Journal of Experimental Botany, Vol. 53, No. 375, pp. 1735±1745, August 2002 DOI: 10.1093/jxb/erf016
Effect of exogenous ¯avonoids on nodulation of pea (Pisum sativum L.) Ï krdleta, Martina KropaÂcÏovaÂ, Ludmila LisaÂ and Marie NeÏmcovaÂ K. NovaÂk1, P. Chovanec, V. S Division of Microbial Ecology, Institute of Microbiology, Academy of Sciences of the Czech Republic, VõÂdenÄskaÂ 1083, Prague 4, 142 20, Czech Republic Received 25 September 2001; Accepted 27 March 2002
Abstract Selected ¯avonoids that are known as inducers and a suppressor of nodulation (nod) genes of the symbiotic bacterium Rhizobium leguminosarum bv. viciae were tested for their effect on symbiosis formation with garden pea as the host. A solid substrate was omitted from the hydroponic growing system in order to prevent losses of ¯avonoids due to adsorption and degradation. The presumed interaction of the tested ¯avonoids with nod genes has been veri®ed for the genetic background of strain 128C30. A stimulatory effect of a nod gene inducer naringenin on symbiotic nodule number formed per plant 14 d after inoculation was detected at concentrations of 0.1 and 1 mg ml±1 nutrient solution. At 10 mg ml±1, the highest concentration tested, naringenin was already inhibitory. By contrast, nodulation was negatively affected by a nod gene suppressor, quercetin, at concentrations above 1 mg ml±1, as well as by another tested nod gene inducer, hesperetin. The deleterious effect of hesperetin might be due to its toxicity or to the toxicity of its degradation product(s) as indicated by the inhibition of root growth. Both the stimulatory effect of naringenin and the inhibitory effect of quercetin on nodule number were more pronounced at earlier stages of nodule development as revealed with speci®c staining of initial nodules. The lessening of the ¯avonoid impact during nodule development was ascribed to the plant autoregulatory mechanisms. Feedback regulation of nodule metabolism might also be responsible for the fact that the naringenin-conditioned increase in nodule number was not accompanied by any increase in nitrogenase activity. By contrast, the inhibitory action of quercetin and hesperetin on nodule number was associated 1
with decreases in total nitrogenase activity. Naringenin also stimulated root hair curling (RHC) as one of the earliest nodulation responses at concentrations of 1 and 10 mg ml±1, however, the same effect was exerted by the nod gene suppressor, quercetin, suggesting that feedback regulatory mechanisms control RHC in the range of nodulation-inhibiting high ¯avonoid concentrations. The comparison of the effect of the tested ¯avonoids in planta with nod gene activity response showed a two orders of magnitude shift to higher concentrations. This shift is explained by the absorption and degradation of ¯avonoids by both the symbionts during 3 d intervals between hydroponic solution changes. The losses were 99, 96.4, and 90% of the initial concentration of 10 mg ml±1 for naringenin, hesperetin, and quercetin, respectively. Key words: ¯avonoids, nod gene, Pisum, Rhizobium, symbiosis.
Introduction A key event in the formation of the symbiosis between legume plants and nodule bacteria (rhizobia) is the activation of rhizobial nodulation (nod) genes. The ®nal product of nod genes are the secreted lipo-chitin oligosaccharides, termed as Nod factors, that represent the agent inducing initial nodule formation in the host root. In parallel, Nod factors elicit a number of other early symbiotic responses, in particular, root hair curling (RHC, for a review, see Spaink, 2000). On the other hand, rhizobial nod genes are known to be activated by ¯avonoid compounds that are present in the plant root exudates. Flavonoids interact speci®cally with
To whom correspondence should be addressed. E-mail: [email protected]
ã Society for Experimental Biology 2002
1736 NovaÂk et al. Table 1. Microbial strains and plasmids used Denotation
Symbiotically competent, wild-type strain of Rhizobium leguminosarum bv. viciae R. leguminosarum bv. trifolii cured of its indigenous Sym plasmid, carrying pMP154 and pMP280, RifR, used as a source of the latter plasmids Asymbiotic derivative of 128C30 carrying plasmid MP154 128C30 with plasmid MP154 MBR6 with plasmid MP280 Laboratory strain of Escherichia coli, F± e14± (McrA±) metB1 lacY1 galK2 galT22 glnV44 hsdR2 (rk±mk+) mcrB; a host for wide-host-range plasmids allowing their conjugative transfer to rhizobia Derivative of the vector pMP190, 15 kb, IncQ, carries promoter nodABCIJ from pSymRL1JI of the standard strain 248 of R. leguminosarum bv. viciae transriptionally fused to lacZ of E.coli, codes for Cmr Smr Derivative of the vector pMP92, 9.3 kb, IncP, carries ¯avonoid receptor gene nodD from pSymRL1JIÐa symbiotic megaplasmid of the standard R. leguminosarum bv. viciae strain, codes for Tcr A derivative of the plasmid RK2, codes for conjugative functions in trans, Kanr, used as a helper plasmid in conjugative crosses
Czech Collection of Nodule Bacteria, Institute of Crop Production, Prague ± RuzyneÏ Zaat et al., 1988
RBL5280 MBR6 MBR7 MBR16 K802
the protein product of the nodD gene and the active form of NodD is believed to activate transcription through promoters of nod operons. This response is even clearer in those rhizobia for which nodD itself is activated by the NodD±¯avonoid complex, such as Rhizobium leguminosarum bv. viciae, a garden pea (Pisum sativum L.) microsymbiont (Burn et al., 1987). The bulk of the information about the activation process has been obtained from mutational analysis of nod genes, construction of nod gene fusions with reporter genes, fractionation of root exudates and detection of compounds with nod gene-inducing or suppressing activity (Firmin et al., 1986; Burn et al., 1987; Peters and Long, 1988). On the other hand, a very limited amount of information is available on the possible limitation of the symbiosis by ¯avonoid compounds in planta (Zaat et al., 1988; Schlaman et al., 1991; Hungria and Phillips, 1993) or about the effect of additional ¯avonoids on host plant nodulation (Kapulnik et al., 1987; Jain et al., 1990; Kosslak et al., 1990; Cunningham et al., 1991; Zhang and Smith, 1995). Therefore, an effort was made to obtain more data on the effect of exogenously supplied ¯avonoid compounds on symbiosis development in the model symbiotic system, pea±R. leguminosarum bv. viciae. To exclude the effects of other organisms and to reduce adsorption and degradation of ¯avonoids on the surface of a solid phase, a simpli®ed semi-sterile hydroponic system has been optimized from which a solid substrate had been omitted. In addition, the response to ¯avonoids at the whole plant level was compared to the response of the used rhizobial strain alone at the level of nod gene induction.
This work This work This work ATCC, Bacteria and bacteriophages, 19th edn. 1996 Spaink et al., 1987a
Spaink et al., 1987b
Figurski and Helinski, 1979
Materials and methods Plants Seeds of pea cv. `BohatyÂr', selected for an average weight of 0.245 g with 10% tolerance, were sterilized with 2% Chloramine B for 20 min, rinsed with sterile distilled water and germinated for 3 d in Petri dishes in the dark at 28 °C with 3.3 ml of sterile distilled water for each seed. When the main root length reached 2±3 cm, the seedlings were transferred onto aluminium foil covers of beakers that had been sterilized with dry air at 170 °C and ®lled with hydroponic solution up to 1 cm below the edge. The root was ®xed in the perforated aluminium cover to allow the plant to continue the growth. The sides of beakers were shielded against light with a layer of black photographic paper and with an outer layer of aluminium foil. Plantlets were kept in a growth chamber at 18/6 h day/night period, 500 mmol m±2 s±1 of photosynthetically active radiation, 22/ 16 °C temperature, and 75/85% of relative humidity. When different cultivation vessels were tested in order to optimize nodulation, the same growth protocol was used. Bacterial inoculum Strain 128C30 of Rhizobium leguminosarum bv. viciae (Table 1) was chosen for experiments as a highly ef®cient wild-type strain that is highly competitive with native soil rhizobial populations (SÏkrdleta et al., 1993). A 4-d-old culture was scraped off the surface of the yeast±mannitol agar plates (Vincent, 1970) and resuspended in sterile water. For pea plant inoculation, 23104 bacteria, according to a microscope count, were added per ml of hydroponic solution, i.e. 106 bacteria were applied per plant when grown in standard 50 ml vessels. Plant treatment with ¯avonoids The composition of the nutrient solution, if not speci®ed to be different, was based on the solution used for the submerged nodulation of pea by Lie (1969a, b; NovaÂk et al., 1994) containing: K2HPO4 360, KH2PO4 120, MgSO4.7H2O 250, CaSO4 250, and ferric citrate 30 mg l±1. The microelement composition was according to SÏkrdleta et al. (1980) and a nodulation-stimulating basic level of nitrate (0.625 mM) was supplied as Ca(NO3)2
Effect of ¯avonoids in pea 1737 (Streeter, 1988). Distilled water and stock solutions were partially sterilized by boiling. The hydroponic solution was changed twice a week by decanting. The continuous supply of ¯avonoid solution to the root, used by Cunningham et al. (1991), was avoided, since the high partition coef®cient (root/nutrient solution) might lead to the uncontrolled retention of ¯avonoids in the root and its functional damage. The plantlets were left uninoculated for 4 d after transfer to adapt. The changes of nutrient solution contained the speci®ed concentration of the tested ¯avonoids, and the ®rst change also contained bacteria for inoculation. For evaluation, the plants were harvested 7, 10 and 14 d after inoculation (DAI). Flavonoids Naringenin (4¢,5,7-trihydroxy¯avanone, from Aldrich) and hesperetin (3¢,5,7-trihydroxy-4¢-methoxy¯avanone, Serva) were used as compounds with very high nod gene-inducing activity for R. leguminosarum bv. viciae (Firmin et al., 1986; Zaat et al., 1988; NovaÂk et al., 1995). Quercetin (3,3¢,4¢,5,7-pentahydroxy¯avone, dihydrate obtained from Serva) was used as a model nod gene suppressor. The ¯avonoid structures are shown in Fig. 1. Although more potent suppressors for pea rhizobia than quercetin are known, such as daidzein (4¢,7-dihydroxyiso¯avone, Firmin et al., 1986; NovaÂk et al., 1995), this compound has been chosen for its commercial availability in suf®cient amounts. Quercetin ®ts the structural requirements of a nod gene suppressor of R. leguminosarum bv. viciae as formulated by Firmin et al. (1986), partly by possessing a C3 hydroxyl, and experimental evidence of its suppressor activity was provided by NovaÂk et al. (1994). A stock solution of naringenin was prepared in methanol (100 mg ml±1) and stored at ±20 °C. Therefore, the methanol concentration was adjusted to the standard level of 0.01% in all treatments. Master solutions of hesperetin (10 mg ml±1) and quercetin (2 mg ml±1) in 0.1 M Na2CO3 were prepared immediately before dilution with nutrient solution. Alkalinization of the nutrient solution was prevented by an equimolar amount of HCl. Dilutions of all ¯avonoids in nutrient solution were prepared immediately before their addition to the growing system. The concentration of ¯avonoids in the nutrient solution was determined by UV spectrophotometry at 283 nm (naringenin), 284 nm (hesperetin in a non-dissociated form), and 425 nm (quercetin). Naringenin was ®rst extracted with chloroform from nutrient solution acidi®ed with KH2PO4. nod gene-inducing activity of ¯avonoids The nod gene-inducing activity was determined in derivatives of the strain 128C30 carrying pMP154, an IncQ plasmid carrying nodABC±lacZ fusion which allows nod gene activity to be monitored as b-galactosidase activity (Table 1). pMP154 was isolated from the strain Rhizobium RBL5280 (Zaat et al., 1988) by alkaline lysis (Sambrook et al., 1989), transformed into Escherichia coli K802 (Inoue et al., 1990), and conjugatively transferred into 128C30 in a triparental mating using pRK2013 as a helper plasmid (Figurski and Helinski, 1979). In the same way, pMP280 carrying cloned ¯avonoid receptor gene nodD from a standard R. leguminosarum bv. viciae strain 248 (Josey et al., 1979) was isolated from RBL5280 and transferred to the rhizobial recipients. The selection levels of antibiotics were as follows (in mg ml±1): tetracycline 2, streptomycin 20 (E. coli) or 250 (Rhizobium), kanamycin 50, and chloramphenicol 15. In addition to the antibiotic markers, the presence of small plasmids in the derivatives of the recipient strain 128C30 was con®rmed by gel electrophoresis of plasmid preparations (Sambrook et al., 1989). The asymbiotic derivatives of the strain 128C30 are supposed to lack the symbiotic megaplasmid as a result of spontaneous losses occurring in this strain (Leyva
Fig. 1. Structure of the tested ¯avonoids.
et al., 1987). The asymbiotic phenotype was con®rmed by inoculation of a small-seeded model plant Vicia tetrasperma (L.) Schreb. grown on agar-solidi®ed nutrient solution under aseptic conditions. The microplate technique was used for nod gene induction and bgalactosidase assay (NovaÂk et al., 1994). The stock solutions of ¯avonoids were prepared as described for in planta assay. Statistical treatment The data were processed using Statgraphics Statistical Graphics System. The means were separated using one-way analysis of variance and a multiple range test based on Tukey intervals at the 95% con®dence level. The number of replicates used is indicated in the individual graphs. Symbiosis evaluation In detached root systems of the harvested plants, total nitrogenase activity (TNA) was determined as acetylene-reducing activity using a closed-chamber system according to SÏkrdleta et al. (1987). Subsequently, the following traits were estimated: nodule number per root, nodule fresh and dry mass, number of leaves, shoot length, nodulated root and shoot dry mass. In the plants harvested 7 and 10 DAI, whole root systems were stained to detect emerging initial nodules by a scaled-up version of the technique of Niehaus and PuÈhler (1988) as described below. Root staining procedure The whole root systems were ®xed in ethanol:acetic acid, 2:1 (v/v) and stored at 4 °C until processed. Before staining, each root was placed in a 50 ml beaker and rinsed three times with distilled H2O for 2 min each time, 1 min with 1 N HCl, and treated for 6 min with 1 N HCl at 60 °C on a water bath in the hydrolysis step. After rinsing with cool 1 N HCl, the roots were incubated for 2 h in Shiff's reagent and destained four times for 5 min with 45% acetic acid. The stained roots were stored without changes in 60% glycerol at 4 °C. The volume of liquids in each treatment was 50 ml per root. To suppress the background staining completely, it was necessary to clear the Shiff's reagent before use by ®ltration through charcoal. The initial nodule number per root was determined by examination under a stereoscopic microscope in 60% glycerol. The degree of RHC was estimated by bright ®eld microscopy in an arbitrary 6ball scale.
1738 NovaÂk et al.
Fig. 2. Effect of the cultivation vessel shape, volume and buffering with CaHPO4 on submerged nodulation of pea. (a) Final pH of the nutrient solution after a 3 d contact with the root of a 21-d-old (14 d after inoculation) pea plant. Initial pH was 7.0. (b) Effect on number of nodules formed per root. (c) Effect on total nodule dry mass. (d) Effect on total nitrogenase activity. Data represent a calculated mean of ®ve replicates (b) or a result obtained for pooled ®ve samples (a, c, d). Tukey con®dence intervals at a 0.05 signi®cance level are indicated in (b), where the data points without a shared letter in the index are signi®cantly different.
Results Submerged nodulation
As no solid substrate was used in the employed hydroponic system, the roots grew and nodulated submerged in the growing medium. Since the development of the symbiotic nodules is hampered under such conditions, optimization of the growing system was required. An emphasis was put on the prevention of losses of the added ¯avonoids by adsorption and degradation and on the obtaining of nodulated plants in a minimum volume of nutrient solution to reduce growing space, handling, and consumption of nutrient solution with supplemented ¯avonoids. The high af®nity of ¯avonoid nod gene inducers to solid surfaces, partly to a wide range of ®lters, was pointed out by Zaat et al. (1988) and Recourt et al. (1991). Glass cultivation vessels of different shapes and volumes were tested with respect to nodule development
under the above regime of nutrient solution changes (Fig. 2). Satisfactory results were obtained with 30, 50, and 100 ml beakers. Since strong acidi®cation of the nutrient solution by pea roots, which might be a limiting factor for nodulation (Lie, 1969b; Richardson et al., 1988), was observed, buffering with solid calcium hydrogen phosphate (CaHPO4.2H2O, 2 g l±1) was attempted. Smoothing of the pH drops, especially in small volumes, was accompanied by improvement of the symbiotic traits (Fig. 2), con®rming this assumption. The inhibitory effect of pH below 5.5 on nodulation is consistent with an almost complete inhibition of nodulation at pH 4.5, as reported by Lie (1969b). However, the addition of solid phosphate was not used in subsequent experiments in view of the tendency of ¯avonoids to adsorb onto solid surfaces (Zaat et al., 1988). The deleterious effect of acidi®cation was diminished by the selection of 50 ml beakers as a standard in further work. Aeration of the fresh nutrient
Effect of ¯avonoids in pea 1739
Fig. 3. Effects of ¯avonoids on the activity of nod genes and growth of Rhizobium leguminosarum bv. viciae 128C30. The activity of nodABC operon is followed indirectly as b-galactosidase activity after the addition of naringenin (a) or hesperetin (b) in 128C30 and its derivatives containing nodABC-lacZ fusions as listed in Table 1. While MBR6 does not contain the nodD gene for ¯avonoid receptor in addition, MBR7 carries nodD from 128C30 and MBR16 a standard nodD. Suppressor activity of rising concentration of quercetin was followed in strain MBR7 supplemented with 0.02 mg ml±1 of naringenin or hesperetin, respectively. (c) Effect of a ¯avonoid concentration on the ®nal density of the suspension of 128C30 after 20 h incubation in the microplate assay of nod gene-inducing activity. Data are means of two independent determinations (a, b) or three incubations (c).
solution with sterile air for 1 h before a change did not improve the symbiosis, indicating that the availability of oxygen was not a limiting factor in submerged nodulation, in contrast to the general opinion (Lie, 1974). Given the absence of a delimited rhizosphere in submerged nodulating plants, disturbances in the root colonization by rhizobia and subsequent infection were expected. In a homogeneous solution, the ¯avonoid gradients required for the bacterial chemotaxis to the root (Bauer and Caetano-AnolleÂs, 1990) may be lacking. However, full independence of the number of resulting nodules from the number of added rhizobia in the range from 102±108 bacteria per plant, i.e. 2±23106 bacteria ml±1 of nutrient solution (not shown) indicated that the normal rhizobial multiplication and colonization of the root surface took place. Careful protection of the root system from visible light turned out to be a very important factor in nodulation. In the unprotected beakers, nodulation was completely prevented, in accordance with the observations reported by Lie (1969a, 1974). In the optimized growing system, the initial lateral roots appeared 4 d after seedling transfer to beakers, i.e. at the time of inoculation. First macroscopically visible nodules appeared 7 DAI at the 4±5 leaf stage. Traces of nitrogenase activity could be detected as early as 10 DAI, while signi®cant activity, ranging from 1 to 5 mmol C2H2 h±1 g±1 nodule dry mass, was measured only 14 DAI. The nitrogenase activity onset corresponded to the pink coloration of nodules that indicates the presence of leghaemoglobin, a marker of maturity of the differentiated nodules. At the time of plant evaluation, i.e. 21 d after germination and 14 DAI, plants had formed 7±8 leaves, including two rudimentary, and approximately 100 nodules, including approximately 60 pink nodules. Since the tested ¯avonoids are not directly soluble in water, some organic solvents were used as agents for their dissolution. The method for testing their potential interference with nodulation followed the procedure described in the Materials and methods section for testing the ¯avonoid effect. Ultimately, methanol was chosen as the most suitable solvent for naringenin since no effects on nodulation and plant development were detectable at a concentration as high as 0.1% (v/v). Ethanol was found to be fully inhibitory for nodulation at 0.1% and to decrease nodule number by 85% at 0.01% concentration. In addition, ethanol at 0.1% decreased shoot fresh mass by 33% and the root fresh mass by 60%. An alternative solvent for ¯avonoids, dimethyl sulphoxide, that has already been used for the delivery of ¯avonoid inducers to rhizobia themselves (Firmin et al., 1986), was found to be toxic to pea plants at 0.01% in the nutrient solution. The other two ¯avonoids, hesperetin and quercetin, had to be dissolved as phenolates in Na2CO3 solution to deliver the maximum concentration required to the nutrient solution.
1740 NovaÂk et al.
Fig. 4. Effect of exogenous ¯avonoids on the development of symbiosis in pea inoculated with Rhizobium leguminosarum bv. viciae 128C30. (a) Effect on root hair curling (RHC) 7 d after inoculation (DAI) expressed in a 6-ball scale (0, no RHC; 5, maximum RHC). (b) Number of nodules per plant formed 14 DAI. (c) Total nitrogenase activity determined as acetylene reduction, 14 DAI. (d) Effect on the growth of the root system, 14 DAI. Means from ®ve replicates (plants) are presented. Tukey con®dence intervals at a 0.05 signi®cance level are indicated for data within each series. The data points without a shared letter in the index are signi®cantly different.
Flavonoid persistence in the hydroponic system
To learn whether the exogenous ¯avonoids persisted in the nutrient solution between changes, the residual ¯avonoid content was determined in the used nutrient solution collected 7 DAI. The presence of added ¯avonoids after 3.5 d of contact with bacteria and the plant root was detectable only at the highest concentration applied, and their contents were reduced to 1, 3.6 and 10% of the input for naringenin, hesperetin, and quercetin, respectively. The ¯avonoid fraction that disappeared from the solution has to be ascribed to the absorption by the plant root, spontaneous degradation, degradation by bacteria or by the plant root. Fortunately, the used system excludes appreciable degradation by other organisms or adsorption on a substrate. Response of the strain 128C30 to ¯avonoids
As shown in Fig. 3a, b, both naringenin and hesperetin act as strong nod gene inducers in the strain 128C30.
Moreover, the character of the response was almost indistinguishable from that mediated by the ¯avonoid receptor NodD from the standard strain 248. While the indigenous b-galactosidase activity of the strain 128C30 was negligible, the introduction of the plasmid MP154 led to the appearance of increased, but still constitutive, bgalactosidase activity. This activity was obviously due to the spontaneous transcription from the nodABC promoter, as already pointed out by Spaink et al. (1987b). Introduction of the cloned ¯avonoid receptor gene nodD on pMP280 in strain RBL16 allowed a marked response to the added ¯avonoid inducers. Both the saturation threshold, around 0.02 mg ml±1, and the maximum induced activity were identical for the two ¯avonoids. The same effect was mediated by the presence of an indigenous copy of nodD borne on the presumed symbiotic megaplasmid of 128C30 in the strain MBR7. The response to hesperetin seemed to be even higher, in spite of a lower copy number
Effect of ¯avonoids in pea 1741
Fig. 5. Roots of pea plants grown submerged in nutrient solution at early stages of nodulation. (A) Uninoculated control root 11 d after germination. (B, C) Roots inoculated with Rhizobium leguminosarum bv. viciae, 14 d after germination and 7 d after inoculation (DAI). (D) Inoculated root 10 DAI. Conical lateral root primordia (lrp) and round-shaped nodule primordia (np) can be discriminated in the taproots. Stained as described in the text. Magni®cation 33.9.
of symbiotic megaplasmid. This suggests the possibility of slight differences in the NodD speci®city. Curiously, the simultaneous presence of both forms of nodD in one strain consistently increased the maximum response to the ¯avonoid inducers by 40% (not shown), probably due to the NodD monomer interactions (Kondorosi, 1992). Although no signi®cant inducer activity of quercetin was detected, this compound ef®ciently suppressed the nod gene activity induced by 0.02 mg ml±1 of naringenin or hesperetin. However, suppression required two orders of magnitude higher concentration of the suppressor than the inducer to be detected. The inhibitory action of all three tested ¯avonoids on the rhizobial growth is shown in Fig. 3c. Since the rhizobia multiplied two or three times during the incubation of microplates at 28 °C in the nod gene-inducing activity assay, the optical density of the suspension after incubation can indicate adverse effects of the compound tested on growth and metabolism in general. Surprisingly, quercetin was the least deleterious for bacterial growth, implying
that the observed nod gene-suppressing activity was not an artefact due to non-speci®c action on bacteria. Effect of ¯avonoids on plant traits
The effect of exogenous ¯avonoids in planta was clear when quanti®ed by the effect on the early symbiotic response RHC. Inoculum of 128C30 combined with naringenin and, surprisingly, quercetin as well, elicited a more intensive RHC response than an uninduced culture (Fig. 4a) in a concentration-dependent manner. However, the maximum of RHC was shifted to higher ¯avonoid concentrations, by two orders of magnitude, in comparison with nod gene induction. On the other hand, the marked effects of exogenous ¯avonoids on bacterial nod gene activity were only weakly re¯ected in nodule number. Nevertheless, a 14 d application of naringenin on roots caused a signi®cant increase in nodule number at concentrations 0.1 and 1 mg ml±1 (Fig. 4b). The second inducer tested, hesperetin, failed to increase nodule number signi®cantly, while the third
1742 NovaÂk et al.
observed closer to the point of inoculation, as shown for naringenin in Fig. 6. A similar promotion of the inhibitory effect at earlier stages was observed for quercetin (not shown). Discussion
Fig. 6. Time-course of nodule initiation in pea by Rhizobium leguminosarum bv. viciae 128C30 as affected by naringenin. Number of developing nodules was revealed by staining (7 d and 10 d after inoculation, DAI) or mature nodules counted directly (14 DAI). The presented data are means of ®ve replicates. Tukey con®dence intervals at a 0.05 signi®cance level were calculated for each sampling time. The data points without a shared letter in the index are signi®cantly different.
compound tested, the nod gene suppressor quercetin, decreased the number of nodules formed. Inhibition of nodule formation was also observed by hesperetin and quercetin at the highest concentrations tested. Increased nodule number by naringenin was not followed by an increase in TNA (Fig. 4c). However, reduction in nodule formation always led to lower TNA. The dry mass accumulation in the root system did not reveal any signi®cant deleterious effects of ¯avonoids on symbiotic root development except for hesperetin (Fig. 4d). The drop in the pea root mass at 10 mg ml±1 probably re¯ects its phytotoxicity or toxicity of its degradation products. This notion is in accord with the earlier onset of TNA drop in hesperetin than in naringenin with increasing ¯avonoid concentration, as observed in Fig. 4c. Flavonoids exerted no effects on shoot dry mass and number of leaves (not shown). Effect of ¯avonoids on nodule initiation The staining technique used allowed nodule initiation to be monitored starting from 4 DAI (Fig. 5). By contrast to the report of Niehaus and PuÈhler (1988), the bacterial infection threads did not stain suf®ciently for their microscopic enumeration. On the other hand, the RHC was preserved well in the ®xed and stained root specimens. When the ¯avonoid effects on nodule number were monitored over time, more pronounced action was
An optimized hydroponic system and the adapted root staining technique allowed changes induced in the rhizobial symbiosis development by the application of exogenous ¯avonoid compounds possessing regulatory activity with respect to rhizobial nodulation genes to be determined. In a system including pea and R. leguminosarum bv. viciae, a nod gene inducer, naringenin, was able to stimulate nodulation, especially during the early stages, while the nod gene suppressor, quercetin, inhibited it. Their effect on the symbiosis corresponded to the behaviour expected from their interaction with nod genes of the strain of Rhizobium used. The deleterious effect of hesperetin in the in planta assay could be ascribed to its phytotoxicity. This assumption is consistent with the reported role of its glycoside hesperidin as an allelopathic agent (Inderjit and Dakshini, 1991). It has been established that rhizobial strain 128C30 responded to the chosen ¯avonoids at the nod gene level like a representative strain of this biovar in spite of the difference in the chromosomal background and in spite of the ¯avonoid perception by a NodD receptor coded for on its indigenous symbiotic megaplasmid. This ®nding is not self-evident as a thorough survey of compounds for inducing or inhibitory activity in nod genes of B. japonicum revealed great interstrain variability for the spectrum of recognized ¯avonoids (Cunningham et al., 1991). Nevertheless, the effect of the ¯avonoids applied in planta was less pronounced than nod gene activation, especially at the later stages of symbiosis. While the effect of exogenous ¯avonoids could be easily distinguished at the level of root hair curling or nodule initiation, at the level of number of mature nodules the effect was barely detectable and disappeared at the level of nodule function such as nitrogen ®xation. This diminishing of ¯avonoid effect on nodule development is consistent with the existence of endogenous regulatory feedback mechanisms controlling nodule number and, presumably, nodule function in legumes as well (Rolfe and Gresshoff, 1988). Regulatory feedback could also explain the observed stimulation of RHC by both the inducer naringenin and the suppressor quercetin. The maximum RHC response for naringenin is shifted to higher concentrations than maximum nodule induction, corresponding to the range of nodulation inhibitory concentrations. Therefore, the feedback response increasing the sensitivity of the root hairs to Nod factors might be responsible for RHC stimulation
Effect of ¯avonoids in pea 1743
rather than the increased production of Nod factors by bacteria. Since the stimulatory effect was clearly observed in the hydroponic system used, it is assumed that the ¯avonoids naturally released from the pea root did not fully saturate the chosen volume of nutrient solution. This notion is consistent with previous direct measurements of nod geneinducing activity in nutrient solution (NovaÂk et al., 1994). In 1-week-old uninoculated pea seedlings, the nutrient solution containing 1 d exudate induced approximately 3% of the maximum nod gene activity in the indicator strain RBL5280, when mixed 1:1 with bacterial suspension (NovaÂk et al., 1994). Zaat et al. (1988) observed 100% of inducing activity only after a 4 d incubation in the same volume of nutrient solution per root. However, upon inoculation the nod gene-inducing activity sharply increases, up to 80±85% of the maximum activity, in a 1 d exudate (unpublished data). This implies that increases in the volume of nutrient solution per root should lead to more pronounced stimulation of nodulation by the addition of ¯avonoids. Consistently, decreases in nodulation ef®ciency in a larger volume (100 ml) of nutrient solution (Fig. 2) would correspond to the dilution of nod gene inducers from the root exudate and Nod factors formed by bacteria. These results are consistent with the observation of Kapulnik et al. (1987) who reported a stimulatory effect of the nod gene inducer, luteolin, on nodulation and even N2®xation in alfalfa±R. meliloti in hydroponic culture. Similarly, Zhang and Smith (1995) found that preincubation of Bradyrhizobium japonicum with a nod gene inducer, genistein, accelerated the development of nodules at limiting low temperatures. The addition of genistein also helped to overcome another physiological block in nodulation, nitrate inhibition of ¯avonoid production (Pan and Smith, 2000). A positive effect on nodulation in the alfalfa±R. meliloti system (Jain et al., 1990) was also reported for naringenin, which otherwise acts as a suppressor of nod genes of a standard strain of this bacterium (Peters and Long, 1988; NovaÂk et al., 1995). In pea, a rhizospheric application of naringenin partially alleviated the deleterious effect of low temperature on nodulation status and nodule ef®ciency (Ahlawat et al., 1998). A conclusion can be drawn that the stimulation of nodulation and nitrogenase activity by ¯avonoids can occur only under conditions where the accessibility of ¯avonoid inducers is limited by cultivation factors, such as during hydroponics where the root exudates are either diluted or quickly degraded. Alternatively, a stimulatory effect could be observed in a case of genetic de®ciency of the host plant. The role of plant genetic factors governing ¯avonoid production on nodulation has been con®rmed by Firmin et al. (1986) who found a negligible nod gene-inducing activity in the root exudate of a white ¯owering mutant of
Antirrhinum, and by Hungria and Phillips (1993) who found less release of inducers from bean seeds with light testa pigmentation. Also Kapulnik et al. (1987) observed a correlation between luteolin synthesis in roots of alfalfa lines and nodulation in breeding experiments. In a related group of plant-associated bacteria from the genus Agrobacterium, a stimulation of the agrobacterial infection, i.e. transformation by a tumorigenic plasmid, has been observed after activation of vir genes by exogenous phenolic inducers (Owens and Smigocki, 1988; Joubert et al., 1995). vir genes play a crucial role in interactions of agrobacteria with a plant host in a similar way to nod genes in rhizobia. The approach used to achieve overexpression of nod genes in the current work is complementary to the genetic approaches, where the ef®ciency of strains with deregulated expression of nod genes is followed (Knight et al., 1986; Burn et al., 1987; Spaink et al., 1989). Generally, treatment of rhizobia with excessive ¯avonoid inducers is supposed to create phenocopies of the genetically deregulated strains. However, such a parallel is valid only for stages preceding the infection of the root tissue, as inside the tissue a stable internal milieu can be assumed. Nevertheless, contrary to the expected improvement in nodulation, constitutive expression of nod genes led to a reduced nodule number and failure in nitrogen ®xation (Burn et al., 1987), while an increased dose of nod genes prevented nodulation (Knight et al., 1986). Therefore, it seems that Nod-factor overproduction had no positive effect at the infection stage in these cases. It is also conceivable that the inability of the deregulated strains to switch off the nod gene activity during nodule development (Schlaman et al., 1991) caused damage to nodules and their function. Surprisingly, the hybrid allele nodD604, consisting of recombined nodD of R. meliloti and R. trifolii and conditioning ¯avonoid-independent nod gene expression in different rhizobia, not only extended the host range but also positively affected nitrogen ®xation (Spaink et al., 1989). However, the gain in nitrogenase activity was not mediated by an increase in nodule number in this case. The inhibitory effect of the highest tested concentration of naringenin in these experiments might be caused by the same mechanism as the failure of the deregulated strains, provided that naringenin present at the very high external concentration penetrates the plant tissue and affects the nod genes at the later stage of symbiosis (Knight et al., 1986; Burn et al., 1987). In addition to compensating feedback responses, low activity of inducers in planta may be partially ascribed to ¯avonoid degradation since a considerable rate of ¯avonoid decay was recorded in the hydroponic system used. However, the residual concentration of naringenin before the change of the nutrient solution at the two highest
1744 NovaÂk et al.
concentrations was obviously still above the concentration found saturating for nod gene induction. Rhizobium rather than the plant root role in the degradation of added ¯avonoids is consistent with the ¯avonoid-degrading activity described in the fast-growing rhizobia (Cooper and Rao, 1995). Unfortunately, in soil, substantially higher losses of ¯avonoids must be expected due to sorption and degradative processes than in a semi-sterile hydroponic system. Both processes eliminated ¯avonoid action in soybean±B. japonicum system when soil, commercial potting mixtures, and vermiculite were used as a growing substrate (Cunningham et al., 1991). Therefore, it seems that, for the above two reasons, the direct application of ¯avonoids can be used for the intensi®cation of nodulation of legumes in agriculture only in speci®ed cases, although the approach is generally considered as realistic (LubrizolGenetics, 1987, 1993). A practical solution might be deposition of ¯avonoids in a bound form from which they would be released continually, for example, in a form of glycosides which probably persist longer in the rhizosphere soil (LeoÂn-Barrios et al., 1993). The application of ¯avonoid analogues recalcitrant to biodegradation, which would retain biological activity, is an approach already exploited in other agricultural preparations (Cunningham et al., 1991). An alternative and, probably, a more sensitive technique for monitoring the whole plant effect of exogenous activation of nod genes has been used by Kosslak et al. (1990) and Cunningham et al. (1991). The technique is based on the competition of two strains for nodule occupancy, where a strain more sensitive to the exogenous nod gene inducer or resistant to a suppressor possesses an advantage. Higher or earlier nod gene activation will allow it to colonize more nodules, as can be subsequently detected. Reported pronounced changes in nodule occupancy upon ¯avonoid application (Cunningham et al., 1991) indicate that ¯avonoid preparations are more suitable for the manipulation of competitiveness of a desirable strain than for the improvement of the nodulation intensity as such. Acknowledgements The authors are indebted to Dr R Okker, Leiden University, The Netherlands, for supplying the strain RBL5280 and to Dr M WeiserovaÂ for supplying the plasmid RK2013. The authors thank Dr M Gryndler for the help with the establishment of the staining procedure and Mr P BrÏichaÂcÏek for taking the photographs. The work was supported by the grant 521/00/0937 of the Grant Agency of the Czech Republic and by the institutional research concept No. A V0Z5020903.
References Ahlawat A, Jain V, Nainawatee HS. 1998. Effect of low
temperature and rhizospheric application of naringenin on pea± Rhizobium leguminosarum biovar viciae symbiosis. Journal of Plant Biochemistry and Biotechnology 7, 35±38. Bauer WD, Caetano-AnolleÂs G. 1990. Chemotaxis, induced gene expression and competitiveness in the rhizosphere. Plant and Soil 129, 45±52. Burn J, Rossen L, Johnston AWB. 1987. Four classes of mutations in the nodD gene of Rhizobium leguminosarum biovar. viciae that affect its ability to autoregulate and/or activate other nod genes in the presence of ¯avonoid inducers. Genes and Development 1, 456±464. Cooper JE, Rao JR. 1995. Flavonoid metabolism by rhizobia: mechanisms and products. Symbiosis 19, 91±98. Cunningham S, Kollmeyer WD, Stacey G. 1991. Chemical control of interstrain competition for soybean nodulation by Bradyrhizobium japonicum. Applied and Environmental Microbiology 57, 1886±1892. Figurski DH, Helinski DR. 1979. Replication of an origincontaining derivative of plasmid RK2 dependent on a plasmid function provided in trans. Proceedings of the National Academy of Sciences, USA 76, 1648±1652. Firmin JL, Wilson KE, Rossen L, Johnston AWB. 1986. Flavonoid activation of nodulation genes in Rhizobium reversed by other compounds present in plants. Nature 324, 90±92. Hungria M, Phillips DA. 1993. Effects of a seed color mutation on rhizobial nod-gene-inducing ¯avonoids and nodulation in common bean. Molecular Plant±Microbe Interactions 6, 418± 422. Inderjit, Dakshini KMM. 1991. Hesperetin 7-rutinoside (hesperidin) and taxifolin 3-arabinoside as germination and growth inhibitors in soils associated with the weed, Pluchea lanceolata (DC) CB Clarke (Asteraceae). Journal of Chemical Ecology 17, 1585±1591. Inoue H, Nojima H, Okayama H. 1990. High ef®ciency transformation of Escherichia coli with plasmids. Gene 96, 23± 28. Jain V, Garg N, Nainawatee HS. 1990. Naringenin enhanced ef®ciency of Rhizobium meliloti±alfalfa symbiosis. World Journal of Microbiology and Biotechnology 6, 434±436. Josey DP, Beynon JL, Johnston AWB, Beringer JE. 1979. Strain identi®cation in Rhizobium using intrinsic antibiotic resistance. Journal of Applied Bacteriology 46, 343±350. Joubert P, Sangwan RS, Aouad MEA, BeaupeÁre D, SangwanNorreel BS. 1995. In¯uence of phenolic compounds on Agrobacterium vir gene induction and onion gene transfer. Phytochemistry 40, 1623±1628. Kapulnik Y, Joseph CM, Phillips DA. 1987. Flavone limitations to root nodulation and symbiotic nitrogen ®xation in alfalfa. Plant Physiology 84, 1193±1196. Knight CD, Rossen L, Robertson JG, Wells B, Downie JA. 1986. Nodulation inhibition by Rhizobium leguminosarum multicopy nodABC genes and analysis of early stages of infection. Journal of Bacteriology 166, 552±558. Kondorosi A. 1992. Regulation of nodulation genes in rhizobia. In: Verma DPS, ed. Molecular signals in plant±microbe communication. Boca Raton: CRC Press, 325±340. Kosslak RM, Joshi RS, Bowen BA, Paaren HE, Appelbaum ER. 1990. Strain-speci®c inhibition of nod gene induction in Bradyrhizobium japonicum by ¯avonoid compounds. Applied and Environmental Microbiology 56, 1333±1341. LeoÂn-Barrios M, Dakora FD, Joseph CM, Phillips DA. 1993. Isolation of Rhizobium meliloti nod gene inducers from alfalfa rhizosphere soil. Applied and Environmental Microbiology 59, 636±639. Leyva A, Palacios JM, Ruiz-ArguÈeso T. 1987. Conserved plasmid hydrogen-uptake (hup)-speci®c sequences within Hup+
Effect of ¯avonoids in pea 1745 Rhizobium leguminosarum strains. Applied and Environmental Microbiology 53, 2539±2543. Lie TA. 1969a. Non-photosynthetic effects of red and far-red light on root-nodule formation by leguminous plants. Plant and Soil 30, 391±404. Lie TA. 1969b. The effect of low pH on different phases of nodule formation in pea plants. Plant and Soil 31, 391±406. Lie TA. 1974. Environmental effects on nodulation and symbiotic nitrogen ®xation. In: Quispel A, ed. The biology of nitrogen ®xation. Amsterdam, Oxford, New York: North-Holland/ American Elsevier, 555±582. Lubrizol-Genetics. 1987. Nodulation inducing composition for legume e.g. clover, alfalfa or pea. European patent, EP-245931. Lubrizol-Genetics. 1993. New soybean inoculating composition. US patent, US5229113. Niehaus K, PuÈhler A. 1988. Light microscopic analysis of entire alfalfa (Medicago sativa) nodules by a fast staining and clearing method. Endocytobiosis and Cell Research 5, 59±68. NovaÂk K, KropaÂcÏovaÂ M, HavlõÂcÏek V, SÏkrdleta V. 1995. Iso¯avonoid phytoalexin pisatin is not recognized by the ¯avonoid receptor NodD of Rhizobium leguminosarum bv. viciae. Folia Microbiologica 40, 535±540. NovaÂk K, SÏkrdleta V, NeÏmcovaÂ M, LisaÂ L. 1994. Optimization of rhizobial nod gene-inducing activity assay in pea root exudate. Folia Microbiologica 39, 208±214. Owens LD, Smigocki AC. 1988. Transformation of soybean cells using mixed strains of Agrobacterium tumefaciens and phenolic compounds. Plant Physiology 88, 570±573. Pan B, Smith DL. 2000. Preincubation of B. japonicum cells with genistein reduces the inhibitory effects of mineral nitrogen on soybean nodulation and nitrogen ®xation under ®eld conditions. Plant and Soil 223, 235±242. Peters NK, Long SR. 1988. Alfalfa root exudates and compounds which promote or inhibit induction of Rhizobium meliloti nodulation genes. Plant Physiology 88, 396±400. Recourt K, Schripsema J, Kijne JW, Van Brussel AAN, Lugtenberg BJJ. 1991. Inoculation of Vicia sativa subsp. nigra roots with Rhizobium leguminosarum biovar viciae results in release of nod gene activating ¯avanones and chalcones. Plant Molecular Biology 16, 841±852. Richardson AE, Simpson RJ, Djordjevic MA, Rolfe BG. 1988. Expression of nodulation genes in Rhizobium leguminosarum biovar trifolii is affected by low pH and by Ca and Al ions. Applied and Environmental Microbiology 54, 2541±2548. Rolfe BG, Gresshoff PM. 1988. Genetic analysis of legume nodule initiation. Annual Review of Plant Physiology and Plant Molecular Biology 39, 297±319.
Sambrook J, Fritsch EF, Maniatis T. 1989. Molecular cloning. A laboratory manual, 2nd edn. New York: Cold Spring Harbor Laboratory Press. Schlaman HRM, Horvath B, Vijgenboom E, Okker RJH, Lugtenberg BJJ. 1991. Suppression of nodulation gene expression in bacteroids of Rhizobium leguminosarum biovar viciae. Journal of Bacteriology 173, 4277±4287. SÏkrdleta V, GaudinovaÂ A, NeÏmcovaÂ M, HyndraÂkovaÂ A. 1980. Symbiotic dinitrogen ®xation as affected by short-term application of nitrate to nodulated Pisum sativum L. Folia Microbiologica 25, 155±161. SÏkrdleta V, LisaÂ L, NeÏmcovaÂ M. 1987. Comparison of peas nodulated with a hydrogen-uptake positive or negative strain of Rhizobium leguminosarum. I. Nodulation, acetylene-reducing, dihydrogen and carbon-dioxide evolving activities. Folia Microbiologica 32, 226±233. SÏkrdleta V, NeÏmcovaÂ M, LisaÂ L, NovaÂk K. 1993. Characterization of indigenous soil populations of Rhizobium leguminosarum bv. viceae using plant symbiotic traits. Soil Biology and Biochemistry 25, 63±67. Spaink HP. 2000. Root nodulation and infection factors produced by rhizobial bacteria. Annual Review of Microbiology 54, 257± 288. Spaink HP, Okker RJH, Wijffelman CA, Pees E, Lugtenberg BJJ. 1987a. Promoters in the nodulation region of the Rhizobium leguminosarum Sym plasmid pRL1JI. Plant Molecular Biology 9, 27±39. Spaink HP, Wijffelman CA, Okker RJH, Lugtenberg BEJ. 1989. Localization of functional regions of the Rhizobium nodD product using hybrid nodD genes. Plant Molecular Biology 12, 59±73. Spaink HP, Wijffelman CA, Pees E, Okker RJH, Lugtenberg BJJ. 1987b. Rhizobium nodulation gene nodD as a determinant of host speci®city. Nature 328, 337±340. Streeter J. 1988. Inhibition of legume nodule formation and N2 ®xation by nitrate. CRC Critical Reviews in Plant Sciences 7, 1± 23. Vincent JM. 1970. A manual for the practical study of the root nodule bacteria. Oxford: Blackwell Scienti®c Publishers. Zaat SAJ, Wijffelman CA, Mulders IHM, Van Brussel AAN, Lugtenberg BJJ. 1988. Root exudates of various host plants of Rhizobium leguminosarum contain different sets of inducers of Rhizobium nodulation genes. Plant Physiology 86, 1298±1303. Zhang F, Smith DL. 1995. Preincubation of Bradyrhizobium japonicum with genistein accelerates nodule development of soybean at suboptimal root zone temperatures. Plant Physiology 108, 961±968.