Common bean lines developed in a biological nitrogen fixation system based on wild germplasm

ABSTRACT

Wild common bean accessions represent an important source of genetic variability for improving biological nitrogen fixation (BNF). The aims of this study were to estimate genetic parameters of lines developed under rhizobial inoculation conditions, obtained by backcross between an elite line and a wild accession, and to select high-yielding carioca common bean lines adapted to these conditions. A total of 95 lines derived from a cross between the elite line CNFC 10762 and the wild accession G2k3499A, followed by one backcross to the elite parent, were evaluated. The lines and three check genotypes were evaluated in the 2019 and 2020 winter crop seasons in Santo Antônio de Goiás, GO, Brazil, under BNF-based systems (rhizobial inoculation and absence of mineral nitrogen) using a randomized block design with three replications. Yield, grain appearance, and 100-seed weight were evaluated. Genetic variability was observed among lines, with some exceeding the CNFC 10762 line (1,877 kg•ha^-1) and the Pérola cultivar (1,724 kg•ha^-1). Heritability estimates were 59% for yield, 78% for 100-seed weight, and 80% for grain appearance, indicating favorable conditions for selection. Notable gains were achieved through simultaneous selection of the best lines for the traits evaluated (6.7% for yield, 15.4% for grain appearance, and 1.5% for 100-seed weight). Six superior lines combining high grain yield (1,835–2,147 kg•ha^-1), favorable grain appearance (1.2–1.0), and 100-seed weight (21.1–22.6 g) were identified, confirming the potential of wild germplasm to broaden the genetic base and contribute to developing carioca common bean cultivars better adapted to BNF.

Key words
Phaseolus vulgaris L.; symbiotic nitrogen fixation; rhizobial inoculation; biological nitrogen fixation; genetic parameters

INTRODUCTION

Nitrogen (N) is the most required nutrient by common bean, and N deficiency is one of the main factors limiting grain yield for this crop (Silva et al. 2024). The nutrient can be supplied to plants through two main sources: mineral N fertilization and biological nitrogen fixation (BNF). Although it is widely used, most mineral N fertilizer applied is lost through leaching, volatilization, and denitrification, resulting in high economic costs and significant environmental impacts (Sousa et al. 2022).

BNF is a sustainable and economically efficient alternative for reducing the use of N fertilizers (Sousa et al. 2022). In this symbiotic process, bacteria known as rhizobia convert atmospheric N₂ into nitrogen compounds for plants in exchange for carbohydrates (Jiang et al. 2020). However, BNF efficiency varies among leguminous plants, and common bean is known as a poor N fixer (Reinprecht et al. 2020). Factors such as promiscuous nodulation with native rhizobia (Moura et al. 2022), the short crop cycle, and the high sensitivity of BNF to abiotic stresses (Ferreira et al. 2013) contribute to the low symbiotic efficiency of common bean.

Genetic variability for BNF has been reported among common bean lines (Farid et al. 2017, Oladzad et al. 2020). However, most cultivars and elite lines when inoculated with bacteria for BNF have shown grain yields lower than the yields obtained with mineral N (Fageria et al. 2014, Pereira et al. 2015, Dias et al. 2020, Dias et al. 2024). The reason for this is that although these genotypes have favorable phenotypes for most agronomic traits, they were developed in systems with high availability of mineral N, which has been proven to reduce BNF capacity in leguminous crops (Reinprecht et al. 2020, Sousa et al. 2022).

Breeding plants for better N₂ fixation has not been a standard practice in common bean breeding programs (Ferreira et al. 2013, Kamfwa et al. 2019). Among the main factors that breeders should consider, there is the development of lines under rhizobial inoculation and in the absence of mineral N fertilization (BNF-based systems), conditions in which the N available to plants relies primarily on symbiotic associations. In soils with high inherent N fertility or where a substantial amount of mineral N fertilizer is applied, distinguishing lines with high and low symbiotic capacity becomes difficult, regardless the efficiency of the inoculated bacteria (Bliss 1993). This occurs because BNF is a process that requires energy from the plants, and less energy is required for taking up mineral N (McKenzie et al. 2001). However, few recent studies have reported the development of common bean lines in systems using BNF as an important N source (Costa et al. 2025).

As it is difficult to identify lines with good symbiotic performance among elite genotypes, an alternative is to turn them to wild germplasm. Wild genotypes constitute an untapped source of genetic variation that can increase the diversity of domesticated genotypes (Acosta-Gallegos et al. 2007). Studies show that some wild common bean accessions have superior nodulation traits (Ferreira et al. 2010, Knupp et al. 2017), suggesting that crosses between elite and wild lines may improve the symbiotic efficiency of the crop.

Due to the low agronomic adaptation of wild germplasm, the use of backcrossing has been recommended to introgress it in breeding (Teran et al. 2020). In common bean, this method has been applied to traits such as grain yield (Blair et al. 2006, Acosta-Gallegos et al. 2007, Teran et al. 2020), increased grain mineral concentrations (Blair and Izquierdo 2012), heat and drought tolerance (Burbano-Erazo et al. 2021), and disease resistance (Singh and Schwartz 2010). However, studies have not been conducted for introgression of BNF-related genes yet.

In 2014, the common bean breeding program of Embrapa Arroz e Feijão began to make backcrosses between elite genotypes and wild genotypes promising for BNF, which had been identified in various studies (Ferreira et al. 2010, Pereira et al. 2015, Knupp et al. 2017, Dias et al. 2020, Dias et al. 2024). Developing populations formed by backcrosses under rhizobial inoculation and reduced N input conditions is essential to promote expression of BNF. Furthermore, evaluating the genetic variability of these populations through estimation of genetic parameters constitutes an important tool for evaluating gains from selection of symbiotically efficient genotypes (Ramalho et al. 2012, Farid et al. 2017).

In this respect, the aims of this study were to estimate genetic parameters of lines developed under rhizobial inoculation conditions, obtained by backcrosses between an elite line and a wild accession, and to select carioca common bean lines with higher yield for these conditions.

MATERIALS AND METHODS

The lines used were obtained from a population derived from the cross between the elite line CNFC 10762 (recurrent parent) and the wild genotype G2k3499A (donor parent), followed by a backcross with the recurrent parent. The elite parent CNFC 10762 is part of the carioca commercial group, characterized by beans with a cream-colored seed coat with light brown streaks (Fig. 1a). It exhibits a semi-erect plant architecture and a normal growth cycle of approximately 90 days, and it has high yield potential, including in systems that use BNF (Pereira et al. 2015). Dias et al. (2024) reported that the CNFC 10762 line not only had a higher mean yield in BNF environments but also high specific weight of root nodules.

Figure 1
Grain appearance of the parental genotypes, the F₂ population before and after visual selection, and the selected lines. (a) Beans of the parental elite line CNFC 10762. (b) Beans of the parental wild accession G2k3499A. (c and d) Beans of the CNFC10762 / G2k3499A // CNFC10762 population in the F₂ generation before and after selection for grain appearance, respectively. (e and f) Beans of the selected lines CNFC 20332 and CNFC 20371, respectively.

The wild parent G2k3499A is of Mesoamerican origin and has beans of gray color with black streaks (Fig. 1b). Ferreira et al. (2010) and Knupp et al. (2017) found that this accession stood out among the 800 evaluated for nodulation parameters, with a high number and high dry weight of nodules.

The population under study was synthesized in 2015, along with nine others, from crosses between elite lines (CNFC 10762, CNFC 15023, and CNFC 15086) and wild accessions (G2k3499A, PHAVUL8169B, G12904, and G23460) that showed promise for BNF (Ferreira et al. 2010, Pereira et al. 2015, Knupp et al. 2017, Dias et al. 2020), followed by a backcross with the elite parent. The ten populations were advanced from the BC₁F₂ to BC₁F₄ generations in the field in the municipality/county of Santo Antônio de Goiás, GO, Brazil, always in the winter crop season (sowing in May) using the bulk method, with rhizobial inoculation and no application of mineral N fertilizer, and fertilization with phosphorus and potassium based on soil analyses.

As generations advanced, larger beans with appearance like those of the carioca commercial group were selected to form the following generation. Among the 10 populations, the CNFC10762 / G2k3499A // CNFC10762 population was selected to develop lines, as it had beans and plants with greater similarity to those of the carioca group. Figure 1c shows the beans of the population selected in the F₂ generation, in which maximum genetic variability is expressed. Figure 1d, in turn, illustrates the beans of this population after the visual selection process, which were later sown to form the F₃ generation. This process was repeated to compose the F₄ generation.

In the F₄ generation in the 2018 winter season, 95 plants were harvested, which gave rise to the 95 lines evaluated in the F_4:5 generation, in the 2019 winter crop season, and in the F_4:6 generation, in the 2020 winter crop season, in Santo Antônio de Goiás. Three check genotypes were also sown: the elite parent CNFC 10762; the Pérola cultivar, which is a good yield standard in systems with BNF (Pereira et al. 2015, Dias et al. 2024); and the non-nodulating line NORH 54. A randomized block experimental design was used with three replications, and plots consisted of three 3-m-length rows at a spacing of 0.5 m.

The climate in the region where the experiments were conducted is Aw, tropical savanna, according to the Köppen’s classification, and the predominant soil is Latossolo Vermelho-Escuro with clayey texture and flat topography (Embrapa 2018). The experimental area had been previously cultivated with Brachiaria brizantha pasture, a system typically characterized by low external N inputs. At planting, P₂O₅ and K₂O fertilizers were applied according to soil analysis. No mineral N fertilizer was applied during the experiments.

Liquid inoculant produced in the soil biology laboratory of Embrapa Arroz e Feijão was applied, consisting of a 1:1:1 mixture of the Rhizobium tropici (SEMIA 4077 and 4088) and Rhizobium freirei (SEMIA 4080) strains, registered with the Brazilian Ministry of Agriculture for common bean. The inoculant had a density of 10⁹ cells⸱mL^-1 and was applied in the furrow at sowing using a directed jet sprayer attached to the plot planter, with a dosage of 200 mL⸱ha^-1.

The traits evaluated were grain yield, 100-seed weight, and grain appearance. Grain yield was calculated based on the weight of the grain harvested from each plot and converted to kg⸱ha^-1. For 100-seed weight (g), a random sample of 100 beans was weighed from each plot. Grain appearance was evaluated after harvest using a 1 to 5 scoring scale, adapted from the scale proposed by Faria et al. (2013), in which score 1 refers to the typical carioca bean (cream-colored beans with light brown streaks, a light-colored seed coat, no halo, and non-flattened beans) and score 5 refers to cream colored beans with dark brown streaks, dark seed coat, yellow halo, and flattened beans.

Individual and combined analyses of variance were performed for all the traits. The genotype effect was considered random, and the environment effect was considered fixed. For combined analysis of the experiments, homogeneity of variances was checked using the ratio between the highest and lowest mean square of the residual (Pimentel-Gomes 2009). Selective accuracy (SA) was estimated to assist evaluation of experimental quality (Resende and Duarte 2007).

The following genetic and phenotypic parameters were estimated: genetic variance (σ²_g), phenotypic variance (σ²_ph), variance of the genotype-by-year interaction (σ²_gxe), heritability at the mean level (h²), and expected gain from direct and simultaneous selection, with selection intensity of 20%. To estimate direct gain, the best lines for each trait were selected. For simultaneous selection, the method of independent culling levels was used (Hazel and Lush 1942), which consists of establishing maximum and/or minimum values that the lines must have for each trait. All the lines that did not reach the overall mean of the experiments for at least one of the traits were eliminated. All statistical analyses were performed using the Genes software (Cruz 2013).

RESULTS AND DISCUSSION

The environmental coefficients of variation (CVe) were 20 and 19% for yield and 4 and 4.5% for 100-seed weight, in the years 2019 and 2020, respectively, indicating high experimental accuracy. For grain appearance, the coefficients of variation (CVs) were high, at 46 and 31%, which can mainly be explained by the low mean values of the trait, which directly affect the magnitude of the CVe. The estimates of SA were high for the three traits evaluated (> 0.70), which reflects the high informativeness of the experiments and the possibility of obtaining reliable genetic parameters (Resende and Duarte 2007).

Significant differences were detected among the lines for all the traits in the combined analyses (Table 1). The introduction of wild germplasm in common bean breeding represents an innovative strategy for increasing the genetic variability and the efficiency of BNF. Detection of genetic variability among the lines derived from crosses between elite and wild germplasm in environments with rhizobial inoculation allows the selection of genotypes with higher yield and 100-seed weight and better grain appearance under these conditions.

Table 1
Summary of the combined analyses of variance for grain yield (kg⸱ha^-1), 100-seed weight (g), and bean grain appearance (scores from 1 to 5) in common bean genotypes of the carioca commercial group (Phaseolus vulgaris) evaluated in the 2019 and 2020 winter crop season in Santo Antônio de Goiás, GO, Brazil, with inoculation for biological nitrogen fixation.

Source of variation		Grain yield			100-seed weight			Bean grain appearance
	DF	MS	P-value		MS	P-value		MS	P-value
Blocks/environments	4	402,309	0.005		20.49	0.001		2.47	0.001
Genotypes	97	265,279	0.001		4.24	0.001		2.72	0.001
Lines (L)	94	259,478	0.001		3.59	0.001		1.63	0.001
Checks (C)	2	656,290	0.002		25.95	0.001		8.00	0.001
C versus L	1	28,608	0.605		22.01	0.001		95.00	0.001
Environments (E)	1	62,326,743	0.001		617.56	0.003		19.11	0.031
Genotypes × E	97	192,379	0.001		1.12	0.016		0.37	0.237
L × E	94	193,353	0.001		0.97	0.119		0.38	0.212
C × E	2	136,240	0.280		8.62	0.001		0.00	1.000
(C versus L) × E	1	213,102	0.158		0.61	0.387		0.60	0.178
Residual	388	106,702			0.81			0.33
Overall mean		1,654			21.1			1.4
CVe (%)		19.7			4.3			41.0
SA (%)		77.3			90.0			93.7

• CVe: environmental coefficient of variation; SA: selective accuracy; DF: degrees of freedom; MS: mean square.

There were significant differences between the years for the three traits evaluated (Table 1). The mean values were 1,329 and 1,980 kg⸱ha^-1 for grain yield, 20.1 and 22.1 g for 100-seed weight, and 1.6 and 1.2 for grain appearance in 2019 and 2020, respectively. The 33% difference between the mean yield values indicates that the climate factors in the two years of evaluation considerably affected expression of the trait. Plant performance under rhizobial inoculation is highly sensitive to environmental variations, such as temperature, water availability, and nutrient content in the soil (Singh et al. 2023), which may have contributed to the differences found.

The effect of the line-by-year interaction (L × Y) was significant only for yield (Table 1). These results indicate a differential response of the lines between the years of evaluation, which is frequently reported for the common bean crop in evaluations with mineral N fertilization (Torres et al. 2022, Cavalheiro et al. 2023) and rhizobial inoculation (Pereira et al. 2015, Farid et al. 2017, Barbosa et al. 2018, Dias et al. 2020, Dias et al. 2024). In other studies, grain appearance and 100-seed weight showed significant interaction (Cavalheiro et al. 2023, Dias et al. 2024), indicating that the lines evaluated in the present study were less affected by specific annual conditions, which may represent an advantage when selecting genotypes with stable performance under BNF-based systems.

The mean values of the 95 lines ranged from 1,059 to 2,227 kg⸱ha^-1 for yield, based on combined analysis. Among the 95 lines evaluated, five (12%) had yields superior to the yield of the CNFC 10762 line, and 19 (36%) were higher yielding than the Pérola cultivar. Pereira et al. (2015) highlighted CNFC 10762 as one of the five highest-yielding lines in trials with rhizobial inoculation. Dias et al. (2024) emphasized that besides exhibiting higher yield in environments with BNF, the line had high specific weight of root nodules. The Pérola cultivar is also recognized for its good performance in BNF-based systems, standing out for its consistent yield under different N sources (Pereira et al. 2015, Steiner et al. 2019, Dias et al. 2020).

The use of a poorly adapted wild accession as a donor parent in the backcross that gave rise to the population did not inhibit identification of lines with yield performance like or higher than that of the recurrent parent (CNFC 10762). These results suggest that wild genotypes can be a potential source of new alleles for breeding with the aim of enhancing BNF and that the advanced backcross strategy can be successful in transferring these genes (Teran et al. 2020).

The selection of genotypes promising for BNF based on grain yield has been reported in the literature (Fageria et al. 2014, Pereira et al. 2015, Dias et al. 2020, Dias et al. 2024). As N is the nutrient of greatest demand and it is associated with vital plant functions, its deficiency directly affects biomass production and grain yield. Thus, in BNF-based systems without mineral N fertilization, plants better adapted to these conditions tend to show higher grain yield (Bliss 1993).

The mean values for 100-seed weight ranged from 15.6 g for the CNFC 20357 line to 23.8 g for the CNFC 10762 parental line. The Pérola cultivar had the second highest mean (22.9 g), indicating that no line exceeded the check genotypes for the trait. For grain appearance, the variation was from 1 for 34% of the lines to 5 for the check genotype NORH 54.

The Brazilian market demands very light beige beans with light brown streaks (score 1) and seed weight higher than 25 g per 100 seeds (Faria et al. 2013, Pereira et al. 2017). This indicates that for 100-seed weight, the commercial goals were not met in the population under study, even though there was selection during development of the lines. However, the large number of lines with good bean appearance highlights the potential of these genotypes in meeting market demands. This is quite advantageous, especially involving crosses with wild genotypes.

The potential of the population under study for selection of promising lines can be confirmed by the estimates of genetic variance (σ²_g) and heritability (h²) (Table 2). The high estimates of genetic variance found may be related to the use of wild germplasm in the cross that gave rise to the population. Hybridizations between wild and domesticated genotypes of P. vulgaris offer high potential for increasing genetic variation in the crop. Wild germplasm contains a vast collection of alleles not present in plants grown as crops, as these alleles may have been lost during domestication (Teran et al. 2020).

Table 2
Estimates of genetic variance (σ²_g), phenotypic variance (σ²_f), variation of the genotype-by-environment interaction (σ²_ge), heritability (h²), and expected gain from selection (GS), direct and simultaneous, based on individual and combined analyses for grain yield (kg⸱ha^-1), 100-seed weight (g), and bean grain appearance (scores from 1 to 5), obtained from 95 common bean lines of the carioca commercial group (Phaseolus vulgaris), evaluated in the 2019 and 2020 winter crop seasons in Santo Antônio de Goiás, GO, Brazil, with inoculation for biological nitrogen fixation.

Parameter	Grain yield				100-seed weight				Bean grain appearance
	2019	2020	Combined		2019	2020	Combined		2019	2020	Combined
σ 2 g	36,381	43,426	25,462		0.44	0.54	0.46		0.28	0.17	0.22
σ 2 f	60,218	90,725	43,246		0.65	0.87	0.60		0.45	0.22	0.27
σ 2 g e	-	-	14,279		-	-	0.05		-	-	0.01
h2(%)	60.4	47.9	58.9		67.2	62.5	77.5		61.5	78.0	79.6
Direct GS (%)	15.2	10.1	10.0		3.0	2.2	2.3		-21.0	-10.0	-19.9
Simultaneous GS (%)	10.9	7.6	6.7		1.6	1.7	1.5		-16.3	-10.0	-15.4

The variance of the genotype-by-environment interaction was 56% lower than the genetic variance for grain yield (Table 2). This suggests that the genotype-by-environment interaction was less important for expression of the trait, indicating good possibilities for successfully selecting genotypes with superior BNF. Similar results for grain yield have been reported by other authors in evaluations of common bean under mineral N fertilization (Torres et al. 2022, Cavalheiro et al. 2023) and rhizobial inoculation (Farid et al. 2017).

Mean heritability estimates (h²) for grain yield were moderate in both environments (60% in 2019 and 48% in 2020) and in combined analysis (59%) (Table 2). During the experimental period, mean air temperature was 22.4 in 2019 and 21.9°C in 2020, with similar maximum and minimum temperatures between years. Precipitation was minimal in both years, as water supply was ensured through center-pivot irrigation. These results indicate that climatic conditions were generally similar between years, suggesting that the lower heritability estimate observed in 2020 does not appear to be associated with major differences in temperature or rainfall. Instead, this year-to-year variation is likely due to non-controllable environmental factors typical of field trials, which increased the residual variance and consequently reduced the heritability in the second year.

The heritability estimates for grain yield obtained in this study were lower than those obtained by Torres et al. (2022) and Cavalheiro et al. (2023) in evaluations of common bean under mineral N fertilization. In one of the few studies estimating genetic parameters in a common bean population evaluated for BNF, Farid et al. (2017) found heritability estimates for grain yield of 32% in environments without stress and 25% in environments with water stress, showing the reduction in heritability under adverse climate conditions in BNF-based systems.

As grain yield is controlled by various genes and is highly affected by the environment, low to moderate heritability estimates are expected for the trait. Variation in these estimates depends on both environmental conditions and the genetic variability of the evaluated genotypes (Ramalho et al. 2012). Despite the year-to-year variation observed, the moderate heritability estimates obtained for grain yield indicate that the experimental conditions with rhizobial inoculation were adequate for discriminating genetic differences among lines.

For 100-seed weight, the heritability estimate was 78% based on combined analysis, ranging from 67 to 63% in individual analyses, confirming that selection may still be successful in increasing 100-seed weight (Table 2). For grain appearance, heritability was 80% in combined analysis, ranging from 62 to 78% in individual analyses. These estimates are similar to those reported by Cavalheiro et al. (2023) upon evaluating two carioca common bean populations. The high estimates found are noteworthy, especially considering that the population arose from a cross with wild germplasm and only one backcross was performed. That shows greater ease of selection of genotypes superior for these traits in BNF-based systems.

Gains from direct and simultaneous selection of the 20 best lines were higher in 2019 for the three traits evaluated (Table 2). That was due to the greater environmental variance observed in 2020. Direct selection provided for gains of 10% for yield, 2.3% for 100-seed weight, and 19.9% for grain appearance based on combined analysis, confirming the possibility of increasing the frequency of favorable alleles controlling these traits for the next generation.

The chances of success in selecting cultivars in a breeding program are maximized in simultaneous selection because it considers the multiple traits of interest in a commercial genotype (Ribeiro et al. 2024). Based on combined analysis, simultaneous selection led to increases of 189 kg⸱ha^-1 for grain yield, 0.4 g for 100-seed weight, and 0.2 for grain appearance in the mean of the selected lines compared to the original mean, which represent increases of 7, 2, and 15%, respectively (Table 2).

The 20 selected lines had a mean grain yield of 1,844 kg⸱ha^-1 (Table 3). Of these, five had a mean value higher than that of the parent CNFC 10762 (1,877 kg⸱ha^-1). In the study by Pereira et al. (2015), the CNFC 10762 line had a mean yield of 1,658 kg⸱ha^-1 in inoculated environments and of 2,076 kg⸱ha^-1 in environments with mineral N fertilization. Despite the differences between the mean values, the genotype was selected for being among the five highest-yielding lines under both N sources. Dias et al. (2024) also evaluated the CNFC 10762 line and found mean values of 2,380 kg⸱ha^-1 in inoculated environments and 2,304 kg⸱ha^-1 in environments with mineral N. This line also had a high mean for specific weight of root nodules in the study, and it was recommended as a source of alleles favorable to BNF.

Table 3
Mean values based on combined analysis of the 20 common bean lines of the carioca commercial group (Phaseolus vulgaris) selected using simultaneous selection, and of the check genotypes for grain yield (kg⸱ha^-1), 100-seed weight (g), and bean grain appearance (scores from 1 to 5), evaluated in the 2019 and 2020 winter crop season, in Santo Antônio de Goiás, GO, Brazil, with inoculation for biological nitrogen fixation.

Genotype	Grain yield	100-seed weight	Grain appearance
CNFC20379*	2,147	21.3	1.0
CNFC20380*	2,051	21.1	1.2
CNFC20321*	1,990	21.3	1.0
CNFC20332*	1,936	21.2	1.0
CNFC20344*	1,901	21.3	1.2
CNFC10762	1,877	23.8	3.0
CNFC20371	1,858	21.3	1.0
CNFC20365	1,844	21.8	1.0
CNFC20378	1,843	21.1	1.2
CNFC20289*	1,835	22.6	1.2
CNFC20301	1,827	21.8	1.2
CNFC20366	1,822	21.5	1.2
CNFC20319	1,803	21.3	1.2
CNFC20358	1,800	21.4	1.0
CNFC20369	1,784	21.6	1.2
CNFC20294	1,773	21.5	1.0
CNFC20306	1,759	21.4	1.0
CNFC20308	1,746	21.4	1.2
CNFC20323	1,742	21.5	1.0
PÉROLA	1,724	22.9	3.0
CNFC20318	1,719	21.7	1.0
CNFC20300	1,710	21.6	1.0
NORH 54	1,243	19.9	5.0
Means of the check genotypes	1,614	22.2	3.7
Mean of the lines	1,655	21.1	1.3
Mean of the selected lines	1,844	21.5	1.1

• ^*
  Most promising lines among the 20 selected; check lines in bold print.

Among the 20 selected lines, 18 showed yield higher than that of the Pérola cultivar (1,724 kg⸱ha^-1) (Table 3). The mean value observed for the Pérola cultivar was lower than the mean values reported by Steiner et al. (2019) and Dias et al. (2024), but higher than those found by Pereira et al. (2015) and Mercante et al. (2017) in experiments with rhizobial inoculation. However, in studies with mineral N fertilization, the mean values of the cultivar were higher (Pereira et al. 2017, Cavalheiro et al. 2023).

Although the elite line CNFC 10762 shows high yield performance in inoculated environments (Pereira et al. 2015, Dias et al. 2024) and the Pérola cultivar has high market acceptance, especially because of its yield and grain quality traits (Cavalheiro et al. 2023), both were developed in environments with mineral N fertilization. Therefore, it is expected that they would have higher yield under these conditions.

Identification of lines with better performance than these genotypes (Table 3) shows the potential of the population under study, which came from crossing elite and wild germplasm for environments with inoculation. In addition to the high BNF capacity inherited from the wild accession, advancing the population through the bulk method in exclusive rhizobial inoculation systems may have contributed to the mean values found. Therefore, natural selection favors the higher-yielding genotypes better adapted to growth under rhizobial inoculation (Pereira et al. 2015).

For 100-seed weight, the mean of the selected lines was 21.5 g (Table 3), and no line exceeded the mean of the check genotypes CNFC 10762 (23.8 g) and Pérola (22.9 g). According to Blair et al. (2006), an important issue in crosses between elite and wild genotypes is the trade-off between smaller seed size or later flowering and higher grain yield. Another possible factor may be related to the N source used. The effect of N management through mineral fertilization and rhizobial inoculation on 100-seed weight is widely reported in the literature (Fageria et al. 2014, Dias et al. 2020, Reinprecht et al. 2020, Dias et al. 2024). In the studies, mineral N fertilization led to positive effects on the mean values of 100-seed weight compared to inoculation. This can be confirmed through the mean of the Pérola cultivar (Table 3), a commercial reference for the trait, which exhibited performance lower than more than 25 g per 100 seeds, as reported in studies with mineral N (Fageria et al. 2014, Silva et al. 2023).

For grain appearance, the mean of the selected lines was 1.1 (Table 3). All the lines had lower mean values and were therefore favorable compared to the check genotypes CNFC 10762 (3) and Pérola (3). One of the biggest concerns in crosses involving wild common bean genotypes is bean appearance, and this is even more essential for the carioca group, which has specific requirements for commercial acceptance (Pereira et al. 2017). For that reason, in advancing the generations of the population using the bulk method, beans with appearance like the one of the carioca group were selected to establish the next generation. Figures 1e and 1f show the beans of the CNFC 20332 and CNFC 20371 lines, respectively, and the uniformity in color and bean shape can be seen, meeting the desired visual standards. This highlights that the selection process used contributed to obtaining lines with very light-beige-colored beans with slow darkening over time in BNF-based systems.

The lines CNFC 20379, CNFC 20380, CNFC 20321, CNFC 20332, and CNFC 20344 stood out among the 20 selected by having mean yield and appearance values higher than those of the elite parent (Table 3). The CNFC 20289 line is also noteworthy, as it obtained the highest mean value for 100-seed weight among the selected lines. Although traits directly related to nodulation were not measured in this study, the high agronomic performance of these lines under exclusive rhizobial inoculation reflects their capacity to develop and yield satisfactorily in the absence of mineral N, representing a reliable indicator of adaptation to BNF-based systems. The lines selected not only show traits like the commercial type but also offer the advantage of incorporating genetic diversity into the crop background. In addition to its potential contribution to improving symbiotic efficiency, this diversity is fundamental for developing more robust and adaptable cultivars that can take on environmental challenges and enhance yield.

CONCLUSION

The advanced backcross strategy using wild germplasm enables the development of carioca common bean lines with adequate agronomic performance under rhizobial inoculation.

Moderate to high heritability estimates for yield, grain appearance, and 100-seed weight indicate favorable conditions for genetic gains through selection in BNF-based systems.

The selected lines (CNFC 20379, CNFC 20380, CNFC 20321, CNFC 20332, and CNFC 20344) represent promising genetic material for advanced trials and for use as parents in breeding programs targeting common bean cultivation under rhizobial inoculation.

ACKNOWLEDGMENTS

We thank all members of the common bean breeding team at the Brazilian Agricultural Research Corporation (EMBRAPA).

• How to cite:
  Costa, N. V., Knupp, A. M., Cavalheiro, S. B., Ferreira, E. P. B., Souza, T. L. P. O., Melo, L. C. and Pereira, H. S. (2026). Common bean lines developed in a biological nitrogen fixation system based on wild germplasm. Bragantia, 85, e20260041. https://doi.org/10.1590/1678-4499.20260041
• FUNDING
  Conselho Nacional de Desenvolvimento Científico e Tecnológico
  Grant No.: 384993/2025-7
• DECLARATION OF USE OF ARTIFICIAL INTELLIGENCE TOOLS
  The authors declare that no artificial intelligence tools were used in the development, writing, or editing of this manuscript.

DATA AVAILABILITY STATEMENT

Data will be available from the corresponding author on reasonable request.

REFERENCES

• Acosta-Gallegos, J. A., Kelly, J. D. and Gepts, P. (2007). Prebreeding in common bean and use of genetic diversity from wild germplasm. Crop Science, 47, S44-S59. https://doi.org/10.2135/cropsci2007.04.0008IPBS
  » https://doi.org/10.2135/cropsci2007.04.0008IPBS
• Barbosa, N., Portilla, E., Buendia, H. F., Raatz, B. and Beebe, S. and Rao, I. (2018). Genotypic differences in symbiotic nitrogen fixation ability and seed yield of climbing bean. Plant Soil, 428, 223-239. https://doi.org/10.1007/s11104-018-3665-y
  » https://doi.org/10.1007/s11104-018-3665-y
• Blair, M. W. and Izquierdo, P. (2012). Use of the advanced backcross-QTL method to transfer seed mineral accumulation nutrition traits from wild to Andean cultivated common beans. Theoretical and Applied Genetics, 125, 1015-1031. https://doi.org/10.1007/s00122-012-1891-x
  » https://doi.org/10.1007/s00122-012-1891-x
• Blair, M. W., Iriarte, G. and Beebe, S. (2006). QTL analysis of yield traits in an advanced backcross population derived from a cultivated Andean× wild common bean (Phaseolus vulgaris L.) cross. Theoretical and Applied Genetics, 112, 1149-1163. https://doi.org/10.1007/s00122-006-0217-2
  » https://doi.org/10.1007/s00122-006-0217-2
• Bliss, F. A. (1993). Breeding common bean for improved biological nitrogen fixation. Plant and Soil, 152, 71-79. https://doi.org/10.1007/BF00016334
  » https://doi.org/10.1007/BF00016334
• Burbano-Erazo, E., León-Pacheco, R. I., Cordero-Cordero, C. C., López-Hernández, F., Cortés, A. J. and Tofiño-Rivera, A. P. (2021). Multi-environment yield components in advanced common bean (Phaseolus vulgaris L.) × tepary bean (P. acutifolius A. Gray) interspecific lines for heat and drought tolerance. Agronomy, 11, 1978. https://doi.org/10.3390/agronomy11101978
  » https://doi.org/10.3390/agronomy11101978
• Cavalheiro, S. B., Melo, L. C., Souza, T. L. P. O. and Pereira, H. S. (2023). Genetic parameters and selection of carioca common bean lines resistant to fusarium wilt. Pesquisa Agropecuária Brasileira, 58, e03313. https://doi.org/10.1590/S1678-3921.pab2023.v58.03313
  » https://doi.org/10.1590/S1678-3921.pab2023.v58.03313
• Costa, N. V., Melo, P. G. S., Pereira, H. S., Ferreira, E. P. B., Knupp, A. M. and Melo, L. C. (2025). Combining ability of black common bean developed under a biological nitrogen fixation system. Pesquisa Agropecuária Brasileira, 60, e03926. https://doi.org/10.1590/S1678-3921.pab2025.v60.03926
  » https://doi.org/10.1590/S1678-3921.pab2025.v60.03926
• Cruz, C. D. (2013). GENES - a software package for analysis in experimental statistics and quantitative genetics. Acta Scientiarum, 35, 271-276. https://doi.org/10.4025/actasciagron.v35i3.21251
  » https://doi.org/10.4025/actasciagron.v35i3.21251
• Dias, P. A. S., Melo, P. G. S., Ferreira, E. P. B., Melo, L. C., Souza, T. L. P. O., Faria L. C., Almeida, V. M. and Pereira, H. S. (2024). Nitrogen sources on agronomic traits in carioca common bean lines developed under mineral nitrogen fertilization. Bragantia, 83, e20230285. https://doi.org/10.1590/1678-4499.20230285
  » https://doi.org/10.1590/1678-4499.20230285
• Dias, P. A. S., Melo, P. G. S., Melo, L. C., Souza, T. L. P. O., Faria, L. C., Ferreira, E. P. B. and Pereira, H. S. (2020). Production and disease resistance of elite black bean lines previously selected using mineral nitrogen fertilization cultivated with natural versus artificial nitrogen supplementation. Genetics and Molecular Research, 19, gmr18491. https://doi.org/10.4238/gmr185
  » https://doi.org/10.4238/gmr185
• Embrapa – Empresa Brasileira de Pesquisa Agropecuária (2018). Sistema brasileiro de classificação de solos. Brasília: Embrapa.
• Fageria, N. K., Melo, L. C., Ferreira, E. P. B., Oliveira, J. P. and Knupp, A. M. (2014). Dry matter, grain yield, and yield components of dry bean as influenced by nitrogen fertilization and rhizobia. Communications in Soil Science and Plant Analysis, 45, 111-125. https://doi.org/10.1080/00103624.2013.848877
  » https://doi.org/10.1080/00103624.2013.848877
• Faria, L. C., Melo, P. G. S., Pereira, H. S., Del Peloso, M. J., Brás, A. J. B. P., Moreira, J. A. A., Carvalho, H. W. L. and Melo, L. C. (2013). Genetic progress during 22 years of improvement of carioca-type common bean in Brazil. Field Crops Research, 142, 68-74. https://doi.org/10.1016/j.fcr.2012.11.016
  » https://doi.org/10.1016/j.fcr.2012.11.016
• Farid, M., Earl, H. J., Pauls, K. P. and Navabi, A. (2017). Response to selections for improved nitrogen fixations in common bean (Phaseolus vulgaris L.). Euphytica, 213, 99. https://doi.org/10.1007/s10681-017-1885-5
  » https://doi.org/10.1007/s10681-017-1885-5
• Ferreira, E. P. B., Barbosa, L. H. A., Knupp, A. M., Mata, W. M., Wendland, A., Didonet, A. D., Melo, L. C. and Del Peloso, M. J. (2010). Identification of high nodulation efficiency among wild genotypes of common beans. Annual Report of the Bean Improvement Cooperative, 53, 170-171.
• Ferreira, E. P. B., Mercante, F. M., Hungria, M., Mendes, I. C., Araújo, J. L. S., Fernandes Júnior, P. I. and Araújo, A. P. (2013). Contribuições para melhoria da eficiência da fixação biológica de nitrogênio no feijoeiro comum no Brasil. In A. P. Araújo and B. J. R. Alves (Eds.). Tópicos em ciência do solo (p. 251-292). Viçosa: Sociedade Brasileira de Ciência do Solo.
• Hazel, L. N. and Lush, J. L. (1942). The efficiency of three methods of selection. Journal of Hered, 33, 393-399.
• Jiang, Y., MacLean, D. E., Perry, G. E., Marsolais, F., Hill, B. and Pauls, K. P. (2020). Evaluation of beneficial and inhibitory effects of nitrate on nodulation and nitrogen fixation in common bean (Phaseolus vulgaris). Legume Science, 2, e45. https://doi.org/10.1002/leg3.45
  » https://doi.org/10.1002/leg3.45
• Kamfwa, K., Cichy, K. A. and Kelly, J. D. (2019). Identification of quantitative trait loci for symbiotic nitrogen fixation in common bean. Theoretical and Applied Genetics, 132, 1375-1387. https://doi.org/10.1007/s00122-019-03284-6
  » https://doi.org/10.1007/s00122-019-03284-6
• Knupp, A. M., Ferreira, E. P. B. and Araújo, A. P. (2017). Variability of nodulation traits in Andean and Mesoamerican common bean gene pools. Pesquisa Agropecuária Brasileira, 52, 252-260. https://doi.org/10.1590/S0100-204X2017000400005
  » https://doi.org/10.1590/S0100-204X2017000400005
• McKenzie, R. H., Middleton, A. B., Seward, K. W., Gaudiel, R., Wildschut, C. and Bremer, E. (2001). Fertilizer responses of dry bean in southern Alberta. Canadian Journal of Plant Science, 81, 343-350. https://doi.org/10.4141/P00-106
  » https://doi.org/10.4141/P00-106
• Mercante, F. M., Otsubo, A. A. and Brito, O. R. (2017). New native rhizobia strains for inoculation of common bean in the Brazilian Savanna. Revista Brasileira de Ciência do Solo, 41, e0150120. https://doi.org/10.1590/18069657rbcs20150120
  » https://doi.org/10.1590/18069657rbcs20150120
• Moura, F. T., Ribeiro, R. A., Helene, L. C. F., Nogueira, M. A. and Hungria, M. (2022). So many rhizobial partners, so little nitrogen fixed: The intriguing symbiotic promiscuity of common bean (Phaseolus vulgaris L.). Symbiosis, 86, 169-185. https://doi.org/10.1007/s13199-022-00831-6
  » https://doi.org/10.1007/s13199-022-00831-6
• Oladzad, A., González, A., Macchiavelli, R., Jensen, C. E., Beaver, J., Porch, T. and McClean, P. (2020). Genetic factors associated with nodulation and nitrogen derived from atmosphere in a middle American common bean panel. Frontiers in Plant Science, 11, 576078. https://doi.org/10.3389/fpls.2020.576078
  » https://doi.org/10.3389/fpls.2020.576078
• Pereira, H. S., Alvares, R. C., Silva, F. C., Faria, L. C. and Melo, L. C. (2017). Genetic, environmental and genotype x environment interaction effects on the common bean grain yield and commercial quality. Semina: Ciências Agrárias, 38, 1241-1250. https://doi.org/10.5433/1679-0359.2017v38n3p1241
  » https://doi.org/10.5433/1679-0359.2017v38n3p1241
• Pereira, H. S., Melo, L. C., Faria, L. C., Ferreira, E. P. B., Mercante, F. M., Wendland, A. and Souza, T. L. P. O. (2015). Common bean elite lines cultivated under nitrogen fertilization and inoculation with Rhizobium tropici Ciência Rural, 45, 2168-2173. https://doi.org/10.1590/0103-8478cr20141135
  » https://doi.org/10.1590/0103-8478cr20141135
• Pimentel-Gomes, F. (2009). Curso de estatística experimental (15th ed.). Piracicaba: FEALQ-USP.
• Ramalho, M. A. P., Abreu, A. F. B., Santos, J. B. and Nunes, J. A. R. (2012). Aplicações da genética quantitativa no melhoramento de plantas autógamas. Lavras: Editora UFLA.
• Reinprecht, Y., Schram, L., Marsolais, F., Smith, T. H., Hill, B. and Pauls, K. P. (2020). Effects of nitrogen application on nitrogen fixation in common bean production. Frontiers in Plant Science, 11, 01172. https://doi.org/10.3389/fpls.2020.01172
  » https://doi.org/10.3389/fpls.2020.01172
• Resende, M. D. V. and Duarte, J. B. (2007). Precisão e controle de qualidade em experimentos de avaliação de cultivares. Pesquisa Agropecuária Tropical, 37, 182-194.
• Ribeiro, N. D., Santos, G. G. and Andrade, F. F. (2024). Enhancing agronomic performance through genetic improvement of key traits in newly development common bean cultivars in southern Brazil. Euphytica, 220, 1. https://doi.org/10.1007/s10681-023-03261-7
  » https://doi.org/10.1007/s10681-023-03261-7
• Silva, C. G. M., Moreira, S. G., Moraes, L. C., Gaudencio, J. R. F. and Pimentel, G. V. (2024). Nutrient extraction and export by determinate and indeterminate common bean cultivars. Pesquisa Agropecuária Tropical, 54, e78404. https://doi.org/10.1590/1983-40632024v5478404
  » https://doi.org/10.1590/1983-40632024v5478404
• Silva, F. C., Martins, S. M., Pereira, H. S., Melo, P. G. S. and Melo, L. C. (2023). Strategies for the selection of common bean lines for yield and commercial grain quality. Pesquisa Agropecuária Brasileira, 58, e03403. https://doi.org/10.1590/S1678-3921.pab2023.v58.03403
  » https://doi.org/10.1590/S1678-3921.pab2023.v58.03403
• Singh, A., Schöb, C. and Iannetta, P. P. M. (2023). Nitrogen fixation by common beans in crop mixtures is influenced by growth rate of associated species. BMC Plant Biology, 23, 253. https://doi.org/10.1186/s12870-023-04204-z
  » https://doi.org/10.1186/s12870-023-04204-z
• Singh, S. P. and Schwartz, H. F. (2010). Breeding common bean for resistance to diseases: A review. Crop Science, 50, 2199-2223. https://doi.org/10.2135/cropsci2009.03.0163
  » https://doi.org/10.2135/cropsci2009.03.0163
• Sousa, W. S., Soratto, R. P., Peixoto, D. S., Campos, T. S., Silva, M. B., Souza, A. G. V., Teixeira, I. R. and Gitari, H. I. (2022). Effects of Rhizobium inoculum compared with mineral nitrogen fertilizer on nodulation and seed yield of common bean. A meta-analysis. Agronomy for Sustainable Development, 42, 52. https://doi.org/10.1007/s13593-022-00784-6
  » https://doi.org/10.1007/s13593-022-00784-6
• Steiner, F., Ferreira, H. C. P. and Zuffo, A. M. (2019). Can co-inoculation of Rhizobium tropici and Azospirillum brasilense increase common bean nodulation and grain yield? Semina: Ciências Agrárias, 40, 81-98. https://doi.org/10.5433/1679-0359.2019v40n1p81
  » https://doi.org/10.5433/1679-0359.2019v40n1p81
• Teran, J. C. B. M., Konzen, E. R., Palkovic, A., Tsai, S. M. and Gepts, P. (2020). Exploration of the yield potential of Mesoamerican wild common beans from contrasting eco-geographic regions by nested recombinant inbred populations. Frontiers in Plant Science, 11, 00346. https://doi.org/10.3389/fpls.2020.00346
  » https://doi.org/10.3389/fpls.2020.00346
• Torres, M. H. R. M., Souza, T. L. P. O., Faria, L. C., Melo, L. C. and Pereira, H. S. (2022). Genetic parameters and selection of black bean lines for resistance to fusarium wilt and yield. Pesquisa Agropecuária Brasileira, 57, e02846. https://doi.org/10.1590/S1678-3921.pab2022.v57.02846
  » https://doi.org/10.1590/S1678-3921.pab2022.v57.02846

Edited by

• Section Editor:
  Carlos Alberto Scapim

Publication Dates

• Publication in this collection
  28 Aug 2026
• Date of issue
  2026

History

• Received
  09 Feb 2026
• Accepted
  13 July 2026