1 Introduction

In the Mediterranean basin, leguminous plants have a primary position in the agricultural sector owing to agronomic, economic, and nutritional interests. Similarly, the cultivation of legumes tends to conserve, or even increase, the availability of nitrogen in soils. The Rhizobium (Rhizobiaceae family) genus bacteria play an important role in agriculture because of their ability to fix atmospheric nitrogen in symbiosis with leguminous plants and hence in soil fertility (Du et al., 2019).

Nitrogen, available in abundance as atmospheric gas (78% N2), plays a vital role in plant metabolism and is a major protein component and a critical component of living matter (Bingham & Cotrufo, 2016; Xu et al., 2018). Several biological alternatives, such as biofertilizers, have been developed to prevent the irrational use of chemical nitrogen fertilization, which is responsible for adverse effects on human and environmental health (Bhattacharjee et al., 2008). Nitrogen-fixing bacteria are microorganisms that are used as biofertilizers. They have the capability of reducing atmospheric nitrogen (N2) to a form that can be assimilated by plants. (Cardoso et al., 2018; Du et al., 2019). Nitrogen-fixing bacteria' capacity to create symbiotic relationships with leguminous plants and fix atmospheric nitrogen has been utilized in the field to meet nitrogen requirements. This phenomenon gives an alternative to the usage of nitrogenous fertilisers, whose excessive use is detrimental to the environment (Bhattacharjee et al., 2008).

Rhizobia, in particular, consumes energy from the sugars provided by the host plant's photosynthesis. Per year, symbiotic nitrogen fixation between Rhizobiaceae and legumes will lead to at least 70 million tons of nitrogen (Brockwell et al., 1995). A large part of this fixed N comes from soya beans (Glycine max (L.) Merr.), which are the most widely grown legumes in the world (Herridge et al., 2008). According to Lodwig et al. (2003), biological nitrogen fixation (BNF) to ammonium accounted for roughly 65 percent of available nitrogen in the biosphere. The legume-rhizobia symbiosis produces a large portion of this nitrogen, which can result in an average annual nitrogen fixation of up to 300 kg N/ha depending on the species (Board, 2013; Ma et al., 2019).

By increasing the use of legumes in farms, rhizobial symbiosis will also alleviate farmers from the expensive and high polluting risk of chemical fertilization. There are many causes that affect this symbiotic relationship such as pesticides (Pati et al., 1984; Shen et al., 2021). From the many chemicals widely used in agriculture, herbicides are of critical importance owing to their growing usage in weed control in all plants (Sousa et al., 2020). As a consequence, although progress toward sustainable agriculture has become more important, caution must be exercised due to the possibly harmful side effects of pesticides on soil microflora. Many researchers have carried out in vitro and/or field experiments to investigate the effects of pesticides on the viability and behavior of soil microorganisms in this context (Ampofo et al., 2009; Angelini et al., 2013). In this context, previous research has demonstrated that glyphosate can change the natural ecosystem by affecting various components of the soil microbial population (Carlisle & Trevors, 1988; Ermakova et al., 2010).

Glyphosate (N-(phosphonomethyl) glycine) is a broad-spectrum systemic herbicide that inhibits the enzyme, 5-enolpyruvylshikimate 3-phosphate synthases (EPSPS), that produces amino acids and proteins in plants (Guijarro et al., 2018; Steinrucken & Amrhein, 1980; Thiour-mauprivez et al., 2019; Zhan et al., 2018). Furthermore, it is the most commonly used herbicide in the world of agriculture (Berman et al., 2020; Tang et al., 2019), the best-selling herbicide in the world, and in Morocco, it is also the most important (Battisti et al., 2021; Maldani et al., 2017a). Glyphosate is important in the growth and cultivation of herbicide-tolerant genetically modified plants (Newman et al., 2016).

It's becoming common in the world of weed control to debate about glyphosate without also mentioning paraquat, which has emerged as a key management tool for glyphosate-resistant species in recent years. (Moretti & Hanson, 2017). Paraquat is a highly efficient and cost-effective herbicide that is commonly used to suppress broadleaf weeds around the world. This herbicide via interaction and translocation inside the plant, rapid, non-selective compounds degrade green plant tissue (Ma et al., 2019; Maldani et al., 2017a; Moretti & Hanson, 2017).

Glyphosate and paraquat are the most commonly used pesticides in weed control in legume production, but most farmers ignore the presence of microorganisms (nitrogen-fixing bacteria) in the soil–plant system and are directly involved in the plant N nutrition, affecting the yield and production of legume crops, and by increasing soil fertility.

On this basis, the concept was born to find bacteria that can fix atmospheric nitrogen while still avoiding herbicides, as well as to determine the impact of paraquat and glyphosate on nitrogen fixation by inoculating Bituminaria bituminosa with four rhizobia strains treated with these two herbicides and evaluating plant growth and nodulation parameters. These four strains were evaluated in vitro previously, for their resistance to glyphosate and paraquat concentration (Maldani et al., 2017b, 2018). The present study was based on the results of these previous studies.

2 Materials and Methods

2.1 Chemicals

The glyphosate used was a generic RoundUp® (containing 360 g active ingredient/L glyphosate, Bayer, Germany). Besides, the paraquat used was a generic Gramoxone® (containing 200 g of active ingredient/L of paraquat, Syngenta, Switzerland). All other chemicals had the highest purity commercially available. The strains were cultured in YMB (Yeast Mannitol Broth, HIMEDIA Laboratories, India), containing 0.5 g/L KH2PO4, 0.2 g/L MgSO4·7H2O, 0.1 g/L NaCl, 0.5 g/L yeast extract, 10 g/L mannitol, in distilled water pH 7.0–7.2.

2.2 Bacterial Strains

Four different diazotrophic bacterial strains, namely Pantoea agglomerans, Rhizobium nepotum, Rhizobium radiobacter, and Rhizobium tibeticum, have been isolated from Bituminaria bituminosa root nodules (Ben Messaoud, 2015). Bacteria were chosen concerning their potential to fix nitrogen and solubilize phosphates and picked to test the effect of pesticides on nitrogen fixation. The bacteria studied were preserved at -80 °C in Yeast Mannitol Broth (YMB) + 40% Glycerol. Before using the four strains in this study, molecular sequencing was carried out to ensure their identification.

2.3 16S rDNA Gene Sequence Determination and Analysis of Phylogenetic Relationships

Analyses of the 16S rDNA gene sequences of bacteria in the study were performed as previously described (Troussellier et al., 2005). Total DNA extraction of bacterial isolates was performed with the CTAB method (Winnepenninckx, 1993). 16S rDNA loci were amplified using 16S rDNA forward domain-specific Bacteria, Bact27_F (5ʹ-AGAGTTTGATCCTGGCTCAG-3ʹ) and reverse primer Uni_1492R (5′-TACGYTACCTTGTTACGACTT-3ʹ) (Lane, 1991). The amplification reaction was performed in a total volume of 50 µl mixture containing 1 × solution Q (Qiagen, Hilden, Germany), 1× Qiagen reaction buffer, 1 µM of each forward and reverse primer, 10 µM dNTPs (Gibco, Invitrogen Co., Carlsbad, CA), 2.0 mL (50–100 ng) of the template and 2.0 U of Qiagen Taq Polymerase (Qiagen). The amplified 16S rDNA was sequenced using Macrogen Service (Korea). SIMILARITY_RANK from the Ribosomal Database Project (RDP) (Maidak et al., 1997) and FASTA Nucleotide Database Query (Pearson & Lipman, 1988) were used to determine partial 16S rDNA sequences to estimate the degree of similarity to other 16S rDNA gene sequences. For each strain and depending on the expected size of the 16S rRNA gene, DNA was amplified to end up with a band of 1500 bp. The bands of each strain were sequenced. Our sequences were between 600 and 1500 bp in size. Analysis and phylogenetic affiliates of sequences were performed as previously described (Cappello et al., 2012; Yakimov et al., 2006). The neighbor-joining method was used to build a phylogenetic tree.

2.4 Soil

To provide a good substrate for the growth of Bituminaria bituminosa, a mixture of three soils was used. All soil samples were collected in the spring following the standard protocol (NM ISO 10381–1). The following soils were used: Vertisols (latitude 33°, 52′, 8″ N; longitude 5°, 36′, 4″ W); Calcimagnesic (33°, 50′, 8″N; 5°, 28′, 3″W); Isohumic (33°, 51′, 6″ N; 5°, 28′, 0″ W) (Table 1). Samples from the three soils were taken for analysis: texture, pH, exchangeable cations, organic matter, and phosphorus were measured. Every soil sample was autoclaved for 20 min at 121 °C and repeated three times with a one-week interval.

Table 1 Physico-chemical characteristics of the studied soils
Full size table

2.5 Preparation of Inoculum and Bacterial Pretreatment

Our experience consisted of exposing the nitrogen fixing strains to different doses of glyphosate and paraquat separately in the laboratory, before inoculating the plants with these strains in a greenhouse (Faculty of Sciences at Moulay Ismail University in Meknes, Morocco).

The inoculum was prepared for all the strains evaluated by cultivation in Yeast Mannitol Broth (YMB) for 7 days at 30 °C under shaking conditions (150 rpm) until the growth reached the exponential phase. The strains were cultivated in 100 ml of YMB for 7 days at 30 °C under shaking conditions (150 rpm) in the presence of different concentrations of each herbicide (0.05, 0.1, 5.4, 10.8 g/L for glyphosate, and 0.05, 0.1, 2, 4 g/L for paraquat) and incubated at 28 °C, (5.4, 10.8 g/L for glyphosate, and 2, 4 g/L for paraquat are the doses recommended by the National Office of Health and Food Safety of Morocco for each pesticide used. In order to eliminate traces of the two herbicides (glyphosate and paraquat) to avoid damaging the plants during inoculation, bacteria were harvested by centrifugation at 6 415 rpm for 5 min (SIGMA 1–115 K, Germany), washed with 0.9 percent sterile physiological solution, and finally re-suspended in sterile physiological solution to a 0.5 McFarland nephelometer standard (optical density of 0.108 at 625 ~ 1.5 × 108 CFU/mL nm measured by Spectrophotometer UV-2005, Selecta; Spain), this suspension was used as inoculum.

2.6 Preparation and Inoculation of Plants

Bituminaria bituminosa seeds (98% viability) were hand-sorted to ensure size uniformity and to discard damaged seeds. They were then disinfected by immersing them in 80 percent (v/v) ethanol for one minute and rinsing them three or four times in sterile purified water; finally, 5 percent (v/v) sodium hypochlorite was applied, gently shook for five minutes, and rinsed six times.

After disinfection, seeds were placed in Petri dishes (140 mm Ø) containing 50 mL of 0.5 percent agar in distilled water (Biokar, France) and incubated at 28 °C for 3 days to facilitate pre-germination. Plastic pots containing 500 g of sterile soil were used for the tests.

In the greenhouse, the experiments used a fully randomized approach of four replicates of each treatment. Plants were irrigated based on the soil's field capacity.

After the appearance of the first true leaf, each bacterial strain's inocula has been added separately. The inoculation has been carried out by injection of 15 ml of bacterial suspension (1.5 × 108 CFU/mL) into the middle of the root using a sterile syringe and needle for each of the four replicates. To increase the likelihood of infection, the inocula were applied three times (Chanway et al., 2000). Plants not inoculated and which have not received N-fertilizer served as negative control while plants not inoculated plus N-fertilizer at a rate of 70 μg N mL−1 applied as KNO3 solution were used as positive control.

2.7 Morphological Traits of Plants Growth

After six months from the beginning (February to July 2019) of our experiment, the size of the plant was measured from the intersection between the stem and the root to the tip of the last leaf of the plant. Also, fresh and dry weights (in the oven at 70 °C for 72 h) were weighed. Roots are cut, well-shaken, and cleaned with water to eliminate the attached soil particles. Nodules were detached and counted, and their dry weight (at 65 °C for 48 h in the oven) was registered. The leaf area (stipules and leaflets) was measured when harvested using Mesurim Pro 3.4 freeware (Aubernon et al., 2015).

2.8 Amino Acids Content Determination

The amino acids content was determined according to the method described by (Yemm et al., 1955). 50 μL of plant extract, 0.5 mL of 80% ethanol, 0.5 mL of citrate buffer (0.2 M, pH 5), 1 mL of ninhydrin acetone solution (1 g of ninhydrin in 125 mL of acetone), were added in a test tube, and incubated in a water bath at 100 °C for 15 min. After cooling, 8 mL of distilled water was added. The absorbance was read at 570 nm, ethanol was used in instead of the extract in a blank. The amino acid content was assessed by reference to a standard curve made with a concentration range of pure glycine.

2.9 Total Nitrogen Determination of the Plant by the Kjeldahl Method

To measure the plants’ amount of nitrogen, we used the Kjeldahl method (Nelson & Sommers, 1973). Samples of 200 mg of ground dry matter were inserted into the digestion tube to which the following reagents were added: 3 mL of concentrated sulfuric acid, 1.1 g of a catalyst consisting of K2SO4, CuSO4·5H2O, and selenium in the proportions of 100:10:1 respectively. The temperature of the digester block was gradually increased to 300 °C (VELP Scientifica Model DK 20 Kjeldahl Digestion Unit, Italy).

Digestion continued until a clear green color was obtained (about 4 h), then allowed to continue for one hour. After cooling the tubes, the contents were transferred to distillation balloons. 5 mL of the 4% boric acid-indicator solution (80 g of boric acid (H3BO3) dissolved in 1500 mL of boiling distilled water. After cooling, 20 mL of Bromocresol green (BCG) solution (100 mg of BCG powder/100 mL of alcohol); and 12 mL of methyl solution (100 mg of red methyl powder/100 mL of alcohol) were added. Then, the solution was adjusted to 2L with the distilled water) was added to an Erlenmeyer flask and placed under the condenser outlet of the distillation apparatus. 20 mL of 10 N NaOH was transferred slowly throughout the use of a funnel into the distillation flask. The distillation was completed when the volume in the flask containing the boric acid indicator solution was about 40 mL (about 4 min). Titration was performed using 0.05 N sulfuric acid (Steam distillation unit: KJELDAHL PRO-NITRO-S, Semi-automatic, JP SELECTA. SPAIN).

Total nitrogen content was computed as follows:

$$\%N\;\mathrm{of}\;\mathrm{samples}=\left[\left(V-V_0\right)\times14\times\mathrm{Na}/\mathrm m\right]\times100$$

Where V = volume (mL) titrating the sample, V0 = volume (mL) titrating a blank, Na = normality of the sulfuric acid used for the titration, m = mass (mg) of the analyzed sample (Bremner & Hauck, 1982).

The nitrogen content of each plant was determined by combining the dry plant biomass weight of the shoot with the overall nitrogen content of the plant.

Symbiotic efficiency was calculated by comparing the effect of each treatment inoculated with the native isolated strains with nitrogen applied control (Plant nitrogen content in inoculated pots/plant nitrogen content in nitrogen application) × 100 (Kawaka et al., 2014).

2.10 Statistical Analyses

The main variables evaluated were: shoot growth, root growth, number of nodules, nodule dry weight, nodule fresh weight, leaf area, amino acid content, nitrogen content, and symbiotic efficiency. The normal homogeneity of the variances (Bartlett’s test) was assessed. All the response variables met the normality assumption. We conducted a one-way analysis of variance (ANOVA) tests to assess the effect of paraquat and glyphosate on nitrogen fixation of Bituminaria bituminosa inoculated with four rhizobia strains treated with these two herbicides and to evaluate plant growth and nodulation parameters. For post-hoc tests, we used Tukey’s test, which is a multiple comparison test and is applicable when more than two means are being compared. The results were expressed as the mean ± standard deviation (SD). The software R (version 3.5.3) for statistical analysis was used. OriginLab software (version 8.6) to create and customize publication-quality graphs was also used.

3 Results

3.1 Characterization and Molecular Identification of the Strains

The 16S rDNA sequences were obtained from Macrogen Service (Korea) and data available in the NCBI data bank. The similarities were identified using the Blast server after the DNA sequences were deposited in the Gene Bank database, and the isolates were compared to available type strains.The isolate NFB 1 showed the maximum base sequence similarity (99.70%) to type strains Pantoea agglomerans KABNA2 (Accession number MT605811.1) and Pantoea agglomerans DSM 3493 (Accession number NR_041978.1) showed the maximum base sequence similarity. Based on maximum relatedness to type strains, isolate NFB 1 was identified as Pantoea agglomerans (Accession number MT605811.1) (Fig. 1). For the isolate, NFB 2 showed the maximum base sequence similarity (100.00%) to type strains Rhizobium nepotum (Accession number FR870230.1) and Rhizobium nepotum 39/7 (Accession number NR_117203.1) showed the maximum base sequence similarity (100.00%). Based on maximum relatedness to type strains, isolate NFB 2 was identified as Rhizobium nepotum (Accession number FR870230.1) (Fig. 1). For the isolate, NFB 3 showed the maximum base sequence similarity (99.04%) to the type strains Rhizobium radiobacter NCBI 9042 (Accession number MT534523.1) and Rhizobium radiobacter ATCC 19358 (Accession number MT534520.1) showed the maximum base sequence similarity (98.74%). Based on maximum relatedness to type strains, isolate NFB 3 was identified as Rhizobium radiobacter (Accession number MT534523.1) (Fig. 1). For the isolate, NFB 4 showed the maximum base sequence similarity (%) to the type strains Rhizobium tibeticum CCBAU 85,039 (Accession number NR_116254.1) and Rhizobium tibeticum (Accession number FR714442.1) showed the maximum base sequence similarity (100.00%). Based on maximum relatedness to type strains, isolate NFB 4 was identified as Rhizobium tibeticum (Accession number NR_116254.1) (Fig. 1).

Fig. 1
figure 1

Unrooted neighbour-joined phylogenetic tree of 16S rRNA genes for the four strains NFB1, NFB2, NFB3, and NFB4 with highest similarity type strain after a BLAST. The tree was generated based on DNA sequences obtained in specific bands references downloaded from NCBI databas. Bootstrap values are given at branching points. The scale bar represents 1 nucleotide substitution per 100 nucleotides

Full size image

3.2 Effect of Glyphosate and Paraquat on Plants Growth

The data indicated that there were no significant differences in plant growth at low glyphosate and paraquat concentrations of 0.05 g/L and 0.1 g/L, respectively, as compared to controls for all treatments (p > 0.05). (Table 2). Whereas we noticed that, compared to the control treatment, the high doses utilized have a significant (p < 0.05) effect on the plant height inoculated with the strains studied. The difference between control and plants treated with 4 g/L paraquat inoculated with Pantoea agglomerans and Rhizobium radiobacter was significant (p < 0.05). However, when used at 10.8 g/L, glyphosate affects all plant heights infected by all strains studied (Table 2).

Table 2 Effect of pesticides on the shoot growth of Bituminaria bituminosa in soil under inoculation with Pantoea agglomerans, Rhizobium nepotum, Rhizobium radiobacter and Rhizobium tibeticum
Full size table

There were no significant differences in dry biomass between treatments at low doses (0.05 g/L and 0.1 g/L for glyphosate and paraquat, respectively) compared to controls (p > 0.05). Additionally, no significant difference in dry biomass was observed between controls and all treatments (paraquat and glyphosate) on plants inoculated with the Rhizobium radiobacter and Rhizobium tibeticum how (Table 2).

There were no significant differences in root length between strains and controls (p > 0.05), except for plants inoculated with Rhizobium tibeticum, which were significantly (p < 0.05) affected by the application of both herbicides at 5.4, 10.8 g/L glyphosate and 2.44 g/L paraquat (Tables 3).

Table 3 Effect of pesticides on the root growth on Bituminaria bituminosa in soil and inoculated with Pantoea agglomerans, Rhizobium nepotum, Rhizobium radiobacter and Rhizobium tibeticum
Full size table

Additionally, no significant differences in dry root weight were observed at low doses (0.05 g/L and 0.1 g/L for glyphosate and paraquat, respectively) compared to controls for any treatment (p > 0.05), and this similar behavior was observed on plants inoculated with Rhizobium radiobacter and Rhizobium tibeticum and treated with high concentrations of paraquat (4 g/L) and glyphosate (10.8 g/L). Under these conditions of high herbicide concentration, the herbicides had a considerable (p < 0.05) effect on plants inoculated with Pantoea agglomerans and Rhizobium nepotum (Table 3).

No significant variations in leaf area were observed at low doses (0.05 g/L and 0.1 g/L for glyphosate and paraquat, respectively) as compared to controls for any treatment (p > 0.05), while the effect of high concentration has similar behavior for dry root biomass (Table 4).

Table 4 Effect of pesticides on leaf area and amino acid content on Bituminaria bituminosa in soil and inoculated with Pantoea agglomerans, Rhizobium nepotum, Rhizobium radiobacter and Rhizobium tibeticum
Full size table

3.3 Effect of Glyphosate and Paraquat on the Amino Acid Content

For amino acid content, statistically, there were no significant differences for glyphosate and paraquat at low doses of 0.05 g/L and 0.1 g/L compared to controls for all treatments (p > 0.05) (Table 4). Also, at high levels of herbicides, no significant difference (p > 0.05) was observed in Rhizobium nepotum inoculated plants, whereas a significant difference (p 0.05) was found in Rhizobium radiobacter. Plants inoculated with Pantoea agglomerans, on the other hand, showed no significant difference (p > 0.05) between control and concentrations of 5.4 g/L glyphosate and 2 g/L paraquat. For Rhizobium tibeticum, a significant difference (p 0.05) was found between treatments and control except for 2 g/L paraquat concentration (Table 4).

3.4 Effect of Glyphosate and Paraquat on the Symbiosis with Atmospheric Nitrogen-Fixing Rhizobacteria

In general, there was a significant difference (p < 0.05) for all parameters compared to controls for all treatments (Table 5). At low doses (0.05 g/L and 0.1 g/L for glyphosate and paraquat, respectively), no significant variations in dry nodule weight were observed compared to controls for all treatments (p > 0.05). While for high doses, the number of nodules in plants inoculated with rhizobacteria strains previously treated with pesticides clearly shows the adverse impact of the pesticides used. Our findings indicated a significant difference (p < 0.05) in all plants compared to the control treatment at all high pesticide dosages applied, with negative impacts associated with the pesticide concentration (Table 6).

Table 5 Effect of pesticides on the nodule on Bituminaria bituminosa in soil and inoculated with Pantoea agglomerans, Rhizobium nepotum, Rhizobium radiobacter and Rhizobium tibeticum
Full size table

The major objective of our work was to quantify the contribution of BNF to the nitrogen nutrition of Bituminaria bituminosa inoculated with various diazotrophic bacteria strains and previous pesticide treatment in the laboratory before inoculation. The findings strongly support the idea that glyphosate and paraquat have a detrimental effect on the symbiosis between atmospheric nitrogen-fixing plants and bacteria. There was a statistically significant difference between the treatments and the control in the analyzed parameters of plant nitrogen content and symbiotic efficiency (p < 0.05). (Table 5).

Table 6 Effect of pesticides on N content by NFB and symbiotic efficiency on Bituminaria bituminosa inoculated with Pantoea agglomerans, Rhizobium nepotum, Rhizobium radiobacter and Rhizobium tibeticum
Full size table

Consequently, the content of N in plants decreased up to 74% for those inoculated with Pantoea agglomerans for the 4 g/L paraquat concentration, 56% for Rhizobium nepotum for the 5.4 g/L glyphosate concentration, 90% for Rhizobium radiobacter for the 10.8 g/L glyphosate concentration, and 62% for Rhizobium tibeticum for the 10.8 g/L glyphosate concentration, Besides, symbiotic efficacy decreased up to 85% for Pantoea agglomerans for the 4 g/L paraquat concentration, 56% for Rhizobium nepotum for the 5.4 g/L glyphosate concentration, 88% and 80% for Rhizobium radiobacter, and Rhizobium tibeticum, successively, for the 10.8 g/L glyphosate concentration (Tables 5).

3.5 Correlation Study

A negative correlation was observed between herbicide concentrations, and all parameters studied: plant growth, nitrogen content, and symbiotic efficiency. This finding was observed for treatments, glyphosate, and paraquat. Therefore, there is a significant relationship (p < 0.05) between herbicide concentration and all parameters studied (Fig. 2).

Fig. 2
figure 2

Relationship between all parameters tested in this study in plants and the treatments. Con concentration, Her herbicide, Ndwt nodule dry weight, Rle root length, La leaf area, Aa amino acid content, NC nitrogen content NQ, Ph plant height, Sdwt shoot dry weight, Rdwt root dry weight, SE symbiotic efficiency. Negative correlations are in red color and positive correlations in dark blue

Full size image

4 Discussion

4.1 Effect of Glyphosate and Paraquat on Plants Growth

Bituminaria bituminosa is a perennial legume species widely spread in the Mediterranean region and near Europe, North African (including the Canary Islands), and Asian countries, with potential as a drought-tolerant pasture species and as a source of pharmaceutical compounds (Pazos-Navarro et al., 2011).

The study clearly shows that using the herbicide at the recommended doses inhibits nitrogen-fixing bacteria from replenishing natural nitrogen released by the soil, leading to lower yields, stunted growth, and a growing need for nitrogen additives (chemical products) to enhance crop production, ultimately contributing to soil degradation. Pesticide residues in soil could not only reduce crop yield indirectly, but also increase the need for N synthetic fertilizers, raising farmers’ costs and contributing to environmental pollution.

Besides, our results showed that both herbicides decreased plant growth and disturbed microbial symbiosis of Bituminaria bituminosa when used at field application of recommended rates. The influence could occur directly or indirectly on the root microflora of the plant itself. The results show that the impact on microorganisms may be the key explanation for plant inhibition of growth. Similar inhibitory effects of herbicides have been recorded previously. According to Abd-Alla et al., 2000, the usage of pesticides (Pyrozophos, Bromoxynil, Paraquat, Profenfos, and Methyl parathion) at the recommended doses has contributed to a decline in growth of cowpea (Vigna sinensis L.), common bean (Phaseolus vulgaris L.) and lupin (Lupinus albus L.). Likewise, Shankar et al. (2012), demonstrated that when 2, 4-d amine salt, Roundup, and Atrazine were applied to the side root system of cowpea, no nodules appeared, which reduced their vigor despite the use of pesticides at the recommended doses. In the current study, when paraquat and glyphosate were applied at double the recommended rate, the plants showed evidence of phytotoxicity. The study by Duke (2011), showed that glyphosate is stable inside plants and has been metabolized very slowly.

4.2 Effect of Glyphosate and Paraquat on Symbiotic N2 Fixation

After processing our results statistically, the total number of nodules on Bituminaria bituminosa was significantly reduced as herbicide doses increased. The inhibitory effect of herbicides on nodulation was observed at the lowest herbicide level and became more negative with the increase in herbicide concentration. Furthermore, total nodule weight per plant was reduced in herbicide-treated treatments in relation to herbicide rate.

Numerous parameters related to the symbiosis of plant with atmospheric nitrogen-fixating rhizobacteria (nodule number, nodule mass, nodule weight, nodule dry and fresh weight, plant nitrogen content, and symbiotic efficiency) were reduced when the paraquat and glyphosate pesticides were applied to Bituminaria bituminosa. After statistical analyses, we observed that both pesticide applications led to a reduction in the number and nodule mass of Bituminaria bituminosa compared to the control. On the other hand, despite the low doses of these two pesticides may not impact significantly the growth parameters but nitrogen-fixing parameters were weakly affected. From this study, it is clear that the soil was an important N source for the plant, because for Rhizobium nepotum for the treatment 5.4 g/L of glyphosate, even with poor nodulation, the plant growth suffers less than nodulation. The impact of glyphosate application on the root nodule mass was also reported by Fan et al. (2017) and Zobiole et al. (2011). Also, Abd-Alla and Omar (1993) reported that even when applied at the recommended doses, paraquat caused a decrease in nodulation, and similar results were observed by Flores and Barbachano (1992), who worked on the effect of paraquat on the growth and nodulation of Rhizobium meliloti, and found that paraquat pre-treatment caused a strong decrease in nodule weight. Moreover, the negative effects of herbicides on nodulation, symbiotic nitrogen fixation, growth, and yield of pea (Pisum sativum) increased with increasing the application rate, as reported by Singh and Wright (1999). Furthermore, the reduction in root mass and nodule mass due to glyphosate treatment has been observed in greenhouse experiments by many authors (King et al., 2001; Reddy et al., 2000; and Zobiole et al., 2011). Besides, Clark and Mahanty (1991) found that paraquat was more toxic than MCP (4-(4-Chloro-o-tolgloxy) and Bentazone on nodulation and growth of white clover. The deleterious effect of glyphosate was shown by Molin (1998), who suggested that the rapid translocation of glyphosate from foliage to roots of treated plants is the key to the inhibition of root growth.

Nitrogen estimation in plants is the most widely reported index of assessing the effect of pesticides on atmospheric biological nitrogen fixation. Our findings revealed a significant influence of both herbicides on the BNF process, as evidenced by a decrease in the amount of nitrogen in the plants and a decrease in symbiotic efficiency. Xia et al. (2017) reported that a change in nitrogen fixation activity may be directly related to the degree of nodulation but, at the same time, related to nitrogen activity regardless of its effects on the degree of nodulation. Likewise, our results were consistent with those of Fox et al. (2007), who showed the negative effect of pesticides on the symbiotic efficiency of nitrogen-fixing rhizobacteria. Besides, Lu et al., 2020 proved in their study on the effect of chlorpyrifos on nitrogen fixation in rice-vegetated soil containing Pseudomonas stutzeri A1501, the toxicity of this pesticide on nitrogen-fixing bacteria, and its mechanisms. Moreover, they confirm that chlorpyrifos caused the production of reactive oxygen species which was directly responsible for decreased nitrogenase activity by inhibiting the expressions of nitrogen-fixing related genes. Although the nitrogen content was negatively correlated with the symbiotic efficiency, these results were similar to the data reported here and also confirmed by Albrecht et al. (2014), who showed that without an active symbiosis there was a decrease in nitrogen intake to the plant, consequently decreasing protein content. The only study on this topic with contradictory findings is that of Bärwald Bohm et al. (2014), who found that glyphosate has no effect on glyphosate-resistant soybean and that glyphosate use has no effect on important microbial processes such as biological nitrogen fixation but promotes high residual levels in soil and seeds. It is important to point out that nowadays in Brazil, most of the soybean area, over 38 Mha, is occupied by glyphosate-resistant soybean varieties, where the BNF process responds to 80% of the total N demand of this crop (Alves et al., 2003).

Needless to say, the symbiotic nitrogen-fixing bacteria are influenced by pesticides. As a result, more research is needed to develop more specific PCR primers for detecting symbiotic N2 fixing activity within root nodules. Additionally, genomics tools must be combined with field studies and microbial selection in order to define and explain the effect of pesticides on nitrogen-fixing bacteria.

5 Conclusion

The results provide compelling evidence that both herbicides used in this study negatively affect the symbiosis of plants with atmospheric nitrogen-fixing bacteria and, as a consequence, cause a reduction in the level of nitrogen in the soil. As a result, this practice gives the green light to the use of chemical fertilization to increase agricultural yield, which in its turn causes a significant disruption of the composition of the diazotrophic bacterial community and also the reduction of nitrogenase of soil diazotrophic bacteria. The results of the study suggested that the declined growth of herbicide-treated plants was due to the direct effects of the herbicides on rhizobia rather than to the indirect effects of the herbicides on Bituminaria bituminosa.