1982. P-antigen-recognizing fimbriae from human uropathogenic Escherichia coli strains. Infect. Immun. 37: 286-291. 81. Korhonen, T. K., V. Vaisanen-Rhen, M. Rhen, A. Pere, J. Parkkinen, and J. Finne. 1984. Escherichia coli fimbriae recognizing sialyl galactosides. J. Bacteriol. 159: 762-766. 82. Kraft, J. K. and T. A. Stamey. 1977. The natural history of symptomatic recurrent bacteriuria in women. Medicine 56: 55-60. 83. Kucherer, C., H. Lother, R. Kolling, M. A. Schauzu, and W. Messer. 1986. Regulation of transcription of the chromosomal dnaA gene of Escherichia coli. Mol. Gen. Genet. 205: 115-121. 84. Kuehn, M. J., D. Haslam, S. Norntark, and S. J. Hultgren. 1994. Structure, function, and biogenesis of Escherichia coli P pili, p. 37-51. In P. Klemm, (ed.), Fimbriae: Adhesion, Genetics, Biogenesis and Vaccines. CRC Press, Boca Raton. 85. Kuehn, M. J., J. Heuser, S. Normark, and S. Hultgren. 1992. P pili in uropathogenic E coli are composite fibbers with distinct fibrillar adhesive tips. Nature (London) 356: 252. 86. Kusters, J. G. and W. Gaastra. 1994. Fimbrial operons and evolution., p. 179- 196. In P. Klemm (ed.), Fimbriae: Adhesion, Genetics, Biogenesis and Vaccines. CRC Press, Boca Raton. 87. Lahue, R. S., K. G. Au, and P. Modrich. 1989. DNA mismatch correction in a defined system. Science 254: 160-164. 88. Lange, R., M. Barth, J. Bohringe, and R. Hengge-Aronis. 1993. Complex transcriptional control of the (78)-dependent stationary phase induced and osmotically regulated osmY (csi-5) gene suggests novel roles for Lrp, IHF, and cAMP-Crp in the stationary phase response of Escherichia coli. J. Bacteriol. 175: 7910-7917. 89. Leffler, H. and C. Svanborg-Eden. 1980. Glycolipid receptors for uropathogenic Escherichia coli on human erythrocytes and uroepithelial cells. Infect. Immun. 34: 920-929. 90. Lin, R., R. D'Ari, and E. B. Newman. 1992. placMu insertions in genes of the leucine regulon: extension of the regulon to genes not regulated by leucine. J. Bacteriol. 174: 1948-1955. 42 91. Lin, R., B. Ernsting, N. Hirshfield, R. G. Matthews, F. C. Neidhardt, R. L. Clark, and E. B. Newman. 1992. The lrp gene product regulates expression of lysU in Escherichia coli K-12. J. Bacteriol. 174: 2779-2784. 92. Lin, R. T., R. D'Ari, and E. B. Newman. 1990. The leucine regulon of E. coli: A mutation in rblA alters expression ofL-leucine-dependent metabolic operons. J. Bacteriol. 172: 4529-4535. 93. Lindberg, F., B. Lund, L. Johansson, and S. Normark. 1987. Localization of the receptor-binding protein adhesin at the tip of the bacterial pilus. Nature (London) 328: 84-87. 94. Lindberg, F., J. M. Tennent, S. J. Hultgren, B. Lund, and S. Normark. 1989. PapD, a periplasnlic transport protein in P pilus biogenesis. J. Bacteriol. 171: 6052. 95. Linder, H., I. Engerberg, I. Mattsby, K. Jann, and C. Svanborg Eden. 1988. Induction of inflammation by Escherichia coli on the mucosal level: requirement for the adherence and endotoxin. Infect. Immun. 56: 1309-1313. 96. Lipsky, B. A. 1989. Urinary tract infections in men: Epidemiology, pathophysiology, diagnosis, and treatment. Ann. Intern. Med. 110: 138-150. 97. Lobner-Olesen, A., K. Skarstad, F. G. Hansen, K. von Meyenboeg, and E. Boye. 1989. The DnaA protein determines the initiation mass of Escherichia coli K-12. Cell 57: 881-889. 98. Lomberg, H., B. Cedergren, H. Leffler, B. Nilsson, A. C. Carlstrom, and C. Svanborg-Eden. 1986. Influence of blood group on the availability of receptors for attachment of uropathogenic Escherichia coli. Infect. Immun. 51: 916- 926. 99. Lomberg, H., H. M., U. Jodal, H. Leffler, K. Lincoln, and C. Svanborg-Eden. 1984. Virulence-associated traits in Escherichia coli causing first and recurrent episodes of urinary tract infection in children with or without vesicoureteral reflux. J. Infect. Dis. 150: 561-569. 100. Low, D., V. David, D. Lark, G. Schoolnik, and S. Falkow. 1984. Gene clusters governing the production of hemolysin and mannose-resistant hemagglutination are closely linked in Escherichia coli serotype 06 and 04 isolates from urinary tract infections. Infect. Immun. 43: 353-358. 101. Low, D., E. N. Robinson, Z. A. McGee, and S. Falkow. 1987. The frequency of expression of pyelonephritis-associated pili is under regulatory control. Mol. Microbiol. 1: 335-346. 102. Low, D. A., B. Braaten, and M. van der Woude. 1996. Fimbriae. p. 146- 157. In F. Neidhardt (ed.), Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology. ASM publications, Washington, D.C. 103. Lund, B., F. Lindberg, B. I. Marklund, and S. Normark. 1987. The PapG protein is the a-D-galactopyranosyl-(1-4)-~-D-galactopyranose-binding 43 adhesin of uropathogenic Escherichia coli. Proc. Natl Acad. Sci. USA 84: 5898- 5902. 104. Lund, B., F. Lindberg, and S. Normark. 1988. Structure and antigenic properties of the tip located P pilus proteins of uropathogenic Escherichia coli. J. Bacteriol. 170: 1887-1894. 105. Madhusudhan, K. T., D. Lorenz, and J. R. Sokatch. 1993. The bkdR gene of Pseudomonas putida is required for expression of the bkd operon and encodes a protein related to Lrp of Escherichia coli. J. Bacteriol. 175: 3934-3940. 106. Marild, S., B. Wettergren, J. Hellstrom, U. J odal, K. Lincoln, I. Orskov, F. Orskov, and C. Svanborg-Eden. 1988. Bacterial virulence and inflammatory response in infants with febrile urinary tract infection or screening bacteriuria. J. Pediatr. 112: 348. 107. Marinus, M. G. 1984. Methylation of prokaryotic DNA. p.81-11l. In A. Razin" H. Cedar and A. D. Riggs (ed.), DNA methylation; Biochemistry and biological significance. Springer-Verlag. New York. 108. Marinus, M. G. 1987. DNA methylation in Escherichia coli. Ann. Rev. Genet. 21: 113-131. 109. Marinus, M. G., A. Poteete, and J. A. Arraj. 1984. Correlation of DNA adenine methylase activity with spontaneous mutability in Escherichia coli K-12. Gene 28: 123-125. 110. Meyer, T. F., C. P. Gibbs, and R. Haas. 1990. Variation and control of protein expression in Neisseria. Ann. Rev. Microbiol. 44: 451-477. 111. Misra, T. K. and D. Apirion. 1979. RNase E, an RNA processing enzyme from Escherichia coli. J. BioI. Chern. 254: 11154-11159. 112. Modrich, P. 1991. Mechanisnls and biological effects of mismatch repair. Ann. Rev. Genet. 25: 229-253. 113. Mudd, E. A., H. M. Krisch, and C. F. Higgins. 1990. RNase E, an endoribonuclease, has a general role in the chemical decay of Escherichia coli mRNA: Evidence showing that me and arm are the same genetic locus. Mol. Microbiol. 4: 2127-2136. 114. Mudd, E. A., P. Pretki, D. Belin, and H. M. Krisch. 1988. Processing of unstable bacteriophage T4 gene 32 mRNA into a stable species requires Escherichia coli ribonuclease E. EMBO J. 7: 3601-3607. 115. Nilsson, P. and B. E. Uhlin. 1991. Differential decay of a polycistronic Escherichia coli transcript is initiated by RNase-dependent endonucleolytic processing. Mol. Microbiol. 5: 1791-1799. 116. Norgren, M., M. Baga, J. M. Tennet, and S. Normark. 1987. Nucleotide sequence, regulation and functional analysis of the papC gene required for cell surface localization of the Pap pili of uropathogenic Escherichia coli. Mol. Microbiol. 1: 169. 44 117. Nou, X., B. Skinner, B. Braaten, L. Blyn, D. Hirsh, and D. Low. 1993. Regulation of pyelonephritis-associated pili phase variation in Escherichia coli: binding of the PapI and Lrp regulatory proteins is controlled by DNA methylation. Mol. Microbiol. 7: 545-553. 118. O'Hanley, P., D. Lark, S. Falkow, and G. Schoolnik. 1985. Molecular basis of Escherichia coli colonization of the upper urinary tract in BALB/c mice. J. Clin. Invest. 75: 119. Ogden, G. B., M. J. Pratt, and M. Schaechter. 1988. The replicative origin of the E. coli chromosome binds to cell membranes only when hemimethylated. Cell 54: 127-135. 120. Peekhaus, N., B. Tolner, B. Poolman, and R. Kramer. 1995. The glutamate uptake regulatory protein (Grp) of Zymomonas mobilis and its relation to the global regulator Lrp of Escherichia coli. J. Bacteriol.l77: 5140-5147. 121. Platko, J. V. and J. M. Calvo. 1993. Mutations affecting the ability of the Escherichia coli Lrp to bind to DNA, activate transcription, or respond to leucine. J. Bacteriol. 175: 11110-11117. 122. Platko, J. V., D. A. Willins, and J. M. Calvo. 1990. The ilvIH operon of Escherichia coli is positively regulated. J. Bacteriol. 172: 4563-4570. 123. Pukkila, P., J. Peterson, G. Herman, P. Modrich, and M. Meselson. 1983. Effects of high levels of DNA adenine methylation on methyl-directed mismatch repair in E. coli. Genetics 104: 571-582. 124. Reid, G. and J. D. Sobel. 1987. Bacterial adherence in the pathogenesis of urinary tract infection: A review. Rev. Infect. Dis. 9: 470-487. 125. Rex, J. H., B. D. Aronson, and R. L. Somerville. 1991. The tdh and serA operons of E. coli: Mutational analysis of the regulatory elements of leucine­responsive genes. J. Bacteriol. 173: 5944-5953. 126. Ricca, E., D. A. Aker, and J. M. Calvo. 1989. A protein that binds to the regulatory region of the ilvIH operon of Escherichia coli. J. Bacteriol. 171: 1658- 1664. 127. Ricca, E., C. T. Largo, M. Sacco, and M. De Felice. 1991. Absence of ace to hydroxy acid synthase TIl in Salmonella typhimurium is due to early termination of translation within the ilvI gene. Mol. Microbiol. 5: 1741-1744. 128. Richet, E. and L. Sogaad-Andersen. 1994. CRP induces the repositioning of MalT at the Escherichia coli malKp promoter primarily through DNA bending. EMBO J. 13: 4558-4567. 129. Rinquist, S. and C. L. Smith. 1992. The Escherichia coli chromosome contains specific, unmethylated Dam and Dcm sites. Proc. Natl. Acad. Sci. USA 89: 4539-4543. 130. Roberts, J. A. 1991. Etiology and pathophysiology of pyelonephritis. Amm. J. Kidney Dis 17: 1-9. 45 131. Roberts, J. A., K. Hardaway, B. Kaack, E. N. Fussel, and G. Baskin. 1984. Prevention of pyelonephritis by immunization with P-fimbriae. 1. Urol. 131: 602-607. 132. Robertson, B. D. and T. F. Meyer. 1992. Genetic variation in pathogenic bacteria. Trends Genet. 8: 422-427. 133. Russel, D. W. and N. D. Zinder. 1987. Hemimethylation prevents DNA replication in E. coli. Cell 50: 1071-1079. 134. Salit, I. and E. Gotschlich. 1977. Hemagglutination by purified type 1 Escherichia coli pili. 1. Exp. Med. 146: 1169-1181. 135. Sandberg, T., B. Kaijser, G. Lidin-Janson, K. Lincoln, F. Orskov, I. Orskov, E. Stokland, and C. Svanberg-Eden. 1988. Virulence of Escherichia coli in relation to host factors in women with symptomatic urinary tract infection. 1. Clin. Microbiol. 26: 1471. 136. Shao, Z. Q., R. T. Lin, and E. B. Newman. 1994. Sequencing and characterization of the sdaC gene and identification of the sdaCB operon in Escherichia coli K-12. Eur. 1. Biochem. 222: 901-907. 137. Skarstad, K., E. Boye, and H. B. Steen. 1986. Timing of initiation of chromosome replication in individual Escherichia coli cells. EMBO 1. 5: 1711- 1717. 138. Sternberg, N. 1985. Evidence that adenine methylation influences DNA-protein interactions in Escherichia coli. 1. Bacteriol. 164: 490-493. 139. Svanborg, C., F. Orskov, and I. Orskov. 1994. Fimbriae and disease, p. 239-254. In P. Klemm (ed.), Fimbriae: Adhesion, Genetics, Biogenesis and Vaccines. CRC Press, Boca Raton. 140. Svanborg Eden, C., R. Freter, L. Hagberg, R. Hull, S. Hull, H. Leffler, and G. Schoolnik. 1982. Inhibition of experimental ascending urinary tract infection by an epithelial cell-surface receptor analog. Nature (London) 298: 560-562. 141. Svanborg-Eden, C., G. Lidin-Janson, and U. Lindberg. 1979. Adhesiveness to urinary tract epithelial cells of fecal and urinary Escherichia coli isolates from patients with symptomatic urinary tract infections or asymptomatic bacteriuria of varying duration. 1. Urol. 122: 185. 142. Svenson, S. B., H. Hultberg, G. Kallenius, T. K. Korhonen, R. Mollby, and J. Winberg. 1983. P-fimbriae of pyelonephritogenic Escherichia coli: Identification and chemical characterization of receptors. Infection 11: 61-67. 143. Tchetina, E. and E. B. Newman. 1995. Identification of Lrp-regulated genes by inverse PCR and sequencing: regulation of two mal operons of Escherichia coli by leucine-responsive regulatory protein. 1. Bacteriol. 177: 2679-2683. 144. Tennent, J. M., S. Hultgren, B. Marklund, K. Forsman, M. Goransson, B. E. Uhlin and S. Normark. 1990. Genetics of adhesion 46 expression in Escherichia coli., p. 97-110. In B. W.lglewski and V. L. Clark (ed.), Molecular Basis of Pathogenesis. Academic Press, San Diego. 145. Van den Bosch, J. F., U. Verboom-Sohmer, P. Postma, J. De Graaff, and D. MacLaren. 1980. Mannose-sensitive and mannose-resistant adherence to human epithelial cells and urinary virulence of Escherichia coli. Infect. Immun. 29: 226-233. 146. van der Woude, M., L. S. Kaltenbach, and D. A. Low. 1995. Leucine­responsive regulatory protein plays dual roles as both an activator and a repressor of the E. coli pap fimbrial operon. Mol. Microbiol. 17: 303-312 147. van der Woude, M. and D. A. Low. 1994. Leucine-responsive regulatory protein and deoxyadenosine methylase control the phase variation and expression of the E. coli sfa and daa pili operons. Mol. Microbiol. 11: 605-618. 148. van der Woude, M. W., B. A. Braaten, and D. A. Low. 1992. Evidence for a global regulatory control of pilus expression in Escherichia coli by Lrp and DNA methylation: model building based on analysis of pap. Mol. Microbiol. 6: 2429-2435. 149. van Die, I. and H. Bergmans. 1984. Nucleotide sequence of the gene encoding the F72 fimbrial subunit of a uropathogenic Escherichia coli strain. Gene 32: 83-90. 150. van Die, I., E. Zuidweg, W. Hoekstra, and H. Bergmans. 1986. The role of fimbriae of uropathogenic Escherichia coli as carriers of the adhesin involved in mannose-resistant hemagglutination. Microb. Pathol. 1: 51-56. 151. van Ham, S. M., L. van Alphen, F. R. Mooi, and J. P. M. van Putten. 1993. Phase variation of H. inJluenzae fimbriae: transcriptional control of two divergent genes through a variable combined promoter region. Cell 73: 1187-1196. 152. Wang, M. X. and G. M. Church. 1992. A whole genome approach to in vivo DNA-protein interactions in E. coli. Nature (London) 360: 606-610. 153. Wang, Q. and J. M. Calvo. 1993. Lrp, a global regulatory protein of Escherichia coli, binds co-operatively to multiple sites and activates transcription of ilvIH. 1. Mol. BioI. 229: 306-318. 154. Wang, Q. and J. M. Calvo. 1993. Lrp, a major regulatory protein in Escherichia coli, bends DNA and can organize the assembly of a higher-order nucleoprotein structure. EMBO J. 12: 2495-2501. 155. Wang, Q., M. Sacco, E. Ricca, C. T. Lago, M. DeFelice, and J. M. Calvo. 1993. Organization of Lrp-binding sites upstream of ilvIH in Salmonella typhimurium. Mol. Microbiol. 7: 883-891. 156. Wang, Q., J. Wu, D. Friedberg, J. Platko, and J. Calvo. 1994. Regulation of the Escherichia coli lrp gene. J. Bacteriol. 176: 1831-1839. 157. Westerlund, B., P. Kuusela, T. Vartio, I. van Die, and T. K. Korhonen. 1989. A novel lectin-independent interaction of P fimbriae of Escherichia coli with immobilized fibronectin. FEBS Lett. 243: 199. 47 158. Westerlund, B., I. van Die, C. Kramer, P. Kuusela, H. Holthofer, A. M. Tarkkanen, R. Virkola, N. Riegman, H. Bergmans, W. Hoekstra, and T. K. Korhonen. 1991. Multifunctional nature of P fimbriae of uropathogenic Escherichia coli: mutations infsoE andfsoF influence fimbriae binding to renal tubuli and immobilized fibronectin. Mol. MicrobioI. 5: 2965. 159. White-Ziegler, C. A. and D. A. Low. 1992. Thermoregulation of the pap operon: evidence for the involvement of RimJ, the N-terminal acetylase of ribosomal protein S5. J. BacterioI. 174: 7003-7012 160. White-Ziegler, C. A. 1993. Thermoregulation of Pyelonephritis-Associated Pili Gene Expression. Ph. D. Thesis. The University of Utah. 161. Willins, D. A., C. W. Ryan, J. V. Platko, and J. M. Calvo. 1991. Characterization of Lrp, an Escherichia coli regulatory protein that mediates a global response to leucine. J. BioI. Chern. 266: 10768-10774. 162. Yoshikawa, A., S. Isono, A. Sheback, and K. Isono. 1987. Cloning and nucleotide sequencing of the genes rim] and rim] which encode enzymes acetylating ribosomal proteins S18 and S5 of Escherichia coli K12. Mol. Gen. Genet. 209: 481-488. 163. Zyskind, J. W. and D. W. Smith. 1986. The bacterial origin of replication, oriC. Cell 46: 489-490. CHAPTER II REGULATION OF PYELONEPHRITIS-ASSOCIATED PILI PHASE­V ARIATION IN ESCHERICHIA COLI: BINDING OF THE PAPI AND THE LRP REGULATORY PROTEINS IS CONTROLLED BY DNA :METHYLATION This chapter is reprinted from Molecular Microbiology (1993) 7:545-553 with permission from coauthers and Blackwell Science Ltd. 49 Molecular Microbiology (1993) 7(4), 545-553 Regulation of pyelonephritis-associated pili phase­variation in Escherichia coli: binding of the Papl and the Lrp regulatory proteins is controlled by DNA methylation Xiangwu Nou, Brett Skinner, Bruce Braaten, Lawrence Blyn, t Dwight Hirsch* and David Low* Division of Cell Biology and Immunology, Department of Pathology, University of Utah Medical Center, Salt Lake City, Utah 84132, USA. Summary Expression of pyelonephritis-associated pili (Pap) in Escherichia coli is under a phase-variation control mechanism in which individual cells alternate between pili+ (ON) and pili- (OFF) states through a process involving DNA methylation by deoxyadenosine methylase (Dam). Methylation of two GATC sites (GATC1028 and GATC11:w) within the pap regulatory region is differentially inhibited in phase ON and phase OFF cells. The GATC1028 site of phase ON cells is non-methylated and the GA TC1130 site is fully methylated. Conversely, in phase OFF cells the GA TC1028 site is fully methylated whereas the GATC1130 site is non-methylated. Two transcriptional activators, Papl and Lrp (leucine-responsive regula­tory protein), are required for this specific methylation inhibition. DNA footprint analysis using non­methylated pap DNAs indicates that Lrp binds to a region surrounding the GATC1130 site, whereas Papl does not appear to bind to pap regulatory DNA. However, addition of Lrp and Papl together results in an additional DNasel footprint around the GATC1028 site. Moreover, Dam methylation inhibits binding of Lrp/Papl near the GATC1028 site and alters binding of Lrp atthe GATC1130 site. Our results support a model in which Dam and Lrp/Papl compete for binding near the GATC1028 site, regulating the methylation state of this Received 6 August, 1992; revised 5 October, 1992; accepted 6 October 1992. tPresent address: Department of Cellular and Developmental Biology, University of Califomia, Irvine, Califomia 94037, USA. tOn sabbatical leave from the Department of Veterinary Microbiology and Immunology, University of California, Davis, Califomia 95616, USA. *For correspondence. Tel. [phone]; Fax [phone]; Email Low@Bio­science. utah.edu. GATC site and, consequently, the pap transcription state. Introduction The pap operon in uropathogenic Escherichia coli encodes pili and adhesin proteins that play important roles in the attachment of E. coli to urinary tract epithelial cells (Normark et al., 1986). The expression of Pap pili is under phase variation control: bacterial cells switch between pili+ (ON) and pili- (OFF) states. Unlike other phase-vari­ation systems examined, switching between ON and OFF Pap pili expression states occurs without pap DNA rearrangements or mutations (Blyn et al., 1989) by a mechanism involving deoxyadenosine methylase (Dam) (Blyn et al., 1990). Dam methylase plays a direct role in Pap phase-variation by methylating two GA TC sites in the pap regulatory region denoted GATC1028 and GATC1130. Methylation analysis of the pap regulatory region showed that in the ON population the GATC1028 site is non­methylated and the GATC1130 site is methylated. Con­versely, in the OFF population the GATC1028 site is methylated but the GATC1130 site is non-methylated (Blyn et al., 1990). We recently identified and cloned a gene denoted mbf (methylation blocking factor) which is necessary for inhi­biting the methylation ofthe pap GATC1028 and GATC1130 sites (Braaten et al .• 1991). This gene is identical to the recently described Irp gene (leucine-responsive regulatory protein) (Braaten et al., 1992) which appears to regulate a number of genes involved in cellular metabolism and will be denoted Irp here. Although methylation protection of the GATC1130 site does not require any pap-encoded proteins, the Papl regulatory protein is necessary in conjunction with Lrp for methylation protection of the GATC1028 site (Braaten et al., 1991). Here we show that Lrp binds to the GATC1130 region in the absence of Papt Binding of Lrp to the GATC1028 region is detected only after Papl addition using non-methylated and hemi­methylated DNAs, but is not detected when using fully methylated DNA. These results indicate that DNA methy­lation patterns control Pap phase-variation by modulating the binding of the Lrp and Papl regulatory proteins. 50 546 X. Nou et al. Table 1. The pap/ and /rp genes are required for transcription initiated at the papBA pilin promoter. Strain Relevant genotype Strain description (3-galactosidase specific activity" DL963 DL967 DL968 DL969 DL970 DL971 DL972 DL973 DL974 MC4100 (pNN387) MC41 00 (pDAL287) DL967/rp- Single copy vector control papBA,:HacZ in pNN387 DL967 mbf-20:.:mTn10b papBAp-/acZ + Papl in trans DL969 mbf-20:.:mTn10 paplp-lacZ in pNN387 2 ± 0.8 98 ± 38 51 ± 19 DL967 (pDAL262)C DL969/rp­MC4100 (pDAL288) DL971/rp- DL971 mbf-20:.:mTn10 pap/p-/acZ + Papl in trans DL973 mbf-20:.:mTn10 1980 134 58 ± 22 58 ±24 9±3 139 ± 68 8 1 DL971 (pDAL262) DL9731rp-a. Values are the mean of four independent measurements ± 1 standard deviation from the mean. Units are as defined by Miller (1972). b. The mbf-20 allele contains a mTn10 insertion within the Irp gene (Braaten 6t al., 1991) and was transduced into E. coli strains by phage P1 transduction (Blyn 8t al., 1990). c. Plasmid pDAL262 expresses Papl constitutively in E. coli MC41 00 since the lacl gene encoding lac repressor is deleted in this strain. Results Papl and Lrp are required for pap transcription To determine the roles of Papl and Lrp in pap phase-vari­ation, we examined the effects of these gene products on pilin transcription initiated from the divergent papBA and papl promoters. For this analysis, we used single-copy plasmids pDAL287 and pDA1288, which contain papBA­lacZ and papl-lacZ operon fusions, respectively, but lack intact PapB and Papl regulatory protein coding sequences. £. coli containing plasmid pDAL287 expressed about 100 Miller units of J3-galactosidase (DL967, Table 1). This J3-galactosidase level increased 20-fold in the presence of plasmid pDAL262, which produces Papl under the control of the lac promoter, indicating that Papl is a positive regulator of the papBA promoter (see strain DL969 , Table 1). However, in the absence of Lrp, Papl did not have any effect on transcrip­tion initiated at papBAp (compare strains DL968 and DL970). Similarly, Lrp had less than a twofold effect on the papBAp promoter in the absence of Papl (compare strains DL967 and DL968, Table 1). These results indicate that both Lrp and Papl are required for stimulation of transcrip­tion from the papBA promoter. £. coli containing plasmid pDA1288 (pap/-lacZ) expressed about 60 units of J3-galactosidase (DL971, Table 1). This level dropped sixfold inthe Irp- strain DL972 , indicating that Lrp is a positive regulator of papl transcrip­tion (compare these results to the twofold effect of Lrp on papBA transcription). Papl appeared to have a two- to threefold autostimulatory effect (strain DL973) which was dependent on the presence of Lrp (strain DL974). Together, these results indicate that Lrp and Papl work together to regulate transcription from both the papBA and papl promoters. In the absence of Lrp, Papl did not stimulate transcription from either of the pap promoters. In contrast, even in the absence of Papl, Lrp had a twofold stimulatory effect on transcription from papBAp and a sixfold stimulatory effect on transcription from paplp. Analysis of the binding of Lrp and Pap/ to pap regulatory DNA Our previous results showed that Lrp and Papl play roles in methylation inhibition of the pap GATC1028 and GATCl130 sites (Braaten et al., 1991). Therefore, we determined if Pap I and Lrp bind to pap regulatory DNA sequences by incubating a 292 bp pap DNA fragment derived from plasmid pDAL336 (Fig. 1), radiolabelled at one end, with protein extracts containing either Lrp, Papl, or both BamHI (955) o 2863 I ~ " " Sau3AI Sau3AI (1028) (1l30) I I " BamHI EcoRI " (291) (312) ,/ " ,/ pDAL336 (3134bp) EcoRI (1245) ,I ,/ ,/ ,/ Fig. 1. Map of plasmid pDAL336. The 292bp BamHI-EcoRI DNA fragment of plasmid pDAL336 was used in gel retardation assays (Fig. 2) and DNase I footprinting analyses (Figs 3 and 5). The plasmid pTZ19U DNA sequence is shown by the light oval and the pap regulatory DNA sequence is shown by the dark line. The pap GA TC1028 and GA TCmo sites are also indicated. 5 1 DNA methylation modulates pap protein-DNA interaction 547 DNA NonmethYiated Methylated Lrp _ _ + + + - + + + Papl _ + _ + _ _ + _ + _ b a Fig. 2. Gel retardation analysis. Protein extracts containing Lrp and Papl were incubated with methylated and non-methylated pap regulatory DNA probes and analysed using high-ionic-strength PAGE (see the Experimental procedures). Additions of extracts containing Lrp and Papl are indicated above each lane. Shown on the left of the figure are DNA fragments: 'a', the 200bp Pvull-BamHI DNA fragment of plasmid pDAL3368 (used as an internal control); 'b', the unshifted 292bp BBmHI-EcoRI pap DNA fragment; 'c', the shifted 292bp DNA fragment resulting from Lrp binding; 'd', the shifted 292bp DNA fragment resulting from Lrp- and Papl-binding. proteins. DNA-protein complexes were separated by polyacrylamide gel electrophoresis (see the Experimental procedures). Addition of a protein extract containing Papl (Lrp-) resulted in only a very slight band-shift of either fully methylated or non-methylated pap regulatory DNAs (Fig. 2, lanes 2 and 7). In contrast, addition of an extract containing Lrp (Papn shifted band 'b' to the more slowly migrating band 'c' regardless of the pap methylation state (Fig. 2, lanes 3 and 8). These results indicate that Lrp binds to pap regulatory DNA sequences and that methylation of the GATC,028 and GATC"3o sites does not block Lrp binding. Addition of Papl and Lrp together resulted in a further shift to band 'd' using methylated or non-methylated pap regulatory DNA sequences (Fig. 2, lanes 4 and 9). A control extract from a strain containing the papl gene in a transcriptional orientation opposite that of the phage T7 promoter and which is phenotypically Papl- did not result in any further shift of band 'c' (Fig. 2, lanes 5 and 10). Together, these results indicated that although Papl does not bind to pap DNA independently, it does appear to associate with the Lrp-pap DNA complex regardless of the pap methylation state. Although the results shown in Fig. 2 indicated that Papl, in the presence of Lrp, bound to both methylated and non-methylated pap regulatory DNAs, they did not pro­vide detailed information about the location of the DNA­binding sites or whether the interaction of Papl and Lrp with methytated DNA was the same as with non­methylated DNA. Therefore, we analysed the interactions of Papl and Lrp with pap regulatory DNA by DNasel footprinting (Galas and Schmitz, 1978). Papl, in the absence of Lrp, did not alter the control DNasel cleavage pattern using either non-methylated or methylated DNAs (Fig. 3, compare lanes 1,2 and 6,7, respectively), which is consistent with the results shown in Fig. 2 indicating that in the absence of Lrp, Pap I does not bind to pap DNA. In contrast, addition of a partially purified Lrp extract resulted in an extended 120bp footprint interrupted by DNasel hypersensitive sites (Fig. 3, lanes 3 and 8; footprints are indicated by brackets 'c' through 'I'). These results sug­gest the possibility that Lrp induces bending of pap DNA (Hochschild, 1991). The methylation state of pap appeared to affect Lrp binding since the GATC"30 site was hypersensitive to DNasel cleavage in methylated, but not non-methylated, pap DNAs (Fig. 3, compare lanes 3 and 8; footprints 'i' and 'j'). Addition of Papl and Lrp to non-methylated pap DNA resulted in two additional protected regions of DNA near theGATC,028 site (Fig. 3,lane4; footprints 'a' and 'b'). The results presented in panel B (Fig. 3) show that footprint 'b' overlaps the GA TC'028 site. These results are consistent with our previous data showing that Papl is necessary for methylation protection of the GATC1028 site (Blyn et al., 1990). In addition to the new footprints 'a' and 'b' near the GATC'028 site. a region near the GATC1130 site became stronger and more extended in the presence of Papl (Fig. 3A, footprints 'i', 'j' and 'k1, These results, which are summarized in Fig. 4, show that although Papl does not bind to pap regulatory DNA sequences in the absence of Lrp, Papl does affect Lrp-pap DNA interactions. Although addition of PapJ and Lrp to non-methylated pap DNA resulted in footprinting of the GATC,028 region, this region was not footprinted when fully methylated pap DNA was used (Fig. 3, compare lanes 4 and 9, footprints 'a' and 'b'), Similarly, footprints 'i', 'j', and 'k' near the GATC"30 site were not altered after Papl addition when fully methylated DNA was used. These results show clearly that DNA methylation alters binding of Lrp-Papl to the pap regulatory region (see Discussion). In our model for how E. coli alternates between phase ON and OFF methylation states, we hypothesized that after one round of DNA replication a fully methylated GATC1028 site (phase OFF state) would give rise to a hemimethylated 'transition state'. If Lrp/Papl binds to this hemimethylated DNA and inhibits binding of Dam, then after one more round of DNA replication the GATC,028 site would be non-methylated (phase ON state) (Slyn et al., 1990). Although Lrp/Papl does not appear to bind to the GATC'028 region of fully methylated pap DNA (Fig. 3), it is possible that binding to hemimethylated pap DNA might occur. Therefore, to test this hypothesis we constructed pap DNA fragments that contained a methylated top strand (orientation shown in Fig. 4) and non-methylated bottom strand (Hemi-1 DNA) as well as pap DNA contain­ing a methylated bottom strand and non-methylated top strand (Hemi-2 DNA) (see the Experimental procedures). 52 548 X. Nou et al. A B UPPER STRANO LOWER STRANO DNA Lrp Papl t>.ONPvETHYLATED + + + - + - + - METHYLATED +++ +-+ DNA Lrp Papl NON\IETHYLATED METHYLATED + + + + + + -+-+--+ + - Fig. 3. DNasel footprint analysis of fully methylated and non-methylated pap DNAs. Results obtained using the upper pap DNA strand (see Fig. 4 for orientation) are shown in (A) and results obtained using the lower DNA strand are shown in (B). Lane 0 is a G+A ladder of the upper (A) or lower (B) DNA strands. Additions of extracts containing Lrp and Papl are indicated above each lane. For lanes 3 and 8. no Papl extract was added; for lanes 5 and 10. extract from E. coli MC41 OO(pDAL288). which contains the papl gene in an orientation opposite to the T7 promoter and is phenotypically Papl-. was added. Arrows and numbers on the left of each gel show the sequence co-ordinates correspond­ing to Fig. 4. Brackets and letters on the right of each gel show protected areas. The GATC'02B site is marked by an asterisk and the GATCl130 site is marked by a triangle. o 1 2 3 4 5 6 7 8 9 10 o 1 2 3 4 5 6 7 8 9 10 We confirmed that these DNAs were hemimethylated by restriction enzyme digestion. Both the Hemi-1 and Hemi-2 DNAs were resistant to digestion with Dpnl, which cuts only fully methylated GATC sites, and Mbal, which cuts only non-methylated GATC sites but were susceptible to digestion with Sau3A, which cuts at all GATC sites regardless of methylation state (data not shown). DNasel footprint analysis of the Hemi-1 and Hemi-2 DNAs was carried out under the same conditions as the analysis of non-methylated and fully methylated DNAs shown in Fig. 3. Addition of Lrp extract to both the Hemi-1 and Hemi-2 DNAs resulted in a DNA footprint pattern similar to that observed for fully methylated DNA (Fig. 5, lanes 1,3 and 6, 8}. The GATC1130 site of fully methylated and hemimethylated DNAs is cleaved by DNasel but is protected in non-methylated DNA (Figs 3 and 5). However, unlike results obtained with fully methylated pap DNA, addition of Lrp and Pap I to Hemi-1 and Hemi-2 DNAs resulted in DNasel protection of the GATCll30 site. Thus, binding of Lrp and Papl to the Hemi-1 and Hemi-2 DNAs was different from that observed for non-methylated and fully methylated DNAs. The GATC,028 region of Hemi-1 and Hemi-2 DNAs was protected from DNasel cleavage only after addition of Lrp and Papl together (Fig. 5). The Lrp/Papl footprint extended over the same base pairs as footprints 'a' and 'b' observed 53 DNA methylation modulates pap protein-DNA interaction 549 P, transcript ~Papl ......, 829 ~T£iTCTCACT GTMCMl\.GT S M 851 TTC-TT-a-;A-A-T MTlIlIAAATC A'ffiCT-CT-CT-G- T-TATCMCG> AMG>TATTT TTATTI:TCTA TGTTTGCTTT ATTTGTTI:M TTTAGTGMT TTOCTTTT'm TTGG.IITT TAT -10 -35 N === M === 961 TTGATGTG'm TCACATTTTG CGTTTTATTT TTCTGCGAM }lGMMTI:dG TMAMTn:A TT'mGACcar CTTTTilrGCT GTAAATTCM TTTGCCATGA 'ffiTTTT'mTC MN== __ _ _ 1071 TGAGTACCCT CTTGCTATTA G'ffiTTTTGTT CT}lGTTTAAT TTTGI'TTTGT GGG'l/TAiIAl># .M.'CGTTTAM TI:MTATTTA iJAACATAMA MCTAAAT~ATTGC PeA transcript -35 -10 ..... PapS ...... 1181 'cTGMGAGTA TrTCCGG;CC GGAAGCAT/l.T ATCCAGGGOC CCGl\CAGMG GGGGMl\.CAT GGCGC 1245 M A Fig. 4. Summary of DNasel footprinting results using the pap regulatory region. The DNA sequence of the pap regulatory region is shown. The RNA polymerase-binding sites and both the transcription and translation start sites for pap/ and papB are indicated. The two 27bp inverted repeats (three mismatches) containing the GATC1028 and GA TC1130 sites (bold letters) are boxed. The bars over the DNA sequence represent areas protected from DNasel cleavage. On the left side, the letters 'N' and 'M' indicate the results obtained using non-methylated and methylated DNAs respectively. Addition of l..rp extracts, in the absence of Papl. is indicated by an open bar. A solid bar indicates that Papl was added to the Lrp extract. The pap DNA sequence used for DNasel footprinting was obtained from plasmid pDAL336B which contains DNA sequences from bp 953 (indicated by solid triangle) to bp 1244. using non-methylated DNA (Fig. 3). However, it is apparent that both hemimethylated DNAs (Hemi-1 and Hemi-2) displayed a weaker footprint in the GATC1028 region than was Observed using non-methylated DNA after addition of the same levels of Papl and Lrp extracts. Similar results were obtained after analysis of the upper DNA strand (data not shown). Thus, Lrp/Papl appears to bind to the pap GATC1028 region of hemimethylated pap DNAs with lower affinity than observed with non­methylated DNA. These results are in contrast to results obtained after addition of Lrp and Papl to fully methylated DNA in which binding to the GATC1028 region was not detectable (Fig, 3). Discussion Recently we identified two regulatory proteins, Lrp and Papl, that are involved in methylation protection of the pap GATC1028 and GATC1130 sites (Braaten et al., 1991), Here we show that Lrp binds near the GATC1130 site in the absence of Papl (Fig. 3). However, after addition of both Papl and Lrp, binding near the GATC1 028 site as well as the GA TC1130 site occurred (Fig. 3). These results are consis­tent with our previous data showing that Lrp is required for methylation protection ofthe papGATC1028 and GATC'130 sites whereas Papl is required in conjunction with Lrp for methylation protection of the GATC1028 site (Braaten et al., 1991). These data suggest the possibility that Lrp inhibits Dam methylation of the GATC1130 site and Lrp/Papl inhibit methylation of the GATC1028 site by binding near these sites and sterically blocking Dam binding. In a previous report, we presented a general model for pap phase-variation which involved competition between Dam methylase and regulatory protein(s) for binding to the GATC1028 site (Blyn et al., 1990). Here we show that binding of Lrp and Pap I to the pap regulatory region is modulated by the pap DNA methylation state. After the addition of Lrp and Papl to non-methylated pap DNAs we detected a 60bp footprint surrounding the GATC1028 site, which was not detected using fully methylated DNA (Fig. 3). We also detected binding of Lrp/Papl to the GATC1028 region of hemimethylated pap DNAs, although the foot­print was weaker than that observed using non­methylated DNA (compare Figs 3 and 5). Based on these results, the OFF methylation state would be maintained because a methylated GATC1028 site prevents binding of Papl and Lrp. Phase switching from OFF to ON could occur a