Isolation and Characterization of Mutants Affecting Carbon Catabolism in Sín O Rh I Zob Ium Melilo Ti

Isolation and characterization of mutants affecting carbon catabolism in Sín o rh i zob ium melilo ti

Nathan lohn Poysti

A Thesis submitted to the Faculty of Graduate studies of the university of Manitoba in partial fulfilment of the requirements of the degree of

Master of Science

Department of Microbiology University of Manitoba Winnipeg, Manitoba Canada

@ Nathan Poysti ,2002 THE UNTVERSITY OF MANITOBA

FACULTY OF GRADUATE STTIDIES tr*rrtr* COPYRIGHT PERMISSION

lsolation and characterization of mutants Affecting carbon catabolism in Sinorhizobium meliloti

Nathan John Poysti

A ThesislPracticum submitted to the Faculty of Graduate Studies of The University of

Manitoba in partial fulfillment of the requirement of the degree

MASTER OF SCIENCE

Nathan John Poysti O 2007

Permission has been granted to the University of Manitoba Libraries to lend a copy of this thesis/practicum, to Library and Archives Canada (LAC) to lend a copy of this thesis/practicum, and to LAC's agent (UMlÆroQuest) to microfilm, sell copies and to pubtish an abstract of this thesis/practicum.

This reproduction or copy of this thesis has been made availabte by authority of the copyright owner solely for the purpose of private study and research, and may only be reproduced and copied as permitted by copyright laws or with express written authorization from the copyright owner. Abstract

Sinorhizobium meliloti is a bacterium capable of participating in a symbiotic relationship with plants in which it forms N2 fixing nodules. The ability to use specific carbon sources has been shotnm to provide S. melíIoti and other rhizobia with competitive ad- vantages in the ability to effectively nodulate host plants. The sugars rhamnose and arabinose were hypothesized to play a role in the abitity to compete for nodule occupancy in S. melilori. To test these hypotheses, the S. meliloti wild type, RmI02l, was subjected to transposon mutagenesis experiments and the resulting mutants were screened for the inability to use rhamnose or arabinose as sole carbon sources. Mutations affecting rhamnose and arabinose catabolism were isolated and subsequent characterization resulted in the identification of a rhamnose catabolism operon, an arabinose catabolism operon and a triose phosphate isomerase encoding gene. A-lthough mutants unable to utilize arabinose were not less competitive than wild type, the ability to utilize rhamnose was found to play a role in competition for nodule occupancy. Furthermore, the roles of two triose phosphate isomerase genes were characterized with respect to symbiosis and the free-living state.

II Acknowledgements

First, I would like to sincerely thank my supervisor, Dr. Ivan Oresnik, for his motivation,

enthusiasm and support during this project. I am very glad to have found such a ded-

icated supervisor! I would also like to thank my advisory committee: Dr. Harry Duck-

worth and Dr. Teri de Kievit, for their thorough consideration and excellent suggestions regarding my work and thesis.

i would like to thank my friends in the lab: Iason Richardson, Mark Miller_Williams,

Brad Pickering, Erin Loewen, and HarryYudistira. Not only did we have excellent scien-

tific discussions, but we had a lot of fun too!

I am very grateful to both my immediate and extended family for being so enthusi- astic and encouraging during my studies.

Last, but not least, Kimberly Poysti has been incredible. I sincerely appreciate her dedication, love, and patience.

I would like to acknowledge the Natural Sciences and Engineering Research Council of Canada fo¡ their generous funding.

III Contents

Abstract II Acknowledgements IiI List of tables. . . . VII List of figures . . . VIII I-ist of abbreviations x Introduction I S. nteliloti symbiosis and nodulation Ja S. meliloti carbon metabolism 7 Bacteroid metabolism 2l Competition for nodule occupancy . . . . 25 ca Research goals J¿

Rhamnose catabolism and transport in S. melílotí 34 Summary 34 Introduction . . . 35 Materials and methods . . . 36 Bacterial strains, plasmids, and media 36 Genetic techniques 3B Rhamnose transport assays 3B Plantassays .... 3B Results 39 S. meliloti rhamnose catabolism Iocus 39 Rhamnose mutants are not sensitive on rhamnose/glycerol media 40 Heterologous complementation of S. meliloti rhamnose mutants 43 R. leguminosa.rum RhaK is capable of complementing S. meliloti rhaK mutations. 46 Testing the S. meliloti rhalocus for uptake of rhamnose. 47 s. meliloti rhamnose mutants are less competitive for nodule occupancy 50 Discussion 52

ry Characterization of S. m.eliloti triose phosphate isomerase genes 57 Summary 57 Introduction 5B Materials and methods . . . 59 Bacterial strains, plasmids, and media 59

DNA manipulations and plasmid constructions . . 60

Genetic techniques and mutant construction. . . 63

Mutant identification . . 64 liiose phosphate isomerase assays 65 Symbiotic competence assays 65 Results 66 Analysis of tpiAand tpiB 66 tpiAand tpiB are needed for gluconeogenesis 69 Both tpíAand tpiB encode triose phosphate isomerases 69 In uiuo complementation of tpiA 72

Constitutively expressed tpiA cannot complement tpiB . /J

S)¡mbiotic competence phenotypes . . . 75 Discussion 79

Arabinose catabolism and transport in S. meliloti B4 Summary B4 Introduction B5 Methods UIoa Bacterial strains and growth conditions 87 Genetic techniques B9 Mutant identification. . . B9 Arabinose dehydrogenase assay. 90

B - galactosidase assays 9l Arabinose and glucose transport assays. 9l Symbiotic profi ciency assays. 92 Symbiotic competition assays. 92 Iìesults 93 Isolation of arabinose catabolism mutants. 93 Analysis of arabinose transcriptional unit. 94 Arabinose dehydrogenase activity is present in an araD mutant. 99

Induction of araoperon by arabinose, seed exudate and galactose...... r00 Expression of the ara opeÍon is moderately repressed by succinate and glucose. I02 The araoperon is required for arabinose transport. . . . 104 Glucose uptake is not compromised in arabinose catabolism mutants. . . 106

V Plant symbiosis and competition for nodule occupancy. . . . . r07 Discussion 108 Conclusion tl3

References tt7

VI List of tables

2.1 Bacterial strains and plasmids . . . J/ 2.2 Homology of S. meliloti and R.leguminosa.rumrhamnose loci 42 1) L.J Growth of s. melilofi rhamnose mutants on rhamnose and glycerol 44 2.4 The A. leguminosarum rharegion does not complement s. meliloti 45 2.5 R. leguminosarum rhaK complements s. melitoti rhaK mutations . . 48 3.1 Bacterial strains and plasmids . . . 6i 3.2 Primers used in this work 62 3.3 Relevant growth phenotypes of triose phosphate isomerase mutants on defined media 70 3.4 Triose phosphate isomerase activities measured for different strains . . . . ZI 3.5 complementation of triose phosphate isomerase mutants by plasmid expressed tpiAandtpiB . Z4 3.6 Dry weights from growth trials of various S. metiloti strains Z6

4.1 Bacterial strains and plasmids . . . BB 4.2 Induction of the arabinose operon by several carbon sources l0I 4.3 The arabinose operon is not subject to catabolite repression by glucose or succinate 103 4.4 Labeled arabinose accumulation bywild tlpe and arabinose mutants . . . 105

VII Líst of figures

o 1.1 Schematic of S. melilori hexose metabolism O

L.2 S. meliloti galactose catabolism via the De Ley-Doudoroff pathway . . . . i5 1.3 Schematic of the pentose phosphate pathway I7 I.4 Schematic of the S. meliloti tricarboxylic acid cycle . I9

2.I S. meliloti rhamnose utilization locus . 4I 2.2 s. meliloti wild type and mutant accumulation of labeled rhamnose . . . . 49

2.3 Rhamnose accumulation by s. melilotí grown in rhamnose and glycerol . 51 2.4 A rhamnose mutant is less competitive than wild type for nodule occupancy

3.t Physical map of S. meliloti triose phosphate isomerase encoding loci . . . 3.2 Nodulation kinetics of S. nteliloti wild type and triose phosphate isomerase mutants

4.r Physical map of S. meliloti arabinose utilization locus . 95 4.2 Schematic of the S. melilotí arabinose catabolism pathway 9B 4.3 Á,n arabinose catabolism mutant is not competitively deficient 109

VIII List of abbreviations

ABC ATP-binding cassette

ATP Adenosine triphosphate

CS Citrate slmthase DHAP Dihydroxyacetonephosphate DME Diphosphopyridine-nucleotidedependentmalicenzyme

DNA Deoxyribonucleic acid ED Entner-Doudoroff

EMP Embden-Meyerhof-Parnas

EPS Exopolysaccharide

FBA Fructose-bisphosphate aldolase

G3P Glyceraldehyde-3-phosphate

GAP Glyceraldehyde-3-phosphate dehydrogenase GC-MS Gaschromatography-massspectrometry

GDH Glucose dehydrogenase Gm Gentamicin

ICD Isocitrate dehydrogenase

x Kan Kanamycin KDPG 2-keto-3-deoxy-6-phosphogluconate

KGD ø-ketoglutarate dehydrogenase LB Luria-Bertani

MDH Malate dehydrogenase mRNA Messenger RNA MCP Methyl-acceptingchemotaxisprotein NAD.Ì Nicotinamideadeninedinucleotide

NADPr Nicotinamide adenine dinucleotide phosphate

Nm Neomycin

NMR Nuclear magnetic resonance

ODcoo Optical density at 600 nm

ORF Open reading frame

PCK Phosphoenolpyruvate carboxykinase

PCR Polymerase chain reaction

PFK Phosphofructokinase

PGDH 6-phosphogluconatedehydrogenase

PHB Polyhydroxybutyrate POD Pyruvateorthophosphatedikinase

PP Pentose phosphate

PYC Pyruvate carboxylase

X Rf Rifampicin

RNA Ribonucleic acid SDH Succinatedehydrogenase Sm Streptomycin Tc l.etracycline

TCA Tricarboxylic acid

TME Triphosphopyridine-nucleotidedependentmalic enzyme

TPI Triose phosphate isomerase

IRNA Transfer RNA

TY Tryptoneyeast

VMM Vincent's minimal medium

XI Chapter I Introduction

Sinorhizobium melilorl is a Gram-negative bacterium which is studied as a model sys-

tenr for plant-microbe symbiosis. S. meliloti lives a saproph¡ic lifestyle and has the

ability to form a symbiotic relationship with leguminous plants. Flost planrs include

species of several plant genera: alfalfa (Medicago); sweet clover (Metitotus) and fenu-

greek (Trigonella) (Jordan I984). The plant-microbe relationship resulrs in atmospheric

nitrclgen being fixed and provided to the plant by the microbe. The plant supplies the bacteria with carbon sources in exchange. In addition to being scientifically fascinat- ing, biological nitrogen fixation by r hizobia is very important in agriculture. Nitrogen is the primary supplement used in agriculture, however, it is expensive and requires large amounts of energy to produce (Bohlool et al. I992). Biological nitrogen fixation is therefore very important in the developing areas of the world where inorganic nitrogen can be prohibitively expensive. The promotion of the use of biological nitrogen fixation in the developed world is also important as it can lead to a number of environmental ben- elits over traditional inorganic nitrogen fertilizers. These include: the reduction of CO2 aud nitrous oxide emissions; decreased fuel dependency during production and transport of fertilizer; and dec¡eased leaching of nitrogenous compounds into water sources

(Iensen and Hauggaard-Nielsen 2003).

S. meliloti possesses a complex genome consisting of a chromosome and two mega- C|]APT'ER 1. INTRODUCTION plasmids named psymA and pSymB (Banfalvi et al. 1985, Finan et al. Ig86, Rosenberg et al. 1982). Extensive genetic mapping and the subsequent sequencing of the complete S. meliloti genome show that it is quite large compared to that of Escheríchia co-

/i (6.69 Mb and 4.64 Mb, respectively) (Barnett et al.2001, Capela et al. 2001, Charles and Finan 1990, Finan et al. 2001, Galibert et al. 200I, Honeycutt et al. lgg3, Riley et al.

2006). The chromosome of S. melilofl appears typical, in that it encodes nearly all of the cell's essential genes such as those encoding the machinery for mRNA to protein trans-

Iation, DNA repair and cell division. The megaplasmid pSymB has been described as chlomosome-like, as it encodes some essential genes: an arginine tRNA, and asparagine and thiamine bioslmthesis genes (Finan et al. Ig86, Galibert et al.200l). pSymB has not beetr completely cured, likely due to the presence of the arginine tRNA, but systematic deletions have sho'o¡n that it also encodes genes necessary for the catabolism of a number of small molecules (Charles and Finan 1991). The pSymA megaplasmid does not encode any functions necessary for viability and has been successfully cured from

S. meliloti in order to understand its role (Oresnik et al. 2000). Although several carbon utilization phenotypes have been associated with the loss of pSymA, it is kno',rm as a symbiotic plasmid because it carries the genes for symbiotic association and nitrogen fixation (Barnett et al. 2001, Galibert et al.200I, Honeycutt et al. I993).

Despite the importance of biological nitrogen fixation and the discovery of the first

Rhizobium species over 100 years ago fWillems 2006), our knowledge of the basic physiology and metabolism of rhizobium species is limited when compared to the model organism E. coli. The ability to utilize a broad range of carbon sources is important for a versatile organism with several growth habits. it is therefore not surprising that the genome of S. meliloti contains a large proportion of genes dedicated to carbon CI II|PT'ER 1. INTRODUCTION catabolism and substrate transport (Galibert et al. 2001). Correspondingly, S. meliloti is capable of catabolism and growth on many carbon sources including a wide range of monosaccharides, polysaccharides, organic acids and some aromatic compounds (Jor- dan 1984,Stowers t985). Thisreviewof theliteraturefocusesonwhatiscurrentlyknonm about S. nteliloti metabolism in relation to its biology, with particular attention paicl to central carbon metabolism.

S. melilofi spnbiosis and nodulation

S. meliloti is able to form a nitrogen fi*i.tg s)¡mbiotic association with a number of legume species. The process of forming the symbiotic association generally consists of plant-microbe recognition, infection, nodule development and bacterial differentiation. Initial mutual recognition of the host and bacterium depends on factors secreted by both (Spaink 2000). Plant roots secrete compounds into the rhizosphere which can be detected by rhizobia. Flavonoids are one such compound, and are the primary inducers of the bacterial nodgenes. The nod genes regulate and synthesize Nod factors, which are [3 - I ,4-linked N-acetyl-D-glucosamine oligosaccharide chains with a fatty acyl group attached to the non-reducing sugar. The nod genes appear to be unique to symbiotic rhizobia and are present on the megaplasmid psymA in S. melilori. The Nod factor synthesis genes are primarily induced by NodD. S. melíloti and some other strains possess multiple variants of nodD, allowing nod expression to be tailored in response to different flavonoids (Demont et al. 1994, Spaink 2000). Host plants detect excreted Nod factor produced by compatible symbionts and respond by allowing root cells in the center of the root to divide and become nodule primordium cells.

Infection and nodule development have been recently reviewed by Gage (200a). Fol- CHAPTER 1. INTRODUCTION lowing detection of Nod factor by the plant and the formation of nodule primordium cells, root hairs curl and trap adhered rhizobia between the root cells as they press together. The plant cell wall near trapped bacteria is then degraded; it is not currently known whether this degradation is caused by the bacterium or by the plant cell itself.

Cell wall degradation is followed by an invagination forming in the plant cell which grows through the root hair and into a root epidermal cell. Infecting rhizobia travel towards the inside of the root along this invagination, called an infection thread. Interest- ingly, studies of mixed populations of fluorescent bacteria in the infection thread have shor,rm that only the bacteria close to the tip of the infection thread seem to be growing

(Gage et al. 1996, Gage 2002). It has been suggested that the observed sectored growth may be due to the plant supplying an energy source only to the growing tip or to the bacteria sensing their relative position in the infection thread and adjusting their growth rate (Gage et al. 1996). Although the reason sectored growth occurs is not yet knor,rm, it usually causes a clonal population to form at the tip which results in only a single tlpe of bacterium being propagated into the nodule. After the infection thread grows through the epidermal cell wall, the bacteria are released into the intercellular space, where they continue moving through an infection thread which forms in an adjacent cortical cell (van Brussel et al. 1992, Gage 2004, Timmers et al. 1999). Infection threads continue to grow and branch off, eventually reaching multiple nodule primordia at the root center. Individual bacteria are brought into the nodule primordium cytoplasm by an endocytosis-like process where they divide and differentiate into bacteroids (Brewin

1991, Oke and Long i999). The extent of nodulation is plant controlled, as many more infection threads form than nodules (Bauer i98l). Although the method of action is not currently understood, the plant is able to abort active infection threads and prevent CHAPTER ]. INTRODUCTION further infections when enough nodules have developed.

Upon entering a nodule primordium's cytoplasm, rhizobia differentiate and express the genes required to become nitrogen fixing symbionts called bacteroids (Brewin l9gl,

Oke and Long 1999). Depending on the plant host, nodules can be of two types: determinate or indeterminate. In indeterminate nodules (such as those formed by S. metito- li), bacteria divide only a few times before differentiating, and undergo no further cell division. The nodules continually grow with new bacteroids forming from undifferentiated rhizobia present in the nodule. Indeterminate nodule bacteroids are often large and pleiomorphic in shape. In contrast, determinate nodules do not continually grow, and they contain bacteroids that appear similar to the infecting bacteria which can continually divide (Oke and Long 1999). The large size of S. melilori bacteroids is due to endoreplication, where cells replicate but do not separate. A study of S. meliloti bacteroids expressing red fluorescent protein showed that they were 5- l0 microns in length

(conrpared to 1-2 microns for free-living S. melilorl cells) and contained up to 24 times the normal amount of DNA per cell (Mergaert et aI.2006).

A significant question about bacteroid development remains to be answered: are bacteroids terminallydifferentiated? A number of studies have been published that have tested the ability of nodule bacteroids to grow after plating on lab media, with mixed results (Oke and Long 1999). However, another group has recently studied the nature of bacteroids in more depth (Mergaert et al. 2006). Firstly, they found that only 0.8% of S. meliloúi cells plated from indeterminate nodules formed colonies on agar plates, which they attribute to undifferentiated cells present along with bacteroids in the nodule (Mergaert et al. 2006). Similar results were found for R. Ieguminosarumbv. uiciae inoculated plants with indeterminate nodules (0.47o of bacteroids formed colonies). in CLIAPTER 1. INTRODUCTION

contrast, determinate nodules from Phaseolus uulgat'¿s (bean) inoculated with R. /eg- uminosarumtx. phaseoll contained viable bacteroids (207o formed colonies) which appeared similar to free-living cells.

Secondly, Mergaert et al. (2006) conducted an elegant experiment using recombinant rhizobia. An R. leguminosarumbv. uíciae strain (normally forming indeterminate nodules) carrying exLra nod genes allowing it to nodulate Lotus japonicus (determinate nodules) was inoculated and bacteroids were harvested. Bacteroids isolated from the determinate nodules were different from indeterminate bacteroids in that they were of a similar size to free-living cells, they had their normal amount of DNA, and they were recoverable from the nodule. The reverse experiment was performed with a modified

R. leguminosarumbv. phaseoli strain (modified to be able to nodulate Pisum satiuurn, indeterminate nodules). It was able to form large, branched bacteroids containing extra l)NA, which did not form colonies when plated on media. From these results, it was concluded that bacteroid differentiation is controlled by the plant, and that bacteroids are terminally differentiated in at Ieast some indeterminate nodules. Although these experiments are consistent with the hypothesis of terminal differentiation, it is still possible that there are some conditions where the bacteroids in question can become viable. It was suggested by Oke and Long (1999) that a system mimicking a natulal environment where plant nodules would degrade may be more appropriate than attempting to plate bacteroids on lab media. Furthermore, it remains to be seen whether these results are common to indeterminate nodules in general, or whether they are specific to the hosts tested. CI IAPTF:II 1. I N'I'RODUCTION

S. meliloti carbon metabolism

S. meliloti is a versatile organism with the ability to survive in the soil, coloni ze thre rhizosphere and differentiate into a symbiotic form. In order to do so, it has evolved an adapt- able and complex metabolic system which allows it to take advantage of many sources of energy. Metabolically, Rhizobia have been divided into two groups - fast growers, and slow growers (Stowers 1985). Fast growing species such as S. meliloti and R. Iegum- i,nosarum share many metabolic phenotypes, including the ability to utilize many car- bohydrates in the free-living state. Although glycolysis is commonly performed by the

Embden-Meyerhof-Parnas (EMP) pathway in many organisms, such as E. coli (Fraenkel i9B7), it may not be as ubiquitous as previously thought. A recent paper surveying glucose catabolism in a number of organisms compared the pathways used by the model systems E. coli and Bacillus subtilis to those of a group of less well characterized, phylo- genetically distinct organisms (Fuhrer et al. 2005). Based on the fact that only B. subtilis and .t-. coli used the EMP pathway, the authors suggest that the metabolism of these organisms may not be representative. In fact, the Entner-Doudoroff (ED) pathway may be more ancient than the EMP pathway and is widely distributed among diverse microor- ganisms (Conway 1992, Romano and Conway 1996).

Hexose metabolism in S. meliloti and other fast growing rhizobium species proceeds differently than in E. coli. Although all of the genes for the EMP pathway except phosphofructokinase (PFK) are predicted to be present in the S. meliloti genome (Capela et al. 2001), experimental analysis indicates that S. meliloti does not use the EMP pathway in a catabolic manner. Instead, the EMP pathway proceeds in a gluconeogenic direction and hexose catabolism is via the ED pathway. See Figure 1.1 for an overview of lrexose catabolism in S. meliloti. C]IAPTER I. I NTRODUCTION

Figure 1.I: Schematic of S. meliloti hexose metabolism including EMP and ED pathways. Arrows represent the direction of carbon flow as currently described in the literature, see text for details. Listed enzlirnes correspond to circled numbers on figure. publications describing relevant mutations are in brackets: I) Glucose kinase; 2) Glucose- 6-phosphate dehydrogenase (Cerveñans\f and A¡ias 1984, Barra et al. 2003); 3) 6- Phosphogluconolactonase; 4) 6-Phosphogluconate dehydratase; 5) 2-keto-3-deoxy-6- phosphogluconate aldolase; 6) Triose phosphate isomerase; 7) Fructose-bisphosphate aldolase; 8) Fructose bisphosphatase; 9) Phosphoglucose isomerase (A¡ias et al. lg79); 10) Glyceraldehyde 3-phosphate dehydrogenase (Finan et al. I9BB, t99t); tt) phosphoglycerate kinase (Finan et al. l99l); I2) Phosphoglycerate mutase; t3) Enolase (Finan et al. I9BB, l99t); 14) Pyruvate kinase; 15) Sorbitol dehydrogenase; 16) Mannitol dehydrogenase; l7) Fructokinase (Gardiol et al. 1980); lB) Glucose dehydrogenase (Gosselin et al. 2001); 19) Gluconate kinase.

(continued on next page) CHAPI-ER ]. INT'RODUCTION

Figure l.l, continued. filr.illtñ**l l@ i tcilr',**=l - gluconate fF--r-,"*-.l . i I O @ ' '\ iglucose6-phosphate 6-phosphogluconolactone 1@ À,-tructose 1@ to f- - Q/ 6-phosphate 6-phosphog I uconate Mannitol---- I --t I 1@ t@ 2-keto-3-deoxy-6- fructose 1,6-bisphosphate phosphog luconate

d i hyd roxyacetone glyceraldehyde phosphate 3-phosphate A @ 1@ 1,3-bisphosphog lycerate i@ 3-phosphoglycerate l@ 2-phosphoglycerate l@ phosphoenolpyruvate

TCA cycle CIIAPTER 1. INTRODUCTION

Phosphofructokinase (PFK) activity has not been detected in enzyme assays of S. meliloti cell-free extracts (Arias et al. 1979, Irigoyen et al. 1990) and the genome does not possess the traditional AlP-dependent pft gene (Capela et al. 2001). Enzyme assays have shonm that fructose-bisphosphate aldolase activity (FBA) is expressed at similar

Ievels during growth with glucose or a gluconeogenic carbon source such as succinate while the ED pathway is highly induced during growth with glucose (Finan et al. 1988,

Irigoyen et al. 1990). Assays of the EMP enzyme phosphoglucose isomerase suggest that it is also constitutively expressed under growth conditions with either glucose or gluconeogenic carbon sources (Arias et al. 1979). The fact that PFK is missing and that the

EMP pathway genes are not differentially induced by hexoses is consistent with the idea that they play a gluconeogenic role. In contrast to the EMP genes, analysis of the S. meliloli transcriptome using microarrays shows that the ED encoding genes are among the most significantly expressed genes during growth with glucose (Barnett et al. 2004).

In addition to enzyrnatic and expression studies, mutations in a number of the EMP encoding genes have been isolated. Mutants deficient in phosphoglucose isomerase

(A¡ias et al. 1979), enolase and glyceraldehyde-3-phosphate dehydrogenase (GAP) (Fi- nan et al. l9BB, l99I) were all able to growwith glucose as a sole carbon source, but the enolase (eno) and gap mtJtants were not able to grow with succinate. Therefore, the genetic evidence is also consistent with EMP functioning as a gluconeogenic rather than a catabolic pathway.

The ED pathway is a widely used alternative to the EMP pathway for hexose catabolism (Conway I992). \Mhen growingwith glucose, the first step is the conversion of glucose to 6-phosphogluconate, which is catalyzed by the enzyrnes glucose kinase, glucose

6-phosphate dehydrogenase, and phosphogluconolactonase. 6-phosphogluconate de-

10 CI.TAPTER 1. I N'TRODUCTION hydratase is one of the two key enzyrnes of the ED pathway, and catalyzes the conversion from 6-phosphogluconate to 2-keto-3-deoxy-6-phosphogluconate (I(DPG). The second key ED enzyrne, KDPG aldolase, catalyzes the cleavage of I(DPG to pyruvate and glyceraldehyde 3-phosphate (G3P). Several modes of ED operation have been observed in diflèrent organisms: a linear mode; a cyclic mode; and modified versions of the ED pathway (Conway 1992). The linear ED pathway is observed in a number of organisms, including Neisseria gonorrheae and .E colí.In the linear ED pathway, further catabolism of

G3P to pyruvate is performed by the lower portion of the EMP pathway, and the resulting pyruvate has different metabolic fates, depending on the organism. In N. gonorrheae, the pyruvate is simply converted to acetate. E. coli uses the linear ED mode as a periph- eral pathway for gluconate catabolism resulting in pyruvate which is fermented or aero- bicallyoxidizedinthe tricarboxylic acid (TCA) cycle. The cyclic mode of ED metabolism is employed by S. nteliloti and pseudomonads. In this mode, EMP does not act gly- colytically, and some of the G3P resulting from hexose catabolism is cycled back up to

6-phosphogluconate via gluconeogenic enzlirnes.

Some evidence for the cyclic operation of the ED pathway by mutational studies has been supplied by the fact tlnat eno and gap mutants are able to growwith glucose (Finan et al. 19BB). Such a block requires that G3P produced by IOPG aldolase is converted back up to glucose to prevent a build up of G3P and DFIAP Additional evidence for cyclic metabolism is supplied by experiments using labeled substrates. r3C-NMR studies of in uiuo carbon labeling has given data that is cclnsistent with catabolism of glucose via

ED resulting in G3P and pyruvate usingr3C labeled glucose (Portais et al. 1999). The results indicate that pyruvate is oxidized in the TCA cycle and that G3P is cycled back up to glucose-6-phosphate in addition to being catabolized to pyruvate as previously

11 C]]APTER 1. INTRODUCTION suggested by Finan et al. (1988). A further study using r:ìC labeled fructose confirmed that ED cycling also occurs for fructose (Gosselin et al. 200I). Analysis of metabolic flux ratios using r3C labeled glucose has also confirmed that glucose is catabolized through the ED pathway, however, the researchers were unable to determine how much G3P is cycled back through the ED pathway (Fuhrer et al. 2005).

In S. meliloúi glucose can be oxidized by the gluconate bypass which consists of a periplasmic pyrrolquinoline quinone-dependent glucose dehydrogenase (GDH) and gluconate kinase (Figure 1.1) (Bernardelli et al.2001, Gosselin et al.2001, Portais et al.

1999). Although this pathway exists, glucose dehydrogenase mutants did not have a slower growth rate than wild fype when grown with glucose as a sole carbon source

(Gosselin et al. 2001). Carbon labeling patterns observed in the glucose dehydrogenase mutant and wild type gro\m with r:rc glucose both showed that the cyclic ED pathway was in operation, indicating that further catabolism of gluconate is by the ED pathway rather than further extracellular oxidation (Gosselin et al. 200I). Growth experiments in chemostat cultures containing glucose and succinate indicate that glucose is pref- erentially oxidized while gluconate accumulates in the culture (Bernardelli et al. 2001).

The gluconate was only catabolized after glucose became limiting. Interestingly, GDH expression was found to be induced under phosphate limiting conditions (Bernardelli et al. 2001). Taken together, these observations suggest that the gluconate bypass provides an advantage under phosphate limiting conditions where gluconate acidifies the extracellular environment to solubilize inorganic phosphate (Bernardelli et al. 2001).

Given that cyclic ED metabolism has been shor¡¡n for growth with hexoses, it is interesting to consider how glycerol is metabolized. Glycerol catabolism has been studied in

Bradyrhizobium japonicum and R. leguminosarum, where it was found that glycerol is

t2 CI1A PI'ER 1. I NTRODUCTION catabolized by glycerol kinase and ø-glycerolphosphate dehydrogenase to DFIAP (Arias and Martinez-Drets 1976). Although glycerol catabolism has not been well studied in

S. meliloti, it is assumed to be the same as in R. Ieguminosarum,based on genetic similarity. The current literature suggests that glycerol is converted to G3P and that some is cycled through the ED pathway, while the rest is catabolized to pyruvate. The fact that eno and gap mutants are able to grow slowly with glycerol as a sole carbon source, indicates that the lower half of EMP is not required but is beneficial, which supports this model (Finan et al. 19BB).

S. meliloti phosphoglucose isomerase mutants have been shornm to grow with a number of carbon sources including glucose, glycerol, arabinose and succinate, but not with fructose, mannose, sorbitol, or mannitol (Arias et al. 1979). This is consistent with a cyclic ED metabolism. The operon required for fructose transport and possibly catabolism has been recently identified in S. meliloti (Lambert et al. 2001). It has previously been shown that fructose kinase is required for fructose catabolism via the

ED pathway (Gardiol et al. l9B0). Mannitol dehydrogenase and sorbitol dehydrogenase activities were both measured in S. meliloti and are hypothesized to catabolize mannitol and sorbitol by conversion to fructose, which is consistent with the fact that phosphoglucose isomerase mutants are unable to utilize these sugars as sole carbon sources

(Gardiol et al. 1980). Similarly, mannose is converted to fructose-6-phosphate by the enzymes mannokinase and mannosephosphate isomerase (Arias et al. l982). A mutation in glucose-6-phosphate dehydrogenase caused a loss of the ability to growwith glucose, fructose, mannose, mannitol, sorbitol, sucrose, maltose, ribose and xylose, but not the gluconeogenic carbon source arabinose (Cerveñansklf andArias 1984, Barra et al. 2003).

Taken together, the genetic evidence suggests that the hexose sugars and sugar alcohols

I3 CI]APTER ]. INTRODUCTION are generally catabolized through the ED pathway, which is expected due to the absence of PI.-K in S. meliloti.

The catabolism of galactose is an interesting exception to the general pattern of hexose catabolism by the ED pathway. The De Ley-Doudoroff pathway of galactose catabolism was first discovered in Pseudomonas saccharophila (Figure 1.2) (De Ley and

Doudoroff 1957, Lessie and Phibbs l9B4). Based on the observation that a glucose-6- phosphate dehydrogenase mutant was unable to grow with most hexoses, but was able to grow with galactose, it was hypothesized that the De Ley-Doudoroff pathway was used in S. meliloti (Cerveñanskli and Arias l9B4). A biochemical analysis of wild-type

S. nteLiloti tested for the presence of four out of five of the enzyrnes required for the De

Ley-Doudoroff pathway (Arias and Cerveñansky 1986). S. meliloti cell-free extracts were found to contain the assayed enzyrnes, consistent with De Ley-Doudoroff pathway operation. A chemically induced mutant unable to grow with galactose was missing one of the enzyrne activities, which explained its inability to utilize galactose. The biochemical and genetic evidence supports the hypothesis that galactose is catabolized via the

De Ley-Doudoroff pathway in S. melilorl, however, none of the genetic determinants involved have yet been identified.

The pentose phosphate (PP) pathway is divided into the oxidative- and the non- oxidative-pentose phosphate pathways (Figure 1.3). In the oxidative pathway, hexoses are converted into ribulose-S-phosphate. The key PP enzyme 6-phosphogluconate dehydrogenase (PGDH) is able to decarboxylate 6-phosphogluconate resulting in ribulose-

5-phosphate. PGDH activityhasbeenfoundinanumberof thefastgrowingrhizobia, including S. meliloti (Martínez-De Drets and Arias 1972). A study of S. meliloti enzlmle activities has found that PGDH activityis found when grown with either hexoses or gluco-

I4 CITAPTER 1. INTRODUCTION

Figure I .2: Schematic of S. meliloti galactose catabolism via the De Ley-Doudoroff pathway. Listed enzymes correspond to circled numbers on figure: 1) Galactose dehydrogenase; 2) y-Lactonase; 3) Galactonate dehydratase; 4) 2-keto-3-deoxy-galactonokinase; 5) 2-keto-3 -deoxy-6-phosphogalactonate aldolase. D-ga lactoselo D-g a I a ctono-y-la ctone @ D-ga lacto nat

I o 2 - keto-3-d eoxy-g a I a cto n ate ATP

ADP

2-keto-3-d eoxy-phosphoga I acton ate

g lycera ldehyde- pyruvate 3-phosphate

t5 CI'IAPTER 1. ]NTRODUCTION neogenic carbon sources suggesting that it is constitutively active (Irigoyen et al. 1990).

The non-oxidative PP pathway is ¡eversible and can interconvert molecules of F6P and

G3P with ribulose-5-phosphate. Complicated labeling patterns were observed during growth with either I3C-NMR glucose or fructose, consistent with the reversible action of the non-oxidative PP pathway (Gosselin et al. 2001, Portais et al. tg99). Metabolic flux ratio modeling has sho',nm that although glucose catabolism is primarily catalyzed by the ED pathway, smaller amounts of carbon flow occurs through the oxidative and non-oxidative branches of the PP pathway to supply pentose-phosphate intermediates for biosynthesis (Fuhrer et al. 2005).

Although a surveyof the literature indicates that the PP pathwaygenes have not been characterized in S. meliloti, the genome contains: three predicted transketolase genes; a predicted transaldolase; two predicted ribose-S-phosphate isomerases; a probable ribulose-5-phosphate epimerase; and a probable G-phosphogluconate dehydrogenase

(Galibert et at. 2001). This is very similar to the situation described for E. coli which is knor¡¡n to possess two ribose-5-phosphate isomerases and two transketolases (Sprenger

1995). Two of the putative S. meliloti pentose phosphate enzyrnes (a probable transketolase and ribose 5-phosphate isomerase) were found to be expressed under free-living conditions and in bacteroids using a proteomic approach (Djordjevic 2004). Further research is needed to positively identifu the PP genes and to study their regulation.

16 s ! rrl Figure 1.3: Schematic of S. meliloti PP pathway. Listed enzyrnes correspond to circled numbers on figure. Publi- :* cations describing relevant mutations are in brackets: l) Glucose kinase; 2) Glucose 6-phosphate dehydrogenase (Cerveñanskf and Arias 1984, Barra et al. 2003); 3) 6-Phosphogluconolactonase; 4) 6-Phosphogluconate dehydro- t¿ genase; 5) Ribose 5-phosphate isomerase; 6) Ribulose 5-phosphate epimerase; 7) Transketolase; B) Transaldolase; o 9) Triose phosphate isomerase; 10) Fructose-bisphosphate aldolase; ll) Fructose 1,6-bisphosphatase; 12) Phospho- Õ glucose isomerase (Arias et al. 1979). I o{ z

qlucosetOô \-/ Ë-Ë nólinate å 6-phosphog I ucono lacto ne t@ 6-phosphog luconate ... phate t@ '1 ribulose-5-phosphate Entner-Doudoroff pathway glyceraldehyde sedoheptulose t1 3-phosphate 6-phosphate 7-phosphate S_phosphare@r1 ll@ xylulose 5-phosphate xylulose erythrose glyceraldehyde 5-phosphate 4-phosphate 3-phosphate CI IAPT'ER 1. INTRODUCTION

S. meliloti is knor¡¡n to contain a complete TCA cycle (Stowers 1985, Dunn i99B).

All the enzyrnes of the TCA cycle have been sho'¡m to be expressed under lab growth conditions, and in s)¡mbiotic association (Djordjevic 2004). Mutations have been generated in the S. nzeliloti genes encoding the following TCA cycle enzymes: citrate s1m- thase (CS) (Mortimer et al. lgg9); isocitrate dehydrogenase (ICD) (McDermott and Kahn

1992); a-ketoglutarate dehydrogenase (KGD) (Duncan and Fraenkel 1979); succinate dehydrogenase (SDH) (Gardiol et al. I9B2); and malate dehydrogenase (MDH) (Dymov et al. 2004). Mutations have also been isolated in the anaplerotic enzyme pyruvate carboxylase (Dunn et al. 2001), the gluconeogenic enzyrnes PCK (Finan et al. l9BB, i991,

Østerås et al. 1995, 1997b) and two different malic enzyrnes (Driscoll and Finan 1993,

1996, 1997). SeeFigure l.4foranoverviewof the S.meliloti TCAcycle.

L-Arabinose is catabolized to oc-ketoglutarate in some bacteria, including S. melilori (Duncan 1979, Duncan and Fraenkel 1979). Mutations in the TCA enzymes KGD,

SDI-I, and MDH were found to be non-permissive for arabinose growth as would be expected (Duncan and Fraenkel 1979, Dymov et al. 2004, Gardiol et al. 1982). Succinate, malate and fumarate are all directly catabolized through the TCA cycle and are important carbon sources during symbiosis. Dicarboxylate catabolism and transport, and its syrnbiotic relevance is discussed in further detail in the next section.

Growth on carbon sources such as acetate, arabinose, malate or succinate requires an enzlirne to allow gluconeogenesis from a TCA intermediate. In S. meliloti and other rhizobia this is performed by the enzyrne phosphoenolpyruvate carboxykinase (PCK) which catalyzes the conversion of oxaloacetate to phosphoenolpyruvate (Dunn l998).

S. metiloti mutants deficient in PCK have been shor¡¡n to grow very slowly with gluconeogenic carbon sources (Finan et al. lgBB). In addition, PCK is encoded by the

1B CI-IAPTER 1. INTRODUCTION

Irigure I.4: Schematic of S. meliloti TCA cycle operation. Listed enzyrnes correspond to circled numbers on figure. Publications describing relevant mutations are in brackets: 1) Pyruvate kinase; 2) þruvate dehydrogenase; 3) Citrate synthase (Mortimer et al. 1999); 4) Aconitase; 5) Isocitrate dehydrogenase (McDermott and Kahn 1992); 6) ø- ketoglutalate dehydrogenase (Duncan and Fraenkel 1979); 7) Succinyl-CoA slmthetase; B) Succinate dehydrogenase (Gardiol et al. t9B2); 9) Fumarase; l0) Malate dehydrogenase (D)¡mov et aI.2004); II) Isocitrate lyase; l2) Malate synthase (García-de los San- tos et aL.2002); 13) NAD+- and NADP+-dependent malic en4rynes (Driscoll and Fi- nan 1993, 1996, 1997); t4) Phosphoenolpyruvate carboxykinase (Østerås et al. 1995, I997b); l5) Pyruvate carboxylase (Dunn et al. 2001); 16) Pyruvate orthophosphate dikinase (Østerås et al. 1997a).

/., phosphoe n o I py ru vate ( @1 lo @vo"i"'i@ (\ @l acetYr-coA "\\""u,ou.1".u,"h citrate \þ ò , isocitrate \ @l I a-ketoglutarate

succinate succinyl-CoA

19 CHAP'TER 1. I NI'IIODUCTION gene pckA which is only induced during growth with a gluconeogenic carbon source

(Østerås et al. 1995). Suppressors were isolated which grew normally on succinate, and further characterization of the suppression event revealed that it was caused by an up- regulation of pyruvate orthophosphate dikinase (POD) activity (Østerås et al. tggTa).

Mutations in pocl alone caused no apparent phenolype, however, pckA pod mutants were completely unable to grow with gluconeogenic carbon sources.

S. melíloti has both NAD+- and NADP+-dependent malic enzyrnes (DME and TME, respectively) which convert malate to pyruvate (Driscoll and Finan 1993, 1996, 1997,

Mitsch er aL.2007). Although neither of these enzyrnes is required for growth with gluconeogenic carbon sources in the free living state, dme mutants do not form a functional symbiosis (Driscoll and Finan 1993). Presumably, gluconeogenesis in the absence of

PCK activity could occur via POD and DME. This appears to be relatively insignificant in free-living wild-lype S. meliloti as mutationsin pod, dme and tme alone have no known phenotypes, and pck{mutants grow only very slowly on succinate. However, dmewas found to be required for gluconeogenesis in the pckA mutant with up-regulated pod mentioned above (Østerås et al. 1997a). Both dme gene expression and DME enzyrne activity are expressed at similar levels during growth with gluconeogenic or glycolytic carbon sources (Driscoll and Finan 1997). DME activity is present at low levels in bacteroids (Driscoll and Finan I996). TME activity is very low in bacteroids and tme was found to be repressed in bacteroids as measured by [3-galactosidase fusions (Driscoll and Finan I996, 1997). Taken together, the physiological role of DME is likely to re- plenish acetyl-CoA used in the TCA cycle when growing with dicarboxylates in the bacteroid state (Driscoll and Finan 1996, 1997). This requirement is consistent with the fact that PCK activity has not been detected in bacteroids (Finan et al. I99I). The carbon

20 CI1APTER ]. INTRODUCTION metabolism of bacteroids is discussed in greater detail in the next section.

Growth with single carbon sources requires anaplerotic reactions to keep the TCA cycle filled with intermediates as they are removed for the synthesis of amino acids and gluconeogenesis (Dunn 1998). The enzyrne pyruvate carboxylase (PYC) has been sho'¡¡n to be required for growth with glucose and pyruvate, but not succinate in S. meliloti

(Dunn et al. 2001). PYC is required to ensure that TCA cycle intermediates are not de- pleted by catalyzing the conversion of pyruvate to oxaloacetate. Growth with acetate as a carbon source also requires an anaplerotic pathway to provide TCA cycle intermediates. In S. meliloti this is done by the glyoxylate shunt (Figure 1.4). The glyoxylate shunt enzyrnes isocitrate lyase and malate slmthase have been detected and are expressed during growth with acetate (Duncan and Fraenkel 1979). Consistent with this, malate synthase has been shor¡¡n to be required for growth with acetate as a sole carbon source in

IL. leguminosarum (García-de los Santos et al.2002). Interestingly, malate synthase was found to be highly inducible by the gluconeogenic carbon source, arabinose, but not by acetate.

Bacteroid metabolism

Very little is knor¡'¡n about the carbon and energy sources that are available to S. meliloti duringinfection thread development. On the other hand, the details of bacteroid metabolism in the nodule have been widely studied. Symbiotic nitrogen fixation requires that large amounts of energy are supplied to the nitrogenase complex. The fact that both the differentiated bacteria and the plant retain functional memb¡ane systems indicates that any carbon supplied for nitrogen fixation must be both transported by the bacteroid and by the plant derived peribacteroid membrane. The carbon sources have

21 CTTAPTER 1. IN'|RODUCTION been studied primarily using bacterial mutants and by isolating root nodule preparations.

Dicarboxylic acids are widely accepted as the primary carbon sources available to rhizobium bacteroids (Poole and Allaway 2000, Lodwig and Poole 2003, Yurgel and Kahn

2004). With few exceptions, sugar catabolism mutants of various rhizobium species have been sho'¡¡n to form functional symbioses (Lodwig and Poole 2003). In addition to mutants unable to utilize individual sugars as sole carbon sources, even the metabolically compromised S. meliloli pyc mutant which cannot utilize any hexoses forms a functional symbiosis (Dunn et al. 2001). Other pleiotropic mutants unable to grow with numerous hexoses and pentoses were isolated which also formed normal nitrogen fixing nodules. One such mutant was found to be unable to transport glucose or galactose due to an insertion in a hypothetical catabolic activator protein (Hornez et al. lg94), and others had insertions in glucose-6-phosphate dehydrogenase (Cerveñansklf and Arias

1984, Barra et al. 2003). It is clear from the literature that the ability to use sugars is not critical for nodulation.

In contrast to sugar catabolism mutants, strains unable to utilize dicarboxylic acids are unable to form a functional symbiosis. The dicarboxylate transport system in S. melilo¿i consists of three genes: dctA, dctB and dctD. DctB and DctD regulate the expression of dctA by sensing dicarboxylates fYurgel and Kahn 2004). DctB acts as a dicarboxylate sensoL phosphorylating DctD, which in turn acts as a regulator responsible for activat- ing transcription of dctAvnth the help of the sigma factor RpoN (Giblin et al. lgg5, Lede- bur et al. 1990). DctA is a dicarboxylate permease which appears to be the sole transport system for malate, succinate, and fumarate in S. meliloti. Interestingly, dct mutants are genelally able to form normal looking nodules with bacteroids, howeve¡ they are un-

22 CHAPTEI| 1. I N'I'RO DUCTION able to fix nitrogen (Finan et al. 1983, l9BB, Jording et al. 1994, Long et al. l9BB, Watson et al. t9BB). A-lthough the ability to use dicarboxylates is required for syrnbiosis, this suggests that dicarboxylates are not necessary during infection and the early stages of nodule development.

Other mutations affecting succinate metabolism cause S. melíloti to become Fix-.

As mentioned in the previous section, a succinate dehydrogenase mutant was isolated which formed Fix- nodules (Gardiol et al. 1982). It is not knonm if this mutant is Fix- due to the inability to catabolize succinate, or the fact that the TCA cycle is disrupted.

TCA cycle mutations in CS, ICD, and MDH were also found to be Fix- (D)¡mov et al.

2004, McDermott and Kahn 1992, Mortimer et al. i999). Mutations affecting gluconeogenesis have also been found to affect symbiosis, which makes sense given that the primary carbon source supplied to bacteroids is dicarboxylates. Enolase, GAR and 3- phosphoglycerate kinase mutants were all found to be Fix-, while PCK mutants had a reduced rate of nitrogen fixation (Finan et al. 1991). As discussed in the previous section, mutations in DME have a Fix- phenotl4pe (Driscoll and Finan 1993). Even though

DME and TME have the same biochemical activity and can functionally replace each otlrer in uiuo, overexpressed tmewas not able to complement dme mutants for the ability to form a functional sl.rnbiosis (Mitsch et al. 2007). The requirement for DME activity in s]¡mbiosis is not currently understood, but it is interesting to note that the E. coli

NAD+ dependent rnalic enzyrne was able to complement the symbiotic phenotype of

S. meliloti dme mutants. Although PCK activity was not detected in bacteroids (Finan et al. l99l ), the fact that pckA mutants have lowe¡ rates of nitrogen fixation suggests that there is some PCK activity present or that the presence of the protein is required. Taken together with the fact that dme mutants are Fix-, the data suggest that gluconeogene-

23 CI-IAPTER 1. INTRODUCTION sis and TCA cycle operation occur via both PCK and DME activities. The reason pckA mutants are able to fix nitrogen at low rates may be due to the fact that some sugars are supplied to the bacteroid.

Bacteroids were previously thought to fix nitrogen and simply export it as ammonium, but new research has revealed an interesting aspect of bacteroid r-netabolism. Mu- tations in the broad-specificity amino acid transporters aap and brawere found to cause

R. leguminosarLtm to become Fix* when both genes were disrupted (Lodwig et al. 2003).

\n/hen either single gene was disrupted, plants inoculated with the mutants were similar to wild type inoculated plants. Root xylem amides were measured from plants grown in the presence of rsN2; double aap bra mutant inoculated plants were found to have much more labeled asparagine than the wild t1pe. Based on these data, a new model was proposed: bacteroids secrete both ammonium and an amino acid precursor to asparagine, probably aspartate. In this model, plants synthesize asparagine using ammonium and aspartate secreted by the bacteroids. The plants then supply the bacteroids with glutamate, some of which is used in a transamination reaction with oxaloacetate to produce aspartate. The aspartate is then exported back to the plant and used as a precursor for asparagine synthesis. Based on the model, the authors proposed two predictions: that mutations in the aspartate transaminase enzyrne antAwould be Fix-; and that amino acid export mutants would accumulate too much carbon due to the lack of carbon removal via transamination and amino acid export. These predictions appear to be correct. First, aalA mutants are Fix-. It was previously shovrm that antA mutants of

S. nzeliloti were Fix- (Watson and Rastogi 1993, Rastogi and Watson l99l), and this was subsequently confirmed for R. leguminosarum (Lodwig et al. 2003). Secondly, plants inoculated with aap bramvtants were found to contain higher levels of plant starch and

24 CHAP'IER I. IN:/]RO DUCTION the bacteroids themselves contained granules of the storage polyrner, polyhydroxybutyrate (PIIB), which the wild type do not (Lodwig et al. 2003). One of the benefits of this updated model is that it helps understand how such a complicated mutualism between plants and bacteria could evolve. Lodwig et al. (2003) suggest that this system prevents either the plant or the bacteria from dominating the relationship because the bacteria must fix and export nitrogen in order to receive amino acids synthesized by the plant, and the plant depends on the bacteroids for ammonium and aspartate synthesis.

Competition for nodule occupancy

Research on the ability of S. meliloll to form a symbiotic relationship with plants has resulted in the generation and testing of many different mutant strains for the ability to form functional nodules. In addition to the identification of genes essential to s)¡mbiosis, there have been many reports of S. melilofl strains that are more or less competitive for nodule occupancy than their respective wild type. When co-inoculated with the wild type, these mutant strains are found in a significantly higher or lower proportion of the plant nodules compared to their presence in the original inoculum. There is a significant interest in what makes specific strains of rhizobium more or less competitive for nodulation than others. From the perspective of using rhizobia as fleld inocula, it would be desirable to engineer or select efficient nitrogen fixing strains so that they can outcom- pete less efficient and non-nitrogen fixing indigenous rhizobia (Triplett and Sadowsky

1992).

A number of rhizobia, including R. leguminosarum strains, produce narrow spec- trum antibiotics called bacteriocins (Oresnik et al. 1999, Triplett and Sadowsky 1992,

Venter et al. 2001). R. Ieguminosarumstrains can produce several different bacteriocins

25 CÍTAP'TER I. INI'RODUCTION arìd some but not all of them have been shonm to influence competition in the rhizosphere by inhibiting the growth of sensitive strains (Oresnik et al. 1999, Robleto et al.

I997, 1998, Triplett and Barta I9B7). Because bacteriocin production can give an inoculant strain a competitive advantage, there is interest in using this property in agriculture. The genes for slmthesis and resistance to the bacteriocin called trifolitoxin have been transferred from a symbiotically ineffective strain of R. leguminosarumto effective strains of R. leguminosarum and R. ¿rll (Robleto et al. 1997, 1998, Triplett 1990). These recombinant tril'olitoxin producing strains have been tested and sho'¡¡n to be more competitive in sterile conditions, non-sterile soil, and even in fields. Antibiotic production is currently among the easiest of determinants to understand which confer a competitive advantage for nodule occupancy. It will be interesting to see if this strategy becomes widely used in agricultural inoculant strains, given the resistance to the use of recombinant organisms.

Recently a link has been found between the abilityto compete for nodule occupancy and salt tolerance in S. melilori (Miller-Williams et al. 2006). Four mutants with a significantly reduced ability to compete were isolated from a screen for mutants unable to grow with higher concentrations of sodium chloride. The disrupted genes have not yet been characterized in S. melilofi, but they are hypothesized to play roles in protein export, membrane and cell wall maintenance. it is not currently kno'¡¡n if the decreased ability to compete for nodule occupancy associated with these mutations is caused by salt sensitivfty or by other effects.

S. meliloti produces two different exopolysaccharides (EPS): EPS I and EPS II. It is well knorn¡n that S. meliloti mutants unable to produce EPS i are sl.rnbiotically ineffective (Finan et al. 1985, Fraysse et al. 2003, Leigh et al. i9B5). Plants forming indetermi-

26 CHAPTER I. INTRODUCT|ON nate nodules require bacterial EPS production for a functional symbiosis, while plants with determinate nodules do not (Gage 2004). However, EPS production mutants of the determinate nodule fbrming lhizobia R. leguminosarumbv. phaseoli, Rhizobium tropici and Bradyrhizobiunt japonicum are syrnbiotically effective, but less competitive than their wild-type strains (Triplett and Sadowsky f 992). A study of the determinate nodule forming strain, Rhizobium etli, isolated a mutation in the regulator rcsR, which caused altered cell surface (Araujo et al. 1994, Bittinger et al. lg97). Mutations in rosR were found to cause a decrease in the ability to compete with the wild-type parent strain. In- terestingly, RosR was sho'ur¡n to negatively regulate many diverse genes including a gene predicted to be involved in EPS production (Bittinger and Handelsman 2000). Mutations in the EPS production gene increased competitiveness compared to its nrsR parent suggesting that altered regulation of EPS production may be responsible for the decreased competitive ability of ¡ osR mutants (Bittinger and Handelsman 2000). Although EPS has been widely studied and a loss of the ability to produce EPS has been shoi¡¡n to prevent infection thread development in host plants with indeterminate nodules, the reason for this is not kno',n¡n. Several hlryotheses have been suggested including the idea that EPS plays a role in modulating the plant defense response and that EPS plays a structural role in infection thread development (Fraysse et al. 2003).

Motility has been shornm to play a role in nodulation in several studies of different rhizobia (Triplett and Sadowsky 1992). Several spontaneous non-motile and non- chemotactic mutants of. S. meliloti were isolated and found to be less competitive than wild type (Caetano-Anollés et al. 19BB). These mutants were all able to form functional symbioses and although they formed nodules at the same rate, there was a four day delay before the first nodules appeared. By plating roots incubated with motile, non-

27 CI]APTER ]. INTRODUCTION motile, and non-chemotactic strains, it was found that far fewer mutant bacteria adhered to the roots. This suggests that strains with altered motility are less competitive because they do not respond and move towards plant signals. Methyl-accepting chemotaxis proteins (MCP) are part of the prokaryotic motilitysystem, playing a role in sensing attractants and repellants as well as rnodulating chemotaxis (Falke et al. 1997). To test if MCPs are involved in competition for nodule occupancy, mutations were constructed in f'our different R. legumínosarumpredicted MCP encoding genes fYost et al. i99B). A1- though a motility phenotlpe was found for only one of the mutants, two of the mutants were found to be significantly less competitive than wild type for nodule occupancy, suggesting a possible link between competitiveness and motiliry. The deletion of a predicted clremotaxis gene cluster was found to cause R. leguminosarutn to move straight forward with no changes in direction (Miller et aI.2007). This mutant strain was also found to be less competitive for nodule occupancy. These experimental results are all consistent with the idea that movement towards plant signals and adherence to the root are critical steps in the early stages of symbiosis (Gage 2004).

Plasmid cured derivatives of R. leguminosarum were found to reach lower growth levels in the rhizosphere compared with the wild type (Moënne-Loccoz and Weaver i995). Because the megaplasmids encode genes required for the catabolism of some carbon sources, it was hypothesized that the inability of plasmid cured strains to use rhizosphere carbon sources may be linked to the abiliry to grow in the rhizosphere and compete for nodule occupancy (Oresnik et al. l99B). A complementation of plasmid cured strains and subsequent mutageneses of the isolated cosmids identified genes involved in several metabolic pathways including the catabolism of rhamnose. Mutants unable to utilize rhamnose were found to have a severe competitive defect compared

2B CI \APTEII ]. ] NTRODUCTION with the wild type (Oresnik et al. l99B). Quantification of the bacteria adhered to plant seedlings indicated that the rhamnose catabolism mutants do not grow at a different rate in the rhizosphere under experimental conditions. It was therefore hypothesized that rhamnose catabolism mutants may be unable to grow at the same rate as wild type during infection thread development, resulting in more wild type occupied nodules being fbrmed. F'urther characterization will be required to understand this phenotype.

Similar to rhamnose, plasmids encode the genes for erythritol catabolism in R. legum- inosrtru¿m and it was shornm that mutants unable to catabolize erythritol are less com, petitive than wild type fYost et al. 2006). These results also require further work to provide an explanation.

Rhizopines are substrates slmthesizedby bacteroids and secreted into the soil environment where they are catabolized by free-living bacteria (Murphy et al. lg95). Not all rhizobia produce rhizopines, and different strains produce unique variations of them.

Based on current surveys, it is hypothesized that bacteroids produce specific rhizopines that are catabolized only by their free-living counterparts; thereby providing them with an aclvantage in the soil (Murphy et al. 1995). S. melíloti mutants unable to synthesize rhizopines were found to be no less competitive than the wild type (Gordon et al. 1996).

I-Iowever, mutants unable to catabolize rhizopines were less competitive than the wild type for nodule occupancy. Gordon et al. (1996) predicted that the ratio of wild type to rhizopine mutants would gradually increase as nodules were formed and started producing rhizopines, giving the wild type a growth advantage in the rhizosphere. Although the wild type had a competitive advantage over the rhizopine catabolism strain, it was almost irnrnediate rather than gradual, and did not increase with the number of nodules.

The reason for this competitive difference is therefore not fully understood and requires

29 CI-|AP'TER I. I N'TRODUCT]ON more research. Because rhizopines are derived from myo-inositol which is found in the soil and some nodules, researchers have also tested myo-inositol catabolism mutants for competitive defects (Fry et al. 200i). Mutants of both R. leguminosctrumand S. fi'edii unable to catabolize myo-inositol were found to have a severe competitive defect (Fry et al. 2001, Iiang et al. 2001). Fry et al. (2001) speculate that myo-inositol catabolism may be required in early infection thread development, but suggest that more work is needed to understand this effect.

S. meliloti mutants unable to catabolize proline were found to have a lesion in the gene putA. The gene putAencodes a proline dehydrogenase, and was found to be associated with a competitive defect when coinoculated with a wild-type parent (Iiménez-

gg5) Zurdo et al. I . Interestingly, during complementation analysis of the putA mutant , it was found that complemented strains with a cosmid carrying putA did not simply restore competitiveness, but became more competitive than wild type. It was hypothesized that increased expression of the cosmid borne putAwas responsible for the increase in competitiveness. Experiments to increase the expression of putAwere performed by cloning putAinto a plasmid with a constitutive promoter (van Dillewijn et al.

2001). The authors hypothesized that proline is secreted by plant roots during drought stress as alfalfa and soybean nodules accumulate proline during drought conditions.

Tlre result of overexpressed putA was therefore tested in greenhouse trials of plants coinoculated with the parent strain and watered either normally or under simulated drouglrt conditions. The overexpressed putAstrainwas not found to have a competitive advantage or disadvantage in the greenhouse under normal conditions, however, when grown under simulated drought conditions, 90% of root nodules contained the overexpressed putAstrain after one month. After another month passed, it was found that the

30 CIIAPTEII I. INTRODUCTION nodule occupancy level of the parent strain reached the same level as the strain overexpressing putA. Similar results were found by inoculating a different strain overexpressing putAover the course of a field trial (van Dillewijn et al.200l).

S. meliloti and other indeterminate nodule forming rhizobia accumulate PHB gran- ulcs during normal growth, which disappear during bacteroid differentiation (Lodwig et al. 200s, Tbmbolini and Nuti I9B9). This has led to the hyporhesis that bacreroids use the stored PHB as an energy source to drive differentiation (Prell and Poole 2006). An alternative hy¡pothesis is that the ability to store a carbon source while free-living helps fuel growth in the infection thread (Charles et al. 1997). Although the exacr role of pHB in synrbiosis is not knoum, S. meliloti mutants unable to store or catabolize PHB have been reported to be less competitive than wild type for nodule occupancy (Aneja et al.

2005, Willis and Walker l99B). The situation is complicated by the fact that the mutant unable to sy'nthesize PHB has a more severe competitive disadvantage than the PHB catabolism mutant (Aneja et al. 2005). It has therefore been suggested that the abitity to remove inhibitory metabolic intermediates by forming PHB may be more important than the ability to utilize PHB as an energy reserve (Aneja et al. 2005). The analysis of these phenotypes is also difficult because rhizobia can accumulate glycogen in addition to PHB (Lodwig et al. 2005), and it is possible that some phenotypes are strain or host dependent. Consequently, the decreased competitive ability associated with PHB mutants is not fully understood and will require further investigation.

Although a number of mutations have been isolated and associated with decreases in competitive abiliry the reasons for these competitive defects are often not understood. Some of them have only been demonstrated in a single strain of rhizobium, and it would be interesting to know if homologous mutations in other species cause similar

31 CHAPTER ]. ] N RODUCTION competitive defects. Several resea¡chers have suggested the hypothesis that the abitity to utilize specific carbon sources may be an advantage during infection thread development, however, there is not yet enough evidence to support this. Ongoing research will be required to determine at which point in nodule developmenr the ability to utilize specific carbon sources is required.

Research goals

Initially, the S. meliloti pathway f'or catabolism of the methyl-pentose rhamnose was a candidate for investigation because it has been studied in R. leguminosarunt, where

Inutants have been shor,r¡n to be less competitive than wild type for nodule occupancy

(Oresnik et al. 1998, Richardson et aL 2004, Richardson and Oresnik2007).In addition, the pathway of rhamnose catabolism was thought to be different than in E. coli . A rhamnose kinase was also determined to play a unique role in rhamnose transport (Richard- son and Oresnik 2007). Several rhamnose mutants had alreadybeen generated in S. me- liLoti, but not characterized.It was hypothesized that rhamnose catabolism would proceed similarly in S. meliloti as in R. leguminosarum,thatit would require a homologous rhamnose kinase for rhamnose transport, and that mutants would be competitivelyde- fective. Part of the reason for studying rhamnose catabolism in S. melílofi is that there are nìote genetic tools available, which would give more flexibility in the type of experiments done.

\¡1/hile screening for S. meliloll rhamnose mutants, a mutant which grew slowly on rhamnose was isolated. Although the mutant strain grew more slowly on plates, single colonies grew and eventually reached wild-type colony sizes. The mutation was found to inferrupt a putative triose phosphate isomerase encoding locus (tpiAl). The mutant

JL.., CI-IAPT'ER ]. I NTIIODUCTION strain was tested for the ability to grow on a number of carbon sources, and was found to lack the ability to grow with glycerol, in addition to its rhamnose growth phenotype.

The genome of S. meliloti predicts a second TPI encoding gene (tpiA2).It was hypothesized that mutations in tpiAl would cause a competitive defect due to the slower rate of growth with rhamnose as a sole carbon source. It was also hypothesized that both genes would encode functional triose phosphate isomerases and that tpiA2 wou\d be inducible by erythritol based on its proximity to an operon predicted to be involved in erythritol catabolism. Experiments to characterize these two genes and test these hypotheses were carried out.

I also decided to investigate the genetics of arabinose catabolism. The catabolism c¡f arabinose to ø-ketoglutarate was initially an interesting pathway because none of the genes required have been identified in S. meliloti or any other organism using the same pathway. Furthermore, an arabinose catabolism mutant was identified as having a compromised ability to form a symbiotic relationship with alfalfa (Duncan 198I). Therefore, it was initiallyhypothesized that arabinose catabolism mutants would be symbiotically ineffective, or at least less competitive for nodule occupancy than wild ty¡pe. It was also hypothesized that the genes for arabinose catabolism would be located on the chromosome, as the megaplasmids pSymA (Oresnik et al. 2000) and most of pS¡rmB (Charles and Finan l99l) have been deleted, and the resulting mutants have been screened for the ability to grow with many carbon sources, including arabinose.

.-).-) Chapter 2 The abitity to cataboltze rhamnose affects competition for nodule occupancy in Sinorhizobium meliloti

Summary

The work in this chapter is a collaboration with Jason Richardson. Nathan Poysti per- forrned mutagenesis, identification of all mutations, complementation, all symbiosis and nodulation competition experiments, and some of the rhamnose uptake assays.

Rhizobium leguminosarummutants unable to use rhamnose as a sole carbon source were found to be less competitive for nodule occupancy. In an effort to determine if

Sinorhizobium melilo¿i would be similarly affected, mutants unable to use rhamnose were generated. The S. nteliloti mutations affecting rhamnose utilization were isolated in operons homologous to those of the previouslycharacterized R. legumínosa.rummu- tants. Although similar, there are some genetic differences between the organisms, as evidenced by complementation experiments. Unlike R. leguminosarum, the ABC-type transporter genes associated with the S. meliloti ¡hamnose utilization locus are clearly not the sole rhamnose transporter in S. meliloti as measured by [3H]rhamnose uptake assays. Despite the differences, we have confirmed that S. meliloti mutants unable to grow with rhamnose are less competitive than wild type for nodule occupancy. C]TAPTER 2. RI-IAMNOSE CATABOLISM AND TRANSPORT

Introduction

The use ol syrnbiotic rhizobia for biological nitrogen fixation is very important in the fìeld of agriculture. In addition to using strains that are efficient at fixing nitrogen, they must form a significant number of nodules on the desired host plants, often compet- ing with indigenous rhizobia. Factors affecting whethe¡ or not a given strain can effectively compete for nodule occupancy include: the quantity of viable bacteria delivered to the plant; the quantity and character of indigenous rhizobium populations; environmental factors; and rhizobial genetic factors affecting competitive ability (Dowling and

Iìroughton 1986, Triplett and Sadowsky 1992).

Plasmid encoded carbon catabolic loci have been identified by systematic deletions and plasmid curing experiments of R. leguminosarum and S. meliloti strains (Baldani et al. 1992, Charles et al. 1990, Charles and Finan 1991, Hlmes et al. 1986, 1989, Oresnik et al. 2000). Rhizobia cured of their plasmids have also been found to be less competitive than their respective wild type when grown together in the rhizosphere and for nodule occupancy (Brom et al. 1992, Moënne-Loccoz and Weaver 1995). Subsequent experi- rnents have linked these two phenotylpes together byshowing that mutations in plasmid encoded genes for the catabolism of myo-inositol, rhizopines, rhamnose, and erythritol cause a decrease in the ability to compete for nodule occupancy (Fry et al. 200I, Gordon et al. 1996, Jiang et al. 200I, Oresnik et al. lg98, Yost et al. 2006). Although the decreases in competitive ability associated with catabolic mutations are poorly understood, it has been suggested that nyo-inositol and rhamnose maybe supplied as energy sources during infection thread development, rendering mutants unable to use them less competitive for nodule occupancy (Fry et al. 2001, Oresnik et al. l99B). Furthermore, there has not been a systematic approach to determine if similar mutations cause competitive

35 C]-IAPTER 2. IIHAIVINOSE CATABOLISM AND TRANSPORT defects in other rhizobia. Of the mutations affecting competition mentioned here, only tlrose affecting myo-inositol catabolism have been tested in more than one species (Fry et aÌ.2001, Jiang et al. 2001).

It has been reported that loss of the ability to utilize rhamnose leads to an inability to compete for nodule occupancy in R. leguminosarumbv. triþlii (Oresnik et al. 1998).

Characterization of the rhamnose utilization genes revealed that they are organized into

2 ope rons containing g genes (Richardson et al. 2004). This work was initiated to address the hypothesis that the ability to catabolize rhamnose also plays a role in allowing S. meliloti to be competitive for nodule occupancy on alfalfa. Here we report the isolation of

S. nteliloti mutants unable to utilize rhamnose as a sole carbon source, their subsequent characterization, and the direct test of the hypothesis that the inability to utilize this sugar affects the ability to compete for nodule occupancy.

Materials and methods

Bacterial strains, plasmids, and media

The bacterial strains and plasmids used and generated in the work are listed in Table 2.1.

S. meliloti strains were routinelygrornm at 30"C on complex (either LB or TY) or defined

(\A4M) media (Sambrook et al. 1989, Richardson et al. 2004). \Mhen required, antibiotics were used at the following concentrations: tetracycline (Tc) either 5 or l0 Ég mL-t; neomycin (Nm), 200 pg ml-r; kanamycin (Kan), 20 'tg mL-r; streptomycin (Sm), 200 pgmL-r;rifampicin (Rf), 100 FgmL-r;chloramphenicol (Cm),20 LrBmL-r;gentamicin

(Gm) 20 or 50 pg mL-r. Growth was routinely monitored spectrophotometrically at 600 nm.

36 CHAPTER 2. RHAMNOSE CATABOLISM AND TRANSPORT

Table 2.I: Bacterial strains and plasmids.

Strain or plasmid Relevant genotype Reference or source

S. meliloti Rml02l SU47 str-21, Sm., wild-type Meade et al. (1982) SRmA102 RmlO2I rhaQl::Tní This work SRmA137 Rm1021 rhaR2::TnS This work SIìmA13B RmIO2l rhaPS::TnS This work SRm,¿\145 Rm102l rhaT4::Tní This work SRmA146 Rm1021 rhaDS::TnS This work SRmAi63 Rm102I rhaK6::Tn5 This work SRmAl86 Rnr1021 rhaTT::TnS This work SRmAl9l Rm1021 rhall2::TnS This work SRm.A,Zl I Rml02l rhaKl6::Tn5 This work SRmAZTB Rm1021 rhaKl7::Tní This work

R. le gumino s er unl bv. tr iþ lii Rlri00 Wl4-2 Sm', wild-t1pe Baldani et al. (I992) Rlr105 Rlt100 rhaDl::Tn5-820 Oresnik et al. (1998) Rlr106 RltlOO rhaT2::Tn5-820 Oresnik et al. (1998) RItI44 RI|ID} rhaK Richardson et al. (2004)

E. coli MM294A pro-82 thi-1 hsdRl7 supE44 Finan et al. (1986) MT6O7 MM294A recA-56 Finan et al. (1986) MT6I6 MT607 (pRK600) Finan et al. (1986)

Plasmids pRK600 pRK20t3 npt::Tn9,Cm' Finan et al. (1986) pRK602 pRK600::Tn5, Cm', Nm' Finan et al. (1985) pRK7B13 Broad host range vector, Tcr Iones et at. (1987) pW3A pRKTB13 based Rltl00 rhamnose Oresnik et al. (i998) complementing cosmid pW3ARl pW3A rhaDl::Tn5-820 Oresnik et al. (1998) pW3AR2 pW3A rhaT2::Tn5-820 Oresnik et al. (1998) pW3Cl pRK7B13 based Rltl00 rhamnose Oresnik et ai. (1998) complementing cosmid pMR53 pRKTBI3 expressed rhaþ Richardson et al. (2004) pMRll0 pRK7B13 expressed rhaK+ Richardson et al. (2007)

J/ CI.TAPTER 2. RI]AMNOSE CATABOLISM AND T-RANSPORT

Genetic techniques

Mutagenesis of Rml02l with TnSwas carried out usingpRK602 as previouslydescribed

(Finan et al. 1985). Putative mutants were routinely single colony purified three times and retested for phenotype. Identification of the location of TnS insertions was per- fbrmed using PCR as previously described (Poysti et aI. 2007) (also see Chapter 4, page

89) . Conjugations were carried out essentially as previously described (Finan et al. I gBB) .

Rhamnose transport assays

'fiansport assays were carried out essentially as previously described (Richardson et al.

2004). Uniformly labeled []I{l rhamnose (l85 GBq mmol-r) was purchased from Amer- ican Radiolabeled Chemicals Ltd. (St. Louis, Missouri). Transport assays were initiated by tlre addition of [:JFI] rhamnose to a final concentration of 2 nM (125,000 dpm) and aliquots of 0.25 or 0.5 mL were withdrar¡¡n at appropriate time points and rapidly fil- tered through a Millipore 0.45 pm Hv filter on a Millipore sampling manifold. Samples were taken every l5 or 20 seconds and continued for up to 2 minutes. The amount of radioactivity retained by the cells was quantitated by a liquid scintillation counter (Beck- man LS6500).

Plant assays

Plant s).rrnbiotic assays were carried out as previously described (Oresnik et al. i 994) . Al- fafia (Medicago saliuacv. Rangelander) seeds were surface sterilized by treating the seeds for 20 minutes withT}Vo ethanol followed by30 minutes in 1% sodium hypochlorite. The seeds were then extensively washed with sterile distilled water and germinated on 1.57o water agar plates. Seedlings were transferred after about 2-3 days to sterile Leonard jar

JC)ao C T.I A I''T E R 2. R H A M N OS E CATA B O LI S M A N D T RA N S P O RT assemblies. Plants were grown in a quartz sand:vermiculite mixture with nitrogen free

Jensen's nutrient solution and water added as required (Glazebrook and Walker l99i).

Plants were inoculated by growing up S. meliloti cultures overnight in LB and diluting the cultures l0-2 in sterile distilled water and inoculating each Leonard Iar assembly with l0 ml (approximately 104-lOs bacteria per plant). Plants were harvested between

28 and 35 days after inoculation. The plants were unearthed and the shoots (above the cotyledon) were dried to determine a plant dry weight. Representative nodules were crushed, sterilized (l minute of lTo sodium hypochlorite), and bacteria were isolated.

Bacteria isolated from nodules were streaked onto LB agar plates and single colonies were checked for relevant phenotypes. Plant competition assays were carried out as previously described (Poysti et aL.2007) (also see Chapter 4, page 92).

Results

SMc0232l -SMc03003 encode genes necessary for rhamnose catabolism in S. meli.Iotí.

Approximately 10,000 Tn5 generated mutants from approximately20 independent mutagenesis experiments were screened for their inability to grow on rhamnose as a sole carbon source while retaining the ability to grow on glucose as a sole carbon source. Ten mutants unable to grow with rhamnose as a sole carbon source were isolated from these screens ('Iable 2.I).

Identification of the sites of the TnS insertions was achieved using a PCR protocol where one primer is homologous to the end of the TnSIS50 element and the other is arbitrarily defìned. The amplification products therefore span a small region of the IS50 and the flanking DNA. Sequencing of the PCR products followed by BTASTN analysis us-

39 CI-IAPTER 2. RHAMNOSE CATABOLISM AND TRANSPORT irrg the S. meliloti database was used for determining the point of insertion. The results of these analyses show that all of the rhamnose negative mutants had inserts that were localized within two putative transcripts. One transcript is composed of SMc02321 and

5Mc02322, whereas the other consists of SMc02323, SMc02324, SMc02325, SMc03000,

SMc03001, and SMc03003 (Figure 2.1). These genes appear to be homologs and have a conserved operon structure to that of the R. leguminosarumbv. tt"ifolii rhamnose catab- olisnr operon that has been previously described (Richardson et aL.2004). Characteriza- tion of the locus necessary for the catabolism and transpofi of rhamnose in R. legumino- sarumhas sho'o¡r that the genes were organized into two transcripts: one contained rhal and rhaD, whereas the other contained ThaRSTPQUK (Richardson et al. 2004). Not only is the S. meliloti locus also required for rhamnose catabolism, but the predicted amino acid sequence similarity of the S. meliloti and R. leguminosarum rhamnose operons is high (Table 2.2). Based on the similarity in function and sequence conservation, the database annotation for the S. melilotigenes should be updated; the systematic names should be renamed as shor,rm in Figure 2.1 to conform with those of their R. legumino- s(rrum homologs.

S. melilof,i rhamnose mutants are not sensitive on rhamnose/glycerol media.

\¡Vhile characterizing the R. leguminosarum rhamnose locus, it was sho'nm that strains carrying inserts in either rhaD or rhal were unable to growwith media containing glycerol and Lhamnose, although they were able to use glycerol as a sole carbon source

(Richardson et al. 2004). This phenotype was hypothesized to be caused by the buildup of phosphorylated rhamnose generated by the rhamnose kinase, RhaK. Consistent with

40 CI]API-ER 2. RIíAMNOSE CATABOL]SM AND TRANSPORT

Figure 2.1: Rhamnose utilization Iocus. A schematic representation of the region necessary for rlramnose utilization in S. melilofi. Boxes represent open reading frames with the pointed end indicating the direction of transcription. Vertical bars represent positions of TnS inserts.

I tbp

SRmÂ186 SRmA I 63

(ìeDc rhal rhaD rha| rhaS rhaT rhaP rhaQ rhaU rhaK Syslcmatic idcntificr SMc182r Sitc02322 SLtú2323 5Mc'2321 sMú2325 SM.0lO00 5Mc03001 5Mc03002 5M.0J003

4I CI1APTER 2. RHAMNOSE CA'IABOLISM AND TRANSPORT

Tab\e 2.2: Amino acid homology of S. meliloti and R. leguminosarum rhamnose catabolism gene products.

Gene product ldentifier % identity % similarity

RhaI SMc0232l 78.2 85.9 RhaD 5Mc02322 74.0 84.5 oo1 RhaQ SMcO300l 72.L OJ. I RhaR SMcO2323 67.8 oJ. / RhaS 5Mc02324 67.4 83.4 RhaT SMcO2325 67.3 BO.B RhaP SMc03000 64.2 78.7 RhaU SMc03002 64.1 73.8 RhaK SMc03003 55.2 67.2

42 C H A PI- E II 2. R I-IAM N OS E CATA B O LI S M A N D TRA N S P O RT this, rhaK, rhaK rhaD and rhaK rhal mutants were all able to grow on media containing both rhamnose and glycerol. Furthermore, RhaK was found to have rhamnose dependent kinase activitywhich was not dependent on RhaD or RhaI (Richardson and Oresnik

2007).läken together, these results suggest that RhaK acts before RhaD and RhaI in the catabolic pathway.

In an effort to test for similar effects in S. melilofl, strains unable to grow on rhamnose were tested for their ability to grow on defined media containing rhamnose and/or glycerol. As expected, all of the strains tested were able to utilize glycerol and unable to utilize rhamnose as sole carbon sources (Table 2.3). In contrast to R. leguminosarum, all of the S. meliloti strains were able to grow on media containing both rhamnose and glycerol. This phenotype suggests several possibilities: the phosphorylated intermediate does not build up to toxic levels in S. melílori; an alternate pathway exists to utilize the phosphorylated rhamnose; or the RhaK enzlirne acts at a different point in the biochemical pathway.

Complementation of S. meliloti rharnnose mutants with R. Iegumíno- sørum rhamnose genes.

To test whether the S. meliloti locus was equivalent to the characterized R. leguminosarum rhamnose locus, a heterologous complementation experiment was carried out. trour different cosmids containing either the R. leguminosarum v,nld-type rha region

(pW3Ct and pW3A), or a transposon in rhaD or rhaT (pW3ARi and pW3AR2, respectively), were mobilized into S. meliloti and R. leguminosarummutants. Complemented strains were tested for the abiliry to use rhamnose as a sole carbon source (Table2.4).

As previouslyshown pW3A complemented both R. leguminosarummutants Rltl05 and

43 CI]API'ER 2. RHAMNOSE CATABOLISM AND TRANSPORT

Table 2.3: Growth of S. meliloti rhamnose catabolism mutants on media containing rhamnose and glycerol.

Growth with carbon source(s),,

Strain Relevant genotype Glyb Gly*Rha Rha Iìm1021 S.meliloti,wild-type + + + Rlt100 R.leguminosarum, wild-type + + + Rlt105 R.leguminosarum, rhaD + RItl06 R. leguminosarum, rhaT + + SRmAl46 S. meliloti, rhaD + + I SRrnAlBG S. meltloti, rhaT ¡ + SRnrAl9i S. meliloti, rhal + + SRnrA2l i S. nzeliloti, rhaK + +

a Grown on VMM media with the given carbon sources. Abbreviations: Rha, rhamnose; Gly, glycerol. t) Growth phenotype indicated by: *, same as wild type; - no visible growth.

44 CIIAPTER 2. RHAMNOSE CATABOLISM AND TRANSPORT

Tab\e 2.4: The R. leguminosat'um rha region does not complement S. meliloti rhamu- tants.

Complementation by plasmido Strain pW3CI pW3A pW3ARl pW3AR2 pMR53 no plasmid Rlrl00++++++ Rlri0s + + + Rlrl06 + + + RmlO2t++++++ SRm^l3B + + SRmAt46 + +

rì + indicates the ability to grow with rhamnose as a sole carbon source when complemented by the plasmid indicated. Lack of growth indicated by -.

45 C I1 A PT E R 2. R I-I A M N O S E CATA BO LI S M A N D T RA N S PO RT

Rlt106 (Oresnik et al. l99B). As expected, strain Rlt105 was complemented by pW3AR2 but not pW3ARl while RIt106 was complemented by pW3AR1, bur nor by pW3AR2. Un- expectedly, the S. meliloti mutants SRmAl3B (rhaP3) and SRmAl46 (rhaDí),were complemented by pW3Cl and pW3A but not by either pW3ARl or pW3AR2.

The data clearly show that although pW3Cl and pW3A were capable of restoring growth to S. melilofi strains unable to use rhamnose, cosmids containing an insertion

(pW3ARI or pW3AR2) were not. It was therefore hypothesized that the R. leguminosarum negative regulator, RhaR, could be preventing heterologous complementation by repressing transcription of the native S. meliloti rhamnose operon through copy- number effects or by acting in a dominant negative fashion. To test these possibilities, pMR53 (R. leguminosarum rhaRl) was introduced into each of the S. melíIotí rhamnose mutants as well as the wild-type Rm1021. The resulting transconjugants were then tested for the ability to grow on rhamnose as a sole carbon source (Table 2.4). If tÌrre

R. leguminosarumregulatorwas repressingthe chromosomal locus due to copy-number effects or a dominant negative activity, Rml021 should not be able to grow on rhamnose as a sole carbon source. The results show that the introduction of pMR53 did not affect the wild type's abilily to grow on rhamnose suggesting that the the inability of R. Iegum- inosaruntbased cosmids containing insertions to complement was not solely due to the presence of rhaR.

R. legumínosa,rum RhaK is capable of complementing S. melíloti rhøK mutations.

The fact that S. meliloti rhamnose catabolism mutants were able to grow on media with rhamnose and glycerol (Table 2.3), and the fact that heterologous complementation re-

46 CHAPTER 2. RHAMNOSE CATABOLISM AND TRANSPORT sults were not straightforward (Table 2.4) suggests that there are some significant differences between rhamnose catabolism mutants of S. meliloti and R. leguminosarum.

We were therefore interested to know if the function of the R. leguminosarumRhaK was interchangeable with the S. meliloti RhaK.

To carry this out, pMRlI0 (À. leguminosarum rhaK+) was conjugated into strains

SIìm4163 (rhaK6), SRmA2lI (rhaK16), and SRnAZTB (rhaKL Z. As controls, the cosmids pW3Cl, pW3A and pW3ARl were also introduced. Transconjugants were tested for the ability to grow with rhamnose as a sole carbon source (Table 2.5). Similarly to the heterologous complementation of the other S. meliloti mutants, pW3ARl did not complement. I-Iowever, pW3Cl, pW3A and pMRll0 were able to complement S. meliloti rhaK mutations. Although we do not currently understand why pW3ARl does not complement, the data suggest that both the S. meliloti and R. leguminosarumRhaK enzyrnes function with the same substrate in uiuo. If RhaK functions prior to RhaD and

RhaI in R. leguminosarum, then it is likely to function the same in S. melilott.

Testing the S. melílotí rha locus for uptake of rhamnose.

Previous characterization of the R. leguminosarumrhamnose utilization operons clearly showed that the associated ABC transporter encoded by rhaSTPQ was responsible for the transport of rhamnose (Richardson et aL.200Ð. More recently it was shor¡¡n that

Il.. leguminosarum rhaK also affects transport into cells (Richardson and Oresnik 2007).

I'o test the likelihood that the components of the S. meliLoti ABC transporter are also responsible for rhamnose transport, whole cells of the wild type, SRmAI4S (rhaT4) and

SRmA2l l (rhaK16) were assayed for labeled rhamnose accumulation.

Initial assays provided results that suggested a decrease in rhamnose accumulation was associated with S. meliloti rhaT and rhaK mutations (Figure 2.2). The first assays

47 CHAPTER 2. RHAMNOSE CATABOLISM AND'TRANSPORT

llable 2.5: R. IeguminosarumrhaK complements S. meliloti rhaK mutations.

Complementation by plasmido Strain pW3Cl pW3A pW3ARl pMRir0 no plasmid Rlrl00+++++ Rlrl44++++ Rm1021 +++++ SRmAl63 + + + SRmA211 + + + SIìmA27B + + +

a + indicates the ability to grow with rhamnose as a sole carbon source when complemented by the plasmid indicated. Lack of growth indicated by -.

48 CLIAPTER 2. RHAMNOSE CATABOLISM AND TRANSPORT

Figure 2.2: Rhamnose transport assays of S. melilorl wild type (Rml02I) and two rhamnose catabolism mutants (SRmA145: rhaT, and SRmA2l I: rhaK) induced with defined medium containing rhamnose*glycerol. Error bars represent standard deviation.

!l .gr 3t.c) -ç.(/l è1.o or L o õ E 50. o oØ c =0.ro -c io.o o tlo0. o- l

Time (seconds)

49 CHAP'.ER 2. RHAMNOSE CATABOLISM AND TRANSPORT were performed with cells grown in rich media, then washed and subcultured into defined media with rhamnose and glycerol. Although S. meliloti rhamnose catabolism mutants can grow in the presence of rhamnose, they were observed to grow significantly more slowly than the wild type (data not shor,rm). Due to differences in the growth phase at the time of assay, further assays were performed using cells gror,rm under different conditions. In order to assay cells that were of the same optical density at the time of assay, rhamnose catabolism mutants were allowed to grow for approximately 24h longer than the wild type. Preliminary assays of cells grown in this manner suggest that there is little difference in rhamnose accumulation between rhaT, rhaK andwild-type strains

(Figure 2.3). Despite the difference in results of these two methods, it is clear that S. meliloti rhaT- and rhaK mutants are still able to accumulate rhamnose, which is not true for the homologous R.leguminosarummutants (Richardson et al. 2004).

S. melilotí rhamnose mutants are uncompetitive for nodule occupancy on alfalfa.

R. legunzinosarumbv. triþIii mutants unable to catabolize rhamnose were found to be severely defective in their ability to compete for nodule occupancy (Oresnik et al. 1998) .

To test the hypothesis that the ability to utilize rhamnose would affect competition for nodule occupancy in S. melílofl a representative mutant was assayed (SRmA146). Pre- liminary plant growth experiments showed that mutants unable to grow with rhamnose were able to form an effective symbiotìc association with alfalfa; plant dryweights were measured and found to be not significantly different from those inoculated with wild type (data not shown). Both approximately equal and high ratios of mutant:wild type were tested for their ability to compete for nodule occupancy by inoculating al-

50 CIIAPTER 2. RHAMNOSE CATABOL|SM AND TRANSPORT

Figure 2.3: Rhamnose transport assays of S. melilofi Rm102l, SRmA211, and SRmA145 grown in defined media containing rhamnose*glycerol.

'oc +J o Rm1021 Lo o- o SRmA145 õ V SRmA2l-l- ¿J uo ..."Y c'' E1 Lo o- õ Y. E -"'í o o o U) co E o Y" -cL o rþ ..""" O-tt- o c) -'t/ l¿ lU -'j2' -l-)o "-4 :f nt/zrlrlrlrlrl "0 10 20 30 40 s0 Time (seconds)

5t CLIAPTER 2. RHAMNOSE CATABOLISM AND TRANSPORT falfa seedlings with mixtures of SRmAl46 and Rml02l. Resulting nodules were sampled after 2B-:J5 days and the isolated bacteria were tested for their ability to grow with rhamnose as well as their antibiotic resistance markers. in each case, the observed nodule occupancy was lower than the inoculum ratio (Figure 2.4). A test for significance by a paired Student's t-test indicates that the ratio of nodules occupied by SRmAI46 is significantly lower (P<0.02) when the inoculum ratio of SRmA146:RmI021 is approximately equal. Similar results were found when SRm,A,l46 made up approximately 707o of the inoculum. Taken together, the ability to use rhamnose as a sole carbon source affects the ability to compete for nodule occupancy in S. melilofi in an analogous manner to

R. legumínosarum.

Discussion

In this work two putative operons were identified that are necessary for rhamnose catabolism in S. meliloti (Figure 2.1). The operons have the same genetic organization, and each gene product has high similarityto the R. leguminosarumrhamnose catabolic genes that have been previously described (Richardson et aL.2004). On the basis of this evidence we suggest that the genes should be renamed to be consistent with R. leguminosarutn.

Although the operons and the sequences of the encoded genes are highly similar, there are differences between R. leguminosarum and S. meliloti. The wild-type region from R. leguminosarum catt complement the inability of the S. meliloti mutants described here to utilize rhamnose as a sole carbon source. However, inserts within this cosmid abolish the ability to complement. This suggests that the whole R. legumino- sarumrhanrtnose system is required for rhamnose transport and/or catabolism in S. me-

52 CHAP ER 2. RHAMNOSE CATABOLISM AND I'RANSPORT

Irigure 2.4: Ability to compete for nodule occupancy is compromised in SRmAI46. The percentage of mutant bacteria is presented as enumerated from plant inocula and nodule preparations after 28-35 days. Plants in Set I were inoculated with approximately equal concentrations and the error bars indicate the standard deviation (P<0.02) of four independent trials. Plants in Set 2 were inoculated with a higher concentration of wild type than mutant and are the average of two trials, where the difference between the mean is less than 5% (marked with asterisk).

C o +J (o o-= o o- t1Fo s

53 CI]APTER 2. RI{AMNOSE CATABOLISM AND TRANSPORT liLoti. We note that there are differences in the deduced amino acid sequences found in the DNA binding motif of the S. meliloti and R. leguminosarumRhaR proteins (data not shor¡¡n). The hypothesis that the R. leguminosarumRhaR negative regulator could be acting as a dominant negative determinant or have excess activity due to copy-number was not supported by experimental data. Based on the fact that R. leguminosarumand

S. nteliloti RhaK proteins were found to be interchangeable and that there is high sequence similarity between the rhamnose genes in the two organisms, it seems unlikely that the catabolic pathway would differ between the two organisms. However, it may be important to note that S. melilotí rhaD and rhal mutants are able to grow on media with rhamnose and glycerol while R. legumínosarummutants are not. This suggests that there are some biochemical differences between the organisms.

In addition to our observation that R. Ieguminosarumcosmids containing insertion elements were unable to complement S. meliloti rhamnose mutants, we have been unable to complement S. meliloti with S. meliloti derived cosmid banks. Despite repeated attempts to isolate a complementing cosmid using different S. meliloti mutants and independently generated S. meliloti cosmid banks, we have been unsuccessful (data not sho',¡¡n). Likewise, attempts to isolate an S. melilofl complementing cosmid by introducing the cosmid bank into R. leguminosarumrhamnose catabolism mutants has also been unsuccessful. Currently, we have no explanation for our inability to isolate a cos- nrid containing the S. meltloti rhamnose region.

The rlramnose operon in R. leguminosarum encodes aABC-type transpofier necessary for rhamnose uptake (Richardson et al. 2004). It was also reported that the ABC transporter in R. legurninosarum does not transport rhamnose into the cell unless a functional RhaK is present (Richardson and Oresnik2007). Given the homology be-

54 CI]APTEII 2. RHAA4NOSE CATABOLISM AND TRANSPORT tween R. leguminosarum and S. meliloti, the results of our assays for rhamnose transport in S. meliloti were very surprising (Figures 2.2 and 2.3) . Itwas unexpected that both the rhaT and rhaK mutants would be able to accumulate rhamnose. Although the data clearly indicate that rhamnose accumulation occurs in S. meLiloti rhaK and rhaT mutants, the data are currently unable to resolve where or not rhaK or rhaT play any role in rhamnose transport. Due to the fact that different rates of transport were measured depending on the method used to grow the assayed cells, further experimentation will be required to resolve these differences.

Despite the inconsistencies in the transport experiment results, the fact that the

S. meliLoti mutants were still able to accumulate labeled rhamnose suggests tvvo possibilities. First, it is possible that rhaSTPQ do not encode a transporter capable of transporting rhamnose. Second, S. meliloti could possess another transporter capable of transporting rhamnose. The possibility of a second rhamnose transporter could be easily tested by generating a strain containing a deletion of. rhaSTPQ and testing for its ability to grow with rhamnose as a sole carbon source. Assays for the ability to accumulate labeled rhamnose could also be performed on this strain. If it is determined that another rhamnose transporter exists, screening for mutants of the rhamnose transporter deletion strain unable to utilize rhamnose should yield mutants in this second transporter.

The original basis of this work was to test the hypothesis that rhamnose catabolism plays a role in competition for nodulation in S. melilofi. Our data clearly shows that

S. meliloti rhamnose mutants, like R. leguminosarum, aÍe compromised in their ability to compete against the wild type for nodule occupancy (Figure 2.4). Compared to

R. leguminosarum, the S. meliloti wild lype does not grow very well on rhamnose as a sole carbcln source (data not shovr¡n). However, mutants unable to use rhamnose as a

55 CI]APTER 2. RI-]AMNOSE CATABOLISM AND TRANSPORT sole carbon source appear to have a similar inability to compete for nodule occupancy against their isogenic wild type (Oresnik et al. l99B). It is possible that at some point during the colonization or infection process rhamnose is available and the inability to use this carbon source may affect the ability to directly compete with the wild t1pe. Con- sistent with this possibility, small amounts of rhamnose have been measured in the root exudates of pea and cowpea legumes (Knee et al. 2001).

Characterizing the competitive phenotype of S. meliloti mutants unable to use rhamnose has shornm that the inabilityto utilize rhamnose is correlated with decreased ability to compete for nodule occupancy. Flowever, the data also highlight that even though the genes necessary for rhamnose catabolism appear very similar to those of R. Iegum- inosarutn, there are significant differences that are not currently understood. Future work should attempt to elucidate the nature of the genetic differences and to address why rharnnose mutants are uncompetitive.

56 Chapter 3 Chara cterizatio n oï Sinorhizob ium meliloti triose phosphate isomerase genes

Sumrnary

The work in this chapter was performed by Nathan Poysti and was published in part in the Journal of Bacteriology (Vol. 189:3445-3451,2007).

A Tn5 mutant strain of. Sinorhizobium meliloti was isolated with an insertion in tpiA

(Systematic identifier: SMc0l023), a putative triose phosphate isomerase (TPI) encoding gene. The tpiA mutant grew slower than wild type on rhamnose and did not grow with glycerol as a sole carbon source. The genome of S. meliloti wild-type Rm102l contains a second predicted TPI encoding gene rplB (SMc016l4). We have constructed mutations and confirmed that both genes encode functional TPI enzymes. tpiA appears to be constitutively expressed and provides the primary TPI activity for central metabolism. l:piB has been shonm to be required for growth with erythritol. TpiB activity is induced by growth with erythritol, however, basal levels of TpiB activity present in t¡tiAmutants allow for growth with gluconeogenic carbon sources. Although tpiAmu- tants can be complemented by tpiB, tpiA cannot substitute for mutations in fprB with respect to erythritol catabolism. Mutations in tpiAor tpiB alone do not cause sgnbiotic defects, howevel, mutations in both tpiAand tpiB caused reduced nitrogen fixation.

57 CITAPTER 3. TRIOSE PITOSPHATE ISOMERASE

Introduction liiose phosphate isomerase (TPI) catalyzes the reversible interconversion of glyceraldehyde-3-phosphate (G3P) and dihydroxyacetone phosphate (DFIAP). This activity makes

TPI a key enzyme of central carbon metabolism, allowing it to play a role in the glycolysis (Embden-Meyerhof-Parnas), gluconeogenesis, pentose phosphate (PP) and Entner-

DoudorolT (ED) pathways. Past research on the symbiotic soil bacterium Sinorhizobium nteliloti and other rhizobia has shown that hexose catabolism proceeds through the ED and PP pathways, while the Embden-Meyerhof-Parnas (EMP) pathwayfunctions atvery low levels, if at all (Stowers 1985). Gluconeogenesis, however, is functional, and mutations in gluconeogenic enz)¡mes have been shoriun to cause complex symbiotic phenotypes, including reduced or abolished nitrogen fixing activity (Bolton et al. 1986, Finan et al. l9BB, 1991, Gardiol et al. 1982, Hornez et al. 1989).

Recent work has confirmed the roles of ED and PP as catabolic pathways in S. meliloti. Analyses of carbon flux using labeled carbon compounds and gas chromatography-mass spectrometry (GC-MS) have confirmed the absence of EMP during growth with glucose (Fuhrer et al. 2005). Another set of experiments using labeled carbon and nuclear magnetic resonance (NMR) is in agreement, in that glucose is primarily degraded through ED (Gosselin et al. 2001, Portais et al. 1999). The GC-MS and NMR experiments also confirm that some of the G3P produced by the catabolism of hexoses through ED is converted back to higher molecular weight compounds through TPI and fructose-bisphosphate aldolase (FBA) in a cyclic pathway. Because FBA requires G3P and DFIAP as substrates, the interconversion of G3P and DFIAP is necessary for the cyclic nretabolism of hexoses as well as gluconeogenesis. An S. melíloti strain missing TPI activity would therefore be metabolically compromised.

5B CHAPT'ER 3. TRIOSE PHOSPHATE ISOMERASE

IJased on gene homology, S. meliloti has two putative chromosomal genes predicted to encode TPI enz]¡mes (Galibert et al. 200I). The gene tp¿A (SMc01023) appears to be transcribed independently, while rptB (SMcO1614) appears to be transcribed along with two other genes: SMcOl6i5 and rpiB. A putative operon upstream of tpiB includes homologs of genes involved in the catabolism of erythritol in Rhizobium leguminosarum and Brt¿cella abortus (Sangari et al. 2000, Yost et al. 2006). The tpiB gene and its proximity to erythritol catabolic genes are conserved in a number of organisms, which led us to hypothesize that tpiB is involved in erythritol catabolism. In this work we describe the iscrlation and characterization of TPI mutants in S. meliloti. We have found that both tpiA and tpiB encode functional TPI enz)¡mes, and that tpiB is induced by and specifically required for erythritol catabolism. Mutations in both TPI genes cause a loss of the ability to use gluconeogenic carbon sources, but viability and growth with other carbon sources or combinations of non-permissive carbon sources remain possible. Mutations in either tpiAor tpiB alone did not affect symbiosis, while mutation of both TPI encoding genes was found to cause reduced levels of symbiotic nitrogen fixation.

Materials and methods

Bacterial strains, plasmids, and media

The bacterial strains and plasmids used and generated in this work are listed in Table 3.I.

Ilacteria were grown at 30"C, using either LB or Vincent's Minimal Medium (\TvIM), as previously described (Sambrook et al. 1989, Richardson et al. 2004). All carbon sources used in \/MM were filter sterilized, and used at a concentration of 15 mM, unless oth- erwise stated. All sugars were in the D-configuration, except L-arabinose, L-fucose, L- rhamnose and L-lyxose. When required, antibiotics were used at the following concen-

59 CHAPTER 3. TRIOSE PHOSPHATE ISOMERASE trations: tetracycline (Tc) 5 FBmL-r;neomycin (Nm),200 pgmL-r;kanamycin (Kan),20 pgmL-r;streptomycin (Sm) 200 ¡"rgmL-r;rifampicin (Rf),50 [rBmL-';chloramphenicol

(Cnr), 20 ptg mL-r; gentamicin (Gm), 20 Vrgml*r for Escherichía coli; Gm,50 pg mL-' tor S. meliloti.

DNA manipulations and plasmid constructions

Standard techniques were used for DNA isolation, restriction enzyrne digests, ligations, transformations, and agarose gel electrophoresis (Sambrook et al. l989).

To construct pNP150, a 350 bp internal fragment of tpiB was PCR amplified using primers I and 2 (Table 3.2) and genomic RmI021 DNA as a template. The PCR product was gel isolated and cloned into pICIIOCK-Gm using the BamHI and KpnI sites provided by primers I and 2. The resulting construct was named pNPl50, and was transformed into competent E colí SI7-1.

The plasmid pNPl55 was constructed to create a chromosomal tpiA deletion in the strain Rml021. A 700 bp region upstream of tpiAwas PCR amplified from RmIOZI genomic DNA using the primers 3 and 4 (Table 3.2), which add the restriction sites NotI and BamHI to the 5' and 3' ends of the fragment, respectively. This region was cloned into pBluescriptll using NotI and BamHI, yielding pNPl52. Similarly, a 660 bp region dornmstream of tpiAwas amplified with primers 5 and 6 (Table 3.2) which add the restriction sites BamI-II and Sall to the 5' and 3'ends of the fragment, respectively. The 660 bp fragment was cloned into pNP152 using BamHI and SalI, yielding pNP154. A 1360 bp NotI-SalI fragment was then subcloned from pNPlS4 into pIQ200SK (Quandt and

I{ynes 1993) and this construct was named pNPl55.

In order to constitutively express tpiAand tpiB,both ORFs were cloned into a broad host range vector, pRKTBl3 (Jones and Gutterson l9B7). Primers 7 and B (Table 3.2) were CHAP'IER 3. l-RIOSE PHOSPHATE ISOMERASE

Table 3.1: Bacterial strains and plasmids.

Strain or plasmid Relevant genorype Reference or source

S. tneliloti Rm1021 SU47 srr-2-1, Sm' Meade et al. (1982) Rnr5000 SU47 rrfS, Rfl Finan et al. (1984) Rrn5439 Rm1021 pckAl:;TnV Finan et al. (1988) SRmAlS5 Rm102t tpiAl::Tní This work SRmr\327 /p¿AI , O(SRmA1B5)+ft¡nlQ2l, Nmr¡r This work SRmA355 Rm5000 fprBl ::pKNOCK-Gm This work SRmA366 tpiAl tpiBl, rÞ (SRmA les¡ *sRmA355, Nm' This work SRnrA449 Rml02l AtpiA This work SRmA5l5 Ltp iA tpiBl, O (SRmA355)'SRmA449, Gmr This work SRmA5B4 RmI02I tpiÙ2;:Tn5-820 This work SRrnA5B5 Ltp iA tpi&L, rÞ (SRmA5B4)-SRmA449, Nm, This work

E. coli MM294A pro-82 thïl hsdRl7 supE44 Finan et al. (1986) MT6O7 MM294A recA-56 Finan et al. (1986) MT6I6 MT607 (pRK600) Finan et al. (1986) st7-l rec,A derivative of MM294A Simon et al. (1983) with integrated RP4-2 (Tc::Mu::Km::Tn Z) Dll5ø 9yr,496, Nal' Invitrogen EcA100 MT607 Rifl derivative This work EcAl01 MT607::TnS-820 Kanr Clark er al. (2001)

Plasnrids plìK600 pRK2013 npt:;Tt1g,Cmr Finan et al. (1986) pRK602 pRK600::TnS, Cmr, Nmr Finan et al. (1985) pllluescriptlf SK+ Cloning vector, Ap' Stratagene pKNOCK-Gm Suicide vector, Gmr Alexeyev (1999) pRKTBl3 Broad host range vector, Tcr Jones and Gutterson (t987) pJQ2o0SK Gene replacemerlt suicide vector, Gmr Quandt and Hynes (1993) pPHIJI IncP plasmid, Gmr Tcr Beringer et al. (1978) pNP150 pKNOCK-Gm with 350 bp ir.rternal This work rpiB fragment pNPì 52 pBluescriptll SK+ with 700bp tpiA This work 5'flanking region pNPl54 pBfuescriptll SK+ with 660bp tpiA This work 3'flanking region pNPl55 plQ200SK with tpiA flanking regions This work pNPl 63 fpiB complementing cosmid This work pNPl66 Constitutive tp¡,4 This work pNPl 67 Corrstitutive rpiB This work

n Notation indicates that SRmA327 was constructed by 0M12 transduction of SRmAl85 into Rnl102l with selection for Nmr.

6l CI IAPTER 3. TRIOSE PHOSPHATE ISOMERASE

Table 3.2: Primers used in this work.

Prirner number Nucleotide sequence (5'---+ 3') I ...... ATATGATCCAGAACATGCACTGGGACGAT 2 ..,. .... TATAGGTACCGCTCGTAAGCCAGAAGAACG 3 ...... AAGCGGCCGCCATGGGCGAGGGCCCA 4 ...... AÀGGATCCGGGATCTCCTCCAGATCA 5 ...... AAGGATCCTTGCCGCGGGACTTGGA 6 ...... AAGTCGACACCGCTCGGAACCTGCG 7 ...... ATGGATCCAGGAGAA,ATAATTATGAGAGGATCGCAT CATCATCATCATCATACGCCCGATATCC GCC C B ...... ATGAATTCTCAGGCAGTCAATTCTTCATA 9 ...... ATGGATCCAGGAGNU\TAì\TTATGAGAGGATCGCAT CATCATCATCAT CATAGCAÄ.GT CT GGT C TC T GGG 10...... ATGAATTCTCAGATCGCCCGGGCTAC II...... CACGATGAAGAGCAGAAG 12...... GGCCACGCGTCGACTAGTCAGNNNNNNNNNNACGCC 13...... TAGGAGGTCACATGGAAGTCAGAT 14...... GGCCACGCGTCGACTAGTCAG

62 CI-ÌAPTDR 3. TRIOSE PHOSPHATE ISOMERASE used to amplify tpiAfrom Rml02l, which adds a ribosome binding site (GGAG), an N- terminal RGS(I{is)6 tag, and BamHI and EcoRI restriction sites to the gene. Primers

9 and l0 (Table 3.2) were used to amplifu tpiB, adding the same features as tpiA. Both tpiAand tpiBwere cloned into pRKTBI3 yielding pNP166 and pNP167 respectively. Both consLructs were entirely sequenced, showing that neither had any deviation in sequence from the published Rm102i genome (data not shonm).

Genetic techniques and mutÍìnt construction

Mutagenesis of Rml021 with TnSwas carried out usingpRK602 as previouslydescribed

(Irinan et al. 1985). Conjugations between E. coli and S. meliloti were carried out as previously described using the mobilizing strain MT6l6 (Finan et al. 19BB). Transduction with phage OM12 (Finan et al. I9B4) was routinely used to move transposon markers from mutant strains into wild-type parent strains, and 50-100 colonies were screened to ensure that phenotypes were 1007o linked in transduction and therefore associated solely with the transposon insertion.

Construction of the rplB single cross-over mutant SRmA355 was accomplished by conjugating pNP150 into Rm5000. Single cross-overs were selected for by plating on

VMM agar containing Rf and Gm. The pICTIOCK tpiB insertion was confirmed by trans- ductional linkage analysis with a nearby marker (data not shor¡¡n) . The tpiA tptBmutant,

SRm,A366, was constructed by transducing a SRmAIBS 0Ml2 lysate into a SRmA355 background. The chromosomal deletion of tpiA in SRmA449 was constructed by mat- ing pNP155 into Rm102l and selecting for single cross-over events using Gm. Several of these were plated onto LB containing 57o sucrose to select for double cross-over events.

A sucrose resistant, Gm' colony was purified and tÌire tpiA deletion was confirmed by

PCR amplifying the deleted region using primers 3 and 6 (Table 3.2), sequencing the CHAP'TER 3. TRIOSE PHOSPLIATE ISOMERASE

gel isolated PCR product, and comparing the result with the expected nucleotide sequence (data not sho'nm). Strain SRmAsl5 was created by moving the Gm marker from

SRmA355 into SRmA449 using transduction. The tpiBcomplementing cosmid, pNP163, was isolated by conjugating an S. meliloti cosmid bank with SRmA515, and plating on a defined medium containing erythdtol as a sole carbon source. The cosmid pNP163 was then mutagenized by passage through the Tn5-820 carrying strain, EcAt0l, selecting for co-transfer of the plasmid and TnS-820 into EcA100. Individual cosmids carrying

Tn5-ts20 were then mated into SRmAS15 and screened for a loss of the ability to complement growth on erythritol. One such mutagenized cosmid was isolated and sho'¡¡n to contain a mutation in tpiB using a protocol described below (data not shorn¡n). Marker exchange of the cosmid borne tpiB mutation into the Rm102l chromosome was per- fbrmed as previouslydescribed (Glazebrook andWalker I99I), yielding strain SRmASB4.

I'he SRmA5B4 transposon marker was subsequently transferred into SRmA449 by transduction, yielding SRmAs85.

Mutant identification

To identify the point of insertion for individual TnS mutations, a modification of arì arbitrary PCR protocol was utilized, as previously described (Miller-Williams et al. 2006,

Poysti et al. 2007) (also see Chapter 4, page B9). Briefly, strains containing Tn5 or TnS-

820 inserts were purified and the genomic DNA was used as a template. Two sequential

PCR reactions were performed, the first using primers I i and 12 (Table 3.2), and the second reaction using primers l3 and 14 with the products of the first reaction as template.

The lìnal PCR product was gel isolated and sequenced using primer 13 (Table 3.2), to locate the point of insertion in the genome.

64 CI]API-ER 3. TRIOSE PHOSPHATE ISOMERASE

Triose phosphate isomerase assays

Iinzyme assays were performed on cells grornrn in rich medium or in defined medium containing either succinate, glucose or erythritol. Cells were grown overnight and resuspended in an extraction buffer containing 100 mM Tris, 5 mM 13-mercaptoethanol, and I rnM MgCl2 at pH 7.6. A French Pressure cell (16,000 psi) was used to disrupt cells with subsequent centrifugation to remove insoluble debris. TPI activity was assayed essentially as described (Lynch et al. 1975). Assays were performed in a total volume of I mI- containing 40 mM Tris (pFI 8.0), 0.2 mM NADH, excess L-ø-glycerophosphate dehydrogenase (5 U) and volumes of cell free extract from 20- I00 pL. The reaction was started by adding G3P to a final concentration of 2.5 mM. The conversion of G3P to DFIAP and then to a-glycerophosphate was monitored spectrophotometrically at 340 nm. Volumes of cell-free extract were assayed such that measured rates of activity increased linearly with increasing volume of extract. Specific activities are expressed as nmol NADH oxidized per minute per milligram protein. Protein concentration was measured as described using a protein determinaticln reagent (Sigma) and bovine serum albumin as a standard (Bradford 1976).

Symbiotic competence assays

Plant tests were carried out as previously described (Oresnik et al. 1994, Poysti et al. 2007)

(also see Chapter 4, page 92). Briefly, surface sterilized alfalfa seeds (Medicago satiua cv.

Rangelander) were sown onto 1.5% water agar plates and left to germinate. Sprouted seeds were then aseptically transferred to pots containing sterile sand and vermiculite with nitrogen-freelensen's plantnutrientsolution (Glazebrook andWalker 1991). Plants were inoculatedwith I0a to 10s bacteriaperplant. Plantswere harvested 28-35 days after

65 CI-IAPI'ER 3. TRIOSE PHOSPHATE ISOMERASE inoculation and dry weights of shoots cut at the cotyledons were used to assess symbiotic competence. Bacteria were routinely isolated fron-r nodules and tested for expected phenotypes. Competition for nodule occupancy was assessed by inoculating plants with a combination of wild t¡rpe and the strain to be tested, followed by enumeration of nodule occupants, as previously described (Poysti et aL.2007).

Nodulation kinetics

To test for differences in the number of nodules or rate of nodule formation on alfalfa plants inoculated with various strains of S. meliloúi, individual sterilized alfalfa seeds were sowrt onto Jensen's media l.5To agar slants in test tubes. After the seeds had sprouted, each test tube containing one plant was inoculated with S. meliloti culture cliluted in water, containing approximately l0:t bacteria. Four independent cultures of each strain were used to inoculate 25 plants, and any tubes containing dead or unusual looking plants were discarded before the end of the experiment. Test tubes were incubated in a growth chamber at a constant temperature and watered as required. Root nodules were counted and marked on each test tube at a number of time points for 25 days.

Results

Analysis of tpíA and, tpiB

As part of our ongoing interest in rhamnose catabolism and associated symbiotic phe- nolypes (Oresnik et al. 1998, Richardson et al. 2004), we have screened many S. meli-

Ioti mutants for the ability to catabolize rhamnose as a sole carbon source. A number of Tn5 induced Rm102l mutants unable to use rhamnose were isolated, including one

66 CHAPT'ER 3. '1'RIOSE PHOSPHATE ISOMERASE strain, SRmAlBs, which was only able to grow slowly on plates containing rhamnose and unable to grown on glycerol plates. All the mutations except for that of SRmAIBS were situated in a contiguous chromosomal locus (Systematic identifiers: SMc0232l - SMc02325 and SMc03000 - SMc03003; see Chapter 2 for details). This locus was found to be homologous to a rhamnose utilization locus previously reported in R. legumino- scuunl (Richardson et al. 2004). An SRmAl85 genomic DNA fragment spanning the TnS insertion and flanking DNA was amplified with an arbitrary PCR protocol. Sequenc- ing the fragment revealed that the TnS insertion was interrupting the putative triose phosphate isomerase encoding gene, tpiA (data not shor¡¡n). The genome annotation of S. meLiloli predicts that there are two chromosomal triose phosphate isomerase encoding genes, which have been annotated as tpiAl and tpiA2 (Systematic identifiers

SMcO1023 and SMcOl614, respectively) (Galibert et al. 2001). In this study, we have shornm that the two genes encode triose phosphate isomerases, but that there is a significant difference in function between them. We are therefore suggesting the names rplA and tpiB to replace tpiAl and tpíA2 respectively, as the new names conform more with the standard genetic nomenclature (Demerec et al. 1966). Our suggested gene names are used throughout this paper.

The gene tpiA is on the S. meliloti chromosome and appears to be isolated rather than transcribed within an operon (Figure 3.1). The gene tpiB is also located on the chromosome, and appears to be in a small operon downstream of a large operon which contains homologs of the erythritol utilization genes, eryA, eryB and eryC from B. abor- f¿¿s and R. leguminosarum (Sangari et al. 2000, Yost et al. 2006). Over the course of our characterization of triose phosphate isomerase in S. melilotl, strains carrying several independent tpiAand/or tpiÙ alleles were constructed (Table 3.1).

67 CHAPTER 3. TRIOSE PHOSPHA'|E ISOMERASE

Figure 3. I : Physical map of two S. meliloti Rm I 02 I chromosomal regions carrying either I) tpiA or 2) tpiB. Boxes represent predicted open reading frames with pointed ends indicating the direction of transcription.

Y0I02l Y01022 tpiA Y01024

PfCdiCted I:lypothetical CytochromeB Triosephosphate Hypolhetical fUnCtiOn t¡ansmembrane isomerase transmembrane protein 2) tc-- rpiB tpiB Y0l6I5 eryc Y0I6l7

Pfedicted fubose 5- Triose phosphate Transcriptional D-erfihrulose- I - Hypothetical fUnCtiOn phosphatc isomerasc rcgulator phosphatc protcin isomerase dehydrogenase

1 kbp

6B CI-IAPI-ER 3. TRIOSE PHOSPHATE ISOMERASE tpíAand, tpiB are needed for gluconeogenesis

Wild-type strain Rml02t and the constructed tpiA and tpiB mutant strains were tested for the ability to grow on a number of types of defined media containing single carbon sources (lhble 3.3). Mutations in tpiA caused a loss of the ability to utilize glycerol for growth as well as slow growth on rhamnose compared with the wild-type parent strain.

Mutations in tpiB only caused loss of the ability to use erythritol as a carbon source.

This correlates with the fact that tpiB is situated dornmstream of a predicted erythritol catabolism operon (Figure 3.1) and that tpiB dependent TPI activity is induced by erythritol (Table 3.4). Not only does the tpiA tpiB strain have a combination of the phenotypes of tpiA and tpiB strains, but having no triose phosphate isomerase genes causes a loss of the ability to utilize fucose, succinate, y-aminobutyric acid (GABA), acetate, arabinose, glutamate and lyxose. Strains with both tpiA tpiB mutations were able to grow with combinations of non-permissive carbon sources such as glycerol and succinate, or glycerol and arabinose. This is consistent with the hypothesis that gluconeogenesis occurs through FBA, as the permissive combinations of carbon sources supply both Di{,AP from glycerol, and G3P from arabinose or succinate (Arias and Martinez-Drets 1976, Fi- nan et al. l9BB, Stowers 1985).

Both tpiA and, tpiB encode triose phosphate isomerases

To confìrn the predicted triose phosphate isomerase functions encoded by tpiA and tpiB, cell-free extracts were prepared and assayed for enzyme activity. Cells were grown in flasks containing a rich medium or a defined medium with either glucose, succinate or erythritol as sole carbon sources. Due to their inability to grow extracts were not prepared for SRmA355 (tpíB) with erythritol or SRmA366 (tpiA tpiB) with either erythritol CHAPTER 3. 'I'RIOSE PHOSPHATE ISOII4ERASE

Täble 3.3: Relevant growth phenotypes of strains used in this work on defined media.

Growth with carbon source Iìclevanl Strain gcrìotypc Glco Suc Gly Rha lìry Iruc GAll^ Acc ,Ara Glu lìill Xyl l.)ry ììrnlO2l wild-rypc+r)++1'-f+-F++++++ SllrììA4l9tpitl++-'+++++++1-+l- Slìnr^stì4tpiß-È+l'+-+++++-l-++ Slìrìì^stìs tpitlrpill + - - + + +

" Grown on VMM media with 15 mM of a given carbon source. Abbreviations: Ery, erythritol; Glc, glucose; Gly, glycerol; Rha, rhamnose; Suc, succinate; Fuc, fucose; GABA, y-amino butyric acid; Ace, acetate; Ara, arabinose; Glu, glutamate; Rib, ribose; Xyl, xylose; Lyx, lyxose. I'Glowth plrenotype indicated by: *, same as wild type; f, slow compared to wild type; - no visible growtl-r.

70 CI-IAPTER 3. TRIOSE PHOSPHATE ISOMERASE

Table 3.4: Triose phosphate isomerase activities measured for different strains.

Activity (nmol/min/mg) " Relevant Strain genotype LBD GIc Suc Ery RmI021 wild-type 879+29 990+55 942+27 1569+117 SRmA327 tpiA 69t4 B5+9 92+4 415+2 SRmA355 tpiB 615+14 816+67 899+44 N.D.' SRnr.A.366 tpiAtpiB 6+I l+1 N.D. N.D.

¿ì Activities reported are the average of three enzyme assays. b Cell-free extracts from cultures growî with: Ery, VMM broth with l5 mM ery'thritol; Glc, VMM with l5 mM glucose; Suc, VMM with t5 mM succinate; LB, Luria-Bertani broth. ' N.D., Not determined, due to inability to grow with carbon source listed.

7l CI-IAP'TER 3. TRIOSE PHOSPI1ATE ISOMERASE or succinate as sole carbon sources.

Assays for TPI activity indicate that tpiA mutants had less than I0% of wild type activity when grown in rich medium or defined medium with either glucose or succinate as shown in Table 3.4. However, activity levels were approximately 25% of wild type activity when grown with erythritol. SRmA355 cell-free extract activity levels were similar to wild type l'or all growth media tested, while SRmA366 extracts had background levels of TPI activity. Taken together these data show that both tpiA and tpiB encode triose phosphate isomerase enzymes, and suggestthat tpiA may be constitutively expressed, providing most of the cell's TPI activity, whereas tpiB is induced by erythritol and expressed at low levels under the other conditions tested.

In víuocomplementation of tpiA

Because tpíBwas identified as a locus necessary for erythritol catabolism, and tpiB dependent TPI activity was induced by erythritol, it was hypothesized that induction of tpiB in a tpiA strain could rescue the ability to grow on glycerol. To test this, SRmA327 was streaked onto a number of plates of defined medium containing 15 mM glycerol and various amounts of erythritol (from 0.1 mM to 5 mM) to determine the lowest concentration of er¡hritol that could rescue the ability to grow on glycerol. SRmA327 was found to be able to grow as well as Rml021 with glycerol when supplemented with at least 0.4 mM er¡hritol. This amount of erythritol alone does not support the growth of

S. metiloti on agar plates and the tpiA tpiB mutant SRmA366 was not able to grow with glycerol supplementedwith er¡hritol. This indicates that tpiBis inducible byerythritol, and has in uiuoTPI activitywith the abilityto complementtpiA. Based on these results it was hypothesized that suppressors to tpiAwould be easily isolated with an up-regulated lplB. Although greater than I x 10r0 SRmA327 cel\s were plated on media with glycerol as

72 CI-IAP ER 3. TRIOSE PHOSPHATE ISOMERASE a sole carbon source, we were unable to isolate any suppressors over the course of several independent experiments.

C o n sti tutive ly exp re s sed tp íA canno t c o mp I ernent tp íB

As both tpiAand tpiBwere found to have ín uitro TPI activity, we were interested in why lplB mutants were unable to grow with erythritol as a sole carbon source. To determine if tpiB was directly required for erythritol catabolism, we complemented tpíAand tpiB mutations with constitutivelyexpressed TPI genes. Both tpiAand tpíB were cloned into a broad host range vector, pRK7Bl3, such that the genes were constitutively expressed

(Jones and Gutterson l9B7). The resulting plasmids, pNPI66 and pNP167 were mated into various strains and tested for their ability to use glycerol, succinate, arabinose, glutamate and erythritol as sole carbon sources (Table 3.5). Constitutively expressed rpiB was able to complement tpiA, tpiB, and tpiA tpiB mutants for growth on all the carbon sources tested. Constitutively expressed tpiA was able to complement the inability of tpiA mutant strains to grow on glycerol and the inability of tpiA tpiB mutants to grow on glycerol, succinate, arabinose and glutamate. In contrast to tpiB, tpiAdid not complement the inability to use erythritol caused by tpiB mutations. This indicates that although tpiA and tpiB have the same enzyrne activity, tpiB is specifically required for erythritol catabolism in S. meliLoti. We note that when the same experiment was performed with SRmASBa GpiB) and SRmASBS (tpiA fpzB), pNPl67 (tpiB+) complementation resulted in growth on erythritol that was slower than wild type. We are attributing this to the Tn5-820 insertion in SRmASB4 and SRmAS85 being polar on rpiB (Figure 3.1) which has been shor¡m to be necessary for erythritol catabolism in R. leguminosarunt

(Yost et al. 2006). The complementing cosmid containing the genomic DNA surrounding tpiB, pNP163, was able to complement each strain, restoringwild-type growth.

/J CI]APTER 3. TRIOSE PHOSPHAT-E ISONIERASE

l'able 3.5: Complementation of TPI mutant strains by plasmid borne tpiAand tpiB.

Relevant genotype Growth with carbon source Strain (chromosomal/plasmid) Glca Gly Ery Suc RmI02t tpiAt tpiB+ I - lìml02l(pNP166) tpiA+ tpiÙt I tpiA+ Rmt02l(pNPt67) tpiA+ tpiÈ I tpiÈ SIìm4327 tpiA tpiÙ+ I - SRmA327(pNPl66) tpiA tpiB+ I tpiA+ SRmA.327(pNPl67) tpiA tpig+ I tpiÈ SRmA355 tpi.# tpiB /- Slìm4355(pNPt66) tpiAl t:piB I tpiA* SRmA355(pNPl67) tpiAt tpiB ltptY SRmA366 tpiA tpiB /- SRmA366(pNP166) tpíA tpiB ltpi,+ SRmA366(pNP167) tpiA tpiB ltpiÈ i' Grown on VMM media with given carbon sources. Abbreviations: Ery, er¡hritol; Gly, glycerol; Suc, Succinate. t' Growth phenotype indicated by: *, same as wild type; - no visible growth.

74 C 11 A P7- E II 3. I'R I OS E P H OS PHATE I SO M ERASE

Symb iotic competence phenotypes

Il. Leguminosarumrhamnose catabolism mutants have been shoi,r¡n to have a deficiency in competition for nodule occupancy (Oresnik et al. l99B). It was hypothesizedthat tpiA mutants would have a similar phenotype, based on their inability to grow with rhamnose at the same rate as wild-type S. meliloti.In addition, there have been many reports of defective symbiosis associated with the inabiliry to use succinate as a sole carbon source (Bolton et al. 1986, Finan et al. 1991, Gardiol et al. 1982, Hornez et al. 19Bg), suggesting thar. ryiA rplB strains may be unable to symbiotically fix nitrogen, or only have parLial activity.

lb test the hypothesis that tpiA mutants would be compromised in their ability to compete with wild type for nodule occupancy, alfalfa were co-inoculated with SRmA327 and RmI02I. The relative amount of each strain recovered from nodules was not significantly different from the amount of each strain found in the inoculum (data not shor¡¡n).

Therefore, the ability to compete for r-rodule occupancy is not compromisedin a tpiA strain, under the conditions tested.

During initial plant growth experiments, it was qualitatively observed that some of thc plants inoculated with strains carrying a tpiA mutation wele growing slightly taller than wild type inoculated plants. The plants were dried and weighed, showing that on average, they were significantly different from wild lype inoculated plants (Table

3.6). Further repetitions yielded similar results with tpiA inoculated plants having an increased dry weight. The average increase in dry weight over four trials was I Ig*3.5%.

A limitation of these studies was that in each case, plants grown in groups of i0 plants per pot were dried and weighed together, concealing the weight of individual plants. In order to rigorously test for a difference in dry weights, we grew many more plants in-

75 C]IAPTER 3. TRIOSE PHOSPHATE ISOMERASE

Table 3.6: Representative plant dry weights from growth trials of alfalfa in nitrogen free media inoculated with various S. meliloti strains. See text for a discussion of the results.

Inoculum Relevant Number Average dry weight 7o wild type strain genotype of plants (mg/plant*S.D.") dry weight

Rm1021 wild-type 30 34.9L4.3 100 SRmA327 tpiA 29 42.9+3.9 r23 SRmAs84 tpiB 24 38.2+6.8 r09 SRmASB5 tpiAtpiB 27 i3.3+2.6 3B SRmA600 tpiAtpiB 29 14.5+1.1 42 Rm5439 pckA 29 14.5+2.2 42 Uninoculated 28 5.6tr.0 t6

¿r Average dry weight t standard deviation

76 C LI A PI'EII 3. TR I O SE PH OS PHATE I SO M ERASE oculated with either Rm1021 or SRmA449 and harvested plants individually, drying and weighing them separately. The data from this assay of approximately g0 individually weighed plants per inoculum strain revealed no significant different between Rml02t and SRmA449 inoculated plants (data not shown).

Plants inoculated with tpiB strains did not show any significant difference in dry weight compared with wild type. However, plants inoculated with tpiA fpiB strains reached dry weights of approximately 50% of wild type inoculated plants (Table 3.6). This reduced nitrogen fixation phenotlpe is very similar to the phenotype described for the succinate mutant, pckA(Finan et al. 1991, Østerås et al. 1995). Dryweights of plants inoculated with Rm5439 (pckA) were not significantly different from the tpiA tpiB mutant

SRmASBS (Table 3.6). In order to test whether the reduced nitrogen fi*ing phenotype was due to bacteria that had undergone a supprcssion or reversion event, SRmASBS was isolated from the nodules of plants from one experiment, and used as an inoculum in another experiment. The SRmAs85 isolated from nodules had all the expected genetic markers and phenotypes intact, and resulted in similar plant dryweights for the second experiment, confirming that tpiAtpiBhas a reproducible s)..rnbiotic phenotype.

Increased nodulatio nby tpíA

To determine if an increased rate of nodulation or number of nodules was associated with the initially observed increased dry weights of plants inoculated with tpiA strains, individual alfalfa plants were grown in glass test tubes. Test tube sprouted plants were inoculated with a preparation of either Rml021, SRmA449, SRmASB4, SRmA5BS or left uninoculated as a control. Our results indicate an increase in the number of nodules per plant inoculated with SRmA449 (Figure 3.2). The number of nodules per plant was not normally distributed, so a two-tailed Mann-Whitney test was used to compare the CHAPT'ER 3. TRlOSE PHOSPHATE ISOMERASE

Figure 3.2: (A) Number of nodules per plant inoculated with either Rm102I, SRmA449, SIìmASB4 or SRmASBS. The number of nodules shorn¡n for each strain is the average of the nodules counted on 19 to 22 alfalfa plants monitored over a period of 25 days. (B) Percentage of plants referred to in (A) with at least one root nodule.

c ão A o o EI o €10o c a- o -a o ! .t E f C o5 o o .: --;-^o') o -:Ä'î2:--"¿ o----o Rm1021 'ida..F.{- E.'..ÐSRmA449 o---ô SRmA584 a- - Á SRmA585

1-0 L5 25 Days after inoculation

oC o f E zo ñ

Days after inoculation

7B CHAPT-ER 3. TRIOSE PHOSPHAT'E ISOMERASE means. Based on l9 to 22 individual plants per strain, the average number of nodules per plant was found to be significantly different (P<0.05) between Rm102I and

SIìm4449 from day 14 onwards. No significant difference in the number of nodules was measured between Rml021, SRmASB4, or SRmA5BS. Plants inoculated with SRmA449 also appeared to nodulate slightly earlier than the other strains, based on the percentage of plants nodulated at each time point sampled. These observations also indicate that the decreased dryweights accumulatedby tpíA tpíBinoculated plants is not due to a decrease in the number of nodules per plant.

Discussion

Despite the central role of TPI in carbon metabolism, very few ryi mutants have been characterized in prokaryotic organisms. In fact, to the best of our knowledge mutations in tpi have only been previously described in three organisms (Anderson and Cooper

1969, Meijer et al. 1997, Pahel et al. 1979, Zhenget al. 2006). Of these, two are relatively closely related (Anderson and Cooper 1969, Pahel et al. i979, Zhenget al. 2006), and the other is associated with autotrophic growth (Meijer et al. 1997). Moreover, well characterized mutations have only been studied in a single case (Zheng et al. 2006). Given that the number of genome sequences predicting multiple metabolic enzyrne activities is increasing, it is imperative that a functional characterization of central metabolic pathways is carried out and not assumed. Our analysis of the two putative S. meliloti genes is only the second study to show two functional tpi genes within any organism, and is the first within the Rhizobiaceae. As many organisms do not possess two TPI encoding genes and their functions were initially unclea¡ we decided to investigate the physiological roles of tpiAand tpí&.

79 CLIAPTER 3. TRIOSE PHOSPHATE ISOMERASE

The errz)¡me activities we have reported are consistent with tpiAbeingconstitutively expressed, and we hypothesize that it primarily plays a gluconeogenic role (Table 3.4).

S. meliloti microarray data are consistent with this and have shor,rm that tpiA is expressed at similar levels during growth with glucose, succinate, rich medium and in the bacteroid (Balnett et al. 2004). Our work indicates that tpiB is induced by growth with erythritol and is required for erythritol catabolism. S. melilofl contains homologs of the e,y operon, which suggests that erythritol is catabolized to DFIAP as it is in B. abortus

(Sperry and Robertson 1975). We have shorn¡n that constitutively expressed copies of tpiA cannot complement the inability of tpíB strains to grow with erythritol (Table 3.5).

Interestingly, Sperry and Robertson (1975) noted that TPI activity could not be completely uncoupled from 3-keto-L-erythronate-4-phosphate decarboxylase activity while characterizing erythritol catabolism in B. abortus. Taken together, this is consistent with the possibility of TpiB and the enzyme(s) of erythritol catabolism forming a metabolic complex. The existence of metabolic complexes and sequential reactions has been proposed f'or a number of pathways, including glycolysis (Srere l9B7) .

Each of the genomes of B. abortus (Chain et al. 2005, Halling et al. 2005) , Brucella melitensis (DelVecchio etal.2002), Brucella suls (Paulsen et al. 2002), Mesorhizobium lo¿i (Kaneko et al. 2000), Rhizobiumetli(Gonzâlez et al. 2006), R.leguminosctrum fYoung et al. 2006), and S. meliloti (Galibert et al. 2001) have homologs of tpiAand tpiB.In each case, tpiBis always followed by rpiB directly dovrrnstream, and in all but the genome of R. etli, IpiB and rpiB are directly downstream of operons containing homologs of the erythritol catabolism genes eryA, eryB, eryC, and eryD.TLris leads us to hypothesizethat tpiB may be necessary for erythritol catabolism in each of these organisms besides R. etli, which does not grow with erythritol and does not carry the erythritol genes (Gonzâlez

BO CIIAPT'EI] 3. TRIOSE PHOSPHATE ISONIERASE et al. 2006, Yost et al. 2006). In addition to our report, tpiBhas also been confirmed to be necessary for growth with erythritol in R. leguminosarum (Yost et al. 2006).

The fact that the S. meliloti genome annotation predicts that there are no TPI encoding genes other than tpiAand tpi&, combined with the observation that TPI activity was essentially absent in SRmA366 (Table 3.4), suggests that tpiAtpiB strains are unable to perform gluconeogenesis. This is consistent with the fact that tpíA tpiB strains are unable to grow with carbon sources that require gluconeogenesis (ie. succinate,

GABA, glutamate, arabinose, acetate, erythritol, glycerol). That the low levels of TPI activity nreasured in lpiAmutants (Table 3.4) are able to allow gluconeogenic growth was sonrewlrat surprising, but is consistent with the report of E. colí tpi revertants growing gluconeogenically with only 5To of wild type TPI activity (Anderson and Cooper 1969).

Although gluconeogenic growth is possible in a tpiA strain, there does not seem to be enough TPI activity to allow catabolism of glycerol, which is thought to proceed through glycerol kinase and glycerolphosphate dehydrogenase forming DHAB requiring TPI for further catabolism (Arias and Martinez-Drets 1976, Stowers 1985).

We hypothesized that tpiBwould be negatively regulated by the adjacent deoR-type regulator SMcO1615. Our inability to isolate spontaneous mutations leading to up-regulation of tpiB by plating on minimal media with glycerol was surprising. The fact that over l0r0 cells were plated and no suppressors were isolated suggests that the regulation of. tpiB may not be solely due to negative regulation by SMcO1615. We note that the erythritol catabolism operon upstream of tpiBhas been sho',rm to be induced by xylitol, adonitol, sorbitol and er¡hritol (Mauchline et al. 2006). As tpiB function is required for erythritol catabolism, these two operons may be coordinately regulated, suggesting a regulatory mechanism involving more than one event.

BI CIlAP'|ER 3. TR]OSE PHOSPHATE ISOMERASE

The first mutation in tpiA was isolated by screening for mutants with an impaired abiliry to utilize rhamnose as a sole carbon source. This led us to test it for a s]¡mbiotic competition defect, as has been reported for R. leguminosarum rha mutants (Oresnik et al. i99B). Data from three independent experiments showed that there is no difference between the two strains for nodulation competition (data not shown). However, tpiA does not cause a complete loss of the abiliry to utilize rhamnose as a sole carbon source, so whether rhamnose catabolism mutants of S. meliloti are less competitive than wild type remains an open question.

The observation that tpiA inoculated plants grew larger than Rm1021 inoculated plants was unexpected (Table 3.6). This prompted several experiments of drying and weighing plants inoculated with SRmA327, Rml}2l and those left uninoculated (data not shonm). Although one large experiment did not support the initial observations, four smaller scale experiments did. in addition, we have measured a small increase in the number of nodules per plant inoculated with SRrrì4449 compared to RmI021 (Fig- ure 3.2). The difficulty in reproducibly demonstrating this phenotype suggests that there may be an uncontrolled variable affecting the outcome, or that we are simply observing an artifact. Growth experiments performed under different conditions or on a larger scale may be able to conclusively test the hypothesis that tpiAinoculated plants grow larger than wild type.

Consistent with the reported literature for S. meliloti mutants unable to grow on succinate, the tpiA rplB mutant was tested and shor¡¡n to have a compromised symbiotic phenoty¡pe. !1/hen inoculated on alfalfa, plant dry weights reached less than 50% of wild type inoculated plants, similar to the phenotype of S. melilori gluconeogenic pckA mutants (l.-inan et al. 199I, Østerås et al. 1995) . As tpiA rpiB strains are unable to perform

82 C]lAPT'ER 3. TRIOSE PHOSPHATE ISOMERASE gluconeogenesis, the fact that symbiotic nitrogen fixation is not abolished suggests that gluconeogenesis is beneficial, but not required in symbiosis with alfalfa, possibly due to small arnounts of plant supplied sugar reaching the bacteroid. Consistent with this hlpothesis, the gluconeogenic enzyme phosphoenolpyruvate carboxykinase (PCK) encoded by pclcA is not found at detectable levels in S. melilori bacteroids, although mutations in pckA cause reduced nitrogen fixation (Finan et al. t99l). On the other hand,

R. Leguminosarum bacteroids were found to have PCK activity, and a PCK deficient mutant was found to have no difference in symbiosis (McKay et al. I9B5). Mutations in pckA of the broad host range species, Rhizobium NGR234, were found to cause several phenotypes, ranging fiom Fix- to low rates of nitrogen fixation, depending on the host species (Østerås et al. 1991). It has been suggested that the strain and host dependent differences in the severity of symbiotic phenotypes of bacterial gluconeogenesis mutants may be due to differences in peribacteroid membrane sugar permeability (Poole and Allaway 2000).

In this study, we have shorn¡n that there are two TPI encoding genes in the S. melilotl genome with different functions, as tpiA cannot substitute for QiB. The gene tpíAis

Irypothesized to be constitutively active in a gluconeogenic role, while tpiB is required f'or erythritol catabolism by an unknonm mechanism. In addition , tpiA tpiBmutants Iose the ability to perform gluconeogenesis but retain the abiliry to s1'rnbiotically fix nitrogen at a lower level than wild rype.

B3 Chapter 4 Sinorhizobium meliloti pSymB carries genes necessary for arabinose transport and catabolism

Summary

This chapter was published in part in Microbiology (Vol. L53:727-736,2007). The work was completed by Nathan Poysti with Erin Loewen andZexi Wang providing technical assistance.

Arabinose is a kno'nm component of plant cell walls and is found in the rhizosphere.

In this work, a previously undeleted region of the megaplasmid pSymB was identified as encoding genes necessary for arabinose catabolism by Tn5-820 random mutagenesis and subsequent complementation. Transcription of this region was measured by

[3-galactosidase assays of Tn5-820 fusions, and shoin¡n to be strongly inducible by arabinose, and moderately by galactose and seed exudate. Accumulation of [3H]arabinose in rnutants and wild type was measured, and the results suggested that this operon is necessary for arabinose transport. Although catabolite repression of the arabinose genes by succinate or glucose was not detected at the level of transcription, both glucose and galactose were found to inhibit accumulation of arabinose when present in excess. To determine if glucose was also taken up by the arabinose transport proteins, ItaC]glucose uptake rates were measured in wild type and arabinose mutant strains. No differences in

B4 CI'IAPTER 4. ARABINOSE CATABOLISM AND TRANSPORT glucose uptake rates were detected between wild type and arabinose catabolism mutant strains, indicating that excess glucose does not compete with arabinose for transport by thc same system. Arabinose mutants were tested for the ability to form nitrogen fixing nodules on alfalfa, and to compete with wild type for nodule occupancy. Strains unable to utilize arabinose did not display any symbiotic defects and were not found to be less competitive than wild tlpe for nodule occupancy in co-inoculation experiments. More- over our results suggest that other loci required for arabinose catabolism exist, including a gene encoding arabinose dehydrogenase.

Introduction

Sinorhizobium melilori is a Gram-negative soil bacterium capable of forming a sym- lriotic nitrogen fìxing relationship with legumes such as alfalfa. S. meliLoti infects and forms nodules on plant roots in a process involving complex signal exchange between thc plant and bacteria (Spaink 2000). Surrounding the roots of plants is a thin zone called the rhizosphere where plant secretions enrich the soil environment with organic compounds (Bowen and Rovira i976). The ability to effectively utilize energy sources found in the rhizosphere may play an important role in the long-term survival of soil microbes such as S. meliloti.

Arabinose supports the growth of many rhizobia (Stowers I9B5), but is catabolized by different pathways in slow and fast growing species. In both groups the first two en- z,yrnatic reactions are conserved. Following the initial reactions the pathways diverge, producing pyruvate and glycoaldehyde in the slow growers, and a-ketoglutarate in the läst growers (Dilworth et al. i986, Duncan and Fraenkel 1979, Duncan 1979, Pedrosa andZancan 1974, Stowers l9B5). As S. melilofi is a member of the fast growing rhizobia,

B5 C I I ¡l P7- E Il 4. A RA B I N O S E CATAB O LI S M A N D TRA N S PO RT arabinose is catabolized to ø-ketoglutarate (Duncan and Fraenkel 1979). To our knowledge, the only characterized gene involved in this pathway of arabinose catabolism has been the arabinose dehydrogen ase of Azospirillum brasíliense (Watanabe et al. 2006) .

Several sugars, including the pentose arabinose have been detected in the exudate of pea plants (Knee et al. 200I). In fact, arabinose makes up a large proportion of the sugars found in pea plant exudate and is present at physiologically relevant concentrations, shor¡.'n to be enough to support the growth of Rhizobium leguminosarum (Knee et al. 2001). As well as being secreted, arabinose has also been found as a component of plant cell walls (McNeil et al. I9B4). Despite the prominence of this sugar around plant roots, genetic characterization of the determinants necessary tor S. meLiloúi arabinose catabolisrn are not described within the literature.

Loss of the ability to catabolize carbon sources has been correlated with defective symbiosis or decreased ability to compete for nodule occupancy in both S. meliloti and

R. leguminosarum. Although non-competitive strains retain the ability to form a symbiotic relationship with the plant host, wild-type parent strains dominate for nodule occupancy when co-inoculated with less competitive mutant strains. Examples of less conrpctitive mutants are the myo-inositol and rhamnose catabolism mutants of R. leg- utninosarum (Fry et al. 2001, Oresnik et al. l99B) and the proline dehydrogenase mutant of S. meliloti (Iiménez-Zurdo et al. 1995, 1997). In addition to arabinose, rhamnose has also been reported to be a constituent of plant secretions and cell walls (Knee et al. 2001,

McNeil et al. 1984), suggesting a possible link between the ability to utilize rhizosphere carbon sources, and nodulation competency. It has also been reported that an S. meliloti TnS arabinose catabolism mutant had a delayed and compromised symbiotic phenotype (Duncan 198t), although the nature of this mutation appears to have remained

B6 CIIAPTER 4. ARABINOSE CATABOLISI/Í AND TRANSPORT uncharacterized.

In an effort to characterize S. melíloti arabinose catabolism genetically, mutants unable to utilize arabinose as a sole carbon source were isolated and we report the existence of a locus necessary for arabinose transport and catabolism on the megaplasmid pSymB. In addition to identifying an arabinose catabolism operon, experiments were performed to determine the effects of catabolite repression on arabinose gene expression, the possibility of gene induction by seed exudate, and the symbiotic competence associated with arabinose catabolism mutations.

Methods

Bacterial strains and growth conditions.

T'he bacterial strains and plasmids used and generated in this work are listed in Table

4.L S. rneliloti strains were routinely grornrn at 30'C on complex (LB) or defined (VMM) media as previously described (Sambrook et al. 1989, Vincent 1970, Richardson et al.

2004). Carbon sources used in defined media were filter sterilized and used at a final concentration of 15 mM. Seed exudate was prepared as previously described, and filter sterilized (Mulligan and Long 1985). Two independent preparations of exudate were used in this work, at lïVo (v/v). \¡Vhen required, antibiotics were used at the following concentrations: chloramphenicol (Cm), 20 Vg ml-r; gentamicin (Gm) 20 or 60 pg ml-'r; kanamycin (Kan), 50 pg ml-r; nalidixic acid (Nal) t5 F.g ml-'; neomycin (Nm),

200 pg ml-r; rifampicin (Rif), 100 pg ml-ì; streptomycin (Sm) 200 or 600 H.g ml-'; tetracycline (Tc) either 5 or 10 pg ml-t. Growth in broth culture was routinely monitored spcctropho tometrically at OD6se.

87 CI.]APTER 4. ARABINOSE CATABOLISM AND TRANSPORT

Table 4.1: Bacterial strains and plasmids.

Strain or plasrnid Relevant characteristics* Reference

S. nteliloti Rml02l su47 srr-21, smr Meade et al. 1982 RmC2l2 RmB501 /ac, Sm' Glazebrook and Walker I991 SRrnA240 RmG2l2 araF I ::Tn5-820, Nm'' This work SRmA482 RmG2l2 araD2::Tn5-820, Nmr This work Slìm4502 Rml02I araD3::Ttt5-820, Nm' This work SRrnA503 Rml02 I araA4::Tt't 5-820, Nm' This work SRrnA505 Rml 02 I araþ-5::Tn5-820, Nm' This work SRrnA506 Rm102l araÛ6::Tn5-820, Nm' This work SRmA507 Rml021 araDT;;Tn5-820, Nm' This work SRmA50B Rml02l araB&:.Ttl,5-B20, Nm' This work

E. coli MM294A pro-82 thi-1 hsdRl7 supE44 Finan et al. 1986 MT607 MM294A recA-56 Fina¡r et al. l9B6 MT6r6 MT607 (pRK600) Finan et al. 1986 M1-620 MT607(pRK600::Tn5- 820) T. Finan via T. Charles EcA100 MT607 Rifl derivative This work EcAt 01 MT607::Tn5-820, Kanr Clark et al. 2001 I.IBIOl hsdS20 recAl3 rpsL20 Lab collection DH5ø gyr,496, Nal' Invitrogen

Plasmids pRK600 pRK2013 npt'.:Tn9, Ctn' Finan et al. 1986 pRK602 pRK600::Tn5, Cm', Nmr Finan et al. 1985 CX-I S. meliloti cosmid bank T. Charles pPHlll IncP plasmid, Gm' Beringer et al. 1978 pZWl CX-l derived arabinose This work complementing cosmid, Tc' pZW2 pZWl araD2'.:Tn5-820 This work pZW3 pZWl araBB'.:Tn5-820 This work pzw4 pZWl araD3'.:Tn5-820 This work pZWs pZWl araA4::Tn5-820 This work pZWT pZWl araFS:'.Tn5-820 This work pZtNB pZWl araE6:.Tn5-820 This work 1lZW9 pZWl araDT::Tn5-820 This work Antibiotic abbreviations: Ap, ampicillin; Cm, chloralnphenicol; Gm, gentamicin; Kan, kanarnycin; Nal, nalidixic acid; Nm, neomycin.

BB CI-IAPTER 4. ARABINOSE CATABOLISIVT AND TRANSPORT

Genetic techniques.

Conjugations of S. melilotí and E. coli were ca¡ried out using the mobilizing strain MT-

616, as previously described (Finan et al. lgBB). Mutagenesis of S. meliloti RmGZ12 was perf'orrned by introducing pTH270 by conjugation from E. coli strain MT620 into

IìnGZ12. Transposon induced mutants were selected with Sm Nm and screened for the inability to grow with arabinose as a sole carbon source. Cosmids complementing the resulting arabinose mutations were isolated from the S. meliloti cosmid bank CXI by conjugating the bank en masse into SRmA240 and plating transconjugants on minimal media containing arabinose and Tc. Cosmids were rescued from complemented arabinose mutant strains by conjugation with DHSa and subsequent Nal Tc selection. Sat- uration mutagenesis was carried out by introducing complementing cosmid pZWl into

EcA101, subsequentlyconjugating the cosmid into EcA100, and selecting for co-transfer of the transposon and the cosmid. Individual mutagenized cosmids were screened for loss of their ability to complement arabinose catabolism mutants. Marker exchange of cosmid borne arabinose Tn5-820 mutations was performed as previously described

(Glazebrook and Walker 1991). Putative arabinose catabolism mutations were subsequently transduced into RmG2I2 or Rml02l prior to being used experimentally. Trans- ductions were carried out using phage OMl2 as previouslydescribed (Finan et al. 1984).

Mutant identification.

'tb identify the point of insertion for individual Tn5-820 mutations, a modification of an arbitrary PCR protocol was utilized (Caetano-Anollés 1993, Raffa and Raivio 2002).

Strains containing Tn5-820 insertions were purified and the genomic DNA was used as a template. A low stringency PCR amplification (annealing temperature of 45"C) using

B9 CITAPl'ER 4. ARABINOSE CATABOLISM AND TRANSPORT the primers IS50(1) (5'-CACGATGAAGAGCAGAAG-3') and DGEN(I) (5'-GGCCACGCG-

TCGACIAGTCAGNNNNNNNNNNACGCC-3') was carried out. The products of this reaction were subsequently amplified with a higher stringency PCR reaction (annealing temperature of 60"C) using the primers IS50(2) (5'-TAGGAGGTCACATGGAAGTCAGI\1--

3') and DGEN(2) (5'-GGCCACGCGTCGACTAGTCAG-3'). The primers IS50(1) and IS50-

(2) are complementary to the end of the IS50 region. The DGEN(2) primer is complementary to the 5' end of DGEN(l), such that the second PCR reaction amplifies weak products generated in the first reaction. The products of the second PCR reaction were gel isolated and sequenced using the IS50(2) primer. Sequencing reactions were per- f'ormed with an r\BI automated sequencer at the University of Calgary Core DNA facili- ties. Sequence datawere trimmed of IS50 sequences and localized within the S. meliloti genome using the BI-¡\STN program (Altschul et al. 1997) and the S. meliloti genome database (http://bioinfo.genopole-toulouse.prd.fr/annotation/iANT lbacterialrhime).

Arabinose dehydrogenase assay.

Cultures were grown in flasks containing defined media with 15 mM glycerol and 15 mM arabinose as carbon sources. Cultures were grown overnight with shaking, pelleted by centrifugation (4,000 x g, l0 min), and resuspended in 2 ml g-t extraction buffer containing 100 mM Tris-l-ICl (pH 7.6), 5 mM B-mercaptoethanol, and I mM MgCl2. Cells were disrupted by two passages through a French Pressure cell (16,000 psi). The extract was subsequently pelleted (10,000 x g, 30 min, 5'C) retaining the supernatant. Aliquots of cell frec extracts were frozen at -70'C. Assays for arabinose dehydrogenase were performed essentially as previously described (Pedrosa andZancan I974). The assay was performed in a total volume of 0.5 mL and consisted of 100 mM Glycine buffer (pH 9.5),

500 pM NAI)+, and a volume of cell free extract from 10-50 pL. The reaction was started CLIAPTER 4. ARABINOSE CATABOLISM AND TRANSPORT by adding L-arabinose to a final concentration of l0 mM, and the absorbance at 340 nm was monitored with a spectrophotometer. The rate of the conversion of arabinose to arabino-y-lactone was calculated using the extinction coeffecient of NADH at 340 nm

(6.22x 10 iì mM-r cm-I). 'Ihe rate of conversionwas thennormalized to the protein content of the cell free extract. Protein content of extracts was nìeasured using Bradford's protein determination reagent and BSA as a standard.

B -galactosidase assays.

Cells to be assayed were induced by growing cultures overnight in defined media containing the appropriate carbon source(s). Overnight cultures at mid log phase were then subcultured into fresh media of the same type, and groinrn to an OD6s¡ of approximately

0.3 bef'ore being assayed as previously described (Miller 1972).

Arabinose and glucose transport assays.

Sugar uptake was measured using [I-3H]arabinose (370 GBq mmol-r; American Ra- diolabeled Chemicals), Il-r4C]arabinose (2.035 GBq mmol-r; American Radiolabeled

Chemicals), and [U-raC]glucose (l l.B GBq mmol-r; New England Nuclear). Assays were performed as previously described (Richardson et aL.2004) with the following modifìca- tions. Cells were grown in defined media containing the appropriate carbon source(s), washed twice, then resuspended in defined media with the original carbon source(s)

(omitting arabinose or glucose as necessary) at an ODooo between 0.1 to 0.3. Uptake of

Iabeled sugars was measured by adding label to bacteria suspensions (final concentrations of either4 nM or I pM [3H]arabinose, I ¡rM ['aC]arabinose, and 120 nM ['aC]glucose) and filtering 0.25 ml samples through a Millipore 0.45-¡rm Hv filter using a Milli- pore vacuum manifold. Samples were withdrawn at time points from l0 to 40 seconds.

9t CIIAPTER 4. ARABINOSE CATABOLISM AND TRANSPORT

Radioactivity retained on the filters was counted using a Beckman LS 6500 liquid scintillation counter. Radioactive uptake was normalized to total cell protein, measured by an enhanced Lowry protein assay (Stoscheck i990).

Syrnbiotic profi ciency assays.

Plant experiments were carried out as previously described (Oresnik et al. 1994). Briefly, alfallä seeds (Medicago satiua cv. Rangelander) were surface sterilized by a 20 minute

70t/o ethanol treatment followed by 30 minutes in 1% sodium hypochlorite. The seeds were then washed with at least 10 volumes of sterile distilled water and sornm onto I .5% water agar plates. After 2 days, germinating seeds were aseptically planted into a sterile sand-verrniculite mixture containing nitrogen-free Jensen's plant nutrient solution

(Glazebrook and Walker I991). Plants were inoculated after another 2-3 days with cultures of S. meliloti that had been grornrn overnight in LB broth and diluted one hundred fold with sterile distilled water to approximately 104 to 10s bacteria per plant. Plants were scored 2B-35 days after inoculation. SSrmbiotic proficiency was determined visu- ally by comparison with wild type inoculated and uninoculated control plants. Bacteria were isolated from several randomly selected surface sterilized nodules (l minute of ITo sodium hypochlorite exposure) to confirm the presence of appropriate genetic markers, ensuring that the inoculated strain was present in plant nodules.

Symbiotic competition assays.

Sl.rnbiotic competition was assessed by germinating and planting seeds as described above, then inoculatingwith a mixture of two bacterial strains containing a total of l0a to 10s bacteria per plant. Cultures of wild type and mutant were grown and diluted to the same OD6ee. A mixed inoculum was made by combining equal volumes of the diluted

92 CI-IAPTER 4. AIIABINOSE CATABOLISM AND TRANSPORT cultures, then diluting the mixture by 10-2 with sterile distilled water. Serial dilutions of the mixed inoculum were spread plated on LB agar plates. At least I00 colonies were patched from the spread plates onto LB agar plates containing the appropriate antibiotics for strain differentiation. A-fter 28-35 days, plants were harvested, and root nodules were manually removed. At least 50 individual nodules per experiment were sterilized

(1 minute of l% sodium hypochlorite exposure), and rinsed in sterile distilled water. In- dividual nodules were then crushed in i00 pl of sterile distilled water. l0 pl aliquots of each nodule extract were spotted onto LB agar plates containing the appropriate antibiotics for strain differentiation. Differences in the percentage of the mutant strain found in the inoculum and nodule extracts were used to assess competitive ability. Data from the competition experiments were evaluated for statistical significance using both bi- nomial probability and a Fisher's Least Significance Difference (LSD) assuming that a P value of less than 0.05 indicated a difference in strain competitiveness.

Results

Isolation of arabinose catabolism mutÍrnts.

As part of an ongoirìg interest in S. melilofi functional genomics and carbon catabolism pathways that may play a role in rhizosphere survival, strain RmG212, a Lac- derivative of Rml02l (Glazebrook and Walker 1991), was subjected to a random mutagenesis with the transposon Tn5-820 that is capable of generating transcriptional lacZ fusions

(Simon et al. t 989). Mutants containing transposon insertions were screened for a lack of the ability to grow on a number of sugars as a sole carbon source, including arabinose. Strain SRmA240 was isolated from this mutagenesis, and the transposon insertion point was identified by sequencing PCR amplified DNA spanning the 3' end of the TnS- CI.IAPTER 4. ARABINOSE CATABOLISM AND TRANSPORT

820 IS50 element and the flanking genomic DNA. A subsequent BL\STN search of the

S. meliloti genome database showed that Tn5-820 is inserted as a transcriptional fusion in iluDí (Systematic identifier SMb2OB90), suggesting that this gene is necessary for the catabolism of arabinose (data not sho'o¡n)

Subsequent complementation of SRmA240 with the CXI cosmid bank, and a satu- ration mutagenesis of the isolated cosmid pZWI with Tn5-820 provided the rest of the arabinose catabolism mutations described in this work (Table 4.I). Each cosmid mutation that was unable to complement SRmA240 for growth on arabinose was integrated into either Rm102l or RmG212 and subsequently retransduced into a wild type to elim- inate the plasmid pPHtII and the possibility that a secondary mutation arose during marker exchange. Transductants were purified and characterized with respect to carbon utilization phenotype. An arbitrary PCR method was used to generate a fragment which was sequenced to verify the point of insertion. Figure 4.1 shows a schematic diagram of the isolated mutational insertion points and the predicted ORFs from the S. meliloti genome sequence database. AII mutants were tested for the ability to grow on defined media with the following sugars as sole carbon sources: arabinose, glucose, fucose, lyxose, xylose, ribose, glycerol, galactose, rhamnose, succinate and erythritol. No growth defects other than those for arabinose were detected in the isolated mutants.

Analysis of arabinose transcriptional unit.

Each non complementing mutagenized arabinose cosmid was mated back into every arabinose mutant isolated, and tested for the ability to complement arabinose utilization. As none of the arabinose mutant strains were complemented by any of the mutagenized cosmids, we conclude that the predicted genes from chuE to iluDí are part of a single transcriptional unit that is necessary for the catabolism of arabinose in S. me-

94 C].IAP ER 4. ARABINOSE CATABOL]SM AND TRANSPORT

Figure 4.1: Predicted open reading frames and isolated mutants in arabinose catabolism transcriptional unit. Boxes indicate predicted open reading frames, with pointed ends indicating the direction of transcription. Triangles indicate positions of TnS-820 insertions for a given allele. Filled triangles indicate that the Tn5-820 insert is a transcriptional fusion. Putative gene functions assigned by homology, see text for details.

I kbp

Itronos¡d ghpR g"ui u"n," dt ltl' uoE aral) aruC ¿ruIl

I'rc\ ¡oils gguc ggul) ggrr;f gclc t¡întc ¡^ 1)i \'l09li I cl¡tIi

I'rc¡liclrd Dclìydrì¡tnsc Hydrûtasc l'crnrcasc l l' binding casscttc l)criplilsrììic lianscri¡tional lirnctioil dchydrogcnasc^ldcl'ydc ^ bindingprotcirt rcgul¡(ù'

95 CHAPT'EII 4. ARABINOSE CATABOLISM AND TRANSPORT

Iiloti. Since this appears to be an operon necessary for arabinose utilization with no other associated phenotypes, we suggest that the current systematic names derived from genome annotation be changed to conform with standard nomenclature (Demer- ec et al. 1966). The suggested replacement names are aTaABCDEF, replacing the names chuE, ggttA, gguB, gguc, Y20B91and iluDí (Systematic identifiers SMb20B95, SMb20B94,

SMbZ0B93, SMb20892, SMb20B9l, and SMb20B9 respectively) that currently exist in the genome database. Throughout this work, these loci will be referred to using our suggested naming scheme. There is another predicted gene less than 200 bp dovrrnstream oï araF (Y2OBB9) in which we did not isolate a mutation. Although Y2OBB9 could be transcribed with the rest of the ara operon, it is also possible that it has an independent promoter. Our data can not resolve whether or not this is part of the ara. operon.

In order to attempt to correlate the components of the araoperon with the biochemically elucidated arabinose catabolic pathway, InterProScan (Quevillon et al. 2005) was used to assign functions to each predicted ORF based on the deduced amino acid se- qucnces. The gene araAis a homolog of. chuE in Agrobacterium tumefacfens, which has been shown to couple sugar binding with the induction of plant virulence genes (Huang et al. 1990, Cangelosi et a-1. 1990) and is thought to be the periplasmic sugar binding protein of an ABC transporter (Kemner et al. 1997). An InterProScan search shows that

AraA is a member of the periplasmic binding protein family (InterPro accession number

IPR00r761).

IloÍh c¿raB and araC appear to encode parts of an ATP binding cassette transport system, specifically, the ÄTP-binding and permease proteins, respectively (iPR003439,

IPR001851). Based on the proximity of these putative transport genes to a locus necessary for arabinose catabolism, we hypothesized that these genes are necessary for arabi-

96 CHAPT'ER 4. ARABINOSE CATABOLISM AND TRANSPORT nose transport. ABC transporters function with two ATP binding cassette proteins and two penneases that contain transmembrane domains (Locher'2004). That this opeïon contains only a single permease gene suggests the permease functions as a homodimer within the ABC transporter.

AraD is part of a family of proteins of unknoi¡¡n function (IPR009645). Conserved domain analysis using Conserved Domain Architecture Retrieval Tool (CDART) (Geer et aI. 2002) show that AraD contains domains that are related to COG01 79 and members of this family have been shonm to have hydratase activity (Arai et al. 1999). AraE is a member of a family of aldehyde dehydrogenases (IPR002086), and araF is a member of a dehydratase family (IPR0005BI). The sequence of Y20BB9 was also queried, since it may be part of the ara, operon; however the results of the searches were uninformative.

Based on the biochemically elucidated arabinose catabolism pathway in S. meliloti and other läst growing Rhizobium species (Dilworth et al. I986, Duncan and Fraenkel

1979, Stowers l9B5), arabinose is likely catabolized to ø-ketoglutarate via five enzymatic steps: L-arabinose 1-dehydrogenase, L-a¡abinonolactonase, L-arabinonate dehydratase, 2-keto-3-deoxy-L-arabinonate dehydratase, and ø-ketoglutarate semialdehyde dehydrogenase (see Figure 4.2). Based on the results of database queries, we hypothesize that araE encodes ø-ketoglutarate semialdehyde dehydrogenase and araF encodes L-arabinonate dehydratase, while araABC are involved in arabinose transport.

It seems plausible that araD also encodes a hydratase. This analysis suggests that the

(rra operoln does not encode all the enzyrnes for arabinose catabolism; in particular, the genes encoding L-arabinonolactonase, and L-arabinose dehydrogenase may be present elsewhere in the genome. Searches of the S. melíloti database did not yield any unam- biguous candidate genes (data not shornm). We did not undertake enzyme assays to con-

97 CI]AI> ER 4. ARABINOSE CATABOLISM AND TRANSPORT

Irigure 4.2: Schematic of S. meliloli arabinose catabolism pathway. Listed enzyrnes correspond to circled numbers on the figure: l) L-arabinose l-dehydrogenase; 2) L- arabinonolactonase; 3) L-arabinonate dehydratase; 4) 2-keto-3-deoxy-L-arabinonate dchydratase; 5) ø-ketoglutarate semialdehyde dehydrogenase.

L-arabinose H3;l O L-a ra bi nono-1,4-l a ctone n,ol Ø L-arabinonate ",ol c 2-dehydro-3-deoxy L-arabinonate ,ro4 @ a-ketoglutarate semialdehyde TCA cycle it3;l c u-ketoglutarate

9B CTlAPTER 4. ARABINOSE CATABOLISM AND TRANSPORT firm the presence of these putative functions in cell-free extracts because the required substrates are not commercially available.

Arabinose dehydrogenase activity is present in an areD mutant.

'I'o test our hypothesis that not all the dehydrogenase enzyrnes involved in arabinose catabolism are encoded by the ara opeÍoî, arabinose dehydrogenase activity was measured in cell-free extracts of SRmA502 and Rm102I. SRmA502 (araD2::Tn5-820) was chosen because it has a mutation predicted to have a polar effect on all donmstream genes, while still allowing the possibility of arabinose transporter (araABC) gene expression. Cultures were grown in defined media with either arabinose/glycerol or glycerol as carbon sources, lysed, and assayed for arabinose dehydrogenase activity. Arabi- nose dehydrogenase activities for Rm 102 I and SRmA502 from arabinose/glycerol grown cultures were 8.710.4 and 33*3 nmol min-l mg-r respectively. Activities for glycerol grown RmI021 were 4.8t1.6 nmol min-r mg-I; measured activities from glycerol gro',nrn

SRmA502 were not significantly different. Consistent with our hypothesis, no loss of activity was detected in SRmA502. Taken together these data provide evidence that arabinose dehydrogenase is encoded by an arabinose inducible locus elsewhere in the gcnome. This is similar to the report by Watanabe et al. (2006) of an arabinose inducible arabinose dehydrogenase in A. brasiliens¿ which does not appear to be within a locus necessary for arabinose catabolism. We note that the level of inducibility of S. meliloti arabinose dehydrogenase, as measured by enzyme activity, appears to be greater in an araDbackground.

99 CHAPTER 4. ARABINOSE CATABOLISM AND TRANSPORT

Induction of arøoperon by arabinose, seed exudate and galactose.

Arabinose has been shor¡m to be a component of pea plant root secretions (Knee et al.

2001) and plant cell walls (McNeil et al. l9B4), leading us to hpothesize that alfalfa seed exudate would induce ara operofi transcription. The araAhomolog in A. tumefacierts, tJtuE, has been shornm to be regulated by a negative regulator, gbpR,and inducible by arabinose, fucose, and galactose (Doty et al. 1993). Moreover the transcriptional regulator, gpbR,that is associated with the ara operon in S. melilorl is also a homolog of theA. tumefaciensgpbR. (F'igure4.1). Thisledustotesttheabilitiesof fucose, galactose, and arabinose to induce araoperon expression. To carry out this experiment, the araFl::Tní-820 fusion strain SRmA240 was grown with different carbon sources, and

the levels of gene expression were measured via 13 -galactosidase activity (Table 4.2).

As with A. t.unzefaciens chuE, our results show that the operon containing araA appears to be strongly induced by arabinose, and moderately induced by seed exudate and galactose (Table 4.2). Unlike in -4. tumeþciens, however, the expression of the operon was unaffected by fucose (Table 4.2). Although galactose induces the operon, none of the arabinose mutants isolated were compromised in their ability to utilize galactose as a sole carbon source, or any other sugar tested. It is interesting to note that the strain

RmG212, which has been used as a Lac- strain (Glazebrook and Walker 1991), had a much higher background level of B-galactosidase activity when it was assayed from exudate grown RmG212. We note that it has been previously shornm that native gels of

S. meliloti induced with lactose and stained for B -galactosidase activity show two bands of activity (Charles et al. 1990) and that galactosides have been shor¡¡n to be present in the seed exudate (Bringhurst et al. 2001). Residual B-galactosidase activity and the presence of galactosides may account for the reproducibly high background activities that

100 CI-IAP'TER 4. ARABINOSE CATABOLISM AND TRANSPORT

Table 4.2: induction of the arabinose operon by several carbon sources.

Strain Carbon source* SRmA240t RmG2l2

Arabinose 2049+181+ 6+2 Exudate 409+45+ 65+17 Galactose 468+58+ 4+2 Fucose 86+10 B+2 Glycerol 87+18 9+1 Glucose 73+B B+.2 Succinate 73+B 10+4

* Strains were grown overnight in either glycerol (15 mM) or glycerol and the carbon source listed at concentrations of 15 mM, except for exudate which was used at I0To (vlv). i The values given represent B-galactosidase activity expressed in Miller units and are the average of three independent replicates t standard deviation. f These values are significantly different (> 95% confidence interval) from the glycerol control value using Student's t- test.

101 CI.IAPT'ER 4, ARABINOSE CATABOLISM AND TRANSPORT were l'ound with RmG212 grornrn on seed exudate.

Expression of the arø operon is moderately repressed by succinate and glucose.

Succinate is a preferred carbon source for S. meliloti and has been sho'o¡n to repress catabolism and/or transport of other carbon sources such as lactose and raffinose (Uck- er and Signer 1978, Gage and Long 1998, Bringhurst and Gage 2002). Howeve¡ glucose has also been shown to repress the expression of catabolic genes in R. leguminosarum

(Oresnik et al. 1998, Richardson et al. 2004). To determine if either succinate or glucose catabolite repression play a role in arabinose catabolism, SRmA240 was grown with arabinose or galactose in combination with either succinate, glycerol or glucose and gene expression was measured by B-galactosidase activity (Table 4.3).

The results show that although transcription appears to be decreased in both suc- cinatc and glucose grown cultures when combined with the inducing sugars arabinose or galactose, transcriptional activity is not abolished. The observed trends for galactose or arabinose induced cells with the addition of glucose or succinate appear similar. The addition of glycerol however did not affect expression suggesting that the effect of the acldition of glucose and succinate was not related to the overall concentration of carbon in the media (Table 4.3). Taken together the data show that although glucose and succinate moderate the transcriptional activity of this operon, it is not subject to complete catabolite repression by either succinate or glucose.

r02 CI-lAPTER 4. ARABINOSE CATABOLISM AND TRANSPORT

Table 4.3: The arabinose operon is not subject to catabolite repression by succinate or glucose.

Strain

Carbon source[s)* SRmA240t RmG2I2 Galactose 593+rBB 1l+3 Glucose 85+17 10+2 Glycerol t02+22 10+4 Succinate B0+2 10+1

Arabinose Glucose 1069+26i+ 9*5 Arabinose Glycerol 2254+299 7+2 Arabinose Succinate 897+92+ 10+1

Galactose Glucose 416+108 9+2 Galactose Glycerol 495+39 B+t Galactose Succinate 274+4s Il+1 * The values given represent ß-galactosidase activity expressed in Miller units following overnight growth in defined media with l5 mM of the carbon source(s) listed. i Values are the average of three independent replicates and are presented as mean i standard deviation. * These values are significantly different from those of the cells grown with arabinose and glycerol at the 95% confidence level using Student's t-test. s This value is significantly different from that of the cells growrì with galactose and glycerol at the 95% confidence level using Student's t-test.

103 CIIAPTER 4. ARABINOSE CATABOLISM AND TRANSPORT

The ørøoperon is required for arabinose transport.

Three putative genes encoding transpofi components were identified in the arabinose operon; a.rctA, araB, and araC. The requirement of this operon for arabinose catabolism suggests that the products of these genes may be involved in arabinose transport. Al- though the first step in arabinose catabolism is thought to lead to the loss of the hydro- gen atom from the C- I carbon by a dehydrogenase (Duncan and Fraenkel 1979, Duncan

1979), we initially attempted to measure arabinose uptake using the only commercially available labeled arabinose, [1-3H]arabinose. Strains Rm1021, SRmA503 and SRmA240 were tested for the abilityto accumulate Ii-3H]arabinose (Table 4.4). Three independent trials were performed and SRmA503 (araAL::Tn5-820) with a disruption at the beginning c¡f the operon consistently showed lowered levels of label accumulation while SRmA240

(arctFl::Tn5-820) which carries an insert at the end of the operon was similar to wild type

(lhbte 4.4). Although we observed measurable arabinose accumulation, the values do not reflect the rate of arabinose transport. We observed that label accumulation reached a maximal level within ten seconds of label addition regardless of the amount of label added to the cells (data not shornm). Therefore we were unable to find a concentration of substrate that showed a linear increase in label accumulation with time. Following our initial work, Il-raC]arabinose became available commercially. Using this label, we rvere able to measure linear rates of uptake for strain Rm102I, however the rates were very low and were only measurable over tens of minutes (remaining linear for at least 20 minutes, data not sho'¡¡n). This suggests that using Il-I4C]arabinose, we measured the incorporation of labeled arabinose into cell materials, rather than rates of transport. We hypothesize that the differences observed between accumulation of [1-:ìH]- and It-'4C]- labeled arabinose are due to the predicted biochemical modifications of arabinose upon

104 CI.IAPTER 4. ARABINOSE CATABOLISM AND TRANSPORT

Table 4.4: Accumulation of [3H]arabinose by S. melilotí Rm102l and arabinose catabolism mutants.

Label accumulation*'l

Growth media RmI02l SRmA503 SRmA240

Glycerol 2.6+1.3 n.d.+ n.d. Arabinose r0.0+2.4 n.d. n.d. Glucose 0.6+0.3 n.d. n.d. Succinate 3.9+1.6 n.d. n.d. Arabinose * Glycerol 7.6+0.5 2.0+0.3 15.1+0.7 Arabinose * Glucose 0.6+0.3 1.3+0.2 0.5+0.1 Arabinose -l- Succinate t2.r+2.4 1.4+0.6 13.0+5.i

* Accumulation reported as uptake of pmol arabinose accumulated per rng total protein over l0 seconds. Formaldehyde killed controls accumulated 0.6+0.1 pmol mg-r. I Values are the average of three independent replicates and are presented as mean t standard deviation. * Not determined.

r05 CITTIPTER 4. ARABINOSE CATABOLISM AND TRANSPORT entry into the cell: dehydrogenation of arabinose by arabinose dehydrogenase and de- carboxylation of ø-ketoglutarate by ø-ketoglutarate dehydrogenase. Despite this, our results clearly indicate that the ara operon encodes arabinose transport components.

\¡Vhether or not the very low levels of arabinose uptake measured for SRmA503 reflect true transport is not knor,rm. It is possible that arabinose is taken up at a low level by some other mechanism, while the ara operon is the cells primary transport mechanism for arabinose.

The fact that glucose grown cells showed the lowest levels of arabinose accumulation was unexpected, and did not correlate with the observation that glucose did not conrpletely repress transcription of the ara operon (Table 4.3) . It was hypothesized that either glucose grown cells expressed a protein capable of inhibiting arabinose accumu-

Iation, or that glucose present in the culture medium may directly inhibit uptake of arabinose. The addition of glucose at the time of assay to arabinose grown cells was found to instantaneously prevent arabinose accumulation, suggesting that it di¡ectly interferes with arabinose accumulation (data not shown). In addition to glucose, this effect was found to be caused by galactose. We noted that this effect was only observed at very high excess concentrations of glucose or galactose compared to labeled arabinose.

Glucose uptake is not compromised in arabinose catabolism mutants.

Iìollowing our initial observation that high levels of glucose abolished arabinose transport, it was hypothesized that perhaps glucose could also be transported by the arabinose transporter, thus competingwith [3H]arabinose. To test this, rates of IraC]glucose uptake were measured for strains Rm1021, SRmA240 and SRmA503. It was reasoned tl-rat if glucose were to be taken up by ATaABC then there should be a clear difference in the measured rate of glucose uptake in the absence of this ABC transporter. To test this,

106 CLIAPI-ER 4. ARABINOSE CATABOLISM AND TRANSPORT

strains were grown up using glucose alone or arabinose plus glycerol as carbon sources.

A linear rate of glucose uptake for Rmt02 t was measured at 8.3+0.9 nmol min-r mg-t total protein. In all cases there was no significant difference in the observed rate of growth or the measured glucose uptake rates for Rm1021, SRmA240 and SRmA503 (P > 0.7).

This suggests that although excess glucose inhibits arabinose transport it is not through competitive uptake by the ABC components of the aratranspolt system.

Plant symbiosis and competition for nodule occupÍrncy.

Delayed an

'lnS arabinose catabolism mutant (Duncan I98i). In addition, we note that tlrre araA homolog in A. tunzefaciens (chuÐ has been sho'nm to be necessary for the induction of plant virulence factors by various sugars (Huang et al. 1990, Cangelosi et al. 1990). f'lrese reports led us to askwhether mutations in the araoperoîwould affect the s)¡mbiotic abilities of S. meliloli. To test this, strains SRmA240 (arat:^::Tn5-820) and SRmA503

(araA:'.1n5-820) were assayed for the ability to form nitrogen-fixing nodules, and in addition, SRmA503 was tested for the ability to compete for nodule occupancy with the wild type.

Plants grown in nitrogen free media and inoculated with SRmA240 or SRmA503 were green and healthy and there were no significant differences in either the plant dry mat- ter accumulated or the number or location of the nodules that were formed on the root system of the host plant. Moreover, bacteria isolated from surface sterilized nodules retained their genetic markers. Examination of the nodule occupancy of plants simul- taneously inoculated with both SRmA503 and wild-type Rm1021 showed a decrease in the ratio SRmA503 isolated relative to the inoculation ratio, when SRmA503 was inoculated at a concentration equal to the wild type. \¡Vhereas SRmA503 made up 52L9To of

107 CI.IAP'il]R 4. ARABINOSE CATABOLISM AND TRANSPORT colonies present in the inoculant for three independent competition experiments, the percentage of SRmA503 recove¡ed from nodule extracts was only 26*9Yo. Although this decrease in nodule occupancy is statisticallysignificant (P:0.0051), further experiments in which SRmAS03 was present at a higher concentration than wild type in the inoculum did not reveal a significant decrease in the proportion of mutant recovered (Figure 4.3).

Taken together, there is no evidence for a competitive difference between SRmA503 or thc wild type, Rml02I.

Discussion

In this work we have suggested new function based nomenclature for an operon of genes that are necessary for growth with arabinose as a sole carbon source. The current name for rtraA in the database is chuE, named for chromosomal virulence (Huang et al. 1990) in A. tumefaciens. We have shornm that mutations in araAhave no apparent defects in syrnbiotic ploficiency or the ability to compete for nodule occupancy with the wild type, indicating that there is no reason to keep the name chuE.Moreover the gga mnemonic, glucose and galactose uptake, is misleading and does not reflect the physiological func- tionwhichwehavedemonstrated(Table4.4). Sincetheonlyphenotypeformutationsin any of the other ara genes is loss of the ability to transport and/or catabolize arabinose, wc suggest that these genes should be annotated as ara genes in S. melilorl. Renaming this region as the araopeÍon is consistent with bacterial nomenclature and reflects the phcnotype of mutations in these genes.

Although arabinose catabolism mutations have been described in the past (Dun- can and Fraenkel 1979, Duncan I98l), to ourknowledge no mutation specific to arabi- n

IOB C[]AP'IEI] 4. ARABINOSE CATABOLISM AND TRANSPORT

Iìigure 4.3: Recovery ratios of the chuE mutant SRmA503 to wild-type Rm102t from alfalfa nodules is not significantly different from inoculum ratios. Bars indicate the average amount of SRmA503 present in the inoculum or recovered from nodules over three indeper-rdent experiments, error bars indicate standard deviation.

C '+to lI' f o- o o_ rt- o s

Trial l- Trial2

109 CI-IAPI-ER 4. ARABINOSE CATABOLISM AND TRANSPORT utilization operon we have identified is present on the megaplasmid pSymB. Although an extensive genetic map and an elegant ín uiuo deletion strategy have been used to characterize this replicon (Charles and F'inan 1990, 1991), some regions were found to bc recalcitrant to deletion (Charles and Finan 1991). Since the utilization of arabinose w¿rs screened in those analysis, we conclude that the aTaABCDEF region we describe is r,r.ithin one of the regions that has not been previously deleted. This is likely since the arct opeton is located only about 12 kbp away from the genome's sole arginine specific

IRNA encoding locus (Finan et al. 2001).

Based on the previous enzymatic characte¡ization of arabinose catabolism in S. me- liLoti, there are five hypothesized steps in the conversion of arabinose to ø-ketoglutarate

(Duncan and Fraenkel 1979). The araoperon includes genes necessaryfor uptake of arabinose, leaving only araDEF left to encode the necessary enzymes for catabolism. Based on our bioinformatic analysis araDEF are likely to encode a semi-aldehyde dehydrogenase and two dehydratases. The fact that arabinose dehydrogenase activity was not affected by mutations in the ara operon, and appears to be inducible in an ara insertion mutant likely to have polar effects on downstream genes, provides evidence that the genes necessary for encoding arabinose dehydrogenase activity, and presumably lactonase activiry are not present at this locus and exist elsewhere in the genome. The location of these genes is unknorn¡n at this time.

Previous work on arabinose utilization was carried out in S. meliloti strain L5-30

(Duncan 19Bl). A mutant unable to utilize arabinose showed a delayed and variable ability to fix nitrogen symbiotically, and its lesion was not in genes that affected either arabinose dehydrogenase or the alpha-ketoglutarate semi-aldehyde dehydrogenase activities. lt was concluded that the lesion must be in a gene that affected either of the

It0 CI{API-ER 4. ARAB]NOSE CATABOLISM AND TRANSPORT dehydratases or the arabinolactonase activity (Duncan IgBt). Taken together, the genomic sequence of Rml02l, the phenotypes of the mutants we have generated, and the demonstration that arabinose dehydrogenase activity was not found at the aTaABCDEF locus suggest that the mutation previously characterized by Duncan (l98i) may be affecting arabinolactonase activity. We note that whereas the phenotypes of the mutants we report here are 100% linked in transduction with the associated transposon markers, the previous researcher did not have access to a generalized transducing phage, and was therefore unable to show that the inability of their mutant to use arabinose was solely due to a transposon insertion (Duncan 19Bl). The isolation and identification of the genes necessary for both arabinose dehydrogenase and arabinolactonase activity would be helpful in this regard.

Expression of metabolic genes is affected by the carbon source upon which they are grown. A very clear example of this is pckA in S. melíloti. The expression of pckA is in- cìuccd when grown using a gluconeogenic carbon source such as succinate or arabinose, whereas no expression is observed if a glycolytic carbon source such as glucose is uti-

Iized (Østerås et al. 1995). it was observed that although pckAexpression was reduced by approximately 50% if the cells were grown on both glucose and succinate, growth on arabinose and glucose resulted in a greater than 90% reduction (Østerås et al. lg95).

Since al'abinose is catabolized to a tricarboxylic acid (TCA) cycle intermediate, it was suggested that glucose may either repress arabinose catabolism or uptake. Our find- ings suggest a combination of both of these effects, as we have observed that glucose can inhibit arabinose uptake when present in excess, and glucose moderately represses transcription of the arabinose operon (Table 4.3,Table 4.4).

A number of mutations that affect competition for nodule occupancy have been re-

llt CTIAPTER 4. ARABINOSE CATABOLISM AND TRANSPORT ported in the literature (Aneja et al. 2005, Fryet al. 2001, Jiménez-Zurdo et al. 1995, Ores- nik et al. 1998, Phillips et al. 1998, Rosenblueth et al. 1998, Soedarjo and Borthakur 1998,

Streit et al. 1996). The magnitude of these phenotypes appears to vary from severe to statistically significant but subtle. Initially, it appeared that the arabinose catabolism mutant SRmA503 was less competitive than wild type when co-inoculated in approximately equal amounts, and the results were statistically significant. Further experiments with SRmA503 present at a higher ratio in the inoculum showed no significant decrease in the number of cells recovered. Performing assays of nodule occupancy competition with several ratios of mutant to wild type is therefore recommended. The ability to thrive and compete in a complex environment such as the rhizosphere, or within the infection thread, is likely affected by many determinants. The challenge ahead is not only to identify which genes affect this complex phenotype, but also to ascribe a proper function to thcm so that the physiology and regulation of the process will be better understood.

Moreover it is imperative that we begin to understand at what point during the interaction with the host plant the effects of different non-competitive mutants are manifested.

TT2 Chapter 5 Conclusion

The goal of this thesis was to characterize the genetics of rhamnose and arabinose catabolism, and the role of triose phosphate isomerase in S. melilori. Now that it has been confirmed that the ability to catabolize rhamnose is an important factor in the ability to compete for nodule occupancy in S. meliloti, the nature of this effect remains to be dcternrined. Infection and nodulation of alfalfa by S. meliloti have been investigated using green fluorescent protein (GFP) expressing strains (Gage et al. Ig96, Gage 2002). A similar approach could be used to provide an explanation for the competitive defect of

S. meliloti rhamnose catabolism mutants. It has been hlpothesized that rhamnose may be supplied in the infection thread, and the abilityto use it would provide a competitive advantage (Oresnik et al. 1998). Plants co-inoculated with rhamnose mutants expressing GlìP and wild type expressing red fluorescent protein should show whether or not there is a difference in infection thread growth rates between the two strains. Another approach would be to construct aplasmid containinggþunder the control of the rhamnose promoter. If rhamnose is present in the infection thread, GFP expression would be expected fì-om the infecting bacteria.

Although it was predicted that the genes for arabinose catabolism would be located on the chromosome, theywere found to be present on a previously undeleted region of pSymB. Based on previous reports, it was also hypothesized that arabinose catabolism

r13 CL]APT'ER 5. CONCLUSION

Inutants would be less competitive for nodule occupancy than wild type. This predic- tion was not supported by the data. Due to the fact that the identified arz operon does not contain all the genes necessary for arabinose catabolism, it is possible that mutations in other genes affecting arabinose utilization would cause competitive defects. A recent study by Watanabe et al. (2006) found an arabinose dehydrogenase in Azospit'il- lunt brasiliense. The arabinose catabolism pathways of S. metiloti and A. brasiliense are homologous, but a homologous arabinose dehydrogenase was not found in the S. me-

Liloti ara operon. Searching through the S. meliloti genome using Bt {STP and the

A. brastliens¿ arabinose dehydrogenase protein sequence shows the gene SMc005BB, which is predicted to encode a protein that is 54% identical. It seems likely that this gerre would also be necessary for arabinose catabolism in S. meliloti, and it would be interesting to see if mutations in SMc005BB affected nodulation or competition. 'l'he characterization of TPI in S. melilorl has provided us with an interesting oppor- tunity to study carbon metabolism. A strain carrying mutations in tpiA, tpiV and pyc should not be able to grow with a gluconeogenic carbon source or a glycolytic carbon soutce. Flowever, this strain was constructed, and was found to be able to grow with both ribose and xylose, but not glucose or gluconeogenic carbon sources. This preliminary observation suggests that ribose and xylose may be catabolized in a novel manner in S. melilofi. lt seems likely that ribose and xylose are catabolized through the pentose phosphate pathway, howeve¡ in order to support the growth of a tpiA tpiB pyc mutant, they must also be converted to a TCA intermediate or acetate. Given the hypothesis of two catabolic end-products, it is not surprising that efforts to isolate non-pleiotropic ribose and xylose mutants have been unsuccessful. A strategythat maywork is to muta- genize t¡tiA tpiB or pyc strains, while screening for a loss of the ability to grow with ribose

rt4 CI-IAP'TER 5. CONCLUS]ON or xylose. I'his should isolate the two proposed pathways allowing mutants in either to be isolated. 'fhe fact that S. meliloti has two functional TPI encoding genes was somewhat sur-

¡rrising, given that this situation has only been characterized and reported for one other prokzrryotic organism (Zhenget al. 2006). Perhaps even more unexpected was that tpiB is specifically required for er¡hritol catabolism in S. meliloti. It is hypothesized that Tpil3 may form a metabolic complex with other gene product(s) involved in er¡hritol catabolism. If this is the case, it should be possible to bind and isolate the partner(s) using recombinant TpiB tagged to bind to a chromatography column matrix. Alternatively, a yeast two-hybrid system could be used to screen an S. meliloti genomic library for binding partners. It would also be worthwhile to construct tpiB mutations in other or- gatrisms with homologs of tpiBandthe erythritol operon. Given thatTpiB is required for erythritol catabolism in S. melilori , it would be expected that tpiB mutations in other organisms would cause a loss of the ability to grow with er¡hritol as a sole carbon source.

It was very surprising that some results suggested that plants inoculated vt-tth tpiA mutants may grow to a higher dry weight than those inoculated with wild t1rye. Several approaches may help to understand whether this is an artifact or a subtle phenotl.pe.

Growing many plants in a large scale experiment such as a greenhouse trial could provide the numbers necessary to resolve a small difference in means. Also, if tpiA inoculated plants are truly growing to a higher dry weight in nitrogen limiting conditions, it would be expected that they are either f,*ittg more N2 per nodule, or that they have more nodules per plant. To partly address this question, plants could be grornrn in a sealed chamber with gas fittings which would allow direct measurements of N2 fixation by undisturbed nodule systems. Plants were grown in glass test tubes to facilitate nodule

115 CI IAPI-ER 5. CONCLUSION observation and the quantification of the average number of nodules per plant and their rate of formation for tpiA and wild rype inoculated plants. The problem with this approach is that the plants are grown in a very non-physiologically relevant manner and it is not possible to correlate the dry weight of a plant to its number of nodules. To address this, the number of nodules would have to be counted on each plant assayed, before it is dried and weighed. If there was both an increase in dry weight and the number of nodules per plant, and both were positively correlated, the argument that plants inoculated with tpiA are being supplied with more fixed N2 would be significantly strengthened.

The work presented in this thesis has contributed to and expanded our knowledge of arabinose and rhamnose transport and catabolism, as well as the roles of the two triose plrosphate isomerases in S. meliloti. The study of carbon metabolism continues to be imprrrtant to further our understanding of the basic physiology of S. meliLori. In addition to discovering the metabolic pathways used by S. meliloli, an ongoing objective is to identify the genes involved in carbon metabolism and to study their regulation and physiological roles.

116 R.eferences

Aìexeyev M., I999. The pKNOCK series of broad-host-range mobilizable suicide vectors for gene knockout and targeted DNA insertion into the chromosome of gram-negative b acteria. Biotechniques, 26: 824-826, B2B.

Altschul S.F., Madden T.L., SchäfferA.A., Zhang I.,ZhangZ., MillerW., and Lipman D.J., 1997. Gapped BI-AST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res., 25: 3389-3402.

Anderson A. and Cooper R., 1969. Gluconeogenesis in Escherichia coli - The role of triose phosphate isomerase. FEBS Lett., 4: 19-20.

Aneja iì, Zachertowska A'., and Charles T.C., 2005. Comparison of the symbiotic and ccrrnpetition phenotypes of Sinorhizobium melílotí PHB synthesis and degradation pathway mutants. Can. I. Microbiol., 5l: 599-604.

Arai H., Yamamoto T., Ohishi T., Shimizu T., Nakata T., and Kudo 1'., 1999. Genetic organization and characteristics of the 3-(3-hydroxyphenyl)propionic acid degradation pathway of Comamonas testosteroni TA44I. Microbiology, I 45: 2813-2820.

Araujo R.S., Robleto E.4., and Handelsman J., 1994. A hydrophobic mutant oT Rltizo- biunt etli altered in nodulation competitiveness and growth in the rhizosphere. Appl. Environ. Microbiol., 60: 1430-i436.

Arias 4., Cerveñansklf C., GardiolA'., and Mdrtínez-Drets G., I979. Phosphoglucose isomerase mutant of Rhizobíum meliloti.I. Bacteriol.,lST: 409-14.

Arias A. and Cerveñansky C., 1986. Galactose metabolism in Rhizobium meliloti L5-30. I. Bacteriol., L67 : 1092-i094.

Arias r\., Gardiol 4., and Martínez-Drets G., 1982. Transport and catabolism of D- rnannose in Rlzizobium meliloti. J. Bacteriol., l5l: 1069-1072.

Arias A. and Martinez-Drets G., 1976. Glycerol metabolism in Rhizobium. Can. J. Micro- biol.,22: r50-153.

It7 RETìERENCES

Baldani j.I., Weaver R.W., Hlmes M.F., and Eardly 8.D., 1992. Utilization of carbon substrates, electrophoretic enzyme patterns, and symbiotic performance of plasmid- cured clover rhizobia. Appl. Environ. Microbiol., 58: 2308-2314.

Banfalvi 2., Kondorosi E., and Kondorosi 4., 1985. Rhizobium meliloti carries two megaplasmids. Plasmid, 13: 129-138.

ßarnett M.J., Fisher R.F., Jones T., Komp C., Abola A.P, Barloy-Hubler E, Bowser L., Capela D., Galibelt E, Gouzy I., Gurjal M., Hong 4., Huizar L., Hyrnan R.W, Kahn D., Kahn M.L., Kalman S., Keating D.H., Palm C., Peck M.C., Surzycki R., Wells D.H., Yeh K.C., Davis R.W., Federspiel N.4., and Long S.R., 2001. Nucleotide sequence and predicted functions of the entire Sinorhizobíum melílotipSymA megaplasmid. Proc. Natl. Acad. Sci. U.S.A.,98: 9BB3-9BBB.

Barnett M.J., Toman C.J., Fisher R.F., and Long S.R., 2004. A dual-genome Symbiosis Chip for coordinate study of signal exchange and development in a prokaryote-host interaction. Proc. Natl. Acad. Sci. U.S.A., I0l: 16636-41. lJarra L., Pica N., Gouffi K., Walker G.C., Blanco C., and Trautwetter 4., 2003. Glucose 6-phosphate dehydrogenase is required for sucrose and trehalose to be efficient os- moprotectants in Sinorhizobium meliloti. FEMS Microbiol. Lett.,229: 183-lBB. lJauer W.D., 1981. lnfection of legumes by Rhizobia. Ann. Rev. Plant Physiol., 32: 407- 449.

Bcringer I.E., Belmon J.L., Buchanan-WollastonA.V, and IohnstonA.W.B., 1978. Transfer of the drug-resistance transposon Tn5 to Rhizobium. Nature, 276: 633-634.

Bernardelli C.E., Luna M.F., Galar M.L., and Boiardi 1.L., 200L. Periplasmic PQQ- dependent glucose oxidation in free-living and symbiotic rhizobia. Curr. Microbiol., 42:310-3i5.

Bittinger M.A. and Handelsman J., 2000. Identification of genes in the RosR regulon of Rhizobium etli. I. Bacteriol., lB2: 1706-1713.

Bittinger M.4., Milner J.L., Saville B.I., and Handelsman I., 1997. ¡?sR, a determinant of nodulation competitiveness in Rhizobium etli. Mol. Plant-Microbe Interact., l0: 180- r 86.

Bohlool Il.B., Ladha I.K., Garrity D.P, and George T., 1992. Biological nitrogeu fixation f'or sustainable agriculture: A perspective. Plant soil, l4l: l-l l.

rlB III]IìERENCES

Bolton E., Lliggisson 8., HarringtonA., and O'Gara F., 1986. Dicarboxylic acid transport in Rhizobium meliloti: isolation of mutants and cloning of dicarboxylic acid transport genes. Arch. Microbiol., 144: 142-146.

Ilowen G.D. and RoviraA.D. ,1976. Microbial colonization of plant roots. Annu. Rev. Phy- topathol., 14: l2I-I44.

Bradford M.M., 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem., 72:248-254.

IJrewin N.J., 199i. Development of the legume root nodule. Annu. Rev. Cell Biol., 7: 19i- 226.

Bringhurst R.M., CardonZ.G., and Gage D.J., 2001. Galactosides in the rhizosphere: uti- Iization by Sinorhizobium meliloti and development of a biosensor. Proc. Natl. Acad. Sci. [J.S.A., 98: 4540-4545.

Brir-rghurst R.M. and Gage D.J., 2002. Control of inducer accumulation plays a key role in succinate-mediated catabolite repression in Sitzorhizobium meLiloti. J. Bacteriol., l84: 5385-5392.

IJrom S., de los Santos 4.G., StepkowskyT., Flores M., Dávila G., Romero D., and Palacios Iì., 1992. Different plasmids of Rhizobium leguminosarumbv. phaseoli are required for optirnal symbiotic performance. I. Bacteriol., 174: 5183-5f 89.

Caetano-Anollés G., I993. Amplifyrng DNA with arbitrary oligonucleotide primers. PCR Methods Appl., 3: 85-94.

Caetano-Anollés G., Wall L.G., Micheli A.T.D., Macchi E.M., Bauer WD., and Favelukes G., tgtì8. Role of motility and chemotaxis in efficiency of nodulation by Rhizobium meliloti. Plant Physiol., 86: I22B-I235.

Car-rgelosi G.4., Ankenbauer R.G., and Nester 8.W., 1990. Sugars induce the Agrobac- teriumvirulence genes through a periplasmic binding protein and a transmembrane signal protein. Proc. Natl. Acad. Sci. U.S.A., 87:6708-12.

Capela D., Barloy-Flubler Fl, Gouzy I., Bothe G., Ampe F., Batut I., Boistard P, Becker 4., I3outry M., Cadieu E., Dréano S., Gloux S., Godrie T., Goffeau 4., Kahn D., Kiss Il., Lelaure V, Masuy D., Pohl T., Portetelle D., Pühler 4., Purnelle 8., Ramsperger U., Renard C., Thébault P, Vandenbol M., Weidner S., and Galibert E, 2001. Analysis of the chromosome sequence of the legume symbiont Sinorhizobíum meliloll strain 102I. Proc. Natl. Acad. Sci. U.S.A., 98:9877-9882.

119 IIEIIERENCES

Ccrveñansky and Arias 4., 1984. Glucose-6-phosphate dehydrogenase deficiency in pleiotropic carbohydrate-negative mutant strains of Rltizobium nteliloti.I. Bacteriol., 160:1027-1030.

Chain PS.G., Comerci D.I., Tolmasky M.E., Larimer F.W., Malfatti S.4., Vergez L.M., Aguero F., Land M.L., Ugalde R.4., and Garcia E., 2005. Whole-genome analyses of speciation events in pathogenic Brucellae. Infect. lmrnun., 73: 8353-8361.

Charles 1..C., Cai G.Q., and Aneja P, 1997. Megaplasmid and chromosomal loci for the PIIB degradation pathway in Rhizobium (Sinorhizobium) meliloti. Genetics, 146: I2tt-t220.

Charles T.C. and Finan T.M., 1990. Genetic map of Rhizobiutn meliloti megaplasmid pIìmeSU47b. I. Bacteri ol., 172'. 2469-7 6.

Charles l'.C. and Finan T.M., 1991. Analysis of a l600-kilobase Rhizobium meliloti megaplasmid using defined deletions generated in uiuo. Genetics, 127: 5-20.

Charles'IC., Singh R.S., and l;inan T.M., I990. Lactose utilization and enzymes encoded

by megaplasmids in Rhizobium meliloti SU47: implications for population studies. J. Gen. Microbiol., 136: 2497-2502.

Cfark S.R., Oresnik LI., and Hynes M.E, 2001. RpoN of Rhizobium leguminosarumbv. uiciae strain VF39SM plays a central role in FnrN-dependent microaerobic regulatior-r of genes involved in nitrogen fixation. Mol. Gen. Genet., 264:623-33.

Conway T., 1992. The Entner-Doudoroff pathway: history, physiology and molecular biology. FEMS Microbiol. Rev., 9: l-27.

De Ley J. and Doudoroff M., 1957. The metabolism of D-galactose in Pseudomonas saccharophila. I. Biol. Chem., 227:745-757.

DelVecchio VG., Kapatral V, Redkar R.1., Patra G., Mujer C., Los T., Ivanova N., Anderson I., Bhattacharyya 4., Lykidis ,{., Reznik G., Iablonski L., Larsen N., D'Souza M., Bernal 4., Mazur M., Goltsman 8., Selkov 8., Elzer PH., Hagius S., O'Callaghan D., Letesson I.J., I{aselkorn R., Kyrpides N., and Overbeek R., 2002. The genome sequence of the facultative intracellular pathogen Brucella melitensis. Proc. Natl. Acad. Sci. U.S.A., 99: 443-448.

I)emerec M., Adelberg 8.4., ClarkA.J., and Hartman PE., 1966. A proposal for a uniform nomenclature in bacterial genetics. Genetics, 54: 6I-76.

t20 IIIT;ETIENCES

I)emont N., Ardourel M., Maillet F., Promé D., Ferro M., Promé J.C., and Dénarié I., Igg4. 'lhe Rhizc¡bium melilorl regulatory nodD3 and syrM genes control the synthesis of a particular class of nodulation factors N-acylated by (r,.r-1)-hydroxylated fatty acids. LÌMIIO I., l3: 2139-2149.

Dilworth M.I-I., Arwas R., McKay I., saroso s., and Glenn 4.R., 1986. Pentose metabolism in Rhizobit¿m leguminosa.rum MNF300 and in cowpea Rhizobium NGR234. I. Gen. Microbiol., 132: 27 33-27 42.

Djordjevic M.4., 2004. Sinorhizobium meliloti metabolism in the root nodule: a proteomic perspective. Proteomics, 4: lB59-1872.

Doty S.1,., Chang M., and Nester E.W., 1993. The chromosomal virulence gene, chuE, of Agrobacterium tuntefaciens is regulated by a LysR family member. I. Bacteriol ., 175: 7BB0-6.

Dowling D.N. and Broughton W.J., 1986. Competition for nodulation of legumes. Annu. Rev. Microbiol., 40: 13I-157.

Driscoll B.T. and Finan T.M., 1993. NADts-dependent malic enzyme of Rhizobium meliloti is required for s]¡mbiotic nitrogen fixation. Mol. Microbiol., 7: 865-873.

Driscoll Il.T. and Finan T.M., 1996. NADP+-dependent malic enzyrne of Rhizobium nteliloti. J. Bacteriol ., l7B: 2224-2231.

Driscoll B.T. and Finan T.M., 1997. Properties of NAD+- and NADP+-dependent malic enzyrnes of Rhizobium (Sinorhízobium) meliloti and differential expression of their genes in nitrogen-fixing bacteroids. Microbiology, 143: 489-498.

Duncan M., 1979. L-Arabinose metabolism in Rhizobia.I. Gen. Microbiol., ll3: 177-179.

I)uncan M.J., 1981. Properties of Tn5-induced Carbohydrate Mutants in Rhizobium meliloti. J. Gen. Microbiol., l2l: 6l-67.

I)uncan M.J. and Fraenkel D.G., 1979. ø-Ketoglutarate dehydrogenase mutant of Rhizo- bium meliloti. I. Bacteriol., 137: 415-9.

Dunn M.Þ1, 1998. Tricarboxylic acid cycle and anaplerotic enzyrnes in rhizobia. FEMS Microbiol. Rev., 22: 105-123.

Dunn M.E, AraízaG., and Finan T.M., 2001. Cloningand characterization of the pyruvate carboxylase îrom Sinorhizobium melilotíRmL021. Arch. Microbiol., 176: 355-363.

T2L REITI]RENCES l)ymov S.I., Meek D.I.I., Steven 8., and Driscoll 8.T., 2004. Insertion of transposon 'l'nStacl in the Sinorhizobium melilotl malate dehydrogenase (mdh) gene results in conditior-ral polar effects on do'utmstream TCA cycle genes. Mol. Plant-Microbe Inter- acL., 17: I3lB-1327.

I;alke J.1., Bass R.8., Butler S.L., Chervitz S.4., and Danielson M.4., I9g7. The two- component signaling pathway of bacterial chemotaxis: A molecular view of signal transduction by receptors, kinases, and adaptation enzymes. Annu. Rev. Cell Dev. Iliol., 13: 457-512.

Finan T., Kunkel 8., De Vos G., and Signer E., 1986. Second symbiotic megaplasmid in Rltizobium meliloti carrying exopolysaccharide and thiamine synthesis genes. J. Bac- teriol., 167:66-72. fìinan T., McWhinnie E., Driscoll 8., and Watson R., I991. Complex symbiotic phenotypes result from gluconeogenic mutations in Rhizobium meliloti. Mol. Plant- Microbe lnteract., 4: 386-392. lìinan T., Oresnik L, and Bottacin A,., l9BB. Mutants of Rhízobium meLiloti defective in succinate metabolism. I. Bacteriol., 170: 3396-3403.

F-inan 1'.M., Flartweig E., LeMieux K., Bergman K., Walker G.C., and Signer E.R., tg84. General transduction in'Rhizobium melilofi. I. Bacteriol., I 59: 120-124.

Finan'l'.M., Flirsch 4.M., Leigh I.4., Iohansen E., Kuldau G.4., Deegan S., Walker G.C., and Signer 8.R., 1985. Symbiotic mutants of Rhizobium meliloti that uncouple plant fronr bacterial differentiation. Cell, 40: 869-877.

Finan T.M., Weidner S., Wong K., Buhrmester J., Chain P, Vorhölter EJ., Hernandez- Lucas I., Becker 4., Cowie A'., Gouzy I., Golding 8., and Pühler A.., 2001. The complete sequence of the 1,683-kb pSymB megaplasmid from the N2-fixing endosymbiont Sinorhizobium meliloti. Proc. Natl. Acad. Sci. U.S.A., 98: 9BB9-94.

Finan'IM., Wood J.M., and Iordan D.C., 1983. Symbiotic properties of Ca-dicarboxylic acid transport mutantsoï Rhizobium leguminosarum.I. Bacteriol.,154: 1403-I4I3.

Fraenkel D.G., 1987. Glycolysis, Pentose Phosphate pathway, and Entner-Doudoroff pathway. In F. Neidhardt, J. Ingraham, K. Low B. Magasanik, M. Schaechter, and Il. Umbarger, editors, Escherichia colí and Salmonella typhimurium. Cellular and Molecular Biology., American Society for Microbiology, 142-150.

Fraysse N., Couderc E, and Poinsot V, 2003. Surface polysaccharide involvement in es- tablishing the rhizobium-legume symbiosis. Eur. J. Biochem., 270: 1365-1380.

r22 I].EFERENCES

Fry I., Wood M., and Poole PS., 2001. Investigation of myo-inositol catabolism in Rhi- zobiunt leguntinosarumbv. uiciae and its effect on nodulation competitiveness. Mol. Plant-Microbe lnteract., l4: l0l6-1025.

Iruhrer'1, F-ischer 8., and Sauer U., 2005. Experimental identification and quantification of glucose metabolism in seven bacterial species. I. Bacteriol., l87: l5Bt-90.

Gage l).J., 2002. Analysis of infection thread development using Gfp- and DsRed- expressing Sinorhizobium meliloti. I. Bacteriol., l84: 7042-7046.

Gage D.I., 2004. infection and invasion of roots by symbiotic, nitrogen-fixing rhizobia during nodulation of temperate legumes. Microbiol. Mol. Biol. Rev., 6B: 280-300.

Gage D.J., Bobo T., and Long S.R., I996. Use of green fluorescent protein to visualize the

early events of symbiosis between Rhizobium meliloti and alfalfa (Medicago satiua). J. IJacteriol., 178: 7159-7166.

Gage D.J. and Long S.R., 1998. ø-Galactoside uptake in Rhizobium meliloti: isolation and characterization of agpA, a gene encoding a periplasmic binding protein required for melibiose and raffinose utilization. J. Bacteriol., lB0: 5739-5748.

Galibert I.l, Finan T'.M., Long S.R., PuhlerA., Abola P, Ampe E, Barloy-Hubler F., Barnett M.J., IJecker 4., Boistard P, Bothe G., Boutry M., Bowser L., Buhrmester J., Cadieu 8., Capela D., Chain P, Cowie Á'., Davis R.W., Dreano S., Federspiel N.4., Fisher R.F,, Gloux S., Godrie f'., Goffeau A,., Golding 8., GouzyI., Gurjal M., Hernandez-Lucas I., HongA., I Iuizar L., Hy¡nan R.W., Jones T., Kahn D., Kahn M.L., Kalman S., Keating D.Iì., Kiss E., Komp C., Lelaure V, Masuy D., Palm C., Peck M.C., Pohl T.M., Portetelle D., Purnelle 13., Ramsperger U., Surzycki R., Thebault P, Vandenbol M., Vorholter EJ., Weidner S., Wells D.H., Wong K., Yeh K.C., and BatutI., 200i. The composite genome of the legume s)¡mbi o n t S ino rhizob i um me liloti. Science, 293:- 668-67 2.

García-de los Santos A'., Morales A,., Baldomá L., Clark S.R.D., Brom S., Yost C.K., ÉIernández-Lucas I., Aguilar ]., and Hynes M.F., 2002. The g/cB locus of Rhizobiunt IeguminosarumVF3g encodes an arabinose-inducible malate synthase. Can. I. Mi- crobiol., 48:.922-932.

GardiolA., Arias 4., Cerveñansklf C., Gaggero C., and Martínez-Drets G., 1980. Biochem- ical characterization of a fructokinase mutantof Rhizobium meliloti. J. BacterioI.,I44: t2-16.

Gardiol4., Arias 4., Cerveñanskli C., and Martínez-Drets G., 1982. Succinate dehydrogenase mutant of Rhizobium melilotl. I. Bacteriol., l5l: 162l-1623.

r23 REF]]RENCES

Geer L.Y., Domrachev M., Lipman D.I., and Bryant S.H., 2002. CDART: protein homology by domain architecture. Genome Res., l2: l6I9-I623.

Giblin L., Boestcn 8., Turk S., Hooykaas P, and O'Gara E, 1995. Signal transduction in the Ilhizobium meliloti dicarboxylic acid transport system. FEMS Microbiol. Lett., 126: 25-30.

Glazebrook I. and Walker G.C., 1991. Genetic techniques in Rhizobium meliLoti. Meth- ods Iìnzym oI., 2O4: 398-4 lB.

GonzâIez V, SautamarÍa R.L, Bustos P, Flernández-Gonzátlez I., Medrano-Soto ,{., Moreno-Hagelsieb G., |anga S.C., Ramírez M.4., Jiménez-Iacinto V, Collado-Vides I., and Dávila G., 2006. The partitioned Rhizobium etli genome: genetic and metabolic redundancy in seven interacting replicons. Proc. Natl. Acad. Sci. U.S.A., I03: 3834- :ì839.

Gordon D.M., Ryder M.H., Heinrich K., and Murphy PI., 1996. An experimental test of tlre rlrizopine concept in Rhizobium meliloti. Appl. Environ. Microbiol., 62: 3991- 3996.

Gosselin I., Wattraint O., Riboul D., Barbotin J., and Portais 1.,2001. A deeper investigation on carbohydrate cycling in Sinorhizobium melilotL FEBS Lett., 499: 45-49. tlalling S.M., Peterson-Burch 8.D., Bricker 8.J., Zuerner R.L., QingZ., Li L.L., Kapur V, Alt D.lì, and Olsen S.C., 2005. Completion of the genome secluence of Brucellaabortus and comparison to the highly similar genomes of Brucella melitetzs¿s and Brucella su.is. J. Bacteriol., lB7: 2715-2726. l loneycutt R.J., McClelland M., and Sobral 8.W., 1993. Physical map of the genome of Rltizobiunt meliloti 102I. I. Bacteriol., 175: 6945-6952.

I-Iornez J.P, El Guezzar M., and Derieux I.C., I989. Succinate transport in Rhizobiunt meliloti: Characteristics and impact on symbiosis. Curr. Microbiol., 19:207-2I2.

HòrnezJ.P, TilninouniM., Defives C., and Derieuxl.C., 1994. Unaffected nodulation and nitrogen fixation in carbohydrate pleiotropic mutantsof Rhizobium meliloti. Curr. Mi- crobiol., 28:225-229.

I'luang M.L., Cangelosi G.4., Halperin W., and Nester E.W, 1990. A chromosomal Agrobacterium tumefacíens gene required for effective plant signal transduction. I. Bacteriol., 17 2: IBI4-1822.

t24 III]T;ERENCES

LIynes M.E, Quandt I., O'Connell M.P, and Pühler .A., 1989. Direct selection for curing and deletion of Rhizobium plasmids using transposons carrying the Bacillus subtilis sacB gene. Gene, 7B: I I l-120. l'lynes M.E, Simon R., Müller P, Niehaus K., Labes M., and Pühler 4., l986.'fhe two nregaplasmidsof Rltizobium melilotiare involved in the effective nodulation of alfalfa. Mol. Gen. Genet., 2O2:356-362.

Irigoyen J.J., Sanchez-Diaz M., and Emerich D.W., 1990. Carbon metabolism enzymes of Rhizobium melíloti cultures and bacteroids and their distribution within alfalfa nodules. Appl. Environ. Microbiol., 56: 2587-2589.

Jensen E.S. and Hauggaard-Nielsen H., 2003. How can increased use of biological N2 lixation in agriculture benefit the environment? Plant Soil, 252: I77-T86.

Jiang G., Krishnan 4.H., Kim Y.W., Wacek T.J., and Krishnan H.8., 2001. A functional nz-yo-inositol dehydrogenase gene is required for efficient nitrogen fixation and competitiveness of Sinorhizobiunt fi'edii USDAIgl to nodulate soybean (Glycine max ll.) Merr.). J. Bacreriol., l83: 2595-2604.

Jiménez-7,urdo I.L, García-Rodriguez EM., and Toro N., 1997. The Rhizobium melilotí putAgene: its role in the establishment of the symbiotic interaction with alfalfa. Mol. Microbiol.,23: 85-93.

Jiménez-Zurdo I.I., van Dillewijn P, Soto M.i., de Felipe M.R., Olivares I., and Toro N., 1995. Characterization of a Rhizobium melilofl proline dehydrogenase mutant altered in nodulation efficiency and competitiveness on alfalfa roots. Mol. Plant-Microbe In- teract., B:492-498.

Jones J.D. and Gutterson N., 1987. An efficient mobilizable cosmid vector, pRK7B13, and its use in a rapid method for marker exchange in Pseudomonas fluoresce¡zs strain I-IV37a. Gene, 6l : 299-306.

Iordan D.C., 1984. Bergey's manual of systematic bacteriology, vol. l, chap. Family III. Rltizobiaceae CONN I938, 3211¿. Baltimore: Williams & Wilkins, 234-256.

|ording D., Uhde C., Schmidt R., and Pühler 4., I994. The Ca-dicarboxylate transport system of Rhizobium nzeliloti and its role in nitrogen fixation during symbiosis with alfalfa (Medicago satiua). Cell Mol. Life Sci.,50: 874-883.

Kaneko '1, Nakamura Y., Sato S., Asamizu 8., Kato T., Sasamoto S., Watanabe A.., idesawa K., Ishikawa ,{., Kawashima K., Kimura T., Kishida Y., Kiyokawa C., Kohara M., Mat- sumoto M., Matsuno 4., Mochizuki Y., Nakayama S., Nakazaki N., Shrimpo S., Sugi-

r25 TIETTERENCES

moto M., Takeuchi C., Yamada M., and Tabata S., 2000. Complete genome structure of the nitrogen-fixing syrnbiotic bacterium Mesorhizobium loti. DNA Res., 7: 331-338.

Kemner I.M., LiangX., and Nester E.W., 1997. Tlire Agrobacterium tumefaciensvirulence gene chuU is part of a putative ABC-type sugar transport operon. I. Bacteriol., l7g: 24s2-8.

Knee E.M., Gong F.C., Gao M., Teplitski M., |onesA.R., FoxworthyA., MortA.J., and Bauer W.D., 2001 . Root mucilage from pea and its utilization by rhizosphere bacteria as a sole carbon source. Mol. Plant-Microbe Interact., 14:775-84.

Lambert 4., Østerås M., Mandon K., Poggi M.C., and Rudulier D.L., 2001. Fructose up- takc in Sinorhizobium meliloti is mediated by a high-affinity ATP-binding casserte transport system. I. Bacteriol., l83: 4709-4717. l.edebur FI., Gu 8., Sojda f., and Nixon 8.T., 1990. Rhizobium meliloti and Rhizobium legumtnosarum dctD gene products bind to tandem sites in an activation sequence lo- catcd upstream of sigma 54-dependent dctApromoters. I. Bacteriol ., 172:3BBB-3897.

Leigh I.4., Signer E.R., and Walker G.C., 1985. Exopolysaccharide-deficient mutants of Illtizobium meliloti that form ineffective nodules. Proc. Natl. Acad. Sci. U.S.A., 82: 6231-6235.

Lessie T.G. and Phibbs PV, 1984. Alternative pathways of carbohydrate utilization in pseudomonads. Annu. Rev. Microbiol., 38: 359-388. l,ocher K.P, 2004. Structure and mechanism of ABC transporters. Curr. Opin. Struct. Iliol., l4: 426-43I.

Lodwig Ii., Leonard M., Marroqui S., \¡Vheeler T.R., Findlay K., Downie I.4., and Poole IìS., 2005. Role of polyhydroxybutyrate and glycogen as carbon storage compounds in pea and bean bacteroids. Mol. Plant-Microbe interact.,lB: 67-74.

Lodwig E. and Poole P, 2003. Metabolism of Rhizobium bacteroids. Crit. Rev. Plant Sci., 22:37-78.

Lodwig E,.M., Hosie A.Il.E, Bourdès 4., Findlay K., Allaway D., Karunakaran R., Do'nmie I.4., and Poole PS., 2003. Amino-acid cycling drives nitrogen fixation in the legume- Rhizobium sl.rnbiosis. Nature, 422: 7 22-7 26.

Long S., McCune S., and Walker G.C., 1988. S)rmbiotic loci of Rhizobium meliloti identi- lìed by random'fnphoAmutagenesis. J. Bacteriol., 170: 4257-4265.

t26 RETTERENCES

Lynch WIl., Macleod J., and Franklin M., 1975. Effect of temperature on the activity and synthesis of glucose-catabolizing enzyrnes in Pseudomonas fluorescens. Can. I. Microbiol., 2l: 156O-157 2.

Martínez-De Drets G. and Arias A.., lg|2.Enzymatic basis for differentiation of Rhizo- bi,untinto fast- and slow-growing groups. J. Bacteriol., 109: 467-470.

MauchlineT.H., FowlerJ.E., EastA.K., SartorA.L.,Zaheer R., HosieA.H.F., Poole PS., and Finan lìM., 2006. Mapping the Sínorhizobium meliloti 102I solute-binding protein- dependent transportome. Proc. Natl. Acad. Sci. U.S.A., 103: 17933-8.

McDermott T.R. and Kahn M.L., 1992. Cloning and mutagenesis of the Rhizobium

meliloti isocitrate dehydrogenase gene. J. Bacteriol ., 17 4: 4790-47 97 .

McKay I.4., Glenn ,\.R., and Dilwo¡th M.I., I985. Gluconeogenesis in Rhizobium leguminosarum MNF3841. J. Gen. Microbiol., l3l:2067-2073.

McNeil M., Darvill4.G., Fry S.C., and Albersheim P, I984. Structure and function of the primarycell walls of plants. Annu. Rev. Biochem., 53: 625-63.

Meade I-I.M., l.ong S.R., Ruvkun G.8., Bro'utm S.E., and Ausubel F.M., 1982. Physical and genetic characterization of symbiotic and auxotrophic mutants of Rhizobium meliloti induced by transposon TnS mutagenesis.I. Bacteriol., l4g: Il4-22.

Meijer W.G., de Boer P, and van Keulen G., 1997. Xanthobacter flauus employs a single triosephosphate isomerase for heterotrophic and autotrophic metabolism. Microbi- ology,143: 1925-31.

Mcrgaert P, Uchiumi T., A-lunni B., Evanno G., Cheron 4., Catrice O., Mausset 4.8., Barloy-Flubler F., Galibert F., Kondorosi 4., and Kondorosi E., 2006. Eukaryotic control on bacterial cell cycle and differentiation in the Rhizobium-legume s)¡mbiosis. Proc. Natl. Acad. Sci. U.S.A., 103: 5230-5235.

Miller J., 1972. Experiments in Molecular Genetics. Cold Spring Ilarbor Laboratory, Cold Spring Harbor', NY.

Miller L.D., Yost C.K., Flynes M.E, and Alexandre G.,2007. The major chemotaxis gene cluster of Rltizobium legumínosarumbv. uiciae is essential for cornpetitive nodulation. Mol. Microbiol., 63: 348-362.

Miller-Williams M., Loewen PC., and Oresnik I.I., 2006. Isolation of salt-sensitive mutants of Sinorhizobium meliloti strain Rm1021. Microbiolo{y,152: 2049-2059.

L27 IIEþ-ERENCES

Mitsch M.]., CowieA., and Finan T.M.,2007. Malic enzyrne cofactor and domain require- nìents for symbiotic N2 fixationby Sinorhizobiutn meliloti.J. Bacteriol., lBg: 160-I68.

Motinne-LclccozY. and Weaver R.W., 1995. Plasmids influence growth of Rhizobia in the rhizosphere of clover. Soil Biol. Biochem., 27: I001-1004.

Mortimer M.w., McDermott T.R., York G.M., walker G.c., and Kahn M.L., ig99. citrate sylìthase mutants of Sinorhizobium meliloti are ineffective and have altered cell sur- fäce polysaccharides. I. Bacteriol., lBl: 7608-7613.

Mulligan J.T. and Long S.R., 1985. Induction of Rhizobium meliloti nodC expression by plant exudate requires nodD. Proc. Natl. Acad. Sci. U.S.A., B2: 6609-13.

Murphy PI., Wexler W., Grzemski W., Rao I.P, and Gordon D., lgg5. Rhizopines-their role irr symbiosis and competition. Soil Biol. Biochem., 2Z: SZ5-529.

Oke V and Lor-rg S.R., 1999. Bacteroid formation in the Rhizobium-legume symbiosis. Curr. Opin. Microbiol., 2: 641-646.

Oresnik L, Pacarlmuk L., O'Brien S., Yost C., and Hlmes M., lgg8. Plasmid-encoded catabolic genes in Rhizobium leguminosarum bv. triþIii: Evidence for a plant- inducible rhamnose locus involved in competition for nodulation. Mol. Plant-Mi- crobe Interact., I l: i I75-I lB5.

Oresnik I.J., Charles T.C., and Finan T.M., I994. Second site mutations specifically sup- press the Fix- phenotype of Rhizobium melíloti nduF mulations on alfalfa: identi- fìcation of a conditional nduF-dependent mucoid colony phenotype. Genetics, 136: 1233-1243.

Oresnik I.J., Liu S.L., Yost C.K., and Hlmes M.E, 2000. Megaplasmid pRme2Olta of Sinorhizobium nteliloti is not required for viability. I. Bacteriol.,lB2:3582-6.

Oresnik I.J., Twelker S., and Hlmes M.E, 1999. Cloning and characterization of a Rhi- zobiunz leguminosarum geîe encoding a bacteriocin with similarities to RTX toxins. Appl. Iìnviron. Microbiol., 65: 2833-2840.

Østerås M., Driscoll 8.T., and Finan T.M., 1995. Molecular and expression analysis of the Rhizobium ntelilotlphosphoenolpyruvate carboxykinase (pckA) gene. J. Bacteriol., 177:1452-1460.

Østerås M., Driscoll8.T., and Finan T.M., lg97a. Increased pyruvate orthophosphate dikinase activity results in an alternative gluconeogenic pathway in Rhizobíum (Sinorhi- zobium) meliloti.Microbiology, 143: I639-t648.

I28 REI:ERENCES

Østerås M., Finan T.M., and Stanley J., 1991. Site-directed mutagenesis and DNA sequence of pclcA of Rhizobiu¡n NGR234, encoding phosphoenolpyruvate carboxyki- uasc: gluconeogenesis and host-dependent syrnbiotic phenotlpe. Mol. Gen. Genet., 23O:257-269.

Østerås M., O'Brien S.'{., and Finan T.M., I997b. Genetic analysis of mutations af'fecting pclc4rcguLation in Illtizobium(sinorhizobium) meliloúi. Genetics, tAZ: IS2L_153I.

Pahel G., Bloom F.R., and Tyler 8., 1979. Deletion mapping of the polA-metB region of the Escherichia coli chromosome. I. Bacteriol., l38: 653-6.

Paulsen I.l'., Seshadri R., Nelson K.8., Eisen I.4., Heidelberg I.F., Read T.D., Dodson R.I., lJmayam 1,., Brinkac L.M., Beanan M.1., Daugherty S.C., Deboy R.T., Durkin A.S., Kolonay I.E, Madupu R., Nelson W.C., Ayodeji 8., Kraul M., Shetty I., Malek I., Aken S.E.V., Riedmuller S., Tettelin H., Gill S.R., \Mhite O., Salzberg S.L., Hoover D.L., Lindler L.8., Halling S.M., Boyle S.M., and Fraser C.M.,2002.The Brucella suis genome reveals fundamental similarities between animal and plant pathogens and sl.rnbionts. proc. Natl. Acad. Sci. U.S.A., 99: 13148-13153.

Pedrosa EO. and Zancan G.T., 1974. L-Arabinose metabolismin Rhizobium japonicum. J. Bacteriol., ll9: 336-8.

Phillips D.4., Sande E.S., Vriezen I.A.C., de Bruijn EJ., Le Rudulier D., and Ioseph C.M., i998. A new genetic locus in Sinorhizobium melilotiis involved in stachydrine utilization. Appl. Environ. Microbiol., 64: 3954-3960.

Poole P and AllawayD., 2000. Carbon and nitrogen metabolism in Rlzizobium. Adv. Mi- crob. Physiol., 43: I 17-163.

Portais J.C., Tavernier P, Gosselin I., and Barbotin J.N., 1999. Cyclic organization of the carbohydrate metabolism in Slnorhizobium melíloti. Eur. I. Biochem., 265: 473-480.

Poysti N.]., Loewen E.D.M., Wang 2., and Oresnik I.1., 2007. Sínorhizobium meliloti pSymB encodes genes necessary for arabinose transport and catabolism. Microbiol- ogy,153: 727-736.

Prell J. and Poole P, 2006. Metabolic changes of rhizobia in legume nodules. Trends Mi- crobiol., l4: I6l-168.

Quandt I. and Hlmes M.E, 1993. Versatile suicide vectors which allow direct selection for gene replacement in Gram-negative bacteria. Gene, 127: 15-21.

Quevillon 8., Silventoinen V, Pillai S., Harte N., Mulder N., Apweiler R., and Lopez R., 2005. InterProScan: protein domains identifier. Nucleic Acids Res., 33: Wi i6-W120.

129 REFERENCES

Raffa R.G. and Raivio T.L.,2002. A third envelope stress signal transduction pathway in Escherichia coli. MoL Microbiol., 45: i599-t6t t.

Rastogi VK. and Watson R.I., 1991. Aspartate aminotransferase activity is required for aspartate catabolism and symbiotic nitrogen fixation in Rhizobíum meliloti. J. Bacte- riol., 173: 2879-2887.

Richardson I.S., Hynes M.F,, and Oresnik I.1.,2004. A genetic locus necessary for rhamnose uptake and catabolism in Rhizobium leguminosa,rumbv. triþlii. J. Bacteriol., 186:8433-8442.

Richardson I.S. and Oresnik LJ., 2007. Rhamnose transport is sugar-kinase (RhaK) dependent in Rhizobium leguminosarumbv. triþlii. J. Bacteriol., In revision.

Riley M., Abe T., Arnaud M.8., Berlyn M.K.B., Blattner ER., chaudhuri R.R., Glasner I.D., Horiuchi T., Keseler LM., Kosuge T., Mori H., Perna N.T., Plunkett G., Rudd K.E., Serres M.H., Thomas G.H., Thomson N.R., Wishart D., and Wanner 8.L., 2006. Escherichia coliK-12: a cooperatively developed annotation snapshot-200S. Nucleic Acids Res., 34: l-9.

Robleto E.4., Kmiecik K., Oplinger 8.S., Nienhuis I., and Triplett E.W., i998. Trifolitoxin production increases nodulation competitiveness of. Rhizobium etli CE3 under agricultural conditions. Appl. Environ. Microbiol., 64: 2630-2633.

Robleto 8.A.., Scupham 4.J., and Triplett E.W., 1997. Trifolitoxin production in Rhizo- bium etli strain CE3 increases competitiveness for rhizosphere colonization and root nodulation of Phaseolus uulgaris in soil. Mol. PIant-Microbe Interact., lO 228-233.

Romano A.H. and Conway T., 1996. Evolution of carbohydrate metabolic pathways. Res. Microbiol., 147: 448-455.

Rosenberg C., Casse-Delbart F., Dusha L, David M., and Boucher C., IgB2. Megaplasmids in the plant-associated bacteria Rhizobium melilotiand Pseudomonas solanacearum. I. Bacteriol., 150: 402-406.

Rosenblueth M., Hynes M.E, and MartÍnez-Romero E., 1998. Rhizobium tropici teu genes involved in specific uptake of Phaseolus uulgaris bean-exudate compounds. Mol. Gen. Genet.,25B: 587-598.

Sambrook I., Fritsch E., and Maniatis T., 1989. Molecular cloning: a laboratory manual. 2nd ed., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

Sangari E|., Aguero J., and Garcia-Lobo I.M., 2000. The genes for erythritol catabolism are organized as an inducible operon in Brucelln abortus. Microbiology, 146: 487-95.

130 REI.-ERENCES

Simon R., Priefer U., and Puhler 4., i983. A broad host range mobilization system tor in uiuo genetic-engineering: Tiansposon mutagenesis in Gram-negative bacteria. Iìio/'Iechnology, I : 784-79I.

Simon R., Quandt I., and Klipp W., 1989. New derivatives of transposon TnS suitable l'or mobilization of replicons, generation of operon fusions and induction of genes in Gram-negative bacteria. Gene, B0: I6t-169.

Soedarjo M. and tsorthakur D., i998. Mimosine, a toxin produced by the tree-legume Leucaenaprovides a nodulation competition advantage to mimosine-degradingRhi- zobium strains. Soil Biol. Biochem.,30: 1605-I613.

Spaink Il.Iì, 2000. Root nodulation and infection factors produced by rhizobial bacteria. Annu. Rev. Microbiol.,54: 257-BB.

Sperry J.F. and Robertson D.C., 1975. Erythritol catabolism by Brucella abortus. J. Bacte- riol., l2l: 6l9-30.

Sprenger G.,{., 1995. Genetics of pentose-phosphate pathway enzyrnes of Escherichia coliK- 12. Arch. Microbiol., 164: 324-330.

Srere PA., 1987. Complexes of sequential metabolic enzyrnes. Annu. Rev. Biochem., 56: t\9-124.

Stoscheck C.M., 1990. Quantitation of protein. Methods Enzlmrol., lB2: 50-68.

Stowers M.D., 1985. Carbon metabolism in Rhizobiurz species. Annu. Rev. Microbiol., 39: B9-108.

StreitW.R., Ioseph C.M., and Phillips D.4., 1996. Biotin and otherwater-solublevitamins are key growth factors for alfalfa root colonizationby Rhízobium meliLoti 1021. Mol. PIant-Microbe Interact., 9: 330-338.

Timmers A.C.J., Auriac M.C., and Truchet G., 1999. Refined analysis of early syrnbiotic steps of the Rhtzobium-Medicago interaction in relationship with microtubular cy- toskeleto n rearran gements. Development, L26: 3617 -3628. lbmbolini R. and Nuti M.P, 1989. Poly (B-hydroxyalkanoate) biosynthesis and accumulation by different Rhizobium species. FEMS Microbiol. Lett., 60: 299-304.

1ì'iplett 8.W., 1990. Construction of a symbiotically effective strain of RhizobiLmt legu-

m ino s ar um bv. t riþ I ii vnth increased no dulation competitiveness. Appl. Environ. Mi - crobiol.,56: 98-103.

131 REIIERENCES liiplett E.W. and Barta T.M., 1987. Trifolitoxin production and nodulation are necessary for the expression of superior nodulation competitiveness by Rhizobium legumi- rlosarutnbv. tr{olií strain T24 onclover. Plant Physiol., 85: 335-342.

'friplett E.W and Sadowsky M.J., 1992. Genetics of competition for nodulation of Iegumes. Annu. Rev. Microbiol., 46: 399-428.

Ucker D.S. and Signer 8.R., i978. Catabolite-repression-like phenomenon in Rhizobiunt meliloti. J. Bacteriol., 136: lL97-1200. van llrussel A.A.N., Bakhuizen R., van Spronsen PC., Spaink H.P, Tak T., Lugtenberg B.J.J., and Kijne 1.W., 1992. Induction of pre-infection thread structures in the leguminous host plant by mitogenic lipo-oligosaccharides of Rhizobium. Science, 257: 70-72. van Dillewijn P, Soto M.J., Villadas PI., and Toro N., 2001. Construction and environ- nrental release of. a Sinorhízobium meliloti strain genetically modified to be more competitive for alfalfa nodulation.Appl. Environ. Microbiol., 67: 3860-3865.

Venter A.P, lwelker S., Oresnik I.I., and Hynes M.F., 2001. Analysis of the genetic region encoding a novel rhizobiocin from Rhizobium leguminosarumbv. uiciae strain 306. Can. J. Microbiol., 47:495-502.

Vincent J.M., 1970. A manual for the practical study of root nodule bacteria. Blackwell Scientific, Oxford.

Watanabe S., Kodaki T., and Makino K., 2006. Cloning, expression, and characterization of bacterial L-arabinose 1-dehydrogenase involved in an alternative pathway of L-arabinose metabolism. J. Biol. Chem., 2Bl:2612-2623.

Watson R.J., Chan Y.K., 'vVheatcroft R., YangA.F., and Han S.H., 1988. Rhizobium meLiloti genes required for C¿-dicarboxylate transport and symbiotic nitrogen fixation are located on a megaplasmid. I. Bacteriol.,lTO: 927-934.

Watson R.I. and Rastogi VK., 1993. Cloning and nucleotide sequencing of Rhizobium meliloti aminotransferase genes: an aspartate aminotransferase required for symbiotic nitrogen fixation is atypical. I. Bacteriol.,lTS: l9l9-1928.

Willems r\., 2006. The taxonomy of rhizobia: an overview Plant Soil, 287: 3-14.

Willis L.B. and Walker G.C., 1998. The phbC (poly-B-hydroxybutyrate slmthase) gene of RltizobiLutt (Sinorhizobium) meliloti and characterization of phbC mutants. Can. I. Microbiol., 44: 554-564.

132 RIiT;EIIENCES

Yost C.K., Rath 4.M., Noel T.C., and Hynes M.E, 2006. Characterization of genes involved in erythritol catabolism in Rhizobium leguminosarumbv. uiciae. Microbiology, 152: 2061-2074.

Yost C.K., Iìochepeau P, and Hyres M.F., 1998. Rhizobium leguminosarum contains a group of geues that appear to code for methyl-accepting chemotaxis proteins. Micro- biology, 144: 1945-1956.

Young I.IìW., Crossman L.C., Johnston A.W.B., Thomson N.R., Ghazoui Z.F.,I{ull K.H., Wexler M., Curson A.R.i., Todd i.D., Poole PS., Mauchline T.H., East 4.K., Quail M.4., ChurcherC., Arrowsmith C., Cherevach I., ChillingworthT., Clarke K., CroninA., Davis P, Fraser 4., Hance 2., Hauser H., ]agels K., Moule S., Mungall K., Norbertczak H., lìabbinowitsch E., Sanders M., Simmonds M., \¡/hitehead S., and Parkhill I., 2006. The genonÌe of Rhizobium leguminosarum has recognizable core and accessory components. Genome Biol.,7: R34.

Yurgel S.N. and Kahn M.L., 2004. Dicarboxylate transport by rhizobia. FEMS Microbiol. Iìev.,2B: 489-501.

7,beng P, Sun J., van den l-Ieuvel I., and Zeng A.P, 2006. Discovery and investigation of a new second triose phosphate isomerase in Klebsiella pneumoniae.I. Biotechnol., 125:462-73.

r33