THE ACTINORHIZAL SYMBIOSIS of the EARLIEST DIVERGENT FRANKIA CLUSTER Nguyen Thi Thanh Van

Home , Actinorhizal plant, Alnus nitida, Casuarina cunninghamiana, Casuarina pauper, Ceanothus, Ceanothus integerrimus

THE ACTINORHIZAL SYMBIOSIS OF THE EARLIEST DIVERGENT FRANKIA CLUSTER Nguyen Thi Thanh Van

The actinorhizal symbiosis of the earliest divergent Frankia cluster

Nguyen Thi Thanh Van

©Nguyen Thi Thanh Van, Stockholm University 2017 ISBN print 978-91-7649-716-6 ISBN PDF 978-91-7649-717-3 Printed in Sweden by US-AB, Stockholm 2017 Distributor: Department of Ecology, Environment and Plant Sciences Stockholm University, Stockholm, Sweden

Con kính tặng Ba

Sammanfattning

Under de senaste åren har allt större tonvikt lagts vid vikten av att utveckla en världsomspännande utvecklingsmässigt ekologiskt hållbar jordbruksnäring. Nödvändigheten av att minska beroendet av syntetiska kvävebaserade gödningsmedel har lett till ett alltmer växande forskningsfält baserat på vikten av biologisk kvävefixering, med emfas på rotknölsymbios. Mina studier har fokuserat på mekanismerna i och med actinorhizalsymbios, i.e den symbiotiska interaktionen mellan olika kvävefixerande actinobacteria -genus Frankia. i förhållande till en mycket skiftande grupp av växter från åtta olika växtfamiljer, som sammantaget kallas actinorhizalväxter. Frankia kluster II har visats vara en systerklad i förhållande till alla andra kluster, vilket gör att denna speciella gren kan ge insikt i hur Frankia kluster II kan tillföra värdefull information om den evolutionära bakgrunden i och med actinorhizalsymbios. Det första fullt sekvenserade genomet av en medlem från detta kluster, Candidatus Frankia datiscae Dg1 med ursprung från Pakistan, visar att genomet infattar nod generna nodABC som svarar för framställandet av lipochitooligosaccharid Nod-faktorer, vilket särskiljer denna medlem från andra grupper av Frankia som saknar dessa gener. Arbetet i och med denna avhandling är grundat på gDNA isolerat från sex inokula, tre från Nordamerika (USA; två från Kalifornien och ett från Alaska), ett från Europa (Frankrike), ett från Asien (Japan) och ett från Oceanien (Papua Nya Guinea). gDNA isolerades från rotknölar från Datisca glomerata (Datiscaceae), Ceanothus thyrsiflorus (Rhamnaceae), Coriaria myrtifolia och Coriaria arborea (Coriariaceae). Sammantaget sekvenserades tretton metagenom. Denna avhandling visar att de olika medlemarna i Frankia kluster II agerar som en grupp, vilket kan förklara förmågan att just kluster II kan nodulera en ovanligt stor och divers grupp av värdväxter, vilka innefattar fyra familjer från två ordningar. Det inokulum som isolerades från Papua Nya Guinea (den enda art från södra halvklotet som hittills sekvenserats), visade sig härbärgera en hittills okänd Frankia-art, som namngavs Candidatus Frankia meridionalis. Alla kluster II-arter studerade i detta arbete bär nod-generna nodABC i sitt genom, med undantag för den art som isolerats från Papua Nya Guinea vars genom enbart innehar nodB’C. Alla tre metagenom isolerade från USA innefattade dessutom en sulfotransferas-gen, nodH. Denna gen har visat sig vara uttryckt värdväxtspecifikt, då proteinet kunde spåras i rotknölar från värdväxten C. thyrsiflorus men inte i de knölar som isolerats från D. glomerata. Fylogenetiska analyser sammantagna med transposasfrekvensdata

i våra analyserade genom ger starkt stöd till hypotesen att den utökade värdkretsen från Coriaria till Datisca för Frankia-arter i kluster II skedde i Euroasien samt att kluster II-arter kom till nordamerika via Berings sund. För att erhålla mer information om och hur värdväxten påverkar mikrosymbiontens molekylära reaktioner jämfördes metabolismen i rotknölar från D. glomerata (Cucurbitales) och C. thyrsiflorus (Rosales) på transkriptionsnivå. Det system som skyddar nitrogenas från de negativa effekterna av syre verkar vara mer effektivt i rotknölar från Ceanothus i jämförelse med hur systemet verkar i knölar från Datisca, enär bakteriernas kvävemetabolism sannolikt är likställd i samröret med båda växtarterna. Aminosyraextrakten från rotknölarna från D. glomerata visar att profilen utifrån kvävehaltiga aminosyror domineras av glutamat and arginin, vilket stöder hypotesen att Frankia utsöndrar kväve, troligen i form av arginin, speciellt anpassat för symbiosen med D. glomerata. Således visar våra data att arter från Frankia II särskiljer sig från andra Frankia-arter i hänseendet att deras repertoar av nod-gener och deras kvävemetabolism i symbios med sin värdväxt är specifikt annorlunda.

List of publication

This thesis is based on the following publications and manuscripts:

I. Nguyen TV, Wibberg D, Battenberg K, Blom J, Vanden Heuvel B, Berry AM, Kalinowski J, Pawlowski K (2016) An assemblage of Frankia Cluster II strains from California contains the canonical nod genes and also the sulfotransferase gene nodH. BMC Genomics 17: 796. II. Nguyen TV, Wibberg D, Vigil-Stenman T, Battenberg K, Demchenko KN, Blom J, Fernandez MP, Yamanaka T, Berry AM, Kalinowski J, Brachmann A, Pawlowski K. Meta- genomes from the earliest divergent Frankia cluster: They come in teams. Manuscript III. Nguyen TV, Hilker R, Wibberg D, Battenberg K, Berry AM, Kalinowski J, Pawlowski K. Cucurbitales vs. Rosales: Anal- ysis of bacterial metabolism based on candidate gene expression profiling in root nodules of Datisca glomerata and Cea- nothus thyrsiflorus. Manuscript IV. Persson T, Nguyen TV, Alloisio N, Pujic P, Berry AM, Nor- mand P, Pawlowski K (2016) The N-metabolites of roots and actinorhizal nodules from Alnus glutinosa and Datisca glomerata: can D. glomerata change N-transport forms when nodulated? Symbiosis 70: 149-157

My contributions to these papers/manuscripts were as follows:

I: Obtained the inoculum, grew and infected the plants, isolated the DNA, participated in the evaluation of the genome data, conducted and evaluated the RT-qPCR experiments, participated in writing the manuscript II: Grew the plants and maintained the inocula, isolated the DNA, participated in the evaluation of the genome data, conducted and evaluated the RT-qPCR experiments, participated in writing the manuscript III: Grew the D. glomerata plants, performed all RNA isolations, conducted and evaluated the RT-qPCR experiments, wrote the manuscript IV: Grew the plants and extracted the metabolites for D. glomerata, participated in the evaluation of the data and the writing of the manuscript

Also partially written by the author but not included in this thesis:

V. Nguyen TV, Pawlowski K (2017). Frankia and actinorhizal plants: symbiotic nitrogen fixation. In: Mehnaz S (ed). Mi- crobes for sustainability - Rhizotrophs, Springer, New York. In press. VI. Normand P, Nguyen TV, Battenberg K, Berry AM, Vanden Heuvel B, Fernandez MP, Pawlowski K. Proposal of Candi- datus Frankia californiensis, the uncultured nitrogen-fixing symbiont associated with a phylogenetically broad group of hosts endemic to California. Int J Syst Evol Microbiol, in preparation.

Contents

1. Introduction ...... 13 1.1 Nitrogen-fixing symbiosis in root nodules ...... 13 1.2 Actinorhizal symbiosis ...... 14 1.3 Actinorhizal nodules ...... 19 1.3.1 Nodule induction ...... 19 1.3.2 Nodule structure ...... 21 1.3.3 Nodule metabolism ...... 23 1.4 Signal exchange between microsymbiont and host plant ...... 24 1.4.1 Signal exchange in the legume-rhizobia symbiosis ...... 24 1.4.2 Signal exchange in actinorhizal symbiosis ...... 26 1.5 Evolution of actinorhizal symbiosis ...... 27 2. Aim of this thesis ...... 29 3. Comments on methodology ...... 30 3.1 Frankia gDNA isolation for genome sequencing ...... 30 3.2 Candidate gene expression profiling of bacteria in nodules of Datisca glomerata and Ceanothus thyrsiflorus ...... 30 3.3 Amino acid profiling of roots and nodules from Datisca glomerata . 32 4. Results and Discussion ...... 33 4.1 Genome sequencing of Frankia cluster II strains...... 33 4.1.1 Frankia cluster II inocula represent strain assemblages ...... 33 4.1.2 Eurasian cluster II Frankia (meta-)genomes contain the canonical nod genes nodABC and North American strains also contain the sulfotransferase gene nodH...... 37 4.1.3 Cluster II genome (in-)stability: transposase frequencies in the different strains ...... 38 4.1.4 Evolution of the symbiosis between actinorhizal plants and Frankia cluster II strains ...... 39 4.2 Candidate gene expression profiling of bacterial metabolism in root nodules of D. glomerata (Cucurbitales) and C. thyrsiflorus (Rosales) .... 40 4.2.1 Frankia nodABC genes are expressed in both types of nodules, and nodH genes are expressed in nodules of C. thyrsiflorus ...... 41 4.2.2 The expression levels of genes encoding bacterial truncated hemoglobin and catalases in C. thyrsiflorus are higher than in D. glomerata ...... 41

4.2.3 Nodules induced by members of Frankia cluster II seem to share the same nitrogen metabolism ...... 42 4.2.4 Which carbon source does the host plant provide for symbiotic Frankia in both systems? ...... 43 4.3 The N-metabolites of roots and actinorhizal nodules from Datisca glomerata ...... 46 4.3.1 Amino acid profiles of roots and nodules ...... 46 4.3.2 Do D. glomerata plants change nitrogen transport forms when nodulated? ...... 47 5. Conclusions ...... 49 6. Future perspectives ...... 50 7. Acknowledgments ...... 51 8. References ...... 53

Abbreviations

AMF Arbuscular Mycorrhizal Fungi AMS Arbuscular Mycorrhizal Symbiosis ANI Average Nucleotide Identity ATP Adenosine triphosphate BNF Biological Nitrogen Fixation CcaMK Calcium- and Calmodulin-dependent serine/threonine protein Kinase CSSP Common Symbiotic Signalling Pathway DMI Does not Make Infection DNA Deoxyribonucleic acid EDGAR Efficient Database framework for comparative Ge- nome Analyses using BLAST score Ratios GABA γ-aminobutyrate gDNA Genomic Deoxyribonucleic acid GOGAT Glutamine oxoglutarate aminotransferase (Glutamate synthase) GS Glutamine Synthetase HPLC High-Performance Liquid Chromatography IS Insertion Sequence LCO Lipochitooligosaccharide LysM Lysin motif Mbp Mega base pairs Mya Million years ago NFR Nod Factor Receptor NFP Nod Factor Perception NIN Nodule Inception Nod Nodulation NORK Nodulation Receptor Kinase NSP Nodulation Signalling Pathway ORF Open Reading Frame PHYLIP Phylogeny Inference Package RNA Ribonucleic acid RT-qPCR Quantitative Reverse Transcription Polymerase Chain Reaction SYMRK Symbiosis Receptor-like Kinase TCA Tricarboxylic Acid Vc Vesicle cluster

1. Introduction

1.1 Nitrogen-fixing symbiosis in root nodules Nitrogen is most often a limiting element for plant productivity even though it is abundant in the Earth’s atmosphere (Bohlool et al. 1992; Graham and Vance 2000). Since the invention of the Haber-Bosch process, which made the industrial manufacture of ammonia economically feasible, the extensive use of synthetic nitrogen fertilizer has been boosting crop yields worldwide (reviewed by Frink et al. 1999). How- ever, the production of nitrogen fertilizer requires significant amount of fossil fuel energy (Galloway et al. 1995). In addition, it has caused several adverse environmental issues such as eutrophication and hypoxia in aquatic ecosystems (Galloway et al. 1995; Tilman et al. 2001), in- creasing greenhouse gas emission (Socolow 1999; Kim and Dale 2008) and contribution to acid rain and photochemical smog (Vitousek et al. 1997). The growing international concern for environmental issues and depletion of natural resources has led to emphasis on alternative nitrogen resources. In nature, plants acquire reduced forms of nitrogen from different sources, organic matter decomposition, conversion of atmospheric nitrogen into forms that plants could utilize by lightning and biological nitrogen fixation (BNF). BNF is a process carried out by a number of prokaryotes that could convert atmospheric nitrogen to ammonia thanks to the nitrogenase enzyme complex (Postgate 1982). Ni- trogenase consists of two proteins, the MoFe protein and the electron- transfer Fe protein, that associate and dissociate in a catalytic cycle in- volving adenosine triphosphate (ATP) hydrolysis (reviewed by Hoff- man et al. 2014). The reaction that this enzyme catalyses can be given as

- + N2 + 8e + 16ATP + 8H  2NH3 + H2 + 16ADP + 16Pi

13 BNF not only enhances agricultural production but also contributes to nitrogen input into natural ecosystems (Giller and Cadisch 1995; O’Hara et al. 2002). The contribution of symbiotic/associative nitrogen fixation to the total fixed nitrogen was estimated to 70%, as compared to 30% by non-symbiotic systems (Paul 1988). Nitrogen-fixing symbioses include the cyanobacterial partnerships, e.g., Azolla/Nostoc, and root nodule symbioses, namely the legume-rhizobia symbiosis and the actinorhizal symbiosis. The legume-rhizobia symbiosis is established between members of the legume family, and one non-legume genus Parasponia (Cannabaceae), with a polyphyletic group of Gram-negative soil proteobacteria called rhizobia. Actinorhizal symbioses are established between members from 25 plant genera belonging to eight families from three orders, collectively called actinorhizal plants, with Gram-positive soil actinobacteria of the genus Frankia (Benson and Sil- vester 1993).

1.2 Actinorhizal symbiosis 1.2.1 Frankia

The genus Frankia includes nitrogen-fixing soil actinobacteria that can establish a symbiotic relationship with a widely diverse group of perennial trees and shrubs, representing more than 200 species from eight families (Lechevalier 1994). The genus name was first used by J. Brunchorst to honour his mentor, the Swiss biologist A.B. Frank who coined the word symbiosis. Becking (1970) affirmed the name and added six more Frankia species based on host plant specificity to the existing four in the family Frankiaceae: Frankia alni, F. elaeagni, F. brunchorstii, F. discariae, F. casuarinae, F. ceanothi, F. coriariae, F. dryadis, F. purshiae, and F. cercocarpi. Based on its morphology, this family was placed in the order Actinomycetales (now: Actinobacteria) (Becking 1970) which later was confirmed based on 16S rRNA sequences (Normand et al. 1996). Phylogenetically, Frankia strains can be divided into four main clusters (Figure 1; Normand et al. 1996; Sen et al. 2014; Persson et al. 2015;

14 Nguyen et al. 2016). Members of cluster IV are either unable to induce root nodules, or induce ineffective, i.e., non-nitrogen-fixing nodules. Strains from cluster I enter nitrogen-fixing symbioses with members of the actinorhizal plant families Betulaceae, Casuarinaceae (with the exception of the genus Gymnostoma) and Myricaceae (except the genus Morella). Strains from cluster II nodulate a wide range of host plants which include four families from two different orders, the Rosaceae and the rhamnaceous genus Ceanothus from the Rosales, and Datiscaceae and Coriariaceae from the order Curcubitales. This cluster is sister to all other Frankia clusters (Sen et al. 2014; Persson et al. 2015). Strains from cluster III induce nodules on members of two families belonging to the Rosales (Rhamnaceae [with the exception of the genus Ceano- thus] and Elaeagnaceae), two genera from the order Fagales (Gymnos- toma, Morella) and occasionally on the genus Alnus.

Figure 1. Comparison of core genomes of 14 sequenced Frankia strains. Outgroups were two actinobacterial genomes, Nocardia farcinica and Mycobacterium gilvum Spyr1. The phylogenetic tree was deduced from concatenated core gene alignments using PHYLIP (Felsenstein 2005). The bar below the phylogenetic tree represents the scale of sequence divergence. The phylogenetic tree was kindly provided by Daniel Wibberg (University of Bielefeld, Germany) and Jochen Blom (Justus Liebig University, Gießen, Germany).

Table 1 summarizes the sequenced Frankia genomes available at the onset of this thesis. The size of Frankia genomes varies significantly, from the 5.0 mega base pairs (Mbp) genome of the strain CeD (cluster I) which infects Casuarina equisetifolia to the 10.45 Mbp genome of

15 R43 (cluster III) which was isolated from nodules of C. cunninghamiana but only infects members of the Eleaegnaceae family. Strains of cluster I have genome sizes between 7 and 8 Mbp (Betulaceae- and My- ricaceae-infective strains) or between 5 and 6 Mbp (Casuarinaceae-infective strains). Cluster II strains have genome sizes between 5 and 6 Mbp. Cluster III genomes range from 6.61 Mbp to 10.45 Mbp while genome sizes of cluster IV range from 6.9 Mbp to 10 Mbp. With only one exception, Frankia strain BMG5.1, members of cluster II could not be cultured despite numerous attempts. In contrast to other known Frankia strains which show optimal growth at pH values of ca. 7, BMG5.1 is alkaliphilic. It achieves optimal growth at pH 9.5 and barely shows any growth at pH 7 (Gtari et al. 2015).

Table 1. Sequenced Frankia genomes with their GenBank accession numbers and the host plants from which they were originally isolated. Size Accession Cluster Strain Host (Mbp) number ACN14a Alnus crispa 7.49 GCA_000058485.1 ACN1ag Alnus viridis 7.52 LJPA00000000 AvcI.1 Alnus viridis 7.74 LJFZ00000000

Allo2 Allocasuarina verticillata 5.35 JPHT00000000

BMG5.23 Casuarina glauca 5.27 JDWE00000000

BR Casuarina equisetifolia 5.22 LRTJ00000000 I CcI3 Casuarina cunninghamiana 5.43 CP000249 CcI6 Casurina cunninghamiana 5.57 AYTZ00000000 CeD Casuarina equisetifolia 5.00 JPGU00000000.1 CpI1-S Comptonia peregrina 7.62 JYFN00000000 CpI1-P Comptonia peregrina 7.61 LJJX00000000 G2 Casuarina equisetifolia 9.53 FAOZ00000000 QA3 Alnus nitida 7.59 AJWA00000000 Thr Casuarina cunninghamiana 5.30 JENI00000000

BMG5.1 Coriaria myrtifolia 5.79 JWIO00000000

II Dg1 Datisca glomerata 5.32 CP002801 Dg2 Datisca glomerata 5.92 FLUV01000001 BCU110501 Discaria trinevis 7.89 ARDT00000000

BMG5.12 Eleaegnus sp. 7.58 ARFH00000000 III EAN1pec Elaeagnus angustifolia 8.98 CP000820.1

EI5c Elaeagnus angustifolia 6.61 LRTK00000000

16 R43 Casuarina cunninghamiana 10.45 LFCW00000000 CN3 Coriaria nepalensis 9.97 AGJN00000000 IV DC12 Datisca cannabina 6.88 LANG00000000 EuI1cT Elaeagnus umbellata 8.81 CP002299

The nitrogenase enzyme complex is very sensitive to oxygen in vitro and in vivo. This is because one of its components, the iron-molyb- denum complex, rapidly denatures upon exposure to oxygen (Gallon 1981). This leads to the so-called oxygen dilemma of nitrogen fixation: on one hand, nitrogenase has to be protected from oxygen, on the other hand, the nitrogenase reaction requires high amounts of ATP which are best supplied by aerobic respiration. Most rhizobial strains can only fix nitrogen inside root nodules because they rely on their host plant to provide protection for their nitrogenase complex from oxygen. Beta-rhizobia can fix nitrogen in the free-living state but only under microaerobic conditions (Patriarca et al. 2002). In contrast with rhizobia, Frankia strains have the capability to protect their nitrogenase enzyme by de- veloping highly specialized structures called vesicles at the end of hyphae or short side branches, and therefore can fix nitrogen in the free- living state under aerobic conditions (Benson and Silvester 1993). Thus, in symbiosis, both partners can contribute to the protection of nitrogenase from oxygen. Frankia strains form vesicles in nodules of most host plants; the shape, septation and subcellular position of the vesicles depends on the host. i.e., the same strain can form different types of vesicles in nodules of different plant species (Huss-Danell 1997). The multi-layered envelopes of vesicles contain a high amount of hopanoids, bacterial lipids (Berry et al. 1993). Like in culture, the thick- ness of the envelope depends on the concentration of oxygen (Parsons et al. 1987; Huss-Danell 1997; Kleemann et al. 1994). In nodules of Casuarina and Allocasuarina species, the microsymbiont does not form vesicles. The infected and adjacent uninfected cortical cells have lignified cell walls which impede the diffusion of oxygen (Berg and McDowell 1988). In addition, a large amount of protein hemoglobin has been found in the infected cells of Casuarina nodules (Fleming et al. 1987; Jacobsen-Lyon et al. 1995). The purified protein possess similar characteristics as legume leghemoglobin’s, suggesting

17 that it plays a role in the transport of oxygen to the respiratory chain (Gibson et al. 1989). 1.2.2 Actinorhizal plants

Actinorhizal plants are widely Table 2. Actinorhizal plants and their distributed on Earth, from the arc- host specificity. Actinorhizal plants consist of 25 genera belonging to eight di- tic (Lawrence et al. 1967) to the verse families from three orders. Host tropics (Burleigh and Dawson plants of cluster I strains are depicted in blue. Hosts of cluster II strains are de- 1994), but primarily located in the picted in red, and hosts of cluster III temperate zone. They belong to strains are given in green. eight families from three different orders (Table 2). The actinorhizal Oder Family Genus species of the order Fagales in- Betulaceae Alnus clude members of the families Bet- Comptonia ulaceae, Casuarinaceae and Myri- Myricaceae Morella Myrica caceae. The actinorhizal Rosales Fagales Allocasuarina include members of the families Casuarina Elaeagnaceae, Rhamnaceae, and Casuarinaceae Ceuthostoma Rosaceae. The actinorhizal Cucur- Gymnostoma bitales include members of the Ceanothus families Coriariaceae and Datisca- Colletia ceae (Swensen 1996). All actino- Discaria rhizal plants are woody with the Rhamnaceae Kentrothamnus exception of two species of Da- Retanilla Talguenea tisca. Datisca glomerata (common Trevoa name: Durango root) has a genera- Rosales Elaeagnus tion time of ca. six months. This Elaeagnaceae Hippophae wind-pollinated, self-compatible Shepherdia species is native to Baja California Cercocarpus in Mexico and California, USA Chamaebatia (Davidson 1973). Another actino- Rosaceae Cowania rhizal genus that has its centre of Dryas distribution in California is Ceano- Purshia Coriariaceae Coriaria thus, belonging to the family Cucurbitales Datiscaceae Datisca Rhamnaceae (Mabberely 1988).

18 Actinorhizal plants are the major source of fixed nitrogen in diverse ecosystems including forests, swamps riparian zones, prairies and de- serts (Silvester 1976; Dawson 1986). Alnus rubra, the red alder, which is native to North America and Alnus glutinosa, the common alder, colonize riparian zones as well as nitrogen-poor wetlands (Dawson 2008). Casuarina and Allocasuarina species play an important role in sand dune stabilization and prevention of soil erosion in tropical climates and warm temperate regions from Africa (Diagne et al. 2013) to Australia (Barritt and Facelli 2001). Members of the genera Alnus, Shepherdia, Hippophae and Dryas probably played a crucial role in soil formation in vast areas of Europe, Asia and North America (Dawson 2008). Sea buckthorn (H. rhamnoides) is currently being domesticated. This is because its seed oil can be used in cosmetics and the fruits are very rich in vitamin C, carotenes and unsaturated fat and can be used in food in- dustry as well as for phytopharmaceutical purposes (Suryakumar and Gupta 2011).

1.3 Actinorhizal nodules

1.3.1 Nodule induction In the legume-rhizobia symbiosis, the best examined way of root infection is via root hairs. Just a few cases have been analysed where this infection occurs via crack entry (Capoen et al. 2010; Goormachtig et al. 2004). In actinorhizal symbiosis there are currently two known ways of root infection, intracellular infection for host plants from the order Fa- gales and intercellular infection for hosts from the order Rosales. The infection pathway in Cucurbitales has not been characterized yet. As in legumes, the host plant of the actinorhizal symbiosis determines the mechanism of infection. Intracellular infection begins with the deformation of root hairs in the presence of their respective microsymbionts (Callaham et al. 1979; Berry et al. 1986). When a Frankia hypha is trapped in a root hair curl, infection can occur (Berry and Torrey 1983). The Frankia hypha enters

19 the root in an infection thread-like structure, embedded in a plant-de- rived pectin-rich matrix. The invasion of cortical cells by infection thread-like structures is always preceded by the formation of a preinfection thread structure (Berg 1999). Concurrently, cell division in the cortex is triggered, similar to the process in legume nodule formation. Some of the newly-divided cells become infected, i.e., infection threads branch and fill the cells from the centre outward. Eventually they form the so-called prenodule which is part of the infection process of Fagales. However, the prenodule is not directly involved in nodule formation. Unlike in legume where the development of a nodule primordium takes place in the cortex, actinorhizal nodule primordia originate in the pericycle like lateral root primordia. The number of the primordia formed per prenodule varies from a few up to a dozen or more (Callaham et al. 1979). Hyphae in infection threads grow from the prenodule to the nodule primordium and infect primordium cells by branching. In the intracellular infection process, infection threads grow from cell to cell (transcellular growth). When the cells have been filled with hyphae in branching infection threads, the microsymbiont starts to differentiate vesicles wherein nitrogenase is formed and nitrogen fixation takes place (reviewed by Pawlowski and Demchenko 2012). The most obvious differences between the intracellular and the intercellular infection pathway is that in the latter, there is no root hair deformation and no prenodule formation. A Frankia hypha enters the young root by penetrating the middle lamella between epidermal cells (Racette and Torrey 1989; Berry and Sunnel 1990). The hyphae grow through intercellular spaces and proceed to colonize the apoplast of the root cortex. During these stages the intercellular spaces become filled with extensive darkly-staining matrix which contains high levels of pectic compounds and proteins (Liu and Berry 1991). Concomitantly, a nodule primordium is formed in the root pericycle. Hyphae infect primordium cells from the apoplast, and infection threads are formed directly in the infected cells. No transcellular growth of infection threads is observed. Berg et al. (1999), in the description of infection threads in actinorhizal Fagales, made a division between those containing invasive hyphae and those containing vegetative hyphae. The former invade the

20 plant cells and perform transcellular growth, while the latter branch within infected cells, and eventually forms vesicles. Both types of hyphae/infection threads are involved in intracellular infection while only vegetative hyphae are found in the intercellular pathway. The mechanism of infection in the order Curcubitales shares features with the processes in both Fagales and Rosales. No prenodules are formed, but the infection threads show transcellular growth (Berg et al. 1999). However, unlike in legumes and actinorhizal Fagales, the formation of preinfection threads does not take place (Berg et al. 1999), and the infected cells are filled with hyphae from the periphery inward. Furthermore, there is no visual difference between infection threads containing invasive or vegetative hyphae (reviewed by Pawlowski and Demchenko 2012).

1.3.2 Nodule structure Root nodules of legume and actinorhizal plants are quite different in many aspects. In legume-rhizobia nodules, the infected cells are located in the central tissue and the vascular bundles are peripheral. The nodules can be determinate or indeterminate depending on the host plant. In determinate nodules, meristemic activity stops in the early stages of development, leading to a rather spherical shape (Newcomb 1981). In- determinate nodules have a more cylindrical shape because they main- tain an active apical meristem which keeps producing new cells over the life time of the nodule. The infected cells of indeterminate nodules are arranged in a spatial developmental gradient, while the developmental gradient of the infected cells of determinate nodules is temporal. All non-legume root nodules, i.e., actinorhizal nodules and nodules of Parasponia sp., are characterized by their perennial coralloid structures. They consist of multiple lobes with each lobe resembling a mod- ified lateral root with infected cells located in the cortex and a central vascular system (Pawlowski and Bisseling 1996). As in indeterminate legume nodules, the infected cells are arranged in a developmental gradient from the apex to the base, allowing the delineation of zones. The apical meristem is considered as zone I. Below it is the infection zone (II) where infection thread-like structures containing Frankia invade

21 and branch in infected cells. The fixation zone (III) is where nitrogen is fixed by the nitrogenase enzyme complex. The last one is the senescence zone (IV) where bacteria are inactive and degraded (reviewed by Pawlowski and Demchenko 2012). In the species of the orders Fagales and Rosales, the infected cells are intermixed with the uninfected cells. However, infected cells in root nodules of the order Cucurbitales form a distinct sector located on one side of the acentric vascular system, not intermixed with uninfected cells (Figure 2) (Newcomb and Pankhurst 1982; Hafeez et al. 1984). When the host plants are grown in wet or waterlogged land, the nodules’ access to oxygen or nitrogen can be limited. To cope with gas deprivation, members of Casuarina, Comptonia, Datisca, Gymnos- toma, and Myrica can form nodule roots at the tips of their lobes due to a change of meristem activity (Silvester et al. 1990). Nodule roots show negative geotropic growth and contain large air spaces in their cortex in order to allow gas diffusion. They are not infected by Frankia thus do not contribute to nitrogen fixation.

C D A B

Figure 2. Comparison of longitudinal sections of lateral root and different nodule types. The vascular system is depicted in black. (A) shows a standard lateral root with central vascular bundle given in black, root hairs (rh), apical meristem (m) and calyptra (c). (B) shows an indeterminate legume nodule with peripheral vascular system and infected cells in the inner tissue. Due to the activity of the apical meristem (I) the cells of the inner tissue (shaded in red) are arranged in a spatial developmental gradient (Vasse et al. 1990): The prefixation zone (II) is followed by the interzone (II-III) where nitrogen fixation commences. Zone III is the nitrogen fixation zone and Zone IV is the zone of senescence. (C) shows a lobe of an actinorhizal nodule formed by a member of the Fagales or Rosales. Like a lateral root, the lobe has a central vascular bundle. The infected cells (shaded in yellow) are located in the expanded cortex, intermixed with uninfected cells. (D) shows a nodule formed by a member of the Cucurbitales. The infected cells (shaded in green) form a continuous section in the cortex, kidney-shaped in cross section, on one side of the acentric stele, and are not interspersed with uninfected cells.

22 1.3.3 Nodule metabolism Root nodules are specialized organs formed on the plant root system after communication with the bacteria. For the host plants, nodules represent strong carbon sinks and nitrogen sources. Berry and Sunell (1990) showed that uninfected prenodule cells contain large amounts of starch granules. The function of this storage is not clear but it has been shown that nodule starch is reduced during carbon-stressed situations (Vikman et al. 1990). The most common form of carbon transported in higher plants is sucrose, a disaccharide in which fructose and glucose are linked via an O- glycosidic bond (Giaquinta 1983). Sucrose can be degraded by cytosolic sucrose synthase (SucS) or by cytosolic, vacuolar or apoplastic invertases. In legume nodules, SucS is mostly responsible for sucrose degradation (Gordon et al. 1999). SucS transcripts have also been shown to be present at high levels in infected cells of actinorhizal nodules of A. glutinosa, C. glauca and D. glomerata (Van Ghelue et al. 1996; Schubert et al. 2011; Schubert et al. 2013). In legume-rhizobia symbioses, dicarboxylates, especially malate and succinate, are the main carbon source supplied to the microsymbiont by the host plant (reviewed by Udvardi and Poole 2013). More research needs to be done to determine the carbon source that Frankia utilizes in the infected cells of actinorhizal nodules. In A. glutinosa, a nodule specific dicarboxylate transporter, AgDCAT1, has been identified and characterized (Jeong et al. 2004). This transporter locates to the peri- symbiont membrane and facilitates the flow of dicarboxylates from the cytoplasm to the microsymbiont (Jeong et al. 2004). Alternatively, amino acids, i.e. Gln or Asp have been proposed as sole carbon source for Frankia strain CpI1, isolated from Comptonia peregrina root nodules (Zhang et al. 1992). In all types of root nodules, bacteria reduce atmospheric nitrogen to ammonium. In the legume-rhizobia symbiosis, ammonium assimilation does not take place inside the microsymbiont. The product of nitrogenase activity is exported to the plant cytosol where it is assimilated in the glutamine synthetase-glutamate synthase (GS-GOGAT) cycle (Temple et al. 1998). In actinorhizal root nodules from plants of two different

23 orders, A. glutinosa in Fagales and Discaria trinervis in Rosales, glutamine synthetase (GS) is highly expressed in infected cells (Guan et al. 1996; Valverde and Wall 1999), implying that the primary nitrogen assimilation mechanism is similar to that in the legume-rhizobia symbiosis. In root nodules of A. incana, bacterial GS protein cannot be detected (Lundquist and Huss-Danell 1990), indicating that ammonium is exported to the plant cytosol where assimilation takes place. However, D. glomerata of the order Cucurbitales shows a nitrogen metabolism significantly different from other types of actinorhizal nodules. GS transcripts and proteins were found in uninfected cortical cells that surrounded the infected tissue but were not detected in the infected cells (Berry et al. 2004). Gln, Glu and Arg were present at high levels in the root nodule extract (Berry et al. 2004). Detailed analyses implied that in nodules of D. glomerata, Frankia carries out primary nitrogen assimilation and exports arginine to the plant cytosol. This arginine is transported to the uninfected cells where it is broken down, and the ammonium is re-assimilated (Berry et al. 2011). Assimilated nitrogen is further processed into amino acids (glutamate, glutamine, aspartate and asparagine) or ureides (citrulline, arginine) for transport via the xylem and/or temporary storage (Schubert 1986).

1.4 Signal exchange between microsymbiont and host plant

1.4.1 Signal exchange in the legume-rhizobia symbiosis The symbiotic interaction between legumes and rhizobia has been studied extensively. Chemical communication must take place between the bacteria and the host plant root. The plant secrets flavonoids which are inducers of rhizobial nodulation genes (Phillips 1991). These compounds bind the NodD protein, a transcriptional regulator of the LysR family. NodD activates the genes that are responsible for the synthesis of rhizobial signal molecules, lipochitooligosaccharide (LCO) Nod (Nodulation) factors (reviewed by Oldroyd 2013). These molecules have a chitooligosaccharide backbone bearing an acyl chain on the non- reducing end (Mylona et al. 1995). The backbone of these compounds

24 is synthesized by proteins encoded by the nod genes nodABC, which are present in all rhizobial species. NodC represents a chitin synthase that synthesizes the N-acetyl-glucosamine oligomer. NodB, a deacety- lase, removes the acetyl group of the sugar at the non-reducing end, while the acyl transferase NodA transfers the acyl chain to the position where the acetyl group was removed. There are also several other nod genes that are involved in chemical modifications of the Nod factors, influencing host specificity (Dénarié et al. 1996). Nod factors are perceived by the plant and stimulate transcriptional changes in the plant roots that lead to nodule formation and are required for bacterial infection (reviewed by Oldroyd 2013). Nod factors are perceived by lysin motif (LysM) domain receptor- like kinases in the root hair plasma membrane. In Lotus japonicus these receptor kinases are Nod Factor Receptor 1 (NFR1) and Nod Factor Receptor 5 (NFR5) while in Medicago truncatula, LysM receptor kinase 3 (LYK3) and Nod Factor Perception (NFP) fulfil the equivalent function (Madsen et al. 2003; Limpens et al. 2003). They form a receptor complex that interacts with Nod factors secreted by the bacteria. Other receptor kinases that also participate in the recognition of Nod factor are Symbiosis Receptor-like Kinase (SYMRK) in L. japonicus and its ortholog Does not Make Infections 2 (DMI2 in M. truncatula) and Nodulation Receptor Kinase (NORK in M. sativa; Endre et al. 2002; Stracke et al. 2002). SYMRK/ DMI2/NORK are presumed to fa- cilitate the perception of Nod factors by LysM receptor complexes. SYMRK/DMI2 and LysM receptor kinases are also required in the signal exchange between arbuscular mycorrhizal fungi (AMF) and their host plant. This is unsurprising in view of the fact that AMF signal factors, the so-called Myc factors, represent LCOs or chitooligosaccha- rides (Genre et al. 2013; Maillet et al. 2011; Op den Camp et al. 2011). Downstream of the plasma membrane receptors in the signal transduction pathway are the cation channels CASTOR and POLLUX in L. japonicus (Charpentier et al. 2008) and DMI1 in M. truncatula (Peiter et al. 2007). They are located on the nuclear membrane and are essential for the symbiotic calcium oscillations that are caused by Nod (and Myc) factor signalling which play the central role in both the infection process

25 and root nodule organogenesis (Capoen et al. 2011). The calcium spiking signatures are decoded by the Calcium- and Calmodulin-dependent serine/threonine protein kinase (CCaMK/DMI3) which phosphorylates LjCYCLOPS/MtIPD3 in L. japonicus and M. truncatula, respectively (Horváth et al. 2011; Yano et al. 2008). CYCLOPS/IPD3 is a transcription factor that together with CCaMK/DMI3 activates the expression of a further transcription factor, Nodule Inception (NIN) (Marsh et al. 2007; Singh et al. 2014). Other nodulation-specific transcription factors functioning downstream of the calcium oscillations include Nodulation Signalling Pathway 1 (NSP1) and NSP2 (Singh et al. 2014). Arbuscular mycorrhizal symbiosis (AMS) were already present in the first terrestrial plants (Redecker et al. 2000). It has been concluded that when root nodule symbioses evolved, they recruited part of the arbuscular mycorrhizal signal transduction pathway (Markmann and Par- niske 2009). The signalling pathway from SYMRK/DMI2 to CY- CLOPS/IPD3 is shared between AMS and legume-rhizobia symbiosis and it is called the “common symbiotic signalling pathway” (CSSP).

1.4.2 Signal exchange in actinorhizal symbiosis There are commonalities between the signalling pathways used by rhizobia and Frankia strains. Homologs of legume SYRMK have been found in C. glauca (CgSYMRK) and D. glomerata (DgSYMRK) and could be shown to be essential for the induction of nodules by Frankia (Gherbi et al. 2008; Markmann et al. 2008). CCaMK, a key component in the CSSP, is also crucial for the formation of nodules on roots of C. glauca (Svistoonoff et al. 2013). Furthermore, an ortholog of NIN, which plays a central role in nodulation in rhizobial symbiosis has been found in C. glauca (Clavijo et al. 2015). Homologs of NIN were also found to be expressed in nodules of A. glutinosa (Hocher et al. 2011) and D. glomerata (Demina et al. 2013). These results show that the CSSP is shared between rhizobial and actinorhizal symbiosis. Since LCOs are used by both rhizobia and AM fungi to signal via the CSSP, one would expect Frankia strains to utilize the same type of

26 molecule to communicate with its host plants. However, all the sequenced genome of members from cluster I and cluster III do not contain the common genes nodABC (Normand et al. 2007). Cérémonie et al. (1999) used high-performance liquid chromatography (HPLC) to purify the putative symbiotic signal of the Alnus-infective Frankia strains ACoN24d. However, the root hair deforming factor was chemically different from rhizobial Nod factors. Recent research shows that the signal molecule produced by Frankia cluster I strain CcI3 is chitinase resistant, suggesting that it does not represent a LCO (Chabaud et al. 2016). This factor was shown to trigger calcium spiking and activate NIN expression in its host plant C. glauca (Chabaud et al. 2016). LysM receptor kinases possess the ability to rec- ognize not only LCOs but also peptidoglycans and exopolysaccharides (Willmann et al. 2011; Kawaharada et al. 2015). Therefore, it is possible that members of Frankia cluster I and III produce signalling molecules that are different from LCOs. The only Frankia genome available so far that contains nodABC genes, Candidatus Frankia datiscae Dg1, belongs to Frankia cluster II, the earliest divergent cluster (Persson et al. 2011, 2015). These genes were shown to be expressed in root nodules of D. glomerata (Persson et al. 2015). It was also shown that the phylum Actinobacteria contains a much more diverse series of nodA homologs than the ones in rhizobia (alpha- and beta-Proteobacteria), which led to the hypothesis that nodA originated and evolved in Actinobacteria and was laterally transferred to rhizobia (Persson et al. 2015). However, the second genome of a cluster II strain, BMG5.1, does not contain the canonical nod genes (Gtari et al. 2015). So the question is whether Dg1 or BMG5.1 represents an exception.

1.5 Evolution of actinorhizal symbiosis All plants able to evolve a root nodule symbiosis contain a common predisposition which is supposed to have emerged in the late Creta- ceous ca. 100 millions years ago (mya) in the common ancestor of the Rosales, Cucurbitales, Fagales and Fabales (Doyle 2016). Actinorhizal

27 plant species belong to eight families from the orders Rosales, Cucur- bitales and Fagales (Figure 2). The oldest eudicot families containing actinorhizal species include the families Betulaceae, Casuarinaceae and Myricaceae with fossils of related species from the Cretaceous (reviewed by Swensen and Benson, 2008). Pollens from the families Elaeagnaceae, Rhamnaceae and Rosaceae were dated much later, 39- 22.5 mya (Muller 1981). Although no fossil record is known for the family Datiscaceae, fossil wood of related species from India suggests that this family may have arisen in the Eocene (55-39 mya; Lankhanpal 1970). The divergence time of the species of the genus Coriaria (Cori- ariaceae) between the Northern and the Southern hemisphere was dated to 63-59 mya using molecular clock data (Yokoyama et al. 2000). Since all Coriaria species form root nodules, the evolution of the symbiosis must have preceded this date. The distribution of the bacterial symbionts has been shown to have a correlation with the distribution of the host plants (reviewed by Ben- son et al. 2004). Since Frankia cluster II is sister to all other Frankia clusters, it is of particular interest to study the evolution of the symbiotic relationships between Frankia strains and actinorhizal plants. Frankia cluster II strains have been suggested to be closely related as cross-infection by crushed nodules have been successful between Ceanothus, Coriaria, Datisca and Dryas species (Kohls et al. 1994; Torrey 1990). In addition, phylogenetic analysis of the glnA and 16S rDNA region suggested that there was a low genetic diversity among Ceanothus of the family Rhamnaceae and Datisca glomerata of the family Datisca- ceae (Vanden Heuvel et al. 2004). Nevertheless, more studies of Frankia cluster II are needed to attain further insights into the evolution of actinorhizal symbiosis.

28 2. Aim of this thesis

My research topic is the nitrogen-fixing actinorhizal root nodule symbiosis, particularly of the Frankia cluster II. This cluster is earliest divergent among Frankia strains (Persson et al. 2015). Even though several genomes of the other clusters have been sequenced (starting with Normand et al. 2007), just two genomes of members of cluster II have been sequenced so far, Candidatus Frankia datiscae Dg1 (Persson et al. 2011) and BMG5.1 (Gtari et al. 2015). While Dg1 contains the canonical nod genes nodABC, BMG5.1 and members of the other Frankia clusters do not. Therefore, one objective of this thesis was to acquire more genome sequences of members of this cluster to (i) find out whether the presence of nod genes in Dg1 is an exception; (ii) gain insights into the evolution of actinorhizal symbiosis based on comparison of these genomes. Members of cluster II nodulate a broad range of host plants belonging to four families from two orders Cucurbitales and Rosales. In order to investigate the diversity of microsymbionts in nodules of host plants from different orders, we compared the expression levels of genes encoding enzymes involved in bacterial metabolism in nodules of a member of the order Cucurbitales (Datisca glomerata) to that in nodules of a member of the order Rosales (Ceanothus thyrsiflorus). To further understand the bacterial nitrogen metabolism in actinorhizal Cucurbitales, the amino acid profiles of roots and nodules of D. glomerata were obtained and analysed.

29 3. Comments on methodology

With only one exception, Frankia cluster II strains cannot be cultured so far. This poses several technical challenges in obtaining sufficient amount of pure, high quality DNA and RNA for next-generation sequencing. Part of this thesis work was dedicated to develop methods to acquire genomic DNA (gDNA) and total RNA from nodules induced by Frankia cluster II strains. The lysis methods that worked on cultures of cluster I Frankia strain ACN14a and strain CcI3 failed on cluster II vesicle clusters; it seems cluster II cell walls are particularly difficult to break.

3.1 Frankia gDNA isolation for genome sequencing Different techniques to break Gram-positive cell wall described in the literature were applied including the traditional grinding with mortar and pestle (Kucho et al. 2009; Alloisio et al. 2010), using a tissue- lyzer (Qiagen, Hilden, Germany) with iron beads or a FastPrep BIO101 benchtop homogenizer (Qbiogene, Carslbad, CA, USA) with glass beads (Bickhart and Benson, 2011). Ultimately, the best results were obtained by grinding in a mortar with a pestle in liquid nitrogen, followed by breaking the cell walls using ultrasonic homogenizer Sono- plus HD 2070 (Bandelin Electronic, Berlin, Germany), in spite of the fact that this method also fragmented the gDNA.

3.2 Candidate gene expression profiling of bacteria in nodules of Datisca glomerata and Ceanothus thyrsiflorus In our group, bacterial and plant symbiotic transcriptomes were analysed in parallel. Therefore, total plant and bacterial RNA was isolated from nodules, sequenced and later separated by bioinformatics means.

30 Prior to transcriptome sequencing, rRNA had to be removed from the total RNA to produce an mRNA-enriched sample. This meant to remove both plant and bacterial rRNA. In the first attempt to deplete rRNA, I used the Ribo-Zero rRNA Removal Kit (Illumina, California, USA). Based on the Technical Application Staff’s advice two kits were combined, the Ribo-Zero bacterial rRNA removal kit and the Ribo-Zero kit for Plants. The two rRNA removal reagents were mixed with a ratio of 1:1 to be used in one standard reaction. However, there was very little RNA left after the experiment. To achieve a sufficient rRNA-degraded preparation without losing too much RNA, finally the Terminator™ 5´- Phosphate-Dependent Exonuclease (Epicentre, Wisconsin, USA) was utilized. This enzyme efficiently digest RNA that has a 5’-monophos- phate end without affecting other groups of RNA. However, in most cases, repeated Terminator treatments were required to remove the rRNA, and in some cases, most of mRNA was lost during these repeated treatments. After rRNA depletion, the bacterial transcriptomes were sequenced at the National Genomic Infrastructure, SciLifeLab (Stockholm, Swe- den). Altogether, five transcriptomes from Datisca glomerata nodules and three transcriptomes from Ceanothus thyrsiflorus nodules were sequenced (Table 3). The amounts of mapped reads vary significantly between different data sets, between 0.9 and 10.3 million reads. This could be due to the fact that the RNA samples had been isolated using a tis- suelyzer since the ultrasonic method had not been established at the time. Hence the transcriptomes did not provide enough reads for reliable statistical analyses. We decided to analyse the metabolic pathways of interest based on quantitative reverse transcription polymerase chain reaction (RT-qPCR) based on RNA isolated using the ultrasonic method to break bacterial structures.

31 Table 3. Bacterial transcriptomes from root nodules of D. glomerata and C. thyrsiflorus. Transcriptome Dg (Millions of reads) Ct (Millions of reads) Total Mapped Total Mapped reads reads 1 31.1 0.9 25.4 6.6 2 28.08 3.0 22.4 9.2 3 65.0 3.3 30.6 5.02 4 65.0 6.0 n.a. n.a. 5 67.6 10.3 n.a. n.a.

3.3 Amino acid profiling of roots and nodules from Datisca glomerata While the metabolism of nitrogen-fixing legume nodules has been studied extensively (reviewed by Oldroyd et al. 2011), less information is available for the situation in actinorhizal nodules. To further understand actinorhizal nitrogen metabolism, the metabolomes of nitrogenous solutes in roots and nodules of D. glomerata were obtained (part of paper IV). For nodulation, D. glomerata plants were infected with Candidatus Frankia datiscae Dg1. Nodulated plants were watered with ¼ strength nitrogen free Hoagland’s solution while uninfected plants were provided with ¼ Hoagland’s solution supplemented with 10 mM KNO3. Six weeks after inoculation, young roots (4 cm from the root tips) from un-infected plants and young roots and nodules from nodulated plants were collected. For statistical analysis, four replicates of each sample were harvested. The free amino acids were extracted according to the protocol in Hacham et al. (2002). The analysis was performed on a Hitachi L-8900 Amino Acid Analyzer (Ibaraki, Japan) at the Molec- ular Structure Facility, UC Davis Genome Center, California, USA.

32 4. Results and Discussion

4.1 Genome sequencing of Frankia cluster II strains We sequenced 13 genomes based on seven different inocula including Candidatus Frankia datiscae Dg1. These inocula were collected from different places on four different continents: Japan (Asia), France (Europe), Papua New Guinea (Oceania) and three members from the USA (North America).

4.1.1 Frankia cluster II inocula represent strain assemblages

Dg1 originated from nodules of Coriaria nepalensis in Pakistan (Persson et al. 2011) whereas BMG5.1 originated from nodules of C. japonica growing in Japan (Gtari et al. 2015). Therefore, obtaining a genome from North America in an area where the genus Coriaria is not present was likely to provide more information about the evolution of Frankia cluster II. A soil-and-nodule sample was obtained from under a growing Datisca glomerata plant in Vacaville, California, USA, and this inoculum was used to infect D. glomerata plants in a growth cham- ber. Bacterial gDNA was isolated from vesicle clusters isolated from root nodules of these plants. While Dg1 represented a single strain, the first member of Frankia cluster II from North America, Dg2, was shown to represent a metagenome consisting of two closely related major strains and one minor strain. It was not feasible to separate the two major strains Dg2a and Dg2b (ca. 98–99 %) due to their strong sequence similarity and similar abundance (ca. 55 % Dg2a, ca. 45 % Dg2b). The third strain, on the other hand, was underrepresented with ca. 1% (Nguyen et al. 2016). The fact that Dg1 represents one single strain and Dg2 consists of an assemblage of Frankia might be due to the origin and maintenance of the inocula. Since cluster II strains cannot be cultured, actinorhizal

33 plants growing in the greenhouse have to be infected by crushed nodules of other plants. Dg1 was propagated in D. glomerata for more than ten years in the greenhouse whereas Dg2 was isolated from nodules of the first series that was infected with an inoculum from the field. Hence, obtaining more inocula from other regions and host plants will help an- swering the questions (i) how many strains are present in the actinorhizal nodules; (ii) whether all the bacterial strains fix nitrogen in the nodules or only the dominant ones and (iii) whether the composition of the assemblage depends on the host plants. Five more inocula were collected and named according to their original host plants and the name of the country if needed: Cj1 from Cori- aria japonica in Japan, Cm1 from Coriaria myrtifolia in France, Crpng1 from Coriaria ruscifolia in Papua New Guinea, Dd1 from Dryas drummondii in Alaska, USA. In all of those cases the host plants represented the only endemic host species of Frankia cluster II in the region. We also added one more inoculum, Cv1 from Ceanothus veluti- nus from California where many host plant species are present (D. glomerata, Ceanothus sp., Purshia sp., Cercocarpus sp., Chamaebatia sp.). Two approaches of acquiring the bacterial gDNA were followed, either from whole root nodules (nod) or from vesicle clusters (vc). The metagenomes were named according to the name of the inoculum, the plant species that was used for bacterial propagation, and the method of gDNA extraction. For example, the metagenome Cj1_Dg_nod was sequenced from gDNA of whole nodules from D. glomerata infected with Cj1 while Cj1_Dg_vc was obtained from gDNA of vesicle clusters of nodules from the same plant species. The features of all sequenced genomes and metagenomes including the previously published Dg1 (Persson et al. 2015) are presented in Ta- ble 4.

34 Table 4. Features of genomes and metagenomes analysed in this thesis. *maintained in D. glomerata over a period of 10 years before isolation of vesicle clusters for sequencing; Persson et al. (2015); **Nguyen et al. (2016). Dg1_Dg_nod2 and Dg1_Cn_nod contain several very similar strains of which the exact relatedness could not be quantified.

A core genome tree of all genomes in Table 4 was calculated using EDGAR (Blom et al. 2009), together with several Frankia genomes from other clusters and two Micromonospora sp. genomes as outgroups (Figure 3). A clear separation could be seen between strains from North America, Eurasia and the Southern Hemisphere. Crpng1_Ca_nod, based on a cluster II inoculum from Papua New Guinea, was phylogenetically basal to all cluster II strains, and to all other Frankia strains. This finding is in accordance with previous studies where Frankia strains from New Zealand usually formed the sister branch to other Frankia strains (Benson et al. 1996; Clawson et al. 2004; Nouioui et al. 2014; Nguyen et al. 2016). Figure 3 also shows two cases of metagenomes based on the same inoculum displaying strong phylogenetic differences (Cm1_Dg_nod/Cm1_Dg_vc and Cj1_Dg_nod/Cj1_Dg_vc).

Figure 3. Core genome tree of sequenced Frankia strains from clusters I, III and IV and of all cluster II (meta-)genomes available thus far. The tree was calculated using EDGAR (Blom et al. 2009). The host plant orders and geographic origins of the original cluster II inocula are color-coded, red for Cucurbitales, blue for Rosales, green for Eurasia, pink for North Amer- ica, purple for the Southern hemisphere. Two actinobacterial genomes were used as outgroups, Micromonospora lupini Lupac 08 (Trujillo et al. 2014) and Micromonospora aurantiaca ATCC 27029 (GenBank accession nr. CP002162). The phylogenetic tree was deduced from concatenated core gene alignments using PHYLIP (Felsenstein 2005). Bootstrap values were 100 for every branch. The bar denotes 0.01 substitutions.

Comparisons of percentages of Average Nucleotide Identity (ANI) show that the Crpng1_Ca_nod genome has less than 86% ANI with genomes of all other members of Frankia cluster II. This means that Crpng1_Ca_nod, the only sequenced cluster II from the Southern Hem- isphere so far, represents a novel Frankia species. We propose this strain as type strain of Candidatus Frankia meridionalis (me.riː.di.oːˈnaː.lis. N.L. masc./fem. adj. meridionalis, southern).

36 4.1.2 Eurasian cluster II Frankia (meta-)genomes contain the canonical nod genes nodABC and North American strains also contain the sulfotransferase gene nodH 4.1.2.1 Nod gene presence in Frankia (meta-)genomes As shown in Table 4, all sequenced (meta-)genomes of Eurasian Frankia cluster II in the course of this study contain the canonical nod genes nodABC and the nltIJ operon, a homolog of the nodIJ operon which, like nodIJ, is linked to nodC (Persson et al. 2015). They are usually present in two genomic regions, the nod1 region being represented by nodA’B1A, the nod2 region by nodB2CnltIJ (Persson et al. 2015; Nguyen et al. 2016). Crpng1_Ca_nod contains nodB2’C as well as nltIJ and a carbamoyl transferase gene that shows homology to rhizobial nodU genes. In addition, all North American strains contain one or two copies of nodH, a sulfotransferase gene which in rhizobia is required for producing sulfated Nod factors. To investigate the origin of Frankia nodH genes, the amino acid sequences of the two putative NodH proteins of Dg2 together with 25 sequences from ten different rhizobial genera were used to build a dendrogram in GARLI 2.0 (Zwickl 2006) with Cy- anothece sp. PCC7424 as outgroup (Figure 4; Nguyen et al. 2016). This tree shows that the two Dg2NodH proteins are sister to each other and all other NodH proteins from Proteobacteria. The tree suggests that nodH was laterally transferred between Frankia and rhizobia. The di- rection of this transfer, however, cannot be determined based on phylogeny.

37 Figure 4. Phylogenetic tree of NodH amino acid sequences in Dg2. The tree shows that the NodH proteins of Dg2 are sister to the rhizobial proteins. The scale bar indicates the number of amino acid substitutions per site (Nguyen et al. 2016).

4.1.3 Cluster II genome (in-)stability: transposase frequencies in the different strains In order to adapt to different conditions of different environment, bacterial genomes need a certain flexibility, which is achieved by mo- bile DNA elements, horizontal gene transfer, genome rearrangements, etc. (Juhas et al. 2009). The frequency and size of insertion sequence (IS) elements of genomes of all cluster II (meta-)genomes are depicted in Figure 5. Dg1 contains more transposases than all other genomes since it is the only fully assembled genome and assembly normally breaks down in repetitive elements. The result shows that strains from North America contain higher numbers of transposases than the Eura- sian strains. The amount of IS elements in Crpng1, the Southern hemisphere representative, seems to be closer to that of North American strains; however, this result is not statistically significant.

Figure 5. Transposase frequencies. Numbers of transposase ORFs per Mbp in the cluster II Frankia (meta-)genomes are given. The genome of Candidatus Frankia Datiscae Dg1 sequenced from DNA isolated from vesicle clusters (Dg1_Dg_vc) contains at least four times more transposases in its genome than all other strains; this discrepancy is explained by the fact that this genome is the only one in the group which is not in draft stage, but fully assembled. Among the draft genomes, the Eurasian strains contain significantly less transposase ORFs than the North American strains (Wilcoxon's p=0.006).

4.1.4 Evolution of the symbiosis between actinorhizal plants and Frankia cluster II strains Cluster II host genus Coriaria currently has a disjunct distribution in four isolated geographical locations of the world. Since all known Co- riaria species can form root nodules, the evolution of the symbiosis must have predated this separation. Therefore, it can be postulated that the symbiosis between Coriaria sp. and Frankia cluster II evolved in the Southern hemisphere. Together with Coriaria species, cluster II Frankia strains probably spread to Eurasia ca. 60 mya (Yokoyama et al. 2000). Together with Datisca sp., Frankia cluster II seemingly reached North America 55-25 mya via the Bering Strait (Nguyen et al. 2016). In the Southern hemisphere, currently cluster II Frankia strains only nodulate Coriaria species. Yet, two scenarios are possible. Cluster II symbioses might have originated in the Southern hemisphere in Gond- wana in the common ancestor of Coriaria sp. and Datisca sp. In that case, cluster II strains would have reached Asia via two routes, once

39 with Coriaria sp. and once with Datisca sp. which evolved in India (Zhang et al. 2006). India separated from Gondwana ca. 120 mya and collided with Eurasia ca. 55 mya (summarized by Ali and Aitchison, 2008). Alternatively, cluster II symbioses would have originated in the Southern hemisphere in the ancestor of Coriaria sp. and extended their host range after India had collided with Eurasia and Da- tisca sp. reached Eurasia. The low level of diversity between the Eura- sian cluster II genomes sequenced in this study suggests a single origin, i.e., points toward an extension of the host range from Coriaria to include Datisca sp. After Frankia cluster II came to North America together with Datisca sp. via the Bering strait, it extended its host range to include plants from the order Rosales. An analysis of genome stability shows that with the exception of Crpng1_Ca_nod, the nod gene regions show numerous signs of earlier transposition events. These regions are highly conserved in the Eurasian strains but less in the North American strains. This is consistent with the significantly higher transposase frequency in North American than in the Eurasian metagenomes. However, the size distribution of transposases indicates that the time of transposition is in the past. The distribution in Eurasia was probably associated with a burst of transposition. Later the distribution in North America required a new burst of transposition, possibly related to the extensions of the host range that required more genomic flexibility.

4.2 Candidate gene expression profiling of bacterial metabolism in root nodules of D. glomerata (Cucurbitales) and C. thyrsiflorus (Rosales) In order to gain information on the bacterial metabolism in host plants from different orders, we compared gene expression levels of enzymes involved in bacterial metabolism from root nodules of D. glomerata (Datiscaceae, Curcubitales) and C. thyrsiflorus (Rhamnaceae, Rosales). Candidatus Frankia datiscae Dg1 was used to infect D. glomerata and Frankia strain Dg2 was used to induce nodule on C. thyrsiflorus. Gene expression levels were normalised against infC, the gene

40 encoding the bacterial translation initiation factor IF3 (Alloisio et al. 2010; Persson et al. 2015).

4.2.1 Frankia nodABC genes are expressed in both types of nodules, and nodH genes are expressed in nodules of C. thyrsiflorus With one exception, all cluster II Frankia strains sequenced so far contain the canonical nod genes nodABC (Persson et al. 2015; Gtari et al. 2015; Nguyen et al. 2016). The Californian strain Dg2 also contained two copies of gene encoding sulfotransferase nodH (Nguyen et al. 2016). In previous study, nodABC were found to be expressed in nodules of D. glomerata but nodH was not (Persson et al. 2015; Nguyen et al. 2016). Therefore, in this study nod gene expression was examined in nodules of C. thyrsiflorus. As shown in Table 5, in C. thyrsiflorus nodules, not only nodABC are expressed, but also both copies of nodH. This result might suggest that nodH is involved in host specificity in the symbiosis between Frankia cluster II and actinorhizal plants.

Table 5. Relative expression levels of Frankia nod genes from root nodules of D. glomerata (Dg) and C. thyrsiflorus (Ct). Expression levels were normalised against the initiation factor gene infC. Three biological replicates were analysed for each type of nodule. Except for nodC, the differences between the relative expression levels of nod genes in D. glomerata and C. thyrsiflorus are not statistically significant (p-value >0.05). n.a., not applicable. nodA nodB1 nodC nodH1 nodH2 Dg 0.0087 0.0130 0.0009 n.a. n.a. Ct 0.0396 0.0240 0.0064 0.0037 0.0036

4.2.2 The expression levels of genes encoding bacterial truncated hemoglobin and catalases in C. thyrsiflorus are higher than in D. glomerata In D. glomerata and C. thyrsiflorus, different oxygen protection mechanisms for the oxygen-sensitive nitrogenase enzyme complex are applied (Pawlowski et al. 2007; Strand and Laetsch 1977). In D. glomerata, the vesicles have thin envelopes but a thick blanket of mitochondria around the zone of vesicles creates a microaerophilic space (Berg et al. 1999; Silvester et al. 1999). In nodules of actinorhizal Rosales,

41 Frankia strains form non-septate vesicles with multilaminate envelopes (Strand and Laetsch 1977). A bacterial truncated hemoglobin trHbO gene was found to be expressed in symbiotic Frankia and could act as an oxygen diffusion facilitator under hypoxic conditions (Pawlowski et al. 2007). The question was which oxygen protection system was more efficient, the physiological protection by positioning of vesicles and mitochondria in D. glomerata or the multi-layered vesicle envelopes in C. thyrsiflorus. Examining of the expression levels of the bacterial hemoglobin gene trHbO showed it was ca. 15-fold higher in nodules of D. glomerata than in nodules of C. thyrsiflorus (Table 6). This is combined with significantly higher expression levels of catalase genes katA and katG in nodules of Datisca than in nodules of Ceanothus. The result suggests that the protection system of nitrogenase from oxygen in C. thyrsiflorus leads to less reactive oxygen species (ROS) stress for the symbiotic bacteria than in D. glomerata. This could be explained by the fact that Frankia metabolism is better adapted to utilizing their endogenous protection system of multilayered vesicles than the plant-controlled system of physiological protection.

Table 6. Relative expression of genes involved in oxygen protection of nitrogenase and in ROS detoxification in root nodules of D. glomerata (Dg) and C. thyrsiflorus (Ct). trHbO, truncated hemoglobin; katA, catalase; katG, catalase-peroxidase. Three biological replicates were analysed for each type of nodule. The relative expression levels of these genes are significantly different between nodules of D. glomerata and C. thyrsiflorus (p-value<0.05).

trHbO katA katG Dg 0.240 0.197 0.560 Ct 0.016 0.039 0.001

4.2.3 Nodules induced by members of Frankia cluster II seem to share the same nitrogen metabolism With the exception of D. glomerata, in all actinorhizal nodules studied so far, the assimilation of ammonium by glutamine synthetase (GS) in the glutamine synthetase/glutamate synthase (GS/GOGAT) cycle takes place in the cytosol of the infected cells. However, in D. glomerata GS accumulates in the uninfected cells surrounding the infected

42 cells (Berry et al. 2004). This result suggested that, unlike in other type of root nodules where the bacteria export ammonia directly to the plant, Frankia in D. glomerata nodules exports an assimilated form of nitrogen, most likely arginine (Berry et al. 2004, 2011). This type of nitrogen metabolism would enhance the need for carbon supply from the host plant to the intracellular bacteria. In order to find out whether the unusual nitrogen metabolism of D. glomerata (Berry et al. 2004, 2011) nodules was shared by C. thyrsiflorus nodules, expression levels of genes encoding enzymes involved in the GS-GOGAT cycle and in arginine biosynthesis were compared. The results showed that the genes encoding the bacterial enzymes of the GS- GOGAT cycle are highly expressed in nodules of both D. glomerata and C. thyrsiflorus. The expression levels of genes encoding the enzymes of the arginine biosynthetic pathway are much lower, and there are no significant differences between the two types of nodules. How- ever, to ascertain that Frankia in C. thyrsiflorus nodules also exports a metabolite of the arginine cycle, analysis of the nodule N-metabolite profile is needed for C. thyrsiflorus.

4.2.4 Which carbon source does the host plant provide for symbiotic Frankia in both systems? In legume nodules and in nodules of the actinorhizal species Alnus glutinosa (Betulaceae, Fagales) dicarboxylic acids, mainly malate and succinate, are supposed to be the main form of carbon source supplied to the microsymbionts by the host plant (Vance and Gantt 1992; Jeong et al. 2004; Udvardi and Poole 2013). These metabolites would be fed into the tricarboxylic acid (TCA) cycle, which would break down the substrates for production of ATP and reduction equivalents and also provide precursors for the primary metabolism of Frankia, e.g., to provide sugars for cell wall biosynthesis. However, in D. glomerata and probably also in C. thyrsiflorus, the plant would also have to provide carbon skeletons (ultimately, alpha-ketoglutarate, the precursor of glutamate) for ammonium assimilation. Therefore, oxaloacetate would be shuttled out of the TCA cycle to be used in gluconeogenesis via phosphoenolpyruvate carboxykinase

43 (pepck; Figure 6, 7). Alpha-ketoglutarate would be used for the synthesis of glutamate via the GS/GOGAT cycle and maybe, in case of high ammonium concentrations, also by glutamate dehydrogenase (gdh; Fig- ure 7). Since the symbiotic bacteria provide nitrogen sources for the entire host plant, a large supply with alpha-ketoglutarate would be required. So its removal from the TCA cycle would be disruptive. How- ever, Figure 7 shows that the expression levels of genes of the pathway between alpha-ketoglutarate and malate are not much lower than in the part between malate and alpha-ketoglutarate. More importantly, all cluster II Frankia strains sequenced so far lack the gene for the first enzyme of the glyoxylate shunt, isocitrate lyase. So it is unlikely that the production of alpha-ketoglutarate could be bypassed.

Figure 6. Comparison of the expression levels of genes encoding enzymes of the glycolysis pathway of Frankia from root nodules of D. glomerata (Dg) and C. thyrsiflorus (Ct). Ex- pression levels were normalised against the infC gene. Glk, glucokinase; pgi, phosphoglucose iso- merase; pfk1, phos- phofructokinase-1; fba, aldolase; gap, glycer- aldehyde 3-phosphate dehydrogenase; pgk, phosphoglycerate kinase; pgm, phosphoglycerate mutase; eno, enolase; pyk, pyruvate kinase; pepck, phosphoenolpyruvate carboxykinase. Three biological replicates were analysed for each type of nodule. Rel- ative expression levels of genes are depicted in red for Dg and in blue for Ct.

Figure 7. Comparison of expression levels of Frankia genes encoding TCA cycle enzymes from root nodules of D. glomerata (Dg) and C. thyrsiflorus (Ct). Expression levels were normalised against the infC gene. AcnA, aconitase; icd, isocitrate dehydrogenase; lsc1, succinate-CoA ligase subunit alpha; lsc2, succinate-CoA ligase subunit beta; sdhA/B/C/D, succinate dehydrogenase subunit A/B/C/D; mdh, malate dehydrogenase; gltA, citrate synthase; gdh, glutamate dehydrogenase; gltB and gltD, glutamate synthase large subunit and small subunit, respectively. Three biological replicates were analysed for each type of nodule. Relative expression levels of genes are given in red for Dg, in blue for Ct.

One explanation could be that in D. glomerata and C. thyrsiflorus, the plant provides a combination of succinate and malate to the bacteria as was suggested for legume nodules (Udvardi and Poole 2013). The citrate cycle would then work as a linear pathway from succinate to alpha-ketoglutarate. The additional input of malate from the host plant would compensate the oxaloacetate that leaves the cycle for gluconeogenesis and acetyl-CoA production. This would also explain the absence of the glyoxylate shunt in cluster II strains. The export of an assimilated form of nitrogen would require more energy-intensive transport processes. The question is why this type of nodule metabolism would evolve in two different plant/bacterial com- binations. As suggested above, Frankia cluster II strains seem to have come to North America from Asia with Datisca sp. via the Bering Strait 55-25 mya (Nguyen et al. 2016). It is possible that by the time cluster II strains had reached North America and expanded their host range to

45 Ceanothus sp. and Rosaceae host plants, they had already lost the glyoxylate shunt and had adapted to the symbiotic metabolism in actinorhizal Cucurbitales.

4.3 The N-metabolites of roots and actinorhizal nodules from Datisca glomerata

4.3.1 Amino acid profiles of roots and nodules The nitrogenous solute profiles of roots of uninfected plants as well as roots and nodules of nodulated plants are shown in Table 7. The results showed that the profiles of nitrogenous solutes in roots of uninfected plants were dominated by glutamine (51.2%), followed by glutamate (16%), γ-aminobutyrate (GABA) (3.4%), aspartate (6.2%), serine (3.4%) and alanine (2.7%). All other amino acids contributed less than 2% of the total amount. In nodulated roots, glutamine (25.1%) and glutamate (24.9%) were still dominant even though the contribution of glutamine was reduced by half in comparison with uninfected roots. As in roots of plants supplemented with nitrate, GABA (5.4%), serine (3.4%) and alanine (4.9%) played minor roles in nodulated roots. While the amount of arginine was insignificant in uninfected roots (0.9%), its contribution was 10.5% in nodulated roots. In the profiles of nodules, glutamate (27.9%) was still dominant while glutamine was present at 9.5%. Aspartate (6.6 %) and alanine (4.8 %) played minor roles, followed by serine (3.6 %) and GABA (2.5 %). In contrast to the profiles in roots, arginine (26.9%) was dominant among the nitrogenous solutes in nodules.

Table 7. Percentage of total nitrogenous solutes in D. glomerata plants. GABA, γ-aminobutyrate. (Taken from Supplementary material, Persson et al. 2016)

Percentage in Percentage in roots of unin- roots of nodu- Percentage in Amino acid fected plants lated plants nodules ala 2.7 4.9 4.8 asn 1.8 5.4 2.8 asp 6.2 10.7 6.6 arg 0.9 10.5 26.9 citrulline 0.0 0.0 0.2 ornithine 0.0 0.0 0.1 ethanolamine 6.8 0.0 0.9 GABA 3.4 5.4 2.5 gln 51.2 25.1 9.5 glu 16.0 24.9 27.9 gly 1.2 1.2 0.3 his 0.6 1.2 1.5 ile 0.7 1.0 1.2 leu 0.4 0.9 1.9 val 1.2 1.7 3.3 thr 1.6 1.7 3.3 phe 0.5 0.6 0.4 pro 0.8 0.7 0.7 ser 3.4 3.4 3.6 tyr 0.0 0.0 0.7 lys 0.5 0.6 1.0

4.3.2 Do D. glomerata plants change nitrogen transport forms when nodulated? Root nodule nitrogen metabolism is a combination of nitrogen export from the microsymbiont, nitrogen assimilation, and nitrogen storage. In all root nodules, the host plant provides the bacteria with carbon sources in exchange for nitrogen. In legume-rhizobia symbiosis, the product of nitrogenase activity, ammonium, is exported to the plant cytosol where it is assimilated (Temple et al. 1998). The primary nitrogen assimilation in root nodules of actinorhizal plants of Fagales (Alnus glutinosa and A. incana) and Rosales (Discaria trinervis) is similar to that

47 in legume nodules (Guan et al. 1996; Valverde and Wall 1999; Lundquist and Huss-Danell 1990). However, the situation was different in D. glomerata where the microsymbiont exported an assimilated form of ammonium (Berry et al. 2004, 2011). In this study, arginine was shown to be one of the major nitrogenous solutes in nodules, suggesting that D. glomerata nodules can export not only amides and their acids but also arginine. Apart from arginine, glutamine and glutamate were the major amino acids in nodules, as reported in previous studies (Berry et al. 2004). The amino acid profiles of roots of uninfected plants were dominated by glutamine and glutamate while aspartate played a minor role and arginine was present in a very low amount (0.9%). The profiles of roots from infected plants were also dominated by glutamine and glutamate followed by aspartate. However, these roots contained a significant higher amounts of arginine (10.5%) compared to uninfected roots, indicating that nodulated plants seemed to add arginine as nitrogen transport form. Hence, for D. glomerata growth conditions may lead to changes in the xylem N transport forms.

48 5. Conclusions

The range of sequenced genomes of Frankia cluster II, the earliest divergent Frankia cluster, has expanded from Asia to three more continents, North America, Europe and Oceania. The results suggest that cluster II inocula represent groups of strains which might explain their wide host range. Since all Eurasian cluster II strains analysed in this study contain the canonical nod genes nodABC and the North American strains also contain the sulfotransferase nodH, it could be concluded that the presence of the canonical nod genes is the standard in Frankia cluster II. The host plant-specific expression of nodH might suggest that this gene is linked to host specificity. Based on the analysis of transposase frequencies, it could be postulated that the distribution of Frankia in Eurasia was associated with a burst of transpositions. The distribution in North America possibly came later and required a new burst of transpositions. These events may have coincided with the ex- pansion of the host range which necessitated a higher level of genomic flexibility. Analysis of the expression levels of genes involved in metabolism of symbiotic bacteria in nodules of Datisca glomerata and Ceanothus thyrsiflorus showed that the nitrogenase protection in Ceanothus sp. nodules seemed to be more efficient than that in Datisca sp. nodules. On the other hand, there was a similarity in expression levels of genes involved in nitrogen metabolism, implying that in both systems an assimilated form of nitrogen is exported to the plant. The result was sup- ported by analyses of amino acid profiles of D. glomerata roots and nodules, which showed that the amounts of arginine in roots and nodules of nodulated plants were significantly higher than in roots of uninfected plants.

49 6. Future perspectives

This thesis has paved the way to understand the evolution of the Frankia cluster II, yet there are several questions left unanswered. - Are the canonical nod genes expressed in other host plants beside Datisca sp. and Ceanothus sp.? To answer this question, the expressions of nod genes in nodules of Coriaria sp. should be investigated. - Is the expression of nodH necessary for the infection of North American host plants other than D. glomerata? It would be necessary to infect North American host plants from the Rosales, e.g., Ceanothus sp., with Eurasian inocula. - How do the genome sequences of the South American strains fit into the phylogeny of cluster II? It is therefore essential to obtain metagenomes from Frankia cluster II inocula from Central and South Amer- ica, i.e. from strains that infect Coriaria ruscifolia which is endemic to this area, presumably brought from New Zealand (Yokoyama et al. 2000; Nouioui et al. 2014). - To obtain reliable evidence on the function of the canonical nod genes in members of cluster II, it is crucial to purify the signalling molecule and conduct bioassays with the purified products. In parallel, it is important to invest effort in determining the culture conditions for nod- gene-containing cluster II strains, e.g., Dg1, in order to set up a transformation protocol and to knock out the nod genes. - To gain additional data of carbon source supplied to the microsymbiont in symbiosis, the metabolome of symbiotic Frankia needs to be analysed. This would require non-aqueous fractionation to isolate bacterial vesicles under conditions where no enzyme activity is possible. - To test the hypothesis that the microsymbiont exports assimilated forms of ammonium in Ceanothus sp. nodules, like in D. glomerata, it is necessary to analyse the amino acid profiles of roots and nodules of a candidate of the genus Ceanothus.

50 7. Acknowledgments

I would like to thank the Department of Ecology, Environment and Plant Sciences (DEEP) at Stockholm University for giving me the opportunity to conduct my doctoral study in this beautiful city and this nice country. For me, it has been a life-changing experience. I would like to express my deep gratitude to my supervisor, Dr. Katharina Pawlowski. “Despite numerous attempts”, I was not successful in convincing her to name Dg2 as Frankia nguyeni. I am still really grateful for her valuable guidance and kind support. Without her, my encounter with the fascinating and challenging research topic of actinorhizal symbiosis would have not happened. I would like to thank Dr. Alison Berry at the University of California, Davis, for providing me the opportunity to work in her lab and her scientific support. I would also like to thank Dr. Daniel Wibberg and other co-authors for their essential collaboration. I am very grateful to Dr. Lisbeth Jonsson and Dr. Edouard Pesquet for their constructive criticism of my thesis. I would also like to thank my opponent Dr. Peter Young and the committee members Dr. Roger Fin- lay, Dr. Anita Sellstedt and Dr. Per-Olof Lundquist for their time and the scientific input. Special thanks go to the current and former members of the Paw- lowski’s group: Marco, Irina, Pooja, Tomas, and Theo for their cooper- ation and stimulating discussions. Their encouragement during my stressful times is very much appreciated. I would like to thank my co-supervisor Dr. Sophia Ekengren, Dr. Ulla Rasmussen, Dr. Sara Rydberg, Denis, Sandra, Andreas, Sara, Aleksan- dra, Lina, Paul, Andrea, and other colleagues for their help during my thesis. I am grateful to Peter, Anna, and Ingela for taking care of my experimental plants in the greenhouse. I would also like to thank all the colleagues at DEEP for fostering a friendly and positive working environment.

51 I am heartily thankful to my beloved ones, Nhật Quang, Lisa, Xiaoyun, Mai Trang, Nguyệt Minh, and Rickard for sharing my joyful and tearful moments. I really enjoy their companionship on my journeys around Europe. Though we are not always together, memories of all of them remain deep within me. I especially want to thank my dear friend Katha who has always been there for me through thick and thin. I would also like to use this opportunity to extend my gratitude to my family and relatives. Con xin chân thành cảm ơn mọi người đã thương yêu, dạy dỗ và chăm sóc con từ lúc nhỏ cho đến lúc trưởng thành. Cảm ơn em Bích Tuyền đã hết lòng chia sẻ và giúp đỡ chị. This thesis would have not been accomplished without the contribution of many of my colleagues, relatives, and friends. To all whose names have not been mentioned in this acknowledgement: Thank You!

52 8. References

Ali JR, Aitchison JC (2008) Gondwana to Asia: Plate tectonics, paleogeogra- phy and the biological connectivity of the Indian sub-continent from the Middle Jurassic through latest Eocene (166–35 Ma). Earth-Sci Rev 88:145-166 Alloisio N, Queiroux C, Fournier P, Pujic P, Normand P, Vallenet D, Médigue C, Yamaura M, Kakoi K, Kucho K (2010) The Frankia alni symbiotic transcriptome. Mol Plant Microbe Interact 23:593-607. Barritt AR, Facelli JM (2001) Effects of Casuarina pauper litter and grove soil on emergence and growth of understorey species in arid lands of South Australia. J Arid Environ 49:569-579. Becking JH (1970) Frankiaceae fam. nov. (Actinomycetales) with one new combination and six new species of the genus Frankia Brunchorst 1886, 174. Int J System Bacteriol 20:201-220. Benson DR, Silvester WB (1993) Biology of Frankia strains, actinomycete symbionts of actinorhizal plants. Microbiol Rev 57:293-319. Benson DR, Stephens DW, Clawson ML, Silvester WB (1996) Amplification of 16S rRNA genes from Frankia strains in root nodules of Ceanothus griseus, Coriaria arborea, Coriaria plumosa, Discaria toumatou, and Purshia tridentata. Appl Environ Microbiol 62:2904-2909. Benson DR, Vanden Heuvel BD, Potter D (2004) Actinorhizal symbioses: Di- versity and biogeography. In: Gillings M (ed) Plant microbiology, pp 99- 129. BIOS Scientific Publishers Ltd., Oxford, UK. Berg RH (1999) Frankia forms infection threads. Can J Bot 77:1327-1333. Berg RH, Langenstein B, Silvester WB (1999) Development in the Datisca- Coriaria nodule type. Can J Bot 77:1334-1350. Berg RH, McDowell L (1988) Cytochemistry of the wall of infected cells in Casuarina actinorhizae. Can J Bot 66: 2038-2047. Berry AM, Harriott OT, Moreau RA, Osman SF, Benson DR, Jones AD (1993) Hopanoid lipids compose the Frankia vesicle envelope, presump- tive barrier of oxygen diffusion to nitrogenase. Proc Natl Acad Sci USA 90:6091-6094. Berry AM, McIntyre L, McCully ME (1986) Fine structure of root hair infection leading to nodulation in the Frankia-Alnus symbiosis. Can J Bot 64:292-305. Berry AM, Mendoza-Herrera A, Guo Y-Y, Hayashi J, Persson T, Barabote R, Demchenko K, Zhang S, Pawlowski K (2011) New perspectives on nodule

53 nitrogen assimilation in actinorhizal symbioses. Funct Plant Biol 38:645- 652. Berry AM, Murphy TM, Okubara PA, Jacobsen KR, Swensen SM, Pawlowski K (2004) Novel expression pattern of cytosolic Gln synthetase in nitrogen- fixing root nodules of the actinorhizal host, Datisca glomerata. Plant Phys- iol 135:1849-1862. Berry AM, Sunell LA (1990) The infection process and nodule development. In: Schwintzer CR, Tjepkema JD (eds) The Biology of Frankia and Ac- tinorhizal Plants, pp 61-81. Academic Press, San Diego USA. Berry AM, Torrey JG (1983) Root hair deformation in the infection process of Alnus rubra. Can J Bot 61:2863-2876. Bickhart DM, Benson DR (2011) Transcriptomes of Frankia sp. strain CcI3 in growth transitions. BMC Microbiology 11:192. Blom J, Albaum SP, Doppmeier D, Pühler A, Vorhölter F-J, Zakrzewski M, Goesmann A (2009) EDGAR: a software framework for the comparative analysis of prokaryotic genomes. BMC Bioinformatics 10:154. Bohlool BB, Ladha JK, Garrity DP, George T (1992) Biological nitrogen fixation for sustainable agriculture: A perspective. Plant and Soil 141:1-11. Burleigh SH, Dawson JO (1994) Occurrence of Myrica-nodulating Frankia in Hawaiian volcanic soils. Plant Soil 164:283-289. Burleigh S, Torrey JG (1990) Effectiveness of different Frankia cell types as inocula for the actinorhizal plant Casuarina. Appl Environ Microbiol 8:2565-2567. Callaham D, Newcomb W, Torrey JG, Peterson RL (1979) Root hair infection in actinomycete-induced root nodule initiation in Casuarina, Myrica, and Comptonia. Bot Gaz (Suppl) 140:S1-S9. Capoen W, Jongho S, Wysham D, Otegui MS, Venkateshwaran M, Hirsch S, Miwa H, Downie JA, Morris RJ, Ané J-M, Oldroyd GED (2011) Nuclear membranes control symbiotic calcium signaling of legumes. Proc Natl Acad SciUSA 108:14348-14353. Capoen W, Oldroyd G, Goormachtig S, Holsters M (2010) Sesbania rostrata: a case study of natural variation in legume nodulation. New Phytologist 186:340-345. Cérémonie H, Debelle F, Fernandez MP (1999) Structural and functional comparison of Frankia root hair deforming factor and rhizobia Nod factor. Can J Bot 77:1293-1301. Chabaud M, Gherbi H, Pirolles E, Vaissayre V, Fournier J, Moukouanga D, Franche C, Bogusz D, Tisa LS, Barker DG, Svistoonoff S (2016) Chi- tinase-resistant hydrophilic symbiotic factors secreted by Frankia activate both Ca2+ spiking and NIN gene expression in the actinorhizal plant Casu- arina glauca. New Phytol 209:86-93. Charpentier M, Bredemeier R, Wanner G, Takeda N, Schleiff E, Parniske M (2008) Lotus japonicus CASTOR and POLLUX are ion channels essential for perinuclear calcium spiking in legume root endosymbiosis. Plant Cell 20:3467-3479.

54 Clavijo F, Diedhiou I, Vaissayre V, Brottier L, Acolatse J, Moukouanga D, Crabos A, Auguy F, Franche C, Gherbi H, Champion A, Hocher V, Barker D, Bogusz D, Tisa LS, Svistoonoff S (2015) The Casuarina NIN gene is transcriptionally activated throughout Frankia root infection as well as in response to bacterial diffusible signals. New Phytol 208:887-903. Clawson ML, Bourret A, Benson DR (2004) Assessing the phylogeny of Frankia – actinorhizal plant nitrogen-fixing root nodule symbioses with Frankia 16S rRNA and glutamine synthetase gene sequences. Mol Phylo- genet Evol 31:131-138. Davidson C (1973) An anatomical and morphological study of Datiscaceae. Aliso 8: 49-110. Dawson JO (1986) Actinorhizal plants: Their use in forestry and agriculture. Outlook Agr 15:202-208. Dawson JO (2008) Ecology of actinorhizal plants. In: Pawlowski K, Newton WE (eds) Nitrogen-Fixing Actinorhizal Symbioses, pp 199-234. Springer, New York, USA. Demina IV, Persson T, Santos P, Plaszczyca M, Pawlowski K (2013) Com- parison of the Nodule vs. Root Transcriptome of the Actinorhizal Plant Datisca glomerata: Actinorhizal nodules contain a specific class of defensins. PLoS ONE 8:e72442. Dénarié J, Debellé F, Promé J-C (1996) Rhizobium lipochitooliosaccharide nodulation factors: Signaling molecules mediating recognition and morphogenesis. Annu Rev Biochem 65:503-535. Diagne N, Diouf D, Svistoonoff S, Kane A, Noba K, Franche C, Bogusz D, Duponnois R (2013) Casuarina in Africa: distribution, role and importance of arbuscular mycorrhizal, ectomycorrhizal fungi and Frankia on plant development. J Environ Manage 128:204-209. Doyle (2016) Chasing unicorns: Nodulation origins and the paradox of nov- elty. Am J Bot 103:1865-1868. Endre G, Kereszt A, Kevei Z, Mihacea S, Kaló P, Kiss GB (2002) A receptor kinase gene regulating symbiotic nodule development. Nature 417:962- 966. Felsenstein J. PHYLIP (Phylogeny Inference Package). 2005. Available from: http://evolution.genetics.washington.edu/phylip.html Fleming AI, Wittenberg JB, Dudman WF, Appleby CA (1987) The purification, characterization and ligand-binding kinetics of hemoglobins from root nodules of the nonleguminous Casuarina glauca-Frankia symbiosis. Biochim. Biophys. Acta 911:209-220. Frink CR, Waggoner PE, Ausubel SH (1999) Nitrogen fertilizer: retrospect and prospect. Proc Natl Acad Sci USA 96:1175-1180. Gallon JR (1981) The oxygen sensitivity of nitrogenase: a problem for bio- chemists and microorganisms. Trends Biol Sci 6:19-23. Galloway JN, Schlesinger WH, Levy H, Michaels A, Schnoor JL (1995) Ni- trogen fixation: anthropogenic enhancement-environmental response. Glob Biogeochem Cycles 9:235-252.

55 Genre A, Chabaud M, Balzergue C, Puech-Pagès V, Novero M, Rey T, Four- nier J, Rochange S, Bécard G, Bonfante P, Barker DG (2013) Short-chain chitin oligomers from arbuscular mycorrhizal fungi trigger nuclear Ca2+ spiking in Medicago truncatula roots and their production is enhanced by strigolactone. New Phytol 198:190-202. Giaquinta RT (1983) Phloem loading of sucrose. Annu Rev Plant Physiol 34:347-387. Gibson QH, Wittenberg JB, Wittenberg BA, Bogusz D, Appleby CA (1989) The kinetics of ligand binding to plant hemoglobins. J Biol Chem 264:100- 107. Giller KE, Cadisch G (1995) Future benefits from biological nitrogen fixation: An ecological approach to agriculture. Plant and Soil 174:255-277. Gherbi H, Markmann K, Svistoonoff S, Estevan J, Autran D, Giczey G, Au- guy F, Péret B, Laplaze L, Franche C, Parniske M, Bogusz D (2008) SymRK defines a common genetic basis for plant root endosymbioses with arbuscular mycorrhiza fungi, rhizobia, and Frankia bacteria. Proc Natl Acad Sci USA 105:4928-4932. Goormachtig S, Capoen W, James EK, Holsters M (2004) Switch from intracellular to intercellular invasion during water stress-tolerant legume nodulation. PNAS 101:6303-6308. Gordon AJ, Minchin FR, James CL, Komina O (1999) Sucrose synthase in legume nodules is essential for nitrogen fixation. Plant Physiol 120:867- 878. Graham PH, Vance CP (2000) Nitrogen fixation in perspective: an overview of research and extension needs. Field Crop Res 65:93-106. Gtari M, Ghodhbane-Gtari F, Nouioui I, Ktari A, Hezbri K, Mimouni W, Sbissi I, Ayari A, Yamanaka T, Normand P, Tisa LS, Boudabous A (2015) Cultivating the uncultured: growing the recalcitrant cluster-2 Frankia strains. Sci Rep 5:13112. Guan C, Ribeiro A, Akkermans AD, Jing Y, van Kammen A, Bisseling T, Pawlowski K (1996) Nitrogen metabolism in actinorhizal nodules of Alnus glutinosa: expression of glutamine synthetase and acetylornithine transam- inase. Plant Mol Biol 32:1177-1184. Hacham Y, Avraham T, Amir R (2002) The N-terminal region of Arabidopsis cystathionine gamma-synthase plays an important regulatory role in me- thionine metabolism. Plant Physiol 128:454-462. Hafeez F, Akkermans ADL, Chaudhary AH (1984) Observations on the ultra- structure of Frankia sp. in root nodules of Datisca cannabina L. Plant Soil 79:383-402. Hocher V, Alloisio N, Auguy F, Fournier P, Doumas P, Pujic P, Gherbi H, Queiroux C, Da Silva C, Wincker P, Normand P, Bogusz D (2011) Tran- scriptomics of actinorhizal symbioses reveals homologs of the whole common symbiotic signaling cascade. Plant Physiol 156:700-711.

56 Hoffman BM, Lukoyanov D, Yang Z-Y, Dean DR, Seefeldt LC (2014) Mech- anism of nitrogen fixation by nitrogenase: The next stage. Chem Rev 114:4041-4062. Horváth B, Yeun LH, Domonkos A, Halász G, Gobbato E, Ayaydin F, Miró K, Hirsch S, Sun J, Tadege M, Ratet P, Mysore KS, Ané JM, Oldroyd GE, Kaló P (2011) Medicago truncatula IPD3 is a member of the common symbiotic signaling pathway required for rhizobial and mycorrhizal symbioses. Mol Plant Microbe Interact 24:1345-1358. Husemann P, Stoye J (2010) r2cat: synteny plots and comparative assembly. Bioinforma Oxf Engl 26:570-571. Huss-Danell K (1997) Actinorhizal symbioses and their N2 fixation. Tansley Review No. 93. New Phytol 136:375-405. Jacobsen-Lyon K, Jensen EO, Jørgensen JE, Marcker KA, Peacock WJ, Den- nis ES (1995) Symbiotic and nonsymbiotic hemoglobin genes of Casua- rina glauca. The Plant Cell 7:213-223. Jeong J, Suh S, Guan C, Tsay Y-F, Moran N, Oh CJ, An CS, Demchenko KN, Pawlowski K, Lee Y (2004) A nodule-specific dicarboxylate transporter from alder is a member of the peptide transporter family. Plant Physiol 134:969-978. Juhas M, van der Meer JR, Gaillard M, Harding RM, Hood DW, Crook DW (2009) Genomic islands: tools of bacterial horizontal gene transfer and evolution. FEMS Microbiol Rev 33:376-393. Kawaharada Y, Kelly S, Nielsen MW, Hjuler CT, Gysel K, Muszyński A, Carlson RW, Thygesen MB, Sandal N, Asmussen MH, Vinther M, Ander- sen SU, Krusell L, Thirup S, Jensen KJ, Ronson CW, Blaise M, Radutoiu S, Stougaard J (2015) Receptor-mediated exopolysaccharide perception controls bacterial infection. Nature 523:308-312. Kim S, Dale BE (2008) Effects of nitrogen fertilizer application on greenhouse gas emissions and economics of corn production. Environ Sci Technol 42:6028-6033. Kleemann G, Kellner R, Poralla K (1994) Purification and properties of the squalene-hopene cyclase from Rhodopseudomonas palustris, a purple non- sulfure bacterium producing hopanoids and tetrahymanol. Biochim Bio- phys Acta 1210:317-320. Kohls SJ, Thimmapuram J, Buschena CA, Paschke MW, Dawson JO (1994) Nodulation patterns of actinorhizal plants in the family Rosaceae. Plant Soil 162:229-239. Kucho K, Kakoi K, Yamaura M, Higashi S, Uchiumi T, Abe M (2009) Tran- sient transformation of Frankia by fusion marker genes in liquid culture. Microbes Environ 24:231-240. Lankhanpal RN (1970) Tertiary floras of India and their bearing on the histor- ical geology of the region. Taxon 19:675-694. Lawrence DB, Schoenike RE, Quispel A, Bond G (1967) The role of Dryas drummondii in vegetation development following ice recession at Glacier

57 Bay, Alaska, with special reference to its nitrogen fixation by root nodules. J. Ecol 55:793-813. Lechevalier MP (1994) Taxonomy of the genus Frankia (Actinomycetales). Int J Syst Bacteriol 44:1-8. Limpens E, Franken C, Smit P, Willemse J, Bisseling T, Geurts R (2003) LysM domain receptor kinases regulating rhizobial Nod factor-induced infection. Science 302:630-633. Liu Q, Berry AM (1991) The infection process and nodule initiation in the Frankia-Ceanothus root nodule symbiosis: A structural and histochemical study. Protoplasma 163:82-92. Lundquist P-O, Huss-Danell K (1990) Immunological studies of glutamine synthetase in Frankia – Alnus incana symbioses. FEMS Microbiol Lett 91:141-146. Mabberely DJ (1988) The Plant-book. Cambridge University Press: Cam- bridge, UK. Madsen EB, Madsen LH, Radutoiu S, Olbryt M, Rakwalska M, Szczyglowski K, Sato S, Kaneko T, Tabata S, Sandal N, Stougaard J (2003) A receptor kinase gene of the LysM type is involved in legume perception of rhizobial signals. Nature 425:637-640. Maillet F, Poinsot V, André O, Puech-Pagès V, Haouy A, Gueunier M, Cromer L, Giraudet D, Formey D, Niebel A, Martinez EA, Driguez H, Bé- card G, Dénarié J (2011) Fungal lipochitooligosaccharide symbiotic signals in arbuscular mycorrhiza. Nature 469:58-63. Markmann K, Giczey G, Parniske M (2008) Functional adaptation of a plant receptor-kinase paved the way for the evolution of intracellular root symbioses with bacteria. PLoS Biol 6:e68. Markmann K, Parniske M (2009) Evolution of root endosymbiosis with bacteria: How novel are nodules? Trends Plant Sci 14:77-86. Marsh JF, Rakocevic A, Mitra RM, Brocard L, Sun J, Eschstruth A, Long SR, Schultze M, Ratet P, Oldroyd GED (2007) Medicago truncatula NIN is essential for rhizobial-independent nodule organogenesis induced by auto- active calcium/calmodulin-dependent protein kinase. Plant Physiol 144:324-335. Muller J (1981) Fossil pollen records of extant angiosperms. Bot Rev 47:1- 142. Mylona P, Pawlowski K, Bisseling T (1995) Symbiotic nitrogen fixation. Plant Cell 7:869-885. Newcomb W, Pankhurst CE (1982) Fine structure of actinorhizal nodules of Coriaria arborea (Coriariaceae). New Zeal J Bot 20:93-103. Newcomb W (1981) Nodule morphogenesis and differentiation. Int Rev Cytol 13:S247-S297. Nguyen TV, Wibberg D, Battenberg K, Blom J, Vanden Heuvel B, Berry AM, Kalinowski J, Pawlowski K (2016). An assemblage of Frankia cluster II strains from California contains the canonical nod genes and also the sulfotransferase gene nodH. BMC Genomics 17:796.

58 Normand P, Lapierre P, Tisa LS, Gogarten JP, Alloisio N, Bagnarol E, Bassi CA, Berry AM, Bickhart DM, Choisne N, Couloux A, Cournoyer B, Cruveiller S, Daubin V, Demange N, Francino MP, Goltsman E, Huang Y, Kopp OR, Labarre L, Lapidus A, Lavire C, Marechal J, Martinez M, Mas- tronunzio JE, Mullin BC, Niemann J, Pujic P, Rawnsley T, Rouy Z, Schenowitz C, Sellstedt A, Tavares F, Tomkins JP, Vallenet D, Valverde C, Wall LG, Wang Y, Medigue C, Benson DR (2007) Genome characteristics of facultatively symbiotic Frankia sp. strains reflect host range and host plant biogeography. Genome Res 17:7-15. Normand P, Orso S, Cournoyer B, Jeannin P, Chapelon C, Dawson J, Evtu- shenko L, Misra AK (1996) Molecular phylogeny of the genus Frankia and related genera and emendation of the family Frankiaceae. Int J Syst Bacteriol 46:1-9. Nouioui I, Ghodhbane-Gtari F, Fernandez MP, Boudabous A, Normand P, Gtari M (2014) Absence of cospeciation between the uncultured Frankia microsymbionts and the disjunct actinorhizal Coriaria species. BioMed Res Int. 2014:924235. O’Hara GW, Howieson JG, Graham PH (2002) Nitrogen fixation and agricultural practice. In: Leigh GJ (eds) Nitrogen fixation at the Millennium, pp 391-420. Elsevier Science, Amsterdam, The Netherlands. Oldroyd GE (2013) Speak, friend, and enter: signalling systems that promote beneficial symbiotic associations in plants. Nat Rev Microbiol 11:252-263. Oldroyd GE, Murray JD, Poole PS, Downie JA (2011) The rules of engage- ment in the legume-rhizobial symbiosis. Annu Rev Genet 45:119-144. Op den Camp R, Streng A, De Mita S, Cao Q, Polone E, Liu W, Ammiraju JS, 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. Parsons R, Silvester WB, Harris S, Gruijters WTM, Bullivant S (1987) Frankia vesicles provide inducible and absolute oxygen protection for nitrogenase. Plant Physiol 83:728-31. Patriarca EJ, Tatè R, Iaccarino M (2002). Key Role of Bacterial NH4+ Metab- olism in Rhizobium-Plant Symbiosis. Microbiol Mol Biol Rev 66:203-222. Paul EA (1988) Towards the year 2000: directions for future nitrogen research. In: Wilson JR (ed) Advances in Nitrogen Cycling in Agricultural Ecosystems, pp 417-425. C A B International, Wallingford, UK. Pawlowski K, Bisseling T (1996) Rhizobial and actinorhizal symbioses: what are the shared features? Plant Cell 8:1899-1913. Pawlowski K, Demchenko KN (2012) The diversity of actinorhizal symbiosis. Protoplasma 249:967-979. Pawlowski K, Jacobsen KR, Alloisio N, Denison RF, Klein M, Winzer T, Sirrenberg A, Guan C, Berry AM (2007) Truncated hemoglobins in actinorhizal nodules of Datisca glomerata. Plant Biol 9:776-785.

59 Peiter E, Sun J, Hecjmann AB, Venkateshwaran M, Riely BK, Otegui MS, Edwards A, Freshiyr G, Hahn MG, Cook DR, Sanders D, Oldroyd GED, Downie A, Ané J-M (2007) The Medicago truncatula DMI1 protein mod- ulates cytosolic calcium signaling. Plant Physiol 145:192-203. Persson T, Battenberg K, Demina IV, Vigil-Stenman T, Vanden Heuvel B, Pujic P, Facciotti MT, Wilbanks EG, O’Brien A, Fournier P, Hernandez MAC, Herrera AM, Médigue C, Normand P, Pawlowski K, Berry AM (2015) Candidatus Frankia datiscae Dg1, the actinobacterial microsymbiont of Datisca glomerata, expresses the canonical nod genes nodABC in symbiosis with its host plant. PLoS One 10:e0127630. Persson T, Benson DR, Normand P, Vanden Heuvel B, Pujic P, Teshima H, Bruce D, Detter C, Tapia R, Han J, Woyke T, Pitluck S, Pennacchio L, Nolan M, Ivanova N, Pati A, Land ML, Pawlowski K, Berry AM (2011) Genome sequence of "Candidatus Frankia datiscae" Dg1, the uncultured microsymbiont from nitrogen-fixing root nodules of the dicot Datisca glomerata. J Bacteriol 193:7017-7018. Persson T, Nguyen TV, Alloisio N, Pujic P, Berry AM, Normand P, Paw- lowski K (2016) The N-metabolites of roots and actinorhizal nodules from Alnus glutinosa and Datisca glomerata: can D. glomerata change N- transport forms when nodulated? Symbiosis 70:149-157. Postgate JR (1982) Biological nitrogen fixation: fundamentals. Phil Trans R Soc Lond B 296:375-385. Phillips DA (1991) Flavonoids: plant signals to soil microbes. Rec. Adv. Phy- tochem. 26:l-33. Racette S, Torrey JG (1989) Root nodule initiation in Gymnostoma (Casuari- naceae) and Shephardia (Elaeagnaceae) induced by Frankia strain HFPGpI1. Can J Bot 67:2873-2879. Redecker D, Kodner R, Graham LE (2000) Glomalean fungi from the Ordo- vician. Science 289:1920-1921. Sen A, Daubin V, Abrouk D, Gifford I, Berry AM, Normand P (2014) Phy- logeny of the class Actinobacteria revisited in the light of complete genomes. The orders ‘Frankiales’ and Micrococcales should be split into co- herent entities: proposal of Frankiales ord. nov., Geodermatophilales ord. nov., Acidothermales ord. nov. and Nakamurellales ord. nov. Int J Syst Evol Microbiol 64:3821-3832. Schubert KR (1986) Products of biological nitrogen fixation in higher plants: synthesis, transport, and metabolism. Annual Review of Plant Physiology 37:539-574. Schubert M, Koteyeva NK, Wabnitz PW, Santos P, Büttner M, Sauer N, Dem- chenko K, Pawlowski K (2011) Plasmodesmata distribution and sugar par- titioning in nitrogen-fixing root nodules of Datisca glomerata. Planta 233:139-152. Schubert M, Koteyeva NK, Zdyb A, Santos P, Voitsekhovskaja OV, Dem- chenko KN, Pawlowski K (2013) Lignification of cell walls of infected cells in Casuarina glauca nodules that depend on symplastic sugar supply

60 is accompanied by reduction of plasmodesmata number and narrowing of plasmodesmata. Physiol Plant 147:524-540. Silvester WB (1976) Ecological and economic significance of the non-legume symbiosis. In: Newton WE and Nyman (eds) Proceedings of the First In- ternational Symposium on Nitrogen Fixation, pp 489-506. Washington University Press, Pullman, Washington D.C, USA. Silvester WB, Harris SL, Tjepkema JD (1990) Oxygen regulation and hemoglobin. In: Schwintzer CR, Tjepkema JD (eds) The Biology of Frankia and Actinorhizal Plants, pp 157-176. Academic Press, San Diego, USA. Silvester WB, Langenstein B, Berg RH (1999) Do mitochondria provide the oxygen diffusion barrier in root nodules of Coriaria and Datisca? Can J Bot 77:1358-1366. Singh S, Katzer K, Lambert J, Cerri M, Parniske M (2014) CYCLOPS, a DNA-binding transcriptional activator, orchestrates symbiotic root nodule development. Cell Host Microbe 15:139-152. Socolow RH (1999) Nitrogen management and the future of food: lessons from the management of energy and carbon. Proc Natl Acad Sci USA 96: 6001-6008. Stracke S, Kistner C, Yoshida S, Mulder L, Sato S, Kaneko T, Tabata S, San- dal N, Stougaard J, Szczyglowski K, Parniske M (2002) A plant receptor- like kinase required for both bacterial and fungal symbiosis. Nature 417:959-962. Strand R, Laetsch W (1977) Cell and endophyte structure of the nitrogen-fixing root nodules of Ceanothus integerrimus H. & A. I. Fine structure of the nodule and its endosymbiont. Protoplasma 93:165-178. Suryakumar G, Gupta A (2011) Medicinal and therapeutic potential of sea buckthorn (Hippophae rhamnoides L.). J Ethnopharmacol 138:269-278. Svistoonoff S, Benabdoun FM, Nambiar-Veetil M, Imanishi L, Vaissayre V, Cesari S, Diagne N, Hocher V, de Billy F, Bonneau J, Wall L, Ykhlef N, Rosenberg C, Bogusz D, Franche C, Gherbi H (2013) The independent ac- quisition of plant root nitrogen-fixing symbiosis in Fabids recruited the same genetic pathway for nodule organogenesis. PLoS One 8:e64515. Swensen SM (1996) The evolution of actinorhizal symbioses: evidence for multiple origins of the symbiotic association. Am J Bot 83:1503-1512. Swensen SM and Benson DR (2008) Evolution of actinorhizal host plants and Frankia endosymbionts. In: Pawlowski K and Newton E (eds) Nitrogen- fixing actinorhizal symbioses, pp 73-104. Springer, Dordrecht, The Neth- erlands. Temple SJ, Vance CP, Gantt JS (1998) Glutamate synthase and nitrogen assimilation. Trends Plant Sci 3:51-56. Tilman D, Fargione J, Wolff B, D’Antonio C, Dobson A, Howarth R, Schindler D, Schlesinger WH, Simberloff D, Swackhamer D (2001) Fore- casting agriculturally driven global environmental change. Science 292:281-284.

61 Torrey JG (1990) Cross-inoculation groups within Frankia. In: Schwintzer CR, Tjepkema JD (eds) The biology of Frankia and actinorhizal plants, pp 83-106. Academic Press, New York, USA. Trujillo ME, Bacigalupe R, Pujic P, Igarashi Y, Benito P, Riesco R, Médigue C, Normand P (2014) Genome features of the endophytic actinobacterium Micromonospora lupini strain Lupac 08: on the process of adaptation to an endophytic life style? PLoS One 9:e108522. Udvardi M, Poole PS (2013) Transport and metabolism in legume-rhizobia symbioses. Annu Rev Plant Biol 64:781-805. Valverde C, Wall LG (1999) Time course of nodule development in the Dis- caria trinervis (Rhamnaceae)–Frankia symbiosis. New Phytol 141:345- 354. Vance CP, Gantt JS (1992) Control of nitrogen and carbon metabolism in root nodules. Physiol Plant 85:266-274. Vanden Heuvel BD, Benson DR, Bortiri E, Potter D (2004) Low genetic diversity among Frankia spp. strains nodulating sympatric populations of actinorhizal species of Rosaceae, Ceanothus (Rhamnaceae) and Datisca glomerata (Datisceae) west of the Sierra Nevada (California). Can J Mi- crobiol 50:989-1000. Van Ghelue M, Ribeiro A, Solheim B, Akkermans AD, Bisseling T, Paw- lowski K (1996) Sucrose synthase and enolase expression in actinorhizal nodules of Alnus glutinosa: comparison with legume nodules. Mol Gen Genet 250:437-446. Vasse J, de Billy F, Camut S, Truchet G (1990) Correlation between ultra- structural differentiation of bacteroids and nitrogen fixation in alfalfa nodules. J Bacteriol 172:4295-4306. Vikman P-A, Lundquist P-O, Huss-Danell K (1990) Respiratory capacity, nitrogenase activity and structural changes of Frankia, in symbiosis with Al- nus incana, in response to prolonged darkness. Planta 182:617-625. Vitousek PM, Aber JD, Howarth RW, Likens GE, Matson PA, Schindler DW, Schlesinger WH, Tilman DG (1997) Human lateration of the global nitrogen cycle: sources and consequences. Ecol Appl 7:737-750. Willmann R, Lajunen H, Erbs G, Newman MA, Kolb D, Tsuda K, Katagiri F, Fliegmann J, Bono JJ, Cullimore JV, Jehle AK, Götz F, Kulik A, Molinaro A, Lipka V, Gust AA, Nürnberger T (2011) Arabidopsis lysin-motif proteins LYM1/LYM3/CERK1 mediate bacterial peptidoglycan sensing and immunity to bacterial infection. Proc Natl Acad Sci USA 108:19824- 19829. Yano K, Yoshida S, Müller J, Singh S, Banba M, Vickers K, Markmann K, White C, Schuller B, Sato S, Asamizu E, Tabata S, Murooka Y, Perry J, Wang TL, Kawaguchi M, Imaizumi-Anraku H, Hayashi M, Parniske M (2008) CYCLOPS, a mediator of symbiotic intracellular accommodation. Proc Natl Acad Sci USA 105:20540-20545.

62 Yokoyama J, Suzuki M, Iwatsuki K, Hasebe M (2000) Molecular phylogeny of Coriaria, with special emphasis on the disjunct distribution. Mol Phy- logenet Evol 14:11-9. Zhang LB, Simmons MP, Kocyan A, Renner SS (2006) Phylogeny of the Cucurbitales based on DNA sequences of nine loci from three genomes: implications for morphological and sexual system evolution. Mol Phylogenet Evol 39:305-322. Zhang X, Benson DR (1992) Utilization of amino acids by Frankia sp. strain CpI1. Arch Microbiol 158:256-261. Zwickl DJ (2006) Genetic algorithm approaches for the phylogenetic analysis of large biological sequence datasets under the maximum likelihood crite- rion. http://hdl.handle.net/2152/2666. Accessed 26 January 2017.