Developmental downregulation of rhizobial ... - Wiley Online Library

Research

Developmental downregulation of rhizobial genes as a function of symbiosome differentiation in symbiotic root nodules of Pisum sativum Blackwell Publishing Ltd.

V. E. Tsyganov1, V. A. Voroshilova1, J. A. Herrera-Cervera2, J. M. Sanjuan-Pinilla2, A. Y. Borisov1, I. A. Tikhonovich1, U. B. Priefer3, J. Olivares2 and J. Sanjuan2 1

All-Russia Research Institute for Agricultural Microbiology, Podbelsky chaussee 3, Saint-Petersburg, Pushkin 8, 196608, Russia; 2Departamento de

Microbiologia, Estacion Experimental del Zaidin-CSIC, Profesor Albareda 1, E-18008 Granada, Spain; 3Ökologie des Bodens, RWTH Aachen, Worringer Weg 1, 52056 Aachen, Germany

Summary Author for correspondence: A. Y. Borisov Tel: +7 812 470 43 92 Fax: +7 812 470 43 62 Email: [email protected] Received: 8 December 2002 Accepted: 11 April 2003 doi: 10.1046/j.1469-8137.2003.00823.x

• The expression of nodA and dctA genes of Rhizobium leguminosarum bv. viciae has been studied in mutant nodules of pea (Pisum sativum L.), blocked at the following developmental stages: infection thread development inside the nodule (Itn); infection droplet differentiation (Idd); bacteroid differentiation after endocytosis (Bad); and nodule persistence (Nop). • With the use of reporter fusions to these symbiotic bacterial genes it was shown that both nodA and dctA were expressed at all developmental stages, with a pattern similar to that of constitutive, symbiosis-unrelated genes. • As well as two constitutively expressed genes, both nodA and dctA genes seemed to be subjected to gradual downregulation in nodule bacteria, correlating with the stage of bacteroid differentiation reached. No such effect was observed for the symbiotic, oxygen-regulated fixN gene. The bacteroid development stage also appeared to be related to the ability of bacteria that have been subjected to endocytosis to resume free-living vegetative growth. • The results support the suggestion that bacteroid differentiation into a nitrogenfixing, organelle-like form, is a gradual process involving several stages, each controlled by different plant genes. Key words: Plant–microbe interactions, Legume–Rhizobium symbiosis, Pisum sativum, Rhizobium leguminosarum bv. viciae, Pea symbiotic mutants, sym genes, nod genes, dct genes. © New Phytologist (2003) 159: 521–530

Introduction Nodule bacteria belonging to the family Rhizobiaceae are capable of establishing symbiotic interactions with leguminous plants (Fabaceae) that lead to the formation of new plant organs, the nitrogen-fixing nodules. The process of legume nodule initiation and early development is triggered by chemical signals (lipo-chitooligosaccharides), synthesized and secreted by proteins encoded by bacterial nodulation (nod ) genes. Expression of nod genes is specifically induced by flavonoid compounds secreted by the plant root (reviewed in

© New Phytologist (2003) 159: 521–530 www.newphytologist.com

Dénarié et al., 1996; Long, 1996; Schultze & Kondorosi, 1996; Spaink, 1996; Downie & Walker, 1999). During the development of nodules in alfalfa (Medicago sativa L.) and pea (Pisum sativum L.), bacteria undergo profound metabolic and morphologically pronounced differentiation into the symbiotic, organelle-like forms, termed bacteroids (Vasse et al., 1990; Brewin et al., 1993). In the nodule, the bacteroids provide the plant with fixed atmospheric nitrogen and, in turn, the plant provides bacteroids with carbon and energy sources (Brewin et al., 1993; Mylona et al., 1995).

521

522 Research

In indeterminate nodules, the transition from juvenile to mature nitrogen-fixing bacteroids is a gradual process, where several bacteroid types (types 1–5, according to Vasse et al., 1990) can be morphologically recognized during nodule development. Perhaps the most significant characteristic of nitrogen-fixing bacteroids is that they appear to be terminally differentiated, with a loss of the capacity for cell division and an inability to resume free-living vegetative growth. It remains to be determined, however, at which stage during bacteroid differentiation these capacities are lost. Together with plant cells containing them, bacteroids are destroyed by the plant after fulfilling their mission. It is assumed that much of the increase in protein synthesis in bacteroids comes from the induction of genes that are necessary for nitrogen fixation and bacteroid metabolism. In conflict, expression of many genes unrelated to nitrogenfixation declines (Perret et al., 1999; Margolin, 2000). Furthermore, there seems to be an active repression of certain early symbiotic genes (e.g. flavonoid-inducible nodulation genes), which appear unimportant at the later stages of the symbiosis. It has, for example, been shown that rhizobial nod-genes are repressed in bacteroids but remain induced in bacteria in infection threads (Sharma & Singer, 1990; Schlaman et al., 1991, 1992; Marie et al., 1994). On the other hand, a more recent report has shown that Nod-factors are still produced by differentiated bacteroids at a low but detectable level, suggesting some unknown function for Nod-factors at the late stages of nodule development (Timmers et al., 1998). However, it is unknown when or how these changes in gene expression take place during the development of nitrogen-fixing bacteroids. Previous morphological or biochemical studies have shown how plant mutations can affect the development of bacteroids. Nevertheless, there are few reports on the effects of plant mutations on the expression of bacterial genes during bacteroid differentiation or on the ability of bacteroids to resume freeliving growth.

In a previous work the expression patterns of oxygenregulated late symbiotic genes of R. leguminosarum bv. viciae, namely nifA, fnrN and fixN, were studied at different nodule developmental stages by using transcriptional reporter fusions and a panel of pea symbiotic mutants. The endocytosis of bacteria into host-plant cell cytoplasm was shown to be a prerequisite for induction of these genes (Voroshilova et al., 2001). The aim of this study was to analyse the expression of the symbiosis-related genes, nodA and dctA, in R. leguminosarum bv. viciae. In particular, the investigation was aimed at how this gene expression is affected by the arrest of differentiation of symbiotic compartments at specific stages of pea nodule development. As a control, nonsymbiotically regulated genes, and also the oxygen-regulated fixN gene, were also studied.

Materials and Methods Plant material Fix− mutants of pea (Pisum sativum L.), blocked at different stages in late nodule development, and their parental lines were used in this study (Table 1). The order of late developmental stages, as revealed by genetic dissection of nodule morphogenesis in symbiotic roots of peas, is: Itn, Idd, Bad, Nop (Borisov et al., 2000). Bacterial strains R. leguminosarum bv. viciae strain VF39 is a wild-type strain that is Nod+Fix+ on pea (Priefer, 1989). Plasmid pGD499 carries an npt-lacZ fusion (Ditta et al., 1985), pMP240 carries a nodA-lacZ fusion (Spaink et al., 1987) and pCR12 is an IncP (TcR) plasmid carrying a dctA-lacZ fusion (Ronson et al., 1987a, 1987b). The plasmids were introduced into VF39 by triparental matings using pRK2013 as helper (Ditta et al., 1980). For constitutive expression of gusA, we used a

Table 1 Plant material used in the study Lines

Phenotype

References

SGE SGEFix−-1 (sym40)

Wild-type line Abnormal (hypertrophied) infection droplet formation (Idd–), leaky phenotype No endocytosis of bacteria (Itn–), leaky phenotype No endocytosis of bacteria (Itn–) Wild-type line Premature degradation of symbiotic compartments (Nop–) Wild-type line No bacteroid differentiation (Bad–) No bacteroid differentiation (Bad–) Wild-type line No endocytosis of bacteria (Itn–)* No bacteroid differentiation (Bad–)*

Kosterin and Rozov (1993) Tsyganov et al. (1994, 1998)

SGEFix−-2 (sym33) RBT3 (sym33, sym40) Sparkle E135f (sym13) Sprint-2 Sprint-2Fix− (sym31) RBT (sym13, sym31) Finale RisFixU (sym33) RisFixL (sym32)

Tsyganov et al. (1994, 1998) Borisov et al. (1997b) Kneen et al. (1990) Kneen et al. (1990) Berdnikov et al. (1989) Borisov et al. (1992, 1997a) Borisov et al. (1997a) Engvild (1987) Engvild (1987); Voroshilova et al. (2001) Morzhina et al. (2000)

*Gene symbols for the genes identified in these mutant lines were assigned after G. Duc and M. Sagan (personal communication).

www.newphytologist.com © New Phytologist (2003) 159: 521–530

Research

derivative of VF39 that contained in the genome a Tn5-gusA insertion that was verified to be constitutively expressed (Voroshilova et al., 2001). The fixN-gusA fusion strain carried a gusA cassette fused to the 5′ part of fixN and integrated into the fixNOQP operon on plasmid ‘c’ (Schlüter et al., 1997). To prepare inocula, rhizobial strains were grown on TY agar medium (Beringer, 1974) at 28–30°C with the required antibiotics, tetracycline (Tc), 10 mg l−1, streptomycin (Sm), 100 mg l−1, gentamycin (Gm), 40 mg l−1 and kanamycin (Km), 100 mg l−1. Plant growth conditions and inoculation Plants were grown in growth chambers (Koxka, Navarra, Spain) at a cycle of 16-h day and 8-h night, a temperature of 21/19°C, a relative humidity of 75% and a photon irradiance of 490 µmol m−2 s−1). Sterile vermiculite was used as substrate. Seeds were surface sterilized with concentrated sulphuric acid for 30 min at room temperature and inoculated at the time of sowing with 1 ml of an aqueous suspension of bacteria (108−109 cells ml−1). Plants were watered with a nitrogen-free nutrient solution, as described previously (Borisov et al., 1997a). Histochemical staining, microscopy and photography Nodules from mutant and wild-type plants were harvested 4 weeks after inoculation. The nodules were sliced into 70 µm sections by means of a vibratom Leica VT 1000S (Leica Microsystems Wetzlar GmbH, Wetzlar, Germany). Sections were stained as described in Boesten et al. (1998). Staining was for 1 h except for nodules from the mutants Sprint-2Fix− (sym31) and RBT (sym13, sym31), in which the staining time for dctA-lacZ fusions was 30 min. For light microscopy, sections were transferred onto glass slides. Microphotographs were taken using an Olympus Om-4 camera mounted onto an Olympus BX50 light microscope (Olympus Optical Co., Europa GmbH, Hamburg Germany). The pattern of gene expression was studied in nodA-lacZ and dctA-lacZ fusions as well as in the constitutively expressed fusion npt-lacZ. Quantitative analysis of expression of the transcriptional reporter fusions lacZ and gusA VF39 derivatives carrying the different reporter fusions were inoculated on pea wild-type and mutant lines as above. For each experiment, three to four pots (five plants per pot) were prepared with each combination of plant genotype and bacterial strain. Nodules (usually > 200) were collected 4 weeks after inoculation from plants grown in the same pot, pooled and crushed in 1 ml ice-cold buffer containing 0.25  mannitol, 0.05  Tris-HCl, pH 7.8 and 100 mg polyvinylpolypyrrolidone (PVPP). To remove insoluble material and nodule cell debris, the homogenate was centrifuged at 119 g for 1 min at 4°C. The supernatant was carefully transferred to


a new tube and centrifuged at 4300 g for 5 min at 4°C. After removing the supernatant, the pellet (containing bacteria and bacteroids) was washed and resuspended in 1 ml of buffer and maintained on ice until use. The specific activity of β-galactosidase was determined in each sample following Miller (1972) and expressed as nmol ο-nitrophenol produced min−1 mg protein−1. β-glucuronidase assays were performed as described previously (Wilson et al., 1992) using 4-nitrophenyl β--glucuronide (pNPG) as the substrate. β-glucuronidase specific activity was expressed as nmol 4-nitrophenol produced min−1 mg protein−1. Total protein in samples was determined by the method of Bradford (1976) using Bio-Rad protein assay reagents. Nodule bacterial cells that were able to resume free-living growth were determined by counting the number of colony forming units after plating serial dilutions on TYGagar plates followed by incubation at 28–30°C for 3–4 d. The stability of plasmid-borne reporter fusions was determined by quantifying the number of colony forming units that maintained the antibiotic resistance encoded by the plasmids. Quantitative data were confirmed in three independent experiments. For quantitative analyses, the fusions nodA-lacZ, dctA-lacZ and fixNc-gusA were used, along with the constitutively expressed fusions npt-lacZ and Tn5-gusA. Statistical analysis The data were subjected to standard methods of variance analyses, the comparison of two groups of the data and Pearson correlation. SigmaStat for Windows, version 2.3 (SPSS Inc, Chicago, IL, USA) was used for statistical analysis.

Results and Discussion Patterns of expression of nodA and dctA genes in pea nodules Expression of the constitutive npt-lacZ fusion was observed at all nodule developmental stages (Fig. 1a–e). Expression of the reporter dctA-lacZ was similar to that of nodA-lacZ and also to that of the symbiosis-unrelated fusion npt-lacZ. Therefore only the results obtained for dctA are presented in the Fig. 1f–j, along with, for comparison, the pattern of expression of the constitutive npt-lacZ fusion (Fig. 1a–e). In mutants SGEFix−-2 (sym33), RisFixU (sym33) and RBT3 (sym33, sym40), which are blocked at the stage of endocytosis of bacteria (Itn–, Table 1), the nodA and dctA fusions were clearly expressed in the infection threads in the same manner as npt (Fig. 1a,f ). This result was consistent with earlier reports on nodA gene expression (Schlaman et al., 1991), but was unexpected in the case of the dctA gene. In alfalfa nodules, expression of Sinorhizobium meliloti dctA was not observed in infection threads (Boesten et al., 1998). The contradiction can be explained in two ways. Firstly, mutants blocked at the

523

524 Research


Research

stage Itn were used in this study and these mutants have slightly ‘wider’ infection threads than those in the wild-type lines (Tsyganov et al., 1998), thereby facilitating the visualization of staining. Secondly, regulation of the dct-gene system may differ between R. leguminosarum bv. viciae and S. meliloti. In mutant SGEFix−-1 (sym40), which is blocked at nodule developmental stage Idd (Table 1), expression of nodA and dctA (Fig. 1g) was observed both in plant nodule cells containing hypertrophied infection droplets (cells of distal part of the nodule) and in endocytosed bacteria (cells of proximal part of the nodule). Again the expression pattern of these genes did not differ from that of the constitutive npt-lacZ fusion (Fig. 1b). In mutant lines Sprint-2Fix− (sym31) and RBT (sym13, sym31), which are blocked in bacteroid differentiation after endocytosis of bacteria by plant cells (Bad– phenotype), expression of nodA and dctA was observed in nodule histological zones II and III (Vasse et al., 1990), with a pattern similar to that of the constitutive fusion (Fig. 1c,h). However, staining of nodule sections from mutant lines Sprint-2Fix− (sym31) and RBT (sym13, sym31), inoculated with rhizobial strains containing dctA reporter fusions, required significantly less time than other lines to achieve similar intensity. Mutant line E135f (sym13) is blocked at nodule developmental stage Nop (nodule persistence) and characterized by a premature appearance of nodule histological zone IV, the zone of degradation of symbiotic compartments. In this line, expression of nodA and dctA was observed in nodule histological zones II, III and in some cells of zone IV (Fig. 1d). A similar pattern was seen for the constitutively expressed lacZ fusion (Fig. 1i). The pattern of expression of nodA and dctA fusions in nodules of the wild-type lines Sprint-2, SGE, Sparkle and Finale was also found to be similar to that characteristic of constitutive genes: the expression of all fusions was observed in histological zones II and III (Fig. 1e,j). It was surprising to observe expression of the nodA-lacZ fusion in plant cells with differentiated bacteroids (nodule histological zone III) and in the plant cells with bacteroids close to degradation (zone IV). Previous reports indicated that repression of rhizobial nod genes occurs immediately after endocytosis of bacteria into plant cell cytoplasm (Sharma & Singer, 1990; Schlaman et al., 1991, 1992). Thus we had expected a sharp decrease in staining from the middle of zone II to zone III in wild-type nodules. One of the possible explanations for our unexpected results could be that, even if repressed, the expression resulting from the nodA-lacZ plasmid in these nodule zones is still high enough to mask

such a decrease. However, constitutive fusions required the same staining time to achieve a similar staining intensity. Schlaman et al. (1991) found only background levels of nod gene expression using various techniques. These authors also demonstrated the existence of a repressor that inhibited the binding of NodD to nod promoters (Schlaman et al., 1992). However, Timmers et al. (1998) have more recently shown that Nod factors are produced by bacteria in infection threads and in the cytoplasm of invaded cells in zone II, and are still detected, although at much lower levels, in bacteroids in zone III. These facts suggest some unknown function for Nod factors at late nodule developmental stages. Thus, although nod genes are repressed in nodule zone III, nod promoters may still maintain some level of expression in differentiated bacteroids. Also in agreement with our data, the results from Timmers et al. (1998) suggest that repression of nod genes may not occur immediately after endocytosis, but may rather be a gradual event associated with the process of bacteroid differentiation. The identical nature of the expression patterns of nodA and dctA in pairs of lines Sprint-2Fix− (sym31), RBT (sym13, sym31) and SGEFix−-2 (sym33), RBT3 (sym33, sym40) demonstrated that the developmental epistasis observed in the mutant gene pairs sym31, sym13 and sym33, sym40 covers not only morphologically pronounced differentiation of nodule symbiotic compartments and bacteroids (Borisov et al., 1997a, 1997b), but also the influence of the mutations on expression of the bacterial genes nodA and dctA. Thus, the results of this qualitative study have confirmed the sequential functioning of these pairs of genes that has previously been demonstrated (Borisov et al., 1997a, 1997b; Voroshilova et al., 2001). Levels of expression of nodA and dctA genes in pea nodules Pea mutants Sprint-2Fix− (sym31), E135f (sym13) and double mutant RBT (sym13, sym31) were chosen to quantify gene expression levels. RisFixL (sym32), a mutant in which bacteroid differentiation is blocked at a stage slightly later than that controlled by sym31 (Morzhina et al., 2000) was also included. The mutants SGEFix−-2 (sym33), RisFixU (sym33) and RBT3 (sym33, sym40) were omitted from this study because no bacteria that had been subjected to endocytosis were found in their nodule cells; similarly, mutant SGEFix−-1 (sym40) was not included due to its mosaic histological zonation. Besides the above lacZ reporter fusions (dctA, nodA and npt), the levels of expression of two other reporter fusions were determined: a constitutively expressed Tn5-gusA fusion,

Fig. 1 Expression patterns of npt-lacZ and dctA-lacZ transcriptional fusions in pea mutant nodules blocked at different developmental stages, and in wild-type nodules. Similar results were obtained with nodA-lacZ fusion (see section ‘Patterns of expression of nodA and detA genes in pea nodules’ in text). (a– e) npt-IacZ fusion; (f– j) dctA-IacZ fusion; (a–f ) RisFixU, (sym33), arrowheads show infection threads; (b,g) SGEFix−-1 (sym40); (c,h) Sprint-2Fix− (sym31); (d,i) E135f (sym13); (e,j) wild-type line SGE. I, II, III, IV: nodule histological zones (marked where identifiable); Bar = 0.4 mm.


525

526 Research Table 2 Levels of expression of nodA-lacZ, dctA-lacZ and fixNc-gusA transcriptional fusions, along with constitutively expressed fusions nptlacZ and Tn5-gusA, in nodules of mutant and wild-type pea lines β-Galactosidase SA

β-Glucuronidase SA

Pea lines

npt-lacZ

nodA-lacZ

dctA-lacZ

fixNc-gusA

Tn5-gusA

Sprint-2 (wild-type) Sprint-2Fix− (sym31) RBT (sym13, sym31) Finale (wild-type) RisFixL (sym32)b Sparkle (wild-type) E135f (sym13)

116 ± 2.4 546 ± 11.8a 320 ± 16.9a 95 ± 5.8 289 ± 28.6a 97 ± 5.8 247 ± 29.0a

9.9 ± 1.08 351.3 ± 73.94a 553.8 ± 28.04a 8.5 ± 0.62 122.1 ± 36.9a 9.6 ± 0.78 30.1 ± 1.90a

183 ± 5.6 1320 ± 58.7a 1082 ± 74.8a 414 ± 32.5 1327 ± 253.1a 274 ± 10.3 372 ± 14.5a

18.3 ± 0.57 21.6 ± 1.15a 28.5 ± 5.88a 10.8 ± 0.33 7.6 ± 0.23a 17.9 ± 0.68 16.2 ± 1.86

203 ± 6.7 2298 ± 40.1a 1885 ± 155.8a 70 ± 2.6 302 ± 80.3a 160 ± 6.7 495 ± 6.8a

Specific enzyme activity is expressed in nmol min−1 mg protein−1 (± standard error). aThe mean values differ significantly (P > 0.95) from those in corresponding wild-type lines (line RBT was compared with both Sprint-2 and Sparkle). bRisFixL (sym32) showed high variability between replicate measurements.

integrated into the genome of strain VF39 and expressed in a constitutive manner, and a genomic fixN-gusA fusion, subject to oxygen-dependent regulation, which, in a previous study, showed a similar expression pattern in nodules of various mutant and wild-type plant genotypes (Voroshilova et al., 2001). As determined by the maintenance of antibiotic resistance, plasmid stability in nodule bacteria did not significantly differ among nodules from the various pea genotypes (mutant or wild-types), although there were differences among plasmids. Thus, culturable nodule bacteria maintaining the nodA-lacZ fusion plasmid ranged between 12 and 20%, whereas the dctA-lacZ fusion plasmid was more stable (60–70%). Thus, any differences in gene expression found between bacteria and bacteroids isolated from the various pea genotypes would not be caused by differences in plasmid stability. The expression of reporter gene fusions in total bacteria isolated from wild-type or mutant pea nodules was compared. Since plant mutations determine a block in bacteroid development at particular stages after endocytosis, any differences in bacterial gene expression should be attributed to the distinct bacteroid developmental stage reached in each nodule type. For a given gene-reporter fusion, there were some differences among wild-type pea lines (Table 2). For instance, βglucuronidase activity derived from the fixNc-gusA fusion was significantly higher in bacteria isolated from nodules of wild-type lines Sprint-2 and Sparkle than from Finale. This variation may be due to specific factors in the host-plant that might affect the level of bacterial gene expression (e.g. Bedmar et al., 1983) thereby enforcing the need to compare each mutant pea line exclusively with its corresponding wildtype genotype. The two nonsymbiotically regulated gene fusions (npt-lacZ and Tn5-gusA) used as controls in this study behaved in a similar fashion, despite the fact that one was plasmid-encoded (pGD499) and the other was integrated into the genome (coefficient of correlation of reporter activities among the

different pea lines was 0.87, P > 0.95). Relatively to expression in corresponding wild-type nodules, expression of both reporter fusions was enhanced several-fold in nodules of Fix− mutants (Table 2). The levels of expression appeared to be inversely correlated with the degree of bacteroid differentiation imposed by the plant genotype, in the following order: Sprint-2Fix− (sym31) > RisFixL (sym32) > E135f (sym13) > wild-type (Table 2, Fig. 2). The npt and Tn5 promoters are constitutive and have no relationship with the symbiotic process. Thus, to some extent they could be considered as bacterial ‘house-keeping’ genes, whose expression may vary according to the physiological and metabolic state of the cells. The enhanced expression of nonsymbiotic genes in juvenile or incompletely differentiated bacteroids accords with the suggestion that housekeeping genes are expressed at low levels in nitrogen-fixing bacteroids (Perret et al., 1999; Margolin, 2000). This would support the suggestion that non-nitrogen fixing, partially developed bacteroids are in a different physiological state from fully differentiated bacteroids. In addition, our results envisage a gradual decline of nonsymbiotic gene expression that would parallel the transition from juvenile to mature bacteroids (Fig. 2). Expression of the fixNc gene in R. leguminosarum bv. viciae VF39, did not show great variation among wild-type or mutant nodules. Expression of the fixNc-gusA fusion in mutant Sprint-2Fix− (sym31) and RBT (sym13, sym31) nodules was, respectively, 1.2- and 1.6-fold higher than in the wildtype Sprint-2. By contrast, in nodules from the RisFixL (sym32) mutant, fixNc-gusA expression was 1.4-fold lower than in the corresponding wild-type, Finale (Table 2). Although these differences were statistically significant, they do not appear to be biologically meaningful and probably fall within the range of plant genotype-specific variation. Expression of fixNc in nodules of the mutant E135f (sym13) was identical to that in wild-type Sparkle nodules (Table 2). It should be emphasized here that although the fixNc-gusA fusion used in this work shows low levels of expresssion, it was induced perfectly


Research

Fig. 2 Ratios (genotype/wild-type) of the levels of expression of the transcriptional fusions nodA-lacZ, dctA-lacZ and fixNc-gusA in the nodules of various pea mutants and wild-types. For comparison, the levels of expression of the constitutively expressed fusions, npt-lacZ and Tn5-gusA are also shown.

under low-oxygen conditions. In free-living cultures, the βglucuronidase activity derived from this fusion was 0.15 nmol min−1 mg protein−1 under normal aerobic conditions, and almost 50-fold greater (7.3 nmol min−1 mg protein−1) in microaerobically grown bacteria. The dctA-lacZ fusion in plasmid pCR12 was inducible by dicarboxylates. In bacteria growing in minimal media with mannitol as the carbon source, the β-galactosidase activity derived from this fusion was 40 nmol min−1 mg protein−1, which rose to 1700 nmol min−1 mg protein−1 when the carbon source was replaced by 10 m sodium succinate. As was also seen for nonsymbiotic genes, expression of the dctA-lacZ fusion in nodules was inversely correlated with the level of bacteroid differentiation in nodules. In comparison with wildtype nitrogen-fixing nodules, expression was increased severalfold in pea genotypes carrying sym31 mutations, increased a few-fold in sym32 mutant nodules, and only slightly enhanced (1.4-fold) in nodules of plants mutated in sym13 (Table 2; Fig. 2). These results are in agreement with previous reports showing that malate uptake is enhanced in bacteroids isolated from sym31 mutant nodules (Borisov et al., 1996; Radukina et al., 1996). The ability of bacteria to take up C4-dicarboxylates (fumarate, malate, succinate), determined by the dctA gene, is essential for nitrogen fixation in symbiotic nodules (Ronson et al., 1981; Watson et al., 1988; Van Slooten et al., 1992; Driscoll & Finan, 1993). In R. leguminosarum bv. viciae and in S. meliloti the Dct system consists of three genes. The dctA gene codes for a high affinity uptake system for dicarboxylic acids. Expression of dctA is controlled by a two-component regulatory system (dctB, dctD) responding to the presence of dicarboxylates (Ronson & Astwood, 1985). The fact that dctA expression is higher in nodules containing undifferentiated bacteroids than in wild-type nodules, despite the fact that dctA function is required for nitrogen fixation, indicates that


dctA induction in the nodule also declines during the transit from juvenile to fully differentiated bacteroids. It appears paradoxical that a gene required for nitrogen fixation by bacteroids is expressed at higher levels in juvenile, nonfixing nitrogen bacteroids. However, it is possible that dctA is also required for bacteroid development and /or that the inducible expression of dctA is also subjected to the general decline in gene expression that seems to occur during bacteroid maturation. This would be consistent with the fact that activation by dicarboxylates is not essential for dctA expression in nitrogenfixing bacteroids of S. meliloti and can be replaced by an oxygen-dependent type of regulation (Boesten et al., 1998). Furthermore, certain dctB mutations that lead to constitutive, dicarboxylate-independent expression of dctA and succinate uptake result in severely decreased nitrogen fixation. This suggests that not only dicarboxylate uptake, but also properly regulated dctA expression are important for bacteroid function (Mavridou et al., 1995). Induction of the nodA-lacZ fusion in plasmid pMP240 by flavonoids was also tested. The β-galactosidase activity under noninducing conditions was 0.212 µmol min−1 mg protein−1, compared to 6 µmol min−1 mg protein−1 seen when 0.4 µ naringenin was added to the bacterial culture. Like the dctAlacZ fusion, the nodA-lacZ construct showed higher expression in mutant nodules containing undifferentiated bacteroids than in wild-type nodules (Table 2). Expression of flavonoidinducible nodulation genes is known to be actively inhibited in bacteroids, although it still remains unclear whether they are fully repressed (Schlaman et al., 1991, 1992; Timmers et al., 1998). It is also unclear whether nod gene repression occurs immediately after endocytosis of bacteria or if this event may be associated with particular stages during bacteroid differentiation. Repression of nodulation genes inside host cells appears to be mediated by NodD2 in Rhizobium sp. NGR234

527

528 Research

and could be important for normal bacteroid development in certain plant hosts (Fellay et al., 1998). Nevertheless, levels of nod gene expression in wild-type nodules probably represent maximum repression. In this case we can envisage a direct correlation between bacteroid development and the extent of nodA gene repression. Expression was greatest in nodules containing mainly juvenile bacteroids (pea lines carrying sym31 mutation), lower in the nodules of line RisFixL (sym32), even lower in the nodules of the sym13 mutant, which contain almost fully differentiated bacteroids (3.1-fold higher than in wild-type Sparkle), while lowest expression levels was observed in wild-type nodules (Table 2, Fig. 2). This would suggest that active repression of flavonoid-inducible nodulation genes is also a gradual event associated with the process of bacteroid differentiation, so that maximum repression is achieved only in fully differentiated, nitrogen-fixing bacteroids. Number of culturable bacteria from pea nodules This is only with difficulty that nitrogen-fixing bacteroids return to the free-living state (Brewin, 1998). Culturable bacteria (i.e. those able to resume free-living growth) recovered from wild-type nodules should correspond mostly to bacteria inside infection threads and, to a certain extent, to those bacteria in the zone of re-invasion (zone V) that could have a saprophytic lifestyle inside the nodule (Timmers et al., 2000). However, zone V appears only in nodules of a certain age (6 weeks in the case of alfalfa), and is not observed in 4-week-old pea nodules, like those used in this study. Some of the culturable bacteria from nodules could also correspond to bacteroids that were not fully differentiated, and that might perhaps retain some capacity for free-living vegetative growth. Indeed, in sym31 and sym32 mutant nodules, where bacteroid differentiation is blocked at an early stage after endocytosis, the numbers of culturable bacteria were, respectively, 60- and 40-fold higher than in the corresponding wild-type nodules (Table 3). By contrast to wild-type nodules, the symbiosomes in these mutant nodules contain several bacterial cells, suggesting that, after endocytosis, bacteria still maintain a capacity for cell division (Borisov et al., 1992; Morzhina et al., 2000). Thus, it is possible that, in these mutant nodules, bacteria that have been subjected to endocytosis are still able to recover free-living growth. This suggests that, even though these bacterial cells are membrane-enclosed, their physiology is very different from that of fully differentiated bacteroids. We do not think that the large numbers of bacteria recovered from these mutant nodules are due to the appearance of a nodule ‘saprophytic’ zone (zone V, according to Timmers et al., 2000) following on from the senescence zone (zone IV). Zone IV can hardly be observed in 4-week-old nodules of wild-type pea plants, although it can be seen in 4-week-old nodules of the line RisFixL (sym32) (Morzhina et al., 2000). Furthermore, lines Sprint-2Fix- (sym31) and RBT (sym13, sym31) were reported to show no premature degradation

Table 3 Number of culturable bacteria from nodules of mutant and wild-type pea lines

Pea lines

Number of cfu* in nodules (104 mg−1 nodule fresh weight)

Sprint-2 (wild-type) Sprint-2Fix− (sym31)b RBT (sym13, sym31) RisFixL (sym32) Finale (wild-type) E135f (sym13) Sparkle (wild-type)

8.2 ± 1.3 525.2 ± 200.4 1058.7 ± 154.3 696.0 ± 110.0 18.1 ± 2.6 144.2 ± 43.8 2.7 ± 0.6

(64) (259) (38) (53)

* cfu, colony forming units. The ratio of mean values between mutant and corresponding wild-type are given in brackets. For the double mutant line, RBT, this ratio is calculated as an average of ratios to Sprint-2 and Sparkle. aThe mean values differ significantly (P > 0.95) from those in corresponding wild-type lines (line RBT was compared with both Sprint-2 and Sparkle). bSprint-2Fix− (sym31) is distinguished by high variability of the data between replicate measurements.

(senescence) of nodule symbiotic structures (Borisov et al., 1992, 1997a). Surprisingly, we found that nodules of early senescent mutant E135f (sym13) also contained significantly more (50fold) culturable bacteria than did wild-type nodules (Table 3). This contrasts with morphological studies, which have shown that zone III of these nodules contains almost fully differentiated bacteroids (Kneen et al., 1990). Our data suggest that these bacteroids are yet not completely differentiated, but perhaps blocked at a stage earlier than the point at which return to free-living growth becomes impossible, as is the case for nitrogen-fixing, mature bacteroids. Nodules of mutant E135f (sym13) are characterized by premature senescence (Kneen et al., 1990) and therefore it cannot be ignored that at least part of the increased numbers of culturable bacteria recovered from these nodules is due to bacterial re-infection of the senescent zone, as suggested by Timmers et al. (2000). It is possible that bacteroids in nodules of mutant E135f (sym13), although showing morphologically advanced differentiation, may retain a greater capacity than nitrogen-fixing bacteroids to recover free-living growth. This suggests that the loss of capacity to resume vegetative growth is perhaps one of the last events in the development of nitrogen-fixing bacteroids.

Conclusions As a whole, our results suggest that, with the exception of oxygen-regulated symbiotic genes, expression of bacterial genes, whether or not related to symbiosis, declines gradually as the programme of bacteroid differentiation advances. This would parallel the gradual morphological changes observed in the bacterial cells during this developmental process. Even actively repressed genes, such as the flavonoid-induced nodulation genes, appear to be subject to this decline in


Research

expression during bacteroid development. It has been speculated that nod genes are repressed immediately after bacterial endocytosis. Our results, however, indicate that active repression is a function of the extent of bacteroid differentiation reached, and that greatest repression is achieved only in nitrogen-fixing bacteroids. Similarly, actively induced genes, such as dctA, appear to be subject to this general decline in gene expression. Regulation of dctA in nodules is a complex, incompletely understood process (Mavridou et al., 1995; Boesten et al., 1998). Induction of dctA inside the nodule may be taken over by an oxygendependent type of regulation, not dependent on the stage of bacteroid differentiation, linking dicarboxylate uptake and energy production to the oxygen-regulated process of nitrogen fixation. Our data agree with previous reports (Vasse et al., 1990) suggesting that development of organelle-like, nitrogen-fixing symbiosomes after endocytosis of bacteria by plant nodule cells is not an obvious consequence of endocytosis, but rather a gradual process involving several additional stages, each controlled by different plant genes. The gradual decline in bacteroid gene expression would parallel the various morphological stages observed in indeterminate nodules. Within this perspective, the process of bacteroid differentiation could resemble an adjustable process of adaptation to the symbiosome environment, believed to be of a stressful nature (Brewin, 1998; Santos et al., 2000; Nogales et al., 2002). The end result of this adaptive process would be the mature, actively nitrogenfixing bacteroid, which has undergone a loss of biological identity (transition to an organelle-like state) as a consequence of the loss of the capacity to resume free-living growth.

Acknowledgements This work was supported by NATO Scientific and Environmental Affairs Division (HTECH.LG 971210; LST.CLG 977076), for VET, AYB, IAT, JMS-P, JS, JO; VolkswagenStiftung, Germany (I/72 935), for VET, VAV, AYB, IAT, UBP; the Russian Foundation for Basic Research (98-0449883; 01-04-48580), for VET, AYB, VAV, IAT; and a grant of the President of Russia (02-15-99408), for VET, AYB, VAV. VET and AYB have been supported by NATO Scientific and Environmental Affairs Division DC Fellowships, and VET by INTAS Fellowship YSF 00-175. VAV has been also supported by DAAD Fellowship 325A/00/02223 and INTAS Fellowship YSF 00-174. The authors are grateful to C. W. Ronson and H. P. Spaink for providing reporter fusions. S. Muñoz is acknowledged for excellent technical assistance.

References Bedmar EJ, Edie SA, Phillips DA. 1983. Host plant cultivar effect on hydrogen evolution by Rhizobium leguminosarum. Plant Physiology 72: 1011–1015.


Berdnikov VA, Rosov SM, Bogdanova VS. 1989. Construction of a series of laboratory pea lines. Abstracts of Conference on Genetics 23–25 May, Vol. II. Genetics of Plants. Kiev, USSR: 26–27. Beringer JE. 1974. R factor transfer in Rhizobium leguminosarum. Journal of General Microbiology 84: 188–198. Boesten B, Batut J, Boistard P. 1998. DctBD-dependent and -independent expression of the Sinorhizobium (Rhizobium) meliloti C4-dicarboxylate transport gene (dctA) during symbiosis. Molecular Plant–Microbe Interactions 11: 878–886. Borisov AY, Barmicheva EM, Jacobi LM, Lebsky VK, Morzhina EV, Tsyganov VE, Voroshilova VA, Tikhonovich IA. 2000. Mendelian genes controlling development of nitrogen-fixing nodules and arbuscular mycorrhiza in pea. Czech Journal of Genetics and Plant Breeding 36: 106– 100. Borisov AY, Lebsky VK, Morzhina EV, Rozov SM, Tsyganov VE, Tikhonovich IA. 1997b. Study of pea (Pisum sativum L.) symbiotic gene interactions using a panel of lines carrying two Fix- mutations and a comparative study of their nodule ultrastructure. In: Elmerich C, Kondorosi A, Newton WE, eds. Nitrogen fixation for the 21st century. Dordrecht, The Netherlands: Kluwer Academic Publishers, 345. Borisov AY, Morzhina EV, Kulikova OA, Tchetkova SA, Lebsky VK, Tikhonovich IA. 1992. New symbiotic mutants of pea (Pisum sativum L.) affecting either nodule initiation or symbiosome development. Symbiosis 14: 297–313. Borisov AY, Rozov SM, Tsyganov VE, Morzhina EV, Lebsky VK, Tikhonovich IA. 1997a. Sequential functioning of Sym13 and Sym31, two genes affecting symbiosome development in root nodules of pea (Pisum sativum L.). Molecular and General Genetics 254: 592–598. Borisov AY, Tsyganov VE, Morzhina EV, Chvabauskene JA, Radukina NL, Skøt L, Rozov SM, Lebsky VK, Romanov VI, Tikhonovich IA. 1996. Search for morphological and biochemical characters which mark the stages of symbiosis development controlled by certain pea (Pisum sativum L.) symbiotic genes. In: Wójtowicz A, Stepkowska J, Szlagowska A, eds. Biological fixation of nitrogen for ecology and sustainable agriculture. Poznan, Poland: Scientific Publishers OWN, 237. Bradford MN. 1976. A rapid and sensitive method for the quantification of microgram quantities of protein utilizing the principle of protein-dye binding. Analytical Biochemistry 72: 248–254. Brewin NJ. 1998. Tissue and cell invasion by Rhizobium : The structure and development of infection threads and symbiosomes. In: Spaink HP, Kondorosi A, Hooykaas PJJ, eds. Rhizobiaceae, molecular biology of model plant-associated bacteria. Dordrecht, The Netherlands: Kluwer Academic Press, 417–429. Brewin NJ, Ambrose MJ, Downie JA. 1993. Root nodules, Rhizobium and nitrogen fixation. In: Casey R, Davies DR, eds. Peas: genetics, molecular biology and biotechnology. Wallingford, UK: CAB International, 237–291. Dénarié J, Debellé F, Promé J-C. 1996. Rhizobium lipochitooligosaccharide nodulation factors: signaling molecules mediating recognition and morphogenesis. Annual Review of Biochemistry 65: 503–535. Ditta G, Schmidhauser T, Yakobson E, Lu P, Liang X-W, Finlay DR, Guiney D, Helinski DR. 1985. Plasmids related to the broad host range vector, pRK290, useful for gene cloning and for monitoring gene expression. Plasmid 13: 149–153. Ditta G, Stanfield S, Corbin D, Helinski DR. 1980. Broad host range DNA cloning system for Gram-negative bacteria. Construction of a gene bank of Rhizobium meliloti. Proceedings of the National Academy Sciences, USA 77: 7347–7351. Downie JA, Walker SA. 1999. Plant responses to nodulation factors. Current Opinion in Plant Biology 2: 483–489. Driscoll BT, Finan TM. 1993. NAD+-dependent malic enzyme of Rhizobium meliloti is required for symbiotic nitrogen fixation. Molecular Microbiology 7: 865–873. Engvild KJ. 1987. Nodulation and nitrogen fixation mutants of pea (Pisum sativum). Theoretical and Applied Genetics 74: 711–713.

529

530 Research Fellay R, Hanin M, Montorzi G, Frey J, Freiberg C, Golinovsky W, Staehelin C, Broughton WJ, Jabouri S. 1998. nodD2 of Rhizobium sp. NGR234 is involved in the repression of the nodABC operon. Molecular Microbiology 27: 1039–1050. Kneen BE, LaRue TA, Hirsch AM, Smith CA, Weeden NF. 1990. sym-13 – a gene conditioning ineffective nodulation in Pisum sativum. Plant Physiology 94: 899–905. Kosterin OE, Rozov SM. 1993. Mapping of the new mutation blb and the problem of integrity of linkage group I. Pisum Genetics 25: 27–31. Long SR. 1996. Rhizobium symbiosis: Nod factors in perspective. Plant Cell 8: 1885–1898. Margolin W. 2000. Differentiation of free-living rhizobia into endosymbiotic bacteroids. In: Brun YV, Shimkets LJ, eds. Prokaryotic development. Washington, DC, USA: American Society for Microbiology. 441–466. Marie C, Plaskitt KA, Downie JA. 1994. Abnormal nodule development in nodules induced by a glucosamine synthase mutant of Rhizobium leguminosarum. Molecular Plant–Microbe Interactions 7: 482–487. Mavridou A, Barny M-A, Poole P, Plaskitt K, Davies AE, Johnston AWB, Downie JA. 1995. Rhizobium leguminosarum nodulation gene (nod ) expression is lowered by an allele-specific mutation in the dicarboxylate transport gene dctB. Microbiology 141: 103–111. Miller JH. 1972. Experiments in molecular genetics. Cold Spring Harbor, New York, NY, USA: Cold Spring Harbor Laboratory Press. Morzhina EV, Tsyganov VE, Borisov AY, Lebsky VK, Tikhonovich IA. 2000. Four developmental stages identified by genetic dissection of pea (Pisum sativum L.) root nodule morphogenesis. Plant Science 155: 75–83. Mylona P, Pawlowski K, Bisseling T. 1995. Symbiotic nitrogen fixation. Plant Cell 7: 869–895. Nogales J, Campos R, BenAbdekhalek H, Olivares J, Lluch C, Sanjuan J. 2002. Rhizobium tropici genes involved in free-living salt tolerance are required for the establishment of efficient nitrogen-fixing symbiosis with Phaseolus vulgaris. Molecular Plant–Microbe Interactions 15: 225–232. Perret X, Freiberg C, Rosenthal A, Broughton WJ, Fellay R. 1999. High-resolution transcriptional analysis of the symbiotic plasmid of Rhizobium sp. NGR234. Molecular Microbiology 32: 415–425. Priefer UB. 1989. Genes involved in lipopolysaccharide production and symbiosis are clustered on the chromosome of Rhizobium leguminosarum biovar viciae VF39. Journal of Bacteriology 171: 6161–6168. Radukina NI, Chvabauskene YA, Borisov AY, Tsyganov VE, Sofin AF, Romanov VI, Tikhonovich IA. 1996. Malate and glucose transport by bacteroids isolated from the nodules of the different plant-determined ineffective pea mutants (Pisum sativum L.). In: Wójtowicz A, Stepkowska J, Szlagowska A, eds. Biological fixation of nitrogen for ecology and sustainable agriculture. Poznan, Poland: Scientific Publishers OWN, 205. Ronson CW, Astwood PM. 1985. Genes involved in the carbon metabolism of bacteroids. In: Bottomley PJ, Newton WE, Nijhoff M, eds. Nitrogen fixation research progress. Dordrecht, The Netherlands: Kluwer Academic Press, 201–207. Ronson CW, Astwood PM, Nixon BT, Ausebel FM. 1987a. Deduced products of C4-dicarboxylate transport regulatory genes of Rhizobium leguminosarum are homologous to nitrogen regulatory gene products. Nucleic Acids Research 15: 7921–7934. Ronson CW, Lyttleton P, Robertson JG. 1981. C4-dicarboxylate trasport mutants of Rhizobium trifolii form ineffective nodules of Trifolium repens. Proceedings of the National Academy Sciences, USA 78: 4284–4288. Ronson CW, Nixon BT, Albright LM, Ausubel FM. 1987b. Rhizobium meliloti ntrA (rpoN ) gene is required for diverse metabolic functions. Journal of Bacteriology 169: 2424–2431.

Santos R, Herouart D, Puppo A, Touati D. 2000. Critical protective role of bacterial superoxide dismutase in Rhizobium-legume symbiosis. Molecular Microbiology 38: 750–759. Schlaman HRM, Horvath B, Vijgenboom E, Okker RJH, Lugtenberg BJJ. 1991. Suppression of nodulation gene expression in bacteroids of Rhizobium leguminosarum biovar viceae. Journal of Bacteriology 173: 4277– 4287. Schlaman HRM, Lugtenberg BJ, Okker RJ. 1992. The nodD protein does not bind to the promoters of inducible nodulation genes in extracts of bacteroids of Rhizobium leguminosarum biovar viceae. Journal of Bacteriology 174: 6109–6116. Schlüter A, Patschkovski T, Quandt J, Selinger LB, Weidner S, Krämer M, Zhou L, Hynes MF, Priefer UB. 1997. Functional and regulatory analysis of the two copies of the fixNOQP operon of Rhizobium leguminosarum strain VF 39. Molecular Plant–Microbe Interactions 10: 605–616. Schultze M, Kondorosi A. 1996. The role of lipooligosaccharides in root nodule organogenesis and plant cell growth. Current Opinion in Genetics and Development. 6: 631–638. Sharma SB, Singer ER. 1990. Temporal and spatial regulation of the symbiotic genes Rhizobium meliloti in planta revealed by transposon Tn5-gusA. Genes and Development 4: 344–356. Spaink HP. 1996. Regulation of plant morphogenesis by lipo-chitin oligosaccharides. Critical Reviews in Plant Sciences. 15: 559–582. Spaink HP, Okker RJH, Wijffelman CA, Pees E, Lugtenberg BJJ. 1987. Promoters in the nodulation region of the Rhizobium leguminosarum Sym plasmid pRL1JI. Plant Molecular Biology 9: 27–39. Timmers ACJ, Auriac M-C, de Billy F, Truchet G. 1998. Nod factor internalization and microtubular cytoskeleton changes occur concomitantly during nodule differentiation in alfalfa. Development 125: 339–349. Timmers ACJ, Soupène E, Auriac MC, de Billy F, Vasse J, Boistard P, Truchet G. 2000. Saprophytic intracellular rhizobia in alfalfa nodules. Molecular Plant–Microbe Interactions 13: 1204–1213. Tsyganov VE, Borisov AY, Rozov SM, Tikhonovich IA. 1994. New symbiotic mutants of pea obtained after mutagenesis of laboratory line SGE. Pisum Genetics 26: 36–37. Tsyganov VE, Morzhina EV, Stefanov SY, Borisov AY, Lebsky VK, Tikhonovich IA. 1998. New pea (Pisum sativum L.) genes sym33 and sym40 control infection thread formation and root nodule function. Molecular and General Genetics 256: 491–503. Van Slooten JC, Bhuvanasvari TV, Bardin S, Stanley J. 1992. Two C4-dicarboxylate transport systems in Rhizobium sp. NGR234: Rhizobial dicarboxylate transport is essential for nitrogen fixation in tropical legume symbioses. Molecular Plant–Microbe Interactions 5: 179–186. Vasse J, De Billy F, Camut S, Truchet G. 1990. Correlation between ultrastructural differentiation of bacteroids and nitrogen fixation. Journal of Bacteriology 172: 4295–4306. Voroshilova VA, Boesten B, Tsyganov VE, Borisov AY, Tikhonovich IA, Priefer UB. 2001. Effect of mutations in Pisum sativum L. genes (sym13, sym31, sym33, sym40 ) blocking different stages of nodule development on the expression of late symbiotic genes in Rhizobium leguminosarum bv. viciae. Molecular Plant–Microbe Interactions 14: 471–476. Watson RJ, Chan Y-K, Wheatcroft R, Yang A-F, Han S. 1988. Rhizobium meliloti genes required for C4-dicarboxylate transport and symbiotic nitrogen fixation are located on a megaplasmid. Journal of Bacteriology 170: 927–934. Wilson KJ, Hughes SG, Jefferson RA. 1992. The Escherichia coli gus operon: induction and expression of the gus operon in E. coli and the occurrence and use of GUS in other bacteria. In: Gallagher SR, ed. Gus protocols. Using the GUS gene as a reporter of gene expression. New York, USA: Academic Press, 8–22.
