The common bean crop (Phaseolus vulgaris) is one of the main crops in Sinaloa, Mexico. Its production is affected by a diversity of factors, mainly fungal diseases such as the southern blight Sclerotium rolfsii (teleomorph Athelia rolfsii), which causes root and stem rot. S. rolfsii is a polyphagous pathogen reported in tropical and subtropical areas, related to significant economic losses (^{Gholami et al., 2019}). Chemical treatments are used to control the disease, although efforts are required to implement eco-friendly strategies for the control of the disease.

Soil microorganisms have been reported with the potential to control S. rolfsii in some crops, although studies performed on the common bean crop are scarce. ^{Volpiano et al. (2018}) reported that Rhizobium spp. (SEMIA 439 and SEMIA 4088) reduce the incidence 18.3 and 14.5%. However, the use of native strains is recommended for the biological control of the pathogen, since they present tolerance to the local environmental conditions and a higher potential in the control of the disease. Thus, the objective of the study was to select bacterial strains from the common bean crop rhizosphere collected in the Municipalities of Ahome, Guasave and Angostura, Sinaloa, Mexico during the 2018-2019 growing season and determine their potential for the control of S. rolfsii.

Bacterial isolation. Samples were taken from common bean plants in a vegetative state from three Municipalities of Sinaloa. Soil samples were taken from the rhizospheres of asymptomatic plants. Four locations were included (with 10 fields each) in each of the Municipalities (Ahome, Guasave and Angostura), for a total of 12 compound samples. In the laboratory, 5 g of soil samples were mixed with 50 mL of a sterile saline solution at 0.85% to carry out serial dilutions up to 10^-4; next, 100 µL of the last two dilutions were distributed in triplicate in Petri dishes with nutrient agar (AN, BD Bioxon^®) and the dishes were incubated at 25 °C for 24 h. The colonies were purified based on their color and shape, and stored at -70 °C in 15% glycerol.

Isolation and identification of S. rolfsii. The S. rolfsii isolates were obtained from sclerotia present in the common bean plants in the fields in the Municipality of Ahome, Sinaloa; they were disinfested, seeded and purified in PDA medium. The pathogenicity of the S. rolfsii isolates used in this study was confirmed under greenhouse conditions with the inoculation in common bean seeds; the symptoms were corroborated 13 days after the inoculation and re-isolation of the fungus. The pathogen was identified preliminarily by carrying out morphological studies on the hyphae, as well as colony and sclerotia morphology (^{Nandi et al., 2017}). The identity was confirmed with the amplification and sequencing of a fragment of the ribosomal region, with oligonucleotides ITS4/ITS5 (5’-TCCTCCGCTTATTGATATGC-3’/5’-GGAAGTAAAAGTCGTAACAAGG-3’) which amplify ~ 587 pb (^{White et al., 1990}).

Antagonism test of bacteria against S. rolfsii. The in vitro evaluation of the antagonism of bacteria was carried out using the dual culture technique in Petri dishes with PDA (^{Yánez-Mendizábal et al., 2011}); Petri dishes with PDA containing only S. rolfsii were used as a control. The dishes were incubated at 25 °C until the growth of the mycelia reached a diameter of 4 cm. The percentage of inhibition of radial growth (PIRG) was calculated following the procedures described by ^{Kumar et al. (2011}), with the formula PIRG = (R1-R2 / R1) multiplied by 100, where R1 is the radial growth of the pathogen and R2 is the radial growth of the pathogen that interacts with the bacteria.

Hemolysis tests. Hemolysis tests were carried out (^{Forbes et al., 2002}) on the bacterial isolates that displayed percentages of inhibition higher than 40%. Agar-blood culture media dishes were used, in which holes with a diameter of 0.5 cm were made. The bacterial isolates were grown in 5 mL of Luria Bertani (LB) medium at 30 °C on a shaker at 250 rpm for 24 h. Out of each bacterial isolates, 1 mL was centrifuged at 13,000 rpm for 5 minutes, 100 µL of the supernatant were taken and placed in the holes of the dishes containing agar-blood. The dishes were incubated for 24 h at 37 °C. The criteria for the hemolysis test were as follows: alpha hemolysis (α-hemolysis) or partial hemolysis when a dark green halo appears in the culture medium; beta hemolysis (β-hemolysis) or total hemolysis when a pale halo appears in the culture medium as an effect of the total lysis of the erythrocytes; and gamma hemolysis (γ-hemolysis) or non-hemolytic, is the culture medium presents no halo (Forbes et al., 2002).

Evaluation of antagonistic bacteria in planta. For this evaluation, the three bacterial isolates with the highest percentage of inhibition, non-hemolytic, were chosen. A 0.5 cm PDA plug with active fungal growth was placed in pots with sterile vermiculite/sand (1:1, v/v), on top of which the previously disinfested common bean seed with 0.5% NaClO was placed. One mL of bacterial suspension (COHUI 06, SANMI 02 or GLS 06) in LB broth, incubated for 24 h (Optical density at 595 nm = 1.0) was added to each seed in the substrate. The control seeds were treated with 1 mL in LB broth without bacteria. Ten plants were used per treatment, which were randomly distributed and kept under controlled conditions (8 h of light/16 h of darkness, 25 °C) for 13 days.

The biological effectiveness of the treatments in the control of the disease was determined based on the incidence (number of plants with symptoms divided by the total number of plants evaluated in each treatment times 100) (Table 2). The severity index was also evaluated with a scale of 1 to 4 (^{Moreno and Acevedo, 2002}),where: 0 = no damage; 1= 1-25% of yellow leaves and initial wilting of the shoot; 2= 26-50% of yellowing leaves and the start of the stem death; 3= 51-75% of necrotic yellow leaves and progressive death of stem, and 4= 76-100% of necrotic tissue, mycelial growth and sclerotia. In addition, the variables of plant height, and fresh and dry weights of roots and the aerial section were registered (data not shown). The data were subject to the Shapiro-Wilk normality test; then they were included in the Kruskal-Wallis and Mann-Whitney non-parametric tests with a value of p<0.05.

Molecular identification. The bacterial isolations with the highest percentage of inhibition were molecularly identified; genomic DNA, extracted with DNAzol^® (Invitrogen, Cat. No. 10503-027), was used as a template to amplify a fragment of ~1400 pb of the 16S rDNA region by PCR, with the oligos F₂C/C (5´-AGA GTT TGA TCA TGG CTC-3´ and 5´- ACG GGC GGT GTG TAC-3´) (^{Shi et al., 1997}). The PCR products were visualized by electrophoresis through an 0.8% (w/v) agarose gel. Afterwards, the PCR products were purified using the QIAquick^® PCR Purification kit (QIAGEN, Cat. No. 28104); the purified products were sequenced in both directions. The sequences obtained were compared with the data bank of the National Center for Biotechnology Information (NCBI; http://www.ncbi.gov), in the BLASTn platform to determine the similarity of the obtained sequences with the organisms that displayed an identity higher than 90%. The phylogenetic tree was created using the software MEGA X (^{Kumar et al., 2018}) with the neighbor-joining method (^{Saitou and Nei, 1987}) and the two-parameter substitution model (^{Kimura, 1980}). The solidness of the topology was evaluated using the bootstrap test with 1000 replications.

A total of 65 bacteria were isolated from the rhizosphere of the common bean plants, with a percentage of inhibition in vitro against S. rolfsii that ranged from 2.5 to 65% (Table 1). Only 14 isolations displayed a percentage of inhibition higher than 40%. In the hemolysis tests (^{Forbes et al., 2002}), six displayed total hemolysis (β), two displayed partial hemolysis (α) and six were non-hemolytic (γ) (Table 1).

The three, non-hemolytic, with the highest in vitro inhibition, were Acinetobacter pittii (COHUI06), Pseudomonas putida (SANMI02) and Burkholderia sp. (GLS06) (Figure 1). For the in planta evaluation, the plants produced from seeds treated with Burkholderia sp. displayed an incidence and severity of 40 and 50%, respectively (Table 2), whereas in the plants in which P. putida and A. pittii were used, even though they did display inhibition of growth in vitro, these did not control the disease under greenhouse conditions, since the severity indices were similar to those of the control plants inoculated only with the pathogen (Table 2). This suggests that the results of the effects of antagonism observed in vitro do not always correlate to the results observed in planta, as in with Pseudomonas fluorescentes against Gaeumannomyces graminis var. tritici, where an effect was observed in vitro, although its efficiency in vivo did not manifest itself (^{Elsherif and Grossmann, 1994}).

The potential of the microorganisms to produce antimicrobial compounds in vitro is not always correlated with results in situ, since the availability of nutrients for bacteria is greater in the artificial medium than in the rhizosphere (^{Köhl et al., 2019}). Therefore, the metabolites in vitro may not be expressed in situ or have a reduced effect (^{Hennessy et al., 2017}). Additionally, the in vitro antagonism experiments without any contact with the plant exclude other modes of bacterial action, such as the ability to induce systemic resistance (Köhl et al., 2019), and therefore, we cannot discard the possibility that isolations that showed no antagonism in vitro can induce other action mechanisms against S. rolfsii in the plant.

Species of both Pseudomonas and Acinetobacter genera have been proposed as plant growth promoters due to their ability to solubilize phosphate, produce indole acetic acid, fixate nitrogen and produce siderophores (^{Daur et al., 2018}; ^{Qessaoui et al., 2019}). A. pittii JD-14 is efficient for increasing growth, fresh and dry weight in alfalfa (Daur et al., 2018); the strain evaluated in the present study promoted the increase of the dry weight of the root (data not shown). P. putida showed no growth promotion in the common bean variety included (data not shown).

Figure 1 Phylogenetic relation of isolations COHUI 06, SANMI 02 and GLS 06, based on the sequence of the gene 16S rRNA and compared with sequences from the GenBank data base in the NCBI. 

Species of Acinetobacter have been reported as antagonists of fungal pathogens. Acinetobacter sp. (5H2) and A. lwoffii (A5) inhibited 69.4 and 62.7% respectively, and in vitro growth of Corynespora cassiicola and significantly reduced the severity of the pathogen in Hibiscus sabdariffa (^{Patricio-Hernández et al., 2020}). The species A. pittii evaluated in this investigation was not efficient in controlling S. rolfsii in planta.

The Burkholderia has a complex taxonomy and is a dominant genus in the microbiota of the rhizosphere. The group B. cepacia includes plant growth-promoting species and phytopathogen control agents (^{Eberl y Vandamme, 2016}; ^{Rojas-Rojas et al., 2019}); B. contaminans UFLA02-27 has been isolated from common bean plants and evaluated. This strain can solubilize phosphate, fixes nitrogen, and it antagonizes Fusarium oxysporum f. sp. phaseoli (^{Da Silva et al., 2012}). Species of this group have a biotechnological potential in agriculture, since they produce a variety of hydrolytic enzymes and bioactive compounds, but their use is restricted, due to the risks the pose as opportunist pathogens in immunocompromised people (Eberl and Vandamme, 2016; Rojas-Rojas et al., 2019; ^{Espinosa-Victoria et al., 2020}).

Burkholderia sp. (strain GLS06), which resulted to be non-hemolytic (Table 1), shares around 95% of homology with sequences of gene 16S of species of the group of B. cepacia (GenBank: MG571668.1, MH022722.1, FJ907187.1); the homology of gene 16S of the species of group B. cepacia is high (>97.5%) (^{Da Silva et al., 2012}). Our analysis limits the isolate isolation GLS06 at genus level; concatenated analysis is required, with different markers to determine the species.

Burkholderia sp. (GLS06), isolated from rhizosphere samples obtained in Guasave, Sinaloa, Mexico, displayed potential in the control of southern blight of common bean caused by S. rolfsii, which justifies future studies at an intermediate level and on the field to determine its potential as a biocontrol agent in an integrated disease management system, in strict adherence to ethical criteria related to care for human health and the environment.