Bradyrhizobium Japonicum Genes for Life in Specific Hosts

Research Collection

Doctoral Thesis

Bradyrhizobium japonicum genes for life in specific hosts

Author(s): Koch, Marion

Publication Date: 2011

Permanent Link: https://doi.org/10.3929/ethz-a-006410064

Rights / License: In Copyright - Non-Commercial Use Permitted

This page was generated automatically upon download from the ETH Zurich Research Collection. For more information please consult the Terms of use.

ETH Library DISS. ETH No. 19577

Bradyrhizobium japonicum genes for life in specific hosts

A dissertation submitted to the ETH Zürich

for the degree of DOCTOR OF SCIENCES

presented by MARION KOCH Dipl. Biol., Universität Konstanz

born on January 1st, 1983 citizen of Germany

Prof. Dr. Hauke Hennecke, examiner Dr. Gabriella Pessi, co-examiner Prof. Dr. Thomas Boller, co-examiner Prof. Dr. Julia Vorholt, co-examiner Prof. Dr. Samuel Zeeman, co-examiner

2011

Table of contents

Thesis summary 1

Zusammenfassung 3

Chapter 1: Introduction 7

1.1 Rhizobia and Bradyrhizobium japonicum 8 Historical overview 8 Bradyrhizobium japonicum 8 Nitrogen fixation 9 1.2 Molecular basis of the rhizobia-legume symbiosis 11 Legume flavonoids 11 Rhizobial Nod factors 11 Nod factor perception in legume roots 13 Plant responses to Nod factors 14 Formation of a root nodule 15 1.3 Global strategies to monitor rhizobial gene and protein expression in symbiosis 18 1.4 Host speciﬁcity 25 NodD proteins 25 Other nod genes as host-specific determinants 26 Polysaccharides and secreted proteins as host-specific determinants 26 Cultivar specificity 27 Host-speciﬁc adaptation 28 1.5 Carbon metabolism in rhizobia 30 Carbon metabolism in free-living rhizobia 30 Carbon metabolism in symbiotic rhizobia 32 Oxalotrophic bacteria 35 1.6 Aim of this work 40

Chapter 2: Characterization of two carbonic anhydrase genes 41

2.1 Abstract 42 2.2 Introduction 43 2.3 Material and methods 46 Bacterial strains, media and growth conditions 46 DNA work 46 Construction of Δbll2065-2066 deletion and bll4865::pRJ6226 insertion mutants 47 Plant growth 49 Quantitative real-time PCR 49 2.4 Results 50 Bradyrhizobium japonicum possesses five carbonic anhydrase genes 50 Construction of mutations in two β carbonic anhydrase genes bll2065 and bll4865 51 Symbiotic phenotype of ∆bll2065-2066 and bll4865::pRJ6226 strains 53 Expression studies of B. japonicum carbonic anhydrase genes 54 2.5 Discussion 56

Chapter 3: Rhizobial adaptation to hosts, a new facet in the legume root- 59 nodule symbiosis

3.1 Abstract 60 3.2 Introduction 61 3.3 Results and discussion 64 Transcriptome analysis of B. japonicum in root nodules of cowpea, siratro, and 64 soybean Proteome analysis of bacteroids from cowpea, siratro, and soybean nodules 65 The stringent data set: where host-specific transcriptomes and proteomes overlap 67 An approved determinant for bacteroid adaptation to life in siratro nodules 67 In search for the substrate of Bll1600-1604 73 Further candidates with a perspective to act in a host-responsive manner 75 The relaxed data set: transcriptomes and proteomes combined 76 3.4 Concluding remarks 78 3.5 Material and methods 79 DNA methods and construction of mutant strains 79 Phenotypic MicroArray tests 81 Growth and sensitivity assays 81 Plant material, inoculation, and growth conditions 81 Transcriptome analyses 82 Proteome analyses 83 3.6 Addendum: Further candidates with a perspective to act in a host-responsive 86 manner Results and discussion 86 Mutant construction 88

Chapter 4: Oxalotrophy in Bradyrhizobium japonicum 89

4.1 Abstract 90 4.2 Introduction 91 4.3 Material and methods 94 Bacterial strains, media and growth conditions 94 Oxalotrophic growth 94 Isothermal calorimetry 95 DNA methods and construction of mutant strains 96 Symbiotic growth analysis: Plant material, inoculation and cultivation 98 Determination of the oxalate content of roots and nodules of soybean, mungbean, 98 cowpea and sirato infected with B. japonicum 4.4 Results 100 Identification and transcriptional analysis of the frc, oxc genomic region 100

Construction of ∆oxlT1+2 and ∆frc-oxc deletion mutants 102 Free-living growth analysis in presence of various carbon sources 102 Oxalate content in roots and root nodules induced by B. japonicum 105

Symbiotic properties of the ∆frc-oxc and ∆oxlT1+2 mutants 106 4.5 Discussion 107 4.6 Addendum: Competitiveness in symbiosis 112

Chapter 5: Future perspectives 113

5.1 Characterization of two β-class carbonic anhydrase (CA) genes 114 5.2 Host-specific adaptation 116 5.3 Oxalotrophy in B. japonicum 118

References 121

Publications 131

Curriculum vitae 133

Acknowledgements 135

Thesis summary

Bradyrhizobium japonicum is able to either persist as a free-living bacterium in soil or laboratory cultures or to enter a symbiosis with various legume plants such as soybean, mungbean, cowpea and siratro. These symbiotic interactions result in the formation of nodules at the roots of their hosts. At the heart of these symbioses is the ability of the bacterial endosymbiont to reduce atmospheric nitrogen to ammonia, which is given to and used by the plant. This natural fertilization enables legume plants to grow on nitrogen-poor soil. In return, the bacterium is supplied with carbon sources such as succinate generated by the symbiotic partner as a consequence of photosynthetic CO2 fixation.

The first part of this thesis deals with the characterization of a carbonic anhydrase gene

(bll2065) which was previously discovered to be specifically and highly expressed during symbiosis of B. japonicum with soybean when compared to free-living aerobically grown cells (Pessi et al. 2007). A complementary proteomics approach recently showed that a second B. japonicum carbonic anhydrase, Bll4865, was also expressed during soybean symbiosis (Delmotte et al. 2010). Therefore, we decided to characterize these two carbonic anhydrase genes. Although these genes are expressed in bacteroids, mutant analysis showed that each of them by itself is not essential for B. japonicum to successfully enter and establish a nitrogen fixing symbiosis with its soybean host.

In the second part, we aimed at monitoring global gene expression changes of B. japonicum in response to different host environments. To achieve this goal, we analyzed the transcriptome as well as the proteome of B. japonicum bacteroids in root nodules from soybean, cowpea and siratro. These analyses revealed that B. japonicum bacteroids indeed

Thesis summary respond partly disparately in gene expression, depending on who is the symbiotic plant partner. In total two genes/proteins for cowpea, five for siratro, and seven for soybean were identified. One gene cluster for a predicted ABC-type transporter (blr1601-1604) plus a monooxygenase (blr1600) was identified to be specifically expressed during symbiosis with siratro. Mutant analysis showed that this operon is indeed more important for the B. japonicum-siratro symbiosis than for the symbiosis with soybean and cowpea.

Complementation analysis revealed that the monooxygenase gene (blr1600) alone cannot compensate the host-specific deletion mutant phenotype on siratro. At least one of the ABC transporter genes is needed for successful complementation. Thus, based on two global studies, we could identify a host-specific adaptation determinant for B. japonicum.

Finally, a third part of this work was dedicated to gain insight into a particular aspect of carbon catabolism of B. japonicum. Based on genome annotation and microarray data, B. japonicum possesses and expresses genes for the uptake and metabolism of oxalate. Similar genes were previously shown to be important for the oxalotrophic lifestyle of bacteria such as

Oxalobacter formigenes. We showed here that B. japonicum is able to use the C2 compound oxalate as the sole source of carbon. Moreover, mutant analysis indicated that the genes frc

(coding for a formyl-CoA transferase) and oxc (coding for an oxalyl-CoA transferase) are essential for oxalate degradation. Interestingly, a preliminary analysis suggested that frc and oxc are important for competition for nodule occupancy. The ability to degrade oxalate might confer an advantage for B. japonicum during the establishment of a symbiosis with legume plants.

Zusammenfassung

Bradyrhizobium japonicum ist einerseits fähig, als frei lebendes Bakterium im Erdboden sowie in Laborkulturen zu existieren, und kann andererseits Symbiosen mit diversen

Hülsenfrüchtlern wie beispielsweise Sojabohnen, Mungbohnen, Langbohnen und Siratro eingehen. Die symbiotische Wechselbeziehung bewirkt die Ausbildung von Knöllchen an den Wurzeln der Hülsenfrüchtler. Von zentraler Bedeutung bei der Symbiose ist die

Fähigkeit des bakteriellen Endosymbioten, atmosphärischen Stickstoff in Ammonium umzuwandeln, welcher dann der Pflanze zur Verfügung gestellt und von ihr verwertet wird.

Dieser natürliche Düngungsprozess ermöglicht es genannten Hülsenfrüchtlern auch auf stickstoffarmen Böden zu gedeihen. Als Gegenleistung wird das Bakterium in der Symbiose mit Kohlenstoffquellen wie beispielsweise Succinat versorgt, welche die Hülsenfrüchtler als

Konsequenz der photosynthetischen CO2-Fixierung erzeugen.

Der erste Teil dieser Arbeit beschäftigt sich mit der Charakterisierung eines Carboanhydrase-

Gens (bll2065), welches man bei einer vorhergehenden Studie identifiziert hat. Jene Arbeit widmete sich der Identifizierung von Genen, welche in B. japonicum während der Symbiose mit Sojabohnen im Vergleich zu freilebenden, aerob gezüchteten Zellen hoch exprimiert waren (Pessi et al. 2007). Bei einer komplementären Studie, welche das Proteom von B. japonicum-Bakteroiden untersuchte, wurde das Genprodukt einer weiteren Carboanhydrase,

Bll4865, entdeckt (Delmotte et al. 2010). Aufgrund dessen wurden Mutanten von beiden

Carboanhydrase-Genen hergestellt und analysiert. Obwohl beide Gene in Bakteroiden hoch exprimiert sind, haben Mutantenanalysen gezeigt, dass jedes der beiden Gene allein nicht essentiell ist, um eine Stickstoff fixierende Symbiose von B. japonicum mit der Sojabohne einzugehen.

Zusammenfassung

Im zweiten Teil dieser Arbeit war die Zielsetzung, die Veränderung der Genexpression von

B. japonicum in unterschiedlichen Wirtspflanzen zu beobachten. Zu diesem Zweck wurde das

Transkriptom sowie das Proteom von B. japonicum-Bakteroiden untersucht. Zur

Proteomanalyse wurden die Bakteroide aus Wurzelknöllchen der Sojabohne, Langbohne und

Siratro isoliert. Diese Untersuchungen haben gezeigt, dass sich die Genexpression von

Bakteroiden von B. japonicum in der Tat dem symbiotischen Partner anpasst. Es wurden insgesamt zwei Gene/Proteine für Langbohnen, fünf für Siratro, sowie sieben für die

Sojapflanze gefunden. Ein Gencluster, der vermutlich für einen ABC-Transporter (blr1601-

1604) sowie eine Monooxygenase (blr1600) kodiert, wurde in der Symbiose mit Siratro spezifisch exprimiert. Mutantenanalysen haben gezeigt, dass dieses Operon für die B. japonicum-Siratro Symbiose wichtiger ist als für die Symbiosen mit der Sojabohne und

Langbohne. Weiterführende Studien ergaben, dass das Monooxygenase-Gen (blr1600) alleine nicht in der Lage ist, den wirtsspezifischen Mutanten-Phänotyp in Siratro aufzuheben.

Mindestens eines der ABC-Transporter-Gene ist unabdingbar für eine erfolgreiche

Komplementation. Zusammenfassend hat man mittels zweier globaler Studien ein Beispiel für wirtspezifische Gene von B. japonicum gefunden.

Im dritten und letzten Teil der Arbeit haben wir uns einem bestimmten Teilgebiet des

Kohlenstoff-Abbaustoffwechsels von B. japonicum gewidmet. Genomannotationen und

Microarray-Studien zur Folge besitzt und exprimiert B. japonicum Gene, welche für Proteine kodieren, die für die Aufnahme und die Verwertung von Oxalat verantwortlich sind.

Ähnliche Gene wurden bereits in dem Oxalat verwertenden Bakterium Oxalobacter formigenes untersucht und als wichtig für den oxalotrophen Lebensstil erachtet. Wir beobachteten, dass B. japonicum Oxalat als alleinige Kohlenstoffquelle verwerten kann.

Darüber hinaus machten Mutantenanalysen deutlich, dass die Gene frc (kodiert für die

Zusammenfassung

Formyl-CoA Transferase) sowie oxc (kodiert für die Oxalyl-CoA Decarboxylase) unabdingbar für einen oxalotrophen Lebensstil sind. Interessanterweise weisen vorläufige

Ergebnisse darauf hin, dass frc und oxc für eine kompetitive Knöllchenbesiedlung wichtig sind. Dies liesse darauf schliessen, dass die Verwertung von Oxalat für B. japonicum von

Vorteil ist, um eine Symbiose mit Hülsenfrüchtlern einzugehen.

Introduction

CHAPTER 1

Introduction

1.1 Rhizobia and Bradyrrhizobium japonicum

Historical overview

Rhizobia-legume symbioses are of great importance as they enable the cultivation of legume plants on nitrogen-poor soil without the need of nitrogen fertilizers. Traditional agriculture increased the supply of nitrogen either by cultivating legumes in crop rotation or co- cultivation of legumes with cereals such as wheat or barley (Downie 2007). In each case, the nitrogen derived from legume plants favored growth of the cereals. The earliest documentation of symbiosis may date baack more than 3000 years in the Book of Odes where the ancient Chinese character shu was introduced to describe soybean plants (Fig 1.1)

(Downie 2007). Soybean is one of the most important crop plants on earth. In 2009, 1.2 million metric tons of soybeans were produced in the US alone

(http://www.ers.usda.gov/news/soybeancoverage.htm). Nowadays soybean is not only cultivated for culinary value to produce soy-derived products, but also is base of ink, glue and textile production.

Fig 1.1. The characters represent the ancient Chinese character ‘shu’ which was introduced to describe soybean plants. On the left-hand side a plant with branched shoots and a root with three tear-drop lines are illustrated. The drops are thought to represent root nodules. Taken from Downie (2007).

Bradyrhizobium japonicum

Bradyrhizobium japonicum is a Gram-negative flagellated, rod shaped bacterium which belongs to the family of Bradyhizobiaceae of the order Rhizobiales within the α-subdivision of Proteobacteria. Bradyrhizobium japonicum resides as a free-living bacterium in soil or as a

Introduction nitrogen-ﬁxing symbiont in determinate root nodules of legume host plants, such as soybean

(Glycine max), mungbean (Vigna radiata), cowpea (Vigna unguiculata), or siratro

(Macroptilium atropurpureum) (Göttfert et al. 1990). For the agricultural cultivation of these legume plants B. japonicum is routinely used as inoculant.

During heterotrophic growth, B. japonicum respires aerobically in the presence of oxygen.

The bacterium can however respire anaerobically as a denitrifier when nitrate is provided as a terminal electron acceptor in the absence of oxygen. Furthermore, under free-living microoxic conditions supplemented with hydrogen and carbon dioxide B. japonicum can grow chemoautotrophically, coupling hydrogen oxidation to aerobic respiration and carbon dioxide fixation.

In 2002, the full genome sequence of B. japonicum became available (Kaneko et al. 2002). It was shown, that the genome of B. japonicum consists of a single circular chromosome which is 9,105,828 bp in length with an average GC content of 64.1%. No plasmid was detected.

The chromosome comprises 8,317 potential protein-coding genes and 52 RNA genes

(Kaneko et al. 2002). This allowed the design of a B. japonicum Aﬀymetrix GeneChip, which also includes intergenic regions thus enabling genome-wide transcriptional analysis (Hauser et al. 2007). Moreover, B. japonicum is model organism for the studies of symbiotic nitrogen fixation since it can be genetically modified using conjugation methods.

Nitrogen fixation

Nitrogen is a major component of proteins and nucleic acids and is therefore an essential nutrient. Even though 78.1% of the earth’s atmosphere consists of the chemically inert nitrogen gas (N2), nitrogen availability is limited in many soils, and N2 is inaccessible for most of the living beings (Ferguson et al. 2010). Hence, nitrogen must be converted either chemically or biologically to a usable form that life on earth can profit from.

Introduction

One possibility is the chemical fixation also termed Haber process which reduces N2 to ammonia. The protonated form, ammonium, may be added directly as nitrogen fertilizer to soil. This process is extremely energy consuming and uses large amounts of fossil fuel with the accompanying production of carbon dioxide. In contrast, biological nitrogen fixation, which has so far only been described for prokaryotes, is innocuous for the environment. The key enzyme for this process is the metalloenzyme nitrogenase. Nitrogenase was shown to consist of two proteins namely the molybdenum-iron (MoFe) protein and the iron (Fe) protein (Dixon and Kahn 2004). The Fe-protein is a dimer of the nifH gene product that carries a 4Fe4S cluster. Two ATP bind to Fe-protein and, after hydrolysis, allow the transfer of one electron. The MoFe-protein is an α2β2 heterotetramer encoded by NifD and NifK and contains the MoFe cofactor which is the substrate-reducing site, and an FeS cluster referred to as the P-cluster (Dixon and Kahn 2004). The P-cluster participates in the electron transfer from the Fe-protein to the MoFe-cofactor. Thus, nitrogenase catalyzes the 6-electron reduction of dinitrogen to ammonium together with the 2-electron reduction of 2H+ to hydrogen driven by the hydrolysis of 16 ATP molecules.

In symbiotic diazotrophs the produced ammonium is transported to the plant cell cytosol and further assimilated into amino acids. The bacterial endosymbiont receives in exchange for ammonium C4-dicarboxylic acids from the plant, which are fueled into the tricarboxylic acid

(TCA) cycle (Prell and Poole 2006). Lodwig and coworkers proposed an extended model of the carbon-nitrogen-exchange which includes amino acid cycling between the plant cell cytoplasm and the bacteroid (Lodwig et al. 2003). The model suggests that amino acids imported from the plant cytosol to the bacteroid could be used to drive the transamination of oxaloacetate or pyruvate to aspartate or alanine. The resulting amino acid products could be exported and the keto acid which was released in the transamination reaction metabolized through the bacterial TCA cycle (Lodwig et al. 2003; Prell and Poole 2006).

Introduction

1.2 Molecular basis of the rhizobia-legume symbiosis

Legume flavonoids

The interaction between rhizobia and their leguminous host plants relies on the exchange of signaling compounds. The first signal is initiated in the plant by the secretion of ﬂavonoids into the rhizosphere (Redmond et al. 1986). Flavonoid compounds (2-phenyl-1,4- benzopyrone derivatives) act as chemoattractants for rhizobia. These signaling molecules are composed of a core structure of 15-carbon atoms, making up two aromatic rings A and B as well as an oxygen-containing heterocyclic ring (Fig 1.2). Previously, flavonoids were classified into various groups according to the oxidation level of the oxygen in the heterocyclic ring. The isoflavone genistein (Fig 1.2) was previously shown to be secreted by soybean and is an inducer of nod genes in B. japonicum (Kosslak et al. 1987).

Fig 1.2. Structural representation of the soybean derived flavonoid genistein which belongs to the group of isoflavons. The molecule is composed of a core structure of 15-carbon consisting of two aromatic rings A and B as well as an oxygen containing heterocyclic ring.

Rhizobial Nod factors

Flavonoids are perceived by rhizobial NodD proteins. These proteins are transcriptional regulators of the LysR family. Upon the presence of the appropriate flavonoid, NodD proteins bind to cis-regulatory elements which are called nod boxes. Many rhizobia possess several nodD homologs; however, each one responds to a certain set of ﬂavonoids (Perret et al. 2000). This binding leads to the activation of transcription of rhizobial genes which are predominantly involved in the production of bacterial signaling molecules called Nod factors

Introduction

(Dénarié et al. 1996). Recently for some photosynthetic Bradyrhizobium strains a Nod factor- independent nodulation strategy was described (Giraud et al. 2007). Giraud and coworkers showed that these strains enter the legume plant by cracks in the epidermis which result from emerging lateral roots (Giraud et al. 2007; Masson-Boivin et al. 2009).

Nod factors are lipo-chitooligosaccaride compounds that induce various responses in the appropriate symbiotic partner. These molecules are composed of a backbone of β-1,4 linked

N-acetyl-D-glucosamine residues which is synthesized with the help of nodBC gene products

(Fig 1.3). nodA encodes an acyltransferase which is required for the N-acylation of the aminosugar backbone with long-chain fatty acids. The products of other nod and nol genes are responsible for Nod factor modification such as the substitution of sulfate, acetate or fucose (Fig 1.3). Other nod gene products like those of nodI and nodJ were shown to be involved in the secretion of Nod factors (Spaink et al. 1995; Spaink 2000). Previously it was shown that several rhizobial species are capable of producing more than one type of Nod factor (Ardourel et al. 1994).

Fig 1.3. Rhizobial Nod factor structure with potential modifications: R1 carbamoyl and acetyl, R2 carbamoyl, R3 carbamoyl, R4 acetyl, sulphuryl, fucosyl, R5 mannosyl and glycerol, R6 arabinosyl. C16, C18, and C20 correspond to variations in the length of the fatty acyl chain. The number (n) of N-acetylglucosamine residues can vary between 1-4. Adapted from Downie and Walker (1999).

In B. japonicum the expression of nod genes is activated not only by NodD but also by the two-component regulatory system NodV and NodW (Göttfert et al. 1990). It was

Introduction experimentally shown that NodVW positively regulate nod gene expression in response to the soybean plant-derived isoflavone genistein (Fig 1.2) (Loh et al. 1997).

Nod factor perception in legume roots

Less is known about the mechanism how Nod factors are perceived at legume roots. Our current knowledge is based on genetic approaches which identified several Nod factor receptors for legume plants such as LjNFR1 and LjNFR5 in Lotus japonicus, PsSYM2A and

PsSYM10 in Pisum sativum, MtLYK3/MtLYK4 and MtNFP in Medicago truncatula, as well as GmNFR1α/β and GmNFR5α/β in soybean (Fig 1.4) (Limpens et al. 2003; Madsen et al.

2003; Radutoiu et al. 2003; Arrighi et al. 2006; Indrasumunar et al. 2010). These receptor- like kinases with a N-acetyl-glucosamine-binding motifs (LysM-type) are composed of a extracellular LysM domain, a transmembrane domain, as well as an intracellular kinase domain (Ferguson et al. 2010). It was shown that Nod factor specificity is determined by these receptor-like kinases (Radutoiu et al. 2007). Recently, a new aspect of Nod factor perception was discovered. Op den Camp and colleagues showed that in the non-legume plant Parasponia a single Nod factor-like receptor is indispensable for symbiotic interactions not only with rhizobia but also for endomycorrhizal symbiosis (Op den Camp et al. 2011).

This result is very interesting since Nod factor receptors were thought to be nodulation specific and not essential for mycorrhization. Consequently, it was concluded that the mycorrhizal fungi-plant symbiosis is controlled by other receptors which are specific for mycorrhizal signals such as Myc factors. Since the mycorrhiza symbiosis is considered to be the more ancient symbiosis, the authors concluded that during evolution, Parasponia has recruited a Myc factor receptor to serve as Nod factor receptor in the rhizobial plant symbiosis. This finding opens up new possibilities to investigate whether the ability to perform symbiotic nitrogen fixation can be transferred to other non-legume plants.

Introduction

Plant responses to Nod factors

Nod factors are powerful signals and induce several complex downstream signal transduction cascades in legumes upon binding to the corresponding receptor. Our current knowledge about the Nod factor signal transduction pathways is derived mainly from studies with the model legumes L. japonicus and M. truncatula. Using genetic methods in these organisms, a detailed genetic dissection of the events during the nodulation process was uncovered. Thus, a set of genes was identified to be necessary for Nod factor-induced signaling events in the leguminous host. These genes encode a receptor-like kinase with extracellular leucine-rich repeats (LLR), two putative cation channels, and two components of the nucleoporin (Fig

1.4) (Oldroyd and Downie 2008). The LLR-receptor-kinase was studied in further depths in

M. truncatula where the DMI2 (does not make infections 2) mutant was shown to be important for entrapping the rhizobia in the root hair (Fig 1.4) (Esseling et al. 2004). One of the earliest responses of Nod factors induced in legume root hairs are changes in the intracellular calcium concentration. The M. truncatula gene DMI1 was shown to encode a ligand-gated ion channel that is homologous to prokaryotic cation channels (Fig 1.4) (Riely et al. 2007). Mutant analysis revealed that DMI1 is essential for mediating calcium spiking

(Riely et al. 2007). Interestingly, Nod-factor-induced calcium spiking is mostly restricted to the nuclear region. Thus, it was proposed that the nuclear envelope and the nuclear-associated endoplasmatic reticulum function as internal calcium stores (Oldroyd and Downie 2008). The necessity of nucleoporins for Nod factor signaling was surprising (Kanamori et al. 2006). The precise role of these nucleoporins is still unknown. The transport of a second messenger or the movement of symbiosis-specific proteins into or out of the nucleus (Kanamori et al. 2006;

Oldroyd and Downie 2008) have been proposed. It has been shown that calcium spiking is the essential component of Nod factor-mediated signal transduction. The observation that

Nod factor perception leads to the activation of several receptor kinases (LysM and LLR)

Introduction suggested that Nod factor-induced signal transduction involves phosphorylation cascades at the plasma membrane (Yoshida and Parniske 2005). Downstream of the calcium signal at least four genes were described to translate the calcium signal. DMI3 encodes the calcium- and calmodulin-depended protein kinase CCaMK (Fig 1.4). This remarkable protein possesses the ability to bind calcium either directly by so-called EF-hand domains (a helix- loop-helix structural domain which is found among calcium-binding proteins) or indirectly by binding calcium in a complex with the calcium-binding protein calmodulin (Levy et al.

2004). Thus it was speculated that these mechanisms allow the protein to decode the signal from the calcium spiking. Downstream of CCaMK several transcription factors are involved in a Nod factor induced gene expression (Kalo et al. 2005; Smit et al. 2005). These proteins are most likely involved in the regulation of gene expression since in some cases it was shown that they bind to the promoters of early nodulin genes whose expression is limited to epidermal root cells and promote the initiation of the outgrowth of root hairs cells

(Andriankaja et al. 2007).

Formation of a root nodule

In the Nod factor-dependent nodulation process, the perception of Nod factors leads to a rhizobial colonization of legume roots via infection threads which results in the development of root nodules. While the infection process is restricted to epidermal cells, nodule formation is controlled in the root cortex to form a root nodule meristem (Fig 1.4). Rhizobia migrate through the infection thread and are endocytosed by the plant plasma membrane forming a so-called peribacteroid membrane (PBM) surrounding the invading bacterium. The resulting complex that consists of a PBM and the enclosed bacteria is called a symbiosome.

Fig 1.4. Molecular basis of rhizobia-legume symbiosis. Legume plants secrete flavonoids such as genistein into the rhizosphere. These signaling molecules are perceived by rhizobial NodD regulatory proteins. Upon binding the appropriate flavonoid, these transcription factors lead to the activation of transcription of various nod genes which induce the synthesis of rhizobial Nod factors (NF). Nod factors (NFs) are now perceived by Nod factor receptors (NFR) at the legume root. DMI2 is required for root hair curling. DMI1 encodes an ion channel that is localized to the nuclear membrane. DMI3 encodes a calcium-calmodulin-depenndent protein kinase which is required for the induction of transcriptional changes due to Nod factor binding as well as cortical cell division. Adapted from Oldroyd and Downie (2008) and Masson-Boivin et al. (2009). Introduction

In order to create an environment suitable for nitrogen fixation the plant induces the synthesis of leghemoglobin which functions to keep free oxygen levels down (Sprent 2001). In fact, low oxygen concentrations were shown to trigger the gene expression of several nitrogen

ﬁxation genes (Fischer 1994). Two different types of root nodules are defined in regard to the nodule meristem. Indeterminate nodules (e.g. Vicia) possess an active apical meristem which continues to grow over time. In contrast, cell division of the meristem of determinate nodule

(e.g. Glycine max) is time limited and growth is maintained due to cell expansion. Several genes were shown to be essential for nodule organogenesis such as several nodulins (Charon et al. 1999).

Introduction

1.3 Global strategies to monitor rhizobial gene and protein expression in symbiosis

The availability of several rhizobial genome sequences opened new perspectives for research on a global basis allowing researchers to enlarge the knowledge about plant-microbe interactions. Powerful techniques such as transcriptomics and proteomics were developed and used to look at the cell more globally.

So far, large scale gene expression analyses of symbiotically grown rhizobia were performed in several species such as S. meliloti, M. loti, B. japonicum and most recently in Rhizobium leguminosarum (Ampe et al. 2003; Berges et al. 2003; Barnett et al. 2004; Becker et al. 2004;

Uchiumi et al. 2004; Pessi et al. 2007; Karunakaran et al. 2009). One challenge linked to transcriptional analysis of endosymbiotic bacteria is the fact that rhizobial cells are encompassed by plant host cells. The different research groups approached this problem from various sides. The transcriptome of bacteroids from S. meliloti, M. loti and R. leguminosarum was determined by isolating bacteroids from root nodules followed by bacteroid RNA isolation (Ampe et al. 2003; Berges et al. 2003; Becker et al. 2004; Uchiumi et al. 2004).

Thus, it was ensured that little or no plant-derived material was further processed and potentially caused cross-hybridization on the rhizobial microarrays. This preparation technique however is time consuming and therefore risks that a large proportion of the bacterial messenger RNA is degraded. The degradation of RNA was shown to be in the range of seconds to minutes (Bernstein et al. 2002).

In our laboratory, total RNA was directly isolated from soybean root nodules infected with B. japonicum thus minimizing the risk of RNA degradation (Pessi et al. 2007). This extraction procedure is much faster but leads, however, to a mixture of bacterial and plant RNA.

Importantly, all the probesets present on the Affymetrix custom made B. japonicum genechip used in this case had been pruned against the soybean database. Moreover, for data

Introduction evaluation, stringent filtering criteria were applied. Thus, only genes were considered which were expressed in at least two out of three biological replicates and at the same time showed a twofold higher expression level than in a control experiment where uninfected root material was hybridized on the microarray (Pessi et al. 2007).

Barnett and coworkers used a so called Symbiosis Chip to study the interaction of both symbiotic partners at the same time (Barnett et al. 2004). Also in this case total RNA was extracted from nodules since the gene expression profile of both symbiotic partners was monitored. This dual microarray contains besides the complete genome of S. meliloti also

10,000 probe sets of its host plant M. truncatula (Barnett et al. 2004). When comparing the transcription profile of root nodules induced by S. meliloti wild type to that of Fix¯ nodules induced by a fixJ mutant (Oke and Long 1999a), the researchers showed that many changes in gene expression of M. truncatula were not caused by nitrogen starvation, as expected, but were derived more from nodule morphogenesis and bacterial occupancy (Barnett et al. 2004).

Transcriptional profiling of B. japonicum bacteroids isolated from soybean root nodules revealed that 2,780 genes are expressed which corresponds to 34% of the genome (Pessi et al.

2007). Among those, 411 genes are exclusively expressed in bacteroids suggesting an important function during symbiosis (Pessi et al. 2007). Table 1.1 compiles a list of the 20 most highly induced genes of B. japonicum bacteroids when the expression profile is compared to that of aerobically grown free-living cells. Among those, a carbonic anhydrase gene was identified which was studied in further depth in chapter 2. Transcription profiling of endosymbiotic bacteria confirmed previous findings such as the increased expression of various fix and nif genes which are known to be important for symbiotic nitrogen fixation.

However new insights were also gained which provided the opportunity to discover genes not yet known to be important for symbiosis.

Introduction

Table 1.1. List of the 20 most highly induced genes in wild-type bacteroids (bact.) when aerobically (aer.) grown wild-type cells were used as reference (Pessi et al. 2007).

Gene no Description Gene name FC bact. vs. aer.

blr2106 L-ectoine synthase ectC 1201 blr2143 similar to cytochrome P450-family protein - 1008 bsr2010 - - 359 blr8234 - - 317 blr2131 probable oxygenase - 298 blr2146 dehydrogenase - 243 bll1767 - - 242 bsl1637 - - 216 bll2047 - - 209 bll1858 - - 208 blr2011 - - 204 blr3719 - - 182 blr1850 - - 178 blr2071 similar to inosamine-phosphate amidinotransferase - 171 bll1777 alkyl hydroperoxide reductase ahpC 170 bll2065 carbonic anhydrase icfA 156 blr1638 - - 155 blr2132 - - 148 bll2004 - - 147 blr1964 putative sugar hydrolase - 145

When comparing the transcriptomic data obtained from endosymbiotically grown B. japonicum to that of S. meliloti (Barnett et al. 2004; Becker et al. 2004) and M. loti (Uchiumi et al. 2004) a considerable overlap was found (Pessi et al. 2007). Notably, the most pronounced concordance was observed with the study of Barnett and coworkers who used a similar type of Affymetrix array and preparation of RNA (Barnett et al. 2004; Pessi et al.

2007). Moreover, out of the 2,780 B. japonicum genes and those differentially expressed in symbiosis, 44% have orthologous genes in S. meliloti and 45% in M. loti (Pessi et al. 2007).

Differences in the outcome of the performed studies could be due to the different nodule types used in these studies which were determinate nodules in the case of B. japonicum and indeterminate nodules in the case of S. meliloti (Pessi et al. 2007).

Introduction

Even though microarray technology is a powerful tool to study global gene expression, several limitations are known. One was already mentioned above and is directly linked to the sample preparation namely the short half-life of RNA (Bernstein et al. 2002). Thus, the extracted RNA pool could represent only more stable mRNA molecules. Additionally, these analyses do not consider post-transcriptional regulatory events which may affect mRNA and protein abundance.

Proteomics provides the chance to measure the abundance of proteins in an organism, thus enabling us to understand biological systems in further depth. So far, proteomic approaches of symbiotically grown rhizobia were carried out for B. japonicum, S. meliloti and R. leguminosarum (Panter et al. 2000; Morris and Djordjevic 2001; Djordjevic et al. 2003;

Sarma and Emerich 2005; Delmotte et al. 2010). In these analyses various aspects of the rhizobium-legume symbiosis were examined. Hereby, several studies investigated the proteome of bacteroids isolated from root nodules (Djordjevic et al. 2003; Sarma and

Emerich 2005; Delmotte et al. 2010), and attention was also drawn to cultivar-specific responses of R. leguminosarum (Morris and Djordjevic 2001). Moreover, Panter and colleagues enriched proteins of subcellular fractions of the PBM of soybean nodules induced by B. japonicum to investigate its proteome and further understand PBM biogenesis and function (Panter et al. 2000).

Two-dimensional (2D) gel electrophoretic proteome analysis was used in the studies performed in B. japonicum (Sarma and Emerich 2005), S. meliloti and R. leguminosarum

(Morris and Djordjevic 2001; Djordjevic et al. 2003). In this technique proteins are separated on 2D gels according to their isoelectric point and molecular mass. This method is not ideal for proteins which are extremely basic or acidic as well as for very large or very small proteins (Delmotte et al. 2010).

Introduction

Recently, a sensitive geLC-MS/MS proteomics approach was carried out with B. japonicum bacteroids isolated from soybean root nodules (Delmotte et al. 2010). Previously, it was shown that this method allows a comprehensive proteome coverage due to its sensitivity and detection of low-abundance and very small proteins (de Godoy et al. 2006). Thus, two independent studies investigated the proteome of symbiotically grown B. japonicum bacteroids using two different proteomic techniques (Sarma and Emerich 2005; Delmotte et al. 2010). Sarma and coworkers identified 149 proteins in B. japonicum bacteroids whereas

Delmotte and colleagues identified 2,173 which correspond to 27.8% of the theoretical proteome (Sarma and Emerich 2005; Delmotte et al. 2010). Notably, 95% of the proteins identified by Sarma were also detected using the shotgun proteomics approach. Among the nine proteins only detected by the 2-D gel electrophoresis approach several hypothetical proteins, a transcriptional regulator, a putative transglycosylase and a transaminase were identified (Delmotte et al. 2010). Interestingly, there was no indication that these proteins are small or basic proteins, which are known to be typically underrepresented in proteomics studies. The authors conclude, that these proteins are either not abundant or not expressed under symbiotic growth conditions employed in this study (Delmotte et al. 2010). The outcome of both studies reflects the enormous advances in the field of proteomic analysis and underlines its enormous power.

Since both techniques, transcriptomics as well as proteomics, have different limitations to comprehensively understand cellular activity of the bacterial endosymbiont, gene expression data should ideally be correlated with protein expression data. This was successfully done in

B. japonicum (Delmotte et al. 2010).

Fig 1.5. Functional classification in 15 categories according to the Kazusa annotation (www.kazusa.or.jp/rhizobase/). Identified genes/proteins were grouped in three data sets, (i) all B. japonicum protein-coding genes (8,317, illustrated in grey), (ii) proteins or protein-coding genes expressed in soybean bacteroids (3,540, illustrated in red) and (iii) proteins or protein coding genes not detected as expressed in our study (4,777, illustrated in blue). The relative frequencies of (ii) and (iii) were calculated by dividing the number of genes/proteins present in each category by the total number of genes/proteins identified in that data set. Asterisks indicate statistical significance for over- represented or under-represented categories which was computed using the Fisher’s test with multiple testing correction (p-valuer0.01). Taken from Delmotte et al. (2010). Introduction

Combining both approaches, an integrated proteomics and transcriptomics reference data set was created which comprises 3,587 genes/proteins (43% of the predicted genome/proteome) and enabled the researcher to expand our current knowledge about bacteroid metabolism

(Delmotte et al. 2010). Using functional classification analysis, based on 15 major functional categories (http://bacteria.kazusa.or.jp/rhizobase), it was possible to reveal over-represented as well as under-represented categories. Several different datasets were examined (i) all protein-coding genes of B. japonicum (8,317), (ii) the genes/proteins expressed in the soybean bacteroid reference dataset (3,587), and (iii) the genes/proteins not detected in these studies (4,678). As indicated in Fig 1.5, several categories were found to be over-represented as well as under-represented with statistical significance. Proteins annotated in the “central intermediary metabolism” and “energy metabolism” categories were over-represented in B. japonicum bacteroids as well as genes/proteins belonging to the “other categories” and

“hypothetical proteins” (Fig 1.5) (Delmotte et al. 2010).

Introduction

1.4 Host speciﬁcity

At the beginning of the 19th century, scientists made the observation that rhizobia vary greatly in their ability to enter symbiosis with legume plants. Later on the term host specificity was introduced to describe the early host-specific recognition and nodulation processes between rhizobia and their corresponding plant host. The molecular basis of host specificity was studied, and key players were identified which include NodD proteins, Nod factors, as well as polysaccharides and secreted proteins.

NodD proteins

Several genetic studies revealed that nodD genes are determinants of host specificity since they specifically recognize legume root-secreted flavonoids. Point mutations in nodD expand the number of recognized flavonoid inducer molecules in R. leguminosarum bv. trifolii and thereby extend the bacterial host range (McIver et al. 1989). Moreover, transferring the nodD1 gene of the broad-host-range Rhizobium sp. strain NGR234 into narrow-host-range rhizobia leads to an extension of the nodulation capacity of the recipients to new host plants

(Dénarié et al. 1992). It was shown that NodD proteins of narrow-host-range rhizobia respond to fewer flavonoids than NodD proteins of broad-host-range rhizobia which possess a larger spectrum of inducing molecules (Dénarié et al. 1992). Besides NodD proteins other species-specific sensor-activator systems were shown to contribute to rhizobial host range.

The two-component regulatory system NodVW of B. japonicum is important for nodulation of mungbean, cowpea, and siratro but only marginally contributes to elicit symbiosis with soybean (Göttfert et al. 1990). It could be shown that genistein induced the activation of

NodW which is required for nod gene expression (Loh et al. 1997).

Introduction

Other nod genes as host-specific determinants

Besides nodD, other bacterium-speciﬁc nod genes, such as nodEFZ confer host plant speciﬁcity by chemically modifying the oligosaccharide backbone of Nod factors (Fig 1.3).

The products of the nodEF genes modify the fatty acids which are N-linked to the Nod factor molecule (Fig 1.3). Mutations in nodEF results in the ability to nodulate new host plants. This was described in R. trifolii where nodEF − mutants acquired the ability to infect and nodulate peas (Djordjevic et al. 1985). Bradyrhizobium japonicum nodZ codes for a fucosyltransferase which adds fucose to Nod factors at position R4 (Fig 1.3). This modification was shown to be nodulation specific for siratro plants (Stacey et al. 1994). Several other studies described the importance of Nod factors in regard to host specificity and are reviewed by Dénarié et al.

(1992) as well as Perret et al. (2000).

Polysaccharides and secreted proteins as host-specific determinants

In addition to nod genes and Nod factors, other factors contribute to host speciﬁcity. The rhizobial cell wall is composed of various extracellular polysaccharides which are also termed exopolysaccharides (EPS), lipopolysaccharides (LPS), capsular polysaccharides

(CPS) as well as cyclic β-glucans (Perret et al. 2000). These polysaccharides accumulate at the surface of the rhizobial endosymbiont. Several studies showed that EPS are determinants of host specificity. In M. loti EPS deletion mutants were fully effective when inoculated onto

Leucaena pedunculatus (determinate root nodules), however, were ineffective when inoculated onto Leucaena leucocephala (indeterminate root nodules) (Hotter and Scott 1991).

Thus it was suggested that EPS in the symbiosis described above are important to initiate a fully functional symbiosis with indeterminate rather than determinate nodules.

Several host-specific and symbiosis-relevant rhizobial proteins were shown to be secreted into the legume host. An example is NodO which was shown to act in a host-specific manner

Introduction with P. vulgaris but not L. leucocephala (Perret et al. 2000). NodO is secreted by a type I secretion system (T1SS) PrsDE encoded by prsD and prsE (Finnie et al. 1997). Several possible functions were assigned to NodO, such as the uptake of Nod factors by the host or the amplification of the Nod factor signal (Fauvart and Michiels 2008). Besides T1SS, other secretion systems such as T3SS, T4SS and most recently T6SS were described, and it was shown that T3SS and T4SS act in a host-specific manner (Fauvart and Michiels 2008).

Cultivar specificity

Rhizobial host-range is generally displayed at the level of plant genera. There are, however, examples of host-specific interactions that can occur at the strain-cultivar level (also called cultivar specificity). The nolA gene of B. japonicum, for instance, is essential to nodulate different soybean cultivars (Sadowsky et al. 1991). So far, several other examples of rhizobial cultivar specific nodulation events were described predominantly for S. fredii (Davis et al.

1988; Meinhardt et al. 1993; Bellato et al. 1997; Jiang and Krishnan 2000; Lorio et al. 2004).

An example of cultivar specificity in a legume plant was recently described for the soybean-

B. japonicum and S. fredii symbiosis. The soybean genes Rj2 and Rfg1 were identified to restrict nodulation with these rhizobial strains (Trese 1995; Yang et al. 2010). Rj2 restricts nodulation with B. japonicum strain USDA122 whereas Rfg1 prevents nodulation with certain fast-growing rhizobial strains such as S. fredii USDA257 (Yang et al. 2010). Rj2 and

Rfg1 encode a Toll-interleukin receptor/nucleotide-binding site/leucine-rich repeat (TIR-

NBS-LRR) class of plant resistance proteins. The identification of plant resistance (R) genes involved in the control of genotype-specific infection and nodulation leads to the conclusion that a common recognition mechanism between symbiotic and pathogenic host-bacteria interactions exists (Yang et al. 2010).

Introduction

Host-speciﬁc adaptation

In contrast to host speciﬁcity, host-speciﬁc adaptation relies on physiological adaptation of the rhizobial endosymbiont according to the host environment. So far knowledge regarding bacterial adaptation to hosts is limited, and advances were mostly made by serendipity.

Several examples suggest that certain rhizobial metabolic features are crucial for adaptation and survival in root nodules of host plants. The pckA gene of Rhizobium sp. NGR234 which encodes the anaplerotic enzyme phosphoenolpyruvate carboxykinase was reported to display a host-dependent phenotype (Osteras et al. 1991). Thus it was proposed that the availability of different carbon sources which are provided by the host plant affect nodule development and nitrogen fixation. Moreover, a host-dependent expression of R. leguminosarum bv. viciae hydrogenase activity was described, and the authors proposed a host-derived substance to cause the host-depended variations (Brito et al. 2008). The B. japonicum gene hsfA (host- specific fixation) is essential for nitrogen fixation on cowpea, but not required for nitrogen fixation on soybean or siratro (Chun et al. 1994). Later on it was shown that HsfA participates in the signal exchange to mediate bacteroid development in cowpea nodules (Oh et al. 2001). Recently in B. japonicum the resistance/nodulation/cell division (RND) multidrug efflux system BdeAB was shown to be more important for the B. japonicum- soybean symbiosis than for the symbiosis with siratro, cowpea or mungbean (Lindemann et al. 2010). Thus the authors speculate that BdeAB specifically exports a toxic substance which is more abundant in root nodules of soybean than in those of the other host plants.

With the aim to study rhizobial adaptation to hosts with a more global perspective, two studies were carried out to assess the transcriptome of endosymbiotically grown rhizobia grown in different host environments (Karunakaran et al. 2009; Koch et al. 2010). The transcription profiles of R. leguminosarum bv. viciae bacteroids isolated from root nodules of pea and vetch were assessed, however, revealed that the rhizobial gene expression does not

Introduction vary according to the host environment (Karunakaran et al. 2009). An integrated transcriptomics–proteomics approach was carried out for B. japonicum during symbiosis with different host plants and is described in further depth in Chapter 3.

Introduction

1.5 Carbon metabolism in rhizobia

Rhizobia are able to utilize and metabolize a broad spectrum of diﬀerent carbon and nitrogen sources. The metabolic diversity of rhizobia was suggested to be reflected in their large and complex genomes (Prell and Poole 2006). Among all rhizobial strains sequenced so far, B. japonicum possesses the largest genome encompassing 9.1 Mbp (Kaneko et al. 2002).

Notably, based on genome annotation, this organism harbors more than 549 different transport systems (http://www.membranetransport.org/). Among those, 45% are predicted

ATP binding cassette (ABC) transport systems. Thus it was speculated that rhizobia are enabled to access a great diversity of nutrients from the soil or the rhizosphere. According to their metabolic capacity, rhizobia are divided into two broad groups: fast growers and slow growers (Stowers 1985). Fast growers such as S. meliloti, R. leguminosarum have a generation time of less than six hours whereas the generation time of slow growers such as B. japonicum exceeds six hours (Stowers 1985).

Carbon metabolism in free living rhizobia

Free living rhizobia in the rhizosphere encounter a great diversity of carbon sources derived from root exudates. The characterization of pea mucilage and root exudates detected a broad range of different sugars such as galactose, arabinose and others (Knee et al. 2001). As none of these compounds is dominant it was suggested that an oligotrophic lifestyle might represent the most successful strategy in terms of competition for energy resources (Prell and

Poole 2006). The bacteria assimilate most of these compounds via the Entner-Doudoroﬀ pathway, pentose phosphate pathway and TCA cycle which constitute the central carbon metabolism (Fuhrer et al. 2005).

Introduction

Fig 1.6. Proposed pathway for L-arabinose degradation in fast- and slow growing rhizobia. Adapted from Watanabe et al. (2006a).

One of the best studied carbon sources for B. japonicum is L-arabinose (Pedrosa and Zancan

1974; Stowers 1985). Notably, the degradation pathway differs between fast and slow growers and was previously used to distinguish both groups (Fig 1.6). In fast growers like S. meliloti L-arabinose is metabolized to α-ketoglutaric semialdehyde which is then dehydrogenated to α- ketoglutarate (Fig 1.6), whereas in slow-growing rhizobia like B. japonicum L-arabinose is converted to pyruvate and glycolaldehyde (Fig 1.6). Genes in the uptake and metabolism of L-arabinose were previous studied in S. meliloti (Poysti et al.

2007). The identified genes are organized in the araABCDEF operon which is located on the megaplasmid pSymB. The araABC gene products are responsible for the uptake of arabinose and araDEF encode enzymes for the catabolism (Poysti et al. 2007). Orthologs of the

Introduction araABC were identified in B. japonicum based on amino acid sequence identity: Blr3208 shares 71% sequence identity to AraA, Blr3209 65% to AraB, and Blr3210 59% to AraC.

Genes coding for enzymes of the L-arabinose catabolism such as L-arabinose-1 dehydrogenase and the L-arabinolactonase were not identified in S. meliloti. However, the genes encoding these enzymes were identified in the nitrogen fixing α-proteobacterium A. brasiliense (Watanabe et al. 2006a; Watanabe et al. 2006b). Azospirillum brasiliense

BAD95974 and BEA94275 encode the L-arabinose-1-dehydrogenase and the L- arabinolactonase, respectively. Amino acid sequence analysis revealed that BAD95974 shares 55% identity to the B. japonicum gene product Blr3205, and BEA94275 33% to

Blr3207. Interestingly, blr3205 and blr3207 as well as genes for the arabinose uptake are located in close vicinity on the B. japonicum genome. All these genes which are potentially involved in the uptake and catabolism of arabinose were found to be expressed in minimal medium supplemented with L-arabinose suggesting a possible involvement in the degradation of this carbon source in B. japonicum (G. Pessi unpublished data). Besides L-arabinose, free- living rhizobia are able to metabolize various other carbon compounds (reviewed by Stowers

1985).

Carbon metabolism in symbiotic rhizobia

The differentiation of a free-living rhizobial cell into a nitrogen fixing bacteroid involves dramatic changes in metabolism especially in regard to energy metabolism (Oke and Long

1999b). During symbiosis bacteroids strongly rely on carbon sources generated by the host via photosynthesis. C4-dicarboxylic acids like malate and succinate were shown to be the primary carbon sources of bacteroids which are directly fed into the TCA-cycle to supply the bacteria with enough energy to perform nitrogen fixation (Prell and Poole 2006).

Introduction

The transport of C4-dicarboxylates was shown in R. leguminosarum to be facilitated by the

Dct system which is composed in total of three genes namely dctA, dctB and dctD (Ronson et al. 1981; Ronson et al. 1984). DctA was shown to be the structural component mediating the

C4-dicarboxlic transport (Ronson et al. 1984). The dctB and dctD genes code for a two- component regulatory system which responds in the presence of C4-dicarboxylates (Ronson et al. 1987). It was proposed that upon binding of the signal (e.g. succinate) to the periplasmic sensor domain of DctB, the signal is transmitted across the membrane initiating phosphorylation and phosphoryl transfer to DctD (Yurgel and Kahn 2004). Mutations in genes responsible for the transport of C4-dicarboxylic acids were shown to be essential for nitrogen fixation in R. leguminosarum (Ronson et al. 1981; Ronson et al. 1984; Yurgel and

Kahn 2004).

Endosymbiotic rhizobia are able to store carbon in the form of poly-β-hydroxybutyrate

(PHB). This polymer can represent up to 70% of the dry weight of bacteroids from determinate root nodules such as soybean or chickpea (Kim and Copeland 1996). The function of PHB during symbiosis may differ between determinate and indeterminate nodules. Mature bacteroids of determinate nodules accumulate large amounts of this storage compound whereas bacteroids of indeterminate nodules break down PHB granules during maturation (Prell and Poole 2006).

Recently, using global approaches new insights into the energy metabolism of bacteroids from B. japonicum were gained (Delmotte et al. 2010). Integrating gene and protein expression data of endosymbiotically grown B. japonicum, the authors showed for the first time that all the enzymes which are involved in catabolism of C4-dicarboxylates were present

(Fig 1.7) (Delmotte et al. 2010). Several important anaplerotic enzymes which sustain the

TCA cycle as well as the gluconeogenesis were identified (Fig 1.7). Moreover all enzymes of

Introduction the pentose phosphate pathway were detected which points towards the utilization of C5 sugars for the biosynthesis of nucleotides (Fig 1.7).

Fig 1.7. Proposed energy pathways in bacteroids of B. japonicum. The Embden-Meyerhof-Parnas (EMP), the Entner-Doudoroff (ED) pathway and the pentose phosphate (PP) cycle as well as the TCA cycle are illustrated. The following symbols were used to indicate genes/proteins that have been detected at the level transcriptomics or proteomics. A white circle represents enzymes that have only been detected by proteomics. Black and white circle represent genes/proteins that have been detected in both approaches. Unlabeled arrows indicate enzymes that have not been detected in these studies. Taken from Delmotte et al. (2010).

Introduction

To sum up, most of our current knowledge about carbon metabolism in bacteroids derived from studies based on C4 and C5 compounds. In contrast only limited knowledge is available about the role of C2-dicarboxylic acid like oxalate during symbiosis.

Oxalotrophic bacteria

Oxalotrophic bacteria are widely dispersed within the bacterial phylogeny and thus do not constitute a homogeneous taxonomic group. Among oxalotrophs some species are considered as "generalists", metabolizing oxalate among many other carbon compounds, other species are considered as "specialists", using oxalate as the sole carbon and energy source (Sahin

2003). Oxalic acid or its salts are widely found in nature and were reported to occur in plants, animals, humans and soil (Sahin 2003). In plants, oxalic acid accumulates with calcium to form highly insoluble Ca-oxalate crystals (Franceschi and Nakata 2005). It was reported that oxalate producing plants can accumulate oxalate in a range of 80% (w/w) of their dry weight (Franceschi and Nakata 2005). Oxalate can also be found in the gut of humans and animals due to ingesting oxalate-containing plant material such as rhubarb and spinach (Abratt and Reid 2010). Massive consumption of oxalate can cause severe health problems such as the formation of kidney stones. People predisposed for kidney stone formation must be obviating food with particularly high quantities of oxalate. Oxalate can as well be found in high concentrations in soil deriving from the decomposition of plant material

(Sahin 2003).

Until now two metabolic routes of oxalate assimilation were described among oxalotrophic bacteria (Fig 1.8) (Sahin 2003). First, oxalate has to be imported and activated to oxalyl-CoA and then reduced to glyoxylate. Glyoxylate can now either be metabolized to tartronate semialdehyde (Fig 1.8) or hydroxypyruvate, using the enzymes glyoxylate carboligase and serine-glyoxylate aminotransferase, respectively. In both cases glycerate is subsequently

Introduction formed which is phosphorylated to 3-P-glycerate. 3-P-glycerate can further be processed and either enters the TCA cycle via pyruvate, or is used in the gluconeogenesis to build up biomass (Fig 1.8).

Fig 1.8. Proposed pathways of the oxalate assimilation. For more information see text. Adapted from Sahin (2003).

In all oxalotrophic bacteria the catabolism of oxalate depends on two enzymes namely the formyl-CoA transferase as well as the oxalyl-CoA decarboxylase encoded by frc and oxc, respectively. In a coupled reaction of the two cytosolic enzymes oxalate is converted to formate and CO2.

Introduction

The Frc catalyzes the transfer of CoA froom formate to oxalate by activating the oxalyl moiety for a thiamine-dependent decarboxylation in a reaction that is mediated by the oxalyl-CoA decarboxylase (Fig 1.9A/B.).

Fig 1.9. Oxalate catabolism of anaerobic (A) and aerobic (B) oxalotrophic bacteria. For more information see text. Adapted from Anantharam et al. (1989) and O. Braissant PhD thesis (2005).

Introduction

Due to the omnipresence of frc among oxalotrophic bacteria, the gene was used as a molecular marker to detect oxalate-oxidizing bacteria in complex environmental samples

(Khammar et al. 2009). The transport of oxalate and formate into and out of the cell is mediated via the oxalate:formate antiporter encoded by oxlT (Fig 1.9A/B.). OxlT was first discovered in O. formigenes (Anantharam et al. 1989). This carrier was shown to mediate the exchange of oxalate:formate without the requirement of energy (Ruan et al. 1992). In O. formigenes OxlT comprises around 10% of the membrane protein (Stewart et al. 2004).

Importantly, the mechanism of oxalate catabolism varies between aerobic (Fig 1.9B) and anaerobic (Fig 1.9A) oxalotrophic bacteria. Most of our current knowledge is derived from the anaerobic bacterium O. formigenes (Fig 1.9A). The decarboxylation of oxalate consumes a proton (Fig 1.9). It was shown that anaerobic oxalotrophs link this reaction to an oxalate2−:formate1− exchange over the membrane mediated by OxlT (Fig 1.9A). Thereby a pH gradient (ΔpH) as well as a membrane potential (Δψ) are generated. Together, they use the proton motive force to support ATP synthesis via the action of a putative F0F1-type ATP synthase (Fig 1.9A) (Stewart et al. 2004). In aerobic oxalotrophic bacteria (Fig 1.9B) such as

Ralstonia eutropha the degradation of oxalate leads to the formation of CO2 from formate by the action of the formate dehydrogenase (FDH) (Fig 1.9B). Thereby, one reduction equivalence is produced which allows energy production in the respiratory chain (Fig 1.9B).

A. Oxidation of oxalate

̶ H2C2O4 + O2 → 2 CO2 + 2 H2O : ∆G°ˈ= ̶ 328 KJ/mol equals 328 KJ/e pair

B. Oxidation of glucose

̶ C6H12O6 + 6 O2 → 6 CO2 + 6 H2O : ∆G°ˈ= ̶ 2873 KJ/mol equals 239 KJ/e pair

Fig 1.10. Reaction of the oxidation of oxalate (A) and the oxidation of glucose (B). Adapted from Olivier Braissant PhD thesis (2005).

Introduction

The catabolism of oxalate provides the bacteria with energy for growth. From a thermodynamical perspective the oxidation of oxalate does not provide as much energy as compared to glucose per mole of substrate (Fig 1.10). Oxalate is however a good electron donor providing 328 KJ per electron pair whereas glucose provides only 239 KJ per electron pair. In this thesis we showed that B. japonicum is an oxalotrophic bacterium and analyzed key enzymes for this life style (Chapter 4).

Introduction

1.6 Aim of this work

The aim of this work was to gain insights into the metabolism of B. japonicum bacteroids in legume root nodules. In Chapter 2 two carbonic anhydrase genes, which were shown to be highly expressed in B. japonicum during symbiosis with soybean were characterized. In

Chapter 3 the transcriptome as well as the proteome of B. japonicum bacteroids isolated from soybean, cowpea and siratro root nodules was analyzed in order to systematically monitor differential gene expression of B. japonicum according to the host environment. Hereby an

ABC type transporter system associated with a monooxygenase was identified and shown to be more important for the B. japonicum-siratro symbiosis than for the other legumes. In

Chapter 4 we showed that B. japonicum is an oxalotrophic bacterium and identified genes that are essential for this lifestyle.

Characterization of two carbonic anhydrase genes

CHAPTER 2

Characterization of two carbonic anhydrase genes

41 Characterization of two carbonic anhydrase genes

2.1 Abstract

A previous global transcriptome analysis identified Bradyrhizobium japonicum genes specifically up-regulated during symbiosis with soybean when compared to free-living aerobically grown cells. A carbonic anhydrase gene (bll2065) was identified amongst the 20 most highly induced genes. Carbonic anhydrases can be found in organisms of all kingdoms of life; they catalyze the reversible hydration of CO2 to carbonate. Based on genome sequence, B. japonicum possesses five carbonic anhydrase genes (blr0500, bll1137, bll2065, bll4863, bll4865) but only the expression of bll2065 was up-regulated during symbiosis. In addition, global protein profiling of B. japonicum bacteroids from soybean root nodules revealed that besides Bll2065, another carbonic anhydrase, Bll4865, is detected during symbiosis. To further understand the function of carbonic anhydrase during symbiosis, two mutant strains were constructed ∆bll2065-2066 (6224/6225) and bll4865::pRJ6226 (6226).

Phenotypic characterization of these mutant strains showed that each of the two genes by itself is not essential for B. japonicum during symbiosis with soybean.

42 Characterization of two carbonic anhydrase genes

2.2 Introduction

Carbonic anhydrases (CA; EC 4.2.1.1) are zinc-containing enzymes catalyzing the reversible hydration of CO2 to carbonate (Fig 2.1) (Smith and Ferry 2000). This metalloenzyme has long been subject of investigation not only because CAs can be found in organisms of all kingdoms of life but also because the efficient catalysis of this process is important in many biological processes such as photosynthesis, respiration, and pH regulation (Badger and Price

1992; Scott et al. 1998). Based on amino acid sequence comparisons, CAs exist in four evolutionary distinct classes, namely the α-, β-, γ-, and δ-class that share no structural identity to one another underlining their independent evolution (Smith and Ferry 2000; Tripp et al.

2001; Park et al. 2007). Alpha-class CA were previously identified in mammals, vertebrates, green algae, protozoa, and bacteria whereas β-class CAs were found in bacteria, archaea, fungi, algae, mono- and dicotyledons (Smith and Ferry 2000; Tripp et al. 2001). While γ- class CAs were identified in archaea and eubacteria, recently a new CA class (δ-class) was discovered in the group of algae called diatoms (Tripp et al. 2001; Park et al. 2007). Despite the major structural differences between the so far characterized CAs, the enzyme-catalyzed reaction is the same and occurs in two mechanistically distinct reactions. In the first reaction, a nucleophilic attack of a zinc-bound hydroxide ion on CO2 takes place (Fig 2.1) followed by the second step leading to the regeneration of the active form of the enzyme by ionization of the zinc-bound H2O molecule and the removal of a proton from the active site (Fig 2.1) (Smith and

Ferry 2000).

+ CO2 + H2O ↔ HCO3- + H

2+ 2+ (1) Zn ̶ OH- + CO2 ↔ Zn + HCO3-

2+ + 2+ (2) Zn + H2O ↔ H +Zn ̶ OH-

Fig 2.1. Proposed reaction mechanism of carbonic anhydrase.

43 Characterization of two carbonic anhydrase genes

In several prokaryotes multiple homologs of CA have been identified (e.g. Escherichia coli: 2

β- and 2 γ- class) suggesting an important physiological role of this enzyme (Tripp et al.

2001). According to genome annotation, the α-proteobacterium B. japonicum strain

USDA110 possesses four genes coding for β-class (blr0500, bll2065, bll4863, bll4865), and one α-class CA (bll1137) (Table 2.1) (Kaneko et al. 2002). So far several α- and β-class CAs were characterized in different α-proteobacteria such as Rhodopseudomonas palustris,

Azospirillum brasiliense and Mesorhizobium loti (Puskas et al. 2000; Kalloniati et al. 2009;

Kaur et al. 2009), and several physiological roles were described in these organisms such as the uptake of carbonate in R. palustris (Puskas et al. 2000) or pH regulation in A. brasiliense

Sp7 and M. loti (Kalloniati et al. 2009; Kaur et al. 2009). Curiously, even though the α-class

CA CAA1 from M. loti was shown to be up-regulated during symbiosis, CAA1 was shown to be not essential for nitrogen fixation or nodule development in Lotus japonicus (Kalloniati et al. 2009).

Table 2.1. Overview of all annotated carbonic anhydrase genes in B. japonicum. Numbers in the first data column indicate the fold change (FC) of expression in wild-type bacteroids when compared to aerobically grown cells (Pessi et al. 2007). Signal intensities (SI) derived from these microarray experiments are also given. In addition number of assigned protein spectra were added when detected in soybean bacteroids (Delmotte et al. 2010).

Soybean bacteroids Carbonic anhydrase genes in B. japonicum FC SI Number of assigned spectra blr0500 (β-class) -2 634 - bll1137 (α-class) -29 154 - bll2065 (β-class) +156 12468 9 bll4863 (β-class) -4 379 - bll4865 (β-class) -3 1364 14

A previous transcriptome analysis of B. japonicum soybean bacteroids and free-living cells showed that the expression of 692 genes was up-regulated during symbiosis, and amongst the

44 Characterization of two carbonic anhydrase genes

20 most highly induced genes a carbonic anhydrase gene (bll2065) was identified (Pessi et al.

2007). Moreover the proteomics analysis on soybean bacteroids detected two carbonic anhydrases, Bll2065 and Bll4865 (Table 2.1). Since in B. japonicum none of the annotated

CA genes had been studied in further depths, we decided to characterize the role of the two β- class CA genes bll2065 and bll4865 during symbiosis.

45 Characterization of two carbonic anhydrase genes

2.3 Material and methods

Bacterial strains, media and growth conditions

Bacterial strains used in this work are listed in Table 2.2. Escherichia coli cells were grown in Luria-Bertani medium (LB) supplemented with the appropriate antibiotics (μg ml-1): ampicillin, 200; kanamycin, 30; tetracycline, 10.

B. japonicum wild type 110spc4 cells were grown in PSY medium (Regensburger and

Hennecke 1983) supplemented with 0.1% L-arabinose. For mutant cultivation the following concentration of antibiotics were added, if needed, to the medium (μg ml-1): spectinomycin,

100; kanamycin, 100; tetracycline, 50.

DNA work

Recombinant DNA work was performed according to standard protocols (Sambrook and

Russel 2001). Escherichia coli S17-1 cells were used for plasmid mobilization into B. japonicum. B. japonicum chromosomal DNA was isolated as previously described (Hahn and

Hennecke 1984).

Fig 2.2. Physical map of the genomic organization of the two β-class carbonic anhydrase genes bll2065 (A) and bll4865 (B) (black arrows) in B. japonicum wild type. The structure and the name of the mutant strains are given at the bottom. The genes bll2065-2066 were replaced byy a kanamycin cassette by double homologous recombination whereas the bll4865 gene was inactivated by inserting the pRJ6226 plasmid. Gene names are derived from Rhizobase. HP stands for hypothetical proteins. Drawn to scale.

46 Characterization of two carbonic anhydrase genes

Construction of Δbll2065-2066 deletion and bll4865::pRJ6226 insertion mutants

Mutagenesis of selected genes (bll2065-2066) (Fig 2.2) was done by marker exchange. PCR fragments of the 5ˈ and 3ˈ flanking regions of the bll2065-2066 genes were amplified using the following primer pairs: bll2065-1R (5ˈ-CTCCCCCGGGAGCTGACGCTAGC-3ˈ) and bll2065-F (5ˈ-GCCGTGG-TTGAAGAGCTTGCTTGCC-3ˈ) for the downstream region and bll2065-3R (5ˈ-CTGCCCCG-GGTTGACAGCATTATG-3ˈ) and bll2065-4F (5ˈ-

CGAAGGCAC-CAGGAAATGAG-3ˈ) for the upstream region. PCR products were cloned in the pGEM-T Easy vector (Promega, Madison, WI, USA) and sent for sequence analysis.

Correct DNA fragments were subcloned into the pSUP202pol4 vector (pRJ6223). A 1.2 kb kanamycin resistance cassette (aphII) derived from pBSL15 (Alexeyev 1995) was introduced in both directions between the up- and downstream regions. The resulting plasmids pRJ6224 and pRJ6225 were transferred into B. japonicum strain 110spc4 for marker replacement using previously described methods (Hahn and Hennecke 1984). The correct genomic structure was verified by PCR. The resulting deletion mutants were named 6224 and 6225 (Table 2.2). The mutants still had coding information for 9 nt of the 3ˈ end (bll2065) and 12 nt of the 5ˈ end

(bll2066).

Mutagenesis of bll4865 was done by insertion of a derivative of the pSUP202pol4 vector in the bll4865 gene using single cross over. An internal 348 bp fragment of bll4865 was amplified using primers bll4865-3-for (5ˈ-GCTCTAGAGCGGACGCGGAGAC-3ˈ) and bll4865-4-rev (5ˈ-GGAATTCCGACTG-CCGCATTG-3ˈ). The PCR product was cloned in the pGEM-T Easy vector, sequenced and subcloned into the pSUP202pol4 vector resulting in pRJ6226. The plasmid was mobilized into B. japonicum wild type strain yielding strain 6226 listed in Table 2.2.

47 Characterization of two carbonic anhydrase genes

Table 2.2. Bacterial strains and plasmids used in this work.

Strain or plasmid Relevant genotype or phenotype Source or reference E. coli DH5α supE44 ΔlacU169-(Φ80lacZΔM15)hsdR1 recA1 gyrA96 thi-1 Bethesda Research, relA1 Gaithersburg, U.S.A.

S17-1 Smr Spr; (RP4-2 kan::Tn7 tet::Mu, integrated in the Simon et al 1983 chromosome) B. japonicum 110spc4 Spr; wild type Regensburger and Hennecke 1983 6224 Spr Kmr; bll2065-2066::ahpII (opposite orientation) This work 6225 Spr Kmr; bll2065-2066::ahpII (same orientation) This work

6226 Spr Tetr; pSUP202pol4 insertion in bll4865 This work Plasmids pGEM-T Easy Apr; cloning vector Promega Corporation, Madison, WI

pBSL15 Apr Kmr; cloning vector Alexeyev 1995 pSUP202pol4 Tcr; (pSUP202) part of the polylinker from pBlueskript II KS+ Fischer et al. 1993 between EcoRI and PstI pRJ6220 Apr; pGEM-T Easy with a 0.83 kb fragment comprising the This work upstream promoter region of bll2065 and 9 nt of bll2065 pRJ6221 Apr; pGEM-T Easy with a 0.9 kb fragment comprising the This work downstream promoter region of bll2066 and 12 nt of bll2066 pRJ6222 Apr; pGEM-T Easy with a 0.35 kb fragment comprising an This work internal part of bll4865 pRJ6223 Tcr Kmr; (pSUP202pol4) comprising the bll2065/2066 5ˈ- and This work bll2065/2066 3ˈ-flanking regions on a 1.73-kb SacII-NotI fragment pRJ6224 Tcr Kmr; (pSUP202pol4) comprising the ahpII cassette derived This work from pBSL15 flanked by the bll2065/2066 5ˈ-upstream and bll2065/2066 3ˈ-downstream region (opposite orientation as the reading frame of bll2065/2066) pRJ6225 Tcr Kmr; (pSUP202pol4) comprising the ahpII cassette derived This work from pBSL15 flanked by the bll2065/2066 5ˈ-upstream and bll2065/2066 3ˈ-downstream region (same orientation as the reading frame of bll2065/2066) pRJ6226 Tcr; (pSUP202pol4) comprising the 0.348 kb internal fragment This work of bll4865 derived from pRJ6222

48 Characterization of two carbonic anhydrase genes

Plant growth

Soybean (Glycine max cv. Williams 82) seed sterilization and plant growth was done as previously described (Göttfert et al. 1990). Symbiotic phenotype was determined using the acetylene reduction assay (Hahn and Hennecke 1984).

Quantitative real-time PCR

For quantitative real time PCR analysis (qRT-PCR), root nodules from soybean plants induced by B. japonicum wild type and 6224, 6225 strains were collected 21 days post infection (DPI) and RNA was extracted and processed as previously described (Pessi et al.

2007). While cDNA was obtained using hexamer primers, primer sequences for qRT-PCR were designed using the CloneManager9 software (Scientific & Educational Software, Cary,

NC, USA). The following primers combinations were used: bll2065-11F

(5ˈTAACCAGCACTGCGCGTTC-3ˈ), bll2065-12R (5ˈ-AGCAGGTTATCTGGCTTGG-3ˈ)

(for bll2065); bll2064-F (5ˈ-CATGCGCTGC-CTGATATTGC-3ˈ), bll2064-R (5ˈ-

TGCCAGTTTCGTCTCGTGTC-3ˈ) (for bll2064); blr0500-F (5ˈ-

AGCATTTGCTGGAAGGCTAC-3ˈ), blr0500-R (5ˈ-CCATCACTTCCGGGAACTGC-3ˈ)

(for blr0500); bll1137-F (5ˈ-GACATCGAGGCGACGATCAAG-3ˈ), bll1137-R (5ˈ-

TTCGGCGAAGTTGAGCTGG-3ˈ) (for bll1137); bll4863-F (5ˈ- AGGTCGCTCATCC-

ACTTCAC -3ˈ), bll4863-2R (5ˈ-TCCATCAGCCGCTTCAAC-3ˈ) (for bll4863); bll4865-F

(5ˈ-ATGTGTGACAAATGCTCCG-3ˈ), bll4865-R (5ˈ-ACCGCAGGCCCGCCAAAC-3ˈ)

(for bll4865); sigA-1533F (5ˈ-CTGATCCAGGAAGGCA-ACATC-3ˈ), sigA-1617R (5ˈ-

TGGCGTAGGTCGAGAACTTGT-3ˈ) (for the constitutive sigA gene). The amplicon length ranged from 90 to 110 bp. The qRT-PCR reaction, cycling and data evaluation were done as previously described (Hauser et al. 2006). Amplicons were visualized as a quality control on a 1% agarose gel.

49 Characterization of two carbonic anhydrase genes

2.4 Results

Bradyrhizobium japonicum possesses five carbonic anhydrase genes

The genome sequence of B. japonicum shows the presence of five carbonic anhydrase (CA) genes (Table 2.1) (Kaneko et al. 2002). Four of them belong to the β-class (blr0500, bll2065, bll4863, bll4865) and one (bll1137) to the α-class. Previous global transcriptome analysis had demonstrated that bll2065, which is the only CA gene located in the symbiotic island, is highly induced in symbiosis when compared to free-living aerobically grown cells displaying a fold change (FC) value of 156 (Table 2.1) (Pessi et al. 2007). All the other four B. japonicum CA genes showed down-regulation in this study (Table 2.1) (Pessi et al. 2007).

Nevertheless, gene expression data derived from microarray analysis of symbiotically grown

B. japonicum revealed that all CA genes are expressed during symbiosis, albeit to different extents (Table 2.1). The highest expression levels are displayed for the bll2065 and bll4865 genes. Interestingly, the products of these genes (Bll2065 and Bll4865) were also identified in a more recent shotgun proteomics study (Delmotte et al. 2010; Koch et al. 2010).

An amino acid alignment of the four B. japonicum β-class CAs revealed that the six amino acid residues shown to be conserved among β-class CAs from various species (Smith and

Ferry 2000) are present in all four β-class CA homologs of B. japonicum (Fig 2.3). Three of those residues, two cysteines and one histidine (Fig 2.3) are expected to serve as ligands of the zinc active site (Smith and Ferry 2000). The remaining three residues, aspartate, arginine and a glycine are important to maintain the catalytic function of the enzyme (Fig 2.3) (Smith et al. 2002). Moreover, in silico topology analysis using the SignalP 3.0 Server

(http://www.cbs.dtu.dk/services/SignalP/) suggested that Bll2065, Bll4863, Bll4865, but not

Blr0500, carry a signal peptide. Using the TatP1.0 server, the signal peptide was proposed to be a twin-arginine motif (http://www.cbs.dtu.dk/services/TatP/) (Fig 2.3). The twin-arginine translocation pathway enables proteins with an N-terminal signal peptide to cross the

50 Characterization of two carbonic anhydrase genes cytoplasmic membrane. Thus it is likely that Bll2065, Bll4863, Bll4865 are exported to the periplasm.

Bll4863 ----MCENCGSVRG--AHIRPSRRSLIHFTVSAIG-LAFAGTAFAKEAKAPPKPQNVLAP 53 Bll4865 ----MCDKCSENLH--QSVAPSRRSMMLFAASALGVAAFGGPAVAKEAKAPPKPQNVLSP 54 Bll2065 MRRHMHAANSPNMNHLTSTARSRRSLLLFAASTVFLRFGNKIADAREAKGTPKPDNLLSP 60 Blr0500 ------MVTFP 5 : *

Bll4863 DAALKRLMEGNARYVEGVSRRHDFKHEREALAGGQNPFAAVLSCADSRIAPEYAFDTGRG 113 Bll4865 DAALKRLMDGNSRYVSGVSRRHDFAHEREALVGGQNPYAAVLSCADSRIAPEYAFDSGRG 114 Bll2065 DAALKRLLMGNDRYVQGTSRADDFRRERSALVEGQNPYAAVLSCADSRVAPELVFDSGLG 120 Blr0500 ----KHLLEGYQAFAT-QRLPTEQTRYRELSVKGQFPEVMVIGCCDSRVSPEVIFDVGPG 60 *:*: * :. : : *. . ** * . *:.*.***::** ** * *

Bll4863 DLFVCRVAGNFA------GTETIASMEYAVAVLGAPLILVLGHDSCGAVDATLKAIKD 165 Bll4865 DLFVCRVAGNFA------GDETVASMEYTVAVLGTPLILVLGHDNCGAVDATIKSLKD 166 Bll2065 DLFVCRVAGNFA------NDDTLASMEYAVAVLNTPLILVLGHDHCGAIDATIKSLHQ 172 Blr0500 ELFVVRNIANLVPVYQPDGNAHGVSAALEYAVTVLKVKHIVVLGHAQCGGIRAFVDKIEP 120 :*** * .*:. * . *::**:*:** . *:**** **.: * :. :.

Bll4863 NTS--PPGHIPSLIDAIAPAAKAAMQQGGDVLDKATRQNVIDNVAKLKSAAPILNAAVEQ 233 Bll4865 DKP--LPGHIPSLVSAIAPAVKTAAQQSGNALDNAIRQNVIDNVAKLKSAAPILNAAVEQ 224 Bll2065 DKP--PPGHISSLVTALAPAVNASLGQAGDISAHATRKNVIDNVNKLRSTGPILTAAVEQ 230 Blr0500 LTPGDFIGKWMQMFIKPGEVVEQRDHESMAQFVERIEKAAVFRSLENLMTFPFVQKAVDA 180 .. *: .:. . ..: :: . .: .: . : : *:: **:

Bll4863 GKLKVMGGIYRLTTGTVDLIAQG------Bll4865 GKLKVVGGIYRLSTGTVELLAQG------Bll2065 NRLKVVGGLYRIGTGKVDIVS------Blr0500 GQMQTHGAYFGVAEGSLFVLDKVAKEFRSVKDAA .:::. *. : : *.: ::

Fig 2.3. CLUSTALW alignment of the deduced amino acid sequence of all β-carbonic anhydrases from B. japonicum. The conserved Zn ligands (two cysteines and one histidine) are indicated in red shaded boxes. Three conserved residues, an aspartate, an arginine, and a glycine, which were found to be present in the active sites of several β-carbonic anhydrases, are shown in grey shaded boxes. The in silico predicted TAT signal is underlined.

Construction of mutations in two β carbonic anhydrase genes bll2065 and bll4865

Transcription profiling of the bll2065 genomic region (Fig 2.4) in B. japonicum wild type either grown in symbiosis with soybean or aerobically free-living (Pessi et al. 2007) confirmed that the genes bll2066 and bll2065 are highly up-regulated during symbiosis and suggested that the gene bll2066 is co-transcribed with the carbonic anhydrase gene bll2065.

The bll2066 gene codes for a hypothetical protein that shares no significant sequence identity with previously identified proteins.

51 Characterization of two carbonic anhydrase genes

Fig 2.4. Transcript analysis of the bll2065 region. Signal intensities are derived from previously performed microarray experiments (Pessi et al. 2007). Genome coordinates are given on the vertical axis and genes are visualized by vertical arrows. Genes were assigned according to the annotation of Kaneko et al. (2002). CA stands for carbonic anhydrase and HP stands for hypothetical protein.

To study the role of the two carbonic anhydrase genes that are highly expressed in B. japonicum during symbiosis with soybean, two mutants were constructed (Fig 2.2). The bll2065 gene and the upstream gene bll2066 were deleted and replaced by a kanamycin cassette by double homologous recombination (Fig 2.2, Table 2.1). In bll2065-2066, 9 nt of the 3ˈend of bll2065 as well as 12 nt of the 5ˈend of bll2066 were kept. The bll4865 gene was mutagenized by a plasmid insertion in the gene using a single-cross-oveer strategy. For this purpose, an internal 348 bp gene fragment of bll4865 was amplified and cloned in a suicide vector (Fig 2.2, Table 2.1). The resulting plasmid pRJ6226 was then mobilized by conjugation into the B. japonicum wild type strain and 6224, 6225 mutant strains. While

bll2065-2066 and bll4865::pRJ6226 were successfully constructed, the attempt to construct a double mutant of both carbonic anhydrase genes failed.

52 Characterization of two carbonic anhydrase genes

Symbiotic phenotype of ∆bll2065-2066 and bll4865::pRJ6226 strains

The fact that Bll2065 and Bll4865 were found to be expressed in B. japonicum during symbiosis with soybean, suggested that these proteins are potentially important for symbiosis.

Thus, bll2065-2066 (6224 and 6225) and bll4865::pRJ6226 (6226) strains were tested for their symbiotic properties on soybean plants. During the phenotypic characterization, attention was paid to the ability to develop nodules and to fix nitrogen. For this purpose, seedlings of soybean were infected with B. japonicum wild-type and mutant strains. After 21 days post infection (DPI), nitrogenase activity was assayed using the acetylene reduction assay. The mutants 6225 and 6226 showed wild-type like phenotypes with respect to nodule number, and the ability to fix nitrogen (Table 2.3). The 6225 strain displayed a decreased nodule dry weight as compared to the parental strain. In the 6224 strain, which has the kanamycin resistant gene inserted in the opposite direction to the bll2066 reading frame, nodule number and dry weight did not alter significantly between wild type and mutant strain implying that the mutant strain is not affected in nodule development. However, nitrogen fixation was decreased to 50% in 6224 as compared to the parental strain (Table 2.3).

Table 2.3. Symbiotic properties of B. japonicum wild type and carbonic anhydrase mutant strains on soybean plants. In total two independent plant infection experiments were performed. At least six plants were measured per strain. Shown are mean values ± standard deviations.

Nodule dry weight Fix activity Strain Relevant genotype Nodule number (mg) (% of wild type) 110spc4 Wild type 27 ± 8 1.62 ± 0.2 100 ± 25 Δbll2066/2065 in the opposite 6224 direction as the bll2065-bll2066 37 ± 1.5 1.13 ± 0.3 55 ± 10 reading frame Δbll2066/2065 in the same 6225 direction as the bll2065-bll2066 35 ± 9 0.77 ± 0.2 94 ± 23 reading frame 110spc4 Wild type 25 ± 3.2 1.56 ± 0.8 100 ± 36 6226 Insertion in bll4865 27 ± 2.8 2.43 ± 0.12 95 ± 22

The wild type strain is listed twice in the table, which reflects the independent experiments performed to assess the symbiotic properties of 6224, 6225 and later 6226, respectively.

53 Characterization of two carbonic anhydrase genes

Expression studies of B. japonicum carbonic anhydrase genes

In order to investigate the expression level of all CA genes in symbiotically grown B. japonicum wild type and 6224, 6225, quantitative real time (qRT) PCR was performed (Table

2.4). The primary sigma factor gene (sigA) which was previously shown to be constitutively expressed in various conditions (Hauser et al. 2006) was used as a reference for relative quantiﬁcation. These expression analysis of CA genes in symbiotically grown B. japonicum wild type revealed that these genes alter in expression (data not shown). bll2065 was confirmed to be the most highly expressed gene followed by bll4865 and blr0500. In contrast bll1137 as well as bll4863 were only slightly expressed during symbiosis. This is in agreement with previously performed microarray analyses (Table 2.1) (Pessi et al. 2007). In the 6224 and 6225 mutant several differences in expression were detected for the other four

CA genes when compared to the gene expression in the wild type (Table 2.4). Interestingly, almost all other CA genes showed decreased expression in the 6224 and 6225 mutant, most apparently the bll4863 gene which was five-fold down-regulated in the 6225 mutant compared to the parental strain.

Table 2.4. Fold change values of gene expression from qRT PCR experiments when comparing the expression profile of symbiotically grown B. japonicum wild-type (wt) to that of symbiotically grown mutant strains 6224 and 6225. In total two independent biological replicates were analyzed. Given are mean values plus standard deviation.

Gene number Description 6224 vs. wt 6225 vs. wt blr0500 β-class carbonic anhydrase 0.8 ± 0.3 0.6 ± 0.2 bll1137 α-class carbonic anhydrase 0.5 ± 0.2 0.5 ± 0.2 bll4863 β-class carbonic anhydrase 0.7 ± 0.6 0.2 ± 0.1 bll4865 β-class carbonic anhydrase 1.1 ± 0.8 0.5 ± 0.1 bsl2064 Hypothetical protein 0.2 ± 0.1 1.9 ± 0.9

54 Characterization of two carbonic anhydrase genes

Downstream of the bll2065-2066 operon bsl2064 is located (Fig 2.2). Bsl2064 shares no sequence identity to previously identified proteins and was shown to be up-regulated in B. japonicum during symbiosis with soybean (Pessi et al. 2007). Thus it was speculated that due to an outward reading activity of the inserted ahpII cassette expression levels of bsl2064 alters in strain 6225. In fact the expression analysis revealed that bsl2064 is two-fold up- regulated in the 6225 mutant as compared to B. japonicum wild type and five-fold down- regulated in a 6224 mutant background (Table 2.4). Thus the orientation of the inserted ahpII cassette influences the expression of bsl2064.

55 Characterization of two carbonic anhydrase genes

2.5 Discussion

Here, we present the characterization of two β-class CA genes (bll2065 and bll4865) whose gene products were detected during symbiosis in a global study (Pessi et al. 2007; Delmotte et al. 2010). Moreover, the expression of bll2065 was shown to be highly up-regulated during symbiosis (Pessi et al. 2007) and the gene is localized in the symbiotic region of the B. japonicum chromosome. Notably, CA genes of M. loti and S. meliloti were also shown to be located on the genome in regions known to be important for nitrogen fixation such as the symbiotic island or the pSymA plasmid, respectively (Kalloniati et al. 2009). Interestingly, no other so far sequenced rhizobial species harbors more annotated CA genes than B. japonicum.

Carbonic anhydrases are enzymes catalyzing the reversible hydration of CO2 to carbonate.

Thus, during symbiosis, a CA could serve the function to convert CO2 gas, which is produced for example during the consumption of C4-dicarboxylic acids in the TCA cycle, to soluble carbonate which could be transported back to the plant. The plant in turn could use carbonate for assimilation or a soybean-derived CA could convert carbonate back to CO2 and either use it for CO2 fixation or release it over the stomata into the atmosphere.

The symbiotic phenotype of the Δbll2065-2066 mutant strains (6224, 6225) (Fig 2.1) revealed a dependence on the orientation of the inserted antibiotic resistance cassette (aphII) possibly caused by an out reading promoter activity from aphII. Tentative conclusions regarding this phenomenon are put forward here, which are solely based on the previously mentioned symbiotic phenotype and expression analysis. An antisense inhibition of bll2067 expression (Fig 2.2) could occur in mutant 6224 (kanamycin cassette inserted opposite to the bll2065-2066 reading frame) which displayed 50% decreased nitrogenase activity. The bll2067 gene encodes the nodule formation efficiency C protein (NfeC) and was previously shown to be essential for nodulation competitiveness in B. japonicum (Chun and Stacey

1993). Another option would be an overexpression of the downstream gene bsl2064 (Fig 2.2)

56 Characterization of two carbonic anhydrase genes in mutant 6225 (kanamycin cassette inserted in the direction of the bll2065-2066 reading frames) which could potentially complement the symbiotic phenotype observed for strain

6224. Indeed, gene expression analysis demonstrated that bsl2064 in a 6225 mutant background is two-fold up-regulated when compared to the wild type (Table 2.4).

Alternatively, the insertion of the kanamycin cassette could cause polar effects on the gene expression of bsl2064 in a 6224 mutant (oriented opposite direction of the bll2065-2066 reading frame). The performed real time PCR analysis showed that in a 6224 mutant the bsl2064 gene is five-fold down-regulated as compared to the parental strain (Table 2.4).

These analyses suggest that the orientation of the ahpII cassette influence the gene expression of bsl2064. Notably, bsl2064 is also up-regulated in symbiotically grown B. japonicum as compared to free-living aerobically grown cells (Fold Change of 14) (Pessi et al. 2007). At the moment, we can only speculate about the reason for the symbiotic phenotype of strain

6224. However, one can rule out that the deletion of the carbonic anhydrase gene per se has caused the phenotype.

The level of expression of all annotated CAs in the wild type alters. The highest expression in nodules is displayed for bll2065 and bll4865 which is in agreement with previously performed microarray experiments (Table 2.1) (Pessi et al. 2007). Moreover these analyses showed that none of the other CA genes is significantly up-regulated in 6224 or 6225 when compared to the expression level of these genes in the wild type strain. Thus the lack of a symbiotic defect in a ∆bll2065-2066 mutant could be due to the fact that the remaining four

CAs together could compensate the absence of bll2065 in a 6224 and 6225 mutant background. Likewise, strain bll4865::pRJ6226 (6226) was not affected in the ability to develop root nodules or fix nitrogen. In M. loti, the α-type CA gene CAA1 was shown to be expressed during symbiosis with L. japonicus, however, was not essential for symbiotic nitrogen fixation (Kalloniati et al. 2009). Following these findings Kalloniati and coworkers

57 Characterization of two carbonic anhydrase genes draw their attention to a more subtle phenotype than nitrogen fixation. They could show that a reduced CA activity in M. loti causes an enrichment of 13C when compared to wild type leaves (Kalloniati et al. 2009). Thus, CAA1 seems to be involved in carbon economy of root nodules of L. japonicus. Following these findings phenotypic characterization of 6224, 6225 and 6226 should be expanded to include phenotypes other than nitrogen fixation.

In fact, several important physiological functions were assigned for CA genes in other α- proteobacteria such as the regulation of intracellular pH (Puskas et al. 2000; Kalloniati et al.

2009; Kaur et al. 2009). Unfortunately, several attempts to construct a mutant strain in which bll2065 and bll4865 were deleted or inactivated failed. A possible reason for this failure is that the absence of both CAs is lethal for B. japonicum. In conclusion, this study has described the characterization of two β-class CA genes bll2065 and bll4865 in B. japonicum which were shown to be substantially expressed during symbiosis with soybean; however each of the two alone is not essential for symbiotic nitrogen fixation.

58 Rhizobial adaptation to hosts

CHAPTER 3

Rhizobial adaptation to hosts, a new facet in the legume root-nodule

symbiosis

Modified, expanded version of a paper published in:

Molecular Plant-Microbe Interactions 23:784-790 (2010)

Supplemented tables (S3.1-5) are provided as electronic files on the enclosed CD.

59 Rhizobial adaptation to hosts

3.1 Abstract

Rhizobia are able to infect legume roots, elicit root nodules, and live therein as endosymbiotic, nitrogen-fixing bacteroids. Host recognition and specificity are the results of early programming events in bacteria and plants, in which important signal molecules play key roles. Here, we introduce a new aspect of this symbiosis: the adaptive response to hosts.

This refers to late events in bacteroids where specific genes are transcribed and translated that help the endosymbionts to meet the disparate environmental requirements imposed by the hosts in which they live. The host-adaptation concept was elaborated with Bradyrhizobium japonicum and three different legumes (soybean, cowpea, siratro). Transcriptomes and proteomes in root-nodule bacteroids were analyzed and compared, and genes and proteins were identified which are specifically induced in only one of the three hosts. We focused on those determinants that were congruent in the two data sets of host-specific transcripts and proteins, 7 for soybean, 5 for siratro, and 2 for cowpea. One gene cluster for a predicted

ABC-type transporter and a monooxygenase, differentially expressed in siratro, was deleted in B. japonicum. The respective mutant had a symbiotic defect on siratro rather than on soybean or cowpea. Complementation analysis revealed that the entire operon could restore the host-specific phenotype on siratro plants to wild-type, whereas the monooxygenase gene alone could not. The identification of a siratro-specific adaptation determinant demonstrates the value of the applied approach and corroborates the host-specific adaptation concept.

60 Rhizobial adaptation to hosts

3.2 Introduction

Certain alpha- and beta-proteobacteria, collectively called rhizobia, are soil bacteria which are able to establish a symbiotic interaction with legume plants. This leads to the formation of a specific plant organ, the root nodule. Inside nodules the bacteria live as nitrogen-fixing endosymbionts (bacteroids) in a protected environmental niche. Rhizobia-legume symbioses develop as a result of a complex molecular interplay between both partners (Broughton et al.

2000; Spaink 2000). The host range of rhizobial species is a trait that allows them to nodulate either a narrow or a broad range of legumes. A rather narrow host range has been described for Sinorhizobium meliloti which induces nitrogen-fixing nodules only on few species of the genera of Medicago, Melilotus, and Trigonella. In contrast, numerous Bradyrhizobium species have a broader host range as they nodulate legumes from different families such as

Papilionoideae, Mimosoideae, and Phaseoleae (Dénarié et al. 1992).

At the beginning of symbiotic events, plant-released phenolic root exudates, also known as flavonoids (Peters et al. 1986), lead to the expression of many of the rhizobial genes (nod) required for nodulation. Induction of nod genes is controlled by the NodD protein, a member of the LysR family of transcription regulators (Schell 1993). Point mutations in nodD may expand the number of recognized flavonoid inducers, thereby extending host range (McIver et al. 1989). Apart from NodD, species-specific transcription factors may be involved in controlling host range. In Bradyrhizobium japonicum, for example, an isoflavonoid- stimulated two-component regulatory system NodVW is important for nodulation of mungbean, cowpea, and siratro, but only marginally contributes to eliciting symbiosis with soybean (Göttfert et al. 1990; Loh et al. 1997).

The products of nod genes are enzymes for the biosynthesis of bacterial nodulation (Nod) factors (Dénarié et al. 1996). These are lipo-chito-oligosaccharides that elicit a series of physiological and developmental responses on host plants leading to nodule organogenesis

61 Rhizobial adaptation to hosts

(Perret et al. 2000; Madsen et al. 2003). The quality and quantity of chemical modifications at the Nod factor are further important determinants that affect host range (Ehrhardt et al. 1995;

Dénarié et al. 1996).

Once the establishment of a symbiosis between both partners approaches completion, rhizobia are expected to adapt their physiology to that of the host (Oke and Long 1999b). In contrast to the remarkable advances in the area of host specificity, current knowledge about bacterial adaptation to hosts is quite limited, and possible processes associated with adaptation have not been investigated systematically. Certain features in the metabolic diversity of rhizobia are probably crucial for adaptation and survival in nodules of disparate host plants. Only sporadic literature exists in support of this notion. For example, a

Rhizobium sp. NGR234 pckA mutant defective in phosphoenolpyruvate carboxykinase was reported to display a host-dependent phenotype (Osteras et al. 1991). More recently, host- dependent expression of Rhizobium leguminosarum bv. viciae hydrogenase activity was proposed to be due to a host-derived substance (Brito et al. 2008), and a host-specific regulation of nif and fix gene expression was shown for R. leguminosarum bv. trifolii (Miller et al. 2007).

The aim of the present study was to systematically find out whether or not the exposure of bacteroids to different host nodule environments is reflected by differential transcript and protein expression patterns. Specifically, we analyzed the possible adaptation of B. japonicum during symbiosis with three host plants (soybean, cowpea, and siratro). We have chosen legumes from two different subtribes of the tribe Phaseoleae to cover the maximal diversity of known host plants for B. japonicum. Soybean belongs to the subtribe Glycininae whereas the other two species, cowpea and siratro, are members of the subtribe Phaseolinae (Lee and

Hymowitz 2001; Delgado-Salinas et al. 2006). The transcription profiles were assessed with a previously established microarray technology, using a custom-made Affymetrix GeneChip

62 Rhizobial adaptation to hosts

(Hauser et al. 2007). The results were compared and combined with a complementary gel-

LC-MS/MS proteomics analysis of B. japonicum bacteroids. Our study culminated in the discovery of sets of B. japonicum adaptation genes that are differentially expressed in each of the three hosts. As a proof of principle, we could show that deletion of a gene cluster that is specifically expressed in siratro nodules produced a significant symbiotic defect only on siratro.

63 Rhizobial adaptation to hosts

3.3 Results and discussion

Transcriptome analysis of B. japonicum in root nodules of cowpea, siratro, and soybean

Little is known about the genetic determinants of host-specific bacteroid performance in a persistent root nodule. With the aim to monitor possible transcriptional changes in B. japonicum cells living in symbiosis with Glycine max (soybean), Vigna unguiculata (cowpea) and Maacroptilium atropurpureum (siratro), microarray experiments were performed using a high-density oligonucleotide GeneChip (Hauser et al. 2007; Pessi et al. 2007). RNA from nodules, isolated at the time of maximal nitrogenase activity in the three hosts (see Materials and Methods), was processed and analyzed as described (Pessi et al. 2007). At least three independent experiments for each host were carried out and subjected to careful data analysis

(see Materials and Methods).

Fig 3.1. Hierarchical clustering of all of the 202 identified host-specific adaptation genes identified by transcriptomics. The gene expression level is shown for B. japonicum bacteroids at 21 days post infection (dpi) for soybean and cowpea, and 31 dpi for siratro. The color code for the expression level is shown at the bottom; red color represents a very high expression, green a very low level of expression (log2 scale).

64 Rhizobial adaptation to hosts

While more than 2,000 genes were commonly expressed, irrespective of the host, 129 B. japonicum genes were specifically up-regulated during symbiotic life in soybean, 38 in siratro, and 35 in cowpea (see Supplemental Tables S3.1, S3.2, and S3.3, respectively), which amounts to 202 B. japonicum genes that were defined here as host-specific adaptation genes.

A hierarchical clustering analysis of these 202 genes is displayed in Fig 3.1. The illustration shows a remarkable exclusiveness of their induction.

Proteome analysis of bacteroids from cowpea, siratro, and soybean nodules

To identify B. japonicum proteins that are detectable in root-nodule symbiosis with soybean, siratro, and cowpea, a sensitive gel-LC-MS/MS proteomics approach was applied (Delmotte et al. 2009). To this end, root nodules were collected at the peak of maximal nitrogenase activity, and bacteroids were isolated by gradient centrifugation before they were processed for proteomics (see Materials and Methods). The method detects most of the expressed soluble proteins, many of the membrane-anchored or -associated proteins, but rarely the truly membrane-integral proteins. Among those 1,620 proteins were detected in bacteroids from all three hosts (Fig 3.2), which corresponds roughly to 20% of the potential B. japonicum proteome as deduced from the genome sequence (Kaneko et al. 2002). Amongst them, we found almost all of the soluble housekeeping proteins such as DNA and RNA polymerases, aminoacyl-tRNA synthetases, and ribosomal proteins, which can be regarded as an excellent quality control of the method employed. If detection in at least one of three biological replicates is considered, the sensitive proteomics approach identified 238 bacteroid proteins that were detectable only in soybean, 222 in siratro, and 89 in cowpea (Fig 3.2). All of these are compiled in Supplemental Tables S3.1, S3.2, and S3.3. If detection in at least two of three biological replicates is considered, the respective number of proteins drop down to 63 for soybean, 78 for siratro, and 12 for cowpea bacteroids.

65 Rhizobial adaptation to hosts

Fig 3.2. Venn diagram representing the total number of proteins detected by whole-cell proteomics. B. japonicum wild- type bacteroids were isolated from cowpea, siratro, and soybean. Protein processing and identification are described in Material and Methods.

We assume that the primary reasons why proteins are sometimes not detected in all three biological replicates are: (i) when they are of small size (i.e., the number of peptides available for detection is inherently low), and (ii) when they are present in low abundance (i.e., they are at a threshold level of detection sensitivity), or (iii) a combination of both reasons. On the one hand, the detection of a protein even in only one out of three biological replicates is a positive

(hence, non-debatable) result, showing that the protein is expressed albeit not abundantly. On the other hand, if it comes to judging if such a low-abundance protein can be regarded as host specific, we run the risk of over-estimating our data, and a cut-off at a minimum of two hits in the three replicates appears compelling. These concerns will be considered in the so-called stringent data set (see separate paragraph below).

Another aspect shown in Fig 3.2 is how many bacteroid proteins are shared between just two of the three plants. These numbers are given in the overlapping sectors of pair-wise comparisons, again on the basis of at least one hit in the three biological replicates. The numbers, however, do not reflect the relatedness of the three hosts (Lee and Hymowitz 2001;

Delgado-Salinas et al. 2006).

66 Rhizobial adaptation to hosts

The stringent data set: where host-specific transcriptomes and proteomes overlap

We looked for congruence between the list of host-specific adaptation genes identified by transcriptomics and the list of host-specific adaptation proteins found by proteomics. For the latter, proteins were considered only if they were detected in at least two out of three biological replicates. A total of 14 genes were found for which both, transcripts as well as proteins are synthesized in a host-specific manner. These are compiled in Table 3.1, showing

2 host-specific determinants for cowpea, 5 for siratro, and 7 for soybean. We regard the data set in Table 3.1 as the most stringent and reliable one in our search for functions involved in the adaptive response. The genes in Table 3.1 can now be tested for functional involvement in the respective hosts.

An approved determinant for bacteroid adaptation to life in siratro nodules

We focused our attention on the genes that are host-specifically expressed in bacteroids from siratro nodules, and noticed two (blr1601, bll7011) that encode periplasmic substrate-binding proteins associated with ABC-type transporters (Higgins 2001; Moussatova et al. 2008).

Being soluble proteins, the blr1601 and bll7011 products were detected by proteomics, whereas the membrane-bound components of the cognate ABC transporters were not detected. The genes for the latter, however, were found to be siratro-specifically induced just like the genes for the binding proteins, suggesting co-regulation of the contiguous genes and their organization in two separate operons (blr1601–blr1604, and bll7011–bll7008). One of the two operons is depicted in Fig 3.3A which, in fact, consists of five genes, including blr1600 at the 5’ end. Gene blr1600, which codes for a putative monooxygenase, was also identified as a siratro-specific adaptation determinant by both, transcript and protein analysis

(Table 3.1). The tiling-like architecture of the GeneChip (Hauser et al. 2007) allowed us to analyze the transcription profile of the entire blr1600–1604 genomic region, which confirmed

67 Rhizobial adaptation to hosts that the five genes of the operon are co-transcribed as a polycistronic mRNA. Downstream of the substrate binding protein gene (blr1601) are two permease genes (blr1602, blr1603) and the gene for an ATP-binding protein (blr1604).

A rhizobial ABC transporter induced in bacteroids and functioning in a host-specific manner, i.e. in siratro, is a novelty. It was necessary, therefore, to obtain support for such a host- related role by mutational analysis. Two B. japonicum deletion mutants were constructed by marker replacements (strains 6202, 6203) in which all five genes of the blr1600–1604 operon were deleted (Fig 3.3A). Soybean, cowpea, and siratro plants were then inoculated with the

B. japonicum wild-type and the two mutants. Phenotypic analyses revealed that siratro root nodules suffered from a statistically secured 50% decrease in symbiotic nitrogen fixation activity, as measured in the acetylene reduction test, whereas mutant-infected soybean and cowpea nodules had wild-type activities (Fig 3.3C). Moreover, only siratro plants inoculated with the mutants showed typical symptoms of nitrogen limitations such as chlorotic leaves and pale nodules (Fig 3.3D). Nodule number and dry weight did not alter when siratro, soybean and cowpea were infected with the wild-type or the deletion mutant. In conclusion, the blr1600–1604 operon is specifically important for an optimal function of B. japonicum bacteroids in siratro nodules, whereas its presence or absence makes no difference to the symbiotic performance in nodules of soybean and cowpea.

Tentative conclusions regarding the possible function of the blr1600–1604 genes are put forward here, which are solely based on the predicted protein properties and the aforementioned phenotype. The presence of a co-induced periplasmic substrate-binding protein, encoded by blr1601, suggests that the cognate ABC transporter is an importer rather than an exporter. Hence, siratro nodules are expected to produce a substance that is taken up by B. japonicum bacteroids. This could be either a beneficial or a detrimental compound which, respectively, is either utilized as nutrient or detoxified.

68 Rhizobial adaptation to hosts

Interestingly, the associated monooxygenase encoded by blr1600 might fulfill either role, i.e., to convert a useful substance into a metabolizable form or to inactivate a toxic compound by oxygenation. The lack of a signal peptide suggests that the monooxygenase is located in the cytoplasm. Intriguingly, it was reported recently that cytoplasmic monooxygenases can be associated with ABC-type transport systems and directly process the substrate after import

(Zhang et al. 2007). Although the Blr1601 protein shares 35% identity with the oligopeptide- binding protein OppA from an ABC transport system of Xanthomonas axonopodis (Moutran et al. 2007), and carries a solute-binding domain (PF00496) that in other bacteria was thought to bind oligopeptides (Tam and Saier 1993), these similarities are not compelling enough to suggest oligopeptides as substrates.

In order to understand whether the siratro-specific plant phenotype of the Δblr1600-1604 strain (6203) is due to the deletion of the monooxygenase (blr1600) or to the deletion of the

ABC transporter (blr1601-1604), the 6203 mutant strain was complemented with two different constructs (Fig 3.3A). One construct exclusively harbors the monooxygenase gene blr1600 (6203-05) and the other construct comprised the whole blr1600-1604 operon (6203-

06) (Fig 3.3A). The correct integration of pRJ6205 and pRJ6206 into the B. japonicum mutant strain 6203 was confirmed by PCR analysis of genomic DNA (Fig 3.3B). As a control genomic DNA of B. japonicum wild type and 6203 mutant were included in this experiment.

Using a primer binding upstream of the bsr1599 gene as well as a primer binding site on blr1600 located on pRJ6205 and pRJ6206 it could be shown that in both cases the plasmid integrated upstream of the kanamycin resistance cassette (Fig 3.3B).

69 Table 3.1. B. japonicum host-specific adaptation gene and protein candidates for symbiosis with cowpea, siratro, and soybean assessed by transcriptomics and proteomics.

FC in transcriptomicsd Number of assigned MS/MS spectra Gene numbera Descriptionb Gene namec Cowpea vs soybean Cowpea vs siratro Cowpeae Siratrof Soybeang Cowpea bll0339 4-hydroxyphenylpyruvate dioxygenase - 3.4 8.3 126 0 0 bll0346 hypothetical protein - 3.2 3.4 11 0 0

Siratro vs cowpea Siratro vs soybean bsr0859 hypothetical protein - 5.7 4.3 0 4 0 blr1600 putative monooxygenase - 13.6 5.1 0 7 0 Siratro blr1601 ABC transporter substrate-binding protein - 7.0 5.7 0 28 0 bll6452 acyl-CoA dehydrogenase acd 9.6 7.5 0 20 0 bll7011 ABC transporter sulfonate-binding protein - 11.5 4.1 0 22 0

Soybean vs siratro Soybean vs cowpea blr1080 enoyl-CoA hydratase - 9.1 7.0 0 0 18 blr1349 hypothetical protein - 2.7 2.9 0 0 7 blr2168 putative transketolase alpha subunit protein - 16.1 23.1 0 0 28 Soybean blr2169 putative transketolase beta subunit protein - 11.4 8.4 0 0 38 blr2606 2-nitropropane dioxygenase - 2.3 3.1 0 0 5 blr3389 ABC transporter substrate-binding protein - 4.7 2.6 0 0 12 blr6576 hypothetical protein - 17.2 22.9 0 0 50

aNomenclature according to Kaneko et al. (2002). bFunctional description according to Kaneko et al. (2002). cGene name according to the EMBL-EBI database. dFold change (FC) of expression, comparing bacteroids of two hosts eSum of all spectra of host-specific adaptation proteins for cowpea measured in 3 independent experiments. Proteins were detected in at least two out of three replicates. fSum of all spectra of host-specific adaptation proteins for siratro measured in 3 independent experiments. Proteins were detected in at least two out of three replicates. gSum of all spectra of host-specific adaptation proteins for soybean measured in 3 independent experiments. Proteins were detected in at least two out of three replicates. Rhizobial adaptation to hosts

As expected the corresponding amplicon was obtained for B. japonicum wild type but not for the 6203 mutant strain.

Siratro plants were infected with the 6203-05 and 6203-06 strains, and the symbiotic properties of these plants were analyzed using the acetylene reduction assay (Fig 3.3E).

Interestingly, the monooxygenase gene alone is not able to complement the host specific phenotype of Δblr1600-1604. Plants infected with 6203-05 behaved like the mutant being affected in nitrogen fixation and only reach 35% activity as compared to the parental strain

(Fig 3.3E). Moreover siratro plants displayed symptoms of nitrogen starvation such as yellowish leaves (Fig 3.3F). In contrast, the presence of the ABC transporter genes plus the monooxygenase gene partly overcomes the host specific phenotype of 6203. Siratro plants infected with 6203-06 were able to fix nitrogen almost to 70% as compared to the wild type strain (Fig 3.3E) and displayed green leaves (Fig 3.3F). Soybean and cowpea plants were also included in this experiment (data not shown); however, nitrogen fixation activity did not alter in plants infected with 6203-05 and 6203-06 as compared to the parental strain. These results demonstrate that the complementation of the monooxygenase alone cannot restore the phenotype displayed on siratro plants. The entire operon in contrast could complement the symbiotic defect on siratro plants. This result indicates that the determinant which confers the host specific phenotype is either due to the action of the ABC transporter or to the combination with the monooxygenase.

To conclude, a siratro plant-derived compound might specifically bind to the periplasmic binding protein Blr1601, and will subsequently be transported to the cytoplasm via the ABC transporter (Blr1602-1604). The associated monooxygenase (Blr1600) could now potentially oxygenate the substance and makes it therefore useable for B. japonicum.

71 Rhizobial adaptation to hosts

bllr1600-7 + blr1600-2

1 2 3 4

1 KB

Fig 3.3 Legend see next page.

72 Rhizobial adaptation to hosts

Fig 3.3. Genotype and phenotype of the Δblr1600-1604 strains 6202 and 6203 as well as the complemented strains 6203-05, 6203-06. A. Physical map of the blr1600-1604 genomic locus. Numbers on the top refer to nucleotide positions on the B. japonicum genome according to Kaneko and coworkers (Kaneko et al. 2002). In addition, positions of primers are indicated with arrows as well as the naturally occurring NdeI restriction site located on blr1602. The genotype of the 6203, 6202 mutants are indicated as well as the orientation of the ahpII cassette (black horizontal arrow). The cloned inserts of plasmid pRJ6205 and pRJ6206 are also indicated which were used to complement 6203. B. PCR analysis of genomic DNA of 1. B. japonicum wild type, 2. 6202-05, 3. 6202-06, and 4. 6203 to verify the correct insertion of pRJ6205 and pRJ6206 in a 6203 mutant. The primer blr1600-7 + blr1600-2 amplified a 990 bp product. C./E. Whole nitrogenase activity was assayed by acetylene reduction in root nodules of soybean, cowpea and siratro induced by B. japonicum wild type (wt), Δblr1600- 1604 strains (6202/6203) as well as the complemented strains 6203-05 and 6203-06. In total at least two independent plant infection experiments were performed. A minimum of six plants were measured per strain. Shown are mean values ± standard error. D./F. Visual inspection of the symbiotic phenotype of siratro plants infected by B. japonicum wild-type (wt), mutants, and complemented strains.

In search for the substrate of Blr1600-1604

With the aim to get an idea about the possible substrate transported by Blr1600-1604,

Phenotype MicroArray (PM) analyses were performed. PM tests were performed in 96-well microtiter plates containing various different nutrients. The cell respiration (cell growth) was measured with a redox indicator over time (Bochner et al. 2001). In this experiment the ability of the 6203 mutant strain to grow in the presence of several nutrients was tested.

According to the PM analysis, the 6203 mutant could metabolize adenosine, pyruvic acid, L- cysteine as well as the dipeptide Met-Ala better than the wild type (Fig 3.4). In addition, these experiments revealed that the 6203 mutant was impaired in growth in presence of ethylenediamine and phosphoenolpyruvate (Fig 3.4). The finding that the 6203 mutant is altered in growth as compared with the parental strain when cultivated with ethylenediamine as sole source of nitrogen (Fig 3.4) was interesting in several aspects. Blr1600 is annotated as a nitrilotriacetate (NTA) and ethylenediaminetetraacetate (EDTA) monooxygenase.

Interestingly, the molecule resulting from a complete oxygenation of EDTA is ethylenediamine.

73 Rhizobial adaptation to hosts

Fig 3.4. Phenotypic MicroArray analysis of the 6203 mutant strain on PM1-8 plates (for more information regarding all tested compounds please consult http://www.biolog.com/pdf/PM1-PM10.pdf). A comparison of B. japonicum wild type (red) with 6203 (green) is shown. When the two strains reveal a similar growth in a well, the kinetic graphs overlap and are presented in yellow. Red kinetic graphs indicate a stronger growth by the parental strain as displayed in the case of ethylenediamine (PM 3_D10), phosphoenolpyruvate (PM 4_C1), whereas green kinetic graphs a stronger growth by the 6203 strain as displayed for adenosine (PM 1_E12), pyruvic acid (PM 1_H08), L-cysteine (PM 3_A11), Met-Ala (PM 3_HH12). For illustration purposes the aforementioned examples were highlighted in frames. The graphic illustrates the summary of two performed experiments.

It was previously reported that EDTA and NTA can be degraded by Mesorhizobium sp. and moreover can serve as sole carbon and nitrogen sources in this organism (Nörtemann 1999;

White and Knowles 2003). However, ouur attempts to cultivate the B. japonicum wild type and strain 6203 in the presence of NTA, EDTA and ethylenediamine as a C source failed. We therefore considered the possibility that these compounds were toxic for the cell. Therefore, we evaluated the sensitivity of the 6203 and wild-type strains in the presence of NTA, EDTA, ethylenediamine, and the structurally related compound L-glutamate-N,N-diacetate (L-

GLDA) in plate diffusion assays. Table 3.2 comprises the results of these analyses. Wild type as well as 6203 mutant strains did not show an altered sensitivity towards the tested compounds. The only compound leading to a clearing zone was EDTA which was similar for all tested strains. In conclusion, the substrate transported by Blr1600-1604 remains enigmatic.

74 Rhizobial adaptation to hosts

Alternative approaches will be needed to fully understand the host-specific phenotype of the

6203 mutant specifically on the siratro host.

Table 3.2. Susceptibility of B. japonicum wild type and 6203 strain towards several tested compounds using agar plate diffusion assays. The compounds tested were EDTA, NTA, ethylenediamine and L-glutamate-N,N- diacetate (L-GLDA) at a concentration of 100 mM. Indicated is the distance (mm) of the growth zone around the place where the tested compound was added to minimal medium. All compounds were applied as carbon source.

Strain EDTA NTA ethylenediamine L-GLDA wt 14 0 0 0 6203 13 0 0 0

Further candidates with a perspective to act in a host-responsive manner

We restrict ourselves to briefly commenting on just two other promising cases, one concerning again siratro, the other concerning cowpea. (i) The siratro-specifically induced operon bll7011–7008 (see above) codes for yet another ABC transporter. Its periplasmic substrate-binding protein (Bll7011) has been annotated as a sulfonate-binding protein

(Sugawara et al. 2011). The transporter might, therefore, be important for the uptake of inorganic or organic sulfur by B. japonicum bacteroids, when living in siratro nodules. The need for adequate sulfur supply to bacteroids has recently been addressed in the Lotus japonicus–Mesorhizobium loti symbiosis (Krusell et al. 2005). A host-plant sulfate transporter SST1 was shown to be important for symbiotic nitrogen fixation and plant development. SST1 was proposed to be inserted to the peribacteroid membrane where it transports sulfate from the plant cell cytoplasm into the peribacteroid space, thus making sulfate available as a nutrient to bacteroids. (ii) The cowpea-specific adaptation genes bll0339, bll0343, and bll0346 (Table 3.1, Supplemental Table S3.3) map in close vicinity on the B. japonicum genome, and two of the three have been annotated as coding for key enzymes of tyrosine degradation (Kaneko et al. 2002). Our proteomics study has revealed that the product of bll0339, a putative 4-hydroxyphenylpyruvate dioxygenase, is the most

75 Rhizobial adaptation to hosts abundant of all proteins listed in Table 3.1, as judged from the number of assigned spectra after LC-MS/MS analysis. By analogy with other work on amino acid utilization (Prell and

Poole 2006) it is attractive to speculate that tyrosine is a nutrient for B. japonicum bacteroids in cowpea nodules. In this context, it is relevant to note that S. meliloti was shown to utilize tyrosine as a carbon and nitrogen source (Milcamps and de Bruijn 1999).

The relaxed data set: transcriptomes and proteomes combined

The results of our transcriptomics and proteomics approaches were combined, and all genes and proteins which had been detected as being host-specifically expressed by at least one of the two methods were kept as an expanded, integrated dataset (Supplemental Tables S3.1,

S3.2, and S3.3). We thus identified 347 genes and proteins for soybean, 253 for siratro, and

120 for cowpea, amounting to a total of 720 potential adaptive-response determinants. There are reasons why entries in the list of transcripts do not appear in the list of proteins, or vice versa. For example, an obvious reason has already been alluded to in the case of the ABC transporters: membrane-integral proteins were missed in the proteome approach, whereas their genes were detected as being host-specifically transcribed. Likewise, rapidly degraded mRNAs might escape detection, whereas their more stable translation products do not.

Another reason is that extraction was done from different biological material, i.e., whole nodules were used for RNA isolation, and bacteroids were used for protein isolation.

Especially the step via bacteroid isolation could lead to alterations in protein turnover.

Nevertheless, the relaxed data set ought not to be neglected. It is a source of further potential functions involved in the adaptation of B. japonicum to the three hosts investigated. All of the

720 determinants were functionally classified according to the 14 categories established by the Kazusa DNA Research Institute (Kazusa website). Several categories were identified that

76 Rhizobial adaptation to hosts are overrepresented with statistical significance (Supplemental Table S3.4). Particularly eye- catching are the high numbers of hypothetical proteins (or genes) expressed in all three hosts.

77 Rhizobial adaptation to hosts

3.4 Concluding remarks

Recent developments in genome-wide expression analyses allow in an unbiased way a systematic search for rhizobial determinants involved in symbiosis. Here, we show the utility of an integrated transcriptomics–proteomics approach which has led to candidates for host- specific adaptation functions in B. japonicum, if one applies restrictive criteria in data analysis. Recently, the transcription profiles of Rhizobium leguminosarum bv. viciae in root nodules from pea and vetch were assessed; there was no change in response to these hosts except for a difference in 23S rRNA processing (Karunakaran et al. 2009). The identified host-specific adaptation genes and proteins are promising targets to study the concept of bacterial adaptation in greater detail and thereby unravel new aspects of bacteroid physiology. As a proof of principle, the blr1600–1604 operon, which had been detected as being host-specifically expressed in siratro, was found to be more relevant for the B. japonicum–siratro symbiosis than for the symbioses with soybean and cowpea. Moreover we could show that the entire blr1600-1604 operon and not the monooxygenase blr1600 is responsible for the host specific adaptation phenotype of Δblr1600-1604 on siratro plants.

Since the approaches we used to identify the substrate transported by Blr1600-1604 failed, new strategies such as the isolation of the periplasmic binding protein Blr1601 directly from siratro root nodules and the identification of the potentially bound substrate using MS/MS analysis could be developed.

As the mutant phenotypes in this example showed, the adaptively responding genes may not be essential, but at least important in the relevant host. The situation is somewhat reminiscent of the early events in the recognition and infection of legumes by rhizobia, where few of the nod genes are essential, whereas most of the host-specificity nod genes are important, but not essential, for the nodulation process (Dénarié et al. 1992; Spaink 2000).

78 Rhizobial adaptation to hosts

3.5 Material and methods

DNA methods and construction of mutant strains

DNA was isolated from B. japonicum wild-type strain 110spc4 as previously described

(Hahn and Hennecke 1984). Plasmid DNA from Escherichia coli strains was obtained by using the NucleoSpin® Plasmid kit (Macherey-Nagel, Düren, Germany). Recombinant DNA work was performed according to standard protocols (Sambrook and Russel 2001). To construct a deletion mutation in the blr1600-1604 gene cluster, PCR fragments of the 5´and

3´flanking regions of the blr1600-1604 genes were amplified using the following primer pairs: blr1600-3_rev (5´-GGACTAGTCCCGAATCAACACGTAGG-3´) and blr1600-4_for

(5´-TTTCTGCAGCACATGCGGTTCGTGTAGAGG-3´) for the 5´ region; blr1600-5_rev

(5´-TTTCTGCAGCTTCCAGGCGGCATAAAATTTC-3´) and blr1600-6_for (5´-

CCGCTCGAGTAGCGATCGCCGAGAGGCGAG-3´) for the 3´ region. PCR products were cloned in the pGEM-T Easy vector (Promega, Madison, WI, USA), and subsequently in pSUP202pol2 vector lacking the cat resistance cassette. A 1.2 kb kanamycin resistance cassette (aphII) derived from pBSL15 (Alexeyev 1995) was introduced in both directions between the up- and downstream regions. The resulting plasmids pRJ6202 and pRJ6203 were transferred into B. japonicum strain 110spc4 for marker replacement using previously described methods (Hahn and Hennecke 1984). The correct genomic structure was verified by PCR. The resulting deletion mutants were named 6202 and 6203 (Fig 3.3A). They still had coding information for 23 N-terminal amino acids (blr1600) and 15 C-terminal amino acids

(blr1604).

To complement the 6203 deletion mutant, two different constructs were used (Fig 3.3A). The first construct contains the monooxygenase gene blr1600 plus 615 bp of its upstream promotor region. This 2.6 kb DNA fragment was amplified using the PhusionTM DNA polymerase (Finnzymes, BioConcept, Allschwil, Switzerland) and the following primers

79 Rhizobial adaptation to hosts pairs: blr1600-3 rev (5´-GGACTAGTCCCGAATCAACACGTAGG-3´) and blr1600-9 for

(5´-GCGGCCGCAAGCTCCGAATGTGATAAAG-3´) (Fig 3.3A). The resulting PCR fragment was cloned in the pGEM T-Easy vector, sequenced and subcloned in the pSUP202pol4 vector. The plasmid was named pRJ6205 (Fig 3.3A) and mobilized in the 6203 strain, yielding in strain 6203-05. 5´

The second construct, harbors the entire blr1600-blr1604 operon plus 615 bp of the upstream promotor region of blr1600 (Fig 3.3A). This 7.3 kb DNA fragment was amplified using the

PhusionTM DNA polymerase (Finnzymes, BioConcept, Allschwil, Switzerland) in two steps.

Using the primers blr1600-3 rev (5´-GGACTAGTCCCGAATCAACACGTAGG-3´) and blr1600-10 rev (5´-AGACCTGAAACCCGCCTAGC-3´) a 4.2 kb fragment was amplified.

The primers blr1600-11 rev (5´-GCCTGGATCATCACCATCAC-3´) and blr1600-12 for (5´-

TGCGCCGCGCCCGCGCAGTCACCCTCCC-3´) were used to amplify a 3.2 kb fragment.

The two PCR fragments were cloned in the pGEM T-Easy vector, sequenced and subcloned in the pSUP202pol4 vector. A natural NdeI restriction site (Fig 3.3A) was used to assemble the two fragments. The resulting plasmid was named pRJ6206 (Fig 3.3A) and mobilized in the 6203 strain yielding in strain 6203-06. The correct integration of pRJ6205 and pRJ6206 in strain 6203-05 and 6203-06 was verified by PCR analysis of genomic DNA (Fig 3.3B).

The following primers pairs were used to analyze the correct vector insertion upstream of the kanamycin cassette: the forward primer blr1600-7 (5´-GGTTCGATACGCGCCCATTG-3´) binds on the genome upstream of bsr1599 and the reverse primer blr1600-2 (5´-

TTGTTTCTCGGGCAATGCTG-3´) binds on the blr1600 gene located on the plasmids pRJ6205 and pRJ6206 respectively. According to the PCR results it was shown that in both cases the integration occurred upstream of blr1600.

80 Rhizobial adaptation to hosts

Phenotypic MicroArray tests

The B. japonicum wild type and strain 6203 were analyzed using Phenotypic MicroArray plates (PM1-8). These experiments were performed by Biolog, Inc. (Biolog, Phenotype

MicroArray, Hayward, U.S.A.) as previously described (Bochner et al. 2001).

Growth and sensitivity assays

The ability to grow in presence of EDTA, NTA and ethylenediamine (10, 50 and 100 mM) was tested in minimal medium (Becker et al. 2004) for the wild type and 6203 mutant. The sensitivity of wild type and 6203 mutant to potential growth-inhibitory compounds like

EDTA, NTA, ethylenediamine and L-glutamate-N,N-diacetate at concentrations of 100 mM was tested on plate diffusion assays as described earlier (Lindemann et al. 2010).

Plant material, inoculation, and growth conditions

Soybean seeds (Glycine max [L.] Merr. cv. Williams) were supplied by the CIFA, Las

Torres-Tomejil (Seville, Spain) and surface-sterilized as described (Hahn and Hennecke

1984; Göttfert et al. 1990). Cowpea (Vigna unguiculata [L.] Walp. cv. Red Caloona) and siratro seeds (Macroptilium atropurpureum [DC.] Urb.) were provided by W. D. Broughton

(University of Geneva, Switzerland) and surface-sterilized according to Lewin et al. (1990).

Germination, inoculation, and growth of the plants were done as previously described (Hahn and Hennecke 1984; Göttfert et al. 1990). Nitrogenase activity in root nodules was determined after 13, 21, 31, and 43 days post inoculation (dpi) and found to be maximal at 21 dpi for soybean and cowpea, and 31 dpi for siratro, which might explain the slower development of siratro plants as compared with the other hosts. These time points were chosen for transcriptome and proteome analyses in at least three independent experiments

81 Rhizobial adaptation to hosts

(see below). The acetylene-dependent ethylene production was measured to monitor nitrogenase activity (Hahn and Hennecke 1984; Göttfert et al. 1990)

Transcriptome analyses

Root nodules collected from plants at the time of maximal nitrogenase activity were immediately frozen in liquid nitrogen, and stored at –80°C for subsequent RNA isolation.

RNA from symbiotically grown bacteria was extracted and processed as previously described

(Pessi et al. 2007). Three to four biological replicates of B. japonicum-infected root nodules from soybean, cowpea and siratro were prepared for transcriptome analyses. Independently, host root-only tissue was collected and processed for GeneChip analysis and used as hybridization control (see below) (Pessi et al. 2007).

The high-density oligonucleotide GeneChip (BJAPETHa520090), custom-designed and manufactured by Affymetrix (Santa Clara, CA, USA), was used for transcriptome profiling

(Hauser et al. 2007). All probe pairs had been pruned against the soybean ESTs represented on the commercial Affymetrix Soybean Genome Array to minimize cross hybridization with plant-derived material in total RNA extracted from nodules. Furthermore, the chip includes

29 host control genes for soybean (including the 18S rRNA), and 4 for cowpea (Hauser et al.

2007). Hybridization signals were pre-processed by employing the Gene-Chip-Operating-

System (GCOS) software from Affymetrix. Signals were scaled to a target signal intensity of

500 for all arrays. Data analysis was performed using GeneSpring GX 7.3.1 software (Agilent

Technologies, Palo Alto CA, USA). Each chip belonging to one experiment was normalized to the median. Values below 0.01 were set to 0.01. As described previously, stringent criteria were applied to minimize false positive results among the bacteroid-induced genes (Pessi et al. 2007). This included filtering on flags (present in at least two out of three of the arrays), and root filtering where the signal in bacteroids was requested to be at least two times higher

82 Rhizobial adaptation to hosts than in the control hybridizations with root-only material. The amount of plant RNA was normalized to the signal of the Glycine max 18S rRNA. Genes differentially expressed in siratro, soybean, and cowpea were selected using an ANOVA analysis with a post-hoc Tukey test. A subsequent filter (2-fold cutoff) for the fold change was applied. Differentially expressed genes were clustered hierarchically using GeneSpring 7.3.1 (Fig 3.1). The

Euclidean distance was used as distance measure between two gene expression profiles, and clusters were aggregated using the complete linkage algorithm. Functional classification of the differentially expressed genes was based on Kazusa functional gene classification for B. japonicum (Kazusa website) which comprises 14 major categories. To identify functional groups that are significantly over- or underrepresented among the differentially expressed genes, Fisher’s Exact Test was used with a p-value threshold of 0.01 (Supplemental Table

S3.4).

Proteome analyses

Notably, the extraction procedure differed between the transcriptomics and proteomics experiments. Total nodule RNA was extracted and used for GeneChip hybridizations, whereas bacteroids were extracted and fractionated prior to proteome analyses as described recently (Delmotte et al. 2010). Between 1 and 4.38 g of root nodules (which were collected at the peak of maximal nitrogenase activity) from at least 40 separately grown plants infected with B. japonicum wild-type were needed per biological replicate. Bacteroid isolation from nodules, and bacteroid fractionation were done essentially as described by Sarma and

Emerich (2005), except that cells were collected by centrifugation at higher g value (8,000xg for 10 min). Proteins were then extracted using the AllPrep DNA/RNA/Protein Mini Kit according to the instructions in the kit manual (Qiagen, Hilden, Germany). Proteins were separated on a Tris-HCl polyacrylamide gel (10.5-to-14% linear gradient, 13.3 x 8.7 cm)

83 Rhizobial adaptation to hosts obtained from Bio-Rad Laboratories (Reinach, Switzerland), and staining was performed with Coomassie blue. Gel lanes were cut in 20 pieces and destained. After reduction and carbamidomethylation, proteins were digested with trypsin (Promega, Madison WI, USA).

Digestion was quenched, and peptides were cleaned up with a C18 ZipTipTM supplied by

Millipore Corporation (Billerica MA, USA).

Peptides were separated by RP-HPLC, and mass spectrometric detection was performed with

LTQ-Orbitrap mass spectrometers (Thermo Fisher Scientific, Waltham MA, USA). The four most abundant double- or triple-charged ions from the high-accuracy survey scan with a minimum ion count of 500 were automatically taken for further MS/MS analysis at the linear ion trap. Precursor masses already taken for MS/MS were excluded for further selection for

60 sec. All mass spectra were recorded in positive ion mode with an electrospray source voltage between 1.5 kV and 1.90 kV. Precursor mass spectra were acquired at the Orbitrap mass analyzer with a scan range from m/z 300.0 to 1,600.0 using internal lock mass calibration on m/z 429.088735 and 445.120025. Resolution was set to 60,000 at m/z 400

(Delmotte et al. 2009).

Mass spectra processing was performed with Xcalibur 2.0.7 (Thermo Fisher Scientific). Peak list generation for database searches was performed with Mascot Distiller 2.1.1.0 (Matrix

Science, London, UK). Protein database was built by combining protein sequences of B. japonicum USDA 110 (8,317 protein sequences downloaded from the NCBI website) with protein sequences of Glycine max (62,199 sequences downloaded from the soybean annotation website). 258 protein sequences of usual contaminants (e.g. human keratin, trypsin) were also appended to the database. All experimental mass lists were tested against a composite version of the database, created by concatenating the target protein sequences with reversed sequences (total of 141,290 sequences, target-decoy searches) as described by Elias and Gygi (2007). Database searches were performed with Mascot 2.2 (Perkins et al. 1999)

84 Rhizobial adaptation to hosts

(Matrix Science, London, UK) and X!Tandem (Craig and Beavis 2004). Results were validated with Scaffold 2.02.01 (Proteome Software Inc., Portland OR, USA). Peptide identifications were accepted if they could be established at greater than 95% probability as specified by the Peptide Prophet algorithm (Keller et al. 2002). Probability greater than 99%

– in sense of the Protein Prophet algorithm (Nesvizhskii et al. 2003) – was required to validate protein identifications. One-hit wonders were removed (i.e., only proteins identified with at least 2 detected peptides were considered), and proteins that contained similar peptides and could not be differentiated based on MS/MS analysis alone were grouped to satisfy the principle of parsimony. The computed false discovery rate was below 0.1%. The complete list of 2,633 B. japonicum proteins that were identified during symbiosis with cowpea or with siratro or with soybean is provided as Supplemental Table S3.5. If only those proteins detected in two out of three biological samples are considered as positive hits, 2,015 proteins were identified.

Two programs from the Expasy website, TMHMMv2.0 and SignalPv3.0, were used to predict transmembrane-spanning domains and signal peptides of secreted proteins, respectively.

85 Rhizobial adaptation to hosts

3.6 Addendum: Further candidates with a perspective to act in a host- responsive manner

Results and discussion

Two operons coding for an annotated sulfonate ABC transporter (bll7007-7011, bll6450-

6455) were identified to be specifically expressed in siratro root nodules (Table 3.1).

Interestingly, the bll6450-6455 operon harbors a gene encoding an acyl-CoA dehydrogenase

(bll6452). The importance of sulfur during symbiosis in general and in particular in the Lotus japonicus - M. loti symbiosis was previously described (Krusell et al. 2005). The L. japonicus sulfate transporter SST1 was shown to be important for nitrogen fixation and plant development. Therefore it was speculated that the two identified ABC transporters were more important for the siratro host than for the other host plants of B. japonicum. One could speculate that in siratro plants sulfonate sources are more abundant than in the other host plants thus influencing gene and protein expression.

Recently Sugawara and coworkers showed that both operons (bll7007-7011, bll6450-6455) are up-regulated in B. japonicum when growing with sulfonate sources such as 4-(2- hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) and taurine (Sugawara et al. 2011).

To assess the capacity of B. japonicum to utilize sulfonate two deletion mutant strains were constructed of the genes encoding the alkanesulfonate monooxygenases bll6451 and bll7010

(Sugawara et al. 2011). Both deletion mutants were tested for their symbiotic properties; however they were not essential for nitrogen fixation with soybean (Sugawara et al. 2011).

To possibly elucidate the importance of sulfonate for the B. japonicum-siratro symbiosis, we aimed on characterizing the two annotated sulfonate transporters in further depths. Therefore, we first intended to construct mutants in each of the transport systems separately followed by the construction of a double mutant strain.

86 Rhizobial adaptation to hosts

In contrast to the work performed by Sugawara, we aimed on deleting the entire ABC transporter bll7007-7011 (Fig S3.1). The resulting deletion mutant strains were named

6230/6231 (Fig S3.1).

Fig S3.1. Genomic overview of the bll7007-7010 region. The genotype of the mutant strains 6230 and 6231 is given including the orientation of the inserted resistance cassette illustrated by horizontal arrows.

The symbiotic properties of 6230/6231 were analyzed in plant infection experiments.

Therefore soybean, siratro and cowpea plants were infected with 6230/6231 (Fig S3.2). The mutants showed wild-type-like nitrogen fixation in all host plants (Fig S3.2). This is in agreement with the results reecently published by Sugawara et al. (Sugawara et al. 2011) showing that the deletion of bll7010 does not cause any symbiotic defect on soybean. It is possible that during symbiosis the homologous sulfonate transporter system encoded by bll6450-6455 takes over the function of Bll7007-7011. Consequently, we aimed at constructing a double mutant of both transport systems. However, the attempt to replace the entire bll6450-6455 operon by an antibiotic resistance gene using double homologous cross over failed. Therefore, another strategy was employed namely to construct an insertion mutant in the gene bll6454 by inserting the pSUP202pol4 suicide vector. Even though several attempts were undertaken, such a mutant could not be constructed. In fact only co-integrates were obtained.

Interestingly, Sugawara and coworkers could successfully construct a deletion mutant in bll6451 whereas we failed at deleting the entire operon (Sugawara et al. 2011). At the

87 Rhizobial adaptation to hosts moment, we can only speculate about the reason why in this study attempts to construct a deletion mutant in bll6450-6455 or to inactivate bll6454 failed. A possible reason could be the different mutant construction strategies applied.

Fig S3.2. Symbiotic properties of B. japonicum wild type and 6231, 6230 on different host plants.

Mutant construction

To construct a deletion mutation in the bll7007-7010 gene cluster, PCR fragments of the 5´ and 3´flanking regions of the bll7007-7010 genes were amplified using the following primer pairs: bll7010-3_rev (TTTCTGCAGTCGTGCCGGTGGTGCAGAGC) and bll7010-4_for

(TCCCCCGGGCGTTTGATCGCCGCAAAGAGG) for the 5´ region; bll7010-5_rev

(TCCCCCGGGTCCAACGGAAAGTCTCAGAAAG) and bll7010-6_for (GGACTAGT-

CGCCGACCAGTTCGACGTG) for the 3´ region. PCR products were cloned in the pGEM-

T Easy vector (Promega, Madison, WI, USA), and subsequently in pSUP202pol4 vector

(Fischer et al. 1993). A 1.2 kb kanamycin resistance cassette (aphII) derived from pBSL15

(Alexeyev 1995) was introduced in both directions between the up- and downstream regions.

The resulting plasmids pRJ6230 and pRJ6231 were transferred into B. japonicum strain

110spc4 for marker replacement using previously described methods (Hahn and Hennecke

1984). The correct genomic structure was verified by PCR. The resulting deletion mutants were named 6230 and 6230 (Fig S3.1).

88 Oxalotrophy in Bradyrhizobium japonicum

CHAPTER 4

Oxalotrophy in Bradyrhizobium japonicum

4.1 Abstract

The occurrence of oxalotropic metabolism in rhizobia has not been investigated yet. In the genome of the nitrogen-fixing legume endosymbiont Bradyrhizobium japonicum, genes encoding enzymes potentially involved in the uptake and catabolism of oxalate were identified. B. japonicum possesses a putative formyl-CoA transferase (frc), an oxalyl-CoA decarboxylase (oxc) as well as two paralogous oxalate:formate antiporters (oxlT1+2) which all share substantial identity with previously characterized proteins of the oxalotrophic bacterium

Oxalobacter formigenes. Here, we show that oxalate is present in roots and root nodules of B. japonicum host plants. Global transcription profiling analysis revealed that these genes are highly expressed in free-living conditions, when B. japonicum grew in arabinose-containing medium, as well as in symbiotic conditions. The frc and oxc genes seem to be organized in an operon together with a gntR-like regulatory gene. To study the importance of oxalate degradation in B. japonicum, frc and oxc were mutated. The characterization of the ∆frc-oxc strain showed that these genes are required for oxalotrophic growth; however, they are not important for symbiotic nitrogen fixation.

Oxalotrophy in Bradyrhizobium japonicum

4.2 Introduction

Within root nodules, bacteroids reduce N2 to ammonium, which is secreted to the plant in return for dicarboxylic acids as carbon and energy sources (Lodwig and Poole 2003). C4- dicarboxylic acids like succinate and malate have been shown to be the primary carbon sources of bacteroids and can actively cross the peribacteroid membrane (Prell and Poole

2006). Rhizobial genes responsible for the transport of C4-dicarboxylic acids were shown to be essential for nitrogen fixation (Ronson et al. 1981; Ronson et al. 1984; Yurgel and Kahn

2004). C4-dicarboxylic acids are directly fed into the TCA-cycle to supply the bacteroid with enough energy to perform nitrogen fixation (Prell and Poole 2006).

However, little is known about the role of the C2-dicarboxylic acid oxalate during symbiosis.

In plants, the vast majority of oxalic acid complexes with calcium resulting in highly insoluble Ca-oxalate crystals (Franceschi and Nakata 2005). Ca-oxalate crystals can be deposited in tissues either intracellularly within vacuoles or extracellularly, there associated with cell walls (Libert and Franceschi 1987; Franceschi and Nakata 2005). Legume plants such as soybean, mungbean and cowpea also showed the presence of Ca-oxalate crystals in the root nodule cortex (Sutherland and Sprent 1984). Oxalate was also present in high concentrations in the cytosol of nodules of the legume plant Vicia faba (Trinchant et al. 1994) which had been infected with Rhizobium leguminosarum. Trinchant and coworkers (1994) hypothesized that oxalate is a potential energy-yielding substrate sufficient for nitrogen fixation in V. faba. Uptake experiments with symbiosomes of V. faba exhibiting a functional nitrogenase revealed that oxalate is taken up by the peribacteroid membrane (PBM)

(Trinchant et al. 1994). Moreover, ex planta nitrogenase activity was as efficient in bacteroids provided with oxalate as with succinate (Trinchant et al. 1994).

Bacteria able to catabolize the C2 compound oxalate as sole source of carbon are called oxalotrophs (Aragno and Schlegel 1991; Dimroth and Schink 1998; Sahin 2003). Some

Oxalotrophy in Bradyrhizobium japonicum species are "generalists" and therefore able to catabolize oxalate and other carbon compounds; other species are "specialists" and use oxalate as the sole carbon and energy source (Sahin 2003). Three enzymes were previously described to be involved in the uptake and catabolism of oxalate in many oxalotrophic bacteria. Extracellular oxalate is taken up into in the bacterial cell by the membrane-associated oxalate:formate antiporter OxlT, encoded by oxlT. Oxalate is then oxidized to CO2 and formate by a coupled two-step reaction catalyzed by the formyl-CoA transferase encoded by frc and the oxalyl-CoA decarboxylase encoded by oxc. Frc transfers the CoA moiety from formyl-CoA to oxalate followed by the decarboxylation of the activated oxalate molecule by Oxc (Lung et al. 1994; Sidhu et al.

1997). Formate can now be converted to CO2 by the action of the formate dehydrogenase.

Hereby one reduction equivalent is released which allows energy production in the respiratory chain. Alternatively, OxlT catalyzes the export of formate from the cell

(Anantharam et al. 1989; Abe et al. 1996).

So far little is known about oxalotrophy in rhizobia or a possible role of oxalate during symbiosis as a potential carbon source. The genome sequence of B. japonicum USDA110, the root endosymbiont of soybean, mungbean, cowpea and siratro, revealed that this organism possesses genes involved in the transport and metabolism of oxalate (Kaneko et al. 2002).

Among them two paralogous oxalate:formate antiporters (bll3149/bll3150), a formyl-CoA transferase (bll3156), and an oxalyl-CoA decarboxylase (bll3157) were identified. In a previous study, using the frc gene as molecular marker to identify oxalotrophic bacteria in complex ecosystems, the presence of frc in B. japonicum strain 110 was reported (Khammar et al. 2009). Our previous studies on the global expression profile of B. japonicum strain 110 showed that these genes are expressed in defined free-living and symbiotic conditions

(Hauser et al. 2007; Pessi et al. 2007). In this study, a ∆frc-oxc strain was constructed. The

Oxalotrophy in Bradyrhizobium japonicum characterization of this strain showed that B. japonicum frc and oxc are essential for growth on oxalate.

Oxalotrophy in Bradyrhizobium japonicum

4.3 Material and methods

Bacterial strains, media and growth conditions

The bacterial strains and plasmids used and generated in this work are listed in Table 4.1.

Escherichia coli cells were cultivated in Luria-Bertani medium (Miller 1972) at 37°C using the following concentrations of antibiotics (in µg/ml): ampicillin (200), kanamycin (30), chloramphenicol (20) and tetracycline (10). B. japonicum cells were routinely cultivated at

30°C on rich peptone-salts-yeast extract (PSY) medium (Regensburger and Hennecke 1983) supplemented with 0.1% arabinose or defined buffered Vincent’s minimal medium (BVM)

(Vincent 1970; Becker et al. 2004). Carbon sources used in defined media were filter- sterilized and used at a final concentration of 20 mM arabinose or 20 mM succinate. The appropriate antibiotic concentrations (in µg/ml) were added: spectinomycin (100), kanamycin

(100). Aerobic cultures for phenotypic growth analysis in PSY or BVM medium were grown in 500 ml Erlenmeyer flasks containing 25 ml medium supplemented with spectinomycin

(100 µ ml-1) and the respective C-source on a shaker (160 rpm) at 30°C. For each strain or condition, the growth of 3 independent cultures was analyzed. The results were obtained in duplicates.

Oxalotrophic growth

B. japonicum wild type and mutant strains, ∆frc-oxc and ∆oxlT1+2, were grown on double- layered Schlegel mineral medium plates supplemented with either Ca-oxalate or Na-oxalate

(Aragno and Schlegel 1991). The first layer contains Na2HPO4×12H2O, 9.0 g/l; KH2PO4, 1.5 g/l; NH4Cl, 1.0 g/l;MgSO4×7H2O, 0.2 g/l; ammonium ferric citrate, 0.005 g/l; CaCl2, 0.01 g/l;

ZnSO4×7H2O, 50 μg/l; MnCl2×4H2O, 15 μg/l; H3BO3, 150 μg/l; CoCl2×6H2O, 100 μg/l;

CuCl2×2H2O, 50 μg/l; NiCl2×6H2O, 10 μg/l; NaMoO4×2H2O, 15 μg/l; agar 15 g/l. The second layer is composed of Schlegel mineral medium to which Ca-oxalate or Na-oxalate

Oxalotrophy in Bradyrhizobium japonicum

(Sigma-Aldrich, Steinheim, Germany) at 7 g/l or 2.7 g/l was added (Aragno and Schlegel

1991). For this test, bacterial cultures of wild type and mutant strains were pre-grown in minimal medium (BVM). Cells were washed and set to an optical density at 600 nm [OD600]

= 4. A cotton swab was used to streak the cells on the plate. Plates were incubated at 30°C for

7 days. The assays were repeated three times.

Isothermal calorimetry

Solid Angle's medium (Angle et al. 1991) was supplemented with different carbon sources at a final concentration of 20 mM arabinose and 54 mM Ca-oxalate. The media was poured into sterile microcalorimetric ampoules to obtain slants. These slants were inoculated with B. japonicum wild type, ∆frc-oxc and ∆oxlT1+2 strains using a loop with an inoculum size sufficient to grow bacteria as a lawn. The ampoules were then sealed and introduced in the microcalorimeter (TAM48, Waters/TA, Delaware). After the thermal equilibration procedure, measurements were taken for at least 7 days. The data were recorded by the microcalorimeter at a constant time interval of 5 minutes. Samples were removed from the microcalorimeter, and visually inspected in order to check that bacteria did form a lawn. Net growth rate (µ) was calculated using the integrated heat-flow data (i.e., heat over time curve) by fitting the

µt equation Nt = N0 • e on these data. Similarly, maximum growth rate (µmax) was calculated using the Richards equation (Zwietering et al. 1990). All calculations were performed using the statistical software R and the grofit package (Kahm et al. 2010). Finally, the maximum activity in µW (i.e., µJ/s) was used to calculate the maximum substrate consumption rates assuming the following equations and ∆G°': oxalate respiration: 2C2H2O4 + O2 → 4CO2 +

2H2O (∆G°' = ̶ 328 KJ/mol); arabinose respiration: C5H10O5 + 5O2 → 5CO2 + 5H2O (∆G°'

= ̶ 2413 KJ/mol).

Oxalotrophy in Bradyrhizobium japonicum

DNA methods and construction of mutant strains

DNA was isolated from B. japonicum wild-type strain 110spc4 as previously described

(Hahn and Hennecke 1984). Plasmid DNA from E. coli strains was obtained by using the

NucleoSpin® Plasmid kit (Macherey-Nagel, Düren, Germany). Recombinant DNA work was performed according to standard protocols (Sambrook and Russel 2001). Mutagenesis of genes was done by marker replacement. To construct deletion mutations in the bll3156-3157 and bll3149-3150 genes, PCR fragments of the 5ˈand 3ˈflanking regions of bll3156-3157 and bll3149-3150 were amplified using the following primer pairs: bll3156-1_rev (5´-

TACGGCTGGCGTCCGGCGAAC-3´) and bll3156-2_for (5´-GTCTCTGGTCGAACCGC-

TTACTCGGCG-3´) for the 3ˈ region of bll3156; bll3156-3_for (5´-TGTCTCCCTGGTC-

TCAATACTG-3´) and bll3156-4_rev (5´-GTCGAGCATCACCGGCTCG-3´) for the 5ˈ region of bll3157; bll3149-3_rev (5´-GGCGCCGTAGATGCAGATCTTC-3´) and bll3149-

4_for (5´-CATTATTGACGTTTGCGGTGCC-3´) for the 3ˈ region of bll3149 and bll3149-

5_for (5´-GCACCATGTCCGTCATAAAG-3´) and bll3149-6_rev (5´-CAACTGGCTT-

AAACGGTCGCC-3´) for the 5ˈ region of bll3150, respectively. PCR products were cloned in the pGEM-T Easy vector (Promega, Madison, WI, USA) and the correct sequence was verified by sequence analysis. Up- and downstream regions were subcloned into the pSUP202pol4 vector (Fischer et al. 1993) and a kanamycin resistance cassette (aphII) derived from pBSL15 (Alexeyev 1995) was inserted in between both regions. The resulting plasmids pRJ6243, pRJ6263 were used for conjugation with B. japonicum strain 110spc4.

The correct genomic integration was verified by PCR. The resulting deletion mutants were named 6243 and 6263 (Fig 4.1). For further information on all used strains see Table 4.1.

Oxalotrophy in Bradyrhizobium japonicum

Table 4.1. Bacterial strains and plasmids used in this work.

Strain or plasmid Relevant genotype or phenotype Source of reference Strains E. coli DH5α supE44 ΔlacU169-(Ψ80lacZΔM15)hsdR1 recA1 gyrA96 thi- Bethesda Research 1 relA1 Laboratories, Inc., Gaithersburg, U.S.A. S17-1 Smr Spr; hdsR (RP4-2 kan::Tn7 tet::Mu, integrated in the Simon et al. 1983 chromosome) B. japonicum strains Regensburger and Hennecke 110spc4 Spr; wild type 1983

6243 Spr Kmr; ∆bll3156-3157::aphII (same orientation) This work

6263 Spr Kmr; ∆bll3149-3150::aphII (same orientation) This work

Plasmids pGEM-T Easy Apr; cloning vector Promega Corporation, Madison, U.S.A. pBSL15 Apr Kmr; cloning vector Alexeyev 1995 pSUP202pol4 Tcr; (pSUP202) part of the polylinker from pBlueskript II Fischer et al. 1993 KS+ between EcoRI and PstI pRJ6240 Apr; (pGEM-T Easy) 623 bp amplicon of the bll3156 This work downstream region comprising the B. japonicum genome sequence from coordinate 3489814 to 3490437 pRJ6241 Apr; (pGEM-T Easy) 636 bp amplicon of the bll3157 This work upstream region comprising the B. japonicum genome sequence from coordinate 3493519 to 3494155 pRJ6260 Apr; (pGEM-T Easy) 648 bp amplicon of the bll3149 This work downstream region comprising the B. japonicum genome sequence from coordinate 3481836 to 3482484 pRJ6261 Apr; (pGEM-T Easy) 637 bp amplicon of the bll3150 This work upstream region comprising the B. japonicum genome sequence from coordinate 3485435 to 3486072 pRJ6262 Tcr; (pSUP202pol4) 1285 bp XbaI/PstI fragment containing This work the up- and downstream region of the bll3149-bll3150 region pRJ6242 Tcr; (pSUP202pol4) 1259 bp EcoRI/NotI up- and This work downstream region of bll3156-bll3157 pRJ6243 Tcr Kmr; (pSUP202pol4) the aphII cassette flanked by the This work up-and downstream region of bll3156-bll3157

Oxalotrophy in Bradyrhizobium japonicum pRJ6263 Tcr Kmr; (pSUP202pol4) the aphII cassette flanked by the This work up- and downstream region of bll3149-bll3150

Symbiotic growth analysis: Plant material, inoculation and cultivation

Sterilization of soybean (Glycine max [L.] Merr. cv. Williams), mungbean (Vigna radiata), cowpea (Vigna unguiculata) and siratro (Macroptilium atropurpureum) seeds was done as described earlier (Koch et al. 2010). Plants were inoculated with cultures of B. japonicum that had been grown for five days in PSY medium, and diluted to approximately 102 bacteria per plant. Nitrogenase activity was determined by using the acetylene reduction assay 21 days post infection (DPI) for soybean, mungbean and cowpea and 31 DPI for siratro. Bacteria were isolated from randomly selected nodules to confirm the presence of appropriate genetic markers.

Determination of the oxalate content of roots and nodules of soybean, mungbean, cowpea and sirato infected with B. japonicum

Root nodules of soybean, mungbean, cowpea and sirato infected with B. japonicum as well as root-only material were collected and immediately flash-freezed in liquid nitrogen and stored at −80°C until use. To assess the soluble oxalate content in nodules, 50 mg of nodule material was homogenized in 200 μL dH2O using a TissueLyzer (Qiagen, Valencia, CA, U.S.A.) (1.3 min at 30 Hertz) with a tungsten carbide bead (3 mm; Qiagen). Hundred mg of root material were ground in liquid nitrogen using a mortar and a pestle. The resulting powder was collected and resuspended in 200 μL of dH2O. To ensure an optimal destruction, the material was further homogenized in a TissueLyzer (1.3 min at 30 Hz). The supernatant of nodule and root samples was collected by a low centrifugation step. The oxalate concentration in each sample was determined using the Urinalysis Diagnostic Kit (TrinityBiotech) according to the

Oxalotrophy in Bradyrhizobium japonicum manufacturer's instructions. Measurements were done in triplicate on three independent plants.

Oxalotrophy in Bradyrhizobium japonicum

4.4 Results

Identification and transcriptional analysis of the frc, oxc genomic region

According to the genome annotation performed by Kaneko and coworkers

(http://bacteria.kazusa.or.jp/rhizobase/), B. japonicum strain 110 harbors genes involved in the transport and metabolism of oxalate (Kaneko et al. 2002). These genes cluster in close vicinity in the B. jaaponicum genome (Fig 4.1).

Fig 4.1. Physical map of the B. japonicum region harboring several genes of the oxalate metabolism. Bll3149 and bll3150 encode an oxalate:formate antiporter (OxlT1+2), bll3156 codes for a formyl CoA-transferase (Frc), and bll3157 for an oxalyl CoA-decarboxylase (Oxc). Upstream of frc and oxc the regulatory gene bll3158 is located. The structure of the B. japonicum mutants are indicated with arrows along with the strain number. Numbers below the vertical lines represent genome coordinates. The map is drawn to scale.

Among them, genes encoding a putative oxalate:formate antiport system (oxlT1+2)

(bll3149/bll3150) as well as a formyl-CoA transferase (frc) (bll3156) and an oxalyl-CoA decarboxylase (oxc) (bll3157) were identified. The products of the aforementioned genes resemble previously characterized proteins: Bll3149/Bll3150 are predicted membrane- spanning proteins sharing 55% and 53% amino acid sequence identity with OxlT of

Oxalobacter formigenes (Ruan et al. 1992) which has been shown to mediate the exchange of oxalate:formate in the cytoplasma. Bll3156 and Bll3157 are predicted cytoplasmic proteins and share 71% and 78% amino acid sequence identity with the previously studied Frc and

Oxc of O. formigenes (Lung et al. 1994; Sidhu et al. 1997). Upstream of oxc the regulatory gene bll3158 is located. Bll3158 shares 26% similarity with the previously identified GntR regulator of Bacillus subtilis which was shown to be a repressor for the gluconate degradation

100

Oxalotrophy in Bradyrhizobium japonicum operon (Hoskisson and Rigali 2009). Analysis of the genomic organization of several oxalotrophic bacteria revealed that only B. japonicum possess a regulatory gene upstream of the oxalate catabolism genes (Abratt and Reid 2010). In order to investigate the frc and oxc expresssion profile of B. japonicum grown in symbiosis, and in minimal (BVM) and complex

(PSY) medium supplemented with L-arabinose as sole source of carbon, previously obtained microarray results were analyzed (Pessi et al. 2007). The so-called tiling analysis showed that frc and oxc are highly expressed under symbiotic and free-living conditions and revealed that these genes are most likely organized in an operon together with bll3158 (Fig 4.2).

Fig 4.2. Transcript analysis of the frc, oxc genomic region. Data illustrated in this graph are derived from individual probe pairs of a microarray experiment from B. japonicum wild type strain grown under different conditions (symbiosis, rich medium and minimal medium supplemented with L-arabinose). Genome coordinates are given on the horizontal axis. For illustration purpose data points were connected by a line. Genes are shown as arrows. frc stands for formyl-CoA transferase and oxc stands for oxalyl-CoA decarboxylase.

101

Oxalotrophy in Bradyrhizobium japonicum

Construction of ∆oxlT1+2 and ∆frc-oxc deletion mutants

To study the function of genes involved in the transport and metabolism of oxalate in B. japonicum, two mutant strains were constructed, strain 6263 (∆oxlT1+2) and strain 6243 (∆frc- oxc). In both cases the deleted genes were replaced by a kanamycin resistance cassette using double homologous recombination. In the ∆frc-oxc mutant 9 nt of the 3ˈend of bll3156 and

12 nt of the 5ˈend of bll3157 were kept, whereas in 6263 the bll3149 gene was entirely removed and 18 nt of the 5' end of bll3150 remained (Table 4.1 and Fig 4.1).

Free-living growth analysis in presence of various carbon sources

Growth of B. japonicum wild type and mutant strains was analyzed in minimal (BVM) and complex (PSY) medium. The minimal medium was supplemented with either 20 mM L- arabinose, or 20 mM succinate, or 20 mM Na-oxalate. L-arabinose and succinate have previously been shown to be good C-sources for B. japonicum (Pedrosa and Zancan 1974;

Prell and Poole 2006). The ∆frc-oxc mutant strain is affected in growth as compared to the wild type when cultivated in minimal medium supplemented with L-arabinose (Fig 4.3A).

The mean generation time of the parental strain was 23.4 h compared to 37.3 h for the ∆frc- oxc mutant. Moreover wild-type cells reached a final optical density of 5.6 while the mutant strain did not exceed 1.9 (Fig 4.3A). The ∆oxlT1+2 strain in contrast showed no growth alteration when cultivated in minimal medium containing L-arabinose (Fig 4.3A).

Interestingly, the strains ∆frc-oxc and ∆oxlT1+2 were growing similarly as the wild-type strain when cultivated in minimal medium with the dicarboxylic acid succinate (Fig 4.3B). No growth difference for wild type and mutant strains ∆frc-oxc and ∆oxlT1+2 was observed when cultivated with L-arabinose in complex PSY medium (data not shown).

102

Oxalotrophy in Bradyrhizobium japonicum

Fig 4.3 Legend see next page.

103

Oxalotrophy in Bradyrhizobium japonicum

Fig 4.3. Growth analysis of B. japonicum wild type and ∆frc-oxc, ∆oxlT1+2 mutant stains in BVM minimal medium supplemented with various carbon sources A. 20 mM of L-arabinose, B. 20 mM succinate. Shown are mean values for all taken time points ± standard deviations. In panel C, oxalotrophic growth of B. japonicum wild type and mutant cells was tested on plates containing Schlegel`s minimal medium supplemented with 20 mM Na-oxalate or 20 mM of L-arabinose (control). D/E. Microcalorimetric thermograms of B. japonicum wild type and mutant strains grown on either 20 mM L-arabinose (D) or 35 mM Ca-oxalate (E).

In order to prove whether B. japonicum is an oxalotrophic bacterium, we compared the ability of wild type, ∆frc-oxc, and ∆oxlT1+2 to catabolize oxalate using plate assays and calorimetry.

We initially aimed to monitor oxalotrophic growth in liquid cultures using Ca- or Na-oxalate as sole source of carbon. Even though various concentrations (5, 10, 20 mM) of oxalate were used these cultivation attempts failed. Therefore we examined growth on Schlegel`s minimal medium agar plates in the presence of 20 mM Na-oxalate (Fig 4.3C) or 35 mM Ca-oxalate as sole carbon source (data not shown). For this test, strains were first grown in minimal medium and then streaked out on agar plates. No growth difference between wild-type and mutant strains was observed on control medium plates supplemented with 20 mM of L- arabinose. This test revealed that B. japonicum wild-type strain is able to grow on plates supplemented with 20 mM Na-oxalate (Fig 4.3C). Growth was also observed for the ∆oxlT1+2 strain which suggests that the lack of OxlT1+2 allows B. japonicum to bypass the uptake of oxalate, possibly using other transport systems. Notably, no growth was displayed for the

∆frc-oxc mutant. Thus, these genes seem to be required for oxalotrophic growth.

For calorimetric analysis, strains were inoculated on Angle`s media supplemented with either

20 mM L-arabinose or 35 mM Ca-oxalate as described in the Material and Methods section.

Microcalorimetric thermograms of the total heat generated by B. japonicum wild type,

∆oxlT1+2, and ∆frc-oxc mutant showed variations depending on the substrate used (Fig

4.3D/E). A maximum heat-flow was detected for L-arabinose. As already shown in liquid cultures supplemented with L-arabinose, the growth rate for wild type and ∆oxlT1+2 is similar whereas the ∆frc-oxc mutant is strongly inhibited in growth (Fig 4.3D). Oxalotrophic growth of B. japonicum wild type and ∆oxlT1+2 mutant was confirmed by microcalorimetry (Fig

104

Oxalotrophy in Bradyrhizobium japonicum

4.3E). Figure 4.3E shows a heat detection delay for the ∆oxlT1+2 strain in comparison to the parental strain (Fig 4.3E). In contrast, no heat production was measured for ∆frc-oxc suggesting that this strain was unable to catabolize Ca-oxalate (Fig 4.3D) which is in agreement with the results obtained by the plate tests. Growth detection varied according to the substrate (Fig 4.3D/E). On L-arabinose, growth (and related heat production) was immediately detected after the experimental set up. In contrast, growth on Ca-oxalate started after approximately 30 h. This could be due to the insolubility of Ca-oxalate which is less diffusible than L-arabinose and thus less accessible for the cells.

Oxalate content in roots and root nodules induced by B. japonicum

The oxalate content of roots and root nodules of B. japonicum hosts was analyzed (Table

4.2). Root and root-nodule material was collected, homogenized and immediately used for measurements of soluble oxalate level. The concentration of oxalate was determined enzymatically using an oxalate oxidase assay. In root nodules the amount of oxalate varies depending on the host plant (Table 4.2). The highest concentration of oxalate was measured in the two Vigna species mungbean and cowpea with 0.37 mg g-1 wet weight (WW) and 0.31 mg g-1 WW respectively (Table 4.2). In contrast, in soybean nodules 0.18 mg g-1 WW oxalate was measured corresponding to half the amount detected in the Vigna species (Table 4.2).

Generally, three times higher oxalate concentrations were measured in root nodules than in roots of these plants.

Table 4.2. Oxalate content of root nodules and root material of cowpea, mungbean, siratro and soybean plants infected with B. japonicum wild type strain. Results are presented as mean ± standard deviations with N = at least 2.

Oxalate mg g-1 wet weight Host plant Root Root nodules Cowpea 0.10 ± 0.01 0.37 ± 0.08 Mungbean 0.09 ± 0.01 0.31 ± 0.05 Siratro 0.09 ± 0.01 0.22 ± 0.01 Soybean 0.03 ± 0.01 0.18 ± 0.02

105

Oxalotrophy in Bradyrhizobium japonicum

Symbiotic properties of the ∆frc-oxc and ∆oxlT1+2 mutants

The fact that B. japonicum host plants displayed different concentrations of soluble oxalate prompted us to investigate the symbiotic efficiency of ∆frc-oxc (6243), ∆oxlT1+2 (6263) on all hosts to possibly monitor a host plant-dependent phenotype.

-1 -1 Table 4.3. Specific nitrogenase activity (% C2H4 min g ) of B. japonicum wild type and 6243, 6263 mutants on soybean, mungbean, cowpea and siratro plants. In total two independent plant infection experiments were performed. At least six plants were measured per strain. Shown are mean values ± standard deviations.

Relevant genotype Soybean Mungbean Cowpea Siratro

110spc4 Wild type 1.9 ± 0.7 1.9 ± 0.9 2.5 ± 0.7 1.7 ± 0.8 6243 ∆frc-oxc 1.7 ± 1.2 2.2 ± 0.8 2.9 ± 1.7 1.2 ± 0.7 6263 ∆oxlT1+2 1.8 ± 0.6 1.6 ± 0.2 3.9 ± 1.5 1.5 ± 0.3

Plants infected with ∆frc-oxc and ∆oxlT1+2 grew as healthy as wild-type plants displaying green leaves. Nitrogenase activity was assayed on root nodules using the acetylene reduction test 21 DPI for soybean, mungbean and cowpea and 31 DPI for siratro. No alteration in the ability to develop nodules or to fix nitrogen (Table 4.3) was observed. Moreover, the same number of nodules was induced by ∆frc-oxc and ∆oxlT1+2 as compared to the parental strain

(data not shown). Re-isolation of bacteroids from plants infected with B. japonicum wild type and ∆frc-oxc revealed comparable viable cell counts for wild type and mutant stain (Table

4.4).

Table 4.4. Viable cell counts of B. japonicum wild type and ∆frc-oxc strain reisolated from soybean, mungbean, cowpea and siratro root nodules which were previously infected with 102 bacteria per plant. In total two independent experiments were performed. At least three plants were analyzed per strain. Shown are mean values ± standard deviations.

Relevant genotype Soybean Mungbean Cowpea Siratro

110spc4 Wild type 2.5 x 106 ± 3.7 x 4.8 x 106 ± 9.8 x 1.8 x 106 ± 3 x 2 x 106 ± 2.6 x 105 105 105 105 6243 ∆frc-oxc 1.4 x 105 ± 1.1 x 1.2 x 106 ± 1.3 x 3 x 106 ± 5.1 x 9.8 x 105 ± 1.3 x 104 105 105 104

106

Oxalotrophy in Bradyrhizobium japonicum

4.5 Discussion

This study revealed, for the first time, that a rhizobial species such as B. japonicum is able to use the C2-diarboxylic acid oxalate as sole source of carbon and has, therefore, to be considered an oxalotrophic bacterium. This finding expands our knowledge about carbon metabolism in rhizobia. It was earlier reported that rhizobia possess a broad metabolic capacity and can use various carbon sources as electron donors for growth (Stowers 1985).

Previously it was hypothesized that an oligotrophic lifestyle is the most successful strategy to survive in the rhizosphere since free-living rhizobia face a great diversity of carbon sources derived from root exudates and soil environment (Prell and Poole 2006). Since oxalate is present in soil either derived from fungi, decomposing plant material, or root exudates, it is not surprising that oxalotrophy was previously reported for several other soil bacteria (Sahin

2003; Khammar et al. 2009). A previous report also showed that oxalate was present in high concentration in root nodules induced by R. leguminosarum suggesting a function of this C2 compound during symbiosis. This hypothesis was supported by the finding that oxalate can be taken up by the peribacteroid membrane of V. faba, and symbiosomes of this legume were as efficient fixing nitrogen in the presence of oxalate as with succinate (Trinchant et al.

1994). Moreover, Ca-oxalate crystals were identified in the root nodule cortex of legume plants like soybean, mungbean and cowpea using electron microscopy (Sutherland and Sprent

1984).

The availability of the B. japonicum genome sequence and previous gene expression studies revealed that three genes known to be essential for the oxalate metabolism and transport frc, oxc and oxlT are present on the genome and highly expressed during symbiosis (Pessi et al.

2007). In addition, Frc and Oxc have been detected in soybean, cowpea and siratro bacteroids using a global proteomic approach (Delmotte et al. 2010; Koch et al. 2010). In this paper, we showed that oxalate is present in roots and root nodules of B. japonicum host plants. Thus,

107

Oxalotrophy in Bradyrhizobium japonicum

the expression of frc, oxc and oxlT1+2 during symbiosis could be explained with the fact that once symbiotically associated, bacteroids face oxalate.

Previous global gene expression analysis also showed that frc, and oxc are highly expressed in free-living cells grown in minimal medium and complex medium containing L-arabinose as carbon source. This result was surprising for us taking into consideration that oxalate was neither present in rich nor in minimal medium. In line with this result, the phenotypic characterization of the ∆frc-oxc strain grown in minimal medium containing L-arabinose revealed that the mutant strain is defective in growth as compared to the parental strain.

Interestingly, the degradation of L-arabinose could potentially be linked to the oxalate metabolism (Fig 4.4). L-arabinose is degraded as previously described for slow-growing rhizobia (Fig 4.4) (Stowers 1985). So far, in B. japonicum, genes responsible for the catabolism of arabinose are unknown. Genes encoding the L-arabinose 1-dehydrogenase and the L-arabinolactonase were identified in the nitrogen fixing α-proteobacterium Azospirillum brasiliense (Watanabe et al. 2006a; Watanabe et al. 2006b). Azospirillum brasiliense

BAD95974 and BEA94275 genes are encoding the L-arabinose 1-dehydrogenase and the L- arabinolactonase respectively. Amino acid sequence analysis revealed that BAD95974 shares

55% identity with the B. japonicum gene product Blr3205, and BEA94275 33% with

Blr3207. blr3205 and blr3207 genes are expressed in minimal medium supplemented with L- arabinose suggesting a possible involvement in the degradation of this carbon source in B. japonicum. It was biochemically shown that arabinose is further processed to the L-2-Keto-3- deoxyarabonate (L-KDA) intermediate using the L-KDA aldolase to yield pyruvate and glycolaldehyde (Fig 4.4) (Novick and Tyler 1982; Watanabe et al. 2006a; Watanabe et al.

2006b). Pyruvate could now directly be fed into the TCA cycle. Glycolaldehyde could be oxidized to glyoxylate and potentially enter three different metabolic pathways (Fig 4.4): (a) the glyoxylate cycle, (b) could be oxidized to oxalate, or (c) get reduced to glycerate by the

108

Oxalotrophy in Bradyrhizobium japonicum action of the glyoxylate carboligase and the tartronate semialdehyde reductase. In case (b) oxalate could be oxidized to formate. Formate in turn, by the action of the formate dehydrogenase, could be oxidized to carbon dioxide. Hereby one reduction equivalent is obtained which could be further used in the oxidative phosphorylation to yield energy in the form of ATP. This suggests that the degradation of L-arabinose creates byproducts which could be fueled into the oxalate degradation pathway (Fig 4.4). String analysis showed that frc and oxc are conserved among slow-growing rhizobial strains such as B. japonicum. A possible reason why orthologs of genes important for oxalotrophic life were not identified in fast-growing rhizobia like Sinorhizobium meliloti could be that fast- and slow-growing rhizobia possess different L-arabinose degradation pathways (Pedrosa and Zancan 1974).

Here, we show that in contrast to growth on L-arabinose, B. japonicum ∆frc-oxc strain was able to efficiently use succinate as carbon source. This C4-dicarboxylate was previously shown to be the major carbon source during symbiosis (Stowers 1985; Prell and Poole 2006).

The oxalotrophic lifestyle of B. japonicum was demonstrated using a plate assay as well as by calorimetry (Fig 4.3C/E). Cultivation attempts of B. japonicum wild type and mutant strains in liquid culture supplemented with either Ca- or Na-oxalate failed. It is possible that free- oxalate ions even in small amounts are toxic for B. japonicum. Since the plate assay does not give quantitative data, we were using a microcalorimetric approach to monitor oxalotrophic growth of B. japonicum strains. Thus, we were able to demonstrate that frc and oxc are essential for B. japonicum to metabolize oxalate as sole source of carbon (Fig 4.3E). The

∆oxlT1+2 strain was shown to be able to consume oxalate. The lack of OxlT1+2 seems to enable

B. japonicum to bypass the uptake of oxalate to other transport systems. However, growth

(and related heat production) of the ∆oxlT1+2 strain was delayed compared to the parental strain (Fig 4.3E). It can be speculated that the mutant strain needs time to be able to overcome the defect in oxalate uptake.

109

Oxalotrophy in Bradyrhizobium japonicum

Fig 4.4: Proposed pathway for the catabolism of L-arabinose and its possible link to the oxalate degradation pathway in B. japonicum. Possible metabolic routes for glyoxylate are given (a) the glyoxylate cycle, (b) via the glyoxylate oxidase, and (c) via the glyoxylate carboligase. The figure was conceptually adapted from Watanabe et al. (2006b). Gene names were assigned to the enzymes according to amino acid sequence similarity to previously identified enzymes.

Transcriptional analysis revealed that the frc and oxc genes are transcribed in an operon with a gene encoding a regulator of the GntR family. The members of this family of transcriptional regulators were shown to be repressors in response to various metabolites. A GntR-like

110

Oxalotrophy in Bradyrhizobium japonicum regulator was first identified and studied in B. subtilis and shown to be a repressor of gluconate degradation operon (Hoskisson and Rigali 2009). A closer look into gene positions in several oxalotrophic bacteria revealed that B. japonicum uniquely harbors a regulatory gene adjacent to the frc and oxc genes (Abratt and Reid 2010). A recent study which analyzed the response to oxalate in Moorella thermoacetica showed that three proteins respond to this stimulus (Pierce et al. 2010). Interestingly those candidates are organized in an operon that includes amongst others an oxalate:formate antiporter, as well as GntR-like regulator (Pierce et al. 2010). The fact that GntR-like regulators were identified in close vicinity to genes involved in the transport and catabolism of oxalate suggests that they might be involved in the regulation of oxalate metabolism in these organisms. These hints are of interest since little is known about the regulatory mechanism of oxalate uptake and degradation in prokaryotes.

The ∆frc-oxc and ∆oxlT1+2 strains are able to successfully initiate and establish a functional symbiosis with mungbean, cowpea, soybean and siratro. Preliminary results show that Frc and Oxc are required for the ability to competitively nodulate host plants (see Table S4.1). It is still not known at which step during the infection and nodulation process oxalate catabolism is of advantage for B. japonicum. Defective nodule occupancy was previously reported for mutants of genes for the catabolism of secondary carbon sources such as myo- inositol and proline (Jimenezzurdo et al. 1995; Fry et al. 2001; Kohler et al. 2010).

Complementation analyses are however required to exclude the impact of the antibiotic resistance cassette (in strain ∆frc-oxc) on the nodulation competition phenotype.

In summary, in this work we expand the knowledge about carbon metabolism in rhizobia.

Here, we describe genes essential for the oxalotrophic lifestyle in B. japonicum. Elucidating the pathway that enables cells to catabolize oxalate and possibly determine its regulation remain major objectives.

111

Oxalotrophy in Bradyrhizobium japonicum

4.6 Addendum: Competitiveness in symbiosis

Table S4.1. Competitiveness in symbiosis. Soybean, mungbean, cowpea and siratro plants were infected with a mixture of wild type and ∆frc-oxc strains containing a total of 102 bacteria per plant. Cultures of wild-type and ∆frc-oxc were grown and diluted to the same colony forming units (cfu) per milliliter. The following ratios were chosen to assess symbiotic competition: 90:10 (∆frc-oxc:wt), 50:50 (∆frc-oxc:wt). Serial dilutions of the mixed inoculum were plated on selective agar to control the number of inoculated cells. All nodules from one plant were harvested at the peak of nitrogenase activity, 21 (soybean, mungbean, and cowpea) and 31 (siratro) DPI. Nodules were surface sterilized (100% EtOH for 5 minutes) and rinsed in sterile distilled water. Nodules were then crushed in 1 ml PSY using a mortar and this suspension was serial diluted and spotted on plates containing the appropriate selection for strain differentiation. The plates were incubated for 4 days at 30°C and the ratio of the mutant strain in nodule extracts was determined and compared to the initial inoculum ratio. Per ratio and per host, at least 3 independent plants were processed. The nodule extracts were spotted in duplicates. Data were evaluated for statistical significance using a student’s t-test and the SPSS 17.0 software.

Ratio Soybean Mungbean Cowpea Siratro wt ∆frc-oxc wt ∆frc-oxc wt ∆frc-oxc wt ∆frc-oxc 90:10 (∆frc-oxc:wt) 80 ± 11 20 ± 11 87 ± 7 13 ± 7 89 ± 8 11 ± 8 70 ± 14 30 ± 14 50:50 (∆frc-oxc:wt) 93 ± 3 7 ± 3 83 ± 2 17 ± 2 78 ± 10 22 ± 10 85 ± 0 15 ± 0

112

Future perspectives

CHAPTER 5

Future perspectives

113

Future perspectives

5.1. Characterization of two β-class carbonic anhydrase (CA) genes

The aim of this work was to identify novel genes important for symbiotic nitrogen fixation.

Global expression studies using microarray technology were used to gain insight into global gene expression patterns of B. japonicum bacteroids isolated from soybean root nodules

(Pessi et al. 2007). In this work, the function of two β-class CAs (bll2065, bll4865), one of which was shown to be highly up-regulated during symbiosis, was studied. The characterization of two mutant strains showed that each of them by itself is not important for symbiosis.

However, only the ability to fix nitrogen in symbiosis was investigated. Phenotypic characterization of mutant strains now ought to be expanded by determining the total C- and

N- content of soybean plants that were previously infected by B. japonicum wild type and mutant strains 6224, 6225 (∆bll2065-2066), 6226 (bll4865::pRJ6226) (Kalloniati et al. 2009).

Additionally, the ultrastructure of root nodules induced by wild type and mutant strains could be analyzed using electron microscopy. Furthermore, soybean plants could be co-inoculated with wild type and mutant strains in order to investigate the nodulation competiveness in soybean.

Additionally, attention should be addressed to free-living phenotypic characterizations which are missing in this study. Since it was previously shown that CAs are involved in pH regulation (Kalloniati et al. 2009; Kaur et al. 2009), the growth behavior in minimal medium as well as the pH of the culture could be investigated in wild type and mutant strains. These analyses could be supported by reporter gene fusions.

A major challenge in this project is the fact that although only two CAs have been detected in the proteome of soybean bacteroids (Delmotte et al. 2010), a total of four β-class and one α- class CA are encoded in the B. japonicum genome. Therefore, these enzymes could potentially replace each other functionally. In order to study the role of CAs during

114

Future perspectives symbiosis, it is probably not sufficient to delete only one of the four genes. Yet, our attempt to construct a double mutant in two β-class CA genes (bll2065, bll4865) failed. To elucidate whether the attempt to construct a double mutant failed due to lethality, a conditional mutant of bll4865 in a 6224 and 6225 mutant (∆bll2065-2066) background could be constructed.

Thus, a plasmid would have to be mobilized into the mutant strains encoding a functional copy of bll4865. The chromosomal copy of bll4865 could then be deleted by marker replacement. Monitoring the presence of the marker gene on the plasmid over various generations would indicate a selective pressure to maintain the bll4865 gene.

Moreover, with the aim to construct a conditional knockout mutant of bll4365 in a 6224 and

6225 mutant background, one could consider using an inducible gene regulation system. In

Bacillus subtilis a tetracycline inducible gene regulation system was successfully constructed to generate conditional mutants. The method is based on the tetracycline repressor (TetR) and the so-called reverse TetR (revTetR) mutant (Kamionka et al. 2005). While TetR mediated gene regulation requires the presence of an inducer in this case anhydrotetracycline for gene expression, reversible silencing of genes is facilitated by revTetR which allows controllable gene expression upon the addition of anhydrotetracycline. Notably, such a system is so far not available for B. japonicum.

One conclusion we can draw regarding global expression studies is that strong expression does not guarantee importance of a gene. For example, high expression of bll2065 in bacteroids was no guarantee for the importance of the encoded CA in nitrogen fixation. This is in line with a global study performed in bacteroids of Rhizobium leguminosarum

(Karunakaran et al. 2009). Characterization of mutants in 37 genes that were shown to be highly up-regulated in bacteroids as compared to free-living cells (FC > 6 to 100) showed that none of them was essential for nitrogen fixation (Karunakaran et al. 2009). As an alternative method of choice, large scale functional screens with a clear read out could be

115

Future perspectives performed. An example is signature-tagged random mutagenesis in order to identify genes involved in nitrogen fixation which was already successfully performed in R. leguminosarum and Sinorhizobium meliloti (Pobigaylo et al. 2006; Karunakaran et al. 2009).

5.2 Host-specific adaptation

In the present work we investigated the change in gene and protein expression of B. japonicum according to the host environment. This resulted in the identification of an ABC type transport system specifically expressed in the siratro host. The respective mutant had a symbiotic defect on siratro rather than on soybean or cowpea.

In order to provide further proof for the concept of host-specific adaptation, it would be desirable to characterize more candidates acting in a host specific manner. Promising candidates are genes encoding key enzymes in the degradation of tyrosine (bll0339 and bll0343) that were identified to be specific for the cowpea host as well as genes encoding two transketolases (bll2168-2169), which act in the pentose phosphate pathway and were specifically expressed in the soybean host. In parallel with the construction of mutant strains, the metabolome in the different host plant nodules could be assessed to possibly monitor changes in the abundance of tyrosine or pentoses depending on the host plant. One could speculate that tyrosine or pentoses are more abundant in the cowpea or soybean host, respectively, thus leading to the induction of gene/protein expression of a related pathway in

B. japonicum during symbiosis with these host plants.

Other promising candidates to function in a host-responsive manner in the siratro host are the two annotated sulfonate ABC transporters (bll7007-7011, bll6450-6455). So far, mutant analysis of ∆bll7007-7011 revealed that these genes are not important for symbiotic nitrogen fixation. Thus, it was speculated that the homologous sulfonate transporter system encoded by bll6450-6455 could functionally replace Bll7007-7011 during symbioses. Recently

116

Future perspectives

Sugawara and coworkers could successfully construct a deletion mutant in bll6451

(Sugawara et al. 2011). One possibility to obtain the desired double mutant strain in both annotated sulfonate transporters would be by using the plasmid pRJ6230/6231 to introduce the bll7007-7011 deletion into strain ∆bll6451. The resulting double mutant needs then to be analyzed in plant infection experiments with siratro plants.

Moreover, several new bioinformatics analyses with the obtained transcriptomics as well as proteomics data could be performed. Throughout this study attention was paid only to genes of B. japonicum that were specifically up-regulated in one host plant. Next, it would be interesting to investigate genes which are specifically down-regulated in one host plant as compared to the others. Since a single high-quality MS spectrum of a peptide is sufficient to consider a protein as being identified, the consideration of one-hit wonders would expand the list of proteins identified in bacteroids of B. japonicum. Moreover, in order to compare protein expression in different samples, a quantitative analysis of the proteome data is needed

(Ong and Mann 2005).

Concerning the siratro host-specific ABC transporter encoded by blr1601-1604, more experiments are needed to identify the substrate it transports. Several strategies could be envisaged such as isolating the periplasmic binding protein Blr1601 directly from bacteroids of siratro nodules with the hope that the unknown substrate will still be bound. Taking into account that enormous quantities of bacteroids are needed to purify Blr1601 in sufficient amounts to perform this experiment, another option would be to express and purify Blr1601 in vitro and use siratro root nodule extracts for binding assays. In both cases the unknown substrate could be analyzed by MS/MS analysis. Alternatively a computational approach

(Matsusaki et al. 2002) could be used to predict substrate specificities of membrane transporter proteins from their amino acid sequence. Once the substrate is known, transport assays could be performed (Dupont et al. 2004). The specificity of the transport could be

117

Future perspectives determined using competition experiments with potential competitors. In regard to the periplasmic binding protein Blr1601, purified protein could be used for binding assays with labeled substrates (Dupont et al. 2004). The monooxygenase Blr1600 should as well be included in further analyzes by determining the enzyme activity as well as its kinetic parameters (Payne et al. 1998).

5.3 Oxalotrophy in Bradyrhizobium japonicum

Our analysis showed that B. japonicum is an oxalotrophic bacterium. Several experiments have been performed to expand the knowledge about the capacity of B. japonicum to use oxalate as a C-source. In order to obtain a more comprehensive view of the oxalotrophic life style in B. japonicum, the proteome of free-living cells grown in minimal medium supplemented with oxalate could be performed and compared to cells grown in the presence of succinate. Succinate is a suitable control since it was shown in chapter 4 that the frc-oxc mutant strain is not affected in growth in the presence of this carbon source. Thus the catabolism of succinate is independent from the catabolism of oxalate. Accordingly proteins could be identified which are expressed under oxalotrophic growth conditions.

Since the frc-oxc mutant was impaired in growth in the presence of L-arabinose the hypothesis that the degradation of L-arabinose is linked to the oxalate energy metabolism, could be followed up by metabolomics analysis. For this purpose the metabolome of B. japonicum wild type and frc-oxc grown in minimal medium supplemented with L-arabinose should be investigated and compared. Thus, the fate of arabinose and glyoxylate could be determined in further depth since it was proposed in chapter 4 that glyoxylate could enter various metabolic routes (Fig 4.4). As mentioned above, cells grown in the presence of succinate could be a suitable control.

118

Future perspectives

We can imagine that the abundance of oxalate changes in the frc-oxc strain as compared to the wild type since in the mutant strain oxalate cannot be further processed to formate and

CO2.

The characterization of ∆frc-oxc showed that the mutant is not affected in terms of nitrogen fixation and nodule development. However, the abundance of oxalate in legume nodules and roots suggested that oxalate could have a role during symbiosis. Preliminary results show that

∆frc-oxc is less competitive in nodule occupancy as compared to the parental strain.

However, to exclude an involvement of the kanamycin resistance gene potentially causing this phenotype, complementation of the ∆frc-oxc needs to be performed. Subsequently, the competiveness of the complemented ∆frc-oxc strain in nodule occupancy needs to be analyzed with various host plants.

The function of the regulatory gene bll3158 coding for a GntR-like regulator upstream of the frc and oxc remains unclear. The determination of Bll3158 regulon could help in gaining insights about the role of this regulator in the context of oxalate metabolism.

119

120

References

Abe, K., Ruan, Z.-S., and Maloney, P.C. (1996) Cloning, sequencing, and expression in Escherichia coli of OxlT, the oxalate:formate exchange protein of Oxalobacter formigenes. J Biol Chem 271: 6789-6793. Abratt, V.R., and Reid, S.J. (2010) Oxalate-degrading bacteria of the human gut as probiotics in the management of kidney stone disease. Adv Appl Microbiol 72: 63-87. Alexeyev, M.F. (1995) Three kanamycin resistance gene cassettes with different polylinkers. Biotechniques 18: 52-56. Ampe, F., Kiss, E., Sabourdy, F., and Batut, J. (2003) Transcriptome analysis of Sinorhizobium meliloti during symbiosis. Genome Biol 4: R15. Anantharam, V., Allison, M.J., and Maloney, P.C. (1989) Oxalate:formate exchange. The basis for energy coupling in Oxalobacter. J Biol Chem 264: 7244-7250. Andriankaja, A., Boisson-Dernier, A., Frances, L., Sauviac, L., Jauneau, A., Barker, D.G., and de Carvalho-Niebel, F. (2007) AP2-ERF transcription factors mediate Nod factor dependent Mt ENOD11 activation in root hairs via a novel cis-regulatory motif. Plant Cell 19: 2866-2885. Angle, J.S., McGrath, S.P., and Chaney, R.L. (1991) New culture medium containing ionic concentrations of nutrients similar to concentrations found in the soil solution. Appl Environ Microbiol 57: 3674-3676. Aragno, M.S., and Schlegel, H. G. (1991) The mesophilic hydrogen-oxidizing (Knallgas) bacteria. Springer Verlag, Berlin, Heidelberg, New York. Ardourel, M., Demont, N., Debellé, F., Maillet, F., de Billy, F., Promé, J.C., Dénarié, J., and Truchet, G. (1994) Rhizobium meliloti lipooligosaccharide nodulation factors: different structural requirements for bacterial entry into target root hair cells and induction of plant symbiotic developmental responses. Plant Cell 6: 1357-1374. Arrighi, J.F., Barre, A., Ben Amor, B., Bersoult, A., Soriano, L.C., Mirabella, R., de Carvalho-Niebel, F., Journet, E.P., Gherardi, M., Huguet, T., Geurts, R., Dénarié, J., Rouge, P., and Gough, C. (2006) The Medicago truncatula lysine motif-receptor-like kinase gene family includes NFP and new nodule-expressed genes. Plant Physiol 142: 265-279. Badger, M.R., and Price, G.D. (1992) The CO2 concentrating mechanism in cyanobacteria and microalgae. Physiol Plantarum 84: 606-615. 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 101: 16636-16641. Becker, A., Berges, H., Krol, E., Bruand, C., Ruberg, S., Capela, D., Lauber, E., Meilhoc, E., Ampe, F., de Bruijn, F.J., Fourment, J., Francez-Charlot, A., Kahn, D., Küster, H., Liebe, C., Pühler, A., Weidner, S., and Batut, J. (2004) Global changes in gene expression in Sinorhizobium meliloti 1021 under microoxic and symbiotic conditions. Mol Plant-Microbe Interact 17: 292-303. Bellato, C., Krishnan, H.B., Cubo, T., Temprano, F., and Pueppke, S.G. (1997) The soybean cultivar specificity gene nolX is present, expressed in a nodD-dependent manner, and of symbiotic significance in cultivar-nonspecific strains of Rhizobium (Sinorhizobium) fredii. Microbiology 143: 1381-1388. Berges, H., Lauber, E., Liebe, C., Batut, J., Kahn, D., de Bruijn, F.J., and Ampe, F. (2003) Development of Sinorhizobium meliloti pilot macroarrays for transcriptome analysis. Appl Environ Microbiol 69: 1214-1219.

121

References

Bernstein, J.A., Khodursky, A.B., Lin, P.H., Lin-Chao, S., and Cohen, S.N. (2002) Global analysis of mRNA decay and abundance in Escherichia coli at single-gene resolution using two-color fluorescent DNA microarrays. Proc Natl Acad Sci U S A 99: 9697- 9702. Bochner, B.R., Gadzinski, P., and Panomitros, E. (2001) Phenotype microarrays for high- throughput phenotypic testing and assay of gene function. Genome Res 11: 1246- 1255. Braissant, O. (2005) Carbonatogénèse bactérienne liée au cycle biogéochimique oxalate- carbonate. PhD Thesis Université de Neuchâtel. Neuchâtel, Switzerland. Brito, B., Toffanin, A., Prieto, R.I., Imperial, J., Ruiz-Argueso, T., and Palacios, J.M. (2008) Host-dependent expression of Rhizobium leguminosarum bv. viciae hydrogenase is controlled at transcriptional and post-transcriptional levels in legume nodules. Mol Plant-Microbe Interact 21: 597-604. Broughton, W.J., Jabbouri, S., and Perret, X. (2000) Keys to symbiotic harmony. J Bacteriol 182: 5641-5652. Charon, C., Sousa, C., Crespi, M., and Kondorosi, A. (1999) Alteration of enod40 expression modifies Medicago truncatula root nodule development induced by Sinorhizobium meliloti. Plant Cell 11: 1953-1966. Chun, J.Y., and Stacey, G. (1993) A Bradyrhizobium japonicum gene essential for nodulation competitiveness is differentially regulated from two promoters. Mol Plant-Microbe Interact 7: 248-255. Chun, J.Y., Sexton, G.L., Roth, L.E., and Stacey, G. (1994) Identification and characterization of a novel Bradyrhizobium japonicum gene involved in host-specific nitrogen fixation. J Bacteriol 179: 6717-6729. Craig, R., and Beavis, R.C. (2004) TANDEM: matching proteins with tandem mass spectra. Bioinformatics 20: 1466-1467. Davis, E.O., Evans, I.J., and Johnston, A.W. (1988) Identification of nodX, a gene that allows Rhizobium leguminosarum biovar viciae strain TOM to nodulate Afghanistan peas. Mol Gen Genet 212: 531-535. de Godoy, L.M., Olsen, J.V., de Souza, G.A., Li, G., Mortensen, P., and Mann, M. (2006) Status of complete proteome analysis by mass spectrometry: SILAC labeled yeast as a model system. Genome Biol 7: R50. Delgado-Salinas, A., Bibler, R., and Lavin, M. (2006) Phylogeny of the genus Phaseolus (Leguminosae): A recent diversification in an ancient landscape. Syst Bot 31: 779- 791. Delmotte, N., Knief, C., Chaffron, S., Innerebner, G., Roschitzki, B., Schlapbach, R., von Mering, C., and Vorholt, J.A. (2009) Community proteogenomics reveals insights into the physiology of phyllosphere bacteria. Proc Natl Acad Sci U S A 106: 16428- 16433. Delmotte, N., Ahrens, C.H., Knief, C., Qeli, E., Koch, M., Fischer, H.M., Vorholt, J.A., Hennecke, H., and Pessi, G. (2010) An integrated proteomics and transcriptomics reference data set provides new insights into the Bradyrhizobium japonicum bacteroid metabolism in soybean root nodules. Proteomics 10: 1391-1400. Dénarié, J., Debellé, F., and Rosenberg, C. (1992) Signaling and host range variation in nodulation. Annu Rev Microbiol 46: 497-531. Dénarié, J., Debellé, F., and Prome, J.C. (1996) Rhizobium lipo-chitooligosaccharide nodulation factors: signaling molecules mediating recognition and morphogenesis. Annu Rev Biochem 65: 503-535. Dimroth, P., and Schink, B. (1998) Energy conservation in the decarboxylation of dicarboxylic acids by fermenting bacteria. Arch Microbiol 170: 69-77.

122

References

Dixon, R., and Kahn, D. (2004) Genetic regulation of biological nitrogen fixation. Nat Rev Microbiol 2:621-631. Djordjevic, M.A., Schofield, P.R., and Rolfe, B.G. (1985) Tn5 mutagenesis of Rhizobium trifolii host-specific nodulation genes result in mutants with altered host-range ability. Mol Gen Genet 200: 463-471. Djordjevic, M.A., Chen, H.C., Natera, S., Van Noorden, G., Menzel, C., Taylor, S., Renard, C., Geiger, O., and Weiller, G.F. (2003) A global analysis of protein expression profiles in Sinorhizobium meliloti: discovery of new genes for nodule occupancy and stress adaptation. Mol Plant-Microbe Interact 16: 508-524. Downie, J.A. (2007) Rhizobia. In Encyclopedia of Life Sciences, pp. 1-8. John Wiley and Sons, Inc. Downie, J.A., and Walker, S.A. (1999) Plant responses to nodulation factors. Curr Opin Plant Biol 2: 483-489. Dupont, L., Garcia, I., Poggi, M.C., Alloing, G., Mandon, K., and Le Rudulier, D. (2004) The Sinorhizobium meliloti ABC transporter Cho is highly specific for choline and expressed in bacteroids from Medicago sativa nodules. J Bacteriol 186: 5988-5996. Ehrhardt, D.W., Atkinson, E.M., Faull, K.F., Freedberg, D.I., Sutherlin, D.P., Armstrong, R., and Long, S.R. (1995) In vitro sulfotransferase activity of NodH, a nodulation protein of Rhizobium meliloti required for host-specific nodulation. J Bacteriol 177: 6237- 6245. Elias, J.E., and Gygi, S.P. (2007) Target-decoy search strategy for increased confidence in large-scale protein identifications by mass spectrometry. Nat Methods 4: 207-214. Esseling, J.J., Lhuissier, F.G., and Emons, A.M. (2004) A nonsymbiotic root hair tip growth phenotype in NORK-mutated legumes: implications for nodulation factor-induced signaling and formation of a multifaceted root hair pocket for bacteria. Plant Cell 16: 933-944. Fauvart, M., and Michiels, J. (2008) Rhizobial secreted proteins as determinants of host specificity in the rhizobium-legume symbiosis. FEMS Microbiol Lett 285: 1-9. Ferguson, B.J., Indrasumunar, A., Hayashi, S., Lin, M.H., Lin, Y.H., Reid, D.E., and Gresshoff, P.M. (2010) Molecular analysis of legume nodule development and autoregulation. J Integr Plant Biol 52: 61-76. Finnie, C., Hartley, N.M., Findlay, K.C., and Downie, J.A. (1997) The Rhizobium leguminosarum prsDE genes are required for secretion of several proteins, some of which influence nodulation, symbiotic nitrogen fixation and exopolysaccharide modification. Mol Microbiol 25: 135-146. Fischer, H.M. (1994) Genetic regulation of nitrogen fixation in rhizobia. Microbiol Rev 58: 352-386. Fischer, H.M., Babst, M., Kaspar, T., Acuña, G., Arigoni, F., and Hennecke, H. (1993) One member of a groESL-like chaperonin multigene family in Bradyrhizobium japonicum is co-regulated with symbiotic nitrogen fixation genes. EMBO J 12: 2901-2912. Franceschi, V.R., and Nakata, P.A. (2005) Calcium oxalate in plants: Formation and function. Annu Rev Plant Biol 56: 41-71. Fry, J., Wood, M., and Poole, P.S. (2001) Investigation of myo-inositol catabolism in Rhizobium leguminosarum bv. viciae and its effect on nodulation competitiveness. Mol Plant-Microbe Interact 14: 1016-1025. Fuhrer, T., Fischer, E., and Sauer, U. (2005) Experimental identification and quantification of glucose metabolism in seven bacterial species. J Bacteriol 187: 1581-1590. Giraud, E., Moulin, L., Vallenet, D., Barbe, V., Cytryn, E., Avarre, J.C., Jaubert, M., Simon, D., Cartieaux, F., Prin, Y., Bena, G., Hannibal, L., Fardoux, J., Kojadinovic, M., Vuillet, L., Lajus, A., Cruveiller, S., Rouy, Z., Mangenot, S., Segurens, B., Dossat, C.,

123

References

Franck, W.L., Chang, W.S., Saunders, E., Bruce, D., Richardson, P., Normand, P., Dreyfus, B., Pignol, D., Stacey, G., Emerich, D., Vermeglio, A., Medigue, C., and Sadowsky, M. (2007) Legumes symbioses: Absence of nod genes in photosynthetic bradyrhizobia. Science 316: 1307-1312. Göttfert, M., Grob, P., and Hennecke, H. (1990) Proposed regulatory pathway encoded by the nodV and nodW genes, determinants of host specificity in Bradyrhizobium japonicum. Proc Natl Acad Sci U S A 87: 2680-2684. Hahn, M., and Hennecke, H. (1984) Localized mutagenesis in Rhizobium japonicum. Mol Gen Genet 193: 46-52. Hauser, F., Lindemann, A., Vuilleumier, S., Patrignani, A., Schlapbach, R., Fischer, H.M., and Hennecke, H. (2006) Design and validation of a partial-genome microarray for transcriptional profiling of the Bradyrhizobium japonicum symbiotic gene region. Mol Genet Genomics 275: 55-67. Hauser, F., Pessi, G., Friberg, M., Weber, C., Rusca, N., Lindemann, A., Fischer, H.M., and Hennecke, H. (2007) Dissection of the Bradyrhizobium japonicum NifA+σ54 regulon, and identification of a ferredoxin gene (fdxN) for symbiotic nitrogen fixation. Mol Genet Genomics 278: 255-271. Higgins, C.F. (2001) ABC transporters: physiology, structure and mechanism. Res Microbiol 152: 205-210. Hoskisson, P.A., and Rigali, S. (2009) Variation in form and function the helix-turn-helix regulators of the GntR superfamily. Adv Appl Microbiol 69: 1-22. Hotter, G.S., and Scott, D.B. (1991) Exopolysaccharide mutants of Rhizobium loti are fully effective on a determinate nodulating host but are ineffective on an indeterminate nodulating host. J Bacteriol 173: 851-859. Indrasumunar, A., Kereszt, A., Searle, I., Miyagi, M., Li, D.X., Nguyen, C.D.T., Men, A., Carroll, B.J., and Gresshoff, P.M. (2010) Inactivation of duplicated Nod factor receptor 5 (NFR5) genes in recessive loss-of-function non-nodulation mutants of allotetraploid soybean (Glycine max L. Merr.). Plant Cell Physiol 51: 201-214. Jiang, G., and Krishnan, H.B. (2000) Sinorhizobium fredii USDA257, a cultivar-specific soybean symbiont, carries two copies of y4yA and y4yB, two open reading frames that are located in a region that encodes the type III protein secretion system. Mol Plant- Microbe Interact 13: 1010-1014. Jimenezzurdo, J.I., Vandillewijn, P., Soto, M.J., Defelipe, M.R., Olivares, J., and Toro, N. (1995) Characterization of a Rhizobium meliloti proline dehydrogenase mutant altered in nodulation efficiency and competitiveness on alfalfa roots. Mol Plant-Microbe Interact 8: 492-498. Kahm, M., Hasenbrink, G., Lichtenberg-Fraté, H., Ludwig, J., and Kschischo, M. (2010) grofit: Fitting biological growth curves with R. J Stat Software 33: 1-21. Kalloniati, C., Tsikou, D., Lampiri, V., Fotelli, M.N., Rennenberg, H., Chatzipavlidis, I., Fasseas, C., Katinakis, P., and Flemetakis, E. (2009) Characterization of a Mesorhizobium loti α-type carbonic anhydrase and its role in symbiotic nitrogen fixation. J Bacteriol 191: 2593-2600. Kalo, P., Gleason, C., Edwards, A., Marsh, J., Mitra, R.M., Hirsch, S., Jakab, J., Sims, S., Long, S.R., Rogers, J., Kiss, G.B., Downie, J.A., and Oldroyd, G.E. (2005) Nodulation signaling in legumes requires NSP2, a member of the GRAS family of transcriptional regulators. Science 308: 1786-1789. Kamionka, A., Bertram, R. Hillen, W., (2005) Tetracycline-dependent conditional gene knockout in Bacillus subtilis. Appl Environ Microbiol 71: 728-733. Kanamori, N., Madsen, L.H., Radutoiu, S., Frantescu, M., Quistgaard, E.M., Miwa, H., Downie, J.A., James, E.K., Felle, H.H., Haaning, L.L., Jensen, T.H., Sato, S.,

124

References

Nakamura, Y., Tabata, S., Sandal, N., and Stougaard, J. (2006) A nucleoporin is required for induction of Ca2+ spiking in legume nodule development and essential for rhizobial and fungal symbiosis. Proc Natl Acad Sci U S A 103: 359-364. Kaneko, T., Nakamura, Y., Sato, S., Minamisawa, K., Uchiumi, T., Sasamoto, S., Watanabe, A., Idesawa, K., Iriguchi, M., Kawashima, K., Kohara, M., Matsumoto, M., Shimpo, S., Tsuruoka, H., Wada, T., Yamada, M., and Tabata, S. (2002) Complete genomic sequence of nitrogen-fixing symbiotic bacterium Bradyrhizobium japonicum USDA110. DNA Res 9: 189-197. Karunakaran, R., Ramachandran, V.K., Seaman, J.C., East, A.K., Mouhsine, B., Mauchline, T.H., Prell, J., Skeffington, A., and Poole, P.S. (2009) Transcriptomic analysis of Rhizobium leguminosarum biovar viciae in symbiosis with host plants Pisum sativum and Vicia cracca. J Bacteriol 191: 4002-4014. Kaur, S., Mishra, M.N., and Tripathi, A.K. (2009) Regulation of expression and biochemical characterization of a β-class carbonic anhydrase from the plant growth-promoting rhizobacterium, Azospirillum brasilense Sp7. FEMS Microbiol Lett 299: 149-158. Keller, A., Nesvizhskii, A.I., Kolker, E., and Aebersold, R. (2002) Empirical statistical model to estimate the accuracy of peptide identifications made by MS/MS and database search. Anal Chem 74: 5383-5392. Khammar, N., Martin, G., Ferro, K., Job, D., Aragno, M., and Verrecchia, E. (2009) Use of the frc gene as a molecular marker to characterize oxalate-oxidizing bacterial abundance and diversity structure in soil. J Microbiol Meth 76: 120-127. Kim, S.A., and Copeland, L. (1996) Enzymes of poly-(β)-hydroxybutyrate metabolism in soybean and chickpea bacteroids. Appl Environ Microbiol 62: 4186-4190. Knee, E.M., Gong, F.C., Gao, M.S., Teplitski, M., Jones, A.R., Foxworthy, A., Mort, A.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: 775-784. Koch, M., Delmotte, N., Rehrauer, H., Vorholt, J.A., Pessi, G., and Hennecke, H. (2010) Rhizobial adaptation to hosts, a new facet in the legume root-nodule symbiosis. Mol Plant-Microbe Interact 23: 784-790. Kohler, P.R., Zheng, J.Y., Schoffers, E., and Rossbach, S. (2010) Inositol catabolism, a key pathway in Sinorhizobium meliloti for competitive host nodulation. Appl Environ Microbiol 76: 7972-7980. Kosslak, R.M., Bookland, R., Barkei, J., Paaren, H.E., and Appelbaum, E.R. (1987) Induction of Bradyrhizobium japonicum common nod genes by isoflavones isolated from Glycine max. Proc Natl Acad Sci U S A 84: 7428-7432. Krusell, L., Krause, K., Ott, T., Desbrosses, G., Kramer, U., Sato, S., Nakamura, Y., Tabata, S., James, E.K., Sandal, N., Stougaard, J., Kawaguchi, M., Miyamoto, A., Suganuma, N., and Udvardi, M.K. (2005) The sulfate transporter SST1 is crucial for symbiotic nitrogen fixation in Lotus japonicus root nodules. Plant Cell 17: 1625-1636. Lee, J., and Hymowitz, T. (2001) A molecular phylogenetic study of the subtribe Glycininae (Leguminosae) derived from the chloroplast DNA rps16 intron sequences. Am J Bot 88: 2064-2073. Levy, J., Bres, C., Geurts, R., Chalhoub, B., Kulikova, O., Duc, G., Journet, E.P., Ane, J.M., Lauber, E., Bisseling, T., Dénarié, J., Rosenberg, C., and Debellé, F. (2004) A putative Ca2+ and calmodulin-dependent protein kinase required for bacterial and fungal symbioses. Science 303: 1361-1364. Lewin, A., Cervantes, E., Chee-Hoong, W., and Broughton, W.J. (1990) nodSU, two new nod genes of the broad host range Rhizobium strain NGR234 encode host-specific nodulation of the tropical tree Leucaena leucocephala. Mol Plant-Microbe Interact 3: 317-326.

125

References

Libert, B., and Franceschi, V.R. (1987) Oxalate in crop plants. J Agric Food Chem 35: 926- 938. Limpens, E., Franken, C., Smit, P., Willemse, J., Bisseling, T., and Geurts, R. (2003) LysM domain receptor kinases regulating rhizobial Nod factor-induced infection. Science 302: 630-633. Lindemann, A., Koch, M., Pessi, G., Müller, A.J., Balsiger, S., Hennecke, H., and Fischer, H.M. (2010) Host-specific symbiotic requirement of BdeAB, a RegR-controlled RND-type efflux system in Bradyrhizobium japonicum. FEMS Microbiol Lett 312: 184-191. Lodwig, E., and Poole, P. (2003) Metabolism of Rhizobium bacteroids. Crit Rev Plant Sci 22: 37-78. Lodwig, E.M., Hosie, A.H., Bourdes, A., Findlay, K., Allaway, D., Karunakaran, R., Downie, J.A., and Poole, P.S. (2003) Amino-acid cycling drives nitrogen fixation in the legume-Rhizobium symbiosis. Nature 422: 722-726. Loh, J., Garcia, M., and Stacey, G. (1997) NodV and NodW, a second flavonoid recognition system regulating nod gene expression in Bradyrhizobium japonicum. J Bacteriol 179: 3013-3020. Lorio, J.C., Kim, W.S., and Krishnan, H.B. (2004) NopB, a soybean cultivar-specificity protein from Sinorhizobium fredii USDA257, is a type III secreted protein. Mol Plant- Microbe Interact 17: 1259-1268. Lung, H.Y., Baetz, A.L., and Peck, A.B. (1994) Molecular cloning, DNA sequence, and gene expression of the oxalyl-coenzyme A decarboxylase gene, oxc, from the bacterium Oxalobacter formigenes. J Bacteriol 176: 2468-2472. Madsen, E.B., Madsen, L.H., Radutoiu, S., Olbryt, M., Rakwalska, M., Szczyglowski, K., Sato, S., Kaneko, T., Tabata, S., Sandal, N., and Stougaard, J. (2003) A receptor kinase gene of the LysM type is involved in legume perception of rhizobial signals. Nature 425: 637-640. Masson-Boivin, C., Giraud, E., Perret, X., and Batut, J. (2009) Establishing nitrogen-fixing symbiosis with legumes: how many rhizobium recipes? Trends Microbiol 17: 458- 466. Matsusaki, S., Watanabe, H., Oshima, T., Kanaya, S., and Mori, H. (2002) Prediction of target substrates of transporters in Escherichia coli. Genome Informatics 13: 394-395. McIver, J., Djordjevic, M.A., Weinman, J.J., Bender, G.L., and Rolfe, B.G. (1989) Extension of host range of Rhizobium leguminosarum bv. trifolii caused by point mutations in nodD that result in alterations in regulatory function and recognition of inducer molecules. Mol Plant-Microbe Interact 2: 97-106. Meinhardt, L.W., Krishnan, H.B., Balatti, P.A., and Pueppke, S.G. (1993) Molecular cloning and characterization of a sym plasmid locus that regulates cultivar-specific nodulation of soybean by Rhizobium fredii USDA257. Mol Microbiol 9: 17-29. Milcamps, A., and de Bruijn, F.J. (1999) Identification of a novel nutrient-deprivation- induced Sinorhizobium meliloti gene (hmgA) involved in the degradation of tyrosine. Microbiology 145: 935-947. Miller, J. (1972) Experiments in molecular genetics. Cold Spring Harbor Laboratory Press Cold Spring Harbor N. Y. USA. Miller, S.H., Elliot, R.M., Sullivan, J.T., and Ronson, C.W. (2007) Host-specific regulation of symbiotic nitrogen fixation in Rhizobium leguminosarum biovar trifolii. Microbiology 153: 3184-3195. Morris, A.C., and Djordjevic, M.A. (2001) Proteome analysis of cultivar-specific interactions between Rhizobium leguminosarum biovar trifolii and subterranean clover cultivar Woogenellup. Electrophoresis 22: 586-598.

126

References

Moussatova, A., Kandt, C., O'Mara, M.L., and Tieleman, D.P. (2008) ATP-binding cassette transporters in Escherichia coli. Biochim Biophys Acta 1778: 1757-1771. Moutran, A., Balan, A., Ferreira, L.C., Giorgetti, A., Tramontano, A., and Ferreira, R.C. (2007) Structural model and ligand interactions of the Xanthomonas axonopodis pv. citri oligopeptide-binding protein. Genet Mol Res 6: 1169-1177. Nesvizhskii, A.I., Keller, A., Kolker, E., and Aebersold, R. (2003) A statistical model for identifying proteins by tandem mass spectrometry. Anal Chem 75: 4646-4658. Nörtemann, B. (1999) Biodegradation of EDTA. Appl Microbiol Biotechnol 51: 751-759. Novick, N.J., and Tyler, M.E. (1982) L-Arabinose metabolism in Azospirillum brasiliense. J Bacteriol 149: 364-367. Oh, H.S., Son, O., Chun, J.Y., Stacey, G., Lee, M.S., Min, K.H., Song, E.S., and Cheon, C.I. (2001) The Bradyrhizobium japonicum hsfA gene exhibits a unique developmental expression pattern in cowpea nodules. Mol Plant-Microbe Interact 14: 1286-1292. Oke, V., and Long, S.R. (1999a) Bacteroid formation in the Rhizobium-legume symbiosis. Curr Opin Microbiol 2: 641-646. Oke, V., and Long, S.R. (1999b) Bacterial genes induced within the nodule during the Rhizobium-legume symbiosis. Mol Microbiol 32: 837-849. Oldroyd, G.E., and Downie, J.A. (2008) Coordinating nodule morphogenesis with rhizobial infection in legumes. Annu Rev Plant Biol 59: 519-546. Ong, S.E., and Mann, M. (2005) Mass spectrometry-based proteomics turns quantitative. Nat Chem Biol 1: 252-262. Op den Camp, R., Streng, A., De Mita, S., Cao, Q., Polone, E., Liu, W., Ammiraju, J.S.S., Kudrna, D., Wing, R., Untergasser, A., Bisseling, T., Geurts, R. (2011) LysM-type mycorrhizal receptor recruited for rhizobium symbiosis in nonlegume parasponia. Science. 331: 909-912. Osteras, M., Finan, T.M., and Stanley, J. (1991) Site-directed mutagenesis and DNA sequence of pckA of Rhizobium NGR234, encoding phosphoenolpyruvate carboxykinase: gluconeogenesis and host-dependent symbiotic phenotype. Mol Gen Genet 230: 257-269. Panter, S., Thomson, R., de Bruxelles, G., Laver, D., Trevaskis, B., and Udvardi, M. (2000) Identification with proteomics of novel proteins associated with the peribacteroid membrane of soybean root nodules. Mol Plant-Microbe Interact 13: 325-333. Park, H., Song, B., and Morel, F.M. (2007) Diversity of the cadmium-containing carbonic anhydrase in marine diatoms and natural waters. Environ Microbiol 9: 403-413. Payne, J.W., Bolton, H., Jr., Campbell, J.A., and Xun, L. (1998) Purification and characterization of EDTA monooxygenase from the EDTA-degrading bacterium BNC1. J Bacteriol 180: 3823-3827. Pedrosa, F.O., and Zancan, G.T. (1974) L-arabinose metabolism in Rhizobium japonicum. J Bacteriol 119: 336-338. Perkins, D.N., Pappin, D.J., Creasy, D.M., and Cottrell, J.S. (1999) Probability-based protein identification by searching sequence databases using mass spectrometry data. Electrophoresis 20: 3551-3567. Perret, X., Staehelin, C., and Broughton, W.J. (2000) Molecular basis of symbiotic promiscuity. Microbiol Mol Biol Rev 64: 180-201. Pessi, G., Ahrens, C.H., Rehrauer, H., Lindemann, A., Hauser, F., Fischer, H.-M., and Hennecke, H. (2007) Genome-wide transcript analysis of Bradyrhizobium japonicum bacteroids in soybean root nodules. Mol Plant-Microbe Interact 20: 1353-1363. Peters, N.K., Frost, J.W., and Long, S.R. (1986) A plant flavone, luteolin, induces expression of Rhizobium meliloti nodulation genes. Science 233: 977-980.

127

References

Pierce, E., Becker, D.F., and Ragsdale, S.W. (2010) Identification and characterization of oxalate oxidoreductase, a novel thiamine pyrophosphate-dependent 2-oxoacid oxidoreductase that enables anaerobic growth on oxalate. J Biol Chem 285: 40515- 40524. Pobigaylo, N., Wetter, D., Szymczak, S., Schiller, U., Kurtz, S., Meyer, F., Nattkemper, T.W., and Becker, A. (2006) Construction of a large signature-tagged mini-Tn5 transposon library and its application to mutagenesis of Sinorhizobium meliloti. Appl Environ Microbiol 72: 4329-4337. Poysti, N.J., Loewen, E.D.M., Wang, Z.X., and Oresnik, I.J. (2007) Sinorhizobium meliloti pSymB carries genes necessary for arabinose transport and catabolism. Microbiology 153: 727-736. Prell, J., and Poole, P. (2006) Metabolic changes of rhizobia in legume nodules. Trends Microbiol 14: 161-168. Puskas, L.G., Inui, M., Zahn, K., and Yukawa, H. (2000) A periplasmic, α-type carbonic anhydrase from Rhodopseudomonas palustris is essential for bicarbonate uptake. Microbiology 146: 2957-2966. Radutoiu, S., Madsen, L.H., Madsen, E.B., Jurkiewicz, A., Fukai, E., Quistgaard, E.M., Albrektsen, A.S., James, E.K., Thirup, S., and Stougaard, J. (2007) LysM domains mediate lipochitin-oligosaccharide recognition and Nfr genes extend the symbiotic host range. EMBO J 26: 3923-3935. Radutoiu, S., Madsen, L.H., Madsen, E.B., Felle, H.H., Umehara, Y., Gronlund, M., Sato, S., Nakamura, Y., Tabata, S., Sandal, N., and Stougaard, J. (2003) Plant recognition of symbiotic bacteria requires two LysM receptor-like kinases. Nature 425: 585-592. Redmond, J.W., Batley, M., Innes, R.W., Kuempel, P.L., Djordjevic, M.A., and Rolfe, B.G. (1986) Flavones induce expression of the nodulation genes in Rhizobium. Nature 323: 632-635. Regensburger, B., and Hennecke, H. (1983) RNA polymerase from Rhizobium japonicum. Arch Microbiol 135: 103-109. Riely, B.K., Lougnon, G., Ane, J.M., and Cook, D.R. (2007) The symbiotic ion channel homolog DMI1 is localized in the nuclear membrane of Medicago truncatula roots. Plant J 49: 208-216. Ronson, C.W., Lyttleton, P., and Robertson, J.G. (1981) C4-dicarboxylate transport mutants of Rhizobium trifolii form ineffective nodules on Trifolium repens. Proc Natl Acad Sci U S A 78: 4284-4288. Ronson, C.W., Astwood, P.M., and Downie, J.A. (1984) Molecular-cloning and genetic organization of C4-dicarboxylate transport genes from Rhizobium leguminosarum. J Bacteriol 160: 903-909. Ronson, C.W., Astwood, P.M., Nixon, B.T., and Ausubel, F.M. (1987) Deduced products of C4-dicarboxylate transport regulatory genes of Rhizobium leguminosarum are homologous to nitrogen regulatory gene products. Nucleic Acids Res 15: 7921-7934. Ruan, Z.S., Anantharam, V., Crawford, I.T., Ambudkar, S.V., Rhee, S.Y., Allison, M.J., and Maloney, P.C. (1992) Identification, purification, and reconstitution of OxlT, the oxalate:formate antiport protein of Oxalobacter formigenes. J Biol Chem 267: 10537- 10543. Sadowsky, M.J., Cregan, P.B., Göttfert, M., Sharma, A., Gerhold, D., Rodriguez-Quinones, F., Keyser, H.H., Hennecke, H., and Stacey, G. (1991) The Bradyrhizobium japonicum nolA gene and its involvement in the genotype-specific nodulation of soybeans. Proc Natl Acad Sci U S A 88: 637-641. Sahin, N. (2003) Oxalotrophic bacteria. Res Microbiol 154: 399-407.

128

References

Sambrook, J., and Russel, D.W. (2001) Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor N.Y. USA. Sarma, A.D., and Emerich, D.W. (2005) Global protein expression pattern of Bradyrhizobium japonicum bacteroids: a prelude to functional proteomics. Proteomics 5: 4170-4184. Schell, M.A. (1993) Molecular biology of the LysR family of transcriptional regulators. Annu Rev Microbiol 47: 597-626. Scott, D., Weeks, D., Rektorschek, M., Sachs, G., and Melchers, K. (1998) Acid-adaptive mechanisms of gastric Helicobacter. Springer Verlag, Dordrecht, Netherlands. Sidhu, H., Ogden, S.D., Lung, H.Y., Luttge, B.G., Baetz, A.L., and Peck, A.B. (1997) DNA sequencing and expression of the formyl coenzyme A transferase gene, frc, from Oxalobacter formigenes. J Bacteriol 179: 3378-3381. Simon, R., Priefer, U., and Pühler, A. (1983) Vector plasmids for in vivo and in vitro manipulation of gram-negative bacteria. Molecular genetetics of the bacteria-plant interaction. A. Pühler (ed). Springer Verlag, Heidelberg, Germany, pp.98-106. Smit, P., Raedts, J., Portyanko, V., Debellé, F., Gough, C., Bisseling, T., and Geurts, R. (2005) NSP1 of the GRAS protein family is essential for rhizobial Nod factor-induced transcription. Science 308: 1789-1791. Smith, K.S., and Ferry, J.G. (2000) Prokaryotic carbonic anhydrases. FEMS Microbiol Rev 24: 335-366. Smith, K.S., Ingram-Smith, C., and Ferry, J.G. (2002) Roles of the conserved aspartate and arginine in the catalytic mechanism of an archaeal β-class carbonic anhydrase. J Bacteriol 184: 4240-4245. Spaink, H.P. (2000) Root nodulation and infection factors produced by rhizobial bacteria. Annu Rev Microbiol 54: 257-288. Spaink, H.P., Wijfjes, A.H., and Lugtenberg, B.J. (1995) Rhizobium NodI and NodJ proteins play a role in the efficiency of secretion of lipochitin oligosaccharides. J Bacteriol 177: 6276-6281. Sprent, J.I. (2001) Nodulation in legumes. Royal Botanic Gardens, Kew, UK. Stacey, G., Luka, S., Sanjuan, J., Banfalvi, Z., Nieuwkoop, A.J., Chun, J.Y., Forsberg, L.S., and Carlson, R. (1994) nodZ, a unique host-specific nodulation gene, is involved in the fucosylation of the lipooligosaccharide nodulation signal of Bradyrhizobium japonicum. J Bacteriol 176: 620-633. Stewart, C.S., Duncan, S.H., and Cave, D.R. (2004) Oxalobacter formigenes and its role in oxalate metabolism in the human gut. FEMS Microbiol Lett 230: 1-7. Stowers, M.D. (1985) Carbon metabolism in Rhizobium species. Annu Rev Microbiol 39: 89- 108. Sugawara, M., Shah, G., Sadowsky, M., Paliy, O., Speck, J., Vail, A., and Gyaneshwar, P. (2011) Expression and functional roles of Bradyrhizobium japonicum genes involved in the utilization of inorganic and organic sulfur compounds in free-living and symbiotic conditions. Mol Plant-Microbe Interact 24: 451-457. Sutherland, J.M., and Sprent, J.I. (1984) Calcium-oxalate crystals and crystal cells in determinate root-nodules of legumes. Planta 161: 193-200. Tam, R., and Saier, M.H., Jr. (1993) Structural, functional, and evolutionary relationships among extracellular solute-binding receptors of bacteria. Microbiol Rev 57: 320-346. Trese, A.T. (1995) A single dominant gene in McCall soybean prevents effective nodulation with Rhizobium fredii USDA257. Euphytica 81: 279-282. Trinchant, J.C., Guerin, V., and Rigaud, J. (1994) Acetylene reduction by symbiosomes and free bacteroids from broad bean (Vicia faba L.) nodules - role of oxalate. Plant Physiol 105: 555-561.

129

References

Tripp, B.C., Smith, K., and Ferry, J.G. (2001) Carbonic anhydrase: new insights for an ancient enzyme. J Biol Chem 276: 48615-48618. Uchiumi, T., Ohwada, T., Itakura, M., Mitsui, H., Nukui, N., Dawadi, P., Kaneko, T., Tabata, S., Yokoyama, T., Tejima, K., Saeki, K., Omori, H., Hayashi, M., Maekawa, T., Sriprang, R., Murooka, Y., Tajima, S., Simomura, K., Nomura, M., Suzuki, A., Shimoda, Y., Sioya, K., Abe, M., and Minamisawa, K. (2004) Expression islands clustered on the symbiosis island of the Mesorhizobium loti genome. J Bacteriol 186: 2439-2448. Vincent, J.M. (1970) A manual for the practical study of root nodule bacteria. IBP Handbook No. 15. Blackwell Scientific Publications, Oxford. Watanabe, S., Kodaki, T., and Makino, K. (2006a) Cloning, expression, and characterization of bacterial L-arabinose 1-dehydrogenase involved in an alternative pathway of L- arabinose metabolism. J Biol Chem 281: 2612-2623. Watanabe, S., Shimada, N., Tajima, K., Kodaki, T., and Makino, K. (2006b) Identification and characterization of L-arabonate dehydratase, L-2-keto-3-deoxyarabonate dehydratase, and L-arabinolactonase involved in an alternative pathway of L- arabinose metabolism - Novel evolutionary insight into sugar metabolism. J Biol Chem 281: 33521-33536. White, V.E., and Knowles, C.J. (2003) Degradation of copper-NTA by Mesorhizobium sp. NCIMB 13524. Int Biodeterior Biodegrad 52: 143-150. Yang, S., Tang, F., Gao, M., Krishnan, H.B., and Zhu, H. (2010) R gene-controlled host specificity in the legume-rhizobia symbiosis. Proc Natl Acad Sci U S A 107: 18735- 18740. Yoshida, S., and Parniske, M. (2005) Regulation of plant symbiosis receptor kinase through serine and threonine phosphorylation. J Biol Chem 280: 9203-9209. Yurgel, S.N., and Kahn, M.L. (2004) Dicarboxylate transport by rhizobia. FEMS Microbiol Rev 28: 489-501. Zhang, H., Herman, J.P., Bolton, H., Jr., Zhang, Z., Clark, S., and Xun, L. (2007) Evidence that bacterial ABC-type transporter imports free EDTA for metabolism. J Bacteriol 189: 7991-7997. Zwietering, M.H., Jongenburger, I., Rombouts, F.M., and van 't Riet, K. (1990) Modeling of the bacterial growth curve. Appl Environ Microbiol 56: 1875-1881.

130

Publications

Koch, M., Delmotte, N., Rehrauer, H., Vorholt, J.A., Pessi, G., and Hennecke, H. (2010) Rhizobial adaptation to hosts, a new facet in the legume root-nodule symbiosis. Mol Plant-Microbe Interact 23: 784-790.

Delmotte, N., Ahrens, C.H., Knief, C., Qeli, E., Koch, M., Fischer, H.M., Vorholt, J.A., Hennecke, H., and Pessi, G. (2010) An integrated proteomics and transcriptomics reference data set provides new insights into the Bradyrhizobium japonicum bacteroid metabolism in soybean root nodules. Proteomics 10: 1391-1400.

Lindemann, A., Koch, M., Pessi, G., Müller, A.J., Balsiger, S., Hennecke, H., and Fischer, H.M. (2010) Host-specific symbiotic requirement of BdeAB, a RegR-controlled RND-type efflux system in Bradyrhizobium japonicum. FEMS Microbiol Lett 312: 184-191.

131