INFORMATION TO USERS

This manuscript has been reproduced from the microfihn master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.

The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.

In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.

Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6” x 9” black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order. UMI A Bell & Howell Information Company 300 North Zeeb Road, Ann Arbor MI 48106-1346 USA 313/761-4700 800/521-0600

STARVATION-INDUCED CHANGES IN MOTILITY AND SPONTANEOUS SWITCHING TO FASTER SWARMING BEHAVIOR OF SINORHIZOBIUM MELILOTI

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy in the Graduate School at The Ohio State University

By

Xueming Wei. B.S., M.S.

Department of Plant Biology The Ohio State University

1999

Dissertation Committee; Approved by

Dr. W. Dietz Bauer, advisor Advisor Dr. David L Coplin

Dr. Richard T. Sayre, advisor Advisor Dr. Olli H. Tuovinen Graduate Program in Plant Biology UMI Number: 9931697

UMI Microform 9931697 Copyright 1999, by UMI Company. All rights reserved.

This microform edition is protected against unauthorized copying under Title 17, United States Code.

UMI 300 North Zeeb Road Ann Arbor, MI 48103 ABSTRACT

Changes in motility and chemotaxis in responses to nutrient limitation have been studied in Sinorhizobium meliloti. Cells lost motility in a strain- specific pattern within 8 h to 4 d after transfer to starvation buffer. A transient,

2- to 6-fold increase in chemotactic responsiveness toward attractants was observed. Nonmetabolizable chemoattractants could prevent motility loss and partially restore the motility of starving cells. Thus, interactions of chemoreceptors with attractants appear to affect motility independently of nutrient availability. Most nonmotile cells retained flagella, indicating that deactivation of flagellar motors was the first important response of S. meliloti to nutrient deprivation.

Several Tn5 mutants of S. meliloti which swarmed twice as faster as the parent in semi-solid agar, moist sand and viscous liquid were identified.

The faster swarming (FS) mutants outgrew the wild type 30- to 40-fold within 2 d in soft agar plates and had a significant growth advantage in all circumstances where nutrient gradients were present. The mutants had higher percentages of motile and flagellated cells, and longer and more flagella than the wild type. Thus FS behavior is likely a result of derepression of flagellar synthesis. Spontaneous variants that behaved exactly like the Tn5 FS mutants

ii were obtained at a frequency of about 1 per 15,000 cells. These FS variants

reverted to wild-type behavior at low, but variable frequencies. The FS

mutants and variants produced less exopolysaccharide (EPS) than the wild

type. Various swarm rates, swim patterns, EPS phenotypes, and restriction

patterns observed in different FS mutants indicate that multiple genetic

configurations could result in FS behavior. The significant competitive

advantage of the FS mutants over the wild type suggests that behavioral

switching may be an important adaptation in natural habitats. Preliminary

molecular characterization have shown that the Tn5 insertion sites in five of

the six FS mutants were all in the same location within about 50 bp. The

sequences flanking the Tn5 did not complement the mutant behavior, nor did

marker exchange of the Tn5 into the parent recreate the FS phenotype.

Complementing sequences have been isolated and are being characterized.

Possible molecular mechanisms for FS-EPS phenotype and for switching are

discussed, and future directions presented.

In many Gram-negative bacteria, expression of certain genes is

regulated by N-acyl-derivatives of homoserine lactone (AHL). We investigated the effect of mutations in AHL production and reception on growth and plant root colonization by Pseudomonas aeruginosa. Mutants defective in either

AHL synthesis or signal perception were found to grow more slowly than the wild type in both moist sand and on roots. Our results indicate that AHL signalling plays a significant role in growth at low cell densities, in root colonization and in the development of a stress resistant subpopulation.

iii ACKNOWLEDGMENTS

Million of thanks goes to my advisor. Dr. W. D. Bauer, for his guidance, patience, encouragement, and support, especially when the project encountered difficulties. I am greatly grateful to my advisory committee members, Drs. David L. Coplin. Richard T. Sayre, Olli H. Tuovinen. Fred Sack, and Ralph E. J. Boerner for their constructive ideas, advice on the project and beyond, and for their willingness to be my future references. The assistance of Catherine Wolkin, Elke Kretschmar and Robert Whitmoyer with electron microscope is greatly appreciated. I wish to thank Drs. Dave Coplin. John Leigh and Graham Walker for providing strains and plasmids, and Drs. Jyan-Chyun Jang, Jayne Robinson. Ms. Doris R. Majerczak and Ms. Debbie Estes for technical helps.

Ohio Agricultural Research and Development Center, and OSU Graduate School Alumni Research Award provided partial support for salaries and experimental supplies.

Chapters 2 and 3 were already published as; Wei. X.. and W. D. Bauer. 1998. Starvation-induced changes in motility, chemotaxis and flagellation of Rhizobium meliloti. Appl. Environ. Microbiol. 6 4 : 1708- 1714.; and as: Wei. X.. and W. D. Bauer. 1999. Tn5-induced and spontaneous switching of Sinorhizobium meliloti to faster-swarming behavior. Appl. Environ. Microbiol. 65:1228-1235.

This dissertation is dedicated to all my family members who have been supporting me with all their hearts and soul throughout all the ups and downs for these long years.

IV VITA

1981: B. S. degree; major: microbiology, minor: soil science and agrochemistry. Nanjing Agricultural University, Nanjing, and Department of Biology, Northwestern University, Xi'an, P. R. China.

1984: M. S. degree; major: soil and agricultural microbiology. Nanjing Agricultural University, P. R. China.

1985—1990: Lecturer, Laboratory of Agricultural Microbiology, Nanjing Agricultural University, P. R. China.

PUBLICATIONS

1. Wei, X., and W. D. Bauer. 1998. Starvation-induced changes in motility, chemotaxis, and flagellation of Rhizobium meliloti. Appl. Environ. Microbiol. 64:1708-1714.

2. Wei, X., and W. D. Bauer. 1999. Tn5-Induced and spontaneous switching of Sinorhizobium meliloti to faster swarming behavior. Appl. Environ. Microbiol. 65:1228-1235.

FIELD OF STUDY

Major: Plant Biology Areas of research and interests: Plant-microbe interactions Microbial ecology Molecular biology TABLE OF CONTENTS

ABSTRACT ...... ii ACKNOWLEDGEMENTS...... iv CURRICULUM VITA...... v LIST OF TABLES...... viii LIST OF FIGURES...... x LIST OF ABBREVIATIONS...... xii SCOPE OF THE STUDY ...... 1

CHAPTER PAGE

1 INTRODUCTION AND LITERATURE REVIEW ...... 3

I. Chemotactic motility as a way of response to environmental signals ...... 3 II. Bacterial locomotion ...... 6 1. Motility of flagellated species ...... 7 2. Motility in nonflagellated bacterial species ...... 13 III. Tactic responses in bacteria ...... 14 1. Types of bacterial taxis ...... 15 2. Signal transduction in E. coli chemotaxis ...... 17 IV. Role of motility and chemotaxis in bacterial fitne ss 21 V. S. meliloti and Its unique features regarding motility and chemotaxis ...... 25 VI. Goals of this study ...... 29

2 STARVATION-INDUCED CHANGES IN MOTILITY AND CHEMOTAXIS IN SINORHIZOBIUM MELILOTI ...... 31

Abstract ...... 31 Introduction ...... 32 Materials and Methods ...... 34 Results ...... 39 Discussion ...... 55 vl 3 TN5-INDUCED AND SPONTANEOUS SWITCHING TO FASTER SWARMING BEHAVIOR OF SINORHIZOBIUM MELILOTI RMB7201 ...... 63

Abstract ...... 63 Introduction ...... 65 Materials and Methods ...... 67 Results ...... 75 Discussion ...... 101

4 MOLECULAR CHARACTERIZATION OF FASTER SWARMING MUTANTS OF SINORHIZOBIUM MEL/LOT/ RMB7201 113

Abstract ...... 113 Introduction ...... 114 Materials and Methods ...... 117 Results ...... 125 Discussion ...... 147

5 EFFECT OF AUTOINDUCER MUTATION ON THE GROWTH, ADHESION AND ROOT COLONIZATION OF PSEUDOMONAS AEROGINOSA PAO^ ...... 156

Abstract ...... 156 Introduction ...... 157 Materials and Methods ...... 161 Results ...... 164 Discussion ...... 175

6 GENERAL DISCUSSION ...... 182

CONCLUDING REMARKS ...... 197

REFERENCES ...... 199

VII LIST OF TABLES

TABLE PAGE

2.1 Motility of S. meliloti L5-30 after transfer to SB containing nutrients, culture filtrate (CF) or non-metabolizable attractants 51

2-2 Restoration of motility to starved cells by addition of glucose, cycloleucine, or both ...... 52

3.1 Strains, phages and plasmids used in this study, and their characteristics...... 69

3.2 Swarm rates of S. meliloti RMB7201 wild type and FS mutants in various media ...... 80

3.3 Chemotaxis responsiveness of S. meliloti RMB7201 wild type and FS mutants to alfalfa seed exudates and nutrients ...... 81

3.4 Chemotaxis responsiveness of S. meliloti RMB7201 wild type and FS mutants to various attractants ...... 82

3.5 Competitive growth of S. meliloti RMB7201 wild type and faster swarming mutant FS1 under different culture conditions ...... 83

3.6 Survival of RMB7201 wild type and FS1 and FS7 during drying in silt-soil matrix ...... 85

3.7 Percentages of motile and flagellated cells, and flagellar number and length of S. meiiloti RMB7201 wild type and FS m utants after transfer to SB ...... 87

3.8 Tn5 FS mutants and spontaneous FS variants obtained by enrichment or screening of RMB7201 wild type ...... 94

3.9 Competitive growth of a 1:1 mixed inoculum of FS17, SV8 or SV69 with the RMB7201 wild ty p e ...... 100

viii 4.1 Strains, phage and plasmids used and constructed in this study, their characteristics and source of reference ...... 118

4.2 Primers used in PCR analysis and sequencing ...... 122

4.3 Primer combinations for PCR tested on selected templates ...... 137

4.4 Complementation of swarming behavior of FS1 and FS7 by cosmid clones ...... 141

4.5 Complementation of swarming behavior of FBI7, EV2 and EV10 by selected cosmid clones ...... 142

4.6 Complementation of swarm phenotype of FBI and FB44 by cosmid clones pME4 and pME2.2 ...... 144

4.7 Effect of the extragenic suppressor gene on the EPB production of S. me//7of/RMB7201, FBI and F B 4 4 ...... 145

5.1 Growth of PA01 (pGB3) (WT) and PAO-JP3 (1 ;1 mixed Inoculation) in 1/4 Hoagland solution ...... 167

5.2 Growth in and adhesion to moist sand of PA01, PA01(pGB3), and PAO-JP3 after 1 day incubation ...... 168

5.3 Growth and adhesion of P. aeruginosa PA01 and PA0-JP2 to sand particles after 1 day incubation ...... 169

5.4 Colonization of soybean roots by PA01 and P AO-J P2 six days after planting ...... 170

5.5 Effect of culture age on colony formation on PIA and LB of P. aerug/nosa PA01, PAO-JP2, and PAO-JP3 ...... 172

5.6 Colony formation of PA01, PAO-JP2, and PAO-JP3 on LB and NM m edia ...... 172

5.7 Effect of cell density on colony forming ability of P. aeruginosa PA01 and FRD1 on P IA ...... 174

IX LIST OF FIGURES

FIGURE PAGE

1.1 Structure of an enteric flagellum ...... 10

1.2 Bacteria! chemotaxis signal pathway ...... 19

1.3 Comparison of motility and chemotaxis of S. me///of/with E. coli...... 27

2.1 Long term survival of S. meliloti L-5-30, RMB7201 and JJ1 cl 0 in starvation buffer ...... 40

2.2 Changes in motility of S. meliloti L5-30, RMB7201 and JJ 1 cl 0 after transfer to starvation condition ...... 41

2.3 Changes in chemotactic responsiveness of S. meliloti L5-30, RMB7201 and JJIcIO after transfer to starvation buffer ...... 44

2.4 Chemotactic responsiveness of S. meliloti L-5-30 to different concentrations of cycloleucine before and after transfer to starvation buffer ...... 45

2.5 Retention of motility and flagella by S. meliloti L5-30 aftertransfertoSB ...... 49

2.6 Retention of motility of E coli strains after transfer to starvation condition ...... 61

3.1 Swarm colony morphology and EPS production on NM plate of S. meliloti RMB7201 wild type and FS mutants ...... 77

3.2 Relationship between nutrient concentration and swarm rates of S. meliloti RMB7201 wild type and mutant FS1 ...... 79 3.3 Flagellar number and length of S. meliloti RMB7201 wild type and mutant FS1 ...... 88

3.4 SDS-PAGE and quantitative Western analysis of flagellin produced by S. meliloti RMB7201 wild type and mutant FS1 ...... 89

3.5 Swarming rates of FS variants of P. aeruginosa...... 112

4.1 Southern analysis of Tn5 non-swarming reduced mutants ...... 126

4.2 Southern analysis of Tn5 reduced motility and FS mutants ...... 126

4.3 Southern analysis of Tn5 and miniTn5/acZFS mutants ...... 127

4.4 Nucleotide sequence of the 3255 bp EcoRl fragment ...... 130

4.5 Potential open reading frames of the 3255 bp EcoRl fragm ent 133

4.6 Presence of sequence homologous to the 3255 bp fragment in other species and S. melilotistrams...... 134

4.7 A schematic view of the primers designed from the 3255 bp sequence for PCR analysis ...... 137

4.8 pLAFR3-based subclones from the E coli 2.9 Kb EcoRV fragment and their complementing nature ...... 146

4.9 The 400 bp sequence from the 2.9 kbp EcoRVfragment ...... 154

5.1 Growth rates of P. aeruginosa PA01, PA0-JP2 andPAO-JP3inLB ...... 165

5.2 Effect of OdDHL on the growth rates of P. aeruginosa PA01 and PAO-JP2...... 166

XI LIST OF ABBREVIATIONS

AHL acyl (acylated) homoserine lactone

Amp amplcillin bp base pair (DNA)

BHL N-butanoyl-L-homoserine lactone

CB chemotaxis buffer (same as SB, starvation buffer)

CF culture filtrate

EOTA Ethylenediaminetetraacetic acid-Naa

EPS exopolysaccharide, extracellular polysaccharide

EV Enriched Variant (FS variants obtained through enrichment)

FS faster swarming

FS-EPS faster swarming-reduced EPS phenotype

HEP ES yV-2-hydroxyethylpiperazine-/V'-2-ethanesulfonic acid

HHL N-hexanoyl-L-homoserine lactone

HSL homoserine lactone

Gm gentamicin

IS insertion sequence

kbp kilo base pair (1,000 bp)

xii Km kanamycin

LB Lauria-Bertani medium

NM New Medium: defined medium suitable for S. meliloti

OdDHL N-(3-oxododecanoyl)-L-homoserine lactone

OHHL A/-(3-oxohexanoyl)-L-homoserine lactone

ORF open reading frame

PC phenotype conversion

FIA Pseudomonas Isolation Agar

RT root tip

SB starvation buffer (also OB)

Sm streptomycin

Spc (Spe) spectinomycin

SV spontaneous FS variant

Te tetracycline

Tp trimethoprim

TY tryptone-yeast extract medium

WT (wt) wild type, wild-type

XIII SCOPE OF THE STUDY

This study includes three sections which address i) the behavioral responses to starvation conditions, ii) switching from normal swarming to faster swarming behavior in Sinorhizobium meliloti, and iii) the effect of quorum sensing on growth of Pseudomonas aeruginosa.

CHAPTER 1 provides a review of bacterial locomotion and chemotaxis, a description of the unique features of S. meliloti, and a brief discussion on quorum-sensing in P. aeruginosa, followed by the specific goals of this dissertation project.

CHAPTER 2 is concerned with the behavioral responses of S. meliloti to nutrient limitation. The basic questions were whether S. m eliloti cells, upon exposure to starvation conditions, would sustain their chemotactic motility to actively search for additional nutrients, or shut off motility to conserve substrates for long term survival, and what mechanisms regulate the behavioral responses of starving cells.

In CHAPTERS 3 and 4, we were trying to understand the regulation of swarming motility of S. m eliloti. We obtained mutants that swarmed faster than the parent, and discovered a switching between the wild-type and faster- swarming (FS) phenotypes in S. meliloti. Phenotypic characterizations of the

FS behavior are presented in CHAPTER 3, and preliminary molecular studies on the possible mechanisms for the FS behavior in CHAPTER 4.

CHAPTER 5 addresses the possible role of quorum sensing on the growth, adhesion to solid surfaces, root colonization, and culturability of P. aeruginosa PA01. Work presented in this dissertation, especially in

CHAPTERS 4 and 5, have certainly opened new areas for future studies.

Some of them are discussed in CHAPTER 6, a general discussion, and in the

Concluding Remarks. CHAPTER 1

INTRODUCTION

I. Chemotactic motility as a way of response to environmental signals

Flagella are the second largest subcellular entity in bacteria. Over 50 genes are required for the biosynthesis of flagella and the functioning of motility and taxis in bacteria (e.g. Macnab. 1996; Sourjik et al. 1998). It is logical to assume that motility and taxis must contribute significantly to the overall fitness of bacteria in natural environments. To understand how tactic motility contributes to bacterial competitiveness is of great importance to microbial ecology. Such understanding may have application in the generation of more efficient inoculum strains (e. g. rhizobia and biocontrol agents), or in the bioremediation of toxic or waste compounds.

Motility, when accompanied by taxis, allows bacteria to respond to environmental conditions in a way favorable to the cells ( Adler et al., 1973;

Macnab, 1987). Potential benefits of motility and taxis include enhanced access to nutrients and migration to favorable colonization sites and hosts.

Tactic motility also allows bacteria to avoid adverse conditions such as toxic substances, or avoid desiccation by moving to water-filled pores in soils.

Motility may also provide ways for bacteria to interact with other species

(symbiosis, commensalism or antagonism) for coordinated activities in an ecosystem. Finally, tactic motility can assist bacterial cells in finding mates

(genetic exchange), and in dispersal in the environment so that a population has a t)etter chance to perpetuate.

However, motility and taxis are costly processes. In addition to synthesis and assembly of flagellar filaments, flagellar motors, and chemotaxis signal transduction components, flagellar operation poses a substantial energetic burden on the cell: flagella rotate at rates up to 15,000 rpm and each revolution consumes up to 1,000 protons (Armitage, 1992).

Flagella are also targets for the immune system of host animals and for bacteriophage attack. Consequently, it is a crucial choice for survival whether or not cells produce flagella and keep flagellar motors active in response to specific environmental conditions. Flagellar synthesis and operation are likely to be under complex and multiple-level regulatory controls, both genetic and physiological.

In most natural aquatic and terrestrial environments, nutrients are sporadically available or in very low concentration. In such habitats, most bacteria are in a starvation-survival state (Morita, 1993). Soil is an environment in which nutrients are highly localized and seasonally fluctuating, depending on plant growth, moisture levels, the proximity of plant roots, and the activities of other soil organisms. Thus soil and rhizosphere bacteria are in a feast-famine state, being starved most of the time. Soil- rhizosphere bacteria may occasionally have a surge of nutrient supply when there are passing plant roots and decaying litter or organisms. When nutrient supplies are depleting, motile bacteria face a critical choice, either to sustain motility and continue their search for nutrients, or turn off motility to save substrates for other adaptations to starvation and long term survival. Either of the choices has its advantages and disadvantages. In Chapter 2, 1 will describe the studies on the changes in motility and chemotaxis behavior of

Sinorhizobium meliloti (formerly Rhizobium meliloti) upon exposure to nutrient-limiting conditions. We have found that bacteria are “smarter” than having just off-on choices. They have subpopulation labor division. One subpopulation continues to search for nutrients, while others shut off motility to conserve energy. Furthermore, different strains or species adopt distinct strategies in dealing with nutrient-limiting conditions.

A second important area of uncertainty is the role of chemotactic motility in competition, dispersal and survival of bacteria in natural environments. If motility is important but energy-consuming, then possessing different degrees of motility and expressing them according to environmental signals will have adaptive value for bacteria. In other words, cells would balance the energy cost with the need for motility under any given condition.

There is evidence that tactically motile strains can outcompete nonmotile ones when there is a gradient of nutrients (Pilgram and Williams, 1976). But whether motile strains have variable degrees of motility and hence different competitiveness is not known. As will be discussed in CHAPTER 2, we have

found that motility is generally downregulated when S. meliloti cells face

nutrient deprivation. In CHAPTER 3, we will describe our discovery that

certain cells in a population can upregulate their motility spontaneously, i. e.,

switch to faster swarming (FS) behavior. FS mutants and variants spread 1.5

to 2 times faster than the wild type in various media and matrices. The

increased swarming ability gives FS mutants and variants a great growth

advantage over the parent when a nutrient gradient is present. In CHAPTER

4, results of preliminary molecular studies on mechanisms for FS behavior

and switching will be presented.

II. Bacterial locomotion

Motile bacterial species are widely distributed among different taxonomic groups, including heterotrophic, chemoautotrophic, and

photosynthetic families, and in archaea. There are motile species in rod­ shaped and spirillar groups, in Gram-positive and Gram-negative, aerobic

and anaerobic, as well as spore-forming and nonspore-forming taxons. The mechanisms for locomotion are also varied with groups. Bacterial motility

can be driven by flagella, or by pill, or by unidentified cell surface components. Motility, together with taxis, allows bacteria to move directionally and coordinately in response to environmental signals. 1. Motility of flagellated species

Fiagellation. Flagellar motility is by far the most common mode of

locomotion among bacteria, and has been studied most extensively (for

reviews, see Macnab, 1996; Stock and Surette, 1996). There are a number

of flagellation patterns in bacteria. Polar flagellation is observed in bacteria

such as pseudomonads (unipolar). Spirillum species (bipolar) and Vibrio

(sheathed monopolar). Flagellation in Escherichia coli, Salmonella

typhimurium. Bacillus, Rhizobium, Sinorhizobium, and other species is

peritrichous.

Swimming motility. Bacteria with flagella as a locomotive

organelle display two common motility behaviors, swimming and swarming.

Swimming in liquid media is an individual cell act. Swarming is the

population level result of swimming. Flagellar swimming has been studied

mostly with E. coli and S. typhimurium as model systems (Macnab, 1996).

Their flagellar filaments are left-handed helices, and when rotating

counterclockwise (CCW), the flagella form a helical bundle. The rotation of

the bundle generates a thrust to push the cell fonward (Macnab, 1987). This

causes cells to swim smoothly forward in a straight line, usually for 1-2

seconds (Silverman and Simon, 1974). When flagella rotate clockwise

(CW), the flagellar bundle comes apart and thus cells tumble and reorient in

a new, random direction for the next swim. The path of a swimming cell is

therefore a three dimensional random walk. In the absence of chemoattractants, a population of bacteria randomly moving will spread slowly, similar to diffusion. Nonmotile bacteria will also spread by Brownian

motion, though much slower. In S. meliloti, flagella rotate only clockwise,

and changes of direction are achieved through pauses of flagellar rotation and consequent random reorientation of cells (Gotz and Schmitt, 1987). The

swimming speeds of flagellated bacteria range from 20 to 60 /ym/sec.

Swarming motility. Bacterial swarming motility is the movement or

migration of a bacterial population on a solid medium, or inside a matrix. The former is referred to as surface swarming. Proteus mirabilis (Williams et al.,

1976) is well known for surface swarming ability. Surface swarming causes a cell mass or colony to spread quickly on a moist surface. These bacterial species are able to swim in liquid as well. Their polar flagella are responsible for swimming and lateral flagella for surface swarming.

Therefore, cells with polar flagella are called swimmers and those with lateral flagella swarm ers. Swarmer cells ( e.g. Rhodospirillum centenum and

P. mirabilis) are commonly hyperflagellated with as many as 1000 flagella per cell (e.g. Eberl et al., 1996).

The migration or spreading of a bacterial population inside any matrix is also defined as swarming (Wolfe and Berg, 1989). We will tentatively call it matrix swarming. This is the type of swarming that is discussed in this dissertation. Laboratory swarming assays are routinely done in semisolid agar media. Matrix swarming differs from surface swarming in that it is driven by the same flagella that power swimming motility. Swarming rates are much slower than swimming speeds because cells change directions frequently to

8 circumvent obstacles in a matrix.

Bacterial flagellar structure and biosynthesis

The flagellar filament. Prokaryotic flagella are composed of a long filament, a hook, and a basal body embedded in the cell membrane and cell wall (DePamphilis and Adler. 1971; Macnab, 1996). FIG. 1.1 shows the structure of a typical enteric bacterial flagellum (from Armitage, 1992).

The long, helical filament in most species consists of only one protein repeating unit called flagellin. In Caulobacter crescentus (Dingwall et al.,

1990), Bacillus pumllis (Oiler et al.. 1971), rhizoblal spp. (Bergman et al..

1991) and archaebacterla Including Halobacterium halobium (Gerl and

Bumper, 1988; Southam et al., 1990), the flagellar filament consists of two or more different flagelllns. The rotation of the helical filaments provides hydrodynamic propulsion to push or pull a cell through media.

In E. coli and S. typhimurium, about 20,000 flagellin protein monomers make up a flagellar filament (Macnab, 1987). The molecular weights of E. coii and S. typhimurium flagelllns are about 55 kilodalton (kd)

(Macnab, 1996), while the flagelllns of Rhizobium leguminosarum and S. meliloti are around 43 kd (Robinson et al., 1992; Pleler and Schmitt, 1989 and 1991). The filament Is the major antigenic determinant of flagella, and of the cell as well.

Two general types of flagellar filaments have been described based on their structural characteristics: plain and complex (Joys, 1988). Species such as E coli and S. typhimurium have plain flagellar filaments, while a HAP2 (FliO) /

filament (FliC)

HAP3 (FigL) MAPI (FlgK)

hook (FIgE) exterior Lring (FIgH) Pring (FIgl) outer membrane peptidoglycan layer & periplasm (FIgB-FlgC WM cell membrane FIgF-FIgG) Mot complex / C W . M ring (FliF) cytoplasm (MolA-MotB) switch complex export apparatus (FliG-FliM-FliN) 20 nm

FIG. 1.1. A diagrammatic representation of the E. coli bacterial flagella In cell membrane. Gene names In brackets. From Armltage (1992).

10 number of common soil bacteria, including Pseudomonas rhodos (Schmitt et a!., 1974), Rhizobium lupini (Maruyama et al., 1978), S. meliloti (Gotzetal.,

1982; Robinson at al., 1992), Bradyrhizobium japonicum {Kape et a\., 1991), and Hyphomicrobium (Tuhela et al., 1998), are found to have complex flagella. Plain flagellar filaments are structurally uniform with smooth surface.

A single protein (FliC in E coli) assembles into the left-handed filament. In contrast, complex flagellar filaments are right-handed helices with alternating ridges and grooves (Krupski et al., 1985).

Flagellar hook, basal body, and flagellar motor. The hook connects the filament to the basal body of flagella. It is also composed of a protein which self assembles into a tubular structure. Unlike the filament, the hook usually has a rather short, defined length. It is curved and rigid yet flexible (Block et al., 1989 & 1991). The presumed function of the hook is to transfer torque generated in the motor to the filament (Macnab and Aizawa,

1984). The hook probably plays a role in flagellar bundle formation

(Macnab, 1996). The basal body of flagella is a multiunit complex embedded in the cell wall and plasma membrane (Depamphilis and Adler, 1971). The basal body is composed of a central rod and four rings. These rings act as a mounting plate for the motor/switch proteins, and as an anchor for the rod/flagellum to rotate in, but they are not directly related to torque generation (Macnab, 1996). The rod transfers torque from a flagellar motor to the external filament. The motor/switch proteins produce torque and control the direction of flagellar rotation.

11 Regulation of flagellar motility. Since motility and chemotaxis are potentially expensive, sophisticated and coordinated controls are expected over tactic motility for better adaptation of bacteria to various environmental conditions. One condition affecting motility is growth phase or culture age. For example, in E. coli batch cultures, swimming motility (swim speed) was low in early-exponential phase, reached maximum in post- exponential phase, and decreased thereafter (Amsler et al., 1993). Flagellin synthesis followed a similar pattern. The authors speculated that the patterns reflected the enhanced role of motility as cells entered post-exponential phase and the decreased energy reserves available during later-stationary phase. E. coli cells grown in adverse growth conditions were unable to produce flagella (Li et al., 1993). High temperature, high concentrations of salts, carbohydrates and alcohols, or the presence of gyrase inhibitors have been found to inhibit motility due to the reduced flagellin synthesis (e.g.

Adler and Templeton, 1967). Mutants that were motile under harsh conditions were isolated, indicating that the wild-type strain could actively repress motility in unfavorable environments. Austin and Austin (1990) reported that Aeromonas media, a species described as nonmotile in taxonomy texts, displayed motility in old cultures after subculturing. This finding implies that some species might not express motility under common culture conditions where there are excess nutrients.

Lateral flagellar production in species such as Proteus mirabilis

(Williams et al., 1976) is another example of gene expression regulated by

12 growth conditions. Surface contact Induces the development of numerous lateral flagella, which are responsible for surface swarming. Dufour et al.

(1998) identified four genes that can upregulate the expression of the flhDC master operon, resulting In the differentiation Into elongated, hyperflagellated swarm cells. Caulobacter spp. represents another type of flagellation, and flagellar synthesis Is closely associated with the cell cycle

(Dingwall et al., 1990). Results from the aforementioned Investigations indicate that the expression of motility is quite flexible, and multiple control mechanisms may exist. Our FS mutants and variants are another type of regulation of bacterial flagellation and motility.

2. Motility in nonflagellated bacterial species

Bacterial motility is not restricted to flagellated species. In nonflagellated species, gliding movement, twitching and swimming motility have been demonstrated. Gliding movement is seen In a number of bacteria such as Cytophaga (e. g. Pate, 1988). Gliding motility requires moist surfaces, but gliding bacteria are Incapable of swimming In liquid (Burchard,

1981). Pate and De Jong (1990) suggested that special membrane proteins were required for such movement. High molecular weight polysaccharides may also play a role In It (Godchaux et al. 1991). Myxococcus xanthus. another typical gliding species which shows two kinds of movement: adventurous (A-motllity) and social (S-motlllty) (Kaiser, 1979). The A-motillty refers to the movement of single, or Isolated small groups of cells, whereas

13 S-motility involves rafts of cells. The two patterns are controlled by two different genetic systems (MacNeil et al., 1994). Pili are probably involved in

S-motility (Kaiser, 1979; Rosenbluh and Eisenbach, 1992).

A second type of nonflagellar bacterial locomotion is twitching motility, which involves pili (Bradley, 1980; Henrichsen, 1983). Twitching motility is a sporadic, outward spreading of cells at the colony edge. It can occur on relatively dry surfaces, and under coverslips. It is interesting that species like

Pseudomonas aeruginosa show both twitching and swimming motility

(Henrichsen, 1983). In a sense, twitching and gliding motilities resemble surface swarming motility, in all of which a cell mass or colony spread on a surface.

These different patterns of motility might reflect the incredible ability of bacteria to live in diverse conditions. Species with non-flagellar motility are adapted to their unique habitats. But they face the same challenges as the flagellated species: frequent changes in growth conditions and nutrient availability, limited nutrient supply, interactions with other organisms. Our findings on the changes in motility and chemotaxis upon nutrient deprivation, and on the FS behavior and switch mechanism will provide useful information to studies on nonflagellar motility and its regulation.

III. Tactic responses in bacteria

Taxis is defined as directed or directional movements towards or away from a stimulus. Tactically motile microorganisms can accumulate at

14 locations the cells perceive as favorable, or move away from unfavorable

sites. Attraction towards a stimulus is positive taxis, repulsion away from a

stimulus is negative taxis. Bacteria can sense changes in stimulus intensity

over time (i.e. a temporal gradient). The stimulus signals are processed to

effect changes in the duration of smooth swimming versus random re­

orientation along the stimulus gradient. Moving bacteria can also detect

spatial gradients of a stimulus. To a cell in motion, a spatial gradient is equal

to a temporal gradient.

When a cell is experiencing an increase in positive stimulus intensity,

it swims for a longer time and tumbles or pauses more briefly. The reverse is

true with decreasing positive stimulus or increasing negative stimulus. As a

result, a population of cells displays net migration towards the source of a

positive stimulus (up a gradient), or away from a negative stimulus (down a gradient) (Macnab, 1979).

1. Types of bacterial taxis

The stimulus for bacterial taxis can be chemical or physical in nature.

A wide variety of environmental factors affect bacterial behavior. Many physical forces can be tactic stimuli. Light at proper intensity is an attractant for some bacteria such as photosynthetic species. At high intensity it acts as a repellent to many bacteria. Thermotaxis has been observed in E. coli and thermophiles. Electrotaxis, or galvanotaxis, was also observed in a number of bacteria, in which microorganisms with charge on their cell surface move

15 towards the opposite charge. Magnetotaxis Is found commonly In aquatic

bacteria (e.g., Blakemore, 1982). Many researchers (e. g. BazyllnskI et ai.

1994) reported that these magnetic bacteria produce Intracellular magnetite

particles aligned In a specific way to form bacterial magnets. Positive

geotaxIs allows aquatic anaerobic bacteria to settle In deep waters with low

oxygen concentration, and on sediment surfaces containing nutrients. The

effect of negative geotaxIs Is to counteract sedimentation of the organisms,

especially aerobes. In aquatic environments, this has the same result as Og

taxis. LI et al. (1988) reported that E. coli cells escaped from chemicals of

high concentration. It Is referred as (negative) osmotaxis. Another novel type

of bacterial taxis Is viscotaxis which was first Identified In spirochetes (Kaiser

and Doetsch, 1975). They found that Leptospira species swam faster In

media of high viscosity than in those of lower viscosity. VIscotactIc

pathogens Inhabiting mucus membrane In animal intestines or other areas

have an advantage over those without vIscotaxIs responses.

Chemotaxis. Chemotaxis, the taxis toward chemicals, is the

predominant form of bacterial taxis. A wide range of chemicals. Including

most amino acids, sugars, and other nutrients, are chemoattractants (Adler et al., 1973; Meslbov and Adler, 1972; Macnab, 1987). Inorganic substances can be attractants or repellents (e. g., Ingolla and Koshland, 1979; Tso and

Adler, 1974). Hydrophobic amino acids like leucine, alcohol and extreme pH are repellents to most bacteria. Virtually all motile bacteria have some type of chemotactic response. Photosynthetic bacteria are attracted to H^S, a

16 substrate for photosynthesis. Aerobic and even facultative anaerobic bacteria are attracted to O^, which is a repellent to anaerobes (Shioi et al.,

1987). Plant symbionts or pathogens are attracted toward root exudates

(Barbour et al., 1991; Reinhold et al., 1985; Hawes and Smith, 1989).

2. Signal transduction In E. coli chemotaxis

Chemoreception. Most of our understanding of motility and chemotaxis comes from the extensive studies on E. coli. The first step in bacterial chemotaxis is the binding of external signal molecules to receptor protein molecules directly, or to periplasmic substrate-binding proteins which then bind to receptors. The receptors are embedded in the plasma membrane, with membrane-spanning domains characteristic of two- component signal transduction systems (Russo and Koshland, 1983). The binding of the signal molecules changes the conformation of the receptors, and the chemotactic signal is transmitted across the cell membrane into the cell, hence the receptors are also known as transducers. Receptors are better known as methyl-accepting chemotaxis proteins (MCPs) because these proteins are methylated and demethylated during chemotactic signal transduction. Based on the compounds they bind, the MCPs in E coli are divided into four types: Tsr (taxis to serine and repellent), Tar (taxis to aspartate and repellent). Trg (taxis to ribose and galactose), and Tap (taxis- associated grotein) (Boyd et al., 1983; Russo and Koshland, 1983; Macnab.

17 1987; Eisenbach, 1991). MCPs have three basic functions: binding of

ligands (attractants, or attractant-periplasmic substrate-binding protein

complex), transmission of excitation signal to cytoplasmic components of

chemotaxis sensors, and adaptation. The E. coli chemotaxis pathway is

summarized in FIG. 1.2. (from Blair, 1995).

Excitation and adaptation. When a cell is experiencing an

increase in attractant concentration, more receptors are occupied by ligands,

resulting in a rapid excitation. The conformation of the intracellular domains

of the MCPs are changed or activated, and signals are transmitted to the

cytoplasmic signal transduction components (i.e. CheA, CheW, CheY and

CheZ, discussed below), which affect flagellar motor activity. The end result

is longer CCW rotation of flagellar motors, leading to longer straight swims. If

the concentration stays unchanged, cells will adapt to that concentration of

attractants within minutes and resume the basal alternation between

swimming and tumbling (Eisenbach, 1991; Stock and Surette, 1996).

The basis for adaptation is the methylation-demethylation of MCPs.

There are four méthylation sites on each MCP. In the resting state, on average only one site is methylated. Ligand binding promotes méthylation of

MCPs by the CheR protein, a methyl transferase. After a certain time

(seconds to minutes), more methyl groups are present on the MCPs. This is the adapted state of MCPs. The adapted MCPs have the same effect as the resting-state ones on switching frequncy and the direction of motor rotation.

18 Attractant Receptor

CW-signaling state CCW-signaling state

inner membrane

1

ATP adaptation (very slow) pathway

signaling pathway little phosphotransfer to CheY, CheB fe

Increased Primarily CW Rotation CCW Rotation

Flagellar Motor

FIG. 1.2. The chemotactic signaling and adaptation pathways of E. coli. Chemoeffectors are detected by membrane receptors. The receptors modulate the activity of the autokinase CheA and thus the level of phosphorylation of its substrate CheY. when Phosphorylated, CheY promotes CW rotation of the flagellar motors. Dephosphorylation of CheY is accelerated by CheZ. The receptors are hypothesized to exhibit at least two conformations, one that activates CheA (and thus signals CW) and one that does not (and thus signals CCW). Binding of attractant stabilizes the CCW state, and repellent (not shown) the CW state. Adaptation involves the transfer of methyl groups to the receptors by CheR and their removal by CheB. (+Me signifies méthylation and -Me déméthylation.) CheB can be phosphorylated by CheA, thereby increasing its demethylating activity. Adaptation to an attractant stimulus occurs as follows: Attractant binding reduces the activity of CheB (via the CheW-CheA pathway) and also makes the receptors better substrates for CheR. Thus, over the course of several seconds the receptors become more heavily methylated. Additional methyl groups stabilize the CW state, which causes the equilibrium between CW and CCW states to shift back toward prestimulus levels. From Blair (1995).

19 The reverse is true for removal of attractants or addition of repellents: MCPs signal a CW rotation of flagella, resulting in tumbling. Then déméthylation restores basal activity of MCPs. In other words, methylation-demethylation eventually resumes the basal activity of the transducers (MCPs) after addition or removal of an attractant or a repellent. Déméthylation is catalyzed by the phosphorylated CheB protein, a methylesterase. CheA mediates the phosphorylation of CheB. Both CheR and CheB are soluble cytoplasmic proteins. CheR mutants are defective in adaptation, they can either smooth swim or tumble for long time (Parkinson and Revello, 1978).

CheB mutants are also unable to adapt (Yonekawa et al., 1983).

Although chemotaxis through MCPs is the predominant system in E coli, an MCP-independent pathway has been discovered. Lengeler et al.

(1982) identified a phosphoenolpyruvate-dependent carbohydrate: phophotransferase system (PTS), which mediates chemotaxis to oxygen and sugars.

Control of flagellar motor activity. The external chemotaxis signals eventually affect the direction of flagellar rotation. The signals are received and processed by MCPs. and relayed to flagellar motors by the integrated activities of the cytoplasmic chemotaxis proteins CheA, CheW,

CheY and CheZ. The process consists of a series of protein phosphorylations. First, unstimulated MCPs and other receptors activate the

CheA protein, an autophosphorylatable kinase (Gegner et al., 1992). CheW is involved in this process (Ninfa et al., 1991). The activated CheA catalyzes

20 the phosphorylation of CheY by transferring its phosphate group to CheY.

CheY-PO^ binds to the flagellar motor switch and causes CW rotation,

resulting in tumbling. Mutants lacking CheY are constant smooth swimmers.

CheZ. a phosphatase, first competes with the flagellar switch for the binding

site on CheY, then mediates dephosphorylation of CheY, leading to CCW

rotation of flagella. It is believed that the binding of attractants to chemotaxis

receptors decreases the rate of phosphorylation of CheA and CheY, hence

prolongs smooth swimming. Conversely, decrease in attractant

concentration promotes phosphorylation, causing more and longer tumbling

(Eisenbach, 1996). There are variations on the CheY story and motor activity

control in different species. We will discuss the control of flagellar motors of

S. meliloti in Section V.

IV. Role of motility and chemotaxis in bacterial fitness

Motility and chemotaxis in relation to starvation. As

discussed in Section I, most bacterial cells in natural habitats are in a

starvation-survival state most of the time. Ocean habitats are considered to

be a model oligotrophic ecosystem where available nutrient concentration is

very low (Moriarty and Bell, 1993). In many cases, soil is also an

environment of nutrient dearth (van Elsas and van Overbeek, 1993). Long term starvation-survival of bacteria has been studied quite extensively

(Morita, 1993). During relatively long periods of nutrient deprivation, a

number of physiological and morphological changes or adaptations were

21 observed among various bacterial species (Morita, 1993; Nystrom, 1993;

Oliver, 1993). For example, the starvation responses of a marine bacterium were studied with respect to viable and culturable cell counts, attachment ability, and fatty acid composition (Rice and Oliver, 1992). Givskov et al.

(1994) reported changes in cell shape and size, viability, protein synthesis and other parameters of Pseudomonas putida under low carbon conditions.

But there have been relatively few studies on the short-term changes in motile and chemotactic behavior of flagellated bacteria in response to starvation or reduced nutrient availability, and none of those studies has examined soil or rhizosphere bacteria.

Changes in motility during starvation have been studied mostly in some marine bacteria, E. coli and pseudomonads. These studies are briefly reviewed in Chapter 2, where we describe the changes in motility, flagellation and chemotaxis of three strains of S. meliloti after transfer to buffer containing no available nutrients.

Role of motility and taxis in bacterial competition, dispersal and host Interactions. The role of motility and chemotaxis in competition and survival of bacteria in natural environments was discussed by Macnab

(1987). There are examples that motility and chemotaxis provide benefits to bacteria. Pilgram and Williams (1976) compared the growth of a chemotactic strain and a motile but chemotactically deficient mutant of Proteus mirabilis.

They found that the two strains grew equally in shake liquid culture, but the tactic strain outgrew the mutant in a semisolid medium, where constant

22 nutrient gradients existed. Lauffenburger et al. (1982) presented a mathematical model to explain the effects of nutrient diffusion and uptake, cell motility and chemotaxis, and cell growth and death on bacterial population growth in confined, unmixed regions. They concluded that motility could increase population density only when accompanied by chemotaxis. Mass flow or increased diffusion of nutrients offset the effects of motility. Skowlund and Kirmse (1988) determined the relative cell numbers of aflagellated, motile but nonchemotactic, and chemotactic bacteria in small pores. They found that the diffusion rate of nonmotile bacteria was several orders of magnitude lower than that of the motile bacteria. In an experiment done in nonsterile soil, Soby and Bergman (1983) demonstrated that motile wild-type cells of S. meliloti spread faster than the nonmotile mutants. In

CHAPTER 3, we report the growth advantages provided by the increased swarming ability over the wild type in media with nutrient gradients.

There is evidence that motility and chemotaxis allow motile microorganisms to colonize host plants or animals. In a review, Kennedy

(1987) pointed out that motility, chemotaxis, and adhesion seemed to enhance colonization by allowing certain bacteria to selectively find the colonization sites. DeFlaun et al. (1990) demonstrated that flagella were the major factor for effective adhesion of Pseudomonas fluorescens to soil and plant seeds. Adhesion to soil might enhance bacterial acquisition of nutrients adsorbed to the soil particle surfaces. A number of studies show that motility and chemotaxis do play an important role in bacteria-host

23 interactions. In laboratory tests, wild-type motile S. m eliloti strains could produce over 20 times more nodules on alfalfa than nonmotile or nonchemotactic derivatives (Ames and Bergman, 1981 ; Caetano-Anolles et al., 1988). Broek et al. (1998) reported that bacterial chemotactic motility was important for root colonization by Azospirillum brasilense. Similarly, Hawes and Smith (1989) reported that chemotaxis was required for pathogenicity of

Agrobacterium tumefaciens on pea roots. In their study of root colonization by a plant-growth-stimulating Pseudomonas fluorescens strain. De Weger et al. (1987) found that flagella were required for root colonization. Tn5- induced aflagellate mutants of P. fluorescens appeared to be impaired in their ability to colonize growing potato roots. It was also demonstrated that flagellar motility was a factor promoting the colonization of plant leaves by

Pseudomonas syringae (Haefele and Lindow, 1987). These previous results show that motility and taxis are important to colonization ability. Yet it is not clear how, and under what regulation, tactic motility contributes to bacterial success in soil and rhizosphere environments.

Roles of motility in bacterial-animal host interactions were reviewed by Ottemann and Miller (1997). Adhesion and chemotaxis were found to be determinants of bacterial association with mucosal surfaces (Freter et al.,

1978; Freter et al., 1981; Freter and O'Brien, 1981). They found that chemotaxis facilitated the association of Vibrio cholera with the mucosal surface. They concluded that chemotaxis was important to the in vivo growth of Vibrio cholera, and only chemotactic motility, not random motility, was

24 beneficial to the bacterial survival. Chemotactic vibrios associated significantly better with the mucosa than the non-chemotactic strains.

OToole et al., (1996) showed that chemotactic motility was essential to the invasion of fish by Vibrio anguillarum. Loss of motility caused a 500-fold decrease in virulence when bacteria were inoculated into water. However, the flagella and motility were not required for pathogenicity if the bacteria were injected intra-peritoneally into fish.

V. Sinorhizobium meliloti and Sts unique features regarding motility and chemotaxis

S. meliloti nodulates alfalfa (Medicago sativa) and fixes atmospheric nitrogen. It is therefore agriculturally important. Wild-type strains of the species produce complex extracellular polysaccharides (EPS), and their colonies are mucoid on solid media. The species is motile with 2 to 8 peritrichously inserted complex flagella. Complex flagella are more rigid and stable than the plain flagellar filament, thus more suitable for producing thrust in viscous media and in soil (Maruyama et al., 1978; Krupski et al.,

1985; Robinson et al., 1992). Two flagellin genes, flaA and flaB, were identified in S. meliloti (Bergman et al., 1991; Pleier and Schmitt, 1989 and

1991). The genes encode two proteins of almost identical amino acid sequences.

Common amino acids, dicarboxylic acids, flavonoids and sugars are chemoattractants for S. meliloti (Ames at al., 1980; Bergman et al., 1988;

25 Gôtz et al., 1982; Gulash at al., 1984; Gaetano-Anollés et al., 1988;

Dharmatilake and Bauer, 1992), and toward some unidentified attractants

secreted at localized sites on the roots (Gulash at al., 1984). Bergman et al.

(1988) provided evidence for the existence of a dual chemotactic pathway in

S. meliloti. They found mutants that lost chemotaxis responses to animo acids and sugars, but still responded to root exudates of alfalfa, whereas another group of mutants were generally chemotaxis defective.

In their review, Armitage and Schmitt (1997) compared enteric

bacteria with Rhodobacter sphaeroides and S. meliloti I n terms of motility and chemotaxis (see FIG. 1.3). Although the chemosensory pathways in S.

meliloti bear many similarities to those in E. coli, the regulation of the expression of chemoreceptors and flagellar motor activity is very different. In contrast to constitutive expression in enteric bacteria, the expression of S.

meliloti and R. sphaeroides chemoreceptors is controlled by environmental signals. Unlike E. coli which swims by CCW rotation of flagella and tumbles by CW rotation, S. meliloti swims exclusively by clockwise rotations of its flagella, and direction change is achieved through asynchronized flagellar rotation, brief pauses or changes in swimming speed (Gôtz and Schmitt,

1987). S. meliloti exhibits chemokinesis, swimming faster upon exposure to higher concentrations of attractants (Sourjik and Schmitt, 1996). It has at least two CheY proteins which interact with the flagellar motors. The swimming speed of S. meliloti \s found to be dependent on the intensity of

26 (a) (b) Environmental signal Signal Signal rT~) Mcpr?~i . œ C O Aer

! ® 1 i

Inner membrane Inner membrane

Motor [ Peptidoglycan layer Motor PeptWoglycan layer

Outer membran* O uter m em brane Hook Hook

Filament Filament

FIG. 1.3. Possible sensory pathways of (a) E. coli, (b) S. meliloti. A, B, R, W, Y and Z represent CheA, CheB, CheR, CheW. CheY and CheZ, respectively. P denotes phosphorylation. Tip is transducer-like-protein. Aer=aerotaxls and other energy taxis response (Rebbapragada et al., 1997). From Armitage and Schmitt (1997).

27 chemosensory stimulation.

A number of investigators have isolated and studied motility and chemotaxis mutants of S. meliloti (Ames et al., 1980; Ames and Bergman,

1981; Bergman et al., 1988). Soby and Bergman (1983) showed that a motile strain of S. m eliloti migrated faster than nonmotile mutants in soil.

Mutants defective in motility or chemotaxis are impaired in their ability to compete for sites of nodule initiation on the host roots (Ames and Bergman,

1981; Caetano-Anolles et al., 1988). Motility played an important role in nodulation efficiency of S. meliloti L5-30 (Malek, 1992). We have isolated nonmotile (Mot"), chemotaxis defective (Che'), reduced motility/swarming

(RS), and faster (increased) swarming (FS) mutants and variants. The increased swarming mutants are a new class of mutants which migrate twice as fast as the wild type in semisolid agar, in sand plates and in viscous liquid medium. Previous studies have reported only comparisons of motile wild type and nonmotile mutants. It is therefore important to find out whether these FS mutants and variants are better at growth and survival, root colonization, and nodulation than the wild type. In CHAPTERS 3 and 4, phenotypic and genetic characterization of the FS mutants and variants are presented. We have discovered that the wild type can switch to FS configuration spontaneously, and vice versa. The ecological significance of the FS phenotype and the switching mechanism Is discussed.

28 VI. Goals of this study

My dissertation research has focused primarily on motility and chemotaxis in S. meliloti, although I have also investigated the effect of homoserine lactone-mediated signalling between bacterial cells on growth, adhesion and root colonization by P. aeruginosa. The first part of my Ph.D. project was to study the behavioral responses of S. meliloti to starvation conditions. There has been no previous study on the changes in motility induced by nutrient deprivation in any soil-rhizosphere species. I wanted to determine whether the cells, when facing nutrient limitation, commit their reserved substrates to sustain their motility in order to search for new nutrient sources, or whether they turn off motility to save energy for long term survival. 1 sought to find out the first responses of S. meliloti to starvation. If the bacterium lost motility, one subsequent objective was to determine the mechanism for the loss of motility; by loss of flagella, or simply by switching off of flagellar motors yet retaining flagella. I was also interested in studying the mechanisms and the importance of such regulation in natural soil and rhizosphere environments.

The second aspect of the motility. project was to generate and characterize faster swarming (FS) mutants. I was interested in finding out whether S. meliloti had the ability to regulate and express various degrees of motility under different conditions, and to assess the ecological value of such ability. It has been reported that isolates with better motility could be obtained by enrichment of bacterial cells in semisolid (swarm) media, but no

29 one has studied such Isolates or their genesis, stability, ecological

consequences or genetic basis. It is the first systematic characterization of

Tn5 bacterial mutants that spread faster than the wild type. Another goal was to find out the possible cellular and molecular mechanisms for the FS

behavior, and to further pursue its role in bacterial competitiveness and

survival in natural habitats.

The third project was to study the possible effect of N-acyl homoserine

lactone (AHL) signalling on bacterial growth, adhesion and root

colonization. There are many reports in the literature that several important cellular processes are cell density dependent, or regulated through AHL cell-cell communication in a bacterial population. But there have been no studies on the possible role of AHL signalling in cell adhesion and growth in a soil matrix and on root surfaces. Mutants defective in AHL production and perception were available to us for doing some preliminary investigations into this subject, which are of considerable ecological significance.

30 CHAPTER 2

STARVATION-INDUCED CHANGES IN MOTILITY, CHEMOTAXIS AND FLAGELLATION OF SINORHIZOBIUM MELILOTI

ABSTRACT

The changes in motility, chemotactic responsiveness, and flagellation of S. meliloti RMB7201, L5-30, and JJ1 c10 were analyzed after transfer of the bacteria to buffer containing no available C, N or phosphate. Cells of these three strains remained viable for months after transfer to starvation buffer (SB), but lost all motility within just 8 h to 4 days after transfer. The three strains of the same species showed different patterns of motility response to starvation, with L5-30 losing motility fastest, and JJIcIO maintaining it the longest. Each strain showed a transient, two- to sixfold increase in chemotactic responsiveness toward glutamine within a few hours after transfer to SB, even though motility dropped substantially during the same period. Strains L5-30, and JJIcIO also showed increased responsiveness to the nonmetaboljzable chemoattractant cycloleucine. Cycloleucine partially restored the motility of starving cells when added after transfer and prevented the loss of motility when included in the SB used for initial suspension of the cells. Thus,

31 interactions between chemoattractants and their receptors appear to affect the regulation of motility in response to starvation Independently of nutrient or energy source availability. Electron microscopic observations revealed that S. m eliloti cells retained flagella as the cells lost their motility. Even after prolonged starvation when none of the cells were motile, about one-third to one-half of the initially flagellated cells retained some flagella, but flagellar integrity deteriorated. Inactivation of flagellar motors therefore appears to be a rapid and important response of S. meliloti to starvation conditions. Flagellar- motor inactivation was at least partially reversible by addition of either cycloleucine or glucose. During starvation, some cells appeared to retain normal flagellation, normal motor activity, or both for relatively long periods while other cells rapidly lost flagella, motor activity, or both, indicating that starvation-induced regulation of motility may proceed differently in various cell subpopulations. The significance and possible ecological role of such responses to starvation conditions are discussed.

INTRODUCTION

In the majority of natural environments, nutrients are limited and nutrient distribution is highly uneven. Bacteria are in either a “feast-famine” or a starvation-survival state. Bacteria have certain phenotypical adaptations in response to diverse environmental conditions. Motility and chemotaxis are mechanism that enables motile bacteria to sense a favorable environment

32 and to migrate to it (Macnab, 1987; Adler et al., 1973). As nutrient availability

approaches zero in natural environments, populations of motile bacteria face

a potentially crucial choice between searching actively for additional nutrients

or shutting off motility to conserve energy and substrates. On one hand,

continued motility and chemotaxis may offer a population its best chance to

find new nutrient supplies, avoid stress, mate, and reproduce. On the other

hand, motility and taxis are inherently costly cellular processes, requiring the

synthesis of perhaps 50 gene products and the performance of considerable

mechanical work, with flagella rotating up to 15,000 rpm and requiring about

1,000 protons per revolution (Macnab, 1996; Schuster, and Khan, 1994;

Sourjik and Schmitt, 1996; Stock and Surette, 1996). Thus survival and

competitive success in natural environments may require sophisticated,

perhaps subpopulation-level regulation of motility and chemotaxis in

response to low-level and fluctuating nutrient availability. Relatively few

studies have examined the changes in behavior of flagellated bacteria in

response to starvation or reduced nutrient availability, and none so far has

involved soil or rhizosphere bacteria. Terracciano and Canale-Parola (1984)

observed that carbon-limited growth of a Spirocheata species in chemostat

cultures resulted in a 10- to 1,000-fold increase in chemotactic

‘ responsiveness to the specific sugars used to support the growth, indicating

that this bacterium selectively enhanced its behavioral sensitivity to low

concentrations of growth-limiting nutrients. Previous studies (Amy and Morita,

1983; Geesey and Morita, 1979; Torrella and Morita, 1981) showed that some

33 marine vibrio isolates increased their chemotaxis responsiveness after 48 h starvation, and shifted to high affinity transport system for nutrients. However, another marine vibrio lost motility and responsiveness within 10 h after transfer to starvation conditions (Malmcrona-Friberg et al., 1990), whereas a marine Pseudomonas strain increased its motility after starvation for 27 h

(Wrangstadh et al., 1990). It is not clear from these limited studies whether there are common patterns of behavioral regulation for flagellated bacteria in response to starvation conditions.

In this Chapter, we examine the starvation-induced changes in the motile behavior of a common soil and rhizosphere bacterium, S. meliloti. The present study describes some of the changes in motility, chemotaxis, and flagellation seen in three strains of S. meliioti after transfer to starvation conditions and seeks to characterize the behavioral strategies used by this soil-rhizosphere species in response to starvation and the variability among various strains of this species in their responses.

MATERIALS AND METHODS

Bacterial strains, media and culture. S. meliloti strains RMB7201

(Hankinson et al., 1990), L5-30 (Caetano-Anolles et al., 1988), and JJIcIO

(Watson et al., 1988) were used in this study. Bacterial stocks were maintained in 15% glycerol at -80°C. Strains were routinely cultured in 1/10 or full strength TY medium (6.0 g of bactotryptone, 3.0 g yeast extract, and 0.5 g

34 CaCI^*2H^O per liter) (Berlnger. 1974). In some experiments minimal (NM) salts medium (Robinson et al., 1992) was used, with 20 mM of succinate-Na^, or mannitol as cartjon source, and 5 mM of KNO^or (NH^)gSO^ as nitrogen source, with pH adjusted to 6.8. After autoclaving of the NM media, and immediately before use, 5 mg of CaCl^ 2 H2 O and 1 ml of vitamin solution (Gotz et al., 1982) were added per liter. E. coli CC118 (Herrero et al., 1990) and

Pseudomonas aeruginosa PA01 (Holloway et al., 1979; Seed et al., 1995) were grown in Luria-Bertani (LB) medium. All liquid cultures of S. m elllotl strains and P. aeruginosa were grown at 28°C, and E. coli at 37°C, on a rotary shaker at 175-200 rpm. Solid media were made by the addition of 1.5% agar

(Difco). All chemicals used for these media were analytical or reagent grade

(Baker Chemical Co., or Sigma Chemical Co.). Cells were routinely harvested for transfer to starvation medium during early exponential phase (/Agg^. ca.

0.15), the period during which the S. m eliloti strains exhibit the greatest motility.

SB. The starvation buffer (SB, or chemotaxis buffer-CB) contained 10 mM of HEP ES (AA2-hydroxyethylpiperazine-A/-2-ethanesulfonic acid), 0.1 mM

EDTA-Na and 0.2 mM CaCl^ (Sigma Chemical Co.), prepared in HPLC grade water (Baker Chemical Co.) or nanopure water, pH 7.0. This buffer was found to be optimal for maintaining the motility of S. meliloti and the integrity of its complex flagella (Robinson et al., 1992). It is also suitable as a starvation

35 medium since it contains no utilizable carbon, nitrogen or energy source.

Motility and chemotaxis assays. For routine analysis of motility

after transfer of cells to starvation medium, early exponential phase cells,

usually 1:500 dilution for 1/10 TY and 1:250 for NM, were grown overnight,

were harvested by centrifugation of cultures at 1,800xg for 10 min. Droplets of

liquid were carefully removed from the tube after decanting. The pellet was

then resuspended in an equal volume of SB by gentle inverting and tapping.

To determine the percentage of motile cells, an aliquot of bacterial suspension

was placed in a Petroff-Hauser counting chamber and observed at a

magnification of 860x with phase-contrast optics (Zeiss I 35 inverted

microscope). Only the cells near the grid surface were counted, and the

counting was performed rapidly (1 to 2 min.) to minimize the settling of

nonmotile cells. Reported percentages are averages of counts from two

independent replicate suspensions for all experiments. A total of 80 to 100

cells from two or three different fields were counted per determination. The

percentage of motile cells determined in this way was normally reproducible

within ±10%. In some cases, video recorded data were analyzed to get motile­

cell percentages and to characterize motility behavior. Chemotactic

responsiveness towards glutamine (10 mM in SB) or other attractants was

determined by capillary assay in Palleroni chambers (Palleroni, 1976; Adler,

1973). Responsiveness is expressed as a chemotaxis ratio (the average number of cells entering attractant-filled capillaries divided by the average number of cells entering buffer-filled capillaries). Relative motility was also

36 measured by entry into SB-filled capillary tubes (Segal et al., 1977). In some experiments, responsiveness was assayed with the attractant present in both the bacterial suspension and chemotaxis capillaries as a control. Entry into capillary tubes for motility and chemotaxis assays were determined after a 30- min incubation at room temperature. The average CPU per capillary tube was determined by plating aliquots of suspensions from four replicate tubes per data point.

Transfer to starvation conditions by membrane filtration and dialysis. Ten-rnilliliter aliquots of a mid-exponential-phase 1/10 TY or NM cultures were filtered through micropore membranes of various type (0.02-pm- pore-slze Anodisc [Alltech Associates, Inc., Deerfield, III.], 0.45-pm-pore-size cellulose acetate and nylon [Magna, Honeye, N. Y.], 0.22-;vm-pore-size mixed cellulose esters [Poretics, Livermore, Calif], or 0.4-pm-pore-slze polycarbonate [Nucleopore Co., Pleasanton, Calif.]). After filtration, the filters were washed twice with SB (20 ml each), and the cells were subsequently resuspended by gentle agitation of the membrane in 10 ml of SB.

Cells were also transferred to SB by dialysis. Mid-exponential phase cultures were transferred into wells of a microdialysis device (Pierce, System

100). Through continuous flow of starvation buffer, in one to two hours, about

99% of the test solute NaCi was gone. Aliquots of cell suspension in the dialysis wells were microscopically observed to determine motile cell percentages and motile behavior. However, it is not known whether large molecules such as proteins in bactotryptone and yeast extracts were removed

37 completely. In other experiments, dialysis tubing (MWCF 6000-8000,

Spectrum Medical Industrial, Inc., L. A., Calif) were substituted for dialysis

device. The tubings were submerged in SB and gently agitated on a stirrer, with a number of SB changes before cells were observed or assayed. By visually inspecting the color of the culture inside the tubings, it seems to take

hours to remove nutrients.

Flagellar staining and electron microscopy. The percentage of flagellated cells was determined at various times after transfer to SB. A 3-^m

droplet of the suspension was placed on a piece of Parafilm, and 1-;yl of 2%

uranyl acetate stain was added. After a few seconds of mixing, 2 fj\ of the

droplet was transferred to a Formvar-coated 300-mesh copper grid. The cells were allowed to react with the stain and to settle onto the surface of the film for

5 min prior to being air dried in a laminar-flow hood. Grids were observed through a Philips model 201 transmission electron microscope. Air drying of the grid surface left areas of rather dense stain, but other areas had a light background and were suitable for observation of bacteria and flagella. The percentage of cells with flagella was determined by observation of 75 to 100 cells per grid on at least two duplicate grids per sample. For quick estimation of flagellation, regular flagellar staining was performed as described (Mayfield and Inniss, 1977; West et al., 1977).

TnS mutagenesis. Transposon insertion mutagenesis was carried out to create insertional mutants of strain L5-30, with E coli strain

WA803(pGS9) as Tn5 donor. Screening of mutants that retain longer motility

38 was done as follows; exconjugant cells were spotted on SB soft agar, after overnight incubation, the outer rings of agar was transferred to fresh NM media containing kanamycin. This process was supposed to enrich cells that retained longer motility, which should spread farther away from the inoculation point. The enrichment was repeated. Putative mutants were purified and tested for motility retention after transfer to SB.

RESULTS

Survival of S. m elllotl In starvation medium. Viable-cell counts of all three S. meliloti strains remained high for months after transfer to SB

(FIG. 2.1). The number of viable cells of strain JJIcIO gradually diminished about 50-fold over a period of 120 days, but counts of strains L5-30 and

RMB7201 remained almost unchanged, at about 10® CFU/ml, during this period.

Changes in motility in response to starvation conditions.

Different strains of S. meliloti were compared to see if they have different strategies for adapting motile behavior to starvation conditions. The visible motility of all three strains of S. meliloti diminished to almost zero within 6 to

96 h after transfer to starvation medium (FIG. 2.2). The length of time required for each of the three strains to reach half-maximal or essentially zero motility was quite reproducible and also considerably different for each strain. These times were not significantly affected by the growth phase prior to harvest. The

39 10"

c > o . RMB7201

i LL o

JJ1C10

10" 0 20 40 60 80 100 120140 time after transfer to SB (d)

FIG. 2.1. Long term survival of S. meliloti RMB7201 (o), L5-30 (V ) and JJIcIO (T) after transfer to SB. CPUs of each strains were determined at various times by plating. Error bars represent sample standard deviations. Each data point was average of triplicates.

40 100

# # e-e (g E

RMB7201

L 5 -3 0

0 5 10 20 3515 time after transfer to SB (h)

FIG. 2.2. Changes in motility of R. meliloti strains L5-30 (o), RMB7201 (V). and JJIcIO (•) at various times after transfer to SB. Cells were transferred to SB and assayed microscopically to determine the percentage of motile cells as described in Materials and Methods. Data points are from a single experiment. The curves shown are representative of those obtained in at least 4 other experiments. Replicate samples normally show less than 10% variation from the mean.

41 loss of motility seen in FIG. 2.2 was not affected by addition of trace metals or

vitamins. Strain JJ1c10 retained good motility (>40% motile) for more than 20

hours, then gradually decreased. Even after sitting in SB for more than 96

hours, a fraction of cells were still motile. Strain RMB7201 had a similar

pattern but the percentage of motile cells dropped more quickly than that of

JJIcIO. Good motility was observed within 5-8 hours after transfer, and very

small percentage of cells were motile after 24 hours. Strain L5-30, on the

other hand, retained motility for a very short period (less than an hour) then

rapidly diminished and dropped to essentially zero after only 2 hours. The

results indicate that very different regulatory responses exist among strains of

the same species.

The bacteria lost their motility after transfer to SB regardless of whether

they were transferred by centrifugal washing (FIG. 2.2). by membrane filtration,

or by dialysis (data not shown). Centrifugal washing and resuspension was

adopted for routine use in subsequent starvation response studies. Dialysis

was considered unsuitable due to the relatively long time interval between the

initiation of dialysis and essentially complete removal of nutrients from the

bacterial cultures. Filtration through micropore membrane filters of various types allowed for rapid transfer to starvation conditions but resulted in large

(>50%) reductions in the percentage of motile cells after resuspension in SB, presumably due flagellar adhesion or damage (data not shown). Centrifugal washing was rapid and resulted in relatively small (5 to 10%) reductions in the

42 motile and flagellated cells during the transition from exponential-phase, 1/10-

strength TY cultures to fresh SB suspensions. The number of detached

flagella present in the supernatants of the SB suspensions was found to be

low, indicating that centrifugation and resuspension caused relatively little

damage to flagella. Centrifugal washing, however, did not give complete

recovery of the motile cells in the original cultures. At the relatively low

centrifugation speeds that were used to pellet the cells as gently as possible,

10 to 15% of the motile cells remained in the supernatant.

Changes in chemotactic responsiveness upon exposure to starvation conditions. All three strains showed significantly (two to sixfold),

but transiently, enhanced responsiveness to 10 mM glutamine within a few

hours after transfer to starvation medium (FIG. 2.3). The enhanced responses occurred despite the fact that the overall motile activity diminished rapidly during the same time period. L5-30 cells suspended in SB showed a transiently increased responsiveness to nonmetabolizable attractant cycloleucine, similar to that obtained with glutamine (FIG. 2.4). Cycloleucine is a potent attractant for S. meliloti (Robinson and Bauer, 1992) but does not support its growth and is not toxic (data not shown). The entry of cells into buffer-filled capillaries was not appreciably affected by the addition of cycloleucine to the bacterial suspension. Thus, enhanced entry of starving S. meliloti cells into cycloleucine-filled capillaries is not due to localized activation of motility in cells exposed to the attractant.

43 (a) LS-30 o (6 )« « S 7 a o i ta S I o

a

o E 0 5 10 25 0 5 <0 IS 20 2S 7080 time after transfer to SB (h) time after transfer to SB (h)

o 2

' o " V o E 2 0 20 «) so 30 100 120 time after transfer to SB (h)

FIG. 2.3. Changes in motility and chemotaxis of S. meliloti strains L5-30, JJIcIO and RMB7201 after transfer to SB. Cultures were transferred to SB and assayed at various times after transfer for percentage of motile cells and for chemotactic responsiveness as described in Materials and Methods. Percentage of motile cells (#, solid line), number of cells entering SB-filled capillary tubes (V, dashed line), or number of cells entering capillary tubes containing either 5 mM glutamine or cycloleucine (o, dashed line). The chemotaxis ratio (T, solid line) was calculated by dividing the number of cells in attractant-filled capillaries by the number in SB-filled capillaries. (3 a ) Strain L5-30, glutamine as attractant; (b ) Strain RMB7201. glutamine as attractant; (c ) Strain JJIcIO, glutamine,as attractant. Data are from a single experiment and are representative of those obtained in at least two independent trials. Replicate samples normally show less than 10% variation from the mean.

44 25

20

o 2 (g X <0 o E 0) SI o

10° CLconcentr. (icg[mM])

FIG. 2.4. Chemotactic responsiveness of S. meliloti L5-30 to different concentrations of cycloleucine (CL) before and after starvation. L5-30 cells freshly transferred from 1/10 TY cultures to SB (o), or starved in SB for 2 h (•) were assayed for entry into capillary tubes filled with 5, 1, 0.2, 0.04, or 0.008 mM of cycloleucine as described in Materials and Methods. Data points represent averages from four replicate capillaries. The chemotaxis ratio was calculated by dividing the number of cells in attractant-filled capillaries by the number in SB-filled capillaries. Curves are from a single experiment that was repeated with similar results. Replicate samples normally show less than 10% variation from the mean.

45 Swimming behavior of ceils exposed to starvation conditions. Light and videomicroscopy revealed clear changes in the character of swimming behavior after transfer of S. meliloticeWs to starvation medium, immediately after transfer to SB, when 40 to 60% of the cells were motile, the cells were very active, swimming quite continuously with almost no discemible pauses. The swimming was relatively fast (ca. 20 to 30 /vm/s) and straight, the bacteria often covering distances of 100 fjm between abrupt changes of direction. The proportion of cells of all three strains that exhibited this normal swimming behavior diminished steadily after transfer to SB. By the time the percentages of motile cells had dropped to 15 to 25% of the initial levels, very few cells exhibited any smooth forward swimming. The average rate of forward movement of these starved cells was slow, perhaps one-third of the speed seen in fresh suspensions, and few cells moved more than 50 ^m from their starting point within 1 min. Motile cells in starved suspensions seemed to change direction more frequently than those in fresh suspensions, and many of the swimming cells were observed to come to a complete stop, with no visible motion. Sometimes these stops lasted for a fraction of a second, while at other times they lasted 1 s or more before swimming was resumed. Near the grid surface, where motion is constrained, many motile cells in these starved suspensions were observed to move in circles of 1 to 3 cell lengths in diameter. With longer periods of starvation, an increasing proportion of the motile cells appeared unable to swim at all and just spun or tumbled weakly in one location.

46 To ensure that light microscopic determinations of the percentage of motile ceils provided an accurate measure of the changes in motility induce by starvation, and to see whether the reduced forward swimming behavior exhibited by starving cells might affect entry into pores (such as those found in soil), motility was measured by capillary assay as described by Segal et al- (1977). Motilities measured by microscopy and capillary assay did not always change in parallel to each other with time after cell transfer to starvation medium (FIG. 2.3). However, both measures of motility seem valid and useful. They provide averages over different time scales, with the microscopic assay giving an average over a short (roughly 1-min) time period and the capillary assay giving an average motility over a longer (30- min) time period.

Changes in flagellation of cells exposed to starvation conditions. Initially, it seemed possible that the loss and disintegration of flagella could explain the observed losses of motility and normal swimming behavior following transfer to starvation conditions. To test this possibility, a simple method for determining the percentage of cells that retained flagella was devised. Grids for electron microscopy were prepared by mixing an SB suspension of bacteria with uranyl acetate stain and drying this mixture directly on the grid film, without washing or blotting. This procedure prevented the potentially selective loss of actively motile cells during blotting as well as the potentially selective retention of any cells that might adhere more strongly to the grid film during blotting or washing. Addition of the

47 uranyl acetate stain was observed to stop all motility within seconds, making it likely that the motile,flagellated cells settled on the surface of the film in much the same manner as nonflagellated cells. When freshly transferred suspensions of bacteria were examined, relatively few loose flagella or fragments of flagella were observed on the grids. Cells from growing cultures and those from freshly transferred suspensions were similar with respect to the number and length of flagella associated with the bacteria.

These observations provide evidence that low-speed centrifugal transfer of cells to SB, exposure to uranyl acetate, and subsequent drying resulted in relatively little breakage of flagella or separation of flagella from their cell of origin. About 10 to 15% of the cells on the grids were present in clumps of five or more cells. Abundant flagella were seen in association with the clumped cells, possibly indicating a tendency for flagellated cells to adhere to each other. The cells present in such clumps were not included in counts of flagellated cells since it was not easy to distinguish between individual cells with and without flagella in such clumps.

The reliability of this direct counting method for determining flagellated cell percentages was tested by mixing a suspension of L5-30 containing a high percentage of flagellated cells (i.e., freshly transferred to

SB) with a suspension containing a low percentage of flagellated cells (i.e.,

20 h after transfer to SB). The percentage of flagellated cells counted in the mixture was in close (±3%) agreement with the percentage predicted from counts of the two initial suspensions. Flagellated-cell percentages

48 50

CO 1 ■D 40

v _ o CO 20 CD Ü CD +-> O E 10

0 4 8 12 16 20

time after transfer to SB (h)

FIG. 2.5. Retention of motility and flagella by strain L5-30 after transfer to SB. Cells were transferred to SB. then assayed at various times for percentage of motile cells and percentage of flagellated cell counts as described in Materials and Methods. Percent motile cells (o), percent flagellated cells (•). Results are representative of two independent experiments.

49 determined from duplicate grids were usually found to agree within 5%, indicating that the method is quantitatively reproducible.

After suspension of L5-30 cells in starvation medium, the percentage of cells with flagella decreased gradually from about 45% to 32% over an 8-h period (FIG. 2.5). The percentage of visible motility, however, decreased from about 40% to essentially zero during this same period. Most of the L5-30 cells that retained flagella after several hours of starvation did not have the normal numt)er of full-length flagella. Indeed, many retained only a short (0.5- to 1.0 pm) remnant of a single flagellum. About one-third of the flagellated cells In 8-h, nonmotile suspensions of L5-30, however, retained two or more full-length flagella. A similar proportion of flagellated cells in suspensions of

L5-30 starved for 20 h or more had two or more full-length flagella, even through the overall percentage of flagellated cells dropped to 15 to 20%.

Strains RMB7201 and JJIcIO behaved similarly: a relatively high proportion

(30%) of the cells from nonmotile suspensions of these strains, starved for 24 to 48 h, still retained flagella, and a considerable faction (approximately one- third) of these flagellated cells had two or more full-length flagella.

Effects of glucose, glutamine, culture filtrates, and nonmetabolizable chemoattractants on retention of motility and reversal of motility loss. As shown in TABLE 2.1, the addition of 5 mM glucose, glutamine, or cycloleucine (nonmetabolizable attractant) to the SB used to wash and resuspend the initial cultures of L5-30 resulted in an almost complete retention of motility for a period of several hours. The

50 TABLE 2.1. Motility of S. meliloti L5-30 after transfer to SB containing nutrients, culture filtrate (CF) or non-metabolizable attractants.

Percentage of motile cellsa Substance added Time after transfer (min) 0 20 30 40 60 90 120 ISO

No addition 38 25 20 18 16 12 12b 10b

5 mM glucose 45 55 ND 60 ND 60 ND 60

5 mM glutamine 60 60 ND 60 ND 60 ND 60

5 mM cycloleucine 60 60 ND 60 ND 60 ND 60

5 f/M cycloleucine 55 55 ND 60 ND 60 ND 60

5 mM rtaconate 35 35 ND 35 ND 40 ND 30

L5-30 CF 75 NO 75 ND 75 75 75 ND

JJIcIO CF 75 ND 80 ND 80 80 80 ND

7201 CF 65 ND 65 ND 65 65 65 ND a Percentages are from single representative experiments that were repeated two or three times. ^ Cells mostly tumble slowly, few swim straight. ND=not determined.

51 TABLE 2.2. Restoration of motility to starved ceils by addition of glucose (Glc) or cycloleucine (CL), or both.

Strain Hours starved Hours elapsed Percent motile cells before addition after addition Comooundfs^added None Glc CL Glc+Cl

L5-30 1.5 0 20 40 40 48 0.5 8 42 40 50 2 5 50 35 55 4 2 48 30 50 8 0 26 23 28

8 0 0 4 3 5 0.5 0 4 3 5 1 0 1 4 8 16 2 0 1 4 8 16 3 0 23 13 26 14.5 0 2 0 10

JJlclO 24 0 35 ★ 38 * 2 35 *• 38 8 25 35 ■*r 24 1 5 ■*- 23 *

*— Glucose caused cells clumping of strain JJIcIO, hence the data are not reported.

52 addition of 0.05 mM cycloleucine was just as effective as 5 mM cycloleucine in blocking motility loss. The addition of 5 mM itaconic acid, another nonmetabolizable (but weaker) chemoattractant, was also effective in blocking the loss of motility during starvation but did so only about half as well as cycloleucine. L5-30 cells washed and resuspended in SB containing either glucose or one of the nonmetabolizable attractants did not grow, and they gradually lost motility after the first few hours. Cells washed and resuspended in SB containing glutamine, which provides both N and 0 for

growth, were able to multiply and maintained motility for longer period of time. L5-30 cells resuspended in SB containing one-fifth-strength culture filtrate from 3-day-old stationary-phase cultures of L5-30, RMB7201, or

JJIcIO grown in NM-succinate defined medium also retained essentially full

motility for several hours.

Because the loss of motility Induced by starvation could be prevented by the addition of either a nonmetabolizable attractant or an energy-C source

such as glucose, it was of interest to determine if these substances could

reverse the loss of motility after a period of starvation. Cells of L5-30 were washed and resuspended in SB and incubated for either 1.5 or 8 h, so that the visible motility of the suspensions was reduced by 50 or 98%,

respectively. Then either glucose, cycloleucine, or both were added to a concentration of 5 mM. As seen in TABLE 2.2, delayed additions of glucose

and cycloleucine were able to restore the motility of starving L5-30 cells to some extent. Both the glucose and glucose-cycloleucine additions quickly

53 restored the motility of 1.5-h-starved ceils to full, pre-starvation levels of

normal swimming activity (TABLE 2.2). The addition of cycloleucine alone

appeared to activate a comparable number of nonmotile cells, but their

motility was largely restricted to circling and tumbling activity, with little

evidence of straight swimming. When glucose and cycloleucine were added to L5-30 cells which had been starved for 8 h and had lost ca. 98% of their

motility, it took 1 h or more for the added compounds to have any appreciable effect on motility. Restoration of motility, even after 6 to 12 h, was only partial,

particularly in terms of straight swimming activity (TABLE 2.2). In similar experiments with strain JJIcIO, the addition of cycloleucine to suspensions, which had been starved for 24 h in SB and had lost about 50% of their motility, was found to immediately, though modestly, increase the percentage of motile cells and the number that entered cycloleucine-filled capillaries. The addition of cycloleucine 24 h after transfer did not prevent the cell motility from gradually decreasing over the next 24 h of starvation. When glucose was added to these 24-h starved JJIcIO suspensions, the cells were found to clump in large aggregates within a few hours, so motility could not be assessed.

54 DISCUSSION

Starvation-induced changes in chemotactic

responsiveness. All three strains of S. meliloti studied here showed increased but transiently expressed chemotactic responsiveness during starvation. Enhanced chemotactic responsiveness may represent a general starvation response by which motile bacteria increase their probability of finding new nutrient supplies without committing internal reserves to continued flagellar synthesis or motor activity. The molecular mechanisms involved in the chemotactic-sensitivity facet of S. meliloti starvation response are unknown. In starving S. meliloti cells, enhanced responsiveness does not appear to involve significant changes in either the threshold of or the optimal concentration for responses to attractants (Figs. 2.3 & 2.4). Since starving S. meliloti cells exhibit increased responsiveness to cycloleucine, a nonmetabolizable attractant, we conclude that this organism's enhanced responsiveness to attractants does not require exposure to new nutrient or energy supplies. It is not clear why, or even whether, the increased responsiveness of S. m eliloti is transient. It is quite possible that responsiveness per se remains at a high level during prolonged starvation and that the reduction in entry into attractant-filled capillaries simply reflects reduced motility.

Regulation of motility In response to starvation. In general, the motility of S. meiiloti cells was down regulated during starvation. Both

55 flagellar maintenance and motor activity seemed to be affected. The bacteria lost flagella as a result of starvation, but this loss of flagella could not fully account for the extensive loss of motility during the initial phases of the starvation response (FIG. 2.5), and many of the nonmotile cells in starving cultures retained flagella. We conclude that one of the first responses of S. m eliloti to starvation is to turn off the rotation of the flagellar motor in certain cells. Flagellar-motor activity in S. meliloti is reported to alternate between an

“on” state and an “off” state during normal swimming behavior (Gotz and

Schmitt. 1987). Similar behavior, with considerably longer pauses between swims, is seen in the related, uniflagellate species Rhodobacter sphaeroides

(Ward et al., 1995). We speculate that starvation may progressively Increase the proportion of time that the individual flagellar motors of a cell spend in the off state. If so, then one would predict that the time Intervals that cells spend in straight swimming would gradually decrease, and that the frequency of directional changes would correspondingly Increase. Later during starvation, the probability that all flagellar motors on a given cell are off at the same time would become appreciable and the cells would stop moving altogether for brief periods. Finally, the proportion of flagellated cells that stay at rest for extended time periods would increase to 100%. Our videomicroscopic observations of motility are consistent with this starvation response model.

After a period of starvation sufficient to reduce the percentage of motile cells to ca. 20% of their initial level, those cells that remained motile often stopped moving altogether for periods of time ranging from about 0.5 s to several

56 seconds. Unstarved cells rarely, if ever, stopped moving for more than a small fraction of a second. The possibility that an increased amount of time in the off position is a starvation-induced response seems in accord with the increased proportion of time that E. coli flagellar motors were reported to spend in the inactive, pausing mode In suspensions without nutrients as opposed to suspensions with added glycerol (Lapidus et al., 1988).

The switching off of flagellar-motor activity in response to starvation seems to be at least partially reversible In all three strains tested. The reactivation of motor activity in starved L5-30 cells by added attractants was effectively instantaneous during the first 1.5 h after transfer to SB but required

1 h or longer when glucose or cycloleucine was added after 8 h of starvation

(TABLE 2.2). It remains to be determined how the cells change between 1.5 and 8 h of starvation to prevent the full and rapid reactivation of flagellar- motor activity. Access to a carbon or energy source was the main limiting factor in reactivation of motility in cells starved for an extended period.

However, the appreciable reactivation of motor activity by cycloleucine alone

(TABLE 2.2) indicates that the binding of a receptor to a nonmetabolite can be an important contributor to activation, even if it is not sufficient to restore normal swimming behavior. It will be of interest to learn whether one of the recently described MotC or MotD proteins (Platzer et al., 1997), which have no counterparts in E. coli, might serve as a receptor for signals corresponding to energy sources or attractants that would subsequently regulate flagellar- motor activity.

57 In contrast to L5-30 and RMB7201, strain JJIcIO remained fairly responsive to reactivation by chemoattractants even after 48 to 100 h of starvation (FIG. 2.2; TABLE 2.2). Such differences between strain JJIcIO and the other two strains in the reactivatability of flagellar motors, as well as the different periods of starvation required to reduce motility to half-maximal or zero levels for the three strains (FIG. 2.2), suggest that all three strains of S. meliloti have adopted somewhat different, genetically programmed patterns of behavioral response to starvation. Strain JJIcIO appears to be significantly more committed to sustained motility and chemotactic responsiveness than either L5-30 or RMB7201. Perhaps as a consequence, strain JJIcIO loses viability (or enters a viable but nonculturable state) significantly faster than the other two strains under conditions of prolonged starvation, at least in vitro (FIG. 2.1).

Loss of flagella In response to starvation. When S. meliloti cells were starved long enough to bring visible motility to essentially zero, roughly half of the flagellated cells retained only short lengths of just one or two flagella. It is likely that cells with only one or two truncated flagella are incapable of normal, smooth forward swimming. Thus, the progressive loss and disintegration of flagella on many of the cells may help to explain why most of those cells which remained visibly motile only circled, twitched, or tumbled slowly. Frequent or extended pausing by some individual flagella on a cell with several flagella might also result in such aberrant swimming behavior (Gotz and Schmitt, 1987; Sourjik and Schmitt, 1996).

58 However, even in extensively starved suspensions in which no

swimming or even slowly spinning cells were detected, about one-third of the

flagellated cells of all three strains appeared to retain a normal number of

full-length flagella. In electron micrographs, such cells were indistinguishable

from flagellated cells taken from actively growing cultures. We conclude that

the loss and disintegration of flagella occur in just a selected subpopulation

of the cells. The molecular mechanisms behind this selective loss or retention

of flagella in certain subpopulations of this bacterium remain to be

determined. The cell-specific, starvation-induced regulation of flagellar loss

on the one hand and of flagellar-motor activity on the other hand appears to

result in the production of five distinct subpopulations: (i) cells that have lost

all their flagella and are non-motile; (ii) cells that retain normal numbers of

intact flagella and have active flagellar motors; (iii) cells that retain normal

numbers of intact flagella, but have inactive flagellar motors; (iv) cells that

have lost most of their flagella, but have active flagellar motors; and (v) cells that have lost most of their flagella and have inactive flagellar motors. After

prolonged starvation, only the first, third and fifth subpopulations of cells

remain. It would be of considerable interest to know whether the different subpopulations generated by these two kinds of starvation response might each have some kind of advantage in surviving starvation under natural environmental circumstances.

Retention of motility by nutrients or attractants during starvation. The loss of motility of L5-30 cells following transfer to starvation

59 medium could be prevented or minimized for an extended period by the early addition of an attractant to the starvation medium during transfer (TABLE 2.1).

The extent to which loss of motility was prevented seemed to be determined by how strongly the attractant elicited tactic responses, and not by whether the attractant was metabolizable or by its concentration per se (TABLE 2.1).

Substances present in the culture filtrates from stationary phase cultures of

L5-30 or other bacteria had a similar ability to prevent motility loss during starvation. Presumably the active substances in these culture filtrates were non-metabolizable attractants, but other possibilities, including quorum- sensing autoinducers (Eberl et al., 1996) remain open. From these observations, it appears that the presence of chemoattractants can at least temporarily override the normal down regulation of behavioral activity in S. meliloti. Based on the effects of non-metabolizable attractants like cycloleucine (1-amino-1-cyclopentanecarboxylic acid) and itaconic acid

(methylenesuccinic acid), the ability of chemoattractants to override starvation-induced down regulation of motility seems independent of any energy or nutrient input from the attractant. Thus, the interaction of such chemoattractants with their receptors would seem to provide one important input to the regulation of behavioral activity under starvation conditions. A second, independent part of such regulation undoubtedly involves nutrient- energy sensing and its connections to flagellar maintenance and motor activity.

60 100

60 s (D o E 40

20

0 40 80 120 160 200 time after transfer to SB (h)

FIG. 2. 6. Percentage of motile cells of E. coli strains ccl 18 ( •) and HB101 (o) at various times after transfer to SB.

61 In the future, the regulatory mechanisms that determine how a bacterium changes its pattern of behavioral activity when it faces starvation

need to be examined at both the molecular genetic and ecological levels. To this end, we have attempted to isolate putative Tn5 mutants of strain L5-30 by serial enrichment in SB soft agar plates for cells that moved furthest from the point of inoculation after several days. Preliminary work indicated that this approach was only modestly successful in isolating mutants with prolonged motility under starvation conditions. Further characterization of the mutants is

required. Since far more is known about the molecular mechanisms of motility and starvation responses in E. coli and Pseudomonas aeruginosa, we considered whether the needed molecular and ecological studies might be more efficiently pursued in these organisms. A brief examination of the behavioral responses of E. co//strains cc118 and HB101, and P. aeruginosa

PA01 after transfer to SB revealed that motility remained relatively high for several days (FIG. 2.6), indicating that E. coli or P. aeruginosa may not be good models for studying the kinds of behavioral regulation found in S. meliloti, although interesting in its own right. In contrast to E coli, S. meliloti \s reported to have two independent CheY proteins controlling flagellar rotation

(Sourjik and Schmitt, 1996). It would be of interest to learn whether one of these two CheYs, or the balance between them, plays a role in controlling flagellar motor activity in response to starvation.

62 CHAPTER 3

TN5-INDUCED AND SPONTANEOUS SWITCHING OF SINORHIZOBfUM MELILOTI TO INCREASED SWARMING BEHAVIOR

ABSTRACT

Three Tn5 mutants of S. meliloti RMB7201 which swarmed 1.5 to 2.5 times faster than their parent in semi-solid agar, moist sand and viscous liquid were identified by screening 6,000 transconjugants. When these faster

(increased) swarming (PS) mutants and the wild type were mixed at a 1 ;1 ratio and inoculated into swarm agar, the FS mutants outgrew the wild type 30- to

40-fold within 2 d. The only circumstance where the wild type was consistently found to grow or survive better than the FS mutants was in a soil matrix subjected to air drying. The three mutants, FS1, FS7 and FS44, came from independent matings, had similar restriction patterns and swarm behavior, but were phenotypically distinct. FS1 and FS7 showed normal swimming behavior, but FS44 changed direction very frequently. EPS synthesis was reduced in each of the three FS mutants when grown on defined succinate- nitrate medium, but the extent of reduction was different for each. At present, it

63 appears that FS behavior likely results from a modest, general derepression of motility. Each of the mutants had a 10% to 20% higher proportion of motile and flagellated cells than the wild type, and both the average number and average length of flagella per cell were about 1.6-fold higher for the mutants than the parent. Chemotactic responsiveness of the mutants was similar to that of the wild type for a variety of attractants. Reasonably stable spontaneous

FS variants of RMB7201 were obtained at a frequency of about 1 per 10,000-

20,000 cells by either enrichment from the periphery of swarm colonies or by screening of colonies for reduced EPS synthesis on succinate-nitrate plates.

The spontaneous FS variants and Tn5 FS mutants were symbiotically effective and competitive in nodulation of alfalfa. Reversion of FS variants to wild-type behavior was sporadic, sometimes securing at high frequency and sometimes not occuring, indicating that reversion is affected by unidentified environmental factors. Based on differences in swarm rate, swimming behavior, EPS phenotypes, reversion rates and restriction patterns between individual FS variants and mutants, it appears that there may be multiple genetic configurations that result in FS behavior in RMB7201. The facile isolation of what appear to be spontaneous FS variants of Escherichia coli and Pseudomonas aeruginosa indicates that switching to FS behavior may be fairly common amongst bacterial species. Based on the substantial growth advantage that FS mutants and variants have wherever nutrient gradients exist, such switching may be an important behavioral adaptation in natural environments.

64 INTRODUCTION

Although flagellar motility and chemotaxis in bacteria have been studied intensively for more than 20 years, relatively little is known about how or when they actually operate in natural environments or how they are regulated to optimize growth and survival under diverse conditions. We are interested in the role of flagellar motility in the ability of S. meliloti to survive in soil, colonize roots and symbiotically infect and nodulate its host, alfalfa.

Strains of S. m eliloti are motile and chemotactic. Nonmotile and nonchemotactic mutants of S. meliloti are substantially less competitive than the wild type at infecting and nodulating host roots (Ames and Bergman, 1981,

Gaetano-Anolles, et al., 1988). Molecular mechanisms and the regulation of behavior in S. meliloti differ In many respects from those seen in enteric bacteria (Armitage & Schmitt, 1997; Mcnab, 1996). Motility in S. meliloti is based on exclusive clockwise rotation of the bacterium’s 2 to 8 peritrichous,

“complex” flagella, flagella which are more rigid and more efficient in propulsion in viscous media than plain flagella (Gôtz et ai., 1982,-Gôtz and

Schmitt, 1987, Krupski et al., 1985, Pleier and Schmitt, 1991). S. meliloti cells are chemokinetic, swimming at higher speeds when exposed to higher attractant concentrations, with brief pauses or asynchronous flagellar rotation resulting in changes in the direction of swimming (Gotz and Schmitt, 1987;

Pleier and Schmitt, 1991, Armitage and Schmitt, 1997). The bacterium Is attracted towards a variety of amino acids, dicarboxylic acids, and sugars

65 (Ames and Bergman, 1981, Bergman et al., 1988, Robinson et al., 1992),

towards the nodulation gene-inducing flavonoids secreted by roots of its host

(Dharmatilake and Bauer, 1992), and towards unknown attractants secreted at

localized sites on the host root (Gulash et al., 1984).

In two previous studies, isolates of S. meliloti with greater motility or

swarming have been obtained, either by serial enrichment for faster moving

cells from the periphery of soft agar swarm colonies (Krupski et al., 1985 ) or

by Tn5 mutagenesis (Ronco, 1988). The faster swarming isolate obtained by

TnS mutagenesis was not studied further. The isolate obtained by serial

enrichment was found to have greater flagellation than the wild type (Krupski

et al., 1985 ), but the genesis, stability and consequences of this enhanced

swarming behavior were not examined. We thought that mutants with

increased motility and taxis might be valuable tools for investigating the

ecological consequences of altered behavioral activity and might also prove

valuable in developing more competitive inoculants. The studies reported

here examine several Tn5 mutants of S. meliloti RMB7201 that have faster

swarming (FS) behavior. Our studies also reveal that flagellar swarming

behavior of S. meliloti is subject to spontaneous, relatively high-frequency

switching between normal and FS activities, and that FS cells differ quite significantly from wild-type cells in their growth and survival under various conditions.

66 MATERIALS AND METHODS

Bacterial strains plasmids, and media and buffers. Bacterial strains, bacteriophages and plasmids used in this study are listed in TABLE

3.1. Stock cultures were kept in 15% glycerol at -80°C. S. meliloti strains were routinely cultured in TY (containing 6 g of tryptone, 3 g of yeast extract, and 0.5 g of CaClg 2 H2 O per L), or in TY containing reduced concentrations of tryptone

and yeast extract, but maintaining the CaCI^ 2 H2 O at 0.5 g/L. S. meliloti was also cultured in a defined medium (NM) containing mineral salts and vitamins with 20 mM succinate and 5 mM nitrate as the carbon and nitrogen sources

(Robinson et al., 1992). E .coli strains were grown in Luria-Bertani (LB) medium. All chemicals were analytical or reagent grade (Sigma Chemical

Co., Baker Chemical Co.). To improve reproducibility of behavior, all liquid cultures of S. meliloti were started from glycerol stocks, grown at 28°C to late exponential or early stationary phase on a rotary shaker at 175 to 200 rpm, then subcultured in fresh medium. Cells from the subcultures were routinely harvested during early exponential phase (AggQ= 0.15 to 0.30), the period of best motility for S. meliloti. Solid media were prepared by the addition of 1.5% agar (Difco). Semisolid (swarm) agar contained 3 g of Bacto agar per liter and either 1/20 strength TY or 1/10 strength NM medium unless otherwise specified. Motility, chemotaxis, and some swarm assays were carried out in chemotaxis buffer (CB) consisting of 10 mM HEPES, 0.1 mM CaCI^ and

67 0.01 mM sodium EDTA in HPLC grade water. pH 7.0 (Robinson et al., 1992).

CPUs of Tn5 mutants were determined by subtraction of counts on TY medium containing streptomycin (Sm) at 100 p.g/mL, spectinomycin (Spc) at

100 jxg/mL and kanamycin (Km) at 200 jxg/mL from total counts on TY containing just Sm and Spc. The ratio of mutant to parent was verified by testing representative colonies on swarm agar. A spiral plater (Model D, Spiral

System. Inc.. Cincinnati. OH) was used for all CPU determinations. Counts from duplicate or triplicate plates were averaged.

Transposon mutagenesis, screening for behaviorai mutants, and Southern analysis. S. me///of/RMB7201 was mutagenized with Tn5 through filter and plate mating with E. coli WA803(pGS9) (Selvaraj and Iyer.

1983) or with CC118^jj. (pUTmini-Tn5-/acZ) (de Lorenzo et al., 1990). Well- isolated Km resistant colonies were screened on swarm agar for swarming behavioral mutants. Genomic DMA preparation, restriction digestion and

Southern blotting were performed as described (Sambrook et al.. 1989). DNA probes (the internal 3.3 kb H/ndlll fragment from Tn5, and the 5.0 kb EcoRI-

SamHI fragment from mini-TnSlacZ) were labelled either with ^^P-ATP

(Amersham, Arlinton Heights. IL). or biotin-14-dCTP (Life Technologies.

Gaithersburg. MD) according to manufacturer’s instructions. The blots were detected respectively with X-ray film, or BlueGene nonradioactive nucleic acid detection system (Life Technologies).

68 TABLE 3.1. Bacterial strains, phages and plasmids and their characteristics

Bacterial strain Relevant Characteristics Source

S. meliloti RMB7201 wild type, Sm^, Spc^, Q insertion in Bosworth chromosomal inositol gene et al., 1994

FS1 Tn5 derivative of RMB7201 (Kmr), This study increased swarming, less EPS in NM medium

FS7 Tn5 derivative of RMB7201 (Kmr), This study increased swarming

FBI 7 mini-Tn5 lacZ derivative of RMB7201 This study Kmr, increased swarming, less EPS in NM medium

FS27 mini-Tn5 lacZ derivative of RMB7201, This study Kmr, increased swarming, less EPS in NM medium

FS28 mini-Tn5 lacZ derivative of RMB7201. This study Kmr, increased swarming, less EPS in NM medium

FS44 Tn5 derivative of RMB7201(Kmr), This study increased swarming, less EPS in NM medium

SCI Tn5 derivative of RMB7201 (Kmr), This study EPS defective, normal swarming

SV8 Spontaneous increased swarming, This study reduced EPS variant of RMB7201

SV68, SV69 Spontaneous increased swarming, This study reduced EPS variant of RMB7201

EV2 Spontaneous increased swarming, This study reduced EPS variant of RMB7201 (continued)

69 (TABLE 3.1 continued)

EV3 Spontaneous increased swarming, This study reduced EPS variant of RMB7201

EV4 Spontaneous increased swarming, This study reduced EPS variant of RMB7201

FS8 Spontaneous increased swarming, This study reduced EPS variant of RMB7201

Rm1021 Smr derivative of SU47 wild type Meade et al-, 1982

Rm7032 exoA, EPS- Leigh et al., 1985

Rm8468 exoP, EPS- Long et al., 1988

Rm8395 exoR, EPS overproduction Doherty et al., 1988 Rm8396 exoS, EPS overproduction Doherty et al., 1988 Rm7020 exoC, EPS overproduction Finan et al., 1988 Rm7103 exoX, EPS overproduction Zhan et al., 1990

E. coli WA803(pGS9) Tn5 donor Selvaraj & lyer, 1983 CO118xpir Tn5 donor de Lorenzo (pUT mini-T n5lac1) et al., 1990

Pseudomonas aeruginosa

PA01 wild type Holloway et al., 1979; Seed et al., 1995 (continued)

70 (TABLE 4.1 continued)

Plasmids:

pD5 complements exoC Leigh et al.. 1985 pM6 exoR cosmid Doherty et al.. 1988

pM13 complements exoS Doherty et al., 1988

pM13-1 5.3 kb subclone complements exoS Doherty et al., 1988

Motility and chemotaxis. The percentage of motile cells was determined in a Petroff-Hauser counting chamber with a Zeiss IM 35 microscope as previously described (Wei and Bauer, 1998). Motile cell percentages were based on observation of 60 to 100 cells in two replicate suspensions and were generally reproducible within +/-10%. In some cases, recorded video data were analyzed to determine motile cell percentages and motile behavior. Chemotactic responses to attractants were assayed according to Adler (1973) and Palleroni (1976) in horizontal chemotaxis chambers using quadruplicate 1 /y| capillary tubes and 30 min assays. To assure stringent comparison of chemotactic responsiveness to various

71 attractants. carefully matched early log phase cultures of FS mutant and wild- type strains were suspended in CB, the suspensions mixed together and diluted to 1 x lo f cells/ml of each strain, and the cells of each strain entering capillaries determined by differential plating .

Swarming in semi-solid agar, sand or viscous media.Swarm colony development in semi-solid agar typically involved inoculating the center of a plate with 2.5 fjL of a suspension of mutant and/or wild type. The plates were then sealed with Parafilm and incubated at 28°C. For analysis of swimming behavior and swarming in viscous liquid medium. 0.4% high viscosity carboxymethylcellulose (CMC) was mixed with an equal volume of

1/10 strength TY. The rate of spreading of swarm colonies in CMC was determined by microscopic detection of cells at the outer edge. To measure swarming in sand, washed quartz sand in a Petri dish was moistened with

1/20 TY to field capacity, inoculated at the center of the plate with 2 jil of bacterial suspension and incubated for 1-3 d at 28°C. Spreading of the swarm colony in sand was determined by replicating onto nutrient agar with a multiprong replicator. To measure the rate of swarm colony spreading in CB swarm agar, agar blocks were removed from points at various distances from the center 3 d after inoculation, homogenized and plated.

Detection and quantification of EPS. The amount of EPS produced by S. meliloti RMB7201 and the FS mutants on soiid NM succinate- nitrate medium was roughly estimated by direct observation of colonies on

72 plates, commonly after two or three days of incubation. To measure the

quantity of EPS produced in liquid NM-succinate medium, cells were removed

by centrifugation of stationary phase cultures. EPS was precipitated from the

supernatants with 3 volumes of ethanol, then rinsed and dried. EPS was then

either weighed directly or suspended in water by thorough sonication and

carbohydrate content determined by the phenol-sulfuric acid method (Robyt

and White. 1987) with glucose as the standard.

Preparation of flagella, SDS-PAGE and Western blotting.

Cells from mid exponential phase cultures were pelleted at low speed

(2000xg), resuspended in CB, and blended in a Waring blender (model

33BL79, Dynamics Cor. of America, New Hartford, CT) at full power for 20

seconds to shear flagella from the cells. Cells were removed by centrifugation

at iS.OOOxg for 15 min. The supernatant containing detached flagella was

centrifuged at 100,000xg for 2 h and the pelleted flagella suspended in CB.

SDS-PAGE analysis of flagellin preparations were performed as described

previously (Robinson et al., 1992), and Western blot as described elsewhere

(Tuhela et al., 1998).

Electron microscopy. Flagella and flagellated cells were observed with a Phillips 201 transmission electron microscope using the direct staining without rinsing method described previously (Wei and Bauer. 1998). Cells for

EM examination were harvested in early exponential phase, resuspended in

CB containing 0.5 mM HEPES with a 3-fold dilution. Five |xl aliquots of the cell

73 susepnsions were mixed with 1 fxl of 2% uranyl acetate and 2 jil of the mixture were transferred to a Formvar-coated grid and air dried. The percentage of

cells with flagella was determined by examination of at least 100

representative cells on two or three independently prepared grids per sample.

The average number and length of flagella determined from EM micrographs of at least 50 cells per strain with a planimeter (Graphics Calculator, Numonics

Co., Lansdale, PA).

Nodulation of alfalfa. Nodulation tests in growth pouches were done as previously described (Bhagwat et al., 1992). Nodule occupancy by

mutant or wild-type strains was determined on 25 to 50 nodules taken from the oldest part of the primary roots of 5 to 40 plants by plating homogenates of surface-sterilized nodules onto selective media and confirming the swarm phenotype of representative colonies on semisolid agar.

Soil drying tolerance. Relative growth and survival of FS mutants and parent during gradual drying in a soil matrix was examined in a sterilized silt-clay soil fraction. This silt-clay material was prepared by sieving fresh silt loam soil through 7 mesh screen, suspending 600 g field-moist soil in 3 liter of tap water, and decanting after 10 min settling to remove floating debris and fine clay. Décantation was repeated twice before blending the suspension in a

Waring blendor in 500 ml portions for 30 sec. The blended suspension was passed through an 80 mesh screen to remove large sand particles, then through a 325 mesh screen to remove particles larger than 45 microns. The silt-clay suspension was then washed 3 times by 20 min. settling and

74 décantation to remove fine clay particles, autoclaved twice for 60 min, then washed twice with sterile tap water before resuspending in a total volume of

600 ml. This suspension was dispensed by pipet into scintillation vials and dried at 100°C for 24 h. Mid-exponential cultures of an FS mutant and the parent were washed and resuspended in CB containing 0.5 mM HEPES, mixed at a 1:1 ratio, then inoculated into the dried silt-clay material to about

10® CPU per strain per gram of dry weight (gdw) at a moisture level of about

75% of field capacity. The vials were kept open in the lab and allowed to dry at approximately 23°C and 50% RH. In a subsequent experiment, the vials were placed in a plastic box with dishes containing saturated K^SO^ solution to maintain relative humidity at 95%. GPU of mutant and parent were determined at various times after inoculation by extracting cells in 3 volumes of sterile water, vigorously vortexing for 1-2 min, then sonicating the suspension at 25% energy output in a Heat System-Ultrasonics, Inc. model w-370 cup horn sonicator for 3 min. After allowing particles to settle for 3 min, the suspension was transferred to a screw-capped 50 ml tube and the sediment extracted 3 more times. The cell recovery rate was 90 to 95%.

RESULTS

Isolation of Faster swarming (FS) mutants. Screening of over

6,000 purified isolates from several Tn5 matings of S. meliloti RMB7201 with

75 E. CO//WA803(pGS9) on 1/20 TY semi-solid agar plates for altered swarm colony development allowed identification of a variety of behavioral mutants.

Nonswarming mutant isolates, with cells which were nonmotile or had severely impaired motility in liquid culture, were obtained at a frequency of 3.0

X 10'^. Isolates with reduced swarm rates and cells that microscopically appeared to have either impaired motility or a lower percentage of motile

cells, or had normal motility but impared chemotaxis, were obtained at a frequency of 3.6 x 10"^. Three mutants, FS1, FS7 and FS44, each from a

different mating, swarmed at rates significantly faster than the parent. The frequency of recovery of FS mutants was 4.5 x 10“^.

Swarming behavior of FS mutants. The FS mutants develop larger swarm colonies than the wild type in semisolid agar media. FIG. 3.1,

Panel A shows typical colony morphology of the strains. TABLE 3.2 shows the relative rates of swarm colony spreading for the wild type and the three FS

mutants. Even in CB agar plates, where nutrient availability was very low and one might expect the swarming capability of the wild type to be fully derepressed, the FS mutants swarmed significantly faster than the parent.

Cells of the FS mutants were present in greater numbers at all locations in these plates, from the center liquid media and through a moist sand matrix, indicating that their FS behavior was not dependent on interactions with a semi-solid agar latice.

76 B

FIG. 3. 1. (A) Swarm colony size and morphology of S. meliloti RMB7201 wt (top left) and fasrter swarming mutants FS1 (top right). FS7 (bottom left) and FS44 (bottom right) in 1/20 strength TY, 0.3% agar swarm plates. 2.5 ul of starter culture was spotted on swarm plates, which were then sealed with Parafilm and incubated at 280G for 40 h. (B) EPS production on solid NM medium by RMB7201 WT, FBI, FS7 and FS44 (same positions as above).

77 Nutrient concentration was found to affect the rate of spreading of both

FS1 and the wild type (FIG. 3.2). However, the wild type was more sensitive to nutrient concentration than the mutant so that, the higher the nutrient concentration, the larger the difference in rate of swarm colony growth between mutant and parent.

Swimming behavior and chemotactic responsiveness of FS mutants. Microscopic analysis revealed that FS1 and FS7 had a pattern of swimming behavior not readily distinguishable from that of the RMB7201 parent, with fairly straight swims interrupted by abrupt changes in direction

(Wei & Bauer. 1998). The percentage of cells that showed active motility was consistently a little higher for the mutants than the parent, and the parent seemed to have a higher percentage of cells that changed directions at high frequency. In contrast to the parent and the other FS mutants, FS44 was observed to change directions at high frequency in liquid media, with no long swims. Interestingly, in viscous medium (0.2% CMC), FS44 did swim long distances without frequent changes in direction.

The chemotactic responsiveness of FS1 to glutamine, succinate, glucose and alfalfa seed exudate was similar to, though perhaps 20-30% higher than that of the wild type (TABLE 3.3). Further assays were conducted to determine chemotactic responses toward other attractants, including phenylalanine, tryptophan, aspartic acid, sucrose, mannitol, cycloleucine, itaconic acid, 4,7-dihydroxy-flavone, and p-hydroxy benzoic acid.

The responsiveness of the wild type, FS1, FS7 and FS44 towards these

78 70 3.0

60 < A ra ti 2.5 I E 50 •FS1 co E U_ u . Vfc- o 2.0 '+ - » E co co 30 k_ (D 7201 wt CO _o 20 o u co 10 $ C/î

0 0 5 10 15 20 25

medium conc (%TY)

FiG. 3. 2. Relationship between nutrient concentration and the swarm rate of S. meliloti RMB 7201 wt (o) and mutant FS1 (•), and the ratio of FS1 swarm rate to wt (A). The two strains were separately spotted into swarm agar containing 0.3% agar and 1/100, 1/40, 1/20, 1/10 or 1/5 strength TY (bactotryptone yeast extract medium). Swarm colony diameters were measured after 2 d incubation at 280C. Each data point is average of 4 replicates. The data shown is from one representative experiment. 79 TABLE 3.2. Rates of spreading of S. meliloti RMB7201 and Tn5 FS mutants In various media.

Medium Soreadlno rate fmm/hl +/- SDa RMB7201 FS1 FS7 FS44

1/20 TY, 0.50±0.019 1.0±0.018 1.0±0.018 1.0±0.012 0.3% agar

Soil extractb, 0.35±0.022 0.50±0 0.50±0 NDc 0.3% agar

CB. 0.29±0 0.40±0.036 0.40±0.036 0.37±0.036 0.3% agar

1/20 TY, 0.64±0.039 1.2±0.076 1.1 ±0.073 0.92±0.14 moist sand

1/20 TY In 3.0±0.43 4.6±0.46 4.6±0.29 3.6±0.31 0.2% CMC solution a Values are averages taken from duplicate plates In a representative experiment after 40 h, except for 1/20 TY, CMC-amended liquid medium, where rates were determined 6 to 9.5 h after Inoculation, b Soil extract solution was obtained by adding 1 liter of tap water to 500 g of soil (Crosby sand loam), mixing, centrifuging at 5,500xg for 15 min and 34,000xg for 15 min to remove particles. The soil solution was then concentrated 9 fold In a rotary evaporator, and finally filter sterilized through 0.2 micron membranes, c Not determined.

80 TABLE 3.3. Chemotactic responsiveness of S. meliloti RMB7201 wild type (WT) and mutant FS1.

Attractant Chemotaxis ratios (concentration) RMB7201 WT FBI glutamine 37±5.8 43±5.2b 10 mM glucose 1.6±0.28 1.8±0.2 30 mM succinate 3.2±0.8 3.7±0.5 30 mM alfalfa seed 93±21 125±17 exudate (1/4) alfalfa seed 53±16 68±4.7 exudate (1/16) alfalfa seed 18±5.8 22±1.9 exudate (1/64) a The chemotaxis ratio is the average number of cells entering attractant-filled capillaries divided by the average number of cells (ca. 3,000 per 30 min) entering buffer-filled control capillaries. Ratios are based on average CPU from 4 replicate capillary tubes, and are from a single experiment that has been repeated with similar results. Che ratio ± sample standard deviation.

81 TABLE 3.4. Chemotaxis response of S. meliloti RMB7201 wild type and mutants towards common chemoattractantsa.

Attractant Che ratio RMB7201 FS1 FS7 FS44 Tn5RM10b

Glutamine (10 mM) 29 43 41 77 34

Phenylalanine 35 22 60 20 15 (1 mM) tryptophan 25 16 32 26 40 (1 mM) aspartic acid 5.8 3.8 10.4 4.9 3.1 (10 mM) glucose (10 mM) 1.38 2.15 1.85 3.7 1.32 sucrose (10 mM) 1.6 3.24 5.3 3.95 1.9 succinate (10 mM) 1.68 5.4 2.36 3.1 NO mannitol (10 mM) 2.46 1.72 1.48 2.16 1.36 cycloleucine 42 17 17 37 6.26 (5 mM)

Itaconic acid 0.72 1.25 1.38 1.72 0.7 (10 mM)

Dihydroxy- flavone 2.0 0.52 3.1 2.36 2.3 (1 mM) p-hydroxy benzoic 0.78 0.46 1.4 0.54 0.87 acid (1 mM) a Owing to the variations between replicates, che ratios of the wild type are not different from those of the mutants at 95% confidence level, b strain Tn5 RNM-10 is a reduced motility mutant. ND=not determined

82 TABLE 3.5. Competitive growth of a 1:1 mixed inoculum of FS1 and the RMB7201 wild type in semi-solid agar, shake and still cultures.

Culture medium and conditions ______Ratio of FBI :WT CFUa

1/10 TY, 0.3% agar swarm plateb 29±6.5c

1/20 TY, 0.3% agar swarm plateb 33±9.6

1/40 TY, 0.3% agar swarm plateb 39

CB, 0.3% agar swarm plateb 10±1.23

1/10 TY liquid, shaken @ 200 rpm for 48 h 1.0±0.12

1/10 TY liquid, still for 5 days 1.83±0.37 a Results are averages from triplicate cultures of a representative experiment. t> After 3 d, the semisolid agar from each plate was vortexed vigorously in a tube prior to plating, c ± sample standard deviation.

83 compounds was not significantly different from each other at a 95% confidence level (TABLE 3.4).

Growth and survival of FS mutants in vitro. Various media were

Inoculated with a 1:1 mixture of FS1 and the wild type, cultured as indicated below, and growth of the cultures determined by seletive plating. The ratio of

FS1 to wild-type cells obtained by selective plating was also confirmed by swarm testing of 50 to 100 colonies. In liquid shake culture on 1/10 strength

TY, where motility can provide no significant advantage, growth of the wild type and the FS mutant were essentially identical (TABLE 3.5). However, in still cultures on the same medium, the number of FBI mutant cells was 1.5 fold higher than the wild type at stationary phase. The FS mutants had an even greater advantage over the parent in semi-solid agar, outgrowing the parent by 30-40 fold in 2 to 3 d.

When FBI and the wild type were incubated separately in still cultures of SB, or in unfiltered pond water (Mirror Lake, OSD campus) for simulation of natural conditions, their survival rates were very similar. CFU of both the parent and FBI after 3 months incubation in the CB “starvation" buffer were only 2 to 3 fold lower than the CFU recovered immediately after inoculation.

CFUs recovered from unfiltered pond water after 4 months were 10'^-fold lower than immediately after inoculation, but the ratio of the wild type to FBI was essentially unchanged (data not shown).

To date, the only circumstance where the wild type has shown a significant growth or survival advantage over the FBI mutant has been in a

84 TABLE 3.6. Survival of S. meliloti RMB7201 (wild type) and increased swanning mutants FS1 and FS7 during drying in silt-soil matrixa.

Strains Day ia Day 3 Day 5 Day 11

CFUb survival% CFU survival% CFU survival% CFU survival%

wt 1.1X107 83 48600 0.40 34900 0.30 4860 0.04 W t& F S I FS1 7.3x106 59 14000 0.13 6800 0.063 530 0.005

wt 2.3x106 16 53200 0.37 20400 0.14 1600 0.011 Wt&FS7 FS7 2.0x106 15 15200 0.12 2600 0.02 130 0.001 a The moisture content of the silt-soil matrix was 8.75%, by weight on day 1, and on days 3, 5 and 11 was below 0.8%. b Values are averages from duplicate samples from a single experiment. CFU per vial = CFU/1.6 g of dry silt matrix.

85 sterile sllt-clay soil that was allowed to dry extensively. Both FS1 and FS7

died out several times faster than the wild type (TABLE 3.6).

In a similar experiment where drying was slower, the populations of

both the parent and FS1 increased more than 100-fold during the first two

weeks, the parent more extensively than the mutant. Then the populations

declined and stabilized at densities of about 10^ cells/gdw, with FS1

populations 7 or 8 times lower than the parent (data not shown).

Flagellation and motility of FS mutants. As shown in TABLE 3.7,

the percentage of flagellated and motile cells was slightly (1.1 to 1.2 fold)

higher for the FS mutants than for the parent. Based on electron microscopic

comparisons, the FS mutants had a 1.6-fold greater average number of

flagella per cell and a 1.6-fold greater average flagellar length (TABLE 3.7,

FIG. 3.3). Thus, overall, EM analysis indicated that the amount of flagellin

attached to FS mutant cells was about 1.2 x 1.6 x 1.6 = 3 times greater than

the amount of flagellin attached to wild-type cells. SDS-PAGE and Western

blot analysis (FIG. 3.4) confirmed that the FS mutants had approximately 2 to 3

times as much flagellin attached to the cells and free In the culture medium as

the parent. EM photographs also showed visible difference in flagellar number

and length between the wild type and FS1 (TABLE 3.7, FIG. 3.3). Growth of

the bacteria at low rates of shaking (50 rpm) had no discernible effect on the percentage of motile or flagellated cells or on the average length of flagella attached to cells. Indicating that shearing of flagella during growth or sample preparation was not a problem affecting the assessment of differences in

86 TABLE 3.7. Motility and flagellation of RMB7201 and T n5 FS mutants.

Strain Time after resusoension in CBa

0 hours 4 hours % Motile % Flagellated #Flagella Length % Motileb %Flac cellsb ceiisc percelM (nm)e

RMB7201 42 47 1.9 2.4 33 44

FBI 55 57 2.4 3.8 43 55

FS7 53 46 NDf ND 35 44

FS44 53 51 ND ND 35 46 a Cells from log phase 1/10 TY cultures were washed and gently resuspended in OB. b Percentages are averages of counts from several fields and two independent cultures from a representative experiment, c Percentages are averages of counts from about 60-100 cells on two separate grids from the same two independent cultures used for motile cell percentage assays. d Average number of flagella per flagellated cell, determined from EM micrographs of at least 100 cells per strain on replicate grids, e Average length of flagella per flagellated cell, determined by planimetry from micrographs of at least 50 cells per strain on replicate grids, f ND = not determined

87 % %

FIG. 3.3. Flagellation patterns of S. meliloti RMB7201 wild type (left) and Tn5 mutant FS1 (right). Cells were negatively stained with 2% uranyl acetate and observed under EM (see Materials and Methods section for details).

88 B

1 2 3 4 5 6 7 8

FIG. 3.4. Amounts of flagellin produced by S. meliloti RMB7201 wild type and mutant FS1, determined by SDS-PAGE and quantitative Western Blot. See Materials and Methods section for details on flagellar isolation, PAGE and Western Blot. Panel A: flagellin stained with Coomassie Brilliant Blue R-250. Lane 1 is protein molecular wt standard, lanes 2 & 3 are sucrose-gradient purified flagellin as control, lane 4 is RMB7201 wt, and lane 5 mutant FS1. Panel B: Flagellin reacted with monoclonal anti-flagellin antibody. Lanes 1 & 2: RMB 7201 wt early log -phase; Lanes 3 & 4: RMB 7201 wt late log -phase. Lanes 5 & 6: FS1 early log -phase; Lanes 7 & 8: FS1 late log -phase.

89 flagellation. After exposing the FS mutants to semi-starvation conditions in CB for several hours, both motiliy and flagellation were reduced, with motility reaching zero in about 24 h. Motility of the FS mutants and the parent was affected to comparable extents under semi-starvation conditions. Motility diminished more quickly and extensively than flagellation (TABLE 3.7), indicating rapid arrest of flagellar motors (Wei and Bauer, 1998).

EPS production. Colonies of the RMB7201 wild-type strain are large and mucoid on both TY agar and NM succinate-nitrate minimal agar.

However, on NM succinate-nitrate agar, colonies of FS1 and FS44 were considerably smaller than those of the wild type, and relatively dry in appearance (FIG. 3.1 B), indicating that these mutants might be defective in

EPS synthesis on this medium. Based on two experiments with duplicate samples, mutants FS1, FS7 and FS44 produced 41%, 68% and 18% as much EPS in liquid NM succinate-nitrate medium, respectively, as the wild type, although they produced the same amounts of EPS as the wild type in TY liquid medium (with a delay of 1 day for FS1 and FS44). On plates containing

Calcofluor white M2R, colonies of the three FS mutants fluoresced strongly under UV light, indicating that they all produced EPSI (Leigh et al., 1985). In order to find out whether the physical presence of EPS affects swarming rate, we isolated an EPS defective Tn5 mutant of RMB7201, designated as SCI.

This strain swarmed at the same rate as the wild type, indicating no direct relation exists between EPS presence and spreading speed.

90 To further explore possible links between regulation of EPS synthesis and regulation of swarming motility in S. meliloti, the rate of swarming of several well characterized EPS overproducing and non-producing mutants of

S. meliloti strain 1021 was studied. Strains Rm7095 (Rm8395) and Rm7096

(Rm8396) are EPSI overproducers caused by insertional mutations in the regulatory genes exoR and exoS (Doherty et al., 1988; Reuber et al.. 1991).

These strains were essentially non-swarming in 1/10 NM semisolid agar for the first 2 d after inoculation. After 2 d, swarm colonies of Rm7095 (in 1/10 NM swarm agar) and Rm7096 (in 1/20 TY semisolid agar) were half as big as those of the wild type. When complementing plasmids pM6 and pM13-1

(Doherty et al., 1988; Reuber et al., 1991) were introduced into Rm7095 and

Rm7096 respectively, the strains swarmed normally. Strain Rm7103 (Zhan and Leigh, 1990), an exoXEPSl overproducer, swarmed at a rate about 20% lower than the wild type. Both of the EPS non-producing mutants (Rm7032 and Rm8468, TABLE 3.1) tested swarmed at the same rate as the parent.

Isolation of spontaneous FS variants and additional Tn5 FS mutants. Our initial molecular analysis of the FS1, FS7 and FS44 mutants raised some uncertainty about the role of the Tn5 insertions in causing the observed FS phenotype. Southern analysis of FS1, FS7 and FS44 revealed that all three mutants had identical restriction patterns, indicating that these three insertions were located at essentially the same site (CHAPTER 4).

However, homologous recombination of the Tn5-containing sequence from

FS1 back into the parent did not yield mutants with either the FS phenotype or

91 altered EPS synthesis, even though the recombinants appeared to have Tn5

inserted In the same location as FS1 (Wei and Bauer, unpublished). In order

to more rigorously test the correlation between FS-EPS behavior and Tn5

insertions at this locus, as well as to identify other possibly critical sequences,

we isolated additional Tn5 FS mutants.

After mating RMB7201 with E. coli CC118,^j^(pUT minl-Tn5/acZ;, the

mating mixtures were cultured on NM succinate-nitrate containing Sm, Spc

and Km, then spotted onto the centers of NM swarm agar plates to enrich for mutant isolates with increased swarming behavior. Based on the growth advantage of the original FS mutants over the parent in swarm agar (TABLE

3.5), cells isolated from the outer edges of such swarm colonies should be enriched roughly 30 to 100 fold for mutants with an FS phenotype. After allowing the swarm colonies from the Tn5 matings to develop for 6 days, cells were transferred from the outer edges of the visible swarm colonies to the centers of fresh NM swarm agar plates. A second swarm colony was allowed to develop for 6 days. Cells were then isolated from the outer edges of these swarm colonies, plated onto selective agar, and purified for phenotypic testing. About 75% of the 117 randomly selected single colony isolates recovered for testing from the mating enrichment plates showed an FS phenotype similar to that of FS1, FS7 and FS44 (TABLE 3.8). The remaining

25% had a swarm phenotype indistinguishable from the parent.

Of the 32 representative isolates selected for further analysis, 28 isolates (from three different matings) were found to be similar to FS1 and

92 FS44 with respect to both rate of swarm colony spreading and reduced EPS synthesis on NM succinate-nitrate. One of the isolates (EV19) was found to be similar to FS7, having normal EPS synthesis on NM succinate-nitrate, but increased swarming rates like FS1. The three remaining isolates had both normal swarming arxj normal EPS synthesis and were presumably wild type.

Southern analysis revealed that only 4 of the 32 strains above had sequences homologous to the miniTnS probe (CHAPTER 4. FIG. 4.3). These four strains, designated WT13, FS17, FS27 and FS28, were the only ones out of the 32 tested which showed p-galactosidase activity corresponding to activity of the

/acZ gene carried in the mini-Tn5/acZ construct.

The great majority (25 of 29) of the isolates obtained from the mini-Tn5 mating enrichments had the FS phenotype, but did not contain mini-Tn5 insertions. Thus, it seemed possible that these non-Tn5 containing, FS isolates from the enrichments were spontaneous variants of the wild type. In order to test the possibility that wild-type cells might spontaneously switch to the FS phenotype at some modest frequency during growth, the RMB7201 wild type was enriched twice on NM succinate-nitrate for fast swarming cells, following the procedure used for enrichment of the mating mixtures above.

Isolates from the periphery of the second RMB7201 swarm colony were purified and tested for swarming and EPS phenotypes. As shown in TABLE

3.8, of the six colonies randomly selected for phenotypic testing, two isolates

(EV2 and EV4) swarmed 1.8 times faster than the wild type (i.e. similar to

FS1), one isolate swarmed at wild-type rates, and the other 3 Isolates

93 TABLE 3.8. Mlni-Tn5 FS mutants and spontaneous FS variants obtained by serial enrichment or direct screening of RMB7201 wild type.

Representative Isolation method Phenotvoe Isolates % of total

Serial Normal swarm rate, EV1 25-50% enrichments normal EPS

Increased swarm rate EV3.EV5, EV6 30-50% (>1.4 -1.6< X wt) and reduced EPS synthesis

Increased swarm rate FS17, FS27 . 30-70% ( > 1 .8 X wt) and FS28, EV2, EV4, reduced EPS synthesis EV8

Increased swarm rate, EV19 <1% normal EPS synthesis

Direct Reduced EPS, 93% screeningb normal swarm rate

Reduced EPS, SV8 1 % increased swarm rate (>1.6 X wt rate)

Reduced EPS, SV68, SV69 2% increased swarm rate (>1. 4-1.6 < X wt rate)

Increased EPS, SV2-5 4% normal swarming a Isolates were obtained from three independent experiments after serial enrichment as described in the text. b Isolates identified after screening approximately 254,000 colonies in two separate experiments.

94 swarmed at rates between 1.3 and 1.5 times faster than the wild type. When this enrichment experiment was repeated, two of the four independent enrichments yielded isolates that were all like FS1 in swarm rate. The other two enrichments yielded isolates that were mostly (65%) like the wild type with others (35%) that swarmed about 1.4 fold faster than the wild type. All FS variants had reduced EPS synthesis on NM succinate-nitrate plates.

The results in TABLE 3.8 indicate that the RMB7201 wild-type strain usually, but perhaps not always, forms FS variants during swarm colony development. Additional experiments were conducted to determine whether such FS variants formed spontaneously at appreciable frequencies in liquid cultures. Since screening many thousands of wild-type colonies for FS variants on semisolid agar would be very laborious, an indirect approach to identification of FS variants in liquid cultures was adopted, plating log phase cultures of RMB7201 onto NM succinate-nitrate and looking for any small, dry colonies corresponding to the reduced EPS phenotype that has been commonly associated with FS behavior.

In an initial screening of about 12,000 colonies of the RMB7201 wild type, 9 colonies with apparently reduced EPS production on NM succinate- nitrate were identified and subsequently tested for swarming behavior. One of these reduced EPS isolates (SV8) formed swarm colonies that spread at same rate as FS1 (TABLE 3.8). SVS showed normal loss of motility after transfer to starvation conditions, but, like FS44, displayed a very high frequency of change in swimming direction in liquid culture. The remaining 8

95 isolates had swarming behavior like the wild type, indicating that they were

EPS variants, but not swarming variants of the parent. When 240,000 wild type colonies on NM succinate-nitrate plates were examined in a similar experiment, colonies with apparently reduced EPS synthesis were identified at a frequency of about 1 per 2,600 (91 of 242,000). As shown in TABLE 3.8,

25 of the 91 spontaneous, reduced EPS variants swarmed at rates faster than

RMB7201 during the initial testing, although none swarmed as fast as FS1. In subsequent swarm tests of single colony isolates from streaks of the 91 reduced EPS variants, only 15 isolates were found to have swarm rates significantly greater than the wild type. When liquid subcultures from single, streaked colonies of the 91 reduced EPS isolates were tested, only 2 isolates

(SV68 and SV69) retained FS behavior, swarming at rates about 1.4 times faster than the wild type. These differing swarm test results indicate that the FS behavior of many of the isolates was not stable during subculture. The FS behavior of SV8, SV68 and SV69 has proven quite stable in repeated tests.

Based on these studies, the initial frequency of switching from normal swarming and EPS synthesis to increased swarming and reduced EPS synthesis was approximately 1 in 10,000 to 20,000. Switching to increased swarming without concomitant switching to reduced EPS synthesis would not have been detected.

Reversion to wild type behavior.Reversion of the FS-reduced

EPS variants to normal swarming and EPS synthesis was examined by both direct and indirect assays. In the first indirect assays, the FS variant SV8 was

96 plated onto NM agar to determine whether any large, mucoid colonies like the wild type developed amongst the smaller, dry colonies characteristic of SV8.

In the first two experiments, involving 5,000 and 37,000 colonies, repectively, no mucoid colonies were detected. In the third experiment, involving 56,000 colonies, only 2 of the 80 plates had mucoid colonies, yet almost 5% of the colonies on these 2 plates were mucoid. Cells from each of the mucoid colonies swarmed at normal, wild-type rates, not the high rate typical of SV8.

Thus, the mucoid colonies appear to be true revertants to normal swarming and EPS synthesis. Cells from the dry colonies near the muciod colonies on these plates swarmed at rates typical of SV8. Most of the 80 plates had 200-

300 colonies per plate, but no mucoid variants. The two plates with mucoid revertants had fewer colonies per plate (ca 50) than most (though not all) of the others. To pursue the possibility that cell density might be a factor in triggering or enhancing the rate of reversion, SV8 was plated at densities ranging from about 75 cells per plate to about 3,000 cells per plate, with 3 or 4 plates at each density (Expt. 4). Mucoid colonies that swarmed at wild-type rates were detected on almost every plate. While there was some variation in the percentage of mucoid colonies on these plates (from about 2% to about

5%). there was no correlation between the frequency of mucoid colony formation and initial plating density. When cells from representative mucoid colonies were purified by single colony isolation and cultures from them retested, the swarm rates were typical of the wild type, not SV8, and both the mucoidy and the swarming phenotypes proved stable.

97 When FS1. FS17. FS44, EV1, EV2, EV3, EV4 and EV8 were plated onto NM succinate-nitrate at a density of about 500 cells per plate and screened for large mucoid colonies, no revertants to normal swarming were detected amongst 1,100 to 20,000 colonies per strain (data not shown).

Likewise when colonies of these strains and SV8 were tested directly for reversion to wild-type swarming on semi-solid agar, no revertants were detected amongst 200-600 total colonies of each strain in either of two experiments (data not shown). Thus, reversion of FS isolates to wild-type behavior at 1% - 5% frequencies is uncommon, at least under the in vitro conditions tested.

In order to establish that the spontaneous FS variants, mutants and revertants described above were indeed S. meliloti derivatives, isolates were inoculated onto alfalfa roots in growth pouches. These isolates all formed normal numbers of normal nodules. This establishes that these isolates are derivatives of S. meliloti, not some contaminating species, and that these FS mutants and variants are symbiotically infective and effective. When FBI and the wild type were co-inoculated onto alfalfa roots at low inoculum dosages to test the efficiency of nodule initiation (Bhuvaneswari et al., 1981 ; 1983), it was found that FS1 generated 70%, 90% and 100% as many nodules as the the parent at inoculum dosages of lO'^/plant, 10^/plant and 10^/plant, respectively

(data not shown). Randomly selected nodules generated by various FS isolates were crushed to determine the swarm and EPS phenotype of the nodule occupants. In the first experiment, putative revertants that swarmed like

98 the wild type were found to occupy 2 of 5 nodules generated by EV8, 0 of 5

nodules generated by SV8, 3 of 5 nodules generated by FS17, and 4 of 5

nodules generated by FS28. All 5 of the nodules generated by FS27 were

oocupled by isolates with swarm rates intermediate between FS27 and the

wild type. From these results it appeared that reversion of FS mutants and variants to wild-type swarming was common during passage through the host

plant. However, in a second experiment, there was no evidence of occupancy by revertants of SV8, EV2, EV4, EV8, FBI, FS7, FB I7, FB27, FB28 or FB44

any of the 10 to 70 nodules examined for each inoculum strain. In a third experiment with EV10, a new FB variant isolated through enrichment,

reversion was detected from three nodules out of 20 tested. This indicates that interaction with the host plants may increase reversion frequency. Growth rate of FS isolates obtained by enrichment or indirect screening. FB17, BV8 and EV69 were chosen to represent the new FS-less EPB mutants and spontaneous FB variants for comparison of growth rates In competition with the wild type under various conditions. As shown in TABLE 3.9, both FBI7 (a miniTnS/acZ mutant) and BV8, a spontaneous FB variant, behaved like FBI (TABLEB 3.5 & 3.9) in growth relative to the wild type in shake, still and swarm agar cultures. These FB strains, like FBI, grew at essentially the same rate as the wild type in shake cultures. While both FBI 7 and BV8 had a 35- to 40-fold growth advantage during 3 d swarm colony development, similar to the growth advantage seen previously for FBI (TABLE 3.9), FB variants BV68 and SV69, with

99 TABLE 3.9. Competitive growth of a 1:1 mixed inoculum of FS17, SV8 or SV69 with the RMB7201 wild type in semi-solid agar, shake and still cultures

Culture conditions Strain Number of FS cells per WT cells

Shake cultured FBI 7 1.0±0.03 in 1/10 TY SV8 1.0±0.04

Still cultures FBI 7 1.6±0.10 in 1/10 TY SV8 1.6±0.09

0.3% agard FS17 30±4.0 1/20 TY SV8 35±2.66

SV68 27±3.95

SV69 20±1.66

EV3e 36±3.30 a Numbers are averages of three replicate cultures from a representative experiment. The relative number of wild type and FS isolate colonies were determined by differential antibiotic sensitivity, colony morphology, and swarm test for FBI and FBI7, and by colony morphology and swarm agar tests, for EV3, BV8, BV68 and BV69. Swarm test was done on at least 100 colonies. The ratio of CFU for the competing strains was corrected for the ratio of CFU for the two strains in the inoculum. b Replicate cultures with 15 mL of medium were inoculated with about 107 cells of the FB strain and wt, cultured at 200 rpm at 280C for 2 d to stationary phase, then plated to determine relative CFU. c Replicate cultures with 6 mL of medium in a 18x150 mm culture tube were inoculated with about 106 cells of the FB strain and wt, cultured at 280C undisturbed for 5 d to stationary phase, then plated to determine relative CFU. d Replicate swarm agar plates were inoculated with about 106 cells of the FS strain and wt, cultured for 3 d, then the agar vortexed vigorously for 2 min and the suspension plated for determination of relative CFU. e The swarm rate of EV3 colonies was about 1.8 fold that of the wild type during tests to determine the ratio of the two strains in the 3 d swarm colony, faster than its normal swarm rate of 1.4 fold faster than wild type.

100 intermediate swarm rates (ca. 1.4 times faster than wild type), had a lower growth advantage (20 to 27-fold) in swarm agar than the faster swarming FS1,

FS17 and SV8 (35 to 40 fold). EV3, another FS isolate with intermediate swarm rate, had a large (36-fold) growth advantage in swarm agar, like that of

FS1. However, we observed that when cells from the 3 d swarm colonies were plated to determine the ratio of EV3 to wild-type cells and representative

colonies tested for swarm behavior, they swarmed at rates equal to FBI and

SV8 rather than intermediate rates like EV3. This indicates that EV3 has

spontaneously switched from intermediate to FS configuration during culture.

DISCUSSION

Reasonably stable FS variants of S. meliloti RMB7201 were readily

isolated from the periphery of growing swarm colonies (TABLE 3.8). Indeed,

after two cycles of enrichment, the majority of RMB7201 isolates from the

periphery of NM swarm colonies displayed FS behavior. In addition,

reasonably stable FS variants were reproducibly obtained (TABLE 3.8) by screening of individual colonies for diminished EPS synthesis on NM agar, a trait associated with FS behavior in almost all of the FS isolates of RMB7201.

The ability to identify FS variants based on EPS phenotype is important

because it demonstrates that the formation of FS variants in a population of

RMB7201 cells occurs independently of growth in a swarm colony or the

presence of sustained nutrient gradients.

101 Based on screening of approximately 3x10^ Individual colonies for reduced EPS variants and subsequent testing of these for FS behavior, we estimate that the formation of FS variants occurs spontaneously at a frequency of about 1 in 10,000 to 20,000 cells. This estimate seems consistent with the frequency of FS variant formation roughly calculated from the 33-fold growth advantage for FS variants on swarm plates (TABLE 3.5), the 50% to 75% recovery of FS variants from the edges of swarm plates after 2 cycles of enrichment (TABLE 3.8), and the 5- to 10-fold higher proportion of FS variants at the edges of swarm colonies compared with the total population in the swarm. Thus, tentatively, it appears that the frequency of FS variant formation of 1 in 10,000 to 20,000 Is similar during swarm colony growth and In shake culture. This frequency seems too high to be readily explained by random mutation. Nonetheless, FS behavior is clearly heritable. Further studies are in progress to determine the genetic mechanisms underlying FS variant formation and reversion.

A spontaneous FS variant of RMB7201 (SVS) was found to revert back to wild-type behavior In a relatively stable manner at frequencies as high as

2% to 5%. However, the reversion of this FS variant to wild-type behavior was very inconsistent, indicating that important, but unknown environmental variables were involved in phenotypic reversion. A number of other FS variants of intermediate swarm rate also appeared to revert to normal swarm behavior during subculture. In an initial and third experiments, a relatively high frequency of reversion of Tn5 FS mutants was observed after inoculation

102 onto the host plant and reisolation from nodules, but there was no evidence of reversion in a second, larger experiment with various FS variants and mutants. Thus, passage through the host plant does not seem sufficient to cause, though it may enhance, reversion. Further studies to clarify the environmental and genetic factors that govern the reversion of FS variants to wild type swarming are needed. The results could be quite helpful in understanding how the switching of this organism's swarming behavior is regulated so as to function optimally in a diversity of environmental circumstances.

Several observations indicate that there may be multiple genetic configurations that result in FS behavior: i) the restriction patterns of FS17,

FS27 and FS28 are different, although they have the same FS-EPS phenotype (see CHAPTER 4); ii) the restriction patterns and swarm behavior of FS1, FS7 and FS44 are the same, but these mutants differ in their EPS regulation or swimming behavior (FS44 vs FS1 and FS7); and iii) spontaneous FS variants (eg. EV3, SV69) have been isolated with swarm rates that are intermediate between that of the parent and the Tn5 FS mutants and spontaneous FS variants like SV8 and EV2 (TABLE 3.8). These genetic and phenotypic differences between FS isolates emphasize the potential complexity of FS switching. They indicate a need for caution in extrapolating the behavior of one FS isolate to others and suggest that there could be a number of cryptic traits (like reduced EPS synthesis on defined media) associated with FS behavior that remain to be identified.

103 Given that cells of S. m eliloti do spontaneously switch between

“nomnal” and increased swarming behavior, it is of interest to make a tentative assessment of the likely costs and benefits of FS behavior to populations of these bacteria in natural environments. The wild type and FS1 mutant were found to grow at the same rate in mixed shake cultures on moderately rich media (TABLE 3.5). This result indicates that the metabolic costs of FS behavior, relative to normal swarming behavior, are modest. Two of the other

FS mutants and variants, FS17 and SV8, were not able to grow quite as fast as the parent in liquid shake cultures. This suggests that there may be significant metabolic costs associated with other changes occuring in these isolates. The parent and FS1 mutant showed very similar abilities to survive for long periods of in vitro storage under conditions of very low nutrient availability, both in buffer and pond water. The comparable rates of survival under low nutrient conditions seem in accord with the notion that FS behavior is not much more costly in terms of metabolic resources than normal swarming behavior. The parent, the Tn5 FS mutants and SV8 were all capable of extensive downregulation of motility under near-starvation conditions (TABLE

3.7, text, Wei & Bauer, 98). Such downregulation might help to explain why the FS mutants survived as well during semi-starvation conditions as the parent.

The FS mutants consistently had a significant growth advantage over the parent under all in vitro conditions where appreciable nutrient gradients developed. When directly competing against one another in still cultures, FS1,

104 FS17 and SV8 formed about 1.5 to 1.8 times as many progeny as the parent upon reaching stationary phase. In actively growing swarm colonies in semisolid agar. FS1 mutant progeny outnumbered wild-type progeny by roughly 33 to 40 fold within 48 h. FS variants (SV68 and SV69) with intermediate swarm rate (ca. 1.3-1.5 fold greater than the wild type) proved to have intermediate (20- to 27-fold) growth advantages in swarm agar, indicating that the competitive growth advantage of FS behavior is proportional to the rate of swarming. Another variant with intermediate swarm rate. EV3, apparently switched to a high swarm rate (= FS1) during swarm colony growth, and had a large growth advantage over the wild type. The faster growth of the FS mutants and variants in swarm colonies should probably be attributed to better access to available nutrients. Populations of

FS cells move outward from the center of the swarm colony faster than the parent, so that the periphery of the swarm colony, where nutrient concentrations are highest, is significantly (over 30 fold) enriched for FS cells.

Thus, on average the FS cells in a population should have better access to nutrients and generate more progeny.

In addition to enhanced nutrient acquisition, FS behavior may provide an appreciable advantage over the wild type with regard to random dispersal of the population. The FS mutants dispersed more rapidly than the parent not only in different physical media (semi-solid agar, moist sand, viscous liquid) but also when nutrient concentrations were low (soil extract, chemotaxis buffer), circumstances of some relevance to natural soil-rhizosphere

105 environments (FIG. 3. 2, TABLE 3.2). The survival value of increased dispersal

in terms of avoidance of local stresses, relocation to protected sites, enhanced

mating opportunities, or encounters with potential hosts depends greatly on

the specifics of a particular microenvironment, but is likely to be appreciable in

many, though by no means all.

The disadvantages of FS behavior are not so apparent or well

documented. Since the great majority of RMB7201 isolates display “normal”

swarming behavior, one must suppose that the genetic changes that confer

FS behavior have a negative selective value in most circumstances. Isolates

that swarm normally predominate regardless of whether one reisolates the

wild-type organism from nodules or native soil or maintains it in laboratory

cultures. Despite the selective disadvantage of FS genetic configurations that this predominance of the normal swarming type implies, the only circumstance

where we have found FS isolates to have an appreciably lower growth or survival rate has been in a sterile soil matrix subjected to air drying (TABLE

3.6). Even here it is important to recognize that the poor survival of the FS mutants relative to the parent during such drying might be due to associated changes in EPS synthesis or some other cryptic, coordinately regulated trait rather than FS behavior per se. EPS synthesis is mentioned specifically because it is downregulated in most of the FS mutants and variants of

RMB7201 obtained, and because EPS synthesis is known to be important to drying tolerance (Potts, 1994). However, FS7, which has only a 30% reduction in EPS synthesis on NM succinate-nitrate media, survived air drying of the

106 matrix just as poorly as FS1, which has a 60% reduction in EPS synthesis.

This suggests that FS behavior, rather than the differences in regulation of

EPS synthesis may be the important determinant of differential survival In a drying soil. If FS behavior per se is indeed responsible for the reduced drying tolerance of the mutants, the reason and mechanism are unclear. All considered, there is still no satisfactory explanation for the persistence of the majority of RMB7201 cells and isolates in the normal configuration rather than an FS configuration.

The cellular-biochemical mechanisms responsible for the 1.4 to 2.5-fold faster spreading rates of the FS mutants and variants are not yet clearly established, although a plausible model can be formulated. No statistically significant differences were observed between the FS isolates and the wild type in their in vitro chemotactic reponsiveness to a variety of attractants. Thus it appears that FS behavior does not involve changes like the 2- to 6-fold increase in general chemotactic responsiveness seen in cells of RMB7201 after transfer to semi-starvation conditions (Wei & Bauer, 1998). Nonetheless, the FS1 mutant did have slightly (20-30%) higher responsiveness than the wild type to every attractant tested (TABLE 3.3). Similarly, both the percentage of motile cells and the percentage of flagellated cells was consistently about

10-15% higher in cultures of the FS mutants than in cultures of the wild type.

While we think that the marginally higher tactic responsiveness and the modestly higher proportion of motile and flagellated cells of the mutants could contribute significantly to increased swarm rates, we do not think that they fully

107 account for the 2-fold increase in spreading rate observed for most FS mutants and variants. The FS mutants synthesized approximately 2-3 times as much flagellin as the parent and this flagellin was primarily on the outside of the cells in longer and more numerous flagellar filaments (TABLE 3.7, FIGS.

3.3 & 3.4). The longer and more numerous flagella of the FS mutants could account for much of the observed 2-fold faster spreading rate, although a linear dependence of swarm rate on flagellation is not certain. As Gotz et al.

(1982) pointed out, complex flagella are more rigid or fragile, thus easier to break. In environments like semisolid agar and soil, flagellar breaking off and regeneration are expected from active bacterial cells. Quick regeneration can be a factor for faster movement through matrices. Also, since S. meliloti has two independently transcribed flagellin genes (Pleier and Schmitt, 1991;

Bergman et al., 1991), it is also possible that the subunit composition of flagella on FS cells differs from that of wild-type cells.

The study on the migration of E. coli strains through semisolid agar revealed that nonmoltile and incessantly tumbler cells swarmed at lowest rates (Wolfe and Berg, 1989). Nonchemotactic cells migrated slower than the wild-type cells. However, swarm colonies of the intermediately-tumbling ceils were close to that of the wild type. Their explanation was that tumbling allowed cells to change direction thus circumvent obstructions in agar matrix.

One of our FS mutants, FS44, tumbles more frequently than the wild type, but it swarms faster than the wild type. It is interesting to have found that FS44 can migrate through viscous media faster than the wild type by long smooth

108 swimming. This indicates that the bacteria can somehow detect and respond to changes in viscosity, which may be important to their ability to swim efficiently through very thin films of liquid on the surface of soil particles and plant roots.

Based on the available data, it seems that the faster spreading rate of

FS mutants and variants is best explained by a modest, general derepression of motility and taxis, with increased flagellar synthesis being perhaps the dominant influence. A model of FS behavior that invokes modest, general derepression of motility and taxis seems consistent with the upregulation of swarming in mutants obtained by Tn5 insertions, normally a loss of function event. Such a model also seems consistent with the gradual increase in the swarm rate of the wild type, approaching that of the FS mutants, when the concentration of added nutrients in the swarm plates approached zero (FIG.

3.2), as though the wild-type swarm rate was repressed at high nutrient concentrations and became gradually derepressed as concentrations diminished. Mutant FS1 also swarmed faster at lower nutrient concentrations, but was not affected as much as the wild type.

All Tn5 FS mutants and most spontaneous FS variants produced less

EPS than the wild type in minimal media. However, these mutants are not

EPS defective since they can still produce large, wild-type amounts of EPS in rich media. The frequent association of altered swarming behavior with altered EPS synthesis in the FS mutants and variants suggests that some relatively high, common regulatory factor(s) may be involved, similar, perhaps

109 to phenotypic conversion (PC) in Ralstonia solanacearum where both EPS and motility were coordinately affected by spontaneous mutations in phcA

(Brumbley and Denny. 1990), or in Vibrio cholerae where spontaneous hyperswarmers and motility are tightly coupled to expression of ToxR- regulated and nonToxR-regulated virulence determinants (Gardel and

Mekalanos, 1996). Kamoun and Kado (1990) also reported a kind of phenotypic switching affecting virulence-related processes including chemotaxis, xanthan (EPS) production in Xanthomonas campestris.

In S. meliloti Rm1021, a contrasting genetic control mechanism was found. Mutations in the EPS regulatory genes exoR and exoS of S. meliloti

Rm1021 result in elevated EPS synthesis (Doherty et al.. 1988; Reuber et al..

1991). Our finding that these mutations also result in reduced swarming suggest that changes in the expression of genes like exoR and exoS could be involved in at least some instances of FS phenotypic switching in S. meliloti. We have also demonstrated that clones complementing EPS synthesis could correct swarming phenotype too. This indicates that motility and EPS production, two important yet expensive cellular processes for symbiotic, pathogenic, and many other species, may be co-regulated. This is probably a high level regulatory mechanism in the genetic hierarchy. It is speculated that under certain growth conditions, higher motility but less EPS may add to survival by active movement to search for favorable habitats; while in other situations lower motility can conserve energy, and more EPS protects cells from predation and desiccation. Studies on this genetic mechanism will

1 1 0 add to our understanding of the responses of rhizobia to environmental

signals of ecological significance.

While the great majority of the Tn5 FS mutants and spontaneous FS variants had reduced EPS synthesis on NM medium, the swarming and EPS

phenotypes are nevertheless separable. For example, many spontaneous

variants of the wild type had altered EPS synthesis but normal swarming

behavior, FS1 formed a “revertant" with normal EPS synthesis but retention of

FS swarm behavior, and some EPS overproducers had normal swarming.

We suspect that switching to FS behavior is a fairly common, though

unappreciated behavioral capability of bacteria. In the literature, it is apparent

that many workers have routinely enriched their strains for more motile

isolates to use in motility and chemotaxis assays (Aguilar et al., 1988, Kaempf

and Greenberg, 1992, Kape et al., 1991, Krupski et al, 1985, Nikata et al.,

1992, and Pleir and Schmitt, 1991), following the early suggestion of Adler

(1973). It seems likely that many of the isolates selected in this way were spontaneous FS variants of reasonable genetic stability. If so, the behavior of the wild-type forms of these strains remains to be characterized. In preliminary tests, it has proven rather easy to isolate putative FS variants of E. coli

CG118^j^ and Pseudomonas aeruginosa PA01 after 3 or 4 cycles of swarm

plate enrichment (FIG. 3.5). The FS variants we obtained of these two model species were stable on storage and plating, swarmed in 0.35% agar at rates of 1.5 to 2 times faster than the wild-type stock cultures, had swimming

111 *1

FIG. 3.5. Swarm rates (measured by colony size) of spontaneous FS variants (bottom) compared to the wild type (top) of P. aeruginosa. The experiment was done in 1/20 strength LB with 0.35% agar.

behavior similar to the wild type under the microscope, and had growth advantages of 37-fold and 83-fold, respectively, over the parental forms after 2 d in 1/20 LB swarm agar (FIG. 3.5). We also obtained FS variants of S. meliloti strain Rm1021. If FS phenotypic switching and behavior are indeed common phenomena, then further studies to examine the ecological consequences of

FS behavior and the genetic mechanisms of variant formation and reversion in a number of representative species are clearly warranted.

1 12 CHAPTER 4

PRELIMINARY MOLECULAR CHARACTERIZATION OF THE FASTER SWARMING BEHAVIOR OF SINORHIZOBIUM MELILOTI RM B7201

ABSTRACT

Faster swarming (FS) Tn5 mutants and spontaneous FS variants of

Sinorhizobium meliloti RMB7201 were isolated and phenotypicaliy characterized (CHAPTER 3). The majority of FS mutants and variants synthesize more flagella in liquid media, but produce less EPS than the wild type on minimal media. This is called the FS-reduced EPS (abbr. as FS-

EPS) phenotype. Southern hybridization analysis of the Tn5 FS mutants revealed that the Tn5 insertion site falls within the same genomic region in all

FS mutants with one exception. This was supported by PGR analyses with the primers designed from the Tn5 terminus and the known wild-type flanking sequences. Six genomic library clones were identified that contained sequences homologous to the Tn5 flanking sequences in FS1. These genomic clones did not complement the FS-EPS phenotype. A 3.25 kbp DNA fragment corresponding to the sequences flanking Tn5 was subcloned and sequenced. A gene bank search shows that this is a unique sequence that

113 shares homology only with an Archaeolglobus fulgidus sequence of known function. Marker exchange of the Tn5 with the flanking sequences back into

the wild-type strain did not recreate the FS-EPS phenotype. Thus, it appears that mutations at the Tn5 insertion site may not be the direct cause of the FS-

EPS phenotype. Several genomic library clones have been identified that correct both EPS and swarm phenotypes. However, these clones do not contain sequences homologous to the Tn5 flanking sequences in FS1. Two

DNA fragments that correct both motility and the EPS phenotype have been identified and subcloned from the above cosmid clones. Sequencing data show that these sequences have homology to a histidine protein kinase involved in flagellar biogenesis and motility in Caulobacter crescentus, and to classic two-component regulator genes. An E. coli sequence containing an extragenic suppressor gene also restored EPS production to the FS mutants, and sequences adjacent to the extragenic suppressor gene complemented swarm behavior. Possible molecular mechanisms for FS behavior and future studies are discussed.

INTRODUCTION

In the proceeding chapter, we described the generation and phenotypic characterization of Tn5 FS mutants. We also reported the isolation of spontaneous FS variants, and the discovery of a switching

114 phenomenon that causes the spontaneous conversion between the wild-type and FS behavior. In this chapter, we will focus on the possible molecular mechanisms causing FS behavior and the switching between the wild-type and FS-EPS phenotypes.

Studies related to switching. Phenomena similar to the S. meliloti

FS motility switch have been observed in Ralstonia solanacearum (Denny et al., 1988; Brumbley and Denny, 1990; Brumbley et al., 1993) and in Erwinia chrysanthemi (Castillo et al, 1998). In R. solanacearum, gene phcA is involved in the regulation of both EPS production and motility. Inactivation of this gene by spontaneous insertions, ranging from 2 bp to as long as 1.2 kbp, caused reduced EPS and extracellular enzyme production, and gain of motility. They called this phenomenon phenotypic conversion. Gene phcA codes for a DNA-binding protein similar to the LysR transcriptional activator family. LysR protein has a typical helix-turn-helix domain (Stragier and Patte,

1983). The PhcA protein regulates the expression of several virulence genes.

It facilitates normal EPS production but represses motility. Since cells with a mutated phcA gene produced less EPS while gained motility, it is clear that

EPS and motility gene expression are related. In E. chrysanthemi, gene pecT codes for a repressor that negatively regulates the expression of virulence genes, including the pectate lyase gene, and both motility and EPS genes

(Castillo et al. 1998; Condemine et al., 1999). In Salmonella typhimurium, phase variation determines that only one of the two types of flagellin was produced at any given time (Okazaki et al., 1993; Silverman et al., 1979).

115 Phase-1 flagellin is encoded by fliC and phase-2 fIjB. Expressing only one of the two genes is achieved by a DNA element that can change orientation.

In one direction, it turns on one flagellin gene, in the other direction, the other

(see ref. Macnab, 1996). The hin gene product is an enzyme involved in the inversion of the element. The Hin protein has homology to the TnpR protein of transposons (Simon et al., 1980; Szekely and Simon, 1983). Tamura et. al.

(1995) reported that spontaneous nonmotile variants of Clostridium chauvaei appeared at frequencies much higher than spontaneous mutation. Reversion also occurred at relatively high rates. They found that one third of the nonmotile variants had flagella, while the rest had none. They concluded that

C. chauvaei possessed a phase variation mechanism controlling motility and flagellation, but the molecular basis is not known. Similar phase variation has been found in other species (e.g. Diker et al., 1992; Nuijten et al., 1995;

Givaudan et al., 1996). In these cases, phase variation was between nonmotile wild type and motile variants, or vice versa. Faster swarming behavior was mentioned in literature, but to our knowledge, our study is the first detailed investigation into FS behavior and switching between a normal motile wild type and hypermotile variants.

Mechanisms governing phenotypic switching or phase variation can be very complex. In this chapter, we will present preliminary results on the possible molecular mechanisms controlling hypermotility and reduced EPS production. The study is far from complete, and future directions will be proposed and discussed.

116 MATERIALS AND METHODS

Bacterial strains, bacteriophages, plasmid and cosmid

vectors, media and buffers. TABLE 4.1 lists the additional bacterial

strains, plasmids used or constructed in this study. Sinorhizobium meliloti, E.

coll and Pseudomonas aeruginosa strains were stored, cultured, harvested and tested in the ways described in CHAPTER 3.

DNA manipulation, cloning, and sequencing. Most of the genomic and plasmid DNA isolation, restriction digestions and mapping,

DNA fragment isolation, ligation, transformation and electroporation, and screening of the genomic library by colony hybridization were performed as described (Sambrook et al.,1989). pBluescript (Stratagene, La Jolla, GA) was routinely used as a vector for cloning. A RMB7201 genomic library was constructed using the strategy and method described previously (Staskawicz et al., 1987). RMB7201 genomic DNA was partially digested with Sau3A.

Fragments in the size range of 20 to 30 kb were isolated by sucrose density gradient centrifugation. The DNA fragments were ligated into the broad host range cosmid vector pLAFR3 (Staskawicz et al., 1987), and recombinant cosmids were packaged and delivered into E. coll DH5

Bauer, unpublished data).

High quality plasmid DNAs for sequencing were isolated with INSTA-

MlDl-PREP kit (5 Prime->3 Prime, Inc. Boulder, CO), or Quantum Prep

Plasmid Miniprep Kit (Bio-Rad, Hercules, CA). For genomic DNA preparation

117 TABLE 4.1. Strains, phage and plasmids used and constructed in this study, their characteristics and source of reference Strains Relevant characteristics Source or reference S. meliloti TcR RMB7201 spontaneous TcR This study derivative of RMB7201 E.coli DH5a F recA 1 endA 1 thi-1 hsdR 17 sup44 Hanahan 1983

Strain H Gartinkel, etal. 1981 HB101 F recA13 hsdS20 thileu arg-14 Boyer et al. 1969

phage X::Tn5 KmR, defective in replication de Bruij'n & Lupski, and integration 1984 cosmid & plasmid pBluescript ApR StrataGene pPHUI GmR, IncP plasmid Gartinkel, etal. 1981 pMSRI -6 Six cosmid genomic clones This study containing sequences homolo­ gous to the Tn5flanking sequ­ ences in FS1 pBSMRI pBluescript clone containing the This study wt 3.25 kb EboRI fragment corre­ sponding to the TnSflanking sequences, isolated from pMSRi pBSMR2 pBluescript clone containing the This study wt 6.8 kb EooRV fragment corre­ sponding to the Tn 5 flan king sequences, isolated from pMSRl pLASMRI pLAFR3 clone containing the This study 3.25 kb EboRI fragment PLASMR2 pLAFRS clone containing the This study 6.6 kb Psfl fragmentfrom pBSMR2 pMSDI-7 Seven genomic cosmid clones This study that complement FS-EPS phenotype (continued)

118 (TABLE 4.1 continued) pMSD8 Genomic done that complements This study EPS phenotype and renders FS mutants nonmotile PBS2.2-3 pBluescript subclone of the 2.2 This study kbp fragmentfrom pMSD3 pBS5.8 pBluescript subclone of the 5.8 This study kbp fragmentfrom pMSD4 pMSD61 pMSD4 with a TnSinsertion, thus This study knocking out complementing ability pBSlIT pBluescript clone of the 5.8 kbp This study fragment plus Tn5from pMSD61 pBSFSWI pBluescript clone containing the This study 9.3 kb EcoR 1 fragment of Tn 5 and flanking sequence from FS1 pLAFSWI pLAFRS clone containing the This study 9.3 kb EcciRl fragment of Tn5 and flanking sequence from FS1 PJQ9.3 pJQ clone containing the 9.3 kb This study EooRI fragment of Tn5 and flanking sequence from FSI pGP9.3 The 9.3 kb EcoRI fragment of This study Tn5and flanking sequence from FSI doned in vector pGP704 pME3 Genomic cosmid done that This study complements only EPS pME4 Cosmid clone (EL coli sequence) This study complementing FS mutants swarming and EPS phenotype. pME17 pME4 inserted with a Tn5, thus This study abolishing complementing ability pBS2.9 The 2.9 kb EooRV complementing This study fragmentfrom pME4, subcloned in pBluescript. pLA2.9 The 2.9 kb EooRV complementing This study fragmentfrom pME4, subcloned in pLAFR3

Note: FIG. 4.8 shows some of the pLAFRS-based subclones not listed above for complementation tests in S. meliloti, and many more subclones used for overlapping sequencing are not listed.

119 from S. meliloti strains, cetyltrimethylammonium bromide (CTAB) was used to

remove excess EPS (Ausubel et al. 1989).

Southern hybridization. For Southern blot analysis of the mutants,

genomic DNAs were digested with various restriction enzymes and

fragments separated by agarose gel electrophoresis. DNA probes were

labelled either with ^^P-ATP with a random hexamer labelling method

(Amersham, Arlinton Heights, IL), or with biotin-14-dCTP (Life Technologies,

Gaithersburg, MD), following the manufacturer's instructions. The probes

were the internal 3.3 kb H/ndlll fragment from Tn5 (Jorgensen et al. 1979;

Auerswald et al. 1981), the 5.0 kb EcoRI-SamHI fragment from mini-Tn5/acZ

(de Lorenzo et al. 1990) and the 3.25 kb EcoRI or 6.6 kb Pstl wild-type

sequence corresponding to the Tn5 flanking sequences in FSI. For

Southern analysis of the mutants created by marker exchange, entire

plasmid clones (vectors plus inserts) were used as probes. Low stringency

hybridization was carried out at 38-40°C in hybridization solution containing

35% formamide. The blots were developed with X-ray film or BlueGene

Nonradioactive Nucleic Acid Detection System (Life Technologies).

Automated DNA sequencing was done by the Biotechnology Center or the Biopolymer Facility at the Ohio State University (Columbus, OH), and by

the Iowa State University Sequencing Facility (Ames, Iowa).

Polymerase Chain Reaction (PGR), Primers for PCR were

designed from the known sequences of the 3.25 kb EcoRI and the 6.6 kb Psfl fragments, and from the Tn5 terminals. Primers used in the study are shown

1 2 0 In TABLE 4.2. The primers were synthesized by integral DNA Technologies,

Inc., (Coralvllle, lA). Random sequence primers ( a gift from Dr. Caetano-

Anollés) were used to amplify the sequences from the known Tn5 sequences outward In seml-speclfic PCR. Cycling conditions recommended by the

manufacturer were followed with slight modifications. DNA amplification was carried out on a thermal cycler model PTC-200 (M J Research, Watertown,

MA) with Taq polymerase, Pfu polymerase (Stratagene, La Jolla, CA), or PCR

SupermIx (Life Technologies, Gaithersburg. MD). Conditions for amplification with one random and one known primer were modified as follows: dénaturation at 94°C for 30 sec, annealing at 48°C for 45 sec, and extension at 72°C for 90 sec. Arbitrary and mlnl-halrpin primers (Caetano-Anollés and

Gresshoff, 1994) were used for DNA fingerprinting of RMB7201 and mutants.

36 cycles were generally carried out for routine amplifications.

Marker exchange through homologous recombination. The

TnS^contalnlng EcoRI fragment from FBI, with flanking sequences of 1.0 and

2.25 kb on each side of the Tn5, was cloned Into pBluescript, pLAFR3 and pGP704. The resulting constructs were designated as pBSFSWI. pLAFSWI and pGP9.3 (TABLE 4.1). The cloned fragment was Introduced back Into the wild type In an attempt to select for mutants by homologous recombination.

Different delivery and selection methods were used for different constructs.

Since pBluescript Is nonmoblllzable and cannot replicate In S. meliloti, pBSFSWI was delivered Into the wild type RMB7201 by electoporatlon.

Putative recombinants or exconjugants were selected for Km resistance. In

121 TABLE 4.2 Primers used in PCR analysis and sequencing^.

Primer # base seouence (5' t o 3')

lb CTTACTTCAC CTCAAACTCG CC 2 TCAGGCCTAC AGTGTACAGA 3 CAAGGTGGCG GTCGTCTCCT CG 4 CGTTCAAGGC CGAGGAGACG AC 5 GCGAAAGTGT AGAAGATGAA GACG 6 TCCTGACCTT CCTTGCCTAT CC 7 GCGACCAGCC TGTCAACGAC TTCG 8 GTCTCCTCGG CCTTGAACGG GTC 9 GATGGCCGAG CAGCGCGATG GAGATC 10 CGCGATCAGC CTGGTGCTCG GCATG 1 1 CACATCCCAT GCCCGCGGTA GTTC 12 CTTTTCAGCG CGGGATAGGC AAG 13 CGTTGTACTG CACATCGGAC TG 14 CATCATGAGC TTACCCGGCT CTAC 19 CACGACGTTG GAGACGGAAG 2 0 CTTCGTCCTC TTCGGAACGC TG a See FIG. 4.7 for relative positions of the primers on the 3.25 kbp fragment, b Primer #1 is an outward primer from the Tn5 terminus.

contrast, cosmid vector pLAFRS or the clone pU\FSW1 is stable in S. meliloti

RMB7201. In order to select for recombinants, the eviction plasmid pPHUI

(Garfinkel, et al., 1981) was mated into RMB7201 (pLAFSWI). pPHUl belongs to the same incompatibility group (IncP) as pLAFRS, thus those cells retaining pPHUI should lose pLAFSWI. The mating mixtures were plated on

122 both gentamicin (Gm) and Km to select for cells having pPHUI and the Tn5 marker. The third vector, pGP704, is mobilizable from E. coli, but cannot replicate in S. meliloti. The pGP704-based clone, pGP9.3, was mated into

RMB7201 and putative recombinants were selected for the acquisition of Km resistance and screened for the loss of the Ap marker carried by the vector.

The final method was to use the sacB vector pJQ200 (Quandt and Hynes.

1993). This sacB gene-containing vector, if retained and the sacB gene expressed, will kill S. meliloti RMB7201 on sucrose-containing media. This method selects for the recombinants generated by double crossovers, not those by whole plasmid integration. The 9.3 kbp EcoRI fragment was subcloned into pJQ200, and the resulting plasmid construct was named pJQ9.3, which was then introduced into the wild-type RMB7201. The putative recombinants created by marker exchange were purified for phenotypic tests.

Genomic DNAs were prepared for Southern hybridization. At least 10 isolates from each marker exchange experiment were tested. The whole plasmid clones were labelled with biotin-dCTP (see above in section on

Southern Hybridization) as probes so that the vectors would be detected if they were integrated in the RMB7201 genome.

Complementation of the FS-EPS phenotype. Genomic library clones or subclones in pLAFR3 were introduced into FS mutants and variants through triparental mating, with pRK2013 as the helper plasmid (Ditta et al.,1980). Exconjugants from spontaneous FS variants were selected and purified on NM media containing 200 fjg/m\ of Sm and Spe, and 10 fjg/m\ To.

123 The above medium was supplemented with 200 ^g/ml of Km to select for exconjugants of Tn5FS mutants.

Complementation of the FS phenotype was technically difficult because spontaneous loss of a complementing cosmid from even 1% of the cells in a swarm colony will result in a fast swarming colony like that of the FS mutants. While Tc can be used in the swarm medium to select for cells which

retain pLAFRS-based clones, it adversely affects motility at the concentration normally used (12-20 jug/ml). Hence in our swarm complementation tests,

low levels (2.5 pg/ml) of Tc were incorporated into the swarm agar to maintain cosmids. Tetracycline at 2.5 pg/ml was effective in maintaining genomic clones for 3 d, and did not significantly alter the ratio of the swarm rate of FSI(pLARFS) to that of RMB7201(pLARF3) or spontaneous Tc^

RMB7201 in control experiments. Two ^1 of late exponential phase cultures were spotted onto swarm agar, cultured at 28°C for 2 days, and swarm colony diameters measured. In all experiments, swarm colony diameters of each strain were averages of at least duplicate plates.

Tn5 insertional mutagenesis of potentially complementing cosmid clones. Stocks of a X::Tn5 phage defective in both replication and integration (de Bruijn and Lupski, 1984) were prepared and titered as described previously (Sambrook et al. 1989). T n5 was introduced into E. coli

DH 5oc carrying cosmid clones by infecting cells with the A.::Tn5 stock. Cells able to grow on LB-Km-Tc plates were collected for cosmid preparation.

124 which was then transformed into E. coli DH5«:. Cells carrying Tn5-containing

cosmid clones were selected on LB-Km-Tc plates. Clones that could no

longer restore swarm rates in FS mutants after Tn5 mutagenesis were

identified by complementation tests (see above). Sequences flanking the Tn5

insertion were mapped and partially sequenced.

RESULTS

Tn5 insertion sites in the motility mutants.As shown in FIG.

4.1, each of the 20 nonmotile mutants obtained during our initial screens has

a unique restriction pattern in Southern blot probed with the Tn5 probe.

Unique restriction patterns were also observed in about 20 reduced motility

mutants (FIG. 4.2), indicating that Tn5 insertion into the S. meliloti genome was not restricted to any particular loci or hot spots. This is consistent with

random Tn5 insertion into a number of chromosomal loci essential to motility in S. meliloti. In contrast, in the three Tn5 FS mutants, FSI, FS7 and FS44, the Tn5 was inserted into the same region of the genome based on the EcoRI and EcoRI-BamHI restriction patterns (FIG. 4.3A). These FS mutants, isolated from independent mating experiments, also gave identical restriction patterns when digested with Sail, Clal, Sall-BamHl, and Cla\-BamH\ (data not shown).

However, the three FS mutants were found to have slightly different phenotypes (see CHAPTER 3). In a later experiment, mini-Tn5/acZ was used to generate additional insertion mutants. As discussed in CHAPTER 3,

125 upper gel 3456789 10 kbp 22

Lower gel

FIG. 4.1. Southern analysis often randomly selected non swarming Tn5 mutants of S. meliloti RMB7201. DNA Bands on top were from EcoRI digestion, and bands on the lower gel BamHl. The membrane was probed with the 3.2 kb /-//bdlll fragmentfrom the wild-type Tn5. Numbers on the left side are DNA molecular size in kbp. This experiment was done by Debbie Estes.

Upper gel 2 3 4 5 6 7 kbp 22

5.1 4.4 3.5

Lower gel

FIG. 4.2. Southern analysis of randomly selected Tn5 mutants of S. me///o&'RMB7201. DNA Bands on top were from EcoRI digestion, and bands on the lower gel BamHl. The membrane was probed with the 3.2 kb HintiWl fragment from the wild-type Tn5. Lane 1=RMB7201 ; lane2=FS1 ; lane 3=Tn5 intermediate swarming mutant; lanes 4~7=Tn5 reduced swarming (RS) mutants. Numbers on the left side are DNA molecular size in kbp. This experiment was done by Debbie Estes.

126 A B ..EcoRI BamHl EcoRV H/ndlll

1 2 3 4 5 6 1 2 3 4 5 6

5.0 :~yrt-PÆÊÊÊmi

■<'V - v !'." r • î - . i. '• •y's.ri. '■ ^5 -4.1

' ■ ■ m “ 2 4

1.0

Fig. 4.3. Southern analysis of six independent FS mutants of S. meliloti RMB7201. Panel A shows FSI (lanes 1, 4), FS7 (lanes 2, 5), and FS44 (lanes 3, 6), probed with the 3.2 kb H/ndlll fragment from the wild-type Tn5. Panel B shows FS17 (lanes 1, 4), FS 27 (lanes 2, 5), and FS28 (lanes 3, 6), probed with the 5.0 kb EcoRl-SamHl fragment from the mini-Tn5/acZ1. The wild-type strain has no homology to the wild-tpye Tn5, but shows a faint band when probed with mini-Tn5/acZ7 sequence.

127 Southern analysis revealed that only 4 of the 32 representative isolates had

sequences homologous to the mini-Tn5 probe. That is, most of the FS

isolates (28/32) from mini-Tn 5 mutagenesis/enrichment experiments did not

have mini-Tn5 insert at all. Unlike FS1, FS7 and FS44, three mini-Tn5/acZ

FS mutants, FSI7, FS27 and FS28, gave different restriction patterns (FIG.

4.3B). Furthermore, the restriction patterns of the regions flanking the mini-

Tn5 in FS17, FS27 and FS28 are different from those flanking the mini-Tn5 in

F S I, FS7 and FS 44. Cloning of Tn5 and flanking sequences and marker

exchange. In order to determine if there was a link between the Tn5

insertion in the Tn5 FS mutants and the FS-reduced EPS phenotype, Tn5 together with the flanking sequences was cloned from FSI for the

identification of the wild-type sequences from a genomic library, and for marker exchange back into the wild type. To do marker exchange, the 9.3 kb,

Tn5-containing EcoRI fragment was cloned into a number of vectors (see

Materials and Methods). The initial marker exchange experiments with pBSFSWI and pGP9.3 did not yield any gene replacement by homologous recombination. Instead, the whole plasmid was integrated into the genome,

regenerating a wild-type copy of the element, the 3.25 kbp EcoRI fragment.

Marker exchange was accomplished with the use of the sacB vectors pJQ200/210. As shown in FIG. 4.3C, each of the three randomly selected marker exchange isolates tested had the same restriction patterns as FSI.

When digested with EcoRI, a single 9.3 kb band was observed when probed

128 with the whole plasmid pJQ9.3, indicating that double crossover had occurred and the vector lost.

Surprisingly, this apparently successful marker exchange did not re­ generate the FS-EPS phenotype. The strains generated by marker exchange swarmed at the rates similar to that of the parent strain. Colonies on NM plates of these recombinants were as mucoid as the wild type.

Isolation of the wild-type sequence flanking the Tn5.A screening of the RMB7201 genomic library was carried out with the Tn5 flanking sequences as probes. The probe was isolated from the cloned 9.3 kb EcoRI fragment (FIG. 4.5). Through colony hybridization of the pooled genomic library with the probes, six cosmid clones were identified containing wild-type sequences homologous to the flanking sequences. Some of these clones have the intact 3.25 kb EcoRI (Rl) fragment, or even the 6.8 kb EcoRV

(RV) fragment corresponding to the sequences flanking the Tn5. Other cosmid clones contained part of the 3.25 kb EcoRI fragment.

When these clones were introduced into FSI, FS7 and FS44, they did not affect either FS swarming or EPS phenotypes.

Sequence of the wild-type 3.25 kb EcoRI fragment.The entire

3255 bp wild-type fragment corresponding to the sequences flanking the Tn5 in FS1 was sequenced in both directions with overlapping sequences (FIG.

4.4). A gene bank search with Blastn, Blastx, and Blastp (National Center for

Biotechnology Information, or NCBI; Anderson and Brass, 1998) revealed that a 474 bp segment (nucleotides 668 to 1141) had high homology

129 FIG.4 .4

1 GAATTCCCATGCCGAGCACCAGGCTGATCGCGGCGATGAGGATCAGCATC

5 1 ATGATGACGTTGCCGCCCGAAATGAACTCGACGAGCTCCGTCATCATCAG

1 0 1 GCCGAGGCCCGTGAGCGTGATCGTTCCGACGACGATACCGGCGGTGGCCG

1 5 1 TGGCGACCGCGATGCCGATCATGTTGCGGGCACCGAGCGCCAGGCCGTCG

2 0 1 GTCAGGTCCCAGGCGGCGGCGCGGACCGAACCAGCAACATCTTCCTTCCG

2 5 1 GAAGATCGCCATCAGCGGCTTGCGGGTGGCGACGATGGCGATGATCGTC A

3 0 1 GCGTTGCCCAGAAGGCCGAAAGCCCGGGCGAAAGCTGCTCGACCATGAGG

3 5 1 CACCAGAGCAGGACGACGATGGGTATGAGGAAATCCAGCCCGGTCCGCGT

4 0 1 CACATCCCATGCCCGCGGTAGTTCCAGGATCGGCGCGTCGGGATCGTCGA

4 5 1 GCTCCAGGTCCGGATAGCGCGACGAATACCAGATGGTCGCGATGTAGAGC

5 0 1 GCCACGCCGGCAATCGCGAGCAGCAGGGGCGCGCTTTCTCCGGATGCGGC

5 5 1 CCGGATGGCAATGATCCCGTAATAAAGAAGGCAGACCGCAAGCACGCTGC

6 0 1 CGGCAAGCCCGAGGCCCATGCGGAGCCATCGCTCGCGCATCGGAGCCGGG

6 5 1 CGCTGGGGGATCGGCTGCATGTTCAGCTTCACCGCCTCCAGATGCACCAT

7 0 1 GTAGAGAAGCGCGATGTATGAGAAGACCGCCGGCAGCAGCGCATGCTTGA

7 5 1 CGATTTCCGAGTAGGGAATGCCGACATATTCGACCATCAGGAATGCCGCG

8 0 1 GCGCCCATGACGGGCGGCATGATCTGGCCGTTAATCGAGGCCGAAGCTTC

8 5 1 GATCGCGCCGGCCTTCACGCCGGAAAGCCCGGTGCGTTTCATCAGCGGAA

9 0 1 TGGTGAAGATGCCGCCCGACACGACGTTGGAGACGGAAGAGCCGGAGACA

9 5 1 GACCCGTTCAAGGCCGAGGAGACGACCGCCACCTTGGCC^GGGCCGCCGCG

1 0 0 1 CAGATGGCCGAGCAGCGCGATGGAGATCTGCATCATCCAGTTTCCGGCAC

1 0 5 1 CGGCCTTCTCGAGCAGCGTTCCGAAGAGGACGAAGAGGAAAACGAAACTC .

1 1 0 1 GTCGACACGCCGAGCGCAATGCCAAAAACGCCCTCCGTCGTCAGCCATTG

1 1 5 1 ATGGTTGATGAACTTAACCAGCGAAGCGCCGCGATGCTGGATGACGTCGG

(continued)

130 (FIG.4.4 continued)

1201 GC ATGTACTGGCCGGCGAAAGTGTAGAAGATGAAGACGCCGGCGACGAAG

1251 ACCATCGGC AGGCCGAGCGCGCGGCGCGTAGCCTCGAGCAGGAGCAGGAT

1301 GCCTGCGGTGCCGGCGACGAGATCGAGCGTCGAGGGCTGGCCGGGCCGTC

1351 CGGCAAGCTCGCCATAAAAAAGGAACAGATAGGCGCCCGCGAAGGCGCCG

1401 AGCGCCCCGAGCACCCAATCGGCGAGCGGCACGCGGTCCCGCGGCGAACT

1451 TTTCAGCGCGGGATAGGCAAGGAAGGTCAGGAACAGCGCGAAACCGAGAT

1501 GGATGGCGCGAGCTTCGGTATCGTTGAGAATGCCGAAGCCGAAGACGAAG

1551 GGCAGGGGAGACGCGTACC AGAGCTGGAACAGCGACCATGCCACAGCGGT

1601 GCAGAGCAATATCTGCCCGGTCAATCCTTTGGCGTTGCGTCCGCCGGTAT

1651 CCGCTTCGGCGACCAGTTGCTCGAGGTCGACATCGACCTTCCGATCTTGA

1701 AGTTCGCTCGTGACACCCTCCCATGCTGATATTGCCGGCGAGCCGCCTTT

1751 CGGCCCGCCGGTAAACGTCAGAATGCGAAGCTTACTTC AACCAGCCCTTT

1801 TCCTTGTAATATTTTTCCGCACCCGGATGCAGTGGTGCAGAGAGGCCGTC

1851 CTTGATC ATCTTGGCGGGCTCGAGACTGGCGAAAGCCGGGT.GCAGCGACT

1901 TGAACTCGTCGAAATTCTCGAAAACCGCTTTCGTCAGCGCGTAGACGCTC

1951 TCTTCCGGAACATTGGCGGAGGTGACCAGTGTCGCCAGAACGCCGAAAGT

2001 TTCCGTGTCTTCCGGATTGTTGTTGTAGAGGCCCCCCGGAATCGTCGCCT

2051 TGGCGTAATAGGGATTGTCGGCGACCAGCCTGTCAACGACTTCGCCGGTG

2101 AGCGGCACCAGCTTCGCCGCGCAGGTCGTCGTCGGATCCTGAATATTGGC

2151 GGAGGGGTGCCCCACGCCGTAGAAGAAGCCGTCGATCTTGCCGTCGCAAG

2201 CGCCGGGCCATGCTCGTCCGCCTTGAGTTCGGATGCGAGCGAGAAGTCGG

2251 CGAGCGTCCAGCCC ATGGCGCCGAGCAGGCGCTCCATGGAAGCGCGCGTT

2301 CCCGAACCCGGATTGCCGACGTTGAAGCGCTTGCCCTTGAAGTCCTCGAA

2351 CTTCGTCACGCCGGCATTTGGATGCGCCAAC ACCGTGAACGGCTCCGGAT

(continued)

131 (FIG.4.4 continued)

2401 GAATCGAGAAGACCGCCCTGAGTTCAGCATGCGCGCCGCCCTCCTTGAAC

2451 GTCTCGTCGCCCTTGAAAGCGTTGTACTGCACATCGGACTGCGCC ATGCC

2501 GAAGTCGAGCTCGCCCTCCTTGATCGTGTTCACGTTGAAACGCTGTATCC

2551 GCCGGTGGTATTCGACCGAAACAGCGGTATGCCGTGCGTCTTGCGGTCCT

2601 TGTTCAGAAGCCGACAGATGGCGCCGCCCGCGGCATAGTAGACGCCGGTG

2651 ACGCCGCCGGTGCCGATCGTCACGAATTTCTGCTGCGCCTCGGCGCTGCC

2701 CGAGAAAAACAGTGCCGCCGAGAGGGCGGCGATCTGCGCTCCGCGCAGGG

2751 TAATCCGTTCATGTTTCATCAATTTCTCCCATAATGCTTCGCGTGCCACC

2801 GAGCCGTCCCGAGATTTATGCCTGTAGCGCCACGCGTCCAGTCGGACGCG

2851 CAAAGGTCGCGGCGGCAGTTAGTATTACACGCAGAAGGGCACGTTTCGAG

2901 CATGTTACGATCGTTGCGCTTTATTCAAGCCAAGGGGGATAACTTTTGCA

2951 AATCGGGCGGCGGAAAAAGCGGTTGATCGGAGCAGCCCCGGATGTCTCGT

3001 TCATGGTAGCGGCCGTTCGCGAATGCCGATAATTGGGGAACCAAGAAAGG

3051 GCGACGGATGTTTTCCTCCTTCAAACGGGAGCATGGCCATGGCGACGATA

3101 CGATATGAAATTGTGGAACACGACGGGGGCTGGGCCTACAAGGTAGACGA

3151 CGTCTTCTCCGAAGCCTTCCCCACCC ATGACGATGC ACTGGCCGC AGCCG

3201 AGATACGCGGCGCGCCATCATGAGCTTACCCGGCTCTACCGAAGCCATCG

3251 AATTC

FIG. 4.4. Nucleotide sequence of the 3255 kbp EcoRI fragment corresponding to the sequences flanking the Tn5 in FSI. Sequence was determined by sequencing both strands with overlaps. The ^ symbol at position 989 indicates the Tn5 insertion site in FSI. The consensus Shine- Dalgano sequence is underlined.

132 VI 2 3 3.255 1______1______I I

% DNA length 10 20 30 40 50 60 70 80 90 100

RF 1

RF 2

RF 3 ------> ------>

RF 4 < -

RF 5 < ------< ------< -

RF 6 < ------

FIG. 4. 5. Potential open reading frames (50 amino acids or greater) of the 3255 bp EcoRI fragment corresponding to the Tn5 flanking sequence. The ^ symbol at position 989 indicates the Tn5 insertion site in FSI.

133 upper gel lane 1 2 3 5 6 7 8 9 10

III 13.0 9.3 6.8

3.3 2

1.2

upper gel

lower gel

13.0

9.0 6.8

3.3

lower gel 11 12 13 14 15 16 17 18 19

FIG. 4.6. Presence or absence of sequence homologous to the 3255 bp fragment in various species and S. meliloti strains as detected by Southern hybridization. Numbers on the right side of the blot are DNA sizes in kbp. Upper gel; lower gel: lane 1=7201 cut with EcoRl(Rl) 2=7201 wt-EcoRV (RV) lane 11=L5-30-Rl 3=7201 wt-eamHl (Bl) 12=L5-30-RV 4=7201 wt-Rl-Bl double ^3=A. tumefaciens-R\ 5=FS44-R1 14=4. tumefaciens-RV 6=FS44-RV 15=4. tumefaciens~B\ 7=JJ1c10-Rl 16=Ps. syringae-R\ 8=JJ1c10-RV 17=Ps. syringae-RV 9=E CO//MV1190 Rl 18=E. CO//WA803(pGS9)-Rl 10=E. CO//MV1190 RV 19=E coli WA803(pGS9)-RV

134 {p(N)se‘^> to an Archaeoglobus fulgidus sequence (Klenk et ai. 1998).

Unfortunately, no function has been demonstrated or described for this

sequence. The rest of the 3255 bp sequence did not have significant homology to any known sequence. Therefore it is a novel sequence.

However, short stretches (25 to 60 bp) of sequence do share low [p(N)>0.01] homology to parts of the exp gene cluster of S. meliloti Rm2011 (Becker et al.

1997), the R. capsulatus fdxD gene (Armengaud et al. 1994), the

Azotobatacter vanelandii mannuronan C-5-epimerase {alaG) gene (Rehm et al. 1996), the Pseudomonas aeruginosa sn-glycerol-3-phosphate dehydrogenase {gIpD) (Schweizer and Po, 1994, 1996) and membrane protein (gIpM) genes (Schweizer et al. 1995). Except the fdxD gene, other genes listed above are all somewhat related to EPS synthesis.

There are a number of possible ORFs in the fragment (FIG. 4.5). In FS1 the Tn5 was located at position #989 (V in FIG. 4.4), disrupting several potential ORFs. There are at least two possible ORFs in the 474 bp A. fulgidus sequence corresponding to the portion of the S. m eliloti sequence with the Tn5 insertion. The consensus Shine-Dalgano (S/D) sequence or ribosome-binding site in prokaryotes is (A)GGAGG. A search for such sequence in the fragment appeared at least 8 times (FIG. 4.4, underlined).

Some S/D sequences are in front of the potential ORFs.

We probed genomic DNAs from a number of bacteria with the 3255 bp sequence. There is no homology in E. co//MV1190 or E coll DH5«: (data not

135 9£1

JO UO!10|0P ‘UO11J0SU! ÂUB SBM 0J0L|) 0U|UJJ0;0p 01 0UOP 0J0M S0SA|BUB

yOd slUBUBA s j snodUBiuods am ;o sisA|BUB AjBUjiunaJd

•SlUBiniU S d 8S0LI1 Uj SUOIU0SU! S U l 0L11 LUOJl 0OUBlSlp

0UJOS pojjnooo 0ABLI mSjUJ ‘AuB 1! ‘0dAlOU0Ljd S d 3 - S d 9MÏ P8SHB0 A|lO0Jip

IBLll S06 uBL|O O1UJOU06 0L|1 IBLIl 0OU0pjA0 SppB SIL|_L 'S0OU0nb0S 6ui)jUB|l

LSd 8 L|1 Uj punoi 0J0M SUOjSJOAUj JO SUOjlJBSUj ‘SUO|l0|0p lU0JBddB

ou ‘JBi OS poujBiqo siins0j 0L|i uo posBg Guiouonbos iBipBd puB S|sA|bub

d O d ‘GujddBLu UOI1OIJ1S0J Aq souonbos dqggz£ 0dAi-p|iM 041 qijM p0jBdaioo

0J0M t't'Sd PUB ZSd ' LSd Uj gU_L 041 6uj)jUB|l S0OU0nb0S 041

■(0861. 1 0 6 j© a ) SjBUjUJJSi 0 0 qi.iM

JO 1 ‘ U0 1 UOO 0 0 4 6 1 4 41|M ©IjS IB UOjPOSUj S U l JO; 0 OU0 J0 1 0 Jd UIB1J0 O P0MO4S

4 0 1 4 M S0 ipnis snojA 0 jd 41|m luaisisuoo si S 1 4 1 0 8 si oouonbos dq os sjqi

10 0+0 %|OUJ 041 •QOVOOOOOOOOOOOO^OOOOJ-LOOVOOOOOV si g u i

6uj>jUB|i 0OU0nb0s dq os @41 '©lis lyoo g liei 041 luoj; dq 686 VS peiBOOj si ^ u i

041 g u i 041 pjBMOi s0OU0nbss 6ui)|UB|i ©41 ujoj; 6uiou0nb0s Aq p0Uiujj0i0p

SBM LSd Uj UOjU0SUj p u i ©41 10 lUjOd UOjUSSUj 041 (^ 861 . ‘UBpjOr)

0UJOU06 !}oi!i3LU s 841 Uj 0+0 %|Ouu 69 oi psjBdujoo '(%|OLu o z ) lusiuoo

0+9 4 b j4 B SB4 0OU0nb©s 6uj>jUBii dq g g z£ 041IB41 6ujiS0J©iuj Sj 11

(Zl^ 11 S0UBI ‘9 > O ld) UJ0HBd UOjlOjJlS0J JBjjLUIS B 41jM

0ou©nb0s sno6o|ouj04 b sb4 08-91 nomsuu s 'A|6ujiS0J©iuj mg ©ouanbBS

041 01 A6010UJ04 0AB4 lou S0op oi.oi.rr IJ 0IH9UJ s U0A3 J©4ij0 0ou©nb0s

041 01 A6o|oluo4 0AB4 lOu ssop ‘/JO///01U g 01 p©iB|0J Ajasojo sajoads b

'suapB^eiLinj LunuajOBqojSy ^ Old) bbBuuAs sBuouiopnasd uj jo '(UM04S 1 < — T n 5 1= / / =

— > 2 11 19 9 5 12 7 13 14

—> —■> —■> —•> —> —> —•> —> — > -V- <— <— <— <— <— <— 10 8 20 6 4 3

FIG. 4.7. A schematic view of the primers designed from the 3255 bp sequence for PGR analysis of the Tn5 FS mutants and spontaneous FS variants. See TABLE 4.2 for oligonucleotide sequences.

TABLE 4.3. Primer combinations for PGR tested on selected templates

DNA template Primer pairs tested

RMB7201 2—8, 2—20, 2—6, 2—4, 11—20, 19—6,19—4, 9—6, 9—4, 5—4, 7—3, 12—4,13—3

FS1. FS7 2—1, 2—8, 11—8, 1—6, 1—4, 1—3, 9—6, 9—4, 5—4, &FS44 7 -3

EV10 2—8, 2—20, 11—20, 19—6,19—4, 5—4, 7—3, 12—4,13—3

Strain created by 2—1, 1 —6, 1 —4, 5—6, 5—4, 7—3 marker exchange

137 rearrangement in the 3.25 kb region in the spontaneous FS variant EV10.

EV10 is a recently isolated spontaneous FS variant of the RMB7201 wild type, obtained by enrichment from the periphery of RMB7201 swarm colonies, which has the same FS-EPS phenotype as FS1. Primers used for the analysis were designed from the Tn5 terminus and the sequences flanking the wild-type Tn5 in FS1, i. e. from the 3.25 kb fragment (TABLE 4.2,

FIG. 4.7). The primer combinations tested on selected strains are shown in

TABLE 4.3. We found that DNA fragments generated with FS mutant and the

EV10 variant genomic DNA templates were same as those with wild-type genomic DNA template (results not shown), indicating that no large insertion, deletion or rearrangement occurred in the 3.25 kb flanking sequence region in the Tn5 FS mutants or spontaneous FS variant. Single base pair or short

(1 to 50 bp) insertions, deletions, or genomic rearrangements that do not affect restriction and primer sites cannot be detected in such experiments.

However, DNA fingerprinting of the entire genome generated with arbitrary primers (Caetano-Anollés and Gresshoff, 1994) did show differences between the wild type and spontaneous FS variants in an initial experiment.

Ten mini-hairpin and ten random octomer primers were used for DNA fingerprinting. A number of DNA bands were obtained only in the reactions with the wild-type DNA template. Further experiments are under way to see if these results are reproducible. Future analysis will include cloning and sequencing the unique PGR fragments, and using these fragments as probes to determine if there are any changes in the FS mutants and variants.

138 Similar PCR analyses were also conducted to compare the Tn5 and

mini-Tn5 mutants with the wild type to find out whether the mini-Tn5 was

inserted in the same location as the original Tn5 in FS1. Amplification was carried out with genomic DNAs of FBI, FS7, FS44, FS17, FS27, and FS28.

The mini-Tn5 mutant FS17 and FS28 template DNAs generated bands of same sizes as seen with the DNA templates from FS1, FS7 and FS44,

indicating that FBI7 and FB28 had Tn5 insertions in essentially the same location as the other FB mutants.

However. FB27 and WT13, a mini-Tn5-containing derivative that behaved like the wild type in swarm assays, did not produce any PCR fragment with the Tn5 outward primer and any of the primers complementary to the 3.25 kbp fragment. Thus the Tn5 insertion in these two strains was

located in loci different from that in FB1, 7, 44, 17 and 28.

Complementation of the FS-EPS phenotype by genomic clones. Bince the sequences from the genomic library that flank the wild- type Tn5 did not complement the mutant phenotype, we sought to identify crucial sequences elsewhere in the genome involved in FB-EPB phenotype by screening the wild-type genomic library for clones that restored the normal rate of swarming and EPB synthesis. The pooled RMB7201 genomic library was introduced into FB1 by triparental mating. Mating mixes were plated on

NM-Bm-Bpc-Km-Tc plates. Colonies with more EPB were picked out, purified, and then tested for swarming rates. Out of about 30 candidate

139 genomic clones (designated as pMSD) that complemented the EPS

phenotype, seven of them also appeared to complement the FS phenotype,

and one rendered FS1 nonmotile. (TABLE 4.4).

To test the effect of these clones on FS7, FS44, and other FS variants,

plasmid DNAs were prepared from cosmid-carrying FS1 and introduced into

E. coli DHSoc. After being verified by restriction mapping, cosmids pMSD1-8

were then introduced into other FS mutants by mating with E. coli

DH5oc(pMSD). The majority of the resultant cosmid-bearing strains swarmed

at similar rates to the wild type, indicating that the swarm phenotype was complemented. But there existed slight differences among the cosmids in the

extent of swarm rate reduction with different FS strains (TABLES 4.4 and 4.5).

We noted that these clones also reduced swarming of RMB7201 wild type.

Southern analysis showed that these potentially complementing or suppressing genomic clones did not share any homology with the original

3.25 Kb fragment flanking the Tn5 in FS1 (results not shown). Restriction analysis showed that these seven clones shared some fragments of similar

sizes, and thus may be overlapping to some extent. By restriction mapping, we have found that cosmids pMSDS and pMSD6 have a common, 2.2 kbp fragment that can correct FS-EPS phenotype. This fragment has been partially sequenced, and it shows homology [P(N)

component regulatory system in Burkholderia pseudomallei (Jones et al.,

1997), the ragA and rag S genes in Bradyrhizobium japonicum (Narberhaus

140 et al., 1997), and a number of other two-component response regulators and transcriptional regulatory proteins. The two-component regulatory system of

B. pseudomallei Is Involved In the Invasion of host cells and heavy-metal resistance. The 6. japonicum ragAB genes are also classic two-component system that Is Involved In the regulation of expression of heat-shock {rpoH ) genes (Narberhaus et al., 1997). The 2.2 kbp sequence also has low homology to the exoS gene of S. meliloti. The exoS gene Is a histidine protein kinase. It regulates EPS production (Doherty et al., 1988). Mutation In exoS caused EPS overproduction and reduced swarming In S. meliloti, the reverse of the FS-EPS phenotype.

TABLE 4.4 Correction of swarming behavior of FS1 and FS7 by cosmid library clones». Strain Bwarm colony Strain Bwarm colony diameter fmm) diameter fmm) 7201 (pLAFRS) 24 7201 (pLAFRS) 2S FBI (pLAFRS) 54 FBI (pLAFRS) 48 FBI(pMBDI) 27 FB7(pMSDl) S3 FBI(pMBDS) 28 FB7(pMBDS) 22 FB1(pMBD4) 29 FB7(pMBD4) SS FBI(pMBDS) 28 FB7(pMBD5) 20 FBl(pMBD6) 26 FB7(pMBD6) 16 FB1(pMBD7) 31 FB7(pMBD7) 29 FBI foMSDS) nsb FS7fDMSD81 ns a The two sets of tests (right and left) were done separately In 1 /20 strength TY swarm agar containing 0.3% agar and Tc at 2.5 mg/I. Diameters reported were averages from two replicate plates. Replicate samples normally show less than 10% variation from the mean. Each plate had either 7201 (pLAFRS) or FBI (pLAFRS) or both as controls. The experiment was repeated with same results. The vector pLAFRS was used as a control. b ns=nonswarmlng

141 Table 4.5 Complementation of swarming behavior of FS17, EV2 and EV10 by selected cosmid clones^.

Strain Swarm colony Strain Swarm colony diameter (mm) diameter (mm)

7201 (pLAFRS) 17.5 7201 (pLAFRS) 1S.5

FS1 (pLAFRS) 32 FS1 (pLAFRS) 25

FS17(pMSDS) IS EVIO(pLAFRS) 25

FS17(pMSD4) 19 EVIO(pMSDS) 12

FS17(pMSD6) 17.5 EV10(pMSD4) 21

EV2(pMSD3) 24.5 EV10(pMSD6) 18

EV2(pMSD4) 29

EV2(pMSD6) 17.5 a The two sets (left and right columns) of tests were done separately in 1 /20 strength TY swarm agar (0.3% agar) containing Sm, Spc and Tc at 25, 25 and 5 mg/I respectively. Diameters reported were averages from two replicate plates. Replicate samples normally show less than 10% variation from the mean. Each plate had either 7201 (pLAFRS) or FS1 (pLAFRS) or both as controls.

142 Cosmid clone pMSD4 has been mutagenized with Tn5. The Tn5- containing fragment and the corresponding wild-type sequences have been subcloned (designated as pBS11.7 & pBS5.8, TABLE 4.1). Sequencing is in progress. The partial sequence data available shows that clone pBS5.8 contains a 870 bp sequence with high homology [p(N)

P. viridiflava (Liao et al., 1994), the lemA gene of P. syringae (Hrabak and

Willis, 1992), the apdA gene of P. fluorescens (Corbell and Loper, 1995), and genes for sensor proteins in E. coli (Nagasawa et al., 1992) and Erwinia carotovora (Frederick et al., 1997). These are two-component sensors.

T he E. coli suhB gene and adjacent sequence correct the

FS-EPS p h e n o ty p e . During the initial library screening, a 2.9 kb E. coli

DNA sequence was accidentally identified that restored the S. meliloti FS-

EPS mutant phenotypes to the wild-type levels. Sequence comparison revealed that this clone contained a sequence identical to the E. coli extragenic suppressor gene suhB (Shiba et al., 1984; Yano et al., 1990;

Chang et al., 1991 ; Matsuhisa et al., 1995). This sequence essentially corrected the mutant phenotype of FS1 and FS44 (FIG. 4.7 and TABLE 4.6), as well as mutant FS7 (data not shown). The DNA fragment also partially restored EPS production of FS1 and FS44 (TABLE 4.7). It caused a moderate increase in EPS synthesis and a slight decrease in swarm rates of the wild type as well. In order to test whether this sequence has any effect on

143 TABLE 4.6. Complementation of swarm phenotypea of FS1 and FS44 by cosmid clones pME4 and pME2.2b.

Strain swarm colony diameter (mm) after incubation for 1 d 2 d

7201 (pLAFRS) 16 SB

FS1 (pLAFRS) 28 62

FS1(pME4) 18 S9

FS1(pME2.2) 18 41

FS44(pLAFRS) SO 60

FS44(pME4) 18 S9 a Swarm test was done in 1/20 TY-0.3% agar containing Tc at 2.5 ;/g/ml. Data reported above were averages of two plates from a representative test. Replicate samples normally show less than 10% variation from the mean. The experiment was repeated with similar results. b pME2.2 is a subclone containing the 2.2 kb complementing fragment from the original cosmid library clone pME4.

144 TABLE 4.7. Effect of an E. coli sequence containing tfie extragenic suppressor gene on the EPS production of S. meliloti RMB7201, FS1 and FS44a.

Strain EPS amount % of control (mg/10 ml culture)

RMB7201 (pLAFRS) 5-4 100

RMB7201(pME4) 6.5 120

FS1 (pLAFRS) 1.6 SO

FS1(pME4) S.O 56

FSI(pMES) S.2 59

FS44(pLAFRS) 1.0 19

FS44(pME4) S.O 56 a Data reported above were averages of triplicate measurements from a representative experiment. The experiment was repeated with similar results.

145 Complement EPS Swarm KprH

EcoRV NrtA PviA\ NsA N coi NriA MliA 0 Smal Apa\ EcoRl EcoRV + + I______LI______I______LI_____I______LI______I____I 0 .4 .45 .85 1.3 1.35 1.7 2.2 2.25 2.75 2.9 kbp

suhB

+

4"

0 ______

FIG. 4.8. pLAFR3-based subclones from the E. coli 2.9 Kb EcoRV fragment and their complementing nature. suhB represents extragenic suppressor gene. Tn5 insertion site is designated with symbol o in the corresponding clone pLA8.7 that abolishes complementation. Complementing is designated by 4- (plus) sign, and noncomplementing by - (minus) sign.

146 internal fragment) and the Tn5-interruptted one were introduced into strain

Rm1021. Interestingly, the wild-type sequence reduced the swarming rate of

Rm1021 too (results not shown).

As summarized in FIG. 4.8, it appears that the extragenic suppressor gene is required for EPS complementation or suppression, while the sequence next to the suhB gene is necessary for swarm complementation or suppression. However, the Tn5 is located outside the suhB gene in the Tn5- knockout cosmid (pME17), which cannot complement either EPS or swarming phenotype. Thus complementation of EPS production and FS behavior are controlled by two closely-linked, but different elements. The 400 bp sequence near the Tn5 insertion site is shown In FIG. 4.9. A gene bank search shows that this short stretch of sequence has high {p(N)se*^^} homology to many known genes or sequences, such as the tldD gene

(Murayama et al., 1996), the lamB-malM intergenic region (Vidal-lngigliardi and Raibaud, 1991), araJ gene (Reeder and Schleif, 1991), and the fadAB genes (Yang et al. 1991).

DISCUSSION

Tn5 insertions and FS behavior. As described above, Tn5 is inserted in the same place in the same DNA fragment in all three FS mutants,

FBI, FS7 AND FS44. However, these mutants all have slight differences in their EPS and swimming phenotypes (CHAPTER 3). This raises the question of whether the observed phenotypic differences between the mutants are due

147 to Tn5 insertions in slightly different locations in the 3.25 kbp region or whether the Tn5 insertions occurred at the same location in this region but have caused slightly different genomic rearrangements elsewhere in the chromosome or megaplasmid DNA. We will attempt to determine which of these two alternatives is correct by sequencing the DNA flanking Tn5 insertions in FS7 and FS44 to determine the exact insertion sites and by looking for evidence of genomic rearrangements in the subcloned sequences that complement the FS and EPS phenotypes.

Since the sequences flanking the Tn5 insertion site(s) failed to complement the FS or EPS phenotypes of the mutants, it seems likely that the genetic changes directly responsible for the FS-EPS phenotype probably lie at least 10 kb away from the Tn5 insertion site, perhaps in the region corresponding to the complementary sequences. However, other explanations for the failure of the Tn5 flanking sequences to complement the

FS-EPS phenotype are possible.

More than 40 nonmotile and reduced motility mutants were isolated from the same batches of mating mixes as FS1, FS7 and FS44. With only one exception, each of these non-motile and reduced motility mutants are due to different, independent Tn5 insertions, as verified by Southern blot

(FIGS. 4.1 & 4.2). It indicates that generally Tn5 insertion is fairly random in

S. meliloti RMB7201. It is highly unlikely that chance alone resulted in Tn5 inserting in the same position in all three FS mutants. We believe that there is likely a relation between the Tn5 insertion site(s), or the event of insertion,

148 and the observed FS-EPS behavior.

Because no functional sequence homologous to our Tn5 flanking

sequences was found in the genebank database, we can only speculate

about the relation leetween the region of Tn5 insertion in the FS mutants and

the FS-EPS phenotype. It is possible that the Tn5 transposition event itself

activated certain native IS elements, which hop around the genome, or

caused certain DNA rearrangements elsewhere, resulting in the FS-EPS behavior. At least ten native IS elements {\SRm) have been identified in S.

meliloti {e.g. Zekri et al. 1999; Selbitschka et al. 1995). These IS elements

have been shown to be involved in regulation of processes such as

nodulation and nitrogen fixation (e.g. Dusha et al. 1987; Ruvkun et al. 1982).

According to the results presented in CHAPTER 3, it appears that there Is a switch element in S. meliloti RMB7201 that controls swarming and

EPS biosynthesis. In one state, swarming (flagellar synthesis) is down

regulated, and EPS up-regulated, i. e. the wild-type configuration. In the other state, the reverse is true, i. e. FS-EPS phenotype. Spontaneous FS variants can switch back to the wild-type (WT) phenotype. Although we do not know the exact molecular nature controlling WT-FS switch, our results suggest that a mechanism analogous to phenotype conversion or phase variation may control such switching.

Genetic and regulatory mechanisms similar to our spontaneous switching have been reported in the literature. In the case of phenotype conversion in Ralstonia solanacearum, both EPS and motility are

149 coordinately affected by spontaneous insertions in phcA (Brumbley et al.,

1990). Similarly, in P. atlantica, the degree of EPS production is controlled by insertion and excision of a native IS element (Bartlett and Silverman, 1989).

Work done by Sokol et al. (1994) showed that a genetic rearrangement, caused by an insertion element, led to mucoidy conversion in Pseudomonas aeruginosa PAO. In B. subtilis, motility and competence are co-regulated by a molecular switch (Liu and Zuber, 1998). Similarly, the hexA gene of Erwinia carotovora regulates the expression of motility and other virulence determinant genes (Harris et al., 1998). Analogous to PhcA, the HexA protein is a LysR homolog and a repressor. Interestingly, hexA mutants were hypermotile. It seems possible that our FS mutants have a similar mutation or change in certain repressor genes. Negative regulatory factors have been identified in S. meliloti, such as the nolR gene which controls the expression of the nod regulon (Kondorosi et al. 1991). NolR is also a LysR family protein. LysR is a DNA binding protein which regulates gene expression or transcription (Stragier et al., 1983, Stragier and Patte, 1983). In another study on plant pathogenic bacterium Erwinia chrysanthemi, Castillo and Reverchon

(1997) found that mutant pec-1 showed not only weak pectate lyase production, but also decreased motility and mucoidicity. Pectinases, motility and EPS are all considered virulence factors, and are apparently co­ regulated. However, in contrast to our FS-reduced EPS behavior and the phcA system, both motility and EPS production were down in pec mutants. It appears that although regulation of EPS and motility can be interrelated, it

150 can be upregulation for one and downregulation for the other, or same for both.

Switching and phase variation in flagella and motility. Since the spontaneous FS variants can switch back to the wild-type configuration, reversion should occur from a mutation or change in a presumed phcA or hexA homolog. However, no reversion had been observed from R. solanacearum PC variants. Therefore a more likely mechanism for the FS-

WT switching can be similar to the phase variation of flagellar synthesis in

Salmonella typhimurium (Simon et al., 1980). Phase variation in

Campylobacter jejuni involves the on and off switch of the flaA gene (Nuijten et al., 1995). The mechanism has not been worked out yet, but they showed that the regulatory element was not at the flagellin locus. Xenorhabdus nematophilus FI undergoes spontaneous phase variation. Phase I variants were motile, whereas phase II variants nonmotile (Givaudan et al., 1996).

Possible role of p le C in motility regulation. One potentially complementing sequence recently identified had strong (ca. >60% similarity) homology to pleC, a histidine protein kinase gene implicated in motility in

Caulobacter crescentus. Mutation in pleC has pleiotropic effects (Wang et al.,

1993). The other sequence has moderate homology to two-component system genes and relatively low homology to the exoS gene of S. meliloti

(Doherty et al., 1988). The PleC protein of C. crescentus controls gene expression through the phosphorylation of certain key regulatory proteins. It is known that phosphorelay system is the main mechanism for chemotaxis

151 signal transduction. Chemotaxis signal transducers, or MCPs, are typical two- component systems. While the correction of the FS-EPS behavior by these

DNA elements may be through actual complementation of the mutation in the

FS mutants and variants, it is equally possible that it is through extragenic suppression, changed kinase-phosphatase activities, titration of certain regulatory proteins or DNA sequences important to gene expression.

Mutation in exoS causes EPS overproduction and reduced motility, hence addition of a functional exoS gene may correct FS-EPS behavior.

Possible link between extragenic suppression and the EPS-

FS phenotype. Mutant alleles of suhB of E. coli have been isolated as extragenic suppressors for mutations in protein secretion, heat shock response, and DNA replication (Shiba et al., 1984; Yano, et al., 1990; Chang et al., 1991). These mutations of suhB led to cold-sensitive growth, indicating that the suhB is required for growth at low temperature. The suhB gene product Is essentially identical to the mammalian inositol monophosphatase.

Matsuhisa et al. (1995) showed inositol monophosphatase activity from the purified SuhB protein. They speculated that such phosphatase activity might have a general effect on other processes involving phosphorylation- déphosphorylation, which affects the activity of many regulatory proteins.

Motility and EPS synthesis can be among them.

A study done by Inada and Nakamura (1995) found that, in the absence E. coli SuhB protein, ribonuclease III (rnc gene product) became lethal by its enhanced RNA processing or cleaving activity. Mutations in rnc

152 (defective In RNA cleavage activity) could restore growth of suhB cells. They discovered that rnc mutations did not restore or change suhB expression or activity. Rather, it was more likely that SuhB controls the activity of RNase III, which itself is potentially lethal to E. coli in the absence of SuhB. Thus the complementation or suppression of the EPS phenotype by suhB could also be due to its inhibition of RNase activity, leading to changes In messenger

levels. Again we can speculate that introduction of suhB into FS mutants probably stabilizes mRNAs for EPS production.

As described in the results section, Tn5 interruption of pME4 abolished apparent complementation of the FS-EPS phenotype. The 400 bp sequence adjacent to the Tn5 is shown in FIG. 4.9. Interestingly, this sequence , which seems to affect FS behavior, showed high homology

[P(N)

(Murayama et al., 1996). The fadAB gene products are proteins that catalyze fatty acid oxidation (Yang et al. 1991). Whereas the araJ gene is part of the

E. coli arabinose regulon, but its function has not been identified yet (Reeder and Schleif, 1991). Other E. coli genes or sequences that share high homology to this sequence include the genes for colanic acid (EPS) synthesis (Stevenson et al. 1996), sucBA (for succinate metabolism)

(Spencer et al. 1984), rhaBAD (Egan and Schleif, 1993), IpcA

(lipopolysaccharide) (Brooke and Volvano, 1996), the hemB, the phn (psiD)

153 (Chen et al. 1990), and genes involved in DNA damage repair such as lexA

(Lewis et al. 1994) and ruv genes (Benson et al. 1988). This sequence appears to be an important regulatory region with highly conserved nucleotide sequence. The complementing nature of this E. coli sequence suggests the potential existence of a similar DNA element in S. meliloti, and of possible mutation or genomic change in the FS mutants or variants. But low stringency Southern hybridization did not detect any in S. meliloti

RMB7201.

1671 GGTAACAAGG CCGAATAACA CCGTTTGGCC TGATGCGAAG 1711 CCTTAATGCG TCTTATCAGG CCTACAGTGA ACAGAACCGT 17 51 AGGTCGGATA AGGCGTTCAC GCCGCATCCG ACAGCCGTTG 17 91 CCTGATGCGA CGCGTAATGC GTCTTATCAG GCCTACAGTG 18 31 AACAGAACCG TAGGTCGGAT AAGGCGTTCA CGCCGCATCC 1871 GACAGCCGTT GCCTGATGCG ACGCGTAATG CGTCTTATTC 1911 AGGCCTACAG TGTACAGACC CCAGGCGGAT ATGCTTCAAC 1951 GCCGCATCCG ACAACAGGTA CAAACGCCAC GATAAAAAAA 1981 TGGCACTGAA GGTTAAATAC CCGACTAAAT CAGTCAAGTA 2 0 2 1 AATAGTTGAC CAATTTACTC GGGAATGTCA AATACTCCGT Î 2060

FIG. 4.9. 400 bp sequence from the 2.9 kbp EcoRV fragment. This sequence has homology to at least 20 genes in the gene bank database.The Tn5 is located between basepairs #2060 and #2061 in the knockout cosmid pME17.

154 According to the position of the sequence in these genes, it is probably a highly conserved regulatory element, e. g. a promoter or some sort of binding site for transcriptional factors.

Proposed near future studies: Further studies are needed to pinpoint the changes directly linked to the FS-EPS behavior. First, subclones described above are being fully sequenced and a genebank search will be done to determine if the DNA sequence has any homology to known genes or DNA elements. The second experiment is to subclone the potentially

complementing sequence interrupted by Tn5, and marker exchange it back into the wild type to see if it can regenerate the FS-EPS phenotype. This will

give a clue to whether the complementation is actually by providing a wild-

type copy of DNA element corresponding to the mutation, or due to

extragenic suppression, or titration of certain regulatory factors. Thirdly, the potentially complementing sequences have been labelled and will be used

as probes to determine if there is any deletion, insertion or DNA rearrangement (over 50 bp) in the corresponding region in the FS mutants and spontaneous variants. Alternatively, a full scale DNA fingerprinting will be necessary to find out the difference between the wild type and the FS mutants or variants. Once unique DNA fragments are obtained from either FS variants or the wild type, further analyses such as Southern hybridization and sequencing can be followed. More discussions on future directions are presented in CHAPTER 6.

155 CHAPTER 5

EFFECTS OF N-ACYL HOMOSERINE LACTONE (AHL) SIGNALLING ON GROWTH, ADHESION AND ROOT COLONIZATION BY PSEUDOMONAS AERUGINOSA

ABSTRACT

In many Gram-negative bacterial species, certain genes are under control of diffusible signal molecules produced by the bacteria. The signal compounds are usually N-acyl-derivatives of homoserine lactone (acyl-HSL or AHL), and the accumulation of such signals is cell density dependent, thus this system is known as autoinduction, or AHL signalling or sensing. P. aeruginosa is one of the model species for studies of AHL signalling. In this preliminary study, the effect of mutations in AHL production and reception genes on growth and plant root colonization were investigated. The results showed that mutant PAO-JP2, defective in AHL synthesis, had a lower initial growth rate than the wild type in LB shake culture. Strain PAO-JP3, a signal- sensing deficient mutant, exhibited normal growth rate in rich media, but grew more slowly in mineral solution or in sand moistened with mineral solution.

Both mutants had lower efficiency in root colonization than the wild type. We also found that the culturability of P. aeruginosa was greatly affected by

156 environmental factors, including salt and antibiotic concentrations, moisture level, and by culture age and initial cell densities plated. The culturability of strain PAO-JP2 was more sensitive than the parent to these conditions. The results indicate that AHL signalling plays an important role in growth, root colonization and resistance to adverse environmental factors.

INTRODUCTION

AHL signalling and quorum sensing. The expression of certain genes of many gram-negative bacteria is subject to regulation by population density ( for recent reviews, see Hussain et al., 1998; Salmond et al., 1995).

This regulatory system, which involves the synthesis, extracellular diffusion and perception of AHL signal molecules, is often called autoinduction, cell-cell communication, or quorum-sensing. It regulates a variety of physiological processes such as bioluminescence, plasmid conjugal transfer, and production of virulence factors like EPS (eg. von Bodman et al., 1998) in pathogens. Many plant-associated bacteria such as Agrobacterium tumefaciens (Zhang and Kerr, 1991), Rhizobium (Gray et al, 1996), and

Pseudomonas produce these compounds (Cha et al., 1998). AHL signalling coordinates cells' physiological and metabolic processes in accordance with population density.

AHL signalling was first discovered in the bioluminescence of

Photobacterium {Vibrio) fischeri (Nealson et al., 1970). The regulation of bioluminescence was found to require small, diffusible signal molecules

157 called autoinducers produced by the bacteria. When cell density reached high

levels, as in the light organ of the host fish, the concentration of autoinducers

increases to a threshold that activates positive transcriptional activator. LuxR.

which in turn triggers the transcription of genes for bioluminescence and the

luxi gene required for synthesis of the AHL signal (hence the term

“autoinducer”, see review by Salmond et al.. 1995).

The signals mediating cell density responses in V. fischeri and other

Gram negative species are N-acyl derivatives of L-homoserine lactone. Side

chain length and the functional groups on the side chain determine their

specificity for receptors. A given bacterial species may make several different

AHLs and have corresponding LuxR-like receptors, controlling different

behaviors, for each.

Quorum-sensing mutants of P. aeruginosa. In P. aeruginosa, the

production of factors involved in pathogenesis requires bacterial cell-cell

communication. In strain PA01. at least three autoinduction systems have been identified. The first two are called the las and rhi systems, which control

elastase and rhamnolipid biosynthesis genes, respectively (Pesci et al..

1997). The main AHL signal molecules are N-(3-oxododecanoyl)-L-

homoserine lactone (OdDHL) and N-butanoyl-L-homoserine lactone (BHL). A third AHL signalling pair consists of the genes vsmR and vsmi (Winson et al..

1995). This system involves the synthesis and sensing of BHL and N- hexanoyl-L-homoserine lactone (HHL).

158 Our experiments have compared growth and adhesion in sand and on root surfaces by the PA01 wild-type strain of P. aeruginosa versus that of two

mutants of PA01 affected in AHL signalling. The first mutant, PAO-JP2, is a double la s i rh il deletion mutant, and is unable to synthesize either BHL or

OdDHL. The second mutant, PAO-JP3, is a double lasR rhIR deletion mutant, and lacks the receptors for BHL and OdDHL (Gambello et al., 1993; Winson et al., 1995; Pesci et al., 1997). The interrelation between the las and rhI systems is complex (Pesci et al., 1997). LasI catalyzes the synthesis of

OdDHL. OdDHL, in turn, activates the positive transcriptional activator LasR.

LasR turns on the expression of genes required for elastase and rhamnolipid biosynthesis , as well as the las! gene, hence amplifying transcription when cell density is high. Similarly, in the rhi system, RhIR is a receptor- transcriptional activator and Rhll controls synthesis of BHL. It was discovered that lasR is expressed constitutively and that rhIR is expressed during late log- phase when cell density becomes high. The rhi system is largely controlled by the /as system.

Bacterial adhesion and root colonization. Bacterial adhesion to solid surfaces is important for certain cell activities such as nutrient acquisition, cell-cell interactions, and biofilm formation. A number of factors have been found to affect bacterial adhesion to surfaces. Cell surface characteristics, such as lipopolysaccharides, charge, hydrophobicity, flagella and pin, are among those factors (Stenstrom, 1985; Stenstrom and

Kjelleberg,1985; Genevaux et al., 1996; Williams and Fletcher, 1996). But so

159 far no study has examined the effect, if any, of autoinduction on soil adhesion and root colonization.

P. aeruginosa is ubiquitous, not only as a pathogen to animals but also a common soil resident (e.g. Green et al., 1974; Olsen et al., 1996; Kelley et al., 1998), and may be common in rhizospheres (Yeung et al., 1989). We were interested in the possible effect of mutations in the las and rhi genes on adhesion to sand or soil particles, and on root colonization. Preliminary results show that these mutations reduce the adhesion and root colonization ability of the mutant strains. We also found that growth rate of the quorum sensing mutants is decreased substantially.

Discrepancies in determination of viable cell counts for P. aeruginosa. Because P. aeruginosa is an opportunistic human pathogen and an important model species for many studies, the organism is routinely cultured in laboratories for identification and cell number counting.

Pseudomonas Isolation Agar (PIA) is designed for easy use and selective isolation of pseudomonads. When amended with glycerol, PIA can be used to specifically identify P. aeruginosa by its characteristic blue color. Because it is convenient, a great many researchers and clinical laboratories routinely use

PIA. The medium contains the antibiotic irgasan, which is reported not to inhibit Pseudomonas strains at concentrations below 25 mg per liter. In our study of P. aeruginosa strains, we noticed that the number of colony forming units (CPU) counted on PIA plates was significantly (10- to 200-fold) lower than that from LB plates. This led us to investigate the possible cause of this

160 great discrepancy in colony forming ability of the strains.

MATERIALS AND METHODS

Strains and media. Strains used in the study, Pseudomonas aeruginosa PA01, PAO-JP2 {Iasi rhll) and PAO-JP3 {lasR rhIR), were provided by Dr. Iglewski (Gambello et al., 1993; Winson et al., 1995; Pearson et al., 1997). Other strains used include P. aeruginosa FRD1 (wt), algD:Jn501, and algT:Jr\501 (Goldberg and Ohman, 1987), a twitching motility defective (Twf) mutant PAOFA3, and a non-flagellated mutant SW1

(Fla')- Strains PA01, FRD1, PAO-FA3 and PAO-SW1 (Fla') were cultured In

LB. PA0-JP2 and PAO-JP3 were also cultured in LB, or when necessary, in

LB containing tetracycline (To) at 20 p.g/ml, or plus HgCl^at 7.5 ^g/ml in solid media, unless specified. Liquid cultures were incubated at 28°G on a rotary shaker at 175-200 rpm. PIA, purchased from Difco (Becton Dickson, Detroit,

Ml), was used according to manufacturer's instructions. In some experiments, colony counts were also obtained from TY or NM media (see CHAPTER 2) for comparison.

Fluorescently marked derivatives PA01(pGB3) and PA01(pSMG2) were constructed by transforming PAG1 competent cells with plasmids pGB3 and pSMC2, in which a green fluorescent protein (GFP) gene is cloned

(Bloemberg et. al.,1997). These marked strains could easily be distinguished from other strains by fluorescence microscopy, which was carried out on a

161 Zeiss 1-35 Inverted microscope with an FITC filter. Fluorescent colonies were directly detected under a low-power objective, and single cells observed on wet mounts at a 860x magnification.

Adhesion to sand and root colonization tests. Quartz sand used in the study was washed in running water overnight, then oven dried at 80°C.

Three hundred g portions of the washed, dry sand was added to plastic

Magenta boxes which were then autoclaved. Bacterial cells were harvested from late exponential phase cultures by centrifugation, and cells were resuspended at an appropriate cell density (ca. 10"^ to 10^ CFU/g of sand) In

1/4 strength N-free Hoagland solution (Hoagland and Arnon, 1950), or 1/4

Jenson’s plant growth medium (full strength containing per liter; 0.2 g of

KgHPO^, 0.2 g MgSO^'THgO, 0.2 g NaCI, 1.0 g CaHPO^, 0.1 g FeClg-GH^O, and 1 ml of Hoagland trace element solution, pH 7.0). Equal numbers of washed wild-type and mutant cells were mixed and inoculated into the sterilized sand. The volume of inoculant suspension was just sufficient to bring the sand to field capacity, so no mechanical mixing was required.

After incubation at room temperature for a specified time, the sand was thoroughly mixed, and replicate 2 to 5 g sand samples were transferred to 15 ml screw-cap tubes. Unattached bacterial cells were recovered by two rinses of the sand with 5 ml of sterile water. After these rinses, bacterial cells that adhered tightly to sand particles were recovered by three rounds of rigorous vortexing and indirect sonication for 3 min at 25% energy output in a cup-horn

162 sonicator ( Model W-370, Heat System-Ultrasonics, Inc.). The pooled extract suspensions were diluted and plated out on media with and without To for the determination of CPUs of the mutants and wild type, respectively. After extraction, some of the remaining sand was briefly blotted on sterile filter paper and about 100-200 particles transferred onto LB plates to determine the percentage of particles with bacteria that remained attached to the sand throughout the recovery process. Usually less than half of the sand particles still had bacteria attached to them, indicating almost complete recovery from the surfaces.

Root colonization assays were done as follows: soybean and pea seeds were surface sterilized with 10% Clorox for 8 min and rinsed with sterile water 15 times. After soaking in water for 6 h, the seeds were planted in the inoculated sand boxes. After incubation for 3 to 5 days, sand and plants were emptied from the Magenta boxes into a sterile pan. After gentle removal of the plants, the sand was mixed, and 2 to 5 g sand samples taken. Recovery of bacteria from the wet sand was carried out as described above. Plant roots were harvested by carefully shaking off the loose sand particles. A sterile razor was then used to excise two 1-cm segments from each root, one that included the root tip and a second one adjacent to this tip segment. Bacteria loosely attached to root surfaces were recovered by rinsing with sterile water three times. Cells bound tightly to the roots were recovered by sonication for 3 min, or by homogenization at full power for 3 min in a Waring Blender (model

33BL79, Dynamics Cor. of America, New Hartford, CT). The suspensions were

163 diluted and plated out for the enumeration of CPU.

RESULTS

Effect of rh I and la s mutations on growth rates. To compare the growth rates of P. aeruginosa PA01, PAO-JP2 and PA0-JP3, the absorbance of the cultures at 590 nm at various times during growth in LB was determined, and CPU on LB plates also determined. As shown in PIG. 5.1, PAO-JP2, the

Iasi rh/7 double deletion mutant, grew more slowly than PA01 or PAO-JP3 at the beginning. But the maximum growth of all three strains was essentially the same. OdDHL was added to the LB medium to test whether exogenous AHL promoted the growth rate of PA0-JP2 in LB shake cultures. As shown in PIG.

5.2, the addition of the autoinducer did not increase the growth rate of either strain. Purther experiments are needed to determine whether there is some other concentration at which OdDHL is effective in stimulating the initial growth rate of JP2. The growth rate of strain PAO-JP2 on LB agar plates was also noticeably slower than the wild type or PAO-JP3. It took 40 h for PAO-JP2 to develop colonies of 2 to 3 mm in diameter, but only 24 h for the wild type and PAO-JP3. Similar results were observed on TY and PIA media.

On rich media, strain PA0-JP3 grew at similar rates to the wild type or

PA01(pGB3) (Pig 5.1). However, in 1/4 Hoagland solution with no added carbon or nitrogen source, PA01(pGB3) outgrew PA0-JP3 at least two fold in one day (TABLES 5.1 and 5.2). Similar results were obtained with PAO-JP2 in

164 12 10

u 10 -□ -V 0 -o o 10 1CO E ■ a CO c 10^ § 0) =) 8- LL 8 Ü o 10

10"

10" 0 10 15 20 25 90 95 100 Time of incubation (h)

FIG. 5.1. Growth rates of P. aeruginosa PA01 (o), PAO-JP2 {rh llrh ll) (V ) and PAO-JP3 {lasR rhlR) (□) in LB. Results are averages of duplicate samples from a single experiment that was repeated with similar results.

165 10

LL O

6 10 0 5 10 1520 25 time of incubation (h)

FIG. 5.2. Effect of OdDHL on the growth rates of P. aeruginosa PA01 wt (o) and PAO-JP2 (v). The strains were cultured in LB liquid medium (open symbols) and In LB amended with 0.01 mM of OdDHL (filled symbols). The experiment was repeated once with similar results.

166 TABLE 5.1. Growth of PA01(pGB3)a (WT) and PAO-JP3 (1:1 mixed inoculation) in 1/4 Hoagland solution^.

Incubation Final CFU ratio of WT CFU time inoculum CFU to JP3CFU WT JP3 day 1 40 19 2.1 day 4 39 17 2.3 a Colony counts were done on LB agar containing To at 10 mg/I. The marked WT was distinguished from PAO-JP3 by fluorescent microscopy, b Inoculum was prepared from overnight cultures (c.a. ICno QFU/ml) by centrifugation and resuspension of the pelleted cells in Hoagland solution. Cell density was adjusted to about 1Q5 CFU/ml of each strain. Cell suspensions were incubated at 28&C on a rotary shaker at 200 rpm. Values reported were averages of triplicate plating from duplicate cultures.

sand moistened with 1/4 Jenson’s solution (TABLE 5.3), showing a growth advantage of the wild type over PAO-JP2 in sand of about 6-fold. Importantly, all three strains were able to multiply at least 5-fold in mineral solution or sand containing no added carbon or nitrogen source.

Relative adhesion to and growth in sand. For the determination of adhesion of the strains to sand particles, 1:1 mixtures of wild type and PAO-

167 JP3, or PA01(pGB3) and PAO-JP3 were inoculated into sand and incubated at room temperature. Cells recovered in the initial rinsates were considered not to be associated with sand particles, while cells obtained by vortexing and sonication were classified as fairly tightly attached to sand particles. As shown in TABLE 5.2, the cell densities of the wild type and the GFP-marked wild-type strain were higher than PAO-JP3 not only in rinsates, but also in the

TABLE 5.2. Growth in and adhesion to moist sand of PA01, PA01(pGB3), and PA0-JP3 after 1 day incubation.

Inoculum Categorya CFUb/o of wet sand Corrected ratio WT PA0-JP3 ofWT/JP3c

Mix of Free 2.75 X 105 1.1 X 104 18 PA01 and PAO JP3 Attached 4.7 X 105 1.58 X 104 21

Mix of Free 3.6 X 105 2.5 X 104 12 PA01(pGB3)d and PAG JP3 Attached 2.33 X 105 1.14 X 104 16 a See Materials and Methods for extraction and classification, b Data are the mean from two boxes. Replicate sand samples were plated to get the average cell counts per g of wet sand from each box. c The inoculation suspension contained 8.4 x 103 wild type cells and 6 x 103 cells of PA0-JP3 per g of wet sand • d PA01(pGB3) was distinguished from PA0-JP3 by fluorescent microscopy.

168 TABLE 5.3. Growth and adhesion of P. aeruginosa PA01 and PAO-JP2 to sand particles after 1 day incubations.

Category CFUbofwt CFU of PAO-JP2 Ratio of wt to PAO-JP2

Free 5-5 X 105 1.5 X 105 3.6:1

Attached 9.6 X 105 1 .2 x 105 7.8:1

Total CFU 1.5 X 106 2.7 X 105 5.5:1

Ratio of attached 1.77 0.82 2.16 to free cells a Equal numbers of cells (ca. 1.7 x 1 os/g) of the two strains were mixed and inoculated into sand moistened to field capacity with 1/4 strength Jenson solution. The inoculated sand was incubated at 22°C for 1 d. Duplicate samples containing about 4.7 g of wet sand were rinsed with sterile water 4 times, then extracted by sonication/vortexing. VCC was determined on LB and LB-Tc plates. b Numbers given are CFU per g of wet sand, and are averages of 2 replicates. The experiment was repeated once with similar results.

169 TABLE 5.4. Colonization of soybean roots by PA01 and PA0-JP2 (las I rhi I) 6 day after plant!nga.

Root Category CFU of wild type CFU of PAO-JP2 Ratio of region per root per root WT/PAO-JP2

Root loosely 4 . 8 X 1 0 4 3 . 5 X 1 0 3 1 4 :1 tip attachedb

boundc 2 . 8 X 1 0 4 1 . 9 x 1 0 3 1 5 :1

Root loosely 1 . 4 x 1 0 5 3 .1 X 1 0 4 4 .5 :1 mature attached region bound 4 . 4 X 1 0 4 6 .5 X 1 0 3 6 .8 :1

Total 2 . 6 X 1 0 5 4 .3 X 1 0 4 6 :1 a See Materials and Methods for details of root harvesting. Data above was from duplicate samples in a single experiment. b Primary roots of 10 seedlings were harvested, cut into segments, then rinsed with sterile buffer four times. The combined rinsate was plated for enumeration of loosely-attached cells. 0 After rinsing, the root segments were sonicated twice, and the extracts were plated for counts of bound cells.

170 extracts obtained by vortexing and sonication. The cell counts of both strains increased substantially from the counts immediately after inoculation. PAO-

JP3 and PAO-JP2 both grew more slowly than the wild-type strains in sand moistened with 1/4 strength Hoagland solution, although PAO-JP3 grew as fast as PA01 in LB medium (FIG. 5.1).

Root colonization by PA0-JP2.Root colonization of soybean by P. aeruginosa PA01 and PAO-JP2 was examined by recovering bacteria from root segments of soybean planted in sand wetted with 1/4 strength Jenson’s solution. On the root tip segments, the cell densities of the wild type, both as loosely-associated and root surface-bound, were over 13 times greater than those of the mutant (TABLE 5.4). On the more mature root segments, the density of wild-type cells was 4.5- to 6.8-fold higher than for loosely- associated and surface-bound cells of PAO-JP2, respectively. The overall cell densities of both strains were higher in mature regions than those on root tips.

In the root tip region, the ratio of bound to loosely-attached cells was 0.58 and

0.54, while in the mature zone, the ratios of bound vs free cells were 0.31 and

0.21 for the wild type and PAO-JP2 respectively (TABLE 5.4).

Inhibition of P. aeruginosa colony formation.A large discrepancy between PA01 colony counts on PIA vs LB plates was observed when using the PIA medium for routine CFU enumeration. As shown in TABLE

5.5, PA01 and both quorum-sensing mutants formed more colonies on LB than on PIA. When the cultures were young, the difference was less than 2- fold for the wild type and PAO-JP3, whereas strain PAO-JP2 formed 7 times

171 TABLE 5.5. Effect of culture age on colony formation on PIA and LB of P. aeruginosa PA01, PAO-JP2, and PA0-JP3.

Strain Ratioa of colonies on LB to colonies on PIA

______Exponential Phase culture^ Stationary phase culturec PA01 (wt) 1. 8 2,300

PAO-JP2 7.6 4,900

PAO-JP3 1.3 1,400 a PA01 was diluted and plated on both LB and PIA plates, PAO-JP2 and PAO- JP3 on LB, LB-Tc, LB-Hg and PIA plates. Data above were taken from a single experiment. The experiment was repeated with similar trend, b Early exponential phase culture; OD 5 9 0 of the three cultures in LB was around 0.10-0.13. c 24 h old cultures, OD 5 9 0 was around 2 .0 .

TABLE 5.6. Colony formation of PA01, PAO-JP2, and PAO-JP3 on LB and NM mediaa.

Strain Ratio of CFU on LB to CFU on NM

PA01 (wt) 1.4:1

PA0-JP2 2 .0 : 1

PAO-JP3 1.9:1 a 3 day-old cultures in LB were diluted in sterile HgO, and plated out on LB and NM (succinate/KNOa). The numbers reported above are averages of two replicates. The experiment was repeated once with similar results.

172 more colonies on LB than on PIA. Stationary phase cultures of these strains, however, showed a difference of 3 orders of magnitude for CFU on LB and

PIA. Cultures from pure single colonies grown on PIA medium gave the same results, therefore the inability of a large fraction of cells to grow on PIA is not inheritable. CFU determinations on PIA plates were not highly reproducible from one experiment to another.

From TABLES 5.5 and 5.6, it can be seen that P. aeruginosa strains formed different numbers of colonies on different media. The strains were able to grow on minimal (defined) medium, though the CFUs were about 2-fold lower than on LB. To test whether glycerol (present in PIA) or NaCI (in LB) were responsible for the differential formation of colonies on LB and PIA, we plated P. aeruginosa PA01 and FRD1 on PIA containing NaCI but no glycerol, and on LB containing glycerol but without NaCI. The results showed that neither glycerol nor NaCI had any effect on CFU for either strain (data not shown). Thus it is the antibiotic irgasan in PIA which inhibited colony formation of P. aeruginosa PA01, FRD1, and the quorum sensing mutants.

In order to find out whether the observed nonculturability of PA01 subpopulations on various media was peculiar to PA01 or a more general characteristic of P. aeruginosa, we repeated the previous experiments with strains P. aeruginosa FRD1, another mucoid wild type, and two mutant derivatives of it, algD: Jr\501 and algT.'JnSOI. The results obtained from LB.

PIA and NM-succinate media confirmed that PIA severely inhibited the growth

173 of each of the above strains (data not shown). With exponential phase

cultures, the colony counts on PIA plates were only 1 to 2% of those on LB or

NM media. With 48 h old cultures, colony counts on PIA were even lower

(<1/1000 of the counts on LB). In contrast, on NM medium and LB plates,

strains FRD1 and algD:Tn501 formed comparable numbers of colonies.

We also found that the colony forming ability of the P. aeruginosa

strains was affected considerably by factors like culture age, salt content,

moisture level of media, and concentration of antibiotics. For example, strains

PAO-JP2 and PAO-JP3 contain a Tn501 insertion that carries Hg and Tc

resistance markers (Pearson et. al., 1997). The recommended concentration

for Tc is at 50 jug/ml. Even when present at concentrations of 10, 20 or 25

/ig/ml, the higher the antibiotic concentration, the lower the CFU (data not

TABLE 5.7. Effect of cell density on colony forming ability of P. aeruginosa PA01 and FRD1 on PIAa.

Strain Cell Colonv # on % of cells forming density LB PIA colonies on PIA

PA01 low 102 0 0 high 9.3x103 5.6 X 102 6

FRD1 low 7.1 x102 10 1.4 high 5 x103 3.3 X 102 6.6 a The data for PA01 and FDR1 were determined in two separate experiments. Cells were spread with a Spiral Plater so were not evenly distributed. Similar results were obtained in other experiments.

174 shown). Different levels of HgClg or NaCI in LB medium also affected colony counts (data not shown).

Cell density plated on PIA agar also greatly affected the percentage of cells that could form visible colonies. When spread with the Spiral Plater,cell density in areas near the center was higher than areas near the plate edge.

We observed that many more colonies formed in the center area, but essentially no colonies were observed near the edge. Results from two typical experiments were combined in TABLE 5.7. It can be seen that when plated at higher cell density, a larger fraction of cells were able to form colonies on PIA plates. It should be pointed out that culture age affected the exact percentage of cells able to grow on PIA, but different experiments gave the same trend.

DISCUSSION

Possible role of la s and rhi genes in growth in shake culture on LB. Strain PAO-JP2, the A/as/ àrhIT double mutant, grew much more slowly at early stages in LB shake cultures than either the wild type or PAO-

JP3, the àlasR Arhl R double mutant (FIG. 5.1). During early exponential phase, the calculated generation time for PAO-JP2 was about 106 minutes, while for PA01 and PA0-JP3 it was around 61 minutes. During mid- and late- exponential phases. PAO-JP2 grew at rates similar to those of the wild type and PAO-JP3 and reached the same maximum cell density as the other two strains. Hence it is unlikely that the mutations in PA0-JP2 caused any defects

175 in metabolism or other AHL-independent processes. The slower growth rate of

PAO-JP2 in the early stages of growth, at low cell densities, was quite unexpected. AHL-mediated regulation of behavior is expected only at high cell densities where threshold concentrations of AHL signals have accumulated- Our results raise the possibility that AHL sensing systems may have important functions in P. aeruginosa at low cell density, a possibility strengthened by several of our other findings.

Since addition of OdDHL to PAO-JP2 cultures did not restore its growth rate to wild-type levels, it is possible that defective synthesis of OdDHL and

BHL was not responsible for the slower growth of PAO-JP2. For example, there may be some artifact of mutant construction that caused delayed growth.

However, given the other effects of mutations in AHL synthesis and perception that we have seen on colony formation on PIA and other media, and the effects of these mutations on growth at low cell density in sand, it seems more likely that the initial slow growth of PAO-JP2 in LB shake cultures is indeed a reflection of AHL deficiency. Further studies are needed to determine whether it is the lasi or rh ll gene that is responsible for the effects on initial growth rates. It will also be important to test whether there is a specific concentration of either OdDHL or BHL (or both) that can restore normal growth. Since PAO-

JP3 did not show slow initial growth on LB. it would also be most interesting to find out whether PAO-JP2’s slow initial growth results from the reduced interaction of these AHLs with a LuxR homolog other than lasR or rhlR.

Several research groups have reported the existence of multiple homologs

176 of LuxI and LuxR in P. aeruginosa (e. g. Latifi et al., 1995).

Role of AHL sensing and responses on growth In sand. The experiments on sand and root colonization are preliminary, but clearly showed that mutations in AHL production and sensing reduced the rate of growth in sand (TABLES 5.2 and 5.3). The 2- to 20-fold differences between growth of PA0-JP3 and the wild type in Hoagland solution shake cultures and in sand wetted with Hoagland solution were quite unexpected, since PAO-JP3 grew just as fast as the wild type in rich medium (FIG. 5.1). These results clearly implicate either the lasR and/or rhIR gene in regulating behaviors necessary for growth under low nutrient conditions. The 5- to 7-fold reduced growth of PA0-JP2 confirms the importance of AHL signalling under these conditions and suggests that synthesis of BHL and/or OdDHL are responsible for a major part of the growth effect. To our knowledge, this is the first time that

AHL signalling has been implicated in the regulation of behavior crucial to growth of a bacterium at low cell density and low nutrient availability.

The wild type had only a 2-fold growth advantage over PAO-JP3 in mixed shake cultures in 1/4 Hoagland solution (TABLE 5.1), while in sand moistened with 1/4 Hoagland solution, PAO-JP3 grew 10- to 20-fold slower than the wild type (TABLES 5.2). This difference indicates that the major role of AHL sensing is not so much in adapting to low nutrient availability at low cell density as in adapting to the microenvironment provided by sand surfaces. It remains to be determined how the sand matrix provides the wild type with substantial growth advantages over AHL sensing mutants. It also

177 remains to be determined why PAO-JP2 did better than PAO-JP3 in moist

sand. One possibility is that PAO-JP2’s defect in AHL synthesis could be

partially compensated for by AHLs synthesized by neighboring wild type cells.

Alternatively, some AHLs other than BHL or OdDHL may be involved.

Effects of AHL sensing and perception on adhesion to sand.

The data in TABLES 5.2 and 5.3 show that the ratio of cells adhering to sand

particles compared to those free in the surrounding liquid was about 20%

higher for the wild type than for PAO-JP3, and about 2-fold higher for the wild

type than PAO-JP2.

Glessner et al. (1999) recently reported that the las and rhi systems are

required for pilus-driven twitching motility of P. aeruginosa. They also found

that the las and rhi mutants are several-fold impaired in adhesion to human

epithelial cells and to sand columns. Since cell-to-cell communication was

involved in the formation and development of P. aeruginosa biofilms (Davies

et al., 1998), and autoinducers were reported to be present in biofilms in

natural habitats (McLean et al., 1997), one can speculate that adhesion and

establishment of bacterial communities on sand surfaces may be similar to

biofilm formation, and that cell-cell signalling may affect the process.

Root colonization. The AHL mutant PA0-JP2 also had a significant

disadvantage compared to the wild type in growth and colonization on root

surfaces (TABLE 5.4). The 14-fold difference in the root tip region was greater

than the 6.8-fold overall growth advantage of the wild type to PAO-JP2 in sand. In the mature region of the root, about 24% of the wild type cells were

178 attached to the surface compared to 17% for PA0-JP2, indicating that mutations in Iasi and rhll had a modest effect on root adhesion. According to the data in TABLE 5.4, the calculated average distance between P. aeruginosa cells on the surface in the root tip region was about 14 p.m, and about 8.5 um on the mature root segments. These cell densities are quite high, comparable to 5 to 10 pim average distances between cells in a stationary phase liquid culture. Thus, AHL-mediated cell-cell interactions are likely to occur on the root surface. Cells on the root surface may form microcolonies in which cells can interact with one another too. Further experiments using Iasi and rhll single mutants are needed to determine which gene(s) is involved in plant root colonization.

Inhibitory effect of PIA on P. aeruginosa. We re-discovered the inhibitory effect of PIA and other growth conditions on P. aeruginosa strains. A literature search identified at least three reports (Marold et al., 1981 ; Fonseca, et al., 1986; Grobe et al.. 1995) that mentioned the discrepancy in enumerating some strains of Pseudomonas on PIA and on media with other selective agents. Grobe et al. (1995) pointed out that only when irgasan content was reduced to 0.8 mg/I from the original 25 mg/I that the discrepancy in colony counts was avoided. We think that PIA, the standard medium for isolation and cell number counts of Pseudomonas spp. and P. aeruginosa in particular, can give very misleading viable cell counts. PIA medium should not be used for the purpose of cell counting at all, especially not in clinical and

179 environmental laboratories where many critical decisions are made on the basis of colony counts. This dependency of colony formation on various conditions Is not observed In species such as S. meliloti, whose colony counts are very reproducible on various media, or on a medium with different concentrations of antibiotics. Therefore caution should be exercised In dealing with Isolation and cell counts of P. aeruginosa strains. One should also be critical when Interpreting the bulk of existing literature regarding VCC of P. aeruginosa strains.

Perhaps the simplest Interpretation of our results Is that PIA and other stress-inducing agents can prevent a variable-sized subpopulation of P. aeruginosa cells from forming visible colonies. From the behavior of PA0-JP2

(TABLES 5.5 and 5.6), culture age (TABLE 5.5) and the effects of cell plating density (TABLE 5.7), It appears that AHL signalling Is a significant factor In determining the size of the stress-susceptible subpopulation. We speculate that the delayed growth of PAO-JP2 In LB shake culture might be due to the presence of the large subpopulation of PA0-JP2 cells that fail to form colonies under stressful conditions. The delayed growth In shake culture might reflect the time required for the “stress-susceptible” subpopulation of PAO-JP2 cells to become like the remaining cells, ready to divide rapidly under favorable conditions. The role of AHL signalling In the determination of subpopulation size remains to be explored, as does the adaptive value of “stress- susceptible” subpopulations, a phenomenon perhaps similar to the viable but nonculturable (VNC) state. Further studies can be done with P. aeruginosa

180 cells being filtered onto membranes, which can be placed on PIA for certain time, then transferred onto LB medium and colony forming capability monitored. Such experiments will provide certain clues to whether exposure of P. aeruginosa cells to PIA will permanently lock them in a VNC-like state.

In summary, although this study is preliminary, and many experiments need to be repeated or supplemented, it has yielded some interesting results and opened new areas for further investigation. It appears that AHL sensing is involved - at low cell densities - in the growth, adhesion, root colonization and subpopulation composition of P. aeruginosa.

181 CHAPTER 6

GENERAL DISCUSSION

Behavioral responses of S. m eliloti to nutrient limitation

(CHAPTER 2). The motility and chemotaxis responses to nutrient deprivation of the S. meliloti strains revealed in this study are different from those previously reported. Three strains of S. m eliioti transiently upregulate chemotaxis responses but generally downregulate overall motility when facing nutrient- limiting conditions. Cells lost their motility within 8 hours to 4 days, depending on strain, after transfer to SB. The motility loss was not caused by cell death or entry into viable-but-nonculturable (VNC) state, because the viable cell counts remained unchanged for weeks. A large portion cf completely nonmotile cells still retained their flagella, indicating that turning off flagellar motors was the first response of these strains to nutrient deprivation.

Motility loss of a marine vibrio strain was observed by Stretton et al.

(1997) under starvation conditions. Majority of the cells were nonmotile after starvation for 3 d, and the authors attributed this loss cf motility to the shedding of flagella. But 3% of cells remained motile even after 16 d starvation. Their recovery experiment showed that within 1 h after the addition of nutrients, up to

182 50% of the maximum motility was restored. It is surprising that the long (16 d) starved, nonculturable (as the authors claimed) cells could recover and make flagella at that rate. It seems to me those cells were readily culturable. Our nutrient add-back results indicate that it took an hour to restore motility of flagellated cells after only hours or a day starvation.

Loss of motility and chemotactic responsiveness was also observed in a marine vibrio after exposure to starvation (Malmcrona-Friberg et ai., 1990).

However, other marine Vibrio species (Amy and Morita, 1983; Torrella and

Morita, 1981) and a Spirochaeta species (Terracciano and Canale-Parola,

1984) showed increased chemotactic responsiveness to nutrients after starvation. As shown in CHAPTER 2, the three S. meliloti strains also showed starvation responses different from one another. It appears that different species or strains adopt very different strategies in response to nutrient limitation. Marine species may have developed strategies very different from soil and rhizosphere species. This is an Important aspect of physiological and behavioral diversity in bacteria.

The effect of low nutrient on motility of S. meliloti strain JJIcIO was also studied in chemostat cultures (Robinson, 1991). It was found that, in continuous cultures with limited carbon source, almost all JJIcIO cells lost motility within 1 d, but 30% of the population regained maximum motility on day

2, then dropped again. Small, coccoid motile cells (termed swarmer cells) appeared after 48 h, comprising about 10-20% of the population. The author suggested that this kind of morphological and physiological changes might be

183 a strategy for survival. Interestingly, the loss of motility of the normal, rod­ shaped cells, and the differentiation into small motile swarmer cells were observed only in chemostat cultures with carbon or energy limitation, not in nitrogen-limiting cultures. In batch cultures with carbon sources at the concentrations similar to those added to the continuous cultures, about 80% of the JJIcIO cells were motile. Thus, continuous dilution had a marked effect on motility. Whether this was related to cell-cell interactions through accumulation of autoinducers or simply to the very low concentration of free mannitol available to cells in a chemostat merits further study. It might also be associated with cells’ physiological status and time after cell division (real age of a single cell). Studies in batch cultures of S. m eliloti and other bacteria showed that motility was growth stage or cell age dependent (Amsler et al.,

1993; Dingwall, 1990; Wei and Bauer, unpublished). One can assume that cells in a batch culture are at same stage, while cells from a continuous culture are heterogenous in terms of cell age and physiological conditions.

Although our study focused on the short term behavioral responses, while collecting data for Fig. 2.1, we did notice that cells became much smaller

(miniaturized) and coccoid in the long-starved cell suspensions in SB.

Interestingly, of the three S. meliloti strains tested, JJIcIO maintained motility for the longest time (over 3 d) after transfer to SB. At the same time we observed that JJIcIO cells became smaller than L5-30 or RMB7201. There were more coccoid cells In JJIcIO suspensions than In the other two strains.

184 The observed behavioral responses of S. melHoti to nutrient or energy deprivation can be due to energy limitation or lack of chemoattractants, or both.

Other species such as P. aeruginosa and E. coll, retained motility for over a week after transfer to SB, thus motility downregulation in S. mellloti is unlikely to be a result of simple nutrient deprivation. Recently energy taxis in £ coli was reported (Zhulin et al. 1997a; Zhulin et al. 1997b; Taylor and Zhulin 1998;

Rebbapragada et al. 1997). They have identified a sensor, Aer, that responds to changes in cells’ internal energy levels. Interestingly, the signal transduction from Aer to flagellar motors works through the same chemotaxis signal transduction system. A similar energy-sensing or metabolism-dependent system may exist in S. mellloti, especially in strain L5-30. On the other hand, the presence of nonmetabolizable attractants at low concentrations prevented

S. mellloti cells from loss of motility during starvation. This indicates that S. mellloti cells can maintain motility in presence of either attractants or energy sources. Our nutrient/attractant add-back experiments showed that energy supply was more effective in restoring motility to the starved cells.

If cells actively turn off flagellar motors after being exposed to starvation conditions, it is theoretically possible to isolate mutants that have prolonged motility by inactivating genes or mechanisms that downregulate motility. These possible mutants should operate their motility “constitutively” as long as there is enough internal energy supply. A preliminary trial with strain L5-30 by several rounds of enrichment in SB swarm agar following Tn5 mutagenesis did not yield mutants that retained motility substantially longer. Since motility is a

185 complex physiological process, downregulation or mutation in any element of the regulation chain, including energy-generating process, will cause

decrease in, or loss of, motility, it may therefore not be feasible to isolate such hypothetical incessant motility mutants.

Significance of the major findings on FS behavior (CHAPTER

3). The FS mutants or variants outgrew the wild type in dual cultures where there is a nutrient gradient (TABLE 3.5). This indicates that better motility (and chemotaxis) in environments with nutrient gradients may add to a strain's

survival. Better motility is expected to help bacteria in soil and rhizosphere environments which are tortuous in nature and generally nutrient-limited, though are much more complex than an agar matrix. At this point, we think that the derepression of flagellar synthesis in FS mutants and variants may be the

main reasons. In laboratory culture conditions, nutrient concentration is much higher than in most natural habitats, where the mutants might not have an advantage over the wild type. We have tried to compare the mutants and the wild type in terms of overall growth, survival, and nodulation in natural soil.

Owing to the numerous native soil bacteria and fungi, the experiments were not very successful.

We speculate that the 2- to 3-fold more flagellin might be the determining factor for the FS behavior. There is a report that mentioned the proportional relationship between swim speed and flagellar length per unit of cell mass (Amsler et al., 1993). They monitored motility and flagellation of E coli throughout a life cycle in shake batch cultures, which is different from our

186 study. However, their results showed that bacterial cells could regulate the

level of flagellar production, and the swim speed is proportional to the amount

of flagella. An analysis of the expression of genes that regulate flagellar biosynthesis and EPS production in the wild type and FS mutants may provide

some useful information.

FS behavior and competitiveness. Pouch nodulation tests in which a bacterial suspension was inoculated directly onto plant roots showed

no significant difference in nodule formation between FS mutants and the wild type. A more natural test in non-sterile soil is likely to show the role of FS

behavior in nodulating efficiency and survival in soil. We have done preliminary tests on the rates of spreading and for nodulation efficiencies in

natural soils. FS mutants spread faster than the wild type and showed no significant difference in survival in non-sterile soil plates (data not shown). But the nodulation test was not successful owing to the severe contamination from native soil organisms.

On the molecular mechanisms of FS behavior (CHAPTER 4).

One of the major unresolved questions arising from our studies is the causal

linkage between Tn5 insertions in the FS mutants and the FS-reduced EPS phenotype. As described in CHAPTER 4, the Tn5 is probably inserted in exactly or nearly exactly the same location in all five FS mutants. It is highly unlikely that all five FS-reduced EPS mutants, all obtained from different matings, have the same position of Tn5 insertion solely by chance. We note that the site of the Tn5 insertion in FS1 is quite unique with very high GC%

187 content This high GC content is typical for insertion sites of Tn5 (Boyd et al.,

1993; Lodge et al., 1988). However, the five Tn5 FS mutants showed certain differences in their phenotypes. Three of the Tn5 FS mutants (FS1) exhibited normal wild-type swim patterns in liquid, but another changes (FS44) swimming direction much more frequently. The fifth one (FS7) swims like the wild type but produces more EPS than the other four mutants. This raises uncertainty about exactly how the Tn5 insertion is related to the FS-EPS phenotype. One possibility is that such phenotypic variability is caused by slightly different points of insertion in the GC-rich region of the 3.25 kb sequence. Alternatively, the phenotypic variability could be due to slightly differing genetic rearrangements elsewhere in the genome caused by Tn5 insertions at exactly the same locus. The inability of the wild-type 3.25 kb sequence to complement the mutants and the failure of homologous recombination of the Tn5 back into the wild type to generate the FS phenotype seem to provide support for the latter mechanism involving an indirect relation between the Tn5 Insertion and the FS-EPS phenotype.

Several specific kinds of genetic rearrangement might be triggered by an insertion of Tn5 in the GC-rich region of the 3.25 kb sequence. One possibility is the induced transposition of one of the S. m eliloti native IS elements, leading to deletion, insertion or genomic rearrangement in a gene crucial to FS behavior. More than ten native IS elements have been identified in S. meliloti (e. g. Zekri et al., 1998). The function of these native IS elements is unknown, but Dusha et al. (1987) found that S. meliloti insertion element

188 \SRm2 was closely associated with the fixX gene, insertion and excision of native IS elements or transposons have been Implicated In changes In EPS synthesis In a number of species, Including Pseudomonas atlantica and P. aeruginosa (Bartlett and Silverman, 1989, Sokol et al., (1994). It Is perhaps equally possible that Tn5 Insertions In the GC-rlch region trigger phase variation(s) at some other loci. Phase variation appears to be a common mechanism for regulation of certain significant cellular activities. Including flagellar and fimbriae type, EPS production or mucoldlcity, and even root colonization ability. Phase variation has been shown to regulate pill of £ coli, which was also controlled by an invertible DNA element (Abraham et al., 1985) together with a site-specific recomblnase (Dorman and Higgins, 1987).

Studies by Dekkers et al. (1998) Identified a site-speclfic recomblnase involved

In root colonization by Pseudomonas fiuorescens.

Phenotype conversion. Studies done by Brumbley and Denny

(1990) showed that In R. solanacearum. mutation In a single gene, phcA, could affect a number of pathogenlclty-related traits, including reduced EPS production and acquisition of motility. Later, they discovered that spontaneous

Insertion of sequences of various lengths Into phcA was the cause of phenotype conversion (Brumbley et al. 1993). Clough et al. (1997) reported that the expression of phcA Is regulated by products of the putative phcBSR(Q) operon. phcB Is essential for the production of a volatile extracellular signal compound, 3-hydroxypalmltlc acid methyl ester (3-0H PAME). The phcSR(Q) genes encode proteins similar to two-component signal transduction systems.

189 Expression of phcA and production of the PhcA-regulated factors such as EPS

I were greatly reduced in phcB mutants. Addition of 3-0H PAME to growing cultures of phcB mutants restored to wild-type phenotype. The authors suggested that 3-OH PAME regulated phcA expression, and hence virulence factor production, in a way analogous to autoinduction, which was later confirmed by Flavier et al. (1997 ). 3-OH PAME is volatile, and it can facilitate long-distance cell-cell communication. The regulatory network for the system is very complicated, an alternative sigma factor (sigma S) is involved (Flavier et al. 1998), as well as a two-component system (the vsrAB system) (Schell et al., 1994; Clough et ai., 1997)

A similar regulation mechanism was identified by Harris et al. (1998) in the regulation of virulence gene expression in Erwinia carotovora subsp.carotovora (Ecc) by HexA. HexA represses motility and other virulence factor production. Interestingly, the hexA system is analogous to the phcA system in that quorum-sensing is part of the regulation chain. The quorum- sensing pheromone was A/-(3-oxohexanoyl)-L-homoserine lactone (OHHL).

Production of OHHL in Ecc hexA mutant cells was higher than in the wild-type cells.

We speculate that a similar mechanism exists in S. meliloti RMB7201.

Our results point to a model of regulation by reversible insertion and deletion, or inversion of a genetic element, although we have not been able to identify the genetic element. However, in case of phenotype conversion in R solanacearum, no reversion has been observed from PC variants. The PC

190 progeny seemed to have irreversibly differentiated into a subpopulation. Thus the mechanism for our FS-WT switching may be somewhat different from that for phenotype conversion. The mechanism causing PC was insertion into the phcA gene. Theoretically, precise excision of the insertions should lead to reversion. It is also possible that they did not screen enough isolates for revenants.

Quorum-sensing is also involved in the regulation of EPS production in a few species. In Pantoea stewartii subsp. stewartii, genes esal and esaR

(homologs of luxi and luxR) code for the AHL signal synthase and the gene regulator respectively (von Bodman et al. 1998). EsaR repressed EPS synthesis at low cell density, and mutation in esaR gene caused constitutive and elevated EPS production at low cell densities, suggesting that EPS genes are expressed only when cell density becomes high in P. stewartii subsp. stewartii. Studies done in Ralstonia solanacearum clearly link autoinducers to both motility and EPS production (Clough et al. 1997; Flavier et al. 1997). We have not found any evidence of involvement of HSL signalling in FS-EPS behavior in S. meliloti RMB7201. Although stationary-phase culture filtrates

(CF) of RMB7201 inhibited the growth of both wild type and FS derivatives, CF did not affect either swarming rates or EPS production of the FS mutants

(results not shown). But we cannot rule out possible involvement of a signal compound in the FS-EPS phenomenon before a thorough search is done.

Inactivation of a repressor as possible mechanism for FS behavior. Perhaps the simplest yet reasonable explanation for the increased

191 swarming behavior In the FS mutants and variants Is a derepression of flagellar biosynthesis and motility. As shown In FIG 3.2, the rate of swarming of the wild type Increased as the concentration of nutrients In the swarm agar decreased. This Is consistent with the concept that S. meliloti represses Its swarming activity when nutrients are plentiful and derepresses It when nutrients are scarce. Since Tn5 Insertions In the FS mutants caused an

Increase In swarming activity. It seems most likely that the Tn5 Insertions somehow caused the Inactivation of a repressor, perhaps a repressor gene which negatively regulates motility (flagellin synthesis) but positively controls

EPS production.

Both positive and negative regulations have been shown involved In motility and EPS synthesis. Schmitt et al. (1996) characterized a negative regulator gene fIgM of flagellar synthesis In S. typhimurium. FIIM protein

Inhibits FIIA, the flagellum-speclfic sigma factor, thus reduces motility and virulence. Similar negative control over flagellar synthesis was demonstrated in Bordetella bronchiseptica (Akerley and Miller, 1993), the BvgAS virulence control system negatively regulated the flagellin gene transcription, while positively regulated other virulence-related genes. Negative regulation Is

Implicated In many other Important processes, Including EPS biosynthesis and

S. meliloti nodulation (KondorosI et al., 1991). For Instance, mutations In the exoR and exoS gene caused EPS overproduction by S. meliloti (Doherty et al.,

1988; Reuber et al., 1991).

192 Spontaneous switching mechanisms. Another important finding from our studies is that the wild type can switch to FS behavior spontaneously, and that FS variants spontaneously revert back to the wild-type configuration.

We believe that this kind of regulation, with the creation of a small, extra motile subpopulation, may occur widely among motile bacteria. As noted in

CHAPTER 3, we have succeeded in isolating FS variants of S. meliloti

Rm1021, E. coli and P. aeruginosa by simple enrichment for more motile progeny cells in swarm agar, suggesting that the spontaneous formation of FS subpopulations is common across different genera.

One of the central questions that remains to be answered from our studies is how the formation and reversion of spontaneous FS variants occurs at the molecular level. At this point, it seems that the spontaneous variants have no genetic rearrangement in the GC-rich sequence of the 3.25 kb fragment where the Tn5 is inserted. However, the same clones and subclones from the wild-type genomic library complement or suppress both the Tn5 FS mutants and the spontaneous FS variants in respect to both the reduced EPS phenotype and the swarming phenotype, indicating a similar change is responsible for the observe phenotype (CHAPTER 4, TABLE 4.5).

Current and future studies on molecular mechanism of FS behavior. We have isolated wild-type genomic clones that complement or suppress FS-EPS phenotype. Sequence analysis is in progress. Once the whole sequence information is available, a genebank database search may identify homologous sequences, if any, and the deduced functions. At present,

193 partial sequence data show that the complementing clones have homology to histidine protein kinase, two-component system, and EPS regulatory genes.

PCR analysis of the region will be carried out to compare the DNA patterns of the wild type with those of the FS mutants, the spontaneous FS variants, and the revertants of the FS variants. Insertions, deletions or DNA rearrangements longer than 50 bp will be detected.

Another approach to study the possible switching mechanism is to do subtraction Northern blot to identify any difference at the messenger level between the wild type and FS mutants and variants. By screening cDNA libraries of RMB7201 and FS mutants and spontaneous variants, clones containing sequences unique to either the wild type or mutants can be identified and characterized. A third future study is to look for changes in genes directly related to flagellin and EPS gene expression. For example, gene fIgM codes for a negative regulator (anti-sigma factor) for expression of the flhD master operon of S. typhimurium, FIgM inhibits FliA, an alternative sigma factor specific for late flagellar opérons. Inactivation of fIgM will probably increase flagellin synthesis. We may look for fIgM homolog in S. meliloti RMB7201, and for any change in FS mutants and variants. It may also be potentially valuable to investigate into any possible difference in lysR homolog and its expression between the RMB7201 wild type, FS1 and EV10, since the published studies relevant to our investigation point to a possible role of a LysR family regulator in FS-EPS behavior. Finally, since reversion of the spontaneous FS variants back to the wild-type behavior is inconsistent, indicating that unidentified

194 conditions or factors are involved in regulating the switching and reversion frequencies. A comprehensive study is needed on potential external

conditions (if any) favoring switching and reversion. As a first step, we can test some HSL signalling compounds for possible effect on switch.

Relation of HSL signalling to starvation and stress resistance

(CHAPTER 5). HSL signalling is now known to be involved in bacterial

starvation and stress responses. Huisman and Kolter (1994) Identified an AHL- dependent starvation signaling pathway in E. coli. Upon nutrient depletion,

many bacteria become more resistant to stress factors. In E. coll, the a® subunit mediates this adaptation process. It was found that synthesis of cP was HSL- dependent. o® then directs the synthesis of many starvation-related proteins.

Similarly, Srinivasan et al. (1998) discovered that stationary-phase culture supernatant induced the synthesis of starvation-specific proteins in Vibrio sp. log-phase cells. The addition of furanones, antagonists of AHLs, Inhibited the synthesis of starvation proteins. Low concentrations of furanone greatly reduced the culturability and stress resistance of the strain during starvation.

They concluded that extracellular signalling molecules (probably HSLs) play a significant role in starvation responses and in the development of bacterial resistance to environmental stresses. These reports are consistent with our finding (CHAPTER 5) that P. aeruginosa PAO-JP2, which is defective in rhil and Iasi genes, formed many fewer colonies on PIA medium, indicating a reduced resistance to stress factors (as measured by their culturability in the

195 presence of the antibiotic irgasan).

Bacterial heterogeneity at subpopulation levels. Finally, it is interesting that bacterial responses to environmental signals can be very much heterogenous at a subpopulation level. A culture of S. meliloti, when facing nutrient limitation, can form several subpopulations in terms of motility and flagellation (CHAPTER 2). In case of FS-EPS behavior, it is only a few cells in a population that genetically switch between normal and FS-EPS phenotypes

(CHAPTERS 3 & 4). In a seemingly uniform culture of P. aeruginosa, only a fraction of cells could escape inhibition and form colonies on PIA (CHAPTER

5). In other words, the culturability of P. aeruginosa is heterogenous among the cells in a culture. This heterogeneity was observed in responses to a variety of culture conditions such as salt concentration, moisture content and antibiotic levels. Heterogeneity is probably a common mechanism for single-celled organisms to adapt to fluctuations in environmental conditions so that a population or a species can survive.

196 CONCLUDING REMARKS

This project Included the first studies on the motility and chemotactic responses of S. m eliloti after transfer from nutrient-rich media to starvation conditions. We found that the first behavioral responses of S. meliloti to nutrient limitations were to transiently enhance its chemotactic responsiveness and, at the same time, turn off flagellar motors to conserve energy. Such responses would apparently be beneficial to the survival of a bacterial population facing nutrient scarcity. The responses of S. meliloti differed from those of E. coli and

P. aeruginosa, which could retain motility over a week after transfer to buffer without any available nutrient. As a soil-rhizosphere species, S. meliloti cells in natural environments will frequently encounter fluctuations in nutrient supply and other growth conditions. The ability to quickly respond to new nutrient sources while to survive long starvation is crucial to the competitiveness of the species. Our study opened an new area for further biochemical and molecular studies on how S. m eliloti actively regulates its motility and chemotaxis in response to nutrient availability, and on the ecological significance of such regulation.

Although increased swarming mutants of S. meliloti had been isolated in two other laboratories, we were the first to systematically investigate the FS

197 behavior and Its effect on growth and competitiveness. The discovery of spontaneous genetic switching between normal and FS behavior is a new addition to the previously known bacterial mechanisms for regulating behavioral activities. It is increasingly clear that the genetic changes involving phase variation, phenotype conversion, recombinatorial DNA rearrangement, insertions and deletions, or transposition of native IS elements and transposons are important in creating subpopulations that play a significant role in bacterial adaptation to, and survival of, adverse conditions.

The molecular mechanisms for the FS-EPS behavior and FS-WT switching remain unsolved. Ultimately, it needs successful cloning and characterization of the genes directly responsible for the FS-EPS behavior and switching, and the elucidation of their regulatory schemes in relation to other genes against diverse environmental backgrounds. It will also be interesting to see if such a switch of motility behavior is common among motile bacteria.

Since switch exists in species like E. coli and P. aeruginosa, molecular characterization of its mechanism may be easier using these best-studied model species. Regulation of behaviors such as motility and EPS production are very sophisticated, generally involving interactions between various regulatory networks at multiple levels. It will be challenging to unravel the underlying mechanisms for the enhanced swarming behavior and switching.

But it will certainly lead to a better understanding of how bacteria respond to the ever-changing environments.

198 LIST OF REFERENCES

Abraham, J. M., 0. S. Freitag, J. R. Clements, and B. I. Elsensteln. 1985. An invertible element of DNA controls phase variation of type 1 fimbriae of Escherichia coli. Proc. Natl. Acad. Sci. USA 82:5724-5727.

Adler, J. 1973. A method for measuring chemotaxis and use of the method to determine optimum conditions for chemotaxis by Escherichia coli. J. Gen. Microbiol. 74:77-91.

Alder, J., and M. M. Dahl. 1967. A method for measuring the motility of bacteria and for comparing random and non-random motility. J. Gen. Microbiol. 46:161-173.

Adler, J., G. L. Hazelbauer, and M. M. Dahl. 1973. Chemotaxis toward sugars in Escherichia coli. J. Bacteriol. 1 1 5 :824-847.

Adler, J., and B. Templeton. 1967. The effect of environmental conditions on the motility of Escherichia coli. J. Gen. Microbiol. 46:175-184.

Aguilar, J. M. M., A. M. Ashby, A. J. M. Richards, G. J. Loake, M. D. Watson, and C. H. Shaw. 1988. Chemotaxis of R. ieguminosarum biovar phaseoli towards flavonoid inducers of the symbiotic modulation genes. J. Gen. Microbiol. 134:2741-2746.

Akerley, B. J., and J. F. Miller. 1993. Flagellin gene transcription in Bordetella bronchiseptica is regulated by the BvgAS virulence control system. J. Bacteriol. 1 75:3468-3479.

Ames, P., And K. Bergman. 1981. Competitive advantage provided by bacterial motility in formation of nodules by Rhizobium meliloti. J. Bacteriol. 148:728-729.

Ames, P., S. A. Schluederberg, and K. Bergman. 1980. Behavioral mutants of Rhizobium meliloti. J. Bacteriol. 141:722-727.

199 Amsler, C. D., M. Cho, and P. Matsumura. 1993. Multiple factors underlying the maximum motility of Escherichia coli as cultures enter post­ exponential growth. J. Bacteriol. 175:6238-6244.

Amy, P. S., and R. Y. Merita. 1983. Starvation- survival patterns of sixteen freshly Isolated open ocean bacteria. Appl. Environ. Microbiol. 45:1109-1115.

Anderson, I., and A. Brass. 1998. Searching DNA databases for similarities to DNA sequences: when is a match significant? Bioinformatics 14:349-356.

Armengaud, J., C. Meyer, and Y. Jouanneau. 1994. Recombinant expression of the fdxD gene of Rhodobacter capsulatus and characterization of its product, a [2Fe-2S] ferredoxin. Biochem. J. 300:413-418.

Armitage, J. P. 1992. Behavioral responses in bacteria. Annu. Rev. Physiol. 54:683-714.

Armitage, J. P., R. Schmitt. 1997. Bacterial chemotaxis: Rhodobacter sphaeroides and Sinorhizobium me/z/of/'-variations on a theme? Microbiology 143:3671-3682.

Auerswald, E. A., G. Ludwig, and H. Schaller. 1981. Structural analysis of Tn5. Cold Spring Harb. Symp. Quant .Biol. 45 Pt 1:107-113.

Ausubel, F. M., R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidman, J. A. Smith, and K. Struhl (Eds.). 1989. Current Protocols in Molecular Biology, Greene Publishing Associates and Wiley-lnterscience, New York.

Austin, B., and D. A. Austin. 1990. Expression of motility in strains of the non-motile species Aeromonas media. FEMS Microbiol. Lett. 68:123-124.

Barbour W. M., D. R. Hattermann, and G. Stacey. 1991. Chemotaxis of Bradyrhizobium japonicum to soybean exudates. Appl. Environ. Microbiol. 57:2635-2639.

Bartlett, D, H., and M. Silverman. 1989. Nucleotide sequence of 13492, a novel insertion sequence causing variation in extracellular polysaccharide production in the marine bacterium Pseudomonas atlantica. J. Bacteriol. 171:1763-1766.

Bazylinski, D. A., A. J. Garratt-Reed, and R. B. Frankel. 1994. Electron microscopic studies of magnetosomes in magnetotactic bacteria.

200 Microsc. Res. Tech. 27:389-401.

Becker, A., S. Ruberg, H. Kuster, A. A. Roxlau, M. Keller, T. Ivashina, H. P. Cheng, G. 0. Walker, and A. Puhler. 1997. The 32- kilobase exp gene cluster of Rhizobium meliloti directing the biosynthesis of galactoglucan: genetic organization and properties of the encoded gene products. J. Bacteriol. 179:1375-1384.

Benson, F. E., G. T. Illing, G. J. Sharpies, and R. G. Lloyd. 1988. Nucleotide sequencing of the ruv region of Escherichia coli K-12 reveals a LexA regulated operon encoding two genes. Nucleic Acids Res. 16:1541- 1549.

Berg D. E., M. A. Schmandt, and J. B. Lowe. 1983. Specificity of transposon Tn5 insertion. Genetics 105:813-828.

Bergman, K., M. Gulash-Hoffee, R. E. Hovestadt, R. 0. Larosiliere, P. G. Ronco II and L. Su. 1988. Physiology of behavioral mutants of Rhizobium meliloti: evidence for a dual chemotaxis pathway. J. Bacteriol. 170:3249-3254.

Bergman, K., E. Nulty, and L. Su. 1991. Mutations in the two flagellin genes of Rhizobium meliloti. J. Bacteriol. 173:3716-3723.

Beringer, J. E. 1974. R-factor transfer in Rhizobium Ieguminosarum. J. Gen. Microbiol. 84:188-198.

Bhagwat, A. A., R. E. Tully, and D. L. Keister. 1992. Isolation and characterization of an ndvB locus from Rhizobium fredii. Mol. Microbiol. 6:2159-2165.

Blair, D. F. 1995. How bacteria sense and swim. Annu. Rev. Microbiol. 49:489-522.

Blakemore, R. P. 1982. Magnetotactic bacteria. Annu. Rev. Microbiol. 36:217-238.

Block, S. M., D. F. Blair, and H. C. Berg. 1989. Compliance of bacterial flagella measured with optical tweezers. Nature (London) 338:514-518.

Bloemberg, G. V., G. A. O'Toole, B. J. Lugtenberg, and R. Kolter. 1997. Green fluorescent protein as a marker for Pseudomonas spp. Appl. Environ. Microbiol. 63:4543-4551.

201 Bosworth, A. H., M. K. Williams, K. A. Albrecht, T. R. Hankinson, R. Kwiatkowski, J. Beynon, C.W. Ronson, F. Cannon, T. J. Wacek, and E. W. Triplett. 1994. Alfalfa yield response to inoculation with recombinant strains of Rhizobium meliloti carrying an extra copy of dctA and/or modified nifA expression. Appl. Environ. Microbiol. 6 0:3815-3832.

Boyd, A., K. Kendall, and M, I. Sim on. 1983. Structure of the serine chemoreceptor in Escherichia coli. Nature (London) 301:623-626.

Boyer, H. W., and D Roulland-Dussoix. 1969. A complementation analysis of the restriction and modification of DNA in Escherichia coli. J. Mol. Biol. 4 1 :459-472.

Bradley, D. E. 1980. A function of Pseudomonas aeruginosa PAO polar pili: twitching motility. Can. J. Microbiol. 26:146-154.

Broek, A. V., M. Lambrecht, and J. Vanderleyden. 1998. Bacterial chemotactic motility is important for the initiation of wheat root colonization by Azospirillum brasilense. Microbiology 144:2599-2606.

Brooke, J. S., and M. A. Valvano. 1996. Molecular cloning of the Haemophilus influenzae gmhA (IpcA) gene encoding a phosphoheptose isomerase required for lipooligosaccharide biosynthesis. J. Bacteriol. 178:3339-3341.

Brumbley S. M., and T. P. Denny. 1990. Cloning of wild type Pseudomonas solanacearum phcA, a gene that when mutated alters expression of multiple traits that contribute to virulence. J. Bacteriol. 172: 5677-5685.

Brumbley S. M., B. F. Carney, and T. P. Denny. 1993. Phenotype conversion in Pseudomonas solanacearum due to spontaneous inactivation of PhcA, a putative LysR transcriptional regulator. J. Bacteriol. 175: 5477-5487.

Burchard, R. B. 1981. Gliding motility of prokaryotes: Ultrastructure, physiology and genetics. Annu. Rev. Microbiol. 35:497-529.

Burton, G. J., G. B., Hecht, and A. Newton. 1997. Roles of the histidine protein kinase pleC in Caulobacter crescentus motility and chemotaxis. J. Bacteriol. 179:5849-5853.

Caetano-Anollés, G., L. G. Wall, A T. De Michel!, E. M. Macchi, W. D. Bauer, and G. Favelukes. 1988. Role of motility and chemotaxis in efficiency of nodulation by Rhizobium meliloti. Plant physiol. 86:1228-1235.

202 Caetano-Anollés, G., D. K. Crist-Estes, and W. D. Bauer. 1988. Chemotaxis of Rhizobium meliloti to plant fiavone luteolln requires functional nodulation genes. J. Bacteriol. 170:3164-3169.

Caetano-Anolles, G., L. G. Wall, A.T. De Micheli, E. M. Macchi, W. D Bauer, and G. Favelukes. 1988. Role of motility and chemotaxis in efficiency of nodulation by Rhizobium meliloti. Plant Physiol. 86:1228-1235.

Caetano-Anollés, G., and P. M. Gresshoff. 1994. DNA amplification fingerprinting using arbitrary mini-hairpin oligonucleotide primers. Bio/Technology 1 2:619-623.

Canale-Parola, E. 1984. The spirochetes. Order I. Spirochaetales Buchanan 1917. p.38-70. In N. R. Krieg (ed.), Bergey’s manual of systematic bacteriology. Vol. I. Williams and Willins, Baltimore.

Castillo, A., and S. Reverchon. 1997. Characterization of the pecTcontrol region from Erwinia chrysanthemi3937. J. Bacteriol. 179:4909-4918.

Castillo, A., W. Nasser, G. Condemine, and S. Reverchon. 1998. The PecT repressor interacts with regulatory regions of pectate lyase genes in Erwinia chrysanthemi. Biochim. Biophys. Acta 1442:148-160.

Cha, C., P. Gao, Y. C. Chen, P. D. Shaw, and S. K. Farrand. 1998. Production of acyl-homoserine lactone quorum-sensing signals by gram- negative plant-associated bacteria. Mol. Plant Microbe Interact. 1 1 :1119-1129.

Chang S. F., D. Ng, L. Baird, and C. Georgopoulos. 1991. Analysis of an Escherichia coli dnaB temperature-sensitive insertion mutation and its cold- sensitive extragenic suppressor. J. Biol. Chem. 266:3654-3660.

Chen, C. M., Q. Z. Ye, Z. M. Zhu, B. L. Wanner, and C. T. Walsh. 1990. Molecular biology of carbon-phosphorus bond cleavage. Cloning and sequencing of the phn (psiD) genes involved in alkylphosphonate uptake and C-P lyase activity in Escherichia coliB. J. Biol. Chem. 265:4461-4471.

Clough, S. J., K. E, Lee, M. A. Schell, and T. P. Denny. 1997. A two- component system in Ralstonia (Pseudomonas) solanacearum modulates production of PhcA-regulated virulence factors in response to 3- hydroxypalmitic acid methyl ester. J. Bacteriol. 179:3639-3648.

Condemine, G., A.Castillo, F. Passed, and C. Enard. 1999. The PecT repressor coregulates synthesis of exopolysaccharides and virulence factors in Erwinia chrysanthemi. Mol. Plant. Microbe Interact. 12:45-52.

203 Corbel!, N., and J. E. Loper. 1995. A global regulator of secondary metabolite production in Pseudomonas fiuorescens Pf-5. J. Bacteriol. 177:6230-6236.

Davies, D. G., M. R. Parsek, J. P. Pearson, B. H. Iglewski, J. W. Costerton, and E. P. Greenberg. 1998. The involvement of cell-to-cell signals in the development of a bacterial biofilm. Science 280:295-298. de Bruijn, F. J., and J. R. Lupski. 1984. The use of transposon Tn5 mutagenesis in the rapid generation of correlated physical and genetic maps of DNA segments cloned Into multicopy plasmids-a review. Gene 27:131-149. de Lorenzo V., M. Herrero, U. Jakubzik, and K. P. Timmfs. 1990. Mini-Tn5 transposon derivatives for insertion mutagenesis, promoter probing, and chromosomal insertion of cloned DNA in Gram-negative eubacteria. J. Bacteriol. 172:6568-6572.

De Weger, L. A., 0. I. van der Vlugt, A. H. Wijfjes, P. A. Bakker, B. Schippers, and B. Lugtenberg. 1987. Flagella of a plant-growth- stimulating Pseudomonas fiuorescens strain are required for colonization of potato roots. J. Bacteriol. 169:2769-2773.

DeFlaun, M. F., A. S. Tanzer, A. L. McAteer, B. Marshall, and S. B. Levy. 1990. Development of an adhesion assay and characterization of an adhesion-deficient mutant of Pseudomonas fiuorescens. Appl. Environ. Microbiol. 56:112-119.

Dekkers, L. 0., 0. C. Phoelich, L. van der Fits, and B. J. Lugtenberg. 1998. A site-specific recomblnase is required fo r competitive root colonization by Pseudomonas fiuorescens WCS365. Proc. Natl. Acad. Sci. USA 95:7051-7056.

DePamphllis, M. L., and J. Adler. 1971. Fine structure and isolation of the hook-basal body complex of flagella from Escherichia coli and Baciiius subtiils. J. Bacteriol. 105:384-395.

Dharmatilake, A. J., and W. D. Bauer. 1992. Chemotaxis of Rhizobium meiiioti towards nodulation gene-inducing compounds from alfalfa roots. Appl. Environ. Microbiol. 5 8 :1153-1158.

Diker K. S., G. Hascelik, and M. Akan. 1992. Reversible expression of flagella in Campylobacter spp. FEMS Microbiol. Lett. 78:261-264.

204 Dingwall, A., L. Shapiro, and B. Ely. 1990. Flagellar biosynthesis in Caulobacter, p. 155-176. /n J. P. Armitage and J. M. Lackie (ed.), Biology of the Chemotactic Response. Soc. Gen. Microbiol. Symp. 46th. Cambridge University Press , Cambridge.

Ditta, G., S., Stanfield, D. Corbin, and D. R. Helinski. 1980. Broad host range cloning system for gram-negative bacteria; construction of a gene bank of Rhizobium mellloti. Proc. Natl. Acad. Sci. USA 27:7347-7351.

Doherty, D., J. A. Leigh, J. Glazebrook, and G. 0. Walker. 1988. Rhizobium meliloti mutants that overproduce the Rhizobium meliloti acidic Calcofluor-binding exopolysaccharide. J. Bacteriol. 170:4249-4256.

Dorman, C. J., and 0. F. Higgins. 1987. Fimbrial phase variation in Escherichia coli: dependence on integration host factor and homologies with other site-specific recombinases. J. Bacteriol. 169:3840-3843.

Dufour A., R. B. Furness, and C. Hughes. 1998. Novel genes that upregulate the Proteus mirabilis flhDC master operon controlling flagellar biogenesis and swarming. Mol. Microbiol. 29:741-751.

Dusha, I., S. Kovalenko, Z. Banfaivi, and A. Kondorosi. 1987. Rhizobium meliloti insertion element lSRm2 and its use for identification of the f/xXgene. J. Bacteriol. 169:1403-1409.

Eberl, L., G. Christiansen, S. Molin, and M. Givskov. 1996. Differentiation of Serratia liquefaciens into swarm cells is controlled by the expression of the flhD master operon. J. Bacteriol. 178:554-559.

Eberl, L., M. K. Winson, 0. Sternberg, G. S. Stewart, G. Christiansen, S. R. Chabra, B. Wycroft, P. Williams, S. Molin, and M. Givskov. 1996. Involvement of N-acyl-L-homoserine lactone autoinducers in controlling the multicellular behavior of Serratia liquefaciens. Mol. Microbiol. 20:127-136.

Egan, S. M, and R. F. Schleif. 1993. A regulatory cascade in the induction of rhaBAD. J. Mol. Biol. 234:87-98.

Eisenbach, M. 1991. Signal transduction in bacterial chemotaxis. p. 137- 208. In J. L. Spudich and B. H. Satir (ed.), Sensory receptors and signal transduction. John Wiley & Sons Inc., New York.

Eisenbach, M. 1996. Control of bacterial chemotaxis. Mol. Microbiol. 20:903-910.

205 FInan, T. M., B. Kunkei, G. F. De Vos, and E. R. Signer. 1986. Second symbiotic megaplasmid in Rhizobium meliloti carrying exopolysaccharide and thiamine synthesis genes. J. Bacteriol. 167:66-72.

Flavier, A. B, S. J. Clough, M. A. Schell, and T. P. Denny. 1997. Identification of 3-hydroxypalmitic acid methyl ester as a novel autoregulator controlling virulence in Ralstonia solanacearum. Mol. Microbiol. 26:251-259.

Flavier, A. B., L. M. Ganova-Raeva, M. A. Schell, and T. P. Denny. 1997. Hierarchical autoinduction in Ralstonia solanacearum: control of acyl- homoserine lactone production by a novel autoregulatory system responsive to 3-hydroxypalmitic acid methyl ester. J. Bacteriol. 1 79:7089-7097.

Flavier, A. B, M. A. Schell, and T. P. Denny. 1998. An RpoS (sigma 8) homologue regulates acylhomoserine lactone-dependent autoinduction in Ralstonia solanacearum. Mol. Microbiol. 28:475-486.

Fonseca, K., J. MacDougali, and T. L. Pitt. 1986. Inhibition of Pseudomonas aeruginosa from cystic fibrosis by selective media. J. Clin. Pathol. 39:220-222.

Frederick, R. D., J. Chiu, J. L. Bennetzen, and A. K. Handa. 1997. Identification of a pathogenicity locus, rpfA, in Erwinia carotovora subsp. carotovora that encodes a two-component sensor-regulator protein. Mol. Plant Microbe Interact. 10:407-415.

Freter, R., and P. 0 . O'Brien. 1981. Role of chemotaxis in the association of motile bacteria with intestinal mucosa: fitness and virulence of nonchemotactic V/br/o cho/erae mutants in infant mice. Infect. Immun. 34:222- 233.

Freter R., P. 0. O'Brien, and S. A. Halstead. 1978. Adhesion and chemotaxis as determinants of bacterial association with mucosal surfaces. Adv. Exp. Med. Biol. 107:429-437.

Freter, R., P. C. O'Brien, and M. S. Macsai. 1981. Role of chemotaxis in the association of motile bacteria with intestinal mucosa: in vivo studies. Infect. Immun. 34:234-240.

G am bello M. J., S. Kaye, and B. H. Ig lew ski. 1993. LasR of Pseudomonas aeruginosa is a transcriptional activator of the alkaline protease gene {apt) and an enhancer of exotoxin A expression. Infect. Immun. 6 1 :1180- 1184.

206 Gardel, C. L. And J. J. Mekaianos. 1996. Altérations in Vibrio cholerae motility phenotypes correlate with changes in virulence factor expression. Infect Immun. 64:2246-2255.

Garfinkel, D. J., R. B. Simpson, L. W. Ream, F. F. W hite, M. P. Gordon , and E. W. Nester. 1981. Genetic analysis of crown gall: fine structure map of the T-DNA by site-directed mutagenesis. Cell 27:143-153.

Geesey, G. G., and R. Y. Morita. 1979. Capture of arginine at low substrate concentrations by a marine psychrophilic bacterium. Appl. Environ. Microbiol. 38:1092-1097. Gegner, J. A., D. R. Graham, A. F. Roth, and F. W. Dahlquist. 1992. Assembly of an MCP receptor, CheW, and kinase CheA complex in the bacterial signal transduction pathway. Cell 70:975-982.

Genevaux P., S. Muller, and P. Bauda. 1996. A rapid screening procedure to identify mini-Tn 70 insertion mutants of Escherichia coii K-12 with altered adhesion properties. FEMS Microbiol. Lett. 1 4 2 :27-30.

Gerl, L., and M. Sumper. 1988. Halobaterial flagellins are encoded by a multigene family: characterization of five flagellin genes. J. Biol. Chem. 263:13246-13251.

Givaudan A., A. Lanois, and N. Boemare. 1996. Cloning and nucleotide sequence of a flagellin encoding genetic locus from Xenorhabdus nematophilus: phase variation leads to differential transcription of two flagellar genes (fliCD). Gene 183:243-253.

Givskov, M., L. Ererl, S. Moller, L. K. Poulsen, and S. Molin. 1994. Responses to nutrient starvation in Pseudomonas putida KT2442: analysis of general cross-protection, cell shape, and macromolecular content. J. Bacteriol. 176:7-14.

Glessner, A., R. S. Smith, B. H. iglewski, and J. B. Robinson. 1999. Roles of Pseudomonas aeruginosa las and rhi quorum-sensing systems in control of twitching motility. J. Bacteriol. 18 1:1623-1629.

Godchaux, W., M. A. Lynes and E. R. Leadbetter. 1991. Defects in gliding motility in mutants of Cytophaga johnsonae lacking a high-molecular- weight cell surface polysaccharide. J. Bacteriol. 173:7607-7614.

Goldberg, J. B., and D. E. Ohman. 1987. Construction and characterization of Pseudomonas aeruginosa algB mutants: role of aigB in high-level production of alginate. J. Bacteriol. 169:1593-1602.

207 Gôtz, R., N. Limmer, K. Ober, and R. Schmitt. 1982. Motility and chemotaxis in two strains of Rhizobium with complex flagella. J. Gen. Microbiol. 128:789-798.

Gôtz, R., and R. Schmitt. 1987. Swims by unidirectional, intermittent rotation of right-handed flagellar helices. J. Bacteriol. 169:3146-3150.

Gray, K. M., J. P. Pearson, J. A. Downie, B. E. Boboye, and E. P. Greenberg. 1996. Cell-to-cell signaling in the symbiotic nitrogen-fixing bacterium Rhizobium Ieguminosarum: autoinduction of a stationary phase and rhizosphere-expressed genes. J. Bacteriol. 178:372-376.

Greek M., J. Platzer, V. Sourjik, and R. Schmitt. 1995. Analysis of a chemotaxis operon in Rhizobium meliloti. Mol. Microbiol. 15:989-1000.

Green, S. K., M. N. Schroth, J. J. Cho, S. K. Kominos, and V. B. Vitanza-jack. 1974. Agricultural plants and soil as a reservoir for Pseudomonas aeruginosa. Appl. Microbiol. 28:987-991.

Grobe, S., J. Wingender, and H. G. Truper. 1995. Characterization of mucoid Pseudomonas aeruginosa strains isolated from technical water systems. J. Appl. Bacteriol. 79:94-102.

Gulash, M., P. Ames, R. C. Larosiliere, and K. Bergman. 1984. Rhizobia are attracted to localized sites on legume roots. Appl. Environ. Microbiol. 48:149-152.

Haefele, D. M., and S. E. Lindow. 1987. Flagellar motility confers epiphytic fitness advantages upon Pseudomonas syringae. Appl. Environ. Microbiol. 53:2528-2533.

Hanahan, D. 1983. Studies on transformation of Escherichia coli with plasmids. J. Mol. Biol 1 66:557-580.

Harris, S. J., Y. L. Shih, S. D. Bentley, and G. P. Salmond. 1998. The hexA gene of Erwinia carotovora encodes a LysR homologue and regulates motility and the expression of multiple virulence determinants. Mol. Microbiol. 28:705-717.

Hawes M. C., and L. Y. Smith. 1989. Requirement for chemotaxis in pathogenicity of Agrobacterium tumefaciens on roots of soil-grown pea plants. J. Bacteriol. 1 71:5668-5671.

208 Henrichsen, J. 1983. Twitching motility. Annu. Rev. Microbiol. 37:81-93. Herrero, M., V. De Lorenzo, and L. N. Timmis. 1990. Transposon vectors containing non-antibiotic resistance selection markers for cloning and stable chromosomal insertion of foreign genes in gram-negative bacteria. J. Bacteriol. 172:6557-6567.

Hoagland, D. P., and D, I. Arnon. 1950. The water culture method for growing plants without soil. Univ. of California Agr. Expt. Sta. Giro. 347 (revised ed.).

Holloway, B. W., V. Krishnaplllai, and A. F. Morgan. 1979. Chromosomal genetics of Pseudomonas. Microbiol. Rev. 43:73-102.

Hrabak, E. M., and D. K. Willis. 1992. The lemA gene required for pathogenicity of Pseudomonas syringae pv. syringae on bean is a member of a family of two-component regulators. J. Bacteriol. 1 74:3011-3020.

Huisman, G. W., and R. Kolter. 1994. Sensing starvation: a homoserine lactone-dependent signaling pathway in Escherichia coli. Science 265:537- 539.

Hussain, N. H., M. Goodson, and R. J. Rowbury. 1998. Recent advances in biology: intercellular communication and quorumsensing in micro­ organisms. Sci. Prog. 8 1 :69-80.

Inada, T., and Y. Nakamura. 1996. Autogenous control of the suhB gene expression of Escherichia coli. Biochimie 78:209-212.

Inada, T., and Y. Nakamura. 1995. Lethal double-stranded RNA processing activity of ribonuclease III in the absence of SuhB protein of Escherichia coli. Biochimie 77:294-302.

Ingolia, T. D., and D. E. koshland, Jr. 1979. Response to a metal ion- citrate complex in bacterial sensing. J. Bacteriol. 140:798-804.

Jones, A. L., D. DeShazer, and D. E. Woods. 1997. Identification and characterization of a two-component regulatory system involved in invasion of eukaryotic cells and heavy-metal resistance in Burkholderia pseudomallei. Infect Immun. 65:4972-4977.

Jordan, D. C. 1984. The Rhizobiaceae. p. 242. In N. R. Krieg and J. G. Holts (ed.) Bergey’s manual of systematic bacteriology, vol. 1. Williams & Wilkins, Baltimore.

209 Jorgensen, R. A., S. J. Rothstein, and W. S. Reznikoff. 1979. A restriction enzyme cleavage map of Tn5 and location of a region encoding neomycin resistance. Mol. Gen. Genet. 17 7 :65-72.

Joys, T. M. 1985. The covalent structure of the phase-1 flagellar filament protein of Salmonella typhimurium and its comparison with other flagellins. J. Biol. Chem. 260:15758-15761.

Kaempf, C., and E. P. Greenberg. 1992. Negative chemotaxis in Spirochaeta aurantia. Gurr. Microbiol. 2 1 :187-192.

Kaiser, D. 1979. Social gliding is correlated with the presence od pili in Myxococcus xanthus. Proc. Natl. Acad. Sci. USA 76: 5952-5956.

Kaiser, G. R., and R. N. Doetsch. 1975. Enhanced translational motion of Leptospira in viscous environments. Nature 255:656-657.

Kamoun, S., and C. I. Kado. 1990. Phenotypic switching affecting chemotaxis, xanthan production and virulence in Xanthomonas campestris. Appl. Environ. Microbiol. 56:3855-3860.

Kape, R., M. Parnlske, and D. Warmer. 1991. Chemotaxis and nod gene activity of Bradyrhizobium japonicum in response to hydroxycinnamic acids and isoflavonoids. Appl. Environ. Microbiol. 57:316-319.

Kelley, T. R., O. C. Pancorbo, W. C. Merka, and H. M. Barnhart. 1998. Antibiotic resistance of bacterial litter isolates. Poult. Sci. 77:243-247.

Kennedy, M. J. 1987. Role of motility, chemotaxis and adhesion in microbial ecology. Ann. N Y Acad. Sci. 506:260-273.

KondorosI, E., M. Pierre, M. Cren, U. Haumann, M. Buire, B. Hoffmann, J.Schell, and A. KondorosI. 1991. Identification of NoIR, a negative transacting factor controlling the nod regulon in Rhizobium meliloti. J. Mol. Biol. 222:885-896.

KrupskI, G., R. Gôtz, K. Ober, E. Pleler, and R. Schmitt. 1985' Structure of complex flagellar filament in Rhizobium meliloti. J. Bacteriol. 162:361-366.

Lapldus, I. R., M. Welch, and M. Elsenbach. 1988. Pausing of flagellar rotation is a component of bacterial motility and chemotxis. J. Bacteriol. 170:3627-3632.

210 Latifi, A., M. K. Winson, M. Fogllno, B. W. Bycroft, G. S. Stewart, A. Lazdunski, and P. Williams. 1995. Multiple homologues of LuxR and LuxI control expression of virulence determinants and secondary metabolites through quorum sensing In Pseudomonas aeruginosa PAO^. Mol. Microbiol. 17:333-343.

Lauffenburger, D., R. Aris, and K. Keller. 1982. Effects of cell motility and chemotaxis on microbial population growth. Blophys. J. 40:209-219.

Lengeler, J., A. M. Auburger, R. Mayer, and A. Pecher. 1981. The phosphoenolpyruvate-dependent carbohydrate phophotransfersae system II enzymes as chemoreceptors In chemotaxis of Escherichia coli K12. Mol. Gen. Genet. 183:163-170.

Leigh, J. A., and D. L. Coplin. 1992. Exopolysaccharides In plant- bacterial Interactions. Annu. Rev. Microbiol. 40:307-346.

Leigh, J. A., E. R. Signer, and G. C. Walker. 1985. Exopolysaccharlde- deflclent mutants of Rhizobium meliioti that form Ineffective nodules. Proc. Natl. Acad. Sci. USA 82:6231-6235.

Lewis, L. K., G. R. Harlow, L. A. Gregg-Jolly, and D. W. Mount. 1994. Identification of high affinity binding sites for LexA which define new DNA damage-lnduclble genes In Escherichia coli. J. Mol. Biol. 241:507-523.

Li, C., C. J. Louise, W. Shi, and J. Adler. 1993. Adverse conditions which cause lack of flagella In Escherichia coli. J. Bacteriol. 175:2229-2235.

Li, C., A. J. Boileau, C. Kung, and J. Adler. 1988. Osmotaxis In Escherichia coli. Proc. Natl. Acad. Sci. USA 85:9451-9455.

Liao, C. H., D. E, McCallus, and W. F. Fett. 1994. Molecular characterization of two gene loci required for production of the key pathogenicity factor pectate lyase In Pseudomonas viridifiava. Mol. Plant- Mlcrobe Interact. 7:391-400.

Lindum, P. W., U. Anthoni, C. Christophersen, L. Eberl, S. Molin, M. Givskov. 1998. N-Acyl-L-homoserlne lactone autoinducers control production of an extracellular llpopeptlde biosurfactant required for swarming motility of Serratia liquefaciens MG1. J. Bacteriol. 180:6384-6388.

Liu, J., and P. Zuber. 1998. A molecular switch controlling competence and motility; competence regulatory factors ComS, MecA, and ComK control sD- dependent gene expression In Bacillus subtilis. J. Bacteriol. 180:4243-4251.

211 Liu, R., V. M. Tran, and E. L. Schmidt. 1989. Modulating competitiveness of a nonmotile Tn7 mutant of Bradyrhizobium Japonicum in nonsterile soil. Appl. Environ. Microbiol. 55:1895-1900.

Long, S., J. W. Reed, J. Himawan, and G. C. Walker. 1988. Genetic analysis of a Cluster of genes required for synthesis of the Calcofluor-binding exopolysaccharide of Rhizobium meliloti. J. Bacteriol. 170 : 4239-4248.

Macnab, R. M. 1979. Chemotaxis in bacteria, p. 310-334. /n W. Haupt and M. E. Feinleib (ed.). Encyclopedia of plant physiology, new series, vol. 7. Springer-Verlag, Berlin.

Macnab, R. M. 1987. Motility and chemotaxis, p. 732-759. In F. C. Neidhardt, J. L Ingraham, K. B. Low, B. Magasanik, M. Schaechter, and H. E. Umbarger (eds.), Eschen'chia coli and Salmonella typhymurium :cellular and molecular biology. American Society for Microbiology, Washington D. C.

Macnab, R. M. 1996. Flagella and motility, p. 123-145. In F. C. Neidhardt, R. Curtiss III, J. L. Ingraham, E. C. C. Lin, K. B. Low, B. Magasanik, W. S. Reznikoff, M. Riley, M. Schaechter, and H. E. Umbarger (eds.), Escherichia coli and Salmonella : cellular and molecular biology, 2nd ed. American Society for Microbiology, Washington D. C.

Macnab, R. M., and S.-l. Aizawa. 1984. Bacterial motility and the bacterial flagellar motor. Annu. Rev. Biophys. Bioeng. 13:51-83.

Maeda, K., and Y. Imae. 1979. Thermosensory transduction in Escher/cri/a coli : inhibition of the thermoresponse by L-serine. Proc. Natl. Acad. Sci. USA 76:92-95.

Malek, W. 1992. The role of motility in the efficiency of nodulation by Rhizobium meliloti. Arch. Microbiol. 58:26-28.

Malmcrona-Friberg, K., A. Goodman, and S. Kjelleberg. 1990. Chemotactic responses of marine Vibrio spc. strain SI 4 (CCUG15956) to low molecular-weight substances under starvation and recovery conditions. Appl. Environ. Microbiol. 56:3699-3704.

Marold, L. M., R. Freedman, R. E. Chamberlain, and J. J. Miyashlro. 1981. New selective agent for isolation of Pseudomonas aeruginosa. Appl. Environ. Microbiol. 41 .-977-980.

Maruyama, M., G. Lodderstaedt, and R. Schmitt. 1978. Purification and biochemical properties of complex flagella isolated from Rhizobium lupini H13-

212 3. Biochim. Biophys. Acta 535:110-114.

Matsuhisa, A., N. Suzuki, T. Noda, and K. Shiba. 1995. Inositol monophosphatase activity from the Escherichia coli suhB gene product. J. Bacteriol. 177:200-205.

Mayfield, C. I., and W. E. Inniss. 1977. A rapid, simple method for staining bacterial flagella. Can. J. Microbiol. 23:1311-1313.

McLean, R. J., M. Whiteley, D. J. Stickler, and W. C. Fuqua. 1997. Evidence of autoinducer activity in naturally occurring biofilms. FEMS Microbiol. Lett. 154:259-263.

Mesibov, R., and J. Adler. 1972. Chemotaxis toward amino acids and some sugars. J. Bacteriol. 112:315-326.

Moriarty, D., J. W., and R. T. Bell. 1993. Bacterial growth and starvation in aquatic environments, p. 25-53. in S. Kjelleberg (ed.). Starvation in bacteria. Plenum Press, New York and London.

Morlta, R. Y. 1988. Bioavailability of energy and its relationship to grov/th and starvation survival in nature. Can. J. Microbiol. 34:436-441.

Morlta, R. Y. 1993. Bioavailability of energy and starvation state, pA -23. In S. Kjelleberg (ed.). Starvation in bacteria. Plenum Press, New York and London.

Murayama N., H. Shimizu, S. TaklguchI, Y. Baba, H. Amino, T. HorluchI, K. Seklmlzu, and T. Mikl. 1996. Evidence for involvement of Escherichia coli genes pmbA, csrA and a previously unrecognized gene tIdD, in the control of DNA gyrase by letD (ccdB) of sex factor F. J. Mol. Biol. 256:483-502.

Nagasawa, S., S. Toklshlta, H. Alba, and T. MIzuno. 1992. A novel sensor-regulator protein that belongs to the homologous family of signal- transduction proteins involved in adaptive responses in Escherichia coli. Mol. Microbiol. 6:799-807.

Narberhaus, P., P. Krummenacher, H. M. Fischer, and H. Hennecke. 1997. Three disparately regulated genes for sigma 32-like transcription factors in Bradyrhizobium japonicum. Mol. Microbiol. 24:93-104.

Nealscn, K. H., T. Platt, and J. Hastings. 1970. Cellular control of the synthesis and activity of the bacterial luminescent system. J. Bacteriol. 104:313-322.

213 Nikata, T., K. Sumida, J. Kato, and H. Ohtake. 1992. Rapid method for analyzing bacterial behavioral responses to chemical stimuli. Appl. Environ. Microbiol. 58:2250-2254.

Ninfa, E. G., A. Stock, 8. Mowbray, and J. Stock. 1991. Reconstitution of bacterial signal transduction system from purified components. J. Biol. Chem. 266:9764-9770. Nuijten, P. J., L. Marquez-Magana, and B. A. van der Zeijst. 1995. Analysis of flagellin gene expression in flagellar phase variants of Campylobacter /e/un/81116. Antonie van Leeuwenhoek 67:377-383.

Nystrom, T. 1993. Global systems approach to the physiology of the starved cell, p. 129-150. In S. Kjelleberg (ed.), Starvation in bacteria. Plenum Press, New York.

Oiler, L., F- Kocka, R. W. Smith, and H. Koffler. 1971. The presence of flagellins A and B in the flagellum of Bacillus pumilis 101. Bacteriol. Proc. 1971:27.

Ohta, N, T. Lane, E. G. Ninfa, J. M. Sommer, and A. Newton. 1992. A histidine protein kinase homologue required for regulation of bacterial cell division and differentiation. Proc. Natl. Acad. Sci. USA 89:10297-10301.

Okazaki, N., S. Matsuo, K. Saito, A. Tominaga, and M. Enomoto. 1993. Conversion of the Salmonella phase 1 flagellin gene fliC to the phase 2 gene /7/Bon the Escherichia coliK-^2 chromosome. J. Bacteriol. 175:758-766.

Oliver, J. D. 1993. Formation of viable but nonculturable cells, p.239-272. In S. Kjelleberg (ed.). Starvation in bacteria. Plenum Press, New York.

Olsen, P. E., W. A. Rice, L. M. Bordeleau, A. H. Demidoff, and M. M. Collins. 1996. Levels and identities of nonrhizobial microorganisms found in commercial legume inoculant made with nonsterile peat carrier. Can. J. Microbiol. 42:72-75.

O'Toole R., D. L. Milton, and H. Wolf-Watz. 1996. Chemotactic motility is required for invasion of the host by the fish pathogen Vibrio anguillarum. Mol. Microbiol. 19:625-637.

Ottomann, K. M., and J. F. Miller. 1997. Roles for motility in bacterial-host interactions. Mol. Microbiol. 24:1109-1117.

Palleroni, N. J. 1976. Chamber for bacterial chemotaxis experiments. Appl. Environ. Microbiol. 32:729-730.

214 Parkinson, J. S., and P. T. Revello. 1978. Sensory adaptation mutants of E. coli. Cell 15:1221-1230.

Pate, J- L. 1988. Gliding motility in prokaryotic cells. Can. J. Microbiol. 34:459-465.

Pate, J. L., D. M. De Jong. 1990. Use of nonmotile mutants to identify a set of membrane proteins related to gliding motility in Cytophaga johnsonae. J. Bacteriol. 172:3117-3124.

Pearson, J. P., L. Passador, B. H. IgiewskI, E. P. Greenberg. 1995. A second N-acylhomoserine lactone signal produced by Pseudomonas aeruginosa. Proc. Natl. Acad. Sci. USA 92:1490-1494.

Pearson, J. P., E. C. Peso!, and B. H. Iglew ski. 1997. Role of Pseudomonas aeruginosa las and rhi quorum-sensing systems in control of elastase and rtiamnolipid biosynthesis genes. J. Bacteriol. 179:5458-5467.

Pesci E. C., Pearson J. P., Seed P. C., and B. H. iglewski. 1997. Regulation of las and rhI quorum sensing in Pseudomonas aeruginosa. J. Bacteriol. 179:3127-3132.

Pilgram W. K., and F. D. Williams. 1976. Survival value of chemotaxis in mixed cultures. Can. J. Microbiol. 22:1771-1773.

Platzer, J., W. Sterr, M. Hausmann, and R. Schmitt. 1997. Three genes of a motility operon and their role in flagellar rotary speed variation in Rhizobium meliloti. J. Bacteriol. 179:6391-6399.

Pleier, E., and R. Schmitt. 1989. Identification and sequence analysis of two related flagellin genes in Rhizobium meliloti. J. Bacteriol. 1 7 1 :1467-1475.

Pleier, E., and R. Schmitt. 1991. Expression of two Rhizobium meliloti flagellin genes and their contribution to the complex filament structure. J. Bacteriol. 173:2077-2085.

Potts, M. 1994. Desiccation tolerance of prokaryotes. Microbiol. Rev. 58:755- 805.

Pratt L. A., R. Kolter. 1998. Genetic analysis of Escherichia co//biofilm formation; roles of flagella, motility, chemotaxis and type I pili. Mol. Microbiol. 30:285-293.

215 Quandt, J., and M. F. Hynes. 1993. Versatile suicide vectors which allow direct selection for gene replacement in gram-negative bacteria. Gene 127:15-21.

Rebbapragada, A., M. S. Johnson, G. P. Harding, A. J. Zuccareili, H. M. Fletcher, I. B. Zhulin, B. L. Taylor. 1997. The Aer protein and the serine chemoreceptor Tsr independently sense intracellular energy levels and transduce oxygen, redox, and energy signals for Escherichia coli behavior. Proc. Natl. Acad. Sci. USA 94:10541-10546.

Reeder, T., and R. Schleif. 1991. Mapping, sequence, and apparent lack of function of araJ, a gene of the Escherichia coli arabinose regulon. J. Bacteriol. 17 3 :7765-7771.

Rehm, B. H., H. Ertesvag, S. Valla. A 996. A new Azotobacter vinelandii mannuronan G-5-epimerase gene (aIgG) is part of an alg gene cluster physically organized in a manner similar to that in Pseudomonas aeruginosa. J. Bacteriol. 178:5884-5889.

Reinhold B., T. Hurek, and I. Fendrlk. 1985. Strain-specific chemotaxis of Azospirillum spp. J. Bacteriol. 162:190-195.

Reuber, T. L., 8. Long, and G. C. Walker. 1991. Regulation of R. meliloti exo genes in free-living cell and in planta examined by using JnphoA fusion. J. Bacteriol. 173:426-434.

Rice, S. A., and J. D. Oliver. 1992. Starvation response of the marine barophile CNPT-3. Appl. Environ. Microbiol. 58:2432-2437.

Robinson, J. B. 1991. Regulation of motility and chemotaxis to C 4 - dicarboxylic acids in Rhizobium meliloti: the effect of divalent cations and nutrient limitation. Ph. D. dissertation. The Ohio State university.

Robinson, J. B., O. H. Tuovinen, and W. D. Bauer. 1992. Role of divalent cations in the subunit associations of complex flagella from Rhizobium meliloti. J. Bacteriol. 174:3896-3902.

Robyt, J. F., and B. J. White. 1987. Biochemical techniques, theory and practice. Waveland Press, Inc. IL

Ronco II, P. G. 1988. Tn5 mutagenesis of the motility and chemotaxis genes. M. S. Thesis, Northeastern University, Boston, Mass.

216 Rosenbluh, A., and M. Esenbach. 1992. Effect of mechanical removal of pili on gliding motility of Myxococcus xanthus. J. Bacteriol. 174:5406-5413.

Russo, A. K., and D. E. Koshland, Jr. 1983. Separation of signal transduction and adaptation function of the aspartate receptor in bacterial sensing. Science 220:1016-1020.

Ruvkun, G. B., S. R. Long, H. M. Meade, R. C. van den Bos, and F. M. Ausubel. 1982. ISRmI: A Rhizobium meliloti insertion sequence that transposes preferentially into nitrogen fixation genes. Mol. Appl. Genet. 1 :405- 418.

Salmond, G. P., B. W. Bycroft, G. S. Stewart, and P. W illiams. 1995. The bacterial 'enigma': cracking the code of cell-cell communication. Mol. Microbiol. 16:615-624.

Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: A laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N. Y.

Schell, M. A., T. P. Denny, and J. Huang. 1994. VsrA, a second two- component sensor regulating virulence genes of Pseudomonas solanacearum. Mol. Microbiol. 1 1 :489-500.

Schmitt, R., I. Raska, and F. Mayer. 1974. Plain and complex flagella of Pseudomonas rhodos\ analysis of fine structure and composition. J. Bacteriol. 117:844-857.

Schmitt, 0, K., S. C. Darnell, and A. D. O'Brien. 1996. The Sa/mone//a typhimurium fIgM gene, which encodes a negative regulator of flagella synthesis and is involved in virulence, is present and functional in other Salmonella species. FEMS Microbiol. Lett. 135:281-285.

Schuster, S. C., and S. Khan. 1994. The bacterial flagellar motor. Annu. Rev. Biophys. Biomol. Struct. 23:509-539.

Schweizer H. P., and C. Po. 1994. Cloning and nucleotide sequence of the gIpD gene encoding sn-glycerol-3-phosphate dehydrogenase of Pseudomonas aeruginosa. J. Bacteriol. 176:2184-2193.

Seed, P. C., L. Passador, and B. H. Iglewski. 1995. Activation of the Pseudomonas aeruginosa Iasi gene by LasR and the Pseudomonas autoinducer PAl: an autoinduction regulatory hierarchy. J. Bacteriol. 177:654-659.

217 Segall, J. E., M. D. Manson, and H. C. Berg. 1982. Signal processing times in bacterial chemotaxis. Nature (London) 296:855-857.

Selbitschka, W., W. Arnold, D. Jording, B. Kosier, N. Toro, and A. Puhler. 1995. The insertion sequence element ISRm2011-2 belongs to the 18630-Tcl family of transposable elements and is abundant in Rhizobium meliloti. Gene 63:59-64.

Selvaraj, G., and V. N. Iyer, 1983. Suicide plasmid vehicles for insertion mutagenesis in Rhizobium meliloti and related bacteria. J. Bacteriol. 1 56:1292- 1300.

Shiba, K., Ito K., and T. Yura. 1984. Mutation that suppresses the protein export defect of the secY mutation and causes cold-sensitive growth of Escherichia coli. J. Bacteriol. 1 60:696-701.

Shioi, J.-l., R. J. Galloway, M. Niwano, R. E. Chinnock, and B. L. Taylor. 1982. Requirement of ATP in bacterial chemotaxis. J. Biol. Chem. 257:7969-7975.

Shioi, J., C. V. Dang, and B. L. Tayior. 1987. Oxygen as attractant and repellent in bacterial chemotaxis. J. Bacteriol. 169:3118-3123.

Silverman, m., and M. Simon. 1974. Flagellar rotation and the mechanism of bacterial motility. Nature (London) 249:73-74.

Silverman, M., J. Zieg, M. Hiimen, and M. Simon. 1979. Phase variation in Salmonella: genetic analysis of a recombinational switch. Proc. Natl. Acad. Sci. USA 76:391-395.

Simon M., Zieg J., M. Silverman, G. Mandel, and R. Doolittle. 1980. Phase variation: evolution of a controlling element. Science 209:1370-1374.

Skowlund, C. T., and D. W. Kirmse. 1988. Comparison of the relative concentration of motile and non-motile bacteria in small pores. J. Theor. Biol. 133:499-512

Soby, S., and K. Bergman. 1983. Motility and chemotaxis of Rhizobium meliloti \n soil. Appl. Environ. Microbiol. 46: 995-998.

Sokol, P. A., M. Z. Luan, D. G. Storey, and P. Thirukkumaran. 1994. Genetic rearrangement associated with in vivo mucoid conversion of Pseudomonas aeruginosa PAO is due to insertion elements. J. Bacteriol. 176:553-562.

2 1 8 Sourjik, V., and R. Schmitt. 1996. Different roles of CheY1 and CheY2 In the chemotaxis of Rhizobium meliloti. Mol. Microbiol. 22:427-436.

Sourjik, V., W. Sterr, J. Platzer, I. Bos, M, Hasibeck, and R. Schmitt. 1998. Mapping of 41 chemotaxis, flagellar and motility genes to a single region of the Sinorhizobium meliloti chromosome. Gene 223:283-290.

Southam, G., M. L. Kaimokoff, K. F. Jarreii, S. F. Koval, and T. J. Beveridge. 1990. Isolation, characterization, and cellular insertion of the flagella from two strains of the archaebacterium Methanospirillum hungatei. J. Bacteriol. 172:3221-3228.

Spencer, M. E., M. G. Dariison, P. E. Stephens, I. K. Duckenfieid, and J. R. Guest. 1984. Nucleotide sequence of the sucB gene encoding the dihydrolipoamide succinyltransferase of Escherichia coliK^2ar\d homology with the corresponding acetyltransferase. Eur. J. Biochem. 141:361-374.

Srinivasan, S., J. Ostling, T. Charlton, R. de Nys, K. Takayama, and S. Kjelleberg. 1998. Extracellular signal molecule(s) involved in the carbon starvation response of marine Vibrio spc. strain S I4. J. Bacteriol. 180:201-209.

Staskawicz, B., D. Dahlbeck, N. Keen, and 0. Napoli. 1987. Molecular characterization of cloned avirulence genes of race 0 and race 1 of Pseudomonas syringaep\/. glycinea. J. Bacteriol. 1 69:5789-5794.

Stenstrom T. A., and S. Kjelleberg. 1985. Fimbriae mediated nonspecific adhesion of Salmonella typhimurium to mineral particles. Arch. Microbiol. 143:6- 10.

Stenstrom T. A. 1989. Bacterial hydrophobicity, an overall parameter for the measurement of adhesion potential to soil particles. Appl. Environ. Microbiol. 55:142-147.

Stevenson, G., K. Andrianopoulos, M. Hobbs, and P. R. Reeves. 1996. Organization of the Escherichia coli K-12 gene cluster responsible for production of the extracellular polysaccharide colanic acid. J. Bacteriol. 178:4885-4893.

Stock, J. B., and M. G. Surette. 1996. Chemotaxis, p.1103-1129. InF .C . Neidhardt, R. Curtiss III, J. L. Ingraham, E. C. C. Lin, K. B. Low, B. Magasanik, W. S. Reznikoff, M. Riley, M. Schaechter, and H. E. Umbarger (ed.), E. coli and Salmonella : cellular and molecular biology, 2nd ed. American society for Microbiology, Washington D. C.

219 Stragier, and J. C. Patte. 1983. Regulation of diamlnoplmeiate decarboxylase synthesis in Escherichia coli. III. Nucleotide sequence and regulation of a lysR gene. J. Mol. Biol. 168:333-350.

Stragier, P, F. RIchaud, F.Borne, and J. C. Patte. 1983. Regulation of diaminopimelate decarboxylase synthesis in Escherichia coli. I. Identification of a lysR gene encoding an activator of the lysA gene. J. Mol. Biol. 168:307-320.

Stretton, S., S. J. Danon, 8. Kjelleberg, and A. E. Goodman. 1997. Changes in cell morphology and motility in the marine Vibrio spc. strain SI 4 during conditions of starvation and recovery. FEMS Microbiol, lett. 146:23-29.

Szekely, E., and M. Simon. 1983. DNA sequence adjacent to flagellar genes and evolution of flagellar-phase variation. J. Bacteriol. 155:74-81.

Tamura, Y., M. Kljima-Tanaka, A. AokI, Y. Oglkubo, and T. Takahashl. 1995. Reversible expression of motility and flagella in Clostridium chauvoei and their relationship to virulence. Microbiology 141:605-610.

Taylor, B. L., and 1. B. Zhulin. 1998. In search of higher energy: metabolism- dependent behaviour in bacteria. Mol. Microbiol. 28:683-690.

Terraclano, J. S., and E. Canale-Parola. 1984. Enhancement of chemotaxis in Spirochaeta aurantia grown under conditions of nutrient limitation. J. Bacteriol. 159:173-178.

Torrella, F., and R. Y. Morlta. 1981. Microcultural study of bacteria size changes and microcolony and ultramicrocolony formation by heterotrophic bacteria in seawater. Appl. Environ. Microbiol. 45:518-527.

Torrella, F., and R. Y. Morlta. 1982. Starvation induced morphological changes, motility, and chemotaxis patterns in a psychrophilic marine vibrio. Publ. Centr. Natl. Exploit. Actes Colloq. (Fr.) Oceans no. 13. pp45-60.

Tso, W. W., and J. Adler. 1974. Negative chemotaxis in Escherichia co//. J. Bacteriol. 118:560-576.

Tuhela, L., J. B. Robinson, and O. H. Tuovinen. 1998. Characterization of chemotactic responses and flagella of Hyphomicrobium strain W1-1B. J. Bacteriol. 180:3003-3006. von Bodman, S. B., D. R. Majerczak, and D. L. Coplin. 1998. A negative regulator mediates quorum-sensing control of exopolysaccharide production in Pantoea stewartii subsp. stewartii. Proc. Natl. Acad. Sci. USA 95:7687-7692.

220 van Elsas, J. D., and L. S. van Overbeek. 1993. Bacterial responses to soil stimuli, p. 55-77. In S. Kjelleberg (ed.). Starvation in Bacteria. Plenum Press. New York and London.

Vidal-lngigllardi, D., and O. Raibaud. 1991. Three adjacent binding sites for cAMP receptor protein are involved in the activation of the divergent malEp- ma/Kp promoters. Proc. Natl. Acad. Sci. USA 88:229-233.

Wang, S. P., P. L. Sharma, P. V. Schoenlein, and B. Ely. 1993. A histidine protein kinase is involved in polar organelle development in Caulobacter crescentus. Proc. Natl. Acad. Sci. USA 90:630-634.

Ward, M. J., A. W. Bell, P. A. Hamblin, H. L. Packer, and J. P. Armitage. 1995. Identification of a chemotaxis operon with two cheY genes in Rhodobacter sphaeroides. Mol. Microbiol. 17:357-366.

Watson, R J., Y. K. Chan, R. Wheatcroft, A Yang, and S. Han. 1988. Rhizobium meliloti genes required for C 4 -dicarboxylate transport and symbiotic nitrogen fixation are located on a megaplasmid. J. Bacteriol. 170:927-934.

Wei, X., and W. D. Bauer. 1998. Starvation-induced changes in motility, chemotaxis and flagellation of Rhizobium meliloti. Appl. Environ. Microbiol. 64: 1708-1714.

Wei, X., and W. D. Bauer. 1999. Tn5-induced and spontaneous switching of Sinorhizobium meliloti to faster-swarming behavior. Appl. Environ. Microbiol. 65:1228-1235.

West, M., N. M. Burdash, and F. Freimuth. 1977. Simplified Silver-plating stain for flagella. J. Clin. Microbiol. 6:414-419.

Williams, F. D., D. M. Anderson, P. S. Hoffman, R. H. Schwarzhoff, and S. Leonard. 1976. Evidence against the involvement of chemotaxis in swarming of Proteus mirabilis. J. Bacteriol. 127:237-248.

Williams V., and M. Fletcher. ^996. Pseudomonas fluorescens adhesion and transport through porous media are affected by lipopolysaccharide composition. Appl. Environ. Microbiol. 62:100-104.

Winson, M. K., M. Camara, A. Latifi, M. Foglino, S. R. Chhabra, M. Daykin, M. Bally, V. Chapon, G. P. Salmond, B. W. Bycroft, et al. 1995. Multiple N-acyl-L-homoserine lactone signal molecules regulate production of virulence determinants and secondary metabolites in Pseudomonas aeruginosa. Proc. Natl. Acad. Sci. USA 92:9427-9431.

221 Wrangstadh M.. U. Szewzyk, J. Ôsling, and S. Kjelleberg. 1990. Starvation-specific formation of a peripheral exopolysaccharide by a marine Pseudomonas spc., strain S9. Appl. Environ. Microbiol. 56:2065-2072 (1990)

Yang, X. Y., H. Schulz, M. EIzinga, and S. Y. Yang. 1991. Nucleotide sequence of the promoter and fadB gene of the fadBA operon and primary structure of the multifunctional fatty acid oxidation protein from Escherichia coli. Biochemistry 30:6788-6795.

Yano, R., H. Nagai, K. Shiba, and T. Yura. A mutation that enhances synthesis of sigma 32 and suppresses temperature-sensitive growth of the rpo/-/15 mutant of Escherichia coli. J. Bacteriol. 1990.172:2124-2130.

Yeung, K. H., M. A. Schell, and P. G. Hartel 1989. Growth of genetically engineered Pseudomonas aeruginosa and Pseudomonas putida in soil and rhizosphere. Appl. Environ. Microbiol. 55:3243-3246.

Yonekawa, H., H. Hayashi, and J. S. Parkinson. 1983. Requirement of the cheB function for sensory adaptation in Escherichia coli. J. Bacteriol. 156:1228-1235.

Zekri, S., M. J. Soto, and N. Toro. 1998. ISRm4-1 and ISRm9, two novel insertion sequences from Sinorhizobium meliloti. Gene 207:93-96.

Zhan, H., and J. A. Leigh. 1990. Two genes that regulate exopolysaccharide production in Rhizobium meliloti. J. Bacteriol. 172:5254-5259.

Zhang, L. and A. Kerr. 1991. A diffusible compound can enhance conjugal transfer of the Ti plasmid in Agrobacterium tumefaciens. J. Bacteriol. 173:1867- 1872.

Zhulin, I. B. , E. H. Rowsell, M. S. Johnson, B. L. Taylor. 1997. Glycerol elicits energy taxis of Escherichia coli and Salmonella typhimurium. J. Bacteriol. 1 7 9 :3196-3201.

222 IMAGE EVALUATION TEST TARGET (QA-3)

y

/ r

%

12.8 1.0 ys m L& B e 2.2 !r 83.6 2.0 l.i 1.8

1.25 1.4 1.6

150mm

V

V ^ IP P L IE D A IIVI/4GE . Inc . = 1653 East Main Street — ■ Rochester, NY 14609 USA _L2 J=.-_= Phone: 716/482-0300 . ^ = ~ . ^ = Fax: 716/288-5989

O 1993. Applied Image, Inc.. Ail Rights Reserved

O 7