Comparative Study of Inoculation Methods to Determine the Aggressiveness of Xanthomonas citri pv. glycines Isolates and to Evaluate the Reaction of Soybean Cultivars to Bacterial Pustule

: Bacterial pustule caused by Xanthomonas citri pv. glycines (syn. X. axonopodis pv. glycines ) is an important bacterial disease for soybean production worldwide. Currently, the planting of resistant soybean cultivars is the main measure for control of the disease. Therefore, the objectives of this study were to determine the aggressiveness of different X. citri pv. glycines isolates and to establish a new protocol to evaluate the reaction of soybean cultivars to bacterial pustule under greenhouse conditions. The molecular analysis based on the entire 16S rRNA gene sequence revealed that the isolates GDM 01, GDM 02, 333 and 87-2 belong to the X. citri pv. glycines pathovar with 100% identity with the pathotype strain K29 of this bacterium. Differences in aggressiveness were observed among the X. citri pv. glycines isolates. There were differences in the reaction of the soybean cultivars to bacterial pustule, depending on the inoculation method. The isolates 333, GDM 01 and GDM 02 were the most aggressive, while 87-2 and GDM 03 were the least aggressive ones. The best conditions for inoculation of the soybean plants were in the V3 vegetative growth stage with the inoculum concentration of 10 8 CFU mL − 1 , and moist chamber for 24 h before and after inoculation. Spraying without injury was the quickest and most practical inoculation method to screen soybean genotypes for resistance to bacterial pustule in breeding programs. Further, this method of inoculation closely simulates the natural bacterial infection under ﬁeld conditions and produces typical symptoms of the disease. The standard brush inoculation method overestimated disease severity, so we do not recommend this method for the evaluation of bacterial pustule resistance in soybean breeding programs.


Introduction
Brazil is the largest soybean (Glycine max L. Merrill) producer in the world, with 154.81 million tons produced in 43.8 million hectares in the 2022/2023 growing season [1]. However, bacterial pustule, a disease caused by Xanthomonas citri pv. glycines (Nakano 1919) [2] (syn. Xanthomonas axonopodis pv. glycines), has been considered an important bacterial threat for soybean production worldwide [3,4]. Nevertheless, yield losses due to the disease are still not well reported [4][5][6]. Although there are no official records of the occurrence of bacterial pustule in Brazil, the presence of the disease has been observed in soybean fields in several regions of the State of Paraná [7].
Bacterial pustule symptoms usually appear on the leaves as small yellowish-green spots with a raised straw yellow center that quickly dies off [8]. The pustules are usually formed in the lower side of the leaves [9].
Currently, the planting of resistant soybean cultivars is the main measure for bacterial pustule control and has been a major issue in breeding programs [10,11]. Further, the resistance to the disease may be conferred by the rxp gene, which is found in most commercial soybean cultivars [12,13]. Although most of the soybean cultivars planted in Brazil are resistant to bacterial pustule [14,15] observed that the number of commercial soybean cultivars resistant to the disease has been decreasing since 2011. Thus, new resistance sources to bacterial pustule should be sought, as the planting of susceptible soybean genotypes at a large scale may trigger epidemics of the disease.
Despite the importance of bacterial pustule for soybean production, there are no studies on the genetic variability of the bacterium in Brazil and different races of the pathogen may be present [14]. Genetic diversity in X. citri pv. glycines populations has been reported elsewhere [4], as well as the presence of physiological races and differences in aggressiveness of the bacterium based on the reaction of different soybean cultivars [11,16].
The establishment of improved disease evaluation protocols for bacterial pustule has been strongly recommended to increase the efficiency and accuracy of selection for resistance to the disease in breeding programs [11,17]. Furthermore, different aspects in plant disease studies may require efficient, practical and low-cost inoculation methods that can be reproducible in both greenhouse and field trials [18]. In addition, the method of inoculation should allow quantitative assessments of the disease [19]. The efficiency of inoculation methods has been reported in studies with several plant pathogenic bacteria. Differences in disease expression levels have also been reported with these bacterial diseases, with consequent establishment of the susceptibility of the plant genotypes to these diseases [20][21][22][23].
In addition to the inoculation method, the bacterial strain, the inoculum concentration, the stage of development of the plant and the environmental conditions certainly play important roles in the reaction of the plant to the disease. The conditions for reproducible disease development, including pathogen infection and colonization, are also important to establish a protocol for efficient evaluations of plant germplasms under artificial inoculation [24].
In soybean breeding programs, a revision of the screening protocols for identification of cultivars resistant to bacterial pustule under controlled conditions is needed. Further, the severity of the disease in a given soybean cultivar in artificially inoculated plants under greenhouse conditions should correlate with the performance observed in the field under natural conditions of disease development. For this purpose, the choice of the inoculation method is extremely important. Furthermore, the method should be practical, quick and low-cost.
In some countries such as Brazil, there are requirements for soybean cultivar registration and protection, including the characteristics of distinguishability, homogeneity and stability (DHE) of the cultivar (www.gov.br/agricultura/pt-br/assuntos/insumosagropecuarios/insumosagricolas/protecao-de-cultivar/agricolas, accessed on 10 May 2023) [25]. Reaction to some diseases is also a requirement [25]. For the bacterial pustule, the brushing inoculation method [15] has been adopted to determine the reaction of soybean cultivars to the disease in trials carried out under greenhouse conditions [25]. However, differences have been observed in the level of resistance of the soybean cultivars to bacterial pustule reported for official registration based on evaluations carried out using the recommended protocol [15] and the field performance of the same cultivar with the disease [26]. Further, the recommended protocol for cultivar registration may not reproduce well the conditions for the bacterial pustule development as observed in the field. The proposed screening protocol for X. citri pv. glycines [25] may be too drastic, overcoming the resistance of the soybean genotype that is expressed under natural conditions in the field.
This study aimed (i) to characterize and to determine the aggressiveness and pathogenic variability of different X. citri pv. glycines isolates; (ii) to evaluate different inoculation methods and conditions to determine the reaction of soybean cultivars to bacterial pustule under greenhouse conditions, including the inoculum concentration, plant growth stage at inoculation and environmental conditions before and after inoculation; and (iii) to characterize the reaction of 16 commercial soybean cultivars to bacterial pustule based on a proposed protocol in comparison with the reaction reported for the official registration of these same cultivars.

Bacterial Isolates and Pathogenicity Test
The study was carried out in the laboratory and greenhouse facilities of the Don Mario Group (GDM Seeds) in Cambé, PR, Brazil. The identification, source and year of isolation of the X. citri pv. glycines isolates included in this study are presented in Table 1. The isolates GDM 01, GDM 02 and GDM 03 were established from soybean fields of the genotype NT 12 2042 with typical bacterial pustule symptoms. The bacterial isolates were initially plated on a nutrient agar (NA) medium and maintained at 28 • C for 48 h. Then, they were transferred to new NA medium plates and kept at 28 • C for 48 h to obtain the bacterial inoculum. For the pathogenicity test, seeds of the cultivar NT 12 2042, susceptible to bacterial pustule, were sown in 1 L plastic pots containing autoclaved clay soil, and thinned to three plants per pot after 10 days. The pots were kept in a semi-climatized greenhouse under natural light at the maximum and minimum temperatures of 30 and 22 • C, respectively, and relative humidity ranging from 60 up to 80%. The soybean plants were inoculated in the V3 vegetative stage, with three fully developed trifoliate leaves, approximately 20 days after sowing, by spraying the foliage with a bacterial suspension adjusted to 1 × 10 8 UFC mL −1 . After inoculation, the plants were kept in a moist chamber condition for 24 h. After removal from the moist chamber condition, the plants were maintained under automatic misting every 20 min to ensure a relative humidity of 80%.

Molecular Characterization of the Bacterial Isolates
The X. citri pv. glycines isolates were subjected to a molecular analysis by sequencing the 16S rRNA gene for genetic identification at the species and pathovar levels [27]. Thẽ 1500 bp 16S rRNA gene comprises nine variable regions interspersed throughout the highly conserved 16S sequence [27]. The isolates were grown in a nutrient broth medium at 200 rpm and 28 • C for 16 h. A pellet was obtained from 1 mL of the bacterial suspension by centrifugation at 14,000 rpm for 10 min. Total genomic DNA was extracted using the Qiagen DNA extraction kit protocol (Qiagen, Hilden, Germany). The DNA quality was examined by gel electrophoresis and quantified using the NanoDrop DN-1000 (Thermo Scientific, Waltham, MA, USA). The DNA was subjected to PCR with the universal primers fD1_ CCGAATTCGTCGACAAC (forward) and rD1_ AGAGTTGATCCTGGCTCAG (reverse) to amplify a 1500 bp fragment of the 16S rRNA gene [28]. Each PCR reaction was performed in a final volume of 25 µL, containing 1 µL genomic DNA; 2.5 µL 10× PCR buffer; 1 µL of each primer at 10 µM; 0.8 µL MgCl (50 mM); 1 µL dNTP (5 mM); 0.1 µL GoTaq DNA polymerase (Promega Corp., Madison, WI, USA) and 17.6 µL sterile MiliQ water. The optimized cycling program consisted of initial denaturation at 94 • C for 3 min, followed by 30 cycles of denaturation at 94 • C for 50 s, annealing at 62.3 • C for 50 s and extension at 72 • C for 1 min and 45 s. A final extension step was carried out at 72 • C for 7 min. The amplified DNA was subjected to horizontal electrophoresis at 6 V/cm 2 in 1% agarose gel (w/v), and stained with ethidium bromide and a TAE buffer (0.04 M Tris-acetate, 0.001 M EDTA, pH 8.0). The gel was photographed under ultraviolet light (UV). The PCR products visualized in the agarose gel were purified using the PureLink Kit (Quick Gel Extraction Kit-Invitrogen, Carlsbad, CA, USA) and run in an automated sequencer using the Sanger method (Applied Biosystems, Foster City, CA, USA) in the Universidade Estadual Paulista (UNESP), Jaboticabal, SP, Brazil.

Assessment of the Aggressiveness of the Bacterial Isolates
The aggressiveness of the X. citri pv. glycines isolates was assessed on the soybean genotypes of NT 12 2042 (susceptible), Peking (susceptible), VMax RR (moderately susceptible) and BRS 316 RR (resistant) ( Table 2). The plants were inoculated in the vegetative stage V3 by the spray inoculation method described previously with a bacterial suspension adjusted to 10 8 CFU mL −1 . The experiments were carried out in a semi-climatized greenhouse under natural light with the mean minimum and maximum temperatures of 28 • C and 30 • C, respectively, and relative humidity ranging from 60 up to 80%. Disease severity evaluations were performed on the first and second trifoliate leaves of the soybean plants, starting on the 4th day after inoculation (DAI), and were carried out every 3 days until the 19th DAI. Disease severity was rated as described earlier for the spray inoculation method and the area under the disease progress curve (AUDPC) was calculated for each isolate based on the severity data. The AUDPC was calculated using the equation proposed by [33]. The values were normalized by dividing the AUDPC by the total experimental period based on the number of days [34], as follows: where X i is the proportion of diseased host tissue at the ith evaluation, t i is the time in days at the ith evaluation and n represents the total number of observations. The experimental design was a randomized block design with five replications. The experiment was performed twice.

Evaluation of Conditions for Inoculation of X. citri pv. glycines
The soybean line NT 12 2042, susceptible to bacterial pustule, and the isolate 333 of X. citri pv. glycines were used in the experiments. Three moist chamber periods before and after the bacterial inoculation were evaluated: 12, 24 and 48 h. The moist chamber consisted of metal trays of 2.89 × 1.10 m, with a 10 cm height covered with transparent plastic film supported by a metal arch. The trays were halfway filled with water to maintain a high-humidity environment.
Three phenological growth stages were tested for each of the moist chamber periods, V2 with two fully developed trifoliate leaves at 15 days after sowing (DAS), V3 with three fully developed trifoliate leaves at 20 DAS and V4 with four fully developed trifoliate leaves at 25 DAS, with two concentrations of the bacterial inoculum, 10 6 and 10 8 CFU mL −1 . The inoculum concentrations were adjusted in a GENESYS™ (Thermo Fisher Scientific, Waltham, MA, USA) spectrophotometer at O.D. 600 : = 0.2. The bacterial suspensions were prepared in sterile distilled water and sprayed on the abaxial and adaxial leaf surfaces of the soybean plants [35]. Control plants were sprayed with distilled water. The minimum and maximum temperatures in the growth chamber were maintained at 26 to 29 • C, respectively, until disease evaluation. Bacterial pustule severity was determined based on the percentage of the affected leaf area (% A.A.) 15 DAI.
The experimental design was a randomized block design with five replications. Two experiments were carried out to establish the best conditions for inoculation of Xanthomonas citri pv. glycines. However, in the second experiment, only the inoculum concentration of 10 8 CFU mL −1 was included. Under this inoculum concentration, the soybean plants had the highest bacterial pustule severity in all tested moist chamber periods and plant growth stages in the first experiment. After inoculation, the plants were maintained at the minimum and maximum temperatures of 27 and 29 • C, respectively.

Evaluation of Methods for Inoculation of X. citri pv. glycines in Soybean Plants
The isolate 333 of X. citri pv. glycines (Table 1) was used in all experiments. Soybean genotypes with different levels of susceptibility to bacterial pustule were included in the study ( Table 2).
Four inoculation methods were evaluated, i.e., leaf piercing, leaf brushing, leaf cutting and suspension spraying. The plants were inoculated in the vegetative stage V3 (20 DAS) with a bacterial suspension adjusted to 10 8 CFU mL −1 for all inoculation methods.
The leaf piercing inoculation was based on the method described by Silva [23]. The device for inoculation consisted of a 2 cm diameter plastic cone containing five equidistant needles. A sponge moistened in the bacterial suspension was placed on the abaxial surface of the leaflet of the first fully developed trifoliate leaf of the soybean plant and the inoculation device was pressed on the adaxial surface to pierce the leaflet. Disease severity was assessed using a modified scale proposed for the common bacterial blight in a common bean [36]. Disease was rated as follows: 1, no symptoms; 2, mild chlorosis around the inoculated area; 3, chlorosis and presence of few bacterial pustules in the inoculated area; and 4, chlorosis, high incidence of bacterial pustules and necrosis around the inoculated points.
The leaf cutting inoculation was based on the method proposed for X. phaseoli pv. phaseoli [37]. Five 1 cm long cuts were made with a pair of scissors at the edge of the leaflet of the first fully developed trifoliate leaf of the soybean plant. The sterilized scissors were immersed in the bacterial suspension before each cut. Disease severity was evaluated in the inoculated leaflet using the scale proposed by Pastor-Corrales [37,38] (apud [25]), with minor modifications: 1, no symptoms; 2, mild chlorosis around the cuts; 3, mild chlorosis and presence of necrosis around the cuts; and 4, well-developed chlorosis around the cuts to a distance of approximately 1 mm and necrosis.
The leaf brushing method was the same described in the protocol for characterization of bacterial pustules recommended by Brasil [25]. A sterilized brush was immersed in the bacterial suspension and passed on the abaxial surface of the leaflets of the first fully developed trifoliate leaf of the soybean plant. For the disease evaluation, the three-level scale established by Brasil [25] was used: 1, resistant plants (no pustule, chlorosis or necrosis symptoms); 2, moderately resistant plants (no pustules or slight chlorosis in the area injured by the brush); and 3, susceptible plants (pustules in the area damaged by the brush, well-developed chlorosis and necrosis).
The spray inoculation method followed the procedure described by Kronka [35], using a 500 mL hand-spraying bottle. The inoculum was sprayed on the abaxial and adaxial leaflet surfaces until runoff, at a distance of approximately 2 cm. Bacterial pustule severity was determined by estimating the percentage of the affected leaf area (% A.A.).
Additionally, a control treatment was included for all soybean genotypes and inoculation methods using only sterile distilled water in order to determine any possible mechanical damage that could lead to leaf injuries. All plants were maintained in a moist chamber for 24 h before and after inoculation. Further, the disease severity evaluations were initiated when the first symptoms showed up in the susceptible genotype, NT 12 2042. The practicality of each inoculation method was also evaluated, based on the time consumed and the bacterial suspension volume required for inoculation.
The experimental design was a randomized block design with five replications. The experiment was carried out twice. The mean minimum and maximum temperatures during the experimental period were, respectively, 26 • C and 29 • C in Experiment 1 and 28 • C and 30 • C in Experiment 2.

Comparison of the Spray Suspension and the Leaf Brushing Methods for Screening for Resistance of Soybean Cultivars to Bacterial Pustule
A total of 18 soybean genotypes, including commercial cultivars registered in the RNC (https://sistemas.agricultura.gov.br/snpc/cultivarweb/cultivares_registradas.php, accessed on 10 March 2022) [39], with different reactions to bacterial pustule were included in the study to compare the spray suspension and the leaf brushing methods for evaluation of the reaction to the disease ( Table 2). The line NT 12 2042 and the cultivar BRS 316 RR were included in the study as susceptible and resistant standards, respectively.
The soybean plants were inoculated by the spray suspension and the leaf brushing methods with the isolate 333 of X. citri pv. glycines in the vegetative stage V3, as described previously, with a bacterial suspension adjusted to 10 8 CFU mL −1 . A control treatment sprayed with sterile distilled water was also included for each method. Disease evaluations were carried out at 16 DAI. Plants inoculated by the spray suspension method were evaluated based on the percentage of the diseased leaf area (% A.A.) and those inoculated with the leaf brushing method were evaluated using the scale proposed by Brasil [25]. The mean minimum and maximum temperatures during the experimental period were 28 • C and 30 • C, respectively, in both experiments.
The experimental design was a randomized block design with three replications. The experiment was performed twice.

Statistical Analyses
Data were subjected to the analysis of variance and means were compared by the Tukey test at 5% probability. Data that did not fit the assumptions for the analysis of variance were Box-Cox transformed. Homogeneity and normality were achieved before the transformation of data with Box-Cox. Data were analyzed using the ExpDes package in R Studio statistical software version 1.1.463 [40].

Molecular Characterization of the Xanthomonas citri pv. glycines Isolates
The~1500 bp sequences of the 16S rRNA gene of the isolates GDM 01, GDM 02, 333 and 87-2 revealed similarities of 100% with the sequence of the strain K29, a pathotype of X. citri pv. glycines (syn. X. axonopodis pv. glycines), deposited in the GenBank ® database under the accession number OL979613.1 (https://www.ncbi.nlm.nih.gov/genbank/, accessed on 10 March 2022). Further, the phylogenetic tree reconstructed with the sequence of the 16S rRNA gene confirmed that the isolates GDM 01, GDM 02, 333 and 87-2 are likely to belong to the X. citri pv. glycines (syn. X. axonopodis pv. glycines) pathovar, as they are in the same branch of the phylogenetic tree of the strain K29 (OL979613.1) (Figure 1). On the other hand, it was not possible to perform the molecular identification of the isolate GDM 03 due to the poor quality of the 16S rRNA sequence. The 16S rRNA sequences of the bacterial isolates were deposited in GenBank under the accession numbers OL415538, OL415539, OL415537 and OL415540 for the isolates GDM01, GDM02, 333 and 87-2, respectively.

Assessment of the Aggressiveness of the Isolates of Xanthomonas citri pv. glycines
All tested isolates of X. citri pv. glycines produced typical symptoms of bacterial pustule in all three genotypes of soybean (Table 3, Figure 2). The disease severity and disease progress curve for the genotypes NT 12 2042 and Peking revealed differences in aggressiveness of the X. citr pv. glycines isolates, particularly for the isolates 333, GDM 01 and

Assessment of the Aggressiveness of the Isolates of Xanthomonas citri pv. glycines
All tested isolates of X. citri pv. glycines produced typical symptoms of bacterial pustule in all three genotypes of soybean (Table 3, Figure 2). The disease severity and disease progress curve for the genotypes NT 12 2042 and Peking revealed differences in aggressiveness of the X. citr pv. glycines isolates, particularly for the isolates 333, GDM 01 and GDM 02 in comparison to the other two isolates, 87-2 and GDM 03 (Table 3, Figure 2). In both experiments, disease symptoms developed quickly in the first days after inoculation in the plants of the genotypes NT 12 2042 and Peking, increasing exponentially until approximately 13 DAI (Figure 2). A factorial analysis (soybean genotype x bacterial isolate) for the last severity evaluation at 19 DAI and the AUDPC revealed a significant interaction between these factors in both experiments ( Table 2). Bacterial pustule severity was higher in the NT 12 2042 and Peking plants inoculated with the isolates 333, GDM 01 and GDM 02 of X. citri pv. glycines compared to those inoculated with the isolates 87-2 and GDM 03 ( Table 3). On the other hand, no differences were observed in bacterial pustule severity in the Vmax RR cultivar when inoculated with any of the isolates (Table 3). The first symptoms of bacterial pustule showed up in the susceptible genotypes NT 12 2042 and Peking 4 DAI (Figure 2A,B). In the resistant cultivar Vmax RR, the first symptoms were observed in the plants inoculated with the aggressive isolate 333 seven DAI ( Figure 2C,D). In plants of Vmax RR inoculated with the isolates GDM 01 and GDM 02, the symptoms appeared 10 DAI, while in those inoculated with the isolates 87-2 and GDM 03, they showed up 13 DAI ( Figure 2D). These reactions of the cultivar Vmax RR when inoculated with the isolates of X. citri pv. glycines were observed in both experiments.
In the evaluation 16 DAI, leaf drop was observed in some plants, mainly for the ones inoculated with the isolates 333, GDM 01 and GDM 02. The leaf drop was more severe in the cultivar Peking and less intense in the genotype NT 12 2042. No leaf drop was observed in the plants of the cultivar Vmax RR.

Evaluation of Conditions for Inoculation of Xanthomonas citri pv. glycines
In preliminary studies, the inoculation of X. citri pv. glycines with bacterial suspension at the concentration of 10 8 CFU mL −1 resulted in the highest number of bacterial pustule lesions for all tested soybean plant growth stages and moist chamber periods. Therefore, all experiments to determine the best phenological plant stage and moist chamber period were performed using this inoculum concentration. Based on the AUDPC of the bacterial pustule, the reaction of the soybean cultivars also varied according to the isolate of X. citri pv. glycines (Table 3). The highest AUDPC values were observed on the cultivar Peking inoculated with the isolates 333, GDM 01 and GDM 02 (Table 3). In contrast, the lowest AUDPC on Peking was observed for the isolates 87-2 and GDM 03 (Table 3). In the genotype NT 12 2042, the isolate 333 induced the highest AUDPC, followed by the isolates GDM 01 and GDM 02 (Table 3). Soybean plants inoculated with the isolates 87-2 and GDM 03 had the lowest AUDPC (Table 3). The AUDPC in the genotype Vmax RR did not differ statistically independent of the bacterial isolate in both experiments (Table 3).

Evaluation of Conditions for Inoculation of Xanthomonas citri pv. glycines
In preliminary studies, the inoculation of X. citri pv. glycines with bacterial suspension at the concentration of 10 8 CFU mL −1 resulted in the highest number of bacterial pustule lesions for all tested soybean plant growth stages and moist chamber periods. Therefore, all experiments to determine the best phenological plant stage and moist chamber period were performed using this inoculum concentration. Some significant interaction was observed between phenological growth stages of the soybean plants of the line NT 12 2042 and the moist chamber period before and after the inoculation with the X. citri pv. glycines isolate 333 (Table 4). Further, the severity of bacterial pustule in the plants inoculated in V3 and V4 growth stages differed significantly depending on the moist chamber period ( Table 4). The highest disease severity level was always for the soybean plants inoculated in the V4 growth stage, independent of the moist chamber period (Table 3). In contrast, the lowest disease severity level was always for the plants inoculated in the V2 growth stage ( Table 4). The condition which provided the highest disease severity level was for the plants in the V4 growth stage under a moist chamber period of 48 h before and after inoculation. Under these conditions, 76% of the foliar area was affected by the disease (Table 4). On the other hand, the lowest amount of disease was observed for the soybean plants inoculated in the V2 growth stage, independent of the moist chamber period before and after inoculation ( Table 4). The amount of disease in the plants inoculated in this growth stage was too low for soybean genotype screening in breeding programs for resistance to bacterial pustule [41]. In contrast, there was a significant increase in the leaf area affected by the disease in the soybean plants inoculated in V3 and V4 growth stages (Table 4). Further, bacterial pustule severity also differed significantly in regard to the moist chamber period before and after the inoculation for these growth stages ( Table 4). The highest disease severity level was observed for the plants inoculated in the V4 growth stage and maintained in a moist chamber for 48 h before and after the inoculation (Table 4).

Evaluation of Methods for Inoculation of the Soybean Plants
Regardless of the inoculation method, the bacterial pustule severity was significantly higher in the susceptible line NT 12 2042 than in the other soybean cultivars (Table 5). On the other hand, the disease severity in the BMX Potência RR and Vmax RR cultivars was very low while the BRS 316 RR was symptomless when inoculated by the piercing and spraying methods, in both experiments ( Table 5). The reaction of these three genotypes to bacterial pustule did not differ significantly from each other. The same was observed in both experiments in response to the brushing inoculation method (Table 5).
CV (%) 7.0 8.4 6.9 10.7 1 Score: 1, no symptoms (pustules, chlorosis or necrosis); 2, mild chlorosis around the inoculated area; 3, chlorosis and presence of few bacterial pustules in the inoculated area; and 4, chlorosis, high incidence of bacterial pustules and necrosis in the inoculated area. 2 Score: 1, resistant plant (no symptoms of pustules, chlorosis or necrosis); 2, moderately resistant plant (no symptoms of bacterial pustules, with slight chlorosis in the area injured by brushing); and 3, susceptible plant (pustule incidence in the brush-injured area, well-developed chlorosis and necrosis in leaflets). 3 Score: 1, no symptoms (cutting does not induce chlorosis and necrosis); 2, slight chlorosis in response to cutting; 3, mild chlorosis and necrotic spots in response to cutting; and 4, well-developed chlorosis in response to cutting (approximately 1 mm) and necrotic spots. 4 % A.A., percentage of affected leaf area. 5 In each experiment, means followed by the same letter in the column do not differ by the Tukey's test at 5% probability. Original means; the data were transformed by √ (x + 1) for statistical analysis.
In regard to the cutting method, a significant difference in disease severity was observed between the susceptible line NT 12 2042 and the cultivars BMX Potência RR and BRS 316 RR, in the first experiment, while in the second experiment, the line NT 12 2042 differed from all three other cultivars (Table 5). In the first experiment, the disease was more severe in BMX Vmax RR than in BRS 316 RR (Table 5). Furthermore, the lowest disease severity in this experiment was observed for BRS 316 RR (Table 5). On the other hand, BMX Potência RR and BRS 316 RR were symptomless and differed significantly from BMX Vmax and NT 12 2042 in the second experiment (Table 5).
When inoculated by the spray method, significant differences were observed between the susceptible line NT 12 2042 and all the other soybean cultivars, which basically showed no symptoms of bacterial pustule (Table 5). Mild symptoms of the disease were observed only for BMX Vmax in the second experiment (Table 5). Disease severity on the susceptible line NT 12 2042 was 32.6% and 35.4% of the affected leaf area in Experiments 1 and 2, respectively, when X. citri pv. glycines was inoculated by the spraying method (Table 5). On the other hand, the resistant cultivar BRS 316 RR showed no disease symptoms when inoculated by any of the inoculation methods (Table 5).
Variations were observed among the soybean genotypes in response to the method of inoculation. For instance, the susceptible line NT 12 2042 developed bacterial pustules and chlorotic halos in response to the inoculation by piercing, brushing and spraying methods ( Figure 3A,B,D). However, the bacterial pustules were limited to the wounded area when inoculated by brushing, without expansion to the rest of the leaf surface ( Figure 3B). Plants inoculated by the cutting method did not develop typical pustule symptoms; only small chlorotic areas around the wounds could be observed ( Figure 3C). The physical damage caused by the inoculation methods on the soybean leaves was examined to distinguish from the symptoms of bacterial pustule. The soybean plants were subjected to the same procedure of the inoculations with the bacterial pathogen suspension, but instead using sterile water ( Figure 3A-D). It is worthwhile to point out that the brushing method caused major physical damage to the leaf tissue ( Figure 3B). The same is true for the piercing and cutting methods, which also injured the leaf tissue ( Figure 3A,C). In contrast, the spraying method did not cause any physical damage to the leaf tissue ( Figure 3D).

Comparison of the Spraying and the Leaf Brushing Inoculation Methods for Screening for Resistance of Soybean Cultivars to Bacterial Pustule
Eighteen soybean genotypes, including the susceptible NT 12 2042 and the resistant BRS 316 RR standards, were evaluated in this study to experimentally characterize the reaction to bacterial pustule. The isolate 333 of X. citri pv. glycines was inoculated by the spraying and brushing methods. Typical bacterial pustule symptoms were produced in eleven soybean cultivars when inoculated by the spraying method ( Table 6). The highest disease severity was observed in the susceptible standard line NT 12 2042 while no disease symptoms were observed in seven soybean cultivars, including the resistant standard BRS 316 RR, when inoculated by the spraying method (Table 6). Among the commercial cultivars included in this study, the highest disease severity level was observed for the cultivar SYN 1561, which is also considered as a bacterial pustule susceptible cultivar based on the Brasil [25] protocol (Table 6).
C.V (%) 11.1 13.8 Most soybean cultivars tested showed bacterial pustule reactions identical to those listed in the cultivar registry [39], when inoculated by the brushing inoculation method (Table 6). Further, the reaction of these cultivars to the disease in response to the inoculations by brushing and spraying methods was also closely correlated, with a few exceptions. No bacterial pustule symptoms were observed in the resistant cultivars BRS 316 RR (resistant standard), M 6210 IPRO, TMG 2286 IPRO, DM 81I84 IPRO and Domínio IPRO inoculated with both methods, confirming the resistance to the disease (Table 6). On the other hand, the cultivars BR-4 and Desafio RR react as resistant to bacterial pustule when inoculated by the spraying method, but they were considered susceptible and moderately resistant, respectively, to the disease when inoculated by the brushing method, the standard method of inoculation [25].
The inoculation by the brushing method, following the protocol recommended by Brasil [25], confirmed the disease severity level of susceptible (S) for BR-4 and moderately resistant (MR) for Desafio RR cultivars (Table 2). Furthermore, this is the level of reaction of the two cultivars for bacterial pustule provided by the companies holding their registration. However, the same two cultivars showed no disease symptoms and can be classified as resistant when evaluated using the spraying method for bacterial inoculation (Table 6), a reaction to bacterial pustule also observed under natural conditions in the field [25].

Discussion
The molecular characterization of the X. citri pv. glycines (syn. X. axonopodis pv. glycines) isolates was based on the 16S rRNA gene, comprising a sequence of~1500 bp. The 16S rRNA gene has been a main sequence for the bacterial analysis providing taxonomic resolution of bacteria at species and strain levels [27,28,42]. Further, our results strongly confirm that the isolates GDM 01, GDM 02, 333 and 87-2 belong to the X. citri pv. glycines (syn. X. axonopodis pv. glycines) pathovar with 100% identity with the pathotype strain K29 of this bacterium, deposited in the GenBank ® database under the accession number OL979613.1 (https://www.ncbi.nlm.nih.gov/genbank/, accessed on 10 March 2022).
The use of aggressive isolates of X. citri pv. glycines to evaluate soybean genotypes for resistance to bacterial pustule is of upmost importance, as it provides a more rigorous screening and improves the discrimination between resistant and susceptible soybean genotypes [4,11,43]. Kaewnum et al. [44] found significant differences in bacterial pustule severity among different X. citri pv. glycines isolates and concluded that the geographic origin of the isolate did not seem to be related to the aggressiveness of the bacterium. Further, genes homologous to avrBs3 may play a role in the pathogenicity and aggressiveness of X. citri pv. glycines to soybean cultivars [4,45]. These results support that pathogens' variability should be taken into account when evaluating soybean genotypes in breeding programs for resistance.
Despite that the highest disease severity level was observed for the plants inoculated in the V4 growth stage maintained under moist chamber conditions for 48 h before and after the inoculation (Table 4), the inoculation of the plants in the V3 growth stage in combination with 24 h of the moist chamber before and after inoculation was chosen for screening the soybean germplasm for resistance to the disease. Under these conditions, disease severity attained 44% in the soybean cultivars susceptible to bacterial pustule, which has been considered an appropriate level to compare the inoculation methods and to characterize cultivar resistance. In addition, these conditions may reduce the evaluation period in breeding programs. It should also be considered that 48 h under moist chamber conditions before and after inoculation may result in plants with incipient wilting due to the long period under this stressing condition. In view of the objective of characterizing the reaction of the soybean cultivars to bacterial pustule, a less conducive condition for disease development may allow a better expression of the potential of the cultivar for resistance to the disease.
Under favorable environmental conditions, inoculation by spraying confirmed that X. citri pv. glycines is capable of infecting soybean plants by entering through natural openings in the leaves, i.e., stomata, and causing typical symptoms of bacterial pustule. Further, there is no need for injuries in the plant tissues for bacterial infection. In addition, this inoculation method has high disease-causing potential and is suitable for large-scale application, as required in screening in breeding programs [18,37]. The spraying inoculation method also simulates more closely the bacterial infection under natural conditions in the field. As a result, the bacterial pustule symptoms obtained are highly characteristic and have a distribution on the leaves closely to that observed under natural field conditions. Plant pathogenic bacteria may exert mechanical forces to overcome barriers such as the cell wall and plant turgor pressure, though they cannot penetrate intact epidermal cells [46]. Therefore, most of these bacteria can only penetrate the plant tissue through wounds and natural openings. Under low relative humidity, the stomata may remain closed to prevent transpiration and water losses, consequently, reducing the chance for bacterial infection [47]. On the other hand, a longer period in a moist chamber, with high relative humidity, may improve the conditions for leaf wetness and stomatal opening, favoring bacterial entry through stomata. Leaf traits such as the stomatal opening are of upmost importance for a successful pathogen infection, and free water on the leaf surface is essential for the movement and penetration of plant pathogenic bacteria in the host cells through stomata [48][49][50].
The quantitative evaluation of the infection response is among the criteria considered to establish an appropriate method for inoculation of plant pathogens [19]. Though, the cutting method does not seem appropriate for evaluation and quantification of the severity of bacterial pustule in soybean. Although the methods of piercing and brushing were efficient in expressing the disease symptoms, they also injured too severely the leaves of the soybean plants. In the case of brushing, the wounds could easily be confused with the symptoms caused by X. citri pv. glycines infection. Moreover, the symptoms' distribution was quite different from a natural infection by the bacterium. Unlike the spraying method, inoculations involving artificial injuries may concentrate infection in specific areas of the leaf surface, resulting in fewer points with lesions [51]. Bacterial suspension spraying was the only method included in this study that did not involve any kind of plant injury. The results obtained with this method of inoculation provided higher uniformity in the symptoms and in the development of the pustules along the leaf blade, with a distribution similar to that observed under natural infection in the field. Similarly, Weindling et al. [52] tested the efficiency of different inoculation methods to evaluate X. citri pv. malvacearum and they concluded that plant infection by spraying inoculation was the most similar to natural infections.
Torres and Maringoni [53] concluded that leaflet inoculation by multiple needles and cutting of Xanthomonas axonopodis pv. phaseoli in common bean plants were effective in discriminating resistance and susceptibility reactions of genotypes, regardless of the phenological stage of development of the plants. The authors do not report leaf injuries or changes in the expression of symptoms resulting from the methods used. Sharma et al. [51] found that the spraying method delivers 10 7-8 cells of Xanthomonas axonopodis pv. punicae per ml and it proved to be the most efficient method for inducing a bacterial blight in the pomegranate. As in our work, spraying inoculation induced symptoms similar to those resulting from natural infections.
The injuries caused by the brushing method allow the bacterial cells to reach the intercellular spaces of the leaves without the need to overcome possible structural barriers that would hinder their progress under natural infection [54]. Furthermore, criteria such as speed, uniformity of results and ease of execution must be taken into account in breeding programs for evaluation and selection of resistant genotypes [37]. These criteria are unlike to be met by the methods of piercing, brushing and cutting. Further, inoculation methods that cause injuries to the host tissues may break some of the plant natural resistance barriers for the invasion of plant pathogens [55]. In this way, the host plant resistance based on some structural condition that prevents the entering of the bacteria in the host tissue, i.e., morphology of the stomatal ostiole, would be bypassed by inoculation methods that cause alternative entry points for the pathogen, such as wounds that damage the plant tissue. In our study, the methods of piercing and brushing injured too severely the leaves of the soybean plants.
In conclusion, screening of soybean genotypes for bacterial pustule resistance using the spraying method provided consistent results, with the cultivar susceptibility ranking turning the same in different evaluations. The spraying and brushing methods provide similar results for bacterial pustule screening, but in some cases, the brushing method induces overestimation of the bacterial pustule severity, as observed for the Desafio RR and BR-4 cultivars in the comparative study of these two methods with eighteen soybean cultivars. Further, bacterial suspension spraying was the quickest and most practical inoculation method for a quantitative evaluation of bacterial pustule in soybean genotypes. This inoculation method reproduces typical symptoms of the disease under controlled conditions. On the other hand, inoculation by the brushing method can mislead disease symptoms in some genotypes, and thus leads to an erroneous classification of the reaction of the genotypes to bacterial pustule.
The expression of bacterial pustule symptoms in soybean plants inoculated by the spraying method was optimized when the plants were inoculated in the V3 stage, approximately 20 DAS, with a bacterial suspension of approximately 10 8 CFU mL −1 and maintained under moist chamber conditions for 24 h before and 24 h after inoculation. It is also important to control and maintain the temperature and humidity of the greenhouse for reproducibility of the results. Furthermore, the isolates 333, GDM 01 and GDM 02 of X. citri pv. glycines were the most aggressive, while isolates 87-2 and GDM 03 were the least aggressive ones, and the most susceptible soybean genotype to bacterial pustule was the cultivar Peking.