Mutations at therug4locus alter the carbon and nitrogen metabolism of pea plants through an effect on sucrose synthase


*For correspondence (fax +44 1603 456844; e-mail


The biochemical and molecular basis of the wrinkled-seeded phenotype ofrug4mutants of pea (Pisum sativumL.) has been investigated. Mutant embryos have reduced starch contents and only 5% of the sucrose synthase activity of wild-type embryos during development. Activities of other enzymes involved in the conversion of sucrose to starch are unaffected. A gene encoding an isoform of sucrose synthase expressed in the embryo co-segregates with therug4locus, and one of the three mutant alleles has been show to carry a point mutation in this gene that converts a highly conserved arginine residue to a lysine residue. It is highly likely that the reduced starch content of the mutant embryo is a direct consequence of the loss of sucrose synthase activity. The mutations reduce the activity of sucrose synthase in the testa and the leaf by 50% or less, but activity inRhizobium-infected root nodules is reduced by 85%. Although the nodules of mutant plants contain metabolically active bacteroids, the N content and δ15N values of these plants in the field indicate that, unlike wild-type plants, they derive little of their N from N2 fixation viaRhizobium. Sucrose synthase thus appears to be essential for the supply of carbon for bacteroid metabolism and/or ammonia assimilation during nitrogen assimilation.


The aim of this work was to investigate the basis of the wrinkled-seeded phenotype caused by mutations at the rug4 locus of pea (Pisum sativum L.). Three mutant alleles at the rug4 locus –rug4-a, rug4-b and rug4-c– were identified by selection of wrinkled-seeded peas ( Wang et al. 1990 ) and allelism testing ( Wang & Hedley 1993) following chemical mutagenesis of a round-seeded line. In common with other wrinkled-seeded genotypes, seeds of the rug4 mutants have lower starch contents than seeds of the round-seeded wild-type. Starch is typically 55% of the final dry weight of wild-type seeds but 40% of that of mutant seeds ( Wang & Hedley 1993).

A wrinkled-seeded phenotype is likely to be due to a reduction in activity of one of the enzymes in the pathway of starch synthesis in developing embryos. Well-characterized mutations at the rugosus loci r and rb cause specific reductions in activity of individual enzymes in the pathway of starch synthesis and hence a reduced rate of starch synthesis. The consequent accumulation of sucrose in the embryo leads, through osmotic effects, to increased fresh weight during development, and thus to wrinkling of the seed during desiccation at maturity ( Wang & Hedley 1991). Mutations at the r and rb loci lie in genes encoding one of the two isoforms of starch-branching enzyme ( Bhattacharyya et al. 1990 ;Smith 1988) and the large subunit of ADP-glucose pyrophosphorylase ( Hylton & Smith 1992; our unpublished data), respectively.

In this paper, we show that mutations at the rug4 locus cause a large and specific reduction in the activity of sucrose synthase (Sus) in the developing embryo and in Rhizobium nodules on the roots, and smaller reductions in Sus activity in testas and leaves. We demonstrate that a gene encoding Sus lies at the rug4 locus. Sus has been implicated in many aspects of plant metabolism. It is thought to be responsible for the mobilization of sucrose in sink organs including pea embryos, and to be a determinant of sink strength in these organs ( Edwards & ap Rees 1986a;Zrenner et al. 1995 ). Sus is believed to catalyse the metabolism of part of the sucrose entering the phloem complex to provide the energy required for phloem loading in leaves ( Geigenberger et al. 1996 ;Martin et al. 1993 ). Sus activity in the testa of pea seeds is proposed to facilitate the interconversion of starch and sucrose and thus to buffer the supply of carbohydrate to the developing embryo ( Rochat & Boutin 1992). The high Sus activity in Rhizobium nodules on the roots of legumes, and its increase during development of nitrogen-fixing capacity in the nodule, suggest that it mediates sucrose metabolism to provide an energy supply for the bacteroids and/or carbon skeletons for the assimilation of ammonia ( Gordon 1995;Streeter 1995). In the light of these suggested roles for the enzyme, we have assessed the effects of mutations at the rug4 locus on the metabolism of the pea plant as a whole.


The rug4 mutations specifically reduce sucrose synthase activity in the developing embryo

Comparison of the maximum catalytic activities of enzymes in the pathway from sucrose to starch in wild-type embryos and embryos of the rug4-b mutant line in mid-development revealed only one significant difference. Activity of Sus in the mutant embryos in both the synthetic and the cleavage direction was only about 5% of that in wild-type embryos ( Table 1). Sus activity in extracts of mixtures of wild-type and rug4 embryos was within 15% of that predicted from separate extracts of the two sorts of embryo, indicating that the low activity of Sus in rug4 embryos was not due to inhibition or degradation of the enzyme during extraction. These results suggest that mutations at the rug4 locus have a specific and direct effect on the activity of Sus in the developing embryo. We therefore investigated whether a Sus gene lies at the rug4 locus.

Table 1.  Comparison of maximum catalytic activities of enzymes involved in starch synthesis in wild-type and rug4 mutant embryos
Enzyme activity (μmol min–1 g–1 fresh weight) 
  1. Extracts of embryos of 200–300 mg fresh weight were assayed for activities of enzymes on the pathway of starch synthesis. Except for total starch synthase and starch-branching enzyme, activities were assayed on the soluble fractions of extracts. Total starch synthase and starch-branching enzyme were assayed on crude homogenates. Starch-branching enzyme activity was measured as the stimulation by the homogenate of incorporation via phosphorylase a of glucose from glucose-1-phosphate into methanol-insoluble polymer, and is expressed as μmol glucose incorporated min–1 g–1 fresh weight. Values are means ± SE of four measurements, each made on an embryo from a different plant.

Sucrose synthase (synthetic) 3.04 ± 0.38
(cleavage) 2.00 ± 0.27
Phosphoglucomutase 21.6 ± 2.4
ADPglucose pyrophosphorylase 1.11 ± 0.54
Starch synthase (soluble)0.082 ± 0.016
(total)0.239 ± 0.036
Starch-branching enzyme 11.7 ± 1.8
Alkaline pyrophosphatase 3.93 ± 0.29

A gene encoding sucrose synthase co-segregates with the rug4 phenotype

To provide a suitable population of plants for linkage analysis, the rug4-b mutant line was crossed with line JI281 (John Innes Centre germplasm collection). A Sus gene was obtained by screening a cDNA library derived from pea leaves with a 1 kb fragment of a Sus gene from Lotus japonicus ( Skøt et al. 1996 ; kind gift from Dr L. Skøt, Welsh Plant Breeding Station, Aberystwyth, Dyfed, UK). The DNA sequence of the pea Sus gene was 98% identical to that of a cDNA obtained from pea root nodules (EMBL database accession number AF079851) and 67% identical to a cDNA obtained from pea testa (EMBL database accession number AJ001071). The gene contained a sequence identical to a second, partial Sus cDNA obtained from pea testa ( Déjardin et al. 1997 ). The pea Sus gene was used to probe DNA gel blots of rug4-b and JI281 DNA digested with various restriction enzymes. A suitable restriction fragment length polymorphism (RFLP) was identified in DNA digested with HindIII. Genomic DNA from 121 F2 plants from randomly selected seeds derived from the cross between rug4-b and JI281 was then digested with HindIII, blotted and probed with a 1.5 kb NcoI fragment of the Sus gene. The plants were also scored according to whether the seeds from which they were grown were round or wrinkled. In some cases the seed shape was difficult to classify unambiguously. We had observed previously that the rug4 phenotype may be mild, with little wrinkling of the seed. Where identification was difficult, 10 developing embryos from the F2 plant were assayed for Sus activity. For the 91 plants for which classification of seed shape was unambiguous, all of the plants from round seed had the RFLP pattern of the JI281 parent or a combined pattern. All of the plants from wrinkled seed had the RFLP pattern of the rug4-b parent. Of the remaining 30 plants, embryos of 25 had the Sus activities expected from the RFLP pattern. In the remaining five cases, enzyme activity was lower than expected. This may have been due to the effects of plant age, growth conditions or genetic background. Overall, therefore, linkage analysis provides strong evidence that the Sus gene lies at or very close to the rug4 locus.

The rug4-b mutation lies in a gene encoding sucrose synthase

To discover whether mutations at the rug4 locus actually lie within the Sus gene at this locus, first-strand cDNA was synthesized from mRNA isolated from embryos of the rug4-b line. Sus cDNAs were isolated by RT–PCR, subcloned and sequenced. The sequences from three independently derived clones from rug4-b were identical to the gene from the wild-type except at position 1757 where the G/C in the wild-type had been changed to A/T. This change converts arginine 578 to lysine. The arginine in this position is completely conserved in all of the putative amino acid sequences so far reported for Sus (not shown) and is thus likely to be very important to the normal functioning of the enzyme.

Effects of rug4 mutations on enzymes of sucrose metabolism in the embryo

Although Sus – in conjunction with UDP-glucose pyrophosphorylase – is the predominant route via which sucrose and hexose phosphates are interconverted in developing pea embryos ( Edwards & ap Rees 1986a, b), embryos also contain invertase, glucokinase and sucrose phosphate synthase activity. To assess the potential contributions of these enzymes to sucrose metabolism, activities were measured through the development of wild-type and mutant (rug4-b) embryos, and compared with activities of Sus (Figure 1A). The activity of alkaline invertase was slightly higher in mutant than in wild-type embryos, but activities of the other enzymes were not affected by the mutation. Activities of alkaline and soluble acid invertase, glucokinase, fructokinase and sucrose phosphate synthase were all considerably lower than that of Sus, whereas activity of UDP-glucose pyrophosphorylase was about 20-fold higher. No insoluble acid invertase activity was detected.

Effect of rug4 mutations on the testa and the leaf

Sus activity in wild-type testas was comparable with that in wild-type embryos on a fresh-weight basis. However, in marked contrast to the situation in embryos, the rug4 mutation reduced activity in the testa by only about 50% (Figure 1B) throughout testa development. Activities of alkaline and soluble acid invertase, glucokinase, fructokinase, UDP-glucose pyrophosphorylase and sucrose phosphate synthase in wild-type testas were comparable with or up to two-fold higher than those in wild-type embryos, and the activity of insoluble acid invertase was readily detectable in testas. None of these activities was significantly affected by mutations at the rug4 locus.

Sus activity in fully expanded leaves of greenhouse-grown plants (supplied with nitrogenous fertilizer) was 30–40% lower in mutant relative to wild-type plants. Activities in wild-type, rug4-a, rug4-b and rug4-c leaves were 0.193 ± 0.012, 0.136 ± 0.009, 0.113 ± 0.009, and 0.109 ± 0.019 μmol min–1 g–1 fresh weight, respectively (degradative direction: means ± SE of measurements on four plants of each genotype). Mutant plants were indistinguishable in appearance from wild-type plants. Chlorophyll contents and activities of the Calvin cycle enzyme NADP-glyceraldehyde 3-phosphate dehydrogenase – both of which are reduced in leaves in which sucrose export is seriously restricted (e.g. Geigenberger et al. 1996 ;Stitt et al. 1990 ) – were the same in wild-type and mutant (rug4-b) leaves (not shown). However, wild-type and mutant plants grown in the field without the addition of fertilizer differed from an early stage. Mutant plants were more yellow and grew more slowly. At a stage when 2–4 nodes were flowering, leaflet area was reduced by approximately twofold, chlorophyll content was reduced by 25–30% and leaf starch content was increased by three- to fourfold in mutants ( Table 2). All the mutant lines were significantly different from the wild-type line for these three characters (paired t tests, 5% level).

Table 2.  Leaf area, chlorophyll content and starch content of wild-type and rug4 mutant plants grown in the field
  1. For each genotype, the basal pair of leaflets was harvested from the leaf at the first flowering node of ten randomly selected plants from field plots, at a stage when flowering had commenced at 2–4 nodes. Leaflet areas were measured with a leaf area meter, then one leaflet was assayed for chlorophyll and the other was extracted in 900 ml l–1 aqueous ethanol and assayed for starch according to Smith (1988). Values for leaf area are the means ± SE of the mean values for the two leaflets from each of the 10 plants. Values for starch and chlorophyll (Chl) content are the means ± SE of measurements on the 10 plants.

Chl content (mg g–1 FW)2.30 ± 0.21 1.56 ± 0.14 1.75 ± 0.13 1.61 ± 0.07
Starch content (mg g–1 FW)3.63 ± 1.9015.25 ± 3.6114.15 ± 2.3610.31 ± 2.74
Leaflet area (cm–2) 20.5 ± 4.6 9.5 ± 3.9 7.7 ± 3.3 9.4 ± 4.5

The phenotypes of the mutants in the field suggested that these plants might be nitrogen-deficient. We therefore investigated the effect of the rug4 mutations on the nitrogen status of field-grown plants and on the metabolism of root nodules.

Effect of rug4 mutations on the nitrogen status of plants in the field

Total nitrogen content (as percentage dry weight) and δ15N were measured on freeze-dried samples of field-grown pea plants. Three weed species growing in the same field plot –Brassica napus, Chenopodium album and Poa annua– were also sampled to provide a non-leguminous comparison. δ15N measures enrichment or depletion of a sample in the nitrogen isotope 15N relative to atmospheric nitrogen (see Experimental procedures).

The nitrogen contents of all three rug4 mutant lines were considerably lower than those of wild-type plants, particularly in the above-ground parts of the plant ( Table 3). From measurements of shoot dry weight (leaves plus stems), we estimated that mutant plants contained three to five times less nitrogen (mg N per shoot) than wild-type plants (data not shown). Values of δ15N for wild-type and mutant plants were also very different, and differences were more pronounced in leaves than in roots ( Table 3). In wild-type plants the greatest value was in the root, whereas in mutant plants it was in the leaf. Shoot values of δ15N for mutant peas were much more similar to those for non-leguminous weeds than those for wild-type peas. Shoot values for the mutant and wild-type pea lines were 3.4–5.9‰ and 0.1‰, respectively, and values for Brassica napus, Chenopodium album and Poa annua were 7.3, 6.9 and 4.1‰, respectively.

Table 3.  Nitrogen contents and δ15N values of field-grown pea plants
Nitrogen content (% dry weight)δ15N value (parts per thousand)
  1. The above- and below-ground parts of ten randomly selected plants of each genotype at the developmental stage described in Table 2 were harvested from field plots and divided into five samples, each consisting of two plants. The leaves, stems, roots (from which nodules had been removed) and nodules from each of the five samples were freeze-dried and analysed for nitrogen content and δ15N via continuous flow mass spectrometry ( Handley et al. 1993 ). Values are the means ± SE of measurements on the five samples.

Wild-type3.19 ± 0.391.24 ± 0.111.97 ± 0.104.22 ± 0.21– 0.05 ± 0.320.38 ± 0.541.59 ± 0.330.36 ± 0.25
rug4-a1.58 ± 0.160.73 ± 0.241.59 ± 0.063.63 ± 0.386.74 ± 1.455.37 ± 1.162.86 ± 0.451.81 ± 0.43
rug4-b1.39 ± 0.150.70 ± 0.111.38 ± 0.073.57 ± 0.514.91 ± 1.653.72 ± 1.854.14 ± 0.503.03 ± 0.26
rug4-c1.33± 0.080.62 ± 0.091.56 ± 0.093.12 ± 0.194.22 ± 0.962.79 ± 0.902.74 ± 0.301.77 ± 0.84

Effect of rug4 mutations on root nodules

Nodules were obtained by inoculation with Rhizobium of the roots of pea plants on a sterile medium lacking mineral nitrogen. At 15 days after inoculation, nodules from mutant plants had much lower activities of Sus than those from wild-type plants ( Table 4). The activities of ATP-phosphofructokinase and PEP carboxylase – enzymes involved in the conversion of sucrose to carbon skeletons for bacteroid respiration and ammonia assimilation – and of alkaline invertase were unaffected by the mutation. The activity of alkaline invertase was very high in both wild-type and mutant nodules. The activity of 3-hydroxybutyrate dehydrogenase – involved in storage product metabolism in bacteroids – was reduced by 40% in mutant relative to wild-type nodules ( Table 4). Sus activity in extracts of mixtures of wild-type and rug4 nodules was within 10% of that predicted from separate extracts of the two sorts of nodule, indicating that the low activity of Sus in rug4 nodules was not due to inhibition or degradation of the enzyme during extraction.

Table 4.  Maximum catalytic activities of enzymes in nodules from wild-type and rug4 mutant plants
Enzyme activityWild-typerug4-arug4-brug4-c
  1. Activities of sucrose synthase (Sus), alkaline invertase, phosphoenolpyruvate carboxylase (PEP carboxylase), ATP-phosphofructokinase (ATP-PFK), and β-hydroxybutyrate dehydrogenase (β-hydroxybutyrate DH) were measured on soluble extracts of nodules harvested from plants grown in a sterile, nitrogen-deficient medium and inoculated with Rhizobium leguminosarum. Values are μmol min–1 g–1 fresh weight and are the means ± SE of measurements made on four extracts, each of nodules from a different plant. The ability of root systems to reduce acetylene to ethylene is presented as arbitrary units g–1 fresh weight. Values are the means ± SE of measurements on the numbers of individual plants shown in parentheses except for the rug4-c line, for which a single measurement was made. nd, not determined.

Sus (synthetic)1.94 ± 0.39nd0.10 ± 0.03nd
(cleavage)1.88 ± 0.210.42 ± 0.060.36 ± 0.120.23 ± 0.07
Alkaline invertase4.71 ± 0.36nd3.76 ± 0.27nd
ATP-PFK5.10 ± 0.37nd4.34 ± 0.94nd
PEP carboxylase0.62 ± 0.09nd0.75 ± 0.20nd
β-hydroxybutyrate DH1.55 ± 0.11nd0.95 ± 0.15nd
Acetylene reduction18.2 ± 8.1 (4)22.4 ± 7.0 (3)16.0 ± 2.3 (3)7.2 (1)

Measurements of activity of the bacteroid nitrogenase responsible for nitrogen fixation cannot be made in extracts of nodules. The enzyme is unstable except at very low oxygen concentrations. However, the conversion of acetylene to ethylene by this enzyme can be measured in whole, nodulated root systems ( Hardy et al. 1968 ). Nodules of all three mutant lines had acetylene reduction activity. The rug4 mutations did not bring about any dramatic change in acetylene reduction activity at this stage in nodule development ( Table 4).

Differences between nodules of mutant and wild-type plants in the appearance of the central, infected region became apparent as the nodules matured. Senescence of cells in the infected region occurred earlier in mutant than in wild-type nodules, and was observed as early as 14 days after inoculation ( Fig. 2).

Figure 2.

Light micrographs of median longitudinal sections of nodules from rug4 mutant (a) and wild-type (b) plants at 14 days after inoculation.

Nodules were harvested from plants grown in a sterile, nitrogen-deficient medium, 14 days after inoculation with Rhizobium leguminosarum. Sections of fixed, embedded material were counter-stained with basic fuchsin. The bar represents a distance of 0.1 mm. Note the presence of disorganized, senescent cells (arrowheads) in the infected region (ir) of rug4 mutant but not wild-type nodules.


Our data provide unequivocal evidence that the mutations at the rug4 locus directly affect a gene encoding sucrose synthase. Co-segregation analysis using a fragment of a Sus gene expressed in embryos revealed that this gene lies at or very close to the rug4 locus. Peas carrying the rug4-b allele have a point mutation within the DNA encoding this sucrose synthase. The mutations result in a dramatic loss of Sus activity in embryos and nodules, and much smaller losses of activity in testas and leaves.

The fact that the rug4 mutations affect Sus activity to different degrees in different organs suggests that there is more than one isoform of this enzyme in the pea plant. The results are consistent with the idea that the isoform affected by the rug4 mutation accounts for most of the Sus activity in embryos and nodules, about half of the activity in testas, and a smaller fraction of the activity in leaves. Multiple, differently expressed, isoforms of Sus have been reported from several other species. For example, a gene encoding one isoform of Sus from potato (sus3) is highly expressed in the stems and roots and a gene encoding a second isoform (sus4) is primarily expressed in the tuber ( Fu & Park 1995). Taken as a whole, results from other species suggest that specific isoforms of Sus are associated with sucrose mobilization in sink organs, and with phloem loading in vascular tissue. The isoform affected by the rug4 mutations in pea plants would fall into the former category.

In several plant organs, Sus is reported to occur both in the cytosol and associated with the plasmalemma. In the latter location it may form part of a cellulose synthase complex ( Amor et al. 1995 ;Carlson & Chourey 1996). There is evidence that partitioning of the enzyme between cytosol and plasmalemma is controlled by phosphorylation of the protein ( Winter et al. 1997 ). Recent studies of maize carrying mutations in one or both of the genes encoding Sus (Sh1 and Sus1) indicate that the product of the Sh1 gene is the more important in cellulose synthesis (sh1 but not sus1 mutants display cell degradation in the developing endosperm consistent with deficiencies in cell wall synthesis), whereas the Sus1 product is the more important in starch synthesis ( Chourey et al. 1998 ). We are investigating whether rug4 mutant plants display deficiencies in cellulose synthesis in the developing embryo.

Given that Sus is the only enzyme on the pathway of starch synthesis in the embryo affected by the rug4 mutations, it is reasonable to assume that the reduced starch content of the mutant embryo is directly attributable to the reduction in Sus activity. Sucrose mobilization in the mutant embryo may be attributable to the residual Sus activity (0.09–0.17 μmol min–1 g–1 fresh weight: Table 1), which is probably greater than the net rate of sucrose mobilization in these embryos. Based on a rate of starch synthesis of 0.08 μmol hexose units min–1 g–1 fresh weight ( Smith et al. 1989 ) and the fact that approximately half of the sucrose metabolized by the embryo is converted to starch ( Edwards & ap Rees 1986a), the net rate of sucrose mobilization in a wild-type embryo is 0.08 μmol min–1 g–1 fresh weight. This may underestimate the flux of carbon from sucrose to hexose phosphate since developing embryos contain sufficient activities of sucrose phosphate synthase to catalyse a significant rate of sucrose synthesis. Alkaline invertase and hexokinase could also make significant contributions to sucrose mobilization in the mutant embryo.

The lack of effect of the rug4 mutations on enzymes other than Sus in the developing embryo is interesting in the light of reports that changes in Sus activity bring about large changes in activities of other enzymes in both potato tubers and the developing endosperm of maize. Potato tubers in which Sus activity has been dramatically reduced by expression of antisense RNA display increases of up to seven- and 40-fold in the activities of alkaline and soluble acid invertase, respectively, and changes in transcript levels of other enzymes of sucrose and starch metabolism ( Zrenner et al. 1995 ). Reductions in Sus activity in developing maize endosperm brought about by sh1 mutations are accompanied by several-fold changes in the activities of several other enzymes of sucrose and starch metabolism and of glycolysis ( Singletary et al. 1997 ). Reductions in Sus in both potato tuber and maize endosperm cause large increases in soluble sugar concentrations, and these may bring about changes in activities of other enzymes through effects on gene expression ( Singletary et al. 1997 ;Zrenner et al. 1995 ). The rug4 mutations cause only small changes in sugar levels in immature embryos (a 20–25% increase in sucrose and smaller changes in hexoses: A. Déjardin, unpublished data). Differences between plant organs in the extent to which reduction of Sus activity affects the activities of other enzymes may thus be due at least in part to differences in the extent of elevation of sugar levels in these organs.

The rug4 mutations have little or no direct effect on the metabolism of leaves. Mutant plants in the greenhouse are indistinguishable from wild-type plants in appearance and in the chlorophyll content and NADP-glyceraldehyde 3-phosphate dehydrogenase activity of their leaves. Activities of Sus in their leaves are reduced by only 30–40%. However, leaves of mutant plants grown in the field show yellowing, starch accumulation and much-reduced expansion. These symptoms are characteristic of nitrogen deficiency, and our studies show that the mutant plants in the field have a considerably lower ratio of nitrogen to carbon than wild-type plants grown under the same conditions. Measurements of δ15N indicate strongly that the reduced nitrogen to carbon ratio of the mutant plants is due to a severe reduction or absence of fixation of atmospheric N2 in the root nodules. Whereas wild-type plants have δ15N values close to that of atmospheric N2 (0‰), and within the range normally attributed to strong reliance on N2 fixation ( Handley & Scrimgeour 1997), mutant plants have values similar to those of non-leguminous plants, which obtain their nitrogen from sources in the soil.

The failure of nodules on mutant plants to assimilate nitrogen is probably attributable to the much-reduced activity of Sus in these organs. Two other key enzymes involved in provision of carbon skeletons to the bacteroids and for ammonia assimilation are unaffected by the mutation. The presence of the bacteroid marker enzyme 3-hydroxybutyrate dehydrogenase and the ability of nodules to convert acetylene to ethylene indicate that metabolically active bacteroids containing nitrogenase protein are present. Mutations leading to the development of ineffective nodules that undergo premature senescence have been described at several loci in a range of legume species ( Novák et al. 1995 ;Pladys & Vance 1993;Vance et al. 1988 ). The similarity between these effects and those caused by the mutations at the rug4 locus suggests that at least some of the mutations affect sucrose mobilization in the nodule, and may directly affect sucrose synthase.

These results confirm the view that Sus is the enzyme primarily responsible for sucrose mobilization associated with nitrogen fixation in Rhizobium nodules ( Gordon 1995). Further research is needed to establish whether the low Sus activity in mutant nodules restricts primarily the energy supply to the bacteroids or the supply of carbon skeletons for ammonia assimilation. It is also unclear why the high activity of alkaline invertase in the nodule is unable to compensate for the reduction in Sus activity, since both enzymes are found primarily in the infected region ( Gordon 1991;Gordon et al. 1992 ). The development of roots and the establishment and early maturation of the nodule are apparently normal in mutant plants. The activities of alkaline invertase and/or an isoform of sucrose synthase unaffected by the rug4 mutations thus appear to be sufficient for normal sucrose mobilization in root cells other than those directly involved in carbon metabolism associated with nitrogen fixation.

Experimental procedures

Plant material

Four near-isogenic lines of peas, carrying the wild-type and three mutant alleles at the rug4 locus (rug4-a, rug4-b, rug4-c) and described in Wang et al. (1990 ) and Wang & Hedley (1993), were used in all experiments. Plants were grown in soil-based compost in a greenhouse at a minimum temperature of 15°C (day) and 10°C (night), fed weekly with a nitrogen-containing fertilizer, and provided with supplementary illumination (16 h day–1) in winter. Material was harvested onto ice and extracted within 30 min.

The four lines were sown in the field as a square block of 400 evenly spaced plants, 10 cm apart. The block contained four squares, each of 100 plants of one line. Plants in the outside rows of the block were not used in experiments. Plants for analysis were selected randomly from within the squares. All measurements and harvests for N analyses were carried out at the same time for the four genotypes, when pods had started to set at 1–2 nodes. No fertilizer was applied during the growing season.

For study of nodule metabolism, plants were grown on agar in sterile conditions and inoculated with Rhizobium leguminosarum strain B556 ( Sindhu et al. 1990 ) according to Smith (1985).

Extraction of plant organs for measurement of enzyme activities

Plant material was homogenized in an ice-cold medium with a pestle and mortar and/or an all-glass homogenizer, and the homogenate centrifuged for 3–10 min at 10 000 g and4°C to produce a soluble extract. Extraction media were: embryos and testas, 50 m m HEPES, pH 7.5, 1 m m DTT, 1 m m EDTA, 100 g l–1 polyvinylpolypyrrolidone; leaves, 100 m m MOPS (pH 7.2), 2 m m DTT, 1 m m EDTA, 100 g l–1 polyvinylpolypyrrolidone; nodules, 100 m m MOPS (pH 7.2), 10 m m DTT, 5 m m MgCl2, 200 ml l–1 ethanediol.

Enzyme assays

All assays were at 25°C and were on soluble extracts except those for total starch synthase and starch-branching enzyme which were on homogenates, and that for insoluble acid invertase for which the pellet after centrifugation of the homogenate was washed twice by resuspension in extraction medium and centrifugation, then resuspended in extraction medium. For measurements in Fig. 1, soluble extracts were first desalted on a column of Sephadex G25. For each enzyme, concentrations of assay components and the pH were optimized to give the maximum rate for extracts of each plant organ. Rates were linear with respect to time and proportional to amounts of extract over the ranges used in the assays. References from which assays were modified are given in parentheses.

Figure 1.

Maximum catalytic activities of enzymes that interconvert sucrose and hexose phosphates in developing embryos and testas of wild-type and rug4 mutant peas.

Activities of sucrose synthase (a), alkaline invertase (b), soluble acid invertase (c), sucrose phosphate synthase (d), UDPglucose pyrophosphorylase (e), glucokinase (f), fructokinase (g) and insoluble acid invertase (h) were measured in extracts of developing embryos (A) and testas (B) from a single batch of wild-type (▪) and rug4 mutant (rug4-b;□) plants. Sucrose synthase was assayed in the synthetic direction. Each symbol represents a value from a single extract, and each extract was made with embryos or testas from a different pod.

Sucrose synthase (synthetic) ( Salerno et al. 1979 ). The assay mixture contained 100 m m 3-[dimethyl(hydroxymethyl)methylamino]-2-hydroxypropanesulphonic acid (AMPSO, pH 9.4), 10 m m fructose, 10 m m UDP[14C]-glucose (11.2 GBq mol–1) and 25 μl extract in a total of 55 μl. The reaction was incubated for 20 min, heated to 100°C for 2 min and processed by the resin method described by Jenner et al. (1994 ) for soluble starch synthase. For wild-type nodules, the buffer was 100 m m AMPSO, pH 9.9.

Sucrose synthase (cleavage) ( Ross & Davies 1992). The assay mixture contained 20 m m Tris (pH 6.5), 100 m m sucrose, 0.5 m m UDP and 25 μl extract in a total of 500 μl. The reaction was incubated for 15 min, heated to 100°C for 2 min, and assayed spectrophotometrically for fructose by a modification of the method of Lowry & Passonneau (1972).

Alkaline invertase. The assay mixture contained 100 m m HEPES (pH 7.5), 200 m m sucrose and 10 μl extract in a total of 100 μl. The reaction was incubated for 180 min, heated to 95°C for 4 min, and assayed spectrophotometrically for glucose and fructose. Nodule extracts were assayed by the direct spectrophotometric method of Gordon (1991).

Soluble and insoluble acid invertase. The assay mixture contained 20 m m Na acetate (pH 4.7), 100 m m sucrose and 30 μl extract in 100 μl. Reactions were incubated for 60 min, mixed with 10 μl 1 m Na phosphate, heated to 95°C for 3 min and assayed for glucose and fructose.

Fructokinase (spectrophotometric). The assay mixture contained 100 m m Tris (pH 8.1), 1 m m NAD, 1.5 m m MgCl2, 1 m m ATP, 0.4 m m fructose, 5 units glucose-6-phosphate dehydrogenase (from Leuconostoc mesenteroides), 3.5 units phosphoglucomutase and 25 μl extract in a total of 1 ml.

Glucokinase. As above except 1 m m glucose instead of fructose.

UDPglucose pyrophosphorylase ( Smith et al. 1989 ; spectrophotometric). The assay mixture contained 100 m m HEPES (pH 7.6), 0.5 m m Na pyrophosphate, 1 m m MgCl2, 0.4 m m NAD, 0.8 m m UDP-glucose, 10 units glucose 6-phosphate dehydrogenase (from Leuconostoc mesenteroides), 4 units phosphoglucomutase and 10 μl extract in a total of 1 ml.

Sucrose phosphate synthase ( Salerno et al. 1979 ). The assay mixture contained 50 m m HEPES (pH 7.5), 8 m m fructose-6-phosphate, 8 m m glucose-6-phosphate, 8 m m UDP[14C]-glucose (11.2 GBq mol–1) and 25 μl extract in a total of 60 μl. The reaction was incubated for 20 min, mixed with 10 μl 100 m m glycine (pH 10) and heated to 95°C for 2 min. The mixture was incubated with 1 unit alkaline phosphatase (calf intestine) for 30 min at 37°C then processed by the resin method as for sucrose synthase.

Alkaline pyrophosphatase. As in Gross & ap Rees (1986).

Phosphoglucomutase. As in Foster & Smith (1993).

ADPglucose pyrophosphorylase. As in Smith et al. (1989 ).

Starch synthase. As in Jenner et al. (1994 ).

Starch-branching enzyme. As in Smith (1988).

Phosphofructokinase, phosphoenolpyruvate carboxylase and β-hydroxybutyrate dehydrogenase. As in Smith (1985).

Isolation of cDNA clones

A λgt10 library prepared from cDNA from pea leaves was kindly provided by Dr Sam Zeeman (John Innes Centre). The library was screened with a 1 kb EcoR1/Xho1 fragment of the cDNA for Sus from Lotus japonicus (Skøt et al. 1996). After hybridization at 60°C, filters were washed twice for 30 min at 60°C with 2× SSC and 5 g l–1 SDS (SSC is 0.15 m NaCl, 0.015 m Na citrate). Seventeen positive clones were identified. Each was either subcloned into pBluescript or amplified by PCR then subcloned into vector pCR2.1 (Invitrogen). The sequence of each clone was determined using a dye terminator cycle sequencing ready reaction kit (Perkin-Elmer, Norwalk, Connecticut, USA) and an ABI sequencer (Applied Biosystems, Warrington, Cheshire, UK). Sequence data were analysed using UWGCG computer programs ( Devereux et al. 1984 ). A full-length clone of 2.8 kb was identified. It contained sequence identical to that of the 800 bp fragment of the pea Sus1 gene reported by Déjardin et al. (1997 ).

Co-segregation analysis

Genomic DNA was extracted from pea leaves, purified and concentrated according to Ellis (1994). DNA digested with restriction enzymes was separated on 8 g l–1 agarose gels and blotted onto Duralon-UV membrane (Stratagene, La Jolla, California, USA). The radioactive probe was a 1.5 kb NcoI fragment of the Sus1 gene of pea, prepared according to Feinberg & Vogelstein (1984). Blots were washed with 0.1× SSC and 5 g l–1 SDS at 65°C.

Isolation and sequencing of cDNA from rug4-b

Total RNA was isolated from 2 g of pea embryos (from seeds of 200–300 mg fresh weight) according to Edwards et al. (1995 ). Polyadenylated RNA was purified using oligo(dT)-cellulose spin columns (Pharmacia Ltd, Milton Keynes, UK). First-strand cDNA was synthesized using a cDNA synthesis kit (Amersham International, Amersham, Buckinghamshire, UK). Three pairs of oligonucleotides were synthesized to generate three overlapping fragments by PCR. PCR was performed using Pfu DNA polymerase (Stratagene) according to the manufacturer's instructions. Each fragment was cloned into vector pGEM-T Easy (Promega Ltd, Southampton, UK). The sequence of three clones from each fragment was determined as above. For each of the three fragments, the sequence of the three clones was identical.

Assay of acetylene reduction activity

The root systems of intact plants, grown as described above for study of nodule metabolism, were sealed into the 250 ml flasks in which they were growing using Silicoset (Ambersil Ltd, Bridgewater, UK) and supplied with acetylene (5% of flask volume) via a syringe needle. After 15 min, samples of gas were withdrawn and assayed for ethylene by gas chromatography ( Hardy et al. 1968 ).

Light microscopy

Nodules were fixed overnight at room temperature in 10 g l–1 glutaraldehyde, 40 g l–1 formaldehyde in 100 m m Na phosphate (pH 7.3). Samples were dehydrated in a graded ethanol series and embedded in LR White resin containing 5 g l–1 Na benzoin methyl ether. Polymerization was by UV irradiation for 24 h at room temperature. Sections (0.5 μm thick) were counter-stained with basic fuschin.

Nitrogen analyses

Freeze-dried material was ground into a fine powder and analysed by continuous-flow mass spectrometry according to Handley et al. (1993 ). Analysis of nitrogen content was done on a Europa Scientific Model 20–20 mass spectrometer, and analysis of δ15N was done on a Europa Scientific Tracer Mass Spectrometer, with a Roboprep combustion unit.

δ15N was calculated as [(δ15Nsample–δ15Nstandard)/δ15Nstandard] × 10–3. The δ15N values of shoots of pea plants were calculated as [(δ15Nleaf × mg Nleaf) + (δ15Nstem × mg Nstem)]/mg Nleaf + stem. δ15N values of non-leguminous weeds were measured on the pooled, freeze-dried shoots of three plants of each species. The value given for each species is the mean of two analyses carried out on the pooled sample.


We are grateful to Drs Tony Gordon (IGER, Aberystwyth, UK) Nick Brewin and Allan Downie (John Innes Centre) for advice on Rhizobium nodules, and to Drs Nick Brewin, Rod Casey and Kay Denyer for their comments on this manuscript. J.C. and C.D.G. thank the John Innes Foundation for Research Studentships, and A.D.'s work was funded via a Biotechnology and Biological Sciences Research Council (UK)/INRA(France) collaboration. The pea mutagenesis programme at the John Innes Centre received support from the Ministry of Agriculture, Fisheries and Food.

EMBL databank accession number AJ01208 (pea sus1 cDNA).