Common bean (Phaseolus vulgaris L.) is one of the main components of the Brazilian population diet. Its grains are excellent sources of protein, calories and minerals. Occurrence of diseases can cause significant damage to the crop, with reduction in grain yield and quality. Anthracnose, caused by the fungus Colletotrichum lindemuthianum (Sacc. & Magnus), can cause losses of up to 100% (Paula-Júnior et al. 2015).

Anthracnose management using resistant cultivars is a strategy that minimizes production costs and reduces damage to the environment. However, the large pathogenic variability of C. lindemuthianum limits disease control and development of new cultivars with durable resistance (Pinto et al. 2012). More than 100 races of C. lindemuthianum have been found in the world, and races 65, 73, 81 and 89 are the most frequent in Brazil (Silva et al. 2007; Pereira et al. 2010; Pinto et al. 2012). In addition, pathogenic variability has been detected within races 65 and 81 in Brazil (Ishikawa et al. 2011, 2013; Costa et al. 2017).

Common bean resistance to anthracnose is predominantly vertical, and a gene-a-gene model has been verified in the C. lindemuthianum-P. vulgaris pathosystem. Twenty resistant genes to anthracnose have been reported in common bean (Ferreira et al. 2013; Zuiderveen et al. 2016). In addition, 12 independent genes have been identified that confer resistance to six different isolates belonging to race 65 of C. lindemuthianum (Costa et al. 2017). Despite the existence of vertical resistance, several genes already identified display a polygenic inheritance of this trait (Costa et al. 2017). Therefore, cultivars with durable resistance should present resistance to the most frequent C. lindemuthianum races in different common bean crop regions.

The reaction of common bean lines and progenies to C. lindemuthianum has been assessed using artificial inoculation of isolates of the most frequent races (Barcelos et al. 2013; Costa et al. 2017). However, when many races are used, a large amount of seeds is necessary, hindering the evaluation of progenies in a breeding program. A strategy could be to use mixtures of different isolates or races. However, literature information about the efficiency of this method for evaluation of anthracnose resistance in common bean is scarce (Chilipa et al. 2016), and there is no information about pathogen virulence when mixtures of isolates are used.

In the present study, anthracnose severity on common bean was evaluated after inoculation of C. lindemuthianum races 65, 73 and 81 individually and also using mixtures of isolates. Implications for common bean breeding to obtain cultivars with resistance to anthracnose are discussed.

Three isolates of C. lindemuthianum were used, which were previously identified as race 65 (differential cultivars Michelite and Mexico 222 are susceptible), race 73 (differential cultivars Michelite, Cornell 49.242 and Mexico 222 are susceptible) and race 81 (differential cultivars Michelite, Widusa and Mexico 222 are susceptible). In addition, inoculum obtained from the mixture of the three races (M1) and from mixtures (M2, M3, M4, M5, M6 and M7) obtained after inoculation on cultivars Pérola (susceptible to the three races) and BRSMG Majestoso (resistant to the races 73 and 81) was used.

Each isolate was grown in sterile pods partially immersed in agar-water medium in test tubes, and incubated for 15 days at 22 °C in a BOD in the dark. A conidial suspension was prepared as described by Pinto et al. (2012) and the concentration was adjusted to 1.2 × 106 conidia/mL. A conidial suspension of the mixture of the three races (M1) was obtained by equally mixing the suspensions at the adjusted concentration of conidia of each race.

Inoculations of each isolate individually and of the mixtures were carried out as described by Pinto et al. (2012). Inoculated common bean seedlings were incubated for 48 h at 100% relative humidity with a photoperiod of 12 h and temperature of approximately 20 °C. After incubation, seedlings were transferred to a greenhouse with a temperature of approximately 24 °C and 95% relative humidity.

Two experiments were conducted in a completely randomized design with subdivided plots, with two replicates. A plot consisted of 36 plants. A previous test was carried out by inoculating each race individually and also M1 on plants of the cultivars Pérola and BRSMG Majestoso, as described above. Isolations of the pathogen from lesions on leaves and stems were carried out in order to obtain mixtures of first generation isolates. The isolate obtained from the cultivar Pérola was denominated M2 and the one obtained from cultivar BRSMG Majestoso was denominated M3.

Mixtures of second generation isolates were obtained by isolation from lesions caused by M2 and M3 inoculated on leaves and stems of cultivars Pérola and BRSMG Majestoso. Thus, isolates obtained from the inoculation of M2 on Pérola and BRSMG Majestoso were named M4 and M5, respectively. Likewise, isolates obtained from the inoculation of M3 on Pérola and BRSMG Majestoso were named M6 and M7, respectively.

In the second experiment, inoculations of each isolate and each mixture (M1, M2, M3, M4, M5, M6 and M7) were carried out on plants of the cultivars Pérola and BRSMG Majestoso. In all experiments anthracnose severity was evaluated at three to 10 days after inoculation (dai) using the scale proposed by Schoonhoven and Pastor-Corrales (1987).

The area under the disease progress curve (AUDPC) was estimated using severity mean scores of expression adopted by Shaner and Finney (1977), according to the following equation:

$$ AUDPC={\sum}_{k=1}^n\left[\frac{\left({Y}_{k+1}+{Y}_k\right)}{2}x\left({T}_{k+1}-{T}_k\right)\right] $$

where Yk e Yk + 1 are disease severity at evaluation times k and k + 1, respectively, and Tk e Tk + 1 are the time of evaluation k and k + 1, respectively, in number of dai.

AUDPC and anthracnose severity at 10 dai data were subjected to analysis of variance (ANOVA), and means were compared by the Scott-Knott test (p = 0.05) using software R (R Core Team 2015).

All sources of variation were significant in both experiments, except for the interaction isolate/mixture x cultivars in experiment 1. Isolates/mixtures that caused susceptibility reaction in ‘Pérola’ displayed scores above 3.0 from the fifth and sixth days after inoculation in experiments 1 and 2, respectively (Figs. 1 and 2). The cultivar BRSMG Majestoso was resistant to all isolates/mixtures, except for race 65 and mixture M1, in which cases it displayed scores above 3.0 from the fifth day on (Fig. 1d-f).

Fig. 1
figure 1

Anthracnose severity progress between 3 and 10 days after inoculation (dai) of Colletotrichum lindemuthianum on plants of cultivars Pérola and BRS Majestoso, comparing individual inoculations of races with inoculations of mixtures of races (M1, M2 and M3)

Full size image
Fig. 2
figure 2

Anthracnose severity progress between 3 and 10 days after inoculation (dai) of Colletotrichum lindemuthianum on plants of cultivars Pérola and BRS Majestoso, comparing individual inoculations of races with inoculations of mixtures of races (M1, M2, M3, M4, M5, M6 and M7)

Full size image

Mean AUDPC was higher in experiment 1 than in experiment 2 (Table 1). Anthracnose severity progress was faster and more intense in experiment 1 for isolates/mixtures of C. lindemuthianum that were evaluated in both experiments. It is possible that environmental conditions were more favorable to the disease during experiment 1. In spite of humidity and temperature control, outside temperatures above 30 °C can make it difficult to keep conditions between 15 and 25 °C (most favorable to disease development) inside the greenhouse (Paula-Júnior et al. 2015).

Table 1 Area under the disease progress curve (AUDPC) for common bean cultivars Pérola and BRSMG Majestoso 10 days after inoculation (dai) with different isolates and mixtures of Colletotrichum lindemuthianum
Full size table

In experiment 1, AUDPC means for both cultivars were grouped in two classes by the Scott-Knott test, and were higher for the isolate of race 65 and mixture M1 in both cultivars (Table 1). However, in experiment 2, the mixtures M3, M4 and M5 presented higher AUDPC estimates for Pérola. Race 81 presents an additional virulence factor compared to race 65. However, isolates of these races presented the same virulence in experiment 2. Race 65, the simplest race evaluated, and the M1 mixture were the most virulent in experiment 1. These results corroborate with those from other authors, who have found that series of international differential cultivars are not suitable to discriminate the pathogenic variability of Brazilian isolates of C. lindemuthianum (Ishikawa et al. 2011; Carbonell et al. 2012; Costa et al. 2017). Considering virulence factors, M1 has factors 20 and 26 in higher proportion since all races evaluated present these factors. Therefore, virulence factors 23 and 24, present in isolates of races 73 and 81, respectively, are in a lower proportion in the mixture. This fact may justify the similar behavior of M1 and of the isolate of race 65.

The greater magnitude of AUDPC and also of anthracnose severity estimates for Pérola using M1, M2, M3, M4 and M5 reveals an increase in race virulence for mixed isolates and also for new generations of isolations, except for M6 and M7. For M1 (mixture of the three races), although the concentration of each race was reduced by 1/3 (4 × 105 conidia/mL), the virulence of the mixture was greater or similar to the other isolates/mixtures. It has been shown that a concentration of 105 conidia/mL provides similar results to those obtained by using the usual recommended concentration of 1.2 × 106 for assessing anthracnose severity on common bean (Davide and Souza 2009). The higher virulence of M1 may be due to a synergistic effect among the three isolates, as already verified in this pathosystem by Chilipa et al. (2016). Another probable explanation would be the occurrence of heterosis, because alleles of different races are contributing to the virulence in the mixture.

The behavior of isolates and mixtures was similar at 3 and 4 dai. After that, differences in anthracnose severity were observed (Fig. 1). This can be confirmed by the estimates of the variation limits, which were of greater magnitude after 4 dai (Table 2). However, mean scores higher than 4.0 were observed at 6 and 10 dai in experiments 1 and in 2, respectively.

Table 2 Mean anthracnose severity scores (on a scale of 1 to 9) and variation limits from 3 to 10 days after inoculation (dai) on cultivars Pérola and BRSMG Majestoso
Full size table

Isolates of races 65 and 81 and mixtures M1, M2 and M3 presented similar virulence for cultivar Pérola, as well as similar anthracnose intensity in experiment 1 (Fig. 1a-c). Most mixtures were more virulent on cultivar Pérola in experiment 2 (Fig. 2a-c). Mixtures were more virulent than isolates of races 65, 73 and 81 in both experiments on cultivar BRSMG Majestoso (Figs. 1d-f and 2f). This cultivar was evaluated as resistant to all isolates and mixtures, as scores were lower than 3.0 in experiment 2.

Disease severity scores at 10 dai were used to carry out ANOVA in experiment 2. All sources of variation were significant, including the interaction that was partitioned; there were significant difference for isolates/mixtures for Pérola. The means test grouped M1, M2, M3, M4 and M5 mixtures that caused higher mean scores and therefore were more virulent than individual isolates and mixtures M6 and M7 (Table 1).

The lower magnitude of AUDPC and of anthracnose severity obtained upon inoculation of M6 and M7 (Table 1) reveals that there was a reduction in virulence using new generations of M1 initially inoculated on BRSMG Majestoso. This reduction may be related to the stabilizing selection action that is effective against races with unnecessary virulence alleles in a pathogen population (Vanderplank 1968), favoring 1 (20) and 64 (25) virulence factors, which were more frequent in M1, in comparison to virulence factors 8 and 16 of races 73 and 81, respectively. Casela et al. (2001) observed differences in competitive capacity of C. sublineolum races after five and six inoculations of mixture generations in sorghum cultivars, indicating a possible action of stabilizing selection against races of distinct virulence in the pathogen population. Our results were obtained with only two generations, but there are no reports of occurrence of conidial anastomosis tubes (CATs) recombinants for C. sublineolum, a mechanism that may be related to the asexual recombination of C. lindemuthianum (Ishikawa et al. 2012; Pinto et al. 2012; Mota et al. 2015).

The inoculation of mixtures resulted in similar or greater virulence compared to the inoculation of individual isolates. In addition, mixtures of isolates of C. lindemuthianum can be more efficient in selecting common bean progenies with different levels of resistance, as they have a larger number of virulence alleles. We recommend the use of mixtures of races at each inoculation. On the other hand, we do not recommend the use of mixtures obtained from new isolations done in plants inoculated with mixtures, due the action of stabilizing selection on the pathogen population.