PLACE IN RETURN Box to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 6/01 cJCIRC/DatoDuepGS-p. 1 5 Regulation of Chromosomal DNA Replication in Escherichia coli: 1. Function of an N -terminal Domain in DnaA Oligomer Formation; II. Biochemical and Genetic Studies of Hyperactive dnaA Alleles By Lyle Adair Simmons A DISSERTATION Submitted to Michigan State University In partial fulfillment of the requirements For the degree of DOCTOR OF PHILOSOPHY Department of Biochemistry and Molecular Biology 2003 ABSTRACT Regulation of Chromosomal DNA Replication in Escherichia coli: I. Function of an N-terminal Domain in DnaA Oligomer Formation; II. Biochemical and Genetic studies of Hyperactive dnaA alleles By Lyle Adair Simmons The initiation of DNA replication in Escherichia coli is a tightly regulated process that results in the completion of one round of replication per cell cycle. DnaA is a 52 kDa sequence-specific DNA binding protein that is required for initiation of replication from the bacterial origin, oriC. DnaA protein has been studied intensely over the last several decades to determine the biochemical functions that are required for initiation of replication. Despite these efforts, several questions remain concerning the biochemical functions of DnaA protein that are required for replication initiation and mechanisms that might exist in the cell to limit the frequency of initiation. The N-terminal region of DnaA protein has been implicated in two biochemical activities (e.g. oligomerization and replicative helicase interaction). In this study, amino acids in the N-terminal region of DnaA protein are identified that are required for oligomerization. These results are the first to show that specific amino acids of DnaA protein are required for oligomerization and that DnaA protein oligomerization is a required activity for initiation from oriC. Because DnaA protein is responsible for initiation of DNA replication, it is expected that dnaA alleles could be obtained that are unresponsive to elements that regulate initiation. The dnaAcos allele is an example of a dnaA allele that causes hyperactive initiation and is inert to negative regulation. In this work, dnaAcos and several novel hyperactive dnaA alleles are examined genetically. We show that many activities of DnaA protein function to set the rate of origin firing during the cell cycle. This work also demonstrates that the wild type protein is negatively regulated in vivo, and that hyperactive dnaA alleles are unresponsive to these regulatory elements. The study of hyperactive initiation was extended to visualize the genome-wide distribution of replication forks by microarray analysis. Hyperactive initiation was induced and the array pattern obtained clearly indicates that replication forks are stacked near the origin, suggesting that replisomes are stalled randomly under these conditions. Cell viability is compromised and absolutely requires homologous recombination, and replication fork restart proteins (RecB and PriA respectively). This data is consistent with Holliday junction formation and Rqu cleavage at stalled replication forks. To Kristi, my wife and my inspiration iv Acknowledgments I would like to thank my thesis advisor Dr. Jon M. Kaguni. Over the years, Dr. Kaguni has provided his expertise in the areas of genetics, biochemistry and molecular biology. He has been very patient, and he has always set aside time to discuss any matter that I felt needed his counsel. He provided me with tremendous leeway, which enabled me to pursue my project without restrictions. This experience, and the environment that Dr. Kaguni promotes in his laboratory has facilitated my growth as an independent scientist. Dr. Kaguni has also been a joy to work with and a great friend. I would like to thank Dr. Laurie Kaguni for recruiting me to the department six years ago. She has also provided many excellent suggestions that have helped the progression of my work. I would like to thank the members of my thesis guidance committee, Drs. John Wang, Michele Fluck, Gregg Howe and Zachary Burton. My committee members have always shown a high level of enthusiasm during for our annual meetings which has allowed for very useful discussions and stimulated the progression of my thesis research. I would like to thank past and present members of both Kaguni laboratories: Kevin Carr, Matias Vicente, Andrew Antzcak, Matthew McNatt, Alec Murillo, Tony Ludlam, Magda Felzcak, Dr. Kasia Hupert-Kocurek , Dr. Sundari Chodavarapu, Carol Farr, Dr. Yuxun Wang, Dr. Li Fan, Dr. Anthony Lagina, Phil Kiefer, Cassandra Campbell, Dr. Yuichi Matsushima, and Ningguang Luo. I have also received help from many of the excellent graduate students that I have had the pleasure of interacting with over the years. I would like to thank my family (Lynne and Roger Caylor, Katie and Fred Simmons, Jon, Diane, and Bart Carlson, Lori and Aaron Mansfield, Kay Osborn, Lila and Adair Simmons) for their love and support over this exciting yet difficult venture. I express my highest level of gratitude to my wife Kristi, for her countless sacrifices and her unwavering love and support. Without her help none of this work would have been possible. I would also like to recognize some family members that have helped my growth over the years, but were never able to see this accomplishment: Dr. Gary Adair S i nmons, Dr. Michael Osborn, Wilmont Osborn and John Bernard Flanagan. vi Table of Contents Li st of Tables .................................................................................. xi Li st of Figures ................................................................................. xiii Li st of Abbreviations ......................................................................... xv Chapter I. Literature Review ................................................................ 1 Introduction .......................................................................... 2 The Cell Cycle ...................................................................... 2 The origin of replication ........................................................... 5 DnaA binding to oriC ..................................................... 8 Other oriC binding proteins .............................................. 9 In vitro replication assay .................................................. 1 1 DnaA protein ........................................................................ 1 l DnaA-ATP binding ........................................................ 20 DnaA-RepA ................................................................ 21 Anionic phospholipid regulation of DnaA protein .................... 22 DnaA C-terminus .......................................................... 23 Structure of DnaA (Domains IHa-IV) protein from Aquifex 23 aeolicus ..................................................................... Molecular chaperones and the activation of DnaA .................... 26 Hyperactive initiation of DNA replication ...................................... 27 DnaAcos 28 Alternative modes of chromosomal replication in Escherichia coli ......... 31 Replication fork collapse and replication fork restart .......................... 33 Similarities between prokaryotic and eukaryotic intiation of DNA 34 replication ........................................................................... References ........................................................................... 39 Chapter II. DnaA protein of E. coli: initiation of DNA replication requires DnaA Oli gomerization and specific N-terminal amino acids .................................... 53 Abstract .............................................................................. 54 Introduction .......................................................................... 55 Experimental Procedures .......................................................... 59 Strains and Plasmids ...................................................... 59 Bacterial transformation .................................................. 59 Site-directed mutagenesis ................................................ 59 Glutaraldehyde crosslinking ............................................. 60 Results .............................................................................. 62 vii Intragenic complementation as an in vivo assay for DnaA oligomerization ............................................................ Immunoblot analysis indicates that the mutant proteins are expressed at levels sufficient to support initiation ..................... Glutaraldehyde crosslinking of DnaA .................................. Discussion ........................................................................... Oligomer formation of DnaA ............................................ T435K and suppression of dnaX(Ts) .................................... Model for DnaA at oriC ................................................... Acknowledgements ................................................................. References ........................................................................... Chapter H1. The dnaAcos allele of E. coli: Hyperactive initiation is due to substitution of A184V and Y271H, resulting in defective ATP binding and aberrant DNA replication control ........................................................... Abstract .............................................................................. Introduction .......................................................................... Experimental Procedures .......................................................... Bacteriological methods .................................................. Separation of dnaAcos mutations ....................................... Quantitative PCR assay ................................................... Results ................................................................................ dnaAcos is dominant-negative to dnaA“ ................................ A184V and Y2711H are required for the overininitiation phenotype of dnaAcos ..................................................... The dominant-negative phenotype requires oriC ..................... DnaAcos and the A184V+Y271H double mutant results ............ In vivo initiation activity as measured by oriC plasmid maintenance Discussion ........................................................................... Hyperactive initiation of the A184V+Y271H mutant ................ References ........................................................................... Chapter IV. Novel dnaA alleles are defective for regulation of initiation from the E. coli replication origin oriC ................................................................ Abstract .............................................................................. Introduction .......................................................................... Experimental Procedures .......................................................... viii 62 67 69 74 74 76 77 79 80 85 86 87 93 93 93 94 98 98 99 101 101 106 114 114 117 122 123 124 131 Bacterial strains and plasmids ............................................ 131 Quantitative PCR .......................................................... 131 Mutant protein overexpression in RTS ................................. 132 In vitro replication assays ................................................ 132 Immunoblot analysis ...................................................... 133 Homology model .......................................................... 133 Results ................................................................................ 135 Isolation of lethal dnaA alleles ........................................... 135 Steady-state levels of protein expression ............................... 140 Lethality requires the presence of functional oriC .................... 140 Novel alleles are dominant negative to dnaA” ......................... 146 dnaA alleles are active for oriC plasmid replication ................... 146 Marker frequency analysis of hyperactive initiation in vivo ......... 147 Mutant DnaA proteins are active for replication from oriC in 149 vitro .......................................................................... Location of missense mutations .......................................... 156 Discussion ........................................................................... 160 Genetic screen and characterization of novel dnaA alleles ........... 160 Hyperactive initiation in viva ............................................. 161 Hyperactive initiation in vitro ............................................ 162 Cdc6 protein and the frequency of initiation ........................... 163 What is the cause of lethality? ........................................... 164 References ........................................................................... 1 66 Chapter V. Hyperinitiation of DNA replication in E. coli leads to replication fork collapse and inviability ....................................................................... 170 Abstract .............................................................................. 171 Introduction .......................................................................... l 72 Experimental procedures .......................................................... 176 Strains and plasmids ...................................................... 176 Preparation of ectopic ter-blocked chromosomal DNA .............. 176 Preparation of chromosomal DNA from cultures induced to express DnaA+ protein ..................................................... 177 Quantitative PCR analysis ................................................ 177 Microarray procedures and data analysis .............................. 178 Pulse field gel electrophoresis ........................................... 178 Results ................................................................................ 180 Do replication forks stall at specific sites? ............................. 180 Hyperactive initiation interferes with viability in the absence of recombinational repair .................................................... 190 Discussion ........................................................................... 198 Replication forks do not appear to stall at specific sites .............. 198 Hyperinitiation induces replication fork collapse ..................... 199 Lethality of dnaAcos ...................................................... 200 The replication restart pathway .......................................... 201 Hyperinitiation in eukaryotes ............................................. 201 Acknowledgements ................................................................. 203 References ........................................................................... 204 Chapter VI. Summary and Perspectives ................................................... 209 References ........................................................................... 2 14 List of Tables Chapter I Table 1. Summary of mutations cloned or isolated classically in the dnaA gene ............................................................... Table 2. Hyperactive dnaA alleles ...................................... Chapter II Table 1. List of strains .................................................... Table 2. L5A, W6A, and C9Y alleles are defective in allelic complementation when combined with dnaA46 ....................... Table 3. Allelic complementation by dnaA46 and T435K in oriC plasmid maintenance demonstrates the formation of mixed complexes .................................................................. Table 4. The T435K allele is defective in maintenance of an oriC plasmid ..................................................................... Table 5. Steady state levels of mutant DnaA proteins ............... Chapter III Table 1. List of strains .................................................... Table 2. List of plasmids ................................................. Table 3. Transformation of W3110 (dnaA", oriC’) with plasmids carrying various dnaA mutations ........................................ Table 4. Transformation of L8] 10 (dnaA‘fl AoriC) with plasmids carrying various dnaA mutations ........................................ xi 12 30 61 63 65 66 68 90 91 100 102 Table 5. Marker frequency analysis of oriC and relE by quantitative PCR ........................................................... Table 6. Function of dnaAcos mutations in oriC plasmid maintenance ................................................................ Chapter IV Table 1. List of plasmids ................................................. Table 2. Plating efficiencies of plasmid-home dnaA alleles in LS 1073 (dnaA46) .......................................................... Table 3. Nucleotide and amino acid substitution in novel dnaA alleles ........................................................................ Table 4. Novel dnaA alleles are lethal when overexpressed in an oriC’ strain .................................................................. Table 5. Dominant-negative effect of dnaA alleles when expression is elevated in a dnaA+ host strain ........................... Table 6. Novel dnaA alleles are active for oriC plasmid replication ................................................................... Chapter V Table 1. Strains and plasmids ............................................ Table 2. Analysis by microarray and quantitative PCR of oriC or adjacent loci and relE ..................................................... Table 3. Replication fork restart genes priA and recB are required for viability under conditions of hyperactive initiation ............... Table 4. Double stranded breaks accumulate in recB strains due to hyperactive initiation ................................................... xii 107 108 127 134 137 139 144 145 175 188 189 195 List of Figures Chapter I Figure 1. Localization of chromosomal DNA sequences during different periods of the cell cycle ........................................ 4 Figure 2. Structural organization of oriC .............................. 7 Figure 3. Schematic representation of DnaA functional domains ...................................................................... 18 Figure 4. Crystal structure of DnaA protein from Aquifex aeolicus ..................................................................... 25 Figure 5. Structural comparison of bacterial and archaeal replication proteins ........................................................ 37 Chapter 11 Figure 1. Glutaraldehyde crosslinking of DnaA confirms that an N-terminal region is required for oligomerization ..................... 71 Chapter 111 Figure 1. Functional domains of DnaA protein relative to mutations of dnaAcos ...................................................... 97 Figure 2. PCR amplification of oriC and relE show that amplification is proportional to the amount of bacterial chromosomal DNA used as the template and is specific for these loci ........................................................................... 105 Figure 3. Relative stabilities of mutant DnaA proteins ............... 1 1 1 xiii Chapter IV Figure 1. Genetic method to isolate hyperactive dnaA alleles ...... 129 Figure 2. Relative expression levels of dnaA alleles ................. 142 Figure 3. Mutant DnaA proteins are active for oriC plasmid replication in vitro ......................................................... 151 Figure 4. Continued DNA replication activity with hyperactive mutant proteins ............................................................. 155 Figure 5. Homology model of E. coli DnaA protein (Domains IIIa-IV) ......................................................... 158 Chapter V Figure 1. Elevated dnaA expression induces hyperinitiation ........ 183 Figure 2. Map of the E. coli chromosome ............................. 186 Figure 3. Pulse field analysis of chromosomal DNA after induced expression of wild type or mutant DnaA protein ............ 192 xiv bp Dam DnaA box DTI‘ EDTA HEPES kDa SDS Tris l 3-mer initiation E. coli B. subtilis List of Abbreviations base pair(s) deoxyadenosine methyltransferase asymmetric 9-mer consensus sequence bound by DnaA dithiothreitol (ethylenedinitrilo)tetraacetic acid N-2-Hydroxyethylpiperazine-N’-2-ethanesulfonic acid kilodalton Sodium dodecyl sulfate Tris (hyroxymethyl)aminomethane a sequence in oriC that is A+T rich and is unwound during Escherichia coli Bacillus subtilis XV Chapter I Literature Review Introduction The process of genome duplication is a fundamental pathway that is required for the perpetuation of all living organisms. One of the model organisms used to study the mechanisms that regulate DNA replication is the bacterium Escherichia coli. This organism contains a single circular duplex DNA molecule of 4,700 kilobase pairs (von Meyenburg and Hansen 1987). The genome contains a genetic locus, oriC, where replication begins and replication forks proceed bidirectionally (Prescott and Kuempel 1972). After replication has proceeded 180° around the chromosome, replication is terminated, the chromosomes are segregated, and cell division ensues, producing two daughter cells each with a single copy of the genetic material. The cell cycle The E. coli cell cycle is composed of two distinct parts, the C period and the D period. The C period represents the time required for events including initiation and termination of DNA replication. The D period represents the time required to complete cell division (Cooper and Helmstetter 1968). A total of 60 minutes is required for the completion of the bacterial cell cycle (40 minutes C period and 20 minutes D period) after its onset. It has certainly been noticed that rapidly growing cultures are able to double in less than 60 minutes. To allow for doubling intervals that are less than 60 minutes, new rounds of replication must be initiated before the first round of replication has been completed. Analysis of DNA content using flow cytometry has revealed that E. coli growth is synchronous, resulting in the number of genomes to be 2n (n2 1 and an integer). This observation indicates that initiation is tightly regulated, even in rapidly growing cultures (Boye, Lobner-Olesen et a1. 1988; Skarstad, von Meyenburg et a1. 1988). Analysis of Figure 1. Localization of chromosomal DNA sequences during different periods of the cell cycle. (adapted from Niki and Hiraga, 1998) This diagram represents a cell growing with a doubling time of 60 minutes. The circle (0) located at the left nucleoid border represents oriC. The circle located (0) at the right nucleoid border represents the terminus. Early in replication oriC is duplicated and one segment begins migrating toward the right nucleoid border while the other segment remains. The terminus migrates to the midcell after one segment of oriC has reached the other nucleoid border. After the terminus is duplicated the chromosomes are decatenated and a septum forms at the midcell. This results in the formation of two daughter cells each with oriC and the terminus located at the nucleoid borders. A Model for the migration of oriC and the terminus to their respective cellular positions has been proposed. A putative cis-acting centromere sequence may be located with oriC in a loop of the nucleoid. This loop is pulled to the other nucleoid border after oriC has been duplicated. This diagram indicates that the bacterial nucleoid has a specific arrangement with in E. coli. This result also suggests that the subcellular localization of oriC and the replication terminus are important functionally during the cell cycle. temperature sensitive E.coli strains have revealed initiation to be asynchronous at the permissive temperature, and at the restrictive temperature DNA replication has failed to initiate (Skarstad, von Meyenburg et a1. 1988). These studies were the first to demonstrate that mutations could result in aberrant replication control. More recently, the use of fluorescently labeled proteins, and the hybridization of fiuorescently labeled DNA probes have indicated that the bacterial nucleoid is well organized (Figure 1). These results support a model for origin migration from one nucleoid border to the other subsequently after duplication of oriC (Niki, Ichinose et al. 1988; Niki, Jaffe et al. 1991; Niki and Hiraga 1997; Webb, Teleman et al. 1997; Niki and Hiraga 1998; Hiraga 2000; Hiraga, Ichinose et al. 2000). These studies not only indicate tight regulation of replication but they also verify a specific arrangement of replication proteins and the bacterial nucleoid during the cell cycle. The origin of replication The E.coli origin of replication is the site for initiation of bidirectional replication (Prescott and Kuempel 1972), located at 84.3 minutes on the chromosomal linkage map (von Meyenburg, Hansen et al. 1978). The minimal sequence required to confer replication of a plasmid is 245 bps (Meijer, Beck et al. 1979; Oka, Sugimoto et a]. 1980). The oriC sequence is highly conserved indicating structural significance (Zyskind, Cleary et al. 1983), and oriC contains several motifs (Figure 2) common among enteric bacteria (Buhk and Messer 1983). The origin contains five 9—mer sequences that serve as binding sites for the replicative initiation protein, DnaA (discussed in more detail below) (Fuller, Funnell et al. 1984). The origin also contains other binding sites for proteins involved in activation or inhibition of initiation (Craig and Nash 1984; Torheim and Skarstad 1999). There are three A+T rich 13-mers [also known as the duplex unwinding element (DUE)] Figure 2. Structural organization of oriC The origin of replication (Matsui et. a], 1985; Gille and Messer, 1991; Kano et. a1, 1991) oriC is located between genes gidA and mioC with the direction of transcription indicated by the black arrows. The oriC region contains five DnaA protein binding sites (Rl-R4, M). The directionality of these DnaA boxes are indicated. The origin also contains binding sites for IHF and Fis proteins ,which are important for the initiation of DNA replication. The left end of oriC contains three A+T rich 13-mer regions which are denatured during processing of the origin by DnaA protein, to form the open complex. The 13—mers are represented by circles and as labeled (L,M,R). 6.8 in; use av 0mm oom oop AmVMEVAAVT _ _a_a% 359 5;. :4. located at the left border of oriC. The DUE constitutes a cis-acting element (Kowalski and Eddy 1989) subsequently denatured during initiation to form the open-complex. The open-complex then serves as the site for assembly of the replication fork machinery (Bramhill and Kornberg 1988; Bramhill and Kornberg 1988). The origin also contains the highest number of GATC sequences for a DNA segment of this size in the E. coli genome. The GATC sequence is repeated nine times in oriC (Zyskind, Harding et a1. 1981). This sequence is the recognition site for Dam methyltransferase, which methylates the adenine base. Methylated oriC can be specifically bound by proteins that regulate initiation. DnaA binding to oriC Five 9-mer sequences (R1, R2, R3 R4, and M) which contain a consensus sequence TTATCCACA (or slight variation) are specifically bound by DnaA protein when DnaA is complexed with ATP (DnaA-ATP) (Fuller and Kornberg 1983; Fuller, Funnel] et al. 1984) The binding of DnaA to oriC has been analyzed using gel shift mobility assays, DNase I foot-printing and surface plasmon resonance (Matsui, Oka et al. 1985; Roth and Messer 1995; Speck and Messer 2001). Comprehensive mutagenesis of the DnaA boxes indicates that all five are required for DNA replication, although box R3 appears to be dispensable under certain conditions (Langer, Richter et al. 1996). For most DnaA boxes, inversion of the sequence, point mutations or position exchange causes a phenotype indicating the important architecture of this region (Langer, Richter et al. 1996). It has been suggested that DnaA-ATP has an affinity for single stranded DNA and the ATP-bound form of DnaA can bind directly to the A-T rich l3-mer region of oriC. This activity aids in the unwinding of the origin and allows for sequential assembly of essential replication proteins (Speck and Messer 2001). Other oriC binding proteins In addition to specific binding of DnaA to oriC, many other essential proteins bind to the origin and regulate this process. Integration host factor (IHF) was originally identified for its role in lambda phage integration into the E. coli genome (Katayama, Takata et al. 1997). Analysis of IHF binding to a specific site within oriC revealed IHF assists in initiation (Hwang and Kornberg 1992). IHF is known to bend DNA upon binding (Moitoso de Vargas, Kim et al. 1989). It is thought that this topological change in the DNA is required for initiation (Hwang and Kornberg 1992). Factor for inversion stimulation (Fis) is another accessory protein that binds to oriC. Like IHF, Fis also assists in initiation by inducing topological distortion of the DNA upon binding (Gille, Egan et al. 1991; Gille and Messer 1991). Fis protein binds in close proximity to DnaA box R3, although the binding is not mutually exclusive to DnaA protein binding to R3 (Margulies and Kaguni 1998). Biochemical data indicates that high concentrations of Fis protein inhibit replication from an oriC plasmid. This result reveals that a limited amount of Fis is required for initiation (Hiasa and Marians 1994). Both IHF and Fis, under physiological conditions, are oriC accessory binding proteins that contribute to initiation. Other oriC binding proteins have been isolated which negatively regulate the initiation process. Inhibition of chromosomal initiation (IciA) protein was identified as a protein that specifically binds the 13-mer region (DUE) of oriC (Hwang and Kornberg 1990; Hwang and Kornberg 1992; Hwang, Thony et al. 1992). This 33 KDa protein has the ability to block open-complex formation leads to a decrease in the formation of ssDNA in P1 endonuclease sensitivity assays correlates with increasing amounts of IciA in vitro (Katayama, T akata et al. 1997). Squ protein is known for its ability to sequester oriC to the E .coli membrane to stop untimely initiation. Squ protein can also bind to the 13-mer region and this activity prevents open-complex formation (Brendler, Abeles et al. 1995). Plasmids which depend solely on oriC as the site where replication is initiated are referred to as minichromosomes. Minichromosomes have served as useful models for the study of replication both in vitro and in vivo. In vivo replication of these plasmids is bidirectional (Meijer and Messer 1980) and replication is coordinated with the cell cycle (Leonard and Helmstetter 1986; Koppes and von Meyenburg 1987), highlighting that its regulation is similar to that of the chromosomal origin. Minichromosomes are randomly distributed at cell division unless a partition sequence is present (Ogura and Hiraga 1983). When the copy number reaches a critical elevation, these plasmids have been observed to integrate into the chromosome suggesting some level of incompatibility with chromosomal oriC (Lobner—Olesen 1999). In B. subtilis, the incompatibility between chromosomal oriC and plasmid oriC is more severe. Minichromosomes carrying B. subtilis oriC are maintained for short periods of time with a copy number of one (Moriya, Fukuoka et al. 1988; Moriya, Atlung et al. 1992). Differences between chromosomal replication and minichromosomal replication have been described. Strains carrying HU and IHF null mutations do survive, although they are unable to maintain minichromosomes, indicating differences in accessory protein requirements between the two systems (Kano, Ogawa et al. 1991). With these differences in mind, minichromosomes have been widely used to distinguish the order of events that take place during initiation in vitro and have been successful in determining protein requirements in vivo. 10 In vitro replication assay The replication process has been efficiently studied using in vitro replication assays dependent on either a crude extract or the reconstitution of the replication machinery with purified components (Fuller, Kaguni et al. 1981; Kaguni and Kornberg 1984). Both in vitro replication assays require exogenously supplied supercoiled oriC plasmid (Fuller, Kaguni et al. 1981; Kaguni and Kornberg 1984). Replication of minichromosomes proceeds bidirectionally which has also been observed during replication from the chromosomal origin [(Kaguni and Kaguni 1992) for review; (Prescott and Kuempel 1972)]. The reconstituted in vitro assay requires many purified proteins to carry out three distinct stages of replication (initiation, replication fork progression, and suppression of nonspecific initiation) and all of these stages are required for propagation of the minichromosome in vitro (Kaguni, Fuller et al. 1982; Kaguni, Bertsch et al. 1985). One of the more effective ways to use the in vitro assays is to measure the replication activity of mutant DnaA proteins (Hwang and Kaguni 1991; Hupp and Kaguni 1993; Hupp and Kaguni 1993). This assay has also been used to assess the effect of mutations present in other replication proteins (DnaC) that may compromise activity (Ludlam, McNatt et al. 2001). DnaA protein DnaA is a 52.5—kDa protein that is required for DNA replication in E. coli (Skarstad, von Meyenburg et al. 1988). DnaA is also required for initiation of replication from many plasmid origins (pSC 101, F, R6K). Several plasmids contain one or more DnaA boxes but do not require DnaA protein for replication (Molin, Diaz et al. 1980; Molin and Nordstrom 1980; Fuller, Funnell et al. 1984; Hansen and Yarmolinsky 1986; Kline, Kogoma et al. 1986; Wu, Goldberg et al. 1992). 11 Table 1. Summary of dnaA mutations cloned or isolated by classical genetics Mutation (dnaA Replication Biochemical or Reference allele) phenotype genetic mechanism Domain I (1-90) M] V Defective in vivo Unknown Sutton and Kaguni, 1997 L35, L1 OS, L1 75 Defective in vitro and Failure of DnaA- Mirna et. al, (dnaA4I3) in vivo DnaA interaction 2002 C9Y Defective in vivo Unknown Sutton and Kaguni, 1997 L1 75 (dnaA427) Defective in vitro and Failure of DnaA- Mirna et. al, in vivo DnaA interaction 2002 E21 K Defective in vivo Unknown Sutton and Kaguni, 1997 W25S Defective in vivo Unknown Sutton and Kaguni, 1997 1265 Defective in vitro Failure to unwind Mirna et.al, oriC 1999 A31T Defective in pSC101, Unknown Sutton and active in oriC in vivo Kaguni, 1995 MOS Defective in vitro Failure to unwind Mima etal, oriC 1999 A62 Defective in vitro and Failure to retain Sutton et. al, in vivo DnaB in 1999 prepriming complex Domain 11 (91-130) A129 Defective in vitro and Failure to interact Sutton et. al, in vivo with DnaB 1999 Domain IIIa (131-296) V157E (dnaA I 67) Defective 42°C in Unknown Hansen et. al, vivo 1992 GI 75D Defective in vivo Poor ATP-binding Sutton and Kaguni, 1997 A184V, H252Y Defective, 42°C in Poor ATP-binding Carr and (dud/146) vitro 12 Kaguni, 1996 Table 1. Continued Domain Hla (131-296) A184V, P296Q Defective 42°C in viva Unknown Hansen, et. al, (dnaA601/602) 1992 A184V, A347V Defective, 42°C in viva Unknown Hansen et. al, (dnaA604/60® 1992 V198M Defective in pSClOl, Unknown Sutton and and oriC in viva Kaguni, 1995 M21 0L Defective(cold Unknown Guo et. al, sensitive) in viva 1999 A213D Cold Sensitive Suppressor of Gines- (dnaA 73) defective dnaX(ts) Candelaria 1219N Defective(cold Unknown Guo et. al, sensitive) in viva 1999 F222L Defective(cold Unknown Guo et. al, sensitive) in viva 1999 DZ35N Non-complementing Poor ATP-binding Mizushima et, and ariC unwinding al, 1998 F240] Defective (cold Unknown Guo et. al, sensitive) in viva 1999 R245Q Defective(cold Unknown Guo et. al, sensitive) in viva 1999 725311 Defective (cold Unknown Guo et. al, sensitive) in viva 1999 E279V Defective (cold Unknown Guo et. al, sensitive) in viva 1999 62875 Defective in pSC101, Unknown Sutton and active in oriC in viva Kaguni, 1995 and In vitro Domain IIIb (297-347) T3011 Defective in pSClOl , Unknown Sutton and active in ariC in viva and in vitro 13 Kaguni, 1995 Table 1. Continued R302C Defective (cold Unknown Guo et. al, sensitive) in viva 1999 R334H Defective (cold Unknown Guo et. al, sensitive) in viva 1999 R342H Defective (cold Unknown Guo et. al, sensitive) in viva 1999 L340P Decreased activity in Decreased DNA- Blaesing et. vitro binding activity al, 2000 Domain IV (348-467) E3616 Defective (cold Unknown Guo et. al, sensitive) in viva 1999 L363K, L3 73R Suppresses acidic Unable to interact Zheng et. al, phospholipid deficient with phospholipids 2001 growth arrest in viva L366K Suppresses acidic Unable to interact Zheng et. al, phospholipid deficient with phospholipids 2001 growth arrest in viva K3725 Decreased activity in Decrease in Yamaguchi vitro Cardiolipin et. al, 1999 interaction V383M Defective in viva Unknown Sutton and Kaguni, 1997 E385K Defective (cold Unknown Guo et. al, sensitive) in viva 1999 I389N (dnaA203/204) Defective 42°C in viva Unknown Hansen et. al, 1992) S396F Defective in vitro DNA-binding Blaesing defect et.al, 2000 K397E Defective (cold Unknown Guo et. al, sensitive) in viva 1999 R401A Defective in vitro DNA-binding Blaesing et. defect al, 2000 14 Table 1. Continued Domain IV (348-467) Continued V403A Defective in vitro DNA-binding Blaesing et. defect al, 2000 R407C Decease in activity in DNA-binding Blaesing et. vitro defect al, 2000 R407A Defective in vitro DNA-binding Blaesing et. defect al, 2000 R407H Defective in viva Unknown Sutton and Kaguni, 1997 Q408R Defective in vitro DNA-binding Blaesing et. defect al, 2000 K415A Defective in vitro DNA-binding Blaesing et. defect al, 2000 K4] 5E Defective in vitro DNA-binding Blaesing er. defect al, 2000 L417P Decrease in activity in DNA-binding Blaesing et. vitro defect al, 2000 S421N Decrease in activity in DNA-binding Blaesing er. vitro defect al, 2000 S4210 Decease in activity in DNA-binding Bleasing et. vitra defect al, 2000 T426A Defective in vitro DNA-binding Blaesing et. defect al, 2000 04260 Defective in viva Unknown Sutton and Kaguni, 1997 R432L (dnaA 721) dnaX(Ts) suppressor Unknown Gines- in viva Candelaria et. al, 1995 H434A Decease in activity in DNA-binding Blaesing et. vitra defect al, 2000 H434Y Decrease in activity in DNA-binding Blaesing er. vitro defect al, 2000 T435A Defective(cold Unknown Guo et. al, sensitive) in viva 1999 15 Table 1. Continued Domain IV (348-467) Continued T435K dnaX(Ts) Unknown Gines- suppressor in viva Candelaria et. al, 1995; Blaesing et. al, 2000 T435M Defective in viva DNA-binding Sutton and defect Kaguni, 1997 T435P Defective in vitro DNA-binding Blaesing et. al, defect 2000 T436A Defective in vitro DNA-binding Blaesing et. al, defect 2000 V437M Defective in viva Unknown Sutton and Kaguni, 1997 A440T Defective in viva DNA-binding Sutton and defect Kaguni, 1997 A440V Defective in viva DNA-binding Sutton and defect Kaguni, 1997; Blaesing et. al, 2000 K443E Defective in vitro DNA-binding Blaesing et. al, defect 2000 L44 7W Defective in Unknown Sutton and ‘ pSC 101, active in oriC in viva, and in vitro Kaguni, 1995 The information in this table is compiled from the work of many laboratories. De letion mutants were omitted from this table unless the study of the mutant protein or a] lele was essential for understanding the role of the domain effected by the deletion. Alleles with multiple mutations were also omitted unless the allele has been well studied and the resulting phenotype clearly indicated. Alleles of the dnaA gene that are hyperactive or cause a phenotype that correlates with hyperactive initiation are listed in Table 2. 16 Figure 3. Schematic representation of DnaA functional domains A genetic assay was developed to select for mutations that inactivated the dnaA gene (Sutton and Kaguni, 1997). The missense mutations obtained (triangles) were localized to distinct regions of the dnaA gene. These mutant proteins were assayed to determine the biochemical defect resulting in a failure of replication activity. The defects observed allowed for the identification of the functional domains described in this diagram. The N-terminal region (Domains I and II) of DnaA protein is important for both interactions with the replicative helicase DnaB and for self interaction (oligomerization). Domain 1113 contains a Rossman fold (Walker A box) and a Sensor I motif. These elements are important for ATP binding. Domain IIIb contains the Walker B box and the Sensor II motifs (Sensor I and II motifs are shared among AAA+ family members) which are important for ATP binding and hydrolysis. Domain IV of DnaA protein contains a region which is required for DNA binding activity. 17 ucmzm Sam qufiSn. A x__e_._ 22¢. 8669a £5“. 20:33.2 3:825. ‘ wEmEoo Ecoficzu m... >_ :ano . 2: 59:0 _ _ Q 6:. 59:00 95:5 _ __ EmEoo _ c5800 mcmEEmE _ _ . 969m <20 86“. cememmog 83555095 . . mcfiE ., ._. . __ a _ 58mm . .m 965m mac 3380 .Ifl . . _. . PE Emficmfi o mMmEtumi 9: E Now Be Clllltll tillllllt observed 111. The 8m 613111011 lgfllllll 1b 111’ 5011/ 18 More than 100 mutations in DnaA protein have been isolated which span the entire protein [many are listed in Table 1 (Hansen, Koefoed et al. 1992; Mima, Yamagachi et al. 1 999; Yamaguchi, Hase et al. 1999; Mirna, Makise et al. 2002)]. Many of these alleles have been assayed in viva or the mutant protein has been assayed in vitro and a phenotype or biochemical defect has been described (Kellenberger-Gujer, Podhajska et a1 - 1978; Sutton and Kaguni 1995; Sutton and Kaguni 1997; Guo, Katayama et al. 1999; B laesing, Weigel et al. 2000). The phenotype caused by most of these mutations results from inactivation of DnaA protein, although a few mutations have been described (Table 2) that activate function [for review (Katayama, Fujimitsu et al. 2001)]. The study of these mutant proteins has been pursued to elucidate the functional domains of DnaA protein, and determine how DnaA protein is regulated (Figure 3). Two biochemical functions have been ascribed to the N-terminal region of DnaA. This region (also referred to as Domain I and II) has been shown to be important for DnaA self-interaction and interaction with the replicative helicase, DnaB. The first 86 amino acids of DnaA, can be used to construct a 7th repressor chimera which dimerizes and retains repression activity. Solid phase protein-binding assays have also been used to demonstrate DnaA-DnaA interaction (Weigel, Schmidt et a1. 1999). Two different portions of the DnaA N-terminal region are responsible for interaction with DnaB, and these two interactions regulate separate steps of prepriming complex formation (Sutton, Carr et al. 1998; Seitz, Weigel et al. 2000). Residues 1-110 at e important for transient interaction (or recruiting) with DnaB resulting in DnaB binding to the 13-mer region of ariC to form the prepriming complex. Residues 130-168 are responsible for retention of DnaB in the prepriming complex (Sutton, Carr et al. 1998). Based on a few studies that have been designed to define the area of DnaA-DnaA and DnaA-DnaB protein-protein interaction, it appears that the functional domains 19 overlap, and specific amino acid residues have not been identified to be important or required for either interaction. DnaA-A TP binding DnaA protein binds ATP (Domain 111a) and this activity is required for activation (Sekimizu, Bramhill et al. 1987; Sekimizu, Bramhill et al. 1988). DnaA-ATP is required for open-complex formation at ariC. DnaA-ADP is unable to form the open-complex; however, DnaA-ATPyS does facilitate oriC unwinding, suggesting that DnaA-ATP provides a conformational change that is required for initiation, but hydrolysis of ATP is not necessary (Sekimizu, Bramhill et al. 1987). Two dnaA alleles which contain the A 1 84V (dnaA46 and dnaA5) missense mutation have reduced ATP binding activity and are temperature sensitive for initiation (Hwang and Kaguni 1991; Hupp and Kaguni 1 993; Hupp and Kaguni 1993; Hupp, Meek et al. 1993). This mutation is solely responsible for the thermolability of these proteins (Carr and Kaguni 1996). The A184V mi ssense mutation is located near a P-loop motif (Walker A box, GXXGXGKT) which is Common among many nucleotide binding proteins. The Walker A box is responsible for interactions with the triphosphate moiety, whereas the Walker B box is important for Chelating the Mg2+ ion (Walker, Saraste et al. 1982; Saraste, Sibbald et al. 1990). Site- directed mutagenesis of the ATP binding region of DnaA protein has shown that Lysl78 and Asp 235 are critical for ATP binding and activation of DnaA protein for open- cornplex formation (Mizushima, Takaki et a1. 1998). Immuno-precipitation of DnaA Protein from soluble extracts during the C and D period of the cell cycle has revealed that the highest level of DnaA-ATP is observed just before the onset of replication. The level 0f IDnaA-ATP drops rapidly after replication is initiated (Kurokawa, Mizushima et al. 1998). This data indicates that the cellular level of DnaA-ATP is closely tied to the state 0f l‘fitplication as the abundance of DnaA-ATP coincides with the cell cycle. It has also 20 been shown that DnaA-ATP has a higher affinity for certain DNA sequences. A 6-mer consensus sequence (AGATCT) has been identified that is only recognized by the ATP bound form of DnaA protein. One of these DnaA-ATP boxes (6-mer) is located in the promoter region of the dnaA gene adjacent to the more well characterized 9-mer DnaA box that is bound by DnaA protein to repress transcription of the dnaA gene (Braun, O'Day et a1. 1985; Lee and Hwang 1997). DNase I protection patterns clearly indicate that DnaA-ATP protects this region more effectively than DnaA-ADP resulting in increased repression of dnaA gene expression (Speck, Weigel et al. 1999). Interestingly, DnaA-ATP is only able to occupy the 6-mer binding site when it is adjacent to a 9—mer site suggesting coopertivity in DnaA-ATP binding to the 6-mer sequence. DnaA-RepA (pSC101) interaction Novel alleles of the dnaA gene have been isolated that are defective for pSC101 maintenance but active for replication from oriC (Sutton and Kaguni 1995). These mutations are located in the region of DnaA protein contained by residues 237-360 (Domain IIIa-b). It has been suggested that this region may be important for protein- protein contact with the pSC101 initiation protein RepA, although this interaction has not been demonstrated directly. Forward and reverse yeast-two hybrid assays have shown interaction between DnaA and RepA utilizes the N-terminus and the C-terminus of DnaA Protein (Shanna, Kachroo et al. 2001). Seitz and coworkers confirmed this observation by Constructing and internal deletion of amino acid residues 137-376 and determined that the mutant protein still functioned in pSC101 maintenance. Based on the data described above, it seems likely that the novel dnaA alleles (V198M, G287S, T3011) are unable to assel'nble a functional initiation complex at the pSC101 origin due to a defect in some recIllired activity other than interaction with RepA protein. 21 Anionic phospholipid regulation of DnaA protein Disruption of the pgsA coding region (encoding phosphatidyl glycerol phosphate synthase catalyzing the step of commitment of phosphatidyl glycerol biosynthesis) causes a decrease in the levels of anionic phospholipids in the E. coli cell membrane, which leads to growth arrest (Heacock and Dowhan 1987; Heacock and Dowhan 1989; Xia and Dowhan 1995). It was observed that activation of DNA replication through a DnaA independent pathway (constitutively stable DNA replication, described later) suppressed the growth arrest previously observed (Xia and Dowhan 1995). It was demonstrated later that DnaA protein interacts physically with anionic phospholipids (Garner and Crooke 1996) and this interaction is important for the reactivation of DnaA protein by stimulating the release of ADP allowing DnaA protein to bind ATP to activate it for initiation of replication (Garner and Crooke 1996). Deletion of this region (360-373) and (or) single amino acid substitutions have been shown to suppress growth arrest of E. coli strains deficient in anionic phospholipid synthesis (Zheng, Li et al. 2001). These results are interesting because they describe a direct role for phospholipids in the initiation of DNA replication. Recent results however, place this Observation under scrutiny (see below). A careful analysis of pgsA null mutants demonstrated that the cells are viable When DNA replication proceeds normally through oriC and is dnaA dependent (Kikuchi, Shibuya et al. 2000). The viability of this strain requires inactivation of the lpp gene (encoding lipoprotein, see below). It was observed that the pgsA null mutants were able to grow in rich media, but were not viable in minimal media or temperatures over 40°C. Other experiments revealed that in a pgsA null strain there is an accumulation of prolipoprotein in the periplasm. One of the important steps in prolipoprotein maturity is a Cliacylglycerol group donated from phosphatidyl glycerol ( a product of the pgsA pathvvay of acidic phospholipid biosynthesis). Apparently, the unmodified prolipoprotein 22 accumulates in the periplasm and is able to make covalent linkages with the outer membrane resulting in a disruption of the cell surface. This has now been attributed to the growth defects that are caused in pgsA null mutants, and confirms the dependence on deletion of the lpp gene. The role of acidic phospholipids in DNA replication is rather unclear at this time. Further experiments will be required to strengthen or weaken the role of DnaA protein and membrane attachment in the regulation of initiation. DnaA C -terminus DnaA protein binds specifically to its cognitive 9—mer sequences. This interaction takes place in the C—terminal domain (also known as Domain IV) residues 387-467 (Sutton and Kaguni 1997). The sequence specific binding of DnaA to oriC is conferred by threonine residue 435 (Sutton and Kaguni 1997). The 94 amino acid DNA binding domain has the capability to bind DNA in viva, random mutations throughout the 94 amino acid region cause a decrease or abolish DNA binding activity (Blaesing, Weigel et al - 2000). To date no known DNA binding domain mutations cause an increase in the binding affinity of DnaA protein to oriC or other DNA sequences. Structure ofDnaA (Domains Illa-IV) protein from Aquifex aeolicus Several labs have tried to crystalize DnaA protein with the goal of elucidating the folded of the macromolecule. These failed efforts included the purification of DnaA Protein from E. coli as well as from bacterial thermophiles. Recently the 23 Figure 4. Crystal structure of DnaA protein from Aquifex aeolicus The structure of DnaA (from Erzberger et. al, 2002) includes Domains IIIa, HIb and IV. Domains I and II were deleted from the full-length protein to circumvent the problem of oligomerization, which has complicated the crystallographic attempts of others. This figure is composed of a ribbon diagram containing Domains IIIa, IIIb and IV. Domain IIIa is a five-stranded B-sheet (Bl-[35) flanked by two sets of a-helices on each side. This fold is similar to the RecA-type fold that is observed in many nucleotide binding proteins. Domain IIIb is an antiparrallel three-helix bundle. This contains the Sensor 11 motif which is common among AAA+ family members. Domain IIIa contains the Walker A and B boxes and the sensor 11 (AAA+) motif. The Walker A and B boxes in conjunction with the Sensor I and II motifs are responsible for the coordinate binding of ADP and the Mg2+ ion. Domain IV begins at helix 12, which is the acidic phospholipid binding region of DnaA protein. The actual DNA binding domain is a helix-tum-helix and is very similar to the Trp repressor DNA-binding fold. A. aeolicus and E. coli DnaA protein share 35% amino acid identity and 65% amino acid similarity, indicating the structure will be important for studies of E. coli DnaA protein. 24 Domain lllb Ill/IV Junction 01. 12 Domain IV ' sensor-2 Walker-A sensor-1 Walker-B 25 Domain Illa structure was solved for Domains IIIa-IV (see below) to a 2.7A resolution with ADP bound to DnaA protein (Erzberger, Pirruccello et al. 2002). Domain I and II were deleted during the cloning procedure to circumvent potential problems caused by DnaA protein 01 i gomerization. DnaA protein from E. coli and A. aeolicus share 35% sequence identity and 65% sequence similarity over the aligned region (Erzberger, Pirruccello et al. 2002). Because E. coli and A. aeolicus share strong sequence identity, the structural information obtained bears significantly on work with E. coli DnaA protein (Figure 4). Domain IIIa of DnaA protein is composed of a five B-strand bundle flanked by two oc-helices (al- a2) on the left and five or helices on the right (a3-0t8). Domain IIIa is a version of the RecA-type fold which is often observed among nucleotide-binding proteins. Domain IIIb is organized with three anti-parallel a-helices (a9-0t11) connected to Domain IIIa by a short linker segment. The region contained by Domain IIIa-IIIb comprises the Walker A / B boxes and the Sensor-1 and 2 motifs (common to AAA” proteins). Domain IV contains DNA-binding activity, and is connected to Domain IIIb by a long a-helix (0:12). This helix is presumed to be used as the site for membrane attachment based on E. coli DnaA protein studies. This connector helix is linked to a l“flix-turn-helix (HTH) motif which is with in a structural fold almost identical to the DNA-binding domain fold of the Trp repressor (Erzberger, Pirruccello et al. 2002). This recent advancement should enable more direct mutagenic studies of E. coli DnaA protein, and will help facilitate the dissection of DnaA protein’s regulatory role. MOIecular chaperones and the activation of DnaA The temperature sensitive mutant proteins DnaA46 and DnaA5 are activated for replication by molecular chaperones (Hwang and Kaguni 1988; Hwang and Kaguni 1988; Hupp and Kaguni 1993; Hupp and Kaguni 1993; Hupp and Kaguni 1993; Carr and Kaguni 1996). Both DnaA5 and DnaA46 contain the A184V missense mutation. This 26 can ses a reduction in the affinity for ATP, and ATP binding is almost abolished at the restrictive temperature. Overexpression of these mutant proteins results in the formation of DnaA aggregates which are inactive in biochemical assays. It is thought that these temperature sensitive DnaA mutant proteins have a greater propensity to form aggregates because of the nucleotide binding defect. Even though chaperones (DnaK, GrpE) have been demonstrated to have a role in activation of temperature sensitive DnaA proteins, a clear role for chaperone activity in DnaA” protein activation has not been established. The synthesis of DnaK protein does not change during the cell cycle in E. coli. ClpB and DnaK proteins have been shown to be important for activation of the initiation protein (TrfA) for plasmid RK2 replication but the action of these chaperones is not required for oriC dependent replication (Konieczny and Liberek 2002). Hyperactive initiation of DNA replication According to the model that DnaA protein is responsible for initiation of rcplication, it would be expected that the frequency of initiation could be increased by either an over-supply of DnaA protein or by isolating mutations in DnaA protein that cause an increase in the frequency of initiation. Both of these expectations have been documented. Increasing the level of DnaA protein in the cell results in unscheduled initiation. In these studies, dnaA gene expression is under the control of an inducible Promoter, resulting in elevated DnaA protein expression (Atlung, Lobner Olesen et al. 1987). This synthetic situation is not lethal to the host, but results in a decreased growth rate. The idea that the cell prefers a steady-state DnaA protein level is confirmed by I‘CSUIts describing a sequence near oriC containing several DnaA-boxes (datA). The datA locus is required for normal growth (Ogawa, Yamada et al. 2002). When this segment is deleted, cells undergo increased initiation. Based on this observation, it has been preposed that the datA locus serves to titrate the initiation potential of excess DnaA 27 protein under normal growth conditions. The other well-characterized condition involves a mutant form of DnaA protein, DnaAcos (see below), which causes a cold sensitive phenotype to the host. The cold sensitive phenotype associated with DnaAcos is due to hyperactive initiation, which is lethal to the host strain at the nonperrnissive temperature (for other hyperactive dnaA allele see Table 2). DnaA cos The dnaAcas allele was isolated as an intragenic suppressor of the temperature sensitive dnaA46 allele (Kellenberger-Gujer, Podhajska et al. 1978). Marker frequency experiments have measured the relative amount of the oriC locus to be 3-5 fold greater than the terC locus after the temperature shift from 42°C to 30°C (Katayama and Kornberg 1994) indicating that DnaAcos induces hyperactive initiation of DNA replication. Cloning and sequencing of the dnaAcas allele revealed four nucleotide substitutions which result in four amino acid changes in the protein (Q156L, A184V, H252Y and Y271H). Elevated expression of strains harboring DnaAcos have a lethal Phenotype at both 30°C and 42°C when the dnaAcas allele is encoded on a plasmid and linked downstream of the lacUV5 promoter (Braun, O'Day et al. 1987). The A184V mutation of dnaAcas does cause hyperactive initiation of replication at 30°C when Plasmid-home; however, the A184V mutation alone does not reconstitute the plasmid- home dnaAcos phenotype of lethality at 30°C and 42°C, indicating that other mutations are required (Nyborg, Atlung et al. 2000). DnaAcos protein has been purified to homogeneity and assayed in vitro. As expected DnaAcos binds ATP poorly (Kd100 pM), and has more replication activity than DnaA in crude extracts. Interestingly, DnaAcos has poor replication activity in replication assays reconstituted with purified components. This result suggests that DnaAcos requires a soluble factor for activation that is present in the crude extract and absent from the reconstituted system. This result 28 prompted others to try and identify proteins in the soluble fraction that regulate DnaA protein activity. The beta clamp of DNA polymerase III holoenzyme (Katayama, Kubota et a]. 1998) and Hda (Hda is a novel protein which negatively regulates DnaA protein) (Kato and Katayama 2001) were purified as components of a soluble fraction (RIDA regulatory inactivation of DnaA) that causes DnaA protein to hydrolyze ATP quickly. Some researchers have tried to isolate mutations in other genes or to determine if overexpression of other proteins will suppress DnaAcos. An increase in the expression of the cedA gene product suppresses DnaAcos. Strains carrying the dnaAcos allele are filamentous suggesting an inhibition or a reduction in cell division. Overexpression of CedA protein restores cell morphology but did not inhibit over-replication. The cedA dependent pathway is thought to be independent of FtsZ (FtsZ oligomerizes to form a ring during cell division), as filamentation of cells bearing dnaAcas is independent of the FtsZ inhibitor SfiA protein (Katayama, Takata et al. 1997). This conclusion seems imprecise as SfiC (e14 gene product) protein can also inhibit cell division through interaction with FtsZ. Deletion of the dam gene also suppresses DnaAcos. It was observed that dam“ Strains fail to overinitiate replication. If Dam methylase is supplied in trans the over- rePlication of DnaAcos was restored. This result indicates that DnaAcos may depend on the activation of oriC by Dam methylase to over-initiate replication (Katayama, Akimitsu Ct al. 1997). The only replication-related protein that is known to suppress DnaAcos is Squ. It was observed that a moderate over-expression of Squ protein allowed for 29 Table 2. Hyperactive dnaA alleles Mutation and allele Replication Mechanism Reference Phenotype Q 1561., A184V, Hyperactive 30°C, in Assumed to be Kellenberger- H252Y, Y27IH viva and in vitro caused by Gujier et. al, (dnaAcas) nucleotide binding 1978 defect V1 5 7E Hyperactive in viva Unknown Hiraga and Saitoh, 1974 GI 75D Dominant Negative to Unknown Sutton and dnaA+ Kaguni, 1997 P43L, GI 77D Dominant Negative to Unknown Sutton and dnaA‘ Kaguni, 1997 G] 7 7D Dominant Negative to Unknown Sutton and dnaA“ Kaguni, 1997 GI 77D, E248K, Dominant Negative to Unknown Sutton and E249K dnaA” Kaguni, 1997 A184V Hyperactive 30°C in Poor ATP binding Nyborg et. al, viva 2000; Carr and Kaguni, 1996 AI 84 V+H252Y Hyperactive at 30°C ATP binding Katayama and (dnaA46) when GroE is Nagata, 1991 overexpressed A1 84 T Dominant Negative to Unknown Sutton and dnaA“ Kaguni, 1997 A184V, H252Y, Hyperactive in viva Unknown Weigel et. al, R342C 1999 R3 34A Hyperactive in vitro Crippled ATPase N ichida et. al, activity 2002 The dnaA alleles listed in this table are either known to cause hyperactive initiation or it has been suggested based on genetic evidence. Most of these alleles encode a mutation near the Walker A box to correlate hyperactive initiation to a defect in nucleotide binding. In the case of DnaAcos it has been demonstrated that ATP binding activity is very poor and DnaAcos is hyperactive in vitro and in viva. The A184V mutant pI‘Otein is hyperactive in viva but not in vitro at 30°C. 30 colony formation of a dnaAcas strain at 30°C (Lu, Campbell et al. 1994). Although the mechanism of this suppression is unknown, an increase in the amount of Squ protein is probably inhibitory for initiation of replication as it may decrease the availability of oriC for DnaAcos, preventing hyperactive initiation. Alternative modes of chromosomal replication in Escherichia coli The primary focus of this review is to summarize DnaA-dependent initiation of DNA replication. Other pathways of chromosomal replication do exist in E. coli and these pathways are important because they allow E. coli to survive under abnormal circumstances, and they utilize many of the proteins that are required for oriC-dependent replication. The SOS response in E. coli is induced after DNA damage, ethanol treatment or thymine starvation. During the SOS response chromosomal DNA must be repaired and replicated. The mode of chromosomal replication that is activated during the SOS response is termed iSDR { (inducible stable DNA replication, reviewed in (Kogoma 1997)} and is referred to as being stable because this pathway is not inhibited by protein Synthesis inhibitors. In contrast, DNA replication from oriC requires an unstable factor as new rounds of replication require protein synthesis (Kogoma, Torrey et al. 1981). The Pathway of iSDR replicates from two known origins (ariMI and ariM2) (Magee, Asai et 31- 1 992). The origin ariMI is located in part of oriC and ariM2 is located at terC where noI‘mal chromosomal replication is terminated. Replication through iSDR is independent of DnaA protein, as dnaA null strains are able to support iSDR. This pathway is dependent on many of the homologous recombination proteins (Asai, Sommer et al. 1993) and is a form of recombination-dependent replication (RDR). It has been suggested that replication activity used for iSDR resembles that required for recombination. It is thought that RecA* protein (an activated form of RecA) promotes 31 strand invasion to form a D-loop. Strand separation facilitated by this D-loop (Asai and Kogoma 1994) provides a site for the loading of DnaB helicase by the primosome (Masai, Asai et a1. 1994) of which all the components are required for iSDR activity. The other method of DnaA independent chromosomal duplication is termed con stitutively stable DNA replication (cSDR). This mode of replication is activated in strains that lack a functional mhA gene (Torrey and Kogoma 1987). The mhA gene codes for RNase H protein, which specifically recognizes DNA-RNA hybrids. RNase H recognizes and degrades DNA-RNA hybrids that arise outside of oriC, and confers specificity of replication to oriC. Deletion of the dnaA gene has no effect on cSDR (Ogawa, Pickett et al. 1984). Initiation of replication through cSDR (like iSDR) is dependent on RecA protein (Torrey and Kogoma 1987) and assembly of the primosome. The initiation of DNA replication takes place at random sites around the chromosome tenned oriK (de Massy, Fayet et al. 1984). The strand opening at these sites is thought to be facilitated by R-loop formation of the DNA-RNA hybrids (Kogoma 1997). Both iSDR and cSDR do not require concomitant protein synthesis (von Meyenburg, Boye et 31- 1 987), unlike DnaA dependent replication (Kogoma 1978). Although iSDR and CSDR may appear similar, they are distinct pathways of chromosomal replication. The pathways of iSDR and cSDR are examples of replication in the absence of DnaA protein and these pathways are either activated in response to adverse conditions (iSDR) or mutations (cSDR). Other modes of DnaA independent replication have been dCScribed. These pathways involve integrative suppression, where the need for DnaA is suppressed by integration of a plasmid origin which replicates independently of DnaA pI‘Otein. The most well—characterized example involves the integration of the mini R1 replicon into oriC (Koppes and Nordstrom 1986). Although DnaA protein is not required for chromosomal replication, other replication proteins are required. 32 Replication fork collapse and replication fork restart As described earlier, the E. coli cell cycle constitutes two distinct parts, the C and D periods. The C period is the time required for the faithful replication of the E. coli genome. During the C period, replication fork collapse, DNA repair and replication fork restart occur in response to DNA damage or other events that block replication fork progression. Pyrimidine dimers, frozen proteins on the DNA, collisions with RNA polymerase have all been cited as frequent events that result in replication fork demise (Sandler 2000; Sandler and Marians 2000). After replication fork collapse the replication fork must be restarted to complete replication and ensure survival of the cell. Replication fork restart is a process that is dependent on proteins that constitute the primosome, and those that support homologous recombination (Marians 2000). The primosome requires the biochemical activities of PriA, lPriB, DnaT, PriC, DnaC, DnaB and DnaG proteins. The primosome is assembled in an ordered manner with the initial binding of PriA, followed by PriB, DnaT and PriC (N g and Marians 1996a; Ng and Marians 1996b). It has been observed that PriA binding is stabilized by the binding of PriB, which facilitates the binding of DnaT. PriA also directs the binding of the DnaC6-DnaB6 complex. After DnaC releases DnaB in an ATP-dependent reaction, DnaG (primase) transiently associates with the primosome to prime replication (Ng and Marians 1996; Ng and Marians 1996). It has been observed that priA null mutations are viable. These mutations, however, are lethal when DNA damage is significantly increased due to UV treatment or growth in the presence of a DNA damaging agent. PriA null mutants are also inviable when the rep coding region is disrupted (Sandler and Marians 2000). Because Rep helicase moves ahead of the replication fork and acts as a “bulldozer” clearing the path for the replication machinery, it is not surprising that priA rep double mutants are not viable (Sandler, McCool et al. 2001). It was also observed that priA null strains have 33 constitutive expression of the SOS response and are impaired for homologous recombination, suggesting an important link between replication fork restart and homologous recombination. Similarities between prokaryotic and eukaryotic initiation of DNA replication E. coli has served as a model system for the study of many biochemical pathways in other organisms. In studies of DNA replication, the extent of information known regarding the steps of initiation and elongation are well beyond that which is known regarding eukaryotic replication. Certainly, there are many limitations associated with studying eukaryotic replication that are overcome rather easily in E. coli (see below). Despite the obvious differences between these two systems, it is apparent that many commonalities exist between prokaryotic and eukaryotic initiation. DnaA protein promotes the initiation of DNA replication in E. coli. The analogous functioning complex in yeast, and metazoans is the origin recognition complex (ORC) (Bell and Stillman 1992). ORC is composed of six polypeptides (Orcl-6) which are required for origin binding activity (Bell and Stillman 1992). S. cerevisiae ORC must bind ATP prior to binding the replication origin for stable association. ATP hydrolysis by ORC is very slow, which is analogous to the slow rate of ATP hydrolysis by DnaA (Klemm, Austin et al. 1997). ATP binding affinity for DnaA is very similar to that observed for ORC (Orcl and Orc5 20-100 nM, DnaA 30 nM). Like DnaA, ORC also remains bound to the replication origin throughout the cell cycle, although the DNase I protection pattern is different for cells that are in GI versus the pattern observed for S and M phase cells (Brown, Holmes et al. 1991; Diffley and Cocker 1992; Diffley, Cocker et al. 1994). Similarly, DnaA protein remains bound to three of the five DnaA boxes throughout the cell cycle, with the fourth and fifth boxes becoming occupied just prior to 34 initiation. DnaA protein seems to have a similar functional role to ORC, although only modest amino acid sequence similarity between ORC subunits and DnaA are observed. The homology that does exist between replication proteins of both systems comes in the form of the AAA+ family (Neuwald, Aravind et al. 1999) of ATPases (ATPases associated with a variety of cellular activities). DnaA protein is a member of this extensive collection of proteins. This family also includes, but is not limited to, DnaC protein, proteins of the 7 complex (Clamp loader) in prokaryotic initiation, Cdc6/Cdc18 (Cdc18 is S. pombe Cdc6), Orc1,4,5, MCM2-7, and RFC family members from eukaryotic replication. Many of the proteins in this group form hexamers (specifically, the 7 complex), suggesting that DnaA and other members of this group may do the same. Recent crystal structure data has revealed significant similarities between eukaryotic and prokaryotic initiation proteins (Erzberger, Pirruccello et al. 2002). Many replication and transcription proteins in archaea are homologous to proteins involved in these processes in eukaryotic cells. Interestingly, the crystal structure of A. aeolicus DnaA is very similar (Figure 5) to the structure of Cdc6 (Liu, Smith et al. 2000), which may be the most DnaA-like eukaryotic replication protein. The role of Cdc6 protein in vivo has been shown to be important for loading a complex of MCM proteins (4,6,7) a complex speculated to function as a helicase to support replication fork movement (Chong, I-Iayashi et al. 2000). However, the helicase activity measured in vitro is weak (You, Komamura et al. 1999; Ishimi and Komamura-Kohno 2001; Lee and Hurwitz 2001). DnaA protein has been demonstrated to be important for the loading of DnaB to form the prepriming complex in E. coli. In S. cerevisae two Cdc6 35 Figure 5. Structural comparison of bacterial and archaeal replication proteins Archaeal Cdc6/Ore 1( from Liu et. al, 2000; Erzberger et. al, 2002) and bacterial DnaA protein have very similar structures even though the sequence homology is limited to 15%. (A) A ribbon diagram of the overall crystal structure and domain organization of DnaA protein and the closest known homologue to DnaA protein archaeal Cdc6/Orc1 protein. The green and orange regions of each structure corresponds to the ATP binding and processing domain. The yellow region corresponds to the domain containing DNA binding activity. Both structures contain helical domains fused to AAA+ motifs. (B) The light gray and pale yellow areas correspond to regions of DnaA and Cdc6/Orc1 that are important for the ATP binding cleft. This picture represent a superimposition of the AAA+ regions of the two proteins. The Sensor I (S-I) Sensor H (S-II) Walker A (W -A) and Walker B (W-B) are labeled and colored in magenta for DnaA, and blue corresponding to Cdc6/Ore]. The arrangement of the AAA+ motif and the Box VII segment are strongly correlated between these two proteins. (C) A comparison of the DNA binding Domains of DnaA and Cdc6/Ore 1. A ribbon diagram is located next to the secondary structure topology diagram. In the topology diagram the bold and yellow regions correspond to the conserved area of the fold; the gray region indicates the differences between the two folds. The DNA binding domain of both proteins is comprised of a HTH (helix-tum-helix) motif. The remaining fold of the C-terminal domains are arranged differently suggesting an evolutionary relationship may exist that is connected by conversion or a rearrangement of modules. 36 F 20 608 SEQ 37 alleles (cdc6-2 and Cdc6-3) were isolated that result in “persistent” initiation. It was determined that the cause of this activity was the constant association of MCM proteins with chromatin, suggesting a constitutively active form of Cdc6 protein continually loads the “helicase” (Liang and Stillman 1997). The molecular cause of hyperactive initiation in E. coli is the major subject of the work presented below. The objectives of this research are to identify amino acid changes in DnaA protein that result in an increased frequency of initiation. This work will be pursued to better understand the regulatory mechanisms that exist to limit the frequency of replication initiation in E. coli. Based on the work of others, we know that hyperactive initiation results in growth interference. The mechanism that results in growth interference after hyperactive initiation has never been identified. Experiments are also presented to examine the fate of replication forks after hyperactive initiation, and mechanisms responsible for lethality are discussed. The work presented in this dissertation are designed to understand the role of DnaA protein in the regulation of initiation or DNA replication. 38 References Asai, T. and T. Kogoma (1994). "D-loops and R-loops: alternative mechanisms for the initiation of chromosome replication in Escherichia coli." Journal of Bacteriology 176(7): 1807-12. Asai, T., S. Sommer, A. Bailone and T. Kogoma (1993). "Homologous Recombination- Dependent Initiation of DNA Replication from DNA Damage-Inducible Origins in Escherichia coli." Euroman Molecular Biology Organization Journal 12(8): 3287-3295. Atlung, T., A. Lobner Olesen and F. G. Hansen (1987). "Overproduction of DnaA protein stimulates initiation of chromosome and minichromosome replication in Escherichia coli." Moleoular and general genetics 206(1): 51-59. Bell, S. P. and B. Stillman (1992). "ATP-dependent recognition of eukaryotic origins of DNA replication by a multiprotein complex." Nature 357(6374): 128-134. Blaesing, F., C. Weigel, M. Welzeck and W. Messer (2000). "Analysis of the DNA- binding domain of Escherichia coli DnaA protein." Molecolar Microbiology 36(3): 557-569. Boye, E., A. Lobner-Olesen and K. Skarstad (1988). "Timing of chromosomal replication in Escherichia coli." Biochimica et Biophysica Acta 951(2-3): 359- 364. Bramhill, D. and A. Kornberg (1988). "Duplex opening by DnaA protein at novel sequences in initiation of replication at the origin of the E. coli chromosome." Ce_ll 52(5): 743-755. Bramhill, D. and A. Kornberg (1988). "A model for initiation at origins of DNA replication." Q11 54(7): 915-918. Braun, R. E., K. O'Day and A. Wright (1985). "Autoregulation of the DNA replication gene dnaA in E. coli K-12." Cell 40(1): 159-169. Braun, R. E., K. O'Day and A. Wright (1987). "Cloning and characterization of dnaA(Cs), a mutation which leads to overinitiation of DNA replication in Escherichia coli K-12." Journal of Bacteriology 169(9): 3898-3903. 39 Brendler, T., A. Abeles and S. Austin (1995). "A protein that binds to the P1 origin core and the oriC 13mer region in a methylation-specific fashion is the product of the host squ gene." Eur ean M le ular Biolo r anizati n ma] 14(16): 4083-4089. Brown, J. A., S. G. Holmes and M. M. Smith (1991). "The chromatin structure of Saccharomyces cerevisiae autonomously replicating sequences changes during the cell division cycle." Molecular and Cellular Biology 11(10): 5301-53] 1. Buhk, H. J. and W. Messer (1983). "The replication origin region of Escherichia coli: nucleotide sequence and functional units." gone 24(2-3): 265-279. Carr, K. M. and J. M. Kaguni (1996). "The A184V missense mutation of the dnaA5 and dnaA46 alleles confers a defect in ATP binding and therrnolability in initiation of Escherichia coli DNA replication." Molecolar Microbiology 20(6): 1307-1318. Chang, J. P., M. K. Hayashi, M. N. Simon, R. M. Xu and B. Stillman (2000). "A double-hexamer archaeal minichromosome maintenance protein is an ATP- dependent DNA helicase." Proceedings of the National Academy of Soience, QSA 97(4): 1530-1535. Cooper, S. and C. E. Helmstetter (1968). "Chromosome replication and the division cycle of Escherichia coli B/r." Journal of Molecular Biology 31: 519-540. Craig, N. L. and H. A. Nash (1984). "E. coli integration host factor binds to specific sites in DNA." Cog 39(3 Pt 2): 707-716. de Massy, B., O. Fayet and T. Kogoma (1984). "Multiple origin usage for DNA replication in sdrA(rnh) mutants of Escherichia coli K-12. Initiation in the absence of oriC." loomal of Moleoolar Biology 178(2): 227-236. Diffley, J. F. and J. H. Cocker (1992). "Protein-DNA interactions at a yeast replication origin." HM 357(6374): 169-172. Diffley, J. F., J. H. Cocker, S. J. Dowell and A. Rowley (1994). "Two steps in the assembly of complexes at yeast replication origins in vivo." Coll 78(2): 303-316. Erzberger, J. P., M. M. Pirruccello and J. M. Berger (2002). "The structure of bacterial DnaA: implications for general mechanisms underlying DNA replication initiation." Bowman Molooular Biology Organization Journal 21(18): 4763-4773. 40 Fuller, R. S., B. E. Funnell and A. Kornberg (1984). "The DnaA protein complex with the E. coli chromosomal replication origin (oriC) and other DNA sites." Qfl 38(3): 889-900. Fuller, R. S., J. M. Kaguni and A. Kornberg (1981). "Enzymatic replication of the origin of the Escherichia coli chromosome." P ocee in of th a i n Academy of Science, USA 78(12): 7370—7374. Fuller, R. S. and A. Kornberg (1983). "Purified DnaA protein in initiation of replication at the Escherichia coli chromosomal origin of replication." Proceedings of the National Aeaoemy of Seience, USA 80(19): 5817-5821. Garner, J. and E. Crooke (1996). "Membrane regulation of the chromosomal replication activity of E. coli DnaA requires a discrete site on the protein." Euroman Molecular Biology Organization Journal 15(9): 2313-2321. Gille, H., J. B. Egan, A. Roth and W. Messer (1991). "The FIS protein binds and bends the origin of chromosomal DNA replication, oriC, of Escherichia coli." Nucleic Aeids Research 19(15): 4167-4172. Gille, H. and W. Messer (1991). "Localized DNA melting and structural pertubations in the origin of replication, oriC, of Escherichia coli in vitro and in viva." Euroman Moleeoler Biology Organization Journal 10(6): 1579-1584. Guo, L., T. Katayama, Y. Seyama, K. Sekimizu and T. Miki (1999). "Isolation and characterization of novel cold-sensitive dnaA mutants of Escherichia coli." FEMS Mierobiology Letters 176(2): 357-66. Hansen, E. B. and M. B. Yarmolinsky (1986). "Host participation in plasmid maintainance: Dependence upon dnaA of replicons derived from P1 and F." Ermeedings of the National Academy of Science, USA 83: 4423-4427. Hansen, F. G., S. Koefoed and T. Atlung (1992). "Cloning and nucleotide sequence determination of twelve mutant dnaA genes of Escherichia coli." Molecular and W 234(1): 14-21- Heacock, P. N. and W. Dowhan (1987). "Construction of a lethal mutation in the synthesis of the major acidic phospholipids of Escherichia coli." Journal of Biologicel Chemistry 262(27): 13044-13039. 41 Heacock, P. N. and W. Dowhan (1989). "Alteration of the phospholipid composition of Escherichia coli through genetic manipulation." al of Bi 1 i a hemi t 264(25): 14972-14977. I-liasa, H. and K. J. Marians (1994). "Fis cannot support oriC DNA replication in vitro." Joamal of Biologieal Chemistry 269(40): 24999-25003. Hiraga, S. (2000). "Dynamic localization of bacterial and plasmid chromosomes." Annoal Review of Cenerics 34: 21-59. Hiraga, S., C. Ichinose, T. Onogi, H. Niki and M. Yamazoe (2000). "Bidirectional migration of Squ-bound hemimethylated DNA clusters and pairing of oriC copies in Escherichia coli." Cenes Cells 5(5): 327-341. Hupp, T. R. and J. M. Kaguni (1993). "Activation of DnaA5 protein by GrpE and DnaK heat shock proteins in initiation of DNA replication in Escherichia coli." Joumal of Biological Chemistry 268(18): 13137-13142. Hupp, T. R. and J. M. Kaguni (1993). "Activation of mutant forms of DnaA protein of Escherichia coli by DnaK and GrpE proteins occurs prior to DNA replication." Journal of Biological Chemistry 268(18): 13143-13150. Hupp, T. R. and J. M. Kaguni (1993). "DnaA5 protein is therrnolabile in initiation of replication from the chromosomal origin of Escherichia coli." Joamal of Bjologiealmrsg'y 268(18): 13128-13136. Hupp, T. R., D. W. Meek, C. A. Midgley and D. P. Lane (1993). "Activation of the cryptic DNA binding function of mutant forms of p53." Nucleic Aeids Research 21(14): 3167-3174. Hwang, D. S. and J. M. Kaguni (1988). "Interaction of DnaA46 protein with a stimulatory protein in replication from the Escherichia coli chromosomal origin." Joomal of Biological Chemistry 263(22): 10633-10640. Hwang, D. S. and J. M. Kaguni (1988). "Purification and characterization of the dnaA46 gene product." Joamal of Biological Chemistry 263(22): 10625-10632. Hwang, D. S. and J. M. Kaguni (1991). "DnaK protein stimulates a mutant form of DnaA protein in Escherichia coli DNA replication." Journal of Biological Chemistry 266(12): 7537-7541. 42 Hwang, D. S. and A. Kornberg (1990). "A novel protein binds a key origin sequence to block replication of an E. coli minichromosome." Cell 63(2): 325-331. Hwang, D. S. and A. Kornberg (1992). "Opening of the replication origin of Escherichia coli by DnaA protein with protein HU or IHF." Journal of Biological Chemisrry 267(32): 23083-23086. Hwang, D. S. and A. Kornberg (1992). "Opposed actions of regulatory proteins, DnaA and IciA, in opening the replication origin of Escherichia coli." Journal of Biological Chemistg 267(32): 23087-23091. Hwang, D. S., B. Thony and A. Kornberg (1992). "IciA protein, a specific inhibitor of initiation of Escherichia coli chromosomal replication." Journal of Biological Chemistry 267(4): 2209-2213. Ishimi, Y. and Y. Komamura-Kohno (2001). "Phosphorylation of Mcm4 at specific sites by cyclin-dependent kinase leads to loss of Mcm4,6,7 helicase activity." Joomal of Biologieal Chemistry 276(37): 34428-34433. Kaguni, J. M., L. L. Bertsch, D. Bramhill, J. E. Flynn, R. S. Fuller, B. Funnell, S. Maki, T. Ogawa, lK. Ogawa and A. van der Ende (1985). "Initiation of replication of the Escherichia coli chromosomal origin reconstituted with purified enzymes." mm: 141-150. Kaguni, J. M., R. S. Fuller and A. Kornberg (1982). "Enzymatic replication of E. coli chromosomal origin is bidirectional." Nature 296(5858): 623-627. Kaguni, J. M. and L. S. Kaguni (1992). Enzyme-labeled probes for nucleic acid hybridization. Methods Birmhem Anal: 115-127. Kaguni, J. M. and A. Kornberg (1984). "Replication initiated at the origin (oriC) of the E. coli chromosome reconstituted with purified enzymes." QB 38(1): 183-190. Kano, Y., T. Ogawa, T. Ogura, S. Hiraga, T. Okazaki and F. Imamoto (1991). "Participation of the histone-like protein HU and of IHF in minichromosomal maintenance in Escherichia coli." Cene 103(1): 25-30. Katayama, T., N. Akimitsu, T. Mizushima, T. Miki and K. Sekimizu (1997). "Overinitiation of chromosome replication in the Escherichia coli dnaAcos mutant depends on activation of oriC function by the dam gene product." Moleeolar Microbiology 25(4): 661-70. 43 Katayama, T., K. Fujimitsu and T. Ogawa (2001). "Multiple pathways regulating DnaA function in Escherichia coli: distinct roles for DnaA titration by the datA locus and the regulatory inactivation of DnaA." Biochimie 83(1): 13-7 Katayama, T. and A. Kornberg (1994). "Hyperactive initiation of chromosomal replication in vivo and in vitro by a mutant initiator protein, DnaAcos, of Escherichia coli." Journal of Biological Chemistry 269(17): 12698-12703. Katayama, T., T. Kubota, K. Kurokawa, E. Crooke and K. Sekimizu (1998). "The initiator function of DnaA protein is negatively regulated by the sliding clamp of the E. coli chromosomal replicase." Ce_ll 94(1): 61-71. Katayama, T., M. Takata and K. Sekimizu (1997). "CedA is a novel Escherichia coli protein that activates the cell division inhibited by chromosomal DNA over- replication." Moleeular Microbiology 26(4): 687-97. Kato, J. and T. Katayama (2001). "Hda, a novel DnaA-related protein, regulates the replication cycle in Escherichia coli." Buromfl Molecular Biology Qrganization Jomal 20(15): 4253-62. Kellenberger-Gujer, G., A. J. Podhajska and L. Caro (1978). "A cold sensitive dnaA mutant of E. coli which overinitiates chromosome replication at low temperature." Molecalar and Ceneral Cenetics 162(1): 9-16. Kikuchi, S., I. Shibuya and K. Matsumoto (2000). "Viability of an Escherichia coli pgsA null mutant lacking detectable phosphatidylglycerol and cardiolipin." Joarnal of Baeteriology 182(2): 371-376. Klemm, R. D., R. J. Austin and S. P. Bell (1997). "Coordinate binding of ATP and origin DNA regulates the ATPase activity of the origin recognition complex." Cell 88(4): 493-502. Kline, B. C., T. Kogoma, J. E. Tam and M. S. Shields (1986). "Requirement of the Escherichia coli dnaA gene product for plasmid F maintenance." Joamal of Bacteriology 168(1): 440-443. Kogoma, T. (1978). "A novel Escherichia coli mutant capable of DNA replication in the absence of protein synthesis." Joamal of Molecalar Biology 121(1): 55-69. Kogoma, T. (1997). "Stable DNA replication: interplay between DNA replication, homologous recombination, and transcription." Microbiology and Molecular Biology Beviews 61(2): 212-238. Kogoma, T., T. A. Torrey, N. L. Subia and G. G. Pickett (1981). An alternative DNA initiation pathway in E. coli. The Ibitiation of DNA Replicarion. D. S. Ray. Los Angeles, Academic Press. 22: 361-374. Konieczny, I. and K. Liberek (2002). "Cooperative action of Escherichia coli ClpB protein and DnaK chaperone in the activation of a replication initiation protein." Joamal of Biolological Chemistm 11: 11. Koppes, L. and K. Nordstrom (1986). "Insertion of an R1 plasmid into the origin of replication of the E. coli chromosome: random timing of replication of the hybrid chromosome." Cell 44(1): 1 17-124. Koppes, L. J. and K. von Meyenburg (1987). "Nonrandom minichromosome replication in Escherichia coli K-12." Journal of Bacteriology 169(1): 430-433. Kowalski, D. and M. J. Eddy (1989). "The DNA unwinding element: a novel, cis-acting component that facilitates opening of the Escherichia coli replication origin." Earooearl Moleealar Biology Organization Journal 8(13): 4335-4344. Kurokawa, K., T. Mizushima, T. Kubota, T. Tsuchiya, T. Katayama and K. Sekimizu (1998). "A stimulation factor for hydrolysis of ATP bound to DnaA protein, the initiator of chromosomal DNA replication in Escherichia coli." Bioehemieal angl Biophysical Research Communieations 243(1): 90-95. Langer, U., S. Richter, A. Roth, C. Weigel and W. Messer (1996). "A comprehensive set of DnaA-box mutations in the replication origin, oriC, of Escherichia coli." Molecalar Microbiology 21(2): 301-31 1. Lee, J. K. and J. Hurwitz (2001). "Processive DNA helicase activity of the minichromosome maintenance proteins 4, 6, and 7 complex requires forked DNA structures." Proceedings of the National Academy of Science, USA 98(1): 54-59. Lee, Y. S. and D. S. Hwang (1997). "Occlusion of RNA Polymerase by Oligomerization of DnaA Protein over the dnaA Promoter of Escherichia coli." Joarnal of Biologieal Cbemistry 272(1): 83-88. 45 Leonard, A. C. and C. E. Helmstetter (1986). "Cell cycle-specific replication of Escherichia coli minichromosomes." Proceedings of the National Academy of Seience, USA 83(14): 5101-5105. Liang, C. and B. Stillman (1997). "Persistent initiation of DNA replication and chromatin-bound MCM proteins during the cell cycle in cdc6 mutants." m Development 11(24): 3375-3386. Liu, J., C. L. Smith, D. DeRyckere, K. DeAngelis, G. S. Martin and J. M. Berger (2000). "Structure and function of Cdc6/Cdcl8: implications for origin recognition and checkpoint control." Molecular Cell 6(3): 637-648. Lobner-Olesen, A. (1999). "Distribution of minichromosomes in individual Escherichia coli cells: implications for replication control." European Molecular Biology Qrganization Journal 18(6): 1712-1721. Lu, M., J. L. Campbell, E. Boye and N. Kleckner (1994). "squ: a negative modulator of replication initiation in E. coli." C_e_ll 77(3): 413-426. Ludlam, A. V., M. W. McNatt, K. M. Carr and J. M. Kaguni (2001). "Essential amino acids of Escherichia coli DnaC protein in an N -terminal domain interact with DnaB helicase." Joamal of Biologioal Chemislry 276(29): 27345-27353. Magee, T. R., T. Asai, D. Malka and T. Kogoma (1992). "DNA damage-inducible origins of DNA replication in Escherichia coli." Burooean Molecular Biology MW 11(11): 4219-4225- Margulies, C. and J. M. Kaguni (1998). "The FIS protein fails to block the binding of DnaA protein to oriC, the Escherichia coli chromosomal origin." N_iiC_l§.iC_Mid_s Researeb 26(22): 5170-5175. Marians, K. J. (2000). "PriA-directed replication fork restart in Escherichia coli." Trends WM 25(4): 185-9. 4__00001565 4_00001565. Masai, H., T. Asai, Y. Kubota, K. Arai and T. Kogoma (1994). "Escherichia coli PriA protein is essential for inducible and constitutive stable DNA replication." European Molecular Biology Organization Journal 13(22): 5338-5345 Matsui, M., A. Oka, M. Takanami, S. Yasuda and Y. Hirota (1985). "Sites of DnaA protein-binding in the replication origin of the Escherichia coli K-12 chromosome." JobmalofMoleeuLanBigmgy 184(3): 529-533. 46 Meiier, M., E. Beck, F. G. Hansen, H. E. Bergrnans, W. Messer, K. von Meyenburg and H. Schaller (1979). "Nucleotide sequence of the origin of replication of the Escherichia coli K-12 chromosome." Prmeedings of the National Academy of Seienee, USA 76(2): 580-584. Meiier, M. and W. Messer (1980). "Functional analysis of minichromosome replication: bidirectional and unidirectional replication from the Escherichia coli replication origin, oriC." Journal of Bacteriology 143(2): 1049—1053. Mirna, S., M. Makise, M. Koterasawa, T. Tsuchiya and T. Mizushima (2002). "Conserved hydrophobic amino acid residues in the N-terrninal region of DnaA protein are involved in DnaA-DnaA interaction." Bimghem Joomal 11. Mirna, S., Y. Yamagachi, T. Kondo, T. Tsuchiya and T. Mizushima (1999). "Role of the arnino-terminal region of the DnaA protein in opening of the duplex DNA at the oriC region." FEMS Microbiol Bet; 176(1): 163-7. Mizushima, T., T. Takaki, T. Kubota, T. Tsuchiya, T. Miki, T. Katayama and K. Sekimizu (1998). "Site-directed mutational analysis for the ATP binding of DnaA protein. Functions of two conserved amino acids (Lys-178 and Asp-235) located in the ATP-binding domain of DnaA protein in vitro and in vivo." Joamal of Biologioal Chemistry 273(33): 20847—20851. Moitoso de Vargas, L., S. Kim and A. Landy (1989). "DNA looping generated by DNA bending protein IHF and the two domains of lambda integrase." Science 244(4911): 1457-1461. Molin, S., R. Diaz, B. E. Uhlin and K. Nordstrom (1980). "Runaway replication of plasmid R1 is not caused by loss of replication inhibitor activity of gene cop." Joamal of Bacteriology 143(2): 1046-1048. Molin, S. and K. Nordstrom (1980). "Control of plasmid R1 replication: functions involved in replication, copy number control, incompatibility, and switch-off of replication." Joomal of Bmteriology 141(1): 111-120. Moriya, S., T. Atlung, F. G. Hansen, H. Yoshikawa and N. Ogasawara (1992). "Cloning of an autonomously replicating sequence (ars) from the Bacillus subtilis chromosome." Molecalar Microbiology 6(3): 309-315. Mariya, S., T. Fukuoka, N. Ogasawara and H. Yoshikawa (1988). "Regulation of initiation of the chromosomal replication by DnaA-boxes in the origin region of 47 the Bacillus subtilis chromosome." Earorman Molecular Biology Qrganization Joilmal 7(9): 2911-2917. Neuwald, A. F., L. Aravind, J. L. Sponge and E. V. Koonin (1999). "AAA+: A class of chaperone-like ATPases associated with the assembly, operation, and disassembly of protein complexes." Genome Researeh 9(1): 27-43. Ng, J. Y. and K. J. Marians (1996). "The ordered assembly of the phiX174-type primosome. 1. Isolation and identification of intermediate protein-DN A complexes." Journal of Biological Chemistry 271(26): 15642-15648. Ng, J. Y. and K. J. Marians (1996). "The ordered assembly of the phin74-type primosome. H. Preservation of primosome composition from assembly through replication." Journal of Biological Chemistm 271(26): 15649-15655. Niki, H. and S. Hiraga (1997). "Subcellular distribution of actively partitioning F plasmid during the cell division cycle in E. coli." Ceb 90(5): 951-957. Niki, H. and S. Hiraga (1998). "Polar localization of the replication origin and terminus in Escherichia coli nucleoids during chromosome partitioning." Cenes Q Developmem 12(7): 1036- 1045. Niki, H., C. Ichinose, T. Ogura, H. Mori, M. Morita, M. Hasegawa, N. Kusukawa and S. Hiraga (1988). "Chromosomal genes essential for stable maintenance of the mini-F plasmid in Escherichia coli." Journal of Bacteriology 170(11): 5272- 5278. Niki, H., A. Jaffe, R. Imamura, T. Ogura and S. Hiraga (1991). "The new gene mukB codes for a 177 kd protein with coiled-coil domains involved in chromosome partitioning of E. coli." Eurorman Molecular Biology Qrganizatiob Journal 10(1): 183-193. Nyborg, M., T. Atlung, O. Skovgaard and F. G. Hansen (2000). "Two types of cold sensitivity associated with the A184-->V change in the DnaA protein." Molecular Microbiology 35(5): 1202-1210. Ogawa, T., G. G. Pickett, T. Kogoma and A. Kornberg (1984). "RNase H confers specificity in the DnaA-dependent initiation of replication at the unique origin of the Escherichia coli chromosome in vivo and in vitro." Proceedings of [he Natiobal Acadamey of Science, USA 81(4): 1040-1044. 48 Ogawa, T., Y. Yamada, T. Kuroda, T. Kishi and S. Moriya (2002). "The datA locus predominantly contributes to the initiator titration mechanism in the control of replication initiation in Escherichia coli." Molecular Microbiology 44(5): 1367- 75. Ogura, T. and S. Hiraga (1983). "Mini-F plasmid genes that couple host cell division to plasmid proliferation." Proceedings of the National Academy of Science. USA 80(15): 4784-4788. Oka, A., K. Sugimoto, M. Takanami and Y. Hirota (1980). "Replication origin of the Escherichia coli K-12 chromosome: the size and structure of the minimum DNA segment carrying the information for autonomous replciation." Molecular and Ceneral Cenetics 178(1): 9-20. Prescott, D. M. and P. L. Kuempel (1972). "Bidirectional replication of the chromosome in Escherichia coli." Proceedings of the N atonal Academy of Sciebees U S A 69(10): 2842-2845. Roth, A. and W. Messer (1995). "The DNA binding domain of the initiator protein DnaA." Earooean Molecular Biology Crganization Joarnal 14(9): 2106-2111. Sandler, S. J. (2000). "Multiple genetic pathways for restarting DNA replication forks in Escherichia coli K-12." Cenetics 155(2): 487-497. Sandler, S. J. and K. J. Marians (2000). "Role of PriA in replication fork reactivation in Escherichia coli." Joarnal of Bacteriology 182(1): 9-13. Sandler, S. J., J. D. McCool, T. T. Do and R. U. Johansen (2001). "PriA mutations that affect PriA-PriC function during replication restart." Molecular Microbiology 41(3): 697-704. Saraste, M., P. R. Sibbald and A. Wittinghofer (1990). "The P-loop--a common motif in ATP- and GTP-binding proteins." Trends in Biochemical Science 15(11): 430- 434. Seitz, H., C. Weigel and W. Messer (2000). "The interaction domains of the DnaA and DnaB replication proteins of Escherichia coli." Molecalar Microbiology 37(5): 1270-1279. 49 Sekimizu, K., D. Bramhill and A. Kornberg (1987). "ATP activates DnaA protein in initiating replication of plasmids bearing the origin of the E. coli chromosome." Cell 50(2): 259-265. Sekimizu, K., D. Bramhill and A. Kornberg (1988). "Sequential early stages in the in vitro initiation of replication at the origin of the Escherichia coli chromosome." Joomal of Bi ologieal Chemistry 263(15): 7124-7130. Sharma, R., A. Kachroo and D. Bastia (2001). "Mechanistic aspects of DnaA-RepA interaction as revealed by yeast forward and reverse two-hybrid analysis." Eampean Moleeular Biology Qrganizatioo Joamal 20(16): 4577-4587. Skarstad, K., K. von Meyenburg, F. G. Hansen and E. Boye (1988). "Coordination of chromosome replication initiation in Escherichia coli: effects of different dnaA alleles." Journal of Bacteriology 170(2): 852-858. Speck, C. and W. Messer (2001). "Mechanism of origin unwinding: sequential binding of DnaA to double- and single-stranded DNA." Burorman Moleeular Biology WW 20(6): 1469-1476. Speck, C., C. Weigel and W. Messer (1999). "ATP- and ADP-dnaA protein, a molecular switch in gene regulation." Euroman Molecular Biology Organization momal 18(21): 6169-6176. Sutton, M. D., K. M. Carr, M. Vicente and J. M. Kaguni (1998). "Escherichia coli DnaA protein. The N-terminal domain and loading of DnaB helicase at the E. coli chromosomal origin." Journal of Biological Chemistm 273(51): 34255-34262. Sutton, M. D. and J. M. Kaguni (1995). "Novel alleles of the Escherichia coli dnaA gene are defective in replication of pSC101 but not of oriC." Joamal of Bageijology 177(22): 6657-6665. Sutton, M. D. and J. M. Kaguni (1997). "Novel alleles of the Escherichia coli dnaA gene." Joamal oi Moleealar Biology 271(5): 693-703. Sutton, M. D. and J. M. Kaguni (1997). "Threonine 435 of Escherichia coli DnaA protein confers sequence- specific DNA binding activity." JotrmaLoLBiolochal Cbmmsoy 272(37): 23017-23024. 50 Torheim, N. K. and K. Skarstad (1999). "Escherichia coli Squ protein affects DNA topology and inhibits open complex formation at oriC." Euromaa Molecolar Biology Qrganization Journal 18(17): 4882-4888. Torrey, T. A. and T. Kogoma (1987). "Genetic analysis of constitutive stable DNA replication in rnh mutants of Escherichia coli K12." Molecular and General Ceneries 208(3): 420-427. von Meyenburg, K., E. Boye, K. Skarstad, L. Koppes and T. Kogoma (1987). "Mode of initiation of constitutive stable DNA replication in RNase H- defective mutants of Escherichia coli K-12." Journal of Bacteriology 169(6): 2650-2658. von Meyenburg, K. and F. G. Hansen (1987). Regulation of chromosome replication. E h richia c li an almonella himurium. ell lar an M lec lar Biolo . F. C. Neidhardt, J. L. Ingraham, K. B. Lowet al. Washington, D. C., American Society for Microbiology. 2: 1555-1577. von Meyenburg, K., F. G. Hansen, L. D. Nielsin and E. Riise (1978). "Origin of replication, oriC, of the Escherichia coli chromosome on specialized transducin g phages lambda asn." Moleealar and Ceneral Ceneties 160(3): 287-295. Walker, J. E., M. Saraste, M. J. Runswick and N. J. Gay (1982). "Distantly related sequences in the 0t- and B-subunits of ATP synthase, myosin, kinases and other ATP-requiring enzymes and a common nucleotide binding fold." Earorman Moleealar Biology Orgaoizalion Journal 1(8): 945-951. Webb, C. D., A. Teleman, S. Gordon, A. Straight, A. Belmont, D. C. Lin, A. D. Grossman, A. Wright and R. Losick (1997). "Bipolar localization of the replication origin regions of chromosomes in vegetative and sporulating cells of B. subtilis." Cell 88(5): 667-674. Weigel, C., A. Schmidt, H. Seitz, D. Tungler, M. Welzeck and W. Messer (1999). "The N-terrninus promotes oligomerization of the Escherichia coli initiator protein DnaA." Molecular Mierobiology 34(1): 53-66. Wu, F., I. Goldberg and M. Filutowicz (1992). "Roles of a 106-bp origin enhancer and Escherichia coli DnaA protein in replication of plasmid R6K." Nucleie Aeids Beseareb 20(4): 81 1-817. 51 Xia, W. and W. Dowhan (1995). "Phosphatidylinositol cannot substitute for phosphatidyl glycerol in supporting cell growth of Escherichia coli." Mail—of Bacteriology 177(10): 2926-2928. Yamaguchi, Y., M. Hase, M. Makise, S. Mirna, T. Yoshimi, Y. Ishikawa, T. Tsuchiya and T. Mizushima (1999). "Involvement of Arg-328, Arg-334 and Arg-342 of DnaA protein in the functional interaction with acidic phospholipids." Biochem J 340(Pt 2): 433-8. You, 2., Y. Komamura and Y. Ishimi (1999). "Biochemical analysis of the intrinsic Mcm4-Mcm6—mcm7 DNA helicase activity." Molecalar and Celldlm Biology 19(12): 8003-15. Zheng, W., Z. Li, K. Skarstad and E. Crooke (2001). "Mutations in DnaA protein suppress the growth arrest of acidic phospholipid-deficient Escherichia coli cells." Embo J 20(5): 1164-72. Zyskind, J. W., J. M. Cleary, W. S. Brusilow, N. E. Harding and D. W. Smith (1983). "Chromosomal replication origin from the marine bacterium Vibrio harveyi functions in Escherichia coli: oriC consensus sequence." Broceedings of the National Academy of Science, USA 80(5): 1164-1168. Zyskind, J. W., N. E. Harding, Y. Takeda, J. M. Cleary and D. W. Smith (1981 ). The DNA replication origin region of the enterobactiriaceae. The Iniliariod of DNA Replication. D. S. Ray. Los Angeles, Academic Press. 22: 13-28. 52 Chapter II DnaA Protein of E. coli: Initiation of DNA Replication Requires DnaA Oligomerization and Specific N -terminal Amino Acids 53 Abstract DnaA box sequences present in multiple copies within the replication origins of bacteria and prokaryotic plasmids are recognized by the replication initiator, DnaA protein. At the E. coli chromosomal origin, ariC, DnaA is speculated to oligomerize in order to initiate chromosomal DNA replication. We developed an in viva assay of oligomer formation at ariC that relies on complementation between two dnaA alleles that are inactive by themselves. One allele is dnaA46; its inactivity at the nonperrnissive temperature is due to a specific defect in ATP binding. The second allele, T435K, does not support DNA replication because of its inability to bind to DnaA box sequences within oriC. We show that the T435K allele carried in a plasmid can complement the temperature sensitivity of the dnaA46 allele. The results support a model of oligomer formation in which DnaA46 is bound to the DnaA box sequences of oriC, and the mutant protein carrying the T435K substitution interacts with the bound DnaA46 to form a complex that is active in initiation. Relying on this assay, leucine 5, tryptophan 6 and cysteine 9 in a predicted alpha helix near the N -terminus of DnaA were identified that, when altered, interfere with oligomer formation. Glutamine 8 and proline 28 are additionally needed for oligomer formation when oriC is carried in a plasmid, suggesting that the structure of the DnaA-oriC complex at the chromosomal ariC locus is similar but not identical to that assembled on a plasmid. These results provide direct evidence that DnaA oligomerization at ariC is required for initiation to occur. 54 Introduction Initiation of DNA replication involves the recognition of the replication origin by an origin binding protein, followed by the stepwise assembly of other proteins to build the replication fork machinery. In E. coli, the origin binding protein is DnaA, recognizing a DNA motif termed the DnaA box (Fuller and Kornberg 1983). In a complex with ATP, DnaA then opens the parental duplex and recruits the replicative helicase, DnaB, to this site to form the prepriming complex (Sekimizu, Bramhill et al. 1987; Bramhill and Kornberg 1988; Marszalek and Kaguni 1994). Two helicase molecules are recruited, one for each replication fork to promote bidirectional fork movement (Fang, Davey et al. 1999; Carr and Kaguni 2001). A similar process occurs in eukaryotes; the ORC proteins (Orcl-6) recognize the replication origin sequence to recruit Mcm2-7, the eukaryotic counterpart to E. coli DnaB (reviewed in (Bell 2002)). Recent results describing the crystal structure of a truncated form of DnaA protein from A. aeolicus reveal a remarkable similarity to that of archaeal Cdc6/Ore] (Erzberger, Pirruccello et al. 2002). The similarities in structure and function of these proteins at the stage of initiation suggest that a common biochemical pathway is followed in all organisms. The E. coli chromosomal origin, ariC, is defined as a 245 base pair minimal DNA sequence, based on mutational analysis of this sequence carried in a ColEl-derived plasmid vector (Oka, Sugimoto et al. 1980). In combination with comparative DNA sequence analysis, conserved sequences were identified that include five DnaA box motifs named R1, M, R2, R3 and R4 (Fuller and Kornberg 1984; Zyskind, Cleary et al. 1983; Matsui, Oka et al.1985). The idea that nonconserved sequences act as spacers to maintain the proper orientation between sites bound by proteins is supported by other mutational studies. Insertion of 4 or 8 base pairs or deletion of 4 base pairs between DnaA boxes R3 and R4 inactivated the cloned replication origin sequence (Woelker and 55 Messer 1993). However, insertion or deletion of one helical turn (10-12 base pairs) between these DnaA boxes was tolerated. These findings support a model of helical phasing, involving the interaction among DnaA molecules bound to oriC. The conclusions drawn from these studies contrast with those of Bates et al. (Bates, Asai et al. 1995). They described that insertion of a 2 kilobase pair fragment carrying the spectinomycin resistance gene (the omega fragment) between DnaA boxes R3 and R4, or deletion of R4 did not interfere with ariC function when these alterations were present in the bacterial chromosome. However, these modifications were inactivating when carried in a plasmid. Although DNA replication of the bacterial chromosome was inefficient and asynchronous, the results suggest that the nucleoprotein complex assembled at the chromosomal ariC locus is different from that which is assembled on an oriC plasmid. We previously described a mutational approach that identified functional domains within DnaA protein of 467 amino acids (Sutton and Kaguni 1997). A set of mutations clustered near the N-terrninal coding region were obtained, using a genetic selection that yielded mutants defective in DNA replication activity. These mutants were also impaired in autoregulation of dnaA expression (Sutton and Kaguni 1997); results from other studies suggest that repression of dnaA transcription involves oligomerization of DnaA molecules bound to the dnaA promoter region (Lee and Hwang 1997). More recent results show that this region (residues 2-86) of DnaA protein can replace the dimerization domain of CI repressor encoded by bacteriophage lambda (Wei gel, Schmidt et al. 1999). Because an N-terminal domain within the first 62 amino acids of DnaA also interacts with DnaB in loading of the helicase at ariC (Sutton, Carr et al. 1998; Seitz, Weigel et al.1999), we wanted to identify amino acid residues that function solely in oligomerization of DnaA to distinguish this functional domain from that which interacts with DnaB. As another objective, we sought to demonstrate that oligomerization of 56 DnaA at ariC is a required function for initiation to occur. At ariC, oligomer formation among DnaA monomers is thought to be required during initiation, but little evidence exists to support this conclusion. To address these objectives, we speculated that we may be able to demonstrate oligomer formation in vivo by using two dnaA alleles that were defective in mutually exclusive functions. Whereas each allele alone is inactive, their combined presence in the same bacterial cell may then result in the formation of mixed complexes that are active in initiation. We chose the dnaA46 allele as one of the pair because the mutant protein is active in binding to oriC, at 42°C, but is inactive in DNA replication due to a defect in ATP binding (Carr and Kaguni 1996). The second allele of the pair was chosen because it is defective in binding to ariC. In other work, we showed that threonine 435 near the C-terminus of DnaA protein is involved in recognition of the DnaA box sequence (Sutton and Kaguni 1997). Based on the crystallographic structure of a truncated form of A. aeolicus DnaA protein, this amino acid resides at the end of an alpha helix and before the start of a loop in a region designated the DnaA signature sequence (Erzberger, Pirruccello et al. 2002). A mutant protein carrying a methionine substitution at this position (T435M) is inactive in DNA replication because it fails to bind to ariC (Sutton and Kaguni 1997). However, T435M can augment limiting levels of wild type DnaA in DNA replication of an oriC plasmid in vitro presumably by the formation of mixed oligomers of T435M and DnaA“. In the study described below, we used a different dnaA allele encoding a lysine substitution at threonine 435 (T435K). As expected, this mutant protein is active in ATP binding, but defective in recognition of the DnaA box sequence (Severson et al., manuscript in preparation). We demonstrate allelic complementation between T435K and dnaA46 at a temperature that inactivates the function of DnaA46 protein when assayed alone. Using an approach to combine the T435K mutation with other mutations that substitute amino 57 acids near the N -terminus of DnaA, we then assayed these alleles for intergenic complementation of the dim/146 allele. Because C9Y and P28L alleles are known to inactivate the function of DnaA protein in DNA replication (Sutton and Kaguni 1997), we focused on amino acid substitutions near the N-terminus of DnaA protein to determine if these and other unique missense mutations (L5A, W6A, Q7A and Q8A) were inactive in oligomer formation. Substitutions L5A, W6A, Q7A, Q8A and C9Y affect consecutive residues that reside in a predicted (rt-helix (Sutton, Carr et al. 1998) proximal to the N-terminus of DnaA, and within Domain 1, a larger conserved sequence shared among bacterial DnaAs (Fujita, Yoshikawa et al. 1990; Messer 2002). 58 Experimental Procedures Strains and plasmids-The E. coli K-12 strains used in this study are listed in Table l, and were grown in LB media (Miller 1992). Plasmid DNAs were purified by cesium chloride-ethidium bromide density gradient centrifugation or by chromatography (Plasmid Midi Kit, Qiagen). Plasmid pRB 100 expresses the dnaA“ gene from its native dnaA promoters and is a derivative of pBR322 (Braun, O'Day et al. 1985). Plasmid pAB2 expresses the T435K mutation from the native dnaA promoters and is a derivative of pACYC177 (Gines, Blinkova et al. 1995). Plasmids carrying both N terminal mutations and the T435K mutation are derivatives of pAB2. Plasmid pKC596 encodes the dnaA+ gene under control of the T7 RNA polymerase promoter (Carr and Kaguni 1996). The ariC plasmid pCM959-CmR is a derivative of pCM959 (Buhk and Messer 1983) constructed by insertion of the chloramphenicol acetyltransferase gene (Simmons and Kaguni, 2003; Hupert-Kocurek and Kaguni, unpublished results). Bacterial transformation-Transformations were by electroporation (E. coli Pulser, BioRad) following the recommendations of the manufacturer (BioRad) with competent cells resuspended at a cell density of ~2 x 10'0 cells/ml, and stored at -70°C. Unless otherwise indicated, 10 ng of plasmid DNA was mixed with 40 pl of competent cells. After electroporation, the cells were quickly resuspended in LB or SOC medium and incubated in the absence of antibiotic selection for 20 min under slight agitation. The cells were diluted and plated on the appropriate selective media and incubated overnight at the temperatures indicated. Site-directed mutagenesis-The N -terminal mutations were introduced into plasmid pAB2 by PCR mutagenesis with overlapping primers carrying the desired mutations as indicated by the underlined sequences. For brevity, only one of the two complementary primer sequences is shown: L5A, GTGTCACTI’TCGGCTTGGCAGCAGTGT; W6A, GTCAC'I‘T'I‘CGCTTGCGCAGCAGTGTCTI‘G; Q7A, 59 CTI'I‘CGC’I'ITGGGCGCAGTGTC'I'I‘GCC; Q8A, TCGCTTTGGCAGGCGTGTCTTGCCCGA; C9Y, TGGCAGCAGTATCTTGCCCGA; and P28L, TGTGGATACGCCTA'I'I‘GCAGGCGGA. Both the N -terminal and C- terminal portions of the dnaA gene were sequenced to verify presence of the respective N-terrninal mutation and to confirm the presence of the T435K mutation. Glutaraldehyde Crasslinking-Protein crosslinking was in 25 mM Tris-HCl pH 7.5, 1 mM EDTA, 0.01% Nonidet P-40, 100 mM NaCl, and 1 mM DTT with purified DnaB (200 ng), DnaAA62 (180 ng), or a derivative of wild type DnaA canying a polyhistidine tag at its N-terminus (190 ng) in a final volume of 25 pl. The DnaA derivative is comparable in activity to wild type DnaA protein in viva and in DNA replication activity in vitro (Walker et al., manuscript in preparation). Glutaraldehyde was added to a concentration of 0.01%, followed by incubation at room temperature for the times indicated. Reactions were quenched by the addition of 5 pl of 50% glycerol, 0.2 M Tris, 1% SDS, 0.1% bromophenol blue and 5 mM DTT. Samples were incubated at 86°C for 5 min, electrophoresed alongside prestained molecular weight markers (Gibco-BRL) in an SDS-polyacrylamide gel, and transferred to a membrane (Protran, Schleicher & Schuell). Protein transfer was confirmed by the presence of prestained molecular weight markers on the membrane. Detection of DnaA protein complexes was with rabbit antiserum prepared against the DNA binding domain of DnaA (residues 370- 467) unless noted otherwise, or with affinity-purified antibody that is specific for DnaB. After incubation with horseradish peroxidase-conjugated goat anti-rabbit secondary antibody (BioRad), the chemiluminescence (Supersignal, Pierce) was detected with Xray film (X Omat, Kodak). 60 Table 1. List of Strains Strain Genotype' Parent Source LS 1073 araDI39 AaraC leu7697 A(lac)X174 MC1061 Simmons and galU galK hst2(rk',mk*) strA mcrA Kaguni, 2003 mchl tna::Tn10 dnaA46 LS 1065 araDI39 AaraC leu7697 A(lac)X174 MC 1061 This work galU galK hst2(rk',mk*) strA mcrA mchI tna::Tn]0 dnaA46 AariC::pKN1562(clockwise) Km', LS 1061 araDI39 AaraC leu7697 A(lac)Xl74 MC1061 Simmons and galU galK hst2(rk',mk*) strA mcrA Kaguni, 2003 mchl AariC::pKN1562(clockwise) Km' LS 1062 araDI39 AaraC leu7697 A(lac)X174 LS 1061 Simmons and gal U galK hst2(rk',mk*) strA mcrA Kaguni, 2003 mchl AariszpKN1562(clockwise) Km’ dnaA85022Tn10 MC1061 araDI39 AaraC leu7697 A(lac)Xl74 Lab Stock galU galK hst2(rk',mk*) strA mcrA mch ‘ The strains listed are derivatives of MC1061 and were constructed by P1 transduction (Simmons and Kaguni, 2003). 61 Results Intragenic complementation as an in vivo assay for DnaA oligomerization. The notion that oligomerization of DnaA at oriC is required for initiation is based initially on estimates from electron microscopy which suggest that 20-40 DnaA monomers are bound to oriC (Fuller, Funnell et al. 1984; Crooke, Thresher et al. 1993). Other results indicate that an N-terrninal domain of DnaA is involved in oligomerization, and that this activity may be required for initiation from oriC (Sutton and Kaguni 1997; Weigel, Schmidt et al. 1999). To obtain direct evidence to support this concept, we chose the T435K and dnaA46 alleles for reasons described above. The experimental approach was to construct merodiploid strains in which the host carried the dnaA46 locus, and different plasmid- bome dnaA alleles were introduced by transformation. When the T435K allele (plasmid- bome) was introduced into the dnaA46 strain, complementation was observed at the higher temperature, confirming that oligomer formation occurs between the two mutant proteins. Likewise, the dnaA+ plasmid complemented the dnaA46 allele at the elevated temperature, in agreement with previous observations (Hansen, Atlung et a1. 1984). However, the latter results do not support the formation of mixed oligomers as DnaA+ protein alone at ariC may support initiation in the dnaA46 host strain. In contrast, the vector for the dnaA+ plasmid (pBR322) or that for the T435K allele (pACYCl77) were ineffective. Double mutations carrying the T435K substitution combined with different missense mutations near the start of the dnaA coding region were assayed in parallel. Q7A, Q8A and P28L were active in complementation whereas L5A,W6A and C9Y were comparable to the empty vector controls (Table 2). These results suggest that the latter group of mutations are inactive due to a defect in oligomerization. Oligomerization of DnaA at oriC is thought to involve interactions among monomers bound to the different DnaA boxes at this site. Because T435K does not bind to the DnaA 62 Table 2. L5A, W6A, and C9Y alleles are defective in allelic complementation when combined with dnaA46 Plasmid dnaA allele Plating Efficiency“ dnaA+ pRB 100 l .0 pBR322 None 2.5X10'3 pAB2 T435K 1 .0 pACYC177 None 1 .9X 104 pLS2 L5A+T435K 2.0x10'3 pLS3 W6A+T435K 6.8X10’3 pLS4 Q7A+T435K 1.1 pLSS Q8A+T435K 1 .0 pLS 1 C9Y+T435K 4.9X10" pLS6 P28L+T435K 0.97 ‘The respective plasmids were transformed into LS 1073 (dnaA46), followed by incubation in SOC media for 20 min before plating on LB media containing 100 pg/ml ampicillin. After overnight incubation at 30°C or 42°C, plating efficiencies were determined as the ratio of transforrnants obtained at the elevated temperature over those obtained at 30°C. Transformation frequencies of the various plasmids at 30°C ranged from 10‘5 to 107 transforrnants per pg of plasmid DNA. 63 boxes but only retains nonspecific DNA binding activity, the results of Table 2 suggest that monomers of T435K bind to DnaA46 monomers complexed to respective DnaA box sequences. In the experiment of Table 2, the activity of the dnaA alleles was measured by initiation from the chromosomal oriC locus. As studies show that DnaA box R4 is required for replication from ariC when this sequence is contained in a plasmid, but that ariC in the bacterial chromosome remains functional when R4 is deleted (Bates, Asai et al. 1995; Weigel, Messer et al. 2001), the possibility arises that the oligomeric DnaA complexes assembled at the chromosomal ariC locus may be different from those assembled on an ariC plasmid. To address the above possibility, allelic complementation between the chromosomal dnaA46 allele and the different plasmid-encoded alleles was measured upon transformation of an oriC plasmid (pCM959-CmR) (Table 3). The host strain chosen for this experiment (LS 1065, relevant genotype: AariszpKN1592, dnaA46) lacked a functional ariC locus, and was integratively suppressed by an R1 derivative (pKN1592). The reason for selecting this host strain was because its viability is independent of dnaA function. Thus, the frequency of obtaining transforrnants at the nonperrnissive temperature for the dnaA46 strain reflects the ability of DnaA46 protein to oligomerize with the form of DnaA encoded by the plasmid on an oriC minichromosome. Interestingly, the results of Table 3 are similar but not identical to those of Table 2. The T435K mutation alone and combined with Q7A were active in maintaining the ariC plasmid. All other mutations including Q8A and P28L (combined with T435K) were inactive. These results suggest that residues involved in DnaA oligomerization at ariC are conditionally dependent on whether this locus is contained in a plasmid or the bacterial chromosome. As a control for the experiment of Table 3, the ariC plasmid was co-transforrned with plasmids that carried either the dnaA“ or T435K alleles, but into two isogenic strains Table 3. Allelic complementation by dnaA46 and T435K in ariC plasmid maintenance demonstrates the formation of mixed oligomers. Resident Plasmid dnaA gene Plating efficiencya 42°C/30°C pRB 100 dnaA+ 1 .0 pBR322 None 9.9X10“t pAB2 T435K 0.97 pACYC177 None 6.7X10" pLSZ L5A+T435K 5.4X104 pLS3 W6A+T435K 6.2X10‘4 pLS4 Q7A+T435K 1 .0 pLSS Q8A+T435K 1 .0 pLS 1 C9Y+T435K 9.0X10‘3 pLS6 P28L+T435K 0.98 aLS1065 (dnaA46 AariC::pKN1592) was transformed with the indicated plasmids harboring either the indicated dnaA allele or the vector as a control, and the transformation mixture was plated at 30°C and 42°C (see Table 5). These competent cells were transformed with ariC plasmid pCM959-CmR and the transformation was plated at 30°C and 42°C on LB media supplemented with 50 pg/ml ampicillin and 25 pg/ml chloramphenicol to select for both plasmids. 65 Table 4. The T435K allele is defective in maintenance of an oriC plasmid. Plasmid“ co—transformed LS 1062 (AariC, AdnaA)" LS 1061 (AariC)b with pCM959-Cm" pRBlOO (dim/1”) 1.0 1.1 pBR322 1.7x10:3 1.0 pAB2 (T435K) 4.4x10‘2 1.0 pACYC177 2.9x10:3 1.1 The indicated plasmids (100 ng each) were separately co-transformed with pCM959-ka (200 ng) into LS1062 (AariszpKN1562, dnaA850zzTn10), LS 1061 (AariszpKN 1562), or MC1061 as described in "Experimental Procedures." After electroporation, the transforrnants were grown in SOC media for 20 min, and dilutions were plated on LB media containing 100 pg/ml ampicillin and 25 pg/ml chloramphenicol, and plates were incubated for 24 hr at 37°C. In separate experiments, 100 ng of pRB 100 DNA was determined to be near saturation for transformation of MC1061. The amounts of the respective plasmids were chosen to favor uptake of both plasmids on co-transforrnation. The number of co-transforrnants obtained for the indicated strains was divided by the number of co-transforrnants obtained (5 x 10’ per reaction) with the control strain MC 1061. 66 that differed from the dnaA46 host strain used in Table 3. One strain was disrupted at the dnaA locus (dnaA850::Tn 10) whereas the other was dnaA". Both strains lacked essential sequences within oriC (hence AariC and integratively suppressed by the R1 derivative pKN1562). Because T435K is defective in initiation from oriC due to its inability to bind to the DnaA box sequence (Walker et al., manuscript in preparation), the result that T435K failed to maintain the oriC plasmid in the null dnaA host was expected (Table 4). In comparison, the wild type dnaA gene maintained the oriC plasmid when either plasmid-bome (in the AoriC, AdnaA host strain) or encoded by the bacterial chromosome (in the AoriC, dnaA+ host strain). These results confirm that the activity of T435K in allelic complementation is due to a physical interaction between T435K and DnaA46 protein. Immunoblot analysis indicates that the mutant proteins are expressed at levels sujficient to support initiation. The assumption in the above experiments is that the steady state levels of the plasmid-encoded mutant proteins are sufficient to sustain DNA replication. Because we have found that some mutant forms of DnaA protein are exquisitely sensitive to proteolysis ((Sutton and Kaguni 1997), data not shown), we performed immunoblot analysis to determine the steady state levels of the N-terminal mutant proteins. To exclude the contribution of chromosomally encoded DnaA protein in the analysis, a AdnaA host strain was used. Control immunoblot experiments were performed with plasmid-free strains, showing that the dnaA+ strain (LS 1061) expressed DnaA but that the isogenic AdnaA strain (LS 1062) did not (data not shown). These results confirm that the levels of the gene products being measured in the AdnaA host strain are indeed plasmid-encoded. Expression of the different plasmid-bome dnaA alleles in the AdnaA host strain (LS 1062) was then analyzed. Compared to the level of wild type DnaA, the level of T435K was elevated (Table 5). Inasmuch as expression of 67 Table 5. Steady state levels of mutant DnaA Proteins Plasmid Protein Relative level‘I pRB 100 DnaA" l .0 pAB2 T435K 1 .7 pLS2 L5A+T435K 0.5 pLS3 W6A+T435K 0.6 pLS4 Q7A+T435K 0.9 pLSS Q8A+T435K 0.6 p151 C9Y+T435K 0.2 pLS6 P28L+T435K 0.6 pKC596 DnaA‘ 0.07 ‘ The steady state levels of mutant proteins were determined by immunoblot analysis as described (Simmons and Kaguni, 2003). Whole cell lysates of LSlO62 (dnaA850::Tn/0) bearing the respective plasmid-bome dnaA alleles were analyzed by immunoblotting. Purified DnaA protein was analyzed in parallel to confirm that the immunoreactive species was indeed DnaA. As the primary antibody, a monoclonal antibody (M43) that is specific for DnaA protein (Marszalek and Kaguni, 1996) was incubated with the membrane. Other steps of the immunoblotting procedure were as described in "Experimental Procedures," but with horseradish peroxidase-conjugated goat anti-mouse antibody as the secondary antibody. The Xray film that captured the chemiluminescence signal was then analyzed quantitatively with Kodak EDAS 120 software, and normalized to the amount of DnaA+ protein encoded by pRB 100. The pBR322 vector of pRB 100 is reported to have a comparable copy number compared to the pACYC177 vector for pAB2, and its derivatives that encode two missense mutations. Plasmid pKC596 is a pETl 1a derivative in which expression of dnaA is from a T7 RNA polymerase promoter. 68 the dnaA gene is autoregulated (reviewed in Messer 2002), the higher level of T435K is consistent with its defect in binding to the DnaA box motif. The difference is not due to the number of the plasmids carrying the T435K or dnaA+ alleles. Both parental plasmids pBR322 and pACYC177 are reported to have comparable copy numbers (Chang and Cohen 1978; Covarrubias, Cervantes et al. 1981). Immunoblot analysis established that the levels of the other mutant proteins relative to wild type DnaA ranged from 0.2 (C9Y + T435K) to near wild type levels (Q7A + T435K) (Table 5). Because the steady state level of C9Y + T435K was the lowest of all of the mutant proteins, its inactivity in complementation may be due to its reduced abundance. To address this possibility, we examined the amount of DnaA protein expressed from a plasmid (pKC596) in which the wild type gene but without the natural dnaA promoter region was placed downstream and under control of the T7 gene 10 promoter (Carr and Kaguni 1996). Compared to the level encoded by pRB 100 (in which dnaA+ expression is from the dnaA promoters) (Braun, O'Day et a1. 1985), the level of DnaA expressed by pKC596 was about 14-fold reduced (Table 5). As the host strain lacks T7 RNA polymerase, we presume that read-through transcription from an upstream promoter accounts for this level of DnaA protein. Given that this amount is sufficient to complement the temperature sensitivity of a dnaA46 strain (LS 1073; data not shown), the results indicate that the inactivity of the C9Y+ T435K allele in complementation is not due to its lower abundance, but its inability to oligomerize at oriC in the bacterial chromosome. Glutaraldehyde crosslinking of DnaA. To measure oligomer formation biochemically, we relied on glutaraldehyde as a crosslinking agent. As DnaB in the presence of Mg2+ is well characterized to exist as a hexamer of identical subunits (Bujalowski, Klonowska et al. 1994), DnaB protein was used as a control to optimize the 69 Figure 1. Glutaraldehyde crosslinking of DnaA confirms that an N -terminal region is required for oligomerization. Panel A. DnaB protein was incubated with 0.01% glutaraldehyde as described in "Experimental Procedures" at room temperature for the indicated lengths of time. As a negative control, DnaB was incubated in parallel but without glutaraldehyde treatment (lane 1). Samples were analyzed by electrophoresis in a 10% polyacrylamide gel under denaturing conditions. After transfer to a membrane, DnaB was detected by immunoblotting. Panel B. Purified DnaA” and an N -terminal deletion (DnaAA62) were incubated at room temperature for 10 min with 0.01% glutaraldehyde, where indicated, as described in "Experimental Procedures." Samples were analyzed by electrophoresis in an 8% polyacrylamide gel under denaturing conditions. DnaA protein was detected by immunoblotting. The electrophoretic mobilities of molecular weight standards are indicated. Arrows indicate the crosslinked complexes. 70 | | + > 9.3 m 035 933% d M o A m m A M w a it: ‘11 I... ii. ‘I .33 I AI 11 | 39m ‘I IV an.— I ‘1 I d S mob I I mob . ‘1 .230 :13 on n a m 3 on 9:83.333 . + + + + + . + . + 71 method. Shown in Figure 1A, incubation of DnaB with 0.01% glutaraldehyde resulted in the time-dependent appearance of crosslinked oligomers. In the absence of glutaraldehyde, these oligomers were not observed. As judged by SDS-polyacrylamide gel electrophoresis, the electrophoretic mobilities of the different crosslinked species of DnaB correspond well to those expected of a dimer, trimer and higher order complexes up to hexameric DnaB. DnaA protein or a truncated form (DnaAA62) lacking the N-terrninal 62 amino acids including those under study was then incubated with glutaraldehyde under the above conditions but for a fixed incubation period (10 min). A representative experiment among several is shown in Figure 18. Glutaraldehyde treatment of DnaA” protein resulted in the appearance of crosslinked oligomers. Their electrophoretic mobilities suggest the formation of a trimer, and a DnaA tetramer. No other crosslinked oligomers were observed in the upper portion of the gel, including the region of the wells. The diffuse appearance of the crosslinked oligomers may be due to intramolecular crosslinking that then influences the shape of the polypeptides and their relative electrophoretic mobilities under the conditions of denaturing electrophoresis. The negative control of DnaAA62 incubated without glutaraldehyde revealed a trace amount of a species expected for dimeric DnaA. We presume that its presence is due to its residual resistance to denaturation under reducing conditions because its appearance was variable among experiments. Glutaraldehyde treatment of the deletion mutant resulted in a slightly increased amount of dimeric protein, but other oligomers were not observed. In other experiments, crosslinking of DnaA+ with ethylene glycobis (sulfosuccinimidylsuccinate) resulted in the formation of complexes whose electrophoretic mobilities were similar to those expected of a dimer, trimer and tetramer, but the bands were not diffuse (data not shown). DnaAA62 treated with this agent gave rise to a complex expected of a crosslinked dimer. These results indicate that a domain 72 within the N-terminal 62 amino acids that is lacking in the truncated protein is required to form oligomers more complex than dimers. These observations provide additional evidence to support the results from genetic assays that N-terminal amino acids function in DnaA oligomer formation. 73 Discussion Oligomer formation of DnaA. The E. coli replication origin contains five DnaA boxes that are recognized by DnaA protein (Fuller and Kornberg 1983; Matsui, Oka et al. 1985). Estimates based on electron microscopy of 20-40 DnaA monomers bound to oriC have been recently refined by quantitative immunoblot analysis (Carr and Kaguni 2001). The latter studies indicate that ten DnaA monomers are bound to oriC in a replication intermediate termed the prepriming complex. Its formation involves the recruitment of DnaB from the DnaB-DnaC complex by DnaA bound to oriC . These ratios of DnaA monomers per oriC suggest an interaction among DnaA monomers. We now have direct evidence that oligomer formation at oriC is required for initiation to occur. Furthermore, the results indicate that leucine 5, tryptophan 6 and cysteine 9 participate in oligomer formation regardless of whether oriC is at its normal chromosomal location, or carried in a plasmid. In contrast, glutamine 8 and proline 28 function in oligomer formation only on an oriC plasmid. Apparently, the DnaA-oriC complex assembled on a plasmid is similar but not identical structurally to that assembled at oriC in the bacterial chromosome. Other studies implicate N-terminal residues 2 through 86 in DnaA oligomerization (Weigel, Schmidt et al. 1999). Two independent approaches were used to support this conclusion. One relied on streptavidin beads to which DnaA or N- terminal deletions were attached via biotinylation, followed by the measurement of the binding of wild type DnaA to the protein-coated beads. The second utilized the E. coli one hybrid system that relies on transcriptional repression of the lambda PL promoter. The cited study involved substitution of the C-terrninal dimerization domain of lambda CI repressor with resides 2 through 86 from the N-terrninal region of DnaA protein. This region was shown to replace the dimerization domain of CI repressor so that the hybrid protein is active in transcriptional regulation. However, the reverse is not true. The 74 dimerization domain of CI repressor cannot replace this region of DnaA to support pSC101 replication (Seitz, Weigel et al. 2000). Apparently, this region of DnaA has other functions in addition to oligomerization. It has been suggested that leucine 3, leucine 10 and leucine 17 near the N- tenninus form a leucine zipper motif, and that this structure functions in DnaA oligomer formation (Schaper and Messer 1997; Mirna, Makise et a1. 2002). Codons for these residues were mutated by Mirna et al. to encode serine, and the respective alleles were assayed in complementation of the temperature sensitive phenotype of a dnaA508 strain. The plasmid-bome L3S and LIOS alleles were as active as dnaA*, suggesting that these amino acids do not form a leucine zipper motif. The difficulty in interpreting these results is that it is not clear if the mutant proteins are active by themselves, or if complementation requires the formation of mixed complexes of either L3S or L10S and DnaA508'. As noted below, we suspect that the defect of the dad/1508 allele is not in oligomer formation. The L1 75 mutation was not active in complementation, and the mutant protein bound poorly to GST-DnaA complexed to glutathione agarose beads (Mirna, Makise et al. 2002). Interestingly, this substitution resides at the end of the predicted or-helix (residues 3-16; (Sutton and Kaguni 1997)) that L5, W6 and C9 reside in. As the (at-helix is formed by 3.6 amino acids per helical turn, W6 and C9 are on the same face of this structure. L10 is on the opposing face. These results support the model that amino acids L5, W6 and C9, which are absolutely required for oligomer formation, reside in an a-helical structure, with L17 positioned at the end and possibly on the same side as W6 and C9. In this model, these amino acids in this structure participate directly in oligomer formation. ' Individual dnaA alleles encoding L3A and L10A amino acid substitutions are active in initiation on an oriC plasmid in viva, indicating that leucine 3 and leucine 10 (30% as active as dnaA“) are not critical for the activity of DnaA protein (M. Felczak and J .M. Kaguni, unpublished results). 75 That this region also functions in interaction with DnaB is based on results with a truncated form of DnaA lacking the N-terrninal 62 residues. This mutant protein failed to retain DnaB in the prepriming complex whereas wild type DnaA protein was active (Sutton, Carr et al. 1998). In support, Seitz et al. concluded that amino acids 24-86 of DnaA are required for interaction with DnaB (Seitz, Weigel et al. 2000). Earlier, we had isolated a collection of nonfunctional missense mutations of dnaA that mapped within or near this N-terminal region . Included in this collection are the C9Y and P28L alleles studied in this report. In our in viva method to measure oligomer formation as reflected by intergenic complementation, we show that cysteine 9 is required, and proline 28 participates only when oligomers assemble on an oriC plasmid. The dnaA508 allele encodes P28L and T801 substitutions, which confer a temperature sensitive phenotype. We had previously shown that P28L by itself was temperature sensitive for pSC101 replication and, for replication of an oriC plasmid in viva (M. Felczak and J. M. Kaguni, unpublished results). The results described above show that P28L is conditionally required for oligomer formation on an oriC plasmid. Weigel et al. had examined the function of the N-terminal region derived from the dud/1508 allele by substitution of the dimerization domain of CI repressor with residues 2 through 86 of DnaA508 (Weigel, Schmidt et al. 1999). The hybrid protein bearing these substitutions was almost as active as a similar construct containing the corresponding region from wild type DnaA protein in transcriptional repression. Although the individual contribution of the T801 substitution was not examined in the cited study, one interpretation is that the biochemical defect of the P28L substitution is not in the failure to oligomerize. T435K and suppression of dnaX( Ts). The dnaX gene encodes the tau and gamma subunits of DNA polymerase III holoenzyme . These subunits function to load the beta clamp, also known as the DnaX complex, of the DNA polymerase, conferring processive 76 DNA synthesis. The T435K mutation was isolated originally as a cold-sensitive suppressor of the dnaX2016 allele . Two models have been proposed to explain the suppression mechanism (Walker eta1., manuscript in preparation). In the model of stabilization of tau and/or gamma subunit function, T435K is proposed to interact and sustain tau and/or gamma activity so that bacterial growth can occur. In the second model, T435K is reduced in its activity in initiation of DNA replication from oriC, resulting in less frequent initiations. The partially active tau and/or gamma can now keep up at a level sufficient to provide viability. Under either model, T435K is active in supporting initiation from chromosomal oriC. The results described in our study are consistent with in vitro results showing that T435K is defective in binding to the DnaA box motif, and can augment the activity of a limiting level of wild type DnaA protein via the formation of mixed complexes. Model of DnaA at oriC. Our results presented here support a model that a monomer of DnaA46 bound to oriC, but impaired in ATP binding activity, interacts with an N -terminal domain of a T435K monomer to localize the latter to oriC. Whereas each mutant protein alone is inert, the mixed oligomer is active in sustaining DNA replication from oriC. Under this model, a separation of labor among monomers occurs in which the function of DNA binding is separated from ATP binding. However, the form of DnaA that is functional in initiation is likely to be more complex than a dimer of DnaA. Based on the glutaraldehyde crosslinking results which suggest that DnaA oligomerizes to form a tetramer (Figure 1), we speculate that each dimer at each DnaA box then interacts with another bound dimer to assemble a DnaA tetramer. Other results support the notion of a DnaA tetramer. In the one-hybrid system, cooperative binding of a pair of CI repressor dimers at two of the three CI operator sequences assembles a tetramer of CI that is required for transcriptional repression of the lambda PL promoter . The ability of residues 2 through 86 of DnaA to substitute for the dimerization domain of CI repressor in 77 llllllé mill; transcriptional repression is consistent with a DnaA tetramer as the active form in initiation at oriC (Weigel, Schmidt et al. 1999). 78 Acknowledgements This research was supported by grant GM33992 from the National Institutes of Health, and by the Michigan Agricultural Experiment Station. We thank Drs. Andrew Wright (Tufts University), and James Walker (University of Texas, Austin) for plasmids pRB 100 and pAB2, respectively. 79 References Bates, D. B., T. Asai, Y. Cao, M. W. Chambers, G. W. Cadwell, E. Boye and T. Kogoma (1995). "The DnaA box R4 in the minimal oriC is dispensable for initiation of Escherichia coli chromosome replication." Nucleie Acids Res 23(16): 3119-3125. Bell, S. P. (2002). "The origin recognition complex: from simple origins to complex functions." Cenes Dev 16(6): 659-72. Blinkowa, A. L. and J. R. Walker (1990). "Programmed ribosomal frameshifting generates the Escherichia coli DNA polymerase III gamma subunit from within the tau subunit reading frame." Nucleic Acids Res 18(7): 1725-9. Bramhill, D. and A. Kornberg (1988). "Duplex opening by dnaA protein at novel sequences in initiation of replication at the origin of the E. coli chromosome." Cell 52(5): 743-755. Braun, R. E., K. O'Day and A. Wright (1985). "Autoregulation of the DNA replication gene dnaA in E. coli K-12." Cell 40(1): 159-169. Buhk, H. J. and W. Messer (1983). "The replication origin region of Escherichia coli: nucleotide sequence and functional units." Cene 24(2-3): 265-79. Bujalowski, W., M. M. Klonowska and M. J. Jezewska (1994). "Oligomeric structure of Escherichia coli primary replicative helicase DnaB protein." J Biol Chem 269(50): 31350-31358. Carr, K. M. and J. M. Kaguni (1996). "The A184V missense mutation of the dnaA5 and dnaA46 alleles confers a defect in ATP binding and therrnolability in initiation of Escherichia coli DNA replication." Mol Microbiol 20(6): 1307-1318. Carr, K. M. and J. M. Kaguni (2001). "Stoichiometry of DnaA and DnaB protein in initiation at the Escherichia coli chromosomal origin." J Biol Chem 276(48): 44919-25. Chang, A. C. and S. N. Cohen (1978). "Construction and characterization of amplifiable multicopy DNA cloning vehicles derived from the P15A cryptic miniplasmid." J Bacteriol 134(3): 1141-56. 80 Covarrubias, L., L. Cervantes, A. Covarrubias, X. Soberon, I. Vichido, A. Blanca, Y. M. Kupersztoch-Partnoy and F. Bolivar (1981). "Construction and characterization of new cloning vehicles. V. Mobilization and coding properties of pBR322 and several deletion derivatives including pBR327 and pBR328." Corie 13(1): 25-35. Crooke, E., R. Thresher, D. S. Hwang, J. Griffith and A. Kornberg (1993). "Replicatively active complexes of DnaA protein and the Escherichia coli chromosomal origin observed in the electron microscope." J. Mol. Biol. 233(1): 16-24. Dadd, I. B., A. J. Perkins, D. Tsemitsidis and J. B. Egan (2001). "Octamerization of lambda CI repressor is needed for effective repression of P(RM) and efficient switching from lysogeny." Cenes Dev 15(22): 3013-22. Erzberger, J. P., M. M. Pirruccello and J. M. Berger (2002). "The structure of bacterial DnaA: implications for general mechanisms underlying DNA replication initiation." Embo J 21(18): 4763-73. Fang, L., M. J. Davey and M. O'Donnell (1999). "Replisome assembly at oriC, the replication origin of E. coli, reveals an explanation for initiation sites outside an origin." Mol Cell 4(4): 541-53. Flower, A. M. and C. S. McHenry (1990). "The gamma subunit of DNA polymerase III holoenzyme of Escherichia coli is produced by ribosomal frameshifting." Proc Natl Acad Sci U S A 87(10): 3713-7. F ujita, M. Q., H. Yoshikawa and N. Ogasawara (1990). "Structure of the dnaA region of Micrococcus luteus: conservation and variations among eubacteria." Gene 93(1): 73-78. Fuller, R. S., B. E. Funnell and A. Kornberg (1984). "The dnaA protein complex with the E. coli chromosomal replication origin (oriC) and other DNA sites." Cob 38(3): 889-900. Fuller, R. S. and A. Kornberg (1983). "Purified dnaA protein in initiation of replication at the Escherichia coli chromosomal origin of replication." Proc. Natl. Acad. Sci. U S A 80(19): 5817-5821. 81 Funnell, B. E., T. A. Baker and A. Kornberg ( 1987 ). "In vitro assembly of a prepriming complex at the origin of the Escherichia coli chromosome.” JBlol Chem 262(21): 10327-10334. Gines, C. E., A. Blinkova and J. R. Walker (1995). "Mutations in Escherichia coli dnaA which suppress a dnaX(Ts) polymerization mutation and are dominant when located in the chromosomal allele and recessive on plasmids." J. Bacteriol. 177(3): 705-715. Gatfried, F. and J. A. Wechsler (1977). "Dominance of dnaA+ to dnaA in Escherichia coli." J Bacteriol 130(2): 963-4. Hansen, E. B., T. Atlung, F. G. Hansen, 0. Skovgaard and K. van Meyenburg (1984). "Fine structure genetic map and complementation analysis of mutations in the dnaA gene of Escherichia coli." Mol Gen Genet 196(3): 387-96. Hansen, F. G., S. Koefoed and T. Atlung (1992). "Cloning and nucleotide sequence determination of twelve mutant dnaA genes of Escherichia coli." Mol. Gen. Genet. 234(1): 14-21. Hu, J. C., E. K. O'Shea, P. S. Kim and R. T. Sauer (1990). "Sequence requirements for coiled-coils: analysis with lambda repressor-GCN4 leucine zipper fusions." Scienee 250(4986): 1400-3. Lee, Y. S. and D. S. Hwang (1997). "Occlusion of RNA polymerase by oligomerization of DnaA protein over the dnaA promoter of Escherichia coli." J Biol Chem 272(1): 83-88. Marszalek, J. and J. M. Kaguni (1994). "DnaA protein directs the binding of DnaB protein in initiation of DNA replication in Escherichia coli." J. Biol. Chem. 269(7): 4883-4890. Marszalek, J ., W. Zhang, T. R. Hupp, C. Margulies, K. M. Carr, 8. Cherry and J. M. Kaguni (1996). "Domains of DnaA protein involved in interaction with DnaB protein, and in unwinding the Escherichia coli chromosomal origin." J Biol Chem 271(31): 18535-18542. Matsui, M., A. Oka, M. Takanami, S. Yasuda and Y. Hirota (1985). "Sites of dnaA protein-binding in the replication origin of the Escherichia coli K-12 chromosome." J Mol Biol 184(3): 529-33. 82 Messer, W. (2002). "The bacterial replication initiator DnaA. DnaA and oriC, the bacterial mode to initiate DNA replication." FEMS Microbiol Rev 26(4): 355-74. Miller, J. H. (1992). A short course in bacterial genetics, Cold Spring Harbor Laboratory Press. Mima, S., M. Makise, M. Katerasawa, T. Tsuchiya and T. Mizushima (2002). "Conserved hydrophobic amino acid residues in the N-terminal region of DnaA protein are involved in DnaA-DnaA interaction." Biochem J 365(Pt 3): 881-7. Oka, A., K. Sugimoto, M. Takanami and Y. Hirota (1980). "Replication origin of the Escherichia coli K-12 chromosome: the size and structure of the minimum DNA segment carrying the information for autonomous replication." Mol, Cen. Genet. 178(1): 9-20. Schaper, S. and W. Messer (1997). "Prediction of the structure of the replication initiator protein DnaA." Proteins 28(1): 1-9. Seitz, H., C. Weigel and W. Messer (2000). "The interaction domains of the DnaA and DnaB replication proteins of Escherichia coli." Mol Mierobiol 37(5): 1270-9. Sekimizu, K., D. Bramhill and A. Kornberg (1987). "ATP activates dnaA protein in initiating replication of plasmids bearing the origin of the E. coli chromosome." Cell 50(2): 259-265. Simmons, L.A. and J.M. Kaguni (2003) “The dnaAcos allele of E. coli: Hyperactive initiation is due to substitution of A184V and Y271H, resulting in defective ATP binding and aberrant DNA replication control”. Mol. Microbiology. 47:755-765 Sutton, M. D., K. M. Carr, M. Vicente and J. M. Kaguni (1998). "E. coli DnaA protein: the N-terrninal domain and loading of dnaB helicase at the E. coli chromosomal origin." J. Biol. Chem. 273: 34255-34262. Sutton, M. D. and J. M. Kaguni (1997). "The Escherichia coli dnaA gene: four functional domains." Mol Biol 274(4): 546-61. Sutton, M. D. and J. M. Kaguni (1997). "Novel alleles of the Escherichia coli dnaA gene." J Mol Biol 271(5): 693-703. Sutton, M. D. and J. M. Kaguni (1997). "Threonine 435 of Escherichia coli DnaA protein confers sequence- specific DNA binding activity." J Biol Chem 272(37): 23017-24. 83 Tsuchihashi, Z. and A. Kornberg (1990). "Translational frameshifting generates the gamma subunit of DNA polymerase III holoenzyme." Proc Natl Acad Sei U S A 87(7): 2516-20. Walker, J. R., J. A. Ramsey and W. G. Haldenwang (1982). "Interaction of the Escherichia coli dnaA initiation protein with the dnaZ polymerization protein in viva." Proc. Natl. Acad. Sci. U S A 79(10): 3340-3344. Weigel, C., W. Messer, S. Preiss, M. Welzeck, Morigen and E. Boye (2001). "The sequence requirements for a functional Escherichia coli replication origin are different for the chromosome and a minichromosome." Mol Mierobiol 40(2): 498- 507. Weigel, C., A. Schmidt, H. Seitz, D. Tungler, M. Welzeck and W. Messer (1999). "The N -terminus promotes oligomerization of the Escherichia coli initiator protein DnaA." Mol Microbiol 34(1): 53-66. Waelker, B. and W. Messer (1993). "The structure of the initiation complex at the replication origin, oriC, of Escherichia coli." Nucleic Acids Res 21(22): 5025- 5033. Zyskind, J. W., J. M. Cleary, W. S. Brusilow, N. E. Harding and D. W. Smith (1983). "Chromosomal replication origin from the marine bacterium Vibrio harveyi functions in Escherichia coli: oriC consensus sequence." Prm Natl Acad Sci U S A 80(5): 1164-1168. 84 Chapter III The dnaAcas Allele of E. coli: Hyperactive Initiation is due to Substitution of A184V and Y271H, resulting in Defective ATP Binding and Aberrant DNA Replication Control 85 Abstract Chromosomal DNA replication is regulated at the level of commitment to this biochemical pathway. In E. coli, DnaA protein appears to regulate this process. A mutant form, DnaAcos carrying four amino acid substitutions, is apparently defective in responding to regulatory signals because it induces hyperactive initiation from the bacterial replication origin (oriC). In this report, the phenotype of hyperactive initiation is shown to be the result of two specific amino acid substitutions. One (A184V) immediately adjacent to a Walker A box (P loop motif) results in a defect in ATP binding (Carr and Kaguni, Mol. Microbiol. 20: 1307-1318, 1996). The second amino acid substitution (Y271H) appears to stabilize the activity of the mutant protein carrying the A184V substitution. The mutant protein carrying both amino acid substitutions (A184V + Y271H) is defective in modulating the frequency of initiation from oriC, as demonstrated by marker frequency analysis of oriC and a locus near the replication terminus. These results indicate that a defect in ATP binding results in aberrant control of DNA replication. 86 IO' Introduction Duplication of the genome in free-living organisms occurs once per cell cycle, and is a highly regulated event. In E. coli, this event occurs at a specific time in the bacterial cell cycle at a particular ratio of chromosomal origins to cell mass ((Donachie 1965) reviewed in ref. (Herrick, Kohiyama et al. 1996)). As evidence indicates that bacterial DNA replication is regulated at the level of initiation of DNA synthesis, the proposed models address the regulation of this process either at the stage prior to its onset, or after semi-conservative synthesis in order to block unscheduled rounds of DNA replication (Hansen, Christensen et a1. 1991; Kitagawa Ozaki et a1. 1998; Katayama 2001). In a model addressing the latter process, the availability of the chromosomal origin sequence (oriC) to the replication machinery is controlled by sequestration (Campbell and Kleckner 1990; Lu, Campbell et a1. 1994). E. coli chromosomal DNA is normally methylated at GATC sites by DNA adenine methylase. As Squ binds preferentially to hemi-methylated DNA (Brendler, Abeles et al. 1995; Slater, Wold et al. 1995), the product of one round of DNA replication, the chromosomal origin is occluded for a period of time to prevent another round of DNA replication until proper partitioning of daughter chromosomes has occurred. Because DnaA protein initiates DNA replication from the bacterial replication origin, other models suggest that its activity is regulated so that DNA replication occurs at the proper time (Sekimizu, Bramhill et a1. 1987; Katayama, Fujimitsu et a1. 2001). Upon binding to specific sets of sequences termed DnaA boxes within oriC, the ATP- bound form of DnaA protein then opens the parental duplex DNA to form an intermediate required in the initiation process (Bramhill and Kornberg 1988). Because DnaA complexed to ADP, and mutants that bind poorly to ATP are defective in initiation from oriC, models have centered on the ATP binding activity of DnaA (Sekimizu, Bramhill et a1. 1987; Carr and Kaguni 1996; Katayama, Fujimitsu et al. 2001). DnaA is a 87 weak ATPase. When bound to an adenine nucleotide, the cofactor is released on interaction with acidic phospholipids (Castuma, Crooke et al. 1993). Based on these observations, it has been suggested that the interaction of DnaA with the cytoplasmic membrane provides a mechanism whereby DnaA in a complex with ADP can be recycled to bind ATP, initiating another round of DNA replication (Sekimizu, Bramhill et al. 1987; Crooke 2001). A related model is based on the observation that ATP bound to DnaA is hydrolyzed on addition of the beta subunit of DNA polymerase III holoenzyme and a newly identified protein named Hda (Kato and Katayama 2001). Specifically, the interaction of DnaA at oriC with beta as a component of the replication fork machinery and Hda stimulates the hydrolysis of ATP bound to DnaA. This event marks the transition between the stages of initiation and elongation of DNA replication. Because the DnaA-ADP complex is less active in initiation, this mechanism provides a means to prevent initiation of DNA replication at an unscheduled time. Another model is based on the availability of DnaA for initiation (Hansen, Christensen et al. 1991). Elevated dnaA expression results in an increased frequency of initiation from oriC (Atlung, Lobner Olesen et al. 1987; Lobner-Olesen, Skarstad et al. 1989; Nyborg, Atlung et a1. 2000). This observation suggests that the level of DnaA protein is normally limiting. In its support, the datA locus appears to be a site that titrates excess DnaA protein. Deletion of either this site, or essential DnaA boxes within it results in unscheduled initiations from oriC (Kitagawa, Ozaki et a1. 1998; Ogawa, Yamada et al. 2002). If DnaA protein acts to regulate the frequency of initiation from oriC, a prediction is that dnaA mutations defective in this regulated pathway should be obtained. Indeed, the dnaAcos allele appears to fail to respond to a regulatory signal(s), resulting in hyperactive initiation (Braun, O'Day et al. 1987). This intragenic suppressor was isolated by its ability to restore growth to a dnaA-46 strain at the nonperrnissive temperature (42 88 °C). At 30 °C dnaAcas is cold-sensitive, a phenotype linked to overinitiation from oriC. Biochemical studies of DnaAcos suggest that it has elevated DNA replication activity in a replication system composed of an oriC-containing plasmid as a template and a crude enzyme fraction as the source of other required replication proteins (Katayama and Kornberg 1994). Structure-function studies correlate functional domains of DnaA protein to its primary structure (Figure 1). To understand the altered biochemical properties of DnaAcos, we sought to identify the mutation that gives rise to its hyperactive initiation activity. The dnaAcas allele encodes four amino acid substitutions (Q156L, A184V, H252Y, Y271H) (Braun, O'Day et al. 1987). Two substitutions are encoded by the dnaA46 allele (A184V, H252Y) (Johanson, Haynes et al. 1986); dnaAcas was isolated as an intragenic suppressor of dnaA46(Ts) (Braun, O'Day et al. 1987). In this study, expression of plasmid-bome dnaAcas was placed under control of the native dnaA promoters. When the dnaAcas plasmid was introduced into a dnaA+ host strain, we observed that dnaAcas is dominant-negative. This phenotype correlates with hyperactive initiation as demonstrated by a quantitative PCR assay used to measure the relative ratio of loci at oriC or near terC. Relying on the dominant-negative phenotype of dnaAcas as a genetic assay, each possible combination of the dnaAcas mutations was then tested. The results show that the combination of A184V and Y2 71H causes the dnaAcas phenotype and hyperactive initiation. We conclude that A184V alone results in overinitiation due to a defect in ATP binding. The Y271H substitution appears to confer stability at elevated temperature to the mutant protein carrying A184V. 89 Table 1. List of Strains Strain Genotype Parent Source LK211 asnA”, asnB32, thi-1, reIA1, spoT1, ER Koppes it“, AoriG:pKN1562 (clockwise), Kmr and Nordstrom, 1986 LK348 asnA“, asnB32, thi-1, re/A1, spoT1, ER K0ppes it“ AariG:pKN1562 (counterclockwise), and Km' Nordstrom, 1986 AQ9433 trpA9605, his-29, praB (or praC), ilv, AQ3519 T- Kogoma metB1, thyA, (deaB or deaC), mhzzcat, dnaABSQzTn 10 W3110 F', x, IN (rmD-rmE)1, Iph-1 Lab Stock L821] asnA“, asnB32, thi-1, reIA 1, spoT1, LK211 This Work it“ , AoriG:pKN1562 (clockwise), Km' dnaA85azTn 10 LS] 10 F', X, lN(rrnD-rmE)1, rph-1, AoriC:: W3110 This Work pKN1562 (clockwise), Krnr LS778 F', x, lN(rmD-rmE)1, rph-1,tna::Tn 10, W3110 This Work dnaA46 LSlO73 araD139, AaraC, leu7697, A(Iac)X174, MC1061 This Work galU, galK, hst2(rk‘,mk*), strA, mcrA, mch1, tnazzTn 10, dnaA46 LSlO61 araD139, AaraC, leu7697, MC1061 This Work A(Iac)X174,galU, galK, hdei2(rk',mk*), strA, mcrA, mch1, AariG:pKN1562(clockwise) Kmr LSlO62 araD139, AaraC, leu7697, A(Iac)X174, L81061 This Work galU, gaIK, hst2(r.;,mk*), strA, mcrA, mch1 , AoriG:pKN1562(clockwise) Km', dnaA850::Tn 10 MC1061 araD139, AaraC, leu7697, A(Iac)X174, Lab Stock galU, galK, hst2(rk"mk*), strA, mcrA, mch1 9O Table 2. List of Plasmids Plasmids Characteristics Vector or Source Parent Plasmid pRB100 Ap’, dnaA‘, dnaA promoters pBR322 Braun et al., 1985 pLS 103 Ap', H252 Y pRB 100 This work pLS 104 Ap', Q156L pRB100 This work pLS105 Ap’, A184V pRB100 This work pLSlO6 Ap', Y271H pRB100 This work pLS 107 Ap', Q156L+A184V pRB100 This work pLS 108 Ap', Q156L+H252Y pRB 100 This work pLS109 Ap', Q156L+Y27IH pRB100 This work pLSl 10 Ap', AI84V+H252Y pRB100 This work pLSl 11 Ap', A184 V+Y2 71H pRB 100 This work pLS112 Ap', H252 Y+Y2 71H pRB 100 This work pLSl 13 Ap', Q156L+A184V+H252Y pRB100 This work pLSl l4 Ap’, AI84V+H252Y+Y271H pRB100 This work pLS115 Ap', Q156L+A184V+Y271H pRB100 This work pLSl 16 Ap', Q1 56L+H252 Y+Y2 71H pRB 100 This work pLSdnaAcos Ap’, dnaAcas pRB 100 This work pLS118 Ap', Y271H, paraBAD ‘ pINGl This work pLS119 Ap’, AI84V+Y2 71H, paraBAD pINGl This work pLS120 Ap', dnaAcas, paraBAD pINGl This work pLS121 Ap', A1 84 V, paraBAD pINGl This work pLS 122 Ap’, Q156L+A184V, paraBAD pINGl This work pDS596 Ap', dnaA”, paraBAD pINGl Hwang and Kaguni, 1988 pBR322 Ap', Tcr Lab Stock pRB37 placUVS'b, dnaAcas pRB25 Braun et al., 1987b pRB25 placUVS', dnaA+ pBR322 Braun et al., 1987 pKC596 Ap’, T7“, dnaA” pET 1 la Carr and Kaguni. 1996C pKC215 Ap', T7, dnaA46 pET 1 la Carr and Kaguni, 1996 pKC-A184V Ap', T7, A184V pET 1 1a Carr and Kaguni. 1996 ¥ pKC-H252Y Ap’, T7, H252Y pET 1 1a Carr and Kaguni. 91 1996 pCM959-Cmr ari C, Cmr unpublished resultsd ‘ paraBAD represents the araBAD promoter under which the respective dnaA alleles is regulated by AraC repressor. These plasmids were constructed by replacing the EcoNI-Xhol restriction fragment of wild type sequence contained in pDSS96 with the corresponding DNA fragment (gel-purified) from the pRB 100 derivative carrying the respective mutation. " placU V5 ’ denotes the lacU V5 promoter which modulates the expression of the respective dnaA alleles. ° In pKC596 and related derivatives, dnaA expression is from a bacteriophage T7 RNA polymerase promoter. dA DNA fragment carrying the gene encoding chloramphenicol acetyltransferase derived from pBR325 (nucleotides 3911 to 5487 of pBR325) was PCR amplified and used to replace an 837 base pair PvuII DNA fragment (nucleotides 3069 to 3905) of pCM959 (Buhk and Messer 1983). The PCR primer sequences inserted in construction of pCM959-Cmr introduced BamHl sites at the junctions between pCM959 and the DNA fragment carrying the chloramphenicol resistance gene (gGATCCGTC and gGACGGATC, respectively, in which the upper case letters represent primer sequences). The structure of the junctions was confirmed by DNA sequence analysis (K. Hupert- Kocurek and J .M. Kaguni, unpublished results). 92 Experimental Procedures Bacteriological methods-Bacteriological media and methods including P1 transduction to construct the strains of Table 1 are as described (Mann, Carrington et al. 1992). Purified DNAs were quantitated spectrophotometrically and electrophoresed on agarose gels to verify that greater than 95% of the DNA was supercoiled. Transformation of the indicated strains was by electroporation (BioRad E. coli Pulser) with 10 ng of plasmid DNA. Various dilutions of the electroporated sample were plated on media containing the appropriate antibiotic and transformation frequencies were assessed after overnight incubation at the indicated temperature. E. coli LS 1062 grows slowly, requiring 24-30 hr incubation at 30 °C. Separation of dnaAcas mutations-The dnaAcas mutations were separated by subcloning and by PCR mutagenesis with overlapping primers containing the desired nucleotide changes. Briefly, A184V, H252Y and dnaA46 were subcloned from pKC215 (dnaA46), pKC-A184 V and pKC-H252Y (Carr and Kaguni 1996) by EcaRl and AvaI cleavage to yield a 1.5 kb fragment containing the respective dnaA mutation(s). These DNA fragments were used to replace the corresponding DNA fragment, but of wild type sequence, carried in the dnaA+ plasmid pRB 100 (Braun, O'Day et al. 1985) to yield the plasmid derivatives in Table 2. dnaAcas was subcloned from pRB37 (Braun, O'Day et al. 1987) as described above in a 1.5 kb EcoRl-Aval restriction fragment. In pRB37, dnaAcas is under lacU V5 promoter control. Substitution of the wild type sequence in pRB 100 by the dnaAcas-containing fragment places the allele under control of the natural dnaA promoters. At this point, the presence of the A184 V mutation in dnaAcas, dnaA46, and the single A184V mutation was confirmed by resistance to SphI cleavage; the mutation that gives rise to the A184V substitution destroys an SphI site. The remaining mutations of dnaAcas were introduced into pRB 100 using the following overlappping primers carrying the respective nucleotide substitutions: Q156L, 93 CG AC unc the 10 mm PC] con mm seq: 165; 11111 I of 11 ban has and PM man 1.11 limp} Erich; luau ACA CGGCGGCTCGCCIGGTGGCGGATAA and the complementary sequence; Y27 1 H, ACC TCGGATCGCCATCCGAAAGAG and the complementary sequence. The underlined nucleotide introduces the corresponding amino acid substitution. To construct the double mutant encoding Q156L and Y271H, the mutagenesis was done sequentially. To construct pLSl 13 encoding substitutions of Q156L + A184V + H252Y, pLSl 10 (A184 V+ H252Y) was used as the template with overlapping primers to introduce the mutation for Q156L. Alternatively, pLS110 served as the template for overlap extension PCR to insert the Y27IH mutation, forming pLS114. pLS115 and pLSl 16 were constructed from pLS 104 (A184V) and pLS 103 (H252Y), respectively, by sequential PCR mutagenesis. DNA sequence analysis of the dnaA coding region including flanking sequences was performed to confirm that each plasmid of Table 2 contained the correct respective mutations. Quantitative PCR assay-Cultures were grown at 30 °C to an optical density at 595 nm of 0.2, then arabinose was added to 1% (w/v) final concentration to induce expression of the indicated dnaA alleles. After continued incubation as indicated, the cells were harvested by centrifugation, frozen in liquid nitrogen, then stored at —70 °C. The isolated bacterial chromosomal DNA (DNeasy, Qiagen) was quantitated spectrophotometrically, and unless indicated 1 ng was used as the template in PCR reactions with respective primers to amplify oriC or relE. PCR reactions were assembled according to the manufacturer’s recommendations (PCR Master mix, Promega), but contained [0:32P]- dATP (1.25 pCi per reaction, 3000 Cmeol, Perkin-Elmer), and amplification was for 23 cycles with Taq DNA polymerase (1.25 units). To measure the amount of PCR- amplified DNA, the radiolabelled product was acid precipitated with 10% (w/v) trichloroacetic acid onto glass fiber filters (GF/C, Whatman), and radioactivity was quantitated by liquid scintillation spectrometry. Primers to amplify oriC are ACAGCGTACAATACGCCACT and ACGGTGAGCACGACGGCTTT. To amplify 94 the relE region which is linked to terC, the primers are GTTGAAGTACTI‘GAGTCACC and CA'I'I‘CAGACTTGAATGCGTG. These primers are identical to those used by Nyborg et al. (Nyborg, Atlung et al. 2000) and amplify relE, relF, and rem. 95 'l. ~41. Figure 1. Functional domains of DnaA protein relative to mutations of dnaAcas. A domain near the N-terminus is involved in oligomerization of DnaA (Sutton and Kaguni 1997; Jakimowicz, Majka et al. 1998) and in retention of DnaB in the oriC prepriming complex (Sutton, Carr et a1. 1998; Seitz, Weigel et al. 2000). A second domain is involved in interaction with DnaB (Sutton, Carr et a1. 1998; Seitz, Weigel et al. 2000). A Rossmann fold constitutes the ATP binding domain (Sutton and Kaguni 1997) followed by a region carrying the Sensor I and II motifs of AAA” family proteins (Koonin 1992; Neuwald, Aravind 1999), and the membrane binding domain (residues 372-381; (Garner and Crooke 1996). The DNA binding domain is near the C-terminus (Roth and Messer 1995; Sutton and Kaguni 1997). Symbols within the functional domains denote amino acid substitutions of DnaAcos (Q156L, A184V, H252Y, and Y271H). The scale at the top indicates length in amino acid residues. The end points for respective domains and the secondary structure prediction by the PHD method have been described (Sutton and Kaguni 1997; Sutton, Carr et al. 1998). 96 m . _ _ _ _ ._ 3o woo woo So now Osmmr 3 Q2 Immm< <93...— . . . a :.....1.-..:..1.......-..ii. iii. .91..-. _ . i... 39038: 0. 03mm 4.8363 3.0 mimic Dambfimfib 3689.0: 02> 5 5o 3 .35 03mm 933 mien 003%: a a 30339:: no.3 maamo. . m. ZoBEmzm m __ msasm 0.603mznmzo: a; is even. 5.. v v .2. v k. fvceem V... ..3v .319 .. e. .55. town? .1: ..,. K as “ 359.03. 003.33 P Oamroom 2.5.3.03 is V 3358 23m 12.x V 30933 were manna 97 Results dnaAcas is dominant-negative to dnaX-The dnaAcas allele that confers hyperactive initiation of DNA replication was originally cloned under lacU V5 promoter control in a pBR322 derivative as attempts to isolate this allele linked to its native dnaA promoters (plasmid-bome) failed (Braun, O'Day et al. 1987). When expression is from the native dnaA promoters, DnaAcos apparently interferes with growth but the wild type protein does not. Likewise, when dnaAcas is under control of the lacU V5 promoter, its induced expression (but not dnaA‘“) also interferes with viability. In contrast, an oriC null strain, in which constitutive stable DNA replication is activated by disruption of mh (Rnase H), did not show growth interference when dnaAcas expression was induced by IPTG (Katayama and Kornberg 1994). Disruption of the oriC locus was required, presumably because of hyperactive initiation from oriC. Because of our interest in identifying the mutation(s) responsible for hyperactive initiation activity, we chose to propagate dnaAcas and combinations of mutations of this allele in an oriC null strain (LS 1061, Table 1). In this host strain, DNA replication is independent of DnaA protein, minimizing the possibility of selecting for mutations in the host genome or in the plasmid-bome dnaA gene that alleviate the lethal phenotype of dnaAcas. As dnaAcas carries multiple mutations, all combinations of them were constructed in a plasmid vector, and their presence verified by DNA sequence analysis (Table 2). These plasmids were then used to transform a wild type strain to assay for dominance over dnaA” (Table 3). The control plasmid (pRB 100) harboring dnaA” (Braun, O'Day et a1. 1985), or pBR322 (data not shown) transformed the duo/1* strain with approximately equal efficiency at both 30 °C and 42 °C, and did not show growth interference. In contrast, the dnaAcas-bearing plasmid (pLSdnaAcos, Table 2) transformed the dnaA“ strain at almost 103-fold lower frequency at both temperatures, indicating that dnaAcas is lethal in a wild type strain when expressed from the dnaA 98 promoters. The dominant-negative effect of dnaAcas over chromosomally encoded dnaA" when expression of DnaAcos is at elevated levels confirms previous observations (Braun, O'Day et al. 1987; Katayama and Kornberg 1994). A184V and Y2 71H are required for the overinitiation phenotype of dnaAcas-This dominant-negative phenotype provided a genetic assay to measure the activity of different combinations of mutations encoded by the dnaAcas allele. Except for the A184 V mutation assayed at 30 °C, none of the single mutations conferred growth interference at either temperature (Table 3). The dominant-negative phenotype of the A184 V mutation at 30 °C that correlates with overinitiation activity confirms previous observations (Nyborg, Atlung et al. 2000). Importantly, this phenotype was not observed at 42 °C to distinguish the behavior of the A184V mutation from that of dnaAcas, suggesting that other mutations contribute to the phenotype of dnaAcas. Analysis of the double mutants showed that H252Y combined with A184V no longer interfered with growth of the dnaA“ host strain (compare lines 5 and 10 of Table 3 under the data for incubation at 30°C). This phenotype was also observed in a dnaA46 strain at 30 °C (data not shown). In contrast, Nyborg et al. observed that a pBR322-derived plasmid carrying these two mutations conferred cold-sensitivity when introduced into either a dnaA+ or dnaA46 strain (Nyborg, Atlung et a1. 2000). The reason for these differences is unknown. Of the double mutants, the A184V + Y2 71H mutant had the same dominant-negative phenotype at both temperatures as dnaAcas. Analysis of the triple mutants supports the conclusion that the combination of A184V and Y271H is responsible as only those triple mutants carrying the A184V and Y2 71H mutations interfered with growth at both temperatures. Growth interference at 30 °C was also observed with the A184V + Y2 71H 99 Table 3. Transformation of W3110 (dnaA", oriC *) with plasmids carrying various dnaA mutations. Plasmid dnaA Allele CFUs‘ at 30°C CFUs at 42°C pRB100 dnaA: 3.3x107 3.2x10" pLSdnaAcos dnaAcas 4.4x 104 b 6.6x 104 b pLSlO3 H252Y 2.1x107 2.4x107 pLS 104 Q156L 3.3x107 3.2x107 pLSlOS A184V 2.8X104 2.3x107 pLSlO6 Y271H 1.8)(107 2.2x107 pLSlO7 Q156L+A184V 3.0X104 3.0)th7 pLS 108 Q156L+H252Y 2.6)(107 2.7x 107 pLS 109 Q156L+Y271H 2.4x107 2.6x107 pLSl 10 A184V+H252Y 2.3x107 1.9)(107 pLSl 11 A184V+Y271H 4.0X104 4.4)(104 pLSl 12 H252Y+Y271H 2.5x107 2.8)(107 pLSl l3 QI56L+AI84V+H252Y 1.6x105 2.4x107 pLSl 14 A184V+H252Y+Y271H 3.2)(104b 8.8)(104b pLSl 15 QI56L+A184V+Y271H 2.4)(104b 6.8X104b pLSl 16 Q156L+H252Y+Y271H 3.3X107 3.2X107 aColony forming units (CFUs) are normalized and expressed as the number of transformants observed per pg of plasmid DNA at the indicated temperature. " Plates of these transformants contained two colony morphologies. Small colonies were assumed to have arisen by spontaneously obtained suppressors and were excluded in the calculation of transformation efficiencies. 100 double mutant and those triple mutants carrying these mutations in a host strain carrying the dnaA46 allele (data not shown). The dominant-negative phenotype requires oriC-We constructed a strain (LS1 10, Table 1) isogenic to the oriC" strain (W3110), but null at the oriC locus and integratively suppressed with a mini-R1 replicon. Such integratively suppressed strains replicate independently of dnaA function (Koppes and Nordstrom 1986; Blinkova, Gines- Candelaria et al. 2000). The different combinations of mutations comprising the dnaAcas allele were used to transform this strain to determine if the dominant-negative phenotype was dependent on initiation from oriC (Table 4). The observation that none of the plasmid-home alleles affected growth of this strain demonstrates that the dominant- negative phenotype of dnaAcas and the double mutant carrying A184V and Y2 71H requires the presence of oriC. Shown below, overinitiation from oriC appears to be the reason for the dominant negative effect. Expression of dnaAcas and the A184V + Y271H double mutant results in hyperactive initiation-The phenotype of dnaAcas and the A184 V + Y2 71H mutant when present in the ariC’ host strain suggests that hyperactive initiation activity from oriC is the cause. To confirm this expectation, a quantitative PCR assay was used. The objective was to determine the relative ratios of the chromosomal origin (oriC) to a locus distant from this site. We chose the relE locus because of its proximity to the replication terminus, terC. Hyperactive initiation with dnaAcas or with A184 V + Y2 71H mutant should result in an elevated ratio of oriC to relE compared to the dnaA” control. We restricted the analysis to only two loci because Nyborg et al. had shown that overinitiation was from oriC as indicated by the reduced abundance of markers with increasing distance from oriC (Nyborg, Atlung et al. 2000). In the following experiments, expression of the respective alleles was induced from a regulated promoter (araB promoter) whereas the 101 Table 4. Transformation of L8] 10 (dnaA", AoriC ) with plasmids carrying various dnaA mutations. Plasmid dnaA Allele CFUs at 30°C CFUs at 42°C pRB 100 dnaA+ 2.9x 10" 2.6x106 pLSdnaAcos dnaAcas 2.0x 106 2.4x 10° pLSlO3 H252Y 1.1x106 1.1x10" pLSlO4 Q156L 1.2x106 1.4x106 pLSlOS A184V 1.9x106 2.1x106 pLS 106 Y271H 5.5x105 5.5x105 pLSlO7 QI56L+A184V 1.1x106 1.1x106 pLS 108 Q] 56L+H252 Y 1.4x106 1.4x 106 pLSlO9 QI56L+Y271H 1.7)(106 1.9X106 pLSl 10 AI84V+H252Y 3.511105 3.7x10’ pLSl ll AI84V+Y271H 1.2X106 1.4x106 pLSl 12 H252Y+Y271H 1.0x106 1.0x106 pLSl 13 QI56L+AI84V+H252Y 2.3x106 2.5x106 pLSl 14 AI84V+H252Y+Y271H 1.0X10t5 1.0X106 pLSl 15 QI56L+AI84V+Y271H 1.1X10‘5 1.0x106 pLSl 16 QI56L+H252Y+Y27IH 1.0x106 1.0X106 102 Colony forming units (CFUs) are expressed as the number of transformants observed per pg of plasmid DNA. transformation experiments relied on elevated expression due to the copy number of the plasmid vector. As one control, we demonstrated that the amount of amplified DNA is proportional to the amount of input template DNA (Figure 2). Both oriC and relE showed a linear relationship for amplification relative to the amount of chromosomal DNA template over at least a 6-fold range. In another control experiment, we found that with increasing time after induced expression of dnaA", the ratio of oriC to relE proportionately increased (data not shown). This observation also confirms previous observations that elevated dnaA+ expression results in an increased frequency of initiation from oriC (Atlung and Hansen 1993). However, the replication forks do not continue to completion but stall at apparently random sites on the chromosome. As a third control, we determined that the ratio of oriC to relE in chromosomal DNA prepared from stationary phase cells was 1.3 (data not shown), consistent with other observations that most cells at this stage of growth contain a single genome. Finally, we showed that PCR amplification generated DNA fragments of the expected sizes (Figure 2). The inclusion of [a32P]-dATP in the PCR reaction mixture resulted in a single radiolabelled DNA for each respective PCR reaction (oriC, or relE) as visualized by autoradiography after agarose gel electrophoresis. The amount of radioactivity measured by liquid scintillation spectrometry in the solubilized gel slice containing the DNA fragment (oriC or relE) was essentially identical to the amount of acid-insoluble radioactivity in an equivalent portion of the reaction after PCR amplification (data not shown). These results indicate that the PCR primers specifically amplify either oriC or relE, with no false priming. The results also show that the respective PCR products can be quantitated directly by the amount of acid-insoluble radioactivity. 103 Figure 2. PCR amplification of ariC and relE show that amplification is proportional to the amount of bacterial chromosomal DNA used as the template and is specific for these loci. Increasing amounts in duplicate of E. coli chromosomal DNA were PCR-amplified with primers specific for oriC (Panel A), or relE (Panel B). The levels of amplification that indicate a linear response over at least a 6-fold range were measured by liquid scintillation spectrometry of acid-insoluble radioactively-labeled DNA (see “Experimental Procedures”). Repeated experiments yielded similar results. In Panel B, the data at 10 ng was omitted in order for the curve derived by linear regression to intersect at zero. In Panels C and D, PCR amplification of oriC and relE, a locus near terC, was performed as described in “Experimental Procedures.” Panel C is an ethidium bromide-stained agarose gel of electrophoretically separated DNAs amplified by PCR. Lane 1 contains DNA molecular weight markers with the sizes of relevant DNAs indicated in kb. Lanes 2 and 3 contain duplicate samples in which oriC was PCR- amplified, yielding the expected 1.3 kb fragment. Lanes 4 and 5 are duplicate samples of the PCR-amplified relE locus, resulting in the expected fragment of 1 kb. Panel D is the autoradiogram of the gel shown in Panel C. The only radioactive DNAs detected correspond to the PCR products visualized by ethidium bromide staining in Panel C. 104 in is he templatenl mosomal DNA anel B). llre ld range were vely-labeled ,imilar results 11 linear regress?! elE. alocusrw C is an ethidium lified by PCR mt DNAs was PCR- )licate sampksl‘i 11. Panel D isle 5 elected in Panel C cat/11110-3 160 -J;- 8120 1 4m oriC relE * 1204 100T 100J 90' 80- 50_ :‘0 - 40— 4o— 20- 20‘ Q r r I n 0 2 4 6 0 1b 12 "1) 2 i e 0 1b 12 105 Genomlc DNA (119) D <— oriC 5_ c oriC 4- relE " d- relE Having validated the conditions for the quantitative PCR assay, cultures of the respective plasmid-bearing strains were grown to early log phase, and expression was induced for the times indicated in order to assess hyperactive initiation. A portion of each culture was analyzed by immunoblotting to confirm that expression of mutant or wild type DnaA protein was dependent on induction, and that the expression levels were roughly comparable (data not shown; see Table 5 legend). Bacterial DNA for quantitative PCR analysis was isolated from cultures grown at 30 °C or 42 °C (Table 5). Compared to the oriC-to-relE ratio after induced dnaA+ expression, the Y2 71H allele was comparable to dnaA" in stimulating initiation at either temperature. The ratio of oriC to relE with dnaAcas was significantly higher by comparison to correlate its dominant-negative phenotype to hyperactive initiation. Results with A184 V + Y2 71H were similar to those ratios for dnaAcas at both temperatures. The A184V mutation is well documented to result in elevated initiation at 30 °C (Nyborg, Atlung et al. 2000). Quantitative PCR analysis showed that this mutation was as active as dnaAcas in stimulating initiation at 30°C, but not at 42°C. Because the Q156L + A184 V allele displayed a dominant-negative phenotype at 30°C but not at 42°C (Table 3), this allele was also examined. Compared to dnaA‘, this pair of mutations resulted in an elevated oriC-to-relE ratio, but this pattern was not observed at 42°C. These observations correlate the dominant-negative phenotype of respective dnaA alleles (Table 3) to elevated initiation activity (Table 5). The most striking result from this set of experiments is that the A184V and Y2 71H mutations when present together appear to be responsible for the hyperactive initiation activity of dnaAcas. In vivo initiation activity as measured by oriC plasmid maintenance-The individual mutations of dnaAcas as well as the different combinations were examined to determine their in viva function in initiation. Because the mutations are plasmid-bome, a 106 Table 5. Marker frequency analysis of oriC and relE by quantitative PCR. Plasmid allele oriC/relE oriC/relE 30°C 42°C pDSS96 dnaA+ 37:01 92:25 pLSl 18 Y27IH 33:06 85:17 pLSl20 dnaAcas 10.1:tl.l 14.0123 pLSl l9 A184V+Y271H 8.7104 15613.6 pLSlZl A184V 11421:] .2 63:02 pLS 122 IQ] 56L+Al 84 V 60:06 4.43:1.1 The indicated plasmids were used to transform MC1061, and grown at 30°C or 42°C to an optical density at 595nm of 015-02. Arabinose was then added (1% final concentration) to induce expression, and incubation was continued for a period equivalent to three generations (240-300 min 30°C and 90 min at 42°C). Bacterial chromosomal DNA was isolated as described in “Experimental Procedures” for use as a template in quantitative PCR reactions. Statistical analysis to calculate the range was in duplicate for the samples from the 30°C cultures, and for 5-7 identically prepared samples from the 42°C cultures with the exception of A184V and Q156L + A184V. For the latter samples, statistical analysis at both temperatures was on duplicate samples. As a control, E. coli MC1061 harboring the respective plasmid (pDSS96, DnaA‘; pLS118, Y271H; pLS121, DnaAcos; and pLS119, A184V + Y271H) was grown in LB media supplemented with 100 pg/ml ampicillin to an optical density of 0.3 at 30 °C, induced by the addition of arabinose (1% w/v final concentration), then incubated further for 30 min. Whole cell lysates prepared from 0.4 optical density units (595 nm) of culture were analyzed by immunoblotting (see “Experimental Procedures”) to demonstrate comparable levels of expression. Purified DnaA protein was analyzed in parallel as a control. As another control, LS 1062 (relevant genotype: AariszpKN1562, dnaA850::Tn10) harboring pDS592 (duo/1*) was grown essentially as described above and whole cell lysates from uninduced and induced cultures were analyzed by immunoblotting. DnaA protein was detected only in the sample from the induced culture, confirming that the host strain does not encode DnaA protein. 107 Table 6. Function of dnaAcas mutations in oriC plasmid maintenance. Plasmid dnaA Allele CFUs‘ 42°C/30°C pRB 100 dnaA“ 1 .0 pBR322 None NDb pLSdnaA cos dnaA cos 1 .0 pLS 103 H252Y 0.94 pLS 104 Q156L 1 .0 pLS105 A184V 4.2x10’2 pLSlO6 Y271H 1.1 pLS 107 Q156L+A 184V 0.92 pLS108 Q156L+H252Y 0.95 pLS 109 Q156L+Y2 71H 1.0 pLSl 10 AI84V+H252Y 5.2x10’3 pLSl 11 A184V+Y271H 0.97 pLSl 12 H252 Y+Y2 711-] 1.0 pLSl 13 QI56L+AI84V+H252Y 1.5x10‘3 pLSl 14 A184V+H252 Y+Y2 71H 0.92 pLSl 15 Q156L+A184V+Y271H 1.1 pLSl 16 9156L+H252 Y+Y271H 1.0 ‘The host strain LSlO62 (AoriszpKN1562, dnaA850::Tn10) carrying the respective plasmids encoding the indicated dnaA alleles or pBR322 as a negative control was transformed with an oriC-containing plasmid (pCM959-Cm‘; 100 ng per transformation reaction). The ratio of colony forming units (CFUs) obtained after transformation and subsequent incubation at 42°C or 30°C is shown at the right. The transformation frequencies at 30°C were about 105 per pg of pCM959-Cm' DNA. Transformants were selected on LB media supplemented with 100 pg/ml ampicillin and 25 mg/ml chloramphenicol. " ND indicates that colonies were not observed when LS 1062 harboring pBR322 was transformed with the oriC-plasmid. This control demonstrates that LS 1062 is dnaA null and unable to maintain pCM959-Cmr at either temperature. 108 host strain (LS 1062, relevant genotype: dnaA810::Tn10, AoriszpKN1562) was used that lacked a functional dnaA gene so that the effect of the mutation could be measured without the complication of host-encoded dnaA. An integrated miniRl plasmid that does not depend on dnaA function for DNA replication serves as an alternate replication origin for the bacterial chromosome (Koppes and Nordstrom 1986). The different plasmid- encoded dnaA mutations were then introduced into this genetic background, and the activity of these alleles in initiation at oriC was assessed by maintenance of an oriC plasmid (Table 6). As controls, the plasmid encoding the wild type dnaA gene maintained the oriC plasmid at both 30 °C and 42 °C whereas pBR322 was inert. The dnaA46(Ts) allele comprises the A184V + H252Y mutations; the temperature sensitive- phenotype conferred by the plasmid encoding this allele is another control for the assay. For the dnaAcas allele and its separated mutations, they were functional at both temperatures with the exception of A184V, which was temperature-sensitive. As described above, the single A184V substitution gives rise to overinitiation activity in viva only at 30°C (Table 5; (Nyborg, Atlung et al. 2000)). Correspondingly, the mutant protein carrying this substitution is active at 30°C but not at 42°C when assayed in a replication system with a crude protein fraction as the source of other essential replication proteins (Carr and Kaguni 1996). Examination of pairwise combinations of mutations revealed that Q156L or Y2 71 H suppressed the temperature-sensitive phenotype of A184V when combined with it (Table 6). As shown in Table 5, Y271H combined with A184V resulted in hyperinitiation activity at 42°C. In contrast, Q156L + A184V was even less active than A184 V in initiation at the higher temperature. The explanation for the anomalous behavior of Q156L + A184V (active in oriC plasmid maintenance at 42°C, but poorly active in initiation at this temperature) may be due to limitations with the transformation assay. 109 Figure 3. Relative stabilities of mutant DnaA proteins. E. coli LSlO62(AoriC::pKN1562, dnaA810::Tn10) was grown at 30 °C in LB media supplemented with 100 pg/ml ampicillin to an optical density (595 nm) of 0.1-0. 15, then a portion of each culture was shifted to 42°C. After continued incubation for 2 hr, 1.5 ml of culture was removed. The resuspended cell pellets corresponding to 2.5 x 107 cells were lysed in Laemmli sample buffer by incubation at 86°C for 2 min, and whole cell lysates were electrophoresed in a 10% SDS-polyacrylamide gel. After transfer, the membrane (Protran, Schleicher & Schuell) was incubated with M43 monoclonal antibody (Marszalek, Zhang et a1. 1996) as the primary antibody, and horseradish peroxidase- conjugated goat anti-mouse secondary antibody (BioRad). Detection with X-ray film (Kodak X-Omat) was by chemiluminescence (SuperSignal, Pierce). Quantitation was with a Kodak EDAS 120 software, and normalized to the level of DnaA+ protein at 30°C and 42°C. 110 amaumqficqo A09 DnaA+ Vector Control DnaAcos H252Y _|IlJ_|.._IIJ_.|J 53:33.2 slut... P ‘1': l:!i::‘: ill-I... ..T.. l,...... we a» «9 an we an we h» D . .. a » gov—n ,o . . . 083320 9% .flg 1.. a... 1. A184V+Y271 H Y271 H A184V A184V+H252Y (DnaA46) Q1 56L+A184V l l l l l on an an an an an an an «6 an . s.» . . . v . . 1!... .. . .4 . 1.... . .3 .. ,7 X, . _ , ..... . > . .. . . . . . . .. a I .q . , t i 3131... xi! .in .1. j . ,V t V.,—o ream 2:352 A 308.: 3156 manna. 14%.. mac—Mus well-2? 800 Duo 0 “can 300 03> .— b a b Dawn‘s Pu. Ob al.-85060 Oh. Oh )aua¢+aflk<+wflmfl< EMM.‘ 6b .— h acigmv Pu Oh a: I a.» PD Ema—1.15 Gh< ...a 6.3 1 1 1 Plasmids that are propagated by initiation from E. coli oriC are segregationally unstable, and show a propensity to become integrated into the chromosome upon antibiotic selection (unpublished observations; (Lobner-Olesen 1999)). In support, the Q156L mutation does not appear to suppress the temperature sensitivity of A184 V when combined with H252Y. The suppression of the temperature-sensitive phenotype of A184V for oriC plasmid maintenance by Q156L or Y2 71H suggests that either the physical form of A184V or its activity is stabilized at the elevated temperature. To discriminate between these two possibilities, immunoblot analysis was performed on the assorted dnaA alleles as they were expressed from the natural dnaA promoters. This method should be able to demonstrate if the steady-state level of A184V (a combination of its rates of synthesis and degradation) is lower at 42°C, and if Y271H counterbalances this instability when combined with A184V. The indicated plasmid-bearing strains were grown at 30°C to midlog phase, then a portion was shifted to 42°C. After two hours of continued growth at both temperatures, an equal number of cells from each culture was analyzed by Western blotting (Figure 3). The levels of the mutant proteins were quantitated and normalized relative to wild type DnaA. We did not detect DnaA protein with the strain carrying the vector , but lacking a cloned dnaA allele, confirming the null dnaA genotype of the host strain (LS 1062 dnaA850::Tn10; data not shown). Compared to the level of DnaA+ protein at either temperature, none of the mutant proteins was dramatically unstable. The relative ratio of DnaAcos (0.7 and 0.4 relative to DnaA+ at 30°C and 42°C, respectively) confirms that its dominant-negative phenotype and elevated initiation activity (Tables 3, 5) is not due to oversupply but hyperactivity. These results also suggest that the inactivity of A184 V (and dnaA46) in initiation at 42°C (Table 6) is not due to its physical instability because its steady-state level at 42°C is near that of wild type DnaA (0.8 relative to DnaA‘). Instead, its enzymatic activity appears to be labile. The relative physical 112 stability of the mutant protein carrying the A184V and Y271H substitutions compared to A184V alone suggests that the Y271H substitution acts by stabilizing an enzymatic function that leads to hyperactive initiation. Because of the resolution of denaturing polyacrylamide gel electrophoresis, we cannot exclude the remote possibility that a few amino acids have been proteolyzed from either A184V or DnaA46 at the higher temperature to result in inactivity. As a control, the same set of samples was analyzed by SDS-polyacrylamide gel electrophoresis, followed by staining with Coomassie blue. An arbitrary portion of the gel is presented in Figure 3, showing that similar amounts of protein were loaded in each lane. The results from the control experiments indicate that, within experimental error, the differences in steady-state levels of mutant and wild type DnaA protein are as represented in Figure 3. 113 Discussion DnaA is a member of the AAA+ family of nucleotide binding proteins (Koonin 1992; Neuwald, Aravind et al. 1999). Other members of this group include the E. coli 8’ and 7 subunits of the DnaX clamp loader complex, the five subunits of the eukaryotic RFC clamp loader complex, Orcl, 4 and 5, Mcm 2-7 of the eukaryotic origin recognition complex, and Cdc6. Cdc6/Orcl from P. aeraphilum is structurally similar to the portion of A. aeolicus DnaA whose structure was recently solved (Erzberger, Pirruccello et al. 2002). The Sensor I and II motifs of AAA+ proteins are speculated to function in regulating enzyme activity by conferring the ability to distinguish whether ATP or ADP is bound. Because both DnaA and Cdc6/Orc1 bind ATP and also function to recruit the replicative helicase (DnaB and Mcm 2-7, respectively) to the replication origin (Marszalek and Kaguni 1994; Bell, 2002), the study of DnaA function may provide important insight into the general mechanism of initiation and its regulation. Due to the interesting phenotype of dnaAcas and the possibility that its study may be revealing, we sought to determine which mutations were responsible. The dnaAcas allele was isolated as an intragenic suppressor of the dnaA46(Ts) allele (Braun, O'Day et al. 1987). Two mutations (A 184V and H252Y) are encoded by dnaA46 (Johanson, Haynes et al. 1986). The remaining two mutations in dnaAcas are Q156L and Y271H (Braun, O'Day et al. 1987). Each of these mutations was separated, and also combined in the different possible combinations to determine the substitution responsible for hyperactive replication activity. Hyperactive initiation of the A184V + Y271H mutant-The genetic assay to measure hyperactive initiation relied on the dominant-negative effect of dnaAcas when its expression was from the dnaA promoters and in a dnaA+ host strain. The assay demonstrated that A184V combined with Y2 71 H accounts for the dnaAcas phenotype at both 30 °C and 42 °C. In contrast, A184V alone is dominant-negative to chromosomal 114 dnaA" only at 30 °C. Previous studies show that this single mutation results in increased initiations at oriC (Nyborg, Atlung et a1. 2000). We showed that the dominant-negative phenotype of the A184 V + Y2 71H double mutant is dependent on the presence of oriC. Hyperactive initiation determined by a quantitative PCR assay correlates with the lethal effect observed with plating efficiencies. In these experiments, induced expression of either dnaAcas (as a positive control) or the A184V + Y27/H mutation resulted in an increased initiation frequency as indicated by elevated ratios of oriC to relE. The A184V substitution resides next to a P-loop motif (also known as a Walker A box; residues 172-179, GXXGXGKT; (Goud, Zahraoui et al. 1990; Gines-Candelaria, Blinkova et al. 1995)) that is absolutely conserved among nearly 100 dnaA orthologues. This sequence motif is found in proteins that bind nucleotide cofactors. For those proteins that have been studied at the structural level, amino acids of the motif interact with the [3 and y phosphates of the bound nucleotide, an event that is presumably important in both nucleotide binding and hydrolysis (Reinstein, Brune et al. 1988: Reinstein, Schlichting et al. 1990; Rehrauer and Kowalczykowski 1993). Genetic studies show that the mutation encoding the A184V substitution shared by classically obtained dnaA alleles (dnaA5, dnaA46, dnaA601/602, and dnaA604/606) (Hansen, Koefoed et al. 1992) is responsible for the temperature sensitive phenotype ((Carr and Kaguni 1996) Table 6). This behavior correlates with both the therrnolabile replication activity of a mutant form of DnaA carrying the A184V substitution, and a defect in ATP binding activity. As expected, DnaAcos is also defective in ATP binding (Katayama and Crooke 1995). The defect in ATP binding of the protein carrying the A184V substitution correlates with overinitiation at 30 oC (Nyborg, Atlung et al. 2000). The failure to detect overinitiation at the elevated temperature is apparently due to 115 instability of its replication activity; the Y2 71H substitution stabilizes this activity at 42 °C, leading to overinitiation (Table 6). How can a mutant protein defective in ATP binding result in overinitiation when current models invoke a requirement for ATP binding so that initiation can occur? Perhaps the asynchronous initiation phenotype of alleles encoding the A184V substitution provides a clue (Boye, Lobner-Olesen et al. 1988). These alleles give rise to odd numbers of chromosomes indicative of asynchronous initiation. In contrast, wild type strains display a pattern of 2D chromosomes (n _>_ 0 and an integer) under conditions where cell division and new rounds of initiation are inhibited, but ongoing rounds are allowed to proceed to completion. The failure to bind ATP apparently results in unscheduled initiations that occur additionally prior to and/or after the normal time of initiation during the bacterial cell cycle. Under steady state growth, these aberrant events (overinitiation) then lead to an abnormally elevated ratio of oriC to ter and lethality. The major conclusions from our study is that a DnaA mutant (A184V) with impaired ATP binding activity is active in initiation of chromosomal replication but displays aberrant replication control. The Y271H substitution stabilizes this activity at 42°C. Jointly, A184V and Y271H account for the overinitiation phenotype of dnaAcas. 116 References Atlung, T. and F. G. Hansen (1993). "Three distinct chromosome replication states are induced by increasing concentrations of DnaA protein in Escherichia coli." Journal of Bacreriology 175(20): 6537-6545. Atlung, T., A. Lobner Olesen and F. G. Hansen (1987). "Overproduction of DnaA protein stimulates initiation of chromosome and minichromosome replication in Escherichia coli." Molecular and Ceneral Cenetics 206(1): 51-59. Bell, S.P. (2002) The origin recognition complex: from simple origins to complex functions. Genes & Development 16: 659-672. Blinkova, A., E. Gines-Candelaria, J. D. Ross and J. R. Walker (2000). "Suppression of a DnaX temperature-sensitive polymerization defect by mutation in the initiation gene, dnaA, requires functional oriC." Molecular Microbiology 36(4): 913-925. Boye, E., A. Lobner-Olesen and K. Skarstad (1988). "Timing of chromosomal replication in Escherichia coli." Biochimica et Bio ica Acta 951(2-3): 359- 364. Bramhill, D. and A. Kornberg (1988). "Duplex opening by DnaA protein at novel sequences in initiation of replication at the origin of the E. coli chromosome." Cell 52(5): 743-755. Braun, R. E., K. O'Day and A. Wright (1985). "Autoregulation of the DNA replication gene dnaA in E. coli K-12." Cefl 40(1): 159-169. Braun, R. E., K. O'Day and A. Wright (1987). "Cloning and characterization of dnaA(Cs), a mutation which leads to overinitiation of DNA replication in Escherichia coli K-12." Journal of Bacteriology 169(9): 3898-3903. Brendler, T., A. Abeles and S. Austin (1995). "A protein that binds to the P1 origin core and the oriC 13mer region in a methylation-specific fashion is the product of the host squ gene." Euro a Mo] ular Bi 10 Or anizati n ma114(16): 4083-4089. Buhk, H. J. and W. Messer (1983). "The replication origin region of Escherichia coli: nucleotide sequence and functional units." Gene 24(2-3): 265-279. 117 Campbell, J. L. and N. Kleckner (1990). "E. coli oriC and the dnaA gene promoter are sequestered from dam methyltransferase following the passage of the chromosomal replication fork." @l 62(5): 967-979. Carr, K. M. and J. M. Kaguni (1996). "The A184V missense mutation of the dnaA5 and dnaA46 alleles confers a defect in ATP binding and thermolability in initiation of Escherichia coli DNA replication." Molecular Microbiology 20(6): 1307-1318. Castuma, C. E., E. Crooke and A. Kornberg (1993). "Fluid membranes with acidic domains activate DnaA, the initiator protein of replication in Escherichia coli." Journal of Biological Chemistry 268(33): 24665-24668. Crooke, E. (2001). "Escherichia coli DnaA protein--phospholipid interactions: in vitro and in vivo." Bigmhimie 83(1): 19-23. Danachie, W. D. (1965). "Control of enzyme steps during the bacterial cell cycle." Namre 205(976): 1084-1086. Erzberger, J. P., M. M. Pirruccello and J. M. Berger (2002). "The structure of bacterial DnaA: implications for general mechanisms underlying DNA replication initiation." Eurorman Molecular Biology Organization Journal 21(18): 4763-4773. Garner, J. and E. Crooke (1996). "Membrane regulation of the chromosomal replication activity of E. coli DnaA requires a discrete site on the protein." Eurorman Molecular Biology Organization Journal 15(9): 2313-2321. Gines-Candelaria, E., A. Blinkova and J. R. Walker (1995). "Mutations in Escherichia coli dnaA which suppress a dnaX(Ts) polymerization mutation and are dominant when located in the chromosomal allele and recessive on plasmids." Journal of Bacteriology 177(3): 705-715. Gand, B., A. Zahraoui, A. Tavitian and J. Saraste (1990). "Small GTP-binding protein associated with Golgi cistemae." Nature 345(6275): 553-556. Hansen, F. G., B. B. Christensen and T. Atlung (1991). "The initiator titration model: computer simulation of chromosome and minichromosome control." Res Mierobiol 142(2-3): 161-167. 118 Hansen, F. G., S. Koefoed and T. Atlung (1992). "Cloning and nucleotide sequence determination of twelve mutant dnaA genes of Escherichia coli." Mom General Genetics 234(1): 14-21. Herrick, J., M. Kohiyama, T. Atlung and F. G. Hansen (1996). "The initiation mess?" Molecular Microbiology 19(4): 659-666. Jakimowicz, D., Majka, J ., Messer, W., Speck, C., Fernandez, M., Martin, M.C., Sanchez, J ., Schanwecker, F., Keller, U., Schrempf, H., and J. Zakrzewska- Czerwinska, (1998). Structural elements of the Streptomyces oriC region and their interactions with DnaA protein. Microbiology 144:1281-1290. Katayama, T. and E. Crooke (1995). "DnaA protein is sensitive to a soluble factor and is specifically inactivated for initiation of in vitro replication of the Escherichia coli minichromosome." Journal of Biological Chemistry 270(16): 9265-9271. Katayama, T., K. Fujimitsu and T. Ogawa (2001). "Multiple pathways regulating DnaA function in Escherichia coli: distinct roles for DnaA titration by the datA locus and the regulatory inactivation of DnaA." Bimhimie 83(1): 13-7. Katayama, T. and A. Kornberg (1994). "Hyperactive initiation of chromosomal replication in vivo and in vitro by a mutant initiator protein, DnaAcos, of Escherichia coli." Journal of Biological Chemistry 269(17): 12698-12703. Kata, J. and T. Katayama (2001). "Hda, a novel DnaA-related protein, regulates the replication cycle in Escherichia coli." Embo J 20(15): 4253-62. Kitagawa, R., Ozaki, T., Mariya, S., and T. Ogawa. (1998) Negative control of replication initiation by a novel chromosomal locus exhibiting exceptional affinity for Escherichia coli DnaA protein. Genes a Development 12: 3032-3043. Koppes, L. and K. Nordstrom (1986). "Insertion of an R1 plasmid into the origin of replication of the E. coli chromosome: random timing of replication of the hybrid chromosome." C_ell 44(1): 1 17-124. Koonin, E.V. (1992) DnaC protein contains a modified ATP-binding motif and belongs to a novel family of ATPases including also DnaA. Nucleic Acids Res 20: 1997 Lobner-Olesen, A. (1999). "Distribution of minichromosomes in individual Escherichia coli cells: implications for replication control." European Molecular Biology Organization Journal 18(6): 1712-1721. 119 Lobner-Olesen, A., K. Skarstad, F. G. Hansen, K. van Meyenburg and E. Boye (1989). "The DnaA protein determines the initiation mass of Escherichia coli K- 12." Ceb 57(5): 881-889. Ln, M., J. L. Campbell, E. Boye and N. Kleckner (1994). "squ: a negative modulator of replication initiation in E. coli." Cell 77 (3): 413-426. Mann, D. L., M. Carrington, M. O'Donnell, T. Miller and J. Gaedert (1992). "HLA phenotype is a factor in determining rate of disease progression and outcome in HIV-l-infected individuals." Aids Res Hum Retroviruses 8(8): 1345-1346. Marszalek, J ., and J. M. Kaguni (1994) DnaA protein directs the binding of DnaB protein in initiation of DNA replication in Escherichia coli. Journal of Biological Chemistry 269: 4883-4890. Marszalek, J ., W. Zhang, T. R. Hupp, C. Margulies, K. M. Carr, S. Cherry and J. M. Kaguni (1996). "Domains of DnaA protein involved in interaction with DnaB protein, and in unwinding the Escherichia coli chromosomal origin." Journal of Biological Chemistm 271(31): 18535-18542. Neuwald, A. F., L. Aravind, J. L. Sponge and E. V. Koonin (1999). "AAA+: A class of chaperone-like ATPases associated with the assembly, operation, and disassembly of protein complexes." Cenome Researeb 9(1): 27-43. Nyborg, M., T. Atlung, O. Skovgaard and F. G. Hansen (2000). "Two types of cold sensitivity associated with the A184-->V change in the DnaA protein." Molecular Mierobiology 35(5): 1202- 1210. Ogawa, T., Yamada, T., Kuroda, T., Kishi, T., and S., Mariya. (2002) The datA locus predominantly contributes to the initiator titration mechanism in the control of replication initiation in Escherichia coli. Molecular Microbiology 44: 1367-1375. Rehrauer, W. M. and S. C. Kowalczykowski (1993). "Alteration of the nucleoside triphosphate (NTP) catalytic domain within Escherichia coli RecA protein attenuates NTP hydrolysis but not joint molecule formation." Journal of Biological Chemislry 268(2): 1292-1297. Reinstein, J., M. Brune and A. Wittinghofer (1988). "Mutations in the nucleotide binding loop of adenylate kinase of Escherichia coli." Biochemistry 27(13): 4712- 4720. 120 Reinstein, J ., I. Schlichting and A. Wittinghofer (1990). "Structurally and catalytically important residues in the phosphate binding loop of adenylate kinase of Escherichia coli." Biochemistry 29(32): 7451-7459. Roth, A. and W. Messer (1995). "The DNA binding domain of the initiator protein DnaA." Edroman Molecular Biology Organization Journal 14(9): 2106-211 1. Seitz, H., C. Weigel and W. Messer (2000). "The interaction domains of the DnaA and DnaB replication proteins of Escherichia coli." Molecular Microbiology 37(5): 1270-1279. Sekimizu, K., D. Bramhill and A. Kornberg (1987). "ATP activates DnaA protein in initiating replication of plasmids bearing the origin of the E. coli chromosome." Cell 50(2): 259-265. Slater, S., S. Wold, M. Ln, E. Boye, K. Skarstad and N. Kleckner (1995). "E. coli Squ protein binds oriC in two different methyl-modulated reactions appropriate to its roles in DNA replication initiation and origin sequestration." CeB 82(6): 927-936. Sutton, M. D., K. M. Carr, M. Vicente and J. M. Kaguni (1998). "Escherichia coli DnaA protein. The N -terminal domain and loading of DnaB helicase at the E. coli chromosomal origin." Journal of Biologieal Chemistm 273(51): 34255-34262. Sutton, M. D. and J. M. Kaguni (1997). "Threonine 435 of Escherichia coli DnaA protein confers sequence- specific DNA binding activity." Journal of Biological Chemistry 272(37): 23017-23024. 121 Chapter IV N ovel dnaA alleles are defective for regulation of initiation from the E. coli replication origin oriC 122 Abstract Chromosomal replication in all organisms is a process that is carefully regulated to ensure that the genome is duplicated once and only once during the cell cycle. In E. coli, this process is regulated by DnaA protein. Many dnaA alleles have been examined that fail to initiate replication. The study of these alleles has allowed for the precise determination of the biochemical activities required for DnaA protein function. There are only two well characterized dnaA alleles that result in an increased frequency of origin firing. The first, DnaAcos, is unable to limit the frequency of initiation because of an ATP binding defect. The second, DnaA R334A, is defective in ATPase activity. Under normal growth conditions the intrinsic ATPase activity of DnaA protein negatively regulates initiation by ATP hydrolysis and formation of the inactive DnaA-ADP complex. The R334A mutant protein is inert to this type of regulation. To probe mechanisms of hyperactive initiation we sought to isolate novel dnaA alleles that are unresponsive to regulation. Many of the regulatory signals that trigger initiation have yet to be identified. In this study, we describe the isolation of dnaA alleles that have a similar phenotype to that of dnaAcas. We also determined that the lethality caused by these alleles appears to be due to increased initiation in viva. The conservation of initiation protein function between DnaA protein (E. coli) and Cdc6 protein (S. cerevisiae) are discussed. 123 Introduction DnaA protein is a sequence specific DNA-binding protein that occupies the origin of replication, oriC (Messer, Egan et al. 1991). There are several reported observations that confirm DnaA protein is the trigger that results in the initiation of DNA replication. When DnaA protein is overexpressed from a lacU V5 promoter the frequency of initiation increases (Atlung, Lobner Olesen et al. 1987; Xu and Bremer 1988). A recently identified locus (datA) serves as a high affinity binding site that is occupied by a few hundred DnaA monomers (Kitagawa, Ozaki et al. 1998; Ogawa, Yamada et al. 2002). This site titrates the initiator protein lowering the pool of cellular DnaA protein available for origin binding. Further analysis demonstrated that deletion of the datA locus results in hyperactive initiation. Another important mechanism that regulates the frequency of initiation is the sequestration of the hemimethylated origin DNA to the cell membrane by Squ protein (Slater, Wold et al. 1995). The action of Squ protein prevents reinitiation for 8-10 minutes just after the onset of the first initiation event. If DnaA is primarily responsible for modulating the frequency of initiation of DNA replication, then two expectations arise. First, mutant forms of DnaA protein should be isolated that result in a failure to begin the process of replication (Sutton and Kaguni 1995; Sutton and Kaguni 1997; Sutton and Kaguni 1997; Sutton and Kaguni 1997). Second, mutant proteins that exhibit an increase in the frequency of initiation should also be isolated (Kellenberger-Gujer, Podhajska et al. 1978). Indeed, both of these expectations are fulfilled. Extensive mutagenic analysis of the dnaA gene has revealed several biochemical functions that are required for replication from oriC (Sutton and Kaguni 1997; Sutton and Kaguni 1997). If any of these functions are crippled, the cell fails to initiate chromosomal replication. Likewise, a mutant form of DnaA protein ODnaAcos) has been isolated, and an increase in the frequency of initiation observed as measured by the abundance of oriC DNA compared to a control locus near the site of 124 replication termination (Kellenberger-Gujer, Podhajska et al. 1978; Katayama, Akimitsu et al. 1997). The increase in replication activity caused by this mutant protein is lethal to the host strain (Simmons and Kaguni 2003). The amino acid substitutions encoded by the dnaAcas allele has been carefully analyzed and it was determined that two amino acid substitutions (A184V-1-Y271H) are required for the increased activity of this mutant protein (Simmons and Kaguni 2003). Biochemical analysis of the A184V bearing- protein has revealed a defect in ATP binding, and it is this defect that results in a failure of DnaAcos to limit the frequency of initiation (Katayama, Crooke et al. 1995; Carr and Kaguni 1996). After a round of replication has been initiated DnaA protein hydrolyzes ATP, and generates the inactive ADP bound form (Sekimizu, Bramhill et al. 1987). It has been shown that the release of bound ADP is very slow in vitro, suggesting that formation of the ATP-bound DnaA protein requires its synthesis. The time required for transcription and translation of DnaA protein might serve as one of the mechanisms that limits initiation frequency. A mutant form of DnaA protein has been studied that does not readily convert to the ADP bound (inactive) form (Nishida, Fujimitsu et al. 2002). DnaA protein-bearing an R334A amino acid change has a defect in ATPase activity. The activity of this mutant protein, like DnaAcos, also results in an increased frequency of replication initiation (Nishida, Fujimitsu et al. 2002). DnaAcos and R334A are examples of mutant DnaA proteins bearing amino acid substitutions that result in a failure of the protein to respond to regulatory signals. In these examples, either a defect in ATP binding or its hydrolysis results in aberrant replication control. These results underscore the importance of ATP binding and hydrolysis on the regulation of replication fork assembly. In our study, we sought to identify novel alleles of the dnaA gene that increase the frequency of initiation in attempt to characterize other regulatory mechanisms that modulate DnaA protein function. The dnaAcas allele was used as a control for the assay and we were able to identify seven 125 Lu" novel alleles that appear to be defective in regulating the frequency of initiation. These alleles are also dominant negative to chromosomally encoded dnaA". The dominant negative phenotype has been shown to result from hyperactive initiation as determined by marker frequency experiments (Simmons and Kaguni 2003). We also measured the replication activity of the corresponding mutant proteins in vitro. Based on a homology model of E. coli DnaA protein derived from the crystal structure of the A. aeolicus protein, the mutations are located in Domains 1, HIa, IIIb and IV. This information suggests we have obtained hyperactive alleles that are defective in ATP binding, ATP hydrolysis, DNA binding and possibly altered interactions with the replicative helicase DnaB. We also discuss hyperactive initiation of S. cerevisiae Cdc6 protein and the implications of our findings and how they relate to the regulation of eukaryotic initiation. 126 Table 1. List of plasmids Plasmid Characteristics Source pDSS96 paraBAD, Ap', dnaA” Hwang and Kaguni, 1988 pDS319 paraBAD, Ap‘, dnaA204(ts) Hwang and Kaguni, 1988 pLS 120 paraBAD, Ap', dnaAcas Simmons and Kaguni, 2003 pL8125 paraBAD, Ap', G79D This work pLSl26 paraBAD, Ap’, S146Y This work pLS127 paraBAD, Ap’, H202Y This work pLS128 paraBAD, Ap’, E244K This work pLS129 paraBAD, Ap’, V292M This work pLS130 paraBAD, Ap’, V303M This work pLSl31 paraBAD, Ap', E445K This work pRB 100 dnaA promoters, Ap‘, dnaA” Braun et. a1, 1985 pLSdnaAcos dnaA promoters, Ap', dnaAcas Simmons and Kaguni, 2003 pLSl32 dnaA promoters, Ap’, G79D This work pLS133 dnaA promoters, Ap', S I46Y This work pLSl34 dnaA promoters, Ap’, H202Y This work pLS135 dnaA promoters, Ap’, E244K This work pLS136 dnaA promoters, Ap', V292M This work pLS 137 dnaA promoters, Ap', V303M This work pLS 138 dnaA promoters, Ap’, E445K This work pCM959-Cmr oriC, Cmr Simmons and Kaguni, 2003 pKC596 T7 promoter, Ap', dnaA“ Carr and Kaguni, 1996 pLS 139 T7 promoter, Ap‘, G79D This work pLS 140 T7 promoter, Ap', $146 This work pLS 141 T7 promoter, Ap', H202Y This work pLSl42 T7 promoter, Ap', E244K This work pLS 143 T7 promoter, Ap', V292M This work pLS 144 T7 promoter, Ap’, V303M This work pLS 145 T7 promoter, Ap', E445K This work PLS145 T7 promoter, Ap', A184V This work pBR322 ColEl, Ap’, Tcr Lab Stock 127 Figure 1. Genetic method to isolate hyperactive dnaA alleles. LS 1073 (relevant genotype, dnaA46) harboring either the plasmid vector or this DNA with the dnaA" or dnaAcas allele was plated on LB media with or without 0.5% (w/v) arabinose, as indicated. A genetic screen was developed to obtain dnaA alleles that are hyperactive in initiation based on the phenotype of growth interference upon induced expression of dnaAcas. As shown in Panel A, growth interference was not observed when dnaA+ expression was induced. In Panel B, mutagenesis to increase the frequency of obtaining hyperactive mutations was by hydroxylamine treatment of pDSS96 (Sutton and Kaguni, 1995). The XhaI-EcoNI restriction fragment carrying dnaA was gel-purified and ligated to the unmutagenized vector fragment, then used to transform LS 1073 by electroporation. Transformants were selected at 42°C on LB media supplemented with 100 pg/mL ampicillin to obtain plasmid-encoded dnaA alleles that complement the temperature sensitive phenotype of the host-encoded dnaA46 allele. Individual transformants were then plated on LB media supplemented with 100 pg/mL ampicillin and 0.5% (w/v) arabinose to identify those alleles that interfere with viability when expression was induced. Plasmid DNA was purified from these isolates by chromatography (Qiagen), and used to transform LS 1073. The transformants were selected as described above and examined to confirm their phenotype of growth interference when plated on LB media containing 100 pg/mL ampicillin and 0.5% arabinose. DNA sequence analysis confirmed the presence and type of mutation. Approximately 1200 colonies were screened. Of these 37 colonies caused growth interference after re-transformation, and 15 putative mutants were analyzed by sequence analysis. After sequencing 7 were verified as mutant dnaA alleles. 128 this DNA 5 ‘7: {ll/)1 les that are induced bsened 1: frequency 596 (Sutton gel-purified 1073 by :nted “uh :m the 1:11 impicilhr u hen . were 15% ion, with A' LS 1073 Plated at 42°C DnaA+ DnaAcos Vector Control NoAm DnaA+ DnaAcos Vector Control 05%Am 129 Xhol and EcoNl digestion ‘._ ..... .1 (__bia/ pDSSSG" ECON/ dnaA+ ”xi" araBAD Agar 0'59 991 Purification *hydroxylamine mutagenesis l Ligation to untreated vector L51073 )Gtol dnaA46 / . Transformation of LSl 073 with randomly mutagenized dnaA gene followed by selection on LB plates . + 100 ug/mL ampicillin at 42°C ”3 -o ,. // 10 25 03. Individual colonies were picked and streaked on plates {7. 6.5 .‘4 containing 0.5% arabinose and 100 ug/mL \8. o 02/ ampicillin at 42°C. 11 10019 ./ All dnaA alleles that are defective as a result of I the hydroxylamine treatment do not surviv e on this plate meUT-bWNd Plasmid isolation and retransfonnation to verify the phenotype. The dnaA gene was sequenced if the plasmid was linked to growth interference. 130 Experimental Procedures Bacterial strains and plasmids E. coli K-12 strains have been described previously (Simmons and Kaguni 2003). E. coli LS 1066 was constructed by P1 transduction of the dnaA850zzTn10 allele from LS 1061 into M81062 (relevant genotype MC1061 but ziazszNSOO). Plasmid DNAs were prepared by either equilibrium centrifugation or by column chromatography (Qiagen, Midi Kit) and are listed in Table 1. All novel alleles (except S 146Y originally isolated in pRB 100) were isolated in vector pDSS96. The novel alleles were cloned into pRB 100 by digestion of both pDSS96 and pRB 100 with restriction endonucleases EcoNI and XhoI. This digestion generates two fragments. The 1544 base pair DNA fragment contains the entire dnaA gene coding region. The larger DNA fragment from pRB 100 (approximately 4000 base pairs) is the empty vector. These DNA fragments were ligated to link the novel alleles down stream of the native dnaA promoters. The DNA sequence of the dnaA coding region was determined for each recombinant plasmid to verify the presence of the respective mutations before use in the genetic assays. Quantitative PCR The method used for marker frequency analysis has already been described (Simmons and Kaguni 2003). Briefly, the indicated cultures were grown to early log phase (0.1-0.15, O.D.595nm) and 0.5% arabinose (w/v) was added to the cultures to induce dnaA expression. The cultures were then incubated for 3 hrs and the O.D.595nm was determined to verify that the cultures were still in log phase growth. Cells were concentrated by centrifugation and frozen in liquid nitrogen. The number of cells processed for column chromatography (Qiagen, DNeasy) was in accordance with the manufactures recommendations (3 2X 109 cells). The concentration of chromosomal 131 DNA was determined by absorbance at A260 and purity (Azgo/A280)was measured between 1.7-1.9 to verify that the concentration determination was accurate. Quantitative PCR analysis was with 1 ng of chromosomal DNA in duplicate as described (Simmons and Kaguni, 2003). The amount of amplified DNA was quantitated by scintillation spectrometry after acid precipitation onto glass fiber filters (Whatman, GF/C). Mutant protein overexpression in RTS Each of the novel alleles was cloned downstream of an T7 RNA polymerase promoter for overexpression in the rapid transcription translation assay (RTS). The alleles were cloned into pKC596 (Carr and Kaguni 1996) by partial digestion with EcaRl and complete digestion with Rsr II. Fragments were gel purified and ligated to construct the recombinant plasmid linking the novel alleles to a T7 promoter. DNA sequence analysis was performed to verify the presence of the respective mutations. RTS reactions were reconstituted according to the manufacturers recommendations (Roche). Reaction volumes were either 50 pl or 25 pl and the amount of plasmid DNA for wild type and mutant dnaA was 500 ng or 250 ng respectively. Samples were centrifuged for 5 minutes at 5000 rpm and the amount of protein that remained soluble was quantitated by western blot against known amounts of purified DnaA protein In vitro replication assays Complementation assays with a crude fraction (Fr 11) prepared from WM433 (relevant genotype: dnaA204) were performed essentially as described (Fuller, Kaguni et al. 1981). Except for DnaA the crude fraction contains all of the proteins required to sustain in vitro replication of an oriC plasmid. The amounts of wild type or mutant DnaA protein added to the reaction are indicated. 132 .1. 7.33.39 «29" ‘ Immunoblot analysis Immunoblot analysis was performed essentially as described (Simmons and Kaguni 2003). Briefly, cells were grown to mid-log phase and expression of the various dnaA alleles was induced by addition of 0.5% arabinose (w/v) followed by incubation for 15 minutes. The optical density of the cultures was then determined. Cells were collected by centrifugation and frozen in liquid nitrogen. An equivalent number of cells were lysed and samples were electrophoresed on 10% SDS-polyacrylamide gels. Proteins were then transferred to nitrocellulose (Protran, Schleicher & Schuell) for immunoblot analysis as described (Simmons and Kaguni 2003). Homology modeling Was accomplished using the SWISS-MODEL web site (W). The DnaA primary structure was submitted along with the coordinates for the crystal structure of A. aeolicus DnaA protein (J .M. Berger personal communication). The model was viewed and manipulated using the Insight II graphics program. Images in this dissertation are presented in color. 133 Table 2. Plating efficiencies of plasmid-bome dnaA alleles in LS 1073‘(dnaA46) Plasmid allele Plating Efficiency Plating Efficiency 30°C 42°C pDSS96 dnaA+ 1 .0 0.9 pLS 120 dnaAcas 2.2 X 10'3 4.2 X 10'3 pDS3l9 dnaA204(ts) 1.0 1.3 X 106 pBR322 None 1 .0 ND ‘ E. coli LS 1073 (relevant genotypezdnaA46) was transformed with pBR322 or the respective plasmids in which dnaA gene expression is under araBAD promoter control. Various dilutions were then plated on LB media supplemented with 100 pg/mL ampicillin, and with or without 0.5% (w/v) L-arabinose to measure colony formation at the indicated temperatures. The frequency of transformation in the absence of arabinose was 106-107 per pg of plasmid DNA. Plating efficiency was calculated as the ratio of colony forming units (CFUs) in the presence of arabinose divided by the CFUs observed in the absence of arabinose. 134 Results Isolation of lethal dnaA alleles We developed a genetic assay based on the following observations using a strain that carries the dnaA46(Ts) allele. We noticed that upon introduction of a plasmid in which the dnaA” or dnaAcas alleles is downstream of the inducible araBAD promoter complementation of the dnaA46 defect at 42°C was observed in the absence of induced expression. When these strains were plated in the presence of arabinose the plasmid- bome dnaA+ allele supported growth of the dnaA46 host at the nonperrnissive temperature. In contrast, the elevated dnaAcas expression interfered with viability (Figure 1A, and Table 2 lines 1 and 2). These observations provided a genetic assay to screen for alleles with properties similar to dnaAcas. We controlled for the assay by transforming the dnaA46 host strain with plasmids-bearing the dnaA” and dnaAcas alleles. We show at 30°C and 42°C overexpression of DnaA+ protein does not interfere with growth. In contrast, the overexpression of DnaAcos protein interferes with viability at 30°C and 42°C. We included a vector control in this assay to demonstrate the temperature sensitivity of the dnaA46 host strain, and a dnaA allele encodes a mutant protein defective in DNA binding (dnaA204). We were concerned that defective dnaA alleles containing missense mutations affecting biochemical functions other than ATP binding could form active replication complexes with DnaA46 protein. The dnaA204 allele provides a control for this possibility because the mutant protein is defective in DNA binding. This allele allows for normal growth at 30°C, but at 42°C it is unable to complement the dnaA46 defect at low levels of expression. Interestingly, it does complement when expression is elevated suggesting oversupply of this mutant protein allows for formation of an active replication complex. This allele serves as an additional control for our assay, and it displays the opposite behavior compared with the alleles isolated in our screen (Table 2). 135 We treated a plasmid bearing the dnaA+ gene with hydroxylamine and subcloned the DNA fragment encoding the mutagenized dnaA gene into the untreated plasmid vector. The ligation mixture was then used to transform the dnaA46 host followed by incubation at 42°C. The transformants that arose represent dnaA alleles that are active in complementation of the temperature sensitive growth of the dnaA46 allele. Individual colonies were then grown on LB media containing 0.5 % arabinose (w/v). Transformants able to grow in the presence of arabinose are similar in behavior to dnaA+ and were not studied further. Those that did not grow under elevated expression were characterized in more detail (Figure 1B). Plasmid DNA was isolated and used to retransforrn the dnaA46 host to verify the phenotype described above. Plasmids that clearly conferred the lethal phenotype were analyzed by DNA sequence analysis. We screened nearly 1200 transformants that apparently complemented the dnaA46 defect in the absence of arabinose. From this collection (Table 3) we obtained seven novel dnaA alleles that each containing a single missense mutation. 136 Table 3. Nucleotide and amino acid substitutions in novel dnaA alleles Nucleotide Nucleotide Amino Acid Frequency of Position“ Change Substitution Conservationb 236 G-A 0790 5/30 604 C-T H202Y 7/30 730 G-A E244K 19/30 874 G-A V292M 1 1/30 907 G-A V303M 8/30 1333 G-A E445K 12/30 437 C-A S 146Y 14/30 1067 C—T Silent ‘ The plasmid pDSS96 which carries the dnaA+ gene was mutagenized by hydroxylamine treatment essentially as described (Sutton and Kaguni, 1997). The 1.5 kbp restriction fragment carrying the dnaA coding region was then isolated and used to replace this same restriction fragment in pDSS96 that was nonmutagenized. The ligation mixture was then used to transform LS 1073 (relevant genotype: dnaA46(Ts). Transformants were selected at 42°C to obtain those dnaA alleles that are active in DNA replication and able to complement the nonfunctional dnaA46 gene product. The transformants were then screened for their ability to grow on LB medium supplemented with 0.5% arabinose. The dnaA+ allele in pDSS96 does not interfere with viability when its expression is induced. We discarded those that remained viable under this condition to eliminate alleles that retained dnaA+ function. In comparison, induced expression of dnaAcas interferes with viability. Hence, those transformants that failed to grow in the presence of arabinose were chosen for further characterization. The map positions of the missense mutations are relative to the first nucleotide of the dnaA coding sequence (Hansen et. al, 1982). b The frequency of conservation was determined by comparison of 30 DnaA orthologs. ‘ The S 146Y mutation was isolated from a screen (Simmons and Kaguni, unpublished data) intended to isolate dnaA alleles that failed to replicate the bacterial chromosome, but were active for replication of an oriC plasmid. In this assay the 137 1H Table 3. Legend continued plasmid-bome allele under native dnaA promoter control failed to maintain a dnaA46 bearing strain (LS 1073) at 42°C, but was able to maintain an oriC plasmid in a AoriC host at the same temperature (LS 1062). Further analysis of this allele revealed that growth interference was observed in the dnaA46 host at 30°C. This result suggested that S 146Y was hyperactive and not defective for replication from the chromosomal origin of replication, oriC. Mutagenesis of the dnaA gene to obtain this mutation was by error- prone PCR (Ludlam et. al, 2001). 138 Table 4. Novel dnaA alleles are lethal when overexpressed in an oriC’ strain oriC" AoriC Plasmid Allele CFUs“ 42°C CFUs 42°C Ara/No Ara/No Ara Ara pDSS96 dnaA+ 1.0 1 .0 pLS 120 dnaAcas 5.0X10'3 1 .0 pLS125 G79D 7.1X10’4 1.0 pLS126 SI46Y 6.8X10“ 0.9 pL8127 H202 Y 4.8X10" 0.9 pLS128 E244K 5.3x 10‘3 0.9 pLS 129 V292M 1.3X10” 1.0 pLS 130 V303M 7.1X10“1 1.0 pLS 131 E445K 2.4X104 1 .0 “Plating efficiency is expressed as the number of CFUs observed in the presence of 0.5% arabinose, divided by the CFUs observed in the absence of arabinose. Transformation efficiencies of the indicated plasmids was between 105-107 per pg of DNA. All colonies on 0.5% arabinose were smaller than those growing in the absence of arabinose but were similar in size to the wild type control. The AoriC strain contains a mini R1 integrated into oriC eliminating oriC dependent replication. The oriC strain is LS 1073 and is isogenic to the AariC strain LS 1062 used to obtain the data presented in the rightmost column. L81073 was the host used to isolate the novel alleles described in this work. 139 Steady-state levels of protein expression Previous studies showed that elevated DnaA” protein expression resulted in more frequent initiations, but this level did not interfere with viability of a wild type strain. Because of the possibility that even higher expression levels may interfere with viability, and that the phenotypes of the mutations we obtained is due to this reason, immunoblot analysis was performed (Figure 2A). To eliminate the contribution by the chromosomally encoded dnaA gene, a dnaA null strain was transformed with the respective plasmids, and expression of the mutant proteins at midlog phase growth was induced for 15 min. The analysis is of an equivalent number of cells revealing a comparable level of expression for most alleles. Whereas DnaAcos and S 146Y proteins were observed at lower levels, the phenotype of these alleles is not due to an increase in protein abundance. Rather, it is the specific amino acid substitution that results in growth interference. Lethality requires the presence of functional oriC Experiments were performed to quantitate the level of growth interference when these plasmid-bome alleles were introduced into a dnaA“ host strain (Table 4. column 3). All of the alleles showed a reduction in viability compared to dnaA+ furthermore, the level of growth interference was more than that observed with dnaAcas except for E244K that was almost identical. This experiment reveals the potent effect of these novel alleles on cell viability. As a control for this assay, we transformed an isogenic strain that is null for both oriC and the dnaA gene. Indeed, the data (Table 4 column 4) indicates that the lethal phenotype is dependent on the presence of oriC. We were also able to suppress growth interference by the presence of a multi-copy plasmid-bearing 140 Figure 2. Relative expression levels of dnaA alleles In Panel A, E.cali LS 1062 (relevant genotype: dnaA 850::Tn10, AoriszpKN1562) was transformed with the respective plasmids, and cultures were grown in LB media containing 100 pg/mL ampicillin to an OD. 595m of 0.4-0.6. Arabinose was added to a final concentration of 0.5% (w/v) followed by incubation for 15 min to induce expression. Whole cell lysates were then prepared and analyzed by immunoblotting as described (Simmons and Kaguni, 2003). The relative abundance of the mutant proteins was normalized to DnaA“. In Panels B and C, LS 1066 (relevant genotype, dnaA850z: Tn10, oriC', zia::pKN500) earring the respective dnaA plasmids listed in Panel A were grown in LB media supplemented with 100 pg/mL ampicillin to early log phase (O.D.595nm ~0.15, 0.2). Then arabinose was added to 0.5% (w/v), followed by incubation for the times indicated in Panel B, or for 3 hrs in Panel C. Genomic DNA was purified (DNeasy, Qiagen) and quantitative PCR analysis was done to determine the relative abundance of oriC to relE as described (Simmons and Kaguni, 2003), with the abundance of relE normalized to 1. The range from duplicate samples is indicated by the error bars. For Panel B and C, we noticed two colony morphologies when LS 1066 was transformed by pLSl30 (V303M). Cells that gave rise to small colonies may lack the plasmid or the V303M mutation because the oriC to relE ratio was comparable to that of plasmid-free LS 1066. Quantitative PCR analysis of DNA from normal size colonies of LS 1066 carrying pLS130 yielded the data in Panel B and C. 141 L81062 Dna A protein Fidafi 119 Protein Tran sfamant Level pD$95 Dna A‘ 1.0 pLS120 DnaAcos [1.4 pLS125 G79D 0.9 pLS126 S146Y 0.6 pLSl 2? H212)! 1.1 pLS128 E244K 1.1 pLS129 vegan/1 [1.9 pLS130 VBEEM 1.0 pLSl 31 E445K 1.0 142 xmwvw 2802 28m.) xnwmw canal >925 Time (min) 0.th mood mco +420 ©0051. ‘4-d—uuun—1uqq—qlnu- 5 U 5 D 5 D 2 2 1 1 .Emk \Uto 05050505050 5443322141 m\ wk >926 B. Q 143 Table 5. Dominant-negative effect of dnaA alleles when expression is elevated in a dnaA“ host strain Plating Efficiency Plasmid Allele 30°C 42°C pDSS96 dnaA+ 1.0 1.0 pBR322 None 1. l 1 .0 pLSlZO dnaAcas 3.0 X 10'3 1.5 X 10'2 pL8125 0790 1.5 X 10’3 1.4 X 10'3 pL8126 SI46Y 9.2 X 10“ 7.0 X 10" pLSl27 H202Y 2.7 X 10'3 2.3 X 10'3 pL8128 E244K 0.3b 3.6 X 10'4 pLSl29 V292M 2.8 X 10“ 2.4 X 10“ pLS130 V303M 0.1b 4.9 X 10'2“ pLS131 E445K 0.2b 6.6 X 10“4 “ The respective plasmids were used to transform MC1061 (dnaA‘) and plating efficiency on media with and without arabinose at 30°C and 42°C was determined as in Table 5. b Compared to the colony size of MC 1061 carrying the dnaA“ plasmid pDSS96, colonies of this plasmid-bearing strain were pinpoint in size. 144 Table 6. Novel dnaA alleles are active for oriC plasmid replication LS 1062 Allele Plating Harboring efficiency 42°C/30°C“ Plasmid pRB100 dnaA” 1.0 pBR322 None ND pLSdnaAcos dnaAcas l .0 pLSl32 G79D 1.4 pLS133 SI46Y 2.4 pLS 134 H202 Y 0.7 pLS135 E244K l .0 pLS136 V292M 1.0 pLSl37 V303M 8.7 X 10“” pLSl38 E445K 2.1 X 10'3 “ E. coli LS 1062 (relevant genotype: AariszpKN1562, dnaA850::Tn10) was transformed with the indicated plasmids followed by selection for growth on LB media supplemented with 100 pg/mL ampicillin. Plasmid DNA was then purified from randomly chosen transformants to confirm the presence of the properly sized plasmid. Electrocompetent cells of each plasmid-bearing strain are then prepared and transformed with the oriC plasmid pCM959-Cm“, followed by selection on LB media containing 100 pg/mL ampicillin and 25 pg/mL chloramphenicol. Plating efficiency was calculated as the ratio of pCM959-Cmr transformants that arose at 42°C divided by the number of transfomants at 30°C. b The colonies observed at 30°C with the V303M allele were very small as compared to dnaA+ and other dnaA alleles used in this assay. 145 oriC (data not shown), indicating that DnaAcos protein can be titrated with multicopies of oriC. These results suggest that growth interference is due to initiation from oriC. Novel alleles are dominant negative to dnaA” We tested the alleles to determine if a dominant negative phenotype was observed at both 30°C and 42°C (Table 5). In this experiment we used an empty vector and a plasmid-bearing the dnaA+ allele as negative controls. As expected, neither the dnaA+ allele or the plasmid vector control (Table 5 lines 1 and 2) resulted in growth interference. The dnaAcas allele displayed growth interference at both 30°C and 42°C, which is consistent with previous observations. All of the novel alleles were dominant negative at both 30°C and 42°C. Three of the alleles (encoding E244K, V303M, E445K) were partially suppressed by the presence of the dnaA+ gene at 30°C; however, a 10-fold reduction in colony formation is still observed suggesting hyperactive initiation. As an additional control we cloned all the dnaA alleles downstream of the native dnaA promoters plasmid-bome (see Table l for characteristics). In this experiment the alleles are expressed from the native dnaA promoters, so elevated dnaA expression results largely from a plasmid copy number effect. We demonstrated that all novel alleles were dominant to dnaA46 (data not shown) and several (S 146Y, H202Y, V292M) were dominant to dnaA*. This experiment demonstrates that the inviability caused by these alleles does not require high overexpression, and can be achieved with a modest increase in protein abundance (data not shown). dnaA alleles are active for oriC plasmid replication We sought to determine the replication activity of these alleles when they are the sole source of DnaA protein in the cell. To achieve this objective we transformed a strain that is AariC and AdnaA with vectors linking the alleles downstream of the native dnaA 146 promoters. We grew these transformants under selection for the dnaA plasmids and then transformed the resultant strain with an oriC plasmid (pCM959-Cm‘). We plated an equal number of cells from each transformation reaction at 30°C and 42°C to measure the replication activity of these alleles. The dnaA” control maintained the oriC plasmid well at both temperatures (Table 6). The host bearing the plasmid vector did not maintain the oriC plasmid, confirming that the host strain is dnaA null. All other strains were transformed at 42°C by the oriC plasmid with comparable efficiency except when the strains harbored the V303M and E445K alleles (Table 6). These results indicate all the alleles are active for oriC plasmid replication as the sole source of DnaA protein when assayed at 30°C. The observation that V303M and E445K are inactive for oriC plasmid maintenance at 42°C suggests that at 42°C, mixed complexes form with the chromosomally encoded dnaA allele causing the interference phenotype observed in previous experiments. Another explanation is that in a oriC plasmid there are different DnaA protomer assemblies that are required when compared to oriC in the bacterial chromosome. It has already been shown that certain dnaA alleles have different replication phenotypes at these two oriC locations (Simmons et. al, submitted). Experiments have also demonstrated directly that oriC plasmid maintenance has different DnaA box spacing requirements than chromosomal oriC. These results indicate that the replication requirements of an oriC plasmid are slightly different than that of the bacterial chromosome (Bates, Asai et al. 1995). Marker frequency analysis of hyperactive initiation in vivo It has been shown previously that an elevated frequency of initiation leads to stalled replication forks affecting sequences located near oriC (Kellenberger-Gujer, Podhajska et al. 1978; Nyborg, Atlung et al. 2000). We used a quantitative PCR assay (Simmons and Kaguni 2003) to demonstrate hyperactive initiation in viva. We 147 transformed the host strains (LS 1073 and LS1066) with plasmids bearing the novel alleles and induced expression of the mutant DnaA proteins. Induction was allowed to continue for the optimal time of 3 hours (Figure 2B) prior to harvesting chromosomal DNA for analysis. We included both of these genetic backgrounds in the analysis to determine the level of hyperactive initiation caused by the dnaA alleles in the original background (LS 1073) and a background where oriC dependent replication can be driven without the complication of a dnaA gene background (LSlO66, relevant genotypeszN500::zia, dnaA850zzTn10). In these experiments we measured hyperactive initiation by harvesting chromosomal DNA and then PCR amplifying oriC and the relE locus for comparison. The PCR amplification was done in the presence of a 0:32P labeled dATP and the incorporation of the radioactive material was determined after acid precipitation and isolation of the labeled material on glass fiber filters. We then quantified the amount of a32P label by scintillation spectrometry. We observed hyperactive initiation in the dnaA46 background as all novel dnaA alleles resulted in at least a two-fold increase (10-fold for V303M) in the oriC to relE ratio compared to the dnaA“ control and the strain growing without a plasmid (data not shown). We decided to examine the affect at 30°C only as the 42°C data may be compromised in light of the oriC plasmid maintenance results (Table 6, lines 9,10). As a control for the 42°C plating efficiency data we examined dnaAcas and observed a 5-fold increase in oriC to relE ratio at 42°C compared to the dnaA+ gene control, and a 10-fold increase compared to the strain only control (data not shown). After confirming an increase in oriC to relE DNA in the dnaA46 background we decided to assay the alleles for hyperactive initiation from chromosomal oriC when the 148 novel alleles are the sole source of DnaA protein in the cell. We performed a similar experiment but this time in a dnaA null host. This strain contains an R1 replicon integrated outside of oriC in the zia locus. Upon disruption of the dnaA coding region (dnaA850::Tn10) viability was maintained by initiation from the integrated R1 replicon. When DnaA protein is supplied by a plasmid-bome allele, initiation can occur from oriC. We did a time course experiment to identify the proper time point for analysis. In this experiment (Figure 2B) expression of the dnaA and V303M alleles was induced and time points were extracted from 0-3 hrs. This experiment revealed that wild type DnaA protein reaches peak oriC to relE values within the first 60 minutes of induction. In contrast, induction of V303M results in an increase in oriC to relE values during the entire time of induction (Figure 2B). This experiment is interesting because it suggests that DnaA protein responds to some negative regulator (perhaps Hda protein), and V303M is unresponsive to this factor. We examined the dnaAcas allele in the same experiment and obtained similar results (data not shown). After establishing the proper induction time for this experiment (3 hrs) we analyzed all the novel dnaA alleles (Figure 2C). This experiment provided similar results to the previous experiment in the dnaA46 background. All the novel alleles showed an increase in oriC to relE DNA compared to LS 1066 and LS 1066 bearing a dnaA+ plasmid (Figure 2C). These results clearly establish an increase in initiation activity resulting from the activity of these mutant proteins in viva. We obtained higher initiation activity in the dnaA null strain, suggesting dnaA46 interferes to some extent. Mutant DnaA proteins are active for replication from oriC in vitro We cloned all seven novel alleles so that their expression is from a T7 RNA 149 Figure 3. Mutant DnaA proteins are active for ariC plasmid replication in vitro The respective mutant proteins were produced in vitro in a coupled transcription- translation system (RTS 100, Roche), and their concentration were determined by quantitative immunoblot analysis relative to proportional amounts of purified DnaA protein. Most mutant proteins were produced at a concentrations a of about 140 ng/pl. Replication activity was measured as described in “Experimental Procedures” by incubation for 20 min at the indicated temperatures. In the upper left panel, the activity of purified DnaA protein is represented by open symbols. The closed circle(.) corresponds to the indicated RTS samples at 42°C. The closed squares(l) correspond to RTS samples incubation at 30°C. 150 9m 9m BED DnaA BED G790 70] 7m 6m ,. : 8CD 5113 ' 3 5m 4m " am A am am lg 2111 o ,. _ 2m . . ' E 1:11 1’" _- «dihfl 1m 1011111171170 v ‘ 'ription- é IJI'I1II "1' II 2‘I 3th 0| "I I1I1'I2II2'I II :I 4m ined by 5 BED led DnaA g em HZOZY it 140 [lg/til ‘2 e5"by g l. the amt“? 3(0) :orrCSPOW' DnaA Protein (ng) 151 DNA Synthesis (pmols) QED 8CD 7CD BED 5CD 4m 3CD 2m 1 111 am BED 7CD BED 5m 4m 3m 2m 1 ED 5m EZ44K I 1II II {in 4II 5II V303M 0 50 1111 15.1 2111 253 313 9C!) BC!) 7CD 6CD 5CD 4m 3m 2m 1 CD V292M 1‘I II 2‘I 311] BED 531" DnaA Protein (ng) 152 polymerase promoter in order to express the corresponding mutant proteins in an in vitro coupled transcription-translation system (Roche RTS). The concentration of the soluble protein was determined by western blot analysis compared to a known amounts of the wild type protein. The replication activity of purified DnaA+ protein was compared to DnaA+ obtained from the RTS system. As a negative control, we assayed a portion of the RTS material that was mock treated with a vector control plasmid for replication activity. This sample was devoid of replication activity, indicating that the stimulation of oriC plasmid replication is due to the DnaA protein in the RTS fraction (data not shown). By comparison, mutant proteins displayed varying levels of replication activity (Figure 3). The G79D mutant protein was very active whereas H202Y protein was about half as active as DnaA“. We did not visualize a robust increase in replication activity in this assay, so we examined the mutant proteins in a time course experiment. Again, in this assay most of the mutant proteins did not show a robust increase in replication activity compared to the wild type control, although, most of the mutant proteins were as active or slightly more active than DnaA+ protein (data not shown). We did however, observe a significant increase in replication activity when comparing DnaA+ and E244K at 42°C and S 146Y at 30°C (Figure 4). In both of these experiments DnaA protein reaches a peak of replication activity. In contrast, E244K and S 146Y show an increase in replication activity in nearly a linear pattern. The data obtained in this experiment is similar to our observations of oriC to relE ratios from the in viva time course experiment. DnaA“ protein activity appears to be regulated negatively while the mutant proteins are unresponsive to this soluble factor. It has been observed that DnaA+ activity is negatively regulated in vitro when in the presence of Hda protein or a crude fraction referred to as IdaB. All of the mutant proteins showed an increase in replication activity at both 153 Figure 4. Continued DNA replication activity with hyperactive mutant proteins DnaA+ or the respective mutant proteins were produced in vitro as described in Figure 3, and DNA replication activity was measured at 30°C or 42°C for the times indicated. The graphs corresponding to the mutant protein or the DnaA” control are also indicated. The DnaA+ protein used in this experiment was generated in the RTS system. For each reaction the amount of DnaA protein was constant at 50 mg. 154 DNA Synthesis (pmols) U1 8 , 30°C 4:11.} 3m; 2m: 1m: 0 I I I I l I I I I l I I I I r I I I I 0 20 40 60 so 1410. ‘ 42°C E244K Time (min) 155 temperatures until at least the 40 min time point, but most increased linearly through the 60 min time point (data not shown). Location of missense mutations The position of the missense mutations relative to the functional domains of DnaA protein suggest that the amino acid substitutions affect different functional domains. The E. coli primary structure shares almost 35% identity and 65% similarity over the aligned region [data not shown,(Erzberger, Pirruccello et al. 2002)]. Recently, the crystal structure of DnaA protein from A. aeolicus (Erzberger, Pirruccello et al. 2002) was reported. The structure was determined for Domain IIIa, which is a abbreviated RecA- type fold, Domain IIIb, which is comprised of an antiparallel three— helix bundle, and Domain IV, which is the DNA-binding domain and contains a helix- tum-helix (HTH) motif. We constructed a homology model of the E. coli protein (Figure 5) based on the structure of A. aeolicus DnaA, and mapped the positions of the amino acid substitutions. The E. coli protein (Figure 5) almost completely overlaps the A. aeolicus structure. The amino acid substitutions S 146Y, H202Y, V292M, and V303M are in close proximity to the ATP binding pocket. Based on the homology model, these mutations may affect ATP binding or hydrolysis. The mutation E445K is located in the HT H motif of the DNA binding domain, suggesting that aberrant complexes are formed at oriC to result in an increased frequency of initiation. The E244K substitution is located in Domain IIIa which contains the Walker A box and Sensor I motif. Because this substitution is located in an outer a-helix and is distant from the site of ATP binding, it seems unlikely that this residue perturbs ATP binding. This residue may be involved in a regulatory mechanism to control initiation. If this residue is solvent exposed in the full length protein, it may 156 Figure 5. Homology model of E. coli DnaA protein (Domains Illa-IV) The predicted structure of E. coli DnaA protein is depicted by the blue ribbon diagram superimposed onto the known structure of Aquifex aeolicus DnaA, shown in red. Amino acid substitutions of the mutant proteins are shown in yellow. Domains IHa, IIIb and IV, the Walker A box and the portion involved in binding to the DnaA box sequence are also shown. The position of the G79D substitution is not pictured because structural information for this portion of A. aeolicus DnaA is not available. 157 ””3 vases “\ I sén46 , , 1” WalkerA \ -Val292 \ i G|u244 158 interact with proteins during assembly of the prepriming complex. We were unable to model the G79D substitution because Domain I and II are not part of the A. aeolicus crystal structure. 159 Discussion DnaA protein has been studied intensely for the last twenty years (Kornberg and Baker 1992). Although many missense mutations have been isolated and characterized, only two appear to bypass a regulatory mechanism that might serve as a checkpoint to limit the frequency of initiation. We sought to isolate other dnaA alleles under the hypothesis that regulatory mechanisms function to limit DnaA protein activity to regulate the frequency of initiation during the cell cycle. Genetic screen and characterization of novel dnaA alleles We devised a genetic method to isolate novel hyperactive dnaA alleles. The method involved two steps. First, we transformed a dnaA46 strain at the nonpermissive temperature with a plasmid carrying a randomly mutagenized dnaA gene, and selected for dnaA alleles that complemented the dnaA46 defect. This step required that complementation was in the absence of induced expression. In the second step, transformants were screened for growth interference when the level of expression was elevated. As the genetic assay was based on the behavior of the dnaAcas allele that is hyperactive in initiation we expected to obtain other dnaA alleles with similar properties. These two genetic experiments have been used as valid assays to discriminate between dnaA alleles that are defective for replication or benign, compared with alleles that are hyperactive (Simmons and Kaguni, 2003). This assay yielded seven alleles that are dominant negative to chromosomally encoded dnaA” and required the presence of oriC for growth interference. As an additional control, we showed that these phenotypes are not due to an increase in the level of protein expression (Figure 2A). The alleles encoded unique single amino acid substitutions that may affect specific biochemical functions of DnaA protein, as indicated by the locations of the amino acid substitutions. 160 Hyperactive initiation in vivo To demonstrate that the alleles we obtained result in an elevated frequency of initiation, we performed quantitative PCR analysis to measure the abundance of oriC and relE, a locus in the replication terminus region. In previous work, we established conditions for quantitative PCR analysis that accurately measured the abundance of specific DNA sequences in the bacterial chromosome (Simmons and Kaguni, 2003). In the cited study, we demonstrated that the ratio of oriC to relE reflects the frequency of initiation. In stationary phase cells, this ratio is near 1, whereas in logarithmically growing wild type cells, this ratio of 4 is consistent with other results that indicate that the majority of cells carry 4 copies of oriC in the replicating chromosome. Using this method, we confirmed that the dnaA alleles, when induced, indeed promote more frequent initiations. Some alleles such as H202Y, E244K and V303M are more active than dnaAcas as reflected by their higher oriC/relE ratios. The S 146Y, V292M and E445K mutations are about half as active as dnaAcas, but more active than dnaA+ in stimulating initiation when expression of these alleles is induced. The activity of the mutant proteins in sustaining in vitro DNA replication of an oriC-containing plasmid was measured, and compared to DnaA+ protein. Assays were performed at 30°C, the standard incubation temperature and also at 42°C. The latter temperature was chosen because the mutant alleles were able to complement the dnaA46(Ts) allele at 42°C in vitro. Except for V303M and E445K the replication activity of the mutant proteins correlates with the ability of the corresponding plasmid-home allele to maintain an oriC plasmid in the dnaA850zzTn10 host stain. The reduced ability of V303M and E445K at 42°C in maintenance of the oriC plasmid contrasts with the ability of these plasmid-bome alleles to complement the host-encoded dnaA46(Ts) allele at 42°C. We do not have an explanation for this apparent discrepancy. 161 .3 Hyperactive initiation in vitro The genetics of these dnaA alleles, and the in vivo analysis indicates hyperactive initiation. We decided to examine the replication activity of the mutant proteins in vitro in a Fraction H complementation assay (Kaguni, Fuller et a1. 1982). Because our mutant proteins behaved so poorly during preparation of the soluble fraction we decided to synthesize the proteins in a rapid transcription-translation coupled assay in vitro. A few of the mutant proteins showed an increase in replication activity, but all were active for oriC plasmid replication in vitro at both 30°C and 42°C confirming they are active for replication from oriC. Even the most well studied hyperactive mutant protein DnaAcos, only shows a modest increase in replication activity (Katayama and Kornberg 1994). Our in vitro replication data and the published data on DnaAcos activity prompted the time course replication experiment (Figure 4). We observed a robust increase in replication with time for the E244K and S 146Y mutant proteins compared with the wild type control. This data is interesting because the activity of the mutant proteins continues linearly with time, whereas DnaA" proteins reaches a peak activity. This data suggests that a soluble factor in the crude WM433 extract is able to negatively regulate DnaA protein but not the mutant proteins. We suggest that this factor could be Hda protein (Kato and Katayama 2001), or possibly the beta subunit of DNA polymerase HI holoenzyme (Katayama, Kubota et a1. 1998). Both of these proteins have already been implicated in negative regulation of DnaA protein. We constructed a homology model of E. coli DnaA protein from the crystal structure of DnaA protein of A. aeolicus (Erzberger, Pirruccello et al. 2002) to determine the relative locations of our amino acid substitutions in the predicted folded structure. To our surprise all of the mutations except for G79D and E445K were located in Domain Illa-IIIb indicating ATP binding or hydrolysis may be affected (Figure 5). We performed the same analysis for DnaAcos and determined that the two missense mutations required 162 for the dnaAcas phenotype (data not shown) are located in Domain IIIa, confirming ATP binding and hydrolysis are important regulatory biochemical mechanisms limiting DnaA protein function. We do believe that G79D, E445K, and probably E244K cause an increase in origin firing by some other biochemical mechanism. This suggests origin activation is not only limited through the availability of DnaA-ATP. Cdc6 protein and the frequency of initiation The recent crystal structure of Domains III and IV of A. aeolicus DnaA protein reveals a remarkable structural similarity to archaeal Cdc6/Ore] from P. aerophilium (Liu, Smith et al. 2000). In yeast, Cdc6 functions in initiation of DNA replication. Two cdc6 alleles (cdc6-2, cdc6-3) have been isolated from S. cerevisiae that result in “persistent initiation” as determined by >2C DNA content, and the constant association of MCM proteins with chromatin (Liang and Stillman 1997). The increased frequency of initiation leads to a failure of these cells to divide, a feature that is also observed under hyperactive initiation in E. coli. We aligned the primary structure of Cdc6 from S. cerevisiae and P. aeraphilum to determine the corresponding amino acids and the structural domains affected by these mutations. The cdc6-2 allele encodes an A3208 amino acid substitution (P. aeraphilum 1223) that affects Domain II and is located in 0t- helix 12. If we use the structure of DnaA protein for comparison, this mutation may affect ATP hydrolysis because it resides in that portion of DnaA protein involved in ATP hydrolysis. The cdc6-3 allele encodes two amino acid substitutions. H144T does not align with a corresponding amino acid of P. aeraphilum Cdc6. L258S is located in B strand 4 and may affect ATP binding, based on the location of this residue in the structure of DnaA protein. From the comparative analysis of DnaA and Cdc6 the location of relative missense mutations to functional domains suggests that both proteins may limit the frequency of initiation through ATP binding. In the case of Cdc6 protein, ATP 163 binding defects appear to result in a constant loading of the MCM helicase complex at the origin of replication. It is possible that hyperactive dnaA alleles are able to continually load DnaB when ATP binding and hydrolysis are reduced. What is the cause of lethality? In earlier studies, we showed that the induced expression of DnaA protein resulted in an elevated frequency of initiation from oriC. Two independent methods of microarray analysis that measured the relative abundance of essentialy every gene sequence of the E. coli chromosome, and quantitative PCR analysis of oriC and relE were used to obtain results supporting this conclusion. Interestingly, an elevated level of DnaA protein does not interfere with viability unless the cell is defective in the repair of double strand breaks. In contrast, dnaAcos expressed at a comparable level interferes with viablility, apparently because of an even higher frequency of initiation. This hyperinitiation results in arrested replication forks, and the accumulation of an unusually high level of double strand breaks. Apparently, the failure to repair these double strand breaks interferes with viability. As shown in Figure 2, some of the dnaA alleles result in even higher frequencies of initiation than dnaAcas. We presume that the growth interference associated with these alleles when expression is induced is due to the accumulation of double strand breaks and their failure to be repaired. Other alleles do not confer as high a frequency of initiation, but the initiation frequency is nevertheless elevated in comparison to the wild type dnaA gene. These alleles interfere with growth presumably for the same reason. One model proposes that DnaAcos causes whole chromosome overreplication, and the cause of lethality is a cell division or chromosome partitioning defect (Katayama, Takata et al. 1997). This observation is supported by results identifying a novel protein CedA, that restores cell morphology (dnaAcas causes filamentation). It is not clear that 164 CedA protein affects cell division, this protein may suppress DnaAcos and restore normal morphology as a result. We propose that dnaAcas and other hyperactive dnaA alleles cause an increase in the frequency of initiation that leads to stalled replication forks, and Holliday junction formation at the site of replication fork collapse. It has been documented that Holliday junctions are cleaved to the point of double strand breaks (DSB) by a fully activated Rqu protein (Michel, Ehrlich et al. 1997). We have shown that dnaAcos, and many of the novel alleles described in this study cause a large increase in linear DNA observed in vivo by PFGE analysis (data not shown). We have also demonstrated that cell viability during DnaA" induced hyperactive initiation requires homologous recombination proteins (Simmons et. al, in preparation). Taken together, when hyperactive initiation is induced many replisomes are assembled which results in a stacking of replication forks near oriC. This causes many of the replisomes to stall, and DSB are generated by Rqu action (Seigneur, Bidnenko et al. 1998). If these breaks are not sufficiently repaired, and the replication forks are not restarted through a primosome dependent pathway, then the host is no longer viable. In the case of DnaA” induced hyperactive initiation the level of linear DNA apparently does not out compete the homologous recombination machinery, but the increase in replication caused by hyperactive dnaA alleles (up to 20 fold) causes more DSB in the cell than can be repaired. 165 References Atlung, T. and F. G. Hansen (1993). "Three distinct chromosome replication states are induced by increasing concentrations of DnaA protein in Escherichia coli." Journal of Bacteriology 175(20): 6537-6545. Atlung, T., A. Lobner Olesen and F. G. Hansen (1987). "Overproduction of DnaA protein stimulates initiation of chromosome and minichromosome replication in Escherichia coli." Molecular and General Genetics 206(1): 51-59. Bates, D. B., T. Asai, Y. Cao, M. W. Chambers, G. W. Cadwell, E. Boye and T. Kogoma (1995). "The DnaA box R4 in the minimal oriC is dispensable for initiation of Escherichia coli chromosome replication." Nucleic Acids Research 23(16): 3119-3125. Carr, K. M. and J. M. Kaguni (1996). "The A184V missense mutation of the dnaA5 and dnaA46 alleles confers a defect in ATP binding and thermolability in initiation of Escherichia coli DNA replication." Molecular Microbiology 20(6): 1307-1318. Erzberger, J. P., M. M. Pirruccello and J. M. Berger (2002). "The structure of bacterial DnaA: implications for general mechanisms underlying DNA replication initiation." Euroman Moleoolar Biology Qrganization Journal 21(18): 4763-4773. Fuller, R. S., J. M. Kaguni and A. Kornberg (1981). "Enzymatic replication of the origin of the Escherichia coli chromosome." Proceedings of the National Academy of Science, USA 78(12): 7370-7374. Guo, L., T. Katayama, Y. Seyama, K. Sekimizu and T. Miki (1999). "Isolation and characterization of novel cold—sensitive dnaA mutants of Escherichia coli." FEMS Microbiology Letters 176(2): 357-66. Hansen, F. G. and T. Atlung (1995). "Initiation of chromosome replication after induction of DnaA protein synthesis in a dnaA(nuII) mh mutant of Escherichia coli." Mol Microbiol 15(1): 149-54. Kaguni, J. M., R. S. Fuller and A. Kornberg (1982). "Enzymatic replication of E. coli chromosomal origin is bidirectional." Naturo 296(5858): 623-627. Katayama, T., N. Akimitsu, T. Mizushima, T. Miki and K. Sekimizu (1997). "Overinitiation of chromosome replication in the Escherichia coli dnaAcas 166 mutant depends on activation of oriC function by the dam gene product." Molecular Microbiology 25(4): 661-70. Katayama, T., E. Crooke and K. Sekimizu (1995). "Characterization of Escherichia coli DnaAcos protein in replication systems reconstituted with highly purified proteins." Molecular Miorobiololgy 18(5): 813-820. Katayama, T. and A. Kornberg (1994). "Hyperactive initiation of chromosomal replication in vivo and in vitro by a mutant initiator protein, DnaAcos, of Escherichia coli." Journal of Biological Qhomistgy 269(17): 12698-12703. Katayama, T., T. Kubota, K. Kurokawa, E. Crooke and K. Sekimizu (1998). "The initiator function of DnaA protein is negatively regulated by the sliding clamp of the E. coli chromosomal replicase." QEH 94(1): 61-71. Katayama, T., M. Takata and K. Sekimizu (1997). "CedA is a novel Escherichia coli protein that activates the cell division inhibited by chromosomal DNA over- replication." Moloculab Miouobiology 26(4): 687-97. Kato, J. and T. Katayama (2001). "Hda, a novel DnaA-related protein, regulates the replication cycle in Escherichia coli." Embo J 20(15): 4253-62. Kellenberger-Gujer, G., A. J. Podhajska and L. Caro (1978). "A cold sensitive dnaA mutant of E. coli which overinitiates chromosome replication at low temperature." Moleoular auu Qenorul @uotios 162(1): 9-16. Kitagawa, R., T. Ozaki, S. Moriya and T. Ogawa (1998). "Negative control of replication initiation by a novel chromosomal locus exhibiting exceptional affinity for Escherichia coli DnaA protein." Genes Q Devevelooment 12(19): 3032-3043. Kornberg, A. and T. A. Baker (1992). W. New York, W. H. Freeman and Company. Liang, C. and B. Stillman (1997). "Persistent initiation of DNA replication and chromatin-bound MCM proteins during the cell cycle in cdc6 mutants." Qenes & Development 11(24): 3375-3386. Liu, J ., C. L. Smith, D. DeRyckere, K. DeAngelis, G. S. Martin and J. M. Berger (2000). "Structure and function of Cdc6/Cdc18: implications for origin recognition and checkpoint control." Molecular Cell 6(3): 637-648. 167 Margulies, C. and J. M. Kaguni (1996). "Ordered and Sequential Binding of DnaA Protein to oriC, the Chromosomal Origin of Escherichia coli." Journal of Biologiofl Qhomisu'y 271(29): 17035-17040. Michel, B., S. D. Ehrlich and M. Uzest (1997). "DNA double-strand breaks caused by replication arrest." European Moleoular Biology Organization Journal 16(2): 430- 438. Nishida, S., K. Fujimitsu, K. Sekimizu, T. Ohmura, T. Ueda and T. Katayama (2002). "A nucleotide switch in the Escherichia coli DnaA protein initiates chromosomal replication: evidnece from a mutant DnaA protein defective in regulatory ATP hydrolysis in vitro and in vivo." J Biol Chom 277(17): 14986-95. Nyborg, M., T. Atlung, O. Skovgaard and F. G. Hansen (2000). "Two types of cold sensitivity associated with the A184-->V change in the DnaA protein." Molecular Microbiology 35(5): 1202-1210. Ogawa, T., Y. Yamada, T. Kuroda, T. Kishi and S. Moriya (2002). "The datA locus predominantly contributes to the initiator titration mechanism in the control of replication initiation in Escherichia coli." Molecular Microbiology 44(5): 1367- 75. Seigneur, M., V. Bidnenko, S. D. Ehrlich and B. Michel (1998). "RuvAB acts at arrested replication forks." §e_11 95(3): 419—430. Seitz, H., C. Weigel and W. Messer (2000). "The interaction domains of the DnaA and DnaB replication proteins of Escherichia coli." Moleoular Microbiology 37(5): 1270-1279. Sekimizu, K., D. Bramhill and A. Kornberg (1987). "ATP activates DnaA protein in initiating replication of plasmids bearing the origin of the E. coli chromosome." Qefl 50(2): 259-265. 168 Simmons, L. A. and J. M. Kaguni (2003). "The dnaAcas allele of Escherichia coli: hyperactive initiation is caused by substitution of A184V and Y271H, resulting in defective ATP binding and aberrant DNA replication control." Molecular Microbiology 47(3): 755-765. Slater, S., S. Wold, M. Lu, E. Boye, K. Skarstad and N. Kleckner (1995). "E. coli Squ protein binds oriC in two different methyl-modulated reactions appropriate to its roles in DNA replication initiation and origin sequestration." Ce_ll 82(6): 927-936. Sutton, M. D. and J. M. Kaguni (1995). "Novel alleles of the Escherichia coli dnaA gene are defective in replication of pSC101 but not of oriC." Journal of Bacteriology 177(22): 6657-6665. Sutton, M. D. and J. M. Kaguni (1997). "The Escherichia coli dnaA gene: four functional domains." Journal of Molecular Biology 274(4): 546-561. Sutton, M. D. and J. M. Kaguni (1997). "Novel alleles of the Escherichia coli dnaA gene." Journal of Molooulgr Biology 271(5): 693-703. Sutton, M. D. and J. M. Kaguni (1997). "Threonine 435 of Escherichia coli DnaA protein confers sequence- specific DNA binding activity." Journal of Biological Chemistry 272(37): 23017-23024. Xu, Y. C. and H. Bremer (1988). "Chromosome replication in Escherichia coli induced by oversupply of DnaA." Molooulur and General Gonetios 211(1): 138-142. 169 Chapter V Hyperinitiation of DNA replication in E. coli leads to replication fork collapse and inviability Lyle A. Simmons“; Adam M. Breierb', Nicholas R. Cozzarelli", and Jon M Kaguni‘ ‘ Department of Biochemistry and Molecular Biology Michigan State University, East Lansing MI, 48824. b Department Molecular and Cell Biology University of Califomia- Berkeley, Berkeley CA, 94720-3204. 170 Abstract Induced dnaA“ expression from an overproducing plasmid stimulates initiation of chromosomal replication from the E. coli replication origin, oriC, but does not interfere with viability. More frequent initiations are also observed with the dnaAcas allele encoded in single copy in the bacterial chromosome, but lethality is observed with dnaAcas despite the lesser level of expression. In this report, we show that hyperinitiation induced by elevated dnaA‘ expression results in replication forks that abort or collapse, the majority of which are within about 15 map units neighboring the replication origin (oriC). As fork collapse occurs at apparently random sites, nucleoprotein complexes such as datA bound by DnaA are not the cause. To maintain viability, the damaged replication forks must be repaired by replication restart proteins including PriA and RecBCD. Viability under dnaA*-dependent hyperinitiation contrasts with the lethality of dnaAcas regardless of its induced expression. When the replication restart pathway is blocked by a recB mutation, the accumulation of collapsed forks, reflected by the level of broken linear chromosomal DNA, is substantially higher under elevated dnaAcas expression compared to dnaA”. These and other results support the conclusion that the lethal phenotype of dnaAcas is due to untimely initiations from ori C and the insufficiency in repairing the elevated level of collapsed replication forks. 171 Introduction DNA replication in E. coli initiates from the chromosomal origin, oriC (reviewed in Kornberg and Baker, 1992; Messer, 2002). Replication forks then progress around the chromosome, terminating in a region opposite the chromosomal origin. As the time required to duplicate the chromosome is relatively constant, one view is that replication forks progress at constant speed. Other recent work suggests that replication forks occasionally arrest or collapse, requiring reassembly of the replication fork machinery in order to complete duplication of the chromosome to maintain viability (reviewed in Cox, Goodman et al., 2000; Jones and Nakai, 2000). A key protein involved in restarting collapsed replication forks is PriA, first discovered in the study of bacteriophage ¢X174 DNA replication (reviewed in Sandler and Marians, 2000). Although priA is not essential for viability, mutants grow poorly, suggesting that replication forks become arrested frequently and must be restarted to reach the terminus region. Replication forks may arrest by several mechanisms. When replication forks fail to advance due to an upshift in temperature of an E. coli dnaB(Ts) or holD(Ts) mutant to the nonpermissive temperature, fork regression and collapse occur (Flores, Bieme et al., 2001; Michel, Ehrlich et al., 1997; Saveson and Lovett, 1997). As another general mechanism, DNA lesions in the template strand of the parental duplex may stall the replicative polymerase, leading to regression of replication forks, their collapse and potentially lethal double strand breaks. A third mechanism involves a protein-DNA barrier. Paused replication forks have been documented when the replication fork machinery collides with transcribing RNA polymerase or with proteins firmly bound to DNA (French, 1992; Liu and Alberts, 1995; Vilette,Uzest etal., 1992) . Indeed, because it takes about twice as long for replication forks to be completed in rep mutants compared to wild type strains, Rep helicase has been proposed to act at the fork to remove proteins bound to DNA (Lane and Denhardt, 1975). In eukaryotic systems, the replication fork 172 barrier may be programmed as in the case with the rDNA gene cluster in S. cerevisiae. FOBl participates in the arrest of replication forks at the 3’-end of the rDNA transcription unit so that DNA replication does not interfere with rRNA synthesis (Kobayashi and Horiuchi, 1996; Kobayashi, Heck, et al., 1998) . DnaA protein regulates the process of chromosomal replication by modulating the frequency of initiation at the chromosomal origin. Indeed, IPTG-dependent dnaA“ expression under lacUV5 promoter control stimulates initiation from oriC, but fully replicated chromosomes are not produced because the newly initiated replication forks appear to be arrested (Atlung and Hansen, 1993; Skarstad, Lobner-Olesen et al., 1989). Elevated dnaA” expression under this condition, or under expression from the araBAD promoter (shown below) did not interfere with viability. In contrast, the dnaAcas allele, originally isolated as an intragenic suppressor of the dnaA46(Ts) allele present in single copy in the genome, interfered with viability at 30°C that was attributed to an increased frequency of initiation (Kellenberger-Gujer, Podhajska et al., 1978). In either circumstance, (elevated dnaA+ expression by lacU V5 promoter induction, or chromosomal dnaAcas expression from the natural dnaA promoters), marker frequency analysis of various loci distributed around the chromosome confirmed elevated frequencies of initiation, based on the abundance of oriC and markers nearby compared to a site near the replication terminus. However, these results did not explain the apparent contradiction that the lethal effect of dnaAcas compared to the nonlethal effect of elevated dnaA+ expression is due to the level of initiation. In the study described herein, we addressed whether the lethality caused by hyperactive initiation by dnaAcos (or dnaA+ under specific conditions described below) is due to the level of hyperinitiation. The results obtained by genomic microarray analysis, quantitative PCR, and pulse field gel electrophoresis support a model in which elevated initiation activity 173 results in replication fork collapse. Lethality results from the failure to repair the collapsed forks. Under conditions of elevated initiation, it was also unclear whether the replication forks were stalled at specific or random sites. We considered the possibility that specific protein-DNA complexes may impede replication fork movement. As just one example, six sites in the bacterial chromosome have been described that are bound by DnaA with high affinity (Roth and Messer, 1998). The datA locus is one of these sites; it is estimated that several hundred DnaA monomers bind to the 1 kb region surrounding datA (Kitagawa, Mitsuki et al., 1996; Kitagawa, Ozaki et al., 1998; Ogawa, Yamada et al., 2002). As DnaA bound to DnaA box sequences in the coding regions of genes has been described to impede transcription by RNA polymerase (Messer and Weigel, 1997). DnaA bound to specific sites in the bacterial genome may likewise induce replication forks to pause or permanently arrest. In the work described here, we also tested the hypothesis that specific sites in the bacterial genome cause replication forks to stop under conditions log-phase growth or hyperinitiation. The results indicate that replication forks do not encounter specific barriers except for the Tus-ter complex, in the terminus region, where cessation of fork advancement is expected. 174 Table 1. Strains and plasmids Strain Genotype Source MC 1061 araDI39 Aara-leu)7696 A(lac)X74 galU Lab stock galK hst2(rk', mk“) strA mcrA mchl This work LS 1064 priA2::kan araDl39 Aara-leu)7696 A(lac)X74 galU galK hst2(rk', mk“) strA mcrA mchl JJC315 leu6 his4 argE3 lacYI galKZ ara14 xy15 mtlI tsx33 rpsL31 supE44 hst recBZ68::Tn10 KHGIOO7 F X IN (rmD-rrnE)l rph-Ihis tus::Tn3 lacZ::terL (Km’) (nonblocking orientation) KHGIOO5 F X IN (rmD-rmE)l rph-I his tus::Tn3 lacZ::terL (Km') (blocking orientation) B. Michel (Michel et al., 1997) Y. Fujimura (Horiuchi and Fujimura, 1995) Y. Fujimura (Horiuchi and Fujimura, 1995) Plasmids Characteristics Source pBR322 ampr tetr Lab stock pDSS96 dnaA+ paraBAD Lab stock (Hwang and Kaguni, 1988) pLSlZO dnaAcas paraBAD Lab stock (Simmons and Kaguni, 2003) pLST435M T435M paraBAD This work pLSG426D G426D paraBAD This work pTus Tus paraBAD (Peter et al., 1998) 175 Experimental Procedures Strains and plasmids. The bacterial strains and plasmids used in this work are listed in Table 1. LS 1064 was constructed by P1 transduction with JC19008 (priA2::kan; (Sandler, Samra et al., 1996) as the donor strain and MC1061 as the recipient. Transductants were selected at 37°C on LB media supplemented with 25 ug/ml kanamycin. Plasmids pLS T435M and pLSG426D were constructed by partial RerI and EcoNl digestion of pDSS96 (in which the dnaA* coding region is controlled by the araBAD promoter) to gel-purify the vector DNA fragment away from dnaA coding sequences. This DNA was ligated to 1.5 kb DNA fragments carrying the T435M and 64260 mutations of dnaA, obtained by RerI and EcoNl digestion of pACHA-49 and pACHA-79. The resulting plasmids are identical in structure to pDSS96, but wild type dnaA is replaced by these specific alleles. Preparation of ectopic ter-blocked chromosomal DNA. The plasmid pTus, encoding the tus gene placed downstream from the araBAD promoter, was used to transform strain KHG1005 which carries a ter site inserted into lacZ in an orientation that blocks replication forks originating from oriC when the site is bound by Tus protein (Figure 1). The resulting plasmid-bearing strain was grown at 37°C in a rotary shaker to early log phase in LB media supplemented with 0.2% glucose. Cells were collected by centrifugation and resuspended in LB media containing 0.2% arabinose to induce Tus expression, followed by incubation with shaking for one hr at 37°C . As a negative control, KHG1007, which is isogenic to KHG1005 but carries the ectopic ter site inserted at lacZ in the nonblocking orientation, was grown to stationary phase, and genomic DNA was prepared as above. Sodium azide was added to the culture to a final concentration of 0.2% (w/v), and the culture was frozen and stored in liquid nitrogen. After concentrating the cells by centrifugation, genomic DNA was prepared using a DNeasy Tissue kit (Qiagen) following the manufacturer’s instructions. The yield and purity of the genomic 176 DNA was determined spectrophotometrically, then digested with AluI endonuclease (Promega). A portion of the sample was electrophoresed on a 1% agarose gel to confirm cleavage, as indicated by the appearance of a range of DNA fragments from about 0.5 to 3 kb in size. The remainder of the cleaved DNA was purified with a PCR Purification kit (Qiagen), and the amount of DNA recovered was determined spectrophotometrically. The DNA was then lyophilized and stored frozen until use. Preparation of chromosomal DNA from cultures induced to express DnaA+ protein. E. coli MC 106] bearing pDSS96 that encodes dnaA+ under araBAD promoter control was grown in LB media at 37°C to 0.2 CD. at 595 run then the culture was divided. Where indicated, L-arabinose was added to one portion to a final concentration of 1% (w/v) to induce dnaA expression. The second portion was transferred to a separate flask, and growth was continued at 37°C without arabinose addition. Both the induced and untreated cultures were harvested by centrifugation after 45 min of continued incubation with shaking, and chromosomal DNA was prepared as described above. To serve as a negative control in the experiment of Table 2 and Figure 1, one portion of the culture was removed just prior to the time of arabinose addition, and genomic DNA was isolated from this sample as described above. Analysis of this sample is shown in Figure 1A. Quantitative PCR analysis. The samples of purified bacterial chromosomal DNA were then analyzed by microarray (see below), or by quantitative PCR analysis as described (Simmons and Kaguni, 2003). In the latter method, 1 ng of DNA was used as the template in PCR reactions with 40 pmol of each pair of respective primers to amplify oriC or relE. PCR reactions were assembled according to the manufacturer’s recommendations (PCR Master mix, Promega), but contained [a32P]-dATP (1.25 uCi per reaction, 3000 Ci/mmol, Perkin-Elmer), and amplification was for 24 cycles with Taq DNA polymerase (1.25 units). To measure the amount of PCR-amplified DNA, the 177 radiolabelled product was acid-precipitated with 10% (w/v) trichloroacetic acid onto glass fiber filters (GF/C, Whatman), and radioactivity was quantitated by liquid scintillation spectrometry (Simmons and Kaguni, 2003). Primers to amplify oriC are ACAGCGTACAATACGCCACT and ACGGTGAGCACGACGGCTTT. To amplify the relE region, the primers are G'I'I‘GAAGTACTI‘GAGTCACC and CATTCAGACTTGAATGCGTG. These primers are identical to those used by Nyborg (Nyborg, et al., 2000) and amplify relE, relF, and rem. Microarray procedures and data analysis. Microarrays carrying nearly every annotated coding region in the E. coli genome were prepared as described (http://www.microarrays.org/orotocols.htrnl). Of the 4403 coding regions encoded by the E. coli genome, 4184 were successfully amplified for printing. Labeling of genomic DNA and hybridization to microarrays were performed essentially as described (http://www.microarrays.org/orotocolshtml), except that labeling reactions were allowed to proceed overnight to increase the amount of labeled probe synthesized. Fluorescence intensities of microarray spots were measured using a Genepix 4000B scanner and Genepix 3.0 software (Axon Instruments). Spots were filtered to require that at least 40% of pixels in each spot be at least one standard deviation above background in both channels. Smoothed trajectories of gene dosage versus chromosome position were calculated as a moving median with a window of 101 loci. Filtering, smoothing, linear regression, and plotting of microarray data were performed using the R language and environment, version 1.3.0 (http://www.r-projectorg). As some experiments involved induced expression of dnaA or tus from a multicopy plasmid, the signals due to annealing of these sequences to the printed array were not included in the analysis. Pulse field gel electrophoresis. To measure the amount of linear chromosomal DNA that reflects the steady-state amount of double strand breaks in the indicated strains, pulse field gel electrophoresis (PFGE) was in 1% agarose gels (Chromosomal Grade 178 Agarose, BioRad) in buffer composed of 45 mM Tris-borate (pH 9) and 1 mM EDTA with a CHEF-DRII apparatus (BioRad). Electrophoresis was for 16-20 hrs at 160 or 180 volts with a switching time ramped from 3 to 30 see, as described (Michel, Ehrlich et al., 1997). Samples were prepared by adding 1.5% agarose (low melting point, Gibco) to an equal volume of cells (3 x 108; 0.4 CD. at 595 nm) in 10 mM Tris-HCl pH 8, 0.1 M EDTA, 10 mM EGTA (TEE buffer). If the samples were not analyzed immediately, they were stored at 4°C and used within 7 days. S. cerevisae chromosomal DNA (BioRad) or a mixture of concatamers of 2» DNA (lambda ladder PFG marker; New England Biolabs) were used as molecular weight standards for each PFGE experiment and also to confirm that the conditions of electrophoresis were appropriate. Gels were stained with 0.5 jig/ml ethidium bromide, extensively destained, then photographed with a Kodak DC290 camera. Image analysis was with Kodak EDAS 120 software and the data was replotted with a graphing program (DeltaGraph Pro 3). The data of Figure 3 is a quantitative analysis of fluorescence staining of the DNA with ethidium bromide, and the background is equivalent to a portion of the gel that lacks DNA. 179 Results Do replication forks stall at specific sites? To test the hypothesis that specific sites in the bacterial genome may function to arrest replication fork movement under the conditions of hyperinitiation, microarray analysis was performed. This quantitative method measures the abundance of specific gene sequences, and is typically used to measure RNA expression levels under regulated conditions. In our experiments, microarray analysis was used to measure the relative copy number of essentially every open reading frame of the E. coli chromosome, offering the possibility of demonstrating in a population of cells whether replication forks pause or arrest at specific sites in the genome. Genomic DNA was purified from cultures of E. coli MC1061 (relevant genotype: AaraBAD) harboring a plasmid with the dnaA+ gene under araB promoter control (pDSS96; Table 1). One sample was isolated from the culture at early log phase, just prior to induction (Figure 1A). The culture was then divided, and arabinose was added to one portion to induce expression of dnaA” (Figure 1B). The second was mock- treated to serve as a control (Figure 1C), and both were incubated for an additional 45 min. In other experiments (data not shown), this length of induced dnaA+ expression gave rise to the highest frequency of initiation as reflected by the ratios of oriC to relE (a site near terC), measured by quantitative PCR analysis (see below). The DNAs were then annealed to microarrays of E. coli genes to measure the relative abundance of each open reading frame. The data from the induced and uninduced samples were normalized relative to data from genomic DNA isolated from stationary phase cells by dividing the abundance of each locus by that from the stationary phase sample. The data were further normalized by expressing the number of copies of a specific locus as a ratio to the averaged abundance of loci in the terminus region where relE is located (see Table 2 legend). The averaged value of markers in the terminus region was used to represent the 180 Figure 1. Elevated dnaA expression induces hyperinitiation. Microarray analysis of chromosomal DNA was performed as described in “Experimental Procedures.” The horizontal axis in each panel represents the position on the chromosome in map minutes (one map minute is 1% of the genome, or 46.4 kb). Except where noted, oriC is at zero minutes to facilitate visualization of replication fork density, resulting from bidirectional fork movement from oriC. The terminus region is at both the left and right margins. The vertical axis represents relative gene dosage on a base 2 logarithmic scale, so that a difference of one unit indicates a two-fold change in abundance. Turquoise points are raw values for individual microarray spots, and plots of smoothed values from a moving median are in purple. Dashed orange lines are least-squares best fit lines to the purple traces, drawn to help show that the gene dosage patterns fall along a straight line. E. coli MC1061 carrying the dnaA+ plasmid, pDSS96, was grown at 37°C with aeration in LB media. The microarray used to determine fork movement was prepared from total E. coli DNA, and of the 4403 expressed genes of E. coli, 4184 (95%) were successfully amplified for printing. The abundance of loci from induced and uninduced samples was normalized relative to their abundance in genomic DNA from a stationary phase sample. These values were then normalized by dividing by the average gene dosage of loci in the terminus region within 92.8 kilobase pairs (2 map minutes) flanking the point on the chromosome diametrically opposite oriC. The mean abundance of loci in the terminus region was used to represent the abundance of relE. Because oriC is a cis-acting site, its abundance could not be determined. As an alternative, the gidA and mioC genes that directly flank oriC (see Figure 2) were measured. The ratios of gidA and mioC relative to the average gene dosage of loci in the terminus region were then calculated, with the gidA/relE and mioC/relE ratios for samples from the stationary phase culture normalized to 1.0. Panel A is the sample from the culture removed at early log phase prior to induced dnaA expression. The gene dosage pattern shows straight lines descending from the origin and is indicative of normal, exponential growth. After this sample was removed, the culture was divided and arabinose was added (1% w/v final concentration) to one portion of the culture followed by incubation for 45 min before the cells were harvested and genomic DNA isolated for analysis (Panel B). The second portion was mock-treated (no arabinose addition) and incubated with shaking at 37°C for an additional 45 min (Panel C). In Panel D, E. coli KH1005 (relevant genotype, lacZ::terL) carrying pTus was grown as above to midlog phase, and arabinose (1% w/v final concentration) was added to induce tus expression (see “Experimental Procedures”). After continued growth for 1 hr, cells were pelleted and DNA prepared as described in “Experimental Procedures” for microarray analysis. This strain has a ter site inserted into lacZ ( 25 minutes) oriented to block clockwise-moving replication forks. The gene dosage pattern shows clear deviations at the lacZ locus (red line) and the natural terminus region where the positions of terB and terD are indicated by black vertical lines. The 181 roarray analysgqf ocedures." 11,: me in map PM": ted. OriC 15 12,7: 3 fT0111 bidirtc’m 1 right max 3 £316. so that; quoise point: 111 L165 from a 1333 ms to the p33: [might line. £3.“ :h aeration 313 rd from 10121:: ccessfully uced samples :3 ary phzse a” sage of loci 3i. 8 point on he in the temtfi Cis-acting shit? 'aC genes this: id mjoC relax-:2 116d. vith the Jlture norm $6 pl’lOi 10 dcscendifiéw nip-16 “'35 11 contt‘imf 1e C6115 W615 d portion “9- Figure 1. Legend continued boundaries of this plot were shifted so that half of the blocked region between lacZ and the terminus region is shown on each side of oriC. 182 tnlaclu 1092 Relative abundance 1092 Relative abundance -4o -20 0 2o 40 -éo -40 -50 6 20 40 Distance from origin (map min) Distance from origin (map min) 183 abundance of relE. Because oriC does not encode a gene product, adjacent genes (mioC and gid for oriC) were measured. In wild type E. coli cultures under exponential growth, the relative abundance of loci diminishes exponentially with increasing distance from oriC (Cooper and Helmstetter, 1968). A semi-log plot of the relative abundance of gene dosage versus chromosomal position should reveal a pair of descending straight lines with the apex originating at oriC. The results obtained with the genomic DNA from the culture before induced dnaA“ expression, or mock-treated confirm this expectation, showing that these samples are from log phase cultures, and that the method of analysis is reproducible (Figure 1). Relative to a vertical line drawn through oriC, the angles of the descending lines are similar, indicating that replication forks move bidirectionally at the same average speed. Of the samples in which DnaA+ was overexpressed, we expected the distribution of loci to differ from that of the controls in one of two ways. If forks were arresting at specific sites, a downward break in the distribution at the point of arrest is expected (see Figure 1D). If sites of arrest are distributed across the chromosome, the distribution should show a global deviation from exponentiality. In Figure 1B, the abundance of loci no longer resembles that observed with the exponential phase cultures (Figure 1A, C), and the ratio of markers flanking oriC to relE is more than twice that of the uninduced culture. The results show that, at elevated DnaA, replication forks do not stall at datA or other specific sites, but appear to stop randomly throughout the chromosome. The steeper slope near oriC indicates that the replication forks induced by hyperactive initiation do not progress very far from this site. Indeed, the sequence abundance of markers near terC in this sample was comparable to that of the uninduced control, providing additional support to the view that replication forks fail to progress to the normal terminus region. Also, no 184 Figure 2. Map of the E. coli chromosome. The locations of oriC, relE and the ectopic ter site inserted at lacZ of strain KHG1005 (lacZ::terL tus::Tn3) are shown . This strain is a derivative of W3110, which carries a large chromosomal inversion between 73.8 and 90.7 min, placing oriC at 79.9 min. The distance from oriC to lacZ is 27.9 min. In KHG1007, terL is inserted at lacZ in the nonblocking orientation. 185 terL (Blocked Strain only) ‘UacZF Kmr :19ch \ l - . ,r' {2'16 3&3?!- : In E. coli chromosome (minutes) 11. 186 specific site appears to hinder replication fork movement when dnaA+ expression was not induced. In order to demonstrate that our method is capable of detecting forks stalled at a specific location, microarray analysis was performed with genomic DNA isolated from a strain bearing an ectopic ter site inserted at the chromosomal lacZ locus (Figure 2). Tus protein bound to ter sites in the vicinity of the replication terminus normally functions to block replication forks from escaping the terminus region (see Bussiere and Bastia, 1999 for a recent review). The ectopic Tus-ter complex is oriented to block replication forks at lacZ in a strain lacking the chromosomal tus gene and carrying a plasmid in which tus is regulated by the araBAD promoter. Genomic DNA isolated from the strain (KHG1005) induced for one hr to express Tus protein was annealed to the E. coli microarray, and the results were normalized as above to microarray data obtained from genomic DNA isolated from an isogenic strain (KHG1007), but carrying lacZ::terL in the nonblocking orientation, and grown in the absence of the tus plasmid to stationary phase. The plot of gene dosage versus chromosomal position (Figure 1D) reveals breaks at the ectopic ter site (red line) and the chromosomal terminus region (black lines). The results from these experiments confirm that the microarray method is capable of detecting forks stalled at specific sites on the bacterial chromosome. The microarray results were confirmed by quantitative PCR assays in which we measured the abundance of oriC and relE, a locus near the replication terminus. In a recent report, we established conditions for quantitative PCR assays (Simmons and Kaguni, 2003). We showed that the DNA amplified was specific for oriC or relE with no false priming, that the level of amplification was proportional to the amount of bacterial chromosomal DNA, and that the ratio of oriC to relE on induced dnaA+ expression 187 Table 2. Analysis by microarray and quantitative PCR of oriC or adjacent loci and relE. Sample Microarraya Microarray' Quantitative PCRb gidA/relE mioC/relE oriC/relE stationary phase 1.0 1.0 1.0 early log phase 2.4 2.4 2.321: 0.1 no arabinose no arabinose 2.1 2.1 2.9: 0.3 45 min growth 1% arabinose 5.0 4.9 51:01 45 min induced expression ‘ Please see the legend to Figure 1. bThe oriC/relE values of genomic DNA isolated from MC1061 (pDSS96) were normalized to the ratio of these loci in genomic DNA from this strain grown to stationary phase, normalized to 1.0. Each PCR reaction was performed in duplicate to calculate the standard deviation. The oriC/relE ratio for DNA from the stationary phase culture was 1.3. Cells in early stationary phase generally contain one complete chromosome, with a small fraction of cells carrying one partially replicated chromosome. 188 Table 3. Replication fork restart genes priA and recB are required for viability under conditions of hyperactive initiation Plating Efficiency’ Plasmid Allele MC1061 (recB*, LS 1064 (recB*, JJC315 priA+) priA2::kan) (recB::Tn10, priA’) pDSS96 dnaA+ 1.0 1.0 x 10'3 2.7 x 10‘3 pBR322 None 0.98 0.97 1.2 pLS120 dnaAcas 9.5 x 10'3 5.3 x 104 4.4 x 10'4 pLST435M T435M 1 .0 0.97 1.0 pLSG426D 04260 1 . 1 1.2 1 .0 8’The indicated plasmids were used to transform the respective host strains by electroporation with a BioRad E. coli Gene Pulser following the manufacturer’s instructions. Incubation was at 37°C until colonies were visible on the plates lacking arabinose (18-20 hr). MC1061 in the absence of arabinose was transformed with an efficiency ranging from 106-107 per ug of the various plasmid DNAs. The recB mutant was transformed with an efficiency of 2 x 105 per ug of the respective plasmid DNAs in the absence of arabinose. We confirmed the observations of others that strains lacking priA grow poorly on rich media compared to minimal media (Nurse, Zavitz et al., 1991). We noted that transformation efficiency of the priA2::kan strain by the dnaA-encoding plasmids in the absence of induction on LB media (6 x 10“ transformants per ug of plasmid DNA) was reduced compared to M9 media supplemented with glycerol (0.5% w/v) and casamino acids (0.5% w/v). Plating efficiency is described as a ratio of colony forming units obtained in the presence of arabinose to the number obtained in the absence of arabinose. 189 increased linearly with time up to 45 min then approached a plateau. The ratios of gidA and mioC (genes flanking oriC) to relE obtained by microarray were normalized by their division by the ratio of these loci in stationary phase cells (Table 2). The ratio of oriC to relE obtained by quantitative PCR analysis was similarly normalized. Their comparison reveals that the results are in good agreement to support the conclusions that the methods reflect the relative abundance of genetic loci, and that induced expression of dnaA+ leads to hyperinitiation. Hyperactive initiation interferes with viability in the absence of recombinational repair. The results described above are presented under the model that replication forks induced by hyperinitiation appear to be “arrested” as opposed to being transiently paused. The evidence that these forks do not progress to completion is based on the following observations. In experiments comparable to those of Figure 1 but under prolonged incubation after induced dnaA’ expression (in MC1061 carrying the dnaA+ expression plasmid, pDSS96), we found that the array pattern was very similar to that of Figure 1B (data not shown). The gradient of loci from oriC toward the terminus region was only slightly reduced. In a control culture (MC1061 bearing pDSS96) grown in parallel for the same time period but not induced, the culture had reached stationary phase as reflected by the lack of increase in culture turbidity, and by the microarray pattern. The latter revealed approximately equivalent abundance of all E. coli genes regardless of their chromosomal locations (data not shown). These results suggest that whereas some arrested forks induced by hyperinitiation can be resuscitated, a substantial fraction do not proceed to completion during this time period. Moreover, the absolute abundance of loci near the terminus region was comparable to the stationary phase sample, strongly suggesting that the forks are indeed arrested. The persistence of these arrested forks suggests that the replication restart pathway is inefficient. Inasmuch as viability as 190 Figure 3. Pulse field analysis of chromosomal DNA after induced expression of wild type or mutant DnaA protein. E. coli JJC315 (relevant genotype, recB::Tn10; (Michel, Ehrlich et al., 1997) carrying the respective dnaA plasmids or pBR322 was grown as described in the legend to Table 4. Chromosomal DNA from 3 x 108 cells (0.4 CD. at 595 nm) was analyzed by pulse field gel electrophoresis as described in “Experimental Procedures.” DNA at “A” corresponds to the circular DNA that remains in the agarose plugs at the migration origin (Birren and Lai, 1993). DNA at“ B” corresponds in size to 4-6 megabase pairs that has been attributed to spontaneous chromosome breakage observed in recB mutants . The region between arrows C and D is linear DNA in the size range from 750 to 40 kilobase pairs, respectively. 191 1 expressionol 11' 3ch :Tnlli, #1113. {2 was grown as cells (0.4 0111' in “Experimeha' 11115 in the agar irresponcls in 533‘ me breakage tear DNA in 111553 +> DnaA Fluorescence Intensity (EtBr) u ' ' ‘ '1th' T' '35 "'3th Electrophoretlc Mobility 192 41D Fluorescence Intensity (EtBr) 5+) DnaAcos 4m 4—o 64280 d—o Electrophoretlc Mobility 193 measured by colony formation, is preserved on induced dnaA" expression (Table 3), we infer that the longer time period for colony growth is sufficient to repair these damaged forks. (Please see “The replication restart pathway” in the “Discussion” also.) It is well-established that blocked replication forks can regress to form Holliday junctions by the annealing of nascent DNA strands (reviewed in Michel, Flores et al., 2001). RuvAB bound to the junctions can support further branch migration; Rqu cleavage induces double strand breaks (DSBs). Current models on DSB repair in E. coli invoke that DSBs are lethal because of the inability to duplicate the DNA beyond the break. In support, linear bacterial DNA accumulates in recB mutants due to impaired homologous recombination and recombinational DNA replication (Seigneur, Bidnenko, et al., 1998). Based on the microarray results, it is likely that prematurely initiated replication forks catch up and collide with more slowly moving or paused replication forks in front (replication fork collapse), as well as with replication forks that have regressed and been processed subsequently into DSBs. If this explanation is correct, induced expression of dnaA” should be lethal in strains deficient in recombinational DNA replication. As a test of this model, dnaA+ expression was induced in strains carrying recB and priA mutations and viability was compared to a wild type control (Table 3). To confirm published observations (Atlung and Hansen, 1993; Skarstad, Lobner-Olesen et al., 1989), elevated dnaA+ expression did not interfere with viability of the wild type strain. In contrast, viability was greatly reduced when dnaA” expression was in a recB (implicating RecBCD) or priA mutant. Elevated expression of dnaA alleles (T435M, and G426D;(Sutton and Kaguni, l997a,b) known to be inactive in DNA replication from oriC had a benign effect on viability in these mutant strains, indicating that growth interference is due to wild type dnaA function and elevated initiation. (As will be shown below, G426D also does not elevate the level of DSBs.) By comparison, induced 194 Table 4. Double strand breaks accumulate in recB” strains due to hyperactive initiation. Plasmid-Bome Plasmid Relevant % Linear Allele Genotype DNA' dnaA+ pDSS96 recB” 3 1.51:2. l dnaAcas pLS 120 recB 42.3:20 G426D pLS G426D recB 16.411 .4 None pBR322 recB” 16510.8 None None recB+ 4. 1:2.2 ‘ Cultures of JJC315 (relevant genotype, rechzTnIO) or MC1061 (recB‘) bearing the indicated plasmids were grown in LB media at 37°C with shaking to mid-log phase (0.2 0D. at 595 nm), and arabinose (1% w/v final concentration) was added followed by incubation for 1 hr to induce expression of the corresponding proteins. At the end of this incubation, the culture turbidity among these samples was similar (0.7-0.8 CD. at 595 nm). Chromosomal DNA was analyzed by PFGE analysis (see “Experimental Procedures”). The amount of linear chromosomal DNA (in percent) in the range from 750 to 40 kilobase pairs was quantitated by ethidium bromide staining intensity and is expressed as a percentage relative to the total amount of DNA in each lane. Three independent experiments were performed to calculate the standard deviation. 195 expression of dnaAcas interfered with viability of the wild type host strain, but growth interference was reduced only by about 100-fold. Our experience with a dnaAcas mutant (encoded by the chromosome) of a reduction in viability at 30°C compared to 42°C by 6 x 10" suggests that the growth interference described in Table 3 is not as dramatic because this activity is partially attenuated at 37°C, even despite overproduction of the mutant protein. Nevertheless, this effect was exacerbated in either a priA2::kan or a recB::Tn10 mutant. These results support the interpretation that hyperinitiation leads to replication fork collapse instead of arrest as discussed above . Furthermore, the lethal effect of dnaAcas or dnaA” in either a recB or priA mutant appears to be due to the failure to repair DSBs, prompting the following experiments. RecBCD protein functions in DSB repair by processing duplex DNA ends, directing the binding of RecA to the resultant single-stranded DNA to promote homologous recombination (Anderson and Kowalczykowski, 1997). If the reduced viability due to increased expression of dnaA" or dnaAcas (Table 3) is due to the failure to repair DSBs, the level of linear chromosomal DNA under these conditions should be elevated. To test this prediction, pulse field gel electrophoresis (PFGE) was performed. This method has been used extensively to quantitate the level of DSBs that result from mutations in genes involved in DNA metabolism (reviewed in Michel, Flores et al., 2001). As a control for the method, chromosomal DNA from log phase cultures of a recB+ strain and recB mutant carrying a plasmid lacking dnaA (pBR322) was analyzed (Table 4). We quantitated the amount of linear DNA by measuring fluorescence intensity of ethidium bromide-stained DNA, as opposed to the method of measuring the amount of radioactively labelled DNA in gel slices (Seigneur, Bidnenko et al., 1998) because the equipment for PFGE was borrowed. DNA from 750-40 kilobase pairs was selected for analysis because the fluorescence intensity of S. cerevisiae chromosomal DNA markers 196 within this range was proportional to their sizes. For larger DNAs that were poorly resolved, we were not confident that their fluorescence corresponded to their respective molecular weights. Below 40 kilobase pairs, very little fluorescent material was detected. The amounts of linear DNA in each sample confirm values previously reported for a wild type and recB mutant (4% and 16%, respectively; (Michel, Ehrlich et al., 1997), confirming that the method of analysis is appropriate. In a recB mutant induced for hyperinitiation by elevated levels of DnaA" or DnaAcos, the elevated amount of linear chromosomal DNA compared to the empty vector control (2- or 2.6—fold, respectively; Table 4, Figure 3) is indicative of DSBs that arise from replication fork collapse and possibly also from the cleavage of regressed replication forks by Rqu. Compared to the recB mutant harboring pBR322, induced expression of a mutant protein bearing an aspartate residue at glycine 426 (G426D) did not affect the amount of linear DNA detected (Table 4, Figure 3). Because the mutant protein is inactive in DNA replication due to its inability to bind to the chromosomal origin (Sutton and Kaguni, l997a,b) this result confirms that growth interference due to elevated dnaA’ or dnaAcas expression is due to the failure to repair DSBs. 197 Discussion Replication forks do not appear to stall at specific sites. Because accumulated evidence indicates that replication forks arrest under normal growth conditions, we examined whether specific sites in the bacterial genome functioned to impede replication fork movement. An attractive candidate is datA at which 370 DnaA monomers are estimated to bind (Kitagawa, Mitsuki et al., 1996; Kitagawa, Ozaki et al., 1998; Ogawa, Yamada et al., 2002). This site is proposed to titrate excess DnaA protein to prevent overinitiation. As additional support that datA may hinder fork movement, others have described that DnaA bound to duplex DNA impedes RNA polymerase as it transcribes DNA (Roth and Messer, 1998). Using microarray technology which measures the abundance of essentially every open reading frame of the E. coli genome, our results support the conclusion that replication forks are not blocked at specific sites except at ter bound by Tus at both an ectopic location and at sites in the terminus region. This conclusion is even under an elevated level of DnaA in which the datA locus should be fully occupied by DnaA. Apparently, replication fork movement is insensitive to nucleoprotein complexes other than the Tus-ter complex. Under the stringent response in B. subtilis, replication forks arrest at replication checkpoints located about 200 kilobase pairs on either side of oriC (Autret, Levine et al., 1999; Levine, Vannier et al., 1991). A more recent study of the replication checkpoint terminator to the left of oriC showed that the sequence is bound by the replication terminator protein of B. subtilis, but that the function of this sequence in vivo is independent of stringent control (Gautam and Bastia, 2001). Although we did not investigate the effect of the stringent response on replication fork arrest, our results suggest that the E. coli genome does not contain specific sites that function as replication checkpoints. 198 Hyperinitiation induces replication fork collapse. We showed that elevated levels of DnaA” protein do not interfere with viability in a wild type host strain, and that this condition induces hyperinitiation from oriC (Table 5, Table 2, Figure 1). When analyzed by microarray, replication forks accumulate in the vicinity of oriC (within about 15 map units flanking oriC, Figure 1). The dependence on priA and recB for viability (Table 3) and the increased amount of linear chromosomal DNA in a recB mutant (Table 4) indicate that replication forks have collapsed and also have regressed to be processed by Rqu. Whereas we did not assess the proportion of linear DNA that accumulates due to fork regression upon elevated initiation in the recB mutant, it is attractive to consider that replication forks formed by hyperinitiation collide from behind with stalled replication forks to cause replication fork collapse. If replication forks do not move at constant speed, it is conceivable that replication fork collapse occurs when a speedy fork catches up to a laggard. As another possibility, replication fork collapse may occur if the sealing of Okazaki fragments is slow. In this model, a newly initiated replication fork collapses when it encounters the discontinuity separating nascent Okazaki fragments synthesized on the lagging strand template (Zieg, Maples et al., 1978). None of these models explains why collapsed replication forks are more abundant near oriC. In a recent report (Grigorian, Lustig et al., 2003), elevated dnaA" expression was described to reduce viability when the host strain carried a mutation in one of a number of genes (recA, recB, recC, recF, ruvA or rqu). These genes function in homologous DNA recombination and in the repair of double strand breaks. In our study, we present direct evidence that DSBs are produced by hyperinitiation, supporting the model that newly initiated replication forks collide from behind with those from previous initiations. This is an alternate mechanism of replication fork collapse as other models invoke a stalled fork due to a lesion in DNA or inactivity of a component of the replication fork machinery. 199 Whereas the observations above show that hyperinitiation leads to replication fork collapse, forks occasionally abort or collapse when E. coli is propagated under normal growth conditions, requiring the reassembly of the replication fork machinery to complete duplication of the genome and to maintain viability . A provocative notion is that the period of sequestration of oriC after initiation serves to allow collapsed replication forks to restart so that a trailing replication fork is unlikely to run into the one ahead. The premise for this idea is based on studies done with yeast. In the DNA replication checkpoint response, Rad53 and Mecl in S. cerevisiae and their orthologues are proposed to protect stalled or slowly moving replication forks from experiencing DSBs and potentially lethal genomic rearrangements until the repair of damaged DNA is finished (Cha and Kleckner, 2002; Sogo, Lopes et al., 2002). Lethality of dnaAcos. In rich media, induced expression of heterologous genes under araBAD promoter control is elevated several hundred-fold (Guzman,Belin et al., 1995). Interestingly, this magnitude of DnaA" overproduction is not lethal in a wild type host strain (Table 3). By comparison, DnaAcos expression when chromosomally encoded interferes with viability at 30°C . Its phenotype has been attributed to hyperinitiation. Because the dnaAcas allele carried in the bacterial chromosome is difficult to manage, often giving rise to spontaneous extragenic suppressors that compensate for its phenotype of growth interference, we placed it under araBAD promoter control . On measuring the level of initiation by the ratio of oriC to relE by quantitative PCR, elevated DnaAcos protein at 30°C yielded an oriC-to-relE ratio of 10.1 :1: 1.1, 2.7-fold greater than that obtained with DnaA“ protein at a comparable level (oriC- to-relE ratio of 3.7 :t 0.1). These observations strongly suggest that the dnaAcas allele fails to respond to a regulatory signal whereas the lesser magnitude of hyperinitiation by DnaA” suggests that it responds to negative regulation. In the present study, we showed that induced expression of dnaAcas interferes with viability (Table 3), and results in an 200 increased amount of broken chromosomal DNA in the absence of recB function (42% linear DNA compared to 16% for the empty vector control, Table 4). These observations support the conclusion that the lethality of dnaAcas is due to a level of hyperinitiation beyond that seen for the dnaA+ allele, followed by the accumulation of collapsed replication forks that are not efficiently processed by the replication restart pathway. The replication restart pathway. Under conditions that cause replication fork collapse and double strand breaks in DNA, it has been proposed that the viability of recA null mutants is due to degradation of linear chromosomal DNA back to oriC where DNA replication can be reinitiated . We examined the effect of DnaA” overproduction in a recA null strain (JJC16, relevant genotype A(recA-srl)::Tn10; (Michel, Ehrlich et al., 1997)), and observed that viability was not affected, in agreement with this mechanism. Despite this, the colonies that arose were smaller than those observed when dnaA” expression was induced in a recA+ host. Colonies obtained in the presence of arabinose are smaller than those obtained in the absence of induced expression. This result suggests that the growth of each colony, the product of multiple cell cycles, is retarded due to the time required to repair collapsed replication forks with each round of DNA replication. Hyperinitiation in eukaryotes. Recently, the structure of the C-terrninal two- thirds of a DnaA orthologue from Aquifex aeolicus was determined and shown to be very similar to the structure of Cdc6/Ore] from Pyrobelicum aeraphilum (Erzberger, Pirrucello et al., 2002). Assuming that the structure of S. cerevisiae Cdc6 (and Cdc18 of S. pombe) is similar to this archael orthologue, it is remarkable that Cdc6 and DnaA share conserved structures and functions. In S. cerevisiae, Cdc6 (and Cdc18 in S. pombe) recruits Mcm 2-7 helicase onto chromatin to assemble the pre-replicative complex (Tanaka, Knapp et al., 1997; You, Komamura et al., 1999). DnaA performs a similar 201 role in recruiting DnaB helicase to oriC (Marszalek and Kaguni, 1994; Sutton, Carr et al., 1998). Liang and Stillman described a cdc6 allele (cdc6-3) that showed an abnormal increase in nuclear DNA content (>2C) in both asynchronously growing cells at nonpermissive temperature, as well as in mutants released from an or factor block at permissive temperature (Liang and Stillman, 1997). These investigators suggested that Cdc6-3 fails to respond to the negative control exerted by mitotic cyclins. It is interesting to consider the similar defects of cdc6-3 and dnaAcas, and that the inviability of the cdc6- 3 mutant is due to the failure to repair collapsed replication forks induced by more frequent initiations. 202 Acknowledgements We thank Jonathan Walton (Michigan State University for the use of the pulse field gel apparatus, Benedicte Michel (Institut National de la Recherche Agronomique) for recB and recA strains, Steve Sandler (University of Massachusetts, Amherst) for the priA strain, and Yohko Fujimura (National Institute for Basic Biology) for the strain carrying terL inserted at lacZ and its isogenic partner. We also thank members of the J. Kaguni lab and the N. Cozzarelli lab for insightful discussions. This work was supported by NIH grant GM33992 and by the Michigan Agricultural Experiment Station to JMK. 203 References Anderson, D.G., and Kowalczykowski, SC. (1997) The translocating RecBCD enzyme stimulates recombination by directing RecA protein onto ssDNA in a chi- regulated manner. QLII 90: 77-86. Atlung, T., and Hansen, EC. (1993) Three distinct chromosome replication states are induced by increasing concentrations of DnaA protein in Escherichia coli. Journal of Bocteriology 175: 6537-6545. Autret, S., Levine, A., Vannier, F., Fujita, Y., and Seror, SJ. (1999) The replication checkpoint control in Bacillus subtilis: identification of a novel RTP-binding sequence essential for the replication fork arrest after induction of the stringent response. Molecular Microbiolology 31: 1665-1679. Birren, B., and Lai, E. (1993) Pulse Field Gel Electrophoresis: A Practical Guide. New York: Academic Press. Bussiere, D.E., and Bastia, 1D. (1999) Termination of DNA replication of bacterial and plasmid chromosomes. Molecular Microbiology 31: 161 1- 1618. Cha, RS, and Kleckner, N. (2002) ATR homolog Mecl promotes fork progression, thus averting breaks in replication slow zones. Scieuog 297: 602-606. Cooper, S., and Helmstetter, CE. (1968) Chromosome replication and the division cycle of Escherichia coli B/r. Janna! of Molecular Biology 31: 519-540. Cox, M.M., Goodman, M.F., Kreuzer, K.N., Sherratt, D.J., Sandler, SJ., and Marians, KJ. (2000) The importance of repairing stalled replication forks. Nature 404: 37-41. Erzberger, J .P., Pirruccello, M.M., and Berger, J .M. (2002) The structure of bacterial DnaA: implications for general mechanisms underlying DNA replication initiation. Buroooan Molecular Biology Qrganization Journal 21: 4763-4773. Flores, M.J., Bieme, H., Ehrlich, S.D., and Michel, B. (2001) Impairment of lagging strand synthesis triggers the formation of a RuvABC substrate at replication forks. Eurooean Molecular Biology Qrganization Journal 20: 619-629. French, S. (1992) Consequences of replication fork movement through transcription units in vivo. Scienoe 258: 1362-1365. 204 Gautam, A., and Bastia, D. (2001) A replication terminus located at or near a replication checkpoint of Bacillus subtilis functions independently of stringent control. Journol 0t Biologicol themistuy 276: 8771-8777. Grigorian, A.V., Lustig, R.B., Guzman, E.C., Mahaffy, J .M., and Zyskind, J .W. (2003) Escherichia coli cells with increased levels of DnaA and deficient in recombinational repair have decreased viability. Journal at Bactoriology 185: 630-644. Guzman, L.M., Belin, D., Carson, M.J., and Beckwith, J. (1995) Tight regulation, modulation, and high-level expression by vectors containing the arabinose PBAD promoter. Journal of Bacteriology 177: 4121-4130. Horiuchi, T., and Fujimura, Y. (1995) Recombinational rescue of the stalled DNA replication fork: a model based on analysis of an Escherichia coli strain with a chromosome region difficult to replicate. Journal of Booterioiogy 177: 783-791. Hwang, D.S., and Kaguni, J.M. (1988) Purification and characterization of the dnaA46 gene product. Journal at Biological (Zhemistuy 263: 10625-10632. Kellenberger-Gujer, G., Podhajska, A.J., and Caro, L. (1978) A cold sensitive dnaA mutant of E. coli which overinitiates chromosome replication at low temperature. Molecular and General Genetics 162: 9-16. Kitagawa, R., Mitsuki, H., Okazaki, T., and Ogawa, T. (1996) A novel DnaA protein- binding site at 94.7 min on the Escherichia coli chromosome. Molooular Microbiology 19: 1 137-1 147. Kitagawa, R., Ozaki, T., Moriya, S., and Ogawa, T. (1998) Negative control of replication initiation by a novel chromosomal locus exhibiting exceptional affinity for Escherichia coli DnaA protein. (2enes& Development 12: 3032-3043. Kobayashi, T., and Horiuchi, T. (1996) A yeast gene product, Fobl protein, required for both replication fork blocking and recombinational hotspot activities. Genes Qells 1: 465-474. 205 Kobayashi, T., Heck, D.J., Nomura, M., and Horiuchi, T. (1998) Expansion and contraction of ribosomal DNA repeats in Saccharomyces cerevisiae: requirement of replication fork blocking (Fobl) protein and the role of RNA polymerase I. Qenesfi Devolooment 12: 3821-3830. Kornberg, A., and Baker, T.A. (1992) DNA Replication Second Edition: W.H. Freeman And Company. Lane, H.E., and Denhardt, D.T. (1975) The rep mutation. IV. Slower movement of replication forks in Escherichia coli rep strains. Journol ot Moleculor Biology 97: 99-112. Levine, A., Vannier, F., Dehbi, M., Henckes, G., and Seror, SJ. (1991) The stringent response blocks DNA replication outside the ori region in Bacillus subtilis and at the origin in Escherichia coli. Journal at Molecular Biology 219: 605-613. Liang, C., and Stillman, B. (1997) Persistent initiation of DNA replication and chromatin-bound MCM proteins during the cell cycle in cdc6 mutants. Genes é. Development 11: 3375-3386. Lin, B., and Alberts, B.M. (1995) Head-on collision between a DNA replication apparatus and RNA polymerase transcription complex. Scionce 267: 1131-1 137. Marszalek, J ., and Kaguni, J .M. (1994) DnaA protein directs the binding of DnaB protein in initiation of DNA replication in Escherichia coli. Journal of Biological Qhemistgy 269: 4883-4890. Messer, W., and Weigel, C. (1997) DnaA initiator--also a transcription factor. Moleoulor Microbiology 24: 1-6. Messer, W. (2002) The bacterial replication initiator DnaA. DnaA and oriC, the bacterial mode to initiate DNA replication. FEMS Miorobiol Rev 26: 355-374. Michel, B., Ehrlich, S.D., and Uzest, M. (1997) DNA double-strand breaks caused by replication arrest. r ean olec lar i l r ani t' n umal 16: 430- 438. Michel, B., Flores, M.J., Viguera, E., Grompone, G., Seigneur, M., and Bidnenko, V. (2001) Rescue of arrested replication forks by homologous recombination. Proo Natl Acad Sci (J SA 98: 8181-8188. 206 Nurse, P., Zavitz, K.H., and Marians, KJ. (1991) Inactivation of the Escherichia coli priA DNA replication protein induces the SOS response. Joumol ot Baoteriology 173: 6686-6693. Ogawa, T., Yamada, Y., Kuroda, T., Kishi, T., and Mariya, S. (2002) The datA locus predominantly contributes to the initiator titration mechanism in the control of replication initiation in Escherichia coli. Molecular Microbiology 44: 1367- 1375. Peter, BJ., Ullsperger, C., Hiasa, H., Marians, K.J., and Cozzarelli, N.R. (1998) The structure of supercoiled intermediates in DNA replication. [_efl 94: 819-827. Roth, A., and Messer, W. (1998) High-affinity binding sites for the initiator protein DnaA on the chromosome of Escherichia coli. Moleculor Miorobiology 28: 395- 401. Sandler, S.J., Samra, H.S., and Clark, AJ. (1996) Differential suppression of priA2::kan phenotypes in Escherichia coli K-12 by mutations in priA, lexA, and dnaC. Genetics 143: 5-13. Saveson, CJ., and Lovett, S.T. (1997) Enhanced deletion formation by aberrant DNA replication in Escherichia coli. Genetics 146: 457-470. Seigneur, M., Bidnenko, V., Ehrlich, S.D., and Michel, B. (1998) RuvAB acts at arrested replication forks. Cell 95: 419-430. Séror, S.J., Vannier, R, Levine, A., and Henckes, G. (1986) Stringent control of initiation of chromosomal replication in B. subtilis. Nature 321: 709-710. Simmons, L.A., and Kaguni, J.M. (2003) The dnaAcos allele of Escherichia coli: hyperactive initiation is caused by substitution of A184V and Y271H, resulting in defective ATP binding and aberrant DNA replication control. Moleoular Microbiology 47: 755-765. Skarstad, K., Lobner-Olesen, A., Atlung, T., von Meyenburg, K., and Boye, E. (1989) Initiation of DNA replication in Escherichia coli after overproduction of the DnaA protein. Molecular and General Genetics 218: 50-56. Sogo, J.M., Lopes, M., and F oiani, M. (2002) Fork reversal and ssDN A accumulation at stalled replication forks owing to checkpoint defects. Scionoe 297 : 599-602. 207 Sutton, M.D., and Kaguni, J .M. (1997a) Threonine 435 of Escherichia coli DnaA protein confers sequence- specific DNA binding activity. Journal of Biological Chemista 272: 23017-23024. Sutton, M.D., and Kaguni, J.M. ( l997b) The Escherichia coli dnaA gene: four functional domains. Journol at Molecular Biology 274: 546-561. Sutton, M.D., Carr, K.M., Vicente, M., and Kaguni, J .M. (1998) E. coli DnaA protein: the N-terrninal domain and loading of dnaB helicase at the E. coli chromosomal origin. Journal of Biologicol Gbomistpy 273: 34255-34262. Tanaka, T., Knapp, D., and Nasmyth, K. (1997) Loading of an Mcm protein onto DNA replication origins is regulated by Cdc6p and CDKs. Gell 90: 649-660. Vilette, D., Uzest, M., Ehrlich, S.D., and Michel, B. (1992) DNA transcription and repressor binding affect deletion formation in Escherichia coli plasmids. European Molecular Biology Qrganizatiou Journal 11: 3629-3634. You, Z., Komamura, Y., and Ishimi, Y. (1999) Biochemical analysis of the intrinsic Mcm4-Mcm6-mcm7 DNA helicase activity. Molecular ond Gollulor Biology 19: 8003-8015. Zieg, J ., Maples, V.F., and Kushner, S.R. (1978) Recombinant levels of Escherichia coli K-12 mutants deficient in various replication, recombination, or repair genes. Journol ot Bocteriology 134: 958-966. 208 Chapter VI Summary and Perspectives 209 The genetic and biochemical analysis of numerous dnaA alleles has provided important details revealing the biochemical functions of DnaA protein that are required for initiation of DNA replication (Hansen, Koefoed et a1. 1992; Sutton and Kaguni 1995; Sutton 1996; Sutton and Kaguni 1997; Sutton and Kaguni 1997; Sutton and Kaguni 1997; Sutton, Carr et al. 1998). It has been suggested that the N-terrninal region of DnaA protein is important for both oligomerization and interaction with the replicative helicase DnaB (Sutton, Carr et al. 1998; Seitz, Weigel et al. 2000; Mima, Makise et al. 2002). Specific amino acids have not been identified that are required for oligomerization, and oligomerization has not been demonstrated as a required function for replication from the bacterial origin. Here, we identified three amino acids at the N-terminal region that are required for DnaA protein oligomerization, and this activity was demonstrated to be required for replication from oriC. These results suggest that the first 10 amino acids of the N-terminal region are important for oligomer formation, and substitution of only certain amino acids in this region results in a failure of DnaA protein to form higher order complexes. Based on homology modeling of putative DnaA protein monomer interaction, it has been suggested that DnaA protomers may form a hexamer at oriC (Erzberger, Pirruccello et al. 2002). We were unable to clearly define a hexamic complex formed by DnaA protein, although oligomeric species corresponding to dimers, trimers and tetramers were observed. The alleles encoding the N-terminal amino acid substitutions described above can be grouped into a category of dnaA alleles that are defective for DNA replication. A few hundred of these alleles have been described, yet only two alleles have been clearly identified that encode a more active form of DnaA protein (Kellenberger-Gujer, Podhajska et al. 1978; Braun, O'Day et al. 1987). The most well characterized is the dnaAcas allele. This allele encodes four amino acid substitutions (Q156L+ A184V+H252Y+Y271H). The contribution of these individual amino acid changes to 210 the increase in replication activity observed with DnaAcos had not been examined. It seemed unlikely that all four missense mutations contributed to the dnaAcas phenotype, and we felt that determining the amino acid substitutions responsible for the dnaAcas phenotype would reveal the requirements for hyperactive initiation. In this work, we determined that the A184V and Y271H amino acid substitutions were responsible for the phenotype elicited by dnaAcas. This work indicated that the major contributing factor is a decrease in ATP binding activity of the mutant protein, and the Y271H substitution appears to stabilize the activity of the A184V bearing protein at elevated temperatures. This result underscores the importance of ATP binding as a regulatory factor that limits the frequency of initiation. After determining the mutations required for the dnaAcas phenotype we decided to identify novel alleles of the dnaA gene with similar behavior. These experiments were intended to determine the possible causes of hyperactive initiation. In this study, we identified seven novel hyperactive dnaA alleles. These alleles were studied genetically and we determined that they are defective in the regulation of initiation from the bacterial origin of replication, oriC. It appeared as though mutant alleles had been isolated that encoded amino acid substitutions affecting several different functional domains of DnaA protein. Homology modeling revealed that several of the amino acid changes were in close proximity to the site for ATP binding. Despite this, amino acid substitutions were isolated in the N-terminus (G79D), and in the helix-tum-helix (HTH) DNA binding (E445K) motif of DnaA protein. These studies indicate that DnaA protein requires at least three biochemical activities to initiate DNA replication with the appropriate frequency. The mechanism that causes lethality after hyperactive initiation has never been identified. It has been suggested that replication forks stall because a processivity factor for the replicase has become limiting. Although this explanation is possible, it seems 211 unlikely because overexpression of the beta clamp (required for processive DNA synthesis of DNA polymerase IH Holoenzyme) does not suppress dnaAcas (data not shown) or restore the chromosome content to wild type cell profiles after DnaA protein overexpression. To determine the pattern of replication fork stalling we used genomic microarrays to analyze the replication fork distribution during normal culture growth and under-conditions inducing hyperactive initiation. This study revealed that hyperactive initiation causes replication fork stacking near oriC. This suggested to us that viability might require replication fork restart and homologous recombination proteins to repair, and then restart stalled replication forks. These expectations were verified by determining that DnaA protein overexpression was lethal in strains deficient in homologous recombination and replication fork restart. We verified these results by visualizing linear DNA accumulation in recB' cells (deficient in homologous recombination and linear DNA degradation by exonucleaseV activity contained in the RecBCD enzyme). These results indicate that, when replication forks are released from oriC with an increased frequency, the consequence is damage to genome integrity. The data obtained in these studies raises several important questions that are relevant to our understanding of replication initiation in E. coli. Determination of the N- terminal structure of DnaA protein will be important to reveal the contacts that are made between DnaA monomers. Structural modeling of DnaA protein monomers to the p97 hexameric complex (human protein required for assembly of the SNARE complex) suggests that DnaA protein could form a hexamer (Erzberger, Pirruccello et a1. 2002). It would also be useful to crystalize DnaA protein bound to ATP to reveal the amount of structural remodeling that occurs. This experiment would confirm that AAA” proteins require ATP binding for a conformational change required for their activity (Koonin 1992; Koonin and Gorbalenya 1992; Koonin 1993). 212 We know that DnaA protein requires ATP binding to be active for replication (Sekimizu, Bramhill et a1. 1987). Many mutant proteins that are defective for ATP binding are also defective for replication. With the exception of DnaAcos protein this conclusion is correct. DnaAcos protein has feeble ATP binding activity yet it is more active than the wild type protein (Katayama and Kornberg 1994; Katayama and Crooke 1995). How is this possible? The mechanism causing hyperactive initiation is still unknown, and experiments need to directed toward revealing the specific biochemical requirements that result in the regulation of DNA replication. In the case of hyperactive cdc6 alleles (S. cerevisiae) they cause a constant loading of an MCM2-7 complex containing helicase activity (Liang and Stillman 1997). It is possible that hyperactive DnaA mutant proteins load DnaB more frequently than the wild type protein. Both primase (DnaG) and DNA polymerase HI holoenzyme have specificity for DnaB. Thus, once DnaB is loaded the other components of the replication fork should assemble. Some of the mutant DnaA proteins may be impaired in their ability to discriminate between ATP and ADP. In this situation the mutant proteins might remain active in the ADP bound form and are inert to the regulation imposed by the intrinsic ATPase activity of DnaA protein. Another potential mechanism that causes the hyperactive mutant DnaA proteins to remain active, is that their ATPase activity is not stimulated by the beta clamp of DNA polymerase III holoenzyme or Hda protein. Under this model, mutant proteins are not responsive to negative regulation by beta clamp and Hda resulting in the continued binding of ATP leaving the mutant protein in a constitutively active form. Future experiments should be directed to discriminate between these possible models for hyperactive initiation. 213 References Bruan, R. E., K. O'Day and A. Wright (1987). "Cloning and characterization of dnaA(Cs), a mutation which leads to overinitiation of DNA replication in Escherichia coli K-12." Journal of Bacteriology 169(9): 3898-3903. Erzberger, J. P., M. M. Pirruccello and J. M. Berger (2002). "The structure of bacterial DnaA: implications for general mechanisms underlying DNA replication initiation." European Molecular Biology Grganization Journal 21(18): 4763-4773. Hansen, F. G., S. Koefoed and T. Atlung (1992). "Cloning and nucleotide sequence determination of twelve mutant dnaA genes of Escherichia coli." Molecular and General Genetics 234(1): 14-21. Katayama, T. and E. Crooke (1995). "DnaA protein is sensitive to a soluble factor and is specifically inactivated for initiation of in vitro replication of the Escherichia coli minichromosome." LoumuloflBuLlogloulflbemlstu 270(16): 9265-9271. Katayama, T. and A. Kornberg (1994). "Hyperactive initiation of chromosomal replication in viva and in vitro by a mutant initiator protein, DnaAcos, of Escherichia coli." Joumul of Biological Ghemisu'y 269(17): 12698-12703. Kellenberger-Gujer, G., A. J. Podhajska and L. Caro (1978). "A cold sensitive dnaA mutant of E. coli which overinitiates chromosome replication at low temperature." What; 162(1): 9-16- Koonin, E. V. (1992). "DnaC protein contains a modified ATP-binding motif and belongs to a novel family of ATPases including also DnaA." Nucleie Acids Researeb 20(8): 1997. Koonin, E. V. (1993). "A superfamily of ATPases with diverse functions containing either classical or deviant ATP-binding motif." Journal of Moleoular Biology 229(4): 1 165-1174. Koonin, E. V. and A. E. Gorbalenya (1992). "The superfamily of UvrA-related ATPases includes three more subunits of putative ATP-dependent nucleases." Protein Beg Data Anal 5(1): 43-45. Liang, C. and B. Stillman (1997). "Persistent initiation of DNA replication and chromatin-bound MCM proteins during the cell cycle in cdc6 mutants." Genes & Development 11(24): 3375-3386. 214 Mirna, S., M. Makise, M. Katerasawa, T. Tsuchiya and T. Mizushima (2002). "Conserved hydrophobic amino acid residues in the N-terrninal region of DnaA protein are involved in DnaA-DnaA interaction." Biochem Joumal 11. Seitz, H., C. Weigel and W. Messer (2000). "The interaction domains of the DnaA and DnaB replication proteins of Escherichia coli." Molecular Microbiology 37(5): 1270—1279. Sekimizu, K., D. Bramhill and A. Kornberg (1987). "ATP activates DnaA protein in initiating replication of plasmids bearing the origin of the E. coli chromosome." Ce_11 50(2): 259-265. Sutton, M. D. (1996). Mutants of the E. coli dnaA gene: Genetic and Biochemical Analysis of its replication activity, Michigan State University. Sutton, M. D., K. M. Carr, M. Vicente and J. M. Kaguni (1998). "Escherichia coli DnaA protein. The N-terminal domain and loading of DnaB helicase at the E. coli chromosomal origin." Journal of Biological Ghemistry 273(51): 34255-34262. Sutton, M. D. and J. M. Kaguni (1995). "Novel alleles of the Escherichia coli dnaA gene are defective in replication of pSC101 but not of oriC." Journal of Bacteriology 177(22): 6657-6665. Sutton, M. D. and J. M. Kaguni (1997). "The Escherichia coli dnaA gene: four functional domains." Journul of Molecular Biology 274(4): 546-561. Sutton, M. D. and J. M. Kaguni (1997). "Novel alleles of the Escherichia coli dnaA gene." Journal of Molecular Biology 271(5): 693-703. Sutton, M. D. and J. M. Kaguni (1997). "Threonine 435 of Escherichia coli DnaA protein confers sequence- specific DNA binding activity." Journul of Biological Ghemistgl 272(37): 23017-23024. 215