THE Pl PLASMID PARTITION PROTEIN ParA: ROLES FOR ATP BINDING AND HYDROLYSIS IN PLASMID PARTITION

Megan Jeannette Davey

Thesis subrnitted in confomiity with the requirements for the Degree of Doctor of Philosophy,

Graduate Department of Molecular and Medical Genetics in the University of Toronto.

O by Megan Jeannette Davey 1997

National Library 191 ,Cm,& Bibliothèque nationale du Canada

Acquisitions and Acquisitions et Bibliographic Services services bibliographiques

395 Wellington Street 395, nie Wellington Ottawa ON K1 A ON4 OttawaON K1AON4 Canada Canada

The author has granted a non- exclusive licence allowing the National Library of Canada to reproduce, loan, distribute or seil copies of this thesis in rnicrofom, paper or electronic formats.

The author retains ownership of the copyright in this thesis. Neither the thesis nor substantial extracts fiom it may be printed or otherwise reproduced without the author's permission.

L'auteur a accordé une licence non exclusive permettant à la Bibliothèque nationale du Canada de reproduire, prêter, distribuer ou vendre des copies de cette thèse sous la forme de microfiche/fïlm, de reproduction sur papier ou sur format électronique.

L'auteur conserve la propriété du droit d'auteur qui protège cette thèse. Ni la thèse ni des extraits substantiels de celle-ci ne doivent être imprimés ou autrement reproduits sans son autorisation.

The P l plasmid partition protein ParA: Roles for ATP binding and hydrolysis in plasmid

partition.

Doctor of Philosophy , 1997

Megan Jeannette Davey

Department of Molecular and Medical Genetics, University of Toronto

The Pl prophage plasmid is stably maintained in its bacterial host despite its low

copy number. This stability is due in part to an active partition systern, par, that encodes the

genes for two proteins, ParA and ParB, as well as a cis acting site, parS. ParA is required

for at least two roles during partition in vivo. First, ParA represses expression of its own

gene as well as pari3 from a promoter upstream of parA. This repression by ParA is

stimulated by ParB. There is a second requirement for ParA that is thought to reflect a

direct, as yet undefmed role for ParA in the partition process. In vitro, ParA is an ATPase

that is stimulated by ParB and by non-specific DNA. In addition, ParA binds the par operon

promoter region, consistent with ParA's repressor activity in vivo. In this thesis I examine

the effects of ATP binding and hydrolysis on ParA, particularly on ParA interaction with

parOP and with ParB. 1 show that adenine nucleoside di- and triphosphates stimulate ParA

DNA binding. ATP hydrolysis is not required for ParA DNA binding and, in fact, inhibits

ParA DNA binding. The efiects of ATP binding and hydrolysis on ParA-parOP interaction

are correlated with changes in ParA conformation including, but not limited to, ParA

dimerization. 1 have also found that ParB stimulates ParA DNA binding in vitro, consistent

with ParB's role as a CO-repressor in vivo. ParB stimulation of ParA DNA binding requires

ATP hy droly sis. However, preliminary results suggest that ATP hydroly sis is not required

on ParA-ParB

repressor

for physical association of ParA and ParB, therefore ATP must exert its effect

interaction via a different mechanism. My results have led to a mode1 of ParA

function in which ATP binding and ATP hydrolysis have separable roles and have led to the

suggestion that ATP binding and hydrolysis have separable roles in ParA's partition function

as well.

iii

ACKNOWLEDGEMENTS

1 am highly appreciative of the support of fnends. family and colleagues d u ~ g my

graduate career. In paaicular, 1 thank my supervisor, Barbara FunneU, for her guidance and

fnendship. Other members of the FunneIl lab, past and present, have also contributed

inteikctuaiiy and otherwise to my graduate life. In panicular, 1 wish to thank Liane Gagnier

for plasmids and humour, and Jennifer Surtees for His-tagged ParB and for periodic

perspective adjustments. I have dso benefited from the input of other members of the

department iocluding Dr. Gold and Mike Moran, who were on my supervisory cornmittee. In

addition, 1 thank Man Davidson and the members of his lab for their assistance with CD. I

aclcuowledge financial support from Ontario Graduate Scholarships and University of Toronto

Open Fellowships.

1 have made many fiends in graduate school and 1 am grateful for their friendship

and shared experiences, especially the not-so- "New Kids" . 1 also thaok my "non-science"

friends for their patience and encouragement. Finally, 1 have always enjoyed the love and

support of my family, Dad, Mom, Fran, Katherine, Chris and Janie and 1 thank them for

their encouragement.

TABLE OF CONTENTS

PREFACE Thesis abstract Acknowledgements Table of contents List of tables List of figures List of abbreviations

CHAPTER 1: GE- INTRODUCTION

PLASMID PARTITION General mechanisms

Limiting site versus plasmid pairing modefs Role of the membrane in partition One pair partition versus equipartition S m

Pl partition The partition complex ParA plays two roles in partition Regdation of par gene expression ParA ATPase ParA and ParB interactions ParA and ParB homologs s m w

Other plasmid partition systems F and P7 plasmids Incm plasmids RK2 Other plasmids

OTHER FACTORS CONTRIBUTING TO PLASMID STABILITY Copy number control

Replication control of copy number Multimer resolution

Kilier sy stems

CHROMOSOME PARTITION Events prior to positioning

Decatenation of the chromosome Resolution of chromosome multimers

ii iv v viii k

TABLE OF CONTENTS (cont.)

Chromosome positionhg The E. coli muk mutants Other mutants

ATP BINDING AND IfYDROLYSIS

THESIS RATIONALE

CHAPTER 2: A ROLE FOR ATP IN SITE-SPECIFIC DNA BINDING BY PUA

INTRODUCTION

EXPER.IMENTAL PROCEDURES Strains and plasmids Reagents and buffers Antibody production and purification ParA purification DNase 1 protection assays

RESULTS Purification of PUA DNA binding is stimulated by, but does not require, ATP Effects of ATP hydrolysis on DNA binding ATP and ADP affect the sedimentation behaviour of ParA ParA forrns dimers

DISCUSSION

CHAPTER 3: MODULATION OF ParA BY ATP, ADP AND ParB

INTRODUCTION

EXPERIMENTAL PROCEDURES Reagents and buffers Proteins Nucleotide binding Protein concentration detennination DNA binding assay Quantification of DNase 1 protection assays

RESULTS ATP and ADP binding by ParA Nucleotide effects on ParA stability ATP binding and hydrolysis alter ParA conformation ParB stimulates ParA DNA binding Nucleotide specificity of ParB stimulation of ParA DNA-binding activity

TABLE OF CONTENTS (cont.)

DISCUSSION ParA conformation with and without nucleotide Differentiation between adenine nucleoside diphosphates and triphosphates Effects of ATP hydrolysis on ParA structure and finction ParA-ParB interactions Roles of ATP in par gene regulation and partition

CHAPTER 4: ISOLATION OF ParA-ParB COMPLEXES AND FUTURE DIRECTIONS

INTRODUCTION

METHODS Preparation of 35S-labelled ParA extracts Preparation of "S-PUA protein Antibodies and Protein A Sepharose beads Co-immunoprecipitation of ParA and ParB

DISCUSSION AND FUTURE DIRECTIONS ParA-ParB interactions ParA interaction with the partition complex Roles of ATP binding and hydrolysis in par gene expression and partition

REFERENCES

LIST OF TABLES

CHAPTER 3

Table 3-1: Relative binding of AT' and ADP by ParA as measured

using equilibriurn gel filtration.

Table 3-2: Effects of adenine nucleotides on ParA helicity, measured by CD. 91

Table 3-3: Effects of adenine nucleotides on heat denaturation of PUA. 95

LIST OF FIGURES

CHAPTER 1 Page

Figure 1-1: Limiting site and plasmid pairing partition models. 6

Figure 1-2: The Pl plasmid partition system. I l

Figure 13: ParA and PqB contain regions of homology to other proteins. 17

Figure 1-4: A mode1 for Pl plasmid partition. 20

Figure 1-5: The plasmid partition systems of F. W, RI and RK2. 23

Figure 1-6: ATP binding and hydrolysis 35

C-R 2

Figure 2-1: Purification of ParA. 45

Figure 2-2: ATP and dATP hydrolysis by ParA. 47

Figure 2-3: ATP stimulation of ParA-purOP interaction. 50

Figure 2-4: Summary of the extent and pattern of protection from DNase 1 cleavage 52

by ParA protein.

Figure 2-5: Nucleotide effects on ParA DNA-binding activity. 54

Figure 2-6: ATP and ADP affect ParA sedimentation rate. 58

Figure 2-7: EGS crosslinking of ParA. 63

CHAPTER 3

Figure 3-1: Equilibrium gel filtration analysis of ATP binding by ParA.

Figure 3-2: Am, ATPyS and ADP stabilize ParA.

Figure 3-3: ParA conformation with and without nucleotide.

Figure 3-4: The effects of adenine nucleotides on ParA stability .

LIST OF FIGURES (cont.) Page

Figure 3-5: ParB stimulates ParA DNA binding in the presence of ATP. 98

Figure 3-6: ParB does not sùmulate ParA DNA binding in the presence of ADP 100

or in the absence of nucleotide.

CHAPTER 4

Figure 4-1: Physical association of PUA and ParB.

Figure 4-2: Association of ParA and ParB in the absence of nucleotide.

LIST OF ABBREVIATIONS

AMP-PNP: adenylyl-imidodiphosphate

ATPyS : adenosine-5 '-O-(3-thiotriphosphate)

b p : basepair(s)

BSA: bovine serum albumin

CD: circular dichroism

DNase 1: bovine pancreatic deoxynbonuclease 1

DTT: 1,4-dithiothreitol

EDTA: disodium ethylenediaminetetra-acetate

EGS: ethylene glycolbis(succinimidyl succinate)

GST: glutathione-S-transferase

IgG: immunoglobulin G

IHF': integration host factor

WïG: isopropyl-0-D-thiogalactopyranoside

K, : dissociation constant

kDa: kilodaltons

parOP: the P l par operon operator/promoter sequences

TLC: thin layer chromatography

CHAPTER 1 : Generai Introduction

Nanirally occurring plasmids such as Pl, F, and R1 are stably maintained in

bacterial populations despite the fact that their copy number is extremely low. Their stability

is dependent on an active segregation or "partition" system. The mechanism of partition is

not howu, however it can be thought of as a process that positions plasmids to ensure that

each daughter ce11 receives a copy of the plasmid. For many plasmids, including P 1, F, and

RI, the plasmid-encoded elements required for partition have been identified, although their

functions in the partition process are not completely understood. Host-encoded factors are

also thought to participate in plasmid partition, but very few host factors have been

identified. In addition. very linle is known about the partition of bactenal chromosomes.

Plasmid partition systerns are thought to mimic the partition system of the bactenal

chromosome and may utilize components of the chromosome partition apparatus. In

particular, P l and F partition systems, studied in Escherichia coli, serve as good rnodels for

chromosome partition because they are the best characterized systems to date and the

plasmids are the same copy number as the E. coli chromosome. These plasmids are easy to

study, because unlike the chromosome, plasmids are not usually essential to the growth of

the cell. Pl and F also serve as models for other low-copy-number plasmids. Of particular

interest are homologous systems encoded by naturally occurring low-copy-nurnber plasmids

that carry virulence factors or antibiotic resistance genes and by the chromosomes of several

pathogenic bacteria.

The partition systems of low-copy-number plasmids such as F, Pl and R1 share

several common characteristics (see below), which suggest some sirnilarities in their partition

mechanism. Many plasmid partition systems encode an ATPase that is requhed for plasmid

partition, however a role for these ATPases in plasmid partition has not yet been identified.

In this thesis, 1 discuss the roles of ATP binding and hycûolysis in the function of the Pl

plasmid partition ATPase, ParA.

PLASMID PARTITION

General mechadans: Partition systems have been identified on many different Iowtopy-

number plasmids. Partition loci usuaily encode one or more tram acting factors that bind to a

cis acting site on the plasmid. This "partition site" is the site of assembly of the partition

apparatus which presumably consists of both plasmid and host encoded factors.

In order for partition to occur, a plasmid must be able to sense its position with

respect to the division plane in the ce11 and with respect to the other copy or copies of the

plasmid in the cell. One can predict therefore that the plasmids will interact via their partition

sites with specific components in the host ce11 and/or with each other. These predictions must

also account for the phenomenon of "partition-mediated incompatibility" expressed by the cis

acting partition sites. Partition-mediated incompatibility refers to the inability of two different

plasmids carrying the same partition site to exist stably in the same cell. This incompatibility

or cornpetition is specific for a particular partition site; plasmids with different partition sites

do not compete. The phenomenon of partition-mediated incompatibility predicts that plasmids

are randomiy selected from a pool of plasmids for partition.

An early mode1 for partition (86) incorporates some, but not dl, of these predictions

and has greatly influenced the development of subsequent models. Jacob et al. (86) proposed

that plasmids are continually bound to specific sites on the ceIl membrane. Both replication

and partition of the plasmid would be mediated by the plasmid's interaction with these sites.

Replication would result in duplication of the plasrnid and of the site to which the plasmid is

bound. Partition would be achieved by the ordered growth of membrane between the two

sites; the daughter plasmids would be moved apart as the ce11 elongates. While most current

partition models favour some type of membrane atmchment, several observations are

inconsistent with the Jacob model. First, this mode1 predicts that replication and partition

would be fûnctionally Iinked . However , plasmid par loci stabilize heterologous replicons

(Le., plasmids with different replication origins; 10, 28, 63, 1 12, 149). In addition, for Pl at

least, replication is not required for partition (188). Second. this model requires that new

membrane growth is selectively deposited at the center of the cell, pushing older material to

the poles. However, w w material is deposited in the membrane dispersively (71),

consequently membrane growth would not necessarily direct partition of plasmids. Finaily,

this model predicts that plasmids would not be selected at random for partition since

replication and partition occur at the same site. This is inconsistent with the observation that

partition sites carried on otherwise compatible replicons mediate incompatibility.

Limiring sire versus plasmid pairing models: Two more recent models incorporate aspects of

the Jacob model, but are more consistent with experimental observations. In one of these

models, plasmids are proposed to interact with specific sites in the host cell, on the inner

membrane for example (Fig. 1-1A). These sites are Iirnited in number (ideally, two per cell

at cell division), specific for the plasmid's partition apparatus, and located on either side of

the division plane. Following DNA replication, sister plasmids bind to these sites via their cis

sites. The location of one site on either side of the division plane ensures that each daughter

ce11 receives a copy of the plasmid. Incompatibility would result from cornpetition among

plasmids carrying the same partition locus for the limited sites. This model aiso predicts that

the ce11 would contain a different type of limiting site for each plasmid type or

incompatibility group. The large number of different plasmid incompatibility groups (more

than 20; 145) makes this model less appealing because the ce11 would require at least that

number of these putative limiting sites. Atternatively, plasmids could pair via their cis sites

limiting site J \ pairing

O lowcopy-number plasmid

Figure 1 - 1. Limiting site and plasmid pairing models. The plasmids (biack

circles) are partitioned by their par sites (blue circles). In A plaunids are partitioned

by interacting with a specific site in the host that is W t e d in number (orange

shape). In B plasmids pair before being partitioned by a general partition

mechanism,

(Fig. 1-1B) (10). The paired plasrnids would then be partitioned by a general partition

apparatus. In this model, the specificity for partition is provided by the plasmid rather than

by the host. Incompatibility would &se from heterologous pairing of plasmids. This

"pairing" model is generally favoured over the f ~ s t , "limiting site" model, however there is

no direct evidence for plasmid pairing.

Role of the membrane Ni panition: These models predict that plasmids will interact with

specific components in the host ce11 via the plasmid's partition system. Host-encoded factors

that spatially confime or move plasmids have not yet been identified. The only known

subcellular structures in E. coli cells are the chromosome and the ce11 membrane. Potentially,

plasmids could be partitioned sirnply by attaching to the chromosome via specific DNA sites.

However, F and Pl, two unit copy plasmids, segregate independently of the chromosome

(49, 62). Therefore, the membrane is a likely site of plasmid interaction with the host cell.

Partition-specific interactions with the membrane have not yet been fully exarnined. Although

it has been reported that plasmids or Par proteins precipitate with ce11 membranes during ce11

fractionation (72, 194), further examination of one of these reports (194) indicated that there

was no membrane association (78). In addition, the purification properties of several partition

proteins show that they are soluble and therefore these proteins are not exclusively membrane

bound. Nevertheless, some type of membrane association is a common characteristic in

partition models.

One pair partition versus equipanition: Another question that must be addressed is how many

copies of the plasmid are actively partitioned. Since plasmid copy number is measured per

bacterial chromosome, in a rapidly growing E. coli ce11 where there are more than two

copies of the chromosome, there wiii be more than IWO copies of a unit copy plasrnid. At

one extreme, only one pair of plasmids is paaitioned, referred to as "one pair partition".

7

Any remaining plasmids would be randomly segregated. At the other extreme, al1 plasmid

copies are actively segregated, referred to as "equipartition" . Alternatively , some copies of

the plasmid may be actively partitioned and some copies may be randomly segregated. One

pair partition is most consistent with a limiting site model and equipartition is most attractive

for the plasmid pairing model. However, an observation that one pair partition or

equipartition occurs would not exclude either the limiting site or pairing models.

Experimentally, it has not been possible to distinguish between one pair partition and

equipartition (eg . 143, 188). This question will most likely be answered when al1 plasmid-

host interactions have been identified.

Swnmary: It has not been possible to distinguish between the limiting site and plasmid

pairing models and between the one pair and equipartition models in vivo. Researchers are

restricted by having to measure populations of cells rather than individual cells and by the

modest amount of information about the factors involved in the positioning process. 1 have

taken the approach of characterizing one of the known components of the Pl partition

system, ParA. This characterization as well as the identification and characterization of other

factors involved in plasmid partition are necessary steps for the development of in vitro tests

of partition models.

Pl Partition: The Pl prophage is a unit copy plasmid that is stably maintained in its

Escherichia coli host; Pl is lost from a bacteriai ce11 less than once in every lû" ce11

divisions (84, 162). This faithful segregation is dependent on the plasmid's partition system

termed par (10). Pl par is included in a 2.6 kb Hindm-EcoRV fragment of Pl (Fig . 1-2A)

(3) that is sufficient to stabilize heterologous replicons (10, 12). The par region contains the

genes for two proteins, ParA and ParB, and a cis-acting site, pars (Fig. 1-2A and B) (3).

parA and parB constitute an operon, and are aanscribed from an autoregulated promoter

Figure 1-2. The P l plasmid partition system. A: The Pl par operon. The genes for ParA

and ParB are marked by the arrows, pars by the grey box and the promoter region @nrOP)

by the white box (3). The positions of restriction sites referred to in this and subsequent

chapters are indicated. The scaie in kbp is shown below the fragment. B: Sequence of the

pars site (3). The Box A and Box B ParB recognition motifs (60) are indicated by the white

orrows and grey boxes, respectively. The position of the IHF binding site (57) is indicated by

the hatched bar and restriction sites are indicated by the square brackets. C: The 388 bp

HindIII-XhoI fragment from Pl that contains the par operon promoter region. The positions

of the -35 and -10 transcriptional signals as well as the ribosome binding site (RBS) and the

parA start codon (ATG) (3) are marked by the boxes. The inverted arrows mark the position

of the 20 bp imperfect inverted repeat (3). The transcriptional start site is indicated by the

star (77). The scale, in bp, is marked above and below the fragment.

- nql n TCGATAAAAAGCCGAAGCCTTAAAC ATTAACTGACTGTTT

L

TTAAAGTAAATTACTCT d

Drsl

upstream of parA (in purOP; Fig 1-2C) (55). Disruption of parA, parB or pars results in

piasmid destabiiization (3). Pl partition is believed to hvolve the assembly of Pl and E. culi

encoded partition factors at parS. This assemblage mediates positionhg of the plasmid in a

process that is not yet understood. The only known host factor for Pl partition is integration

host factor 0 which contributes to partition but is not required (58). Other host factors

are believed to be involved, but they have not yet ken identified. Since Pl segregates

independently of the chromosome, the host factor is not a DNA site on the E. coli

chromosome (62).

The Paniriorz Compla: The earliest recognized step in partition is ionnation of the partition

complex1 by binding of ParB and IHF to pars (38. 57, 58). ParA is no& required for

formation of t h i s complex and the protein is thought to act at a later undefmed step in

partition (see below) (38, 57, 58). The pars site is the only element that is required on a

low-copy-number plasmid to ensure its proper segregation (as long as ParA and ParB are

provided Ni tram; 11). Its action is likened to eukaryotic centromeres; it foms the site at

which the segregation machinery is thought to assemble and at which pairing of the plasmids

is thought to occur (1 1).

pars is located in a 109 bp TaqI-Sv1 fragment from Pl (Fig. 1-2B). This region

includes two sets of ParB binding sites separated by an MF binding site. As determined by

footprinting assays, DMS interference assays and mutagenesis, ParB binds two different

sequence motifs, denoted "Box A" (A'MTCAA/C) and "Box B" (TCGCCA) (Fig. 1-2B) (40,

60). There is one copy of each Box to the left of the MF binding site and three copies of

Box A and one copy of Box B to the nght of the MF binding site. The limits of a functional

-- - - - . - - - - - -

'In this thesis, "partition complex" refen specifically to the complex formed when Pl ParB and MF bind to Pl parS. This term should not be confused with partition apparatus or partition mechanisrn which refer to generai complexes of partition factors.

11

pars site can be defined by its partition and incompatibility phenotypes. The 109 bp Ta@-

SgI fragment is referred to as pars' and is wild-type for partition and incompatibility.

However, pars-srnail, the 34 bp DraI-StyI fragment which does not contain the left half of

pars or the IHF binding site (Fig. 1-2B). is also active for partition (117). However,

pars-small is a less efficient par site than pars and is unable to compete with (ie. exert

incompatibility against) pars+ in an MF-dependent manner.

A role for IHF in P 1 partition was fmt discovered by its ability to stimulate ParB

binding to pars in vitro (58). IHF also stimulates, but is not required for, partition in vivo.

In E. coli mutants lacking MF. Pl is relatively stable. However, the plasmid is less stable in

MF* mutants than it is in wild-type cells (58). In addition, the

pars-small which lacks the DIF binding site, supports partition NI vivo (57, 118). Aithough

MF is not essential for partition, the incompatibility phenotype of pars-small bearing

plasmids indicates that MF is a component of the wild-type partition apparatus: In wild-type

cells, P l is destabilized by heterologous plasmids bearing pars+ whereas pars-small-bearing

plasmids are unable to compete (117). In MF- mutants, pars' behaves essentially as

pars-small and pars-small plasmids c m now compete with parSc (58). These observations

show that in wild-type cells the partition complex contains MF and that partition complexes

that do not contain MF cannot compete with partition complexes that do contain IHF,

probably because of their differing aff i t ies for ParB (58).

Several observations suggest that the partition cornpiex has a specific three

dimensional structure in which pars sequeaces are wrapped around a core of ParB and MF.

Fint, IHF binds to its site between the left and nght anns of pars, inducing a large bend

(57). In the absence of MF, ParB binds better to its sites in the nght half than in the Ieft half

of pars (57, 60). However, when IHF is present both arms are bound equally well by ParB.

12

Since IHF increases ParB a"nity for purs? this suggesu that IHF stimulates ParB binding by

aüowing ParB to contact both the left and right halves of pars simultaneously (57, 60). In

addition, the ParB-MF-purs complex prefen to form on supercoiled DNA (57); supercoiling

would favour a wrapped complex. Finally, helical phasing between ParB sites in the lefi and

right halves of pars is important suggesting that these sites interact with each other via ParB

(60, 74).

ParA plays a? l e m two roles in panirion. The fmt function of ParA is to repress

transcription of the par genes (55). A second function for ParA is ioferred from the

foilowing genetic experiments. First, mutations in parA are not complemented by plasmids

carrying only pari3 expressed from a heterologous promoter. even though such plasmids

efficiently complement parB mutants. These parA mutants require both parA and parB for

plasmid stability (55). Second, a mutant par promoter was constmcted that is not regulated

by ParA but expresses parA and parB at levels that allow partition. Under these conditions

ParA is not required for its regulatory function but is still required for partition (41). ParA's

second function in partition, referred to as its partition function, is assumed to be a direct

role in the positioning process.

Regularion of par gene expression: The genes for ParA and ParB form an operon, the

transcription of which is initiated from a promoter upstream of parA (Fig. 1-2A and C).

ParA represses transcription from this promoter in vivo (55). ParA repressor activity is

stimulated by ParB, however ParB has no repressor activiq on its own (55). A second

putative promoter lies between parA and parB (3) and there is some evidence that ParB may

also be expressed (unreylated) from this promoter (3, 55) but this has not yet been cleariy

demonstrated. ParA and ParB levels are important for proper segregation; overexpression of

either, or in some cases both, ParA ancilor ParB destabilizes P l (3, 56, 77). Overexpression

13

of partition proteins in other plasmid partition systerns has similar effefts (e.g . 9 1, 103). In

addition, many other plasmid partition genes are autoregulated, including those encoded by

F, P7, RIINRI, RK2, and pTAR (42, 47, 63, 65, 77, 91, 132, 182). PUA and ParB

provided from a mutant par promoter that is not autoregulated and expresses less protein than

the unregulated wild-type promoter support partition (41). It seems likely therefore that

autoregulation is required to maintain a low, but sufficient, level of protein. ParA binds site-

specificaliy to the par operon promoter region (par00 in virro (39). This binding activity is

thought to mediate ParA repressor activity in vivo. The sequence requirements for ParA

DNA binding are not well understood, but the information for ParA binding is probably

included in the 20 bp imperfect, inverted repeat sequence located between the -10

transcriptional signal and the ribosome binding site (Fig. 1-2C). The region of purOP

protected from DNase 1 attack by ParA includes these repeats (Chapter 2; 35, 39) and

mutations in these repeats affect ParA's ability to repress par gene expression in vivo (77).

ParA DNA binding was reported to require ATP (39), but more recently my results

demonstrate that ATP is not required for ParA DNA binding, although ATP greatly

stimulates ParA DNA binding (Chapter 2; 35).

The region protected fiom DNase 1 cleavage by ParA corresponds to the region

protected by RNA polymerase at other promoters (-45 to +20; Ref. 70) and includes the

major transcription start site (Chapter 2; 35, 39, 77). This suggests that ParA represses

transcription by preventing RNA polymerase from binding to the promoter rather than

affecting RNA polymerase a c t i v i ~ after it has bound to the promoter. It is not clear why a

repressor that seems to act by preventing RNA polymerase from binding would utilize ATP.

ATP also affects the repressor activity of TyrR, a negative and positive regulator of the tyr

regulon in E. coli (156, 200). TyrR interaction with tyrosine requires ATP, and TyrR

interaction with tyrosine is in tum required for repression by TyrR (9, 156).

ParA ATPase: ParA ATPase activity was fmt suggested by sequence homology to the

Walker nucleotide binding motifs A and B (Fig . 1-3A) (133, 193). ParA ATPase activity is

quite weak (about 1@ times lower than the RecBCD ATPase, for example Refs. 39 and 100).

It is stimulated by DNA of no specific sequence or topology and by ParB (39). The ATPase

activity does not seem to reflect either a protein kinase activity or a topoisornerase activity

(39). Deletion or mutation of the Walker A motif results in loss of ParA's regulatory and

partition functions (41). Partition is a process that requires energy and one can speculate that

ParA provides this energy via its interaction with ATP. The steps at which the ParA ATPase

may act are not known, however putative steps in the partition process, such as plasmid

movement andfor plasmid pairing, may require ParA's ATPase activity.

Very few site-specific DNA binding proteins are directly affected by ATP. Some of

these proteins, such as the replication initiation protein ORC of Saccharomyces cerevisiae and

T antigen of SV40, are involved in processes that are regulated with respect to the ce11 cycle.

It has been suggested that ATP may regulate the activities of these proteins with respect to

the ce11 cycle (16). Shce partition is likely to be similarly regulated, perhaps ATP also

regulates ParA activity with respect to the ce11 division cycle.

ParA-ParB interachu: ParB stimulates ParA's repressor actvity in vivo (55) and ParA's

ATPase activity in vitro (39) suggesting that the two proteins interact. There is presumably a

direct role for this interaction in partition since ParB is a component of the partition complex

(57, 58) and ParA is also required for partition (41, 55). A direct role for ParA in partition

requires that ParA interact with the partition complex. Further characterization of the

interaction between ParA and ParB is required to understand its potential roles in partition.

15

ParA and ParB homologs: Homologs to ParA and ParB are encoded by other plasmids as

well as by several bacterial chromosomes. The ParA homologs have been grouped in a

superfamily by vimie of their sequence similanties, particularly in the A motif for nucleotide

binding (Fig. 1-3) (99. 133, 193). This superfarnily includes proteins of diverse fiinctions

including ce11 division, nitrogen fixation, membrane transport as well as partition. Some of

these homologs, including some chromosomally encoded homologs, have not been

characterized functionally but are proposed to have roles in partition because (i) they are

more similar to the partition proteins than to proteins of other known functions and (ii) many

of these homologs are encoded by genes upstrearn of a gene for a ParB homolog (99).

Bacillus subtilis encodes ParA and ParB hornologs, Soj and SpoOJ, respectively, that are

involved in sporulation (146). SpoOJ is required for chromosome partition in Bacillus,

however Soj is not (85). E. coli encodes no ParA-ParB homolog pair (the entire E. coli

chromosome has been sequenced) although it dws encode homologs to ParA. For example,

these homologs include MinD, a ce11 division protein that is not thought to be directly

involved in partition (see below) .

The proteins in the ParA superfamily are most sirnilar in motif A, motif B and

another motif between the two, referred to as either "motif 2" or "motif A"' (Fig. 1-3A) (99,

133). One member of this superfamily . NifH, has been crystallized (64). Its structure

suggests that motif 2 may also form part of the nucleotide binding site. Some of the ParA-

like proteins including the putative and known partition proteins, share similarities in another

region called "motif 3" (133). The functional significance of these homologies in ParA is not

completely understood.

Figure 1-3. ParA and ParB contain regions of homology to other proteins. A: The

positions and sequences of the nucleotide binding motifs (motif A and matif B) as well as

morifAD and motif 3 in ParA (398 amino acids) are shown. The one letter code for the amino

acid sequence is used. Underlined residues in motifs A, A' and B are invariant residues in

the PUA superfmily (99). The underlined amino acid in motif 3 is conserved in the ParA-

like partition proteins (133). B: Some of the conserved regions in ParB (333 amino acids) are

indicated. The putative helix-tum-helix motif sequence is sho wn (single letter amino acid

code) (46). The underlined amino acids are conserved in at least five of six proteins

compared (1 10). The positions of the conserved acidic residues, potentially phosphorylation

sites, are indicated by the stars (133).

18

ParB shares short regions of homology with other proteins in the database (Fig . 1 -

3B) (1 10, 133). The conserved regions include a putative helix-nim-helix DNA binding motif

(46) and amino acids that could be substrates for phosphorylation (133). Four acidic residues

at positions 68, 204, 250 and 314 are conserved among ParB homologs. By analogy to

Salmonella fyphimuBum CheY, these amho acids might form an acidic pocket, a potential

site for phosphorylation (133). Although ParB phosphorylation has not been detected in vitro

(39; my unpublished observations), mutation of two of these amino acids, at positions 204

and 250, result in a Par- phenotype in vivo, but do not affect any of the ParB in vitro

activities tested, such as DNA binding and dimerization (110).

Swnmary: P l partition is summarized in schematic form in Fig. 14. The f i t recognized

step in partition is formation of the partition complex by interaction of ParB and MF with

pars. The subsequent steps in partition are not defmed, however they must allow the plasmid

to recognize where it and its cognate plasmid are within the cell. The latter can be achieved

by pairing of the plasmids via their partition complexes. The former c m be achieved by

plasmid interaction with specific sites in the host. 1 have drawn plasmids interacting with the

developing septum, however this is only one possible way to orient plasmids within the cell.

Either or both of these events could require the action of ParA a.nd/or host factors. For

example, ParA interaction with the partition complex may mediate interaction of the partition

complex with host factors. Plasrnid interaction with specific factors in the host could in tum

position plasmids. As part of, or subsequent to, the positioning step, plasmid pairs must be

separated so that each daughter ce11 receives a copy of the plasmid. Al1 of these events must

be CO-ordinated with ce11 division.

Other plasmid partition systems: Partition loci have been identified on many different low-

copy-number plasmids. As with the P l partition system, more is known about the plasmid

Figure 1-4. A model for P l plasmid partition. This schematic depicts a general description

of Pl partition. Briefly, after formation of the partition complex, ParA andor host factors

may associate with the partition complex to assist in positioning of the plasmid and perhaps

plasmid pairing. The model is descnbed in greater detail in the text.

ParB and IHF

ParA and 0 h,

c d t division

1 KEY

ParB

* IHF

0 ParA

1 H host

21

encoded components of the partition apparatus ihan the host encoded components and the

mechanism of partition is not known. Some partition systems encode proteins that share

sequence homology to the P l proteins and some partition systems have analogous

components, however other partition systems exhibit very little resemblance to the Pl

partition system. Many of the partition systems have, or are postulated to have, a cis acting

site that expresses incompatibility. As well, these systems encode a protein(s) that acts at this

site. The genes for these proteins are usually autoregulated. Most of the plasmids encode a

putative ATPase that is required for partition, although the role(s) of ATP binding and

hydrolysis in partition remains unclear. These similarities suggest that the different partition

systems utilize similar partition mechanism, at least at sorne steps .

F and P7pZasmidr: Both the F plasrnid partition system, sop, and the P7 partition system,

par, are functionally analogous to Pl par. The F sop region stabilizes otherwise unstable

plasmids without altering the copy number of the plasmid (149). F sop encodes two tram

acting factors, SopA and SopB, that have limited identity to Pl ParA and ParB respectively,

as well as a cis acting site downstream of the genes for the two proteins, sopC (Fig. 1-5)

(13 1). As with P 1 pars, sopC exerts incompatibility against F (13 1). sopC however is very

different from parS. sopC contains 12 copies of a 43 bp repeat sequence (105, 13 1) to which

SopB binds (73, 132). One copy of the 43 bp repeat is sufficient to support partition (19,

115). The SopB-sopC complex may fonn a high order protein-nucleic acid complex of

defiwd three dimensional structure, analogous to the P l partition complex (20, 114).

Putative host factors for F plasmid partition have k e n identified, however they have not

been well characterized. These host factors include 75 kDa and 33 kDa proteins that were

isolated fiom E. coli extracts by v h e of their ability to bind sopC in the presence of SopB

(73). Like P l ParA, SopA is an ATPase that is stimulated by its cognate B protein

Figure 1-5. The plasmid partition systems of F, P7, RI and RIC2. The genes for proteins

thought to be involved in the partition of the various plasmids are indicated by the arrows.

The purple arrows indicate the (putative) ATPases . Cis-acting sites (where known) are

indicated by the blue boxes. Promoter regions (P) are indicated. The binding site of the site-

specific RK2 ParA recombinase is marked by the black box. The scaie is indicated at the top

of the figure in kilobase pairs. The W partition system is most closely related to the P l

partition system. The maps are derived from Refs. 27, 1 12, 13 1 and 196.

24

(1 95) and SopA binds specificaliy to the sop promoter region (132). The latter interaction

presumably mediates regulation of expression from this promo ter (1 32). SopB s tirnulates

SopA interaction with the promoter region in vitro, suggesting that SopB may also have a

role in autoregdation (132).

The partition system of the P7 plasmid, par, is homologous to Pl par (112) and

their components (ParA, ParB and par9 behave similarly (Fig. 1-5) (39, 77). Although

homologous, the Pl components cannot substitute for the P7 components and vice versa (39,

75). Hybrîd proteins and sites have been used to defme parts of the proteins and of the sites

that determine the species specificity (75, 157). For example, P l and W pars Box A

sequences are interchangeable, but the Box B sequences are not (76). A hybrid Pl ParB

protein that contains P7 sequences in its C-terminus binds to P7 pars, suggesting that the C-

terminus recognizes Box B sequences (157).

The sirnilarities shared by the partition systems of F, Pl, and P7 suggest that these

systems may utilize similar mechanisms. These plasmids are compatible so their partition

systems must have different specificities, perhaps determined at a pairing step.

ZncFIIpl~smidr: R1 and NR1 belong to the Incm incompatibility group (a different group

from Pl) and have identical partition loci. The partition locus of RI parA shares some

superficial similarity to the Pl par locus but has no sequence similarity to Pl par. The R1

parA region stabilizes a Sop' F plasmid without affecting plasmid copy number or the growth

of the host ce11 (65). R1 parA encodes two tram-acting factors, ParR and ParM, the genes

for which are coexpressed from a promoter upstream of parM (Fig. 1-5) (129, 9 1). In

addition, a centromere-like site parc is located upstream of parR and parM rather than

downstream of the par genes as in Pl, F, and P7 (Fig. 1-5) (33, 65). ParR binds to two sets

of five direct repeats in parc located on either side of the parA promoter (33). ParR

25

interaction with pu< has two functions. It represses transcription of the parA operon (91)

and is aiso required for a partition function (27). There are conflicting reports as to whether

ParM stimulates ParR repressor activity in vivo (91, 182). No interaction between ParM and

parc has k e n detected, nor has any effect of ParM on ParR DNA binding been detected

(33, 182). Like P l pars, parc expresses incompatibility , however ody weakly (33). It has

been suggested that this weak uicompatibility results from the preference of ParR and

perhaps ParM to bind parc in cis. ParM is an ATPase (as cited in Ref. 27) and it has been

proposed that the ParM ATPase like P l ParA provides the energy for partition. ParM is

homologous to other ATPases (24), but it does not share any sequence homology with the

ParA superfamily suggesting that partition ATPases may have evolved more than once.

RK2: RK2, also known as RP4, is a promiscuous transmissible medium copy number

plasmid (5-8 copies per host chromosome; 52). Two putative partition regions have been

identified on the plasmid (67, 133, 160). Very Little is understood about the function of these

partition systems. They are fairly cornplex, encoding other functions in addition to theu

partition functions.

The korABF region encodes several genes, korA, incC, korB, korFI, kkoFZZ and @-A

(Fig. 1-5) which have been shown to have regulatory roles in control of expression of several

RK2 genes including their own (14, 89, 90, 185, 186, 201). Roles for at least some of these

genes in partition are based on the foilowing observations. First, the korABF region stabilizes

a heterologous medium-copy-number replicon without changing plasmid copy number or

affecting ce11 growth (133). Second, IncC and KorB share sequence homology to P l ParA

and ParB, respectively (133). Finally, disruption of IncC destabilizes plasmids canying the

rest of the region (133). IncC expresses incompatibility against RK2 (172). however no cis

26

acting site has yet been identifed. It is not clear what contribution, if any, the other genes in

this operon make to plasmid partition.

A second RK2 partition region is encodeci by the parCBA operon (Fig. 1-5) which is

divergently aanscribed from a post-segregational killing system parDE (also involved in

plasmid stability, see below) (42, 47, 160, 161). Both operons are autoregulated (42, 47). A

role for the parCBA region in partition is suggested by its ability to stabilize medium copy

number plasmids without altering the copy number of the plasmids (174). ParA encodes a

site-specific recombinase, that promotes multimer resolution using a site in the promoter

region (67). Multimer resolution systems can stabilize plasmids (see below), however ParA's

resolution function is not sufficient to account for the stabilizing effect of the parCBA region

(67, 159, 174). ParA also regulates transcription from the parC'A promoter (42, 47). ParB

has an endonuclease activity whose function is not hown (as cited in Ref. 174). The role of

parc in RK2 partition is also not known. The parCBA region exerts incompatibility against

RK2, however no cis acting site has been identified so far (169).

Other plasmidi: There are several other putative partition loci that have been identified on

various plasmids (eg. 28, 63, 109, 123, 183). Many of these loci are uncharacterized except

for their sequences and were identified as partition systems by virtue of their homology to

the systems I have described. In al1 cases the mechanism of partition is not known. The

similarities between some of these systems, such as a cis site and an ATPase, suggest

similarities in their mechanisms. Further characterization of these systems as well as

identification of host factors, is required to determine if a general partition mechanism exists.

OTHER FACTORS CONTRIBUTING TO PLASMID STABILITY

Stable maintenance of a plasmid is effected by several different systems. In addition

to partition these systems include mechanisms that maintain the optimal copy number of the

27

plasmid and impair the growth of plasmid-free segregants. With the exception of replication

control, the contributions that these mechanisms make to plasmid stability are not as large as

the contribution of the partition systerns to plasmid stability (3, 13, 23).

Copy number control: Both the average plasmid copy number and the distribution of

plasmid copy number will affect the stability of a plasmid in a cell. For proper segregation,

the ce11 must contain at least two copies of a plasmid at the time of ce11 division. In addition,

many naturally occurring plasmids are quite large (P 1 is about 90 kbp; 84) and their copy

number is kept Iow so as not to drain the host's resources. The distribution of copy number

in a ce11 population also affects plasmid stability. If the distribution is relatively broad, then

there is a greater frequency of plasmid free cells (144). Consequently, plasmids are more

stable when their copy numbers are maintained within a narrow range. The number of copies

of a plasmid is controlled by replication and recombination systems.

Replication control of copy number: Three different mechanisms have k e n shown to

regulate initiation of plasmid replication: regulation by a repressor protein, by antisense RNA

and by iterons. Iterons are multiple repeats of DNA sequences to which a plasmid-encoded

Rep protein binds (reviewed in Ref. 142). Plasmids such as F and P l are maintained at very

low copy number, about one per host chromosome (53, 84). Pl copy number is controlled,

at least in part, by iterons located in a region called inc . . Deletion of this region results in

an 8 to 10 fold increase in Pl copy number (153). RepA binds to the iterons in ind as well

as to sites in oriR, the plasmid origin of replication (1, 29). Binding of RepA to these sites

mediates interaction between incA and oriR, presumably via RepA-RepA interactions (152).

It has k e n suggested that this interaction between onR and incA interferes with initiation of

plasmid replication, perhaps by steric hindrance (2, 134, 142, 152). As the copy number of

the plasmid increases the number of possible interactions that RepA molecules bound at

different sites cm make with each other increases, thereby inhibithg replication.

Multimer resolution: In Rec+ ceiis, plasmids can recombine to form multimers, in effect

lowering the copy number of the plasmid. Many plasmids encode site-specific recombination

systems that catalyze the resolution of plasmid multimers to monomers. This maximizes the

number of independently partitionable molecules. Deletion of plasmid-encoded multimer

resolution systems results in an increase in plasmid multimea and a concomitant decrease in

plasmid stability (13, 180). Pl encodes a multimer resolution system consisting of the loxP

site and the Cre recombinase (176). Pl plasmids without a functioning loxPICre system are

less stably maintained in Rec+ strains than plasmids that do have a functioning recombinase

(13). A bias in the loxPICre reaction towards resolvase function, that is towards formation of

monomers, would be preferred to perform this stabilizing function. Although in vitro snidies

suggest that Cre îünctions equally well to promote and resolve dimers, some observations

suggest that it may function predominantly as a resolvase in vivo (4).

m e r systems: Killer systems contribute to the stability of a plasmid by kiiiing cells that

lose the plasmid. In al1 the systems characterized to date, the plasmid encodes a toxin and an

antidote for the toxin's lethal action (for reviews see Refs. 66 and 92). The antidote is less

stable than the toxin so when the plasmid is lost from a cell, the antidote is degraded before

the toxin and the ce11 is killed. The toxin is usually a protein, but antidotes c m be either a

protein, or an antisense RNA that prevents expression of the toxin (66, 92).

Pl has a kiiler system encoded by the genes pM and doc Qrevents b s t &ath and

death on curing, respectively; 106). Doc is lethal to E. coli in the absence of Phd, however - the cellular target of the Doc toxin is not known (106). Phd is a target for the E. coli ClpXP

protease in vivo. In ClpXP strains, Phd is stable and the killer system does not function

(107). Analogous killer systems have been characterized on F, R1 and RK2 (148, 161, 189).

29

The F killer system, ccd, encodes two proteins, CcdA and CcdB (88, 124, 126, 148). CfdB

is toxic to E. d i and its target is DNA gyrase (18, 125). DNA gyrase is not thought to be

the target of Pl Doc (106).

Curiously, the E. coli chromosome encodes homologs to the RI pem kiiler locus

(1 19) and at least three different homologs to the R1 Hok toxin (25. 65). The pem homologs

ChpAK and ChpBK inhibit cellular growth when overexpressed in the absence of ChpAI or

ChpBI proteins, respectively. The chpA and chpB genes are located near stringent response

genes and it is suggested that the ChpA and ChpB I and K proteins may serve to inhibit

growth under conditions where rapid growth might be hannful to the ce11 (1 19). The roles of

the other homologs are not known.

CHROMOSOME PARTITION

Little is known about the partition mechanism of the bacterial chromosome.

Chromosome partition has been examined to a lirnited extent cytologically, by following the

movement of stained chromosomes. These types of experiments were initially interpreted to

suggest that chromosomes moved rapidly fiom mid-ceIl positions after DNA replication was

complete to positions that are 114 and 314 along the length of the cell (15, 80). In these

experiments chromosome position was measured from the center of the chromosome to the

ce11 poles. When the position of the chromosome was subsequently measured from the edge

of the chromosome to the ce11 poles a more gradua1 movement of the chromosome,

concomitant with ce11 growth and DNA replication was observed (192). The seemingly rapie

movement of the chromosome in the fust set of measurements argued for a positionhg

mechanism similar to mitosis in eukaryotic cells (80). The second set of measurements

suggested that chromosome segregation may occur by a more passive, non-motive m d~..m

30

(192). Regardless, there must be a mechanism to ensure proper distribution of the

chromosome as less than 0.03 % of E. coli ceils are anucleate (79).

Events prior to positioning: Very few of the "partition" mutants that have been identified in

E. coli are thought to directly affect the positioning of chromosomes. Many of the mutants

that have k e n identified affect steps prior to this event, such as replication, multimer

resolution and decatenation. Defects affecting these earlier steps produce a different

phenotype (called Pd-) from defects affecting the positioning step of partition (cailed ParII-).

The ParI' phenotype is characterked by elongated cells with one large centrally located

nucleoid (203). Sometirnes anucleate cells of variable length will be produced. The ParI-

phenotype includes decatenation and recombination defecu; chromosomes are entangled or

dimeric so that the positioning mechanism cannot separate them. The ParII- phenotype is

typified by an increased number of anucleate cells the same length as newbom cells. The

nucleoids are of normal size (203). ParII' mutants are thought to directly affect the

positioning reaction.

Decatemtion of the chromosome: DNA replication causes two topological problems. First,

unwinding of double stranded DNA by the advancing replication fork introduces tighter

twists ahead of the fork which must be removed to allow continued fork movement. Second,

catenanes are formed that interfere with separation of the chromosomes during partition.

There are two type 2 topoisornerases in E. col& D N A gyrase and topoisomerase IV (topo

IV), that can resolve catenanes in vitro and in vivo (5, 101, 1 16, 154). The genes encoding

DNA gyrase, gyrA and gyrB, and the genes encoding topo IV, parc and parE, are essentia'

genes. Decatenation of the products of DNA replication is thought to be carried out

predorninantly by topo IV, since plasrnid catenanes that result from replication accum ..,G in

strains defective in topo IV and these strains have the Pari- phenotype. In additic ., topo IV

31

more efficiently unlinks catenanes in vitro than DNA gyrase unlinks catenanes (190). It has

been suggested that DNA gyrase makes a minor contribution to this activity in vivo, but that

the primary function of DNA gyrase is to introduce negative supercoils ahead of the

replication fork (5, 202). However, some mutations in the DNA gyrase genes result in a

ParI- phenotype and catenated chromosomes. This observation is unexplained since these cells

contain wild-type topo IV (95, 150, 175). Whether one or both enzymes catalyse

decatenation in vivo, this activity is clearly a necessary step before positioning of

chromosomes.

Resoiution of chromosome multimers: The E. coii chromosome possesses a multimer

resolution system that presumably contributes to chromosome segregation by generating

monomers from dimeric daughter chromosomes that arise because of homologous

recombination. This site-specific recombination system. located in the replication terminus

region of the chromosome is not directly involved in the positioning process. Site-specific

recombination at the dif site in the E. coli replication terminus region by the XerCD

recombinase was discovered concurrently by three different labs (21, 31, 102). The 33 bp

site, dif. is located in the replication terminus region of the bacterial chromosome and is

sufficient for either inter- or intramolecular recombination by the XerC and XerD

recombinases (2 1, 3 1, 32, 102, 108, 184). Deletion of dif or xerC results in füamentation

and aberrant nucleoid distribution and size (Pa& phenotype ; 2 1, 102). XerCDldif can be

replaced by Pl CrelloxP without disrupting the ce11 (108). The dif locus is homologous to tk

cer site on ColEl (21) which is also a substrate for XerCD (22, 32). Unlike at cer, the ac:

of XerCD at dif is not biased towards resolution (monomer formation) in vitro. It has bet

suggested that resolution bias occurs in vivo simply because chromosomes are segrega-

away from each other (21, 102).

Chromosome positioning:

nie E. coli muk mu?mzts: The muk mutants were isolated using a screen that selected for

non-lethal mutants that gave rise to a high frequency of anucleate ceils (ParII- phenotype; 79,

140). The m M , mukB, mukC and mukD mutants were identified in this screen. mukC and

mukD have not yet been characterized. MukA is identical to TolC, an outer membrane

protein (79). The effects of tolC mutations on E. coli are pleiotropic, affecting resistance to

detergents, antibiotics, colicins as well as expression of other outer membrane proteins (37).

TolC's role in partition is not well understood. One suggestion is that it helps to foxm

attachrnent of the chromosome to the ce11 membrane via a cotranscriptional complex whereby

concomitant transcription, translation and membrane insertion of an integral membrane

protein gene, such as tolC, form a transient link between the chromosome and the inner

membrane (1 13, 192).

mukB encodes a 177 kDa protein with lirnited homology to rat dynamin D 100 (79).

The predicted secondary structure of M W suggests that it resembles eukaryotic motor

proteins. MukB is predicted to have two globular domains, one at the amino temiinus and

one at the carboxyl terminus, separated by a central rod domain that contains a flexible

"hinge" region (140). MukB binds ATP and GTP as well as DNA of no specific sequence

(139). It was originalïy proposed that MukB propels the chromosomes to their proper

positions in the ce11 dong some as yet undefmed network of fdaments (140). More recently

however it has been suggested that M W may have a role in chromosome condensation

based on the observations that (i) its proposed structurai organization is similar to eukaryotic

proteins that are involved in chromosome condensation (82, 155, 178) and (ii) mukB

mutations are suppressed by genes that affect chromosome condensation when these genes are

expressed Born high-copy-number plasmids (82, 198). Chromosome condensation may be

33

required for partition to occur or may be a motive force in chromosome positionhg (82).

The genes for two other proteins, MukF and M U E . were identified upstream of

muk, and their disruption causes anucleate ce11 formation (199). The functions of these

proteins in partition are not known. MukF, which is similar to histone-like proteins, may

have a role in chromosome condensation since a mukF nul1 mutant strain has irregularly

s haped diffuse nucleiods ( 1 99).

Other mutants: Mutation of sorne chromosomal genes that have functions in other cellular

processes in E. coli give rise to anucleate cells. Mutations Ui r e d for instance, give rise to

anucleate cells without affecting DNA replication (171, 203). However, this phenotype was

shown to be the result of chromosome degradation (170). Certain mutations in min, a ce11

division locus, result in aberrant chromosome segregation (6, 87, 135). The min mutants

were originally identified as mutations that give nse to minicells resulting from aberrant

septum placement at the poles of cells (7, 36). Interestingly, one of the proteins encoded by

this locus, MinD, is a member of the ParA ATPase superfamily (99, 133). However more

recent results suggest that the effect of min mutations on chromosome partition are indirect

via their effects on septum formation (34).

No other ParA homologs, particularly ones encoded by genes upstream of the gene

for a ParB homolog, are encoded by the E. coli chromosome. Other bactena such as Bacillus

subtilis, Pseudornonas putida and Mycobacterium leprae encode ParA and ParB homologs on

their chromosomes (59, 99). Of these homologs, ody the B. subrilis ParB homolog, SpoOJ,

has so far been shown to have a role in chromosome partition (85).

Clearly more information is required to understand chromosome segregation in

bacteria. This will require an understanding of the fûnctions of the genes already identifid as

well as identification of other factors involved in partition.

ATP BINDING AND HYDROLYSIS

As noted earlier, many plasmid partition systems encode an ATPase that is required

for plasmid partition. Sorne of these partition ATPases have been grouped into a superfamily

by vimie of their sequence sirnilarities (99, 133). The Waker A and B motif sequences

found in the ParA superfarnily of ATPases are also found in many but not al1 ATPases and in

GTP-binding proteins (99, 165, 193). The general consensus for motif A is G/AX,GKS/T

and is X,KGGX,KT/S for the ParA superfamily (one letter amino acid code, X is any amino

acid; Fig. 1-3A) (99, 193). The less-well conserved motif B contains an invariant aspartic

acid preceeded by hydrophobie arnino acids (193). Some of the proteins that contain these

motifs have been characterized struchirally, including one member of the ParA superfamily,

the nitrogenase Fe-protein (Nia) from Azotobacter vinlandii (64). These structural studies as

well as mutagenesis of proteins containing these motifs implicate these motifs in interaction

with the phosphate moiety of nucleotides and consequently in nucleotide hydrolysis (eg. Refs.

64, 15 1 and 177). The structure formed by motif A is similar in the different proteins

containing it and foms a loop, called the P-loop, containing the partially conserved glycines

(64, 127, 128, 136, 177). The highly conserved lysine residue in motif A (the second lysine

in the ParA superfamily consensus) is in close proximity to the f l and y phosphates of the

bound nucleotide in many structures (Fig. 1-6) (eg . Refs. 64, 165 and 177). In addition,

mutation of this residue usually affects ATP binding andfor hydrolysis in these

proteins, including Pl ParA (eg . Refs. 41, 45, 165 and 168). The invariant aspartic acid of

motif B is also implicated in interaction with the phosphate moiety either directly or via

coordination of Mg++ (Fig. 1-6) (93, 177, 197). The consemecl serine or threonine of the

motif A may also coordinate Mg++ (Fig. 1-6) (15 1, 177). A sequence motif that interacts

Figure 1-6. ATP binding and hydrolysis. Interaction between the phosphate moieq of the

bound nucleotide and the conserved residues of the A and B nucleotide binding motifs,

denved from the structure of RecA bound to ADP (177) is shown. Only the conserved

residues of the motifs that interact with the phosphate groups are shown. The Thr and Lys

residues are from motif A and the Asp residue is from motif B. The black circle represents a

water molecule. Not drawn to scale.

36

with the sugar and base of ATP has not been identified in proteins containing motif A. There

is some similarity in terms of shape and hydrophobicity in the regions that interact with the

adenylyl group but no sequence similarity, suggesting that many aitemative sequences are

capable of forming the correct binding environment (130).

Hydrolysis of ATP or GTP is thought to occur by an in-line attack of a water

molecule on the y-phosphate. The phosphate is either transferred directiy to the water

molecule or via formation of a phosphoprotein intermediate (eg. Refs. 48, 5 1 and 68).

THESIS RATIONALE

The two roles that ParA plays in partition, repression of par gene expression (55)

and a second as yet undefmed role (41, 55). both require an intact nucleotide-binding motif

in ParA (41). The steps at which ATP binding and hydrolysis function are not well

understood. ParA interaction with parOP, which is affected by ATP (39), Iikely mediates

ParA repressor activity in vivo. 1 have examined the interaction of ParA with adenine

nucleotides and the par promoter region in vitro to address the roles of ATP binding and

hydrolysis in ParA function. ParA interaction with ParB was also examined since (i) ParB

affects ParA repressor activity in vivo (55), (ii) ParB stimulates ATP hydrolysis by ParA (39)

and (iii) ParB interaction with ParA very likely plays a role in the positioning reaction. My

results point to separable roles for ATP binding and ATP hydrolysis in ParA repressor

activity and perhaps in the positioning process.

CHAPTER 2: A role for ATP in ParA site-specific DNA-binding activity

1 performed al1 of the experiments presented in this chapter except for construction of pMD4 and pMD9 which were made by Liane Gagnier.

This chapter is a modified version of the paper:

Davey, M. J., and B. E. Funneil. 1994. The Pl plasmid partition protein ParA: A role for ATP in site-specific DNA binding. J. Biol. Chem. 269:29908-29913 Reproduced with permission fkom the American Society for Biochemistry and Molecular Biology , Inc.

38

INTRODUCTION

The bacteriophage Pl prophage is a plasmid that is maintaineci in Escherichia coli at one

or two copies per host chromosome (162). Consequentiy, stable inheritance of the plasrnid in

a bacterial population is dependent on an active partition system, which ensures that plasmids

are positioned correctly so that each daughter ce11 receives a copy of the plasmid.

P l encodes two proteins, ParA and ParB, and a cis-acting site, parS. that are required

for partition (3). An excess of either protein (3, 56), or in some cases both proteins (77),

interferes with partition. The genes for the Par proteins comprise the par operon and are

transcnbed from a promoter upstream of parA (55). ParB and the E. coli-encoded protein,

integration host factor (IHF), bind to pars to form the partition complex (38, 56-58, 60,

117). The partition complex is thought to mediate cognate plasmid pairing prior to ce11

division (1 1, 149). ParA is not required for formation of the partition complex (38, 56, 58)

and it is not known when, or in fact whether, ParA interacts with the partition cornplex.

ParA regulates par gene expression by repressing transcription from the par

promoter, rneasured using lac2 fusions to the par operon (55). ParB inhibits expression even

further, although ParB alone has no repressor activity (55). The partition systems of other

low-copy-number plasmids such as F, P7, pTAR, RllNRl and RK2 are also autoregulated,

suggesting that this may be a general feature of partition system regulation (42, 47, 63, 65,

77, 91, 132, 182).

Genetic evidence indicates that the regulatory function of ParA is not its only role

(41, 55). When ParA's regulatory function is bypassed by expressing the par genes from an

unregulated promoter, ParA is s t i l l required for plasmid stability. It is assurned that this

requirement reflects a direct role for ParA in plasmid positioning.

Two biochemical activities have been associated with ParA. ParA has a weak

ATPase activity that is stimulateci two-fold by nonspecific DNA and about five-fold by ParB

(39), however the step at which ParA's ATPase activity acts in partition is not known. In

vitro DNase 1 protection assays indicate that ParA binds to the par promoter region in a site-

specific, ATP-dependent manner (39). Over 100 bp of DNA are protected from DNase 1

cleavage by ParA (39) and Little is known about the sequence detenninants of ParA binding.

Protection from DNase 1 cleavage centers on a 20 bp imperfect inverted repeat sequence

located upstrearn of parA (39). Mutations in the inverted repeat sequence that interfere with

ParA's ability to repress expression from the pur promoter have been described (77). No

effect of ParB on ParA DNA-binding activity has yet been detected (39).

The interaction between ParA and the par promoter is somewhat unusual; it is a

site-specific DNA-binding activity that requires ATP. Here . 1 have further examined this

ATP requirement for DNA binding. My results in this chapter lead me to propose that the

effects exerted on DNA binding by ATP, ADP, and ATP analogues are mediated through

ParA oligomerization.

40

EXPERIMENTAL PROCEDURES

Strains and Plasmids: The Escherichia coli KI2 strain DH5 (F e n d l hsdUI7 (rK-mK+)

supE44 thi-I recAl gyrA96 relA1) was used to maintain d l plasmids. The strain BUl(XDE3,

pLysS) ( h a gal (hcIts857 indl Sam7 nin5 lacWS-T7 gene 1)) (Novagen; 179) was used to

express ParA protein from pBEF198.

Restriction digests, ligations , and transformations were performed using standard

methods (164). The plasmid used for ParA protein production, pBEF198 (33, contains the

parA gene under the control of a bacteriophage T7 promoter in the vector pET17b

(Novagen) .

The plasmids pMD6 and pMD9 contain the par promoter region. pMD6 contains the

388 bp Pl HindIII-Hz01 par fiagrnent (Fig. 1-2) in pBlueScript SK+ (Stratagene), and pMD9

contains the 153 bp Rra1 fragment (Fig . 1-2) in the SmaI site of pBlueScript SK + . The

plasmids pMD3 and pMD4 were constnicted to express TrpE-ParA and GST-ParA fusion

proteins, respectively . The TaqI site at the start of the parA coding h e (Fig . 1-2) was

changed to BamHI without destroying the start codon. The BamHI-Bgm fragment encoding

al1 but the f d amino acid of ParA was ligated into the pATH2 (98) BamHI site, creating an

in frame fusion with the 323 amino terminai residues of the E. coli rrpE gene product. The

1202 bp SmaI-XbaI fragment from pMD3 was inserted into the Sm1 site of pGEX-3X

(Pharmacia) to create pMD4.

Reagents and Buffers: Sources for reagents and resins were as follows: S-Sepharose resin

and Mono Q column, Phannacia; ATP, GTP, glutathione-Sepharose, bovine s e m albumin

(BSA), Sigma; isopropyl-0-D-thiogalactopyranoside (IPTG), ADP, adenylyl-

imidodiphosphate (AMP-PNP) and adenine-5'-O-(3-thiotriphosphate) (ATPyS), Boehringer-

Mannheim; ethy lene glycolbis(succinllnidylsuccinate) (EGS) , Pierce; [g2P]ATP, [d2P] dATP

41.

and [d2P]dCTP, NEN-Dupont; lUI-labelled donkey-anti-rabbit-immunoglobulin antibody,

Amersham; polyethyleneimine ceilulose thin layer chromatography (TLC) sheets, Machery-

Nagel; and Bradford concentrated dye, BioRad. Enzymes were purchased from the following

companies: enzymes for cloning, New England Biolabs or Boehringer-Mannheim; bovine

pancreatic DNase 1, Boehringer-Mannheim.

The following buffers were used for ParA purification: Sonication buffer, 50 m M

Tris HCl pH 7.5, 50 mM (m&S04, 1 mM EDTA and 0.5 mM 1,4-dithiothreitol @TT); S

buffer, 25 mM imidazole pH 6.5, 0.1 mM EDTA, 10% (vh) glycerol; Q buffer, 25 mM

HEPES pH 7 3, 0.1 mM EDTA, 10 % (vlv) glycerol. ParA assay buffer consists of 50 mM

Tris-acetate pH 7.5, 100 rnM NaCl and 10 m M MgCI,.

Antibody Production and Purification: Antibodies were raised in rabbits against TrpE-ParA

fusion protein, which was expressed from pMD3 and prepared essentially as previously

described (97). Injection into rabbits and subsequent blood sampling were performed by the

Division of Comparative Medicine at the University of Toronto. Anti-ParA antibodies were

affinity purified from crude sera using the GST-ParA fusion protein immobilized on

nitrocellulose (164). GST-ParA fusion protein was isolated from cmde ce11 lysates using

glutathione-Sepharose affinity chromatography as described (173).

ParA Purification: The strain BL2 1 (XDE3 ,pLysS ,pBEF198) was grown at 37 OC in 2 liters

LB medium (164) containing 40 pg cchlramphenicollml and 100 pg ampicillin/ml to an %,

of approxirnately 0.4. IPTG was added to 0.5 mM, and the culture was incubated at 37OC

for another hour. Cells were chilled on ice, collected by centrifugation, washed and

resuspended in sonication buffer and fiozen in liquid nitrogen. The cells were thawed on ice,

lysed by sonication, and cenmfuged at 25,000xg for 1 h at 0°C. T ' e n 0.35 g of (NH4),S04

per ml of supernatant were added and the mixture was re-centîifuged. AU subsequent steps

42

were d e d out at 4°C. The protein pellet was resuspended in, and then didysed against, S

buffer with 1 rnM DTT. This protein (FrII) was loaded on a 20 ml S-Sepharose column

equilibrated with S buffer containing 25 mM NaCl and 1 mM DTT, and bound proteins were

eluted with a linear 400 ml, 25 mM to 1 M NaCl gradient in S buffer and 1 mM Dm.

Fractions containing ParA were c o d m e d by Western blotting. Peak fractions were pooled

(FrIII) and loaded on a 1 mi Mono Q column equilibrated with Q buffer containing 100 mM

KCl and 1 mM DTT. Bound proteins were eluted with a 20 ml, 100 mM to 1 M KCI

gradient in Q buffer containing 1 mM DTT. The fractions containing ParA were pooled

(FrIV) and used in al1 assays. ParA was stored in Q buffer containing 300 mM KCl and 10

m M DTT at -80°C.

DNase 1 Protection Assays: Substrates were DNA fragments derived from pMD6 and pMD9

(see text). Fragments were labelled on their 3' ends with [d2P]d~TP or -dCTP using DNA

polymerase I large fragment (164). Footprint assays contained (in 25 pl), 10-50 fino1 (see

text) 32P-end-labelled par promoter fragment, 10-2000 ng ParA FrIV, 2 pg sonicated salmon

sperm DNA, 20 pg BSA, and 2 rnM CaCI, in ParA assay buffer. Assays were incubated for

10 min at 30°C and then 1 pl of 1.25 pg bovine pancreatic DNase I/ml was added. M e r

further incubation at 30°C for 60 sec, cleavage was stopped with 75 pl of 1.6 M ammonium

acetate, 200 pg sonicated salmon spem DNA/ml and 0.1 M EDTA. After phenol/chloroform

extraction of proteins, the DNA was precipitated with ethanol and resuspended in 4 pl of

formamide dye (164). Samples were electrophoresed on a 6% polyacrylamide urea gel. The

gel was dried on Whatman DE81 paper and exposed to film or storage phosphor screens for

quantification on a PhosphorIrnager (Molecular Dynamics).

43

RESULTS

Purifkation of ParA: ParA protein was purifiai to over 99% homogeneity, as judged by

Coomassie Blue stained SDS-polyacrylamide gels (Fig. 2-1). Afier S-Sepharose and Mono Q

chromatography (ParA FrIV), I detected both the Pa*-stimulateci ATPase activity and the

DNA-binding activity reported for ParA (39). DNA binding to the par promoter was not

detected in less pure pooled ParA fractions, although binding was detected in the most pure

peak fraction from the S-Sepharose column. ParA's ATPase activity is relatively weak (39)

as compared to other cellular ATPases (e.g., approximately 2xlP-fold lower than the

RecBCD ATPase; lm), so ATPase activity was not assayed in samples less pure than FrIV.

ATPase time courses of ParA (FrIV) with or without ParB confirxned that ParB

stimulates ParA's ATPase activity (Fig. 2-2A) (39). In addition to ATP, ParA also

hydrolysed dATP, albeit at a lower rate (Fig. 2-2B). My experiments showed that ATP

hydrolysis by ParA continues to increase linearly for at least 120 minutes at 30°C with and

without ParB (Fig. 2-2A). Since ParB does not affect ParA's KM for ATP (39), these data

suggest a direct stimulation of the catalytic activity of ParA by ParB rather than an indirect

effect via stabilization of ParA by ParB. In the latter case, one might expect a stimulation of

the extent of hydrolysis (ie., a non-linear increase in product) rather than the rate of ATP

hydrolysis, particularly after a 90 minute incubation.

DNA binding is stimuiated by, but does not require, ATP: Previous snidies (39) indicated

that ParA binds to the par operon promoter in an ATPdependent, sequence-specific manner.

When binding by ParA FrIV was assayed in the absence of ATP at an approximate 1000-fold

excess of ParA monomers to DNA molecules, binding to the promoter fragment was easiiy

detected (Fig. 2-3A). ATP (1 mM) stirnulated binding by about 10 fold; that is, ten-fold less

ParA was necessary to see the same footprint (Figs. 2-3, A and B) . It was formaily possible

Fig. 2-1. hification of ParA. Fractions from a ParA purification were analyzed on a 12%

SDS-polyacrylamide gel (104) stained with Coomassie Blue. Lmes with total cells are: in the

"-" lane (no ParA), 0.15 %, units of BUl(XDE3,pLysS) cells; and in the " +" lane (with

ParA), 0.15 %, units of BL2 1 (XDE3, pLysS, pBEF 198) cells after a one-hour induction. The

protein fraction lanes are 10 pg offrnctiom I (crude lysate), II ((rn4)2S04 precipitate), III

(pooled S-Sepharose eluate) and N (pooled Mono Q eluate) as indicated. ParA was identified

as the major band at approximately 45 kDa (its predicted size from the DNA sequence is 44

kDa; 3), because it reacted with affity-purified anti-ParA antibody (data not shown), and

was absent from cells containing no ParA. The positions of molecular size standards are

indicated on the fefl in kilodaltons @Da).

total protein ceils fractions

Figure 2-2. ATP and dATP hydrolysis by ParA. A: T h e course of ATP hydrolysis by

ParA with and without ParB. ParA (1000 ng) with (A) or without (i) 1400 ng ParB was

incubated in a total volume of 20 pl ParA assay buffer containhg 1 mM [~'*P]ATP (1

gCi/pmol) and 100 ng sonicated salmon spem DNNpl at 30°C. At the indicated times 2 pl

was removed and spotted ont0 a 0.6 x 6 cm polyethyleneimine cellulose TU: strip. The TLC

strips were developed in 0.5M LiCl in 1 M formic acid (8) and exposed to a Phosphor

storage screen overnight. The amount of ATP hydrolysis in each sample was determined

using ImageQuant software (Molecular Dynamics) by cornparing the amount of 32Pi (which

migrates faster than 3 2 ~ - ~ T ~ through the TLC strips) on each sûip to known amounts of "P-

ATP on undeveloped strips. The amount of ATP hydrolysis in the absence of ParA and ParB

( 7 ) as well as with ParB in the absence of ParA ( 0 ) are also shown. Each point is the

average of quadruplicate assays. B: ParA hydrolyses dATP. ATP ( i ) and dATP (*)

hydrolysis by ParA was measured using [y32P]ATP and [d2P]dATP (respectively, 1 mM, 1

pcilpmol) in 20 pl of PUA assay buffer containing 100 ng sonicated salmon sperm DNA/pl

and the indicated amounts of ParA. The assays were incubated for 90 min at 30°C before 2

pl of each assay was spotted on a TLC strip. Thin Iayer chromotography and quantification

of the assays were carried out as in A, except that for the dATP assays, the amount of dATP

hydrolysis was determined from the amount of [ d 2 ~ ] d ~ ~ ~ produced in each assay. Each

point is the average of duplicate assays.

no protein

time (min)

ATP

48

that the ParA fi-action contained a low level of nucleotide that supported DNA binding when

a large volume of ParA was used. A ParA sample, which had been purified over two ion-

exchange columns, was dialysed for 18 h against Q buffer containhg 300 mM KCl and 1

mM D'IT. This highly dialysed sample behaved identically to the original FrIV in the

absence and presence of ATP (data not shown). I concluded therefore, that ParA site-specifc

DNA binding is stimulated by, but not dependent on, ATP.

The protection pattern obtained with ParA in the absence of ATP at high protein

concentrations was identical to that observed at lower ParA levels with ATP. For example,

compare lane 6 (700 ng ParA, no ATP) with lane 8 (70 ng ParA, with ATP) in Fig. 2-3A.

The extended footprint seen at high ParA with ATP (lane 11, Fig. 2-3A), was dso observed

at very high ParA (approximately 3 pg) without ATP (data not shown; summarized in Fig. 2-

4). At high ParA concentration the protected region extended from upstrearn of the -35

region into the parA coding sequence, including the 20-bp inverted repeat. The footprint

covered approxirnately 150 bp (Fig. 2-3). 1 consistently observed a pattern of ParA-

dependent enhanced cleavages and poorly protected residues as compared to most nearby

residues that were fully protected from DNase 1 cleavage (Figs. 2-3 and 2-4).

Effects of ATP hydrolysis on DNA binding: 1 tested the effects of the non-hydrolysable

ATP analogues, ATPyS and AMP-PNP, and of ADP on ParA DNA-binding activity. Al1

tkee nucleotides had a stronger effect than ATP. Fig. 2-SA shows the results of a nucleotide

titration experiment at a level of ParA that gave a partial footprint with ATP. Under these

conditions, the ParA footprint with ADP, ATPyS and AMP-PNP was much more extensive,

and equivalent to that at higher ParA concentrations with ATP. The differences observed

were probably not due to a difference in ParA affinity for the nucleotides, since ParA

required a similar concentration of each nucleotide to bind parOP at this concentration of

Figure 2-3: ATP stimulation of ParA-pdP interaction: A: DNase I protection assays

were performed using 16 h o 1 of the pMD9 BamHI-Rra1 169 bp fragment labelled at BamHI

on the coding strand (upper strand in Fig . 2-4). Where added, ATP was present at 1 mM.

nie amount of ParA added was: lanes 2 and 7 , none; lanes 3 and 8 , 70 ng; ianes 4 and 9,

140 ng; fanes 5 and 10, 280 ng; lanes 6 and 11, 700 ng. Lane 1 is the same fragment treated

with dimethyl sulphate where cleavages at G are stronger than cleavages at A (120).

B: DNase 1 protection assays performed at lower ParA concentrations, with ATP, than in A.

Each assay contained 7 h o 1 of pMD6 HindIII-XhoI fragment labelled at HindIII on the

non-coding strand. The amount of ParA was: fane 1, none; fane 2, 55 ng; lane 3, 110 ng

and fane 4 , 220 ng. In both A and B. the inverted 20 bp repeat (inverted arrows), the -35 and

-10 promoter signals, ribosome binding site (RBS) and parA start codon (ATG) are shown on

the lefr. Sequences upstrearn of the &ai site in pMD9 (in A) have been replaced with vector

sequences resulting in a partial -35 promoter signal.

Figure 2 4 . Summary of the extent and pattern of protection from DNase 1 cleavage by

ParA protein. The results from several DNase 1 protection assays of the par promoter are

summarized on the DNA sequence (GenBankTM accession X02954). The brackets indicate the

lirnits of protection by ParA from DNase 1 cleavage. Positions of enhanced cleavage are

indicated by the solid arrowheads and positions that were poorly protected are indicated by

the open arrowheads. The upper strand is the coding strand for parA and parB. The -35

and -10 regions, ribosome binding site (RBS), and parA start codon (ATG) are indicated by

the boxes. The inverted, imperfect repeat is delineated by the inverted arrows. The major

transcription start site is indicated by the star (77). The patterns depicted here are from

several different experiments using both pMD6 and pMD9 labelled on the coding and non-

coding strands as substrates. No difference in the protection pattern was detected using either

plasmid as a source of substrates. The protection pattern upstream of the &a1 site is from

f o o t p ~ t s of pMD6 oniy; however, the Iength of the protected region was similas on pMD9

where the sequences upstrearn of the Rra1 site were replaced with vector sequences (data not

shown).

GGAGGATGCCAAAGCATGTTGTG CCTCCTACGGTTTCGTACAACAC

JI Rsa I

CATGCAGAGAATGCT GTACGTCTCTTACGA + -1 0

* JI JI A

TTTATGCAGCATTTTTAATTAAATTCAAAAATACAGCATAAAGGA ~ A A G T T T T T A T G T C G T A T T T C C T

4' + +RBS

# CACAAGGTTGCTCAA GTGTTCCAACGAGTT

parA start codon

Figure 2-5. Nucleotide effects on ParA DNA-binding activity: A: ATP, ADP, ATPyS,

AMP-PNP and GTP were added to DNase 1 protection assays at the concentrations indicated

(millirnolar). The fmt lane (G> A) is a dimethyl sulphate cleavage ladder of the fragment.

Each reaction mixture contained 14 fmol of the 388 bp XhoI-Ifindm fragment from pMD6,

labelled on the non-coding s ~ d and 220 ng of ParA, except ime C which is a control assay

that contained neither ParA nor nucleotide. B: ParA binding in the presence of 1 mM ATP

or ADP. The substrate was 30 h o 1 of pMD6 Hindm-XhoI fragment labelled at HindIII end

(the non-coding strand). Each assay contained: Ianes 1 and 8 , no ParA; lanes 2 and 9 , 70 ng;

lnnes 3 and 10, 140 ng; lunes 4 and 1 1 , 280 ng; iane 5, 700 ng; lane 6, 1400 ng and lane 7 ,

2800 ng of ParA. C: ParA interaction with parOP in the presence of either 1 mM ATP or

dATP. Each assay contained 10 mol of the pMD6 388 bp Hindm-XhoI fragment, labelled at

Hindm. The amount of ParA in each assay was: Iane 1 and 5 , none; lunes 2 and 6 , 26 ng;

lanes 3 and 7 , 52 ng; and lanes 4 and 8 , 104 ng. In A. B and C, the location of the -35 and -

10 promoter signals, inverted repeat (inve~ed arrows) ribosome binding site (RBS), and parA

start codon (ATG) are shown on the lep.

A ATP ADP A T m AMP-PNP QW -II--- A

ATG 0

ATG RBS

ATP ADP

ATG =

56

ParA (Fig. 2-5A). In fact, 1 later show that ParA has a lower affinity for ADP thm for ATP

(Chapter 3). ParA protein titrations in the presence of various nucleotides confmed that

ADP (Fig. 2-SB), A m S and AMP-PNP (data not shown) stimulated ParA binding about 5-

fold more than ATP. The ADP used in these assays contained less than 0.5 % ATP, as

determined by thin layer chromatography (as in the legend to Fig. 2-2). AMP, adenosine,

adenine and sodium pyrophosphate (data not shown), or GTP (Fig. 2-4A) had no effect on

binding activity. dATP, which is hydrolysed to a lesser extent than ATP by PUA (Fig. 2-2B)

stimulated ParA DNA-binding activity slightiy better than ATP (Fig . WC). Therefore,

adenine nucleoside di- or tri-phosphates are required to stimulate ParATs DNA-binding

activity. My observation that ADP stimulated DNA binding to the same degree as the non-

hydrolysable analogues, and the previous fiiding that ADP and phosphate are the products of

the ParA ATPase (39), suggest that a bound ADP or ATP is required for stimulation, but

that the act of hydrolysis itself inhibits DNA binding. The observation that the ability of a

cofactor to stimulate ParA DNA binding is inversely related to the rate at which it is

hydrolysed by ParA supports this conclusion.

ATP and ADP affect the sedimentation behaviour of ParA: One possible explmation for

the effect of ATP and ADP on ParA's DNA-binding activity is that they affect the

conformation or the oligomerUation of the protein. The extensive protection from DNase 1

cleavage of the par promoter region suggested that more tban one ParA molecule was bound

to the site. 1 used glycerol gradient sedimentation to examine whether ATP or ATP

analogues affected the hydrodynamic properties of ParA. ADP was used as the cofactor in

most experiments, since ADP had a stronger effect than ATP on the DNA-binding activity.

Without ADP, ParA sedimented at a rate consistent with a molecular size slightly larger

than predicted for a monomer of ParA (44 kDa; Fig. 2-6A). The presence of ADP (1 mM)

Figure 2-6. ADP and ATP affect ParA sedimentation rate. Purified ParA (FrIV) was

analyzed by glycerol gradient sedimentation. Each panel depicts the results from two parallel

gradients, one with no ADP (O) and the other with 1 m M ADP included throughout the

gradient (e). Glycerol gradients (3.9 ml of 10 to 40% (v/v) glycerol in ParA assay bufier)

were poured in seven steps, allowed to diffise for 1 h at 21°C and then equilibrated to 4 T

for 1 h. A mixture of the indicated arnount of ParA and 300 pg BSA in 78 pl was layered

ont0 the top of each gradient. The gradients were centrifuged at 55,000 rpm for 18 h at 5°C

in a TST60.4 swinging bucket rotor (Dupont). Two drop fractions (80-100 pl) were collected

from the bottom of the tube using a syringe needle. Fractions were analyzed on Western

blots using anti-ParA antibody and '=I-labelled donkey-anti-rabbit IgG antibody as the

primary and secondary antibodies, respectively. The blots were exposed to a storage

phosphor screen and bands were quantified on a Phosphorhager (Molecular Dynamics).

BSA, which was included with ParA in the gradients, was detected using Bradford assays

(A,,, dashed fine). At each concentration of ParA, gradient profdes with and without ADP

were aligned by their BSA peaks (relative fraction number=O) to correct for one or two

fraction differences that resulted from variations in collection needle insertion. The positions

of other size standards, run in parallel gradients, are also indicated. The standards were

bovine IgG (158 kDa), chicken ovalbumin (ovd; 44 kDa) and equine myoglobin (myo, 17

kD a).

ov I myo tw

BOTOM Relative fraction number TOP

Figure 2-6 cont. B: The sedimentation of ParA in the presence of 1 mM ATP (m) or 1 mM

ADP (a). The gradients were perfomed as described in A using 8.6 pg of ParA for each

gradient. BSA (A5,, dashed line) was included in each of the gradients and its sedimentation

was used to align the gradients. C: The sedimentation of ParA through gradients with (+ )

and without (O) 1 mM GTP are shown. The experiments were performed as described in A

using 8.6 pg PUA. The scale on the lef: is in arbitrary units (fiom the PhosphorIrnager

values) since the protein was not quantified to ParA standards in this experiment.

Relative ParA concentration 1 mM GTP

0-0 none

ParA concent ration (nglpl) H 1 mM ADP H 1 mM ATP A tu cd O O O

P , O O

61

increased ParA's sedimentation rate to a position intermediate to monomer and dimer shes at

low concentration, and to dimer size at high ParA concentration. These sizes were estimated

by comparing ParA sedimentation to BSA (66 kDa) in the same gradient, or to other size

standards nin in parallel gradients. The change in sedimentation rate caused by ADP was

sensitive to the concentration of PUA such tbat the sedimentation rate increased with

increasing ParA (Fig. 2-6A). ATP had the same effect as ADP in these experirnents, whereas

GTP had no effect on sedimentation rate (Fig. 26B and C).

These observations indicate that ADP and ATP affect ParA structure. Since

sedimentation is sensitive to size and shape, ATP and ADP could be influencing one or both

properties of ParA. However, the concentration dependence of the sedimentation rate

suggested a size effect; that ADP was influencing a monomer-dimer transition. Proteins that

exist in a rapid monomer-dimer (or oligomer) equilibnum can sediment as a single peak

whose size is closer to the predominant species. This hypothesis makes a couple of

predictions: First, ParA must be able to form dimers (or oligomers), and second, that this

transition is reversible. ParA, collected from a gradient containing ADP, sedirnented at the

same rate as a previously untreated sample of ParA when run through separate gradients

without ADP (data not shown), suggesting a reversible effect of ADP on ParA sedimentation.

ParA forms dimers: 1 used chemical cross-linking to analyze the oligomeric state of ParA.

ParA was treated with EGS, a homobifunctional chemical cross-linker, in the presence (1

mM) or absence of ADP. EGS reacts with available primary amine groups, mainly at

lysines. Higher molecular weight forms of ParA were observed after addition of EGS both in

the presence and absence of ADP, consistent with the formation of dimers (Fig. 2-7). A very

small amount of tetramer may also be formed, although the majority of the cross-linked

products were dimer-sized. The doublet bands in the samples treated with cross-linker

Figure 2-7. EGS crosslinking of ParA: ParA (80 nglpl) in ParA assay buffer, with or

without 1 mM ADP (as indicated), was treated with EGS (15 pg/ml). At the indicated times,

500 pl were removed and quenched in 150 mM lysine. Samples were precipitated in 15%

(w/v) trichloroacetic acid, washed w ith acetone, resuspended in Laemmli sarnple buffer and

analyzed by electmphoresis on an 8 % SDS-polyacrylamide gel (104) which was stained with

Coomassie Blue. In this experiment, there was a slight difference in recovery at each time

point. The no EGS lane contains 10 pg ParA FrIV that has not been treated with EGS or

precipitated with trichloroacetic acid, diluted in Laemmli buffer. Migration of size standards

is indicated on the nghr in kDa. Identical patterns were observed when cross-linking was

performed at lower ParA concentration (40 pglml).

na ADP 1 mM ADP na I

EGS O 1530 O 15 30 min

64

indicate intramolecular cross-links. Both the monomer- and dimer-sized populations consisted

of two bands suggesting that these intramolecular cross-links did not interfere with

dimerization. These experirnents were performed at low concentrations of ParA to ensure

that monomers did not cross-link, and these conditions may not have k e n sensitive enough

to detect a kiwtic effect of ADP on dimer formation. Alternatively. the cross-linking

reaction might push the equilibriurn towards dimer formation by irreversibly forming dimers.

Nevertheless, this experiment demonstrates that ParA forms dimers in solution.

DISCUSSION

P l ParA protein binds specifically to a site in the par operon promoter region (39). 1

have found that its DNA-binding activity was dramatically infiuenced by several adenine

nucleotides. ATP and dATP, which are both hydrolysed by ParA, stimulated ParA's DNA-

binding activity approxirnately 10-fold. ADP, one of the products of ATP hydrolysis, and the

non-hydroly sable analogues, ATPy S and AMP-Pm , stimulated ParA binding about 50-fold.

ParA bound to its specific site on DNA in the absence of ATP suggesting that ATP

stabilises a less stable ParA conformation that represents the more active DNA binding fom.

Non-hydrolysable ATP analogues, or the product of ATP hydrolysis (ADP), showed a much

greater stimulation of ParA DNA-binding activity than ATP alone. These observations

suggest that bound nucleotide (ATP or ADP) is required for stimulation of ParA's DNA-

binding activity, and that the act of ATP hydrolysis by ParA inhibits DNA binding, perhaps

by causing dissociation of ParA-DNA complexes. Initial attempts to measure the off-rate of

ParA from DNA indicated that ParA dissociation was too rapid to measure in my assays

( < 30 sec. ; unpublished observations). Alternatively , the act of ATP hydrolysis by ParA

could be inactivating ParA before it binds to DNA, in effect reducing the concentration of

active protein. ATP hydrolysis has been demonstrated to cause dissociation of the E. coli

RecA protein from DNA (121, 122). RecA is a DNA-dependent ATPase (147, 158) that

binds DNA non-specifically (122). RecA's DNA-binding activity is also stimulated by ATP.

However, unlike ParA's DNA-binding activity, the product of hydrolysis, ADP, promotes

RecA-DNA complex dissociation (121). 1 cannot determine from my results whether ATP

hydrolysis infiuences the binding or the release of DNA by ParA.

ATP and ADP also had ciramatic effects on the hydrodynamic properties of ParA.

Both nucleotides increased the sedimentation rate of ParA in glycerol gradients, suggesting

66

either oligomerization or a change to a more spherical, less asymmetric, shape. ParA forrned

at l e s t dimers as indicated by chernical cross-linking experiments with both EGS (Fig. 2-7)

and dithiobis(succinimidylpropionate) @SP; data not shown). A rapid monomer-dimer

equilibrium will yield an average sedimentation rate between true monomer and dimer;

shifting the equilibrium towards dimer formation will increase the average sedimentation

rate. 1 believe that ATP and ADP are promoting dimer formation, or siabilking existing

dimers, such that the equilibrium is shifted towards dimer formation. I favor this idea

because the increase in sedimentation rate caused by nucleotide was also influenced by

protein concentration. More complicated interpretations of the sedimentation data are of

course possible, perhaps involving a combination of shape and size changes. However, und

firther physical analyses of ParA structure are performed (see Chapter 3), 1 think that the

simplest interpretation is that ADP and ATP affect the oligomerization state of ParA.

I predict therefore, that the most active DNA-binding form of ParA is an oiigomer,

and most likely a dimer, and that adenine nudeoside di- or tri-phosphates stimulate DNA

binding by promoting or stabilizing dimer formation. Both ADP and ATP increased ParA

sedimentation rate and both stimulated DNA-binding activity (Figs. 2-3, 2-5 and 2-6). GTP

neither affected sedirnentation behaviour nor stimulated ParA's DNA-binding activity (Figs.

2-6C and 2-5A). Finally, although the sequence requirements for ParA recognition have not

been completely defmed, binding probably requires the inverted repeat sequence (77), which

is often diagnostic of dimeric DNA binding proteins.

ParA bound to the par promoter, although weakly, in the absence of nucleotide,

under conditions where 1 detected no significant change in ParA sedimentation behaviour.

Since ParA cm dimerize Ui the absence of nucleotide (Fig. 2-7), a low concentration of

67

dimer may be present but undetectable by sedimentation. Alternatively, DNA may stabilize

dimer formation, or ParA monomers may bind parOP weakly.

The same increase in sedimentation rate was observeci when ATP was added to the

gradients as when ADP was added (Fig. 26%). ATP was hydrolysed under these conditions

(18 hr at 4°C; data not shown). Why do ATP and ADP have sunila. effects on ParA

dimerkation, when these nucleotides affected DNA binding differentiy? Several explanations

are possible. These gradients may not be sensitive enough to detect small differences in

dimer formation caused by ATP and ADP. Alternatively, the dBerence between the ATP

and ADP foms of ParA may be manifested ody in the presence of DNA. For example,

ATP hydrolysis could cause dimer dissociation only in the presence of DNA. Or f d l y , the

act of ATP hydrolysis by ParA may be coupled to some other change in ParA, inhibithg

DNA binding. In the next Chapter (Chapter 3) 1 will extend my analyses of these complex

roles of ATP binding and hydrolysis on ParA activity.

About 150 bp of DNA were protected from DNase 1 cleavage by ParA (Fig. 2-3).

Although the more active ParA DNA binding form may be a dimer, the footprint length

argues that several dimers bind to this region. The extent and pattern of cleavage could be

explained by ParA binding in a linear array, or by DNA wrapping around ParA. The pattern

of poorly protected residues partially resembled the protection pattern observed in wrapped

complexes, such as nucleosomes (96, 141). However, the poorly protected residues in the

ParA footprint did not strictly confonn to a helical periodicity observed in footprints of

nucleosomes (10 bp, staggered by 5 bp on opposite strands).

There are very few proteins whose sequence-specific DNA bimiing activities are

directiy affected by ATP. Two such factors, aside from ParA, are the origin recognition

complex (ORC) isolated from Saccharomyces cerm'siae (16) and the SV40 and polyoma

virus large himour antigens (T antigens) (43, 44, 1 1 1). These proteins are involved in

initiation of DNA replication as well as other processes. SV40 T antigen is the better

understood protein biochemically and ATP has similar effects on T antigen and ParA. Aside

from stimulating T antigen DNA-binding activity, ATP also stimulates T antigen

oligomerization into structures as large as hexamers (69, 163). The involvement of ATP in

the activities of T antigen and ORC, as well as other replication initiators, led to the proposa1

that ATP may be involved in regulating the activities of these proteins with respect to the ce11

cycle, although not necessarily their DNA binding activities (16). If ATP-mediated regulation

or coordination of Pl partition with respect to E. coli ce11 division does exist, it could be

through the control of ParA's activities.

1 have focused on the effects of ATP, ADP, and ATP analogues on ParA,

specifically on the ParA DNA-binding activity . Presumably , this binding activity mediates

the ParA repression of transcription from the par promoter. How do these activities relate to

ParA activities during partition, and to ParA interaction with ParB? ParB stimulates ParA

ATPase in virro (39) yet increases repression at the par promoter by ParA in vivo (55) . In

agreement with previous studies (39), Pa* had no effect on ParA DNA-binding activity in

my assays (under the conditions assayed, see Chapter 3). it may be that the effect of ParB in

vivo is mediated through a third as yet unidentified protein. Perhaps stimulation of the ParA

ATPase by ParB reflects a role for the ATPase and PUA-ParB interaction in the positioning

process. The formation of the partition complex at the partition site, pars, requires ParB and

M F but not ParA (38, 56, 58). 1 infer that partition complex assembly is an early event in

partition, with ParA acting later. For example, ParA could be acting to position plasmids via

an interaction with this ParB/IHFlparS complex. An amactive possibility is that ParA

dimerkition mediates plasmid pairing and that ATP hydrolysis causes dissociation of plasmid

69

pairs at ce11 division. Association of ParA and ParB in this context may stimulate the ATPase

activity which may serve to regulate assembly, disassembly or positioning of these

complexes. Clearly , a better understanding of ParA/ATP and ParMarB interactions will

help in d e f d g the role of ParA in P l plasmid partition.

CHAPTER 3: Modulation of the P l Plasmid Partition Protein ParA by ATP, ADP and Pl ParB

1 performed al1 of the experiments in this chapter.

A version of this chapter has been submitted for publication to the Journal of Biological Chemistry .

mODUCTION

ParA is an essential component of the Pl plasmid partition system (3). The prophage

of bacteriophage P l exists as an autonomously replicating plasrnid and its copy number is

about the same as that of the bacterial chromosome (84). The P l partition system, par,

directs proper segregation and thus stable maintenance of the plasmid. The mechanisrn of

partition is unknown, but cm be thought of as a positioning process that via interaction with

the E. coli host ensures proper distribution of the plasmid. par encodes two essential

tram-acting proteins, ParA and ParB, and contains a DNA site called p a s that is required in

cis (3). In addition, the E. coli integration host factor, MF, also participates, dthough it is

not absolutely required (58). ParB and MF bind to pars to form the partition complex (38,

57, 58). Assembly of this complex is assumed to be an early step in the partition pathway.

Formation of this complex does not require the action of ParA (38, 57, 58), and I infer that

ParA acts during a later step in the partition process.

ParA has at least two roles in partition. First, ParA represses transcription of its

own gene and parB from a prornoter upstrearn of parA (55). ParA repressor activity is

stimulated by ParB, however ParB has no efiect on par gene expression on its own. A

second role for ParA in partition is inferred from genetic data that show a requirement for

ParA in partition even when its regulatory role is bypassed (see Chapter 1) (41, 55).

Therefore, 1 consider ParA's repressor activity its "regulatory" function, and this second, as

yet undefmed role as ParA's "partitiont' function because 1 think it reflects a direct role of

ParA in the positioning reaction. The latter function requires that ParA interact, directly or

indirectly, with the ParB-MF-pars partition complex.

ParA is an ATPase and a site-specific DNA binding protein (39). The ATPase is

stimulated by Par!3 and non-specific DNA (39). ParA is one of the better characterized

proteins in a s u p e r f d y of ATPases defined by a modified Wallcer type A motif in the

protein sequence (99, 133, 193). This s u p e r f ' y includes other plasmid partition proteins

such as F SopA as well as plasmid and chromosomally-encoded proteins from various

bacteria species whose functions have not yet k e n determined (59, 99, 133). Many of the

plasmids and bacterial chromosomes also encode a ParB homolog adjacent to the ParA

homolog. The similarities of these ParA-like proteins with P l ParA and F SopA have led to

the suggestion that these homologs are also involved in plasmid or chromosome segregation

(99, 133).

ParA binds to a specific DNA site in the par promoter region (39), and this binding

is thought to mediate ParA repressor activity in vivo. The recognition site for ParA is likely

a large inverted repeat in the par promoter region (caiied parof), since (i) several mutations

in these repeats reduce or elirninate ParA repression in vivo (77) and (ii) in DNase 1

foo tp~ t ing experiments, protection centers over these inverted repeats, especially at low

ParA concentrations (39; Chapter 2). ParA DNA-binding activiw is strongly stimulated by

ATP (39; Chapter 2). ATP hydrolysis is not required for this stimulation; in fact, it appears

to be inhibitory since ATP./S, AMPPNP and ADP stimulate ParA DNA binding about 5-fold

better than ATP (Chapter 2). In addition, ATP and ADP promote ParA dimerization

(Chapter 2), suggesting that ParA prefers to bind DNA as a dimer.

1 am interested in how ParA performs both its regulatory and partition roles.

Mutation of the ParA nucleotide binding motif interferes with both ParA repressor and

partition functions (41). However, the steps at which ATP binding and hydrolysis modulate

ParA function are not understood. To address these questions, I have further probed ParA

interaction with ATP, ADP and ParB, by examining their effects on ParA structure and

function. 1 show that ATP binding and hydrolysis alter ParA conformation and that these

conformational changes can be correlatecl to changes in ParA activity. In addition, 1

demonstrate that ParB stimulates ParA site-specifc DNA-binding activity , consistent w ith

ParB's role as a corepressor. ParB modulation of ParA DNA-binding activity requires ATP

hydrolysis. 1 conclude that ATP binding and hydrolysis have different roles in repression and

perhaps partition, presumably mediated by their effects on the conformation of ParA.

74

EXPERIMENTAL PROCEDURES

Reagents and Buffers: Sources for reagents were as foiiows: Mono Q and Mono S columns,

GTP, Sephadex G-25 M, Pharmacia; bovine senun aibumin (FrV), ATP and ADP, Sigma;

ATPyS , bovine pancreatic DNase 1, Boehringer Mannheim; heparin agarose, Bio-Gel Pd,

dithiothreitol (DTT) and Bradford protein assay dye, BioRad; and [d2P] ~ATP, [$'Pl ATP

and [2,8 'HIADP, NEN. Restrictions enzymes and E. coli DNA polymerase I large fragment

were purchased from New England Biolabs.

Buffer A contains 10 mM Tris-acetate, pH 7.5, 30 mM NaCl and 5 mM MgCl,.

Buffer B is 50 mM HEPES pH 7.5, 0.1 mM EDTA and 20% glycerol. Dilution buffer is 50

rnM Tris-HC1, pH 7.5, 100 rnM NaCl and 10% glycerol.

Proteins: ParA was prepared as described (Chapter 2) except that the glycerol concentration

of FrIV was adjusted to 50% (vlv) and the protein was stored at -20°C. 1 consistently

maintain ParA in buffen containing 10 mM DTT (Chapter 2) because 1 have observed that at

lower DTî concentrations ParA oxidizes and forms disulphide cross-linked multimers,

particularly after long-term storage (data not shown).

ParB (FrTV) was prepared as descnbed (57, 58) except that ParB FrIII was

concentrated using a 1 ml Mono S column prior to gel filtration. ParB FrIII was diluted to 50

mM KCl in buffer B containing 2 m M DTT and then loaded on a Mono S column

equilibrated in the same buffer . ParB was eluted in one step with 600 rnM KCI in buffer B

containing 2 mM DTT. Fractions containing P a s were then purified by gel filtration (57).

ParB (FrIV) occasionally contained a contaminiiting ATPase activity that could

potentially interfere with the ParA assays. This ATPase activity was removed by

chromatography through a heparin agarose column. ParB (FrlV) in buffer B containing 50

mM KCI and 2 m M DTT was applied to a 5 ml column equilibrated in the same buffer and

75

eluted with a 100 ml, 50 mM-1 M KCl gradient in buffer B. ParB was stored at -80°C.

For DNA-binding assays, ParB W o r ParA were equiiibrated in dilution buffer by

gel fikation through either 1 ml Sephadex G 25 M or Bio-Gel P-6 columns.

Nucleotide binding: Hummel1 and Dreyer equilibnum gel fdtration experiments were used to

measure ATP and ADP binding by ParA (83, 137). Binding was determined as the amount

of ATP andfor ADP that CO-eluted with ParA through a gel filtration column. The column

was equilibrated in the same buffer and "P-ATP ancilor 3 H - ~ D P concentration as the ParA

sample. A 5 ml Bio-Gel P-6 column was equilibrated in bufier A containhg [y'2P]~TP

(10-20 pCiIpmol) andfor [2 ,83Hl~DP (10 pCiIpmo1). Two hundred microliters of ParA in

the same buffer with the same concentration of nucleotide was applied to the column. Four

drop fractions were collected (150-180 pl) and the radioactivity in 100 gl of each fraction

was measured by liquid scintillation counting. The amount of ParA in the peak fractions was

determined by Bradford assay, using a ParA standard whose concentration was determined

by arnino acid analysis (next section). The amount of ATP bound by ParA was detemined as

the arnount of ATP in the peak minus the background ATP (in the buffer). Experirnents were

perfomed with bufien containing 0.5, 2, 5, 10, 15, 20, 30, or 50 pM ATP, using either 31

pM or 62 p M ParA in the sample loaded on the column. At ieast two experiments were

performed at each concentration of ATP. Concentrations used in the ADP binding assays are

in Table 3-1.

Protein concentration determination: Protein concentration was routinely measured by

Bradford assay (26). The accuracy of this assay for ParA was determined by comparing the

measurement of protein concentration by amino acid analysis to the measurement by

Bradford assay. For amino acid analysis, ParA was equilibrated to 20 mM NH,HCO, by gel

filtration through a 1 ml Bio-Gel P-6 column. ParA hydrolysis and amino acid analysis were

performed at the HSC Biotech Service Centre (Universify of Toronto). ParA molar

concentration is expressed as a monomer and ParB as a dimer.

DNA binding assay: ParA DNA binding to parOP was measured using DNase 1 protection

assays (Chapter 2). The substrate was the 388 bp Hindm.-Bo1 fragment from pMD6

(Chapter 2) and was labelled at H i n m on the 3' end with [ d 2 P ] d ~ T P using DNA

polymerase 1 large fragment (164). A typical binding assay contained between 0.1- 10 pmol

ParA (monomer), 10 fmol pMD6 HindIII-XhoI fragment and 1 pg sonicated salmon spem

DNA in 15 pl buffer A. Nucleotide, when added. was present at 1 mM. The mixtures were

incubated for 10 minutes at 30°C, 1.25 pg of bovine pancreatic DNase 1 was added and the

mixtures were incubated for a further 60 sec at 30°C. DNase 1 digestion was stopped by the

addition of 5 pl of "stop" solution (98% forniamide, 10 m M Na2EDTA pH 8, 0.025% xylene

cyanol, 0.025 % bromphenol blue; 164). The samples were heated to 90 OC for 3 minutes and

loaded on a 6 % acry lamide urea gel. After electrophoresis , the gel was dried on Whatman

DE81 paper and exposed to a PhosphorIrnager screen (Molecular Dynamics) and/or to Kodak

XRP film.

Quantification of DNase 1 protection assays: DNase I protection assays were quantified

using a Phosphorhager and ImageQuant software (Molecular Dynamics). 1 defined binding

as protection fiom DNase 1 digestion of the inverter? repeat sequences in parOP by ParA. To

quantify this binding in each iane 1 drew a box in the area that corresponded to the inverted

repeats and another box around a set of bands outside of the protected region but in the same

lane. The volume ("counts") of the box correspondhg to the inverted repeats was divided by

the volume of the box outside of the footprint to yield the relative DNase 1 sensitivity of the

inverted repeats. The sensitivity values obtained were normalized so that the assays that did

n not contain ParA had a DNase 1 sensitivity of 1. As the sensitivity to DNase 1 decreases, the

protection or binding by ParA increases.

RESULTS

ATP and ADP bhding by ParA: Since ATP affects ParA activity I decided to directiy

measure ATP binding by ParA. I used equilibrium gel fdtration (83, 137) which measures

ATP binding as the amount of ATP that co-elutes with protein through a gel filnation

column. Since this is an equilibrium technique, one can deduce ATP stoichiometry as well as

binding affinity. ParA was loaded on a Bio-Gel P-6 column that had been equilibrated in the

same 32P-~TP concentration as the ParA sample. ATP that bound to ParA CO-eluted with the

ParA peak and there was a corresponding trough of ATP concentration after ParA eluted

(Fig 3-1A) (83). ParA concentration was detexmined by amino acid analysis (see

Experimental Procedures). A typical gel filtration experiment is shown in Fig. 3-1A. The

experiment was repeated multiple times, varying ATP concentration from 0.5 to 50 pM and

using two different concentrations of ParA (31 or 62 PM). The ratio of ATP bound per

monomer of ParA was calculated for each experiment and each expriment represents one

point on a binding curve (Fig. 3-1B) or a Scatchard transformation of the data (Fig. 3-1C). 1

was unable to measure ATP binding by ParA to saturation (i.e. at higher nucleotide and

protein concentrations) because of limitations of ParA solubility at high concentration.

Regression analysis on the Scatchard transformation of the data resulted in a & of 33 p M

ATP and 0.8 ATP binding sites (n) per ParA monomer. This K, is close to the K, for ParA

ATPase (50 PM; 39).

Mixing experiments indicate that ParA has a slightiy lower affinity for ADP than for

ATP. When [YZP]ATP and [2,8 3H]ADP were present at equai concentrations in an

equilibrium gel fdtration experiment , ParA bound approximately 1.5 times less ADP than

ATP (Table 3-l), suggesting that the ParA affinity for ADP is about 1.5 times weaker than

ParA affinty for ATP. When the ADP concentration was 1.5 times that of ATP,

Figure 3-1: Equilibrium gel ritration analysis of ATP binding by ParA: A: The profde

from a typical equilibrium gel filtration experiment is shown. In this experirnent a 200 pl

sampie containing 50 pM ATP and 62 pM ParA was loaded on a 5 ml Bio-Gel P-6 column

and eluted as descnbed in "Experimental Procedures". The ATP concentration in each

fraction (a) and the ParA concentration in peak fractions (O) is plotted against the fraction

number. Note that the fractions containing proteîn have a slightly larger volume than the

fractions without protein, thus the peak appears srnaller than the trough. B: Binding curve of

the results from many equilibrium gel filtration experiments performed at 31 pM (*) and 62

p M (i) ParA (in the sarnple load) at various ATP concentrations in the column buffers (see

"Experimental Procedures") is shown. AU data points (ie both data sets) were used to derive

the curve. C: The data from B are plotted on a Scatchard curve (166).

O 10 20 30 40 50

Fraction Number

pmol ATPIpmol ParAJpM ATP pmol ATPIpmol ParA

Table 3-1: Relative binding of ATP and ADP by ParA as measured using

equilibrium gel filtrationa.

Nucleotide concentrationb

(PM)

nucleotide bound per ParA monomer

(prnoYpm01)

ATP ADP ATP:ParA ADP:ParA

a ParA concentration was 60 pM

These assays measured binding of '*P-ATP and 'H-ADP (alone or together) to ParA

as described in " Experimentai Procedures " .

83

approxirnately equal amounts of each nucleotide were bound by ParA (Table 3-1). In the

presence of ADP, less ATP was bound by ParA than when the same concentration of ATP

alone was used, suggesting that ATP and ADP are binding to the same site. 1 estirnated a K,

of 50 p M for ADP, approximately 1.5 times weaker than the K, for ATP (33 PM). The

different affinties of ParA for ATP and ADP indicate that ParA can distinguish between

adenine nucleoside diphosphates and triphosphates.

Nucleotide effects on ParA stabüity: Adenine nucleoside di- and tri-phosphates strongly

stimulate ParA DNA bindkg (Chapter 2; 39), however ADP and ATPyS are better cofacton

than ATP (Chapter 2). To test whether the ability of adenine nucleotides to modulate ParA

activity was a consequence of their effects on ParA conformation 1 examined ParA stability

with and without nucleotide. 1 found that although ParA was quite stable at O°C (for at l e s t

4 days), it was rapidly inactivated by short heat treabnents at 52 OC (Fig . 3-2) and 1 used this

treatment to detect nucleotide-dependent changes in ParA conformation.

ParA stability at 52OC was measured in the presence of various nucleotides (Fig. 3-

2). When ParA was heated to 52°C for 3 minutes in buffer containhg 1 mM ATP or ATPyS

the protein retained significantly more DNA-binding activity than when heated in buffer with

1 m M GTP (Fig. 3-2A and B). ADP (1 rnM) also stabilized ParA, although not to the same

extent as did ATP and ATPyS (Fig. 3-2A and B). After this heat treatment, ParA with GTP

was essentially inactivated (Fig. 3-2A and B). After a shorter heat treatment (1 min), ParA

with GTP retained the same activity as ParA that was heated without any nucleotide (Fig. 3-

2C), consistent with cther assays that indicate that ParA does not bind GTP (Chapter 2; 39).

Therefore ATP, ADP and ATPyS stabilize ParA, suggesting that ParA interaction with these

nucleotides alters ParA conformation. However ADP is less effective than ATP or ATPyS.

Figure 3-2: ATP, ATPyS and ADP stabilize ParA. A: Activity of heat-treated ParA

assessed by DNase 1 footprinting at the parOP site. ParA was heated for 3 min at 52°C in

buffer A containing 100 mM NaCl, 10% (vlv) glycerol, 100 ng bovine serum albumidpl and

1 mM of the indicated nucleotide (ATP, ATPyS, ADP or GTP; above each set of lanes).

Following the heat treatment, ParA was diluted and titrated into DNA binding assays

containing 1 mM ATP at 30°C. as described in "Experimental Procedures". Each lane is

labelled with the amount of PUA in pmol added to the DNA binding mixture. Note that this

ParA scale varies; ParA concentration was lower in assays containing ATP or ATPyS than

ones containing ADP or GTP. The labels and boxes on the lep indicate the -10 and -35

transcriptional signals, the ribosome binding site (RBS) and the parA start codon (ATG) . The

irnperfect inverted repeat sequence is delineated by the inverfed arrows. B: ParA DNA-

binding activity was quantified from the data in A as described in "Experimental

Procedures". The relative sensitivity to DNase 1 cleavage of the inverted repeat sequence in

each lane was plotted against the amount of ParA (pmol), heated in buffer containing 1 rnM

ATP (a), 1 rnM ATPyS (O), 1 m M ADP (a) or 1 mM GTP (O). C: DNase 1 footprints of

ParA following a heat treatment with (+) or without (-) 1 rnM GTP. ParA was heated for 1

minute at 52°C in buffer A c o n t a k g 100 mM NaCl, 100 ng bovine semm albumidpl, 10%

(v/v) glycerol with or without 1 m M GTP before measuring DNA binding. ParA DNA

binding was assayed in the presence of 1 rnM ATP at 30° C as described in " Experimental

Procedures". The heating step was shorter in this experirnent than in A because ParA is less

stable in these conditions. The positions of the DNA sequence elements are represented as

described in A.

Relative DNase I cleavage of the inverted repeats

GTP - +

87

In these experiments, ParA was heated with different nucleotides and then assayed

for DNA-binding activity in 1 mM ATP. Although different adenine nucleotides stimulate

ParA DNA binding to different extents (Chapter 2). the differences in DNA-binding activity

observed here were not due to carry over of nucleotide fiom the heating step into the DNA

binding step for the following reasons. First, the nucleotide specificity for stability (Fig 3-2)

was different than the nucleotide specificity for DNA binding (Chapter 2). Second. when

ParA was incubated at 0°C rather than 52°C in buffer containhg 1 mM ATP, ADP or

ATPyS and then assayed for DNA-binding activity in 1 m M ATP, there was no difference in

activity compared to each other or to untreated ParA (data not shown) . Finally , when ATP

was added to ParA after rather than before the incubation at 52"C, the ATP had no

stabilizing eEect on ParA (not shown). Therefore, 1 have rneasured the effect of nucleotide

on protein stability, not on DNA-binding activity, in these assays (Fig. 3-2).

ATPyS and ATP stabilized ParA to the same extent in this assay, suggesting that

ATP hydrolysis is not required for maximum ParA stability under these conditions. However

because some ATPases can hydrolyse ATPyS, 1 tested whether ParA hydrolysed ATPyS at

52°C. I did not detect any ParA-dependent hydrolysis of ATPyS at 52°C under conditions

where 1 could detect 1 % or more hydrolysis (data not shown). Under the same conditions,

2-596 of ATP (at 1 mM) was hydrolysed, which is approximately 2- to 5-fold more than the

rate of hydrolysis at 30°C (data not shown). The sirnilarity between the results with ATP and

those with ATPyS was not due to ATP contamination of ATPyS. When a sample of ATPyS

was fractionated on a Mono Q column, using conditions where ATP and ATPyS elute

differently, no ATP was detected in the sample (as littie as 1 % could be detected; data not

shown) .

These experirnents show that adenine nucleotides stabilize ParA, however the

triphosphate-bound forms are more stable than the diphosphate-bound form. This difference

may reflect the difference in ParA affinity for adenine nucleoside diphosphates and

triphosphates, but may also reflect a difference in ParA conformation. These observations

suggest that ParA interaction with adenine nucleotides alter ParA conformation, resulting in

increased stabiiity of ParA.

ATP binding and hydrolysis alter ParA conformation: I tested the conclusion that adenine

nucleotides alter ParA conformation using circular dichroism (CD). Addition of 1 m M ATP

elicited a decrease in ParA molar ellipticity (8) indicating that ATP alters ParA conformation

(Fig. 3-3). ADP or ATPyS (1 mM) each had a smaller, but measurable effect in this assay

(Fig . 3-3). AMP, which does not stimulate ParA DNA-binding activity (Chapter 2), had no

effect on ParA conformation (Fig. 3-3). The high absorbance of light in the samples

containhg nucleotide prevented accurate measurement of ellipticity below 219 m. In this

assay 1 used AMP rather than GTP as a negative control so that the absorption spectra of al1

nucleotides in the buffers were identical. The decrease in molar ellipticity at 222 xm

indicated an increase in helicity upon ParA interaction with nucleotides (Table 3-2) (30). This

increase in helicity corresponded to approximately 16 peptide bonds changing conformation

in a ParA monomer (398 amino acids) with ATP and approximately 8 peptide bonds

changing conformation with ADP or ATPyS. ParA, therefore, assumes three different

conformations as detected by CD: ParA without nucleotide, ParA bound to nonhydrolysable

nucleotide (ADP or ATPyS) and ParA bound to hydrolysable nucleotide (ATP). 1 did not

detect a change in ParA helical content that could be attributed to binding of adenine

nucleoside diphosphates versus triphosphates. 1 conclude therefore that ParA interaction with

nucleotide alters ParA structure and that hydrolysis M e r alters that structure. Hydrolysable

Figure 3-3: ParA conformation with and without nucleotide: A: CD spectra of ParA

(20 pM) in buffer A containhg 100 mM NaCl without nucleotide (@) or with 1 mM AMP

(O), 1 mM ADP (A), 1 mM ATPyS (A) or 1 mM ATP (i) are shown. Spectra were

measured on an Aviv 62A DS circular dichroism specîmmeter. Measurements were read

from 300 to 219 nm in 1 nm intervals with a 1 sec averaging t h e . The spectrum of a buffer

blank with or without 1 m M nucleotide was subtracted Born the ParA spectrum with or

without the corresponding nucleotide. Each experiment is the average of five scans and each

experiment was performed at least three times. B: The data between 219 and 225 nm from A

replotted on an expanded scale.

240 260 280 wavelength (nm)

Table 3-2: Effects of adenine nucleotides on ParA helicity,

measured by CD.

nucleo tide % helicity'

none

AMP

ADP

ATPyS

ATP

' Calculated from the molar ellipticity at 222 nm as described in Chen

et al. (30). The data are the average of at least three different

experiments .

92

and nonhydrolysable nucleotides also have different effects on ParA DNA binding (Chapter

2). It follows therefore that one role for the conformational changes induced by ATP

hydrolysis as well as those induced by ATP binding is to modulate ParA DNA binding and

likely ParA repressor function in vivo.

1 aiso used circular dichroism to examine the effects of adenine nucleotides on ParA

stability and confmed rny initial results (Fig. 3-2). 1 measured temperature dependent

changes in ParA conformation by monitoring ParA molar ellipticity at 220 nm between 25°C

and 65 OC (Fig . 3-4). Since ParA precipitated when it was heated, denaturation was not

reversible and 1 cannot make any conclusions about the thennodynamics of ParA

conformation in response to heat, such as the free energy change. However, 1 can estimate

the temperatures at which ParA denatured with and without nucleotide (Table 3-3). As in the

initial stability assay, adenine nucleotides increased ParA stability as indicated by the

increased temperature at which ParA denatured in the presence of nucleotides (Fig. 3-4,

Table 3-3). However, in this assay ATPyS did not stabiiize ParA as well as ATP did (Table

3-3). This result suggests that ATP hydrolysis modifies ParA conformation, as was also

suggested by the ATP hydrolysis dependent increase of ParA helical content (Fig. 3-3).

However, ADP-bound ParA was still less stable than both adenine nucleoside triphosphate

bound forrns.

ParB sümulates ParA DNA binding: ParB stimulates ParA repressor activity in vivo,

however ParB has no repressor activity on its own (55). Simplistically one might expect that

ParB stimulates ParA repressor activity in vivo via a direct interaction of ParB with ParA at

parOP. 1 predicted that such an interaction would influence ParA DNA binding to the

operator, either by increasing ParA affrnity for parOP or by aitering ParA activity, making it

a more effective block to RNA polymerase (resulting, for example, in a change in the DNase

Figure 3-4: The effects of adenine nucleotides on ParA stability: ParA conformation was

monitored by CD at 220 nm (averaging time 15 sec) from 25 "C to 65 OC. The temperature

was increased in 2°C increments, and the sample was equilibrated to each temperature for 1

minute before measurement of the signal. ParA (20 pM) was in buffer A containing 100 mM

NaCl and 1 rnM ATP (A). 1 mM ADP (i) or no nucleotide (a). The experiment shown is

one of three experiments performed.

temperature (OC)

Table 3-3: Effects of adenine nucleotides on heat denaturation of ParA.

none

ADP

ATPyS

ATP

a Concentration of nucleotide, when present, was 1 mM

Temperature at which the signal at 220 nm had changed by 50% of

total change. The data are the averages from three experiments.

%

1 footprint). In initial experiments no ParB effect on ParA DNA binding was detected

(Chapter 2; 39). However, 1 show here that a large molar excess (10-fold as ParB dimers to

ParA monomers) of ParB over ParA stimulated ParA DNA-binding activity up to 5-fold in

the presence of ATP (Fig. 3-SA). At low ParB to ParA ratios, ParB stimulation of ParA

DNA binding was observed when ParA and ParB were preincubated together More assaying

for DNA binding. Less stimulation of ParA DNA binding by ParB was observed under these

conditions (about 2-fold; Fig. 3-5B). Without this preincubation, no effect of ParB on ParA

DNA binding was observed, consistent with previous results (Fig. 3-5B) (Chapter 2). Since

ParB stimulates ParA DNA binding at lower ParB to ParA ratios after preincubation 1 think

it reasonable that the initial requirement for a large excess of ParB over ParA is the result of

a weak ParB-ParA interaction rather than a high stoichiometric ratio. ParB did not protect

purOP from DNase I cleavage on its own (Fig. 3-SB) nor did ParB alter the pattern of

DNase 1 cleavage in the presence of ParA (e.g. compare the last lane without ParB to the

fifth lane with ParB in Fig. 3-5A). This implies that ParB does not directly contact the DNA

and that ParB probably acts by increasing ParA affinity for parOP.

Nucleotide specifcity of ParB stimulation of ParA DNA-binding activity: ParB stimulated

ParA DNA-binding activity only in the presence of ATP. I did not detect any stimulation of

ParA DNA binding by ParB in the presence of ADP (Fig. 3-6A) or in absence of nucleotide

(Fig. 3-6B). In the absence of nucleotide, ParA bhds parOP weakly, about 10-fold less well

than ParA-ATP binds parOP (Chapter 2). Because a large amount of protein was required in

these assays and because non-specific DNA binding interfered, 1 could not add a 10-fold

molar excess of ParB to ParA DNA binding assays without nucleotide. Instead, less ParB

was used and the two proteins were preincubated on ice before assaying for ParA DNA-

binding activity (as in Fig 3-SB). Under these conditions, ParB did not affect ParA activity

Figure 3-5: ParB stimulates ParA DNA binding in the presence of ATP. A: DNase 1

footprints of ParA binding to parOP in the absence of ParB (-) or in the presence of a

10-fold molar excess of ParB (+) were performed as describeci in "Experimental

Procedures". Each assay contained the indicated amount of ParA and 1 mM ATP. The

amount of ParB, when present, was 10 pmol ParB (dimer) for every 1 pmol ParA

(monomer). In the absence of ParB the same molar concentration of BSA in ParB buffer was

added to the assays. The parA start codon (ATG), the ribosome binding site (RBS) and the

-10 and -35 transcriptional signais are indicated by the boxes to the lep. The position of the

inverted repeat sequence is indicated by the inverted arrows. B: DNA binding was measured

in the presence or absence of ParA and a 2-fold molar excess of ParB (ParB dimers to ParA

monorners). The proteins were preincubated together (b) or separately (a, c and d) on ice in

dilution buffer for 1 hour before assaying for DNA binding at 30°C. In the third set of lanes

("c") ParA and ParB were pre-incubated separately before being added to the DNA binding

assays. DNase 1 footprints were performed in the presence of 1 mM ATP as described in

"Experimental Procedures". The positions of the DNA sequence elements are labelled as in p

part A.

Figure 3-6: ParB does not stimulate ParA DNA bindhg in the presence of ADP or in the

absence of nucleotide: A: DNase 1 footprints of ParA binding to parOP in 1 mM ADP with

or without a 10-fold molar excess of ParB (dimer), as indicated. The positions of the ParA

start codon (ATG), the ribosome binding site (RBS) and the -10 and -35 transcriptional

signals are indicated by the boxes on the f@. The invened arrows mark the position of the

inverted repeat sequences. B: ParA binding to parOP in the absence of nucleotide, with and

without ParB, was measured using DNase 1 footprints as descnbed in "Experimental

Procedures". ParA was mixed in dilution buffer with a 2-fold molar excess of either bovine

serum albumin ("-") or ParB (dimer) and preincubated on ice for 1 hour before assaying for

ParA DNA-binding activity at 30°C. The positions of the various DNA sequence elements

are marked as described in A.

ParB ParA i-

ATG O

f

ATG =

101

(Fig. 3-6B). In the presence of ADP or ATPyS (but without ParB), ParA binds to parOP

very well; about 5-fold better than in the presence of ATP. When a 10-fold molar excess of

ParB was added to ParA DNA binding assays in the presence of 1 mM ADP, no effect of

ParB was detected (Fig. 3-6A). ParB also did not stimulate ParA DNA binding in the

presence of ATPyS (data not shown).

Since ParA-ADP binds to DNA better than PUA-ATP binds (Chapter 2), and ParB

stimulates ParA ATPase (39), one explanation of my results is that ParB stimulates ParA

DNA binding via ParB stimulation of ParA ATPase and hence production of ADP. This

seemed a remote possibüity since approximately 0.1 m M ADP is required to stimulate ParA

DNA binding (Chapter 2), ParA has a lower affinity for ADP than ATP (Table 3-1) and

previous measurements of ParA ATPase with a large molar excess of ParB indicated that

insufficient ADP was produced to stimulate ParA DNA binding (39). Nevertheless, I

measured ATP hydrolysis by ParA under DNA binding conditions with a 10-fold molar

excess of ParB. Insuficient ADP was produced ( f u l concentration < 1 pM; data not shown)

to account for ParB stimulation of ParA DNA-binding activity. It is also unlikely that ADP

was a contaminant in my ParB preparations since ParB was purified in several steps and

ParB did not stimulate ParA DNA binding in the absence of nucleotide (Fig. 3-6B).

1 suggest therefore that ParB stimulates ParA DNA binding via a direct interaction

with ParA. Interestingly , ParB stirnulated ParA DNA binding (with ATP) to the same extent

that ADP does over ATP without ParB (up to 5-fold; Chapter 2; Fig. 3-SA). Since ParB

stimulates ParA ATPase (39) and ATP hydrolysis by ParA is inhibitory to ParA DNA

binding (Chapter 2), 1 propose that ParB stimulates ParA DNA binding by circumventing the

negative action of ATP hydroly sis on ParA DNA- binding activity .

DISCUSSION

P l ParA binds (Fig . 3- 1) and hydrolyses ATP (39), and one or both of these

activities are essential for both its regulation and partition activities (41). Nucleotide-binding

experiments (Fig. 3-1) suggest one ATP-binding site per ParA monomer, which agrees well

with the presence of one putative ATP-binding site defined by sequence d y s i s (133), and

the observation that deletion of this region of the protein destroys ATPase activity (41). I

have used different adenine nucleotide cofactors to probe the structure and function of ParA.

My results suggest that ATP binding and ATP hydrolysis by ParA induce different

conformations of ParA that in nini have different roles in ParA's activity. In addition, I

suggest that ATP binding and hydrolysis play separable roles in ParA activities in vitro and

consequently in partition and par gene expression in vivo.

ParA conformation with and without nucleotide: Upon interaction with ATP, ADP and

ATPyS, ParA changes conformation (Fig. 3-3) and ParA dimerization and DNA binding are

stimulated (Chapter 2; 39). The nucleotide specificities of the changes in ParA helical content

detected by CD and of the stimulation of PUA DNA binding are the same, irnplying that the

effects of ATP binding and hydrolysis on ParA DNA-binding activity are mediated by their

effects on ParA conformation. However, the nucleotide specificities for dimerization and

helical content are different, suggesting that dîmerization is not the ody factor detennining

the change in ParA conformation upon interaction with nucleotides. The concentration of

intracellular nucleotide pools is sufficiently high (eg, ATP is in the millimolar range; 54) that

ParA would rarely be free of nucleotide in vivo. However, differences between ParA-ATP

and ParA-ADP or perhaps the interchange between these conformations are likely to be

important for ParA function in partition.

103

Differentiation between adenine nudeoside di phosphates and triphosphates: 1 found it

intriguing that ADP and ATPyS behaved similarly in my assays, such as site-specifc DNA

binding (Chapter 2) and conformation (Fig . 3-3). This raised the formal possibility that ParA

did not discriminate between nucleoside di- and triphosphates, except via the act of ATP

hydrolysis. Here, measurements of nucleotide binding indicate a slightly better affinty for

ATP than for ADP, and a large difference in the ability of adenine nucleoside tri- and

diphosphates to stabilize ParA. The differences in stability may be a function of the

differences in affhity for nucleotide, andlor due to differences in ParA conformation that 1

cannot detect by CD. Differences in ParA stability mediated by nucIeotide could also

potentially play roles in partition or reguiation.

Effects of ATP hydrolysis on ParA structure and function: ParA consistently behaves

differently when bound to a hydrolysable nucleotide than when bound to a nonhydrolysabie

nucleotide. The ATP bound form of ParA has a lower affinity for DNA (Chapter 2), it has a

different conformation than the other fonns (Fig. 3-3 & 3-4) and it is the only form

detectably modrilated by ParB (Fig. 3-5 & 3-6). My current results provide further support

that ATP hydrolysis plays roles in ParA function that are distinct and separable fiorn ATP

b ind ing .

ParA-ParB interactions: ParB stimulates ParA binding to parOP, but only in the presence

of ATP. ParB did not stimulate DNA binding by ParA lacking nucleotide (Fig. 3-6B), a form

of ParA that binds parOP weakly (Chapter 2). 1 detected no ParB effect on DNA binding by

ParA-ADP or ParA-ATPyS (Fig. 3-6A; data not shown), forms of PUA which bind to DNA

about bfold better than ParA-ATP binds to DNA (Chapter 2). ParB stimulates ParA-ATP

DNA-binding activity to the same level as ParA-ADP DNA-binding activity with or without

ParB. Since ATP hydrolysis by ParA in the absence of ParB is inhibitory to DNA binding

104

(Chapter 2), 1 think that the simplest explanation is that ParA association with ParB prevents

the inhibitory action of ATP hydrolysis on PUA DNA binding. This mode1 also explains the

observation that ParB stimulates both ParA ATPase and ParA DNA binding, since one might

otherwise expect increased ATP hydrolysis by ParA to further interfere with ParA DNA

binding. Mutations have been identified in other ATPases that can bind but not hydrolyse

ATP (165, 191). My hypothesis predicts that a similar ParA mutant would repress

transcription from purOP as well as wild-type ParA represses in the presence of ParB (S),

and that this repression would be independent of ParB.

Skce ParB stimulates ParA DNA-binding activity only in the presence of ATP (Fig.

3-5) and not with ADP, ATPyS or without nucleotide (Fig. 3-6; data not shown) 1 conclude

that ParA-bound nucleotide is required for ParB stimulation of ParA DNA binding and this

bound nucleotide must be hydrolysable. Why is hydrolysis required? One possible

explanation is that ATP hydrolysis is required for physical association of ParA and ParB.

Preliminary CO-immunoprecipitation experiments suggest that ParA and ParB associate in the

absence of ATP (Chapter 4), so 1 favor other interpretations of these results. Perhaps a

ParA-ParB complex foms regardless of the nucleotide present but is detectable by my DNA

binding assay only in the presence of ATP hydrolysis. ParA has no detectable kinase activity

(39; my unpublished data) so 1 think it unlikely that phosphorylation of either protein by

ParA is required for activity.

Roles of ATP in par gene regulation and partition: The ATP site in ParA is essential for

both ParA's regulatory and partition activities in vivo (41). One role of ATP binding is to

promote ParA binding to parOP for ParA repressor activity (Chapter 2; 39). A role for ATP

hydrolysis in repression is suggested by its ability to make ParA DNA binding and thus ParA

repressor sensitive to ParB levels. Alternatively, one rnight argue that the role of ParB in

105

repression is to make ParA DNA binding insensitive to ATP hydrolysis. I do not yet kuow

what other roles ATP binding and hydrolysis have for ParA function in vivo. It is possible,

for example, that ATP hydrolysis is necessary only for partition and not regdation. Thus a

putative ParA mutant that c m bind but not hydrolyse ATP would function as a repressor, but

would not be able to support partition in vivo. InhiitiveIy, 1 think it likely that ATP

hydrolysis is required for the positioning reaction, to fulfd the energy requirement inherent in

the partition process. For example, confoxmational changes induced by ATP binding and

hydrolysis in some proteins are thought to mediate the displacement of other structures in the

cell, such as tramport across membranes (MalK; 167), passage of one DNA strand through

another (topoisornerase II; 17) or movement dong microtubules (kinesin; 8 1). Perhaps the

conformational changes induced in ParA by ATP binding and hydrolysis result in the

movement or orientation of the partition apparatus. As a component of the partition cornplex,

ParB could assist with ParA function via ParB interaction with ParA and by ParB stimulation

of ParA ATPase.

ParA-ParB interactions are likely important for ParA7s activity in the positioning

reaction. PUA interaction with the partition cornplex via interaction with ParB may be

required for subsequent steps in the positioning process, such as interaction with host factors.

Host factors rnight Uiteract with ParA, ParB or both proteins. These host factors are still

undefmed although they are not DNA sites on the host chromosome (62). 1 assume that they

are proteins or membrane components that position partition complexes andlor time

positioning events with respect to the cell division cycle. A careful analysis of ParA-PUB

interactions at pars, including the roles of ATP binding and hydrolysis in these interactions,

as well as identification of host factors, are crucial to the dissection of the P l plasmid

partition pathway .

CHAPTER 4: Isolation of ParA-ParB complexes and Future Directions

This chapter presents prelirninary experiments to isolate ParA-ParB complexes.

INTRODUCTION

ParA-ParB imractions have a role in repression (Chapter 3; 55) and are potentiaiiy

important for ParA hinction in the positionhg process. ParB's role in par gene expression is

to stimulate ParA repressor activity (55). My results in Chapter 3 suggest that at least part of

ParB's stimulatory effect on ParA DNA-binding activity is the result of ParB's abüity to

overcome the inhibitory effects of ATP hydrolysis on ParA-parOP interaction. ParA's second

role in partition, inferred from genetic experiments (Chapter 1 ; 55, 77), is likely mediated by

ParA interaction with ParB in the partition complex (ParB-MF-pars complex). Further

characterization of ParA interaction with ParB and the partition complex as well as

examination of the effects of adenine nucleotides on these interactions are crucial to

understanding ParA function in partition.

Although ParA and ParB interact (Chapter 3; 39. 5 3 , a complex of the two proteins

has not previously been isolated. ParB effects on ParA activity have been detected only in the

presence of ATP hydrolysis by ParA (Chapter 3; 39). suggesting that ATP hydrolysis may be

required for physical association of ParA and ParB. Here. 1 isolate ParA-ParB complexes

using CO-immunoprecipitation assays and present pre1imhx-y results that suggest that ATP

hydrolysis is not required for formation of these complexes. The results in this chapter as

well as in the previous chapters make certain predictions about ParA and its interactions with

ATP, ADP and ParB in partition.

METHODS

Preparation of 35S-labelled ParA extracts: ParA-encoding ceIIs (BL2 1 (XDE3 pLysS

pBEF198); 35). which express parA under control of the T7 RNA polymerase promoter,

were grown in M9 minimal medium (164) containing 1% (wlv) glucose, 100 pg

arnpicillin/mi and 25 pg chloramphenicoYml. When the A, of the culture had reached 0.5,

ParA expression was induced with 0.5 mM IPTG (BioRad). After 15 minutes at 37OC, 10

pCi 35S-methionine (Amersham)/rnl was added and the cells were grown for another 30

minutes at 37°C. The cells were harvested, excess label was removed by two washes in M9

media and the cells were resuspended in 50 mM Tris-HCI pH 7.5, 10% (wlv) sucrose and

frozen to -80°C. Cells were lysed by lysosome lysis (58) and ce11 debris was removed by

centrifugation. The amount of ParA in the ce11 lysate was estimated by comparing to known

arnounts of pure ParA on a Western blot using anti-ParA antibodies (Chapter 2).

Preparation of 3 5 S - ~ a r ~ protein: Labelled crude extracts were prepared from BL21 (hDE3

pLysS pBEF198) cells as described above except that rifampicin was added 7 minutes after

IPTG induction to suppress transcription by E. coli RNA polymerase. 35S-methionine was

then added afier a further 8 minutes incubation at 37°C. Cells were incubated for 30 minutes

before processing as described above. "S-ParA (FrJII) was isolated from these crude extracts

as described in Chapter 2. except on a smailer scale.

Antibodies and Protein A Sepharose beads: Anîi-ParB antibody (61) was affinity purified

against a version of ParB tagged at the amino tenninus with 6 histidine residues and a heart

muscle kinase recognition site (a gift from J. Surtees; 181). The ParB hision protein was

immobilised on nitrocellulose and used to affiiinity puri@ anti-ParB antibody using standard

methods (164). The Protein A Sepharose beads (Sigma) were swelled in IP buffer (10 mM

109

Tris-acetate, pH 7.5, 100 mM NaCl, 5 mM MgCl, and 0.1 % (dv) Triton X-100) for at least

one hour and washed twice in IP buffer before use.

Co-immunoprecipitation of ParA and ParB: ParA was precipitated with ParB using affin@

punfied anti-ParB antibodies and Protein A Sepharose beads. Each 100 pl assay contained

approximately 3 pg ParA in crude extract or 3sS-ParA ( F r a , 8 pg ParB (FrN; 57, 58) and

40 pg BSA (Sigma FrV) in IP buffer (10 mM Tris-acetate, pH 7.5, 100 mM NaCl, 5 mM

MgCl, and 0.1 % (vlv) Triton X- 100). Five microliters of 10 % (w/v) Protein A Sepharose in

IP buffer was added to each assay. Following a 30 min incubation at room temperature, the

mixtures were cenaifbged at 12,000 rpm for 10 min. Approxirnately 5 pg of a f f i t y purified

anti-ParB antibody was added to the supernatant and the mixture was incubated for 1 hou at

room temperature with constant mixing. Twenty-five microliters of 10% (w/v) Protein A

Sepharose in IP buffer were added. The mixtures were incubated for 1 hr at room

temperature with constant end-over-end rnixing. The Sepharose beads were then pelleted by

centrifugation and the supernatant was removed. The beads were washed three times with IP

buffer and then resuspended in 50 pl L a e d i sample buffer (104). The proteins were

separated by SDS-PAGE (10%) and exposed to a PhosphorIrnager (Molecular Dynamics)

screen.

RESULTS

Using a CO-immunoprecipitation assay 1 have isolated a ParA-ParB complex from

crude E. coli extracts that contained ParA (Fig. 4-1). ParB was incubated with 35S-labelled

crude E. coli lysates containing ParA, and ParB was precipitated using anti-ParB antibodies

and Protein A Sepharose. Co-precipitation of ParA depended on the presence of ParB (Fig.

4-2) and anti-ParB antibodies (Figs. 4-1 and 4-2). There was some background binding of

ParA to the Protein A Sepharose beads, however signifcandy more ParA was precipitated

when antibody and ParB were present (approxirnately 8-fold more, Fig. 4-1B). When

partially purified, 35S-labelled ParA (Frm) was used identical results were obtained (Fig. 4-

2). These data c o n f i i that ParA and ParB interact directly and show that a ParA-ParB

complex can be isolated.

IR vitro, effects of ParB on ParA activity have only been detected in the presence of

ATP hydrolysis (Chapter 3; 39), suggesting that ATP hydrolysis may be required for

formation of a ParA-ParB complex. However, I detected CO-irnrnunoprecipitation of ParA

( F r a with ParB in the absence of any added ATP. Although it has not yet been fomally

demonstrated, ParA (FrIII), ParB, a f f ~ t y purified antibody and the Protein A Sepharose

beads are unlikeIy to contain nucleotide; ParA and ParB were purified using at least one ion-

exchange column and the antibody was af f i ty purified. Therefore since ParA and ParB

interact in the absence of nucleotide, ATP and ATP hydrolysis do not seem to be required

for association of ParA and ParB.

Relatively large amounts of both ParA and ParB were present in these assays.

Although the precipitation of ParA was specific and ParB-dependent, only a small percentage

(about 2-5%) of the total ParA present was precipitated. This low recovery may reflect a

weak ParA-Pd interaction. Consistent with a weak ParA-ParB interaction, a relatively high

Figure 4-1. Physical d a t i o n of ParA and ParB. A ParA-ParB complex was isolated by

precipitation of ParA by ParB and anti-ParB antibodies. An SDS-polyacrylamide gel of "S-

labelled ParA in crude extracts before ("Lw) or after ("IP") CO-immunoprecipitation is shown.

Antibody (no Aby) or Protein A Sepharose beads (no beadr) was omitted in the indicated

lanes. The migration of rnolecular size markers (New England Biolabs) is indicated to the le#

of each gel. The band correspondhg to ParA is marked by the arrow. B: ImageQunt

(Molecular Dynamics) software was used to quanti@ the relative amount of ParA precipitated

from crude lysates by ParB. anti-ParB antibody and Protein A Sepharose beads ("complete")

compared to assays from which antibodies ("no A b " ) or Protein A Sepharose beads ("no

beadr") had been omitted. Quantification of a line through each of the three "IP" lanes in A

was used to obtain the arbitrary values in the Y-axis.

ParA

47.5 'w 1 ParA

Figure 4-2. Association of ParA and ParB in the absence of nucleotide. Co-

immunoprecipitation of purified ParA (FrIII) and ParB using anti-ParB antibodies and Protein

A Sepharose beads ("cornpiete"). Anti-ParB antibodies ("no Aby") or ParB ("no ParB") was

omitted fiom the indicated lanes. The faster migrating species in the FrIII ParA sample are

likely degradation products of ParA since they reacted with anti-ParA antibodies on a

Western blot (data not shown). The migration of molecular size standards through the gel is

indicated to the Z e j t of the gel.

concentration of ParB is required to rnaximdiy stimulate ParA DNA binding and ParA

ATPase in vitro (Chapter 3; 39). Perhaps nucleotide stimulates ParA-ParB interaction and

more complex would be isolated in the presence of nucleotide. The amount of ParB

immunoprecipitated in these assays has not k e n quantified and it is possible that not al1 of

the ParB was precipitated. In addition, the anti-ParB antibodies are polyclonal and may

interfere with ParA-ParB interactions. Therefore fuIther optimi7irtion of these preliminary

assays is required to make any conclusions about the strength of the interaction between ParA

and ParB using these assays (see Discussion and Future Directions).

1 15

DISCUSSION AND Fü'ïURE DIRECTIONS

In this thesis, 1 have presented experiments that address the roles of ATP binding and

hydrolysis in ParA function, particularly in ParA repressor function. In addition to its role in

repression, ParA is required for an undefmed role in the partition process (41, 55). The

importance of ParA repressor activity for P l plasmid stability is illustrated by the

destabilizing effect of ParA andor ParB overexpression on Pl (3, 56, 77). ParA repressor

activity is probably mediated by ParA interaction with parOP. 1 have s h o w that ATP is not

required for ParA interaction with parOP (Chapter 2) as was previously believed (39),

although ATP does greatly stimulate ParA-parOP interaction (Chapter 2 ) . 1 have

demonstrated that either adeniae nucleoside di- or tn-phosphates are required to stimulate

ParA DNA binding, and that ATP hydrolysis by ParA inhibits ParA DNA binding (Chapter

2). These effects of adenine nucleotides on ParA are mediated by their effects on ParA

conformation, including but not limited to, stimulation of ParA dimerization (Chapters 2 and

3). My results have led to a mode1 of ParA repressor function in which ATP binding and

ATP hydrolysis have separable roles (Chapter 3). ParA interaction with ParB is also

important for ParA function in repression (55) and, 1 believe, for ParA's partition function. 1

have s h o w that ParB stimulates ParA interaction with parOP in vitro (Chapter 3), consistent

with ParB's role as a CO-repressor in vivo (55). This stimulation by ParB requires ATP

hydrolysis (Chapter 3) which has implications for ParA-ParB interactions (see below). As a

first step towards addressing the roles of ParA-ParB interactions in partition 1 have begun to

examine the physical association of the two proteins.

ParA-ParB interactions: What are the potential roles for ParA-ParB interactions in

partition? I think it likely that ParB interaction with ParA mediates ParA interaction with the

partition cornplex, an interaction that must occur if ParA has a direct role in the positionhg

116

process. ParA interaction with the partition complex may be required for plasmid pairing or

for specific interactions with host components that in tum mediate positioning of the

plasmids. Alternatively, ParATs effect on the partition complex may be negative, for example

by causing dissociation of paired or attacheci complexes.

The CO-immunoprecipitation assays provide the f ~ s t evidence that a complex of ParA

and ParB can be isolated. This information wii1 be important in developing assays to detect

and measure ParA-ParB interaction at pars (see below). In addition, these assays are a good

qualitative assay for ParA-ParB interaction and will be useful for assessing ParA and ParB

point mutations and deletions for their abilities to interact.

Only a small percentage of the total ParA added to the assays was precipitated, and

although this may reflect a weak ParA-ParB interaction, it may also be due to lirniting

quantities of anti-ParB and Protein A Sepharose beads in the assays. In addition, the

antibodies used in these assays are polyclonal and may interfere with ParA-ParB interaction.

Optimization of the amounts of ParA, ParB, antibody and beads in these prelirninary assays

will address some of these questions. If these assays are to be used to examine nucleotide

effects then the reagents must be assayed for contaminating ATPase activity and nucleotides.

However, further characterization of ParA-ParB interactions, specifically measurement of the

strength of the interaction and measurement of the effects of adenine nucleotides on ParA-

ParB interaction, may require a different approach.

ParA interaction with ParB in vitro has only been observed in the presence of ATP

hydrolysis (Chapter 3; 39). Using CO-immunoprecipitation assays 1 have shown that ATP

hydrolysis does not seem to be required for physical association of ParA and ParB (Fig . 4-2),

so ATP hydrolysis must have some other effect on ParA-ParB interactions at parOP. ATP

hydrolysis may affect ParA affinïty for ParB and without this increase in affinity, ParB

117

stimulation of ParA DNA binding could not be detected. Alternatively, ParB rnay alter ParA-

ATP conformation so that it resembles ParA-ADP and hence binds parOP with greater

aff i ty. These possibiiities have implications for ParA function in partition. If ATP

hydrolysis stimulates PUA-ParB interaction then ATP hydrolysis may also stimulate ParA

interaction with the partition compIex. Thus ParA-ADP would be less active for partition

than ParA-ATP. Quantification of the interaction between ParA and ParB as weIl as

rneasurement of the effects of different factors such as ATP or ADP on this interaction can

be measured using surface plasmon resonance (related to refiactive index) on a BIAcore

Biosensor system (eg., 138). BIAcore measures interactions between proteins or proteins and

DNA (94) and can be used to determine the aff ï ty and kinetic parameters of the interactions

kg., 50).

If ParB alters ParA-ATP conformation then perhaps the transition from one

conformation to the other plays a role in positioning. Conformational changes hiduced by

ATP binding and hydrolysis in other proteins mediate movement of molecules in the ce11

(eg . , Refs . 17, 8 1 and 167). ParB interaction with ParA would facilitate ParA's function in

this capacity by stimulating ParA ATPase, by inducing the conformational change in ParA

and by anchoring ParA to the partition complex. One possible way to examine

conformational changes in ParA and/or ParB associated with ParA-Pa* interactions is

limited proteolysis of the proteins.

ParA interaction with the partition complex: If ParA interacts directiy with ParB at pars,

then one might expect ParA to affect ParB interaction with parS. There are several different

DNA binding assays for ParB binding to pars (38, 57, 60) that can be used to examine ParA

effects on ParB DNA binding using conditions similar to those used to detect ParB

stimulation of ParA DNA binding (Chapter 3) and those used to isolate ParA-ParB

118

complexes (Figures 4-1 and 4-2). Conversely, effects of the partition complex on ParA

activities, such as its ATPase activity, can also be examined.

Roles of ATP binding and hydrolysis in pur gene expression and partition: An energy

requirement is inherent in the positioning process ; it is an entropically unfavourable event.

Many plasmid partition systems encode ATPases that are required for partition (Chapter 1),

suggesting that these ATPases are important for plasrnid partition. For example, these

ATPases may be required to move or orient plasmids in the ce11 (see Chapter 3). Davis et. al

(41) show that mutating the Walker A motif of the putative ATP binding site interferes with

ParA repressor and partition functions. However, their mutant was inactive for al1 activities

tested and they did no: test their mutant for ATP binding, therefore one cannot make any

fm conclusions about the differential roles of ATP binding and hydrolysis by ParA in

repression and partition from this mutant. It wodd be interesting to search for ParA mutants,

created by random or site-directed mutagenesis, that affect ATP hydrolysis and/or binding in

vitro and test thern for their abilities to support par gene repression and partition in vivo.

In addition to these approaches, specific roles for ParA in positioning c m be tested.

For instance, plasmid interaction with the ce11 membrane has been proposed in many

partition models (Chapter 1). 1s ParA the anchor that holds Pl at the membrane? This

activity would require that ParA is localized, at least some of the tirne, to the membrane.

There is no evidence to suggest that ParA is an integral membrane protein (the purification

properties of ParA suggest that it is soluble), however ParA may associate with specific

factors at the membrane. Some members of the ParA superfarnily are associateci with the

membrane (45, 187). If ParA interacts with the membrane then specifc membrane

components rnay also affect ParA biochemical activities in vitro, such as ParA ATPase

activity or ParA interaction with ParB. Further analysis of ParA interaction with ATP, ADP

119

and ParB, as well as examination of potentiai interactions of ParA with host factors, are

crucial to derstanding the function of ParA, and partition ATPases in general, in the

positioning process.

REFERENCES

121

Abeles, A. L. 1986. P l Plasmid replication. mirification and DNA-binding activity of

the replication protein RepA. J. Biol. Chem. 261:3548-3555.

Abeles, A. L., and S. J. Austin. 1991. Antiparallel plasmid-plasmid pairing may

control P l plasmid replication. Proc. Natl. Acad. Sci. USA. 88:9011-9015.

Abeles, A. L., S. A. Friedman, and S. J. Austin. 1985. Partition of unit-copy

miniplasmids to daughter cells III. The DNA sequence and functiod organization of

the Pl partition region. J. Mol. Biol. 185261-272.

Adams, D. E., J. B. Bliska, and N. R. CozzarelIi. 1992. Cre-lox recombination in

Escherichia coli cells. Mechanistic differences from the in vitro reaction. J . Mol.

Biol. 226:66l-673.

Adams, D. E., E. Shekhtman, E. L. Zechiedrich, M. B. Schmid, and N. R.

Cozzarelli. 1992. The role of topoisornerase IV in partitioning bacterial replicons and

the structure of catenated intermediates in DNA replication. Cell. 71:277-288.

herlund, T., R. Bernander, and K. Nordstrom. 1992. Cell division in

Escherichia coli minB mutants. Molec. Microbiol. 6:2073-2083.

Alder, H. I., W. D. Fisher, and A. A. Hardigree. 1969. Ce11 division in

Escherichia coli. Tram N.Y. Acad. Sci. 31:1059-1070.

Arai, K., S. Yasuda, and A. Kornberg. 1981. Mechanism of dnaB protein action 1.

Crystallization and properties of dnaB protein, an essentiai replication protein in

Eschenchia coli. J . Biol. Chem. 2565247-5252.

Argaet, V. P., T. J. Wilson, and B. E. Davidson. 1994. Purification of the

Escherichia coli regulatory protein TyrR and analysis of its interactions with ATP,

tyrosine, phenylalanine and tryptophan. J. Biol. Chem. 269:5 17 1-5 178.

122

Austin, S., and A. Abeles. 1983. Partition of unit-copy miniplasmids to daughter

cells 1. Pl and F miniplasmids contain discrete, interchangeable sequences sufficient

to promote equipartition. J. Mol. Biol. 169:353-372.

Austin, S., and A. Abeles. 1983. Partition of unit-copy miniplasmids to daughter

ceils II. The partition region of miniplasmid P l encodes an essential protein and a

centromere-like site at which it acts. J. Mol. Biol. 169~373-387.

Austin, S., S. Friedman, and D. Ludtke. 1986. Partition functions of unit-copy

plasmids can stabilize the maintenance of plasmid pBR322 at low copy nurnber. J.

Bacteriol. 168: 1010-1013.

Austin, S., M. Ziese, and N. Sternberg. 1981. A novel role for site-specific

recombination in maintenance of bacterial replicons. Cell. 25729-736.

Bechofer, D. H., J. A. Kornacki, W. Fiihein, and D. A. Figurski. 1986. Gene

control in broad host range plasmid RK2: expression, polypeptide product, and

multiple regulatory functions of korB. Proc. Natl. Acad. Sci. USA. 83:394-398.

Begg, K. J., and W. D. Donachie. 1991. Experiments on chromosome separation

and positioning in Escherichia coli. New Biologist. 3:475-486.

Beil, S. P., and B. Stillman. 1992. ATP-dependent recognition of eukaryotic origins

of DNA replication by a multiprotein cornplex. Nature. 357: 128-1 34.

Berger, J. M., S. J. Gambh, S. C. Harrison, and J. C. Wang. 1996. Structure

and mechanism of DNA topoisornerase 11. Nature. 379:225-232.

Bernard, P., and M. Couturier. 1992. Ce11 killing by the F plasmid CcdB protein

involves poisoning of DNA-topoisornerase II complexes. J. Mol. Biol. 226:735-745.

123

Biek, D. P., and J. Shi. 1994. A single 43-bp su& repeat of plasmid mini-F is

sufficient to allow assembly of a functiond nucleoprotein cornplex. Proc. Nati. Acad.

Sci. USA. 91:8027-803 1.

Biek, D. P., and J. Strings. 1995. Partition functions of mini-F affect plasmid

topology in Escherichia coli. J . Mol. Biol. 246:38840û.

Blakely, G., S. Colloms, G. May, M. Burke, and D. Sherratt. 1991. Escherichia

coli XerC recombinase is required for chromosomal segregation at ce11 division. New

Biologist. 3:789-798.

Blakeiy, G., G. May, R. McCulloch, L. K. Arciszewska, M. Burke, S. T. Lovett,

and D. J. Sherratt. 1993. Two related recombinases are required for site-specific

recombination at dif and cer in E. coli K12. Cell. 7 5 3 5 1-361.

Boe, La, K. Gerdes, and S. Molin. 1987. Effects of genes exerting growth inhibition

and p lasmid stability on plasmid maintenance. J. Bacteriol . l69:4646-4650.

Bork, P., C. Sander, and A. Valencia. 1992. An ATPase domain comrnon to

prokayotic ce11 cycle proteins, sugar kinases, actin and hsp7O heat shock proteins.

Proc. Natl. Acad. Sci. USA. 89:7290-7294.

Borodovsky, M., E. V. Koonlli, and K. E. Rudd. 1994. New genes in old

sequence: a strategy for fmding genes in the bacterial genome. Trends Biochem. Sci.

19:309-3 13.

Bradford, M. M. 1976. A rapid and sensitive method for the quantitation of

microgram quantities of protein utilizing the principle of protein-dye binding. Anal.

Biochem. 72:248-254.

Breüner, A., R. B. Jensen, M. Dam, S. Pedersen, and K. Gerdes. 1996. The

centromere-like parc locus of plasrnid R1. Mol. Microbiol . 20: 58 1-592.

124

28. Cerin, Hm, and J. Hackett. 1993. The parVP region of the Salmonella typhimurium

virulence plasmid pSLT contains four loci required for incornpatibility and partition.

Plasmid. 30:30-38.

29. Chattoraj, D. K., K. M. Snyder, and A. L. Abeles. 1985. P l plasmid replication:

multiple functions of RepA protein at the ongin. Proc. Natl. Acad. Sci. USA.

82:2588-2592.

30. Chen, Y. II., J. T. Yang, and H. M. Martinez. 1990. Determination of the

secondary structure of proteins by circular dichroism and optical rotary dispersion.

Biochern. ll:4120-4l3l.

3 1. Clerget, M. 199 1. Site-specific recombination promoted by a short DNA segment of

plasmid R1 and by a homologous segment in the terminus region of the Escherichia

coli chromosome. New Biologist. 3:780-788.

32. Coiioms, S. D., P. Sykora, G. Szatmari, and D. J. Sherratt. 1990. Recombination

at ColEl cer requires the Escherichia coli xerC gene product, a member of the

lambda integrase family of site-specific recombinases. J. Bactenol. 172:6973-6980.

33. Dam, M., and K. Gerdes. 1994. Partitioning of plasmid RI. Ten direct repeats

flanking the parA promoter constitute a centromere-like partition site parc, that

expresses incompatibility . I. Mol. Biol. 236: 1289- 1298.

34. Dassain, M., and J.-P. Bouche. 1994. The min locus, which confers topological

specificiv to ce11 division, is not involved in its coupling to nucleoid separation. J.

Bacteriol. l76:6 143-6145.

35. Davey, M. J., and B. E. Funneil. 1994. The P l plasmid partition protein ParA: A

role for ATP in site-specific DNA binding. J. Biol. Chem. 269:29908-29913.

Davie, E., K. Sydnor, and L. 1. Rothfeld. 1984. Genetic basis of minice11

formation in Eschenchia coli K12. J . Bactenol. 158: 1202- 1203.

Davies, J. K., and P. Reeves. 1975. Genetics of resistance to colicins in Escherichia

coli K-12: Cross-resistance among colicins of group A. J. Bacteriol. 123: 102-1 17.

Davis, M. A., and S. J. Austin. 1988. Recognition of the Pl plasmid centromere

analog involves binding of the ParB protein and is modified by a specific host factor.

EMBO 1. 7~1881-1888.

Davis, M. A., K. A. Martin, and S. J. Austin. 1992. Biochemical activities of the

ParA partition protein of the Pl plasmid. Moi. Microbiol. 6: 1 141-1 147.

Davis, M. A., K. A. Martin, and S. J. Austin. 1990. Specificity switching of the

Pl plasmid centromere-like site. EMBO J. 9:99 1-998.

Davis, M. A., L. Radnedge, K. A. Martin, Fm Hayes, B. Youngren, and S. J.

Austin. 1996. The Pl ParA protein and its ATPase activity play a direct role in the

segregation of plasmid copies to daughter cells. Mol. Microbiol. 21: 1029-1036.

Davis, T o La, D. R. H e b k i , and R. Cm Roberts. 1992. Transcription and

autoregulation of the stabilizing functions of broad-host-range plasmid RK2 in

Escherichia coli, Agrobactenum tumefaciens and Pseudomonas aemginosa . Mol.

Microbiol. 6: 1981-1994.

Dean, F. B., M. Dodson, Hm Echols, and J. Aurwitz. 1987. ATP-dependent

formation of a specialized nucleoprotein structure by simian virus 40 (SV40) large

hmor antigen at the SV40 replication origin. Proc. Natl. Acad. Sci. USA.

84:8981-8985.

Deb, S. P.. and P. Tegtmeyer. 1987. ATP enhances the binding of simian vims 40

large T antigen to the origin of replication. J. Virol. 61:3649-3654.

126

45. de Boer, P. A. J., Ra E. Crossley, A. R. Hand, and L. 1. Rothï~eld. 1991. The

MinD protein is a membrane ATPase required for the correct placement of the

Escherichia coli division site, EMBO J. lO:4371-4380.

46. Dodd, 1. B., and J. B. Egan. 1990. Improved detection of helix-tum-helix

DNA-binding motifs in protein sequences. Nucleic Acids Res. l8:5019-5027.

47. EberI, L., M. Givskov, and H. Schwab. 1992. The divergent promoters mediating

transcription of the par locus of plasmid RP4 are subject to autoregulation. Mol.

Microbiol . 6: 1969- 1979.

48. Eckstein, F. 1985. Nucleoside phosphorothioates. Annu. Rev. Biochem. S4:367-402.

49. Ezaki, B., To Ogura, H. Niki, and S. Hiraga. 1991. Partitionhg of a rnini-F

plasmid into anucleate cells of the mukB nul1 mutant. J. Bacteriol. 173:6643-6646.

50. Fagerstam, L. Go, A. Frostell-Karisson, Ra Karisson, B. Persson, and 1.

Ronnberg. 1992. Biospecific interaction analysis using surface plasmon resonance

detection applied to kinetic, binding site and concentration analysis. I. Chromatog.

597:397-410.

5 1. Feuerstein, J., R. S. Goody, and M. R. Webb. 1989. The mechanism of guanosine

nucleotide hydrolysis by p21 c-Ha-ras. The stereochemical course of the GTPase

reaction. J. Biol. Chem. 264:6188-6190.

52. Figurski, D. Hm, R. J. Meyer, and Dm R. Helinski. 1979. Supression of ColEl

replication by the IncP1 plasmid RK2 in hybrid plasmids constructed in vitro. J.

Molec. Biol. l33:295-3 18.

53. Frame, R., and J. O. Bishop. 1971. The number of sex-factors per chromosome in

Escherichia coli. Biochem. J . l21:93-103.

127

Franzen, J. S., and S. B. Binkley. 1961. Cornparison of the acid-soluble nucleotides

in Escherichia coli at different growth rates. J. Biol. Chem. 236515-519.

Friedman, S. A., and S. J. Austin. 1988. The Pl plasmid partition system

synthesizes two essential proteins from an autoregulated operon. Plasmid. 19:103-112.

Funneil, B. E. 1988. Mini-P 1 plasmid partitioning : Excess ParB destabilizes plasrnids

containing the centromere parS. J. Bacteriol . 170: 954-960. .

Funneli, B. E. 1991. The Pl plasmid partition complex at pars. The influence of

Eschenchia coli integration host factor and of substrate topology . J. Biol. Chem.

266: 14328-14337.

Fumeli, B. E. 1988. Participation of Escherichia coli integration host factor in the

P 1 plasmid partition system. Proc. Nati. Acad. Sci. USA. 85:6657-666 1.

Funnell, B. E. 1996. The Role of the Bacterial Membrane in Chromosome

Replication and Partition. R.G. Landes, Austin, Texas.

Funneii, B. E., and L. Gagnier. 1993. The Pl plasmid partition complex at pars II.

Analysis of ParB protein binding activity and specificity . J. Biol. Chem.

268:36 16-3624.

Funneli, B. E., and L. Gagnier. 1994. Pl plasmid partition: Binding of Pl ParB

protein and Escherichia coli integration host factor to altered pars sites. Biochimie.

76: 924-932.

Funnell, B. E., and L. Gagnier. 1995. Partition of P l plasmids in Escherichia coli

mukB chromosomal partition mutants. J. Bacteriol. l77:2381-2386.

Gallie, D. R., and C. 1. Kado. 1987. Agrobacterium tumefaciens pTAR parA

promoter region involved in autoregulation, incompatibility and plasmid partitioning.

J. Mol. Biol. 193:465-478.

128

Georgiadis, M. M., H. Komiya, P. Chakrabarti, Do Woo, J. J. Kornuc, and D.

C. Rees. 1992. Crystallographic structure of the nitrogenase iron protein from

Azotobacter vinelandii, Science. 257: 165 3- 1659.

Gerdes, K., and S. Molin. 1986. Partitioning of plasmid RI: Structural and

functional analysis of the parA locus. J. Mol. Biol. lgO:269-279.

Gerdes, K., L. K. Poulson, T. Thisted, A. K. Nielsen, J. Martinussen, and P. Ho

Andreasen. 1990. The hok killer gene family in gram-negative bacteria. New

Biologist . 2:946-956.

Geriitz, M., O. Hrabak, and Ho Schwab. 1990. Partitioning of broad-host-range

plasmid RP4 is a cornplex system involving site-specific recombination. J. Bacteriol.

172:6194-6203.

Gerlt, J. A., J. A. Coderre, and S. Mehdi. 1983. Oxygen chiral phosphate esters.

Adv. Enzymol. 55:291-380.

Gidoni, Do, A. Scheiler, B. Barnet, P. Hantzopouios, M. Oren, and C. Rives.

1982. Different foms of simian virus 40 large tumor antigen varying in their

aff i t ies for DNA. J. Virol. 42:456-466.

Graiia, J. Do 1990. Promoter recognition and mRNA initiation by Escherichia coli

Eu70. Meth. Enzymol. 185:37-54.

Green, Eo W., and M. Schaechter. 1972. The mode of segregation of the bacterial

ce11 membrane. P m . Natl. Acad. Sci. USA. 69% 12-23 16.

Gustafsson, P., H. Wolf-Watz, L. Lind, K.-Eo Johansson, and K. Nordstrom.

1983. Binding between the par region of plasmids R1 and pSClOl and the outer

membrane fraction of the host bacteria. EMBO J. 2:27-32.

l29

Hayakawa, Y., T. Murotsu, and K. Matsubara. 1985. Mini-F protein that binds to

a unique region for partition of plasmid DNA. J. Bacteriol. 163:349-354.

Hayes, F., and S. Austin. 1994. Topological scanning of the P 1 plasmid partition

site. J. Mol. Biol. 243:19&198.

Hayes, F., and S. J. Austin. 1993. Specificity detenniDants of the P l and W

plasmid centromere analogs. Proc. Natl. Acad. Sci. USA. 90:9228-9232.

Hayes, Fe, M. A. Davis, and S. J. Austin. 1993. Fine-structure analysis of the W

plasmid partition site. J. Bacteriol. 175(1 l):3443-345 1.

Hayes, F., L. Radnedge, M. A. Davis, and S. J. Austin. 1994. The homologous

operons for Pl and P7 plasmid partition are autoregulated fiom dissimila operator

sites. Mol. Microbiol. 11:249-260.

IIiraga, S. 1992. Chromosome and plasmid partition in Escherichia coli. Annu. Rev.

Biochem. 61:283-306.

Hiraga, S., H. Nüci, T. Ogura, C. Ichinose, H. Mori, B. Ezaki, and A. Jaffé.

1989. Chromosome partitionhg in Escherichia coli: Novel mutants producing

anucleate cells. J. Bacteriol. 171:1496-1505.

Hiraga, S., T. Ogura, H. Ni, C. Ichinose, and H. Mori. 1990. Positioning of

replicated chromosomes in Escherichia coli. J. Bacteriol. l72:3 1-39.

Hirose, K., A. Lockhart, R. A. Cross, and L. A. Amos. 1995.

Nucleotide-dependent angular change in kinesin rnotor domain bound to tubulin.

Nature. 376:277-279.

Hu, K. H., E. Liu, K. Dean, M. Gringas, W. DeGraff, and N. Je Trun. 1996.

Overproduction of three genes leads to camphor resistance and chromosome

condensation in Eschenchia coli. Genet. 143: 152 1- 1532.

Hummeil, J. P., and W. J. Dreyer. 1962. Measurement of protein-binding

phenornena by gel filtration. Biochim. Biophys. Acta. 63530-532.

Ikeda, H., and J.4. Tomizawa. 1968. Prophage P 1, an extrachromosomal

replication unit. Cold Spring Harbor Symp . Quant. Bi01 . 33:79 1-798.

ireton, K., N. W. Gunther N, and A. D. Grossman. 1994. spdlJ is required for

normal chromosome segregation as well as the initiation of sporulation in Bacillus

subrilus. J. Bacteriol. 1765320-5329-

Jacob, F., S. Brenner, and F. Cuzin. 1963. On the regulation of DNA replication

in bactena. Cold S p ~ g Harbor Symp. Quant. Biol. 28:329-336.

Jaffé, A., R. D'Ari, and S. EIiraga. 1988. MiniceIl-forrning mutants of Escherichia

d i : production of minicells and anucleate rods. J. Bacteriol. 170:3094-3 101.

Jaffé, A., T. Ogura, and S. Riraga. 1985. Effects of the ccd function of the F

plasmid on bacterial growth. J. Bacteriol. 163:841-849.

Jagura-Burdzy, G., J. P. Ibbotson, and C. M. Thomas. 1991. The korF region of

broad-host-range plasmid RK2 encodes two polypeptides with transcriptional repressor

activity . J. Bactenol. 173:826-833.

Jagura-Burdzy, G., and C. M. Thomas. 1992. @kA gene of broad host range

plasmid RK2 encodes a novel DNA-binding protein. J. Mol. Bi01 . 225: 65 1-660.

Jensen, R. B., M. Dam, and K. Gerdes. 1994. Partitionhg of plasmid RI. The

parA operon is autoregulated by ParR and its transcription is highly stimulated by a

downstream activating element. J. Mol. Biol. 236: 1299-1309.

Jensen, R. B., and K. Gerdes. 1995. Prograrnmed ce11 death in bacteria: proteic

plasrnid stabilization systerns. Mol. Microbiol. 17:205-210.

131

Jurnak, F. 1985. Structure of the GDP domain of EF-Tu and location of the amino

acids homologous to ras oncogene proteins. Science. 230:32-36.

Karisson, R., A. Michaelsson, and L. Mattsson. 1991. Kinetic analysis of

monoclonai antibody-antigen interactions with a new biosensor based analytical

system. J. Immunolog . Meth. 145229-240.

Kato, J., Y. Nishimura, R. Lmamura, H. Ni, S. Hiraga, and H. Suzuki. 1990.

New topoisornerase essential for chromosome segregation in E. coli. Cell.

63:393-404.

Kerrigan, L. A., and J. T. Kadonaga. 1992. Periodic binding of individual core

histones to DNA: inadvertant purification of the core histone H2B as a putative

enhancer-binding factor. Nucleic Acids Res. 20:6673-6680.

Kieid, Dm G., O. Yansura, B. Srnail, D. Dowbenko, D. M. Moore, M. J.

Grubman, P. D. McKercher, D. O. Morgan, B. H. Robertson, and H. L.

Bachrach. 1979. Cloned viral protein for foot-and-mouth disease: Responses in cattle

and swine. Science. 214: 1125-1 129.

Koerner, T. J., J. E. Hill, A. M. Myers, and A. Tzagoloff. 1991. High-expression

vectors with multiple cloning sites for construction of trpE hision genes: pATH

vectors. Meth. Enzymol. 194:477-490.

Koonin, E. 1993. A superfamily of ATPases with diverse functions containhg either

classical or deviant ATP-binding motif. J. Mol. Biol. 229: 1 165- 1 174.

Korangy, F., and D. A. Julin. 1992. Aiteration by site-directed mutagenesis of the

conserved lysine residue of the ATP-binding consensus sequence of the RecD subunit

of the Escherichia coli RecBCD enzyme. J . Biol. Chem. 267: 1727-1732.

132

Kreuzer, K. N., and N. R. Cozzareili. 1980. Formation and resolution of DNA

catenanes by DNA gyrase. Cell. 20:245-254.

Kuempel, P. L., J. M. Henson, L. Dircks, M. Tecklenburg, and D. F. Lim.

1991. dif. a recA-independent recombioation site in the terminus region of the

chromosome of Eschenchia coli. New Biologist. 3:799-811.

Kusukawa, N., H. Mori, A. Kondo, and S. Hiraga. 1987. Partitioning of the F

plasmid: overproduction of an essential protein for partition inhibits plasmid

maintenance. Mol. Gen. Genet. 208:365-372.

Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the

head of bacteriophage T4. Nature. 227:680-685.

Lane, D., R. Rothenbuehler, A.-M. Mermat, and C. Aiken. 1987. Analysis of

the F plasmid centromere. Mol. Gen. Genet. 207:406-4 12.

Lehnherr, Hm, E. Maguin, S. Jafri, and M. B. Yarmolinsky. 1993. Plasmid

addiction genes of bacteriophage P l : doc, which causes death on curing of prophage,

and phd, which prevents host death when prophage is retained. J. Mol. Biol.

233:4 14-428.

Lehnherr, H., and M. B. Yarmolhky. 1995. Addiction protein Phd of plasmid

prophage Pl is a substrate of the ClpXP se&e protease of Escherichia coli. Proc.

Nat1 . Acad. Sci. USA. 92:3274-3277.

Leslie, N. R., and D. J. Sherratt. 1995. Site-specific recornbination in the

replication terminus region of Escherichia coli: hinctional replacement of dif. EMBO

J. 14:1561-1570.

Lin, Z., and L. P. Maliavia. 1994. Identification of a partition region carried by the

plasmid QpHl of Coxella bumeîîi. Mol. Microbiol. 13:s 13-523.

133

1 10. Lobocka, M., and M. Yarmolinsky. 1996. P l plasmid partition: A mutational

analysis of ParB. J. Mol. Biol. 259:366-382.

111. Lorimer, H. E., E. H. Wang, and C. Rives. 1991. The DNA-binding properties of

polyomavirus large T antigen altered by ATP and other nucleotides. J. Virol.

65:687-699.

112. Ludtke, D. N., B. G. Eichorn, and S. Je Austin. 1989. Plasmid-partition functions

of the P7 prophage. J. Mol. Biol. 209:393-406.

113. Lynch, A. S. , and J. C. Wang. 1993. Anchoring of DNA to the bacterial

cytoplasmic membrane through cotranscriptional synthesis of polypeptides encoding

membrane proteins or proteins for expoa: a mechanisrn of plasmid hypernegative

supercoiling in mutants deficient in DNA topoisornerase 1. J. Bacteriol.

175: 1645-1655.

114. Lynch, A. S., and J. C. Wang. 1995. SopB protein-mediated silencing of genes

linked to the sopC locus of Escherichia coli F plasmid. Proc. Nat. Acad. Sci. USA.

92: 1896-1900.

115. Lynch, A. S., and J. Cm Wang. 1994. Use of an inducible site-specific recombinase

to probe the structure of protein-DNA complexes involved in F plasrnid partition in

Escherichia d i . J. Mol. Biol. 236:679-684.

116. Marians, K. Je 1987. DNA gyrase-catalyzed decatentation of multiply linked DNA

dirners. J. Biol. Chem. 262: 10362-10368.

1 17. Martin, K. A., M. A. Davis, and S. Austin. 199 1. Fine-structure analysis of the P 1

plasmid partition site. J. Bacteriol. 173: 3630-3634.

118. Martin, K. A., S. A. Friedman, and S. J. Austin. 1987. Partition site of the P l

plasmid. Proc. Natl. Acad. Sci. USA. 84:8544-8547.

134

119. Masuda, Y., K. Miyakawa, Y. Nishimwa, and E. Ohtsubo. 1993. chpA and chpB,

Ercherichia coli chromosomal homologs of the pem locus responsible for stable

maintenance of R100, I. Bacteriol. l75:6850-6856.

120. Maxam, A. M., and W. Gilbert. 1977. A new method for sequencing DNA. Proc.

Nati. Acad. Sci. USA. 74560-565.

121. McEntee, K., G. M. Weinstock, and 1. R o Lehman. 1981. Bhding of the recA

protein of Escherichia coli to single- and double-stranded DNA. J. Biol. Chem.

256:8835-8844.

122. MeEntee, K., G. M. Weinstock, and 1. R. Lehman. 1979. Initiation of general

recombination catalyzed in vitro by the red protein of Escherichia coli. Proc. Natl.

Acad. Sci. 76:2615-2619.

123. Meacock, P. A., and S. N. Cohen. 1980. Partitioning of bacterial plasmids during

ce11 division: a cis-acting locus that accomplishes stable plasrnid inheritance. Cell.

20:529-542.

124. Miki, T., 2.-T. Chang, and T. Horiuchi. 1984. Control of ce11 division by sex

factor F in Escherichia coli II . Identification of genes for inhibitor protein and trigger

protein on the 42.84-43 -6 F fragment. J. Mol. Biol. 174:627-646.

125. Miki, T., J. A. Park, Ko Nagao, N. Murayama, and T. Horiuchi. 1992. Control

of segregation of chromosomal DNA by sex factor F in Escherichia coli: Mutants of

DNA gyrase subunit A supress letD (cc@) product growth inhibition. J. Mol. Biol.

225:39-52.

126. Miki, Tm, K. Yoshioka, and T. Horiuchi. 1984. Control of ce11 division by sex

factor F in Escherichia coli 1. The 42.84-43.6 F segment couples ce11 division of the

host bactena with replication of plasmid DNA. 3. Mol. Biol. 174:605-625.

135

127. Milner-White, E. J., J. R. Coggios, and 1. A. Antoo. 1991. Evidence for an

ancestral core structure in nucleotide-binding proteins with the type-A motif. J. Mol.

Biol. 221:751-754.

128. Mimura, C. S., S. R. Holbrook, and G. F.-L. Ames. 1991. Structura1 mode1 of the

nucleotide-binding consemed component of periplasmic permeases. Proc . Natl. Acad .

Sci. USA. 88:84-88.

129. Min, Y.-N., A. Tabuchi, D. D. Womble, and R. H. Rownd. 1991. Transcription

of the stability operon of IncFII plasmid NRl. J. Bactenol. 173:2378-2384.

130. Moodie, S. L., J. B. O. Mitchell, and J. M. Thorton. 1996. Protein recognition of

adenylate: An example of a fuzzy recognition template. J. Mol. Biol. 263:486-500.

131. Mon, a., A. Kondo, A. Ohshima, T. Ogura, and S. Hiraga. 1986. Stnicture and

function of the F plasmid partitioning genes essential for partitioning. J. Mol. Biol.

192: 1-15.

132. Mori, H., Y. Mori, C. Ichinose, H. Niki, T. Ogura, A. Kato, and S. Hiraga.

1989. Purification and characterization of SopA and SopB proteins essential for F

plasmid partitioning. J. Biol. Chem. 264:15535-15541.

133. Motaliebi-Veshareh, M., D o A. Rouch, and C. M. Thomas. 1990. A family of

ATPases involved in active partitioning of diverse bacterial plasmids. Molec .

Microbiol. 4: 1455- 1463.

134. Mukhopadhyay, G., S. Sozhamannan, and D. K. Chattoraj. 1994. Relaxation of

replication control in chaperone-independent intiator mutants of plasmid Pl. EMBO J.

13:2089-2096.

136

Mulder, E., M. EI'Bouhaii, E. Pas, and C. L. Woldringh. 1990. The Eschenchia

coli minB mutant resembles gyrB in defective nucleoid segregation and decreased

negative supercoiling of plasmids. Mol. Gen. Genet. 221:87-93.

Munin, A. G., S. E. Brenner, T. Hubbard, and C. Chothia. 1995. P-loop

containhg nucleoside triphosphate hy drolase fold and superfamiiy . 0 strand ordering

within the sheet. J. Molec. Biol. 247536440.

Myles, G. M., J. E. Hearst, and A. Sancar. 1991. Site-specific mutagenesis of

conserved residues within Walker A and B sequences of Escherichia coli UvrA

protein. Biochem. 30:3824-3834.

Naktinis, V., R. Onrust, L. Fang, and M. O'Donneli. 1995. Assembly of a

chrornosomal replication machine: Two DNA polymerases, a clamp loader, and

sliding clamps in one holoenzyme particle. J. Biol. Chem. 270: 13358-13365.

Niki, H., R. Imamura, M. Kitaoka, K. Yamanaka, T. Ogura, and S. Hiraga.

1992. E. coli MukB protein involved in chromosome partition forms a homodimer

with a rod-and-hinge structure having DNA binding and ATPIGTP binding activities.

EMBO J. 115101-5109.

Niki, A., A. Jaffé, R. Imamura, T. Ogura, and S. Hiraga. 1991. The new gene

mukB codes for a 177 kd protein with coiled-coi1 domains involved in chromosome

partitionhg of E. coli. EMBO J. 10: 183-193.

Noll, M. 1974. Internai structure of the chromatin subunit. Nucleic Acids Res.

1: 1573-1578.

Nordstriun, K. 1990. Control of plasrnid replication-how do DNA iterons set the

replication frequency? Cell. 63: 1121-1 124.

137

143. Nordstrom, K., and S. J. Austin. 1989. Mechanisms that contribute to the stable

segregation of plasmids. AMU. Rev. Genet. 23:37-69.

144. Nordstrom, K., S. Molin, and H. Aagaard-Hansen. 1980. Partitionhg of plasmid

R1 in Escherichia coli 1. Kinetics of loss of plasmid derivatives deleted for the par

region. Plasmid. 4:2 15-227.

145. Novick, R. P., R. C. Clowes, S. N. Cohen, R. Curtiss III, N. Datta, and S.

Fakow. 1976. Uniforni nomenclature for bacterial plasmids: A proposal. Bactenol.

Rev. 40:168-189.

146. Ogasawara, N., and A. Yoshikawa. 1992. Genes and their organization in the

replication ongin region of the bacterial chromosome. Mol. Microbiol . 6:629-634.

147. Ogawa, T., H. Wabiko, T. Tsurimoto, T. Horii, H. Masukata, and H. Ogawa.

1978. Characterization of purified recA protein and regulation of its synthesis in vivo.

Cold Spring Harbor S ymp. Quant. Bi01 . 43: 909-9 16.

148. Ogura, T., and S. Hkaga. 1983. Mini-F plasmid genes that couple host ce11 division

to plasmid proliferation. Proc. Natl. Acad. Sci. USA. 80:4784-4788.

149. Ogura, T., and S. Riraga. 1983. Partition mechanism of F plasmid: Two plasmid

gene-encoded products and a cis-acting region are involved in partition. Cell.

32:35 1-360.

150. Orr, E., N. F. Fairweather, 1. B. Hoiiand, and R. H. Pritchard. 1979. Isolation

and characterisation of a strain carrying a conditional lethal mutation in the cou gene

of Escherichia coli K12. Molec. Gen. Genet. 177: 103-1 12.

151. Pai, E. F., U. Krengel, G. A. Petsko, R. S. Goody, W. Kabsch, and A.

Wittinghofer. 1990. Refmed crystal structure of the triphosphate conformation of

138

H-ras p21 at 1.35 A resolution: implications for the mechanism of GTP hydrolysis.

EMBO J. 9:2351-2359.

Pal, S. K., and D. K. Chattoraj. 1988. Pl plasmid replication: Initiator

sequestration is inadequate to explain control by intitiator binding sites. J. Bactenol.

170:3554-3560.

Pal, S. K., R. J. Mason, and D. R. Chattoraj. 1986. Pl plasmid replication: Role

of initiator titration in copy number control. J. Molec. Biol. 192:275-285.

Peng, H., and K. J. Marians. 1993. Decatenation activity of topoisornerase IV

during oriC and pBR.322 DNA replication in vitro. Proc. Nati. Acad. Sci USA.

90t8571-8575.

Petenon, C. L. 1994. The SMC farnily : Novel proteins for chromosome

condensation? Celf . 79: 3 89-392.

Pittard, A. J., and B. E. Davidson. 199 1. TyrR protein of Escherichia coli and its

role as repressor and activator. Molec. Microbiol. 5: 1585-1592.

Radnedge, L., M. A. Davis, and S. J. Austin. 1996. P l and P7 plasmid partition:

ParB protein bound to its partition site makes a separate discriminator contact with the

DNA that detemines species specificity . EMBO J. 15: 1 155-1 162.

Roberts, J. W., C. W. Roberts, N. L. Craig, and E. M. Phizicky. 1978. Activity

of the Escherichia coli red-gene product. Cold Spring Harbor Symp. Quant. Biol.

43:917-920.

Roberts, R. C., R. Burioni, and D. R. Helinski. 1990. Genetic characterization of

the stabilizing functions of a region of broad-host-range plasmid RK2. J. Bacteriol.

l72:62O4-62 16.

Roberts, R. C., and D. R. Helioski. 1992. Defuition of a minimal plasmid

stabilization system from the broad-host range plasmid RK2. J. Bacteriol.

174:8119-8132.

Roberts, R. C., A. R. Strom, and D. R. Helinksi. 1994. The parDE operon of the

broad- host range plasmid RK2 specifies growth inhibition associated w ith plasrnid

loss. J. Molec. Biol. 237:35-51.

Rosner, J. L. 1972. Formation, induction and curing of bactenophage Pl lysogens.

Virology . 48:679-689.

Runzler, R., S. Thompson, and E. Fanning. 1987. Oligomerization and ongin

DNA-binding activity of Simian Virus 40 large T antigen. J. Virol. 61:2076-2083.

Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular Cloning: a

Laboratory Manual, Second ed. Cold Spring Harbor Laboratory , Cold Spring Harbor,

NY.

Saraste, M., P. R. Sibbald, and A. Wittinghofer. 1990. The P-loop- a common

motif in ATP- and GTP-binding proteins. Trends Biochem. Sci. 15430-434.

Scatchard, G. 1949. The attractions of proteins for small molecules and ions. An..

New York Acad. Sci. 513660-672.

Schneider, E., S. Wilken, and R. Scbmid. 1994. Nucleotide-induced conformational

changes of MalK, a bacterial ATP binding cassette transporter protein. J. Biol. Chem.

269: 2O456-2O46 1.

Seefeldt, L. C., T. V. Morgan, D. R. Dean, and L. E. Morteason. 1992. Mapping

the site(s) of MgATP and MgADP interaction with the nitrogeaise of Azotobacter

vinlandii. J. Biol. Chem. 267:668O-6688.

140

169. Sia, E. A., R. Cm Roberts, C. Easter, D. R. Helinski, and D. H. Figurski. 1995.

Different relative importances of the par operons and the effect of conjugal transfer

on the maintenance of intact promiscuous plasmid RK2. J. Bacteriol. l77:2789-2797.

Skarstad, K., and E. Boye. 1993. Degradation of individual chromosomes in recA

mutants of Escherichia coli. J. Bacteriol. 174: 5505-5509.

Skarstad, K., and E. Boye. 1988. Perturbed chromosomal replication in recA

mutants of Escherichia c d . J. Bacteriol. 170:2549-2554.

Smith, C. A., and C. M. Thomas. 1984. Molecular genetic analysis of the r fB and

korB region of broad-host range plasmid RK2. J. Gen. Microbiol. 130: 165 1-1663.

Smith, D. B., and K. S. Johnson. 1988. Single-step purification of polypeptides

expressed in Escherichia coli as fusions with glutathione-S-tramferase. Gene.

67:3 l-4O.

Sobecky, P. A., Cm L. Easter, P. D. Bear, and D. R. Helinski. 1996.

Characterization of the stable maintenance properties of the par region of

broad-host-range plasmid RK2. J. Bactenol. 178:2086-2093.

Steck, T. R., and K. Drlica. 1984. Bacterial chromosome segregation: evidence for

DNA gyrase involvement in decatenation. Cell. 36:1081-1088.

Sternberg, N., and D. Hamilton. 198 1. Bacteriophage P 1 site-specific recombination

1. Recombination between loxP sites. J. Mol. Biol, lS0:467-486.

Story, R. M., and T. A. Steitz. 1992. Structure of the recA protein-ADP complex.

Nature. 3SS:374-376.

Strunnikov, A. V., V. L. Larionov, and D. Koshland. 1993. SMCI: An essentid

yeast gene encoding a putative head-rod-tail protein is required for nuclear division

and defmes a new ubiquitous protein family. J. Ce11 Biol. 123:1635-1648.

Studier, F. W., and B. A. Moffatt. 1986. Use of bactenophage Ti RNA

polymerase to direct selective high level expression of cloned genes. J. Mol. Biol.

189: 113-130.

Summers, D. K., and Do J. Sherratt. 1984. Multimerization of high copy number

plasmids causes instability: ColEl encodes a detenninant essential for plasmid

rnonomerization and stability . Cell. 36: lO97- 1 103.

Surtees, JO A. 1996. MSc. University of Toronto.

Tabuchi, A., Y. N. Min, D. D. Womble, and R. H. Rownd. 1992. Autoregdation

of the stability operon of IncFII plasmid NR1. J. Bacteriol. 174:7629-7634.

Taghavi, S., A. Provoost, M. Mergeay, and D. van der Leiie. 1996. Identification

of a partition and replication region in the Alcaligenes eutruphus megaplasmid

pMOL28. Mol. Gen. Genet. 250: 169-179.

Tecklenburg, M., A. Naumer, O. Naggapan, and P. Kuempel. 1995. The dif

resolvase locus of the Escherichia coli chromosome can be replaced by a 33-bp

sequence, but function depends on location. Proc. Natl. Acad. Sci. USA.

92: 1352-1356.

Theophilus, B. D. M., and C. M. Thomas. 1987. Nucleotide sequence of the

transcriptional repressor gene korB which plays a key role in the regulation of the

copy number of broad host range plasmid W . Nucleic Acids Res. 157443-7450.

Thomas, C. M. 1986. Evidence for the involvement of the incC locus of broad host

range plasmid RK2 in plasrnid maintenance. Plasmid. 16: 15-29.

Tisa, L. S., and B. P. Rosen. 1990. Molecular characterization of an anion pump.

ArsB protein is the membrane anchor for the ArsA protein. J. Biol. Chem.

265: 190-194.

142

188. Treptow, Na, R. Rosenfeld, and M. YarmolinsJcy. 1994. Partition of nonreplicating

DNA by the par system of bacteriophage Pl. J. Bacteriol. 176: 1782-1786.

189. Tsuchimoto, S., H. Ohtsubo, and E. Ohtsubo. 1988. Two genes, p e M and peml,

responsible for stable maintenance of resistance plasmid R1 00. J. Bacteriol.

170: 1461-1466.

190. UUsperger, C., and N. R. Cozzareiii. 1996. Contrasting activities of topoisornerase

IV and DNA gyrase frorn Escherichia coli. J. Biol. Chem. 27l:3 1549-3 1555.

191. van der Woik, J., M. Klose, E. Breukink, R. A. Demel, B. de Kruijff, R.

Freudl, and A. J. M. Driessen. 1993. Characterization of a Baciïïw subrilu~ SecA

mutant protein deficient in translocation ATPase and release from the membrane.

Mol. Microbiol. 8:3 1-42.

192. van Helvoort, J. M. L. M., and C. L. Woldringh. 1994. Nucleoid partitionhg in

Escherichia coli during steady-state growth and upon recovery from chlorarnphenicol

treatment. Molec . Microbiol. 13577-583.

193. Walker, J. E., M. Saraste, M. J. Runswick, and N. J. Gay. 1982. Distantly

related sequences in the alpha and beta-subunits of ATP synthase, myosin, kinases

and other ATP-requiring enzymes and a common nucleotide binding fold. EMBO J.

l:945-95 1.

194. Watanabe, E., S. tuamoto, M.-H. Lee, S. U. Kim, T. Ogura, H. Mori, S.

Hiraga, M. Yamasaki, and K. Nagai. 1989. Purification and characterization of the

sopB gene product which is responsible for stable maintenance of mini-F plasmid.

Mol. Gen. Genet. 2l8:43 1-436.

143

195. Watanabe, E., M. Wachi, M. Yamasaki, and K. Nagai. 1992. ATPase activity of

SopA, a protein essential for active partitioning of F plasmid. Mol Gen Genet.

234:346-352.

196. Williams, D. R., and C. M. Thomas. 1992. Active partitioning of bacterial

plasmids. J. Gen. Microbiol. 138: 1-16.

197. Wolle, D., D. R. Dean, and J. B. Howard. 1992. Nucleotide-iron-sulfur cluster

signal transduction in the nitrogenase iron-protein: The role of Asp'? Science.

258:992-995.

198. Yamanaka, K., T. Mitani, T. Ogura, H. Niki, and S. Hiraga. 1994. Cloning,

sequencing and characterization of multicopy suppressors of a mukS mutation in

Escherichia coli. Mol. Microbiol. IWO 1-3 12.

199. Yamanaka, K., T. Ogura, 8. Niki, and S. Hiraga. 1996. Identification of two new

genes, mukE and mukF, involved in chromosome partition in Escherichia coli. Mol .

Gen. Genet. 2SO:241-25 1.

200. Yang, J., S. Ganesan, J. Sarsero, and A. J. Pittard. 1993. A genetic analysis of

various functions of the TyrR protein of Escherichia coli. J. Bacteriol.

175: 1767-1776.

201. Young, C., R. S. Burlage, and D. H. Figurski. 1987. Control of kiLl gene of the

broad-host-range plasmid RK2: Involvement of korA, korB and a new gene, korE. J.

Bacteriol. 169: 13 15- 1320.

202. Zethiedrïch, E. L., and N. R. Cozzarelli. 1995. Roles of topoisornerase IV and

DNA gyrase in DNA unlinking during replication in Eschekhia coli. Genes Dev.

9:2859-2869.

144

203. Zyskind, J. W., A. L. Svitil, W. B. Stine, M. C. Biery, and D. W. Smith. 1992.

RecA protein of Escherichia coli and chromosome partitioning. Mol. Microbiol.

6:2525-2537.

IMAGE EVALUATION TEST TARGET (QA-3)

APPLiEi IMAGE . lnc 1653 East Main Street

Q 1993. Applied Image. lm.. Ail Riqhts ReserWd