1 Introduction

Mungbean (Vigna radiata L.) is a common grain legume, cultivated on about seven million hectares worldwide, mainly in Asia (Nair et al. 2019). Around 37–99% of farmers in several Asian countries apply mineral fertilizer as basal dressing or via foliar application during mungbean cultivation (Schreinemachers et al. 2019). This can result in higher production costs and overuse of mineral fertilizer, which poses environmental hazards (Shahzad et al. 2019). The use of plant growth promoting rhizobacteria (PGPR) as biofertilizers is, however, known to enhance the plant’s capacity to take up nutrients from the soil, potentially reducing the need for mineral fertilizer and increasing productivity (Ahmad et al. 2018). The underlying direct mechanisms of PGPR that increase nutrient availability or plant uptake are, for instance, phosphate solubilisation (see review by Richardson et al. 2009), siderophore production (see review by Mimmo et al. 2014), phytohormone production, for instance auxins (see reviews by Vacheron et al. 2013; Tsukanova et al. 2017; Sati et al. 2023), and biological nitrogen fixation (BNF) through symbiotic N2 fixers like rhizobia (Ahmad et al. 2018). Indirect mechanisms of PGPR that increase nutrient uptake refer to enhanced root growth, or increased root surface and root morphology (Calvo et al. 2019). PGPR strains producing 1-aminocyclopropane-1-carboxylate (ACC)-deaminase are of particular importance, since they can lower ACC, an intermediate of the plant hormone ethylene (Mayak et al. 1999; Vandenbussche et al. 2012; Vanderstraeten et al. 2019). Ethylene is involved in regular plant development processes. It regulates plant growth under optimal and stressed cultivation conditions, but it is also known to generally inhibit primary root growth (Qin et al. 2019) and nodulation (Ma et al. 2002). Lower ACC levels and, by extension, ethylene are therefore linked more particularly to root elongation and branching, but also to increased nodulation. This affects nutrient uptake and accumulation (Cao et al. 2008; Glick 2014; Dubois et al. 2018; Lin et al. 2020). Several studies observed that mungbean inoculated with ACC-deaminase producing bacterial strains showed lowered ethylene levels and an increase in number of roots, root length, total root biomass, and nodulation under both non-stress and stress conditions (Mayak et al. 1999; Ahmad et al. 2011, 2012, 2013). However, literature on nutrient uptake with the inoculation of ACC-deaminase producing strains is very scarce, since previous studies focused rather on the effect of inoculation on productivity under environmental stress (e.g. soil salinity) or inoculation as a biological control agent (see reviews by Glick 2014; Ilangumaran and Smith 2017).

In addition to single-strain inoculation, co-inoculation of two strains or multi-strain inoculation of several strains is practiced, as some studies had shown that the adaptability and competitiveness of biofertilizers consisting of several bacterial strains were enhanced (see review by Vassilev et al. 2015). Synergistic effects of different strains as consortia were also shown to increase nodulation or BNF (Tilak et al. 2006; Elkoca et al. 2010). In this regard, the formulation of biofertilizers and the choice of strains for combined inoculation is crucial. A study by Ahmad et al. (2012) indicated, for instance, that co-inoculation of ACC-deaminase producing Pseudomonas fluorescens and Rhizobium phaseoli increased mungbean root biomass and total plant dry matter over single inoculation of both strains separately. Furthermore, multi-strain inoculation of mungbean with Rhizobium phaseoli, Pseudomonas fluorescens, and Bacillus subtilis had already been tested in a field study, resulting in increased root dry matter and nodulation (Zahir et al. 2018). However, the effect on plant-nitrogen (N) content or BNF had not been investigated. Nonetheless, Ahmad et al. (2012) did show increased mungbean seed-N with co-inoculation of Pseudomonas fluorescens and Rhizobium phaseoli compared to a control and the inoculation of Rhizobium phaseoli alone.

We, therefore, hypothesized that ACC-deaminase producing bacteria could enhance the plant’s soil N uptake by increasing root growth which may also promote BNF through a rise in the number of potential infection sites for nodulation. Hence, the objective of this study was to investigate the effect of ACC-deaminase producing PGPR (Bacillus subtilis, Pseudomonas fluorescens, and Rhizobium phaseoli) as a mixture and as single strains on productivity and N accumulation (BNF and soil-N uptake) of an improved mungbean cultivar (NM11) in two field studies. Proline content in leaves and greenness of leaves were additionally assessed, since both can be correlated to N availability (Sánchez et al. 2001, 2002; Xiong et al. 2015; Reinprecht et al. 2020). But they may also be indicators for environmental stress and plant health (Shahenshah and Isoda, 2010; see review by Hayat et al. 2012). Moreover, to determine whether inoculation effects were related to environmental stress, carbon isotopic composition of seeds and leaves was also measured, with δ13C as an indicator of long-term plant response to water stress (Behboudian et al. 2000).

The underlying research questions in this study were:

  • Does mungbean root dry matter increase with inoculation of ACC-deaminase producing PGPR and does that affect N accumulation and productivity?

  • Does inoculation with PGPR affect N acquisition, thereby promoting either BNF or soil-N uptake?

  • Does the multi-strain inoculum show increased plant-N accumulation and productivity over single-strain inoculation?

2 Materials and Methods

2.1 Experimental Site and Meteorological Conditions

Two studies were conducted in fields at the research station of the University of Agriculture, Faisalabad, Pakistan (31.4°N, 73.1°E). The first study comprised a trial with multi-strain inoculation, which was conducted in two seasons of 2017, from March 13th to July 1st and from July 11th to September 14th. The second study consisted of a trial with the same multi-strain inoculation and an additional single-strain inoculation, established in the second season only. The first study was conducted on site I in both seasons, the second study was conducted on site II, at a distance of one hundred meters.

A meteorological station at site I measured climatic conditions during the experimental period (Fig. S1). In the first season, average air temperature and accumulated total precipitation were recorded at 30.7 °C and 80 mm, respectively. Average maximum temperatures were highest in May and June at around 33.5 °C. In the second season average air temperature and accumulated total precipitation were 32.8 °C and 187 mm. The highest average temperatures and rainfall amount were recorded in July – 33.7 °C and 117 mm, respectively.

The cropping history before trial establishment consisted of wheat and fallow for site I and maize for site II. Before the establishment of the field trials, soil chemical characteristics were determined (Table 1), following the protocol by Ryan et al. (2001). Both sites were slightly alkaline, and low in organic matter and total N.

Table 1 Soil properties of the two trial sites at the University of Faisalabad, Pakistan in 2017
Full size table

2.2 Seed Material and Bacterial Inoculation

Mungbean variety NM11 was chosen for both studies. NM11 is a popular mungbean cultivar, covering around 29% of the area cultivated with mungbean in Pakistan (Schreinemachers et al. 2019). It contains genetic material developed at the World Vegetable Centre and is resistant to Mungbean Yellow Mosaic Disease. It was released by the Nuclear Institute for Agriculture and Biology in Pakistan in 2011.

Maize was used as the reference plant in the 15N natural abundance method to assess BNF of mungbean, since non-nodulating legumes were not available.

Three pre-isolated and characterized ACC-deaminase producing PGPR strains (M9 = Rhizobium phaseoli; Y16 = Bacillus subtilis; Mk20 = Pseudomonas fluorescens) were obtained from the Soil Microbiology and Biochemistry Laboratory, Institute of Soil and Environmental Sciences, University of Agriculture, Faisalabad, Pakistan. P. fluorescens strain Mk20 has been shown to exhibit an ACC deaminase activity of around 478 nmol α-ketobutyrate mg− 1 biomass h− 1, while B. subtilis strain Y16 has an ACC deaminase activity of approximately 269 nmol α-ketobutyrate mg− 1 biomass h− 1. Further details about the two PGPR strains can be found in Ahmad et al. (2011) and Khan et al. (2017). Rhizobium phaseoli in general, and the strain M9 in particular, have been shown to effectively nodulate (Ahmad et al. 2011; Mahmood et al. 2021) and fix N2 (Ramirez et al. 2020) in mungbean and had previously been isolated from mungbean nodules (Ahmad et al. 2011). In a preceding experiment, the mungbean cultivar NM11 showed maximum number of nodules with R. phaseoli inoculation strain M9 (Zahir et al. 2018). Moreover, the M9 strain exhibits also ACC deaminase activity, which was previously evaluated qualitatively (Ahmad et al. 2011).

A yeast extract mannitol broth as the media was prepared for R. phaseoli following a procedure by Fred et al. (1932) and a lysogeny broth was prepared for P. fluorescens and B. subtilis with 5 g L− 1 NaCl, based on the formulation by Bertani (1951). The inoculated broth was incubated for 72 h at 28 °C in a shaking incubator (SI9R-2, Sheldon). After incubation, uniform population density (OD540 = 0.45; 107-108 cfu mL− 1) was achieved by dilution. To prepare the multi-strain inoculation, equal proportions of each strain were mixed and vortexed to ensure homogenized cell density. Seeds of mungbean and maize were coated with a slurry, consisting of the broth (with either a mixture or single strains), a previously sterilized peat-clay mixture, and a sugar solution in a ratio of 4:5:1. Seeds of the control treatment were coated with autoclaved carrier and broth.

2.3 Experimental Design

The first study on site I was set up with a two-factorial randomized complete block design with three replicates. The 6 m2 plots consisted of NM11 or maize, inoculated with a mixture of the rhizobacterial strains, i.e. R. phaseoli, P. fluorescens, B. subtilis, or with sterilized broth as the control. The second study on site II was laid out in a split-plot design to focus on the effect of the single bacterial strains on mungbean productivity and N accumulation. The trial was established in three blocks, each with two main plots, consisting of either mungbean or maize. The 9 m2 sub-plots were assigned to different inoculations (single inoculation of R. phaseoli, P. fluorescens, or B. subtilis, a mixture of all strains, or a control with sterilized broth). On both sites, plants were sown on ridges with around 15 cm between plants in a row and around 0.75 m between each row.

Before sowing, both sites received a hand-broadcasted fertilization of 20 kg N ha− 1 (urea) as a starter dose to promote initial plant growth and initiate rhizobium symbiosis (Giller and Cadisch 1995), 60 kg P ha− 1 (diammonium phosphate), and 60 kg K ha− 1 (potassium sulfate), following the recommendations in Zahir et al. (2018). Weeds were controlled by manual weeding and herbicides. Insecticides and irrigation with furrow irrigation were applied when needed.

2.4 Sampling and Analytical Methods

In the first season, plants were harvested at physiological maturity, 108 days after sowing (DAS). In order to evaluate the temporal impact of the PGPR during plant development, plants grown in the second season were harvested at flowering, mid-podfilling, and physiological maturity (site I: 39 DAS, 58 DAS, 67 DAS; site II: 37 DAS, 55 DAS, 64 DAS). At harvest, composite samples consisting of three plants per plot or sub-plot were collected, including the surrounding soil in a block with a volume of 15 × 15 × 30 cm to recover roots. Mungbean plants were separated into stems, leaves, and pods, and the corresponding soil block was wet-sieved with a 2 mm sieve to collect roots. At flowering in the second season, the nodules were also evaluated (weighed and counted). Reference plants (maize) were harvested without roots and not separated into individual plant parts. Plant samples were oven-dried at 70 °C until they achieved a constant weight and were then weighed. Pods were threshed and seed yield and 100-seed weight were determined. Plant samples were ground with a ball-mill and analyzed for δ15N, atom% 15N, and δ13C, using an Euro Elemental analyzer coupled to a Finnigan DELTAplus XP continuous-flow isotope ratio mass spectrometer (Thermo Fisher Scientific) at the University of Hohenheim.

In the first season, N deriving from the atmosphere (Ndfa) in % was calculated based on δ15N of the seeds. This may be a suitable method to estimate integrated levels of BNF (Barbosa et al. 2018). In the second season, %Ndfa was calculated based on δ15N of the shoots at mid-podfilling.

In order to calculate BNF, the 15N natural abundance method illustrated in Unkovich et al. (2008) was applied. δ15N of whole shoots and N content of whole shoots and single plant parts were calculated. %Ndfa was calculated with an equation provided by Unkovich et al. (2008):

$$\eqalign{& \% Ndfa \cr & = {{{\delta ^{15}}N\,of\,reference\,plant\, - \,{\delta ^{15}}N\,of\,{N_2}\,fixing\,legume} \over {{\delta ^{15}}N\,of\,reference\,plant\, - \,B}} \cr & \times 100 \cr} $$

where B is the δ15N value of a mungbean plant cultivated in N-free media to adjust for isotopic fractionation during BNF (B value). We used a B value of -2.50‰, which was suggested for mungbean in Pakistan (Peoples et al. 1997).

Least square means of δ15N values of the reference plants were computed for each inoculation treatment (multi-strain, single-strain, control) and across blocks, to obtain one δ15N value for each study and inoculation treatment. The δ15N values of the reference plants did not show any inoculation or block effect in either study.

To assess the amount of N deriving from the soil (Ndfs) in kg ha− 1, the equation based on Barbosa et al. (2018) was used:

$$\eqalign{& Ndfs\,\left( {kg\,h{a^{ - 1}}} \right)\, = \,Total\,N\,in\,shoot\,\left( {or\,seed} \right) \cr & - \,total\,fixed\,N\,in\,shoot\,\left( {or\,seed} \right)\left( {kg\,h{a^{ - 1}}} \right) \cr} $$

In the second season, proline accumulation and greenness of leaves were also assessed between flowering and podfilling (46 and 47 DAS, respectively) on plants of both sites. Greenness of leaves was assessed using a SPAD-502 meter (Konica Minolta, Japan). This is a hand-held device that is widely used for the rapid, non-destructive, and indirect measurement of leaf chlorophyll concentrations (Ling et al. 2011). The SPAD readings were measured in the youngest, fully expanded trifoliate of two randomly selected plants per subplot.

Proline content was measured in the youngest, fully expanded trifoliate of one randomly selected plant per plot. Around 0.5 g of fresh leaf biomass was used to determine proline accumulation with the procedure illustrated in Bates et al. (1973).

2.5 Statistical Analysis

All data was analyzed with SAS (v 9.4), using a general linear model (SAS, Cary, NC, USA). Least significant differences were calculated at p = 0.05. Data of the first study was analyzed using the following model:

$${y_{ijk}} = {\rm{ }}\mu {\rm{ }} + {\rm{ }}{s_j} + {\rm{ }}{b_k} + {\alpha _i} + {\rm{ }}{(\alpha.s)_{ij}} + {e_{ijk}}$$

where µ = intercept, sj = j-th season effect (j = 1, 2), bk = k-th block effect, αi = inoculation (i = 1, 2), (α.s)ij = i-th inoculation by j-th season interaction, and eijk = heterogeneous residual error variance. The data of the single-strain trial was analyzed using the following model:

$${y_{ijk}} = {\rm{ }}\mu {\rm{ }} + {\rm{ }}{b_k} + {\rm{ }}{{\\\beta}_j} + {\rm{ }}{F_{ik}} + {\rm{ }}{e_{ijk}}$$

Where µ = intercept, bk = k-th block effect, ßj = j-th inoculation effect, Fik = random error associated with main plot, and eijk = random error associated with sub plot. To test for significant differences, the means were compared using the PDIFF option of LSMEANS. Additionally, Pearson correlations (PROC CORR) were computed to identify relationships between dry matter and N accumulation.

3 Results

3.1 Study 1: Multi-strain Inoculation over Two Seasons

3.1.1 Morpho-physiological and Yield Traits

The assessment of morpho-physiological and yield traits showed significant treatment effects only when comparing SPAD readings (Table 2), where plants inoculated with multi-strain inoculum showed an increase of 21% over the control in 2017. Additionally, season or the interaction season*inoculation resulted in significant effects on 100-seed weight or the accumulation of vegetative aboveground dry matter. The 100-seed weight in the first season was, in general, 16% higher than in the second season. Plants inoculated with the multi-strain inoculum showed no difference in 100-seed weight across seasons, whereas the control plants differed significantly with higher seed weight in the first season and lower seed weight in the second season (first season: 5.8 g control > 5.3 g multi-strain > second season: 5.1 g multi-strain > 4.5 g control). Moreover, plants in the second season generally accumulated around 100 kg ha− 1 more root dry matter, 0.9 t ha− 1 more vegetative aboveground dry matter, and 0.9 t ha− 1 more total dry matter than plants in the first season. However, root dry matter, root: shoot ratio, seed yield, total dry matter, proline content or δ13C in seeds were not affected by inoculation.

Table 2 Morpho-physiological and yield traits of mungbean inoculated with a multi-strain inoculum consisting of Rhizobium phaseoli, Bacillus subtilis, and Pseudomonas fluorescens (= multi) or autoclaved broth (= control) at physiological maturity, grown in two seasons in 2017 (S1 and S2) at the University of Agriculture, Faisalabad. Proline content and SPAD readings were assessed in the second season only at mid-podfilling. AG = aboveground
Full size table

The assessment of dry matter accumulation in different plant parts, as well as dry matter partitioning in % to different plant parts at physiological maturity of both seasons and on two additional dates (flowering and mid-podfilling) in the second season did not reveal any difference when plants were inoculated with multi-strain inoculum (Table S1, Fig. S2).

3.1.2 N2 Fixation and N Accumulation

Similar to productivity and dry matter accumulation, BNF (%Ndfa, N fixed in kg ha− 1), Ndfs (kg ha− 1) and total N content in the plant material were not affected either by multi-strain inoculation in both seasons in 2017 (Table 3). Rather seasonal effects were visible with regard to %Ndfa, Ndfs (kg ha− 1), and N accumulation in vegetative aboveground dry matter. However, no significant interaction between season and treatment occurred. The plants in the second season accumulated, on average, 138% more N in aboveground vegetative plant parts (13 vs. 31 kg ha− 1) and increased the amount of soil derived N by 106% (17 vs. 35 kg ha− 1) in comparison to plants in the first season. In contrast, plants in the first season took 32% more N from the atmosphere (38 vs. 70%Ndfa) in comparison to plants in the second season and accumulated correspondingly also 73% more fixed N (22 vs. 38 kg ha− 1). > The number and weight of nodules did not differ between treatments.

Table 3 BNF (biological nitrogen fixation) characteristics, N accumulation, and N uptake of mungbean inoculated with a multi-strain inoculum consisting of Rhizobium phaseoli, Bacillus subtilis, and Pseudomonas fluorescens (= multi) or autoclaved broth (= control) grown in two seasons in 2017 at the University of Agriculture, Faisalabad. N accumulation (kg ha− 1 or mg g− 1) was assessed at physiological maturity, BNF characteristics (%Ndfa, N fixed kg ha− 1, Ndfs kg ha− 1) were assessed either at podfilling (second season) or maturity (first season). The nodules (weight and number) were only assessed in the second season at flowering. AG = aboveground. Nc = nitrogen concentration
Full size table

3.1.3 Relationship of Dry Matter and Nitrogen Accumulation

Across both seasons in 2017, total N content of the plants correlated positively with root dry matter and Ndfs (kg ha− 1), but not with %Ndfa or fixed N2 (kg ha− 1) (Table 4). Ndfs (kg ha− 1) was, moreover, also positively related to total plant dry matter, but not to root dry matter, indicating no major effect of root dry matter on soil N uptake.

Although Ndfs (kg ha− 1) was negatively correlated with %Ndfa, there was no relationship between Ndfs (kg ha− 1) and fixed N (kg ha− 1). Both BNF characteristics, %Ndfa and fixed N (kg ha− 1), did, however, correlate positively with proline content in the leaves. Moreover, %Ndfa and fixed N (kg ha− 1) also correlated positively with δ13C in the seeds, suggesting the less negative the δ13C values were, the higher the share of %Ndfa.

SPAD readings, although significantly affected by multi-strain inoculation, showed no correlation with any of the traits. Seed yield only correlated with total plant-N content and total dry matter: yield rose with increasing total N content and increasing total dry matter.

Again, seasonal differences rather than differences between treatments were visible when depicting the relationship between total plant-N, root dry matter, Ndfs (kg ha− 1), and yield in graphs (Fig. S3).

Table 4 Pearson correlation coefficients between dry matter and nitrogen accumulation, BNF traits, proline and SPAD of mungbean with multi-strain inoculation consisting of Rhizobium phaseoli, Bacillus subtilis, and Pseudomonas fluorescens and a control in two seasons of 2017 (March – July, 2017 and July-September 2017) at maturity. Only proline and SPAD were measured at the mid-podfilling stage and only in the second season. TDM = total dry matter (t ha− 1), yield = seed yield (t ha− 1), root = root dry matter (kg ha− 1), TN = total nitrogen content per plant (kg ha− 1), ndfa = nitrogen derived from the atmosphere (%), nfix = symbiotically fixed nitrogen (kg ha− 1), Ndfs = nitrogen derived from the soil (kg ha− 1), proline content in fresh leaves (µmol g− 1), SPAD readings, and δ13C of seeds (‰)
Full size table

3.2 Study 2: Single-strain Inoculation in the Second Season

3.2.1 Morpho-physiological and Yield Traits

Assessing the effect of single bacterial strains on productivity traits showed significant effects on root dry matter, root: shoot ratio, seed yield, total dry matter, vegetative aboveground dry matter, and SPAD readings. In general, B. subtilis inoculation had the largest effects on these traits in comparison to the control treatment (Table 5).

Root dry matter increased significantly by 41% and 102% with R. phaseoli and B. subtilis inoculation, respectively. Similar to the first study, multi-strain inoculation did not affect root dry matter build-up significantly. Furthermore, B. subtilis inoculation increased the root: shoot ratio over control and multi-strain inoculation. Seed yield differed significantly only between B. subtilis and P. fluorescens inoculation, showing around 45% lower yields with the latter. Total dry matter and vegetative aboveground dry matter were both higher with B. subtilis inoculation than with the control treatment, increasing by around 1.7 t ha− 1 and 1 t ha− 1, respectively. However, plants inoculated with multi-strain inoculum also showed higher vegetative aboveground dry matter (+ 0.7 t ha− 1). Plants inoculated with B. subtilis outperformed plants inoculated with P. fluorescens with regard to total dry matter and vegetative aboveground dry matter, both showing increases of 1.9 t ha− 1 and 0.9 t ha− 1. SPAD readings were significantly elevated for plants inoculated with R. phaseoli at 0.1 significance level. Leaves at podfilling showed significant differences in δ13C between plants inoculated with P. fluorescens (less negative values) and plants inoculated with multi-strain or R. phaseoli inoculation.

Table 5 Productivity traits (root dry matter kg ha− 1 [RDM], root: shoot ratio [R: S], seed yield t ha− 1 [yield], 100-seed weight [HSW] g, total dry matter t ha− 1 [TDM], vegetative aboveground dry matter t ha− 1 [VDM], proline µmol g− 1, SPAD readings, and δ13C in leaves) of mungbean with inoculation of either Rhizobium phaseoli, Pseudomonas fluorescens, Bacillus subtilis, multi-strain inoculation with all aforelisted strains (= multi), or autoclaved broth (= control). Data was collected in one season (July–September, 2017) on one site, at the University of Agriculture, Faisalabad, Pakistan at physiological maturity (except proline, SPAD and δ13C in leaves, which were measured at the mid-podfilling stage)
Full size table

In additional to the observed effects on dry matter accumulation at physiological maturity (64 days after sowing), an inoculation effect on aboveground dry matter at flowering (37 days after sowing) was also observed (Fig. 1, see also Table S2). Plants inoculated with the multi-strain inoculum showed significantly more aboveground dry matter than plants inoculated with P. fluorescens. However, none of the strains significantly increased dry matter over the control treatment at flowering or mid-podfilling (55 DAS). Dry matter partitioning in % of total dry matter showed significant differences at maturity only (Table S2). B. subtilis inoculation resulted in significantly larger dry matter partitioning to roots than control and multi-strain inoculation.

Fig. 1
figure 1

Dry matter accumulation in g plant− 1 of mungbean with inoculation of either Rhizobium phaseoli (M9), Pseudomonas fluorescens (Mk20), Bacillus subtilis (Y16), multi-strain inoculation with all aforelisted strains (= multi), or autoclaved broth (= control, C) at flowering (37 days after sowing [DAS], mid-podfilling (55 DAS) and physiological maturity (64 DAS) of one season in 2017 at the University of Agriculture, Faisalabad. Different letters indicate significant differences between treatments on each harvest date at 0.05 significance level. ns means not significantly different

Full size image

3.2.2 N2 Fixation and N Accumulation

The evaluation of BNF and N accumulation identified a significant treatment effect (Table 6). Only the single-strain inoculation with R. phaseoli resulted in a significant greater share of %Ndfa in comparison to the control (+ 24%). This was an indication of the efficacy of the chosen strain in combination with this mungbean cultivar. None of the other treatments differed from the control. Although R. phaseoli inoculation led to an increased %Ndfa, the amount of fixed N (kg ha− 1) was not affected and did not significantly differ between the treatments. The plants sampled at flowering were examined for the presence of nodules; however, with the exception of one plant, no nodules were observed. Therefore, the data did not allow a statistical analysis of nodulation. The amount of Ndfs (kg ha− 1) did, however, show differences between treatments: R. phaseoli inoculation resulted in a reduction of soil N uptake of around 26 kg ha− 1 in comparison to the inoculation with B. subtilis.

The higher share of %Ndfa with R. phaseoli inoculation did not, however, affect total plant-N. Only inoculation with B. subtilis significantly increased total plant-N content in comparison to the control (+ 36 kg ha− 1). Moreover, plants inoculated with B. subtilis showed a higher total N content than plants of all other treatments, except the multi-strain inoculation. It showed a similar absolute amount of N. N accumulation in the different plant parts also differed between strains: root-N was increased with multi-strain, R. phaseoli, and B. subtilis inoculation, whereas N in vegetative aboveground plant parts increased only with multi-strain and B. subtilis inoculation. Seed-N differed only between the two strains, B. subtilis and P. fluorescens, with the latter accumulating around 20 kg N ha− 1 less than B. subtilis. The assessment of N concentration in mg g− 1 dry matter in the different plant parts showed that only the N concentration in roots differed between inoculation treatments, with the highest concentrations in multi-strain inoculated plants, followed by plants inoculated with R. phaseoli and P. fluorescens. Although B. subtilis increased the total amount of root dry matter and root-N, the N concentration in roots was not higher than in control plants.

Table 6 Ndfa (nitrogen deriving from the atmosphere) in %, biologically fixed nitrogen (N fixed) in kg ha− 1, Ndfs (N deriving from the soil) in kg ha− 1, total plant-N in kg ha− 1, root-N in kg ha− 1, N in vegetative aboveground dry matter (veg-N) in kg ha− 1, seed-N in kg ha− 1, N concentrations in mg g− 1 dry matter in roots (Root-Nc), in vegetative aboveground dry matter (Veg-Nc) and seeds (Seed-Nc) of mungbean with inoculation of either R. Phaseoli, P. fluorescens, B. subtilis, multi-strain inoculation with all strains (= multi) or an autoclaved broth as control. Data was collected in one season (July–September 2017) on one site, at the University of Agriculture, Faisalabad, Pakistan at mid-podfilling (Ndfa, N fixed, Ndfs) or physiological maturity (total-N and root-N, veg-N, seed-N in kg ha− 1 and mg g− 1)
Full size table

3.2.3 Relationship of Dry Matter and Nitrogen Accumulation

Total plant-N content correlated positively with Ndfs (kg ha− 1), but not with %Ndfa or fixed N (kg ha− 1), indicating that the driver of N accumulation was soil-N uptake rather than BNF (Table 7). Additionally, total plant-N content also correlated positively with root dry matter, seed yield, total dry matter, and proline content.

Neither %Ndfa nor Ndfs (kg ha− 1) showed any correlation with other morphological traits. As expected, the higher the Ndfs (kg ha− 1) content in plants, the lower the %Ndfa. Root dry matter did not correlate with %Ndfa, N fixed (kg ha− 1) or Ndfs (kg ha− 1), although it did correlate with total plant-N.

SPAD readings only correlated with %Ndfa, fixed N (kg ha− 1), and proline content, but not with total plant-N or Ndfs (kg ha− 1).

The assessment of relationships between total plant-N content and root dry matter, Ndfs (kg ha− 1), and seed yield showed graphically that B. subtilis inoculation generally outperformed the other inoculation treatments (Fig. S4).

Table 7 Pearson correlation coefficients between dry matter and nitrogen accumulation, BNF traits, proline and SPAD of mungbean with inoculation of either Rhizobium phaseoli, Bacillus subtilis, Pseudomonas fluorescens, a combination of all strains or an autoclaved broth as the control in one season (July–September 2017) at the University of Agriculture, Faisalabad, Pakistan at physiological maturity. Only BNF traits (%Ndfa, N fixed kg ha− 1, Ndfs kg ha− 1), proline (µmol g− 1), SPAD readings and δ13C (‰) in leaves were assessed at mid-podfilling. TDM = total dry matter (t ha− 1), yield = seed yield (t ha− 1), root = root dry matter (kg ha− 1), TN = total plant nitrogen content (kg ha− 1), ndfa = nitrogen derived from the atmosphere (%), nfix = biologically fixed nitrogen (kg ha− 1), Ndfs = nitrogen derived from the soil (kg ha− 1), proline content in fresh leaves (µmol g− 1), SPAD readings and δ13C in leaves (‰)
Full size table

4 Comparison of both Studies

In both studies, multi-strain inoculation did not affect biomass production, total plant-N content or N acquisition (BNF or soil-N uptake) in comparison to the control (Fig. 2). Only single-strain inoculation had an effect on root dry matter (R. phaseoli and B. subtilis), %Ndfa (R. phaseoli), total dry matter, and total plant-N (B. subtilis). Similar to multi-strain inoculation, P. fluorescens had no significant effect on dry matter or N accumulation.

Fig. 2
figure 2

Changes in root DM (dry matter) (kg ha− 1), seed yield (t ha− 1), total DM (t ha− 1), nitrogen (N) deriving from the soil (Ndfs) (kg ha− 1), N deriving from the atmosphere (%Ndfa), and total plant-N (kg ha− 1) of mungbean with (A) multi-strain inoculation consisting of R. phaseoli, P. fluorescens, and B. subtilis over two seasons (March–July 2017 and July–September 2017) (B) multi-strain and single-strain inoculation in one season (July–September 2017) at the University of Agriculture, Faisalabad, Pakistan. Data was compared to an uninoculated control. * indicates significant difference to the control at 0.05 significance level

Full size image

5 Discussion

5.1 Only Single Strain Inoculation of B. Subtilis and R. Phaseoli Affected Root Dry Matter and Nitrogen Acquisition

We hypothesized that inoculation with ACC-deaminase producing PGPR enhanced N accumulation in mungbean through increased root growth. Our study confirmed increased root growth with inoculation of B. subtilis or R. phaseoli, suggesting ACC-deaminase activity of the two strains. A statistically significant increase in both root dry matter and plant-N was, however, only observed with single-strain inoculation of B. subtilis. Inoculation with B. subtilis may, therefore, have stimulated plant growth through ACC-deaminase activity, leading to an increased N uptake rate driven by the increased N demand and more effective N capture, as shown by the increased total plant-N. This effect was also observed in studies by Calvo et al. (2017, 2019), where the inoculation of a mixture of different Bacillus strains increased root dry weight, N uptake or N content of maize or Arabidopsis thaliana. Although root dry matter and total plant-N showed a positive correlation in our studies, the impact of extended root growth on the N acquisition mechanism remained unclear, since Ndfs (kg ha− 1) and root dry matter were not strongly positively correlated in both studies and B. subtilis inoculation did not show any significant increase in Ndfs (kg ha− 1) in comparison to the control. This indicated that not only an increase in root dry matter based on ACC-deaminase activity, but additionally other mechanisms of B. subtilis leading to increased N availability, could have supported plant-N accumulation. Several studies showed that Bacillus inoculation could increase N availability or N uptake by increasing nitrification or upregulating nitrate and ammonium uptake genes in plants (Ahmad et al. 2017; Calvo et al. 2019; Sun et al. 2020a). This was also in line with a study by Khan and Zaidi (2006), who observed increased N content in mungbean roots and shoots with single strain inoculation of B. subtilis although root dry matter was not enhanced.

Inoculation with R. phaseoli also increased root dry matter. In addition, it enhanced %Ndfa, but showed no effect on total plant-N. The increase in root biomass with R. phaseoli inoculation was in line with previous studies (Ahmad et al. 2011, 2012). However, root dry matter did not correlate with %Ndfa or fixed N (kg ha-1) in our study. Consequently, any increase in root biomass did not result in an increase in fixed N (kg ha-1). This suggested that nodulation may not have been affected either by increased root biomass, thus negating the second hypothesis. Surprisingly, %Ndfa of plants inoculated with R. phaseoli did not differ from %Ndfa or N fixed (kg ha-1) of the other inoculated plants (multi-strain and single–strain). Other studies ascertained that inoculation with PGPR other than rhizobia could enhance the symbiosis with native rhizobia in the soil (Elkoca et al. 2010). We concluded that ACC-deaminase producing rhizobia did not present any advantages over other rhizobia strains in this study, as root dry matter (although increased) and %Ndfa did not correlate.

Inoculation with P. fluorescens had no effect on productivity or N accumulation. Furthermore, it showed opposite effects on productivity traits when compared to plants inoculated with B. subtilis. This difference could be linked, on the one hand, to stronger growth promotion by B. subtilis, and, on the other hand, to weaker interaction between the plants or this specific genotype and P. fluorescens. Effects of this P. fluorescens strain (Mk20) were discussed contrastingly in the literature. Several studies showed that this strain interacted effectively with mungbean, thereby increasing for instance root growth (Ahmad et al. 2011, 2012). However, another study observed instead that mungbean seedlings inoculated with P. fluorescens under both non-stress and stress conditions did not present any significant difference in total dry matter in comparison to a control (Ahmad et al. 2013). Although the cited study was conducted under controlled conditions and it was, therefore, difficult to transfer the results to field studies, this could indicate that the effect and the interaction of both the mungbean and the Mk20 strain were rather unstable. Furthermore, the assessment of δ13C values and proline content during both seasons revealed a significant difference only in δ13C between P. fluorescens inoculation and multi-strain inoculation, as well as between P. fluorescens and R. phaseoli strain inoculation in the second season. This suggests a lower water use efficiency (Condon et al. 2004) with P. fluorescens and, therefore, less beneficial growth effects with this single strain. Yet, as no water stress was imposed during the experiment we did not expect to see strong differences between the inoculations.

5.2 Multi-strain Inoculation was Less Efficient than Single-strain Inoculation due to Potential Interspecific Competition

Assessing the effect of inoculation on productivity traits and N accumulation showed that the multi-strain inoculation did not outperform the impact of single-strain inoculation. Similar results were shown by another study which came to the conclusion that the efficacy of the multi-strain inoculation was in between the most and the least effective single-strain inoculation (Somasegaran and Bohlool 1990). Multi-strain inoculation, for instance with Bacillus subtilis, Bacillus megaterium, and Rhizobium leguminosarum bv. Phaseoli in a trial by Elkoca et al. (2010) did not show an effect on common bean yield in comparison to single-strain inoculation either. Moreover, nodulation was not enhanced when R. leguminosarum was applied in a mixture with the two other strains. Similarly, a study by Bullied et al. (2002) showed that the stimulating effect of Bacillus cereus on soybean growth and N accumulation was greater without the co-inoculation of Bradyrhizobium japonicum. Also Zahir et al. (2018) could show that multi-strain inoculation with R. phaseoli (M9), B. subtilis (Y16), and P. fluorescens (Mk20) did not affect shoot dry weight, yield or 1000-grain weight of mungbean genotype NM11 in a field trial. The co-inoculation of only two of these strains (Mk20 and M9) did, however, show an increase of total dry matter and root biomass in a pot trial with untreated field soil under non-stressed conditions (Ahmad et al. 2012). This could indicate that B. subtilis as part of the mixture affected the efficacy of the other strains or vice versa. In a study by Chakraborty et al. (2016) it was observed that Pseudomonas sp. produced inhibitors that affected the growth of Bacillus cereus. This strategy helped Pseudomonas to survive and compete against the faster growing Bacillus cereus, present in the same ecological niche. Although this strategy led in general to a co-existence of both strains, B. cereus strains experienced stress caused by the increased release of this inhibitor, which partly led to the differentiation into spores. Since only B. subtilis affected plant-N content in our study, any inhibition of this strain in a mixture with other strains might have affected plant-N content in the end. On the other hand, inoculation with B. subtilis favored soil N uptake over fixed N (kg ha− 1), which potentially equalized the positive effect of other inoculants on N2 fixation, as shown for example in a study by Diatta et al. (2020): With increasing soil N uptake, less N was fixed through BNF in mungbean.

Although the majority of the assessed plant traits were not affected by multi-strain inoculation, SPAD readings (study 1), N content in roots, aboveground vegetative dry matter (study 2), and root-N concentration (study 2) increased with multi-strain inoculation. SPAD meter readings were supposed to be related to leaf-N and were already shown to be increased with PGPR inoculation (Ahamd et al. 2014; Calvo et al. 2017, 2019). Despite the fact that root-N and veg-N on an area base (kg ha− 1) were significantly increased with multi-strain inoculation, this had interestingly no effect on total N accumulation. The control was able to accumulate the same amount of seed-N as the multi-strain inoculation, although it presented around 28% less N in the aboveground vegetative dry matter. This was surprising since most of the seed-N is usually remobilized from sources in leaves, instead of being transported from the roots to the seeds (Masclaux-Daubresse and Chardon 2011; Havé et al. 2017). Lower accumulation of N in vegetative aboveground dry matter would, therefore, mean less N being able to be remobilized to seeds. However, the assessment of N concentrations in the different plant parts also showed that N concentrations in vegetative aboveground dry matter did not differ between the treatments. Significant differences in N accumulation in kg ha− 1 were, therefore, related rather to increased dry matter per ha.

All in all, different mechanisms may have affected the efficacy of the multi-strain inoculum, including inhibitory mechanisms of certain strains (e.g. Pseudomonas). Yet, the success of a biofertilizer will also be dependent on the interaction of bacteria with the plant or rather the specific genotype and the native bacterial community in the rhizosphere. In a field trial by Zahir et al. (2018) with ten different mungbean genotypes, it was shown that multi-strain inoculation of R. phaseoli, B. subtilis, and P. fluorescens increased the yield only of three genotypes. Specific genotype x bacteria interaction was also observed in a study by Mirza et al. (2007), where two genotypes of chickpea showed different responses to the co-inoculation of combinations of Enterobacter and Rhizobium strains. This suggested that the effects observed in our studies may not necessarily be the same with a different mungbean genotype.

6 Conclusion

Currently, only up to 14% of the mungbean farmers in Asia inoculate seeds with biofertilizers (Schreinemachers et al. 2019). Reformulating mixtures and recommending certain strains for different purposes could increase the adoption potential of biofertilizers. In our study we found that the application of multi-strain inoculation offered little advantage. Even though some traits were affected, these effects were not consistent throughout both studies. However, multi-strain inoculation in the second study resulted in the largest N concentration in root dry matter. An increase in the N concentration could constitute a valuable contribution to the soil N pool when plant residues, i.e. roots, remain in the field after harvest and decay (Carranca et al. 2015). This is of particular importance, since mungbean is a short-season grain legume. It is a good fit for many cereal-based cropping systems in the global South and could, thus, improve the yield of the following crops through greater N input.

Single strain inoculation of B. subtilis did, however, increase dry matter and N accumulation. This was probably due to not only the significant increase in root dry matter and soil N uptake, but also to other mechanisms, whereas inoculation with R. phaseoli increased %Ndfa over the control. The choice of one of these strains should be made on the basis of current soil N conditions. With a significantly larger amount of soil N uptake in comparison to plants inoculated with R. phaseoli, the lowest root-N concentration in comparison to all other inoculation treatments, and at the same time no effect on yield in comparison to the control, the inoculation of B. subtilis in a crop rotation might result in less N for the following crops. This would reduce the N sparing effect of legumes. However, under fertilized or medium to high soil N conditions, inoculation with B. subtilis could help to reduce N losses through improved soil N recycling efficiency, as already indicated in other studies (Sun et al. 2020a, b).

Since not all of the ACC-deaminase producing strains affected plant growth in similar ways, future studies should focus on the interaction of different mungbean genotypes and bacterial strains in single-, co- or multi-strain inoculum, and assess additionally in competition assays the efficacy of strains in a mixture in avoiding antagonism.