A Legume Ethylene-Insensitive Mutant Hyperinfected by Its Rhizobial Symbiont

See allHide authors and affiliations

Science  24 Jan 1997:
Vol. 275, Issue 5299, pp. 527-530
DOI: 10.1126/science.275.5299.527


Development of the Rhizobium-legume symbiosis is controlled by the host plant, although the underlying mechanisms have remained obscure. A mutant in the annual legume Medicago truncatula exhibits an increase of more than an order of magnitude in the number of persistent rhizobial infections. Physiological and genetic analyses indicate that this same mutation confers insensitivity to the plant hormone ethylene for multiple aspects of plant development, including nodulation. These data support the hypothesis that ethylene is a component of the signaling pathway controlling rhizobial infection of legumes.

In contrast to pathogenic plant-microbe interactions where persistent infection is correlated with cellular dysfunction and disease, compatible rhizobia trigger morphogenesis of a nodule organ and symbiotic nitrogen fixation on their legume host plant. Despite the beneficial aspects of this symbiosis, rhizobial infection is regulated by the plant host. One mechanism for controlling infection by compatible rhizobia, referred to as feedback inhibition of nodulation, is evidenced as a transient susceptibility to rhizobial infection in root hair cells (1). This transient susceptibility results in a narrow zone of infection and nodule differentiation (Fig. 1A). Plant mutants defective in feedback inhibition of nodulation continue to produce nodules from newly developed root tissue (2). A possible second mechanism for controlling rhizobial infection involves the early arrest of rhizobial infections within the nodulation zone; in fact, only a minority of rhizobial infections persist to colonize differentiating nodule tissue. Vasse et al. (3) observed that many such infections arrest after infection structures (infection threads) penetrate one to several cells, and plant cells containing arrested infections often display characteristics of induced host defense mechanisms (3, 4).

Fig. 1.

Distribution of R. meliloti infections and nodules on M. truncatula roots. (A) Nodulation zone on roots 10 days after inoculation with R. meliloti. The inset shows an expanded view of the nodulation zone. (B to D) Arrested [(B) and (C)] and successful (D) infections, respectively, within the nodulation zone 120 hours after inoculation with R. meliloti. Blue coloration identifies bacteria expressing the lacZ gene within infected root hair cells (arrow). In (D), P denotes a nodule primordium colonized by bacteria from a single infection event (arrow). (E) Distribution of lacZ-stained infections 48 hours after inoculation with R. meliloti.

A possible clue to the physiology underlying rhizobial infection arrest comes from the observation that inhibitors of ethylene biosynthesis, such as aminoethoxyvinyl glycine (AVG), cause an increase in persistent rhizobial infections (5, 6). Certain rhizobia produce rhizobitoxine, an analog of AVG, although roles in nodulation or regulation of ethylene synthesis in plants have not been demonstrated (7). Conversely, application of ethylene reduces nodulation in wild-type pea (8), and ethylene has been implicated as a second signal in the inhibition of nodulation by both light and nitrate (9). Thus, ethylene may provide an endogenous signal for regulation of rhizobial infection, and plant mutants with defects in production or transduction of the ethylene signal might be expected to have correspondingly altered infection phenotypes.

We screened a population of ethylmethane sulfonate (EMS)-mutagenized seedlings of M. truncatula for defects in symbiotic interactions. Using a visual assay for nodulation, we identified putative symbiotic mutants, including non-nodulators and those with altered nodule development (10). To determine whether any of these mutants were defective in rhizobial infection, we transformed the bacterial symbiont Rhizobium meliloti with a constitutively expressed lacZ gene (11). This modified strain converts the colorless X-Gal substrate to an insoluble blue precipitate and thus provides a visual assay for rhizobial infection (12). Initially we characterized infection in wild-type M. truncatula. We determined that initiation of infection was complete by 48 hours, resulting in 100 to 300 visible infections within the nodulation zone (Fig. 1E). At 72 hours, an average of eight macroscopic nodule primordia were evident, each of which were associated with extensive rhizobial infection (Fig. 1D) and subsequently developed into mature nitrogen-fixing nodules. Infection events not associated with nodule morphogenesis were typically arrested in the root epidermis (Fig. 1, B and C). Thus, the infection efficiency of wild-type M. truncatula is low, with 3 to 8% of the infections persisting to colonize a nodule organ.

We used this modified Rhizobium strain to assess infection in the M3 and F2 backcross generations of previously identified nodulation mutants. One such mutant, named sickle for its unusual sickle-shaped zone of nodulation (Fig. 2A), contained an increase of more than an order of magnitude in the number of sustained infections relative to the wild type (Fig. 2, B and C). The sickle mutation affects the number of persistent infections without altering the transience of root susceptibility (Fig. 2A); thus, the corresponding gene is implicated in arrest of rhizobial infection within the nodulation zone, but not in feedback inhibition of nodulation. Sustained infections in sickle were characterized by infection threads that ramified into the root cortex toward developing nodule primordia (Fig. 2D), similar to sustained infection in wild-type plants. Cytological, physiological, and molecular analyses (Fig. 2, E and F, Table 1, and Fig. 3) revealed apparently normal differentiation of both symbionts, although nodule morphogenesis in sickle was retarded relative to the wild type (13).

Fig. 2.

Infection and nodulation phenotype of the M. truncatula mutant sickle. (A) Nodulation zone on sickle roots 10 days after inoculation with R. meliloti. Note that the zone of infection and morphogenesis in sickle (inset) is confined to a narrow region, similar to the wild type (Fig. 1A). (B and C) Nodulation zone on sickle and wild-type plants, respectively, 96 hours after inoculation. Persistent infections are visible as blue-staining regions. (D) Hand section through a 96-hour infection on a sickle root comparable to that shown in (B). The branched infection thread (arrow) has ramified into the root inner cortex. X, xylem tissue in the root vasculature. (E) Bright-field micrograph showing tissue differentiation typical of 21-day-old sickle nodules. X, root xylem tissue; C, nodule central tissue; E, nodule endodermis; M, nodule meristem. (F) Enlargement of (E) showing a gradient of cell differentiation from the infection zone through the adjacent nitrogen fixation zone. The arrow indicates an infection thread in a cell of the infection zone; “i” identifies an infected cell within the nitrogen fixation zone. Scale bars: 20 μm (E) and 120 μm (F).

Table 1.

Nitrogenase activity in wild-type M. truncatula and in the nodulation mutants, sickle and domi. domi is a non-nodulating mutant of M. truncatula that is resistant to infection by Rhizobium (10). Nitrogenase activity was determined by the acetylene reduction assay (27), where nitrogenase enzyme reduces substrate acetylene to ethylene. Acetylene reduction was measured on nine roots for each genotype, 19 days after inoculation with Rhizobium.

View this table:
Fig. 3.

Leghemoglobin expression in wild-type M. truncatula and nodulation mutants. Lanes contain total RNA from roots 23 days after inoculation with Rhizobium [lbg, leghemoglobin (25); H3, control histone expression (26)]. Comparable root tissue was assayed for nitrogenase activity (Table 1) and tissue differentiation (Fig. 2, E and F).

In a manner consistent with a defect in ethylene production or perception (14, 15), sickle plants were pleiotropic for delayed petal (Fig. 4, A and B) and leaf senescence and for decreased abscission of seed pod and leaves. We assayed the sensitivity of wild-type and sickle seedlings to 1-aminocyclopropane carboxylic acid (ACC), the immediate precursor of ethylene, and to ethylene gas. Both ACC and ethylene induced the “triple response” (14, 15, 16) in wild-type seedlings, including inhibition of root and hypocotyl elongation and formation of a hypocotyl hook (Fig. 4C), whereas sickle seedlings were insensitive to both compounds even at >10 times their median effective dose (ED50) values in the wild type (Fig. 4, D and E). Taken together, the ethylene and ACC results indicate that sickle, like mutants in Arabidopsis and tomato (14, 17, 18, 19), is defective in perception of the ethylene signal.

Fig. 4.

Ethylene-related phenotypes in wild-type and sickle genotypes of M. truncatula. (A and B) Normal and delayed petal senescence (arrows) in the wild type and sickle, respectively, 7 days after pollination. P, immature pod. (C and D) Sensitivity of 5-day-old wild-type (C) and sickle (D) seedlings to exogenous ACC (0 to 100 μM, at the values shown in parentheses). The triple response of the wild type to ACC is evidenced by shortened hypocotyls and roots and by hypocotyl hook formation. (E) Hypocotyl growth response of wild-type (solid bars) and sickle (open bars) seedlings to exogenous ACC and to ethylene gas.

To determine the sensitivity of nodulation in wild-type and sickle plants to ACC, we grew seedlings in growth pouches and added ACC directly to the growth medium at various times after inoculation with Rhizobium. When ACC was added 24 to 48 hours after inoculation, nodulation of wild-type plants was effectively blocked (ED50 ≤ 5 μM; Fig. 5A). Similar treatment of sickle failed to inhibit nodulation even at 300 μM ACC (Fig. 5, B to E). After macroscopic nodule primordia were evident (72 hours), continued nodule development on the wild type was largely insensitive to exogenous ACC (Fig. 5A).

Fig. 5.

Effect of ACC on nodulation in wild-type and sickle genotypes of M. truncatula. (A) Suppression of nodulation in the wild type by ACC as a function of symbiotic development. Nodulation is efficiently suppressed when ACC is applied during the primary infection phase (24 and 48 hours), but not when ACC is applied after the appearance of macroscopic nodule primordia (72 hours). Stippled bars, no ACC; hatched bars, 5 μM ACC; open bars, 25 μM ACC; solid bars, 100 μM ACC. (B to E) Nodulation in the absence (B and D) or presence (C and E) of 300 μM ACC. ACC treatment inhibits nodulation in the wild type (B and C) but not in sickle (D and E). Sustained infections are visible as blue-staining tissue.

Genetic analysis indicates that the hyperinfectable and ethylene-insensitive phenotypes of sickle are determined by a single, recessive allele. We assayed cosegregation of hyperinfectability and ethylene-related phenotypes by testing F2 progeny sequentially for nodulation and ethylene-induced chlorophyll loss (20). From a total of 201 F2 individuals, 50 were hyperinfected and ethylene-insensitive, and 151 were normally infected and ethylene-sensitive (P = 0.97, χ2 = 0.0017). We have designated the corresponding M. truncatula gene as skl1 and the sickle allele as skl1-1. The pleiotropic nature of skl1-1 indicates that skl1 acts in the ethylene perception pathway, similar to mutants in Arabidopsis, such as ein2 and ein3 (19). The recessive nature of skl1-1 indicates that skl1 is probably not a member of the ethylene receptor gene family (including ein1/etr1), because all known mutants in this family have dominant phenotypes (17, 18).

Our results support the hypothesis that ethylene is involved in controlling the persistence of rhizobial infection. Ethylene is known to control differentiation of root hair cells (21, 22), the cell type infected by Rhizobium. Thus, endogenous ethylene may affect the persistence of rhizobial infection by controlling the formation of infectable root hair cells. Alternatively, ethylene may act as a diffusible signal for activation of mechanisms that arrest rhizobial infection. ACC is inhibitory to nodulation when applied after the initiation of rhizobial infection (Fig. 5A, 24 and 48 hours). Similarly, the ethylene biosynthesis inhibitor AVG can increase nodule number when applied after the initiation of infection (5). These observations are consistent with a model wherein endogenous ethylene acts subsequent to infection initiation and root hair differentiation.

If ethylene provides a signal for induction of infection arrest, then plant cells containing persistent Rhizobium infections either must avoid the ethylene signal or must be insensitive to the signal. Localized production of ethylene at sites of infection arrest could facilitate avoidance of ethylene by infections destined for nodule colonization. A model for cell-specific regulation of ethylene synthesis during root hair cell differentiation has been proposed in Arabidopsis (22). In wild-type M. truncatula, all rhizobial infections can be blocked by treatment with ACC as late as 48 hours after inoculation (Fig. 5A), indicating that infections destined for nodule colonization are not inherently insensitive to ethylene. However, after macroscopic nodule primordia appear, nodulation is largely insensitive to exogenous ACC (Fig. 5A, 72 hours); thus, sustained rhizobial infections may acquire insensitivity to ethylene.

In plant-pathogen interactions, ethylene has been implicated as an endogenous cue for induction of host defense-related genes (23). Despite extensive correlative data, however, a causal role for ethylene in resistance to pathogens has not been established (24). In M. truncatula, the sickle mutation causes extensive developmental abnormalities and hyperinfection by Rhizobium, which indicates that skl1 encodes a function common to both plant development and control of rhizobial infection.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
View Abstract

Navigate This Article