Which steps are essential for the formation of functional legume nodules?

Nodulation is reviewed in terms of the phenotypes proposed by Vincent (1980). Individual legumes may be infectible by one or more of the three bacterial genera (collectively known as rhizobia) Rhizobium, Bradyrhizobium, or Azorhizobium. The type of infection process by which rhizobia gain entry is largely governed by the host genotype. In addition to the widely studied root‐hair pathway, infections may be associated with lateral root emergence or occur between root epidermal cells. The exact chemical and physical nature of the root hair/epidermal cell wall is likely to be a critical factor in determining whether infections can proceed. In addition to differing with species, wall composition may be influenced by soil chemical (e.g. Ca²⁺) and biotic factors (e.g. bacteria). Rhizobial features essential for infection include particular surface polysaccharides and the induction of nodulation genes by plant root exudates. Neither of these is likely to be a major barrier to the extension of nodulation to new hosts. Dissemination of rhizobia within developing nodules may be intercellular, via infection threads or by division of a small number of infected cells. All functional symbioses eventually have ‘intracellular’ bacteria, in the sense that rhizobia are geographically located within the boundary of the host cell walls. However, they remain extracellular in the sense that they are always confined by a membrane which is largely of host cell origin. In some genera they are also surrounded by infection thread walls, probably modified forms of ‘invasive’ infection thread walls, which allow differentiation of rhizobia into the nitrogen‐fixing form. Thus, natural, functional, symbioses may (a) never involve a stage in which bacteria are confined within tubular infection threads or (b) never release bacteria from infection threads. These features are determined by host genotype. The one feature of legume nodules so far found never to vary is the stem‐like character of a peripheral vascular system. This contrasts with the central vascular system of actinorhizas and the rhizobial‐induced nodules on the Ulmaceous genus Parasponia. Although of great intrinsic interest, this character is unlikely to present an insurmountable barrier to the extension of nodulation to new species. Other features, such as the ability to produce haemoglobin are now known to the in the genetic makeup of many higher plants. The discovery of the wide range of nodule structures occurring in nature, together with work on mutant rhizobia which may bypass critical stages in the nodulation process, suggest various ways in which the extension of nodulation to non‐nodulated legumes and to other (initially at least, dicotyledonous) plants may be engineered. CONTENTS Summary 129 I. Introduction 130 II. The symbionts 130 III. Stages in nodulation 132 IV. Stems and nodules 143 V. Prospects for finding/making new symbioses 144 VI. Conclusions 145 Acknowledgements 147 References 147