Journal of Experimental Botany, Vol. 49, No. 320, pp. 599–609, March 1998

Effect of oxygen availability on nitrogen fixation by two Lotus species under flooded conditions E.K. James1 and R.M.M. Crawford Plant Science Laboratories, Sir Harold Mitchell Building, University of St Andrews, St Andrews, Fife KY16 9AL, UK Received 11 August 1997; Accepted 3 November 1997

Abstract The pasture legumes Lotus uliginosus (Schk.) and Lotus corniculatus (L.), known to differ in their tolerance to flooding, were inoculated with Rhizobium loti and flooded for 60 d while subjected to two levels of dissolved pO : 0.241 and 0.094 mmol ml−-1. L. ulig2 inosus showed significantly greater growth (shoot and root) and N fixation under both pO s, compared to2 2 L. corniculatus, although growth and N fixation by L. 2 corniculatus was not affected by the low pO . 2 Surprisingly, in L. uliginosus., growth, nodulation and N fixation were all increased by low pO while nodula2 2 tion of L. corniculatus was unaffected by low pO . The 2 highest rates of endogenous ethylene production were found with L. corniculatus where low pO plants 2 showed a significant increase over that of the higher pO plants while L. uliginosus plants showed a decline. 2 Root porosity of L. uliginosus doubled in the low pO 2 treatment from a mean of 14.5% in high pO roots to 2 28.5%, whereas that of L. corniculatus was relatively unaffected by pO , being 7% and 9% for high and low 2 pO plants, respectively. The structure of nodules 2 differed little between species and treatments, although nodules/nodulated roots from the L. uliginosus plants had particularly profuse lenticels and aerenchyma. However, L. corniculatus nodules, especially those grown in the lower pO showed signs 2 of early senescence with vacuolation of infected cells and green coloration when cut open. Leghaemoglobin (Lb) concentrations in nodules from both species were unaffected by low pO , although that of L. corniculatus 2 nodules, regardless of pO , was significantly greater 2 than L. uliginosus. Concentrations of the intercellular glycoprotein recognized by the monoclonal antibody MAC265 were significantly reduced in nodules from 1 To whom correspondence should be sent. © Oxford University Press 1998

the low pO treatment in both species. Immunogold 2 labelling showed that the MAC265 antigen was localized primarily within intercellular spaces within nodule cortices from both Lotus species. A marked decrease in deposition of the MAC265 antigen within the cortices of L. uliginosus nodules grown in the lower pO . 2 is discussed in terms of the relative abilities of the two Lotus spp. to maintain an O supply to the N 2 2 fixing bacteroids within submerged nodules. Key words: Lotus uliginosus, Lotus corniculatus, N 2 fixation, flooding, oxygen.

Introduction Nitrogen-fixing legume/rhizobial symbioses are normally adversely affected by flooding or by long-term exposure to artificially-lowered rhizosphere pO ( Witty et al., 2 1986). This is particularly true of common crop species such as soybean (Glycine max), lucerne (Medicago sativa), cowpeas (Vigna unguiculata), peas (Pisum sativum), and Vicia faba (Bond, 1951; Minchin and Pate, 1975; Minchin and Summerfield, 1976; Gallacher and Sprent, 1978; Dakora and Atkins, 1990, 1991). The poor performance of nodulated legumes under flooded conditions has been attributed mainly to a reduction in the supply of O to 2 the nodules (Minchin and Summerfield, 1976; Gallacher and Sprent, 1978; Walker et al., 1983; Shiferaw et al., 1992; Arrese-Igor et al., 1993; Pugh et al., 1995) as, although the enzyme nitrogenase is denatured by O 2 concentrations above 5 mmol m−3 (Gallon, 1992), nevertheless, an adequate supply of O to legume nodules is 2 still essential for the aerobic respiration necessary for high rates of nitrogenase activity ( Witty et al., 1986; Hunt and Layzell, 1993). However, flooding does not always affect nodulated legumes adversely. For example, Pugh

600

James and Crawford

et al. (1995) have shown growth increases after long-term flooding of white clover (Trifolium repens), suggesting that even in temperate regions some legumes of economic importance can grow, nodulate, and fix high amounts of N while flooded. Such wetland symbioses are not 2 restricted to the tropics (Loureiro et al., 1994, 1995). Indeed, flood-tolerant pasture legumes, such as clover and Lotus species, may increase in importance as water levels in temperate highland pastures rise due to the possible effects of global warming (Pugh et al., 1995). However, more information on flooding-tolerant forage legumes is required if they are to be exploited for agricultural use. A legume of potential importance in this respect is L. uliginosus, or ‘greater birdsfoot trefoil’, which is native to northern Europe and the British Isles. L. uliginosus is considered by Grant and Small (1995) to be a parent of the more commonly used (in agricultural terms) L. corniculatus (L.) (birdsfoot trefoil ), which is a tetraploid hybrid of L. uliginosus, L. tenuis, and possibly other Lotus spp. L. uliginosus has also been called ‘L. pedunculatus’ (Grant and Small, 1995), although sensu stricto the name L. pedunculatus is restricted to plants originating from the mountains of Spain. L. uliginosus, in contrast to L. corniculatus, is highly tolerant of flooding and thrives in poorly-drained soils (Justin and Armstrong, 1987). In this present study the growth and N -fixing ability 2 of the symbiosis between L. uliginosus and Rhizobium loti strain WPBS3011 was examined when it was subjected to flooded conditions with two levels of dissolved pO , and 2 was compared with the closely-related symbiosis between L. corniculatus and the same R. loti strain. As well as growth and N fixation, root porosity was also estimated, 2 i.e. aerenchyma and lenticel production (Smirnoff and Crawford, 1983; Jackson et al., 1985; Justin and Armstrong, 1987), in order to assess whether aeration pathways from the aerial stem to the submerged nodules were affected by the pO of the flooded growth media. 2 Endogenous ethylene production was also measured, as ethylene is known to increase aerenchyma production by some species under low pO conditions (Jackson, 1985; 2 Zook et al., 1986; Crawford, 1992), as well as adversely affecting nodulation in some legumes (Goodlass and Smith, 1979; Jackson, 1985; Hirsch, 1992). In the final part of the paper, the O relations of the 2 submerged nodules themselves were examined more closely. This involved first, examining the effects of dissolved pO on concentrations of the O -carrying protein 2 2 leghaemoglobin (Lb). Leghaemoglobin is essential for maintaining an O supply to the N -fixing bacteroids in 2 2 the near-anoxic conditions within the centre of legume nodules (Appleby, 1984), and its expression is closely correlated with nitrogenase expression and N fixation 2 (Bisseling et al., 1980; Dakora, 1995). Secondly, relative concentrations of the glycoprotein recognized by the

monoclonal antibody MAC265 ( VandenBosch et al., 1989) in nodule extracts were measured using enzymelinked immunosorbent assays ( ELISA), and its localization was assessed using immunogold labelling of nodule sections. Previous studies have shown that this glycoprotein occludes intercellular spaces in the nodule cortex surrounding the N -fixing tissue ( VandenBosch et al., 2 1989), including those on Lotus spp. where MAC265 recognises two epitopes at 100 and 110 kDa (James EK, unpublished results). In some legume nodules, e.g. those on soybeans (Glycine max), lupins (Lupinus albus) and Sesbania rostrata (James et al., 1991, 1996, 1997; de Lorenzo et al., 1993; Iannetta et al., 1993a, b, 1995), the glycoprotein(s) recognized by MAC236 and MAC265 have been shown to be involved in the formation and operation of the variable diffusion barrier which finely regulates O diffusion through the cortex into the infected 2 zone ( Witty et al., 1986; Hunt and Layzell, 1993).

Materials and methods Plant material and growth conditions Seeds of L. uliginosus and L. corniculatus were obtained from Chiltern Seeds ( Ulverston, Cumbria, UK ) and germinated in Petri dishes on wet sterile filter paper. After germination, seedlings were placed three to a 12 cm pot in well-drained vermiculite and grown in a greenhouse under natural light (supplemented with artificial light from 400 W mercury vapour bulbs with a 16 h day) and a temperature of 20 °C. All plants were inoculated with Rhizobium loti strain DUS341 ( WPBS3011) at the time of sowing and at weekly intervals for 3 weeks. After 70 d, nodules were well established on all plants and, after thinning to two plants per pot, 20 pots were placed in two tanks (64×57×15 cm) in which they were flooded to 1–2 cm above the hypocotyl for 60 d. For the duration of the experiment the nodulated roots of the flooded plants were subjected to one of two levels of pO : 2 (a) Tank 1. Water fully aerated by bubbling with air, giving a dissolved O concentration of 0.241 mmol ml−1. 2 (b) Tank 2. Water continuously bubbled with N , giving a 2 dissolved O concentration of 0.094 mmol ml-1 (maintained 2 at this low level by oxygen diffusion from shoot to root). The dissolved O concentration in the tanks was monitored 2 daily using a Jenway 9070 portable O meter. 2 Plant growth, nodulation, and C and N content Shoot, root and nodule dry weights were measured after drying in an oven at 80 °C. The C and N content of shoots and roots was examined using a Carlo Erba 1106 CHN analyser (Department of Chemistry, University of St Andrews) according to James et al. (1991). Acetylene reduction activity and endogenous ethylene production As it is not possible to use flow-through gas analyses with flooded pots (Minchin, personal communication), comparative estimates of the nitrogenase activity of intact plants were obtained using the ‘closed’ acetylene reduction assay described by Witty and Minchin (1988). Briefly, whole plants were enclosed in 500 ml jars and subjected to 10% acetylene for 1 h,

Oxygen availability, nitrogen fixation and flooding in Lotus species 601 after which the ethylene concentration in the head spaces was analysed by gas chromatography using a Pye Unicam 104 gas chromatograph fitted with a flame ionization detector. Endogenous ethylene production by whole, intact plants was examined at 21 d into the treatments, and at the end of the experiments, just before harvesting. A modified version of the method described by Jackson et al. (1985) was used where plants were sealed in 20 ml vials for 2 h in the dark, after which 1 ml samples of head space gas were analysed by gas chromatography for the presence of ethylene (see above). Root porosity, microscopy and immunogold labelling Gas-space content of the crown nodulation zone of the roots (0.5 g approximate fresh weight sampled per plant) was estimated (after nodules were removed) using the water displacement method (Jackson et al., 1985; Curran et al., 1996). Submerged hypocotyls, roots and nodules attached to pieces of root were prepared for light microscopy according to James et al. (1992a). Pink, leghaemoglobin (Lb) containing nodules were selected for immunogold labelling for optical microscopy using the monoclonal antibody MAC265 ( VandenBosch et al., 1989). Parallel sections were immunogold-labelled with a polyclonal antibody raised against nitrogenase component II to confirm that the nodules expressed the nitrogenase-enzyme complex. The labelling procedures with both antibodies were as described by James et al. (1996, 1997). MAC265 was a gift from Dr NJ Brewin (John Innes Centre, Norwich, UK ) and the nitrogenase component II antibody was a gift from Dr PW Ludden (Madison, Wisconsin, USA). Nodule soluble protein, Lb and glycoprotein Nodule soluble protein and Lb were extracted in Drabkins solution (+10% polyvinylpyrrolidone; PVP) and their concentrations measured using the techniques described by James et al. (1991). ELISA using MAC265 was performed on the same extracts (after denaturation) using the methods of Iannetta et al. (1993a). Statistical analyses Significant differences between treatments were determined using analysis of variance followed by Duncans multiple range test.

Results Plant growth, C and N content, nodulation, and nitrogenase activity After 60 d flooding, the dry weight of L. corniculatus was unaffected by the dissolved pO in the growth media. 2 However, dry weights of the L. uliginosus plants from the aerated treatment were nearly twice those of L. corniculatus from either treatment ( Table 1). Moreover, compared to the aerated L. uliginosus, there was a significant increase in the dry weight of the L. uliginosus from the N -bubbled treatment. The treatment differences in 2 L. uliginosus total dry weights were primarily due to differences in shoot dry weight (root dry weights were not significantly different), whereas species differences were due to both higher L. uliginosus shoot and root dry weights ( Table 1). The total carbon and nitrogen data closely paralleled the dry weight data ( Table 1). The number, fresh and dry weights of nodules per L. uliginosus plant tended to increase with nitrogen bubbling with the fresh weight of the nodules showing a significant increase ( Table 2). In both species, a substantial proportion of the nodules were clustered at the top of the tap root near the hypocotyl. Nitrogenase activity (ARA) per L. uliginosus plant at the time of harvest was significantly greater than per L. corniculatus plant but, in contrast to the dry weight/total N data, the treatments made no apparent difference to activity comparisons within species ( Table 2). This suggests that errors due to an ‘acetylene-induced decline’ in nitrogenase activity found with other legumes (Minchin et al., 1983), are also likely to occur when Lotus spp. are examined using the ‘closed’ ARA, and hence the actual values should be treated with caution. Nevertheless, the ARA did show that there were large variations in the nitrogenase activity of L. corniculatus from the N -bubbled treatment 2

Fig. 1. (a) Longitudinal section of a Lotus corniculatus nodule and the subtending root (arrow). This nodule was grown in vermiculite flooded with air-saturated water for 60 d. Note the nodule lenticel (L) above the vascular bundle. IZ=infected zone. Bar=100 mm. (b) Higher magnification view of the root/nodule connection from (a). Note the large air spaces in the root (*) and that many of the infected cells are vacuolated (arrows). Bar=50 mm. (c) Longitudinal section of a Lotus corniculatus nodule and the subtending root (arrow). This nodule was grown in vermiculite flooded with deoxygenated (N -bubbled ) water for 60 d. Again, nodule lenticels (L) tend to be situated over the vascular bundles, and note that lenticel/ 2 aerenchyma tissue is more evident at the root/nodule junction than in the nodulated roots from the air-saturated treatment (a, b). As with the airsaturated treatment (a, b), many of the infected cells are vacuolated (small arrows), suggesting bacteroid senescence. IZ=infected zone. Bar= 200 mm. (d) Higher magnification view of the root/nodule connection from (c). This illustrates the root aerenchyma (A) and also some vacuolated infected cells at the base of the nodule (arrows). Bar=50 mm. Fig. 2. (a) Longitudinal section of a Lotus uliginosus nodule and the subtending root (arrow). L. uliginosus nodules are essentially similar in structure to those on L. corniculatus (Fig. 1a–d) except that for both the treatments used in this study, N -fixing (Lb-containing) L. uliginosus 2 nodules tended to be larger and vacuolation of infected cells was rarely evident (b). This nodule was grown in vermiculite flooded with N 2 saturated water for 60 d. Lenticellular tissue (L) covers much of the nodule and is contiguous with aerenchyma tissue (A) on the roots. IZ= infected zone. Bar=200 mm. (b) High magnification view of the root/nodule connection from a Lotus uliginosus nodule grown in vermiculite flooded with air-saturated water for 60 d. Note the profuse aerenchyma (A) on the subtending root. IZ=infected zone. Bar=100 mm. (c) Section through a Lotus uliginosus nodule grown in vermiculite flooded with air-saturated water for 60 d. The section was immunogold labelled (followed by silver-enhancement) with the monoclonal antibody MAC265, and spans the tissue between the infected zone (IZ ) and the outer cortex (OC ) and/or nodule lenticels. Most of the intercellular spaces in the mid-cortex (MC ) are occluded with gold-labelled material (arrows). IC=inner cortex. Bar=25 mm. (d) Section through a Lotus uliginosus nodule grown in vermiculite flooded with N -bubbled water for 60 d. The section was 2 immunogold labelled (followed by silver-enhancement) with the monoclonal antibody MAC265, and spans the tissue between the infected zone (IZ ) and the outer cortex (OC ) and/or nodule lenticels. Gold-labelled intercellular occlusions in the mid-cortex (arrow) are much less evident than in the air-saturated treatment (c). IC=inner cortex. Bar=25mm.

602

James and Crawford

Oxygen availability, nitrogen fixation and flooding in Lotus species 603

604

James and Crawford

Table 1. Effect of 60 d aeration with air or N on the growth (n=15), C and N fixation (n=5) of flooded Lotus corniculatus and 2 Lotus uliginosus Results are means±SE. Values within rows followed by the same letter are not significantly different at P=0.05

Total DW (g) Shoot DW (g) Root DW (g) Total Ca (mg) Total Na (mg)

Lotus corniculatus +Air

Lotus corniculatus +N 2

Lotus uliginosus +Air

Lotus uliginosus +N 2

0.89±0.13 c 0.71±0.11 c 0.17±0.02 b 494±74 c 24±3 c

0.88±0.16 c 0.73±0.14 c 0.14±0.02 b 575±61 c 26±3 c

1.58±0.17 b 1.26±0.13 b 0.27±0.03 a 791±64 b 41±4 b

2.12±0.14 a 1.76±0.13 a 0.30±0.02 a 1043±24 a 55±3 a

aRoots plus shoots.

Table 2. Effect of 60 d aeration with air or N on nodulation and acetylene reduction activity (ARA), of flooded Lotus corniculatus 2 (n=8) and Lotus uliginosus (n=10) Results are means±SE. Values within rows followed by the same letter are not significantly different at P=0.05.

Nodule numbers Nodule FW (mg) Nodule DW (mg) ARA ( mmol C H plant−1 h−1) 2 4

Lotus corniculatus +Air

Lotus corniculatus +N 2

Lotus uliginosus +Air

Lotus uliginosus +N 2

68±17 b 170±30 c 28±7 b 0.284±0.085 b

58±17 b 230±40 c 29±11 b 0.369±0.177 b

252±43 a 600±50 b 90±16 a 1.008±0.174 a

292±36 a 810±80 a 113±15 a 1.108±0.156 a

( Table 2), with some plants showing relatively high activities and some none at all. Mean specific nitrogenase activities (g−1 nodule dry weight) were similar for all the species/treatment combinations (data not shown). Endogenous ethylene, root porosity and nodule structure Twenty-one days into the flooding treatments, the amount of endogenous ethylene produced by all treatments/ species did not differ significantly, except for L. corniculatus from the N -bubbled tank which produced 2–3 times 2 more ethylene than the other plants ( Table 3). By the end of the experiment (after 60 d flooding) the amount of ethylene produced by the N -bubbled L. corniculatus was 2 still greater than any of the other treatments, but only significantly greater than that produced by N -bubbled 2 L. uliginosus ( Table 3). The mean porosity of L. corniculatus tap roots was increased only slightly by N -bubbling, 2 whereas that of the N -bubbled L. uliginosus was doubled 2 compared to the aerated plants ( Table 3). Flooded hypocotyls from all species/treatment

combinations, but particularly L. uliginosus, had epidermal tissues expanded to form aerenchyma (results not shown). Qualitative analyses of lateral root/nodule junctions (Figs 1a–d, 2a, b) showed prominent root aerenchyma. However, at the root/nodule junctions only L. corniculatus from the N -bubbled treatment had an 2 extensive covering of lenticellular tissue (Fig. 1c). The nodule structure from the two Lotus species ( Figs 1a–d, 2a, b) differed little from each other and was similar to that reported previously for L. corniculatus ( Vance et al., 1982) and for L. uliginosus/L. pedunculatus (Pankhurst et al., 1979). However, unlike L. uliginosus, most of the larger L. corniculatus nodules, particularly from the N 2 bubbled treatment appeared to be senescent, i.e. green inside when cut open. Therefore, the Lb-containing L. corniculatus nodules that were sectioned were relatively small ( Fig. 1a–d) compared to the larger, Lb-containing L. uliginosus nodules (Fig. 2a). In addition, unlike the L. uliginosus nodules (Fig. 2a, b), many of the infected cells within the L. corniculatus nodules had prominent vacuoles

Table 3. Effect of 21 d and 60 d aeration with air or N on the production of endogenous ethylene (nmol g–1 plant FW h–1) (n=4) 2 and root porosity (n=4) of flooded Lotus corniculatus and Lotus uliginosus Results are means±SE. Values within rows followed by the same letter are not significantly different at P=0.05.

C H (21 d) 2 4 C H (60 d) 2 4 Root porosity (%)

Lotus corniculatus +Air

Lotus corniculatus +N 2

Lotus uliginosus +Air

Lotus uliginosus +N 2

0.036±0.009 b 0.061±0.017 ab 7±5 b

0.098±0.011 a 0.088±0.008 a 10±1.8 b

0.041±0.002 b 0.053±0.014 ab 14.5±1.8 b

0.031±0.006 b 0.042±0.007 b 28.5±6.3 a

Oxygen availability, nitrogen fixation and flooding in Lotus species 605 Table 4. Effect of 60 d aeration with air or N on nodule protein, Lb (n=8) and glycoprotein (MAC265 antigen) content (n=36) of 2 flooded Lotus corniculatus and Lotus uliginosus Results are means±SE. Values within rows followed by the same letter are not significantly different at P=0.05.

Nodule soluble protein (mg g−1 FW ) Lb (mg g−1 FW ) % Protein as Lb Amount of MAC265 (AMD g−1 FW )a

Lotus corniculatus +Air

Lotus corniculatus +N 2

Lotus uliginosus +Air

Lotus uliginosus +N 2

7.64±0.66 a

5.31±0.28 b

6.15±0.21 b

5.89±0.14 b

1.99±0.26 a 26.0±2.5 ab 750±110 a

1.62±0.15 ab 31.0±3.4 a 400±80 b

1.39±0.07 b 22.5±0.9 b 430±40 b

1.21±0.07 b 20.9±1.2 b 130±20 c

*Average minimum dilution.

( Fig. 1a–d); such vacuoles in determinate nodules (e.g. those on soybean) are often an indicator of stress-induced senescence (James et al., 1991, 1993). The L. corniculatus nodules had small, discrete lenticels situated above the vascular bundles ( Fig. 1a, b), whereas most of the L. uliginosus nodules tended to be covered in aerating tissue which had spread from the lenticels over the surface of the nodules and was contiguous with the aerenchyma on the subtending roots ( Fig. 2a, b). There were no obvious differences in the extent of the lenticellular tissue on the L. uliginosus nodules due to the treatments (Fig. 2a, b, results not shown). Nodule soluble protein, Lb and glycoprotein L. corniculatus nodule soluble protein concentrations were significantly decreased by the low O treatment, whereas 2 those of L. uliginosus were unaffected, being similar to that of N -bubbled L. corniculatus (Table 4). Leghaemoglobin 2 concentrations (and the proportion of nodule soluble protein as Lb) in nodule extracts from both species were not significantly affected by the treatments. However L. uliginosus Lb concentrations were significantly less than those of L. corniculatus (Table 4). Compared to the N -bubbled 2 treatment, concentrations of the MAC265 antigen were significantly higher in extracts from nodules of both species subjected to aeration (Table 4). Aerated extracts of L. uliginosus, were remarkable as they had more than 3-fold the concentration of the MAC265 antigen as compared with the nitrogen bubbled treatments. Qualitative analyses of sections of nodules immunogold labelled with MAC265 showed that the antigen was primarily localized within cortical intercellular spaces ( Fig. 2c, d). Moreover, in comparison with those nodules grown in N -bubbled media ( Fig. 2c), there were more 2 spaces in the inner and mid-cortices of aerated L. uliginosus nodules which contained MAC265-labelled occluding material (Fig. 2d ). Parallel sections which were immunogold labelled with an antibody raised against nitrogenase component II indicated that the nodules from which the sections in Fig. 2c and d were taken, were expressing nitrogenase activity (data not shown). There

were no obvious treatment differences L. corniculatus nodules (data not shown).

with

the

Discussion Effect of dissolved pO on growth, nodulation and N 2 2 fixation Preliminary studies in our laboratory have shown that the wetland legume L. uliginosus can form N -fixing 2 nodules on flooded roots, as well as on adventitious roots arising from flooded hypocotyls and stems. Moreover, it has been demonstrated that growth, nodulation and nitrogenase activity (acetylene reduction activity; ARA) of plants with nodules developed under flooded conditions is equal or greater than that of unflooded plants. The present study has now demonstrated increases in shoot dry matter, C-content, N-content, and nodulation of L. uliginosus when grown in deoxygenated water compared with plants grown in aerated water, and that this species has therefore a metabolic preference for being flooded by water with low dissolved O concentrations. 2 This is the opposite to the results of most studies on the effects of flooding/low pO on nodulated legumes (see 2 Introduction), although a decrease in growth/N fixation 2 after aeration of the hydroponic rooting medium has been shown with the tropical wetland legume Neptunia plena (James et al., 1992b), and positive effects of flooding, as compared to free-drainage, have been demonstrated with Aeschynomene fluminensis (Loureiro et al., 1995), and the ‘non-flooding-tolerant’ legumes, soybean and white clover (Trifolium repens) (Nathanson et al., 1984; Pugh et al., 1995). The decreases in N fixation reported in most flooding 2 studies have usually been attributed to reduced nodulation (Shiferaw et al., 1992) and/or reductions in O supply to 2 the nodules (Bond, 1951; Minchin and Pate, 1975; Minchin and Summerfield, 1976; Walker et al., 1983; Pugh et al., 1995). However, experimental conditions can have a profound effect on the relative ability of nodulated legumes to survive, grow and fix N under low pO and/or 2 2

606

James and Crawford

flooding. For example, although the present study has confirmed the high flooding tolerance reported for L. uliginosus (syn. L. pedunculatus; Grant and Small, 1996) by a number of workers (Justin and Armstrong, 1987; Shiferaw et al., 1992), Shiferaw et al. (1992) reported varying decreases in growth, nodulation and N 2 fixation by a number of L. pedunculatus accessions, including cv. Maku which they classified as having ‘good’ tolerance. The apparent difference in flooding tolerance between the L. uliginosus used in this present study and that of Shiferaw et al. (1992), could be due to the fact that in the latter study, the redox potential of the soil had dropped to −300 mV after 14 d waterlogging and hence the soil would have been both highly anaerobic and reducing. Even stagnant and deoxygenated (N 2 bubbled ) water is unlikely to be anaerobic (Brix and Sorrell, 1996) and hence the actual O -deprivation stress 2 imposed on the flooded vermiculite-grown plants in the present study would have been less than that imposed by the anaerobic flooded soil used in the study of Shiferaw et al. (1992) and more likely to match that found under natural flooded conditions. In addition, anaerobic conditions result in abundant ferrous iron and manganese (Crawford, 1992; Shiferaw et al., 1992), and hence toxicity caused by these ions could have had an additional detrimental effects on the plants of Shiferaw et al. (1992). In the present study, the results will have largely been due to the dissolved pO level rather than to indirect factors, 2 such as an increase in toxic ions. Although L. corniculatus is generally considered to be a non-wetland species and will not germinate under flooded conditions an exhaustive study of 91 species from well-drained to wetland habitats, (Justin and Armstrong, 1987) concluded on the basis of a number of growth parameters, that L. corniculatus was wrongly classified among ‘non-wetland’ species and should be reclassified as an ‘intermediate’ species, even although its growth was reduced by half when flooded. In addition, Shiferaw et al. (1992) determined the tolerance of nodulated L. corniculatus to flooding as being ‘fair’ when compared to a number of other pasture legumes (in agreement with Justin and Armstrong, 1987) with growth being halved compared to well-drained controls. The present study has not only confirmed the flooding tolerance of L. corniculatus, but has also shown that it is even moderately tolerant of being flooded with water with a greatly reduced dissolved pO . 2 Unsurprisingly, as nodule N fixation is essentially an 2 aerobic process ( Witty et al., 1986; Hunt and Layzell, 1993), most studies of flooding/low O tolerance in N 2 2 fixing legumes have emphasized the importance of maintaining an O supply to the nodules (see Introduction). 2 In most legumes with determinate nodules, such as soybean (Pankhurst and Sprent, 1975; Parsons and Day, 1990), cowpea (Minchin and Summerfield, 1976; Dakora

and Atkins, 1990) and Lotus (Pankhurst et al., 1979), lenticels are typically present on nodules from both flooded and non-flooded plants. However, nodule, root and stem lenticels/aerenchyma tend to increase after imposition of flooding/reduced rhizosphere pO and it 2 has been suggested that as long as continuous networks of air pathways from the aerial stem to submerged nodules are maintained then N fixation can still take place (James 2 et al., 1992b; Arrese-Igor et al., 1993), albeit sometimes at reduced rates, e.g. Viminaria juncea ( Walker et al., 1983). The present study suggests strongly that such air pathways are also important in flooded Lotus species. For example, there was profuse stem, root and nodule lenticellular tissue/aerenchyma on both species, particularly on L. uliginosus, and nodules also tended to be concentrated on the tap root near the hypocotyl, the latter being an adaptation which Zook et al. (1986) and James et al. (1992a, b) reported as being important in reducing pathlengths of gaseous diffusion. Moreover, tap roots from both Lotus species had relatively high porosities ( Table 4), that of L. corniculatus (7–10%) being characteristic more of an ‘intermediate’ than a ‘nonwetland’ plant (Justin and Armstrong, 1987), and that of L. uliginosus (14.5–28.5%) being classically characteristic of wetland and flooding-tolerant plants (Justin and Armstrong, 1987; Crawford, 1992). In flooded L. uliginosus, probably as a result of high root/hypocotyl porosity and profuse nodule lenticels, O 2 supply to the nodules was presumably not limiting. Indeed, this is well demonstrated by the fact that aeration resulted in less dry matter and C and N accumulation than deoxygenation of the flooded medium. However, this was apparently not the case with L. corniculatus as the N -bubbled treatment resulted in nodule senescence, 2 particularly in the larger nodules, as evidenced by the decline in their soluble protein content ( Vance et al., 1982; James et al., 1991), and by the lack of nitrogenase activity (ARA) exhibited by some of the plants. Moreover, the vacuolation of infected cells observed in the present study is similar to that reported in soybean nodules subjected to reduced external pO (James et al., 1991) or 2 to stresses which reduce nodule O permeability (James 2 et al., 1993). The early nodule senescence was probably linked to the inability of the N -bubbled L. corniculatus 2 plants to maintain an O supply to the nodules equal to 2 that of the aerated plants, and this would have been a particular problem in larger nodules with consequently longer diffusive pathways to the N -fixing cells (Dakora 2 and Atkins, 1990). Not only did the N -bubbled L. uliginosus plants 2 evidently maintain an O supply to the submerged nodules 2 but they also produced more nodule mass than the aerated treatment. Indeed, the greater mass of nodules was probably one of the main reasons for the higher N fixation 2 shown by the N -bubbled plants. The increase in nodula2

Oxygen availability, nitrogen fixation and flooding in Lotus species 607 tion by the N -bubbled L. uliginosus could be due to two 2 reasons. First, more sites may have been available for infection by R. loti, particularly on the highly lenticellular tap roots/hypocotyls or secondly, as a consequence of lower concentrations of the nodulation-inhibiting hormone, ethylene compared to the other species/treatment combinations (Goodlass and Smith, 1979; Jackson, 1985; Hirsch, 1992). Interestingly, there appeared to be no positive link between endogenous ethylene and aerenchyma production in L. uliginosus as the N -bubbled 2 plants produced less ethylene in their roots and had a higher porosity than the aerated plants. This is the opposite to the effect seen with many non-hydrophytic plants (Jackson, 1985; Jackson et al., 1985), including legumes ( Zook et al., 1986), when subjected to stresses caused by flooding or lowered rhizosphere pO . This 2 could be due to the fact that, far from being stressed by the treatment, hydrophytic L. uliginosus actually benefits from being grown in deoxygenated water and, like many hydrophytes (Jackson, 1985), will always produce extensive aerenchyma irrespective of the growth conditions. By contrast, N -bubbled L. corniculatus produced more 2 endogenous ethylene than all the other species/treatments, and in this respect was behaving as though it was stressed. However, due to large variations these findings are not conclusive for either Lotus spp. and further studies are needed to establish whether there are links between the dissolved pO of the flooded growth medium on the one 2 hand, and endogenous ethylene, aerenchyma production and nodulation on the other. O diffusion within submerged nodules 2 Further circumstantial evidence that the N -bubbled 2 L. uliginosus, but not L. corniculatus, nodules obtained sufficient O comes from the Lb data. In legume nodules, 2 the main function of Lb is to convey dissolved O rapidly 2 through the near-anaerobic cytoplasm of the infected cells to the rapidly-respiring, N -fixing bacteroids (Appleby, 2 1984; Witty et al., 1986; Hunt and Layzell, 1993; Dakora, 1995). If the external O supply to the nodules is reduced 2 it would be expected that Lb levels would consequently increase and/or nodule size be reduced to decrease the length of diffusive pathways (Dakora and Atkins, 1990). The present study seems to support this contention, in that the Lb concentration of the L. uliginosus nodules (expressed on both fresh weight and soluble protein terms) was always less than that of the L. corniculatus nodules and, moreover, was unaffected by the low O treatment. 2 On the other hand, expressed as a percentage of soluble protein, the Lb content of L. corniculatus nodules was increased by N -bubbling and most of the active nodules 2 were small. In both these respects, L. corniculatus responded in a similar manner to flooding/lowered rhizosphere pO to other non-wetland symbioses. For example, 2

James et al. (1991) showed an increase in Lb concentration in soybean nodules grown in 10% O , and Dakora 2 et al. (1991) and Arrese-Igor et al. (1993) respectively, showed that, on a per bacteroid basis, Lb increased in soybean and lucerne nodules subjected to 1% O . On the 2 other hand, the Lb content of nodules from N. plena did not differ if they were grown in freely-drained or flooded conditions and, as in the present study of L. uliginosus, James et al. (1992b) concluded that O supply to the 2 flooded nodules on this wetland legume was adequate without aeration of the medium. Oxygen diffusion into the central (Lb-containing) zone of legume nodules is regulated by a variable O diffusion 2 barrier located within the nodule inner cortex ( Tjepkema and Yocum, 1974; Witty et al., 1987; Parsons and Day, 1990). Along with Lb, this barrier is essential in allowing nodules to balance the potentially antagonistic requirements of O for aerobic bacteroid respiration on the one 2 hand, with the need for a low pO to prevent denaturation 2 of the nitrogenase enzyme on the other ( Witty et al., 1986; Gallon, 1992; Hunt and Layzell, 1993). Analyses of Lotus nodules using non-invasive Lb spectrophotometry (Denison et al., 1992) and O -specific micro-electrodes 2 (Skøt et al., 1996) suggest that they are no different to other nodules (e.g. those on pea; Witty et al., 1987) in this respect. Recent evidence suggests that the variable O diffusion barrier in some legume nodules, such as 2 those on soybean (James et al., 1991; Iannetta et al., 1993a), lupin (de Lorenzo et al., 1993; Iannetta et al., 1993b, 1995; James et al., 1997) and Sesbania rostrata (James et al., 1996), may be partly regulated by the variable occlusion of intercellular ‘air’ spaces in the inner and mid-cortices by material which includes a glycoprotein recognized by the monoclonal antibodies MAC236 and MAC265 ( VandenBosch et al., 1989). The present study supports this hypothesis in that ELISAs of extracts of L. uliginosus nodules demonstrated that nodules in deoxygenated water contained significantly less MAC265 antigen than those grown in aerated water. The ELISA data is likely to have given an accurate estimation of cortical glycoprotein as the MAC265 antigen in the extracts must have originated from cortical intercellular spaces (Fig. 2c, d), rather than from infection threads/ infection droplets within the infected cells ( VandenBosch et al., 1989). However, in the case of L. corniculatus, although ELISAs showed a reduced concentration per plant of the MAC265 antigen in nodules grown in deoxygenated water, the lack of any obvious differences in immunogoldlabelled nodule sections suggested that this was not necessarily due to an overall decrease in cortical intercellular occlusions. Indeed, the fact that nodule soluble protein concentrations per plant were also reduced by the N 2 bubbled treatment, suggests that the decrease (per plant) in the glycoprotein recognized by MAC265 was more a

608

James and Crawford

reflection of the increased nodule senescence caused by this treatment, rather than low pO per se. James et al. 2 (1997) have shown a positive relationship between the expression of nitrogenase and levels of intercellular glycoprotein (MAC236/MAC265 antigens) in developing lupin nodules and concluded that only those nodules that contain nitrogenase need an O -diffusion barrier to pro2 tect it, and hence require high levels of the intercellular glycoprotein. Therefore, it is reasonable to suggest that if a plant has a high proportion of large senescent nodules containing no nitrogenase (as on the N -bubbled 2 L. corniculatus nodules), they no longer require a glycoprotein-containing O diffusion barrier to protect them 2 from O toxicity, and hence glycoprotein concentrations 2 per plant will be reduced.

Conclusions The present study has demonstrated that the wetland species, L. uliginosus is well adapted to tolerate flooded conditions with a low dissolved pO and, indeed, may 2 even prefer being grown within deoxygenated water. Nodules readily form on submerged roots and stems, and oxygen is supplied to the submerged nodules via a network of extensive stem, root and nodule lenticels/aerenchyma. The nodules themselves are able to further adapt to lowered pO via a reduction in glycoprotein-containing 2 intercellular space occlusions in the inner and midcortices. The present study has also confirmed the ‘moderate’ tolerance of L. corniculatus to flooding/reduced pO 2 (Justin and Armstrong, 1987; Shiferaw et al., 1992). However, this species is clearly stressed by being grown in a deoxygenated flooded medium, as evidenced by relatively high endogenous ethylene levels and early nodule senescence. This stress is probably due to a lesser ability (compared to L. uliginosus ) to transport O to the 2 N -fixing bacteroids within the submerged nodules. 2

Acknowledgements We thank Dr PPM Iannetta for help with ELISAs, Professor JI Sprent for use of microscopy facilities at the Dept. of Biological Sciences, University of Dundee, and Dr FR Minchin, Dr R Pugh, Dr JF Witty for helpful discussions. M Gruber, H Hodge and M Kierans are thanked for their technical assistance. This work was funded by NERC.

References Appleby CA. 1984. Leghaemoglobin and Rhizobium respiration. Annual Reviews in Plant Physiology 35, 443–78. Arrese-Igor C, Royuela M, de Lorenzo C, de Felipe MR, Aparicio-Tejo PM. 1993. Effect of low rhizosphere oxygen on growth, nitrogen fixation and nodule morphology. Physiologia Plantarum 89, 55–63. Bisseling T, Moen AA, Van den Bos RC, Van Kammen A. 1980. The sequence of appearance of leghaemoglobin and nitro-

genase components I and II in root nodules of Pisum sativum. Journal of General Microbiology 118, 377–81. Bond G. 1951. Symbiosis of leguminous plants and nodule bacteria. IV. The importance of the oxygen factor in nodule formation and function. Annals of Botany 15, 95–110. Brix H, Sorrel BK. 1996. Oxygen stress in wetland plants: comparison of de-oxygenated and reducing root environments. Functional Ecology 10, 521–6. Crawford RMM. 1992. Oxygen availability as an ecological limit to plant distribution. Advances in Ecological Research 23, 93–185. Curran M, James P, Allaway WG. 1996. The measurement of gas spaces in the roots of aquatic plants—Archimedes revisited. Aquatic Botany 54, 255–61. Dakora FD. 1995. A functional relationship between eghaemoglobin and nitrogenase based on novel measurements of the two proteins in legume root nodules. Annals of Botany 75, 49–54. Dakora FD, Atkins CA. 1990. Morphological and structural adaptation of cowpea to functioning under sub- and supraambient oxygen pressure. Planta 182, 572–82. Dakora FD, Atkins CA. 1991. Adaptation of nodulated soybean (Glycine max L. Merr.) to growth in rhizospheres containing nonambient pO . Plant Physiology 96, 728–36. 2 De Lorenzo C, Iannetta PPM, Fernandez-Pascual M, James EK, Lucas MM, Sprent JI, Witty JF, Minchin FR, de Felipe MR. 1993. Oxygen diffusion in lupin nodules. II. Mechanisms of diffusion barrier operation. Journal of Experimental Botany 44, 1469–74. Denison RF, Hunt S, Layzell DB. 1992. Nitrogenase activity, nodule respiration, and O permeability following detopping 2 of alfalfa and birdsfoot trefoil. Plant Physiology 98, 894–900. Gallacher AE, Sprent JI. 1978. The effect of different water regimes on growth and nodule development of greenhousegrown Vicia faba. Journal of Experimental Botany 29, 413–23. Gallon JR. 1992. Reconciling the incompatible: N fixation and 2 O . New Phytologist 122, 571–609. 2 Goodlass G, Smith KA. 1979. Effects of ethylene on root extension and nodulation of pea (Pisum sativum) and white clover (Trifolium repens L.). Plant and Soil 51, 387–95. Grant WF, Small E. 1995. The origin of the Lotus corniculatus (Fabaceae) complex: a synthesis of diverse evidence. Canadian Journal of Botany 74, 975–89. Hirsch AM. 1992. Developmental biology of legume nodulation. New Phytologist 122, 211–37. Hunt S, Layzell DB. 1993. Gas exchange of legume nodules and the regulation of nitrogenase activity. Annual Review of Plant Physiology and Plant Molecular Biology 44, 483–511. Iannetta PPM, James EK, McHardy PD, Sprent JI, Minchin FR. 1993a. An ELISA procedure for quantification of relative amounts of intercellular glycoprotein in legume nodules. Annals of Botany 71, 85–90. Iannetta PPM, De Lorenzo C, James EK, Fernandez-Pascual M, Sprent JI, Lucas MM, Witty JF, De Felipe MR, Minchin FR. 1993b. Oxygen diffusion in lupin nodules. I. Visualization of diffusion barrier operation. Journal of Experimental Botany 44, 1461–7. Iannetta PPM, James EK, Sprent JI, Minchin FR. 1995. Timecourse of changes involved in the operation of the oxygen diffusion barrier in white lupin nodules. Journal of Experimental Botany 46, 565–75. Jackson MB. 1985. Ethylene and responses of plants to soil waterlogging and submergence. Annual Reviews of Plant Physiology 36, 145–74. Jackson MB, Fenning TM, Drew MC, Saker LR. 1985. Stimulation of ethylene production and gas-space (aeren-

Oxygen availability, nitrogen fixation and flooding in Lotus species 609 chyma) formation in adventitious roots of Zea mays L. by small partial pressures of oxygen. Planta 165, 486–92. James EK, Iannetta PPM, Nixon PJ, Whiston AJ, Peat L, Crawford RMM, Sprent JI, Brewin NJ. 1996. Photosystem II and oxygen regulation in Sesbania rostrata stem nodules. Plant, Cell and Environment 19, 895–910. James EK, Minchin FR, Iannetta PPM, Sprent JI. 1997. Temporal relationships between nitrogenase and intercellular glycoprotein in developing white lupin nodules. Annals of Botany 79, 493–503. James EK, Minchin FR, Sprent JI. 1992b. The physiology and nitrogen-fixing capability of aquatically and terrestrially grown Neptunia plena: the importance of nodule oxygen supply. Annals of Botany 69, 181–7. James EK, Sprent JI, Hay GT, Minchin FR. 1993. The effect of irradiance on the recovery of soybean nodules from sodium chloride-induced senescence. Journal of Experimental Botany 44, 997–1005. James EK, Sprent JI, Minchin FR, Brewin NJ. 1991. Intercellular location of glycoprotein in soybean nodules: effect of altered rhizosphere oxygen concentration. Plant, Cell and Environment 14, 467–76. James EK, Sprent JI, Sutherland JM, McInroy SG, Minchin FR. 1992a. The structure of nitrogen fixing root nodules on the aquatic Mimosoid legume Neptunia plena. Annals of Botany 69, 173–80. Justin SHFW, Armstrong W. 1987. The anatomical characteristics of roots and plant response to soil flooding. New Phytologist 106, 465–95. Loureiro MF, De Faria SM, James EK, Pott A, Franco AA. 1994. Nitrogen-fixing stem nodules of the legume, Discolobium pulchellum Benth. New Phytologist 128, 283–95. Loureiro MF, James EK, Sprent JI, Franco AA. 1995. Stem and root nodules on the tropical wetland legume Aeschynomene fluminensis. New Phytologist 130, 531–44. Minchin FR, Pate JS. 1975. Effects of water, aeration, and salt regime on nitrogen fixation in a nodulated legume–definition of an optimum root environment. Journal of Experimental Botany 26, 60–9. Minchin FR, Summerfield RJ. 1976. Symbiotic nitrogen fixation and vegetative growth of cowpea (Vigna unguiculata (L.) Walp.) in waterlogged conditions. Plant and Soil 45, 113–27. Minchin FR, Witty JF, Sheehy JE, Muller M. 1983. A major error in the acetylene reduction assay; decreases in nodular activity under assay conditions. Journal of Experimental Botany 34, 641–9. Nathanson K, Lawn RJ, De Jabrun PLM, Byth DE. 1984. Growth, nodulation and nitrogen accumulation by soybean in saturated soil culture. Field Crops Research 8, 73–92. Pankhurst CE, Sprent JI. 1975. Surface features of soybean root nodules. Protoplasma 85, 85–98.

Pankhurst CE, Craig AS, Jones WT. 1979. Effectiveness of Lotus root nodules. I. Morphology and flavolan content of nodules formed on Lotus pedunculatus by fast-growing Lotus rhizobia. Journal of Experimental Botany 30, 1085–93. Parsons R, Day DA. 1990. Mechanisms of soybean nodule adaptation to different oxygen pressures. Plant, Cell and Environment 13, 501–12. Pugh R, Witty JF, Mytton LR, Minchin FR. 1995. The effect of waterlogging on nitrogen fixation and nodule morphology in soil-grown white clover (Trifolium repens L.). Journal of Experimental Botany 46, 285–90. Shiferaw W, Shelton HM, So HB. 1992. Tolerance of some subtropical pasture legumes to waterlogging. Tropical Grasslands 26, 187–95. Skøt L, Minchin FR, Timms E, Fortune MT, Webb KJ, Gordon AJ. 1996. Analysis of the two nodulins, sucrose synthase and ENOD2, in transgenic Lotus plants. Plant and Soil 186, 99–106. Smirnoff N, Crawford RMM. 1983. Variation in the structure and response to flooding of root aerenchyma in some wetland plants. Annals of Botany 51, 237–49. Tjepkema JD, Yocum CS. 1974. Measurement of oxygen partial pressure within soybean nodules by oxygen microelectrodes. Planta 119, 351–60. Vance CP, Johnson LEB, Stade S, Groat RG. 1982. Birdsfoot trefoil (Lotus corniculatus) root nodules: morphogenesis and the effect of forage harvest on structure and function. Canadian Journal of Botany 60, 505–18. VandenBosch KA, Bradley DJ, Knox JP, Perotto S, Butcher GW, Brewin NJ. 1989. Common components of the infection thread matrix and the intercellular space identified by immunocytochemical analysis of pea nodules and uninfected roots. EMBO Journal 8, 335–42. Walker BA, Pate JS, Kuo J. 1983. Nitrogen fixation by nodulated roots of Viminaria juncea (Schrad. and Wendl.) Hoffmans. (Fabaceae) when submerged in water. Australian Journal of Plant Physiology 10, 409–21. Witty JF, Minchin FR. 1988. Measurement of nitrogen fixation by the acetylene reduction assay: myths and mysteries. In: Beck DP, Materon LA, eds. Nitrogen fixation by legumes in Mediterranean agriculture. ICARDA, 331–44. Witty JF, Minchin FR, Skøt L, Sheehy JE. 1986. Nitrogen fixation and oxygen in legume root nodules. Oxford Surveys in Plant Molecular and Cell Biology 3, 275–314. Witty JF, Skøt L, Revsbech NP. 1987. Direct evidence for changes in the resistance of legume root nodules to O 2 diffusion. Journal of Experimental Botany 38, 1129–40. Zook DM, Erwin DC, Stolzy LH. 1986. Anatomical, morphological, and physiological responses of alfalfa to flooding. Plant and Soil 96, 293–96.