The response of soybean to nod factors and a bacteriocin

ABSTRACT

Microbe-to-plant signals can enhance the growth of a wide range of crops. The responses by soybean (Glycine max var. 91M01) to 2 signal molecules were investigated: Bradyrhizobium japonicum 532C lipo-chitooligosaccharide (Nod Bj V [C:18, MeFuc]) (LCO); and Bacillus thuringiensis strain NEB17 bacteriocin thuricin 17 (Th17). The objective was to assess and quantify the response by soybean, in terms of factors that contribute to yield, to the experimental signal molecules in germination experiments and field experiments. Soybean germination was stimulated by the experimental concentrations of Th17 under controlled 15°C and 22°C conditions, and 10⁻⁶ M LCO under 15°C. There were negative relationships between Th17 concentration and both the number of trifoliate leaves and the dry weight of nodules: lower concentrations resulted in plants with more leaves and nodules while higher concentrations resulted in plants with fewer leaves and nodules. The 10⁻⁸ M LCO treatment had a significant effect on the dry weight of nodules at the flowering stage of plant development (F_4,21 = 6.06, p = 0.0019). Considering the harvest stage data from both field trials of 2011, the lower experimental concentrations of Th17 resulted in taller plants. The study of Th17 has the potential to expand our understanding of this relatively recent and unexpected finding; and to understand how best to apply this finding, to allow increased production of soybean. Collectively, these results indicate that Th17 has potential in this regard.

Introduction

Plant growth promoting bacteria (PGPB), are the free living bacteria which exist in the rhizosphere and have beneficial importance in agriculture. They can be found either in the soil near plant roots, on the surface of plant roots, or inside the cells of root nodules¹⁰ and are able to stimulate the growth of plants through a wide array of mechanisms. The more widely recognized mechanisms through which PGPB increase plant growth include: production of phytohormones; production of metal chelating siderophores; induction of induced systemic resistance; suppression of disease through antibiosis.³⁶ In the broadest sense, PGPB include the legume-nodulating rhizobia, and as such can be separated into those that reside outside plant cells (extracellular – ePGPB) and the rhizobia that reside, in the context of symbiosis, inside plant cells (intracellular – iPGPB).¹⁰ Application of PGPB to crop production systems, with the exception of rhizobia, has met with mixed results. PGPB are often quite unreliable in the field, causing increases in crop growth in some agricultural systems, at some times, and not others.²³

The nodulation of legumes, during establishment of the N₂-fixing symbiosis, is a multi-step process. Initially, phenolic compounds are produced, which trigger the activity of nod genes in rhizobia. The nod genes initiate the production of nod factors, which are lipo-chitooligosaccharides (LCOs).²⁹ All nod factor LCOs have similar structures: a 3–5 chitin unit backbone (a linear chain of β−1, 4-linked N-acetylglucosamines) linked to an acyl side chain.²² LCOs produced by rhizobia such as Bradyrhizobium japonicum 532C are involved in the preliminary exchanges of plant-to-microbe and microbe-to-plant signals that trigger legume symbiosis. The initial discovery that LCOs stimulate plant growth directly^15,25,33 was confirmed by others;²⁴ confirmed LCO stimulation of root growth in Medicago truncatula;³ showed that LCO spray on tomato accelerates flowering (a typical response to stress), and increases yield. Enhanced germination and seedling growth, along with the mitogenic activity of LCOs,³⁵ suggest accelerated meristem activity.²⁵ Several compounds secreted by other rhizobacteria cause similar effects,¹⁷ although chemically they are quite different (proteins - Gray et al.^8,9 from the chitin-based LCOs.

The PGPB Bacillus thuringiensis NEB17 was isolated from soybean nodules¹ and was shown to increase growth and nodulation when applied as a co-inoculant with B. japonicum 532C.² This bacterium produces the bacteriocin Th17, with molecular weight 3.1 kDa, which is not toxic to B. japonicum 532C.⁹ Bacteriocins are bacteria-produced peptides which generally kill bacteria that are closely related to the producer strain,¹³ which provides a competitive advantage for the producer strain.³⁷ It has been already demonstrated that the application of Th17, to either leaves or roots, can enhance plant growth. Use of this bacteriocin can enhance early seedling growth, photosynthetic rate, soybean nodule number, and total fixed N.¹⁷ However, a great deal of research remains to be done regarding matters such as the range of crops affected, interaction with crop stress, and their specific effects on crop physiology and development.

Preliminary experiments established that the effects of Th17 and LCO to stimulate growth are greater under stress and that, for soybean, the effects are slight if there is no stress. For soybean, the optimum temperature for growth and nitrogen fixation is around 25 – 30°C.¹⁴ Studies have shown that low RZTs decrease growth and nitrogenase activity.^{11,14,18,21,27} Therefore, the experiments involved a moderately stressful low temperature condition (22 ± 2°C), and a severely stressful low temperature condition (15 ± 2°C). The hypothesis was that the beneficial effects of the 2 signal compounds would be greater under severe stress than moderate stress.

The general objective of this study was to assess the effect of Th17 and LCO on germination in controlled environment experiments, and on factors that contribute to soybean yield in field experiments. The experiments were conducted under a range of growing conditions, to determine the responses of soybean to Th17, in comparison to a water only negative control, with LCO as a positive control.

Materials and methods

Extraction of lipo-chitooligosaccharide

Lipo-chitooligosaccharide Nod Bj V (C:18, MeFuc) was extracted according to the method described in Souleimanov et al.³²

Extraction of thuricin 17

Thuricin 17 was extracted according to the method described in Gray et al.⁸ (Bacillus thuringiensis NEB17 is available from the KemX Global, Boone, Iowa, USA.)

Germination assays

Germination experiments were conducted under controlled environment conditions in germination chambers (Model TC30, Controlled Environments Ltd., Winnipeg, Canada) provided by the Department of Plant Science, Macdonald Campus, McGill University. The experiments involved a moderately stressful low temperature condition (22 ± 2°C), and a severely stressful low temperature condition (15 ± 2°C). Germination was carried out under dark conditions for both temperatures. Soybean (Glycine max variety Absolute RR, provided by Belcan, Canada) was used to determine the effects of Th17 and LCO on the progression of seed germination. The treatments were comprised of 2 concentrations of Th17 (10⁻⁹ and 10⁻¹¹ M) and 2 concentrations of LCO (10⁻⁶ and 10⁻⁸ M) plus a distilled water control treatment. Seeds were screened for uniformity of size and overall condition (no obvious damage to the seeds). Ten seeds were placed in each sterile Petri dish (9 cm diam.) lined with filter paper (Qualitative P8, 9 cm diam., Fisher Scientific). Four mL of treatment solution were introduced into each Petri dish. Each dish was then sealed using parafilm. The resulting sets of Petri dishes, with seeds and added treatment solutions, were incubated under dark conditions in germination chambers at the appropriate temperatures.

The experiment was organized following a completely randomized design (CRD) with 5 replications of each treatment. The seeds that had completed germination were counted approximately simultaneously, every 6 h, beginning 24 h after establishment of the experiment. The data were therefore interval data that represented the numbers of germination events occurring within the various time intervals.²¹ Each set of experiments was repeated 4 times.

Field trials

Field experiments were conducted for 2 years, in 2010 and 2011, at the E.A. Lods Agronomy Research Center (45° 25′ 45″N latitude and 73° 56′ 00″ longitude) of the Macdonald Campus of McGill University, Ste- Anne-de-Bellevue, Quebec, Canada, from May to September. In 2010, one field experiment was conducted while in 2011 2 field trials were conducted. In 2010, the soybean was planted on May 12 and relatively low temperatures were expected during seedling emergence. In 2011, 2 soybean trials were planted at separate dates: June 3 and June 22. The first 2011 trial was an attempt to capture temperatures lower than the optimum growing temperatures for soybean seeds (low temperature stress for soybean).

The experiments were organized following a Randomized Complete Block Design (RCBD), in both years, with 4 blocks; each block contained 7 plots in 2010 and 8 plots in 2011. Each block measured 4 × 18 m, with a 1.3 m wide pathway between adjacent blocks. Each plot was comprised of 9 rows of soybean plants; rows within plots were 18 cm apart.

Soybean (G. max variety 91M01, Pioneer) was used in all field trials. Three concentrations of Th17 (10⁻⁹, 10⁻¹⁰ and 10⁻¹¹ M) and LCO (10⁻⁶, 10⁻⁷ and 10⁻⁸ M) were used. A water treatment was included as a control. The field site was plowed to a depth of 20 cm in the autumns of 2009 and 2010 and harrowed during springs of 2010 and 2011, to provide a good seedbed. Fertilizer (4:20:20 NPK at 250 kg ha⁻¹) was applied and the site was harrowed. The rows were prepared using a Plot Man seeder (Swift Current Saskatchewan, Canada). Once the rows were established, sowing was done manually. The seeds were sown (approximately 660,000 seeds ha⁻¹) into the open rows and then covered immediately, using a garden hoe.

The soybean seeds were soaked in the treatment solution for 20 min. prior to planting and then air-dried. After applying the treatment solutions, the seeds were treated with the commercially available Histick N/T soybean inoculant (Becker Underwood, Saskatoon) containing 2 × 10⁹ viable B. japonicum cells and 1 × 10⁸ viable cells of Bacillus subtilis (MBI600 strain) per gram of inoculum. The inoculant was applied at 400 g per 125 kg of soybean seed.

To determine the percentage emergence, data were collected on emergence until no new seedlings were seen. During the course of the growing season, data were collected at 3 crop growth stages. The first sampling was done at the mid-vegetative stage (V(N) stage), when there were more than 3 nodes on the main stem beginning with the unifoliate node. The second sampling was done at the flowering stage (R2 or R3 stage), and last sampling at the harvest stage (R8 stage, when 95% of the pods are brown, harvest maturity).

For each sampling, 10 plants were selected randomly from each plot. At first sampling, data were collected on fresh weight of the plants, number of trifoliate leaves, leaf area, and the dry weight of the plants. At the flowering stage, data on fresh weight per plant, number of trifoliate leaves, leaf area, dry weight, number of nodules, dry weight of nodules, and dry weight of roots were collected.

When the soybean plants reached harvest maturity (R8 stage), 10 plants were selected randomly from each plot to collect data on yield components. The rest of the plot was harvested with a small-plot combine (Wintersteiger Classic, Utah, USA) to determine the total yield from each plot. From the final sampling, the following data on yield components were collected: pods per plant, seeds per plant, weight of seeds per plant, and seed weight.

Statistics

The germination data were analyzed using SAS PROC LIFETEST SAS Institute, according to the method of McNair et al.²¹: the periodic simultaneous observation times were adjusted so that the values entered as event times were slightly less than the observation times; p-values were calculated using the log-rank test with Dunnett-Hsu adjustments for multiple comparisons; this p indicates the probability that the treatment group was the same as the control.

The function S(t) is the probability that the germination time is greater than t. The estimator at the observation times a_i is:

$${\displaystyle \stackrel{\backslash ˆ}{S}\left(a_j\right)=\prod _{i=1}^j(1-D_i/N_i)}$$

Where D_j is the number of germination events occurring within interval I_j. N_i is the number of seeds at risk of germination (i.e. not yet germinated) at the beginning of interval I_i, with N₁ = N. It is most reasonable to assume events occur more or less uniformly during each interval I_i and therefore to estimate S(t) for a_i < t < a_i+1 simply by linearly interpolating between ${\displaystyle \hat{S}\left(a_i\right)}$ and ${\displaystyle \hat{S}\left(a_{i+1}\right)}$. In the accompanying figures, 1.0 indicates 100% probability of not germinating, which is expected at the beginning of every experiment, the probability of not germinating is expected to decrease as the experiment progresses, while 0.0 indicates 0% probability of not germinating, which is expected for healthy seeds given sufficient moisture, heat, and time. The pale, broken lines in Fig. 1 and Fig. 2 are point-wise 95% confidence intervals that assume a normal distribution of ${\displaystyle \hat{S}\left(a_i\right)}$.

$${\displaystyle \hat{S}\left(a_i\right)\pm 1.96\times {\text{SE}}_i}$$

Figure 1.

The probability that the germination time is greater than t under moderately stressful temperature conditions (22 ± 2°C). The pale lines indicate the point-wise 95% confidence intervals.

Figure 2.

The probability that the germination time is greater than t under severely stressful low temperature (15 ± 2°C). The pale lines indicate the point-wise 95% confidence intervals.

Where SE_i is the standard error of ${\displaystyle \hat{S}\left(a_i\right)}$

For the field trials, the means per plot were used as input data. Years and trials were considered both separately and together. The normal distribution of the model residuals were confirmed using normality tests by SAS PROC UNIVARIATE and the homogeneity of variance was confirmed using Lavene's test from SAS PROC GLM. Ratios, based on the average of the controls for the corresponding site-year, were calculated for the variables that were measured from the same plant developmental stage in 2010 and 2011. These ratios were not normally distributed and were therefore analyzed using SAS PROC GLIMMIX with the DIST=T option in the MODEL statement. SAS PROC CORR and SAS PROC MIXED were used to analyze the other field data. For correlation analysis, the concentration was coded as the absolute value of the exponent. Therefore, more dilute concentrations were coded with larger values, and a positive correlation between the treatment concentration and the variable of interest indicated a negative relationship, where a negative correlation would indicate a positive relationship.

When the year and trial data were considered together, a random component was included in the model. In all cases, the mixed model included a categorical variable that indicated the experimental signal, and a nested variable that represented the concentration of the signal.

Results

Effect of Th17 and LCO on soybean germination

The results indicated that with the use of Th17 and LCO, the germination process can be accelerated for soybean under both evaluated temperature conditions, although the effects were greater when the conditions were severely stressful.

While the experimental LCO concentrations did not affect soybean germination under 22°C conditions, soybean responded to both experimental Th17 concentrations by germinating earlier. The initial log-rank test showed that the germination probability differed among the 5 treatment groups (p = 0.0256). Subsequent comparisons using the Dunnet-Hsu adjustment showed that the experimental data did not provide evidence for difference between the probabilities of germination for the 10⁻⁶ M LCO and control groups under moderately stressful temperature conditions (p = 0.1212); nor for 10⁻⁸ M LCO and control groups (p = 0.5743). Nevertheless, the experimental data provided moderate evidence for a difference between the probabilities of germination for the 10⁻⁹ M Th17 and control groups (p = 0.0188); and a difference between the probabilities of germination for the 10⁻¹¹ Th17 and control groups (p = 0.0190) (Fig. 1).

While the 10⁻⁸ M LCO treatment did not affect soybean germination under 15°C conditions, soybean responded to the other experimental treatments by germinating earlier. Using the log-rank test, the experimental data provided convincing evidence for differences in the probabilities of germination between the 5 treatment groups (p = 0.0001). The data provided no evidence of any difference between the probabilities of germination for the 10⁻⁸ M LCO and control groups under severely stressful low temperature conditions (p = 0.1179). The data provided very convincing evidence for a difference between the probabilities of germination of the 10⁻⁶ M LCO group versus the control group (p < .0001); and for the difference between the probabilities of germination for the 10⁻¹¹ M Th17 group vs. the control group (p = 0.0006); and for the difference between the 10⁻⁹ M Th17 group and the control group (p = 0.0089) (Fig. 2).

When the temperature was very stressfully low (15°C), treatment with Th17 or LCO enhanced the germination of soybean. Germination at 15°C began much later than at 22°C. Both Th17 and LCO treatments can accelerate germination relative to the water control.

Effect of Th17 and LCO on soybean growth and yield under field conditions

The germination experiments showed positive effects due to the experimentalsignals, therefore field trials were conducted for 2 y to examine the effects of the Th17 and LCO under field conditions. One trial was conducted in 2010 and 2 trails were conducted in 2011. The 2011 trials were planted at different dates, to determine the response of soybean under early season low temperature during germination. The earlier seeded first trial and the later seeded second trial in 2011 were compared. Daily temperature (minimum and maximum reading) and precipitation values for the months of May-June 2010 and June-July 2011 are presented in Figs 3, 4, 5, and 6 (Source: www.climate.weatheroffice.gc.ca). Data were collected on various growth and development variables, as indicated in the materials and methods, from all the trials of 2010 and 2011.

Figure 3.

Daily temperature reading (maximum and minimum) for the month of May and June, 2010 (growing season for soybean). The temperature is presented in °C (Source: www.climate.weatheroffice.gc.ca).

Figure 4.

Daily total precipitation for the month of May and June, 2010 (growing season for soybean). Total precipitation is presented in millimeters (Source: www.climate.weatheroffice.gc.ca).

Figure 5.

Daily temperature reading (maximum and minimum) for the month of June and July, 2011 (growing season for soybean and potato). The temperature is presented in °C. (Source: www.climate.weatheroffice.gc.ca).

Figure 6.

Daily total precipitation for the month of June and July, 2011 (growing season for soybean and potato). The amount of total precipitation is presented in millimeters. (Source: www.climate.weatheroffice.gc.ca).

Considering the data from the flowering stage of soybean development (from 2010 and both trials from 2011 together) there was a negative correlation between (the absolute value of the exponent for) LCO concentration and nodule number (Pearson r = −0.538, p = 0.0714), therefore the data suggested a positive relationship between LCO concentration and nodule number at the flowering stage of plant development.

With varying degrees of statistical significance, the (inverse-coded) concentration of the Th17 treatment was positively correlated to soybean trifoliate leaf number (r = 0.636, p = 0.0263); leaf area (r = 0.504, p = 0.0949); total dry weight (r = 0.539, p = 0.0705); and the dry weight of nodules (r = 0.677, p = 0.0156); but not the number of nodules at the flowering stage of plant development (r = 0.156, p = 0.6290). There was, therefore, when the whole data set was considered, a negative relationship between Th17 concentration and both the number of trifoliate leaves and the dry weight of nodules: lower concentrations of Th17 resulted in more leaves and nodules.

The experimental molecules had a significant effect on the number of trifoliate leaves (F_2,21 = 3.55, p = 0.0470). The Bonferroni-adjusted limits at 95% confidence indicated that the number of trifoliate leaves on plants treated with LCO would be 0.1 fewer to 4.6 more than on plants treated with Th17.

Regarding total dry weight, the treatment means of the ratios (experimental plot value : average site-year control value) were not equal (F_6,901 = 5.30, p < .0001) (Table 1). Based on the Dunnett-adjusted limits, the total dry weight would be 26 to 2% less for soybean treated with 10⁻⁹ M Th17.

Table 1.

Least Squares-Means for Ratios (Plot Value : Average Site-Year Control Value) Based on the Experimental Soybean Variables^†.

		Flowering Stage of Plant Development
		Number of Trifoliate Leaves		Leaf Area		Total Dry Weight		Number of Nodules		Dry Weight of Nodules		Root Dry Weight		Harvest StageTotal Dry Weight per Plant
Treatment	Concentration	Estimate ± se	p‡	Estimate ± se	p	Estimate ± se	p	Estimate ± se	p	Estimate ± se	p	Estimate ± se	p	Estimate ± se	p
Control	NA	1.04 ± 0.08	.	1.14 ± 0.16	.	1.09 ± 0.12	.	1.02 ± 0.16	.	1.07 ± 0.16	.	1.11 ± 0.11	.	1.00 ± 0.06	.
LCO	10−6	1.10 ± 0.08	0.5192	1.25 ± 0.16	0.1213	1.16 ± 0.12	0.4425	1.23 ± 0.16	0.0014	1.10 ± 0.16	0.996	1.17 ± 0.11	0.7738	0.85 ± 0.06	0.0898
LCO	10−7	1.09 ± 0.08	0.7312	1.2 ± 0.16	0.2911	1.16 ± 0.12	0.4128	1.08 ± 0.16	0.7397	1.01 ± 0.16	0.8806	1.12 ± 0.11	1	0.96 ± 0.06	0.9604
LCO	10−8	1.05 ± 0.08	1	1.16 ± 0.16	0.9925	1.09 ± 0.11	1	1.12 ± 0.15	0.2284	1.26 ± 0.16	0.0121	1.06 ± 0.11	0.6177	0.92 ± 0.06	0.5834
Thuricin 17	10−9	0.94 ± 0.08	0.0637	1.02 ± 0.16	0.0806	0.95 ± 0.12	0.0171	1.05 ± 0.16	0.9899	0.92 ± 0.16	0.1493	0.96 ± 0.11	0.0121	1.03 ± 0.06	0.9945
Thuricin 17	10−10	0.98 ± 0.08	0.3985	1.13 ± 0.16	0.9997	1.04 ± 0.12	0.8604	1.04 ± 0.16	0.9996	0.92 ± 0.16	0.1685	1.00 ± 0.11	0.11	0.98 ± 0.06	0.9996
Thuricin 17	10−11	1.11 ± 0.08	0.33	1.22 ± 0.16	0.4068	1.14 ± 0.12	0.6054	1.09 ± 0.16	0.6753	1.19 ± 0.16	0.3519	1.16 ± 0.11	0.8647	0.96 ± 0.06	0.9681

^†The estimated values are ratios of the measured values from the field plots to the average of the control for the same site-year. Therefore, values above one indicate an increase due to the treatment. Values below one indicate that the variable was decreased in the treatment group, relative to the control group.

^‡p values are Dunnett-adjusted, p values less than 0.05 indicate that the mean of the treatment group was unequal to the mean of the control group.

Regarding the number of nodules per plant, the treatment means of the ratios (experimental plot value : average site-year control value) were not equal (F_6,901 = 3.09, p = 0.0053) (Table 1). Based on the Dunnett-adjusted limits, the number of nodules would be 5 to 36% more for soybean treated with 10⁻⁶ M LCO.

The concentration of the experimental molecules had a significant effect on the dry weight of nodules at the flowering stage of plant development (F_4,21 = 6.06, p = 0.0019). Based on the Dunnett-adjusted limits at 95% confidence, the total dry weight of nodules from plants treated with 10⁻⁸ M LCO was 10 to 190 mg heavier at the flowering stage. The treatment means of the ratios (experimental plot value : average site-year control value) were not equal (F_6,901 = 8.13, p < .0001) (Table 1). Based on the Dunnett-adjusted limits, the total dry weight of nodules would be 3 to 36% more for soybean treated with 10⁻⁸ M LCO.

The F-test also indicated that concentration of the experimental molecules had a significant effect on the dry weight of roots at the flowering stage of plant development (F_4,21 = 3.40, p = 0.0272). The treatment means of the ratios (experimental plot value : average site-year control value) were not equal (F_6,901 = 5.01, p < .0001) (Table 1). Based on the Dunnett-adjusted limits for the ratios, the total dry weight of roots would be 27 to 2% less for soybean treated with 10⁻⁹ M Th17.

Considering the harvest stage data from both trials of 2011, the concentration of the Th17 treatment was correlated to soybean plant height at the harvest stage of plant development (r = 0.611, p = 0.0015), however, due to the inverse-coding of the concentration variable, this result indicated that lower concentrations of Th17 resulted in taller soybean plants.

The pooled harvest stage data, from both experimental trials, indicated that the treatment concentrations affected soybean plant height (F_4,57 = 2.76, p = 0.0364), but after adjustment for multiple comparisons, the treatment concentrations were not distinguishable.

The experimental data from the first sampling (soybean flowering stage) in 2010 provided moderate evidence that the molecule treatments affected the number of trifoliate leaves (means per plot), F_2,21 = 3.55, p =0.0470, but after adjustment for multiple comparisons the data was merely suggestive of an effect due to LCO. According to the Dunnet-adjusted limits at 95% confidence, the LCO-treated plants produced from 1 fewer to 5 more trifoliate leaves than controls (p = 0.0817). However the data provided no evidence that the different concentrations of the molecules affected the mean trifoliate leaf counts, F_4,21 = 2.19, p = 0.1050.

A Kruskal-Wallis test with adjustment for multiple comparisons (Dunn's test)⁷ was applied to the rank ordered values for root dry weight. Dunn's test indicated that, during the soybean flowering stage in 2010, the 10⁻⁶ M LCO treatment group produced heavier dried soybean roots than the 10⁻⁹ M Th17 group (α = 0.05). Furthermore, comparison of concentration levels of Th17 (Dunn's test, α = 0.05) indicated that 10⁻¹¹ M Th17 resulted in heavier dried roots roots than the 10⁻⁹ M Th17.

The data provided convincing evidence for the effect of the concentrations on the dry weight of nodules during the soybean flowering stage in 2010 (F_4,21 = 6.20, p = 0.0019). According to the Dunnett-adjusted limits at 95% confidence, the soybean plants treated with 10⁻⁸ M LCO produced 7.9 mg to 190.6 mg heavier dry weight of nodules, during the flowering stage in 2010, compared to the controls (p = 0.0299).

Dunn's test indicated that treatment with 10⁻⁸ M LCO resulted in heavier dry weights of nodules than either 10⁻⁹ M or 10⁻¹⁰ M Th17 during the soybean flowering stage in 2010 (α = 0.05). The comparison of concentration levels within the LCO treated group (Dunn's test, α = 0.05) indicated that the rank level of nodule dry weight was lower for 10⁻⁷ M LCO than 10⁻⁸ M LCO.

The data was suggestive but inconclusive regarding whether or not the treatment molecules affected the dry weight of pods at the fruiting stage of plant development in 2010 (F_2,21 = 3.30, p = 0.0566). The concentrations, however, did not have a statistically significant effect on the dry weight of pods (F_4,21 = 0.77, p = 0.5587).

From the second sampling (fruiting stage) of 2011 from the early-planted trial, lower concentrations of Th17 were positively related to total dry weights of soybean (Pearson r = 0.58112, p = 0.0475). The comparison of concentration levels within the Th17 treated group (Dunn's test, α = 0.05) indicated that the rank levels for the leaf area were higher for 10⁻¹¹ M Th17 than 10⁻¹⁰ M Th17. Similarly, Dunn's test indicated that the rank levels for the dry weight of roots were higher for the plants treated with 10⁻¹¹ M Th17 than plants treated with 10⁻¹⁰ M Th17.

From the third sampling (harvest stage) of 2011 from the early-planted trial, lower concentrations of Th17 were positively correlated with greater numbers of soybean pods (r = 0.58829, p = 0.0442).

From the combine-yield data of 2011, from the early-planted trial, lower concentrations of Th17 were negatively correlated with the moisture content of soybeans (r = −0.57642, p = 0.0498), and therefore higher concentrations of Th17 were correlated with higher moisture contents of soybeans.

From the third sampling of 2011, from the late-planted trial, lower concentrations of Th17 were negatively correlated with the height of soybeans (r = −0.70406, p = 0.0106), and therefore higher concentrations of Th17 were correlated with taller soybean plants. Dunn's test (α = 0.05) confirmed that the ranks of the ordered height values were higher for plants treated with 10⁻⁹ M Th17 than those treated with 10⁻¹¹ M Th17.

From the third sampling (harvest stage) of 2011, from the late-planted trial, the data provided moderate evidence for the effect of the experimental molecules, but not the concentrations of the treatments, on hundred seed weight (F_3,24 = 3.91, p = 0.0210; F_4,24 = 0.56, p = 0.6973), but the Dunnett-adjusted multiple comparsions showed no significant differences between the treatments.

Discussion

The experimental results showed that both 10⁻⁹ M Th17, 10⁻¹¹ M Th17, and 10⁻⁶ M LCO, but not 10⁻⁸ M LCO accelerated the germination of soybean relative to the water control in a controlled environment. When the temperature was low (15 ± 2°C), the soybean response to both Th17 and LCO was less likely to be similar to the control. Under moderate (22 ± 2°C) temperature conditions, the p-values were less extreme: under lower temperature conditions, the response by soybean to the signals was relatively pronounced. These results showed that both Th17 and LCO can enhance plant growth, which agree with the previous findings of Lee et al.¹⁷ for Th17 and Prithiviraj et al.²⁵ for LCO. Previous experiments with Th17 under field conditions showed that Th17 enhances germination of corn seeds but, at that time the temperature during the germination was around 13 to 15°C.¹² The effect of 10⁻⁶ M LCO was also previously reported to affect non-legume germination,³⁰ found that 10⁻⁶ M LCO increased Brassica napus cv. Polo germination by 75% during the 5 – 15 growing degree day period.

For soybean, the germination process is much slower at 15°C than at 22°C. It has been shown that a low root zone temperature (RZT) not only delays germination but also decreases nodulation and nodule function.²⁰ The time between soybean inoculation and the start of nitrogen fixation was delayed by 2 d for each degree decrease in temperature between 25 and 17°C, and this delay in time increased to one week for each degree decrease in temperature between 17 and 15°C.³⁸ As indicated above, with the use of Th17 and LCO treatments, the soybean germination process can be enhanced at low RZT (15°C). Based on recent publications, Th17 and LCO treatments show greater effects under stressful conditions, such as salinity,³⁴ low temperature, or drought.²⁶

Except for the imposed moderate and severe low temperature stress conditions, the experiments were carried out under otherwise ideal growing conditions, so all the seeds that could germinate eventually did germinate, however, under field conditions this is unlikely to be the case; slower germination can result in less emergence. Hence, under such stressful conditions, Th17 and LCO are expected to help crops overcome stress and maintain good seed germination.

Growing conditions in the field are unlikely to be as ideal as in the growth chambers. There are various factors in the field that cannot be controlled and that have potentially direct effects on how the treatments work. Thus, the field trials were conducted to examine the effect of Th17 and LCO. The 2010 field trial was started on May 12 and the average maximum and minimum daily temperature for 2 weeks after sowing seeds was 23°C and 9.5°C respectively. Due to higher temperatures (around 23°C), the emerging soybeans were not subjected to low temperature conditions. There was relatively little precipitation during May 2010 (Fig. 4) and the emergence of the soybean plants was only ∼45 - 65% one month after seeding.

The previous findings of Lee et al.¹⁷ and Prithiviraj et al.²⁵ also indicated an effect due to the experimental signals on leaf area, total plant weight, and nodule weight. Overall, the data was suggestive of a negative relationship between Th17 concentration and leaf area, which indicated that the low-concentration experimental Th17 treatments stimulated greater leaf area per plant. The comparison of leaf areas - during the fruiting stage of plant development in 2011, between concentration levels of Th17 in the early-planted trial - indicated that 10⁻¹¹ M Th17-treated plants had larger leaf areas than plants treated with 10⁻¹⁰ M Th17. Similarly, the data overall was suggestive of a negative relationship between Th17 concentration and plant dry weight, but statistically significant at the p < 0.05 level during the fruiting stage of plant development in 2011, in the early-planted trial. At the flowering stage of plant development, soybean plants that were treated with 10⁻⁸ M LCO had 10 mg heavier to 190 mg heavier total dry weight of nodules, but 4.6 fewer to 10.6 more nodules than controls (95% confidence Dunnett-adjusted limits); and the comparison of concentration levels within the LCO treated group (Dunn's test, α = 0.05) indicated that the rank level of nodule dry weight was lower for 10⁻⁷ M LCO than 10⁻⁸ M LCO. Lower concentrations of Th17, however, were correlated with heavier dry weights of nodules. Lower concentrations of Th17 were also correlated to taller soybean plants at the harvest stage of plant development in 2011. Previously,³¹ found that B. napus plants grown with 0.2 M NaCl and 10⁻⁹ M Th17 were taller than controls after 38 d of growth in plant culture vessels.

In 2011, the soil temperatures for the early-planted trial were expected to be in the range of 15 to 17°C during the week following seeding, based on long-term norms. Nevertheless, the temperature conditions were generally above 22°C for that week. Due to the unanticipated weather conditions, low temperature conditions did not occur during germination.

Previous field experiments showed that the use of LCO increased the yield of soybean by up to 25 %,¹⁵ but in that experiment the LCO was repeatedly sprayed on soybean leaves, constituting a chronic application. Different methods of application may have different effects, but year-to-year environmental factors affect yield to a statistically significant degree, and meta-analysis could be done to differentiate the sources of variation.

The increase in plant growth, either germination or dry matter accumulation in roots or pods may be attributed to hormone-like effects of LCOs.²⁵ Nod factors are known to induce the cell division²⁸ not only in legumes but also in non-legumes.^4,5 LCO-like molecules also stimulate early somatic embryo development in Norway spruce.⁶ Enhanced germination and seedling growth, along with the mitogenic nature of LCOs,³⁵ suggest accelerated meristem activity. This may lead to increased sink demand and the observed increases in mobilization of seed reserves²⁵ and increased photosynthetic rates¹⁵ for more developed plants; both of which lead to increased growth.¹⁶ LCOs are known to activate nod genes and act as mitogens.¹⁹ The results reported are consistent with this view however, the interaction with low temperature conditions adds an additional component to the earlier findings, one that is now being clearly demonstrated, but that is not well understood at this time.

In this study, the effects on soybean growth of acute Th17 and LCO seed treatments were investigated, under a range of growing conditions, from growth chambers to the field. This is the first study conducted to determine the effect of acute exposure of Th17 on soybean growth under field conditions. Soybean is sub-tropical in origin and requires temperatures of 25 – 30°C for the optimum growth. Th17 was demonstrated to stimulate germination under both moderate (22°C) and low (15°C) temperature conditions, while LCO stimulated germination under low temperature conditions. Based on the limits at 95% confidence, the 100 seed weight from the early-planted trial (in the absence of low temperatures during germination) would be 3.6 g heavier to 7.6 g heavier compared to late-planted trials (p < 0.0001), therefore earlier planting would increase soybean yield in this region. The experimental treatments affected factors that contribute to yield. Lower concentrations of Th17 resulted in plants with more trifoliate leaves and more nodules (dry weight) while higher concentrations of Th17 resulted in plants with fewer leaves and nodules. The 10⁻⁸ M LCO treatment had a significant effect on the dry weight of nodules at the flowering stage of plant development (F_4,21 = 6.06, p = 0.0019). Considering the harvest stage data from both trials of 2011, the lower experimental concentrations of Th17 resulted in taller plants. Soybean plants treated with 10⁻⁸ M LCO produced 7.9 mg to 190.6 mg heavier dry weight of nodules (p = 0.0299) at the flowering stage of plant development in 2010, compared to the controls. Lower concentrations of Th17 were correlated with higher total dry weight of plants at the fruiting stage of development from early planted soybeans in 2011 (Pearson r = 0.58112, p = 0.0475); treatment with 10⁻¹¹ M Th17 resulted in greater leaf area and heavier root dry weight than 10⁻¹⁰ M Th17 during the fruiting stage of plant development in 2011 in the early-planted trial. At the harvest stage of plant development in trial 1 in 2011, lower concentrations of Th17 were positively correlated with the number of pods (r = 0.58829, p = 0.0442). Higher concentrations of Th17 were correlated with higher moisture content of soybeans from the early-planted trial in 2011. However, in the late-planted trial in 2011, higher concentrations of Th17 were correlated with taller soybean plants: in the late-planted trial in 2011 plants treated with 10⁻⁹ M Th17 were taller than those treated with 10⁻¹¹ M thuricin (α = 0.05). As Th17 is relatively new to agricultural research, in comparison to LCO, which is available commercially, Th17 was compared to LCO in these experiments. Collectively, this study showed that Th17 has potential to be used as a commercial technology, as a crop growth enhancer. The experimental treatments showed potential to increase the weight of root nodules while not decreasing the number of nodules, in which case future work should explore whether or not the treatments affect soybean yield in nitrogen-depleted soils.

Diclosure of potential conflicts of interest

Kaberi Gautam declares that she has no conflict of interest. Timothy Schwinghamer declares that he has no conflict of interest. Donald Smith declares that he has no conflict of interest.

Funding

Financial support for the conduct of the research was provided by a CRD grant (231686) from NSERC and Blueleaf.

Ethical approval

This article does not contain any studies with human participants or animals performed by any of the authors.