Search

Nodulation experiments using Alnus glutinosa L. Gaertn. (black alder) seedlings as hosts and soil suspensions as inoculants were made to determine the presence and relative amounts of ineffective (non-nitrogen fixing but infectious) Frankia strains in 10 waterlogged soils of natural alder stands in the Netherlands. Ineffective nodules were formed with soil from six locations, while effective nodules were found at all sites. From three of these locations the majority of the nodules formed were of the ineffective type. These results suggested that ineffective strains form an important fraction of the Frankia population of wet soils under black alder vegetation. No correlation (r(2), P

Ineffective Frankia endophytes were retrieved from various wet soils by using Alnus glutinosa clones as trapping plants. No pure cultures could be isolated from these ineffective nodules. Therefore, the phylogenetic position of these endophytes was determined by sequence analysis of cloned PCR products of bacterial 16S rDNA, derived from nodules. The results showed that all nodule endophytes belong to a hitherto undescribed cluster of the Fuankia phylogenetic tree. The position of these uncultured ineffective Frankia nodule endophytes is different from that of the ineffective Fuankia isolates derived from A. glutinosa nodules, even when originating from the same geographical location. This suggests a bias in current isolation techniques. [KEYWORDS: ineffective Frankia; Alnus; 16S rDNA; phylogeny 16s ribosomal-rna; oligonucleotide probes; datisca-cannabina; genetic diversity; genus frankia; strain type; database]

We isolated and characterized an Alnus glutinosa cDNA clone, pAg13, which corresponds to a gene expressed at higher levels in nodules induced by Frankia than in roots. The deduced polypeptide sequence is rich in glutamic acid and proline and contains a putative signal peptide indicating an extracellular location of Ag13. In situ hybridization showed that ag13 is expressed in the pericycle of the nodule vascular bundle and in infected cells that exhibited degradation of the endosymbiont.

Two different types of nitrogen-fixing root nodules are known- actinorhizal nodules induced by Frankia and legume nodules induced by rhizobia. While legume nodules show a stem-like structure with peripheral vascular bundles, actinorhizal nodule lobes resemble modified lateral roots with a central vascular bundle. To compare carbon metabolism in legume and actinorhizal nodules, sucrose synthase and enolase cDNA clones were isolated from a cDNA library, obtained from actinorhizal nodules of Alnus glutinosa. The expression of the corresponding genes was markedly enhanced in nodules compared to roots. in situ hybridization showed that, in nodules, both sucrose synthase and enolase were expressed at high levels in the infected cortical cells as well as in the pericycle of the central vascular bundle of a nodule lobe. Legume sucrose synthase expression was studied in indeterminate nodules from pea and determinate nodules from Phaseolus vulgaris by using in situ hybridization.

Expression of Frankia genes involved in nitrogen fixation was studied in Alnus glutinosa nodules using the in situ hybridization technique. The results show that high level expression of nif genes does not occur immediately upon infection of cortical cells by Frankia. Also, only in the infected cells near the tips of the nodule lobes, nif genes are expressed at high levels. In the majority of infected cells, nif gene expression is rather low.

Summary Nine Frankia -like actinomycetes were isolated from root nodules of Coriaria nepalensis . All isolates formed hyphae and sporangia typical of Frankia but failed to induce nodules on Coriaria seedlings or to reduce acetylene in pure culture. Hybridization with a 16S rRNA targeted oligonucleotide probe indicated that the isolates did belong to the genus Frankia . This was confirmed for one of the isolates (Cn7) by comparison of partial 16S rRNA sequences of the cloned amplification products that showed 94.2 and 96.1% homology with those of the confirmed Frankia strains from Alnus (Ag45/Mutl5) and Casuarina (ORS 020606), respectively. Similar values were obtained when 16S rRNA sequences of Cn7 were compared with those of the amplification products obtained directly from nodules of Coriaria nepalensis . Analysis of two variable regions of 15 cloned amplification products from nodules showed two types of 16S rRNA sequences, slightly different from each other. In both variable regions, none of the isolates showed identical sequences to those obtained from nodules. The results argue strongly that the isolates are members of the genus Frankia . However, their identity as endophytes of Coriaria nodules could not be proved.

Ineffective, non-infective actinomycetous isolates obtained from actinorhizal nodules of Coriaria nepalensis and Datisca cannabina were identified as Frankia using whole cell fatty acid analysis. The isolates exhibited fatty-acid patterns very similar to those of confirmed Frankia strains from other host plants ( Alnus, Casuarina, Colletia, Comptonia, Elaeagnus and Hippophae ). All Frankia strains, including Coriaria and Datisca isolates, showed fatty-acid profiles very distinct from those of other actinomycetes used as controls ( Actinomyces, Geodermatophilus, Nocardia, Mycobacterium and Streptomyces ). For the genus Frankia , a characteristic pattern of five fatty acids (15:0; 15:1; 16:0 iso; 17:0 and 17:1) was found. These fatty acids comprised 75% or more of the total content. All Frankia strains could be placed into three subgroups. Coriaria isolates were found in the largest subgroup which contained most Frankia strains from other hosts while ineffective strains from Alnus, Elaeagnus and Datisca were distributed in all three subgroups of Frankia .

Dissociated sponge cells are easily confused with unicellular organisms. This has been an obstacle in the development of sponge-cell lines. We developed a molecular detection method to identify cells of the sponge Dysidea avara in dissociated cell cultures. The 18S ribosomal RNA gene from a Dysidea avara specimen was sequenced and compared to eukaryotic 18S rDNA sequences picked up from a proliferating cell culture that originated from a dissociated Dysidea avara specimen. Our method proved unambiguously that this was not a sponge-cell culture. Therefore, it provides a valuable tool for further research on sponge-cell cultures.

Summary Several Frankia strains have been shown to induce ineffective, i.e. non-nitrogen fixing nodules, sometimes in a host-plant dependent manner. Previous studies have demonstrated that the resistance to nodulation of Alnus glutinosa by ineffective Frankia strains is genetically determined. In this study, ineffective nodules induced on susceptible Alnus glutinosa clones by soil suspensions from a local swamp were analysed cytologically. Comparisons with effective nodules showed that ineffective nodules contain higher amounts of polyphenols than effective nodules, indicating a plant defense reaction. Polyphenols were found even in the infected cortical cells. In situ hybridization with a Frankia antisense 16S rRNA probe showed that Frankia is degraded at an early stage of development of infected cells. The mRNAs of two plant genes, ag12ar\6 ag13, which had been found to be expressed in the infected cells of effective nodules, were localized in ineffective nodules. Their expression patterns seemed to be ...

A cDNA clone, pAgthi1, encoding a homologue of yeast Thi4, which is involved in thiazole biosynthesis, was isolated from a library made from poly(A) RNA from actinorhizal nodules of Alnus glutinosa by differential screening with nodule and root cDNA, respectively. The corresponding gene, agthi1, was shown to be expressed at high levels in nodules and shoot tips of A. glutinosa, while having low expression levels in roots, flowers, and developing fruits. The function of AgThi1 was demonstrated by yeast complementation studies, in which AgThi1 was able to rescue a yeast thi4 mutant when fused to the yeast Thi4 signal peptide. In A. glutinosa nodules, high levels of agthi1 mRNA were detected in the infected cortical cells and in the pericycle of the nodule vascular system. A homologue of this cDNA, ara6/tz, was identified in Arabidopsis thaliana. ara6 maps in a region of chromosome 5 of Arabidopsis containing the tz locus which is consistent with the fact that ara6 transcription is disturbed in two tz mutant lines. ara6/tz is expressed at high levels in chloroplast-containing parenchymatic cells of leaves, inflorescence shoots and flowers of Arabidopsis, and at lower levels in the vascular system.

Two nodule cDNA clones representing genes involved in Alnus glutinosa nitrogen metabolism were analysed. ag11 encoded glutamine synthetase (GS), the enzyme responsible for ammonium assimilation, while ag118 encoded acetylornithine transaminase (AOTA), an enzyme involved in the biosynthesis of citrulline, the nitrogen transport form in Alnus. GS mRNA was found at highest levels in root nodules, where it was present in the infected cells as well as in the cells of the pericycle of the vascular system. AOTA transcripts were found at high levels in nodules, confined to the infected cells, suggesting that in nodules of A. glutinosa, citrulline biosynthesis takes place mainly in the infected cells.

The presence of Frankia strains in soil samples collected from northern areas of Pakistan was detected by inoculating Coriaria nepalensis and Datisca cannabina plants. The abundance of compatible Frankia strains in some areas was indicated by profuse nodulation of the host plants, whereas soil samples from other localities failed to result in nodulation. An oligonucleotide probe (COR/DAT) directed against the 16S rRNA gene of the endophytes of Coriaria and Datisca spp. that did not cross-react with the RNA gene of Frankia strains isolated from other hosts was developed. Genetic diversity among Frankia strains nodulating D. cannabina was determined by sequence analysis of the partial 16S rRNA gene amplified from nodules induced by soil samples from different localities by PCR. Four types of Frankia sequences and one non-Frankia sequence were detected by hybridization with a Frankia genus probe and the COR/DAT probe as well as by sequence analysis of the cloned PCR products.

Part of the 16S rRNA gene was amplified directly from uncultured endophyte populations within the root nodules of Datisca cannabina and three strains isolated from nodules of Alnus glutinosa (AgKG′84/4), Coriaria nepalensis (Cn3), and D. cannabina (Dc2). Sequence comparison based on 930 nucleotides indicated that the endophyte of D. cannabina nodules belongs to the genus Frankia and is highly related to the endophyte of C. nepalensis root nodules. The relatedness of the endophytes of C. nepalensis and D. cannabina nodules was also reflected by closely related nifH sequences amplified from the nodules. 16S rRNA sequence analysis of the noninfective strains obtained from both D. cannabina (Dc2) and C. nepalensis (Cn3) nodules also revealed the close relationship of these strains to the genus Frankia.Key words: nitrogen fixation, Frankia, 16S rRNA, nifH.

It is now accepted that the microbial communities living within the gastro-intestinal tract (GIT) can influence pig health, especially at stressful times such as weaning. Careful design of the animal diet is considered to be an easy way to influence the microflora positively. In this work, fermentable carbohydrates were selected after an in vitro test (cumulative gas production- Bosch et al., 2002) for both rate and extent of fermentability. The ingredients were chosen both for their end-products and in terms of their fementation kinetics so as to stimulate fermentation along the entire GIT. They were incorporated into a test diet (TEST), and compared with a control diet (CONT) which contained minimal fermentable carbohydrates. Digesta were analyzed both in terms of microbial species (molecular techniques) and their end-products (e.g. lactic acid), to determine whether fermentation had, in fact, been stimulated by the TEST diet.

The presence of nitrogenase in vesicles and hyphae of Frankia EAN1pec was investigated by using immunogold labelling on ultrathin cryosections for electron microscopy. These studies resulted in the specific labelling of nitrogenase in the vesicles of nitrogen-fixing cultures. No significant label could be found in the hyphae, indicating a strong repression of nitrogenase in the hyphae.

The occurrence and localization of enzymes involved in energy supply and biosynthesis was studied in root nodules of Alnus glutinosa (L.) Vill. Vesicle clusters of the endophyte, Frankia sp., contain NADP-dependent isocitrate dehydrogenase, succinate dehydrogenase, fumarase and malate dehydrogenase. The data indicate that both the endophyte and the host are capable of metabolizing carbon compounds via the tricarboxylic acid cycle. Both vesicle clusters of the endophyte and root nodule cells contain glutamate-oxaloacetate transaminase which can function in a malate-aspartate shuttle. This might enable transport of reducing equivalents from the host cell cytoplasm to the endophyte.

The occurrence and the infectivity of Frankia, the root-nodule endophyte ofAlnus glutinosa, were studied in different kinds of soil in the Netherlands. Both field and pot experiments indicated that many soils, on which alders have not been grown before, had low numbers of endogenous Frankia or none at all. Inoculation of these soils usually enhanced growth and nodulation of alders.

Frankia AvcI1 utilizes only Tweens and fatty acids as sole carbon sources but N2, NH4+, NO3−, and various amino acids as nitrogen sources. The yield in propionate medium was increased by CO2. Cells grown in media with Tween 80 or acetate as C source contain the glyoxylate-cycle enzymes, isocitrate lyase and malate synthase. These enzymes are repressed in cells grown in media with propionate as well as in the symbiotic stage of Frankia in the nodules. The nature of the carbon compounds utilized by Frankia in the nodule symbiosis is discussed. Vesicle clusters derived from root nodules of Alnus respire succinate (however, only at a very low rate, which probably is due to damage of the membrane of the endophyte). Vesicle clusters derived from Alnus and Hippophaë nodules also take up O2 when supplied with a mixture of glutamate, malate, and NAD, indicating that the reduction equivalents could be transported from the host to the endophyte via the malate–aspartate shuttle. Vesicle clusters derived from Datisca respire succinate at a relatively high rate. Electron micrographs of Datisca nodules showed the presence of clusters of mitochondria embedded in the endophyte material, which might be of importance in the energy supply of the endophyte. These mitochondria are partly disintegrated during the preparation of the vesicle-cluster suspensions.

Nodulation tests onin-vitro propagated clones ofAlnus glutinosa ecotypes (forest ecotype, pioneer ecotype) withFrankia strains originating from both ecotypes indicated differences in host-plant compatibility. Inoculated plants of the pioneer ecotype clone were not infected by strains, that were unable to fix nitrogen in pure culture. Nodulation could only be induced on the clone of the forest ecotype, but no nitrogen-fixing activity could be detected. Ultra-structural observations of the nodules by SEM and TEM indicated that ineffectivity of these strains was correlated with the lack of vesicles in the infected cells. Cells were only filled with hyphae: neither sporangia nor vesicles could be detected. In contrast, effective nodules could be obtained on both alder clones after inoculation with an effective strain, showing normal development of vesicle clusters in infected cells. In pure culture the ineffective strains produced no vesicles; sporangia were found only during early stage of growth. The results demonstrate the existence ofFrankia strains which were either non-infective or ineffective on different clones ofAlnus glutinosa.

The occurrence and localization of enzymes involved in glycolysis, tricarboxylic acid cycle and glyoxylate cycle in root nodules of Alnus glutinosa (L.) Vill. and Hippophae rhamnoides L. ssp. rhamnoides were studied. The following enzymes, catalyzing reversible steps in the glycolysis, were found in both the endophyte Frankia spp. and the plant cytosol of Alnus nodules: fructose-1,6-diphosphate aldolase, glyceralde-hyde-3-phosphate dehydrogenase, phosphoglycerate kinase and enolase. The enzymes catalyzing irreversible steps in glycolysis, viz. hexokinase and pyruvate kinase, were detectable only in the plant cytosol. Similar results were obtained with nodule homogenates of Hippophae. This indicates the absence of a complete glycolysis in the endophyte. Vesicle clusters of the nodule endophyte of Alnus contained various dehydrogenases of the tricarboxylic acid cycle and showed activity of glutamate oxaloacetate transaminase. Respiration studies showed that vesicle clusters take up oxygen when supplied with NAD, glutamate and malate together. No oxygen uptake was found when any of these compounds was omitted. Vesicle clusters from both Alnus and Hippophae nodules showed no detectable activity of the glyoxylate cycle enzymes isocitrate lyase and malate synthase. Since these enzymes are known to be present in Frankia Avcll, when grown in a medium with Tween 80 as carbon source, it is suggested that the glyoxylate cycle enzymes are repressed in the root-nodule symbioses.

Summary Citrulline is shown to be the predominant free amino acid in alder root nodules, while serine also occurs in relatively large amounts. Citrulline and glutamic acid are prominent free amino acids in alder xylem tissue. The activities of N2-ase, GS, GDH, and OCT in root nodule homogenates are reported. From the Km values of GDH for NH4+ (16mM) and glutamate (0·9 mM) and the concentrations in the nodule of NH4+ (1·5 HIM) and glutamate (0·5 mM) it is concluded that GDH in alder nodules is responsible for the deamination of glutamate. The important function of GDH in the nitrogen metabolism of alder nodules is confirmed by the much higher activity of this enzyme in homogenates of nodules as compared to that in homogenates of root-tips and leaves. The vesicle clusters, which contain the N2-ase activity, did not show activity of GS, GDH and OCT. The cytoplasm of the host cells was shown to possess the GS activity, while GDH and OCT probably are localized in the organelles of the host cells.

To identify genes specifically expressed during early stages of actinorhizal nodule development, a cDNA library made from poly(A) RNA from root nodules of Alnus glutinosa was screened differentially with nodule and root cDNA, respectively. Seven nodule-enhanced and four nodule-specific cDNA clones were isolated. By using in situ hybridization, two of the nodule-specific cDNAs were shown to be expressed at the highest levels in infected cells before the onset of nitrogen fixation; one of them, ag12 (A. glutinosa), was examined in detail. Sequencing showed that ag12 codes for a serine protease of the subtilisin (EC 3.4.21.14) family. Subtilisins previously appeared to be limited to microorganisms. However, subtilisin-like serine proteases have recently been found in archaebacteria, fungi, and yeasts as well as in mammals; a plant subtilisin has also been sequenced. In yeast and mammals, subtilases are responsible for processing peptide hormones. A homolog of ag12, ara12, was identified in Arabidopsis; it was expressed in all organs, and its expression levels were highest during silique development. Hence, our study shows that subtilases are also involved in both symbiotic and nonsymbiotic processes in plant development.

The effects of dairy cattle slurry management on soil biota, soil respiration and nitrogen (N) mineralization were evaluated in a farm trial across 12 farms and a field experiment on 2 farms located in a dairy farming area in the north of the Netherlands. The slurry management consisted of slit injection or surface application of slurry; the use or no use of additives [Euromestmix® (MX) and Effective Microbes® (EM)] and the type and level of inorganic N fertilization. Slit injection negatively affected epigeic earthworms whereas its effect on anecic and endogeic earthworms was absent or even positive. Enchytraeids were not affected in a consistent way, whereas numbers of nematodes indicative of nutrient-enriched conditions increased. Inorganic N fertilizer had similar effects. Bacterial diversity was not different among the treatments. Nitrifier diversity, however, was high at one of the farms in the field experiment, and was negatively affected by inorganic N fertilizer. The use of MX was usually associated with higher numbers of earthworms. EM affected numbers of earthworms and numbers of bacterial and plant-feeding nematodes, but only in specific combinations of field history, slurry type and slurry application method. We found no effects of EM on the composition of the microbial community. Soil respiration was increased when slurry was surface-applied. The calculated N mineralization by earthworms was in the order of 70–200 kg N ha−1 year−1. It was highest under farm-characteristic surface application of slurry with MX and lowest under farm-characteristic slit injection of slurry without additives. Compared with the N mineralization by earthworms, that by enchytraeids and nematodes was quantitatively insignificant. Negative treatment effects on earthworms led to corresponding reductions in calculated N mineralization.

Summary Fourteen Pseudomonas strains causing canker of hazelnut in Greece and Italy were compared by analysis of plasmid DNA, whole cell fatty acids, whole cell proteins, and restriction fragment patterns of 16S rRNA. Moreover, fatty acid and protein profiles, 16S rRNA sequences and percentages of DNA similarity were compared with those of other Pseudomonas bacteria, including P. aeruginosa, P. cichorii, P. corrugate, P. viridi f lava, P. syringae pv. syringae, P. s. pv. porri, P. s. pv. phaseolicola, P. s. pv. maculicola, P. s. subsp. savastanoi and P. fluorescens. The bacterium causing canker of hazelnut originally was named P. syringae pv. avellanae on the basis of biochemical tests. Strains of the hazelnut organism were very homogeneous, which was confirmed by our work. Only in the number of plasmids, especially in Italian strains, some variation was observed. Fatty acid and protein profiles, as well as 16S rRNA sequence analysis and percentage of DNA-DNA hybridization with other Pseudomonas strains indicated that the bacterium causing canker of hazelnut is very much apart from all pathovars of P. syringae tested, including pv. syringae. The strains were only very distantly related to P. syringae and other Pseudomonas species tested. All other Pseudomonas species tested were closer to P. syringae than the bacterium causing canker of hazelnut. We therefore propose to reclassify P. syringae pv. avellanae into P. avellanae.

This study describes culture-dependent as well as culture-independent strategies to monitor diversity and abundance of Bacillus benzoevorans-related soil bacteria. These bacteria are wide spread around the world, inhabiting a variety of terrestrial environments. A distinguishing feature of the family Bacillaceae is their ability to form endospores. This capacity makes the genus Bacillus amenable to retrospective and biogeographical studies, as their endospores provide the means for survival under environmental conditions of stress, allowing their detection later on. For the rapid detection of B. benzoevorans-re\aiea populations in soil samples selective cultivation media and group-specific primers and probes were developed. Using these techniques the global distribution of this group was demonstrated and indicated their adaptive capacity to diverse soil ecosystems. A unique soil sample archive provided insight in the microbiological impact of land reclamation and flooding. Using multivariate statistical approaches were used to assess the microbial community dynamics over time and in response to the changing environmental conditions.Found in a variety of soil ecosystems, microorganisms related to B. benzoevorans seem to be able to populate a broad range of niches, which indicates a high degree of metabolic versatility and strong adaptive capability. Moreover, they account for a significant part of the total bacterial community (up to 30 %). Based on our exploratory study the importance of their role is just indicated, but not acknowledged yet. Therefore, it is proposed to regard B. benzoevorans relatives as 'kings without crowns'; as this group of bacteria deserves more scientific attention in future studies aiming to unravel their eco-physiology and functionality as major players of the soil microbiota.

Monogastricanimals including pigs coexist with a diverse and densecommensalmicrobiotain their gastrointestinal (GI) tract. Most of these microbes are beneficial and provide the necessary nutrients or protection against harmful pathogens for the host. The microbial colonization of the porcine intestine begins after birth and follows a rapid succession during the neonatal and weaning period. In the immediate post-weaning period, the balance between the development of beneficialmicrobiotaorthe establishment of bacterial intestinal pathogens can be easily tipped toward disease expression. In order to enhance the animal growth and suppress the activity of the gutmicrobiota, antimicrobial compounds have been fed to weaning piglets for more than four decades. Nowadays, however, the emergence of antibiotic resistance in the humancommensalbacteria has raised concerns about the impact of antimicrobial compounds for agricultural use, and accelerated the search for alternative nutritional strategies, such as the addition ofprobioticsandprebiotics. The aim of the current study was to understand the interactions between the intestinalmicrobiota, feed components and the host in the porcine GI tract and to validate some novel dietary strategiesin vivoandin vitro.The porcine intestinalmicrobiotadevelopment wasanalysedby 16S ribosomal RNA gene (rRNA) targeted approaches in relation to differences in nutritional strategies, age and pathogenic challenge. The GI tract microbial diversity was assessed using PCR-analysis of 16SrRNAgenes by cloning and denaturing gradient gel electrophoresis. In addition, methods such as real-time PCR and fluorescent in situhybridisationwere applied to quantify the predominant bacterial community and the major Lactobacillus populations. Furthermore, we developed a novel strain detection system based on isolation of specific genomic fragments by representative difference analysis and their further quantification by real-time PCR. The data obtained during the course of the study indicated that a stable and complexcommensalbacterial community might be considered as prerequisite of a healthy intestinal ecosystem. Moreover, the addition of fermentable carbohydrates to the diet of weaning piglets was found to enhance the stability and diversity of themicrobiotain the large intestine of piglets, and to stimulate the growth of strainsclosely related to Lactobacillusamylovorus. The phenotypic and molecular taxonomic characterization of these strains, however, revealed significant differences with the type strain of L.amylovorus (DSMZ 20531 T ) and the name Lactobacillussobrius sp.nov.wasproposed. L.sobrius strain 001 T was found to exert a significant protective effect against theenterotoxigenicEscherichia coli K88 promoted intestinal damages in vitro using porcine intestinal cell cultures. Furthermore, when L.sobrius was dietary fed to piglets subjected to E. coli K88 challenge, improved daily weight gain (+ 74%, P

The human gastrointestinal (GI) tract is a complex ecosystem in which host and microbial cells live in close contact with each other. The microbial community in the human GI tract has an important nutritional and protective function and mainly consists of anaerobic bacteria. After birth, the germ-free GI tract is colonized by the first invading microbes and this community is successively changing in time, ultimately resulting in a stable climax community. One of the major problems that hampers studying the bacterial diversity in the GI tract is the inability to obtain the predominant bacteria in culture. Estimates of culturability vary between 10% and 50%. One way to overcome these culturability problems is the use of culture-independent approaches to detect, monitor, and quantify bacteria in the human GI tract. Approaches using the sequence variability of the 16S ribosomal (r)RNA gene have become widely applied, since the 16S rRNA gene is an ideal phylogenetic marker. In the year 2001 there were up to 20,000 nucleotide sequences of different 16S rRNA genes available in several databases, which is far more than for any other gene.In this thesis, the bacterial communities in the human GI tract were investigated using a culture-independent approach based on the sequence variability of bacterial 16S rRNA genes. Our main questions were related to the composition of the bacterial community in the GI tract, its stability in time, and the effect of the host on the GI tract community.Our first approach was based on direct DNA isolation from GI tract samples followed by PCR of 16S rDNA. Temperature and denaturing gradient gel electrophoresis (TGGE and DGGE, respectively) combined with cloning and sequence analysis of amplified 16S rDNA was performed to analyze and monitor the bacterial diversity in human GI tract samples. Before starting this culture-independent analysis of varying GI tract samples from humans, we optimized and validated carefully the nucleic acid isolation from feces and biopsy samples, the subsequent reverse transcriptase (RT) and regular PCR amplification, and the generation of TGGE and DGGE profiles from the bacterial communities. In addition, a specific primer was developed and validated to specifically detect Lactobacillus and related lactic acid bacteria in human GI tract samples.TGGE and DGGE analysis of the V6 to V8 regions of fecal 16S rRNA and rDNA revealed that the predominant bacterial communities were host-specific and stable in time. Only slight differences in band intensities were observed when DNA and rRNA derived profiles were compared, indicating that the predominant bacterial 16S rRNA genes are also predominantly expressed in the colon. Since TGGE profiles revealed the bacterial composition to be host-specific, the bands of a fecal profile from one individual were identified by a cloning and sequencing approach. More than half of the clones corresponded to one of the predominant amplicons in the TGGE profile and most of them appeared to originate from novel Gram-positive species. Bacteria related to Ruminococcus obeum were found frequently in DGGE and TGGE profiles and 16S rDNA clone libraries. A specific 16S rRNA targeted probe was developed to target this group in a PCR-independent approach. FISH analysis using this probe confirmed that these uncultured bacteria are predominant in feces.Since the predominant bacterial community in the human GI tract was found to be stable and host-specific, the impact of the host and of environmental factors on the GI tract community of human adults was studied. A significant positive relationship was found between the genetic relatedness of the hosts compared and the similarity between the bacterial communities in feces, while no significant effect was found for hosts sharing similar environmental conditions. These data suggest that host genotype-related factors have a major impact on the bacterial composition in the GI tract of human adults. In addition, it was observed that the predominant bacterial community attached to the colonic mucosa is equally distributed along the colon, is significantly different from the fecal community, and is also host-specific. These results indicate that host-related factors have a major impact on the bacterial composition in the human GI tract. The Lactobacillus group-specific DGGE profiles supported these results, although it was noteworthy that a Lactobacillus gasseri -like bacterium was observed as the predominant lactic acid bacterium of the mucosa-associated community in most individuals.Most cultivation studies have been performed by plating fecal bacteria on very rich media. A cultivation approach, in which anaerobic basal liquid medium with cellobiose as the main carbon source was used, was employed to isolate strain Cello T, which belongs to an unknown genus. The difficulty for strain Cello Tto grow on solid media and its limited substrate utilization are two explanations why this and related bacteria have never been isolated before. Therefore, it can be concluded that alternative cultivation approaches may result in the isolation of novel bacteria.The results obtained in this study provide valuable knowledge about the bacterial diversity in the human GI tract. The power of applying 16S rRNA-based techniques to study bacterial communities in the human GI tract has been demonstrated and ideas for follow-up research to study the ecology have been proposed.

Haloorganics such as chlorophenols and chlorinated ethenes are among the most abundant pollutants in soil, sediments and groundwater, mainly caused by past and present industrial and agricultural activities. Due to bioaccumulation and toxicity, these compounds threaten the integrity of the environment, and human and animal health. A recently discovered, phylogenetically diverse, group of anaerobic so-called halorespiring bacteria is able to couple the reductive dehalogenation of various haloorganic compounds to energy conservation and hence to growth, significantly contributing to in situ dehalogenation processes in anoxic environments. The observed persistence of halogenated pollutants in untreated ecosystems and the accumulation of degradation intermediates during bioremediation, however, indicated the need for engineering of process conditions and/or augmentation with efficient degraders. This thesis describes genetic approaches towards a thorough understanding of the molecular basis of anaerobic reductive dehalogenation in order to enable the further optimization of clean up procedures for contaminated anoxic environments.The Gram-positive Desulfitobacterium dehalogenans , capable of degrading o -chlorophenols, PCE and hydroxylated PCB's, was used as model organism throughout a major part of this study. The key enzyme o -chlorophenol reductive dehalogenase (CPR) was isolated and characterized at the biochemical and molecular levels, and comparison with known chlorophenol- and chloroalkene-reductive dehalogenases indicated that these enzymes constitute a yet unknown, but evolutionary ancient family of corrinoid-containing iron-sulfur proteins, which in addition share a twin-arginine signal sequence. Transcriptional analysis of the CPR-encoding gene cluster revealed that dehalogenation activity is strongly regulated at the transcriptional level. Efficient gene cloning, and random and specific gene inactivation systems were developed to enable (i) the elucidation of additional components involved in the anaerobic dehalogenation process and (ii) the study of their structure and function within the respiratory network. Halorespiration-deficient mutants were isolated following random chromosomal integration, and their characterization lead to the identification of genes encoding proteins possibly involved in structure, maturation and regulation of respiratory complexes. Reductive dehalogenase-encoding gene targeted (RT-)PCR-based molecular approaches were developed that were useful for the detection of reductive dehalogenation potential and activity in pure cultures of potentially halorespiring microorganisms.The results obtained in this study provide valuable knowledge on the molecular basis of anaerobic reductive dehalogenation and might serve as a sound basis for the further exploitation of halorespiring bacteria as dedicated degraders in biological remediation processes.

Microbial Ecology studies aim to describe and assess the behavior and activity of microorganisms in their natural environments (Brock 1966). Nowadays it is clear that the large number of existing microorganisms has surpassed our capabilities to rapidly characterise them by traditional culturing methods. This has resulted in a poor understanding of the structure and composition of microbial communities. As an alternative, microbial communities can be described on the basis of 16S rRNA sequence diversity, without the bias-introducing step of cultivation.In the present thesis a molecular analysis is given of two ecosystems that harbour several uncultured bacteria. The first part of the thesis is focused on the detection and characterisation of Frankia in actinorhizal nodules and soil. Elucidation of the actual diversity within the family Frankiaceae was hampered by the inability to obtain isolates from all known actinorhizal plants. So far, the Nod +/Fix +Frankia symbionts in root nodules of plants from the families Coriariaceae, Datiscaceae, Rosaceae and Rhamnaceae (with exceptions reported by Carú 1993, Carú et al . 1990, and Carrasco et al . 1995) have resisted isolation. Best opportunities to characterise those uncultured endophytes require molecular methods that relay heavily on an easy and efficient technique to extract DNA from the respective actinorhizal nodules. Chapter 2 describes the techniques to isolate DNA from root nodules of different actinorhizal plants such as Casuarina sp, Alnus sp and Ceanothus sp. The procedure has also been successfully applied by Wolters et al . (1997b) in the minuscule ineffective nodules on Alnus glutinosa .Several attempts to characterise the uncultured endophytes from Coriaria sp. and Datisca sp. plants pointed on the one hand, to the presence in those actinorhizae of Frankia -related actinomycetes. This assumption was based mainly on the repeated isolation from those nodules of Nod -/Fix -Frankia -like strains (Hafeez 1983, Mirza et. al . 1994b, c). On the other hand, the effective (Fix +), non-isolated symbionts showed to be phylogenetically closely related (Mirza et al. 1994a), forming a separate lineage within the genus, in spite of the distant geographical distribution of the plants (Nick et al. 1992).The work described in Chapter 3 is focussed on the localisation and phylogenetic position of the nitrogen-fixing Frankia and Nod -/Fix -actinomycetes, both present in root nodules of the Mexican actinorhizal plant Ceanothus caeruleus . Application of the TGGE technique allowed localising the Nod -/Fix -actinomycete in the outer layers of the C. caeruleus nodules. Similar bacteria were also detected in Hippophaë rhamnoides nodules induced with soil inoculum that was collected in the vicinity of the former plant. The fact that a second nodule inhabitant was commonly present in these nodules containing recalcitrant endophytes may allow some speculations about their possible role in the symbiosis. However, it seems worthwhile to apply the same TGGE methodology to other actinorhizal nodules, even to those containing Frankia strains that are easy-to-isolate (i.e. Elaeagnus,Casuarina, Alnus spp.), since the detection of Frankia -related actinomycetes, in addition to the Fix +endophyte, would provide further evidence about the need for their presence. Coincidentally, the Nod -/Fix -isolates from Coriaria, Datisca and Ceanothus are phylogenetically related, pointing again to a certain specificity for their presence in the nodules. This relatedness has also been supported by analysis of low molecular weight RNA (i.e. 5S rRNA and tRNA's) using staircase electrophoresis (Velázquez et al. 1998).The 16S rDNA sequence from the non-isolated Fix +endophyte in C. caeruleus root nodules (Chapter 3), was the first full sequence obtained from a field-collected Ceanothus symbiont. Parsimony and phylogenetic distance analyses grouped it within the Dryas cluster that originally contained only the uncultured endophytes from Dryas , Coriaria and Datisca as proposed by Normand et al. (1996). Benson et al . (1996) redefined this cluster by adding other uncultured endophytes present in Ceanothus griseus (Rhamnaceae), Purshia tridentata and Dryas drummondii (Rosaceae) root nodules. Since the determined partial 16S rDNA sequences were almost identical, they suggested that the Frankia diversity from these actinorhizal plant families might be low. However, Clawson et al. (1998) demonstrated that Frankia isolates obtained from several genera within the Rhamnaceae (i.e. Talguenea , Colletia , Discaria, Retanilla and Trevoa ) were phylogenetically different than those in Ceanothus, grouping in the Elaeagnus cluster. These findings were consistent with morphological differences of the endophytes in planta , since the vesicles found in the Ceanothus symbionts resemble more to those in the Rosaceae, while all the latter host plants in the Rhamnaceae family have endophytes like those in Elaeagnus .The results reported in the first part of the thesis have demonstrated that TGGE and sequence analysis of 16S rDNA provide an accurate picture for the identification of recalcitrant endophytes in root nodules of actinorhizal plants. It has also been demonstrated that besides the N 2 -fixing endophyte, root nodules of C. caeruleus also harbour Frankia -related actinomycetes. Since these have also been observed in other actinorhizae, a further study is needed to understand the possible function of these co-symbionts.The work described in the second part of the thesis was addressing the changes occurring under chlorobenzoate stress in the soil bacterial community and other selected groups of bacteria present in peat soil collected from a natural Alnus glutinosa stand. A combination of culturing and non-culture based approaches was used for the assessment. Among the latter approaches, the possibilities offered by TGGE were exploited in several ways. Profiling of complex communities and subsequent analysis of specific bacterial groups has been one of the major applications of TGGE (Felske et al . 1996). With this approach, major population shifts induced by either 3CBA or 2,5DCB were detected in the uncultured bacterial community (Chapter 4). Although only the former compound was readily metabolised in soil, both xenobiotics promoted similar changes. Several bacterial populations were reduced or suppressed, while few others were enriched in time, as assessed by shifts in the TGGE banding patterns of the total bacterial community.To characterise the soil-enriched bacterial populations, 3CBA-degrading enrichment cultures were obtained and their composition was addressed by TGGE. Further isolation attempts were directed by this means to prove that the isolated strains were indeed the same enriched organisms as detected in soil. One of the enrichment cultures contained two of the soil-enriched bacteria as predominant components. Although isolation was not achieved, both bacteria were identified as belonging to the genus Burkholderia . The bacterial group detected as predominantly enriched in both spiked soils was not present in any of the enrichment cultures, suggesting that the microorganisms belonging to this groups are either unable to degrade 3CBA or not growing under the used culturing conditions. In any case their fitness to the soil conditions imposed by the addition of chlorobenzoates was high, but the mechanisms involved were not elucidated. These bacteria were also identified as Burkholderia by partial 16S rDNA sequence analysis (Chapter 4).The diversity (H) and the equitability (J) indices are important parameters used by ecologists to assess the species richness and the species evenness, respectively, within a community. As the estimation of such indices relies heavily on species definition and individuals enumeration, their application in microbial ecology studies is seldomly possible. Furthermore, assessment of H and J in uncultured bacterial communities must rely on the interpretation of community fingerprints, which should provide means to distinguish between species or operational taxonomic units (OTU), and to estimate their abundance. TGGE community profiling offers both possibilities, and the community changes occurring in the model soil system were evaluated with this original approach (Chapter 5).In addition, H and J indices were also estimated for the fluorescent pseudomonads group, a selected culturable fraction of the bacterial community. OTU recognition was addressed by using TGGE as a ribotype-fingerprinting technique for the isolated fluorescent pseudomonads. Estimation of H and J at the community level without culturing by TGGE profiling, and at the group level by a combination of culturing and TGGE ribotyping should allow to address and compare the population changes occurring, since the target molecule used in both TGGE was the same. Such comparison was only partially possible since most of the bands corresponding to the fluorescent pseudomonads could not be assessed in the community profiles. However, estimation of H and J indices indicated a clear reduction of species richness and individuals abundance in the uncultured community, which was related to the presence of chlorobenzoates in soil. Evaluation of population shifts by indexed values as H and J proved to be a useful means for analysing the community structure in time, and may be used to assess short and long-term responses of a bacterial community to environmental perturbations.Chapter 6 describes the changes in the total frankiae in soil and in the fraction of the population that is able to produce root nodules in Alnus glutinosa seedlings . Culture-independent approaches based on the most probable number concept were used, one in combination with a Frankia -specific PCR detection and another in combination with a plant-nodulation bioassay. After 15 days of incubation in the presence of chlorobenzoates both fractions of the soil Frankia populations were reduced in more than one order of magnitude, while the populations in the unspiked control soil were not affected. The results indicated that 3CBA and 2,5DCB both had a negative effect on the size of the native Frankia population from the used peat forest soil. This negative effect was also evident during in vitro experiments using Frankia strains isolated from Alnus sp. The presence of 1 mM 3CBA in the culture medium, in addition to the normal carbon source, resulted in reduction or suppression of biomass yield.The influence of Alnus glutinosa on the dechlorination of 3CBA by Pseudomonas sp. strain B13 was assessed in hydroponic cultures. It was expected that root exudates could enhance the dechlorination activity of Pseudomonas B13. When the bacteria were incubated in the presence of an alder plant, only a slight increase in the dechlorination rate of 3CBA was registered in comparison to the control without plant. The main observed effect in the alder plants appears to be a protection against 3CBA toxicity, as the alders inoculated with Pseudomonas B13 showed a better survival rate and grew more vigorously than the non-inoculated plants.The toolbox for microbial ecology studies is increasing constantly by means of developing new techniques or by adapting foreign tools into the field, such as indices to evaluate species diversity and eveness. Although the information provided by these community parameters facilitates comparisons and assessment of changes, their suitability to evaluate bacterial communities is still uncomplete. Estimation of diversity indices requires the recognition of bacterial species as discrete units, and this condition is far from real in natural environments. Species-independent approaches to evaluate diversity must be developed, that consider the bacterial diversity as a continuous range of phylogenetically related taxonomic units.In conclusion, the work described in the first part of the thesis strengthen the current phylogenetic division of the Frankiaceae, by adding new evidence supporting two of the already described clusters ( Dryas cluster, Nod -/Fix -cluster). Although the current taxonomic status of the latter cluster must be better evaluated in order to assess its pertinence to the genus Frankia . In addition, the common occurrence of Nod -/Fix -Frankia -like actinomycetes in nodules containing recalcitrant endophytes, mainly from the Dryas cluster, was also demonstrated for Ceanothus actinorhiza. In the second part, the suitability and versatility of molecular tools such as TGGE were demonstrated by their application in community profiling and estimation of bacterial diversity. Community changes occurring in stressed and unstressed soil systems were easily detected and assessed by this means.In addition, specific populations such as the Burkholderia -like bacteria that strongly reacted to the addition of chlorobenzoates to soil were further characterised. Moreover, TGGE was shown to be a fast ribotyping technique that may enable its use in combination with the community profiles to address shifts of specific groups within bacterial communities. It is tempting to suggest that this general approach will be of importance to direct the isolation of hitherto uncultured bacteria from soil.

The research described in this thesis was aimed to provide insight into the effects of grassland succession on the composition of the soil bacteria community in the Drentse A agricultural research area. The Drentse A meadows represent grassland succession at different stages. Since 30 years particular plots have been taken out of intense agricultural production and were not fertilized anymore. However, the grass-vegetation was continuously removed once per year. This caused a progressive depletion of nutrients in the soil. In order to reveal the effect of grassland succession on soil microbes, the main bacteria in Drentse A grassland soils were identified by a molecular strategy based on detection and quantification of 16S rRNA. Instead of only using genomic 16S rDNA to reveal present sequences, we focused on rRNA to quantify the activity of the predominant bacteria. The ribosome is considered to be a useful marker for the overall metabolic activity of bacteria. In bacterial cultures the amount of ribosomes per cell has been found to be roughly proportional to growth activity. In our approach the activity is defined as total activity of one ribotype in relation to the bacterial community and not as activity per cell. Hence, bacteria of low activity per cell but extraordinary high cell number might be assessed as very active. Following direct ribosome isolation from soil, several different methods like RTPCR, separation of amplicons by temperature gradient gel electrophoresis (TGGE), different hybridization methods, cloning and sequencing were applied simultaneously to reveal the predominant 16S rRNA sequences for taxonomic identification. Quantitative dot blot hybridization with taxonspecific oligonucleotide probes revealed dominance of low G+C Gram-positives while other important groups appeared to be aProteobacteria and high G+C Gram-positives (Chapter 5). However, this approach did not demonstrate clear tendencies of community structure shifts by quantifying the rRNA of the major taxa. Therefore, a more sensitive method has been chosen, based on RT-PCR amplicons of bacterial 16S rRNA. The sequence-specific separation of these amplicons by TGGE reproducibly yielded characteristic band patterns from hundreds of soil samples (Chapter 3). Although the TGGE signals were very complex due to the high bacterial diversity in soil, different 16S rRNA fingerprints from a single plot were highly similar, while reproducible differences between plots of different history were observed. A parallel approach with PCRamplified genomic 16S rDNA led to similar results. The presence and activity of prominent bacteria in test fields of several hundreds m 2 were found to be quite homogeneous. Only one gram of soil was found to be representative for the predominant bacteria in large homogeneous grassland areas. After the high reproducibility of presence and activity was demonstrated for the main soil bacteria, representative TGGE fingerprints were compared to TGGE signals from a clone library of genes coding for 16S rRNA (Chapter 5). Cloned 16S rDNA amplicons matching the intense bands in the fingerprint were sequenced. The relationship of these sequences to those of cultured organisms of known phylogeny were determined. Approximately one half of the amplicons represented sequences closely related to those of cultured Bacillus -species, indicating that most of the active bacteria apparently belonged to the genus Bacillus . Other sequences similar to Gram-positive bacteria with high G+C-content were only rernotely related to those of cultured bacteria, as is illustrated by clone DA079 that could be affiliated to uncultured Actinobacteria from peat (Chapter 6). Another important group of sequenees was related to Proteobacteria, mainly theα-subclass. Several sequences could not be related to cultured organisms hut to the Holophaga/Acidobacterium - or the Verrucomicrobiales -cluster (Chapters 7 and 8).The parallel application of RT-PCR/TGGE and 16S rDNA-cloning to reveal the most abundant 16S rRNA sequences was found to be a powerful combination. The clone screening on TGGE was convenient and efficient, offering access to the almost complete 16S rRNA sequence. The subsequently performed V6-hybridization was a relatively simple approach to prove the identity of bands even in complex fingerprints (Chapter 6). The most predominant Bacillus -like ribotype DA001 in Drentse A grassland soils could also be detected by fluorescent whole-cell in situ hybridization (Chapter 10). Prominent rod-shaped cells of approximately 2μm length could be identified in bacterial suspensions from soil with a multiple 16S rRNA probing approach. The specific DA001-signals represented about 5% of all microbial particles, which were visualized by the universal DNA-dye DAPI. Indeed, the sequences detected by the PCRbased methods represented abundant bacteria in soil. The most predominant Bacillus -like 16S rRNA sequence DA001 apparently originated from active, vegetative cells and not from endospores.The possibility to draw quantitative information about the microbial community from the complex TGGE fingerprints has been explored. A novel approach has been developed to quantify rRNA sequences in complex bacterial communities by multiple competitive RT-PCR and subsequent TGGE analysis (Chapter 4). The used primer pair (U968-GC and L1401) was carefully tested and found to amplify with the same efficiency 16S rRNAs from bacterial cultures of different taxa as well as the cloned 16S rDNA amplicons from soil samples. The sequence-specific efficiency of amplification was followed by monitoring the amplification kinetics via kinetic PCR. The primer-specific amplification efficiency was assessed by competitive PCR and RT-PCR, and identical input amounts of different 16S rRNAs were found to result in equal amplicon yields. We applied this method as multiple competitive RTPCR to TGGE fingerprints from soil bacteria to estimate the ratios of their 16S rRNAs (Chapter 9). This was done for different stages of grassland succession in the Dutch Drentse A area. The 16S rRNA amounts g -1 soil of 20 predominant ribotypes were monitored via multiple competitive RT-PCR in six plots of different succession stage. The 20 monitored 16S rRNA levels represented approximately half of all bacterial soil rRNA. The different Drentse A meadows, representing progressing stages of grassland succession, showed highly reproducible shifts of ribotype composition. In general, the rRNA levels were found to be doubled after the first years without fertilization. During the further progression of grassland succession the rRNA amounts were found to decline again. The 20 ribotypes showed remarkably different succession histories, causing the differences in TGGE fingerprints from different plots. While organic carbon and available nitrogen were declining during grassland succession, some bacteria were apparently suffering much more than the average. However, other bacteria showed an increased contribution to the bacterial rRNA pool, indicating that some bacteria could improve their position when less nutrients were available. The general increase in bacterial ribosomes in the first years after fertilization-stop was correlating to the increase of other parameters related to bacterial activity, i. e. carbon mineralization and microbial biomass. This suggested a true correlation between the total activity of bacterial communities in soil and the amount of ribosomes. This study provides extended information about uncultured bacteria in soil and describes the application and evaluation of several novel approaches in molecular microbial ecology. The following six conclusions are highlighting the major achievements and findings:1. Representative rRNA and rDNA fingerprints can be generated for homogeneous landscapes of large scale. This demonstrates that the often tiny sample size in molecular studies has not to be a limitation for microbial ecology. Nucleic acid extraction from small soil samples can be applied to characterize a several magnitudes larger environmental matrix. The composition of bacterial communities might be quite homogeneous for kilometers of grassland with heterogeneous vegetation and cultivation history. This conclusion is of general importance for molecular rnicrobial ecology and landscape ecology.2. The rRNA cycle (see Introduction, Fig. 2) for the predominant soil Bacillus recognized as ribotype DA001 has been completed. Its rRNA not only has been identified as predominant in the isolated fraction of soil ribosomes, but was also detected in an abundant type of bacterial eells by whole-cell hybridization to a fluorescently labeled, 16S rRNA-targeted oligonucleotide probe.3. The genus Bacillus appears to be dominant in the microbial community of Drentse A grassland soils. Uncultured members of the B. benzoevorans -line of descent are predominant among these bacilli. This group of Bacillus -ribotypes accounted for approximately 20% of all bacterial ribosomes in Drentse A soil. Such a predominant cluster of very closely related bacteria has never been observed before in soils. The reasons and circumstances of this special community composition remained unexplored.4. Prominent clusters of hitherto uncultured environmental bacteria were also detected in Drentse A soils. The Holophaga/Acidobacterium -cluster, the Verrucomicrobiales and a peat-related Actinobacteria-cluster had already been known from different locations all over the world. lt is the first time that these hitherto uncultured bacteria were identified as prominent contributors to the ribosome fraction in soil. Since these organisms apparently contain considerable amounts of intact ribosomes they are likely to be metabolically active. Some of these novel ribotypes belong to the most intense bands in the TGGE fingerprints, which may suggest a major role in environmental nutrient fluxes. This finding also contributes to the discussion about the unculturability of environmental bacteria, since based on the presented results it appears unlikely that the reason for their unculturability is a lack of viability. lt is more probable that suitable culture conditions are still not found yet. Since these bacteria are abundantly detected all over the world, future research should be aimed to attempt to culture them.5. A novel approach has been developed to quantify the predominant rRNA molecules of environmental bacteria communities. The multiple competitive RT-PCR allowed to quantify in a highly-specific way many different rRNA molecules within one assay. The RT-PCR-related possibilities of bias were investigated and excluded for the applied primer pair. Therefore, amplification by RT-PCR could be excluded as a major source of bias in this study. Other uncertainties are the selectivity of the applied primers and probes and the cell lysis efficiency. The selection of all the oligonucleotide probes and primers is based on only a few thousand 16S rRNA sequences of different length and quality. Although the available 16S rRNA sequence data are limited, the presenee of hitherto unknown bacteria with novel 16S rRNA sequences not matching the used primers is not indicated. The combined results of cloning, TGGE and dot blot hybridization, all achieved with different probes or primers, do not reveal possibly neglected groups of organisms. A serious bias caused by incomplete cell lysis may not be excluded. However, the majority of ribosomes originated from Gram-positives, indicating the lysis of bacteria with resistant cell walls. The possibility that highly resistant resting stages like endospores might have been missed is not relevant, since this study aimed to detect the most active bacteria.6. Multiple competitive RT-PCR revealed activity shifts for the predominant soil bacteria during Drentse A grassland succession. Some species responded to the nutrient depletion during grassland succession. Though the depleting nutritious matter and the changing vegetation, the overall impact of grassland succession did not cause a correspondingly drastic impact on the microbial community composition. Reproducible shifts of ribosome levels could be demonstrated, but the composition of the bacterial community remained remarkably stable. Evidence for major competition or replacement of species could not be found.

The research described in this thesis was directed to the development of molecular identification and detection techniques for studying the ecology of Rhizobium, a nitrogen- fixing bacterium of agricultural importance. Competition of inoculant strains with indigenous microbes is a serious problem in agricultural practice and was therefore addressed in this work using the developed tools. Furthermore, various molecular techniques have been applied to analyse rhizobial populations nodulating common bean and a new species was characterized.In this chapter the results obtained are summarized and potential future applications are discussed.Development of gusA - and celB -minitransposons and their use in rhizobial competition studiesThe use of marker genes in rhizobial competition studies is reviewed in Chapter 1. Specific attention is given to the gusA gene, encodingβ-glucuronidase (GUS). This gene is a highly suitable marker for studying plant-microbe interactions due to the absence of GUS activity in plants and in most bacteria that are of relevance in agriculture. In Chapter 2 the construction of several GUS transposons containing the marker gene in combination with different regulation systems to be used for ecological and genetic studies is described. The minitransposon mTn5SS gusA 20 contains the aph promoter which was demonstrated to be expressed in a wide variety of Gram-negative bacteria (de Bruijn and Lupski, 1984) and its use in Rhizobium resulted in high-level constitutive expression. Transposon mTn5SS gusA 21 is similar to mTn5SS gusA 20, except that it contains a unique site for Spe I, a rare-cutting enzyme in bacteria with high G+C contents such as rhizobia (Sobral et al., 1991). The use of mTn5SS gusA 2I therefore may be instrumental for the genetic mapping of insertions. The tac promoter was used to drive the expression of the gusA gene in the transposon mTn5SS gusA 11 resulting in high GUS activity in the free-living state. The transposons with constitutive gusA expression are optimal for studying rhizosphere colonization and for studying nodule occupancy in young plants. In order to reduce any metabolic load due to GUS production, the mTn5SS gusA 10 transposon with regulated gusA expression was constructed. It contains the tac promoter in combination with the lacl q repressor gene and gusA gene expression is repressed until an inducer, such as IPTG, is added. This regulation should avoid possible effects on the ecological fitness, Two transposons carrying symbiotically activated gusA genes, mTn5SS gusA 30 and mTn5SS gusA 31, were made by using the nifH promoter of a Rhizobium and a Bradyrhizobium strain, respectively. The nifH gene encodes the Fecomponent of nitrogenase and is only expressed in symbiotic or microaerobic conditions (Fischer, 1994). These constructs are recommended for longer-term nodule occupancy experiments. Furthermore, a promoter-less GUS transposon,mTn5SS gusA 40, is described that should be of use for molecular genetic studies as well as for screening bacteria for their response to specific environmental conditions or signals. The developed transposons carry a gene conferring resistance tospectinomycin and streptomycin that proved to be an appropriate marker for many strains. Nevertheless, few strains with an endogenous resistance exist. Therefore, the development of additional transposons conferring an alternative resistance would be advantageous. This may be realized by inserting the developed GUS expression cassettes into minitransposons: containing other antibiotic (de Lorenzo et al., 1990) or natural resistance markers (Herrero et al., 1990). Chapter 2 also addresses also theapplication of the different transposons in studies on root colonization and nodule occupancy, while various GUS assays are described in detail.Potential effects on the fitness of a strain due to insertion of a GUS transposon were evaluated in Chapter 3. Only few data exist on the impact of foreign genes on the fitness of an organism (Doyle et al., 1995) and before using any marker system for ecological studies its ecological effects have to be studied rigorously. In the case of Rhizobium, it is essential that the nodulation behaviour and competitive ability are maintained. The competitive abilities, nodulation characteristics, and growth rates of five independent derivatives of R. tropici strain CIAT899 marked with the gusA gene on minitransposon mTn5SS gusA 10 were determined relative to the parent strain. Insertion of mTn5SS gusA 10 did not affect the nodulation or nitrogen fixation efficiency of the wild-type strain. Nevertheless, the competitiveness index of the different gusA derivatives relative to the parental strain CIAT899 varied between isolates. One isolate was less competitive than the wild-type strain in three independent experiments, while the other isolates proved to be either equally competitive or more competitive. The utilization of this methodology to assess competitivity resulted in highly significant calculations as all the nodules on each plant were analysed for nodule occupancy. The results showed that the insertion of mTn5SS gusA 10 may have an impact on the ecological behaviour of a strain, but derivatives indistinguishable from the parent strain can be obtained. A primary selection of marked strains is recommended, which may be achieved by coinoculating the parent and the marked derivative in a one to one ratio and ensuring that the proportion of blue nodules does not differ significantly from the expected 50%.Furthermore, in Chapter 3, the detection of dual nodule occupancy is discussed. The appearance of partially stained nodules in mixed inoculum treatments but not in single strain treatments led to the conclusion that these were due to mixed infections. This was confirmed by nodule isolation and plating experiments. Partially stained nodules were also observed by Krishnan and Pueppke (1992) who reported that nodules were occupied by either a lacZ- marked or a non-marked R. fredii strain and X-gal was used for detection. However, the lacZ marker system has several disadvantages due to high background activity in plant and rhizobia whereas the gusA marker gene can be used to readily detect dual nodule occupancy on plant.A new marker gene system based on the celB gene is presented in Chapter 4. The celB gene has been isolated from the hyperthermophilic archaeon Pyrococcus juriosus and it encodes a thermostable and thermoactiveβ-glucosidase with a highβ-galactosidase activity (Voorhorst et al., 1995). The latter enzyme activity can be used for the detection of rhizobia as endogenous background activity in plants as well as in bacteria can be easily eliminated by a heat treatment. Moreover, cheap histochemical substrates are available to determineβ-galactosidase activity. The E. coliβ-galactosidase gene, lacZ , has been used to monitor engineered soil bacteria (Drahos et al., 1986; Hartel et al., 1994) but was only found to be appropiate when used with Lac- bacteria. Transposons containing the ce1B gene were constructed in E. coli , based on the existing gusA transposons. The first transposon, mTn5SS celB 10 , contains the tac promoter which is regulated by the lacI q gene product and should reduce any metabolic stress to the marked strain due to marker gene activity. Transposon mTn5SS celB 31 carries celB expressed from a Bradyrhizobium nifH promoter and is active in nitrogen-fixing legume nodules. A third celB minitransposon, which contains the marker gene constitutively expressed is described elsewhere (Sessitsch et al., submitted). The celB marker gene system has several advantages in rhizobial competition studies over conventional techniques as the assay is simple to perform and the histochemical substrates are cheap. However, the greatest advantage is that gusA and celB marked strains can be localized simultaneously on a plant and a combined gusA/celB assay will enable studies of multi-strain rhizobial inocula competing with indigenous rhizobial populations. Although simultaneous detection of differently marked strains has been reported (Thompson et al., 1995; Bauchrowitz et al., 1996), the celB gene encoding the thermostable marker is better suited for double staining. In addition, a procedure has been described allowing detection of gusA and celB marked strains on plates (Sessitsch et al., submitted). In Chapter 4 the application of the celB marker gene is demonstrated for Rhizobium. However, because of the wide host range of the Tn5 based transposons and the portable expression signals this marking system is suitable for use in a variety of Gram-negative bacteria.The advantages of the different gusA and celB marker gene cassettes are discussed in Chapter 1, but also other reporter genes and their applications in studies on microbial ecology are presented. Moreover, the development of a GUS Gene Marking Kit is reported. This kit was made particularly for agronomists and microbiologists in developing countries who are not familiar with molecular techniques and who do not have the resources to establish this methodology in their laboratories. Meanwhile, a CelB Gene Marking Kit is also available that can be used either in combination with or instead of the GUS Gene Marking Kit (FAO/IAEA, 1992-1997).Ecology of rhizobia nodulating common beanIn the fields around the Seibersdorf laboratory common bean has not been cultivated during the last decades but is well nodulated. In earlier studies, common bean rhizobia populations in this soil have been found to be very competitive in nodulation (see Chapter 1). They were shown to outcompete R. tropici strain CIAT899 when inoculating Phaseolus vulgaris at an inoculation level of 10 5cells per seed. When increasing inoculation to 10 8cells per seed, 65% of the nodules were still occupied by the native strains. In Chapter 5, rhizobial populations isolated from common bean nodules grown in the Seibersdorf soil were characterized. Molecular methods targeting the whole genome such as PCR with repetitive primers were used, and specific chromosomal loci such as the 16S rRNA gene or the 16S- 23S rDNA intergenic spacer were analyzed. Plasmid profiles and Southern hybridization with a nifH probe gave information on symbiotic regions. In addition, the nodulation host range was determined. Two distinct groups were found, one of them was classified as R. etli according to the RFLP analysis of the 16S rRNA gene and because of the presence of three copies of the nifH gene. The members of the second group could not be assigned to any recognized common bean nodulating Rhizobium species, i.e. R. leguminosarum bv. phaseoli, R. etli and R. tropici, but showed high similarity to Rhizobium sp. (Phaseolus) strain R602sp isolated in France (Laguerre et al., 1993). Isolates of this group also formed nodules on cowpea, Leucaena and Gliricidia. For a long time, R. leguminosarum bv. phaseoli was believed to be the only microsymbiont in Europe, but recently R. tropici and two new species have been found in French soils (Laguerre et al., 1993; Amarger, 1994). These studies and the results obtained in Chapter 5 indicate that strains originating in Mesoamerica could establish well in European soils. However, diversity was not high among the Austrian isolates due to the long absence of the host plant in this soil.The focus of Chapter 6 is on the taxonomy and phylogeny of the Austrian isolates showing high similarity to Rhizobium sp. R602sp. A Mexican common bean isolate, FL27, was included in this study since Laguerre et al. (1993) found the partial 16S rRNA gene sequence of R602sp to be identical to FL27. Sequence analysis of the 16S rRNA gene, determination of the copy number and heterogeneity of ribosomal genes, plasmid profiles and DNA-DNA hybridization resulted in valuable taxonomic information on these strains. Based on these results it was proposed that these strains belong to a new species that was named R. pueblae sp. nov., referring to the state Puebla, Mexico, where FL27 was isolated. The Mexican, French and Austrian isolates showed very similar 16S rDNA sequences with a maximum of two nucleotide substitutions. Comparison of the 16S rDNA sequences with those of other bacteria revealed highest similarity to R. leguminosarum strain IAM 12609, R. sp. OK50 and to R. etli. Although phylogenetic dendrograms always positioned R. pueblae sp. nov. strains in the vicinity of the above-mentioned species, the new species was found to belong to a lineage different from those of described Rhizobium species. The whole DNA relatedness among the European isolates was very high but showed lower levels with FL27, probably due to the presence of different plasmids. The DNA homology to other bean-nodulating species was very low. R. pueblae sp. nov. strains possess at least three copies of the 16S rRNA gene and the ribosomal gene organization is different to other species. Despite the high competitive ability of some strains in the Seibersdorf soil, little is known on the agronomic value of this species.Concluding remarksTo improve biological nitrogen fixation, adapted efficient nitrogen-fixing plant genotypes, effective rhizobial inoculants and appropiate agricultural management practices are needed. Consequently, plant breeders, microbiologists, soil scientists, agronomists as well as farmers have to cooperate in order to achieve this goal.The selection of superior sources of natural plant genetic variability and plant breeding in the presence of rhizobia instead of applying nitrogen fertilizers may lead to the identification of high-fixing lines. Other desirable traits such as disease resistance or stress tolerance could be transferred by appropiate breeding methods or by genetic engineering. Soils, in which legumes are cultivated, vary greatly and can be opposed to various environmental stresses such as low pH or high temperature. Efforts have been undertaken to develop appropiate plant genotypes whereas for a long time the stress tolerance of inoculant strains has not been considered. Inoculant strains have been recommended based on good symbiotic performance in a particular environmentbeen recommended based on good symbiotic performance in a particular environment while the soil status or the agroecological zone of the final application has not been taken into account. The vast genetic pool of natural soils containing not yet identified strains and species can provide a variety of inoculant strains that may show better performance in the field. A strong correlation between the indigenous population size and the nodule occupancy of the inoculant strain has been established (Thies et al., 1991). However, the effect of the diversity of indigenous rhizobia on competition has not been determined. Probably, different strategies are needed in order to outcompete highly dominant field isolates or to achieve successful competition of an inoculant strain with a variety of different indigenous strains that are present in low numbers. In addition, inoculation practices have to be developed that are convenient for the farmer and that allow distribution of the introduced strain into the entire rooting zone.It seems that the various aspects important for efficient nodulation and nitrogen fixation are presently not linked sufficiently. A more rational selection of efficient and competitive strains could be realized when a database existed containing data on soil properties, environmental conditions, rhizobial diversity and population size, as well as on the competitive ability and effectiveness of rhizobial strains in combination with particular plant genotypes. Furthermore, these data could be of use for the development of new strains or plants by genetic engineering.In this thesis, the development of new methods to assess rhizobial competition is presented. These marker gene-based techniques are appropiate for the large-scale screening of inoculant strains but can be also used for genetic analysis of a variety of Gram-negative bacteria. Molecular methods have facilitated the analysis of strains nodulating common bean and resulted in the description of a new Rhizobium species that includes strains with possible beneficial properties.

The research described this thesis concerns the diversity and phylogeny of syntrophic propionate-oxidizing bacteria and their ecology in granular sludge, from which they were obtained. 16S rRNA was used as a molecular marker to study both the phylogeny and the ecology of these bacteria. Sequence analysis of the 16S rRNA gave information on the phylogeny of the syntrophic bacteria, while specific oligonucleotide probes based on these sequences enabled quantification and detection of these bacteria. In situ hybridization techniques were applied to study the architecture of granular sludge. Granular sludge was also enriched for propionate-oxidizing bacteria and studied with dot blot and in situ hybridization to combine their presence in the sludge to the physiological processes observed.In this chapter the results are summarized and discussed in a broader context, starting with the diversity of the syntrophic propionate-oxidizing bacteria. Subsequently, the application of the ribosomal approach for the detection of syntrophs in the sludge is discussed. Next, the ecology and the localization of the microbes in granular sludge fed with different substrates is presented in models of sludge architecture, and finally the main conclusions of the research are presented.Diversity of syntrophic propionate-oxidizing bacteria. For more than one decade Syntrophobacter wolinii was the only well-described syntrophic propionateoxidizing bacterium (Boone and Bryant, 1980). Although other syntrophic propionateoxidizing cocultures were enriched, no other bacterium could be isolated (Koch et al., 1983; Boone and Xun 1987; Mucha et al., 1988; Stains et al., 1992).It was initially assumed that S.wolinii was an obligately syntrophic bacterium, which only used propionate as carbon source, and only grew in coculture with H 2-and/or formate-utilizing microbes such as Desulfovibrio G11 or Methanospirillum hungatei (Boone and Bryant, 1980). However, at the beginning of the nineties, when the research of this thesis was initiated, this view changed. Stams et al. (1993) enriched a syntrophic propionate oxidizer on malate, which could grow in pure culture on fumarate. This bacterium, strain MPOB, later named Syntrophobacter fumaroxidans (Chapter 8), could oxidize propionate using fumarate as alternative electron acceptor. Meanwhile, Schink and coworkers, succeeded to grow S.wolinii in the absence of H 2-and/or formate-utilizing microbes on propionate and sulfate, and on pyruvate (Dörner, 1992; Wallrabenstein et al., 1994). These findings showed that these syntrophs were not obligately syntrophic, but could be grown in pure culture with other carbon sources as well. The phylogenetic analysis of the 16S rRNA of the syntrophs showed that S. wolinii, S. fumaroxidans and the later isolated Syntrophobacter pfennigii, are not only closely related to each other, but are related to Gram-negative sulfate-reducing bacteria (Chapters 2 and 3). Moreover, they are sulfate-reducing bacteria themselves, as all three Syntrophobacter spp. could oxidize propionate by the reduction of sulfate.The increasing knowledge of this group of syntrophs may lead to the isolation of more related syntrophic propionate-oxidizing species. However, there are presently arguments that syntrophy is not a physiological characteristic of this phylogenetic group. Two sulfate-reducing bacteria have been isolated that were related to the genus Syntrophobacter, but which did not use propionate syntrophically under the conditions tested. These include, Desulforhabdus amnigenus, isolated by the growth on acetate (Oude Elferink et al., 1995), and Desulfoacinicum infernum, a thermophilic bacterium isolated on lactate from a petroleum resevoir (Rees et al., 1995). 16S rRNA sequence analysis revealed a close and a moderate relation, respectively of D. amnigenus and D. infernum, with the genus Syntrophobacter, and both most likely share a common ancestor with the latter genus. Whether these bacteria lost the ability to grow syntrophically or that Syntrophobacter gained this ability, remains unclear.16S rRNA analysis of two other cultures enriched from granular sludge showed that syntrophic propionate-oxidation was not restricted to the genus Syntrophobacter. The SYN7 bacterium was not related to Syntrophobacter, but moderately related to the syntrophic benzoate oxidizing genus Syntrophus (Wallrabenstein et al., 1995a) (Chapter 6). Furthermore, the sporeforming syntrophic propionate-oxidizing bacteria SporeA and SporeB (Chapter 4), are related to Gram-positive sulfate-reducing genus Desulfotomaculum. Although widely spread throughout the domain of Bacteria, the syntrophic propionate oxidizers that are phylogenetically analyzed so far have one feature in common, they are related to sulfate reducers. It is even possible that they are all sulfate reducers. Such a linkage between syntrophic propionate oxidation and sulfate reduction may originate from common enzymatic pathways, such as the methyl-malonyl CoA pathway, present in the members of the genus Syntrophobacter (Houwen et al., 1990; Stams et al., 1993). Also Desulfobulbus possess such a pathway (Stams et al. 1984), but so far, has not been shown to grow syntrophically on propionate. It is interesting to find out why Desulfobulbus spp. can not oxidize propionate syntrophically, while Syntrophobacter spp. are able to do so. Furthermore, other sulfate-reducing bacteria need to be tested or tested more extensively for the capacity to grow syntrophically on propionate, for instance Desulforhabdus amnigenus or members of the genus Desulfotomaculum.The purification of strain MPOB, resulting in the description of Syntrophobacter fumaroxidans sp . nov., allowed for the testing of growth on many carbon sources in the presence or absence of sulfate. Preliminary experiments indicated that this S.fumaroxidans strain could also grow, although after a long lag-phase, on alcohols and amino acids by the reduction of sulfate. As a consequence it is quite a versatile sulfate reducer and resembles other sulfate reducers in their growth on many different substrates (Widdel and Bak, 1992). However, in methanogenic ecosystems such as granular sludge, propionate appears to be the only substrate available for S.fumaroxidans. This bacterium seems to be specialized in the syntrophic oxidation of propionate, and this feature may be of advantage over other sulfate-reducing bacteria present in methanogenic granular sludge.1 6S rRNA sequence analysis of cocultures and enrichment cultures. Phylogenetic analyses of syntrophic propionate-oxidizing bacteria were performed with mixed cultures. The cultures from which the 16S rRNA genes were obtained, had an increasing complexity. The first culture was a defined coculture consisting of Syntrophobacter wolinii and Desulfovibrio G11 (Boone and Bryant, 1980). Following the analysis of the 16S rRNA gene of Desulfovibrio G11, it was obvious that the other 16S rRNA gene obtained from the culture had to be derived from S.wolinii (Chapter 2). This was not the case for the phylogenetic analysis of the highly purified, but not defined cultures MPOB and KOPROP1, later named S. fumaroxidans and S.pfennigii, respectively (Chapter 8; Wallrabenstein et al., 1995b). Although no other 16S rRNA genes were cloned, it had to be shown that the sequences obtained really came from the dominant syntrophic bacteria. The required evidence was given by in situ hybridization with specific oligonucleotide probes (Chapter 3). Because of the higher complexity of the cultures, more elaborate work was needed for the phylogenetic analysis of the dominant syntrophic propionate oxidizers in the less defined enrichments described in the Chapters 4 and 6. In the case of SYN7, several 16S rRNA genes were cloned and characterized, but the SYN7 sequence was the most abundant one. In situ hybridization with SYN7-specific probes showed that this sequence was indeed abundantly present in the culture, next to sequences that hybridized with a Desulfobulbus- specific probe. However, these Desulfobulbus -like 16S rRNA sequences were not present in the clone library, probably caused by preferential amplification or cloning of the 16S rRNA-genes. This shows that this method of amplification and cloning of 16S rRNA genes may cause biases and should always be followed by feedback methods like in situ hybridization. In the case of the sporeforming syntrophs, only one bacterial morphology was observed, while at least two 16S rRNA sequences were derived from the culture, which were derived form bacteria indicated as SporeA and SporeB. In situ hybridizations with specific probes for the two sequences were performed to investigate from which bacteria these sequences were derived. Unfortunately, this did not result in unambiguous fluorescent signals, probably because the Gram-positive cell wall could not be made penetrable for the probes with our methods. Therefore, it still can not be ruled out that both sequences were obtained from the same organism. However, dot blot hybridizations showed that only the probe specific for the SporeB sequence hybridized with another enrichment of propionate oxidizers by pasteurization, indicating that the sequences are present in separate organisms.Despite of these problems, the 16S rRNA sequence-analysis done on these enrichment cultures gave valuable information on the phylogenetic status of the dominant bacteria in the cultures, which may help in the isolation of these and other syntrophic propionate-oxidizing bacteria in the future. Furthermore, specific oligonucleotide probes were designed on the basis of these sequences, which have helped to study their presence and their dynamics in granular sludge as described below.Detection of syntrophic propionate-oxidizing bacteria in granular sludge. Fluorescently-labeled specific 16S rRNA-based oligonucleotide probes were used to detect the syntrophic propionate-oxidizing bacteria in anaerobic granular sludge.For this, two types of granular sludge were studied, (i) granular sludge that originated from an upflow anaerobic sludge bed (UASB) reactor treating sugarbeet-processing wastewater (Chapter 5) and (ii) sludge from an UASB reactor treating potatoprocessing wastewater (Chapters 6 and 7). S. wolinii and S. pfennigii were never detected in these two types of granular sludge, not even after enrichment on propionate. S. fumaroxidans -like bacteria were present in large numbers in sludge that originated from the sugarbeet wastewater-treating UASB-reactor. SYN7 was present in the potato-wastewater sludge, but could only be detected significantly after enrichment on propionate. Both S.fumaroxidans and SYN7 were only detected in the sludge from which they were originally enriched. From the two sporeforming syntrophic bacteria, just the SporeB sequence was found in potato-processing wastewater treating sludge, but only after pasteurization and enrichment on propionate .It seems that each type of sludge possesses its specific syntrophic propionateoxidizing bacterium, depending on the wastewater to be treated. This could be due to selection caused by the wastewater itself, or due to the different sources of inoculum used to start up the reactors, such as activated sludge or manure. To test this, granular sludge could be transferred from one reactor to another reactor that treats a different type of wastewater, and then follow the population dynamics of the syntrophs. Furthermore, the abundance of S. wolinii and S.pfennigii could be determined in the sludge from which they were isolated, and more importantly, it could be investigated if also other syntrophs are present in that sludge.Architecture of granular sludge. The architecture of the three investigated sludge types appeared to be dependent on the influent of the reactor. The granules obtained from the UASB reactor treating sugarbeet wastewater, fed with either sucrose or volatile fatty acids (VFA), and the granules from the UASB reactor treating potato wastewater, all had a layered structure in which both syntrophic and methanogenic bacteria were observed. However, there are marked differences in the architecture and number of layers between the three sludge types. Moreover, the abundance of the different trophic groups varied, and the organization of the syntrophic consortia and the microbes present were quite distinct. To illustrate these differences, models are designed that represent active granules of a moderate size. However, the centre of large granules often consisted of large cavities, probably due to substrate transport limitation, which cause lysis of the microbes (Alphenaar et al., 1993).The sludge fed with sucrose and VFA that originated from the UASB- reactor treating sugarbeet wastewater, had been operated with a defined influent for several months. This presumably led to a reduction of the diversity of the organisms in the sludge. Based on the results obtained by in situ hybridization (Chapter 5), the following schematic architecture of the sucrose-fed sludge is proposed. This sludge probably contains only four trophic groups of microorganisms, that form granules of a very simple architecture. Fermentative bacteria are present in the loosely bound outer layer and form propionate and acetate from sucrose. Inside the granule syntrophic microcolonies are present in which the syntrophic acetogens utilize the propionate to form acetate and hydrogen, and carbon dioxide and/or formate. These are then used by the Methanobacter- like methanogens also present in the syntrophic microcolonies. The acetate is utilized by Methanosaeta sp. microcolonies present adjacent to the syntrophic microcolonies. The results obtained from the VFA-grown sludge (Chapter 5) showed an architecture different from that of the sucrose fed sludge, which leads to the following model. The outer layer contains bacteria that are more inside the granule, underneath this layer there is a thick layer of mainly aceticlastic Methanosaeta sp., but also Methanosarcina microcolonies are seen in this layer. These latter methanogens have a lower affinity for acetate than Methanosaeta and will only by present if the concentrations are sufficiently high (Jetten et al., 1990). Underneath this layer there is a layer of syntrophic microcolonies and microcolonies of Methanosaeta sp. , degrading propionate and acetate, respectively. The results obtained from the in situ hybridizations of granular sludge of the UASB-reactor treating potato-processing wastewater (Chapter 7) was quite different to the granules described above, which leads to the proposal of the following model. This sludge was fed with industrial potato wastewater containing at least starch, lactate, butyrate, propionate and acetate as substrates. Therefore, the microbial community in the sludge is more diverse. This is clearly shown in the many different morphologies present in the thick outer layer of the sludge. Underneath the outer layer there is a layer present with small syntrophic microcolonies and Methanosaeta microcolonies and maybe other kinds of microcolonies. Remarkably, these microcolonies were all concentrically orientated. In situ hybridizations indicated that the microcolonies in the centre were less active than those between the centre and the outer layer, probably due to substrate limitation (Chapter 7). This suggests a development of the granule from the centre towards the outer layer.The diversity of the potato-starch grown granules was also shown by the enrichment on propionate and on propionate with sulfate, which selected for syntrophic and sulfidogenic. propionate-oxidizing bacteria, respectively (Chapter 7). This research showed that both trophic groups of bacteria were present in the inoculum sludge. The sulfidogenic propionate oxidizing Desulfobulbus sp. colonized the outer layers of the granules, while the syntrophic SYN7 bacteria formed active microcolonies inside the granule. The morphological diversity in both types of granules decreased dramatically, due to cultivation on one substrate. This was seen by the disappearance of the diverse outer layer, and by the decrease of bacterial rRNA in the dot blot hybridizations.The most interesting difference between the sludge that originated from the sugarbeet- wastewater UASB-reactor and the potato-wastewater treating reactor was the organisation and the types of microbes in the syntrophic microcolonies. The sludge grown on sucrose and VFA contained syntrophic microcolonies consisting of thick MPOB-like rods juxta-positioned with rods that hybridized with a Methanobacteriales -specific probe, morphologically resembling Methanobrevibacter cells. A similar morphological organization of syntrophic microcolonies was previously observed (MacLeod et al, 1990; Grotenhuis et al., 1991; Prensier et al., 1988; Wu et al., 1992). Methanobrevibacter only uses hydrogen as substrate, which has a low solubility in the water, and because every MPOB-like cell is surrounded by methanogens, interspecies hydrogen transfer might the mechanism by which the reducing equivalents are transferred. However, MacLeod et al. (1990) measured a high formate utilizing activity in sludge with syntrophic microcolonies containing Methanobrevibacter -like cells, although they did not measure hydrogen utilizing activity of the sludge. In contrast to the granules mentioned above, the granules treating potato-wastewater contained syntrophic microcolonies with densely packed small SYN7 rods that were intertwined with chains of cells morphologically resembling Methanospirillum hungatei. These cells hybridized with a probe specific for the genus Methanogenium and related organisms, including M. hungatei. This type of microcolonies have not been reported before. In the SYN7 microcolonies not every syntrophic acetogen is surrounded by methanogens. Since Methanospirillum can also utilize formate, it is possible that formate transfer might occur in these microcolonies.A layered structure of granules was proposed and shown in models by others (MacLeod et al., 1990; Macario et al., 1991). Macleod et al. (1990) suggested for sucrose-fed granules also the presence of four trophic groups of microbes organized as follows; a central core of Methanosaeta (Methanothrix) surrounded by a layer of hydrogen-producing acetogens and hydrogen-consuming methanogens, and an outer layer of fermentative bacteria. In the model described here the two inner layers are always mixed, and in the case of VFA grown granules Methanosaeta even form a layer around the syntrophic microcolonies. The earlier observation that only Methanosaeta cells were present in the central core (MacLeod et al, 1990) could be due to the size of the granules, so that only acetate is able to reach the central core. A general model for the architecture of granules converting organic matter based on previous models and the observations presented here suggests the presence of two active layers and a substrate-transport limited central core. In the outer layers mainly microorganisms are present which degrade the substrates in the surrounding medium or influent, independently from other microorganisms. The inner layers consist of microbes that degrade substrates which are produced by the primary layer, e.g fatty acids, acetate and hydrogen and/or formate, and which are dependent on each other for growth, such as the syntrophic consortia. The presence and size of the central core depends on the size of the granule. In large granules there will be a central core consisting of large cavities occupied by Methanosaeta sp. which degrade acetate, produced by the second layer as the terminal intermediate in the breakdown of organic matter. The methane which is produced throughout the granule will disappear from the granule through the numerous cavities and channels in the granule.Conclusions. The main conclusions of the research presented in this thesis are:i There are at least three phylogenetically different groups of syntrophic propionate- oxidizing bacteria which are all related to sulfate-reducing bacteria. The first group is the genus Syntrophobacter , consisting of the species S.wolinii S. pfennigii and S. fumaroxidans, which are closely related to each other and belong to the delta subclass of Proteo-bacteria. These bacteria were shown to be sulfate reducers themselves. The second group is related to the genus Syntrophus also belonging to the delta subclass. The third group are sporeforming syntrophic bacteria related to the Gram-positive genus Desulfotomaculum.ii Granular sludge originating from different sources treating various wastewaters contained specific syntrophic propionate-oxidizing subpopulation. Sludge originating from the UASB reactor treating sugarbeet-processing wastewater, mainly contained S. fumaroxidans as propionate-oxidizer, which are juxta-positioned with Methanobrevibacter -like cells in syntrophic microcolonies. Sludge from the UASB reactor treating potato-processing wastewater, contains SYN7-like bacteria growing in microcolonies which are intertwined with chains of Methanospirillum -like methanogens.iii The architecture of granular sludge and the presence of specific microbial subpopulations in this sludge is dependent on the substrates and electron acceptors available in the influent. Fermentative bacteria and sulfate-reducing acetogens are located in the outer layers of the granule, while syntrophic consortia and aceticlastic methanogens are located in the inner layers.

Decline and succession in dune scrub vegetation of sea buckthorn, Hippophaë rhamnoides L. in the Netherlands was previously found to be stimulated by biotic soil factors. In the present study, the role of plant-parasitic nematodes and root fungi was investigated. A field survey showed that during the ageing of H. rhamnoides vegetation there was a change in the plant-feeding nematode fauna. Certain nematode species, such as Tylenchorhynchus microphasmis Loof, were present in different stages and their population increased from stage to stage when expressed per unit of root length. The fraction of root length infected by VA-mycorrhizal fungi increased, as did soil concentrations of nitrogen and organic matter. Pot experiments with selective inhibition of groups of soil organisms indicated that soil fungi could be harmful to H . rhamnoides in the presence of nematodes, although fungi were not harmful on their own. The impact of various densities of T . microphasmis, either alone or in combination with the nitrogen-fixing actinorhizal symbiont Frankia or the fungus Cylindrocarpon destructans Zinnsm. Scholten, was assessed in pot experiments with controlled soil moisture conditions. Results showed inhibitory effects of the nematode on growth and on the acquisition of certain nutrients (especially phosphate). Although root nodulation and nitrogen fixation by Frankia were reduced, this did not appear to be the primary cause of growth reduction. No synergic interaction was found between T . microphasmis and C. destructans in their effect on growth of H . rhamnoides. The results of this study support the hypothesis that the increase in the concentration of soil nitrogen in the field stimulates nematode activity which in its turn impedes the acquisition of phosphate by H. rhamnoides. This reduces the competitive ability of this shrub and gives way to other plant species in the vegetation which are less sensitive to root damage.

Studies on the root nodules of Coriaria nepalensis using light and electron microscope revealed that the nodule structure and morphology of the endophyte is unique among the actinorhizal plants and similar only to those of Datisca cannabina. Plant cells containing the endophyte were relatively enlarged as compared to the non-infected cells. Transverse sections of the nodules revealed that the infected cells form a compact kidney-shaped zone around the stele which is located on one side of the nodule lobe. Inside the host cell, the hyphae of the endophyte were located in the peripheral region of the cytoplasm whilst the elongate vesicles were present near the central vacuole. Host cell mitochondria were abundant and mostly found in close vicinity of the hyphal/vesicular junction of the endophyte. In situ hybridization studies showed localization of the nifH mRNA within the elongate vesicles of the endophyte. The results indicate that the elongate vesicles of the Coriaria nodule endophyte are functionally identical to the spherical vesicles of the Alnus endophyte.Acetylene reduction (C 2 H 2 reduction) and uptake hydrogenase activity of the root nodules of Coriaria nepalensis and Datisca cannabina were measured to study seasonal fluctuations in the enzyme activities. Nitrogenase activity of the nodules of both plants showed biphasic curves with peaks in spring and late summer. Less than 15% of the enzyme activity was retained by the nodules in winter. Results showed that maximum nitrogenase activity coincides with the period of vigorous plant growth in both hosts. Uptake hydrogenase activity was detected in nodules of both plants throughout the year. The peak enzyme activity in both Coriaria and Datisca nodules was recorded in May. Vesicle cluster fraction of the crushed nodule suspensions showed the highest uptake hydrogenase activity, indicating that the enzyme is associated with the endophyte.Nine Frankia-like a ctinomycetes were isolated from root nodules of Coriaria nepalensis. All isolates formed hyphae and sporangia typical of Frankia, but failed to induce nodules on Coriaria seedlings, or reduce acetylene in pure culture. The relationship of these strains to atypical and typical Frankia strains isolated from other actinorhizal plants and various other actinomycetes was investigated by comparing fatty acid patterns. All Frankia strains, including atypical isolates, showed fatty acid profiles distinct from those of Actinomyces , Geodermatophilus , Nocardia , Mycobacterium and Streptomyces. ForFrankia strains a characteristic pattern of five fatty acids was found that comprised 75% or more of the total content. Three subgroups were identified among 30 Frankia strains compared. Atypical strains isolated from Coriaria were found in the largest subgroup which contained most Frankia strains from other hosts, while ineffective strains from other hosts were distributed in all three subgroups.Non-infective, atypical strains isolated from Coriaria nodules were characterized further by partial 16S rRNA sequence analysis. Ribosomal RNA isolated from all atypical strains hybridized strongly with a Frankia genus probe and indicated that the isolates did belong to the genus Frankia. This was confirmed for two of the isolates (Cn3 and Cn7) from Coriaria nodules by comparison of partial 16S rRNA sequences of the cloned amplification products. Both isolates showed about 95% sequence homology to those of the confirmed Frankia strains from other hosts. Similar high homology values were obtained when 16S rRNA sequences of Cn3 and Cn7 were compared with those of the amplification products obtained directly from nodules of Coriaria nepalensis. In two variable regions compared, none of the isolates showed identical sequences to those obtained from nodules. These results suggest that the isolates are members of the genus Frankia , but different from the strain identified within the root nodules of Coriaria by 16S rRNA sequence analysis.The endophyte of the root nodules of Datisca cannabina was identified as Frankia by sequence analysis of the partial 16S rRNA gene, amplified directly from the nodules by polymerase chain reaction. Moreover, 16S rRNA sequence analysis indicated that the endophyte of Datisca nodules is closely related to that of the Coriaria nodule endophyte, to which it also resembles morphologically in the symbiotic state. Relatedness of the endophytes of Coriaria and Datisca nodules was further proved by the closely related nifH sequences obtained from the nodules by PCR. 16S rRNA sequence analysis of the non-infective atypical strain Dc2 obtained from the Datisca nodules revealed its close relationship to the genus Frankia.Nodulation in Datisca was achieved with crushed nodule suspensions from both Coriaria and Datisca , whereas various pure Frankia strains failed to induce nodulation. The origin of the endophyte in Datisca nodules induced by crushed nodules of Coriaria collected from Murree (Pakistan) was investigated by comparing partial 16S rRNA sequences with those of the endophytes of both plants. The sequences in the variable region were found to be identical to those of Coriaria nodule endophyte, confirming that the endophyte of Coriaria can cross-nodulate Datisca plants. Acetylene reduction by the root nodules indicated effectiveness of the nodules. Coriaria seedlings could only be nodulated by crushed nodule suspensions of Coriaria nepalensis. All pure cultures of Frankia including atypical strains used as single inoculum or in combination with the nodule filtrate, failed to induce nodulation on Coriaria seedlings.Sequence information obtained from the root nodules of Coriaria and Datisca by PCR amplification of partial 16S rRNA gene of the endophyte, was used to develop an oligonucleotide probe. The probe was designed against the V6 variable region of the 16S rRNA that was identical in the 16S rRNA of the endophytes of Coriaria and Datisca. The probe did not cross-react with RNA of typical and atypical Frankia strains isolated from Alnus , Casuarina , Colletia , Comptonia , Elaeagnus and Hippophae.To investigate the presence of Frankia strains infective on Coriaria and Datisca , soil samples were collected from different areas of Pakistan and used as inoculum. Three types of soil samples i.e. rhizosphere soil, non-rhizosphere soil and soil from sites effected by erosion were used to inoculate seedlings of Coriaria and Datisca. Abundance of compatible Frankia strains in most areas was indicated by profuse nodulation of the host plants, while soil samples from Athmokam and Puddar failed to induce nodulation on Coriaria seedlings. Exceptionally good nodulation of Coriaria and Datisca with soil samples collected from Abbottabad and Kaghan respectively, reflects abundance of compatible Frankia strains in the soils. This also indicates that these soils can be used effectively as an inoculum for large scale inoculation programmes instead of nodules that are difficult to obtain throughout the year.Successful nodulation of plants inoculated with soil suspensions can confirm the existence of compatible Frankia cells, but the possibility of the presence of more than one infective strain cannot be ruled out. Genetic diversity among the Frankia strains in Datisca nodules originaly induced by soil inoculum collected from different areas of Pakistan, was determined by sequence analysis of 16S rRNA. Part of the 16S rRNA gene was amplified from nodules and the cloned PCR products were screened initially with a Frankia genus probe and then with the specific probe. Three categories of the clones were identified i.e. those reacting with both probes and those hybridizing with only the Frankia genus probe, or only with the probe specific for the Datisca endophyte. Four types of Frankia sequences and one non -Frankia sequence were identified by sequence analysis of the cloned PCR products. The results suggest the presence of more than one Datisca- compatible Frankia strains in the soils collected from different areas.

The research described in this thesis focusses on the role of biotic factors encountered with the establishment of the symbiosis between black alder plants ( Alnus glutinosa ) and introduced Frankia strains. A selection of plant clones and Frankia strains that gave optimal nodulation and nitrogen fixation in forestry was made. For this reason nodulation tests with increasing complexity were set up. An attempt was made to investigate whether introduced strains behaved differently on plants grown under axenic and non-axenic conditions. Since Frankia strains were difficult to identify by conventional techniques, special attention was given to the development of new molecular techniques for identification of the strains at the nucleic acid level.Inoculation testsInitially, plant material of two physiologically different ecotypes of Alnus glutinosa, the forest ecotype "Bentheim" and the pioneer ecotype "Weerribben", respectively, was selected. Using tissue culture techniques plant material of both ecotypes was cloned in order to obtain genetically identical plants (Chapter 2). These micropropagated plants were used to set up a standardized inoculation system under axenic conditions in order to study the genetically determined nodulation ability and nitrogen-fixing capacity of Frankia strains and to select superior Frankia strains as source of inoculum (Chapter 3). The usefulness of selected Frankia strains as inoculum was further tested under more practical conditions, in perlite as model environment for nitrogen-limited conditions and in two soils, representing natural environments with different nutritional factors and different microbial populations. The results of the inoculation tests under axenic conditions were confirmed by studies under greenhouse conditions.The performance of the symbiosis was effected by many variables, e.g. the plant genotype, the Frankia strain and environmental conditions. The influence of the environmental conditions became more pronounced when plants were grown on either a sandy loam ("Bentheim") or a peat ("Weerribben") soil and inoculated with Frankia strains. Plant growth was positively influenced, e.g. by mycorrhizal fungi in "Bentheim" soil, or negatively influenced, e.g. by oomycetes in "Weerribben" soil. The effects of inoculation with Frankia on plant growth remained minimal. The establishment of the introduced Frankia strain was also dependent on the soil conditions. The introduced spore(-) Frankia strain was only able to compete with the natural spore(+) population of the "Weerribben" soil. Introduction of this strain to "Bentheim" soil did not show any establishment of the introduced strain. In contrast to the sandy loam of "Bentheim" which was rich in nutrients, the peat of "Weerribben" was a representative of poor soils. It could therefore be used for feasability studies in inoculation programmes. The use of pure cultures of Frankia as inoculum instead of soil or crushed nodules, has the advantage to prevent the contamination of the plant with root pathogens. Pure cultures did not result in a better symbiosis.Atypical Frankia strainsScreening of several isolates obtained from nodules of both alder ecotypes indicated the existence of atypical, ineffective Frankia strains. The alder clones used showed variable resistance against infection of the ineffective strains (Chapter 4). When compared with growth after the addition of a single strain dual inoculation of typical, effective Frankia strains and an ineffective Frankia strain to both alder clones showed growth increment of the plants (Chapter 5). The growth enhancing effect of the ineffective Frankia strain was not paralleled by increased number of nodules. Nothing is known yet about the growth stimulation by atypical Frankia strains. The results indicate that simultaneous inoculation of different Frankia strains to Alnus plants can be profitable for the host plant.Ribosomal RNABecause the ineffective Frankia strains lacked morphological and physiological characteristics of typical Frankia strains and because nodule formation on actinorhizal plants might be reduced or even absent, detection of the ineffective strains and studies on their competitive abilities were quite difficult. Reliable markers which could be used to detect both types of Frankia in nodules and in soil without reisolation had not been available at that moment. An attempt was made to find specific markers in a molecule which was commonly used to unravel evolutionary relationships: the 16S ribosomal RNA. New sequencing techniques allowed the rapid determination of total or almost total 16S rRNA sequences. Total 16S rRNA sequences indicated the presence of conserved and variable regions. Conserved regions had been used to investigate quantitative evolutionary relationships amoung bacteria. The conserved regions of the total 16S rRNA sequence of the effective Frankia strain Ag45/Mut15 were compared with aligned sequences of other actinomycetes and used to determine the position of the family Frankiaceae in the phylogenetic tree of the actinomycetes (Chapter 6).Analyses of variable regions of 16S rRNA of closely related organisms indicated sufficient variation, despite the fact that DNA/DNA homology studies suggested these two species might actually be one and the same. Large differences in DNA/DNA homology studies of Frankia which were also obtained between strains of one compatibility group suggested chances on large variation within the variable regions of different strains. Sequence analyses of variable regions of 16S rRNA of two ineffective Frankia strains (i.e. AgB1.9 and AgW1.1) and the effective strain Ag45/Mut1 5, all belonging to the Alnus -compatibility group, showed large differences in base composition. These sequences were used to design complementary synthetic oligonucleotides that could act as specific probes in hybridization experiments. The specificity of these probes was shown in hybridization experiments against immobilized rRNA from 23 Frankia strains belonging to different compatibiliy groups and of several related soil actinomycetes. The probes were able to distinguish between Nif +and Nif -strains, between several Nif -strains and between several Alnus compatible Nif +strains and strain AgKG'84/4 also belonging to the Alnus -compatibility group (Chapter 7). Strong strain specific sequences, however, were not obtained. The design of oligonucleotide probes opens up the possibility to investigate competitive abilities of selected strains under defined conditions, e.g. in model systems with perlite and defined Frankia strains. The question whether competition studies under these controlled conditions are ecologically relevant needs further investigations because little basic knowledge on Frankia population dynamics is yet available. The application of probes to identify introduced strains in soil remains restricted, due to the low specificity for strains. Up to now we are not able to design reliable strain specific probes that can be used to follow the establishment of introduced Frankia strains in natural environments. A much more . promising application of probes towards rRNA is concerned with the development of a genus-specific oligonucleoticle probe against Frankia (Chapter 9) that theoretically enables quantitative detection of total Frankia populations.RNA extractionThe application of oligonucleoticle probes in the detection of specific Frankia strains does not only depend on specificity of the probes but also on the development of a reliable isolation method for target sequences. Ribosomal RNA is preferable to DNA as target because of its relative abundance in large amounts in metabolically active cells. Actinorhizal nodules represent enrichments of Frankia, which are metabolically highly active and consequently contain large amounts of Frankia RNA. Our investigations resulted in the development of a rapid RNA extraction method that was sensitive enough to investigate strain composition also from very small nodules or lobes (Chapter 8). The detection of target sequences, however, remained limited by the design of specific probes and the ratios of different target sequences in one sample. For reliable signal expression in hybridization experiments quite similar amounts of target sequences per sample were needed.So far, the usefulness of rRNA sequences as targets for oligonucleoticle probes was only shown in combination with pure cultures of Frankia (Chapter 7) or in metabolically highly active enrichments, e.g. nodules (Chapter 8). Terrestrial environments like soil contain populations of many different microorganisms. These populations normally grow under suboptimal nutrition conditions. Bacteria adapt to these conditions by forming special starvation cells, which are metabolically inactive and contain only low amounts of rRNA. The starvation respons often results in viable, but non-culturable populations. The recalcitrant character of Frankia, which are difficult to isolate, makes it a useful model microorganism of soil bacteria. The application of oligonucleoticle probes for detection of Frankia in soil depends on the development of an extraction method for RNA. RNA directly isolated from soil as target for Frankia specific oligonucleoticle probes was useful in detection of Frankia (Chapter 9). Quantification of the obtained signals, however, is still unreliable because Frankia is a hyphae forming organism. It is also quite difficult to correlate cell numbers (theoretical estimation) to the amount of RNA. The concentration of these molecules in an organism is a function of the activity of the individual cell. Quantification of hybridization signals therefore depends on the availability of basic information of the metabolic activity of Frankia cells in soil. This information, however, is very difficult to obtain for recalcitrant microorganisms like Frankia. It is much easier for other microorganisms, e.g. for Streptomyces . Streptomyces spores are quite easy to isolate from soil and the establishment of Streptomyces cells, i.e. as spores or as mycelium in soil, is well studied. Quantification based on hybridization signals must be possible when this basic knowledge is available. In case of Frankia methods that enable quantification must still be developed. Similar to Streptomyces these quantification methods for Frankia can be of direct character, e.g. quantitative extraction of spores, or of indirect character, e.g. determination of mycelium by phage counts.The development of rapid and sensitive methods to detect Frankia on the basis of rRNA sequences opens up new ways to study other recalcitrant microorganisms in the environment. This molecular approach in microbial ecology can definitely further be explored when the advantages of rRNA as stable target and the rapid extraction of RNA from soil can be combined with in vitro amplification methods commonly used with DNA or mRNA. Promising approaches can also be expected in in situ studies using hybridization signal intensity of fluorescent dye labelled oligonucleotides and the amount of rRNA as criterium for bacterial activity.

The actinomycete Frankia is a symbiotic nitrogen fixer, living in root nodules of many non-leguminous plants. A typical characteristic of this endophytic organism is the formation of specialized swollen cell structures, called vesicles. Frankia vesicles have been brought In relation to nitrogen fixing activity, but many questions about their formation and their function still had to be answered. The present study was done to investigate the function of these vesicles. First, an electronmicros copy study has been initiated to investigate the ultrastructure of Frankia cells. The vesicle envelope was studied in detail, and shown to be a multilaminate structure, which was in accordance with other published results. The vesicle envelope was continuous with the hyphae cell wall. It was also shown that no difference was found in the ultrastructure of vesicles with, or without induced nitrogenase (chapter II). The formation of vesicles in different Frankia strains was investigated by comparing a strain with, and one without vesicle formation in the presence of ammonia. In the presence of ammonia, nitrogenase is generally repressed. The vesicles that were formed in the medium with ammonia did not contain nitrogenase (chapter III).Frankia nitrogenase could be detected by using antisera against nitrogenase from Rhizobium leguminosarum or from Azotobacter vinelandii. The latter appeared to be very specific, and was used for localization of nitrogenase on ultrathin cryosections. It could be shown that nitrogenase is present uniquely in the vesicles in Frankia strains Ccl.17 and EAN1pec (chapters IV-1, IV-2). In these cells, probably a reducing environment is maintained, thus preventing inhibition of nitrogenase by oxygen. In accordance herewith, it has been reported that one Frankia strain (Cc13) was able to fix nitrogen without forming vesicles, under microaerobic conditions. The Influence of microaerobic conditions on the localization of nitrogenase in strain Ccl. 17 was tested. In this strain, no change of the localization of nitrogenase in the vesicles with reduced oxygen partial pressure has been observed (chapter IV-3). From these results, it can be concluded that nitrogenase is generally located within the vesicles, although vesicle formation and nitrogenase induction can be controlled independently in some strains. The mechanism of regulation of the strict localization is still unknown. One of the possibilities of repression of nif- genes is the occurrence of DNA rearrangement during vesicle differentiation. This possibility has been investigated by comparing DNA from vesicle clusters from nitrogen fixing root nodules with DNA from hyphae from the corresponding pure culture. No differences could be detected in the nif -genes from these two preparations. The results mean that the question, how nif- gene expression i s repressed in the hyphae of Frankia, still need to be answered. The progress made in the past years in Frankia molecular genetics promises that many still open questions about regulation mechanisms in Frankia will be answered in the near future.

The research reported in this thesis deals with the symbiosis of Frankia spp. and Alnus glutinosa. Frankia spp. are actinomycetes giving rise to the formation of nitrogen-fixing nodules on the roots of a number of non-leguminous plants. In these nodules Frankia spp. live within the plant cells and obtain all sources of carbon and energy from the plant, giving fixed nitrogen in exchange.To answer the question what compounds Frankia spp. obtain from the plant, insight into the metabolism of this microorganism is required. To obtain this insight, researches have been made with the free-living Frankia AvcI1, isolated from root nodules of Alnus viridis ssp crispa by Baker et al. (1979). This organism has been isolated and cultivated on complex media. In order to obtain insight into the C- and N-source requirements of Frankia AvcI1, a simple and well-defined growth medium is needed. The composition of this medium and the C- and N-metabolism of Frankia AvcI1 are the subjects of Chapter II of this thesis.In Chapter II.1 it is shown that Frankia AvcI1 is able to grow on a medium containing Tween 80 (an oleate ester of polyethyleneglycol sorbitan) as sole C-source and either glutamic acid or NH 4 Cl as sole N-source. The growth yield of Frankia AvcI1 on various media is reported. It is shown that the doubling time of Frankia AvcI1 growing on QMOD/Tween medium is about 2 days, which is slow, even for an actinomycete. When growing with Casamino acids as nitrogen source, Frankia AvcI1 selectively takes up glutamic acid and aspartic acid, leaving the concentrations of the other detectable amino acids unchanged. The mechanism of this selection is still unclear.Chapter II.2 contains further data on the C-sources utilized by Frankia AvcI1 It is shown that Frankia AvcI1 can utilize as C-source also other Tweens, viz. Tweens 20, 40, 60 and 85, and in addition several fatty acids, viz. acetic, propionic, butyric, valeric, caproic, caprylic, capric, palmitic and stearic acids. No growth of Frankia AvcI1 was observed on media containing triglycerides as C-sources.The dependence of the growth yield on the nature of the carbon source and the concentration of NH 4+ in the media is shown. Utilization of NH 4+ as nitrogen source of Frankia AvcI1 growing in the Tween/NH 4+ medium is confirmed by incorporation experiments with 15 N-NH 4 Cl.The results reported in Chapter II.3 show that Frankia AvcI1 does not take up glucose from a medium containing both glucose and Tween 80. In agreement with this observation, it is demonstrated that the activities of isocitrate lyase and malate synthase in cells grown on the QMOD/Tween medium (containing both glucose and Tween 80) are of the same order ofmagnitude as in cells grown on the Tween/NH 4+ medium (containing Tween 80 as sole C-source).Organisms growing on fatty acids, and degrading these compounds to acetyl-CoA, start gluconeogenesis with the action of the glyoxylate cycle (Kornberg and Krebs, 1957) leading to the conversion of 2 acetyl-CoA molecules to 1 molecule of succinate. The presence of the glyoxylate cycle enzymes isocitrate lyase and malate synthase in Frankia AvcI1 cells grown with Tween 80 as C-source is therefore not surprising. The succinate formed in the glyoxylate cycle can be converted to phosphoenolpyruvate by the subsequent action of the enzymes succinate dehydrogenase, fumarase, malic enzyme and pyruvate orthophosphate dikinase, which are found in cell-free extracts of Frankia AvcI1 From phosphoenolpyruvate, gluconeogenesis can continue with the action of phosphoglycerate kinase and 3-phosphoglyceraldehyde dehydrogenase.For energy generation, Frankia AvcI1 can oxidize the acetyl-CoA derived from fatty acid breakdown in the citric acid cycle, although the low activity of the enzyme succinyl-CoA synthetase leaves room for the presumption that the citric acid cycle is not very operative in Frankia AvcI1 This is not impossible since the oxidation of fatty acids yields many reducing equivalents.The data contained in Chapter II.4 show that Frankia AvcI1 does not take up succinate from a medium containing acetate plus succinate or succinate alone. In accordance, no repressing effect of succinate on the activities of the glyoxylate cycle enzymes was observed, whereas in cells grown on propionate these enzymes were not found, indicating that they are hot constitutive in Frankia AvcI1Frankia AvcI1 is able to utilize several amino acids as sole nitrogen source, viz . alanine,γ-aminobutyric acid, aspartic acid, glutamic acid, glycine, leucine, phenylalanine, serine, threonine, tyrosine and valine.No differences in C- and N-source requirements were observed for the three Frankia strains AvcI1, CPI1 (isolated by Callaham et al., 1978, from Comptonia peregrina root nodules) and AgSp+1 (isolated by Quispel and Burggraaf, 1981, from Alnus glutinosa spore-(+) type root nodules). These three strains were shown to utilize either Tween 80 or acetate, but no ethanol, lactate, glucose or succinate as sole C-source, and either NH 4 Cl, Casamino acids, aspartic acid or alanine as N-source.From the data given in Chapter II it will be clear that Frankia AvcI1 is able to grow on a well-defined medium, which is a prerequisite for obtaining a clear insight into the physiology of the organism. The alcoholic root extract (Quispel and Tak, 1978) or the soybean lecithin (Lalonde and Calvert, 1979), added to the medium as growth factors, can be replaced by Tween 80 or fatty acids which are utilized as carbon source, while NH 4 Cl or amino acids can be utilized as nitrogen source.It is unknown whether Frankia spp. grow with the same growth rate on media with different constituents. The growth yields presented in Chapter II are usually small as compared to the amount of carbon present in the media. This suggests that either an additional component of the medium is present in limiting concentrations or that Frankia spp. can only grow for a limited time after inoculation.The only group of compounds known so far to be utilized as C-source by free-living Frankia AvcI1 are fatty acids, either free or esterified. The question what carbon source Frankia spp. living symbiotically in the nodule obtain from the plant, still remains to be answered. Based on the data obtained for free-living Frankia spp. it is not unlikely that some fatty acid functions as C-source under such conditions. Other possible candidates for this function are plant lipids like the alcoholic root extract (Quispel and Tak, 1978) or the soybean lecithin (Lalonde and Calvert, 1979). It is unlikely that sugars like glucose or dicarboxylic acids like succinate are playing this role, unless the ability of Frankia spp. to take up one or more of these compounds should alter during the transition from the free-living to the symbiotic stage.The cultivation of free-living Frankia spp. is a powerful tool in discovering symbiotic interactions of the endophyte and the host. In Chapter II.4 of this thesis it is shown that by replacing acetate as C-source of the medium by propionate, the activities of the glyoxylate cycle enzymes are repressed. Other authors (Tjepkema, Ormerod and Torrey, 1980 and 1981, Gauthier, Diem and Dommergues, 1981) reported a medium in which free-living Frankia spp. show vesicle formation and N 2 -ase activity. The present knowledge thus enables one to influence regulation in freeliving Frankia spp., which is important in studying the symbiotic interactions mentioned above.In Chapter III attention is paid to the assimilation of the ammonia produced by the endophyte living symbiotically in the root nodules of Alnus glutinosa grown in the greenhouse from seeds collected in Wageningen. In Chapter III.1 the composition of the pool of free amino acids in root nodules and the xylem tissue of stems is reported. It is shown that citrulline is the predominant free amino acid in nodules, while serine also occurs in relatively large amounts. In xylem tissue citrulline and glutamic acid are prominent.The activities of N 2 -ase, GS, GDH and OCT in root nodule homogenates are reported. From the K m values of GDH for NH 4+ (16 mM) and glutamate (0.9 mM) and the concentrations in the nodule of NH 4+ (1.5 mM) and glutamate (0.5 mM) it is concluded that GDH in alder nodules probably is responsible for the deamination of glutamate and not for the synthesis of this key amino acid. The important function of GDH in the nitrogen metabolism of alder nodules is confirmed by the much higher activity of this enzyme in homogenates of nodules as compared to that in homogenates of root-tips and leaves.The vesicle clusters, which contain the N 2 -ase activity, did not show activity of GS, GDH and OCT. The cytoplasm of the host cells was shown to possess the GS activity, while GDH and OCT are localized in the organelles of the host cells. No activity of NADH-dependent GOGAT was observed.In Chapter 111.2 it is shown that the activity of NADH-dependent GOGAT from root nodules of lupins is inhibited by some compound in the homogenate of alder nodules.Simultaneously with and independently of the present research, Schubert et al. (1981) analyzed the composition of the pool of free amino acids in nodules and the xylem tissue of Alnus glutinosa grown in the field in East Lansing, Michigan. The only difference with respect to these amino acid pools between the American alders and the European alders studied in the present research, is the relatively high amount of serine in nodules of the latter, whereas in nodules of the former this amino acid is not found. The results of both Schubert et al. (1981) and the present research confirm the hypothesis that in alder,citrulline is the main transport vehicle of fixed nitrogen (Miettienen and Virtanen, 1952; Leaf, Gardner and Bond, 1958; Wheeler and Bond, 1970).From the results shown in Chapter 111.2 it can be concluded that our failure to find any GOGAT activity may be ascribed to the presence of an inhibiting compound in the homogenate of alder nodules, so that it is not excluded that GOGAT is active in alder nodules in vivo .The data reported in Chapter III are in accordance with the following model: The nitrogen fixed as ammonia in the vesicle clusters of the endophyte, is assimilated in the cytoplasm of the host cell into glutamate by the action of GS and presumably GOGAT. Glutamate is in part deaminated in the plant organelles by the action of GDH to supply the NH 4+ required for the synthesis of carbamyl phosphate. Another part of the glutamate is converted to ornithine, which in the organelles of the host cellreacts with carbamyl phosphate to form citrulline according to the OCT reaction. Citrulline is excreted from the nodule and serves as nitrogen source for the plant.