Background: Common bean is the most important staple grain legume for direct human consumption and nutrition. It complements major sources of carbohydrates, including cereals, root crop, or plantain, as a source of dietary proteins. It is also a significant source of vitamins and minerals like iron and zinc. To fully play its nutritional role, however, its robustness against stresses needs to be strengthened. Foremost among these is drought, which commonly affects its productivity and seed quality. Previous studies have shown that photosynthate remobilization and partitioning is one of the main mechanisms of drought tolerance and overall productivity in common bean.

Results: In this study, we sought to determine the inheritance of pod harvest index (PHI), a measure of the partitioning of pod biomass to seed biomass, relative to that of grain yield. We evaluated a recombinant inbred population of the cross of ICA Bunsi and SXB405, both from the Mesoamerican gene pool, to determine the effects of intermittent and terminal drought stresses on the genetic architecture of photosynthate allocation and remobilization in pods of common bean. The population was grown for two seasons, under well-watered conditions and terminal and intermittent drought stress in one year, and well-watered conditions and terminal drought stress in the second year. There was a significant effect of the water regime and year on all the traits, at both the phenotypic and QTL levels. We found nine QTLs for pod harvest index, including a major (17% of variation explained), stable QTL on linkage group Pv07. We also found eight QTLs for yield, three of which clustered with PHI QTLs, underscoring the importance of photosynthate remobilization in productivity. We also found evidence for substantial epistasis, explaining a considerable part of the variation for yield and PHI.

Conclusion: Our results highlight the genetic relationship between PHI and yield and confirm the role of PHI in selection of both additive and epistatic effects controlling drought tolerance. These results are a key component to strengthen the robustness of common bean against drought stresses.

Population development

To investigate the genetic basis of PHI [defined as seed dry weight per pod at harvest or whole-seed weight (WSW) over average dry pod weight, including seeds, at harvest or whole-pod weight (WPW)] and other yield-related traits, a bi-parental population was evaluated consisting of 226 F₉ Recombinant Inbred Lines (RILs) from the cross of ICA Bunsi and SXB405, developed at the Centro Internacional de Agricultura Tropical (CIAT), Colombia. This population was selected for a QTL analysis because a previous study using a subset of this population (78 lines, F_4:6) found large phenotypic variation and transgressive segregation for PHI, as well as relatively high heritability and correlation with yield [27]. This study provides a more extensive analysis on the entire population; it also combines phenotypic and genotypic data, leading to an as yet unperformed QTL analysis in this population.

ICA Bunsi is a navy type bean developed in 1968 by the Instituto Colombiano Agropecuario (ICA) from a cross between Magdalena 8 and Japón 3 [66]. It has been used extensively in Canadian breeding programs [67], due to its high productivity and resistance to white mold (Sclerotinia sclerotiorum) [68,69]. ICA Bunsi carries the I gene, which confers Bean Common Mosaic Virus resistance [70].  SXB405 is a cream-colored type experimental line developed at CIAT from a four-way cross (A 686/A 774//NXB 80/SEA 15) and selected for its high productivity under drought and common bacterial blight (Xanthomonas axonopodis pv. phaseoli) resistance [27,71]. Both genotypes belong to the Mesoamerican gene pool, are photoperiod-neutral, and have a type II (indeterminate, bush) growth habit [27].

Trial design

Field experiments were carried out at the Plant Sciences Farm at the University of California, Davis (38.53 °N, 121.78 °W) for a total of five environments in two years and three water regime treatments. The soil type of the site belongs to the Yolo series, a member of fine-silty loam, mixed, nonacid, thermic Mollic Xerofluvents, considered well-drained, with slow to medium runoff and moderate permeability (https://soilseries.sc.egov.usda.gov). The seeding was carried out on the 5^th of June in 2013 and 8^th of June in 2014. The plants were harvested on the 10^th of September in 2013 and 12^th of September in 2014. All water was provided by irrigation as there were no rain events during the experiments. In 2013, there were three irrigation treatments: terminal drought (TD), intermittent drought (ID), and full irrigation (well-watered). The full irrigation treatment received four irrigations while the intermittent drought received the first and third irrigation, and the terminal drought received the first and second irrigation. Only terminal drought and full irrigation treatments were carried out in 2014, receiving 4 irrigations for the full irrigation treatment, and the first two irrigations for terminal drought. The second irrigation of the terminal drought was applied during early flowering. The timing of the irrigation was decided according to weather and evapotranspiration (Fig. 4). The agricultural management was according to standard practices [72]. The parental lines and 226 of the RIL lines were planted in a split plot design with 3 blocks and one repetition per block. The experimental unit was a plot of 60 plants grown in two 3 m-long rows and 76 cm between rows (density of 131,578 plants per hectare).

Phenotyping

From each plot, 15 to 20 pods were collected at random across the plot during harvest and dried for 5 days at 50^OC. Seed and pod number and pod mass were recorded. The seeds and pod walls were weighed separately. Pod harvest index was calculated as the proportion of seed biomass to the overall pod biomass (sum of pod wall + seed biomass) [17]. Days to flowering (DTF) was assessed when at least 50% of the plants in the plot had an open flower. Agronomic yield was weighed after four days of drying at 40⁰C to standardize grain water content. Yield was not measured on the intermittent drought treatment in 2013 as the plots were mixed after cutting due to unexpectedly high winds.

There are two worksheets in this dataset. Worksheet A provides segregation data for phenotypic traits in the ICA Bunsi x SXB405 recombinant inbred population. Values are averages  across replications. C: control; ID: Intermittent drought; TD: Terminal drought. DTF: Days to flowering; PWW: Pod whole-weight; WSW: whole-seed weight; WPW: Whole-pod weight; SW100: 100-seed weight; SPP: Seeds per pod; PHI: Pod harvest index; YLD: yield.. Values are averages over replicates.

Worksheet B provides SNP segregation in the recombinant inbred population ICA Bunsi x SXH405. A: ICA-Bunsi allele; B: SXB 405 allele.

Missing data are shown with a dash.