The symbiosis genes of pea

The ability to form a symbiosis with nitrogen fixing bacteria occurs in 10 families of dicotyledonous plants. The best known of these of course is the Leguminosae; approximately 85% of legumes can form nodules (1). In agronomy, the nitrogen fixing ability of soybean, lucerne and clover are of major research interest because of the great areas planted with these crops. However, the species most often studied for basic knowledge is the pea, which has served as an experimental plant from the earliest investigations. The early work is reviewed in detail by Fred et al. (10).

Boussingault in 1838 provided the first quantitative data that clover and peas increased their nitrogen content from the air, but he suspected it was an absorption of ammonia. Atwater in 1884 had data indicating that the N increment of peas was from fixation of nitrogen gas ("against the best evidence and opinion at the present time"). However, he could not fathom the means. It was in 1886, that Hellriegel clearly demonstrated with pea that N fixation required a symbiosis with microbes in the root nodules. Use of organisms isolated from pea nodules was central when, in 1889, Prazmowski showed that the symbiotic associations were species-specific.

The bacteria that form a symbiosis with legumes are called rhizobia. At the present time, the strains that nodulate Pisum, Lens and Vicia are classified as Rhizobium leguminosarum biovar viciae. In this species, about 20% of the DNA is in large plasmids. On one of these, the sym plasmid, are found most of the known genes involved in symbiosis.

Nodule development

As microscopic methods have improved, nodule development on pea has been repeatedly described (2, 24, 28). Briefly, competent rhizobia induce the root hairs to curl ("shepherd's crook") and a "thread" of cellulosic material forms in the hair. The rhizobia enter this infection thread and multiply as the thread elongates. After passing from the base of the epidermal cell into the outer cortex, the thread branches. As the branches penetrate towards the inner cortex, anticlinal cell divisions occur in advance of it, and a nodule primordium is established adjacent to the pericycle near the xylem pole. The thread, branching further, penetrates some of the dividing cells, releasing bacteria from the tips of the threads, where they are surrounded by a peribacteroid membrane of plant origin. The infected plant cells enlarge, the bacteria multiply, nitrogen fixing enzymes are formed in the bacteria and leghemoglobin appears in the cytoplasm of the infected cells.

In rhizobia, the nitrogen fixing enzyme nitrogenase, containing Fe and Mo, is coded for by nif genes, which are homologous to the nif genes found in free living nitrogen fixing bacteria. In addition, mutational analysis of R. leguminosarum has identified nod and fix genes, which are involved in nodule formation or function, respectively. Some of these genes are host-specific, while others (the "common" genes) are conserved and occur in all species and biovars of rhizobia. The symbiosis also may be delayed by mutations in exo genes, which are involved in production of the exopolysaccharide of the bacteria (2).

The rhizobia bind to the root. There is evidence that the specificity of binding is due to the plant's lectin. A gene for a pea lectin was introduced into Trifolium repens with Agrobacterium rhizogenes as a vector. The resulting transformed clover roots could then be nodulated by R. leguminosarum bv. viciae (4). Binding may be to the polysaccharide capsule of the rhizobium, but solid evidence for this is lacking. An alternate theory is that the bridge is by a calcium requiring binding protein produced by the rhizobia (29).

The nod genes are induced by compounds secreted by the roots. In pea, these may include naringenin (34), apigenin-7-0-glucoside, eriodictyol, 7-hydroxyflavone, 3, 5, 7, 3' tetra hydroxy-4' methoxy flavone and 7, 3' hydroxy-4'methoxyflavone. The nod genes D, ABC, EF and L are induced to form a tetrasaccharide of D-glucosamine on which a hydroxy and 3 amino groups are acetylated and another amino group is acylated by a C 18:4 unsaturated fatty acid. This compound causes root hair curling and cortical cell divisions in roots (30).

Genetic analysis

Two approaches have been made in studying the host genes involved in symbiosis. One is to identify proteins (nodulins) whose concentration is enhanced in developing or functional nodules, compared with uninoculated roots. The other method is to characterise genes identified in naturally occurring or induced mutants which have an altered symbiosis.

The organ-specific proteins which appear first are the "early nodulins" (Table 1), which were identified by differential screening of a pea nodule cDNA library. The protein ENOD 12 (28) is induced in root hairs by cell-free culture filtrate of appropriate rhizobia. It is found also in cells containing the infection thread, in cells in advance of the infection thread, and in the developing nodule. The protein is rich in hydroxyproline, and may contribute to the cell wall reformation required for the infection thread.

ENOD5 is also hydroxyproline rich, but is similar to the arabinogalactans (28). Its distribution suggests that it is part of the plasma membrane of the infection thread.

ENOD2 which is also rich in hydroxyproline is found in the nodule parenchyma. This layer of cells is thought to control the flux of oxygen into the nodule (12, 34).

Table 1. Nodulins and nodule enhanced proteins in P. sativum

Leghemoglobin, which in legumes is coded for by a multi-gene family, appears just before nitrogen fixation is detectable. A feature of all legume nodules is that they contain enhanced amounts of a form of glutamine synthetase (GS_n), different from the isoforms found in cytosol (GS_c) or plastids (GS_p). At present, the pea is unique in having two nearly identical forms of GS_n (31). These "twins" differ by only a few amino acids (32).

Some naturally occurring symbiosis variants are known for pea, but the genes involved have seldom been backcrossed into standard lines for experimental studies. Fortunately, it has proven easy to obtain induced mutants with altered symbiotic properties. Our group (17), Jacobsen (26), and Duc (5) have made mutants of the cultivars Sparkle, Rondo and Frisson, respectively. We are now testing these three collections for allelism. These mutants include peas with excessive nodule numbers (nod⁺⁺), no or few nodules (nod^-), or nodules which do not adequately fix nitrogen (nod⁺, fix^-).

Table 2. Genes of P. sativum involved in nodule formation or function.

Table 2 lists pea genes involved in the symbiosis which have been defined by inheritance tests. Allelism tests for mutants without gene designations are still in progress. About 30 genes have already been identified; this number approximates the number of nod and fix genes known in the microsymbiont.

In most cases, the phenotype is independent of the strain of R. leguminosarum tested. However, some genes condition an altered specificity, sym-2, originally described in the primitive line 'Afghanistan' (15, 21), results in nodulation by only a few strains of R. leguminosarum. Most of these strains carry a gene nod X on the sym plasmid, a gene not found in non-infective strains (3). To date, this is the only example of a "gene for gene" effect in nodule formation. Loci Sym-4 and sym-6 in primitive lines and induced mutations at sym-18 also condition restricted strain specificity, but the rhizobial genes involved are not yet studied.

One of the factors in the host determining Rhizobium strain specificity is the lectins produced in roots (4). However, Lu et al. (unpublished) have demonstrated that two of the strain-specificity mutants in pea (sym-2 and sym-18) do not reflect mutations in the lectin genes but map to different chromosomal locations. The strain-specificity mutations may still involve lectins by way of genes producing post-translational modifications. Alternatively, strain-specificity may be influenced by other factors as yet not identified.

The pea controls nodule number, in part, by aborting infection. The biochemical nature of this auto-regulation is unknown, but some mutants now available may provide useful experimental material, for these mutants hypernodulate. Our laboratory has been unable to obtain high nodule numbers on the cv. 'Parvus', so the existence of nod-1 and nod-2 is unconfirmed. The nod-3 mutant of Rondo (27) and the 190F line (sym-28) of Frisson (5) both have higher than normal nodule numbers. Allelism tests have not been reported, but the phenotypes are different. Unlike nod-3, 190F is pleiotropic; the shoot has short internodes. Grafting experiments indicate that the hypernodulation of nod-28 is controlled by the genotype of the shoot (5) while that of nod-3 is controlled by the genotype of the root (27).

We have only begun to characterise the many nod^- mutants. The large number indicates that it should be possible to "dissect" nodule formation by blocking it at many stages. R25 (sym-8) does not react at all to the presence of rhizobia; lacking even root hair curling. R72 (sym-9) shows root hair swelling in the presence of R. leguminosarum, but no curling or infection. These two mutants are obviously blocked at the earliest steps of infection (22).

A mutational "hot spot" must occur at sym-5, for seven independently isolated mutants were obtained at this locus (17). Except for forming few or no nodules, these mutants appear normal. There are the normal number of infections, but nodule primordia are seldom formed in advance of the infection threads (13). The sym-5 mutants may involve an increased sensitivity to the hormone ethylene, for primordia initiation and nodule formation are increased by lowered root temperature (which lowers ethylene formation) (9) or by exogenous inhibitors of ethylene formation or action (8).

The sym-5 mutants have an altered 66 kD peptide, constitutive to all plant parts (7). It is not yet known if this peptide is the product of sym-5.

Duc (5) has identified six non-allelic genes required for nodule function. One of these (g) is allelic to the previously described sym-13. Line E135F (sym-13) is characterised by small nodules which senesce early (19). The nodules do not have nitrogenase or leghemoglobin.

Table 3. Location of genes in P. sativum linkage map.

An important observation was that some of the nod^- mutants do not form a symbiosis with vesicular arbuscular mycorrhiza (myc^-) (6). This suggests that the two different symbionts must share some process.

Gene mapping

Whereas the nif, nod and fix genes of rhizobia are closely clustered on the sym plasmid, the plant genes are widely distributed (Table 3). We have found that each chromosome carries at least one gene involved in nodulation (33). The genes seem randomly distributed with one notable exception. Within a relatively small region (10 cM) of chromosome 1 are found the genes for leghemoglobin and sym-2, sym-5, sym-18 and sym-19 (33). sym 2, described in the primitive line 'Afghanistan' (15), and sym-18 condition narrow strain specificity. Mutants at sym-19 are nod^- and myc^-.

The existence of such a cluster of functionally related genes is unparalleled in higher plant genomes. Gene families, such as the genes encoding the small subunit of ribulose bisphosphate carboxylase, the 45S ribosomal array, or leghemoglobin, are often found in very tightly linked clusters. However, genes related by function, such as those involved in carbohydrate metabolism, chloroplast function, or anthocyanin synthesis are usually widely scattered in the genome. Hence, the "sym" cluster on chromosome 1 is particularly intriguing. Initial mapping data place sym-5 at one end of this cluster, sym-2, sym-18 and sym-19 at the other end, and the leghemoglobin genes in the middle. Although sym-2, sym-18 and sym-19 have been shown to be non-allelic, recombination between pairs of these loci has not been demonstrated.

The positions of other sym loci are not particularly remarkable. The lectin genes map in the vicinity of sym-11, but there is no reason to suspect that sym-11 is a mutation in the expressed lectin gene because there is no change in strain specificity. The recently described dominant low nodule number mutant, Sym-22, maps very close to Fum (chromosome 2), as does sym-14. It is possible that these represent alleles at the same locus, despite one being a dominant mutation and the other recessive. A non-nodulating mutant that also exhibited short lateral roots, sym-16, maps near a known mutant, coh, displaying short root laterals; however the two mutants complement each other, the hybrid displaying normal roots and a normal nodulation response.

Future Research

The genetic knowledge of nodulation in P. sativum is far in advance of that of any other species. This information should be extended and exploited in at least two ways. Firstly, it is desirable to bridge the gap between the sym genes and the nodulins; i.e., to identify the products of the sym genes on one hand and to subject the nodulins to classical genetic analysis.

Secondly, it may be time to test the utility of sym mutants. The genes controlling high nodule number and restricted strain specificity should be applied by pea breeders and agronomists to design cultivars which will nodulate profusely, but only with superior strains of rhizobia which are designed to be compatible with the cultivar.

1. Allen, O.N. and Allen, E.K. 1981. The Leguminosae. Univ. Wisconsin Press.

2. Brewin, N.J. 1991. Ann. Rev. Cell Biol. 7:191-226.

3. Davis, E.O., Evans, L.J. and Johnston, A.W.B. 1988. Mol. Gen. Genet. 212: 531-535.

4. Diaz, C.L., Melchers, L.S., Hooykaas, P.J.J., Lugtenberg, B.J.J, and Kijne, J.W. 1989. Nature 338: 579-581.

5. Duc, G. and Messager, A. 1989. Plant Sci. 60:207-213.

6. Duc, G.A., Trouvelot, A., Gianinazzi-Pearson, V. and Gianinazzi, S. 1989. Plant Sci. 60:215-222.

7. Fearn, J.C. and LaRue, T.A. 1990. Plant Mol. Biol. 14:207-216.

8. Fearn, J.C. and LaRue, T.A. 1991. Plant Physiol. 96:239-244.

9. Fearn, J.C. and LaRue, T.A. 1991. Plant, Cell and Environ. 14:222-227.

10. Fred, E.B., Baldwin, I.L. and McCoy, E. 1932. Root Nodule Bacteria and Leguminous Plants. Madison.

11. Gelin, O. and Blixt, S. 1964. Agri Hort. Genet. 22:149-159.

12. Govers, F., Gloudemans, T., Moerman, M., van Kammen, A. and Bisseling, T. 1985. EMBO J. 4:861-867.

13. Guinel, F.C. and LaRue, T.A. 1991. Plant Physiol. 97:1206-1211.

14. Guinel, F.C. and LaRue, T.A. 1992. Plant Physiol. 99:515-518.

15. Holl, F.B. 1975. Euphytica 27:767-770.

16. Kneen, B.E. and LaRue, T.A. 1984. Heredity 52:383-389.

17. Kneen, B.E. and LaRue, T.A. 1988. Plant Sci. 58:177-182.

18. Kneen, B.E., LaRue, T.A., Welch, R.M. and Weeden, N.F. 1990. Plant Physiol. 93:717-722.

19. Kneen, B.E., LaRue, T.A., Hirsch, A.M., Smith, C.A. and Weeden, N.F. 1990. Plant Physiol. 94:899-905.

20. Lee, K.H. and LaRue, T.A. 1992. Plant Physiol. 100:1326-1333.

21. Lie, T.A. 1984. Plant Soil 82:415-425.

22. Markwei, C.M. and LaRue, T.A. 1992. Can. J. Microbiol. 38:548-554.

23. Nap, J.-P. and Bisseling, T. 1990. Science 250:948-954.

24. Newcomb, W. 1976. Can. J. Bot. 54:2163-2186.

25. Ohlendorf, H. 1983. Z. Pflanzenzuchtung 91:13-24.

26. Postma, J.G., Jacobsen, E. and Feenstra, W.J. 1988. J. Plant Physiol. 132:424-430.

27. Postma, J.G., Jager, D., Jacobsen, E. and Feenstra, W.J. 1990. Plant Sci. 68:151-161.

28. Scheres, B., van Engelen, F., van der Knaap, E., van der Wiel, C., van Kammen, W. and Bisseling, T. 1990. Plant Cell 2:687-700.

29. Smit, G., Tubbing, D.M.J., Kijne, J.W. and Lugtenberg, B.J.J. 1991. Arch. Microbiol. 155:278-283.

30. Spaink, H.P., Sheeley, T.M., van Brussel, A.A.N., Glushka, J., York, W.S., Tak, T., Geiger, O., Kennedy, E.P., Reinhold, V.N. and Lugtenberg, B.J.J. 1991. Nature 354:125-129.

31. Tingey, S.V., Walker, E.L. and Coruzzi, G.M. 1987. EMBO J. 6:1-9.

32. Walker, E.L. and Coruzzi, G.M. 1989. Plant Physiol. 91:702-708.

33. Weeden, N.F., Kneen, B.E. and LaRue, T.A. 1990. In Nitrogen Fixation: Achievements and Objectives. Chapman and Hall, London, pp. 323-330.

34. van der Weil, C., Sheres, B., Franzen, H., van Lierop, M.-J., van Lammeren, A., van Kammen, A. and Bisseling, T. 1990. EMBO J. 9:1-7.