2.1. Compatibility with Fungicides

Fungicides are chemical products formulated to prevent the infection of plant tissues by phytopathogenic fungi, and in some cases, capable of extending the control of diseases caused by bacteria and viruses. The control exerted by fungicides can be mediated by killing the pathogen, temporarily inhibiting its germination and growth, or by affecting the production of spores ^[32]. Over the years, several fungicides have made it to the market; some have stood out and remained at the top in the list of the most used fungicides until today, more effective products have replaced others, and some have been banned. Fungicides of contact generally do not have a specific mode of action, are highly toxic, and when applied to seeds, soil, and plant leaves limit the pathogen survival. The most common are thiram (dithiocarbamate), captan (quinone and heterocyclic), exon (aromatic), and guazatine. Upon entering microbial cells, the molecules promote a series of chemical reactions in nucleic acids and their precursors, and metabolic routes, affecting cell survival ^[32].

The majority of studies on compatibility have been performed with fungicides and microbial inoculants carrying rhizobia. Fungicides may affect several steps of the symbiosis, from the survival of the rhizobia on the seed to nodule formation and N₂ fixation efficiency; in general, studies have been performed with soybean (e.g., ^{[33][34][35][36]}). The use of pesticides intensified in the past two decades, and so did concerns about their compatibility with inoculants ^[17].

Brikol et al. ^[21] evaluated the effects of applying different concentrations of the fungicide Thiram (10 to 750 µg mL⁻¹) on soybean seeds inoculated with B. japonicum, which were grown under greenhouse conditions for 75 days. They observed that concentrations above 100 µg mL⁻¹ reduced nodule number and dry weight, as well as the activity of the bacterial enzyme nitrogenase, responsible for the nitrogen fixation process. Similarly, there are reports ^[24][37] of decrease in soybean nodulation and N accumulation in plants when seeds were inoculated with B. elkanii (SEMIA 5019) and B. japonicum (SEMIA 5079) and treated with different fungicides, benomyl, captan, carbendazin, carboxin, difenoconazole, thiabendazole, thiram, and tolylfluanid. Changes in nodule number were also observed in a field trial by Zilli et al. ^[38] when soybean seeds were treated with either carbendazin + thiram or carboxin + thiram.

Interestingly, results of some studies indicate differences between strains in their tolerance to fungicides. For example, in soybean seeds treated with carbendazim + thiram, the lowest tolerance was observed in B. elkanii amongst four soybean Bradyrhizobium strains used in commercial inoculants in Brazil [B. elkanii (SEMIA 5019 and SEMIA 587), B. japonicum (SEMIA 5079), and B. diazoefficiens (SEMIA 5080)] ^[38]. In another study, Gomes et al. ^[39] observed no effects on nodulation when seeds were inoculated with B. japonicum SEMIA 5079 + B. diazoefficiens SEMIA 5080 and treated with carbendazin + thiram. More recently, when the compatibility of B. japonicum SEMIA 5079 and B. elkanii SEMIA 587 was verified with Standak^® Top, composed of a mixture of two fungicides and one insecticide (piraclostrobin, thiophanatemethyl, and fipronil), the effects were also more drastic for B. elkanii ^[25]. Altogether, indications are that B. elkanii is less tolerant to fungicides than B. japonicum or B. diazoefficiens.

It is reasonable to postulate that the main effects of the fungicides used in seed treatment would be the decrease of rhizobial survival or inhibition of the root infection process, consequently affecting nodulation and BNF, and as a result grain yield, as observed by Zilli et al. ^[38]. In that study, the grain yield reduced by 20%, in addition to a decrease in the N content in grains when seeds were treated with B. elkanii SEMIA 587 + carbendazin + thiram (Figure 2). However, some reports have indicated that the effects of fungicides may appear later. For example, in a study by Gomes et al. ^[39], although fungicides (carbendazin + thiram) did not affect nodulation, plants had lower number of pods per plant, grains per plant, and yield. Intriguingly, in two field experiments performed with soybean in Brazil, seed treatment with Standak^® Top affected the total N accumulated in the grains of plants relying on both BNF and N fertilizer, indicating the negative impact of the pesticide on N metabolism ^[25]. In another study, a decrease in both protein and oleic acid contents was observed in soybean inoculated and treated with mefenoxan + fludioxonil ^[40].

Figure 2. Representation of the effects reported on the incompatibility between pesticides and inoculants, from the moment of contact with microbial cells to the damage to plant development.

As mentioned previously, negative effects may be related to the toxicity of fungicides on microbial cells, followed by impacts on microbial metabolism, reducing the effectiveness of the inoculant. Ahmed et al. ^[41] evaluated the growth of Bradyrhizobium sp. and Rhizobium sp. in Petri dishes containing solid culture medium and different concentrations of fungicides (captan, thiram, luxan, milcurb, and frernasan-D), soaked in discs placed on the medium. The bacteria were tolerant to concentrations below 100 µg L⁻¹, but at 1000 µg L⁻¹ inhibition of growth and a decrease in the colony diameter of the surviving cells were observed (Figure 2). Rathjen et al. ^[42] also reported that Rhizobium leguminosarum bv. viciae (WSM1455) was sensitive to the fungicides Thiram 600 and P-Pickel T (PPT), and their active ingredients (thiram and thiabendazole) at concentrations above 200 µg disc⁻¹, with growth halos greater than 10 mm around the disks containing the fungicide.

To verify the survival of B. elkanii (SEMIA 5019) and B. japonicum (SEMIA 5079) on seeds treated with fungicides, Campo et al. ^[24] recovered and counted living cells from seeds inoculated and treated with benomyl, captan, carbendazin, carboxin, difenoconazole, thiabendazole, thiram, or tolylfluanid. After only 2 h of contact with carbendazin + thiram, viable bacterial cells were reduced by up to 64%, and reached 83% after 24 h (Table 1). Mortality was verified with all other fungicides, reaching 95% with the mixture of thiabendazole and tolylfluanid after 24 h of contact. In addition to rhizobia, fungicides can also impair the contribution of other PGPB. The toxic effect of the combination of carbendazin + thiram was also verified for A. brasilense strains Ab-V5 and Ab-V6 by Santos et al. ^[27]. In that study, a decrease in the recovery of viable cells from seeds only after 2 h of contact was observed, and the viable cell count drastically decreased after 24 h, in comparison with the seeds not treated with fungicides (Figure 2).

2.2. Compatibility with Insecticides

Many herbivorous insects feed on plants during their larval and adult stages, and/or some are important vectors of plant diseases. In both cases, insects may cause serious damages to crops. Insecticides, usually synthetic chemicals, acting as ovicidal, larvicidal, and adulticidal agents are used to prevent growth or kill insects ^[45]. Neonicotinoids, organophosphates, diamides, pyrethroids, and carbamates act on nerves and muscles; phosphides, cyanides, and carboxamides on respiration, and cyclic ketoenols and ecdysone are agonists that interfere with insect growth and development ^[46].

Rathjen et al. ^[42] evaluated the in vitro toxicity of an imidacloprid-based insecticide on R. leguminosarum bv. viciae (WSM1455), and Mesorhizobium ciceri (CC1192). The strains were applied on solid culture medium in Petri dishes and sterile filter discs containing different concentrations (0, 100, 200, and 300 µg discs⁻¹) of the insecticide. The inhibition of R. leguminosarum growth was observed at all concentrations. In alfalfa (Medicago sativa L.), Fox et al. ^[22] reported that seed treatment with the insecticides methyl parathion, DDT and pentachlorophenol affected the symbiosis with Ensifer (syn. Sinorhizobium) meliloti. The insecticides not only reduced nodule number and dry weight but also nitrogenase activity in nodules and plant biomass production. Ahemad ^[23] also reported several negative effects with the application of pyriproxifene, at the recommended dose of 1300 μg kg⁻¹ soil, in chickpeas (Cicer arietinum L.), peas (Pisum sativum L.), mung beans (Vigna radiata L. Wiclzek), and lentils (Lens esculentus, = Lens culinaris Medik) grown in pots that remained in an open field. Although the plants had not been inoculated, inferring that nodulation was related to indigenous rhizobia, pyriproxifer resulted in a 44% decrease in nodule number in peas, 14% in mung beans, and 5% in chickpeas and lentils, resulting in decreases in nodule dry weight compared with the controls not treated with insecticide. There was also a 17% decrease in the concentration of nitrogen in the roots of chickpeas, 15% in peas, 14% in mung beans, and 18% in lentils, and a 5% decrease in the protein contents in grains of chickpeas, 4% in mung beans, 3% in lentils, and 1% in peas, compared with the controls not treated with insecticide.

Insecticides also affect PGPB other than rhizobia. For example, Fernandes et al. ^[44] studied the effects of five insecticides (imidacloprid, fipronil, fenamethoxam, endosulfan, and carbofuran) indicated for sugarcane (Saccharum spp.) on the diazotrophic bacterium Herbaspirillum seropedicae. In vitro evaluations of cell growth after 33 h indicated that the insecticides that most interfered with bacterial growth were endosulfan and carbofuran. Madhaiyan et al. ^[43] evaluated the effect of different insecticides (monocrotophos, malathion, chlorpyriphos, diclorvos, lindane and endosulfan) on Gluconacetobacter diazotrophicus, another PGPB found in association with sugarcane. Except for malathion, all other insecticides reduced cell concentration, and lindane lysed the cells (Figure 2). In the same study, nitrogenase activity was fully inhibited by monocrotophos, dichlorvos, and lindane, 83.3% by chlorpyriphos, 80.9% by melation, and 33.4% by endosulfan. Concerning the synthesis of indoleacetic acid (IAA) and gibberellin A3 (GA3) by G. diazotrophicus, inhibition was observed with lindane, with decreases of 93.2% and 96.5% for IAA and GA₃, respectively. The authors also described that the insecticides dichlorvos, chlorpyriphos, and lindane completely inhibited the solubilization of phosphate (P) and zinc (Zn) ^[43] (Figure 2).

2.3. Compatibility with Herbicides

Another class of important pesticides for agriculture are herbicides used for weed control. Herbicides have different degrees of specificity based on differences in biochemical pathways in certain plant groups. The mode of action of herbicides has specific degrees of toxicity, depending on the biochemical differences of the plants, and is generally related to the cell division process ^[47]. Souza and Guedes ^[47] gathered studies on the action of several herbicides on Allium cepa and Vicia faba plants used as bioindicators, and described how most of these studies reported that herbicides induced both a decrease in the mitotic index, and the occurrence of mitotic chromosomal aberrations.

Examples of herbicides applied worldwide include glyphosate, paraquat, and diuron. Among these, the most well-known is glyphosate; since its introduction in the 1970s, its use spread quickly, facilitated cropping, but also implied in the growing appearance of resistant weeds, resulting from a natural process of plant adaptation, and decreasing its efficacy ^[48][49][50]. An important alternative to minimize this problem, in addition to integration with other control methods, would be diversification in the use of herbicides, including others with different mechanisms of action ^[51].

Few studies have investigated the compatibility between herbicides and inoculants. Madhaiyan et al. ^[43] evaluated the effects of different herbicides (butachlor, alachlor, atrazine, and 2,4-D) in liquid culture medium on the growth and metabolism of G. diazotrophicus. Cell growth was impaired only in the presence of 2,4-D, but all herbicides reduced nitrogenase activity, the production of IAA and GA₃, and the solubilization of P and Zn. The highest inhibition of nitrogenase activity (73.6% and 65.3%) was observed with alachlor and atrazine, respectively, while butachlor mostly affected production of IAA (53.3%), whereas 2.4-D mostly affected the production of GA₃ (78.8%). Butachlor was also responsible for the strongest reduction in the solubilization of P and Zn. Therefore, although herbicides do not affect cell survival, they significantly affect metabolism of G. diazotrophicus ^[43].

In an assay performed under greenhouse conditions, Angelini et al. ^[52] evaluated the effects of imazetapir, imazapic, S-metachlor, diclosulam, and glyphosate on diazotrophic bacteria in the soil during the cultivation of peanuts (Arachis hypogaea L.). The seeds were sown and the herbicide was sprayed on the soil surface. All herbicides caused reduction in cell concentration in both free and symbiotic diazotrophic bacteria, and this negative impact was confirmed under field conditions even one year after the application. Nitrogenase activity also reduced due to herbicides, except for glyphosate ^[52] (Table 2).

The ability of cyanobacteria to fix atmospheric nitrogen in flooded soils suitable for rice cultivation make this group an important ally in maintaining soil fertility and contributing to cereal yield. Thus, Dash et al. ^[53] evaluated the responses of cyanobacteria in rice plantation soil to the exposure of different agrochemicals, including the herbicide benthiocarb that was applied in one dose at the time of puddling. The herbicide decreased cell growth, which was even worse when combined with urea (used as a fertilizer). Benthiocarb reduced nitrogenase activity by between 13–27%, compared with the control without herbicide, and its combination with urea resulted in an even greater reduction in addition to a decrease in the nitrogen accumulation that reached 47% at 60 days.

Concerning the symbiosis between legumes and rhizobia, in general, herbicides have been considered less toxic than fungicides and insecticides ^[37], with glyphosate being the one with lower toxicity ^[54][55]. Although the negative effects of glyphosate on B. japonicum growth were reported by King et al. ^[56], the doses in the experiment were far higher than those recommended for field application. With the release of genetically modified (GM) genotypes tolerant of herbicides, studies on the compatibility with the GM genotypes and herbicides have begun. In soybean, which represents the most used herbicide-tolerant species, glyphosate-resistant (Roundup Ready) pairs of nearly isogenic cultivars were evaluated in six field sites in Brazil for three crop seasons. Although the transgenic trait negatively affected some BNF variables, these effects had no significant impact on soybean grain yield, and no consistent differences between glyphosate and conventional herbicide application were observed on BNF-associated parameters ^[57]. Similar results were reported in 20 field trials performed with soybean with the ahas transgene, imidazolinone, and conventional herbicides ^[58]. However, it is worth mentioning that with the increasing doses of the herbicides, BNF can be reduced, especially under abiotic stressing conditions, as shown for glyphosate in soybean ^[59][60]. Interestingly, it has been long shown that several members of the family Rhizobiaceae were able to degrade glyphosate ^[61], ability that has been few explored, but that it can contribute to decrease the toxicity effects. Indeed, the search for indigenous and engineered microorganisms, used as isolated microbial species or microbial consortia can result in the complete mineralization of herbicides such as atrazine ^[62].

2.4. Compatibility with Mixtures (Fungicides, Insecticides, and Herbicides)

Approximately 70% of the pesticides available in the market contain mixtures of two or more types of fungicides and insecticides ^[63] and are often combined with herbicides at the time of application, aiming to facilitate the combined control of pests, diseases, and weeds. However, the damage to microbial inoculants increases with the number of combined chemicals. As previously mentioned, Standak^® Top, one of the most used for treatments of seeds in several countries, especially for soybean, is composed of two fungicides and one insecticide; it affects the survival of B. japonicum and, especially, B. elkanii cells, with a drastic decrease verified after 7 days of contact ^[25]. It is worth mentioning that pre-inoculated soybean seeds have been in contact with Standak^® Top for up to 90 days, often resulting in zero recoveries of rhizobial cells from seeds ^[63].

Another interesting observation in the study by Rodrigues et al. ^[25] was the changes in colony morphology, smaller with the increase of the exposure to the pesticide. However, regular colonies were recovered after the bacteria were grown under optimal conditions, indicating an adaptive mechanism to the stressful conditions when exposed to the pesticides.

Santos et al. ^[26] evaluated the compatibility of Standak^® Top with A. brasilense strains Ab-V5 and Ab-V6. First, differences were observed between strains, with lower tolerance of Ab-V5, so that after 24 h of exposure the recovery of viable cells dropped from 4.56 × 10⁵ to 4.37 × 10² cells seed⁻¹. In a greenhouse experiment with the combined strains, Standak^® Top decreased the number of lateral roots and root hairs and resulted in shorter root hair length.

Pereira et al. ^[28] also reported the mortality of A. brasilense strains Ab-V5 and Ab-V6 in maize seeds treated with a mixture of fungicides and insecticides (metalaxyl-M + fludioxonil + thiamethoxam + abamectin). The cell survival rate after 12 h of inoculation of seeds treated with the pesticide was only 13.56%, whereas in untreated seeds it was 65.87%.