The occurrence of coralloid roots on the South African species of the Cycadales and their ability to fiX nitrogen symbiotically

N. Grobbelaar, W. Hattingh and J. Marshall Department of Botany, University of Pretoria, Pretoria

Thirty-three species of Encephalartos, which include all the Introduction known species indigenous to the Republic of South Africa, Since Reinke (1872) first described coralloid roots on Cycas together with Stangeria eriopus, were found to form revoluta, the presence of these unusual roots has to date been coralloid roots_ In all cases, specimens were found in which the coralloid roots were infected with Cyanobacteria (blue­ reported on only about 49 of the 134 known cycad species green algae) and these roots in all cases reduced acetylene - this includes 13 of the 45 described Encephalartos species to ethylene, and fixed nitrogen. as well as Stangeria eriopus (Allen & Allen 1965; Wittman S. Afr. J. Bot. 1986, 52: 467-471 eta/_ 1965; Halliday & Pate 1976; Lamont & Ryan 1977; Grilli Caiola 1980). In most cases a single endophyte was found Daar is vasgestel dat al 33 die Encephalartos-spesies wat which was identified as an Anabaena or Nostoc species ondersoek is, sowel as Stangeria eriopus koraalvormige (Cyanobacteria). wortels vorm. Die ondersoekte spesies het al die beskrewe spesies wat inheems aan die Republiek van Suid-Afrika is Because heterocystous Cyanobacteria generally are able to ingesluit In aile gevalle is eksemplare gevind waarvan die fix nitrogen, it is usually assumed that the infected cycad koraalvormige wortels met Cyanobacteria (blougroen alge) coralloid roots also have this ability_ In only a few cases ge·lnfekteer was. Sulke wortels het in aile gevalle oor die (Bergersen eta/. 1965; Bond 1967; Grobbelaar eta/_ 1971; vermoe beskik om asetileen tot etileen te reduseer en om Halliday & Pate 1976; Grove et a/_ 1980) has this been \ - stikstof te bind. experimentally substantiated. S.-Afr. Tydskr. Plantk. 1986, 52: 467-471 The indigenous flora of the Republic of South Africa Keywords: Acetylene reduction, coralloid roots, cycads, includes 28 Encephalartos species and Stangeria eriopus Encephalartos sp., nitrogen fixation (Giddy 1974). Twelve of these Encepha/artos species as well as S. eriopus have previously been shown to form coralloid roots but none of the plants has apparently been tested for nitrogen fixation. For this reason, it was decided to investigate all the indigenous Cycadales as well as a few other available cycad species in this regard. This paper reports on the results of this investigation.

Materials and Methods Although plants, especially seedlings, were examined in their natural habitat, most of the work was done on plants growing in botanic gardens and the gardens of private collectors. Coralloid root samples were sectioned transversely and examined microscopically for Cyanobacteria. Only roots that were so heavily infected with Cyanobacteria that their presence could be seen as a dark band with the naked eye were used in nitrogen fixation studies. The roots (ca. 3 g) were brushed free of soil in running tap water and blotted dry before being incubated with either acetylene or 15N in different experiments. In the case of the acetylene reduction experiment, the roots of several species were also merely shaken free of most of the adhering soil and incubated without scrubbing in order to ascertain whether the rhizosphere micro-organisms con­ tributed significantly to the rate of acetylene reduction. N. Grobbelaar*, W. Hattingh and J _ Marshall Department of Botany, University of Pretoria, Pretoria, 0002 Acetylene reduction Republic of South Africa The basic method used was that of Hardy eta/_ (1968). Each 3 *To whom correspondence should be addressed root sample was incubated in a 40 cm conical flask which was fitted with a rubber serum bottle stopper. Four cm 3 Accepted 14 May 1986 acetylene was mixed by syringe with the air in the flask 468 S.-Afr. Tydskr. Plantk., 1986, 52(5)

whereafter a 2,5 cm3 gas sample was withdrawn from the flask Table 1 The rate of acetylene reduction by cyanobac­ 3 and stored in a serum bottle (2,5 cm ) by water displacement teria-infected coralloid roots of cycads (Grobbelaar & Rosch 1981). The remainder of the gas in the 3 incubation flask was brought to atmospheric pressure where­ mm C2H4 formed Cycad species (g dry mass) - 1h - 1 after the flask was kept at ambient temperature (about 25°C) in dim light. Exactly 1 h after the first gas sample was •cycas thouarsii Gaud. 35 withdrawn, a second 2,5 cm3 gas sample was taken from the Encephalartos altensteinii Lehm. 20 E. arenarius R.A. Dyer 30 incubation flask and stored in a serum bottle. E. cajjer (Thunb.) Lehm. 340 As a control, a cleaned root sample of each species was •E. concinnus R.A. Dyer & Verdoorn 50 similarly incubated but in air to which no acetylene had been E. cupidus R.A. Dyer 10 added. Coralloid roots of E. jerox and E. transvenosus, which E. cycadifolius (Jacq.) Lehm. 10 appeared to be devoid of a cyanobacterial endophyte, were E. eugene-maraisii Verdoorn 30 also incubated with acetylene. E. jerox Bertol. f. 40 E. jriderici-guilielmi Lehm. 30 The gas volume of the incubation flasks containing the root E. ghellinckii Lem. 60 samples was determined by the method of Grobbelaar & E. heenanii R.A. Dyer 10 Meyer (1986) whereafter the dry mass (drying at 80°C) of the •E. hildebrandtii A. Braun & Bouche 80 roots was determined. E. horridus (Jacq.) Lehm. 30 The concentration of the ethylene in the gas samples was E. humilis Verdoorn 10 determined on 1 cm3 sub-samples by means of a gas chro­ E. inopinus R.A. Oyer 250 E. laevifolius Stapf & Burtt Davy 10 matograph containing a 2 m column (internal diameter 3 mm) E. /anatus Stapf & Burtt Davy 30 of Poropak Nat 150°C fitted with a flame ionization detector. E. latijrons Lehm. 150 E. lebomboensis Verdoorn 220 15N-fixation E. /ehmannii Lehm. 40 The fresh root samples were incubated in 25 cm3 round­ E. longifolius (Jacq.) Lehm. 400 bottom flasks which were attached to an all-glass manifold. •£. manikensis Gilliland 120 E. munchii R.A. Dyer & Verdoorn 90 The apparatus was evacuated to half atmospheric pressure E. natalensis R.A. Dyer & Verdoorn 30 and brought to atmospheric pressure again by the addition E. ngoyanus Verdoorn 10 of argon. This was done five times in succession whereafter E. paucidentatus Stapf & Burtt Davy 240 the system was again evacuated to half atmospheric pressure. E. princeps R.A. Dyer 20 Nitrogen-15 gas was now added until the pressure in the E. transvenosus Stapf & Burtt Davy 10 system was 800Jo of atmospheric whereafter oxygen was added E. trispinosus (Hook) R.A. Dyer 10 E. umbe/uziensis R.A. Dyer 20 to bring the pressure up to atmospheric. E. villosus Lem. 10 After 20 h in the dark at ambient temperature, a sample E. woodii Sander 20 of the incubation gas mixture was withdrawn and stored over Stangeria eriopus (Kunze) Nash 10 alkaline pyrogallol (Umbreit eta/. 1957) to absorb the COz •Exotic to the Republic of South Africa and oxygen it contained, before determining the isotopic ratio of the nitrogen in the remaining gas by mass spectrometry. After the root samples were dried at 80°C, the nitrogen aerenchyma of the cortex (Figure 4). of a 100 mg sample of each species was prepared for mass spectrometry according to the procedure recommended by Acetylene reduction Smith eta/. (1963). A Consolidated Engineering Corporation The Cyanobacteria-infected coralloid roots of all the species Model 21 - 103 mass spectrometer was used to determine the reduced acetylene to ethylene at a significant rate. In no case 28 29 N2/ N2 ratio of the gas samples. did the rate of acetylene reduction by the scrubbed and unscrubbed roots differ materially and in no case could Results ethylene production be detected in the absence of acetylene. Coralloid roots The Cyanobacteria-free coralloid roots of E. jerox and E. All the cycad species investigated (Table 1) were found to be transvenosus also did not reduce acetylene to ethylene. able to produce coralloid roots which grew apogeotropically Although the roots of certain species tested (E. caffer; E. and often emerged a few mm above soil level. Seedlings (one inopinus; E. latijrons; E. lebomboensis; E. longijolius and year old and older) invariably bore clusters of coralloid roots E. paucidentatus) reduced acetylene at a remarkably high rate, at the base of their taproots (Figure 1) - each cluster ap­ this is probably not a characteristic of the species concerned parently arising from a single initial lateral root that emerged but rather the result of one or more of several factors such opposite a protoxylem group of the taproot. For several as the specific endophyte and the physiological state of the species either three or four such clusters therefore formed per roots. Although only the apical 40 mm of the roots were used, plant. Even in the seedling stage, scattered clusters of coralloid it is known from personal observations that these roots have roots also develop from the ordinary roots, apparently at a limited life span and there was no way in which one could random - either as lateral outgrowths (Figure 2) or through ensure that roots of the same age were being used. From other the modification of ordinary root tips. studies (Grobbelaar et al. 1986) it is also apparent that in at Although the coralloid roots 9f all the species were basically least some cases, the individuals of a cycad species may be similar (Figure 3), they differed in thickness, degree of infected by different cyanobacterial symbionts which could branching and in the prominence of their lenticels. In all differ in their nitrogenase activity. cases, at least some clusters of coralloid roots of some individuals of a species, were found to be infected with 15N-fixation Cyanobacteria. The endophyte invariably appeared to be From the atom OJo 15N excess of the coralloid roots at the limited to the intercellular spaces of the centrally located end of the experiment (Table 2) it is clear that in all cases S. Afr. J. Bot., 1986, 52(5) 469

A

~· =c· ~· =r· ~·======~lOcm

Figure 2 Two-year-old Encephalartos /anatus plant with clusters of coralloid roots developing laterally from 'ordinary' roots.

significant amounts of nitrogen had been fixed. A value of 0,015 atom OJo 15N excess or more is usually considered to be significant (Virtanen et a!. 1954). Because the acetylene reduction and 15 N experiments were not carried out simultaneously and on the same clusters of coralloid roots, the results of Tables I and 2 cannot be used to calculate the ratio of molecules of acetylene reduced, to molecules of dinitrogen reduced for a given symbiotic as­ sociation.

Discussion From the results it would appear that most if not all cycad species can produce coralloid roots which can become infected with nitrogen fixing Cyanobacteria. From the work of Lamont & Ryan (1977), it is known that coralloid roots can be produced under aseptic conditions. Their formation is there­ fore not induced by other organisms. Wittman eta!. (1965) suggested that these roots can be considered to be pneuma­ tophores. The fact that many of them develop well below the soil surface and never reach the upper soil layers would however seem to argue against this theory. The nitrogen fixing ability of infected coralloid roots also appears to be a common attribute which may reach high values (Halliday & Pate 1976). Apart from the benefit which the Figure 1 Clusters of coralloid roots arising from the base of the taproot cycad derives from such an association (Bergersen eta!. 1965), and growing apogeotropically to 15 mm above the soil level. A: seedling the non-nitrogen fixing plants in the ecosystem must eventually of Encepha/artos transvenosus; B: two-year-old Cycas thouarsii plant. also gain considerably from this activity (Grove et a!. 1980). 470 S.-Afr. Tydskr. Plantk., 1986, 52(5)

Figure 3 Samples of the coralloid roots used for the 15N-flxation experiment. From left to right, top row: Cycas thouarsii, Encephalartos altensteinii, E. arenarius, E. cajjer E. concinnus, E. cupidus, E. cycadijolius, E. eugene-maraisii, E. jerox; second row: E. jriderici-guilielmi, E. ghellinckii, E. heenanii, E. hildebrandtii, E. horridus, E. humilis, E. inopinus, E. laevijolius, E. lanatus; third row: E. latifrons, E. lebomboensis, E. lehmannii, E. longijolius, E. manikensis, E. munchii, E. natalensis, E. ngoyanus, E. paucidentatus; bottom row: E. princeps, E. transvenosus, E. trispinosus, E. umbeluziensis, E. villosus, E. woodii and Stangeria eriopus.

Table 2 15-N enrichment of cyanobacteria-infected coralloid roofs of cycads during a 20-h exposure to an atmosphere containing 74 atom % 15N under aerobic conditions in the dark

Atom o/o Cycad species 15N excess •cycas thouarsii Gaud. 0,072 Encephalartos altensteinii Lehm. 0,340 E. arenarius R.A. Dyer 0,837 E. cajjer (Thunb.) Lehm. 0,835 •£.. concinnus R.A. Dyer & Verdoorn 0,260 E. cupidus R.A. Dyer 0,143 E. cycadijolius (Jacq.) Lehm. 0,311 E. eugene-maraisii Verdoorn 0,385 E. jerox Bertol. f. 0,239 E. jriderici-guilielmi Lehm. 0,072 E. ghellinckii Lem. 0,122 E. heenanii R.A. Dyer 0,211 Figure 4 Coralloid roots of Encephalartos transvenosus. Note the •E. hildebrandtii A. Braun & Bouche 0,163 circular dark band of Cyanobacteria in the cortex of the transversely E. horridus (Jacq.) Lehm. 0,716 sectioned roots to the left. E. humilis Verdoorn 0,773 E. inopinus R.A. Dyer 0,830 E. laevifolius Stapf & Burtt Davy 0,332 In the present study, two treatments were included to try E. lanatus Stapf & Burtt Davy 0,271 and establish whether the observed nitrogenase activity could E. latifrons Lehm. 0,168 in part be ascribed to either the activity of free living rhi­ E. lebomboensis Verdoorn 0,486 E. lehmannii Lehm. 0,882 zosphere micro-organisms and/or the cycad's own root tissue. E. longijolius (Jacq.) Lehm. 0,488 In both cases the results were negative. •£. manikensis Gilliland 0,083 E. munchii R.A. Dyer & Verdoorn 0,273 Acknowledgements E. natalensis R.A. Dyer & Verdoorn 0,791 The financial support of the Council for Scientific and E. ngoyanus Verdoorn 0,309 Industrial Research and the University of Pretoria to the senior E. paucidentatus Stapf & Burtt Davy 0,447 E. princeps R.A. Dyer 0,208 author is gratefully acknowledged. Thanks are also due to E. transvenosus Stapf & Burtt Davy 0,260 the Botanical Research Institute, Pretoria, the Hartbeespoort E. trispinosus (Hook) R.A. Dyer 0,696 Nursery of the Transvaal Provincial Administration, Mr W.G. E. umbeluziensis R.A. Dyer 0,319 Myers, Poinsettia, White River, Mr Cas Swanepoel of the E. villosus Lem. 0,607 farm Botanica, Pretoria and Mr Emden Pienaar, Plot 33, E. woodii Sander 0,536 Derdepoort, Pretoria, for making coralloid roots available for Stangeria eriopus (Kunze) Nash 1,584 this study. •Exotic to the Republic of South Africa S. Afr. J. Bot., 1986, 52(5) 471

References nitrogen fixation of the cycad Macrozamia riedlei. Aust. J. ALLEN, E.K. & ALLEN, O.N. 1965. Nonleguminous plant Bot. 28: 271 - 281. symbiosis. In: Microbiology and soil fertility, eds Gilmour, HALLIDAY, J. & PATE, J.S. 1976. Symbiotic nitrogen fixation C.M. & Allen, O.N. Oregon State University Press, Corvallis. by coralloid roots of the cycad Macrozamia riedlei: BERGERSEN, F.J., KENNEDY, G.S. & WITTMANN, W. 1965. Physiological characteristics and ecological significance. Aust. J. Nitrogen fixation in the coralloid roots of Macrozamia Pl. Physiol. 3: 349 - 358. communis L. Johnson. Aust. J. bioi. Sci. 18: 1135- 1142. HARDY, R.W.F., HOLSTEN, R.D., JACKSON, E.K. & BOND, G. 1967. Fixation of nitrogen by higher plants other than BURNS, R.C. 1968. The acetylene- ethylene assay for N2 legumes. A. Rev. Pl. Physiol. 18: 107- 126. fixation: Laboratory and field evaluation. Pl. Physiol. 43: GIDDY, C. 1974. Cycads of South Africa. Purnell and Sons, 1185 - 1207. Cape Town. LAMONT, B.B. & RYAN, R.A. 1977. Formation of coralloid GRILLI CAIOLA, M. 1980. On the phycobionts of the cycad roots by cycads under sterile conditions. Phytomorphology 27: coralloid roots. New Phytol. 85: 537-544. 426 - 429. GROBBELAAR, N. & MEYER, J.J.M. 1986. A rapid and inexpensive REINKE, J. 1872. Parasitische Anabaena in Wurzelen der method for measuring gas volumes. S. Afr. J. Sci. 82: 383 . Cycadeen. Gdttingen Nachrichten 107. GROBBELAAR, N. & ROSCH, M.W. 1981. Biological nitrogen SMITH, J.H., LEGG, J.O. & CARTER, J.N. 1963. Equipment fixation in a northern Transvaal savannah. Jl S. Afr. Bot. 47: and procedure for N-15 analysis of soil and plant material with 493-506. the mass spectrometer. Soil Sci. 96: 313 - 318. GROBBELAAR, N., SCOTT, W.E., HATTINGH, W. & UMBREIT, W.W., BURRIS, R.H. & STAUFFER, J.F. MARSHALL, J. The identification of the coralloid endophytes 1957. Manometric techniques, Burgess Pub!. Co., of the southern African cycads and the ability of the isolates to Minneapolis. fix nitrogen. S. Afr. J. Bot. (in press). VIRTANEN, A.I., MOISIO, T., ALLISON, R.M. & BURRIS, GROBBELAAR, N., STRAUSS, J.M. & GROENEWALD, E.G. R.H. 1954. Fixation of nitrogen by excised nodules of the 1971. Non-leguminous seed plants in southern Africa, which fix alder. Acta. chem. scand. 8: 1730 - 1731. nitrogen symbiotically. Plant Soil (special volume), 325- 334. WITTMANN, W., BERGERSEN, F.J. & KENNEDY, G.S. 1965. GROVE, T.S., O'CONNELL, A.M. & MALAJCZUK, N. 1980. The coralloid roots of Macrozamia communis L. Johnson. Effects of fire on the growth, nutrient content and rate of Aust. J. bioi. Sci. 18:1129 - 1134.