E.P. Journet, M. Pichon, A. Dedieu, F. de Billy, ... the host plant during the earliest stages of the ... a developmental stage preceding root hair formation.
The Plant Journal (1994) 6(2), 241-249 SHORT COMMUNICATION
Rhizobium meliloti Nod factors elicit cell-specific transcription of the ENOD12 gene in transgenic alfalfa E.P. Journet, M. Pichon, A. Dedieu, F. de Billy, G. Truchet, and D.G. Barker* Laboratoire de Biologie Moldculaire des Relations Plantes-Microorganismes, INRA-CNRS, BP 27, 31326 Castanet- Tolosan C6dex, France Summary Extracellular lipo-oligosaccharides of Rhizobiurn, known as Nod factors, play a key role in the molocular signal exchange which leads to the specific nitrogen-fixing symbiotic association between the soil microbe and its host legume. The biological activity of Nod factors and their perception by the host plant during the earliest stages of the Rhizobiurrdlegume intel;sction have been studied using transgenic alfalfa carrying a fusion between the promoter of the early nodulin gene MtENOD12 and the ~-glucuronidase (GUS) reporter gene. Histochemical staining has shown that GUS accumulates specifically in the differentiating root epidermis, prior to and during root hair emergence, within 2-3 h following the addition of purified Rhizobium meliloti Nod factors. This precocious transcriptional activation of the MtENOD12 gene, reminiscent of that observed after inoculation with intact Rhizobium, implies that the Nod factor signal can be perceived at a developmental stage preceding root hair formation. GUS activity can be detected following treatment with a wide range of R. meliloti Nod factor concentrations down to 10-13 M, and furthermore, this rapid response to the bacterial elicitor appears to be non-systemic. Significantly, MtENOD12-GUS expression is not observed after inoculation with a R. meliloti nodH mutant which synthesizes exclusively non-suIphated Nod factors. Indeed purified Nod factors which lack the sulphate substituent are approximately 1000-fold less active than their suIphated counterparts. Thus, the triggering of ENOD12 transcription in the alfalfa root epidermis is a rapid molecular response which is subject to the same host-specificity determinant (Nod factor suIphation) that governs the interaction between alfalfa and its bacterial symbiont. Received 15 December 1993; revised 24 March 1994; accepted 28 April 1994. *For correspondence (fax +33 61 28 50 61).
Introduction The symbiotic association between leguminous plants and the soil bacteria known as rhizobia solicits the formation of novel root organs (nodules) in which atmospheric nitrogen is converted to ammonia by the microsymbiont, thereby providing the host plant with an alternative to soil-based inorganic nitrogen. An important characteristic of symbioses involving temperate legumes is the high degree of specificity; alfalfa is nodulated by strains of Rhizobium meliloti, and pea, vetch and clover are nodulated by biovars of R. leguminosarum. Current interest in this beneficial plant-microbe interaction has focused on the discovery that there is an exchange of signals between the two organisms during very early stages of the interaction and that this is critically important in ensuring the specificity of the association and in preparing the plant for subsequent infection and nodulation. The molecular dialogue between symbiotic partners is initiated when root-secreted plant flavonoids act as transcriptional activators of the Rhizobium nodulation (nod) genes. Following this step, proteins encoded by the activated nod genes contribute to the synthesis of extracellular lipo-oligosaccharide molecules known as Nod factors, which subsequently play a decisive role in determining the outcome of the symbiotic interaction (for reviews see D~nari~ and Roche, 1992; Fisher and Long, 1992). The purification of Nod factors from different rhizobial species has shown that the common structural feature of these molecules comprises an oligosaccharide (chitin) backbone of 6-1,4-1inked N-acetyI-c)-glucosamine (3-5 units) with a partially unsaturated fatty acid chain (C16 or C18) N-acyl linked to the terminal non-reducing sugar residue (Figure 1; for review see Spaink, 1992). Rhizobium strains which are mutated in the so-called common nod genes (nodA, nodB or nodC) no longer produce Nod factors and are unable to elicit the usual symbiotic responses such as root hair curling, infection thread formation and nodule induction. In the case of R. meliloti, the host-specificity nodH (Debell~ et aL, 1986) and nodPQ (Schwedock and Long, 1989) genes are responsible for the sulphation of the terminal reducing sugar of the Nod factor (Roche etaL, 1991a), while nodL determines O-acetylation of the non-reducing sugar (Baev and Kondorosi, 1992; Ardourel et aL, in preparation) and nodFE the structure of the N-acyl derivative (Demont 241
242
E.P. Joumet et al. t de-O.mcetyl~on m~/L"
[ r /
14o
C16:0
H2
~
~ c.=p~
!
nodH"
.o-I1- s%. .
c'. 2
|
/
i
t
c.%' "'
J,
:o
C16:2/ t ; =)'
F i g u r e 1. Structures of lipo-oligosaccharide Ned factors tested for activity in eticiting ENOD12-GUS expression in roots of transgenic alfalfa. Chemical modifications to the predominant R. meliloti NedRm-IV(Ac, S) factor involve the de-O-acetylation of the non-reducing sugar residue and the hydrogenation of the C16:2 lipid moiety. Also shown are those modifications which result from mutations in various nod genes (nodFE, nodH and nodL). The C18:1 lipid which is synthesized by the nodFE mutant is cis-vaccenic acid (Demont et aL, 1993).
et aL, 1993). The nature and position of such sugar decorations differ between Rhizobium species and appear to be responsible for the biological activities and specificities of these molecules (reviewed in D~nari~ and Roche, 1992). For example, the R. meliloti host-specificity nodH mutant is unable to infect and nodulate the normal host alfalfa, but has acquired the ability to interact symbiotically with a non-homologous host, common vetch (Faucher et aL, 1988). Since non-sulphated Nod factors, produced by noclH- strains, are inactive in biological assays on alfalfa, but are active on common vetch, it has been concluded that the nodH gene mediates host specificity in R. melilotias a direct consequence of its role in the sulphation of the extracellular Nod factors (Roche et al., 1991a). Purified Nod factors are able to elicit morphogenic changes in the host plant both at the root surface (root hair deformations; Lerouge et aL, 1990; Spaink et aL, 1991), and in the root outer cortex (formation of preinfection thread structures; van Brussel et aL, 1992), as well as inducing mitotic activity within the root inner cortex (Spaink et al., 1991; Truchet et aL, 1991). In the case of the interaction between alfalfa and its microsymbiont R. meliloti, the hair-deformation response can be evoked at NodRm factor concentrations as low as 10 -12 M (Roche et aL, 1991a). The same factors, added at significantly higher doses (10-9--10 -7 M) can also stimulate mitotic activity in the normally quiescent cortical cell layers, with the formation of non-nitrogen fixing nodular structures (Truchet et aL, 1991). In order to investigate the mechanisms by which the plant perceives and responds to these symbiotic bacterial elicitors, we have attempted to identify plant genes which could act as molecular markers for Nod factor activity. We have focused on the early nodulin gene, ENOD12,
originally identified in pea (Scheres et aL, 1990), and more recently in two Medicago species (Allison et al., 1993; Pichon et aL, 1992). This gene encodes a putative proline-rich cell wall protein and in situ hybridization experiments have shown that ENOD12 is expressed during the infection process, both in cortical cells traversed by the infection thread and in the infection zone of the nodule (Pichon et aL, 1992; Scheres et aL, 1990). It was further shown that ENOD12 transcripts were present in total RNA preparations extracted from pea root hairs following inoculation with the pea microsymbiont, R. leguminosarum bv. viciae (Scheres et aL, 1990), and more recently after treatment with purified preparations of Nod factors (Horvath et al., 1993). The symbiotic interaction between R. meliloti and plants of the genus Medicago was chosen as our experimental system by virtue of the highly specific nature of the association, the availability of both purified Nod factors and bacterial nod gene mutants, and the possibility of transforming and regenerating certain genotypes of alfalfa such as Medic,ago varia (Deak et al., 1986). By cloning an analogous gene (MtENOD12) from the diploid self-fertile M. truncatu/a (Barker et aL, 1990), we were able to construct transgenic alfalfa plants which expressed the Eschedchia coil ~-glucuronidase gene (uidA) under the control of the MtENOD12 promoter (Pichon et a/., 1992). This strategy allowed us to monitor the spatiotemporal expression of the ENOD12 gene throughout the entire root system following bacterial inoculation and showed that ~-glucuronidase (GUS) activity could be detected as early as 3-6 h after inoculation within a zone of differentiating root epidermal cells which lies just behind the growing root tip. Using the same transgenic alfalfa we now demonstrate that the early activation of the ENOD12 gene in the differentiating zone of the plant root epidermis can also be elicited by the exogenous addition of purified R. me/i/oti Nod factors. These symbiotic signalling molecules are active down to 10 -13 M, and localized application of Nod factors to the root reveals that this rapid molecular response is non-systemic. We have investigated the effect of modifying various structural components of the Nod factor, and have found that the removal of the sulphate group on the reducing sugar (a key determinant of host specificity for R. me/i/otl) causes a dramatic reduction in the capacity to trigger ENOD12 transcription in the host epidermis. Results
NodRm factors elicit ENOD 12 gene expression in the alfalfa root epidermis In order to study the effect of R. me/iloti Nod factors on early nodulin gene expression, purified factors were
Rhizobium Nod factors elicit ENOD 12 transcription added directly to the growth medium of 8-12 day-old seedlings of transgenic M. varia in growth pouches, and the entire root system was harvested for histochemical staining for GUS. Initial experiments were carded out using R. meliloti Nod factors purified from a wild-type overproducing strain (see Experimental procedures; Roche eta/., 1991 a), essentially comprising a mixture of NodRmIV(Ac,S) (majority species; see Figure 1) and NodRmIV(S) (minority species), and subsequently referred to as 'NodRm'. As seen in Figure 2(a), a relatively short treatment (6 h) with 10-6 M NodRm is sufficient to elicit expression of the chimeric MtENOD12-GUS fusion in a region of the root which starts just behind the growing root tip and extends throughout the zone of root hair emergence and maturation. This response occurs in both primary and secondary roots, although GUS staining is generally more intense in the primary root (results not shown). Since the activation of the ENOD12 promoter by R. meliloti Nod factors primarily takes place in the epidermal call layer (Figure 2b), both the spatial distribution and the cell-specificity of this response are very similar to that observed within the 24 h period following bacterial inoculation (Pichon et al., 1992). Nod factor treatment for longer periods of time (up to 2 weeks) did not significantly alter the spatial expression pattern of the MtENOD12--GUS fusion. GUS activity could be detected in roots following incubation with Ned factors for times as short as 2-3 h, although, in this case, histochemical staining is confined to epidermal cells prior to and during the eadiest stages of root hair emergence (Figure 2c and d). Since purified Ned factors have been shown to be active in root hair deformation assays at concentrations as low as 10-12-10 -11 M (Roche eta/., 1991a; Schultze et
al., 1992; Spaink et al., 1991 ), experiments on transgenic alfalfa plants were carried out with a wide range of NodRm concentrations down to 10-14 M. The results presented in Table 1 show that reporter gene activity can be detected when plant roots are treated with picomolar concentrations of Ned factors over a 24 h period. Despite the variability in the intensity of the response between individual plants resulting from the use of the $1 seed population (see Experimental procedures), it is important to note that 100% of the plants tested with NodRm factor concentrations down to 10-12 M scored positive for GUS activity (Table 1). Although the results presented in Table 1 do not attempt to quantify reporter gene activity, it should be emphasized that there is a clear reduction in overall GUS activity in the root epidermis as the Nod factor concentration is reduced (compare Figure 2e (10-11 M) and f (10-6 M)). At the lowest concentrations the plant response was often limited to relatively few calls on the root surface. It is worth remarking that, in contrast to the hair-deformation assay (Roche et al., 1991 a; Schultze et al., 1992), there does not appear to be inhibition of expression of the ENOD12-GUS gene fusion at high (10-6-10-6 M) Nod factor concentrations (Table 1). When alfalfa seedlings were allowed to grow for 24 h on the surface of solid agar medium containing NodRm factors (see Exparimental procedures), GUS activity in the epidermis was found to be strictly limited to the underside of the root in contact with the agar surface (Figure 2g). Local application of Ned factors to the exposed surface close to the root tip by means of a small agar block showed that the corresponding epidermal cells were still perfectly reactive to the bacterial elicitors (Figure 2h). Furthermore, staining was not observed on the upper surface of the root not in contact with the agar
Table 1. Relationship between Nod factor concentration and the expression of the ENOD12-GUS fusion in the differentiating root epidermis of transgenic alfalfa Nod factor concentration (M) Elicitor
10-s
10-7
10-e
10-9
10-10
10-11
10-12
10-13
10-14
+
+
+
+
+
+
+
+
_
NOdRm-IV(Ac)a
+
+
+
+
+
-
-
nd
nd
Hydrogenated NOdRm-IV(Ac,S)a
+
+
+
+
+
+
[+]
[_]
_
NodRm-IV(S)
+
+
+
+
+
+
+
[_]
_
N, N',N", N"'-tetraacetylchitotetraose (chitin tetramer)
.
nd
nd
nd
nd
nd
NodRm-IV(Ac,S)a
.
243
.
.
Plants (between 10 and 20 for each concentration tested) were treated with both wild-type and modified NodRm factors (see text and Figure 1 for details) for a 24 h period prior to histochemicai staining (see Experimental procedures). Symbols indicate the percentage of plants which gave a positive response: + = 100%; [+] 70-100%; + = 30-70%; [-] = 0-30%; - = 0%; nd= not determined. aNod factor preparations comprising a mixture of O-acetylated (majority species) and non-O-acetylated (minority species) molecules (see text). =
244
E.P. Journet e t al.
Figure 2. Histochemical localization of GUS activity in roots of transgenic alfalfa after treatment with NodRm factors. (a) Secondary root following 6 h treatment with 10-8 M NodRm. Note that the blue coloration is present in regions of the root prior to and during root hair emergence, and extends into the zone of root hair maturation. Bar = 250 p.m. (b) Transverse section (80 Hm thick) taken from within the zone of root hair emergence of a primary root following a 6 h treatment with 10-9 M NodRm. Note that the MtENOD12-GUSfusion is strongly expressed within all epidermal cells. Bar = 100 p.m. (c and d) Secondary root after 3 h treatment with 10-8 M NodRm. Note that GUS activity is now limited to the region of the root prior to and during root hair emergence. (d) Is a magnification of (c). Bars = 250 p.m. (e and f) Secondary roots after 24 h treatment of transgenic plants with 10-11 M and 10-8 M NodRm, respectively. Bars = 250 p.m. (g) Primary root of a transgenic alfalfa plant which has grown for 24 h on solid agar medium containing 10-9 M NodRm. Note that only the underside of the root which was in contact with the agar has stained positive for GUS activity. Bar = 250 p.m. (h) Similar root to that shown in (g), except that a small agarose block containing 10-9 M NodRm was placed on the top of the root (close to the tip) so that both surfaces of the root were in contact with Nod factor during the 24 h period of treatment. Bar = 250 p.m.
Rhizobium Nod factors elicit ENOD 12 transcription block (results not shown). Taken together, this shows that the Nod factor-dependent triggering of the ENOD12 gene
is not a systemic response and most probably requires direct contact between the signal molecule and the target plant cell at the root surface. Nod factor treatment of plant organs other than roots and of cell suspension cultures derived from transgenic roots has so far failed to elicit detectable GUS activity, further demonstrating the cellspecific nature of this response (results not shown). is ENOD 12 gene transcription a host-specific response ?
As stated earlier, the major host-specificity genes in R. meliloti are nodH and nodPQ, resulting from their role in the suiphation of the NodRm factor (Roche et aL, 1991a). As a first step towards examining whether the early expression of the ENOD12 gene in differentiating epidermal cells is dependent on the same host-specificity determinants, transgenic alfalfa plants were inoculated with a R. meliloti strain carrying a Tn5 insertion in the nodH gene (see Experimental procedures). Under these conditions no trace of GUS activity could be detected in the root epidermis even 3-4 days following inoculation with the host-specificity mutant (results not shown). When experiments were then carried out using high concentrations of non-sulphated NodRm factors (purified from a nodH- strain; Roche et aL, 1991a), vCe found that such factors were indeed able to trigger ENOD12 transcription. However, subsequent analysis showed that non-sulphated factors were not active at concentrations below 10-1°-10 -° M (Table 1). This shows that the absence of the sulphate group reduces NodRm factor activity in the ENOD12 bioassay by approximately 1000-fold (statistically significant at the 5% probability level). Apart from this difference, the cell-specific nature of the response, the extent of the reactive zone on the root surface and the kinetics of the response all appeared to be identical for sulphated and non-sulphated factors (results not shown). Biological assays carried out on purified Nod factors from R. leguminosarum (Spaink et al., 1991 ) had shown that solubility problems might arise when dealing with particularly hydrophobic lipo-oligosaccharides, and that these could be overcome by the inclusion of a mild detergent. However, the addition of the detergent CHAPS at concentrations ranging from 0.01 to 0.1% during both the dilution and incubation steps (see Experimental procedures) failed to modify the concentration/activity profiles of either sulphated or non-sulphated NodRm factors (results not shown), thus arguing that the 1000-fold difference in biological activity is not an artefact due to differences in the hydrophobic properties of the two factors. Interestingly, the chitin oligomer backbone alone, comprising four [[3-1,4 linked N-acetyI-D-glucosamine residues, is not able to elicit ENOD12-GUS gene transcrip-
245
tion in transgenic alfalfa roots even at concentrations as high as 10-6 M (Table 1), thus showing that the activity of the non-sulphated NodRm factors is not merely a background response. Activities of NodRm factors modified in either the N-acyl or O-acetyl substituents
Since the early activation of the ENOD12 gene can readily distinguish between the activities of normal and non-sulphated R. meliloti Nod factors, we examined the effects of introducing other chemical modifications to the sulphated signalling molecule. It has been shown that the degree of saturation of the fatty acid chain and the presence of the O-acetyl group on the terminal nonreducing sugar can influence certain of the biological activities described for both the NodRm factors of R. meliloti (Roche et aL, 1991a; Truchet et aL, 1991) and the NodRI factors of R. leguminosarum bv viciae (van Brussel et aL, 1992; Spaink et aL, 1991). However, neither the conversion of the C16:2 fatty acid of NodRm to a saturated C16:0 chain (Figure 1; Truchet et al., 1991), nor the conversion of the majority O-acetylated fraction of the NodRm mixture to the non-acetylated form (Figure 1; Roche et al., 1991b) resulted in a major change (> 10-fold) in factor activity as assessed by the ENOD12 gene transcription assay (Table 1). The data suggest that both of these modifications lead to a slight reduction in activity in the picomolar concentration range; however, it is important to note that such differences are at the limits of statistical significance at the 5% probability level. Since deletions of either the nodE or nodF genes of R. meliloti result in the replacement of the polyunsaturated C16 fatty acid of NodRm by a C18:1 chain (Demont etal., 1993), and mutations in the nodL gene lead to the synthesis of non-O-acetylated Nod factors (Spaink et al., 1991; Ardourel et aL, in preparation), we carried out inoculation experiments on transgenic alfalfa plants using strains of R. meliloti having deletions at these two loci (see Experimental procedures). For both mutants, epidermal cell activation of the ENOD12-GUS gene was clearly observed 24 h after inoculation (results not shown), as for the wild-type strain (Pichon et aL, 1992). Thus, taking into account the negative result obtained with the nodH mutant (see above), there is a good correlation between the capacities of mutant strains secreting Nod factors with altered structures to elicit ENOD12 transcription in the root epidermis and the activities of purified factors which have been modified in the corresponding structural decoration. In contrast to the delayed nodulation phenotype previously described for R. meliloti nodFE mutants when using M. sativa cv. Gemini as host plant (Debell~ et aL, 1986), the nodFE deletion strain
246
E.P. Journet et al.
used in these experiments nodulated M. varfa with kinetics comparable with the wild-type (results not shown).
Discussion Mono-N-acylated chito-oligosaccharides (Nod factors) play a pivotal role in the establishment of the specific symbiotic association between the soil bacterium Rhizobium and its respective host legume. In this paper we describe how one particular plant gene (ENOD12) can be used as a molecular marker for an early Nod factordependent host response. By using a transgenic plant/ reporter gene strategy we have been able to show that a Medicago ENOD12 gene (originally isolated from M. truncatula; Pichon et al., 1992) is transcriptionally activated in differentiating root epidermal cells of alfalfa in response to purified R. meliloti Nod factors. A similar pattern of gene expression had been observed during the 24 h period which followed the inoculation of such transgenic alfalfa plants with R. meliloti (Pichon et al., 1992). Our results therefore provide conclusive evidence that extracellular Nod factors are indeed the chemical signals which trigger ENOD12 gene expression in the differentiating epidermis during early stages of the symbiotic interaction. Recently, Horvath et al. (1993), using a PCR-based assay, have shown that PsENOD12 mRNA accumulates in pea root hairs following a 12-48 h treatment with NodRI factors purified from the pea symbiont, R. leguminosarum bv. viciae. Our results are consistent with these findings, but show that, at least in the case of alfalfa, gene expression is not limited to root hair-bearing epidermal cells. The transgenic plant approach further reveals that the ENOD12-GUS gene is activated in a zone of the alfalfa root which corresponds to the earliest stages of root hair development (Figure 2a). The fact that detectable GUS activity is restricted to this zone when transgenic plants are treated to very short exposures (2-3 h) of NodRm factors (Figure 2c and d) suggests that maximum reactivity to Nod factors occurs in differentiating epidermal cells which lie just behind the growing root tip. Significantly, it has been shown that the series of events which eventually lead to nodule formation are initiated in the root segment that lies between the root tip and the zone of root hair emergence (Bhuvaneswari et al., 1981 ; Caetano-Anolles and Gresshoff, 1991). Our results now provide direct evidence that Nod factors elicit a very early plant response in precisely this region of the root and further show that a direct interaction between surface epidermal cells and exogenously supplied Nod factors is a prerequisite for eliciting early ENOD12 expression (Figure 2g and 2h). It remains to be seen whether putative Nod factor receptors locate predominantly to the zone of differentiating root epidermal
cells. We have previously speculated that, following bacterial inoculation, a possible role for putative cell wall proteins such as ENOD12 might be to render the root hair susceptible to Rhizobium attachment and/or infection (Pichon et al., 1992). Since it has been shown that Nod factors can stimulate root hair development specifically (Roche et al., 1991a) one plausible interpretation of our results would be that perception of the appropriate Nod factor leads to the differentiation of root hair populations which have specific properties enabling them to serve as sites for bacterial infection. As mentioned earlier, the sulphation status of NOdRm factors is a major determinant which ensures the specificity of the symbiotic interaction between alfalfa and R. meliloti (Roche et al., 1991a), as well as the activity of these factors in stimulating cortical cell divisions in the roots of alfalfa (Truchet et al., 1991). We have shown that non-sulphated factors synthesized by a R. meliloti nodHstrain are approximately 1000-fold less active in eliciting ENOD12-GUS gene transcription in transgenic alfalfa roots as compared with the normal sulphated NodRm (Table 1). This demonstrates that the activation of the ENOD12 gene in the root epidermis is a rapid molecular response to Nod factors which is subject to the main host-specificity determinant which governs the R. melilot#alfalfa interaction. Previous studies using root hair deformation as a biological assay have shown that, unlike their sulphated counterparts, non-sulphated factors were unable to elicit a Had response on alfalfa when tested in the concentration range 10-12-10 -9 M (Roche et al., 1991a). However, more recent experiments have shown that higher concentrations (10-8-10 -7 M) of non-sulphated molecules do indeed elicit hair deformations on alfalfa (Maillet and D6narid, personal communication). This demonstrates that there is a good correlation between the two biological assays on alfalfa in terms of specificity, and comparison with the data presented in Table 1 suggests that the ENOD12/transgenic plant assay is approximately 10-fold more sensitive. Using their PCR-based assay to evaluate mRNA levels, Horvath et al. (1993) were able to show that the PsENOD12 gene is transcribed in pea root hairs in response to a range of concentrations of Nod factors down to 10-1° M. The factors tested were derived not only from the normal microsymbiont /3. leguminosarum bv. viciae (producing non-sulphated factors; Spaink et al., 1991) but also from the non-homologous R. meliloti (sulphated factors). The authors note that the response to NodRm factors is delayed, and they suggest that this apparently non-specific reaction might be a consequence of an enzymatic modification leading to the loss of the sulphate group. In the case of alfalfa, it is important to emphasize that ENOD12 gene expression is not elicited by the host-specificity mutant nodH (even after lengthy
Rhizobium Nod factors elicit ENOD 12 transcription exposures to this non-nodulating strain) and that only assays at concentrations below 10-1° M clearly distinguish between the activities of the sulphated and nonsulphated Nod factors (Table 1). This suggests that the local concentration of Nod factors produced by Rhizobium at the root surface may well be below 10-1° M. These findings imply that considerable prudence should be taken in ascribing significance, in terms of host specificity, to root surface responses which are elicited at higher Nod factor concentrations. In contrast to the removal of the sulphate group, deacetylation of the O-acetylated fraction of NodRm and chemical reduction of the C16:2 fatty acid to the fully saturated C16:0 lipid chain did not lead to major changes (> 10-fold) in activity based on the transgenic plant bioassay (Table 1). This is consistent with previous findings that O-acetylation and the degree of fatty acid saturation for NodRm and NodRI have relatively little influence on hair-deformation activities (Roche et a/., 1991a; Schultze eta/., 1992; Spaink et al., 1991). On the other hand, both the presence of the O-acetyl group (in the case of NOdRI) and the unsaturation status of the lipid (C16:2 for NodRm and C18:4 for NodRI) seem to be important for eliciting cortical cell division activity in the respective host legumes (Spaink et aL, 1991; Truchet eta/., 1991). This suggests that the very early responses to Nod factors at the root surface differ from those elicited within the root cortex in terms of the structural requirements of the factor. At present very little is known about how Nod factors are perceived by the plant and how this signal is transduced by the cellular machinery, although it has recently been shown that Nod factors can cause membrane depolarization in root hair cells (Ehrhardt eta/., 1992). The rapidity with which the MtENOD12-GUS gene is expressed in the root epidermis of transgenic plants in response to Nod factors makes this an attractive system with which to study the role of secondary messengers which are involved in the signal transduction pathway. Finally, since the use of the gusA reporter gene offers the possibility of making quantitative measurements of gene expression (Jefferson eta/., 1987) it should be possible to determine dose-response curves for NodRm factors and structurally modified variants. Unfortunately, such an approach is currently impractical because of the considerable variability in GUS expression levels within the transgenic M. varia $1 seed population (see Experimental procedures). However, the recent introduction of the MtENOD12-GUS gene fusion into the diploid autogamous M. truncatula (Chabaud, Larsonneau and Huguet, in preparation), and the subsequent isolation of lines homozygous for the transgene should provide a potent bioassay for future analyses of the structure/activity relationship of the Rhizobium Nod factor.
247
Experimental procedures Transgenic plant material A number of independent primary transformant lines of Medicago varia A2 (Deak et al., 1986) had previously been obtained following the introduction of the MtENOD12-GUS gene fusion using an Agrobacteriurn tumefaciens transformation protocol and subsequent regeneration via somatic embryogenesis (Pichon eta/., 1992). Transformed lines were propagated vegetatively, grown under the appropriate conditions to induce flowering (3000 lux; 16 h/8 h photoperiod; 20°C), and then manually self-fertilized. The resulting pods contained relatively few seeds (1-4), but germination frequencies as high as 8090% were routinely obtained. Preliminary experiments had shown that Nod factors elicited qualitatively identical responses on the roots of seedlings derived from different transgenic lines. The dose-response experiments presented in Table 1 were performed using the descendants from a single transgenic line.
Bioassay for Nod factors Transgenic seedlings were placed in plastic growth pouches (throe plants per pouch; purchased from Mega International, Minneapolis) containing Fahraeus medium (Vincent, 1970) supplemented with 1 mM ammonium nitrate as starter, and grown for 8-12 days at 25°C with a 16 h/8 h photoperiod. Following the development of secondary roots, appropriate concentrations of Nod factors were added to the medium in the growth pouch and mixed by gentle agitation. After the removal of excess liquid, the pouches were returned to the growth chamber until harvestingof the root systems for GUS histochemical staining (see below). For certain experiments, plants which had been initially grown in pouches were transferred to the surface of solid agar slants containing beth Fahraeus medium and Nod factors, and growth was continued for a further 24 h before harvesting the roots for histochemical staining. Nod factors were applied to the upper surface of the root tip during growth on agar slants by means of a 2 mm thick/5 mm long block of 1.5 % agarose. Routine observations of stained roots were carried out using either binoculars (unmounted samples) or the light microscope (after mounting on slides). Both methods were found to be perfectly adequate for evaluating the presence of reactive cells in the root epidermis. Self-fertilization of the transgenic tetraploid M. varia leads to an $1 population exhibiting considerablevariability, both in terms of genetic background and in the expression levels of the introduced chimeric MtENOD12-GUS gene fusion. This variability is probably responsible for the fact that only a certain proportion of the plants scored positive when tested at the lowest Nod factor concentrations (e.g. 10-13 M for NodRm-IV(Ac, S); Table 1). It was possible to control for the active expression of the MtENOD12-GUS fusion in individual plants by virtue of the fact that this early nodulin gene is also transcribed at sites of secondary root emergence (Pichon, 1993). This transient, nonsymbiosis-related expression could be detected following histochemical staining for GUS activity in approximately 90-95% of the transgenic seedlings examined (results not shown). Plants which scored negative in response to Nod factors and which also scored negative at sites of secondary root emergence were therefore not included in the data presented in Table 1. The statistical significance of the results presented in Table 1 was evaluated using the 2 x 2 contingency table test, adapted for small frequencies (Pearson and Hartley, 1966).
248
E.P. Journet et al.
Bacteria/strains and purified Nod factors Bacterial inoculations were performed using Rhizobium meliloti strains mutated for the nodH gene (Rm2011 nod/-k.:Tn5#2313; Debell~ et aL, 1986), the nodFE genes (Rm2011 A(nodFE)4; Debell(~ etal., 1988) and the nodL gene (Rm2011 AJB109; Renalier et aL, 1987; Ardourel et aL, in preparation). After growth on selective media, a suspension containing between 5 x 10s and 108 bacteria was added to each growth pouch in such a way as to ensure efficient contact with the entire root system. Preparations of purified Ft. meliloti Nod factors were kindly provided by colleagues working in the research groups of JeanClaude Prom~ (LPTF, CNRS, Toulouse) and Jean D~nari~ (LBMRPM, Castanet Tolosan). NOdRm-lV(Ac,S) and NOdRmIV(Ac) were derived from overproducing strains Rm2011 (pMH682) and Rm2011 nodH::Tn5#2313 (pMH682) respectively, as described in Roche etal. (1991a). Catalytic hydrogenation of NOdRm(Ac,S) was performed as described in Truchet et aL (1991 ) and deacetylation of the O-acetyl fraction of NodRm(Ac,S) as described in Roche et al. (1991b). Purified Nod factors were serially diluted in water using sterile glassware prior to addition to the growth pouches. Chitin tetramers (N, N', N", N'"-tetraacetylchitotetraose) and the detergent CHAPS (3-((3cholamidopropyl)dimethyl-ammonio)- 1-propanesulphonate) were purchased from Sigma. Histochemical staining for ~-glucuronidase (GUS) activity GUS activity was detected and localized in whole root segments of transgenic alfalfa using the histochemical substrate X-Gluc (5bromo-4-chloro-3-indolyl glucuronide, cyclohexylammonium salt; Biosynth AG, Staad, Switzerland) and the protocol described in Pichon eta/. (1992), excepting that tissues were not routinely prefixed following 24 h treatment with Nod factors. Stained whole root segments and transverse sections (80 ~m thick; Microcut H 1200; Bio-Rad) were observed with an Olympus Vanox light microscope using bright-field optics.
Acknowledgments We are grateful to Nathalie Demont, Patrice Lerouge, Philippe Roche and Jean-Claude Promd (LPTF, CNRS, Toulouse) and Fabienne Maillet and Jean Ddnari~ (LBMRPM, Castanet Tolosan) for kindly providing us with purified Nod factors. Financial support for this work was provided by The Human Frontier Science Programme entitled 'Rhizobium Nodulation Signals: Structure, Synthesis and Role as Plant Developmental Elicitors'.
Bhuvaneawari, T.V., Bhagwat, A.A. and Bauer, W.D. (1981) Transient susceptibility of root cells in four common legumes to nodulation by rhizobia. Plant Physiol. 68, 1144-1149. van Brussel, A.A.N., Bakhuizen, R., van Spronsen, P.,
Spaink, H.P., Tak, T., Lugtenberg, B.J.J. and Kljne, J. (1992) Induction of pre-infection thread structures in the leguminous host plant by mitogenic lipo-oligosaccharides of Rhizobium. Science, 257, 70-72. Caetano-Anoll6s, G. and Gresshoff, P.M. (1991) Alfalfa controls nodulation during the onset of Rhizobium-induced cortical cell division. Plant Physiol. 95, 366-373. Deak, M., Kiss, G.B., Koncz, C. and Dudits, D. (1986) Transformation of Medicago by Agrobacterium-mediated gene transfer. Plant Cell Rep. 5, 97-100.
Debell6, F., Rosenberg, C., Vasse, J., Maillet, F., MarUnez, E., D6nari6, J. and Truchet, G. (1986) Assignment of symbiotic developmental phenotypes to common and specific nodulation (nod) genetic loci of Rhizobium meliloti. J. Bacteriol. 168, 1075-1086. DebelM, F., Maillet, F., Vasse, J., Rosenberg, C., de Billy, F., Truchet, G., D6nari6, J. and Ausubel, F.M. (1988) Interference between Rhizobium meliloti and Rhizobium trifolii nodulation genes: Genetic basis of Ft. meliloti dominance. J. BacterioL 170, 5718-5727.
Demont, N., Debelld, F., Aurelle, H., D6nad6 J. and Prom6, J.-C. (1993) Role of Rhizobium meliloti nodF and nodE genes in the biosynthesis of lipo-oligosaccharidic nodulation factors. J. Biol. Chem. 268, 20 134-20 142. D6nari6, J. and Roche, P. (1992) Rhizobium nodulation signals. In Molecular Signals in Plant-Microbe Interactions (Verma, D.P.S., ed.). Boca Raton, FL: CRC Press, pp. 295-324. Ehrhardt, D.W., Atkinson, E.M. and Long, S.R. (1992) Depolarization of alfalfa root hair membrane potential by Rhizobium meliloti Nod factors. Science, 268, 998-1000.
Faucher, C., Maillet, F., Vasse, J., Rosenberg, C., van Bruesel, A.A.N., Truchet, G. and D~nari6, J. (1988) Rhizobium meliloti host range nodH gene determines production of an alfalfa-specific extracellular signal. J. Bacteriol. 170, 5489-5499. Fisher, R.F. and Long, S.R. (1992) Rhizobium-plant signal exchange. Nature, 357, 655-660.
Horvath, B., Heidstra, R., Lados, M., Moerman, M., Spaink, H.P., Prom6, J.-C., van Kammen, A. and Bisseling, T. (1993) Lipo-oligosaccharides of Rhizobium induce infectionrelated early nodulin gene expression in pea root hairs. Plant J. 4, 727-733. Jefferson, R.A., Kavanagh, T.A. and Bevan, M.W. (1987) GUS fusions: 13-glucuronidase as a sensitive and versatile gene fusion marker in higher plants. EMBO J. 6, 3901-3907.
Lerouge, P., Roche, P., Faucher, C., Maillet, F., Truchet, References Allison, L.A., Kiss, G.B., Bauer, P., Poiret, M., Pierre, M., Savour0, A., Kondorosi, E. and Kondorosl, A. (1993) Identification of two alfalfa early nodulin genes with homology to members of the pea ENOD12 gene family. Plant Mol. Biol. 21, 375-380. Baev, N. and Kondorosi, A. (1992) Nucleotide sequence of the Rhizobium meliloti nodL gene located in locus n5 of the nod regulon. Plant Mol. Biol. 18, 843-846. Barker, D.G., Bianchi, S., Blondon, F., et aL (1990) Medicago truncatula, a model plant for studying the molecular genetics of the Rhizobium-legume symbiosis. Plant Mol. Biol. Rep. 8, 40-49.
G., Prom6, J.-C. and Ddnad6, J. (1990) Symbiotic hostspecificity of Rhizobium meliloti is determined by a sulphated and acylated glucosamine oligosaccharide signal. Nature, 344, 781-784. Pearson, E.S. and Hadley, H.O. (1966) Biometrika Tables for Statisticians, Volume 1. Cambridge: Cambridge University Press. Pichon, M. (1993) Le gdne MtENOD12: Marqueur moldculaire des stades pr&coces de I'interaction symbiotique Rhizobium meliloti/Medicago. Thesis, Universit~ d'Orsay, Paris, France.
Pichon, M., Joumet, E.P., Dedieu, A., de Billy, F., Truchet, G. and Barker D.G. (1992) Rhizobium meliloti elicits transient expression of the early nodulin gene ENOD12 in the differentiating root epidermis of transgenic alfalfa. Plant Cell, 4, 1199-1211.
Rhizobium Nod factors elicit ENOD 12 transcription
Renaller, M.-H., Batut, J., Ghai, J., et al. (1987) A new symbiotic cluster on the pSym megaplasmid of Rhizobium meliloti 2011 carries a functional fix gene repeat and a nod locus. J. Bacteriol. 169, 2231-2238.
Roche, P., Debell6, F., Maillet, F., Lerouge, P., Faucher, C., Truchet, G., D6nari6, J. and Prom~, J-C. (1991a) Molecular basis of symbiotic host specificity in Rhizobium meliloti: nodH and nodPQ genes encode the sulphation of lipo-oligosaccharide signals. Cell, 67, 1131-1143. Roche, P., Lerouge, P., Ponthus, C. and Prom6, J-C. (1991b) Structural determination of bacterial nodulation factors involved in the Rhizobium meliloti-alfalfa symbiosis. J. Biol. Chem. 266, 10 933-10 940. Scheres, B., van de Wiel, C., Zaleneky, A., et al. (1990) The ENOD12 gene product is involved in the infection process during the pea-Rhizobium interaction. Cell, 60, 281-294.
Schultze, M., Quiclet-Sire, B., Kondorosi, E., Virelizier, H., Glushka, J.N., Endre, G., G6ro, S.D. and Kondoroei, A. (1992) Rhizobium meli/oti produces a family of sulphated lipo-
249
oligosaccharides exhibiting different degrees of plant host specificity. Proc. Natl Acad. ScL USA, 89, 192-196. Schwedock, J. and Long, S.R. (1989) Nucleotide sequence and protein products of two new nodulation genes of Rhizobium meliloti, nodP and nodQ. Mol. Plant-Microbe Interact. 2, 181-194. SpaJnk, H.P. (1992). Rhizobial lipo-oligosaccharides: Answers and questions. Plant Mol. BioL 20, 977-986.
Spaink, H.P., Sheeley, D.M., van Brussel, A.A.N., et al. (1991) A novel highly unsaturated fatty acid moiety of lipo-oligosaccharide signals determines host specificity of Rhizobium. Nature, 354, 125-130.
Truchet, G., Roche, P., Lerouge, P., Vasse, J., Camut, S., de Billy, F., Prom~, J.-C. and D6nari6, J. (1991) Sulphated lipooligosaccharide signals of Rhizobium me/i/oti elicit root nodule organogenesis in alfalfa. Nature, 351,670-673. Vincent, J.M. (1970). A Manual for the Practical Study of RootNodule Bacteria, IBP Handbook No. 15. Oxford: Blackwell Scientific Press.
