Introduction

The establishment of nodulation and nitrogen fixation in legumes is affected by different soil factors such as pH, availability of nitrogen, calcium, and interactions with rhizosphere microorganisms (Solans et al. 2009). Although N is required in significant amounts (~30 kg/t DM) to optimise plant growth (Pattison et al. 2010), larger amounts (>50 kg/ha) of available N in soil generally have a negative effect on legume nodulation and nitrogen fixation (Streeter 1985, 1988; Puiatti and Sodek 1999; Lucínski et al. 2002; Herridge et al. 2005), with many stages of the symbiotic process inhibited (Dogra and Dudeja 1993; Dusha 2002; Mortier et al. 2012). However, a low, static concentration of ammonium (0.5 mM NH4 + or 7 mg/kg N) can result in stimulation of nodulation in pea (Pisum sativum) (Gudden and Vessey 1997; Fei and Vessey 2003).

Plant mutation and breeding approaches have been used to produce nitrate tolerant soybean (Carroll et al. 1985a, 1985b; Reid et al. 2011), but the results of that work have not been widely adopted due to other agronomic shortcomings. Manipulation of the soil microbiology and in particular the application of bacteria, including actinobacteria, is another strategy that is currently being investigated (Li and Alexander 1988, 1990; Le et al. 2014). The potential for actinobacteria to modify legume symbioses was raised by (Sharma et al. 2005) who found a high abundance of actinobacteria in the rhizosphere of three different legumes, faba beans (Vicia faba L., cv. Scirocco), peas (Pisum sativum L., cv. Duel) and white lupin (Lupinus albus L., cv. Amiga). Recent studies on the endophytic actinobacteria show that they also colonise and live within the root nodules of Lupinus angustifolius (Trujillo et al. 2006; Trujillo et al. 2007; Trujillo et al. 2010) and nodules of pea (Pisum sativum L.) (Carro et al. 2007; García et al. 2010). A number of actinobacteria have been shown to improve the growth, nodulation and N2-fixation of soybean (Gregor et al. 2003; Soe et al. 2012; Nimmnoi et al. 2014), pea (Pisum sativum L.) (Tokala et al. 2002), bean (Phaseolus vulgaris L.) (El-Tarabily et al. 2008), chickpea (Gopalakrishnan et al. 2015b) and Lucerne (Medicago sativa L.) (Samac et al. 2003; Martínez-Hidalgo et al. 2014). With regard to overcoming N inhibition of the symbiosis, Solans et al. (2009) and Solans et al. (2015) found that some actinobacteria isolated from the root nodule surface of Discaria trinervis promoted the nodulation (42 % increase in nodule mass) of Lucerne and enhanced the plant biomass and nodulation of Lotus tenuis inoculated with Mesorhizobium loti when grown in the presence of high levels of soil N (7 mM NH4NO3). This study examines the application of three endophytic actinobacteria (Streptomyces spp.) isolated from wheat and Lucerne on the nodulation and growth of Lucerne plants growing in potting media supplied with three levels of NH4NO3.

Materials and methods

Actinobacteria, rhizobia and Lucerne seeds

Seeds of the Lucerne cultivar ‘SARDI Ten’ and Sinorhizobium meliloti strain RRI 128 (the strain used to produce commercial rhizobia inoculants for Lucerne in Australia) were used in all experiments. The three actinobacteria used were Streptomyces sp. EN23, isolated from healthy wheat roots, and Str. spp. LuP30 and LuP47B isolated from healthy Lucerne roots. All three actinobacteria had been shown to increase the growth of Lucerne nodulated with the S. meliloti strain RRI 128 in previous studies, under low soil N conditions (3 mg NH4NO3/kg soil) (Le et al. 2014).

Growth of actinobacteria and rhizobia

Actinobacteria were grown on Mannitol Soya Flour agar (MS) plates and incubated at 27 °C for 7–14 days to obtain spores. The colony-forming units (CFU) of the actinobacterial spores were enumerated as described by the Miles and Misra (1938) technique. Sinorhizobium meliloti strain RRI 128 was streaked onto Yeast Mannitol Agar (YMA) and a single pure colony selected, transferred and maintained on YMA plates or slants for use in the experiments. A standard curve of the relationship between S. meliloti CFU (Miles and Misra 1938) and OD600nm was determined and used to standardise the titre of rhizobial cells in each experiment to approximately 108 cells/ml.

Plant establishment

Seeds of Lucerne cultivar ‘SARDI Ten’ were surface sterilised following the protocol of Coombs and Franco (2003). Seeds were removed from the final water rinse after 10 min, and placed in a laminar flow cabinet overnight to dry. Actinobacteria were applied to the surface of the sterilised Lucerne seeds as a suspension in 0.3 % xanthan gum to provide 108 CFU/g seed and the seeds sown (10 seeds/pot) into a pasteurised (by autoclaving) potting mix (1 kg/pot of 50:50 by volume of sand:vermiculite) contained in 1.25 l self-watering pots (Décor Watermatic™). Seedlings were thinned to 4/pot after 6 days, before inoculation with rhizobia strain RRI 128.

Experimental treatments and design

The factorial experiment comprised (i) 3 strains of actinobacteria, (ii) ± inoculation with S. meliloti RRI 128 and (iii) 3 levels of soil NH4NO3 (3, 25 and 50 mg/kg) which were added with 200 mL of McKnight’s solution on planting day (McKnight 1949). Milli-Q (MQ) water was added to the pots as required in the following weeks. Plants designated to rhizobia treatments were inoculated with an aqueous suspension (1 ml per plant containing 108 CFU) of rhizobia 6 days after sowing. Each treatment was replicated eight times. Pots were arranged in a completely randomised design in a greenhouse and moved weekly to reduce spatial effects. Four pots (replicates) of each treatment were harvested at 4 and 7 weeks after inoculation. The experiment was carried out in the winter of 2013 (approx. Mean 10 h/14 h day/night, temp. range 5-20 °C).

Experiment 2

The 15N experiment to determine the relative contribution of soil and fixed N sources at different stages of plant growth was carried out as described in the above experiment except that only two strains, LuP30 or LuP47B, were applied separately to the seeds prior to inoculation with rhizobia strain RRI 128. A control treatment (rhizobia strain RRI 128 only) was also included. Each treatment was replicated 12 times. Plants were supplied with 25 mg 15NH4 15NO3 (98 %, Cambridge Isotope Laboratories, Inc.), per kg of soil mix with 200 mL of McKnight’s solution (McKnight 1949) on planting day. The experiment was carried out in the winter of 2014 (approx. Mean 10 h/14 h day/night temp. Range 5-20 °C). Four pots (replicates) of each treatment were harvested at 1, 3 and 5 weeks after inoculation with rhizobia strain RRI 128.

Parameters measured

After harvest the height of shoots, dry weight of shoots and roots, and the number and dry weight of nodules per plant were measured. The height of the shoots was measured as the distance from the base of the dominant shoot at sand level to the youngest leaf, and an average calculated for the four plants in each pot. Shoots and roots were dried separately in an oven at 60 °C for 48 h to constant weight and average dry weights calculated for the four plants in each pot. The number and total mass of nodules per plant were determined from sampling 2 plants per pot. Nodules in the top 5 cm of roots and large nodules (diameter ≥ 1 mm) were counted. There were four pots for each treatment.

Elemental and isotopic composition analyses

Four shoots per pot and four pots for each treatment were sampled. The whole shoots were dried at 60 °C for 48 h and ground using a clean mortar and pestle to about 1 mm in size and sent to the Waite Analytical Services, University of Adelaide, Australia to determine N content (%), P, K and other macro and trace elements.

In experiment 2: the N content of the shoots and roots was analysed by mass spectrometry to determine the proportions of plant N derived from the atmosphere and soil. All four shoots and roots from each pot were ground separately to about 1 mm in size. About 3 mg of each sample were analysed for total N and 15 N on an isotope ratio mass spectrometer (Sercon, Crewe, Cheshire, UK).

The At% of 15N has been calculated taking account of the 15N in air with the formula below

$$ At\%=\left[100\times \left(0.3663/\left(100-0.3663\right)\right)\times \left(\left(\mathrm{deltaN}/1000\right)+1\right)\right]/1+\left[\left(0.3663/\left(100-0.3663\right)\right)\times \left(\left(\mathrm{deltaN}/100\right)+1\right)\right]. $$

Statistical analysis

All data were assessed for normality and log transformed prior to ANOVA where necessary. Data was analysed using ANOVA and significant differences between means determined using Duncan’s Multiple Range test. All references to significance in the text imply statistical significance at P < 0.05, unless otherwise stated.

Results

Effect of actinobacteria on Lucerne growth in the absence of rhizobia

In the absence of the rhizobial partner, the three actinobacteria EN23, LuP30 and LuP47B significantly increased the shoot dry weights by between 19 % to 33 % in the 25 mg/kg NH4NO3 treatment; and 23 % to 24 % in the 50 mg/kg NH4NO3 treatment (Table 1). In contrast, root dry weight was reduced significantly by LuP30 (−21 %) and LuP47B (−29 %) at 25 mg and EN23 (−18 %) at 50 mg NH4NO3 after 7 weeks. The actinobacteria significantly increased the shoot:root ratio at 25 mg/kg and 50 mg/kg NH4NO3 while there were no effects at 3 mg/kg NH4NO3. Treatment with all 3 actinobacteria significantly reduced the iron (~50 %), copper (~40 %) and manganese (~30 %) concentrations in the plant shoots (Table 2). None of the actinobacteria increased shoot N concentration. However, total shoot N was increased (+28 %) by LuP47B at 50 mg NH4NO3 compared with the control treatment (Table 2).

Table 1 Effect of endophytic actinobacteria (Streptomyces spp. EN23, LuP30 and LuP47B individually or in combination with S.meliloti strain RRI 128) and soil N on Lucerne shoot and root weight at 7 weeks after inoculation
Full size table
Table 2 Effect of inoculation with Streptomyces spp. (individually or in combination with RRI 128) on the concentration (mg/kg) and total accumulation of nutrients (mg or μg) in Lucerne shoots at 4 and 7 weeks after inoculation
Full size table

Effect of actinobacteria on Lucerne growth when applied with rhizobia

The application of actinobacteria with rhizobia significantly increased the shoot weight of Lucerne (inoculated with rhizobia) at the 4 (data not shown) and 7 week harvests at 25 mg/kg NH4NO3 by 19 % to 36 % and 27 % to 46 %, respectively. The effect of LuP30 was consistent across N levels (7 week harvest). LuP47B had no effect on shoot weight at 50 mg/kg NH4NO3, while EN23 had no effect at 3 mg/kg NH4NO3 (Table 1). Effects of the actinobacteria on root weight were variable. Whilst EN23 and LuP30 increased the root dry weight in the 25 mg/kg NH4NO3 treatment at 7 weeks, LuP30 reduced the root dry weight at 3 and 50 mg/kg NH4NO3 (Table 1). LuP30 and LuP47B both significantly increased the shoot: root ratio at 3 mg/kg NH4NO3 but only LuP47B showed an increase at 25 mg/kg NH4NO3 while LuP30 increased the ratio at 50 mg/kg NH4NO3 at 7 weeks.

Application of actinobacteria affected the nutrient status of the Lucerne shoots (Table 2). In general, the application of actinobacteria with rhizobia increased or had no effect, on the accumulated amount of several nutrients (N, P, Fe, Bo Cu and Mn). Other nutrients (e.g. K) were never affected (data not shown). At 4 weeks after inoculation with rhizobia the Streptomyces spp. EN23, LuP30 and LuP47B significantly increased the total N (μg), P (μg), Bo (μg) and Mn (μg) at 25 mg NH4NO3. For example, co-inoculation with either EN23, LuP30 or LuP47B and rhizobia RRI 128 increased total N in shoot by 26, 50 and 27 % respectively though the Streptomyces spp. did not increase significantly the N concentration of shoots (Table 2). Moreover, where actinobacteria were applied, the numbers of positive/neutral/negative responses (across all NH4NO3 treatments at the 7 week harvest) were for N (4/4/1), for P (6/3/0), for Fe (4/4/1), for Bo (2/7/0), for Cu (2/6/1) and for Mn (0/9/0). Of the 18 positive responses, 7 were attributable to inoculation with LuP47B, 6 to LuP30 and 5 to EN23.

Nutrient accumulation is the product of shoot weight and nutrient concentration. In general, nutrient concentration in the Lucerne shoots was not affected or decreased, with the application of actinobacteria. The numbers of positive/neutral/negative responses (across all NH4NO3 treatments at the 7 week harvest) were for N (0/6/3), for P (1/7/1), for Fe (1/6/2), for Bo (0/5/4), for Cu (1/5/3) and for Mn (1/6/2). Of the 15 negative responses, 5 were attributable to each of the actinobacteria. Three of the 4 positive responses (not P) were attributable to EN23 in the 3 mg/kg NH4NO3 treatment.

Impacts on nodulation

The number of nodules per plant increased with the increasing N supply. This was associated with improved plant growth at the higher N levels. However, number of nodules per mg of root decreased significantly from 2.6 to 1.2 to 0.9 nodules in the 3, 25 and 50 mg/kg NH4NO3treatments at 4 weeks, respectively, but they were similar at the 7 week harvest (Table 3). The promptness of nodulation (proportion of nodules on the top 5 cm of the root) was also decreased from 14.5/22, 12.3/24 and 2.5/40 in the 3, 25 and 50 mg/kg NH4NO3 treatments (Table 4).

Table 3 Effect of actinobacteria and soil N on nodule number per plant and number of nodules per mg of root at 4 and 7 weeks after inoculation with RRI 128
Full size table
Table 4 Distribution of total nodules in the top 5 cm of roots and large nodules (diameter ≥ 1 mm) due to treatment with actinobacteria at different rates of NH4NO3. Data from harvests at 4 and 7 weeks after inoculation with RRI 128 (n = 4)
Full size table

The addition of actinobacteria increased nodule number per plant at each level of N compared to the rhizobia treatment alone, after 4 weeks plant growth. For example, at 25 mg/kg NH4NO3 all three actinobacteria increased the number of nodules per plant by more than 75 %. EN23 and LuP30 also increased the number of nodules at 50 mg/kg NH4NO3 (Table 3). By 7 weeks, the increased nodulation associated with the actinobacteria only persisted in the 3 mg/kg NH4NO3 treatment for LuP30 and LuP47B. Inoculation with actinobacteria and N level also affected the position of nodules on the root system (Table 4). The number of nodules in the top 5 cm of the root was reduced by increasing the amount of N, but the effect was moderated by the application of EN23 or LuP30 especially in the 3 mg/kg NH4NO3 treatment after 4 weeks. Nodule size was also increased with actinobacterial treatment (Table 4). There were several examples where total nodule mass/plant increased with the addition of actinobacteria (Fig. 1). At the 4 week harvest with 25 mg/kg NH4NO3, strains EN23, LuP30 and LuP4B increased nodule mass per plant by 125 %, 140 % and 78 % respectively. At 50 mg/kg NH4NO3, strains EN23 and LuP30 increased nodule mass/plant by 63 % and 42 % respectively. At the 7 week harvest, EN23 at 50 mg/kg NH4NO3 was the only treatment that increased nodule mass.

Fig. 1
figure 1

Effect of actinobacteria and soil N (NH4NO3) on Lucerne nodule weight (mg DM/plant) at 4 and 7 weeks after inoculation (n = 4). Asterisks indicate significant differences from RRI 128 treatment at P < 0.05 (*) or P < 0.01(**)

Full size image

15N experiment

The addition of LuP30 and LuP47B generally increased shoot weight and number of nodules per plant in the 3 and 5 week harvests, confirming previous measures of their efficacy (Table 5). None of the actinobacteria increased total plant N (roots and shoots) at 3 weeks, although there was a small increase in the amount of 15N for the LuP47B treatment (Fig. 2). At the 5 week harvest, the addition of LuP30 or LuP47B increased total plant N by 40 % and 60 %, respectively. This was mostly due to greater accumulation of 14N (derived from N2-fixation) which was increased by LuP30 or LuP47B by 47 % and 72 %, respectively. The total N and 14N were distributed more in the shoots (about 70 %) and roots (~30 %) at the 5 weeks after inoculation with rhizobia (Table 6). LuP47B significantly increased the total amount of 15N in the shoots at 3 weeks and roots at the 3 and 5 weeks harvest (Table 6).

Table 5 Effect of endophytic actinobacteria (Streptomyces spp. EN23, LuP30 and LuP47B in combination with S.meliloti strain RRI 128) on Lucerne shoot and root weight at 1, 3 and 5 weeks at 25 mg 15NH4 15NO3 per kg sand and vermiculite after inoculation (n = 4)
Full size table
Fig. 2
figure 2

Accumulation of N (14 N and 15N) in each Lucerne plant (shoot and root) inoculated with rhizobia and actinobacteria. Asterisks indicate significant differences at P < 0.05 (*) or P < 0.01(**)

Full size image
Table 6 Effect of inoculation treatment on the distribution of accumulated N (14 N and 15N) between plant roots and shoots
Full size table

Discussion

In the absence of rhizobia, the three strains of Streptomyces increased Lucerne shoot weight by 19 % to 33 % compared to untreated plants after 7 weeks growth at the two higher N levels, indicating some of the benefits were independent of the rhizobial symbiosis. Non-symbiotic benefits of actinobacteria have been reported for non-legumes such as cucumber (El-Tarabily et al. 2010) and been attributed to the solubilisation of phosphate, and in wheat to the control of soil borne disease (Franco et al. 2007; Hamdali et al. 2008). LuP30 and LuP47B have been shown to produce siderophores and IAA, and LuP30 has the ability to solubilise phosphate (Le et al. 2015). Nimmnoi et al. (2014) found that the siderophore producing actinobacteria increased the iron content of plants. Martínez-Hidalgo et al. (2014) conducted a similar experiment with Lucerne plants and concluded that Micromonospora spp. enhanced nitrogen uptake efficiency and/or improved nitrogen availability in soil. EN23, LuP30 and LuP47B have characteristics that are consistent with those reported in other actinobacteria that improved the growth of non-legumes, and so some non-symbiotic benefits are likely to contribute to their efficacy.

Plant growth and nodule number/plant were increased by the level of applied N and microbial treatment. Increases associated with N application were attributable to increased plant vigour. Although the larger plants had more nodules overall, they were observed to have delayed nodulation and fewer nodules/mg of root. N levels in the experiment were not high enough to reduce nodulation to the extent reported by others (Heichel and Vance 1979), but were still sufficient to delay nodulation and provide the opportunity for improvement with actinobacterial inoculation.

When LuP30 and LuP47B were co-inoculated with rhizobia strain RRI 128, Lucerne growth and nodulation were increased. Most improvements in nodulation were measured in the first 4 weeks of plant growth, indicating early involvement of the actinobacteria in the regulation of nodulation. The isotope experiment provides evidence that actinobacteria strain LuP47B increased nitrogen (15N) uptake from the soil, but this benefit was relatively small. By comparison, increases in early nodulation (number and density) were large. Martínez-Hidalgo et al. (2014) proposed that Micromonospora play a role as rhizobia helper bacteria (RHB) and since the actinobacteria were coated onto Lucerne seeds prior to inoculation with rhizobia, they could have primed the plant root for rhizobial infection and nodulation (Tokala et al. 2002). Measures of earlier root hair curling and reduced time to nodule appearance (Le, unpublished data) suggest that the effects of the actinobacteria occur very early and therefore may be associated with rhizobial colonisation, plant/rhizobia signalling or Nod factor synthesis. More generalised PGPR effects associated with the production of phyto-hormones such as auxins, cytokinins, and gibberellins could also be important (Glloudemans and Bisseling 1989) because of their effects on root growth and plant development, even if they are not specifically involved in the complex signalling that regulates nodulation. Changed hormone levels resulting from the application of actinobacteria have been postulated to be significant to increased nodulation of both Lucerne (Solans et al. 2015) and chickpea (Gopalakrishnan et al. 2015a; Gopalakrishnan et al. 2015b).The large body of work on hyper- and super-nodulating soybeans that are tolerant of soil nitrate has shown nodulation in the presence of nitrate is strongly controlled by a feed-back signal from the plant shoot (Li et al. 2009; Ferguson et al. 2010; Reid et al. 2011). Changes to the concentration of nutrients in the Lucerne shoots when inoculated with actinobacteria suggests they had some effect on the broader plant physiology and so modification of the feedback loop for control of nodulation needs to be considered. Nimmnoi et al. (2014) proposed that co-inoculation with both actinobacteria and rhizobia increased plant growth by improving the uptake of nutrients and El-Tarabily et al. (2008) also found similar results with actinobacteria in mungbean (Phaseolus vulgaris L.). Although changes to the nutritional status of the Lucerne plants were measured, it is unlikely that nutrient levels per se were critical to the improvements in plant growth, because with the exception of N, they should not have been limiting to growth.

Equalisation of nodule number by the latter harvest at 25 and 50 mg NH4NO3 was probably the result of compensatory nodulation in the control treatment (RRI 128) as soil N declined through the course of the experiment. Whilst it shows that the effects of the actinobacteria may be quite transient, their potential value is not diminished, because early plant vigour is critically important to successful Lucerne establishment and its subsequent persistence and production.

Conclusion

This study confirms that the selected actinobacteria can improve the growth, nodulation and nitrogen fixation of Lucerne plants inoculated with rhizobia strain RRI 128. Both non-symbiotic and symbiotic effects are likely to be contributing to the improvement. The actinobacteria were most effective at 25 mg NH4NO3/kg of soil and in the first four weeks of growth, indicating their use would be best targeted to encourage early plant vigour and aid pasture establishment in soils with low/moderate N levels. The efficacy and durability of the actinobacteria in field soils containing complex micro-flora including different strains of rhizobia is still to be determined, as are the mechanisms of action that promote early nodulation and the enhancement of the plant rhizobial symbiosis. Overall, actinobacteria strain LuP30 showed most promise for use as a co-inoculant with rhizobia based on its benefits to nodulation and plant growth. Strain LuP30 provided the most consistent benefit to nodulation and Lucerne growth across the three N levels in this study.