Nitrogen fixation

Nitrogen Fixation refers to that property of some taxa of eubacteria or archaebacteria to enzymatically reduce atmospheric N2 to ammonia. The ammonia produced can then be incorporated by means of other enzymes into cellular protoplasm. Nitrogen fixation only occurs in prokaryotes and not in higher taxa. 

I. SIGNIFICANCE OF BIOLOGICAL NITROGEN FIXATION 

Although nitrogen is an essential nutrient for life, little available nitrogen is present in mineral form. Above every hectare of soil at sea level, there are 78 million kg of inert N2 (dinitrogen) gas. Plants, as eukaryotic autotrophs, need either an oxidized or reduced form of nitrogen for anabolism. Only certain prokaryotic organisms “fix” N2 at physiological temperatures. Humans can mimic biological nitrogen fixation using the Haber–Bosch process of chemically reducing nitrogen at high temperatures. However, the process consumes precious fossil fuels, contributes to global warming, and pollutes the environment. Worldwide industrial production of NH3 is annually about 100 million metric tons, threefourths of which is manufactured for fertilizer to grow crop plants. Thus, humans produce about 75 million mt/year of nitrogenous fertilizer, whereas the annual contribution of reduced nitrogen from the biological process called nitrogen fixation is roughly estimated to be about two to three times that amount. As shown in Table 60.1, nitrogen fixation occurs in every natural environment, including the sea. The energy cost for nitrogen fixation is significant. Perhaps 10% of the available fossil fuel energy is used for the production of fertilizer nitrogen, whereas Hardy (1980) estimated that between 1 and 2 billion tons of plant carbohydrates derived from photosynthesis fuel the biological process of nitrogen fixation. There are many genetically diverse nitrogen-fixing eubacteria falling into 27 families and 80 genera, and there are at least three thermophilic nitrogen-fixing genera of Archaebacteria. These nitrogen-fixing families and genera are listed in Table 60.2; the “bluegreen algae” or cyanobacteria are listed in Table 60.3. With the exception of the Azotobacteraceae, there is no genus (or family) whose species are all nitrogenfixing. The potential amount of nitrogen fixed by these bacteria depends largely on the ecosystem in which the organisms are active, as will be discussed in later sections. The amount of nitrogen fixed ranges from only trace amounts for some free-living soil bacteria to 584 kg/ha/year for the tropical tree legume Leucaena. 

II. THE BIOLOGICAL NITROGEN FIXATION PROCESS Dinitrogen gas is both chemically inert and very stable, requiring much energy to break the triple bond and reduce the N_N to ammonia in an endothermic reaction: 3H2_N2 → 2 NH3. This can be accomplished chemically by the Haber–Bosch process or biologically by prokaryotic organisms using adenosine triphosphate (ATP) energy to initiate the bond-breaking reaction. The prokaryotes able to fix nitrogen are extremely heterogeneous, with representatives that are autotrophic, heterotrophic, aerobic, anaerobic, photosynthetic, single-celled, filamentous, free-living, and symbiotic. Phylogenetically, these organisms are extremely diverse and yet the nitrogenfixing process and enzyme system are similar in all of these organisms and depend on a nitrogenase enzyme complex, a high-energy requirement and availability (ATP), anaerobic conditions for nitrogenase activity, and a strong reductant. Nitrogenase has been purified from all known nitrogen- fixing eubacteria. It consists of two components, dinitrogenase reductase (an iron protein) and dinitrogenase (an iron–molybdenum protein). Both enzymes are needed for nitrogenase activity. Nitrogenase reduces N2 and H_ simultaneously, using 75% of the electron flow in the reduction of nitrogen and 25% in H_ reduction. A key characteristic of this enzyme complex is that both components are quickly and irreversibly lost on interaction with free oxygen, regardless of the oxygen requirements of the microbe. The protein usually has a molecular mass of 57,000–72,000 Da and consists of two identical subunits coded for by the nifH gene. It has a highly conserved amino acid sequence. The iron–molybdenum protein has a molecular mass of about 220 kDa and has four subunits, which are pairs of two different types. The _-subunit is coded for by the nifD gene and is 50kDa in size. The _-subunit is coded for by the nifK gene and has a mass of 60 kDa. These enzymes have been sequenced and demonstrate considerable similarity. The nitrogenase complex is large and may amount to 30% of total cell protein. 

Theoretically, as much as 28 moles of ATP are consumed in the reduction of 1 mole of N2. Depending on the method used and the nature of the organism in question, in vitro studies show that the energy requirements vary between a minimum of 12–15 moles and 29 moles of ATP/N2. This energy is not only required for the reduction process, but also to maintain the anaerobic conditions needed for the reaction. Various strategies are used to exclude oxygen from the reaction in nonanaerobic bacteria. Facultative organisms fix nitrogen only under anaerobic conditions. Aerobic organisms exhibit a wide range of methods for protecting the enzyme complex from oxygen. Many grow under microaerophilic conditions accomplished by scavenging free oxygen for metabolism or sharing the ecosystem with other organisms that consume the excess O2. Some evidence indicates that some free-living aerobes, such as Azotobacter, can change the conformation of the nitrogenase protein to form a less oxygen-sensitive protein. Photosynthetic aerobes such as cyanobacteria can form special cells called heterocysts in which nitrogen fixation occurs and which lack the O2-evolving mechanism that is part of photosynthesis. The bestknown protective mechanisms are found in symbiotic nitrogenfixing associations (i.e., Rhizobium–legume), in which the nitrogenase system in the endophyte is protected from excess oxygen by a nodule component, leghemogloblin. Nonleguminous nodules (with Frankia or cyanobacteria) are probably protected by other, as yet undescribed, oxygenrestrictive mechanisms. There are other protective mechanisms as well, but all of these require the diversion of energy from the nitrogen-fixation process itself to maintain a favorable environment for fixation. In addition to the limitations to biological nitrogen fixation (BNF) caused by free oxygen, generally the presence in the environment of combined nitrogen, such as ammonia or nitrate, strongly inhibits both nodulation and N2 fixation. Thus, it is unnecessary to use chemical fertilizer in large quantities for legume crop production. 

III. FREE-LIVING NITROGENFIXING BACTERIA There are many diverse nitrogen-fixing prokaryotes, representing 22 families and 52 genera of eubacteria (excluding the cyanobacteria), as well as three thermophilic genera of archaebacteria. Most of these genera fix nitrogen as free-living diazotrophs. These were first described by Winogradsky in 1893 (Clostridium pasteurianum), and Beijerinck in 1901 (Azotobacter). The discovery of other free-living nitrogen-fixing bacteria lagged until the availability of 15N stable isotopic and acetylene-reduction techniques became common. Thus, most of the freeliving nitrogenfixing bacteria listed in Bergey’s Manual of Systematic Bacteriology have been discribed using these techniques. It is generally accepted that diazotrophs obtain their carbon and energy supplies from root exudates and lysates, sloughed plant cell debris, and organic residues in soil and water. They are found “completely free-living” or in loose associations as a result of root or rhizosphere colonization (the associative bacteria are discussed separately). The quantity of nitrogen fixed is a matter of some controversy. Russian workers have estimated that C. pasteurianum or Azotobacter contribute perhaps 0.3 kg N2/ha/year, compared with about 1000 times that amount provided by a good leguminous association. Associative organisms such as Azospirillum have been estimated to fix from trace amounts to 36 kg/ha. The limitations of the terrestrial BNF system is due not only to the difficulty in obtaining sufficient energy and reductant but also because of the need to divert substrate for respiratory protection of nitrogenase. The cyanobacteria can overcome the environmental constraints faced by most other nitrogen fixers. Being photosynthetically active, these prokaryotes use sunlight to fix CO2 and are thus independent of external energy needs. The families and genera of nitrogen-fixing cyanobacteria are listed in Table 60.3. The free-living cyanobacteria are distributed widely in humid and arid tropical surface soils. Extensive studies, especially under rice paddy conditions, have been conducted in India, Japan, and the Phillippines. Of 308 isolates from Philippine paddy soils, most were identified as Nostoc or Anabaena. Reports of fixation rates ranged from 3.2 to 10.9 kg N/ha/year. Reports of 15–20 kg N/ha/year in rice in the Ivory Coast, 44 kg N/ha/year in Lake George in Uganda, and 80 kg N/ha/year in paddy fields in India have been published. Under temperate conditions, BNF by cyanobacteria has been a major problem in the eutrophication of lakes. About 125 strains of free-living cyanobacteria representing 10 families and 31 genera in all taxonomic groups have been shown to fix nitrogen, but the extent to which fixation occurs and conditions needed for fixation vary greatly. There are four types of nitrogenfixing cyanobacteria—heterocystous filamentous, nonheterocystous filamentous, unicellular reproducing by binary fission or budding, and unicellular reproducing by multiple fission. Heterocysts are thickened, specialized cells occurring at regular intervals in some filamentous cyanobacteria. 

These cells lack the oxygen-evolving component of the photosynthetic apparatus. Heterocysts appear to be the only cells capable of fixing nitrogen aerobically as well as anaerobically. The main function of the heterocysts seems to be to compartmentalize nitrogen fixation because there is little or no photosynthesis in the heterocysts, and all of the energy translocated is available for nitrogen fixation. The other groups of cyanobacteria need to be examined as well because they appear quite active as fixers under anaerobic conditions. Nitrogen fixation in paddy rice could thus be improved. One should not overlook the potential for the improvement of nitrogen fixation in unique environments. There are many reports of intimate associations between nitrogen-fixing bacteria and animals (e.g., termites and ruminants). Because nif genes can be transferred and expressed between the enteric bacteria Klebsiella and Escherichia coli, it may be possible to increase nitrogen fixation by rumen bacteria. Perhaps ultimately in ruminants, it will be practical to substitute engineered diazotrophic enterics for the plant protein now required by these animals. 

IV. ASSOCIATIVE NITROGEN-FIXING BACTERIA Since the 1960s, another type of plant–bacterial interaction has been described as “associative.” This interaction was shown to result from the adhesion of the bacteria to the root surfaces of wheat, corn, sorghum, and other grasses. The major associative nitrogen-fixing systems are summarized in Table 60.4. In vitro studies show that many of these bacteria can achieve high rates of N2 fixation under optimum conditions. However, in the rhizosphere, the ability to survive, grow, and colonize plant roots is a precondition limiting the potential for nitrogen-fixation. The characteristics required for an organism to flourish in the rhizosphere are the ability to withstand the changing physical and chemical soil environments, grow well and obtain all the energy needed from carbon and mineral supplies in the root zone, and compete successfully with other rhizosphere organisms for the limited energy and nutrients available. Estimates of in vivo fixation are extremely variable and give rise to a recurring question as to whether energy substrates in the rhizosphere are sufficient to support growth and nitrogen fixation by these associative bacteria at sufficient levels. The nitrogenase system is repressed by bound nitrogen. Therefore, in the presence of nitrogen fertilizer, BNF by free-living or associative bacteria is reduced. Thus, a mixed BNF-nitrogen fertilizer system would work better under field conditions if the microorganisms could be genetically engineered so that the expression of nitrogenase is derepressed when bound nitrogen is present. V. SYMBIOTIC NITROGEN-FIXING BACTERIA Only three groups of nitrogen-fixing bacteria have evolved mutually beneficial symbiotic associations with higher plants: (1) the filamentous bacterium Frankia, forming root nodules with a number of plants such as alder, Purshia, and Russian olive; (2) heterocystous cyanobacterium with a number of diverse hosts from Cycads to Azolla; and (3) Rhizobium and allied genera with legumes. A. Frankia The genera of dicotyledonous plants found to be nodulated by Actinorhizae (Frankia) are listed in Table 60.5. These include more than 180 species in more than 20 genera, representing at least eight families and seven orders of plants. There is no obvious taxonomic pattern among these hosts. Most of the species described are shrubs or trees and are found in temperate climates, but they have a wide growth range and could be grown in the tropics. In fact, Purshia tridento already is an important rangeland forage crop in Africa, and other Purshia species are harvested for firewood. Casuarina, a vigorous nitrogen fixer, has been planted in Thailand where it can be harvested for construction lumber after 5 years of growth. Almost all the hosts are woody, ranging from small shrubs to medium-sized trees.  Frankia, the nitrogen-fixing endophyte found in nodules of these nonleguminous plants is a genus of prokaryotic bacteria closely related to the actinomycetes. In pure culture, these endophytes behave as microaerophilic, mesophilic, or heterotrophic organisms, usually with septate hyphae that develop sporangia. 

Isolates vary morphologically and nutritionally. Most strains can fix atmospheric nitrogen in pure culture. Nitrogenase genes are highly conserved, and Frankia nitrogenase enzymes closely resemble those of other nitrogen-fixing bacteria. Frankia was first reliably isolated in 1978. The organism grows slowly, requiring 4–8 weeks for visible colonies to be formed in culture. Frankia is similar to other aerobic actinomycetes, producing a separate filamentous mycelium that can differentiate into sporangia and vesicles. The cells are routinely grampositive, but unlike other gram-positive bacteria, Frankia has a discontinuous membranous layer. Molecular methods for taxonomy such as DNA–DNA hybridization have demonstrated considerable genetic diversity between isolates within this family, but only a limited number of isolates have been analyzed. Although the majority of actinorhizal microsymbionts apparently infect their plant hosts via a root hair-mediated mechanism, some of them can infect by the direct penetration of the root by means of the intercellular spaces of the epidermis and cortex. As a result of infection, a root meristem is induced as the Frankia hyphae penetrate the cells, but the hyphae remain enclosed by a host-produced polysaccharide layer. Frankia grows in the nodule occupying a major part of the host-produced nodular tissue. Root nodule tissue usually composes 1–5% of the total dry weight of the plant. Frankia can form vesicles whose walls evidently adapt to O2 in such a way that they can fix nitrogen at atmospheric O2 levels, but the precise mechanism is obscure. It appears that the infective abilities of each Frankia strain are limited to one or a few plant genera. Currently, four host-specificity groups have been identified: (1) strains infective on Alnus, Comptonia, and Myrica; (2) strains infective on Casuarinaceae; (3) flexible strains infective on species of Elaeagnaceae and the promiscuous species of Myrica and Gymnostoma; and (4) strains infective only on species of the family Elaeagnaceae. Strains in specificity groups 1 and 2 infect the host plant via the root hair, whereas group 4 isolates infect by intercellular penetration. Some strains can use both modes of infection, and these are in the “flexible” specificity group 3. Because these organisms were only recently isolated and cultured, only in the 1990s have taxonomic, physiological, and genetic studies been started, and not all isolates have been successfully cultured. In contrast to Rhizobium, it is possible to routinely obtain nitrogen fixation ex planta in pure culture. It has been possible to inoculate the host plant with pure cultures and obtain effective nodule formation. B. Cyanobacteria Cyanobacteria form symbioses with the most diverse group of hosts of any other nitrogen-fixing system known. Interestingly, whereas Rhizobium can form associations with highly evolved plants, cyanobacteria associate with more primitive plants, such as lichens, liverworts, a pteridophyte (Azolla), gymnosperms (e.g., Cycads), and angiosperms (i.e., Gunnera). 

The cyanobacterium Nostoc forms symbioses with all of the taxa other than the ferns, indicating a potential for genetic manipulation to increase host range. The habitat range is also wide, from tropics to arctic, and includes freshwater, soil, saltwater, and hot springs. As previously discussed, cyanobacteria are an important source of biologically fixed nitrogen. Whereas the free-living cyanobacteria fix up to 80 kg N2/ha, the Azolla–Anabaena symbiosis can produce three times that amount. This value is based on multiple cropping, but rates in rice paddies have been reliably reported from 1.4–10.5 kg N/ha/day as daily averages over the whole growing season. Nitrogen fixation by cyanobacteria was first reported in 1889 shortly after the Rhizobium–legume symbiosis was described. Until the 1970s, research with this system was intermittent, but, partly due to the recently recognized importance of these eubacteria in agriculture and the environment, research has been steadily accelerating. Cyanobacteria often have been classified by botanists using mainly morphological and anatomical criteria rather than by bacteriologists, who use molecular methods. The recognized families and genera of cyanobacteria containing nitrogen-fixing species are summarized in Table 57.3. Because morphological characteristics can vary greatly depending on growth conditions, only limited physiological, biochemical, and molecular genetic studies have been conducted. These indicate that there is considerably more diversity in this taxon than is accounted for by the traditional taxonomy. Cyanobacteria possess the requirements for the higher plant type of photosynthesis—water is the ultimate source of reductant and oxygen is evolved with CO2 fixation via the Calvin cycle. The photosynthetic pigments are located in the outer cell regions. Among filamentous, heterocyst-forming cyanobacteria, the ability for BNF appears universal. In nonheterocystous filamentous forms and unicellular forms, the ability to fix nitrogen is much less common. The heterocysts, which are found spaced along the cell filaments, appear colorless. These cells cannot fix CO2 or evolve O2, but can generate ATP by photophosphorylation and can fix nitrogen aerobically. This is apparently due to a modified thickened cell wall, which interferes with oxygen diffusion. In these organisms then, photosynthesis and nitrogen fixation occur in different cells, which protects nitrogenase from excess oxygen stress. As stated previously, cyanobacteria associate with almost every group of the plant kingdom, forming symbiotic nitrogen-fixing associations. These associations, however, occur with the more primitive plants. Specificity is less tight than it is with either Rhizobium or Frankia. Generally, cyanobacteria can fix nitrogen and grow independently of their plant partner. However, when they live symbiotically, photosynthesis is often diminished in favor of increased nitrogen fixation, sufficient for both partners of the symbiosis. The cyanobacteria usually invade existing normal morphological structures in the host plant, such as leaf cavities, rather than evoking specialized structures such as the nodules caused by Rhizobium or Frankia, although in some cases, as with the roots of cycads, infection is followed by morphological change. One reason for the renewed interest in this group of bacteria is the Azolla–Anabaena symbiosis. Azolla is a free-floating fern commonly found in still waters in temperate and tropical regions and often found in rice paddies. There are seven species. Azolla is a remarkable plant and under suitable conditions can double in weight in about four days. The plant forms a symbiotic relationship with the cyanobacterium Anabaena azollae. 

Azolla provides nutrients and a protective leaf cavity for Anabaena, which in turn provides nitrogen for the fern. Under optimum conditions, this symbiosis results in as much or perhaps more nitrogen fixed than does the legume–Rhizobium symbiosis. If inoculated into a paddy and intercropped with rice, this symbiosis can satisfy the nitrogen requirement for the rice. In Southeast Asia, where there is a wet and dry season, Azolla is recommended for intercropping with paddy rice, but there are problems because Azolla is an extremely efficient scavenger of nutrients and will compete with the rice for phosphate, so careful management is required. Anabaena azollae symbiotically fixes about triple the amount of nitrogen fixed by free-living Anabaena. The high photosynthetic rate of Azolla no doubt supplies energy in this symbiosis. The Anabaena forms more heterocysts when growing in association with Azolla than when growing alone. Similarly, when freeliving heterocystic cyanobacteria are grown in a nitrogen-starved environment, they form extra heterocysts. The hypothesis, then, is that, when sufficient energy and nutrients are present for metabolism, Anabaena can increase its nitrogen-fixation capacity to maximize growth. It has also been reported by several researchers that nitrogen fixation in this symbiotic system is not repressed by the presence of bound nitrogen. Although Azolla has been cultivated in China and Vietnam for centuries, its use represents a new technology for most areas. Even at a relatively low technological level, there is a great potential for the Azolla–Anabeana system as a green manure supplement or animal feed. Little is known about the plant–bacterial interactions, but the fact that Anabaena can be induced to increase nitrogen fixation gives promise that the system can be optimized. The lack of repression of nitrogenase by bound nitrogen of Anabaena nitrogenase works in favor of a crop rotation following legumes. C. The Rhizobium–legume symbioses With worldwide distribution, the Leguminosae is one of the largest plant families. It consists of about 750 genera and an estimated 18,000 species. Members of the Leguminosae have traditionally been placed into three distinct subfamilies based on floral differences—Mimosoideae, Caesalpinoideae, and Papilionoideae. 

Although only about 20% of the total species have been examined for nodulation, these species are representative of all three subfamilies of legumes. Virtually all species within the Mimosoideae and Papilionoideae are nodulated, but about 70% of the species in the subfamily Caesalpinoideae are not nodulated. It is important to note that Bradyrhizobium strains of the cowpea type have been shown to form effective symbioses with the nonlegume Parasponia, which is a member of the Ulmaceae. This is the only verified nitrogen-fixing association between a Rhizobium and a nonlegumous plant. The soil-improving properties of legumes were recognized by ancient agriculturalists. For example, Theophrastus (370–285 BC) in his “Enquiry into Plants” wrote, “Of the other leguminous plants the bean best reinvigorates the ground,” and in another section, “Beans are not a burdensome crop to the ground: they even seem to manure it.” However, it was only in 1888 that Hellriegel and Wilfarth established positively that atmospheric nitrogen was assimilated by root nodules. This was quickly followed by the experiments of Beijerinck, who used pureculture techniques to isolate the root-nodule bacteria and proved that they were the causative agents of dinitrogen assimilation. He initially called these organisms Bacillus radicicola, but they were subsequently named Rhizobium leguminosarum by Frank (1889). Early researchers considered all “rhizobia” to be a single species capable of nodulating all legumes. Extensive cross-testing on various legume hosts led to a taxonomic characterization of these special bacteria based on bacteria–plant cross-inoculation groups, which were defined as “groups of plants within which the root-nodule organisms are mutually interchangeable.” The concept of cross-inoculation groupings as taxonomic criteria held for a very long time and has only gradually fallen into disfavor, although much of this philosophy is retained in the taxonomic scheme. There is a wide range in the efficiency of the Rhizobium–legume symbiosis. Estimates for the amounts of nitrogen fixed are summarized in Tables 60.6 and 60.7. As cells without endospores, bacteria of the family Rhizobiaceae are normally rod-shaped, motile, with one polar or subpolar flagellum or two to six peritrichous flagella, aerobic, and gram-negative. Considerable extracellular slime is usually produced during growth on carbohydrate-containing media with many carbohydrates used. Some Rhizobium and Agrobacterium evidently overlap; their DNA homology is very high. 16S RNA sequence analysis indicates very similar molecular phylogeny. Also, there is an almost complete lack of distinguishing characteristics other than those that are carried on extrachromosomal elements or plasmids. Some bacterial taxonomists are proposing the amalgamation of Agrobacterium, Allorhizobium, Rhizobium, and Sinorhizobium. Traditionally, legume-nodulating bacteria have been recognized as falling into two major phenotypic groups according to growth rate. 

The term “fast growers” commonly refers to strains associated with alfalfa, clover, bean, and pea because, in culture, these organisms grow much faster (less than one-half the doubling time of slow growers, or _3 hr) than the “slow growers” exemplified by soybean and cowpea rhizobia (generation time _6 hr). Although there is phenotypic and genotypic diversity within these major groupings, and some overlap, numerous studies demonstrated the validity of this approach. Mesorhizobium strains, however, have intermediate growth rates between 3 and 6 hr (See Table 60.8.) The relative fastidiousness of the slow growers has been substantiated by recent studies. Although the major biochemical pathways seem to be similar, evidence suggests that the preferred pathway may be different. 16S RNAanalysis of the fast- and slow-growing symbionts have confirmed that these groupings indeed represent very distinct genetic phyla because the similarity coefficient (SAB) of the RNA is only 0.53. Thus, with modern gene analysis, the fast-and slowgrowing Rhizobium fall into widely separate groups. Jordan (1982) transferred the slow-growers to the new genus Bradyrhizobium. Recent findings using numerical taxonomy, carbohydrate metabolism, antibiotic susceptibilities, serology, DNA hybridization, and RNA analysis all demonstrate the validity of the fastand slow-growing groupings. Bradyrhizobium is transferred to the new family Bradyrhizobiaceae in the new edition of Bergey’s Manual of Systematic Bacteriology. 

A summary of some of the differences is found in Table 60.8. Thus, whereas in the first edition of Bergey’s Manual, the slow-growing strains were placed in the new genus Bradyrhizobium, they now fall into their own family along with close relatives such as Afipia and Nitrobacter. The genus Bradyrhizobium now comprises three species, B. japonicum, B. elkanii, and B. liaoningense, all of which nodulate soybean (Table 60.9). Other bradyrhizobia are known to occur (e.g., the peanut bradyrhizobia) but these have not been classified. Researchers suggest that until further taxa within the genus are proposed, these should be described with the appropriate host plant given in parentheses [i.e, the peanut rhizobia–Bradyrhizobium sp. (Arachis)]. The fast-growing legume-nodulating bacteria (sometimes still called “rhizobia,”) were all originally placed within the genus Rhizobium. There were only a few species, R. leguminosarum, R. meliloti, R. loti, R. galegae, and R. fredii. The first three species, R. phaseoli, R. trifolii, and R. leguminosarum were amalgamated into the single type species R. leguminosarum as biovars. The biovar phaseoli had tremendous genetic diversity and a number of new bean-nodulating species have been named, R. tropici, R. etli, R. gallicum, R. giardinii, or R. mongolense. Rhizobium fredii was the first of a series of species consisting of fastgrowing rhizobacteria that effectively nodulate Chinese soybean cultivars, originally thought to be nodulated only by B. japonicum. R. fredii was reassigned to a new genus, Sinorhizobium, in 1988. This controversial new genus also contains S. meliloti and several close relatives, which all share a very close phylogenetic relationship with the type species, S. fredii. Alfalfa plants are nodulated by S. fredii, S. meliloti, and S. medicae. Soybean is nodulated by B. japonicum, B. elkanii, B. liaoningense, S. fredii, and Mesorhizobium tianshanense. The latter new genus and species are closely related to Rhizobium loti, which is now appropriately named Mesorhizobium, along with other newly described close relatives because they are intermediate, both in growth rate and molecular phylogeny, between Bradyrhizobium and Rhizobium. Thus, they now belong to the family Phyllobacteriaceae. 

Allorhizobium is a newly proposed genus for the microsymbiont of an aquatic legume, Neptunia natans. The taxonomic scheme is summarized here in a list of the recognized species of legume-nodulating rhizobacteria. 1. Allorhizobium undicola fixes nitrogen with Acacia spp., Faidherbia spp., and Lotus arabicus. Most A. undicola strains are closely related to Agrobacterium. 2. Azorhizobium caulinodans forms stem nodules on Sesbania rostrata and readily fixes nitrogen ex planta when microaerobic conditions are provided. It belongs to the family Hypomicrobiaceae. 3. Bradyrhizobium japonicum forms root nodules on species of Glycine (soybean) and on Macroptilium atropurpureum (siratro). Some strains of B. japonicum express hydrogenase activity with the soybean host and are hence more efficient in symbiotic nitrogen fixation. 4. Bradyrhizobium elkanii normally forms root nodules on species of Glycine (soybean), the “nonnodulating” rj1rj1 mutant soybean that fails to nodulate with B. japonicum, black-eyed peas (Vigna), mung bean, and Macroptilium atropurpureum (siratro). Unlike B. japonicum, B. elkanii often produces rhizobitoxine-induced chlorosis on sensitive soybean cultivars. Strains of B. elkanii often are hydrogenase positive on Vigna but not on Glycine, suggesting that they possess more symbiotic affinity or compatibility with the former than the latter. 5. Bradyrhizobium liaoningense is an extra-slowgrowing, soybean-nodulating Bradyrhizobium isolated from alkaline Chinese soils. 6. Mesorhizbium amorphae was isolated from nodules of Amorpha fruticosa growing in North China. 7. Mesorhizobium ciceri was isolated from chickpeas (Cicer) grown in uninoculated fields over a wide geographic range, including Spain, the United States, India, Russia, Turkey, Morocco, and Syria. 8. Mesorhizobium huakuii was isolated from Astragalus sinicus, a green manure crop grown in rice fields in southern parts of China, Japan, and Korea. 9. Mesorhizobium loti nodulates L. corniculatus, L. tenuis, L. japonicum, L. krylovii, L. filicalius, and L. schoelleri. 10. Mesorhizobium mediterraneum is exclusively a Cicer-nodulating bacterium. 11. Mesorhizobium tianshanense isolates were obtained from Glycyrrhiza pallidiflora, G. uralensis, Glycine max, Sophora alopecuroides, Swainsonia salsula, Caragara polourensis, and Halimodendron holodendron growing in Xinjiang Region of China. Most of the host plants are wild and indigenous to that region, except G. max, which is of course a cultivated crop that originated in northeastern Asia. 12. Mesorhizobium plurifarium nodulates Acacia senegal, A. tortilis, A. nilotica, A. seyal, Leucaena leucocephala, and Neptunia oleracea, but most strains do not nodulate Sesbania rostrata, S. pubescens, S. grandiflora, Ononis repens, or Lotus corniculatus. 13. Rhizobium etli nodulates and fixes nitrogen in association with P. vulgaris exclusively; it includes nonsymbiotic strains. 14. Rhizobium galegae nodulates Galega orientalis and Galega officinalis and is specific to this plant genus. 15. Rhizobium gallicum nodulates and fixes nitrogen in association with Phaseolus spp., Leucaena leucocephala, Macroptilium atropurpureum, and Onobrychis viciifolia. 16. Rhizobium giardinii nodulates Phaseolus spp., Leucaena leucocephala, and Macroptilium atropurpureum. 17. Rhizobium hainanense is found in nodules of Desmodium sinuatum or Stylosanthes guianensis, Centrosema pubescens, Desmodium triquetrum, D. gyroides, D. sinatum, D. heterophyllum, Tephrosia candida, Acacia sinicus, Arachis hypogaea, Zornia diphylla, Uraria crinita, and Macroptilium lathyroides. 18. Rhizobium huautlense nodulates Sesbania herbacea, S. rostrata, and Leucaena leucocephala. 19. Rhizobium leguminosarum nodulates with some, but not necessarily all Pisum spp., Lathyrus spp., Vicia spp., Lens spp., temperate species of Phaseolus (P. vulgaris, P. angustifolius, and P. multiflorus), and Trifolium spp. 20. Rhizobium mongolense was recently isolated from Medicago ruthenica, but it also nodulates Phaseolus vulgaris. It is a very close relative of Rhizobium gallicum. 21. 

Rhizobium tropici forms nodules on Phaseolus vulgaris and Leucaena spp. The type strain, CFN299, nodulates Amorpha fruticosa. 22. Sinorhizobium fredii effectively nodulates Glycine max cv. “Peking,” Glycine soja, Vigna unguiculata, and Cajanus cajan. Also nodulates alfalfa. In 1985, Dowdle and Bohlool reported new strains that are symbiotically competent with North American cultivars of soybean. Their molecular phylogeny, however, has not been determined yet. 23. Sinorhizobium medicae is alfalfa-nodulating, with a close phylogenetic relationship to S. meliloti. 24. Sinorhizobium meliloti forms nitrogen-fixing nodules on Melilotus, Medicago, and Trigonella. 25. Sinorhizobium saheli is found in nodules of Sesbania spp. growing in the Sahel and can nodulate Acacia seyal, Leucaena leucocephala, and Neptunia oleracea. 26. Sinorhizobium terangae is also found in nodules of Acacia spp., Sesbania spp., Leucaena leucocephala, and Neptunia oleracea. 27. Sinorhizobium xinjiangense nodulates soybean and is a close relative of S. fredii. The taxonomy of the nitrogen-fixing bacteria is in a dynamic state of change. As molecular information accumulates, the cataloging of phenotypic data has not kept pace and further revision, and “reversions” will be necessary. New approaches to classification are needed because the scheme, unfortunately, does not function to allow the identification of isolates without DNA-sequence analysis. 1. Nitrogen fixation The nitrogenase complex is a highly conserved enzyme system and, as stated earlier, is basically common to all of the dinitrogen-fixing prokaryotes. The evidence is conclusive that there are differences in location of nitrogenase genes of Rhizobium and Bradyrhizobium. In all of the fast-growing Rhizobium species, which have a chromosome size of about 3500 kb, the structural nifH, D, and K genes are localized on extremely large plasmids or megaplasmids. In Bradyrhizobium, nif genes have been mapped on the 8700-kb chromosome. In Mesorhizobium, on the other hand, the nif genes are located on the chromosome in some species and on megaplasmids in other species. Plasmids and megaplasmids, present in a wide variety of the nitrogen-fixing bacteria, control many phenotypic and genetic characteristics of the bacterial cells. Those in Rhizobium, Sinorhizobium, and Azorhizobium species carry genes controlling symbiotic functions, which are clustered on a single large plasmid termed symbiotic plasmid, “pSym.” Two or more plasmids that carry genes controlling symbiotic functions, nod, fix and the nitrogenase structural (nifHDK) genes, have been found in certain strains of the nitrogen-fixing species. In addition to symbiotic plasmid(s), rhizobia strains may carry 1–10 plasmids that range in size from 30 to more than 1000MDa. 

These plasmids are highly stable and have beneficial roles in the soil environment, and plasmid profile analysis is sometimes used to discriminate among Rhizobium strains. The analysis and comparison of nif DNA in the fastand slow-growing “rhizobia” have established the affinity coefficient (SAB) for the nucleotide sequence (nifH, nifD, and nifK) from the two groups. The same analysis was done comparing amino acid sequences of the nitrogenase Fe and FeMo protein polypeptides. A considerable sequence conservation reflects the structural requirements of the nitrogenase proteins for catalytic functions. The SAB nif values (based on nifH sequences) between fast- and slow-growing organisms indicated that these are almost as distant from each other as they are from other gram-negative organisms. The results suggest that nif genes evolved in a manner similar to the bacteria that carry them rather than by a more recent lateral distribution of nif genes among microorganisms. Again, the phylogenetic difference between fast and slow groups is apparent. Although this general system is common to all of nitrogen-fixing prokaryotes, several concomitant alternative systems have been described; but these evidently lack biological significance and they have not been shown to be present or active in the “rhizobia.” 2. Nodule formation Nodule initiation and subsequent maturation is an interactive process involving the eukaryotic host legume and the prokaryotic Rhizobium. The process is complex, resulting in biochemical and morphological changes in both symbionts and leading to the capacity to reduce atmospheric nitrogen. Initially, the proper Rhizobium species proliferates in the root zone of a temperate leguminous plant and becomes attracted and attached to the root hair. A chemotactic response attracts the bacteria to the root surface. At the surface, the bacteria secrete Nod factors, which are certain chitolipooligosaccarides (CLOS) that alter the growth of epidermal root hairs so that they are deformed. CLOS molecules chemically induce nodule organogenesis in extremely low concentrations (~10_10 M). In some tropical legumes such as peanuts (Arachis), root hairs are not the primary invasion sites. However, the alternative invasion process, “crack entry,” has been well-documented because infection occurs at the site of lateralroot emergence. 

The root hair infection process consists of several events leading to nodule formation: (1) recognition by the “rhizobia” of the legume, (2) attachment to the root hair, (3) curling of the root hair, (4) roothair infection by the bacteria, (5) formation of an infection thread, (6) nodule initiation, and (7) transformation of the vegetative cells in the nodules to enlarged pleomorphic forms, called bacteroids, which fix nitrogen. Based on morphology, there are two kinds of nodules, determinative and indeterminative. In general, indeterminative nodules are formed by fast-growing “rhizobia” and are characterized by a defined meristem during nodule growth. Determinative nodules arise from cortex tissue. Legumes nodulated by Bradyrhizobium form determinative nodules close to the endodermis, which is near the xylem poles in the root. The formation of nodules on legumes is the result of a coordinated development involving many plant and bacterial genes. Studies of the nodulation (nod) genes of rhizobia have depended on the development of molecular genetic tools. Many of the genes involved in the nodulation process have been located and identified. Legume roots grown axenically do not appear as morphologically distinctive from other plant roots, so that the abilities of these plants to respond to microbial signals and then alter their metabolism to form nodules are not explained by morphology alone. It has been conclusively demonstrated that genetic information from both symbionts controls nodulation and the host range of nodulation by a Rhizobium species. Metabolically, there are three types of nodules (often termed effective, inefficient, and ineffective). Effective nodules contain a high density of bacteria actively fixing dinitrogen. Inefficient nodules may contain a similar density of the bacteria, but only a relatively low level of fixed dinitrogen results from the symbiosis. Ineffective symbiosis occurs with bacteria not able to nodulate or fix nitrogen normally. Because the regulatory roles of the plant and bacterial genes in nitrogen fixation have not been generally elucidated, the reasons for differential nitrogen-fixing ability of nodules continues to be unclear. 

There is, however, an already observed compatibility in legume host– Rhizobium interaction that can be technologically exploited to enhance dinitrogen fixation. Such an interaction makes it possible to optimize dinitrogen fixation under field conditions in a cultivar through inoculation with an effective Rhizobium strain. Nodulation genes are defined by their effect on the bacteria’s ability to generate the nodulation process on the proper legume host. Because most individual species of legume-nodulating rhizobacteria can each nodulate a limited number of host legumes and plant genes also limit the symbiosis, it follows that a recognition exists between bacteria and host. Thus, there are two types of nod genes, a common nod region, which consists of a structurally and functionally conserved cluster of genes; and host-specific nodulation genes, which cannot complement nodulation defects in other genera or species. Nod genes have been studied in varying degrees in different species (usually in the fast-growing rhizobia). In these organisms, four genes have been identified in two transcription units (nodD and nodABC). Two additional genes, apparently on the same transcriptional unit, nodl and nodJ, have been identified. Genetic maps of the common nod cluster, drawing together the information from many sources, have been published. The nodABC appears to be functionally interchangeable among all rhizobia, and mutations in these genes cause complete nodulation failure. These genes are involved in cell division and roothair deformation. The nodlJ genes cause a delay in the appearance of nodules. The second group of nod genes are termed “hostspecific.” These genes are not conserved because alleles from various rhizobia cannot substitute for each other on different hosts. Bacteria carrying mutations in these host-specific genes cause abnormal root-hair reactions. Many genetic nod loci have been identified in a variety of the symbiotic nitrogen-fixing species. The list includes at least 15 nod genes. In many cases, these have been cloned and sequenced and the gene product associated with a step in nodulation. Although the amino acid sequences of many of the nodulation-gene products have been described, the biochemical functions of these genes have not been fully determined. Possible exceptions are the nodD genes, which are positive gene regulators. The centenary of this first demonstration of biological nitrogen fixation occurred during 1986. During that period, many papers were published expanding the knowledge base, both basic and applied. In 1991, the U.S. Department of Agriculture patented an improved soybean inoculant Bradyrhizobium japonicum strain that results in a significant increase in growth and soybean yield. Since then, this improved strain has been commercially produced. 

Currently, the subject of BNF is of great practical importance because of the use of fossil fuels in the manufacture of nitrogenous fertilizers. The increased scarcity and higher costs of fossil fuel have made it important to optimize biological nitrogen fixation as an alternative to chemical nitrogen. In addition, the increasing usage of nitrogen fertilizer has resulted in unacceptable levels of water pollution, which occurs only to much lesser extent when the biologically fixed forms of nitrogen are used. With the additional research capabilities resulting from the developing field of biotechnology, it is evident that interest in this field will continue and that we may reach a level of accumulated knowledge that will allow full use of BNF as an alternative to the Haber–Bosch chemical industrial process. 3. Rhizobiophages Rhizobiophages occur commonly in the rhizosphere of legumes and are often associated with susceptible Bradyrhizobium, Rhizobium, or Sinorhizobium strains. They reduce rhizobial populations in soils and negatively affect the nitrogen-fixing abilities of these bacteria with the host legume plant. Rhizobiophages can be used to distinguish between rhizobial strains through “phage typing.” Furthermore, rhizobiophages are potential biocontrol agents useful for reducing the number of susceptible rhizobial cells in the soils, thus decreasing nodule occupancy by the undesirable indigenous bacterial strain, and thereby increasing the nodule occupancy by superior strains used as inoculant. The use of specific bacterial viruses or bacteriophages as biocontrol agents requires the identification of symbiotically competent, rhizobiophage- resistant Rhizobium or Bradyrhizobium strains that have a demonstrated ability to promote the growth and yield of their specific legume hosts.