Analysis of nodulation kinetics in Frankia–Discaria trinervis symbiosis reveals different factors involved in the nodulation process

Physiologia Plantarum 133: 776–785. 2008 Copyright ª Physiologia Plantarum 2008, ISSN 0031-9317

Analysis of nodulation kinetics in Frankia–Discaria trinervissymbiosis reveals different factors involved in thenodulation processLuciano Andres Gabbarini and Luis Gabriel Wall*

Programa Interacciones Biologicas, Departamento de Ciencia y Tecnologıa, Universidad Nacional de Quilmes, Roque Saenz Pena 352, Bernal B1876BXD,

Buenos Aires, Argentina

Correspondence

*Corresponding author,

e-mail: [email protected]

Received 7 January 2008

doi: 10.1111/j.1399-3054.2008.01096.x

The induction of root nodule development in actinorhizal symbiosis would

depend on the concentration of factors produced by the bacteria and the plant.

A detailed analysis of nodulation description parameters revealed different

factors related to the nodulation process. The initial time for nodulation (t0), theinitial nodulation rate (v0) and the total time of nodule development (tNOD)

were defined and consequently quantified in different experimental con-

ditions: co-inoculation of Discaria trinervis with increasing concentrations of

different non-infective bacteria together with the full compatible infective

Frankia strain (the indicator strain) used at a limiting concentration or by

changing plant factor(s) concentration. All the above nodulation parameters

were modified by changing doses of full compatibility infective strain Frankia

BCU110501; v0 appears to be an expression of symbiotic recognition betweenpartners as only fully symbiotic indicator Frankia BCU110501 was able to

change it; t0 seems not to reflect symbiotic recognition because it can also be

modified by non-infective Frankia but suggest the existence of a basic level of

plant microbe recognition. The initial time for nodulation t0, reflecting the time

required for the early interactions toward nodulation, is an inverse measure of

the ability to establish early interactions toward nodulation. The increase in

plant factors concentration also reduces t0 values, suggesting that a plant factor

is involved and favors very early interactions. Increases in plant factorsconcentration also modify the final number of nodules per plant and the nodule

cluster profile along the taproot as an expression of the autoregulation

phenomenon. Meanwhile, Frankia inoculums’ concentration, either infective

or not, modified tNOD in an opposite way plant factors did. In conclusion, the

analysis of nodulation kinetics appears to be an appropriate tool to investigate

factors involved in the symbiotic interaction leading to the formation of

nitrogen-fixing nodules.

Introduction

The nitrogen-fixing nodules induced by Frankia inDiscaria trinervis roots are an example of an actinorhizal

symbiosis (Wall 2000) with intercellular root invasion

pathway (Valverde and Wall 1999b) that implies no root

hair deformation at all during recognition and infection,

clearly different to the intracellular infection pathwaythrough root hair (Wall and Berry 2007). Diffusible signal

factors (DSFs) involved in early interactions in actino-

rhizal plants have been studied in symbioses requiring

Abbreviations – cfu, colony forming units; DSF, diffusible signal factor; pcv, packed cell volume; RT, root tip.

776 Physiol. Plant. 133, 2008

root hair infection, such as Alnus sp. by the use of a root

hair deformation bioassay (Ceremonie et al. 1999, Van

Ghelue et al. 1997). From those studies, Frankia root hair-

deforming factor seems not to be the factors that express

symbiotic recognition as the lipochitinoligosaccharides

produced by rhizobia in response to root legume signalsin the Rhizobium–legume symbiosis. Also, plant factors

related to flavonoids have been detected to condition,

either stimulating or inhibiting, nodulation in Alnus rubra

(Benoit and Berry 1997). Nevertheless, it is not clear from

those experiments if those effects are consequences of

a plant–microbe interaction or a plant–plant interaction

mediated by flavonoids or related substances.

Infection and nodule development in the root ofactinorhizal plants is a dose–response phenomenon that

attains saturation at high doses of bacterial inoculum

either in the case of root hair-infected plant such as Alnus

incana (Wall and Huss-Danell 1997) or in the intercel-

lular invaded root such asD. trinervis (Valverde and Wall

1999b). This suggests that the induction of the plant

response depends on the concentration of putative factors

produced by the bacteria and therefore on the affinity ofthe corresponding receptor at the plant side as it occurs

in any biochemical interaction between molecules (i.e.

enzyme–substrate interaction).

Similar to what occurs in legumes, nodulation in

actinorhizal plants is under regulatory control involving

mechanisms such as autoregulation (Valverde and Wall

1999a, Wall and Huss-Danell 1997), feedback inhibi-

tion by N (Valverde et al. 2000, Wall et al. 2003) andstimulation by P levels in plant tissues (Valverde et al.

2002). These inhibitor factors can be organized in a model

of regulation of actinorhizal symbiosis regardless of the

root infection pathway (Wall 2000). Taking in account

these regulatory phenomena, it is possible to understand

the final level of nodulation and the nodule distribution in

the taproot in a particular experimental condition (Valve-

rde and Wall 1999a). Nothing is known about the natureof the molecular signal(s) involved in these phenomena.

The number of nodules developed per plant can be

studied using a plant growth method that allows to follow

nodulation in the undisturbed root along the time lapse of

the experiment. In this way, different nodulation param-

eters can be defined and consequently quantified in

different experimental conditions. Examples of these

parameters have been already reported as indicators ofspecificity at the genus level within a cross-inoculation

group of actinorhizal plants (Nesme et al. 1985) or at the

species level, suggesting intra-genus degrees of symbiotic

recognition (Chaia et al. 2006).

Competition assays with compatible or non-compati-

ble bacteria have been used to study symbiotic specificity

in legumes at the root colonization step (Caetano Anolles

and Favelukes 1986). This idea was adapted to the study

of actinorhizal symbiosis analyzing the modification of

nodulation parameters after co-inoculation with increas-

ing concentrations of different non-nodulating bacteria

together with the nodulating Frankia strain used at

a limiting – non-saturating – concentration or by changingplant factor concentration.

Materials and methods

Plant growth

Seeds of D. trinervis (Hooker et Arnot) Reiche were kindly

provided by E Chaia (Universidad Nacional del Comahue,Argentina). Mature fruits were collected at Pampa de

Huenuleo (41� 10# S, 71� 12# W, Bariloche, Rio Negro,

Argentina) and kept at 220�C. Surface sterilization was

performed by scarification (3-min immersion in 96%

H2SO4) and then rinsed (10 times) with sterile Evans

solution (27.88 mg l21 K2SO4; 49.28 mg l21 MgSO4 7

H2O; 2.3 mg l21 KH2PO4; 14.5 mg l21 K2HPO4;

10.33 mg l21 CaSO4; 7.35 mg l21 CaCl2 2H2O; 0.66 mgl21 FeNaEDTA; 0.143 mg l21 H3BO3; 0.0773 mg l21

MnSO4 H2O; 0.0022 mg l21 ZnSO4 7H2O; 0.0079 mg

l21 CuSO4 5H2O; 0.005 mg l21 Na2MoO4 2H2O and

0.0141 mg l21 CoSO4 7H2O) diluted to 1/10 strength (E 1/

10). Germination was performed on wet perlite in Petri

dishes. Twelve days after germination, the seedlings were

transferred to sterilized seed growth pouches (Mega

International, Minneapolis, MN). Randomly selected fourseedlingswere placed intoeach pouch containing 10 ml of

nitrogen-free nutrient solution (1/10 full-strength Evans

solution; Valverde and Wall 1999a). Germination and

further plant growth were carried out in a greenhouse at

the University of Quilmes (34� 7# S, 58� 3# W) with sup-

plemental artificial light (400 W, K048; Osram, Osasco,

Brazil) so that the photoperiod was 16-h light:8-h dark,

temperature range between 18 and 25�C and a relativityhumidity of 40–95%.

Bacteria culture and preparation of inocula

Frankia BCU110501 was isolated from D. trinervis

nodules and is infective (Nod1) and effective (Fix1) in

D. trinervis plants (Chaia 1998). The Frankia ArI3 (K

Huss-Danell, Umea, Sweden) was isolated originally inpure culture from A. rubra (Berry and Torrey 1979).

Frankia CpI1 (K Huss-Danell, Umea, Sweden) was

isolated from Comptonia peregrina (L.) Coult (Torrey

and Callaham 1982). Frankia Cj82 (J Dawson, Chicago,

IL, USA) was isolated from Casuarina japonicum. Frankia

was grown for 6 weeks at 28�C in static culture in

minimal BAP medium (Murry et al. 1984). Culture media

Physiol. Plant. 133, 2008 777

for strains BCU110501, CpI1 and Cj82 were supple-

mented with 0.055 M glucose, while in the case of strain

ArI3, it was supplemented with 0.005 M Na propionate

as the carbon source. Streptomyces coelicolor A3(2) M

145 (Gernot Vobis, Bariloche, Argentina) was grown in

40 ml of BAP medium supplemented with 0.055 Mglucose for 96 h at 28�C. Bacillus thuringiensis isolate

(L Del Federico, UNQ, Argentina),Bacillus subtilis isolate

(L Del Federico, UNQ, Argentina), Pseudomonas putida

GR12-2 (Glick et al. 1997) and Azospirillum brasilense

Cd pRKLACC (Holguin and Glick 2001) were grown in

10 ml of nutrient broth for 72 h at 28�C.

For inoculum preparation, Frankia and Streptomyces

cells were washed twice with E 1/10, resuspended in 5–8 ml of E 1/10 and gently homogenized by passage five

times through needles (21G). The amount of biomass in

the suspension was estimated by the packed cell volume

(pcv) after centrifugation for 5 min at 1100 g (Nittayajarn

and Baker 1989). The other bacteria were grown over-

night at 28�C and 120 rpm and washed twice with and

resuspended in 10 ml of E 1/10. The amount of cells was

estimated by standard plate count on nutrient broth agarplates.

Plant inoculation

Different experimental designs were assayed on different

sets ofD. trinervisplants. At the moment of inoculation, 20

plants distributed in 5 pouches were separated randomly,

and all pouches belonging to the same treatment wereplaced together in a separate bigger plastic bag. Appro-

priate dilutions of the different bacterial suspensions were

prepared in E 1/10 solution, and 200 ml of the suspensions

were used as inoculum and applied onto the taproot of the

seedlings from the root tip to the root base. In order to study

the dose–response effect operating in D. trinervis symbi-

osis, different Frankia BCU110501 doses were used: 0.01,

0.1, 1, 5 and 20 ml pcv root21.In different experiments, the effect of the co-inocula-

tion of Frankia BCU110501 together with non-compat-

ible (non-infective) bacteria was studied. In those cases,

both bacteria were mixed in the same vial and 200 ml of

the combined inoculum was applied onto the taproot of

the seedlings; the following experiments were performed:

(1) Frankia BCU110501 (0.01 ml pcv root21) together

with different Frankia ArI3 doses (0.1, 1, 5 and 20 mlpcv root21 and 0 pcv root21 for control); (2) Frankia

BCU110501 (0.01 ml pcv root21) together with Frankia

CpI1 doses (1 and 10 ml pcv root21 and 0 ml pcv root21

for control); (3) Frankia BCU110501 (0.01 ml pcv root21)

together with Frankia Cj82 doses (10 ml pcv root21 and

0 ml pcv root21 for control); (4) Frankia BCU110501

(0.01 ml pcv root21) together with different S. coelicolor

A3(2) M 145 doses (0.1, 1 and 4 ml pcv root21 and 0 ml pcv

root21 for control); (5) Frankia BCU110501 (0.01 ml pcv

root21) together with differentA. brasilenseCd pRKLACC

doses (1 � 109 and 1 � 106 cfu root21 and 0 cfu root21

for control); (6) Frankia BCU110501 (0.01 ml pcv root21)

together with different P. putida GR12-2 doses (1 � 1010

and 1 � 107 cfu root21 and 0 cfu root21 for control); (7)

Frankia BCU110501 (0.01 ml pcv root21) together with

differentB. thuringiensis doses (2 � 108 and 2 � 105 cfu

root21 and 0 cfu root21 for control); and (8) Frankia

BCU110501 (0.01 ml pcv root21) together with different

B. subtilis doses (5 � 106, 5 � 103 and 0 cfu root21).

The effect of plant factors operating in D. trinervis

symbiosis was assayed by placing different numbers ofplants per pouch: 4, 8 or 12 distributed as 1, 2 or 3 plants

in each of the 4 holes in the growth pouch. These were

watered always with 12 ml of E 1/10 and refilled as

necessary in the same way for all treatments. In all

treatments, plants were inoculated unloading 200 ml of

Frankia BCU110501 inoculum (0.05 ml pcv ml21) on the

roots growing from each of the four holes of the growth

pouch. Thus, the total amount of Frankia per pouch wasthe same in all treatments independent of the total

amount of plants per pouch.

Pretreatment of Frankia inoculum with bacterialDSF(s)

Frankia inoculum was pretreated before being inoculated

to the plants as follows: 6-week-old Frankia BCU110501culture was washed twice and resuspended with E 1/10

solution to reach 50 ml pcv ml21. About 5 ml of this

bacterial suspension was kept in a sterilized dialysis

cellulose tubing (12-kDa cutoff) of 15 mm diameter

(Sigma, St. Louis, MO) and immersed for 96 h at 28�C, in

darkness, in 25 ml of an inoculum suspension made of

washed cells of Frankia BCU110501 from the same

original culture, diluted at a concentration of 0.05 mlpcv ml21 in E 1/10 solution. After the pretreatment, the

dialysis bag was removed from the inoculum suspension

and 200 ml of the pretreated inoculum was inoculated on

the taproot ofD. trinervis as described above. As negative

control, the inoculum was pretreated with a similar

dialysis tubing containing just 5 ml of E 1/10 solution.

Each inoculum pretreatment was assayed on 20 plants

distributed in 5 pouches (4 plants per pouch).In another experiment, the diluted Frankia BCU

110501 inoculum (0.05 ml pcv ml21), prepared as stated

before, was pretreated with 5 ml of different concen-

trated bacterial suspensions (10 ml pcv ml21) contained

in the dialysis tubing. To study the effect of DSF from

different origins, we used the following bacteria in the con-

centrated suspension: 6-week-old Frankia BCU110501,


6-week-old dead Frankia BCU110501 by previous

exposure to 3 kGy of g-ray, 6-week-old Frankia ArI3

and 1-week-old S. coelicolor. In this experiment, the

pretreatment of the inoculum was for 24 h. The inocu-

lation of plants after the pretreatment of the inoculum and

the negative control was performed as stated above.

Nodulation scoring

The position of each root tip at the moment of inoculation

was marked (RT) on the pouch plastic surface witha waterproof marker (Valverde and Wall 1999a). D. tri-

nervis roots growing in the pouches were recorded each

2–3 days looking for nodules development with naked

eye. The positions of nodules on the taproot were recorded

with a waterproof marker pen in the plastic bag and

counted. The distance (in mm) from each nodule to the

reference mark RT (zero position) was measured and the

frequency of nodules appearance was grouped each 5 mmalong the root and plotted as nodulation profile. Positions

above RT mark were considered as positive values (1),

while positions below RT mark were considered as neg-

ative values (2) (Valverde and Wall 1999a).

Statistical analysis

Significant differences between nodulation parameter

values of the experimental treatment against the appro-

priate control were obtained through Student’s unpaired

t-test using SIGMAPLOT� (version 8.0).

Results

Nodulation analysis

When the appearance of nodules was analyzed in

relation to time in each individual plant, a regular pattern

as depicted in Fig. 1A appeared. Different nodulation

parameters were defined after this kind of plots: (1) t0, thelag time for first nodule appearance; (2) v0, the initial

nodulation rate; (3) vR, a second reduced nodulation rate

that does not always appear; (4) tR, the time at which

nodulation rate is reduced; (5) tF, the final time after which

there is no more nodule development; (6) t1 ¼ tR 2 t0,

the time lapse that nodules develop at the initial

nodulation rate; (7) t2 ¼ tF 2 tR, the time lapse that

nodules develop at the reduced nodulation rate; (8)tNOD ¼ tF 2 t0, the total time of nodule development;

and (9) number of nodules per plant, the final number of

developed nodules per plant. The parameters could be

determined for each individual plant (Fig. 1B) and then

means and SE could be compared between different

treatments. These data were more precise than the

estimation that arose after the analysis of the average

plot of number of nodules per plant vs time in the whole

graph of plant (Fig. 1C). In this average plot (Fig. 1C),each point represents the average number of nodules per

plant at a time considering total number of plants in the

treatment to calculate the average. For instance, the t0 in

this average plot (Fig. 1C) is always misestimated because

it is determined by the first plant in the group that presents

a visible nodule. Meanwhile, the t0 of the other plants got

masked in the following recorded data of the curve.

Effects of different doses of infective FrankiaBCU110501

Nodulation in D. trinervis roots was dependent on the

level of Frankia inoculum (Fig. 2). This dose–response

Fig. 1. Nodulation kinetics analysis. (A) Schematic plot of nodulation

kinetics showing the parameters analysed per plant: t0 is time needed for

the first nodule appearance, v0 is the initial rate of nodulation, vR is the

regulated rate of nodulation, tF is the time for the last nodule appearance

and tNOD is the total time of nodulation (tNOD ¼ tF 2 t0). (B) Group of

plots where each one corresponds to a single plant in a set of plants

belonging to the same treatment assay. (C) Average plot where each point

represents the average data for all the plants at each corresponding date

of recording.


behavior could be easily visualized as an increment in the

final number of nodules per plant (Fig. 2A) and also as

different nodulation rates shown in the initial slope of

each individual plot that increased with the inoculum

dose (Fig. 2B). At variance, nodule distribution along the

taproot, a phenotype that expresses autoregulation of

nodulation in this symbiosis (Valverde and Wall 1999a),

did not show a clear dependence of its pattern with thedose of the inoculums as it is shown with the % of nodules

above or below RT mark (Fig. 2C).

Nodulation parameters were determined by the

analysis of each individual plant’s kinetic as described

above, and mean values of those parameters were

analyzed. The increase in the inoculum dose of Frankia

BCU110501 reduced t0, the time lag for first nodule

apparition (Fig. 3A), increased v0 (Fig. 3B), increasedthe number of plants with a second reduced rate of

nodulation (Fig. 3C), increased tNOD, the total time of

nodule development (Fig. 3D) and also increased the

final number of nodules per plant at the time of harvest

(Fig. 2A). This experiment was repeated twice by the

present authors and it was performed before at the lab and

Fig. 2. Nodulation of Discaria trinervis inoculated with different Frankia BCU110501 doses. (A) Nodulation kinetics at different Frankia doses: 0.01(s),

0.1 (d), 1 (n), 5 (:) and 20 (;) ml pcv (200 ml)21 of bacterial suspension per plant. Values are the means � SE for n ¼ 16–20 plants. (B) The small graph

inserted shows the same graph in decimal logarithmic scale and the arrow points to the limiting inoculum dose referred to in the text. (C) Nodule

distribution profiles along D. trinervis taproot. Different Frankia BCU110501 doses as microliters of Frankia pcv per root are indicated at the right of each

plot. The gray line indicates the RT position at time of inoculation. Values in percent of nodules above and below RT mark are shown within the graph at

left and right, respectively. Position data of all nodules at 79 days after inoculation from 16–20 plants were used to plot these distribution profiles. The

plots represent the probability of a nodule to develop at a certain distance to the position of the root apex at the moment of inoculation (RT mark). Root

growth direction is from left to right.

Fig. 3. Nodulation parameters in Discaria trinervis inoculated with

different Frankia BCU110501 doses. (A) t0, time lag for nodulation; (B)

v0, initial nodulation rate; (C) vR, percentage of plants with regulated rate

of nodulation and (D) tNOD, total time of nodulation. Values are

means � SE for n ¼ 16–20 plants.


published regarding the phenomenon of autoregulation

of nodulation (Valverde and Wall 1999a). The analysis of

the data from the different experiments gave similar

results and we show here the more complete set of data.

Effects of Frankia DSF(s) on nodulation

Frankia BCU110501 inoculum suspension was pre-

treated for 96 h, before being added to the plants, with

DSF dialyzed from a concentrated Frankia BCU110501

suspension (see Materials and methods). This experimentwas repeated independently at three different moments.

The pretreatments of the inoculum with DSF from Frankia

BCU110501 significantly reduced t0 by 2.77, 5.56 and

1.58 days in each experiment with P values of 0.064,

0.003 and 0.452, respectively. The nodulation profile

was clearly modified by DSF pretreatment of the ino-

culum (Fig. 4), although the inoculum dose of Frankia

BCU110501 was not different between treatments. The

nodulation profile was shifted to the upper region of the

root (Fig. 4).

Effects of different doses of D. trinervis rootexudates

To study the potential effects of other factors in the

nodulation assay, we decided to work at a limiting

inoculum dose of Frankia BCU110501 (value indicated

by the arrow in Fig. 2B) and to vary other factors. Thus, we

explored the effects of plant factors by increasing the

number of plants per pouch. Because the amount of

mineral solution per pouch was not changed, the root

exudate concentration should vary between treatmentsaccording to the number of plants present per pouch. The

increase in root exudates concentration reduced t0, the

lag time for first nodule appearance (Fig. 5C) and ap-

parently reduced tNOD, the total time of nodule develop-

ment (Fig. 5D). Nodulation profile was relatively shifted

basipetally as if feedback inhibition of nodulation was

Fig. 4. Effect of DSF from Frankia BCU110501. Nodule distribution

profile along Discaria trinervis taproots inoculated with Frankia

BCU110501 [0.05 ml pcv ml21 E 1/10) pre-incubated with DSF from

Frankia BCU110501 in high concentration (50 ml pcv ml21 E 1/10) (d)

and pre-incubated with DSF from E 1/10 solution (s). Position data of all

nodules from 20 plants were used to plot these distribution profiles. The

plots were normalized because of differences in total nodule number per

plant between treatments and represent the probability of a nodule to

develop at a certain distance to the position of the root apex at the

moment of inoculation (RT mark). Dt0 denotes the reduction in time lag

for nodulation with the inoculums pre-incubated with DSF from Frankia

(d) compared with the time lag for nodulation with the control inoculum

pre-incubated with mineral solution (s). A, B and C are results from

independent experiments.

Fig. 5. Nodulation of Discaria trinervis with different amounts of plants

per pouch. (A) Kinetics of nodulation for 4 plants per pouch (s), 8 plants

per pouch (n) and 12 plants per pouch (:). All treatments were

inoculated with 800 ml of bacterial suspension per pouch distributed in

four doses of 200 ml containing 0.01 ml pcv of Frankia BCU110501 on

each root group of one, two or three roots per hole of the pouch. (B)

Nodule distribution profiles along D. trinervis taproot. Different numbers

of plants per pouch are indicated at the right of each plot. Nodule position

with respect to RT mark was recorded at 70 days after inoculation. A

nodulation profile along the taproot was derived for each set of plants as

described in Materials and methods. Root growth direction is from left to

right. The gray line shows the RT position at time of inoculation, and

values in percent of nodule above and below RT mark are shown within

the graph at left and right, respectively. (C) Lag time (t0) change with

doses of plants. (D) Total time of nodulation (tNOD) change with doses of

plants. Values are the means � SE for n ¼ 16–20 plants.


established earlier, producing nodule cluster with shorter

extension along taproot (Fig. 5B, Table 1). This last effect

produced an earlier stop of nodule appearance and

consequently final number or nodules per plant were

reduced (Fig. 5A). No differences were detected in other

nodulation parameters as v0, the initial nodulation ratewas 0.60, 0.71 and 0.67 (number of nodules plant21

day21) for 4, 8 and 12 plants per pouch, respectively.

Effects of different doses of co-inoculated non-infective bacteria

When high doses of bacteria belonging to different groups

of rhizobacteria, and non-infective to D. trinervis, wereco-inoculated together with a limiting amount of the

infective Frankia BCU110501 (Fig. 2B, arrow), a stimula-

tion of the nodulation was found only when the co-

inoculated bacteria were Frankia belonging to a different

cross-inoculation group by comparison to control treat-

ment inoculated with only Frankia BCU110501 (Fig. 6,

Table 2, lines 1–4). Frankia strains ArI3, CpI1 and Cj82

were not able to induce nodules in D. trinervis wheninoculated alone at high doses. On the contrary, when

either Gram2 or Gram1 rhizobacteria were co-inocu-

lated in very high doses (up to 107 cfu ml21), no effect at

all was observed on the nodulation of D. trinervis by

limiting doses of Frankia BCU110501. No change,

neither inhibition nor stimulation, of any nodulation

parameters was found (Table 2, row 5–13).

A more detailed analysis was performed after co-inoculation of Frankia ArI3 doses with diluted Frankia

BCU110501 (Fig. 6A, C). Frankia ArI3 promoted nodula-

tion by Frankia BCU110501. The main effects were

reducing t0, the lag time for the first nodule appearance

(Fig. 6C) and increasing tNOD, the time lapse of nodule

development (Fig. 6D). There was no effect, either stimu-

lation or inhibition, of the initial rate of nodulation [v0: 1.11,

1.31, 1.29, 1.11 and 1.17 (number of nodules plant21

day21) for 0, 0.1, 1, 5 and 20 ml of ArI3, respectively].

Nodulation profile did not shift when increasing the dose of

ArI3 (Fig. 6B). Similar effect about increasing total time of

nodule development had been observed in Table 2 for the

co-inoculation with other non-infective Frankia as strainsCpI1 and Cj82 (Table 2, line 1–4).

Effects of DSF(s) of non-infective bacteria onnodulation

The stimulation of nodulation of D. trinervis by co-

inoculation of BCU110501 with other Frankia belonging

to different cross-inoculation groups (Fig. 6, Table 2) was

Table 1. Analysis of nodulation cluster in the taproot. Xi is the position of

the uppermost nodule in the nodule cluster along the taproot. XL is the

position of the last nodule in the root tip direction. Xi 2 XL is the length of

the nodule cluster along the main root, that is, the distance between first

and last nodule in the distribution. aP � 0.05, unpaired t-test (SIGMAPLOT

8.0) and statistically significant differences between 4 plants compared

with 8 and 12 plants per pouch. bP � 0.001, unpaired t-test (SIGMAPLOT

8.0) and statistically significant differences between 12 plants compared

with 4 and 8 plants per pouch. Positions are in mm, values are means � SE

for n ¼ 16–20 plants.

Nodule position

Number of plants per pouch

4 8 12

Xi 8.2 � 2.8a 14.9 � 1.6 14.2 � 1.6

XL 215.4 � 4.0a 24.9 � 2.2 20.7 � 1.8

Xi 2 XL 23.6 � 3.4 19.8 � 2.0 12.2 � 1.4b

Fig. 6. Nodulation of Discaria trinervis co-inoculated with Frankia

BCU110501 and different Frankia ArI3 doses. (A) Nodulation kinetics

for 0 (s), 0.1 (d), 1 (n), 5 (:) and 10 (;) ml pcv of Frankia ArI3 per root;

Frankia BCU110501 was inoculated at 0.01 ml pcv root21. Both bacteria

were mixed in appropriate final concentration before the inoculation and

each root was inoculated with 200 ml of this suspension. (B) Nodule

distribution profiles along D. trinervis taproots inoculated with Frankia

BCU110501 (0.01 ml pcv) and different Frankia ArI3 doses indicated as

microliters of Frankia pcv inoculated per root at the right of each plot.

Nodule position with respect to RT mark was recorded at 35 days after

inoculation. A nodulation profile along the taproot was derived for each

set of plants as described in Materials and methods. Root growth direction

is from left to right. The gray line shows the RT position at time of

inoculation, and values in percent of nodule above and below RT mark are

shown within the graph at left and right, respectively. (C) t0, time lag for

nodulation variation with doses of Frankia ArI3. (D) tNOD, total time of

nodulation variation with doses of FrankiaArI3. Values are means � SE for

n ¼ 17–20 plants.


also observed when Frankia BCU110501 inoculum

(0.05 ml pcv ml21) was pretreated for 24 h, before being

added to the plants, with DSFs from high-concentration

suspensions of Frankia BCU110501, Frankia ArI3 and

also S. coelicolor. DSF from all these actinomycetes wereactive to stimulate nodulation by pretreated Frankia

inoculum (Fig. 7). Mineral solution E 1/10 was used as

a negative control (Fig. 7). Dead Frankia BCU110501 did

not show any effect, suggesting that DSF was actively

produced by live cells. Lag time t0 was significantly

diminished by pretreatment with Frankia ArI3 DSF

(P < 0.188), and with S. coelicolor DSF (P < 0.00037),

compared with the negative control (Table 3). DSF from

dead Frankia BCU110501 cells did not modify t0compared with control (Table 3). The initial nodulation

rate v0 of the inoculum pretreated with DSF from Frankia

BCU110501 was higher than control (P < 0.05), but this

effect was not observed for pretreatments with DSF ofFrankiaArI3 or S. coelicoloror dead FrankiaBCU110501

(Table 3).

Discussion

In the course of actinorhizal symbiotic root nodulation,

interactions between partners may be expected to occur

through signal exchanges and physical interactions includ-ing root surface and intercellular spaces colonization,

which would probably depend on a particular bacterial

phase state regulated by the local concentration of bacteria,

but these are presently unknown. In the legume–rhizobia

symbiosis, plant signals (flavonoids) activate the production

of the rhizobial Nod factors (lipochitooligosaccharides),

Table 2. Nodulation parameters of Discaria trinervis co-inoculated with Frankia BCU110501 and different bacteria. Frankia BCU110501 was inoculated

with 0.01 ml pcv root21 of bacteria in all cases. Co-inoculated bacteria were mixed at the appropriate final concentration as indicated in the column,

before the inoculation, and each root was inoculated with 200 ml of this mixed suspension. Rows 1 and 5 are the test control for the data in the

corresponding below rows 2–4 and 6–13, respectively, as indicated by the same superscript designator ‘‘a’’ or ‘‘d’’. Statistically significant differences

at bP � 0.01 and cP � 0.05, t-test unpaired (SIGMAPLOT 8.0) compared with corresponding test control.

Number Co-inoculation Number of nodules plant21 t0 v0 tNOD

1 None (control)a 5.6 � 0.8 20.0 � 1.5 0.29 � 0.03 24.0 � 2.6

2 Frankia CpI1a 10 ml 10.9 � 1.0b 23.5 � 2.1 0.44 � 0.04b 32.2 � 2.6c

3 Frankia CpI1a 0.1 ml 10.6 � 0.9b 22.5 � 1.5 0.40 � 0.04c 32.7 � 2.7c

4 Frankia Cj82a 10 ml 10.3 � 1.4b 21.1 � 2.1 0.40 � 0.07 36.6 � 2.8b

5 None (control)d 7.8 � 1.0 25.2 � 2.1 0.23 � 0.04 33.0 � 2.8

6 Bacillus thuringiensisd 2 � 108 cfu root21 6.9 � 1.0 28.0 � 2.5 0.30 � 0.04 32.6 � 3.6

7 B. thuringiensisd 2 � 105 cfu root21 7.1 � 1.1 28.5 � 1.4 0.26 � 0.04 33.6 � 3.3

8 Azospirillum brasilensed 1 � 109 cfu root21 10.2 � 1.4 29.1 � 2.6 0.34 � 0.05 35.8 � 3.4

9 A. brasilensed 1 � 106 cfu root21 8.3 � 0.7 22.0 � 2.0 0.22 � 0.03 35.8 � 4.6

10 Bacillus subtilisd 5 � 106 cfu root21 8.1 � 1.0 35.8 � 3.8c 0.29 � 0.05 34.0 � 4.0

11 B. subtilisd 5 � 103 cfu root21 8.8 � 1.7 31.4 � 4.1 0.31 � 0.05 34.6 � 4.6

12 Pseudomonas putidad 1 � 1010 cfu root21 8.4 � 1.2 28.3 � 2.9 0.22 � 0.03 36.3 � 3.8

13 P. putidad 1 � 107 cfu root21 5.1 � 0.8c 28.4 � 3.6 0.20 � 0.04 38.1 � 3.8

Fig. 7. Nodulation in Discaria trinervis plant inoculated with Frankia

BCU110501 pre-incubated with DSF from different bacteria. Nodulation

kinetics for the same original inoculum pretreated with DSF from different

bacterial suspensions or solutions contained in a dialysis tubing (see

Materials and methods): negative control, E 1/10 (s), 10 ml pcv of Frankia

BCU110501 ml21 E 1/10 (d), 10 ml pcv of dead Frankia BCU110501

ml21 E 1/10 (1), 10 ml pcv of Frankia ArI3 ml21 E 1/10 (n) and 10 ml pcv

of Streptomyces coelicolor ml21 E 1/10 (:); pretreated Frankia

BCU110501 was inoculated at 0.01 ml pcv root21.

Table 3. Nodulation parameters of Discaria trinervis inoculated with

Frankia BCU110501 pre-incubated with DSF(s) from different bacteria.

Frankia BCU110501 was inoculated at 0.01 ml pcv root21 in all cases. The

DSF origin and preparation was described in Materials and methods.

Statistically significant differences at aP � 0.005 and bP � 0.05, unpaired

t-test (SIGMAPLOT 8.0) compared with test control.

DSF origin t0 (days)

v0 (number of nodules

plant21 day21)

None (Evans 1/10) 26.4 � 1.4 0.77 � 0.10

Frankia BCU110501 19.9 � 1.5a 1.29 � 0.18b

Frankia BCU11050 dead 27.1 � 1.4 0.71 � 0.06

Frankia ArI3 23.3 � 1.8 0.95 � 0.09

Streptomyces coelicolor 19.7 � 1.0a 0.96 � 0.10


which in turn signal back to the plant inducing infection

and nodule development (Mulder et al. 2005). This model

for legume–rhizobia recognition has been established

through genetic analysis in many studies with rhizobia

and plant mutants (Geurts et al. 2005). Although all

available techniques have been tried, Frankia has not yetbeen transformed, and the overlapping of the legume–

rhizobium recognition model with actinorhizal symbi-

osis has not been found yet (Wall 2000, Wall and Berry

2007). Thus, in the absence of nodulation mutants, the

present work makes use of alternative approach to study

the factors in the plant and in the microsymbiont, which

participate and modulate the course of symbiosis. A

detailed analysis of nodulation kinetic (Fig. 1) revealeddifferent factors with dose–response effect related to

nodulation (Figs 3, 5C, D and 6C, D).

All the nodulation parameters were consequently

modified when the doses of full compatibility infective

strain Frankia BCU110501 were varied (Fig. 3).

Because nodulation is a developmental process, each

step can be considered as a consequence of the previous

step. Thus, the change in all parameters because of theincrement of BCU110501 doses could be simply inter-

preted as a series of consequences of modification of

the first limiting step in the nodulation process. Alterna-

tively, many independent plant–microbe interactions

could be occurring in parallel and concurrently.

The fact that non-compatible Frankia (Table 2), even

a different actinomycete genus as Streptomyces (Fig. 6),

can replace BCU110501 for changing some of theparameters but not all of them strongly suggest that there

is more than one independent limiting step. At the same

time, those results suggest at least the existence of more

than one signal of bacterial origin involved in the process. It

has been reported that non-actinomycetal helper bacteria

increased nodulation during infection of Alnus glutinosa

root hair by increasing root hair deformation (Knowlton

et al. 1980). We found no similar effects when assayedwith rhizospheric non-actinomycetal bacteria (Table 2).

This difference could be attributed to the different infection

pathway between D. trinervis and A. glutinosa or could

be that the strains that we used were not expressing the

‘helper function’ observed by Knowlton et al. (1980).

The time lag for nodulation t0 seems not to reflect

symbiotic recognition because it can be modified by DSF

from non-infective Frankia or even from S. coelicolor(Tables 2 and 3). Moreover, it seems to be a step where

the actinomycete nature of Frankia is expressed because

other non-actinomycete bacteria could not replace this

effect (Table 2).

The initial nodulation rate v0 appears to be an expression

of symbiotic recognition between partners as only the

infective Frankia BCU110501 was able to vary v0. Neither

other Frankia nor other bacteria did it. This result suggests

that a recognition signal would be of bacterial origin.

The variation of the concentration of plant factors up to

3� (Fig. 5) reduced the t0, suggesting that a plant factor is

also involved in the very early interaction that conditions

the beginning of nodulation process. Plant factors concen-tration also modified the final number of nodules per plant

and the nodule cluster profile along the main root (Fig. 5,

Table 1), an expression of the autoregulation phenome-

non. This effect suggests the existence of a plant–plant

interaction mediated by DSFs from the plant. From our

assays, we could not know if this observation is related to

flavonoid effects on nodulation described for A. rubra

(Benoit and Berry 1997). The autoregulation of nodulation(Valverde and Wall 1999a, Wall and Huss-Danell 1997)

could be considered as a plant defense response to

infection to prevent hypernodulation and it implies the

inhibition of new infections in the growing root and the

control of nodule development once the infection took

place. In the experiment to study the effects of plant factors,

a variable bacteria–plant ratio was introduced between

treatments because we kept the bacterial concentration inthe pouch constant but changed the number of plants per

pouch. Taking into account the following observations, we

consider the variation in bacteria–plant ratio not to be

relevant. The increase in plant factors shifts the nodule

cluster basipetally in the root by changing the positions of

the extreme nodules in the cluster and the length of the

nodule cluster (Table 1). Those changes in nodulation

pattern could not be assigned to bacterial effect on thebasis of 3� reduction in bacteria–plant ratio within

treatments because we did not find corresponding

variations in the nodulation profile when the bacterial

inoculum concentration was varied from 0.01 to 20 ml of

Frankia pcv per plant, producing higher changes in

bacteria–plant ratio than 3� (Fig. 2). Moreover, when the

bacteria concentration and consequently the bacteria–

plant ratio was increased, the position of the first nodulewas shifted basipetally (Xi ¼ 13 and 18 for 0.01 and 0.1 ml

pcv of FrankiaBCU110501, respectively). Similar shift was

found when plant factorwas increased (Table 1, lineXi) but

in that case, bacteria–plant ratio was reduced by 3�.

Nevertheless, we cannot discard that working at a limiting

dose of Frankia in the inoculum (Fig.2B), a 3� reduction in

the bacteria–plant value could enhance the effect of plant

factors on nodulation.It is worth noting that the increase in the dose of Frankia

BCU110501 did not show a strict shift of the nodules

basipetally toward the mature zone in the root (Fig. 2C) as

it would be expected considering the reduction in t0. On

the contrary, total time of nodule development tNOD was

increased with the concentration of bacterial inoculums

and the length of the nodule cluster was also increased


(Xi 2 XL ¼ 13.4 and 21.9 mm for inoculums of 0.01 and

20 ml pcv of Frankia BCU110501, respectively). It seems

that plant autoregulation of nodulation could be counter-

acted by Frankia factors without expressing necessarily

symbiotic recognition because both BCU110501 and

ArI3 produced similar effects (Figs 2, 3 and 6).In conclusion, we think that the analysis of nodulation

kinetics is an appropriate tool to investigate factors, either

signals or physical components, involved in the symbiotic

interaction leading to the formation of nitrogen-fixing

nodules. At the same time, we think that this analysis

will be useful to investigate the role of particular subst-

ances as signals or molecular components involved in

the nodulation process.

Acknowledgements – Financial support was obtained

through grants from Universidad Nacional de Quilmes

(Argentina) PPUNQ 0340/03, Agencia Nacional de Promocion

Cientıfica y Tecnologica (Argentina) PICT 10006 and PICT

20568 and CONICET PIP 5812. We would like to thank

Professor Gabriel Favelukes for critical reading of the manu-

script and his valuable comments. L. A. G. holds a fellowship

from CONICET (Argentina) and L. G. W. is a member of the

Scientific Research Career of CONICET (Argentina).

References

Benoit LF, Berry AM (1997) Flavonoid-like compounds from

seeds of red alder (Alnus rubra) influence host nodulation

by Frankia (Actinomycetales). Physiol Plant 99: 588–593

Berry A, Torrey JG (1979) Isolation and characterization in

vivo and in vitro of an Actinomycetous endophyte from

Alnus Rubra Bong. In: Gordon J, Wheeler C, Perry D (eds)

Symbiotic Nitrogen Fixation in the Management of

Temperate Forest. Forest Research Laboratory, Oregon

State University, Corvallis, pp 69–83

Caetano Anolles G, Favelukes G (1986) Quantitation of

adsorption of rhizobia in low numbers to small legume

roots. Appl Environ Microbiol 52: 371–376

Ceremonie H, Debelle F, Fernandez MP (1999) Structural

and functional comparison of Frankia root hair deforming

factor and rhizobia Nod factor. Can J Bot 77: 1293–1301

Chaia E (1998) Isolation of an effective strain of Frankia

from nodules of Discaria trinervis (Rhamnaceae). Plant

Soil 205: 99–102

Chaia EE, Valverde C, Wall LG (2006) Local adaptation of

Frankia to different Discaria (Rhamnaceae) host species

growing in patagonia. Curr Microbiol 53: 523–528

Geurts R, Fedorova E, Bisseling T (2005) Nod factor signaling

genes and their function in the early stages of Rhizobium

infection. Curr Opin Plant Biol 8: 346–352

Glick B, Changping L, Ghosh S, Dumbroff E (1997) Early

development of canola seedlings in the presence of the

plant growth-promoting rhizobacterium Pseudomonas

putida GR12-2. Soil Biol Biochem 29: 1233–1239

Holguin G, Glick B (2001) Expression of the ACC deaminase

gene from Enterobacter cloacae UW4 in Azospirillum

brasilense. Microb Ecol 41: 281–288

Knowlton S, Berry AM, Torrey JG (1980) Evidence that

associated soil bacteria may influence root hair infection

of actinorhizal plants by Frankia. Can J Microbiol 26:

971–977

Mulder L, Hogg B, Bersoult A, Cullimore J (2005) Integration

of signalling pathways in the establishment of the

legume-rhizobia symbiosis. Physiol Plant 123: 207–218

Murry M, Fontaine M, Torrey G (1984) Growth kinetics and

nitrogenase induction in Frankia sp. HFPArI 3 grown in

batch culture. Plant Soil 78: 61–78

Nesme X, Normand P, Tremblay F, Lalonde M (1985)

Nodulation speed of Frankia sp. on Alnus glutinosa,

Alnus crispa, and Myrica gale. Can J Microbiol 63:

1292–1295

Nittayajarn A, Baker D (1989) Methods for the quantification

of Frankia cell biomass. Plant Soil 118: 199–204

Torrey J, Callaham D (1982) Structural features of the vesicle

of Frankia sp. CpII in culture. Can J Microbiol 28: 749–757

Valverde C, Wall LG (1999a) Regulation of nodulation in

Discaria trinervis (Rhamnaceae)–Frankia symbiosis. Can J

Bot 77: 1302–1310

Valverde C, Wall LG (1999b) Time course of nodule

development in the Discaria trinervis (Rhamnaceae) –

Frankia symbiosis. New Phytol 141: 345–354

Valverde C, Wall L, Huss-Danell K (2000) Regulation of

nodulation and nodule mass relative to nitrogenase activity

and nitrogen demand in seedlings of Discaria trinervis

(Rhamnaceae). Symbiosis 19: 167–182

Valverde C, Ferrari A, Wall LG (2002) Phosphorus and the

regulation of nodulation in the actinorhizal symbiosis

between Discaria trinervis (Rhamnaceae) and Frankia

BCU110501. New Phytol 153: 43–51

Van Ghelue M, Løvaas E, Ringø E, Solheim B (1997) Early

interactions between Alnus glutinosa and Frankia strain

Arl3. Production and specificity of root hair deformation

factor(s). Physiol Plant 99: 579–587

Wall LG (2000) The actinorhizal symbiosis. J Plant Growth

Regul 19: 167–182

Wall LG, Huss-Danell K (1997) Regulation of nodulation in

Alnus incana-Frankia symbiosis. Physiol Plant 99: 594–600

Wall L, Berry A (2007) Early interactions, infection and

nodulation in actinorhizal symbiosis. In: Pawlowski KN,

William E (eds) Nitrogen-Fixing Actinorhizal Symbioses.

Springer, Dordrecht, the Netherlands, pp 147–166

Wall LG, Valverde C, Huss-Danell K (2003) Regulation of

nodulation in the absence of N2 is different in actinorhizal

plants with different infection pathways. J Exp Bot 54:

1253–1258

Edited by J. K. Schjørring


Documents

Analysis of nodulation kinetics in Frankia–Discaria trinervis symbiosis reveals different factors involved in the nodulation process